RIETTA ENERGY SYSTEMS LIBRARIES IR yvou wish someone else to s ihis docment, i werie with document and the horary il Viaiigs o sumii, This report was prepcred as an occount of Government sponsored work, Neither the United ,S}utes_, nor the Commission, ner a‘ny’pe‘rscm acting on behalf of the Commission: L - A. Makes any warranty or represerntoflon, express or implied, with respect ‘to-the q{:;fu_rocy, completeness, or usefulness of the informotion contained in this report, or ,thdi:}_h_e use of any -~ information, apparatus, method, or proces_s' disclosed in this repott may not infringe privately owned rights; or _ o As_sum.es any liabilities with respect to the use of, or for damages. resulting from the use of = any informcflon, apparatus, method, or process disclesed in this report. As used in the above, '‘person acting on behalf of the _Co,rprpis_s'ion_'_'Vrin'clip'dé;f. any employee of contractor of the Commission to the extent that such employee or contractor prepares, handles or distributes, or provides access to, any information pursuant to his employment or confract with the Commission, Part 1 AIRCRAFT REACTOR ENGINEERING S. J. Cromer " —— e b +¥ 1.1. AIRCRAFT REACTOR TEST DESIGN A. P. Fraas STATUS OF ART DESIGN Design work on the Aircraft Reactor Test (ART) reactor, heat-exchanger, pump, and pressure-shell assembly is nearing completion. Layouts on all the major subassemblies have been completed, along with the major portion of the drawings of the detailed parts. Drawings for the remaining parts should be completed during the coming quarter. The applied mechanics and stress analysis work is accompanying the design, with rough first ap- proximations being completed, usually, shortly after completion of the layouts and with better, second approximations following closely, in most instances, upon completion of the detail drawings. In cases in which component tests have been deemed essential, the results are being anclyzed and modifications in details made when essential. Of course the analyses have not been completed for many very complex situations, and many key component tests have not yet been run. It is believed that modifications that will be required as the results of this work become available will probably involve only relatively minor reworking of partially fabricated parts. Such a calculated risk- is necessary and inherent in design work in- volving such exceptional extrapolahons of avail- able technology. The preliminary ‘layouts for the shield have been modified to include provision for a substantial amount of instrumentation and special equipment. A one-half-scale model of the top portion of the reactor {(commonly referred to as the “*north head''), including the NaK manifolding, has been completed, and models of the instrumentation components, the lead shielding, and associated parts are well. under way. Such models are used to investigate assembly ‘and interference problems. - Preliminary layouts have also been prepared for the arrange- ment of the lube-oil, hydraulic-fluid, water, gas, ~ electrical, and instrumentation lines in the reactor cell. A one-sixth-scale model of the entire ART assembly is being kept closely abreast of this work to ensure accessibility, freedom from inter- ferences, etc. The fuel fill-and-drain tank design has been completed, preliminary layouts for the associated supports, shielding, plumbing and in- strumentation have been prepared, and the con- sequent assembly, accessibility, etc. problems cre being studied in the one-sixth-scale model. The detail design of the plumbing and equipment installation outside the cell is well along and should be largely completed during the coming quarter; APPLIED MECHANICS AND STRESS ANALYSIS R. V. Meghreblian North-Head Pressure Stresses The stress analysis of the composite double- deck structure of the north head, mentioned in the previous report,! was completed. The analysis was based on the pressure loads to which the structure will be subjected during full-power oper- ation. Since the actual design consists of two - circular flat-plate decks joined by a complex pattern of vertical baffles and walls arranged both radially and circumferentially (Fig. 1.1.1), it was not possible to carry out an analysis of this com- posite structure which would yield an exact dis- tribution of the elastic stresses. Moreover, this structure is to be exposed to various operating conditions at temperatures of 1200°F and above _for about 1000 hr, and it is expected that thermal distortions and creep will cause redistributions of stresses which will differ markedly from any predicted elastic stresses. It is not entirely meaningful therefore to think in terms of an exact stress distribution, and, for this reason, precise analyses of this structure were not attempted. So long as the proposed design is capable of supporting the operating loads at relatively low stress levels, the details of the exact distribution are not important. From the viewpoint of creep limitations, it would suffice to know the general focation and magnitude of the highest stresses in the system, This information has beéen obtamed from a series of calculations based on simplified geometric con- figurations of the north-head structure, and these results will eventually be checked by an experi- mental stress analysis of a full-size aluminum model, The calculations consisted of three parts: a very elementary analysis in which the various IR. V. Meghreblian, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 22. 19 ANP PROJECT PROGRESS REPORT of shells and plates subjected to a complex pattern 20 SRe— jl PHOTOD 26471 . A X - Fig. 1.1.1. Plastic Model of ART North-Head Structure. _ : deck and shell areas were treated as individual , plate segments with assumed edge conditions, : ORNL-LR- WG 14944 an analysis of two circular flat plates of annular _l_ - PRESSURE kL - “shape joined by a continuous circumferential baffle UPPER DECK - N CONTOUR (Fig. 1.1.2), and an analysis of a composite system / ! E S ™~ of plates and shells to represent the primary struc- . - ) . . - q N Y tural members, which was based on the require- / N j H H ‘\\ \\ ment that the deformations of adjoining members / 4 N ' A D) N\ be matched along the various junction lines (welds). _ \ The first two analyses served to give a very crude CIRCULAR PLATES LOWER DECK estimate of the stress levels involved. The pur- BAFFLE (CYLINDRICAL SHELL) pose of the third analysis was to determine the * stresses produced in the north head duve to com- ¢ patibility requirements of the various segments . Fig. 1.1.2. Circular Plates with Baffle. I LT of pressure loads. The idealized model used in this calculation is shown in Fig. 1.1.3, and the net pressure loads at various points are indicated, This model represents approximately the cross section indicated in Fig. 1.1.4., It includes some ‘of the longest spans which appear in the design, and the results obtained from this model are there- fore conservative, The configuration of Fig. 1. l .3 was analyzed by writing the deflection equation for each member in terms of its load and edge conditions and solving the resulting system of eleven simultaneous equa- tions on the Oracle. in terms of the moments and reactions at the gclnts' and edges of the members, The stresses due to these loads were then computed. The largest stress, as indicated on Fig. 1.1.3, was found to be 2100 psi. Since the highest temperature which will occur in the north-head structure during full- power operation will be approximately 1300°F, this 2100-psi stress value is to be compared to FUEL EXPANSION TANK WALL - VERTICAL AXIS OF REACTOR Sy "'/4 in. I | % W//////////// 50 psi (PRESSURE) +— 8in.R UPPER -DEC_K_ ' _‘_\fi% wm [ \w&\\\\\&m SRR zOps. (PRESSURE) FUEL SWIRL CHAMBER —+— 9in. R o uowsn DECK ' - b _ 50ps.(pnessuas.) 1225ps: : o~ R 890 psi Fig. 1.1.3. The results were obtained 7 PERIOD ENDING JUNE 10, 1956 the creep properties of Inconel in the fuel mixture at about 1300°F, Creep tests have indicated that ‘the tensile stress required to produce rupture in 1000 hr at 1300°F is about 10,000 psi. The design criterion for creep which has been selected for the ART requires that the total deformation in any member not exceed 0.2% strain in 1000 hr, At 1300°F in the fuel mixture this corresponds to a tensile stress of about 2000 psi. The experimental program, designed as a check on the calculations for the northehead structure, is under way at the University of Tennessee. It is believed that the combined results of the ana- tytical studies and the model tests will reveal any defects in the propesed design. North-Head Thermal Stress With the completion of the analysis for the mechanical stresses in the north-head structure, attention is now being directed to the determinae- tion of the thermal stress disfl:ibutions. For this -ECRER ORNL-LR=DWG 14948 1500 p;\\\\ CONTOUR OF PRESSURE | < \SH\ELL AND LINER 30psi (PRESSURE) N _ ~ | 20 psi { PRESSURE}\ : ~ 2100 psi .~ 1600 psi \ ';} | %// %’%//// /// ////////////////// - o™ PN —— e 4Tin R ‘/z in - aaoo psi : 1o psi (PRESSURE) ',, ////% 3 Va in. 1270 psi— Ideclized Configuration of North-Head Composite Structure (Sect%on A-A of Fig. 1.1.4). 21 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 14946 Fig. 1.1.4. Cross Section Used in Composite- Deck Analysis. purpose, calculations have been undertaken to obtain preliminary estimates of the energy- deposition rates throughout this region (see Chap. 1.2, **Art Physics’). The locations in the north head at which these rates might lead to relatively large temperature rises are indicated in Fig. 1.1.5. The calculations of the associated temperature distributions and the thermal stresses are presently under way, |f these calculations yield excessive temperature gradients, it is planned to review the initial energy-deposition estimates and, if neces- sary, perform more precise analyses. Sodium Expansion Tank Design The design analysis of the sodium expansion tank was completed, and the configuration which was found to be acceptable from both the pressure and thermal stress viewpoints is shown in Fig. 1.1.6. The proposed design consists of a vertical wall of more or less elliptical shape joined to an end cap of slightly cylindrical curvature. This cap (or roof) is welded to the control-rod thimble, “which passes through the center of the ellipse. The stress analysis of this design was based on an idealized model consisting of a short, elliptical cylinder with a flat-plate cap subjected .to an 22 internal pressure of 30 psi. The stresses in the cap were computed from the relations for ‘an elliptical plate with various edge conditions. The stresses at the joint and in the vertical walls were computed from the relations for an equivalent circular cylinder with a flat head. These calcu- tations indicate thet the moximum stress (1000 psi) is due to the bending moment at the cylmder- cap junction. : During full-power - operation the tank wfll be partially filled with sodium at 1270°F to an assumed depth of 2 in., and the upper portion of the side walls and the roof of the tank will be ex- posed to direct gamma radiation from the sodium (Fig. 1.1.7). The outer surfaces of the tank are to be surrounded by insulating material, and, if no cooling is provided for these surfaces, the metal temperature in the roof will rise to 1420°F. Since the side walls are to be welded to the pressure shell, which will be at 1250°F, there will be differential thermal growth between the roof and the shell, which will give rise to a'thermal stress of 80,000 psi (based on an elastic analysis) at the roof-wall intersection. This stress is con- sidered to be excessive; and therefore cooling is to be provided for the roof. Sodium at a tempera- ture of 1250°F will be taken from the pressure- shell cooling circvit and fed into a system of tubes welded to the bottom surface of the roof. A total flow of 3 gpm will pass through this circuit at a pressure drop of 15 psi. The coolant sodium will leave the roof circuit at 1270°F and spill into the expansion tank volume, With this supply of coolant, the average roof temperature will be reduced to 1280°F.. The temperature profiles in the roof in the vicinity of a’cooling tube are shown in Fig. 1.1,8. The maximum thermal stress due to this temperature structure is 10,000 psi. ALUMINUM NORTH-HEAD MOCKUP FLOW STUDIES E. R. Dytko? R. E. MocPherson D. Ward A fullesize aluminum mockup of the fuél"syr;tem R. Curry? components in the north head of the ART has been set up with external piping to complete the 20n assignment from Pratt & Whitney -A_ircr-clft. M PERIOD ENDING JUNE 10, 1956 scomet ORNL—LR-DWG 14947 " Fig.r 1.1.5. Locations in ‘N-orth Hécd at Which Enargy bepdsition Might Cause Large Tempe Increases. PRESSURE SHELL/ TANK "‘/ LINER — ] : @___ FUEL EXPANSION ! TANK UPPER DECK : —\ 5:) -3 P | T ! > _ 'II/II/I//II//IZ c M 1 ¢ % . == Na EXPANSION rature - 23 | | ANP PROJECT PROGRESS REPORT ——— i —— — ——— - - — - —— -—— - - by —— - j N\ TTTTOAN 1 A \ \ ~ ——— e e = Iy — / / / # / K i ! /d.____.__._.( COOLING TUBES e —_——— —_—- [ iy N \ N N N N \ON NSNS \\ \ “ q\ N Y L k \ \ UNCLASSIFIED ORNL-LR-DWG 14948 —— - - ~ .( S p—————— - —_————— - \ \ \c 1 I \ \ 1 \ \ \ — o —— RELAXATION SECTION / (SEE FIG. t.1.8) T“T LEVEL OF SODIUM IN TANK | \1270°F Na SIDE WALLS — —— LINER PRESSURE SHELL EXPANSION - LOOP TANK ROOF COCLING TUBES 1270°F Na \CONTROL ROD HOUSING 1250°F Na SECTION A-A Fig. 1.1.6. Sodium Expansion Tank Design. circuits,® ond flow tests with water being pumped by the two fuel pumps under simulated reactor flow conditions are under woy. Twin fuel pump operation has been demonsitrated up to 3000 rpm, although 2400 rpm produced approxi- mately rated conditions of head and flow, De- gassing of the system with pump speeds of 2400 rpm required less than 30 sec with the water level 1 in. or more above the bottom of the surge chamber. When the speeds of both pumps were lowered in- unison from 2400 rpm, the first signs of ingassing 3p. R. Ward, ANP Quar, Prog. Rep. Dec. 10, 1955, ORNL-2012, p 65, Fig. 2.29. 24 occurred at 1200 rpm. The amount of gas entrained during operation with the pump speeds matched was believed to have reached a maximum, not exceeding 0.1 vol %, at 500 rpm, At speed levels above 1200 rpm, ingassing could not be detected with mismatching of up to 5%. Mismatching the speeds by 20% produced ingassing estimated to be 0.5 vol %. When the power to one of the hy- draulically driven fuel pumps was cut off while both pumps were running uniformly at 2400 rpm, about 13 sec was required for the one pump to stop. This one-pump-stopped condition created ingassing of roughly 2 vol %. - " tw UNCLASSIFIED ORNL~LR—-DWG 14949 pt———— INCONEL TANK RQOF ———= SODIUM LEVEL 06 N ENERGY DEPOSITION (w/¢m3) 02 X 0 0.5 1.0 1.5 20 DISTANCE FROM BOTTOM SURFACE OF ROOF {cm) Fig. 1.1.7. Energy fieaosition in .Sadi.am Ex; pansion Tank Roof Due to Gamma Rediation. Surging and insfabillity' were noted in pump speeds, flow rates, pumpmg power, and mlef and discharge pressures. It is believed that both the . centrifuge hardware and the ‘unfavorable pump ‘inlet pipe conflgurahons ‘were - conmbutmg to the ‘ob-" served mstabihty. The ‘addition of stralghtemng, vanes to’ the pump inlet. reglons rnaternally reduced the surging. - ~This, ‘coupled with some modlflca-':.L_i;:‘surface immeduately below the guide vanes and tions_to the xenon-removal system," _appears to again on the shell surface immediately below ____,_:"‘j__have reduced the* pressure fluctuaflons ‘to an-ace . the equafor, ‘however, the iimprovement was only jceptable level fhat |s, about 0 5 psl m fhe full-_j_.- S *’scale reactor; ” : e Water bypassecl mta fhe surge chamber for sumu-,_- : o lated Xenon removal was very turbuient, and splash- wetting~ of “all- _surfaces ‘and entrainment of fine “bubbles ‘resulted.” Methods ‘being . cansndered for__,, S reducing the extreme turbulence in this region include the addition of baffles, o reduction in PERIOD ENDING JUNE 10, 1956 bypass flow rates, and o redesign of the passages from the pumps to the surge chamber. CORE FLOW STUDIES W. T. Furgerson W. J. Stelzman D. B. Trauger Changes in design of both the center volute of the axial-flow type of header and the island expan- sion bellows located within the header of the proposed ART core resuited in an unsatisfactory core flow pattern being generated by the previously satisfactory inlet guide vane, designated GS-2, -and the conical baffle plate, designated GS-2- P3 (ref. 4). Under the revised design, this par- ticular guide vane and baffle plate combination generated flow reversal ot the island surface in the region of the equator. Systematic relocation of the conical baffle plate only ond analysis of the resulting core flow pattern yielded baffle plate GS-2.P10, which, in combination with guide vone GS-2, again generated a flow pattern containing no flow reversal along either the outer core shell or the islond surfaces. Brief periods of miner flow reversals did occur at the equator; however, ‘these occurred in midstream ond seemed to be caused by the turbulent condition of the fluid mass in this region. In general, the flow generated in the upper half of the core by this combination was extremely unstable, but it exhibited excellent sur- - face scrubbing, with very good transfer of fluid from the walls, and excellent fluid niixing. Below - the ‘equator, some improvements were noted in ~ the fluid flow properties; however, the streamlines "again tended to hug the inner and outer surfaces " as they approached the core outlet, - Attempts had - previously been made with the criginal axial-flow - “header - to improve the flow in the lower half of the core by means of turbulators on the island - minor, : This latest design is being evaluated in ~_the half-scale ART volume-heat-source apparatus _{see Chap. 4. 'I,'v,"Heat Transfer and Phys:cal - Propemes") - ' S o | . ‘GUb-'Whumaa,‘ w. J. Ste'liz‘l_'han,"an‘iiiw.'T.. Fdrgerson, AI;P Quar, Prog. Rep. March 10, 1955, 0RNL-206'|, p 24. 25 ANP PROJECT PROGRESS REPORT DEVELOPED COOLING TUBE 1250°F UNCLASSIFIED ORNL—LR—-DWG {4950 UNDERSURFACE OF ROOF _~ZERQ HEAT TRANSFER SURFACE 1250°F 1254 1260 1265 1270 1275 SYMMETRY SURFACE —3| - 1280 1283 1285 4287 - SYMMETRY SURFACE 1299°F 1298 1297 1296 1295 1293 UPPER SURFACE / 1288 1290 1292 INSULATED SURFACE Fig. 1.1.8. Temperature Profile in Sodium Expansien Tank Roof. PERFORMANCE REQUIREMENTS FOR PUMPS IN THE NaK SYSTEMS M. M. Yarosh A study was completed to determine the speed and power range requirements for the main and auxiliary NaK pump motors. Because of unequal system lengths, the pressure drops in each of the four main NaK systems for an equal NaK flow are different, and thus o different pump speed is re- quired for each system. In addition to the vari- ations between systems, the individual system resistances will vary as a function of operating time and operating conditions. These variations will be attributable to the. mass-transfer buildup anticipated - in the colder sections of the NaK systems, principally in the radiator tubes. Pump speed requirements were established from the pump performance curves., The NaK system curves (head vs flow) showing expected ranges of system resistance as a function of flow were plotted on the pump performance curves, and the operating speeds were established at design flow to be those speeds falling between the minimum 26 and maximum system resistance curves. Thus, a range of operating speeds was established for each individual NaK system. Power requirements were then computed for the corresponding speed, head, and flow requirements. In order to reduce warmup time for the NaK systems, it will be desirable to operate the NaK pumps ot, or near, full speed during the warmup period. Stress considerations, however, will pro- hibit operation at full NaK pump speed for extended periods of time when there is no fuel in the reactor. Therefore an intermediate pump speed was estab- lished at which reduced warfnup periods can be attained ot reduced stress. The operating speeds established and corresponding power requirements for the main and auxiliary NaK pumps are given in Table 1.1.1. CONTROL~-ROD COOLING SYSTEM J. Foster The ART control rod is designed to move verti- cally in a well containing static sodium, this sodium being deep enough to cover the control W e g ——_ 17 TABLE 1.1.1. NaK PUMP SPEEDS AND HORSEPOWER REQUIREMENTS Main System Auxiliary System Pump Power Pump Power Speed Required Speed Required {(rpm) (hp) (rpm) (hp) 1900 27 1900 7 2300 . 42 2300 14 2650 61 2800 26 2800 72 2950 N 3000 87 3100 37 3200 102 3250 41 3400 118 3400 46 3500 127 3550 55 rod in its fully withdrawn position. This places the level of the sodium-free surface in the well just a few inches above the top of the reactor north head. The well extends up about 5 ft above this fevel so as to place the control-rod drive mechanism outside the reactor shield. The sodium at and near the free surface must be cooled to below 500°F to minimize the vapor pressure and hence the diffusion and deposition of sedium vapor on the components of the control-rod drive mecha- nism, where such deposits might create operating difficulties such as shorting of electrical circvits. Tests have shown that sodium vapor evolution and deposition are negligible at 500°F. The lower portions of the sodium in the well will be exposed to temperatures of about 1200°F, and therefore the cooling system includes con- vection baffles to still the upper few inches of the sodium and @ water jacket around this baffled sodium zone. As a precaution to ensure against any possibility of water entering the sodium - chamber, the jacket will be a completely water- - R 'hght assembly, -An Inconel sleeve will be shrunk - over the outside of the controlerod well pipe to form a double wall. Water will be circulated. at 220 to 240°F (sodium melts at 208°F) through the jacket and will serve to remove heat or supply PERIOD ENDING JUNE 10, 1956 heat as required by the condition of the reactor system. No flow or pressure controls will be provided other than an orifice in the water line, designed to give a flow of 1 gpm, The water-jacketed and baffled sodium zone will be separated from the hot sodium in the lower well by a solid Inconel plug inserted in the sodium as a heat dam to keep the thermal gradient along the Inconel well to .a reasonable value from the thermal - stress standpoint. This Inconel plug is a 4-in.- high cylinder with a central hole drilled along the cylindrical axis, through which the 9-|n.-d|a control-rod drive is free to move and posmon the rod as required. The effluent hot water from the jacket will pass through an economizer, in which it will heat the entering water stream. This will reduce the water heating load and cool the effluent stream to prevent flashing in the drain. SODIUM SYSTEM STUDIES R. |, Gray Recent tests showed that the flow resistance in the annuli around the core in which sodium will be circulated will be somewhat smaller than ex- pected with the spacers in place. This will effect a slightly lower over-all pressure drop and more nearly balanced flow between the cooling holes and the core annuli. Pressure drop calculations indicated the need for increasing the thickness of the control-rod cooling annulus, in which sodium will circulate, from 0,080 in. to about 0.125 in. Stress calculations indicate the need for cooling the top of the sodium expansion tank and for the ~ addition of a flexible bellows to the island sodium - inlet line (see previous section of this chdpter 'on "“Applied Mechanics and Stress Analysis’’). auxiliary sedium expansion tank of upprommately 0.6 ft has been added to the system so that in the event of a major reactor shutdown sodium can be added as the sodium temperature is lowered from 1200°F to 300°F to avoid a loss of prime in the sodium pumps (which would otherwuse occur at about 800°F). 27 ANP PROJECT PROGRESS REPORT 1.2, ART PHYSICS A. M. Perry RADIATlo_fl’ HEATING ON THE ART 'EQUATORIAL PLANE IN THE VICINITY OF THE FUEL-TO-NaK HEAT EXCHANGER H. W. Bertini The results of calculations - of the radiation heating on the ART equatorial plane in the outer 3 cm of the beryllium reflector and in the Inconel and the boron-containing shells on both sides of the fuel-to-NaK heat exchanger are presented in Figs. 1.2.1 end 1.2.2. The total gamma-ray heating in- each region is given in Fig. 1.2.1, as well as the heating from the sources which are the main contributors to the total in each shell. The encircled numbers on Fig. 1.2.1 refer to the sources described in Table 1.2.1. The data on heating in the copper-boron layer by alpha particles from the B19(n,a)li? reaction are plotted in Fig. 1.2.2. The heating goes to infinity at the face of the layer closest to the core because the heating at various points is governed by an E function, | o daA Eyw) = [ e —, 1 where A is the mean free path, The integral under the curve will be finite. TABLE 1.2.1. SOURCES OF RADIATION HEATING CONSIDERED IN CALCULATING THE RESULTS PRESENTED IN FIG. 1.2.1 Source No. Source Source Strength 1* Prompt gamma rays in the fuel region of the core of the 28.3 W/C,"‘a reactor _ 2 Decay gamma rays in the fuel region of the core of the 6.84 w/cm? reactor 3 Gamma rays from inelastic scattering of neutrons in the 10.1 w/cm® fuel region of the core 4 Capture gamma rays in the outer core shell 41.4 w/cm? Capture gamma rays in the reflector (average) "~ 0.5 w/emS Capture gamme rays in first Incone! shell outside of 22.5 w/cm> beryllium reflector 7 _ Beron capture gamma rays in copper-boron layer 1.8 w/em? 8 Alpha particles from the Bw(n,a)Li7 reaction in the 42 V_l_/cm3 copper-boron layer (average) 7 ’ 9 Decay gamma radiation from the fuel in the heat exchanger 2.3 w/cm® 10 Gaemma rays from inelastic scattering of neutrons in first 0.7 w/cm3 9 cm of reflector (average) ' 1N Capture gamma rays from delayed neutrons in the heat ex- <0.1 w/cm3_ changer and Inconel shells (including the pressure shell) 12 Capture gamma rays in the copper of the copper-boron layer 0.5 w/cm? 13 Gamma rays from inelastic scattering in both core shells ~4 w/cm? 14 Capture gamma rays in the island core shell 41.4 w/cm? *In Fig. 1.2.1 the data for heating from sources 1, 2, 3 are combined and labeled a. 28 T4 CALCULATED HEATING (w/cm3) w ’ ‘ ( SRR ORNL—-LR-DWG 14916 |~ STAINLESS STEEL \ TOTAL HEATING., D= HEATING FROM SOURCE x. (SEE FUEL 6AP —— . TABLE 1.2.4). g , =—{(@)—— HEATING FROM SOURCES 1, 2, AND S = o 2 3 (SEE TABLE 1.2.1). = - = 2 W 5 S 2 W ALL DIMENSIONS IN cm. w - L w Q I @ B - . 2 = 51| @ x x | w 3 & a1 15 S .5 o o z : 3. 7 E : 5 2 5 le— HEL UM & | & S @ onp | l—s0DIUM 8 o Y @« | L@ e S |a Ll 0157 I~ | S = 2 N ‘ n L 0.457 0.318 g lwlal /1 I ™[N o b o S | o w o o =2 @ : 0 x & ‘_ 2 | = Wl T —+ \ % @ ul I % w @ ¥ \, W o N a Q o a . B = STAINLESS STEEL Ysl \ g = & e, rS o o e ey . ' . S © G \’9“ \ ' | & & roFo— ~o——| #\ ——— ‘_@_ Fi‘g.‘l 1.2.1. Gamma-Ray Heating in the Vicinity of the Fuel-to-NaK Heat Exchanger on the Equatorial Plane of the ART. 9§61 ‘01 INNF ONIGNI aol¥3d ANP PROJECT PROGRESS REPORT SEORETY ORNL=LR~-DWG 14917 103 3 "N o CALCULATED HEATING {w/cm) o 0 0.04 0.08 a2 016 0.20 0.24 THICKNESS OF LAYER {cm} Fig. 1.2.2. Heating in Copper-Boron Layer By Alpha Particles from the B'%(n,a)Li7 Reaction. The heating from sources 10 to 14 was neg- lected. Their combined contributions to the heating in the region being considered was estimated to be about 5% of the total heating. The fuel region of the core of the reactor was assumed to be o spherical shell 5.125 in. thick with an outside radius of 10.5 in.! This region was divided into 13 spherical shells of thicknesses varying from 0.27 cm to 2 em, The source strength - (in watts/cm3) of the prompt gamma rays in each shell was assumed to be proportional to the average fission power in each shell, which was calculated (from ref. 2) at the equatorial plane of w. L. Scott, Jr., Dimensional Data for ART, ORNL CF-56°'I =186 {March 13, 1956). 25, M. Perry, Fission Power Distribution in the ART, ORNL CF+56+1-172 {Jan. 25, 1956). 30 the reactor. The source strength of the decay gamma rays was assumed to be the same for each fuel shell,? The average source strength of gamma rays resulting from inelastic neutron scattering in the fuel was calculated4 by using the output of multigroup calculations® performed by the Curtiss- Wright Corp. on ART-type reactors with spherical symmetry. The inelastic cross sections used for the fuel were those reported in ref, 3. This calculation had been performed before all the data in the latter reference had been accumulated, so it was assumed that, for each inelastic collision, one-half the average neutron energy in each energy group was given off as 1-Mev gamma radiation, Caleulations made by using the more recent data indicate that the source strength used was too high by about 50%. The total heating values given in Fig. 1.2.1 may therefore be about 5% too high. This error is partially compensated for, however, by the neglect of sources 10 through 14. The prompt and decay gamma-ray spectra’ were divided into four energy groups with average energies for each group of 0.5, 1, 2, and 4 Mev, The last four groups listed in ref, 3 were combined into one group with an average energy of 4 Mev. The heating at the vorious places described in Fig. 1.2.1 was calculated by summing the contri- butions from each energy group from every fuel shell. It was assumed for the calculations that each fuel shell was replaced by an infinitely thin spherical-shell source embedded in an infinite homogeneous medium so that the standard transfor- mation from a spherical-shell source to two infinite-plane sources would apply, that is, so that the heating at R, b(R), would be given by r (1) A(R) =-E- [H(R ~ 1) - HR + 1) ] ’ 3H. W. Bertini et al.,, Basic Gamma-Ray Data for ART Heat Deposition Calculations, ORNL-2113 {in press). ‘ 4Calculations performed by R. B. Stevenson, Pratt & Whitney Aircraft, private communication to H. W. Bertml. 5s, Strauch, Curtiss-Wright Corp., prlvate communi- cation to H. W. Bertini. H. Reese, Jr.,, S. Strauch, and J. Michalcozo, Geometry Study for an ANP Circulating Fuel Reactor. WAD-1901 (Sept. 1, 1954). TH, w. Bertini, C. M. Copenhaver, and R, B, Stevenson, ANP Quar, Prog, Rep. March 10, 1956. ORNL-2061, p 35. 9 [ where _ r = radius of source (taken as the average radius of each fuel shell), R = distance from center of fuel shell to field point, heating at a field point due to an infinite plane source of monocenergetic gamma rays a distance x away from the field point, The second ferm of Eq. 1 was dropped for these calculations because of the large radii involved. The assumptions made were certainly not consistent with the geometry, inasmuch as the region between the source shells and the field point ‘is not everywhere homogeneous. The justi- fication for this approach was that it appeared to be as good as could be done without going to a much more detailed numerical integration aver all source points for each energy group to calculate the heating at one field point. , In deriving H, it was assumed that the attenu- ating medium between the plane source and field point consisted of infinite slabs of materials whose thicknesses were determined by their thicknesses at the equatorial plane of the ART.! The buildup factor used was of the form H(x) ' | al pit, _ (2) Aje ? -1) +1, where , A,a = parameters of the equation, p; = linear total gammo-ray absorption cross section {cm=1) for material i, t, = thickness of material { (cm) 1 Wl Under fhese condlhons, s | - +(]—A)E(Ep.itt)} wheré oy ' m : m “source strength (w/cmz), T . t . “ - ! o " section (cm=1) for the field pomt fhlckness of the ith siab (cm) LY m The s term was determmed for each fuel sheII - and each energy group by multiplying the source strength for each group (in w/cm3) by the thickness heuf;ng (w/cm3) ot the fleld po:nt '. PERIOD ENDING JUNE 10, 1956 of the shell. The buildup factor parameters, A and a, were taken to be those for beryllium. The i's, A's, and o's were evaluated at each energy group. : The spectrum of prompt gamma rays. reported in ref. 7, 8.8 e=10VE photons/Mev-fission, is different from that reported in ref, 3, 9.61 e~1-01E ‘photons /Mev+fission, The former value, which neglects the variation of (¢_/0 ) with energy, was used in these. calculchons be?ore the cor- rection was pointed out. Use of the latter spectrum would change the results repor_ted here by less than 2%. . The heahng by the capture gamma rays in the outer core shell was calculated with the use of the same assumptions as those used for the calcu- lations of the heating by the fuel-region sources, that is, a sphere-to-plane transformation was made and slab geometry was used for the inter- vening mediums betweeh the plane source and field point. The absorption rate in the outer core shell was calculated from the output date of multigroup calculations.’ : The spectrum of capture gamma rays in Inconel was divided into seven energy groups in ref. 3, but, for this calculation, the first three groups were combined into one group, and the average energy of this combined group was taken to be 2 Mev. The fourth ond fifth groups in ref. 3 were taken as the second and third energy groups for this calculation, and the average energies were tcken to be 4 and 6 Mev, respectively. The sixth and seventh energy groups in ref, 3 were combined into one group with an average energy of 8 Mev. Thus a total of four energy groups was used in this calculation. The buildup factor used was that for beryllium at ecch energy group. The outer - core shell has an msnde radius of 26 ¢m and ¢ ' thlckness of 0.381 cm. B ~The capture gamma rays in the island core shell - were neglected because of the shielding properties . of the fuel. The heating by the _capture gamma “rays in the beryll:um reflector was calculated by - using the sphere-to-plane tranisformation and the “7 " other assumptlons given above. ‘The reflector_ linear gomma-ray energy absorphon cross region was divided into five spherlcal shells, that ~ is, the same shells as those used by the Curtiss- Wright Corp._m their mulhgroup culculanons of reactor No. 675 6 The capture gamma-ray source 8Calcu|cflons performed by C. M, Copenhaver, ORNL, private communication to H. Bertini. 31 "ANP PROJECT PROGRESS REPORT strengths in each region were calculated® from the output - of the multigroup calculations, The spectrum of the capture gamma rays in beryllium is divided into two energy groups in ref, 3, and these groups were combined into one group with ‘an average energy of ‘6 Mev, for this calculation, The bunldup factor for - berylllum at 6 Mev was used, In calculahng the gamma-ray sources in the first Inconel shell around the beryllium reflector, - it- was assumed that 32% of all neutrons born in the core ‘escape from the reflector as thermal neutrons.® It was further assumed that these ‘neutrons esc¢ape with an isotropic angulor distri- bution from the surface of the reflector. A source strength, S, in thermal neutrons/cm2.sec escaping from the surface of the reflector, was calculated by using a reflector radius of 55.04 cm (ref. 1). Because of the large radii of curvature in this (6)h-% ]f{ [1-@2! ] 2”1 a =2 neutron absorption cross sectlon is 0. 18 cm=1, Then, for the Inconel, (5) absorption rate in Inconel S[1 -~ P_(Na)1P, (lnconel) The capture gamma-ray source strength per unit volume was then calculated by using the di- mensions given in ref. 1, the energy per capture given in ref. 3, and the assumptions given above, The macroscopic neutron absorption cross sections at average velocity at 700°C were cclculafed from values given in BNL-325.° It was assumed that the gamma-ray source was constant in the Inconel. Also, because of the large radii involved, slab geometry was assumed for the source and for the mediums between the source and field points. o For @ slab source and the buildup factor inén in Eq. 2, the heatmg for monoenergetlc gamma rays is given by - Ey [““a) L ‘i#i}} i=1 +(|_A)[E2<§2 t,-#,-) ~ E, (E} t#,)} ' region of the reactor, the neutron absorption rates were calculated on the basis of slab geometry. Between the Inconel shell and the reflector there is a %-in. layer of sodium coolant.! If it is assumed that a neutron leaving the surface source will be absorbed only on a first-flight absorption collision, the probability of absorption in the s_odium,_P'a(Na), is derived to be (49 PNa) =1 - E,[Z (Na):t], where oy =" e | Ll A2 t = thickness of sodium layer (in cm), 2 J(Na) = macroscopic thermal-neutron ab- sorption cross section for sodium = 5.6 x 10~3cm=1, It was assumed that the angular distribution of the neutrons reaching the Inconel was still isotropic, and a similar expression was obtained for the Inconel, for which the macroscopic thermal- 32 - where ¢ = source strength (w/cm3), p; = linear total gamma-ray cross section (cm=1) for the ith slab, t, = thickness of ith slab (cm); i = 1 designates the source region. The spectrum of capture gamma rays from Inconel was divided into seven energy groups?® with average energies for each group of 0.5, 1, 2, 4, 6, 8, and 10 Mev. In calculating the heating in the shells odjocent to the Inconel source, the buildup factor for inconel was uvsed. Test calcu- ~ lations have shown 10 that the heating is relatively insensitive to the type of buildup factor used for materials which have nearly the same equivalent Z, The heating in the beryllium reflector was calculated by using the buildup factor for be- ryllium. For the shells on the pressure shell side of the heat exchanger, the bui!dup fccfor for th‘e . J. Hughes and J. A, Harvey, Neutron Cross Sect:ons, BNL-325 (July 1, 1955). H. W. Bertini, C. M. Copenhaver, and R, B. Stevenson, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 36. ) o © heat exchanger was used., The heating at all field points was found by summing over the contributions of every energy group. The self-heating of this Inconel shell for each energy group was calculated by using the ex- pression (7) H ="€'%i {2 - (Ez(l-’t) + E2 [ l’-(to - t)])} ’ where H = heating at ¢ (w/cm3), ¢ = source strength of the shell (w/cm3), B, = linear energy absorption cross section ~ (ecm=1) of the Inconel shell, p = linear total cross section (cm"‘) of the inconel shell, t, = thickness of the Inconel shell (ecm), t = distance from the surface of the Inconel shell (cm). No buildup factor was used for this self-heating calculation. _ S Sample calculations were made in order to compare the hedting from this shell when it was assumed to be a plane source with the heating when the shell was assumed to be a slab source; the results indicated a difference of 30% between the two values of the heating at o distance of 2.5 mean free paths from the source.1® Thus the approximation of a plane source was not used in this case. The source strength of the gamma rays resulflng from neutron captures in the boron of the copper- boron layer was calculated by assuming that oll the neutrons. leaving the reflector and escaplng from the sodium and above are obsorbed at the fronf surface of \‘he copper-boron layer, - that only 90% of the ‘neutrons smkmg the" layer are cbsorbed there, and they also indicate. ‘that the distribution of. 1he sources of gamma rays would ‘be the ‘same as: the dlstrlbu’non of . hecmng by alpha pcrhcles, as tllusfrated in- Flg. 1.2.2. - - The source strength of the gamma rays-used for - lthls caiculatlon -was ‘therefore ™ about ]0% ‘too o rh_lgrh It was ‘assumed that there was ‘a0, 48-Mev ‘gamma sorphons.‘? e Vg, Ajzenberg and T. Lauritsen, Rev. Mod. Phys. 24, 351(1952). “Actual calcuhnons :ndlccte-, ray. ossocmted wnh 93% of the ab--. PERIOD ENDING JUNE 10, 1956 The heating in the shells for this source was calculated by using Eq. 3. The buildup-factor parameters and the gammaeray cross sections were evaluated .at 0.5 Mev.3 The buildup factors used were the same as those used in the previous calculation of the sources in the Inconel shell outside the reflector. The hecting in the copper-boron layer by alpha particles from the B10(n,a)Li? reaction was calculated by using techniques similar to those used in estimating the neutron captures in Inconel. It was assumed that all neutrons escaping from the reflector, sodium, and Inconel were incident ~on the surface of the copper-boron layer with an isotropic angular distribution. By assuming slab geometry and by assuming that the neutrons make only first-flight absorption collisions, the proba- bility of absorption at ¢ per unit thickness in the ith constituent of the layer is given by the ex- pression ‘ i T (8) P,=320 E, (X, 1), where Ei‘ = macroscopic thermal-neutron absorption cross section of the ith constituent, 2: = total macroscopic thermal-neutron ab- sorption cross section of the copper-boron layer, t = distance from the front face of the slab (cm). omt ; In this case the copper-boron layer was assumed to consist only of copper and B19, and the atomic density of each was calculated for a matrix con- _;s:shng of 16 vol % natural B 4C ond 84 vol % ccappt‘:r..],2 The mlcroscoplc cross sechons used - were. those for the. average velocaty at o temper- ;_oture of 700°C ‘and they were calcuiated from the room-temperafure cross sections given in BNL-325.? ~The neutron. absorphon rate per cublc cenhmeter in B‘o czf tis glven by (‘9_)5:““.7"__*' 23 g (th) , '_where S is glven in neufrons/cm sec tmpmgmg on the front face of the matrix, The heating resulting: - from- neutron cupfures in B10 was ‘calculated by ‘- i'-assumlng thot in 93% of the absorphons in- Bw-, 12A, M. Perry, ORNL, H. W. Bertini. private communication to. 33 ANP PROJECT PROGRESS REPORT @ 2.32-Mev alpha particle is given off and that in the remaining 7% a 2.8-Mev alpha particle is given off.1! The heating from beta decay resulting from neutron captures in copper was calculated, and it was found to be negligible compared with the alpha-particle heating. The gamma-ray heating resulting from the emission of decay gamma radiation in the heat exchanger was calculated on the basis of slab geometry by using Eq. 6 for each energy group. The spectrum of gamma rays was divided into the seven energy groups given in ref. 3, with average energies of 0.5, 1, 2, 4, 6, 8, and 10 Mev. The buildup factor for the heat exchanger was used for each energy group. The total heating at various points was calculated by summing over the contributions of each energy group. The source strength of the gamma rays resulting from inelastic neutron scattering in the beryllium reflector was also calculated. These sources are appreciable only in the first 9 cm of beryllium, However, their contributions to the heating in the regions around the heat exchanger are small compared with those of the beryllium capture gamma-ray sources, The reason for this is that the gamma rays from inelastic scattering were taken to be 1-Mev gamma rays, which are at- tenvated much more strongly by the beryllium than are the 6-Mev capture gamma rays. {n addition the total source strength of the gamma rays from inelastic neutron scattering is much smaller than that of the capture gamma rays because of the volumes of beryllium involved in each, The source strength of the capture gamma rays resulting from delayed neutron captures was calculated by assuming that 50% of the delayed neutrons were given off in the heat exchanger and that ‘all these neutrons were captured in the neighborhood of the heat exchanger. By assuming that an 8Mev gamma ray was given off for each capture, the average source strength per unit volume was found to be small as compared with the other sources in this region. The capture gamma-ray sources in the copper were calculated in the same general way that the alpha-particle heating in the copper was calcu- lated, with the source converted to a surface source of gamma rays. In itself this is not an insignificant source, but the heating resulting from this source would be only cbout 2% of the total from all the sources considered here, 34 The gamma.rays resulting from inelastic neutron scattering in the core shells were calculated by using the Curtiss-Wright multigroup fluxes> and an inelastic microscopic cross section of 1.5 barns (ref. 13) for the constituents of Inconel. It was assumed thaf a 2-Mev gamma ray was given off for each inelastic collision. The results . gave a source strength of about 10% of that resulting from the capture gamma rays in these shells, As has oalready been remarked, the capture gamma rays in the island core shell were neglected - because of the shielding properties of the fuel. Where it has not been stated otherwise, the dimensions used in these calculations were those from ref. 1 for the equatorial plane of the reactor. RADIATION HEATING IN VARIOUS REGIONS OF THE NORTH HEAD - H. W. Bertini D. L. Platus 14 Calculationis of the radiation heating to be . expected in various regions in the north head of the |ART were undertaken in order to supply numbers from which thermal-stress calculations could be made. Because of the complexity and the time that would be involved in calculating accurately the heating in all the regions of the north head, it was decided to make preliminary estimates of the deposition rates, More accurate values calculated for other regions of the reactor were used as guides. In all cases the tendency was to overestimate the heating. These estimates can be used to identify the thermal-stress problems, and where the calculated thermal stresses are marginal, the heating will be recalcu— lated. ' Calculations were mcde of the heuf-deposifion rate in a slab of Inconel bounded on one side by on infinite fuel region containing the sources of radiation. This heat-generation rate was used in all regions in the north head which are bounded by finite fuel volumes, The heat-deposition rates in a slab of Inconel bounded on one side by slabs of sodium of various thicknesses were calculated, and the resuits were extrapolated and interpolated to obtain the heat- g’enerm‘ion rates in the Inconel regions of the n‘orth By, L. Taylor, O. Lonsjo, and T. W, Bonner, Phys. Rev. 100, 174 (1955). 140n assignment from USAF, +) v ek head which are bounded by various thicknesses of sodium, Fairly accurate calculahons have been made of the heat-deposition rates in the -Inconel filler plates below the island and in the vicinity of the fuel-to-NaK heat exchanger on the equatorial plane of the reactor. These results were used as a guide in estimating the heating in some north-head regions, and new values were obtained by compensating (by simple exponential attenu- ation) for decreased beryllium thicknesses, penetrations through additional fuel layers, increased thermal-neutron leakage currents into the north head, etc. A neutron current of 7 x 1013 neutrons/cm2-sec was assumed to be escaping uniformly from the upper portion of the core,'? and it was assumed that 1 Mw of fission power was being generated in the fuel regions of the north head by neutrons escaping into this region. The latter increased the gamma rays in the fuel by about 30%. The sources of gamma radiation considered were those from the sodium and fuel in the north head, the heat exchangers, boron, the fuel in the core, core shells, beryllium, and Inconel shell capture gammas. The sources of beta particles considered were those from the gases in the fuel-expansion tank, and the sources of alpha particles were taken fo be those from boron captures. The average values of heat generation obtained in these calculations are presented in Table 1.2.2. The configuration of the north head is shown in Fig. 1.1.5 of Chap. 1.1 of this report, BETA- AND GAMMA-RAY ACTIVITY IN THE. FUEL EXPANSION CHAMBER AND THE OFF-GAS SYSTEM R. B. .'5tev¢;enson“5 | The power-source dzstrlbuhon of the ochwty of the gases in the space above the fuel in the fuel expansion chomber and in the off-gas - line has CA5A, oM Perry, ORNL, prlvafe communicahon to : _H- W- Berflrflo T 6On ass:gnmenf from Prafl & Whltney Anrcroft. : , 75,30 Newgard, Fission Product Activity and Deéay | , Heat Distribution in the Circulating Fuel Reactor with _F:sszon Gas Strippmg, TIM-205 (Sept. 28, 1955) . {approximately) PERIOD ENDING JUNE 10, 1956 been determined. The results obtained are to be used . in the calculation of the radiation heating and the thermal stresses in this region of the reactor, _ The radicactive constituents of the gas in this space are the gaseous fission products, xenon and krypton, and their daughter products. There is also a possibility that some volatile fission- product fluorides will be formed in the fuel and will escape into this area. However, it has been shown17 that if all the fission-product fluorides entered this space, they would add very little activity to that already caused by the gaseous fission products and their daughters, Thus, their effect has been neglected here. Also, there is some question as to whether the daughter products of the fission gases will actually be carried downstream by the off-gas system or whether they will be deposited on the enclosing walls as they are formed. In order to get a conservative estimate of the power-source distribution, it was decided to treat the daughter products of xenon and krypton as gases (except insofar as their purging from the fuel into the fuel expansion chamber is concerned). The equilibrium amount of the gaseous fission products in the fuel expansion chamber is given by ‘yl‘Ap N, = , oA 4 Ay (A +A) where N is the total number of atoms of type 7 in the gas volume per fission per second, y, is the saturation fission yield of the ith nuclide, A_ is the ‘‘purging’’ constant, A_ is the ‘'sweeping” constant, and A. is the decay constant of the ith nuclide. The **purging’’ constant is determined by the volume flow rate of the fuel through the - purging pumps and the total volume of fuel. This constant determines the amount of the gaseous fission products which are purged from the fuel into the expansion chamber. " The "sweepmg constant is determined by the flow rate of helium through the expansion chamber and the volume of the gas space. This constant determines the dwell time of the radicactive nuclides in the gas "space and thus the number of disintegrations they suffer there. The *‘purging’’ and . ‘'sweeping”’ .. ".constants are given by volume flow rate of fue| through purging pumps = p 7 total volume of fuel volume flow rate of helium through expansion chamber s volume of gas space 35 ANP PROJECT PROGRESS REPORT TABLE 1.2.2. AVERAGE HEAT-GENERATION RATES IN MEMBERS OF ART NORTH HEAD Member Deserioti 'Heat Generation No.* escription (w/cma) 1 Pressure shell 4 (below sodium expansion tank) 2 Liner 6 w/cm® + 16 w/cm? on ex- pansion-tank surface due to beta rays 3 Fuel-expansionstank baffle 3 4 Fuel-expansion-tank wall 6 Upper deck (regions with sodium on both sides) é Upper deck 15 (regions with fuel on both sides) 7 Swirl=chamber baffle 3 Swirl-chamber wall 8 Lower deck 8 {regions with fuel below and sodium above) 10 Lower deck (regions with fuel on both sides) 12 1" Copper-boron tiles 25 w/cm2s + 6 w/cms, where t = thickness of tiles (cm) 12 Filler block 3 13 Beryllium support struts 10 14 Filler block 1 15 Copper-boron tile 30 16 Flat section of lower support ring 15 17 Strut part of lower support ring 3 18 Lower support ring 1.5 *See Fig. 1.1.5, Chap. 1.1, this report, for location of member. The power-source distribution of the radiocactive nuclides is found by multiplying their equilibrium concentration by their decay constant A, and their average energy per disintegration. The total power and the power density in the gas space of the fuel expansion tank as a function of the volume of the gas and the helium flow rate are given in Fig. 1.2.3. For these calculations, A, was taken to be 5.82 x 10=3 sec, which corresponds to a fuel flow rate of 22 gpm through the purging pumps. If the purging device is 36 assumed to be 100% efficient, this means that the reactivity effect of the xenon is reduced to about 0.1% at equilibrium.'® The sweeping constant, A, is dependent on the helium flow rate and the gas volume, and thus it is different for each point on the curves. In converting the STP values of the helium flow rate, the temperature of the gas was assumed to be 1200°F, and the pressure in the expansion chamber and off-gas line was 18, . Meem, The Xenon Problem in tbe ART, ORNL CF+54-5-1 (May 3, 1954). ‘. 1) SaeReT ORNL—-LR—-DWG 14918 35 — - — 140 POWER DENSITY \ o - TOTAL POWER 30 [ [ 120 HELIUM FLOW RATE, 1000 liters/day(STP) L - // o 25 . — = 100 5 - z £ - 3 _~T1000 liters/day (STP) >~ = % N . 80 & = L 5 7 g tzu -t Q. a5 z 3000 Ju==" 60 ° A - 5 § \ 74/ <] _>_Zooo - F - © §< S T e | A [ ] 4o 5000 T T———1 5 , — - 20 o // o 0 0 100 200 300 400 500 600 TOO 8OO VOLUME OF GAS SPACE (ind) Fig. 1.2.3. Total Power and Power Density in the Gas Space of the ART as a Function of the Gas Volume and the Helivm Flow Rate for a Fuel Flow Rate of 22 gpm. taken to be 1.3 atm. The power of the reactor was assumed to be 60 Mw. In the calculation of the curves, the very short- and very Iong-llved nuclides of xenon and krypton (along with their decay products) were neglected,. Since the fuel circulation time in the ART is less than 3 sec, nuclides with half lives less than this value will decay mosfly in the fuel before it reaches the purging pumps. Thus very few atoms with half lives less than about 3 sec would get into the gas space. Also, for nuclides with long half lives (greater than about 100 ht), the number of disintegrations tokmg ploce in the fuel expansion chamber .and off-gas"line WOuld be small, since the dwell time at the assumed helium - flow rates is very short, Thus, these nuchdes' - may be neglected. - “In this' study, 32 nuclides were cons;dered N 16 of these being tSotopes of xenon and krypton “and 16 bemg their daughter products. The main “contributors fo the power " distribution: are the_k,_' : daughter products and not ‘the nuchdes of xenon - and krypton themselves, -In all cases, the daughter products contribute about 50 to 60% of the total power dlstrlbutlon. Of the total power, about 90% is due to the beta-ray decays, with only 10% being due to gamma-ray decays. Thus, in determining the heating caused by these gases, it can be PERIOD ENDING JUNE ‘10, 1956 assumed that the heat deposition will occur mainly in a small surface layer of the materials surrounding the gases in the fuel expanswn chamber and the off-gas line. The power density in the off-gas line as a function of time and gas veolume for helium flow rates of 1000 and 3000 liters/day (STP) is given in Fig. 1.2.4. The time axis can be converted into lengths along the off-gas line by dividing the volume flow rate of the helium gas by the cross- sectional area of the off-gas pipe. Thus, Fig. 1.2.4 gives the power-source density of a cubic centi- meter of the gas at any position in the off-gas line. These plots were made by using the well-known equations of the decay of parent products and the buildup of their daughters as a function of time. The initial conditions at the beginning of the off-gas line were taken as the equilibrium con- ditions prevailing in the fuel expansion tank. ACTIVITIES OF NIOBIUM, MOLYBDENUM, RUTHENIUM, AND THEIR DAUGHTER PRODUCTS AFTER SHUTDOWN R. B. Stevenson The activities of the materials which will plate- out on the walls surrounding the fuel channel in the ART during reactor operation will, along with other factors, determine how long @ period must elapse before reactor disassembly can proceed. The activities of the various radicactive nuclides of niobium, molybdenum, and ruthenium, and their daughter products have been determined for 100 and 300 days after shutdown and for reactor oper- _ation periods of 500 and 1000 hr. These three fission products are expected to plate-out in large ‘quantities, and therefore it has been assumed in ~this . study. that all the atoms of these elements 'rkcreated as fission products are plated-—out There is’ evidence that other fission products may plate- out; however, it is felt that the three taken into ; ‘_.uccount here are the primary ones. - The only isotopes of these elements that need " to be considered at times greater than 100 days ‘after shutdown ‘are Nb?3, Ru193, and Rul06, All the other isotopes have sufficiently short" ~ half lives that they will decay appreciably in this “time, and thus they may be neglected. The only -daughter products that will have large activities after shutdown for more than 100 days will be Rh103m gnd Rh196 (daughters of Ru193 and Ru 106, respectively). 37 ANP PROJECT PROGRESS REPORT SEORET ORNL—~LR—DWG 14219 [ TTTITT T 1 P \\ —— — HELIUM FLOW = 3000 liters/day (STP) \\\ HELIUM FLOW = 1000 liters/day (STP) 30 N L GAS VOLUME =20in3 . - N _ ‘——__'-.____.- 90 ina \ 0 St £ - B \ 2 \ =~ 20 \ 3> b N\ 2 NN a \ N , x N wl N 2 T , \ ' \\ ' e L \ \ Ny T L 20003 \\ \ [T S~ N 300in3 T TN b % s ol \\ \ \ -—-_-___-.. l\.\~~ \ =~ e \\N\ - §\ U NN ) "'--§-':\ \\-.\..::\ S 0 - - 1 2 5 10 2 5 102 2 -5 1003 - 2 5 108 TIME (sec) - ' Fig. 1.2.4. Power Density in the Off-Gas Line as @ Function of Time and Gas Volume in the Ex- pansion Tank for ¢ Fuel Flow Rate of 22 gpm. The activities to be expected at shutdown and 100 and 300 days aofter shutdown for reactor oper- ating periods of 500 and 1000 hr at 60 Mw are presented in Table 1.2.3, which also gives the average beta-ray energies, the gamma-ray energies, and the gammc-ray yields in photons per 100 disintegrations. It has been assumed that, at shutdown, the fuel is dumped instantaneously so that no more plating takes place. Also, the activities of Rh1937 gnd Rh196 gre zero at shut- 38 down, since these nuclides are carried away with the fuel and have a chance to build up only when their precursors decay. The gamma activity at 100 days after shufdown for 500 hr of operation at 60 Mw is approximately 2 x 10° curies, and at 300 days after shutdown, for the same .operating conditions, it is about 8 x 103 curies. These high activities at times fong after shutdown must be contended with in any disassembly procedure suggested for the ART. 7] PERIOD ENDING JUNE 10, 1956 Q TABLE 1.2.3. ACTIVITY OF PLATED-OUT MATERIALS IN THE ART AFTER 500 AND 1000 hr OF OPERATION AT 60 Mw . . Activity 100 Days Activity 300 Days Average - Average Ga Ray Yield Activity at_ Shutdown After Shutdown After Shutdown . : Beta-Ray Gamma-Ray mma-Ray Yie ______fl’is)_____ {curies) (curies} Nuclide Energy ‘Energy {photons per 100 After After ‘ (Mev) (Mev) disintegrations) 500 hr of 1000 hr of After After After After Operation Overation 500 hr of 1000 hr of 500 hr of 1000 hr of ' . pe! Operation Operation Operation Operation Nb%5 0.053 0.745 100 129105 4.08x105 1.82x10% 5.65x104 3.36x102 - 1.06x10% Ry103 0.074 0.498 %9 5.65 x 10° 9.57 x105 1.00x10% 1.68x105 3.06x103 5.16 x 103 Ru106 0.0131 0 0 7.84x10% 1.53x10* 6.49%10% 126x104 4.44x103 8.65x103 Rh103m 0 0,04 100 0 0 9.20x104 1.60x105 292x103 4.92x10% . Rh106 1.05 2.4) 025 0 0 6.49x10° 126x104 4.44x103 865x10° 1.55 0.5 ' 1.045 2 0.87 : 1 z - 0.624 12 ‘ - 0.513 25 - & 39 ANP PROJECT PROGRESS REPORT 1.3. ART INSTRUMENTS AND CONTROLS E. R. Mann LOUVER CONTROL SIMULATION J. M. Eastman! F. P. Green E. R. Mann : The ART simulator was used for further study of the problem involved in closing the main heat- dump louvers to prevent the fuel from freezing as the result of a fast insertion of the control rod, Im- proved transport-lag simulators were incorporated in the system. The control technique to be used will be that of closing the louvers to {imit the minimum NoK temperature. Two louver-closing speeds have been found to be necessary to keep the temperature of the NaK returning from the radiator from dropping below the fuel freezing temperature. One speed will auto- matically close the louvers from the open position to the 10% open position in 9 sec after a fast controlerod insertion. With the louvers at the 10% open position and with the four main blowers at design point speeds, 15 Mw of power will be removed. The slow-speed louver actuators will operate on o temperature signal from the radiator outlets, When this temperature drops below 1070°F, the slow actuator will start to close the louvers, The slow actuator will be capable of closing the 10% open louvers in 3 sec. For the simulation study it was assumed that all four blowers remained in operation, two being supplied from the TVA circuit and two from the diesel circuit. With the four blowers operating, a fast control-rod insertion will automatically shut off the power to one of the blowers on the TVA circuit and one on the diesel circuit. A power failure of either the TVA circuit or the diesel circuit would then leave only one blower in opera- tion. The NaK-temperature undershoots would then be somewhat less, the overshoots somewhat more, and the ultimate cooling rate lower. Although this NaK-temperature limit system seems to be func- tionally acceptable, the system will be further ex- plored on the simulator. The fail-safe character- istics are being studied, and modifications may be made which will require further simulator work for evaluation, | 10n assignment from Bendix Products Division. 40 C. S. Walker LOUVER ACTUATOR J. M. Eastman Specifications for the louver actuators were established that are based on the following safety considerations: first, if the actuators should fail, the louvers must lock into the position they are in at the time of failure; second, the components which are not considered to be of first-order re- liability must be located in an area accessible for repair without the NaK systems having to be cooled or drained; and, third, the over-all system relia- bility and dependability must be high. Since the actuators and any locally mounted associated equip- ment will not be accessible during the test, these components must be of first-order reliability, Hydraulic-actuator cylinders were selected. They will be operated in conjunction with spring-loaded clamps arranged to grip the actuator-output shafts, For steady-state louver conditions, both ends of the actuator cylinder will be vented to the hydraulic return, or drain, The louvers will be moved by valving hydraulic pressure to either end of the hydraulic cylinder. This pressure will be simul- taneously vented to a piston which will release the spring-loaded clamp. The rate of louver movement will then be controlled by an crifice in the line through which return hydraulic fluid will flow from the cylinder to the drain. The fast louver-closing rate will be obtained by opening an orifice in paralle! with the one used for slow closing. The hydraulic pressure supply unit is to be located in an area accessible for repair - work. Two pumps will be used that will operate in parallel and be driven from different electric power sources. They will be arranged so that one can be repaired or replaced without interrupting the fluid-pressure supply. Solenoid valves will be used for control of the hydraulic fluid. When not energized, these valves will hold the louvers fixed, and they will be energized to move the louvers, Low-viscosity hydraulic fluid and thin-plate orifices will be used so that the flow rates and the corresponding louver-actuation rates will be rela- tively insensitive to fluid temperature., Valves ond orifices will be located near the hydraulic- power-supply unit in an area accessible for repair work, 't ENRICHER ACTUATOR J. M. Eastman- Specifications were also established for the actuator for the fuel enricher. The actuator is to consist of a gear-motor drive equipped to provide a multiple-synchro indication of the position of the enricher piston, Enrichment is to be at the rate of 1 Ib of U235 jn approximately 22.5 sec. Three synchros will be used to indicate 1, 10, and 150 Ib of U235 per revolution. Thus the synchro dials will be read in gas-metér fashion to note the pounds of U235 displaced (equivalent volumetric measure- ment of enriched fuel mixture, Na UF6) in terms of piston position. _ _ Data obtained from calibration tests of the enricher during operation of the high-temperature critical assembly have been adjusted for changes in the geometry of the ART, The data indicate that the enriching uncertainty which will result from a fuel-nmieniscus effect at the spillover weir will be equivalent to approximately 4% of 1 Ib of U235, The enricher will be able to deliver about 130 Ib. - FLUX SERVO SIMULATION F. P. Green E. R. Mann ‘ C. S. Walker The ARE micromicroammeter and servo amplifier were used in the preliminary simulation of the ART flux servo, but the system is unstable because of the faster reactor response that results from the lower delayed-neutron contribution. The latter effect is caused by the differences between the ART and the ARE in core-residence time of the fuel and in ratio of core-réesidence time to loop-flow time, The system is being studied to ascertain methods for correcting the mstabllcty CONTROL ROD AND DRIVE MECHANISM 'S, C. Shuford? The design of the control-rod drive mechanism ‘was completed, and 95% of the parts to be supplied by vendors have been obtained. -A mechanical - shakedown test of the mechanism at operating tem- “perature is plonned Ruggedness, reliability, and ~ ease of servicing were the prime objectives in the design of the mechanism. A single rod that will operate - in a thimble passage through the north head and down the center line of the island to 20n assignment from Pratt & Whitney Alrcraft, PERIOD ENDING JUNE 10, 1956 10 in. below the mid-plane is to be employed to furnish normal shim control for changing the fuel mean temperature and to provide for emergency reactor shutdown, , _ - The neutron-absorbing section of the rod will consist of twenty-three l-in.,0 NS~ — e - B _"'";-v_/;—" 1225 rpm / . '//’ 0 . o 200 400 600 800 1000 1200 1400 1600 1800 - - 2000 FLOW (gpm) Fig. 1.4.3. 'ARTv Primary NaK Pump (Model PK-P) Performance Curves of Head vs Flow Showing Efficiency and Volute Pressure Balance Lines. 0.030 in. and the optimum radial clearance to be seal radial clearance of 0,015 in. should be suffi- 0.015 in. These clearances will result in a bypass cient to prevent rubbing of the impeller hub against flow rate of 3 to 6 gpm at 1220-gpm pump-discharge the top labyrinth seal. ' flow. - ‘ ' ' These water tests of the primary NaK pump indi- , R cate that the pump will meet the head and flow ‘Static deflection tests were made on the pump conditions required for the ART. A pump tank gas - shaft used in the water test rig and on the Inconel pressure of 10 psigshould be sufficient fo suppress pump shaft to be used in the final pumps. The cavitation at the design flow rate and up to a NaK .information obtained, along with the calculated temperature of approximately. 1380°F. The design deflection of the Inconel pump shaft during high- conditions are in the range of minimum volute temperature operation, indicated that a top labyrinth hydraulic unbalance. ' 52 o £S HEAD {ft) T SOMNADENTHY, - ORNL—LR—DWG 14956 200 500, . ".)0 (9‘“ 5% & _ 450 P — 180 : 95?’ h/ o o _a00 v . _ _|HEAD AT 3550 rpm | ] 2 o | ——l T | 7 HORSEPOWER / T . == == = HEAD ' ' / 340 250 = e 00 TOML 140 }_300 rpm . 3200 rpm 300 - 120 250 fo 10 2610 rpm / - ) —-——-—-———n——_—._d___- 200 — 80 2400 rpm —_— ‘ — S e — ——'_—_--“—_?-——7—&_ 150 2610 ("] 60 " ____.—-———7(._._.__‘_2_0_0.9@)”1 '___ | ) | / --h)):" _ 0o | -~ 100 -+ 240 — 0 : m ot '2.000 P / 1540 rpm | . _—-—b__--——_ -—-——uq-—_--—_-.-- 50 1228 rpr ' L 20 s 1225 rpm 0 - - 0 200 300 400 500 600 700 800 900 1000 1100 1200 130C 1400 1500 1600 1700 ' o FLOW {gpm) Fig. 1.4.4. ART Pi‘iinory NaK Pump (Model PK-P) Performance Curves Showing Shaft Horsepower vs Flow. SHAFT HORSEPOWER (BASED ON WATER FLOW} 956L 01 INNF ONIGNI QOId3d ANP PROJECT PROGRESS REPORT Auxiliary NaK Pump Development Tests H. C. Young J. G, Teague M. E. Lackey Water tests were conducted on the auxiliary NaK pump (model PK-A), The volute used for these tests consisted of two brass volute halves bolted together. The volute flow passage was accurately milled in each volute half, The test impeller was fabricated of brass, with vanes silver-soldered to the hub and soft-soldered to the inlet shroud. The pump assembly was tested in the test loop used for the water tests of the primary NaK pump.? A 60-hp d-c drive motor was used. No changes were necessary in the instrumentation, except that a new orifice was installed to measure the lower flow rates, Performance. and cavitation data were obtained for as wide a flow range as possible within the practical limits of volute hydraulic unbalance, and the final head vs flow curves are presented in Fig. 1.4.5. A plot of pump shaft horsepower versus flow for some head and flow conditions is shown in Fig..1.4:6. The volute hydraulic unbalance over the design range was found to be quite satisfactory, SR . ORNL-LR-0WG 14957 450 J PUMP EFFICIENCY RANGE ! VOLUTE PRESSURE 400 _ BALANCE LINE SPECIAL rpm DESIGN POINT S : 350 3450rpm 3350rpm 300 3250 rpm 3MS0rpm = 250 3 _ 3050rpm T 200 2950rpm 2850rpm 150 7 om 2250rpm 100 - : / 1900 pm 7 50 000 rpm 0 100 200 300 490 500 600 700 800 FLOW (gpm}) Fig. 1.4.5. ART Auxiliary NoK Pump (Model PK-A) Performance Curves of Head vs Flow Show- ing Efficiency and Volute Pressure Balance Lines. 54 and therefore this pump will be satisfactory for use as the special NaK pump for cooling the ART fuel fill-and-drain tank. This pump requires 300-gpm flow at a 275-ft head. The impeller diameter, the dynamic seal vanes, and the top labyrinth seal are the same as those for the primary NaK pump, and the speed ranges of both pumps are the same. Therefore the radial and axial clearances determined by the primary pump tests were used for the auxiliary pump. The pump showed satisfactory degassmg characteristics with these clearances. After the initial performance tests were com- pleted, a 0.023-in. spacer was installed between the volute halves, and tests were conducted to determine whether the addition of the spacer would change the volute hydraulic unbalence. The pur- pose of these tests was to ascertain the allowable tolerances for use in welding the lnconel volute halves together, The volute hydraulic unbalance was measured during these tests by eight static pressure taps equally spaced in the volute near the periphery of the impeller. The static pressure readings at any of the eight taps, for a given discharge head and a set flow rate, did not change over 1 psi from the measurements taken without the spacer, In general, the water tests indicate that the pump will meet the head and flow conditions required for use in the ART, both for the auxiliary and the special NaK pumps., The design conditions are in the range of minimum volute hydraulic unbalance. Cavitation characteristics of the pump, bosed on extrapolations from water to NaK, indicate that a pump tank gas pressure of 14.4 psia, or ~0.3 psig, should be sufficient to suppress cavitation at design flow for a NaK temperature of 1282°F, as compared with a required pressure of 6 psig for the primary NaK pump at the same NaK temperature. Primary and Auxiliary NaK Pump Test Stands H. C. Young J. G. Teague Fabrication of all parts for two primary and two auxiliary NaK pump high-temperature test loops is under way. At present, the pump volutes are the most critical delivery item. A method of profile- machining the volute halves from a pattern is being used. Several primary pump volute halves have been received, and the vendor is now setting up for profile-machining of the first auxiliary pump volute halves. A series of welding tests were conducted to determine the joint preparation and the welding 1] Cy & qS HEAD (ft) 450 (—— 400 a50 200 - 100 50 ERNEIBEAT b ORNL—~LR—DWG 14958 HEAD A — o 21 3550 rpm .| ~———HORSEPOWER| .| === == HEAD 3450rpm o — ——v—— AUXILIARY PUMP ] n | DESIGN RANGE— s, l—-.__- P S 80 i —~~. 70 300 fi anE- ; S 250" f L s e s _‘ “. oo ~3050 rpm o sl o 60 2850 rem 1T I 850 M .. SPECIAL PUM .\~ DESIGN RANG A ——— o o— — 50 2250 ?fim gl Uy s oS 40 150" rpm A 30 ] _— \ J / 1900 rpm \ 20 e - 1000 rpm | -—."‘—-- 1000 er . 200 250 300 350 FLOW (gpm) 400 500 550 " Fig. 1.46. ART Auxiliory NaK Pump (Model PK-A) Performance Curves Showing Shaft Horsepower vs Flow. SHAFT HORSEPOWER (BASED ON WATER FLOW) 9561 ‘0L INNr ONIGNT QOl1¥3d ANP PROJECT PROGRESS REPORT techniques required to control weld shrinkage so that final volute dimensions would be within ac- ceptable tolerances (see Chap. 3.4, *‘Welding and Brazing Investigations’’). The first set of primary pump volute halves has been welded. ' In the first tests in the new loops the pumps will be operated with water in order to check the pump assembly, test-stand vibration, loop resistance, and operation of the system throttle valve, and, of primary importance, to correlate performance and cavitation data with the data previously ob- tained on the water-test rig. Some engineering changes are being made in order to lengthen the piping on the primary NaK pump endurance-testing loop and also on the guxiliary NaK pump endurance-testing loop and to relocate these loops for use in calibrating electromagnetic flowmeters. - HEAT EXCHANGER DEVELOPMENT E. R, Dytko R. E. MacPherson J. C. Amos Intermediate Heat Exchanger Tests J. W. Cooke H. C. Hopkins' The information obtained from intermediate heat exchanger (IHE) test stand operation during the quarter is summarized in Table 1,4,2, York radiator No. 9, which is being tested in intermediate heat exchanger test stand A, is currently undergoing a thermal-cycling program consisting of 16 hr of power operation followed by 8 hr of isothermal operation. During the power phase of the cycle, the radiator NaK inlet and cutlet temperatures are 1500 and 1100°F, respectively. Isothermal opera- tion is at 1500°F. Transition from isothermal to power operation is accomplished by dropping the NaK outlet temperature at a rate of 20°F/sec for TABLE 1.4,2, SUMMARY OF INTERMEDIATE HEAT EXCHANGER TEST STAND OPERATION Hours of b f R Test Unit? Nonisothermal Total Hours Number o eason f'or . of Operation Thermal Cycles Termination Operation Test Stand A .York radiator No. 9 471 973 20 Test continuing (revised design) Circulating cold trap No. 2 1218 Test continuing (4 in. in digmeter®) NaK screen filter 264 Test continuing Test Stand B Black, Sivalls and Bryson heat 362 403 31/2 Test continuing - exchangers Nos. 1 and 2 ' " (type IHE-3) Cambridge radiators Nos. 1 ‘ 820 1345 6'/2 Test continuing and 2 {modification 3) ' Circulating cold trap No. 3 630 " Removed for exami- (4 in. in diameter) nation after heat exchanger failure Circulating cold trap No. 5 715 Test continuing (4 in. in diameter) %ncludes only units tested during this report period. bFor tests in progress the total operating time is shown as of May 15, 1956. ©This type of cold trap was praviously referred to as 80-gal system. 56 ) 4/ " @k e o rmcknin 1 re 3 . $ “-r.. & the first 200°F and 8°F/sec for the next 100°F, The final 100°F drop is at a slower rate in order to allow the outlet tempersture to level out properly. The transition from power to isothermal operation is accomplished at a rate of 10°F/sec for the first 100°F, 5°F/sec for the final 100°F, The radiator has been cycled 20 times, and the, cycling program is continuing. The ait pressure drop datal? obtained for York radiator No. 9 substantially agree with the data for York rodiators Nos. 1 and 2, and the heat transfer data substantially agree with the data!® for Cambridge radiators Nos. 1 and 2. Thus far the NaK pressure drop has increased 119% above the initial level (219% of the initial value). This radiator is the first 500-kw radiator of the revised design!! to be tested, and it is identical to York radiator No, 7 illustrated in Fig. 1.4.8 of the sub- sequent section, ‘‘Small Heat Exchanger Tests,"’ of this chapter, _ ' . ORNL heat exchangers Nos. 1 and 2, type IHE-3 (ref. 12), were removed from intermediate heat exchanger test stand B and are undergoing metallurgical inspection. Preliminary results of this inspection reveal that heat exchanger No. 1 (NaK-to-fuel heat exchanger) failed in the hot end ~ between the header weld and tube bends (see Chap. 3.4, *“Welding and Brazing Investigations®'). Severe corrosion was evident on the fuel side of the tubes, Five tubes had obvious cracks in the tension. side. The frequency and severity of the cracks were most pronounced in the end row of tubes where the distance between the tube bends and header was the shortest. A maximum of 15 mils of mass-transferred deposit was measured on the NaK side of this heat exchanger {(Fig. 1.4.7). No evidence of mass transfer was detected in heat ex- - changer No. 2 {fuelto-NaK heat exchanger), Mass ~ transfer buildup would account for the NaK pres- sure drop increase experienced during nonisothermal operation. Pressure drop was not measured in the individual heat exchangers. However, if it is assumed, on the basis of the mass-transfer evi- dence, that all the pressure-drop increase occurred ~in heat exchanger No. 1, the total pressure drop 10y, €. Amos, ANP Quar, Prog. Rep. March 10, 1956, ORNL-2061, p 54. B . - VIE, R. Dytko et al,, ANP Quar. Prog. Rep, March 10, 1956, ORNL-2061, p 52. 12, D. Pedk et al., ANP Quar, Prog. Rep., Dec. 10, PERIOD ENDING JUNE 10, 1956 Fig. 1.4.7. Maximum Mass Transfer Found in NaK Circuit of NaK-to-Fuel Heat Exchanger (ORNL No. 1), Type IHE-3, Which Operated- 1825 hr in Intermediate Heat Exchanger Test Stand B, (Seeret with-eaption) increase for this heat exchanger was approximately 180%. Black, Sivalls and Bryson Nos. 1 and 2 heat ex- chdngers, type IHE-3, were installed in test stand B, and test operotions were resumed. The stand is currently operating on a constant-power endurance run to provide corrosion information for comparison with the data for ORNL heat exchangers Nos. 1 and 2. During shutdown of the test stand, instrumenta- tion was added to allow separate measurements of heat exchanger pressure drop. After 362 hr of power operation, the NaK pressure drop has in- creased 140% in heat exchanger No. 1 (NaK-to-fuel). No increase has occurred in heat exchanger No, 2, Cambridge radiators Nos. 1 and 2, which have ~ operated for 820 hr under nonisothermal conditions, - have experienced a NaK pressure drop increase of 84%. A tcbulation of the NaK pressure drop- variations that have occurred in the various heat exchanger tests is presented in Table 1.4,3. Construction work on stand C, to be used as an ART prototype radiator test stand, is continuing. 57 ANP PROJECT PROGRESS REPORT Small Heut Exchanger Tests P stand B was shut down when the ORNL heat ex- u L. H. Devl ind : J. G, Turner!3 changer No. 1, type SHE-2, had operated success- fully £ total of 2071 hr, and th hat hanger A summary of -small heat exchanger (SHE) test vily for adotal o f, and The heal exchang stand operanon is presented in, Tabie 1. 4 4, Test '30n assignment from Pratt & Whltney Aircraft. .. TABLE 1.4.3. SUMMARY OF NaK PRESSURE DROP VARIATIONS THAT HAVE OCCURRED ~IN HEAT 'EXCHANGER TESTS Maximur_n NaK Minimum NaK Pressure Drhop ~ T;st'Unif ' : . Test PTe':"d | Operufing‘ Temperature Temperature : . Change _ S . Stand - —::; Condition _ in System in Test Unit (%, based on ' , r A - (°F) (°F) j,niiia! level) . . ORNL radiator No. 3~ SHE-B 200 Nonisothermal 1500 1270 ' York radiator No. 4~ SHE-B 140 Nonisothermal 1270 ms 4 | B | - R 580 Thermal cycling 1270 1005 : .‘8._5, - - 90 Nonrisothermcl 1255 1020 ‘ 8.2 50 Nonisothermal 1500 1300 EREN: 20 isothermal - 1270 1270 : 3.2 - York radiator No, 9 IHE-A 174 Nonisothermal 1500 1100 - 107 o * 72 isothermal 1200 1200 -0 o 160 Isothermal 1400 ' 1400 -16 25 Isothermal “1500 1500 , 0 - 30 Isothermal 1400 1400 0 512 Thermal cycling 1500 1100 28 Cambridge radiator IHE-B 552 Nonisothermal 1600 A 1100 56 Nos. 1and 2 274 Nonisothermal - 1600 1100 28 ORNL heat exchanger IHE-B 360 Nonisothermal 1600 1275 st ' No. ']o type |HE"3 . NaK dumped and leop cooled 62 to room temperature _ o 448 Nonisothermal 1600 12752 - Black, Sivalls . IHE-B 274 Nonisothermal 1600 1275 140 and Bryson heat ' ~ exchanger No, 1 Black, Sivalls IHEsB 274 Nonisothermal 1600 1100 0 and Bryson heat exchanger No. 2 “Percentage shown is within experimental error. bDurmg this run the cold trap and the plugging indicator were inoperative. Oxide contamination level of the NaK was below 150 ppm in all other cases, as determined by a plugging indicator. ' €This culculnted value was based on the assumption that all the heat exchanger pressure drop chonge occurred in ‘ e heat exchanger No. 1. , . U 58 ey A g 5 PERIOD ENDING JUNE 10, 1956 TABLE 1,4,4, SUMMARY OF SMALL HEAT EXCHANGER TEST STAND OPERATION Hours of Total Number of Test Unit® Nonisothermal Hours of Thermal Reason for Termination Operation Operafionb Cycles " Test Stand B - ORNL Heut exchanger No. 1 . 1041 2071 36 Test completed {type SHE-2) ' Process Englneering heat exchanger : 0 120 o Test continuing No. 1 (type SHE-2) York radiator No. 4 748 1356 31 Test completed (medificction 2) v Yorl.('.irodiotor No. 7 (re_vised design) - 0 - '120 Test continuing . Circulating cold trap No. 1 995 - Test completed (4 in. in diameter)® Circulating cold trep.“No. é 216 Test continuing {4 in. in diameter, modification 1) Test Stand C ORNL heat exchanger No. 2 102 634 15 Removed when restriction (type SHE-2) to fuel flow developed Struthers-Wells heat exchanger No, 1 4 280 2 Test terminated because {type SHE-2) of restriction to fuel flow York radiator No. 5 106 914 17 Test continving (modification 2) Circulating cold trap No, 4 914 Replaced by new cold trap (4 in. in diameter) “lncludes only units tested durmg fl'us report permd. ) - bFer fesfs in progress ‘lhe total operating fime ls shown as of May 15, 1956. | :and York radlafor No. 4 were removed from the stand for metallurgucoi exammohon. ~Heat hcnsfer data, fuel pressure dl'op data, and ' irNaK pressure drop data for the heat exchanger -~ _.were in substantial agreement with data previously . reported'? and did not change throughout the test, . The radlotor heat transfer data’ substantially agreed o mth the datc for ORNL radiater No. 3, and the air. .pressure ‘drop data: subsfanhally agreed with the " data' for Cambridge radiators Nos, 1-and 2, afl . previously reported.19 The radiator NaK pressure’ drop increased approximately 30% during the test operation. Much of the test program on these N c"l'hu; type of cold frap previously referred to es 80-ga| system. units consisted of thermal cycling operahoris. A - complete cycle consisted of 16 hr of power opera- tion, with maximum and minimum NaK_temperatures of 1275 and 1005°F, -respectively, and 8 hr of -|sothermal operation at 1285°F, The rate of NaK ‘temperature change durmg the transition from one .condition to the other was approximately 7°F/sec._ The heat exchanger log-mean temperature difference changed from 0°F during isothermal operuhon to 74°F during power operation, York radiator No. 7 (Fig. 1 A, 8) und Processl Engineering Co. heat exchanger No, 1, type SHE-2, were installed in stand B, and test operations were 59 ANP PROJECT PROGRESS REPORT . UNCLASSIFIED - PHOTO 23875 Fig. 1.4.8. York Radiator No. 7 (Revised Design). resumed. A new, 4-in.-dia cold trap was installed in this loop. This cold trap is identical to the previously used cold traps except that the copper cooling coif has been replaced by a stainless steel coil and additional thermocouple wells have been provided to determine the temperature profile inthe cold trap. Operation thus far has been devoted to cold-trap evaluation under simulated ART operating conditions, ‘ ' _ Operation of small heat exchanger test stand C was terminated cfter 634 hr when ORNL heat ex- changer No. 2, type SHE-2, developed a high re- sistance to fuel flow during power operation. This heat exchanger was replaced by Struthers-Wells Corp. heat exchanger No. 1, type SHE-2, and test operations were resumed. When heat transfer con- ditions were initially established, this heat ex- changer also began to develop a high resistance to fuel flow, similar to that experienced by ORNL “heat exchanger No. 2. At this time, @ leak de- veloped in the resistance heater and the stand was 60 shut down. The resistance heated section was removed for metallurgical examination, Examination of ORNL heat exchanger No. 2 disclosed extensive buildup of metal particles throughout the fuel side of the heat exchanger that would account for the marked increase in resistance to fuel flow. It is felt thot examination of Struthers-Wells heat exchanger No. 1 will reveal the same condition but to a lesser degree, ‘Evidence of mass transfer and self-welding in certain arecs of the resistance heater indicated that extreme hot spots had occurred in the heater, This would account for the buildup of metal particles in the heat exchanger. A section of the resistance heater at which severe overheating occurred is shown in Fig. 1.4.9. The metal particles which built up on one of the heat exchanger tubes are shown ‘in Fig. 1.4.10. : L A new heater design has been completed, and o heater is being fabricated. The test stand is currently being operated as a NaK loop to obtain @ of - PERIOD ENDING JUNE 10, 1954 UNCLASSIFIED § “PHOTO 26114 8 Fig. 1.4.9. Section of Resistoance Heater Removed from Small Heat Exchanger Test Stand C Showing Effects of Overheating. cold-trap-evaluation data. Upon receipt of a new ‘25-tube’ small heat ‘exchanger (type SHE-?),“ the | Struthers-Wells heater exchanger No. 1 will be- replaced, ‘and- the test stcmd wufl be put back into full operanon. Cold Trap Evuluaflon in Hect Exchanger Test Stands J. C. Amos | Six sumllar 4-m.-dm cwculaflng cold trups huve"' been operated in the various heat exchanger test . stands. These cold. traps wete previously referred to as 80-ga|-sysfem curculuhng cold traps.'5 - l_t _ ,MJ. C.','.-A'mbs, .’L._H. ,'De\i.iin,“ rund J; G.. Turnér,'_ANP Quar, Prog. Rep, Dec, 10, 1955, ORNL-2012, p 49. 15k, A, Anderson and J. J. Milich, ANP Quar., Prog. Rep: Sept, 10, 1955, ORNL-1947, p 54. has been found that care must be exercised duwring initial cold-trap operation to prevent oxide plugging during main system heatup. The original design specified copper cooling coils which necessitated -maintaining the cold-trap temperatures below 700°F ot all hmes to prevent excessive oxidation of the . -copper. This meant that at system temperatures’ above 800°F, with high contaminant saturation levels (800 to 1100°F), ‘it was possible to pre- cipitate material in the economizer which tended to settle out and plug the cold-trap inlet line, One cold trap.with a stainless steel cooling coil has been tested. Thls trap was brought up to 1200°F with the main system: The temperature of the trap was then gradually reduced by supplying maximum - air cooling and reducing the cold-frap NeK flow. By using this method of startup, no indication of serious plugging was observed.” No 61 ANP PROJECT PROGRESS REPORT Fig. 1.4.10. Fuel-Side Wall of Tube Removed from ORNL Fuel-to-NaK Heat Exchanger ORNL No. 2, Type SHE-3, Showing Mass-Transferred Metal Particles. (Seeret~withcaption) difficulty with cold-trap circvit plugging has been encountered with any of these cold traps during routine loop operation after initial heatup by the procedure described here. In all cases, the cold traps have been capable of reducing the NaK contomination level of a 77-gal NaK system to less than 100 ppm and, in several instances, to below 50 ppm, as determined by a plugging indicator. All cold traps tested on 18-gal NaK systems have been capable of reducing the contamination to below 50 ppm and, in most cases, to below the sensitivity of the plugging indicator (<30 ppm). A longer period of time is required to clean up a 77-gal system than an 18-gal system. The cleanup time appears to be related to the surface area in the system, and it is appreciably reduced when a large percentage of the system surface area has previously been cleaned at maximum system tem- perature, The rate of cleanup appears to increase with an increase of flow through the cold trap. The available cold-trap cooling limits the NaK flow through the cold traps to approximately 1.5 gpm while maintaining a 300°F cold-trap outlet tempera- ture with a system temperature of 1500°F, There- fore, the effects of cold-trap flow rates of above 1.5 gpm have not been studied. The minimum attainable contamination level has been found to be dependent on the minimum attginable cold-trap temperature. 62 However, these two variables are not necessarily closely related. In some instances the equilibrium plugging temperature of the plugging indicator may be as much as 400°F above the minimum cold-trap temperature, while, in other instances, " ihls dlf- ference may be less than 100°F, At operating temperatures above 1350°F the con- ~ tamination level has been found to increase rapidly when the system temperature is increased to beyond ‘the temperature at which the system was previously cleaned. Sudden variations in system operating " conditions, such as flow stoppage or q'brupf'flpw changes, also appear to increase the contamination level. After a system has been cleaned at the maximum system temperature, there is slow con- ~ tamination buildup if the cold trap is turned off. However, this increase appears to become slower with system age. Cold-trap evaluation tests are contmmng on the heat exchanger stands in an attempt to obtain quantitive data regarding the effects of cold-trap NoK flow, cold-trap temperature, rate of cooling, etc, Various methods of cold-trap operations are being investigated under simulated ART operating conditions to establish an operating procedure for the ETU and ART cold traps. AUXILIARY COMPONENT DEVELOPMENT D. B. Trauger J. J. Keyes Dump Yalve L. P. Carpenter Prototype dump valve No. 2 was tested to de- termine its adequacy for use as the dump valve between the reactor and the fuel fill-and-drain tank in the ART, The variables investigated were seat leakage, self-welding of seat materials, and galling of the stem on the stem guides, Seat leakage requirements are less than 2 cm3/hr with a 90-psig differential pressure across the valve. The valve temperature durmg power operation is expected to be 1250°F. The seat matericls were the some as those for prototype valvelé No. 1, that is, Kennametal 152B (64% TiC~30% Ni~6% NbTaTiC, for the plug and Kennametal 162B (64% TIC—257 Ni-5% Mo—6% NbTaTiC,) for the seat ring. Prototype valve No. 2 differed from valve No. 1 in that the stem-to- stem-guide clearance was increased from 0.001 to 16} . P. Carpenter, J. W. Kingsley, and J. J. Milich, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 60- it 0.005 in. on the diameter and @ 30-deg conical bevel was ground on the seat ring, The valve was tested for seat leakage at 1200°F in the fuel mixture (No, 30) NaF-ZrF +UF (50-46-4 mole %). The minimum leakage rote obtamed was 2.13 cm3/hr, with a 5-psig ‘differential pressure across the seat and a valve stem thrust of 1200-1b force, which gave an estimated seating pressure of 12,000 psi. - Failure of the seat ring braze occurred when the valve was opened after having been closed for 72 hr with the above mentioned force applied to the Stem.__ Examination of the seat ring and plug revealed that self-welding had occurred, and the seat ring had pulled from the valve at the brazed joint. The brazing technique has been revised for future valve fabrication. There was no evidence of galling of the stem guides, such as that. which occurred in the tests of prototype valve No. 1,. | Prototype valve No. 3 has been recewed from the manufacturer for testing. The flame plating on the stem cracked during a stress-relief operation, and the leakage past the seat, as measured with helium, increased. The valve will be accepted, however, for seat material testing. Cold Trap and Plugging ln:hcator R. D. Peak!” The cold-trap evaluation test stand, prewously described, 18 has been operated only intermittently because of continual equipment difficulties. Two different types of plugging indicators were tested, and the results correlated poorly with the data obtained with the Argonne sampler, described previously, 19:. Two methods have ‘been’ fested for ‘coolmg the ‘cold. trap - ‘with ‘air, but it is ‘opparent that air: coohng clone is msuff;clent. A ‘combina- - - tion of ‘air: with_water |n|ect|on und finally, full - -+ water flow should achleve the - Iow-temperuture;;' :..';;.ZrF . The helium ‘and- ZrF, vapor then passed R '-cold-trap operaflon des;red for the ART. systems-}l_”f-jthrough a section “of. y-m.-dm tubing extending S Other mefhods of °°°l""9 are belng mveshgcted through ‘@ hole in- fhe center of a stamless steel The second cold-trap evaiuuf!on ‘fest stand. W"s::flf-fblock whnch was 6 in. in.diometer and 10 in,. Iong. completed and will ‘be placed in operation s°°“°"i“'i-i'i_The _entrance face of the stainless jocket was ~ heated by a Colrod heater and the rest of the trap, 't'___',except ‘the exit face, .was msulofed o prowde o “linear - thermal gradlent. - The ‘temperature of -the ““tube_at the point where the ZrF, first premputated _ -~ This_test. srond is -much . more -versatile than the - . first stand, in that 'several means have been pro-- o ._,_.vnded for sampling and for. recontommanng the NaK' = o -wnh the various |mpurttles of mterest. j-jf nOn nssignment from Protf & Whltney A;rcraft. S ;-‘."_:.7 184,73, Milich and R. D, Peak,-ANP Quar, Prog, Rep, Marcb 10, 1956, ORNL-2061, p 62. %A, s, Meyer, W. J. Ross, and G. Goldber ANP Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 1 2 PERIOD ENDING JUNE 10, 1956 Zirconium Fluoride Yapor Trap M. H. Cooper20 - The program for the development of a trap for zirconium fluoride vapor has included an investi- gation of high-temperature adsorbents for ZtF , a determination of the plugging temperature of ZrF in helium in equilibrium with the fuel as a function of the fuel temperature, and the testing of prototype traps. High-temperature adsorbents were tested by ~passiig helium partially saturated with ZrF, - through a l-in.~dia pipe packed for 24 in. with the test material. The temperature of the test section was controlled by clamshell heaters. From the trap, the helium flowed first through water bottles to ‘collect any residual ZrF 4+ through wet-test gas meters, and then to the atmosphere. The results of these tests are summarized in Table 1.4.5. Alumina is the most promising adsorbent tested, while UF,, which has not yet been tested, may also prove satisfoctory, since ZrF, and UF, form a solid solution with a high melting point. Charcoal, when ~used in a trap, enhanced the formation of ZrF nuclei, which grew into large individual crystals, Charcoal thus may be a svitable packing for a thermal trap, since the formation of individual crystals does not-plug the trap. Suitable operating “conditions have not yet been established for char- coal traps, - The saturation temperature of ZrF, in helium in equilibrium must be known in order to design the fuel off-gas system of the ART for tempera- tures which will prevent the precipitation of ZrF A schematic drawing of apparatus for de- 3 fermmmg the plugging temperature of ZrF, in ~ helium- pcssmg ‘through - '/-m. tubing is shown in ~“Fig. ‘1.4 In each run, helium was bubbled ~through - “the - ‘molten fuel mixture at ¢ measured ~'temperature, and. the hellum became saturated with was assumed to be the p!uggmg temperature of ZrF, d correspondlng to the fuel femperature.- The 200, assignment from Pratt & Whitney Aircraft. - - 63 ANP PROJECT PROGRESS REPORT TABLE 1.4.5. SUMMARY OF TESTS OF ZrF‘ ADSORBENTS Ad Fuel Trap Helium - Weight of Zr MS:TT Temperature Temperature Flow Rate o r:r: 4 Not Trapped Comments ateria : erate : ©Rn R hos) " (mg) NeF 1300 1300 0.0573 260 5 NaF pellets melted in front part ' ' 0.105 119 5 of trap because of formation of 0.214 98 67.1 ZrF 4-NaF eutectic 1320 1400 1o 8007 0.0520 96 18.4 NaF peliets melted in inlet - 1200 to 800% 0.0910 112 section; contents of water trap of second run lost because of leok in water trap ' ALO, 1320 1400 0.0568 96 28.4 Plug in off-gas line believed to ‘ 0.0930 112 46.0 be Al F3 from HF reaction with A|203: water found in helium , supply during this run 1400 1400 10 800° 0,098 480” SHll in operation; no sign of plugging ' CaCl, 1315 1285 to 13355 0.0978 40 CaCl, melted and hence is un- satisfactory Charcoal 1315 1385 to 1415° 0.0978 40 62 Inspection of trap showed that 1308 to 1321° 0.0978 66 703 ZrF4 had precipitated in large 1110 to 1160° 0.0978 96.4 1496 individual crystals; plugs formed in offegas line from trap in all tests 2Temperature gradient afong the trap. bHours operated as of May 15, 1956. € Tomperature varlations during test. experimental plugging temperatures obtained as a function of fuel temperature, as well as the satu- ration temperatures predicted from both ORNL and Battelle Memorial Institute vapor pressure data, are given in Fig. 1.4.12. - The experimental plugging .temperatures agree to within 5% with the tempera~ tures predicted from the vapor-pressure data. The experimental plugging temperatures and the pre- dicted temperatures differ because the helium was not actually saturated with ZrF , in the experiments, Two prototype vapor traps, of the type described previously,2! have been tested. Prototype trap No. 2, which was connected to the fuel pump of intermediate heat exchanger test stand B and was cooled by natural convection and radiation, plugged 21, 3, Milich and J. W. Kingsley, ANP Quar. Prog. Repo DeCQ 10. 1955; ORNL'20]2, P 60- 64 with entrained fuel after 12 hr of operation. The pump off-gas line had been installed to closely approximate one designed for the ART. The fuel entrainment indicated the need for redesign of the off-gas line. Prototype trap No. 3, which is water cooled and mounted on a fuel sump through which helium is bubbled, was placed in operation on April 27, The pressure required to maintain the helium flow at 3.0 liters/min increased from 3.0 psi to 5.0 psi in 50 hr, There was no further rise in the pressure required to maintain the flow until the test was terminated after 420 hr, Two layers of ZrF, were found in the trap entrance just up- stream of the tube header. The first layer was rather dense and amorphous, and the second layer was more porous and crystalline, A total of 984 g of ZrF, was found in the layers, which totaled o at PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED "ORNL~LR—-DWG 4959 f—— INSULATION CALROD HEATER— CALROD HEATER CLAMSHELL HEATERS THERMOCOUPLE WELL —] La——STAINLESS STEEL JACKET —=—TO WET-TEST GAS METER Yo-in. DIA TUBING HELIUM-\ ZrF,-CONTAINING MIXTURE \\\\\\ R T R T A T T =SS USRS S R R \\\\\\\\\\‘\\ R AN RN N AN AR N R Y, AN \\\\\\\\\\\\\_——*FIRE BRICK > . xxxxx Fig. 1.4.11. Apparatus for ZsrF, Tube Plugging Test. - SEORETe ORNL—LR~DWG 14960 1600 & 1500 2 ® RESULTS OF TUBE-PLUGGING TESTS Q : .z 1400 uy S N & 1300 i L2 W BMI VAPOR PRESSURE DATA .~ g ' L2 g L @ // [ . é o | ## ORNL VAPOR PRESSURE DATA ~ | .~ - g 0 A | | g 1 - . 9 1100 . 1000 s e . — o 100 | 1200 _ 1300 - 1400 1500 1600 - FUEL TEMPEhATURE F) Flg. '|4 12., Plugging Temperoture of Zl'F4 in Hehum as a Funchon of Fuel errure Temperature. 23’ in. in thrckness. The werghf of ZrF, was only 30% of the amount that would have vaporized if the helium had been saturated with ZrF,. The dense deposn found in this trap near the inlet renders the design unsuitable for the ART, OUTER CORE SHELL THERMAL STABlLlTY TEST R. Curry A M. Sm:th The first test of ihe one-quarter-scale model ofk - the ‘lower half of the outer core shell23 was termi- - _nated ‘ofter - the completion of 57 thermal cycles,‘- - because a sodium leak developed in the Ioop piping . near the-test piece. "It wds decided fo partially - B 'examme the shell before continuing the test. The . center island was cut loose and lifted out of the ‘holsing | assembly to expose the inner surface of the ';h_ell f_qr,i_nSpectio'n; “A,rdye' ;beck ,bf_ 1he inrrer 220:1 nssignment from Prcm & Whifney Aircraft. o 23p, W, Bell, ANP Quar. Prog. Rep. Sepz. 10, 1955, -’.‘,r:ORNL-1947. p 65 ANP PROJECT PROGRESS REPORT sutface revealed no cracks or other flaws. Xeray photographs taken through the shell wall and the thick outer housing wall likewise did not indicate cracks or flaws, but it is judged that this inspec- tion technique was not sufficiently sensitive to ‘reveal hairline cracks, if such existed, on the exterior surface of the core shell. The internal contour of the shell was mapped by centering the shell on a lathe and measuring 264 radii. For these measurements, 15 axial stations ]/2 in, apart were designated, The radius was then measured for each 30 deg of rotation from an assigned zero position at the eight stations nearest the smail-diameter end. At the seven stations nearest the large-diameter end, the radius was measured for each 15 deg of rotation from the zero position. The measurements obtained were com- pared with measurements taken after the initial machining of the model. It was found that the average radius at each station in no case differed from the original radius by more than 0.012 in. At no station did the spread from minimum to maximum radius exceed 0,021 in., and the average spread for the 15 stations was 0,014 in. The ““crests’’ and ‘‘valleys’’ in a plot of radius vs rotation indicated that the basic deformation at each axial station was from circular to elliptical or pear shaped. This comparison between final and initial shape is invalid to some degree, because the initial measurements were taken before the shell was welded at euch end and installed into the outer housing. The shell is now back in the test rig and it is to be cycled a total of 300 cycles or to failure, Parts are being fabricated for a second 300-cycle test on another core-shell model, FORCED-CIRCULATION CORROSION AND MASS-TRANSFER TESTS W. B. McDonald - Fused Salts in Inconel and Hastelloy B J. W. Kingsley P. G, Smith : A. G, Smith24 Nine forced-circulation loops, seven fabricated of Inconel and two of Hastelloy B, were operated with fused salts as the circulated fluids during 240, assignment from Pratt & Whitney Aircraft. 66 this quarter, The operational data for these loops are summarized in Table 1.4,6. Eight of these loops were electrically heated and one (loop 4935-6) was heated with a natural-gas furnace. Two of the Inconel loops circuloted a new fuel mixture (No. 70) NaF-ZrF ,.UF, (56-39-5 mole %), and the Hastelloy B Ioops cnrculated the fuel mixture (No. 107) NaF-KF-LiF-UF, (11,2-41-45.3-2.5 mole %). = All other tests were made with the fuel mixture (No. 30) - NaF-ZrF ,-UF , (50-46-4 mole %).- Other parameters mvestlgoted were operating temperature of the fluid, surface-area-to-volume ratio, temperature differential, and Reynolds number, When com- pleted, the results of metallurgical examinafions of the loops are reporfed in Chap.-3. Dynomlc Corrosion Studies.”’ The first Hastelloy B loop failed duri_ng attempts to put it into operation. Therefore no corrosion data were obtained, but valuable mformahon per- taining to the engineering, fabrication, and opera- tion of Hastelloy B loops was obtained. As o consequence, the second Hastelloy B loop operated satisfactorily for the scheduled 1000 hr, The one gas-fired loop was terminated at 8319 hr as a result of a leak in the cooled section, This leak occurred while the cooled section was being thawed out aofter a freezeup which occurred when the pump-drive unit failed. Liguid Metals in Inconel and Stainless Steel J. W, Kingsley P. G. Smith A. Ge Smith Twelve forced-circulation loops were operated with sodium or NaK as the circulated fluid, and one was operated with water, Ten of these loops were electrically heated and three were heated by natural-gas furaces. The operational data for . these loops are summarized in Tab!e 1.4.7. The three loops that are heated by gas furnaces are - being operated as life tests.” Small centrlfugal , pumps (model LFB) are used on the gas-heated loops; the other loops that circulate liquid metals - have electromagnetic pumps. The model LFB centrifugal pump was used on the loop that circu- lated water. : When completed, the results of metallurglcal examinations of the loops are re_ported in Chup. 3.1, **Dynamic Corrosion Studies,”’ L9 A o 4 T CM o T .M . R ’fiéi;s 1.4.6. SUMMARY OF OPERATING CONDITIONS OF INCONEL AND HASTELLOY B LOOPS IN WHICH U | FUEL MIXTURES WERE CIRCULATED *Time with '.fii"np.e_rafuro dif#ér_enfi&l ‘elé't‘nlyjli’s‘héd'. Co A : T Maximum Maximum Op & '.. L ; Gl o pproximate emperature Fluid . Tube Wall erating . : Loop ; oop R Circ'inldtéd'.Fluld.f "~ Reynolds Differential o ahe T4 . Time* Comments ‘ No.’«‘w' Mahna! b bt T _ o Temperature Temperature ‘ 74259 inconel NoF-ZrF‘-UF4 - 5750 1300 1600 1700 3000 Life test; terminated LT (50_46,4 mole % i ' at 3000 hr to make Na,,’,‘o . ‘ T roomforuddifuonoi .. : L S o ‘ tests ) "':"’C “v‘ : ' .‘ s \ J ' ‘ . ‘ ‘ ! . " ' . o ! 10,000 200 ']500‘\:.‘,‘.7‘ . 1580 o , 786 . Loop‘vlnl_.urrlef._fém."’_"ti\;'nas: T e T that of Sid'ndord'l'pdp' ~ Nd F'K F-LI F-UF - 10,000 200 1500 ’ Three vnsuccessful ey (1].2*4‘-4‘5-3-2¢5"1‘*7; - . ' aftempts were made to ',:<.’ L mole 9’), No. 107 _ - | ‘ operate this loop 12 Hastelloy B. - 13 Hastelloy B ‘Same;grs;ui:‘:_t_:v‘e_‘ 710,000 200 1500 1555 1000 Trouble-fras operation T e ' | terminated on schedule -14 inconel - NaF-ZrF 15,000 300 1250 Sodium Box of beryllium plates inserted in hot leg; operated 1000 hr -8 Yes Inconel >15,000 300 1250 Sodium Box of beryllium plates inserted in hot leg; coldstrap temperature was 300°F; pump failed after 900 hr of operation <9 Yes Inconel >15,000 300 1300 Sodium Beryllium insert in hot leg; operated 1000 hr -10 Yes lncone! >15,000 300 1250 Sodium Beryllium insert in_hoi leg; cold«trap temperature was 280°F; operated 1000 hr 11 Yes Inconel >15,000 300 1350 Sodium Cold-trap temperature, 280°F; in test 12 Yes {ncone! >15,000 400 1500 Sodium Cold«trap temperature, 300'0 F; in test =14 Yes Type 316 >15,000 300 1650 Sodium Coldetrap temperature, 260°F; in test stainless . : steel =51 Yes Inconel >15,000 750 1500 Sedium Life test; 1824 hr accumulated ‘ 7432-'IA Yeos lnconel Variable Water Operated to obtain date on flow vs ' : pump speed; surface=to-volume ratio . of main NaK circuit of ART simulated - 7;\4‘39‘-51 VYes ~|ncone_'.‘l ' >15,000 665 1500 - _ Noneutectic NaK | Llfo_-tes.t; 2192 he accumulated <52 Yes Inconel © >15,000 875 1600 ' Noneutectic NaK ' * Life test; 1968 hr accumulated C . .C LY0d3IY SSTHI0Vd LID3F0¥d dNY 't ‘\ 5! PERIOD ENDING JUNE 10, 1956 1.5. PROCUREMENT AND CONSTRUCTION W. F. Boudreau "ART FACILITY F. R. McQuilkin Construction work is nearing completion on the contract portion of the Aircraft Reactor Test {ART) facility in Building 7503. Package 1 work on the building additions, building alterations, and cell installation was approximately 7% behind schedule at the 92,5% completion point on June 1, 1956. About one-half the deficiency may be accounted for by lack of breckers and reactors for the elec- trical switchgear. During the quarter the principal work accomplished included final erection of the cell tanks; erection of the stack; placement of concrete for the main air duct, the penthouse, the radiator pit, the blower house floor and equipment bases, the cell encasement, and the main building floor at an elevation of 852 ft; painting; grading; fencing; and general job and site cleanup. Package ‘2 work on the installation of diesel generators and facility, electrical control centers, and Spectrometer-room electrical and air-condition- ing equipment progressed to the 79% completion point during the quarter. The installation of the electrical control centers and the diesel generators will complete this work, The estimated date for shipment of the diesel generators is June 25, 1956. The contract for Package A work on the installa- tion of auxiliary piping was negotiated with the Y. L. Nicholson Company at a contract price of - -$50,351,62, with completion scheduled for June 15, . 1956, unless material delivery schedules interfere, Some of the Packdge ‘1 and Package 2 work may - be seen in Fig. 1.5.1, which is a view from a - location southwest of the bu:ldmg. The dark-sided portion of the main bualdmg is the addition which will house the reactor cell, the heat dumps, and - , the spectrometer tunnel; The appurtenances to = the main: bmldlng ShOW" in"this view ore, ‘from . for supporting these vnits and the radrators, and left to right, the generator house, blower house, ~vent house, ond stack. ‘In front of the vent house " can be seen the top portlon of the 10-ft-dia-and - - 23-ft-deep concrete. tank that wnll Contaln the water-submerged off-gcs piping. - o - " The west-and north faces of the main- bmldmg'l ' _and the generator house may be seen in Fig. 1.5.2. * On the left may be seen the newly installed 32-ft- wide door that is required in order that the reactor and the tops of the cell tanks may pass into the building. The cylinder bank in the left foreground, which served as helium storage for the Aircraft Reactor Experiment (ARE) will be used for storage - of nitrogen for the ART. ' In the foreground, in front of the truck, may be seen the 1500-kva trans- former and substation that will serve the purchased power (TVA) portion of the power service to the facility. The openings in the generator house, on the right, will receive the radiators for the five diesel generators that will have 300-kw con- tinuous ratings and will supply the locally gener- ated power service to the facility. In Fig. 1,53 may be seen the pressure vessel and water tank that make up the containment cell. The 24-ft-dia, 36-ft-high pressure vessel is within the 30-ft-dia water tank. The top portion of the water tank had been removed and was sitting on the main floor when the photograph was taken. Although the pressure vessel manhole is 5 ft in diameter and will pass most of the items to be installed within the vessel, the contractor will remove the top head to provide the required access for the reactor. The pressure vessel was field stress-relieved and hydrostatically tested at 300 psi during the quarter. Inspection of the welded joints between the 24.in.- dia junction panel sleeves and the 4-in.-thick shell walls revealed o failure of a joint in the heat- “affected zone of the sleeve base metal. Therefore -all eight of the one-piece sleeves were replaced - with two-piece units and a revised joint design and welding procedure was used. The field stress- relief procedure and the hydrostatic pressure tests - w:il be repeated “The coricrete penthouse ad|ocent to the cell is. ‘shown in Fig. 1.5.4. This structure will house. the NaK- pumps, pump motors, and the mechanisms ~will _also house the NaK piping to the main air “duct and to-the radiator pit below, The special ;‘equnpment room -is dlrectly ‘below and covered by fhe roof plugs shown in the open doorway. “Design ‘work continued on Package 3, which - ‘l,comprsses ‘the process equipment, process piping, etc. to be installed by ORNL, The design will probably be completed in July 1956. 69 0L R R UNCLASSIFIED PHOTO 17604 W 3 X ] s L3043y SSIII03d 133rodd dNV " ” a - PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED PHOTO 17548 Fig. 1.5.2, View of West and Norfh Faces of Buiiding 7503 and fhe Genemtor House. Photograph taken May 4, 1956, ETU FACILITY ;- M. Bender W, R Osborn G. D. Whnmon A - Design work on the facuhty in Bunldmg 920]-3‘4'_ for the. Engineering Test Unit {(ETU) is well under way, and some preliminary ‘drawings of service piping have been: completed, Line and equipment. | heating requirements have been established, and ‘the design of the NaK: piping and structural sup-: ports has been started. Preliminary layouts have been completed for the heat-dump system, includ.’ - “ing ducting, louvers, barrier doors, and the blower, The building structurcxl ‘modifications were com- ‘pleted, but additional design information has shown the need for further support for the floor structure. The lube-oil pumps, the NaK tanks, the ETU - fuel dump ‘fdnk,'-an‘d"miseelluneoows,?Ificonel ma- terials have been ordered. A decision to expedite assembly of the ART with respect to ETU com- ~pletion -has resulted in procurement of uddmonal items of equipment so that ART construction can - proceed concurrently wnh ETU operatlon. ART-ETU REACTOR PROCUREMENT AND ASSEMBLY ' _ W R, Osborn G.D. Whlfman ) Detmled study is under way on - methods for assembling the reactor. A preliminary procedure for assembling the reflector-moderator has been completed, and work is continuing on procedures M. Bender 71 ANP PROJECT PROGRESS REPORT - UNCLASSIFIED.§. - . - e PHOTO 17605 8 : ~ Fig. 1.5.3." Pressure Yessel and Water Tank That Make Up the Reactor Cell. Photograph taken May 11, 1956. L - SR 72 PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED PHOTO 17547 - Fig. 1.5.4. Concrete Penthouse Adjacent to Reactor Cell. Photograph taken Mdy 4, 1?56. for assembling other parts of the reactor. Of par- ticular concern in these studies are the problems caused by weld shrinkage during assembly, Recent data obtained from test weldments (see Chap. 3.4, **Welding and Brazing Investigations’') have shown the need to modify some assembly techniques originally considered for the reactor in order to control the dimensions of the final assembly, Stress problems which will occur in handling com- "ponents durmg assembly have also imposed re- - strictions on assembly methods and are now being ‘evaluated with respect to the|r effect on the tech- niques.’ Virtually all of the long-delwery materials for the - reactor . ‘assemblies have been ordered, but actual. fabncatlon of most items has been delayed by lack of firm drawings. ,A review of the design indicated the need for flow tests on the reflector-moderator and island assemblies. While the extent of this test program has not been cléarly defined, it is expected to cause a significant. delay. in the c0mp|etlon of the reactor assembly. | As stated above, construchon of the ART i being accelerated with respect to construction of the ETU. A co!culated risk is being ‘taken that the ETU will not show a need for correction of the ART reactor, and thus-the ART reactor will be assembled as soon as possible after the ETU assembly is completed " Parts for a third reactor “are being procured as insurance qgamst excessive time loss if the risk fails. . Fabrication of the north head of 1he ETU has been started. Some delay has occurred because of redesign, and welding progress appears to be " slower than anticipated. Delays in the receipt of the sodium-to-NaK heat exchanger and some forged parts may hold up work on this item. All 73 ANP PROJECT PROGRESS REPORT the necessary information for two of the six Inconel shells is available, and Lycoming is expediting work on these two shells. Additional design in- formation is still needed for the remaining shelis, but mandrels are being rough machined for pro- ducing them. The beryliium reflector for the ETU is being machined by The Brush Beryllium Co., but some information on this item is still lacking because of design modifications. The islond holes are now being drilled at ORNL and the island will be sent to The Brush Beryllium Co. for contouring when this operation is completed. The heat exchanger suppliers are still studying assembly methods and techniques for fabricating the channel sections of the fuel-to-NaK heat ex- changers. Potentially, this equipment still appears to be one of the controlling factors in the assembly schedule. ‘ Orders for the boron carbide tiles and boron- copper cermets have been placed with the Norton Company aond the Allegheny Ludlum Steel Corp., respectively. A fabricator has not yet been se- lected to form the boron-copper cermets into the desired shapes, Fabrication of the heovy l«in.sthick pressure shells, which must be specially formed, wili re- 74 quire equipment that is not widely available. Because of the special dies which will be re- quired, the delivery of these heavy shells may also affect the reactor completion date. Knepp Mills, Inc., has been sele;ted as the supplier of the lead shielding, which will be pre- formed and bolted to the reactor. It is proposed that only the side pieces be mounted on the ETU because of space and time considerations. No design drawings are available to describe the water shield, but it is antncapated that this, too,. will be prefabricated. A number of jigs and fixtures for use in assembly have been designed and are being procured, These include fixtures for assembling the reflector- moderator, assembly and checking fixtures for the main heat exchangers, templates and a gage for checking the shells, and several handlmg devices. An area in Building 9201-3 is now being modi- fied for use as a reactor assembly area. The area will be enclosed and air-conditioned, and it will be serviced by an overhead crane. It is anticipated that the area will be ready for use when the first reactor parts arrive in August, ¥ PERIOD ENDING JUNE 10, 1956 1.6. ART, ETU, AND IN-PILE LOOP OPERATION ART OPERATING MANUAL W. B, Cottrell An Operations Committee, consisting of repre- sentatives of the physics, design, control, engi- neering, construction, and operahng groups, wdas formed for the purpose of preparing an Operating Manual for both the ART and the ETU, Thus far, the efforts of the committee have been directed toward the preparation of the ART operating pro- ‘cedures, which, in first rough-draft form, is about 20% complete. Since the ETU is to serve as ¢ proving ground for the ART, the ETU operating procedures will subsequently be adapted from the ART operating procedures and will include other tests which may be desired. The bases for the operating procedures are the experimental objectives of the ART, the reactor design, the reactor characteristics, as determined from simulator data, and the flow diagrams and instrumentation lists for the various systems. The information being used in the formulation of the procedures is presently available in the form of design memorandums and data sheets, reports prepared by stoff members, and mmutes of desrgn and Operations Commmee meetings. The monual is to ‘provide detailed procedures, including check lists, for all operations beginning after the installotion of the ART equipment and terminating with the orderly shutdown of the re- actor. As presently outlined the manual will in- clude a check list of equipment and detailed 'procedures for cleaning the ART, for loading the ‘NaK- systems, for heating the reactor, for loading the sodium and fuel systems, for shakedown check- ~ing the fluid systems, for enriching the fuel , for ~ operating ‘at low power (10 to 100 w), for operating at mtermedlate power (2 to-20 Mw), for operatmg : -at high power, and for termingating the test,’ S By detailed consideration of procedures,: the _._‘Commnttee has - found sncompahbllmes in the - performance of “some system - ‘components - and - - limitations -in ‘the design of others which would ‘impair ‘operation during unscheduled *shutdowns. These. d:fflcuhles fall ‘into two categories: those 'V'wh:ch will require. design changes to permit opera- - - ‘tion and those which will not require changes but which will restrict the method of operation. Pro- posed design modifications for resolving these difficulties are currently being evaluated. IN-PILE LOOP DEVELOPMENT AND TESTS D. B. Trauger Operation of Loop No, 4 C. C. Bolta? R. A. Dreisbach’ J. A, Conlin W. T. Furgerson C. W. Cunningham D. M. Haines! W. L. Scott In-pile loop No. 4, which was described in the previous report,2 was cut from the shielding plug, and the nose end is being sectioned for metallo- graphic examination. This loop was operated for a total of 501 hr, including 316 hr at the maximum design temperature differential of 200°F and 80 hr at lower temperature differences. The high- temperature point in. the loop was maintained at 1500°F throughout the periods of operation with a temperature differenticl imposed on the system. The loop was operated under isothermal conditions at lower temperatures during the remaining 105 hr, while the reactor was shut down for refueling and for other maintenance. The power density in the loop was 778 w/cm3 during operation with the maximum temperature differentiol, as determined from the heat removal through the heat exchanger and the total volume of fuel in the nose from the heat exchanger outlet to inlet. Functionally, the loop operation was good. The heater circuits, which had better insulation than ~-the insulation used for the heater circuits of loops ‘tested previously, were free of failure. The pump _operated smoothly “throughout- the test, and con- trols for the heat exchanger and other components' _were satisfactory. However, 7 of the 14 nose ‘thermocouples failed in operahon, _e:ther by the “junction coming off the pipe or by breakage of one of the wires (usually the Chromel wire). These failures are now believed to have been caused by poor mstallahon ,techmques.' Slx of the seven 1on osmgnment frorn Profl & Whhney Alrcraft, 2C C. Bolta et al,, ANP Quar. Prog Rep. March 10, 1956, ORNL.-2061, p 41. 75 thermocouple lead wires which failed had, in- advertently, been clamped rigidly to the end of the outer wall of the double-walled heat exchanger. This could have resulted in strains on the thermo- couple wire from the thermal expansion of the pipes. The loop design has been modified to correct this condition. As with loop No. 3, both the bearing-housing and the pump-sump purge outlets plugged. The bearing purge outlet plugged five days after startup, ond the sump purge outlet plugged four days later. The cause of this plugging is still being investi- gated, The fission-gas absorption traps from loop No. 3 have been sectioned. An extensive black deposit was found in the inlet of one and a brownish film was found in the other. This black deposit, believed to have come from the pump lubricating oil in the bearing housing, is to be given radio- assay and spectrographic analysis to determine its composition (see Chap. 4.2, ‘‘Radiation Damage'’). It is planned to operate future loops with little or no purging of the bearing housing and reduced purging of the pump sump to minimize the probability of plugging. Previously, the higher purging rates were considered necessary to mini- mize hydraulic-motor-0il contamination, Experience indicates that, although some contamination results from operating with the purges plugged, such con- tamination is not a serious problem during dis- assembly. Other major difficulties encountered were a leak in the pump bulkhead, probably through a glass heater or thermocouple-wire seal, and an excessive radiation level in the cubicle after shutdown. A new type of electrical seal for the pump bulkhead is being investigated. The excessive activity of the cubicle after shutdown was caused by the deposition of material on the purge-outlet-tube . walls during operation. Radiation levels as high as 50 r/hr were measured on the l/‘-in. pump-purge- outlet tube. This caused considerable difficulty in loop removal because it necessitated a relatively large amount of preparatory work. The design has 76 of 200°F. now been modified so that all tubes may be pinched leak-tight, cut, and removed remotely to make possible the removal of the activity prior to the entry of personnel into the cubicle. The only work which will require entry to the cubicle will be the removal of the air and water lines. Loop No. 5 D. M. Haines In-ptle loop No. 5 was completed- and mserted in the MTR, but could not be filled. This loop was to have operated for the duration of two MTR operating cycles with a maximum fuel tem- perature of 1600°F and a temperature differential ‘The improved thermocouple installa- tion mentioned above was used in the fabrication of this loop, and the rear section was modified to simplify removal and to provide a second hermetic seal to back up the seal at the intermediate bulk- head to prevent fission-gas leakage. The loop is being returned to ORNL for salvage. Horizontal-Shaft Sump Pump for In-Pile Loops W. So Kcrns An improved prototype {Mark ll) of the horizbntal- ' shaft sump pump designed for in-pile {oop operation completed 1000 hr of a 2000-hr endurance test at 4500 rpm and 1500°F, The test was terminated by shaft seizure, and disassembly showed that the seizure was caused by the buildup of zirconium fluoride on the shaft slinger. This was the first pump operated at 1500°F, the previous pumps.being operated at 1400°F. The vapor pressure of zir- conium fluoride in the fuel mixture (No. 44) NaF- ZrF -UF ;- (53.5-40-6.5 mole %) used in this test is 2 ‘mm Hg at 1400°F and 4.5 mm Hg at 1500°F, The increase in vapor pressure is considered to have been the cause of the .buildup of zirconium fluoride on the slinger. A new pump is to be buiit with increased clearances at the slinger. 30n assignment from Pratt & Whitney lAirért:l-ft. ) Part 2 - CHEMISTRY W. R Grimes 2.1. PHASE EQUILIBRIUM STUDIES! C. J. Barton R. E. Moore R, E. Thoma H. Insley, Consultant The several methods described in previous reports of this series were used for further phase equilibrium studies of a variety of binary, ternary, and quaternary systems, Although the phase diagrams are not considered to be final in every respect, a compilation of the diagrams of binary systems consisting of UF, or ZrF, with each of the alkali fluorides is presenteg here. The striking differences observed in the diagrams for these systems indicate a need for detailed study of some of the complex crystal structures that characterize certain of these materials. The RbF-ZrF, and RbF-UF, systems now appear to be moderately well described., A thermal effect in the RbF-ZrF, system at 375°C that was previously reported to be a eufectic temperature has been shown to be a lowered inversion temper- oture of the 2RbF.ZrF, compound; the eutectic contains 42 mole % ZrF , and melts ot 410°C. Solvents with compositions near 35 mole % NaF, 25 mole % RbF, and 40 mole % ZtF , apparently | dissolve up to 4 mole % UF, at Ilquldus temper- atures below 480°C and up to 7 mole % UF, below 510°C. Such mixtures are of interest os reactor fuels, _ Detailed examinations of some ternary systems containing BeF, are being made. While extremely low melting pomts can be obtained (the lowest eutectic observed to date in the NaF LiF-BeF, system melts at 318°C), it is not yet npparent that these materials offer fuel mixtures that are substanhully better -than those avallable in the, ZrF (~containing systems. | Sfudles of the phase. behavior of CeF in bmary systems with alkali flucrides have been conflnued . These systems are of interest because of concern - over -the behavior of flssmn-product fluondes in _.' high-power long-duranon reactor OPe"G*W“- - GENERAL COMPARISONS OF THE BINARY SYSTEMS MF-ZrF "AND MF-UF4 i R E Thoma L . The two porullel famllles consnstmg of bmory systems of UF, and of ZrF, with each of the alkali metal fluorides have been under investi. gation here for several years, and the character- istics of these systems are relatively well known, In the course of this research mony people have contributed to an understanding of these phase relationships, and the information on phase behavior is scattered throughout a large number of reports in this series. Accordingly it has seemed worth while to prepare a brief summary of the differences and similarities in these several phase systems, No attempt has been made in this concise compilation to give credit to those who did the work, Both these families of binary systems are much more complex then a cursory examination would indicate.2 It is obvious that fundamental studies of the structures of the complex compounds involved would be of general value. The large number and wide variety of complex compounds observed in these systems are indicated in Table 2.1.1. (The ratios shown are the ratios of the alkali fluoride component to the ZrF, or the UF, component.,) Comparisons of the phase dmgrams presented in Figs. 2.1.1, 2.1.2, ond 2.1.3 display several of the striking character- istics of these materials. A stable, congruent, high-melting-point compound (Fig.2.1.1) with the formula 3MF-ZrF, or 3MF UF, characterizes all the diagrams, except those for LiF-UF , and NaF-UF,. For the LiF-UF system, which represents the lowest ratio of radlus of M* ion to radius of M4t ion, the 3:1 compound does not exist; for this system, and this system alone, an incongruent compound 4LiF-UF, is ~observed, In the NaF-UF, system the compound 3NaF.UF, is congruent but relatively low melting, and it |s unstable at temperatures below about - 530°C. MThe- petrographic _examinations reported here were performed by G. D. White, Metallurgy Division, and -~ Te. Ne McVay and H. Insley, consultants, The x-ray examinations were performed by R. E. Thome and B. A. Soderberg, Materials Chemlstry Division. 2The case of the related system KF«ThFy is similar. The early concept of the system reported by E. P. Dergunov, Doklady Akad. Nauk S.5.5.R. 60, 1185—1188 (1948), was @ simple one, but a more recent report by w. J. ASkGf, E. R, Segnlt, and A. W, Wylse, ] Chem, Soc. 1952, 4470, shows the system to be relatively complex. 79 ANP PROJECT PROGRESS REPORT TABLE 2.1.1. COMPOUNDS* or'z:g‘on UF, WITH ALKALI FLUORIDES 3:1 2:1 3:2 5:3 5:4 3LiF°ZrF4 {c) 3N¢:F-ZrF‘ (<) SKI"'-ZrF4 (<) 3RbF-ZrF4 (<) 2LIFZrF () 2KF-ZeF , (1) 2RbF-ZrF, (1) ALiF-UF o 3NaF-UF‘ {c) 3KF4UF4 () 3RbF-UF, (c) 2NaF-UF, (1) 2KF-UF, (1) 2RbF-UF, (1) 2NaF-ZeF, (1) 3NaF-2ZrF, (S) 3KF-2ZeF (1) TNaF+6ZrF , (c) .'SRbF-AZrF4 {c) 7LiF-6UF, (1) 7NaF+6UF , (c) 7|(F'6UF4 {c) 7RbF-6UF , (f) S5NaF-3U F4 {1 1 3:4 2:3 132 1:3 14 1:6 3LIF-4ZeF , (1) NaF-ZrF, (M) 3NaF-4ZeF, (1) KF-ZrF, (c) RbF-ZeF, (c) RbF-22rF, (1) | LiF-4UF, (1) NaF+2UF , (S) KF-2UF, (1) RbF-UF, (c) 2RbF+3UF (1) RbF+3UF , (1) RbF+6UF , (1) *(c) ~ Congruent compound, () Incongruent compound. (M) Metastable compound. (S) Subsolidus compound. A series of incongruent compounds, 2MF.ZrF and 2MF-UF4,' are observed in all the diagrams, except those containing LiF. The compound 2LiFUF, does not exist, and 2LiF.ZrF, has a relatwely low melting point but is congruent. All the incongruent 2MF.ZrF, and 2MF UF, com- pounds, except the 2NaF.UF, material, melt to form the compounds 3MF ZrF or 3MF-UF, and liquid. : The |ow-me|fmg-pomt central portion of the binary mixtures, that is, the mixtures with 35 to 55 mole % ZrF , or UF4, as shown in Fig, 2.1.2, exhibits many vorlctlons in compound types. The occurrence of the unusual compound 7MF-6ZrF, “or 7MF-6UF , is remarkable, The compound hos' 80 the same crystal system wherever it oceurs, and it displays no polymorphism. One member, the compound 7NaF 6ZrF forms an extensive interstitial solid solutlon series with the -next lower ZrF 4 compound. Compounds formed from mixtures with high concentrations of ZrF, or UF, are invariably incongruent, but they display wide variations in degree of incongruency, These compounds, as shown in Fig. 2.1.3, occur at widely varying compositions in the range 55 to 95 mole % ZrF, or UF,. Several of the diagrams in this series are considered to be preliminary, and the systems those diagrams depict are still being studied. Al o PERIOD ENDING JUNE 10, 1956 SEORET" ORNL-LR-DWG 14626 1000 CsF-ZrF, RbF—ZrFy KF~Zrf, NaF-Zrf, LiF-ZrF4 P a ¥ 900 AMINZN NN 800 ? AN o \ lfil:j 700 / ‘ ‘ 2 g \ \ o —— W : \ Z 600 _\/ ’... N N u 500 j 3 400 ] ol St — &1 S N N 5| & N N Gl A ol @ 5l % “ Qf o | x x| ¥ Z} = 3 M} N " o MmN mf N o 300 | i [ I I O 10 20 30 0 0 20 30 O {0 20 30 O 10 20 30 O 10 20 30 ZrF, (mole %) 1000 CsF-UF, /\ RbF —UF, /\ KF -UF, NaF -UF, LiF - UF, 900 f 1\ / \\ / / L | \ & 700 - - e NI | \ .— « & a - ! \\ '1"\1/ \/ 500 \ N \ \/‘. \/' [ so0 [ ot | s N ol & & of o b w w ™ 'i'i ué 300 o ml 5 35 40 45 50 35 40 45 50 35 40 45 50 35 40 45 50 35 40 45 50 ZrF4(mole %) 1000 CsF-UF, RbF -UF, KF-UF, NoF-UF, LiF-UF, a a a a 4 900 800 \\ —\ ~ \// \ e [\ s ~ i w 700 N . e = V =2 - . k2 o & 600 B S 7 M Ww o = 500 .0 - - L < w |.I? 400 - o 1= 5 2 5 o wl> o b w U gl% £ z 5 o ~|lx ~ ~ ~ 300 ] | | | 45 50 35 40 45 50 50 35 40 45 50 35 40 Uf, {mole %)} H o, 35 40 45 S0 35 40 Fig. 2.1.2. The Systems MF-ZiF, and MF-UF,, Where MF Is an Alkali Fluoride, in the Composmon Range 35 to 55 mole % Z«F , or UF4. 82 PERIOD ENDING JUNE 10, 1956 PEGRES- ORNL~LR-0OWG 14628. CsF-2rfy RbF-ZrFy KF-ZrFy T NaF—ZrfFy LiF-2ZrFq 900 - ' : ‘ 800 // . // / o & 5 / / < o & / = 600 < ¥ / 500 / : : o = - N . " T 400 " H s %'t . : 3 o - 300 : & ' n m 55 65 75 85 55 65 75 85 85 65 75 85 55 65 75 85 55 65 75 B85 2rf, {mole %) 1000 _ ' CsF-UF, / RbF-UF, / KF-UF, / / NaF~UF, LiF-UF, / 900 / 7/ / / / e / / 800 7 - / / /| / / - & - /[ 1&' 700 ) — . ] J S « Lt £ ) . = 600 | - 500 s — a0 & " 400: —— et ' 5 5 5 u¥ g ‘g .l mi© 2 - o < wif wl o o u o 300 ® : "55 '65-75.8 55 65 75 85 .55 65 75 B85 55 65 75 8 55 65 75 85 : , - UF, (mole %) o Fig. 2.1 3 The Systems MF- ZrF und MF. UF4, Where MF Is an Alkuh Fluorlde, in the Composmon Range 55 to 95 mole % ZrF, or UF‘. : o "ANP PROJECT PROGRESS REPORT THE SYSTEM RbF-UF, H. A, Friedman R. E. Moore The system RbF-UF, has been investigated recently because of its importance in the qua- ternary fuel system NaF-RbF-ZtF .UF,. The earlier work of Blakely et al.3 has been largely confirmed by thermal analyses of slowly cooled melts, In addition to the compounds originally reported, however, there is an incongruent com- pound 7RbF-6UF . : Quenching expenments on the system are now under way, but the work is not yet far enough advanced to warrant presenting a phase diagram of the system, Petrographic examinations of a series of quenched samples containing 33.3 mole % UF, indicated that a rapid inversion of the compound 2RbF.UF, occurs at some temperature below 572°C, The congruent melting point of RbF.UF, (727°C) was confirmed by petrographic examination of a series of gradient-quenched samples containing 50 mole % UF,. The compo- sition of an incongruent compound previously thought to be RbF-4UF has now been established os RbF:3UF, by the observahon of single-phase material in u series of equilibrated and quenched samples containing 75 mole % UF,. This phase is not stable above 714°C. In mixtures con- taining between 50 ond 100 mole % UF,, two phases, in addition to RbF.3UF,, have been observed in both slowly cooled and quenched samples. One phase contains more than 80 mole % UF ,, while the other phase contains between 50 ond‘66 7 mole % UF . THE SYSTEM RI:F--ZIF4 R. E. Moore Several revisions have been made in the tenta- tive diagram presented previously for the RbF- ZrF, system,® The revised diagram is given in Fig. 2.1.4. Visual observation experiments have shown that the eutectic between 2RbF.ZrF 4 and 5RbF 4ZrF contains about 42 mole % ZrF and melts at about 410°C. Results of quenchlng experiments had indicated previously that the eutectic contained 3. P, Blakely et ale, ANP Quar, Prog. Rep. June 10, 1951' ANP.65' P 87' Flg. 4.'. 4H. A. Friedman and R. J. Sheil, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 72, Fig. 4e1s 84 38 mole % ZrF, and melted at 375°C. It is now apparent that 375°C is the lowered rapid inversion temperature of 2RbF.ZrF , rather than the eutectic temperature, In the region between 37 and 44.5 mole % ZrF,, 2RbF.ZrF, appears optically to be like quench growth at all temperatures above the lowered inversion temperature. This accounts for the misinterpretation of the quenching data. . Petrographic examination of a series of samples containing 48.3 mole % ZrF, established that 5RbF4ZrF , is the primary phuse, the liquidus temperoture is about 396°C, and the solidus temperature is 390°C, Thus, the eutectic between 5RbF-4ZrF, and RbF.ZrF, contains about 48.5 mole % ZrF and melts at 390°C Exammchons of slowly cooled preparations have shown the existence of a compound containing more than 50 mole % ZrF . This compound has been tentatively |dent|hed as RbF-2ZrF , by the observation of nearly single-phase matenal in a series of equilibrated and quenched samples containing 66.7 mole % ZrF,. This compound melts incongruently to ZrF , and liquid ot 447°C, THE SYSTEM Nt‘IF'--RhF-ZI'F'4 R. E. Cleary? H. A, Friedman A phase diagram of the fuel solvent system NaF-RbF-ZrF,, based on thermal analysis of slowly cooled melts, differential thermal analysis, and quenching experiments, is shown in Fig. 2.1.5. Some of the compatibility triangles and quasi- binary mixtures were reported previously.® The present status of the information on the com- patibility triangles, eutectics, peritectics, and boundary curves is shown in Fig. 2.1.5. The locations of the eutectics and the exact paths of the boundary curves are not considered to be final in this diagram; they are contingent on confirmation by quenching experiments. The phase relationships of the recently dis- covered compound RbF2ZrF, in the termary system have not yet been determmed The ter- nary compounds of the solvent system are: NaF -RbF.ZrF ,, which melts congruently at 615°C; 3NaF 3RbF+4ZrF ,, which melts incongruently at 450°C; and NaF:RbF.2ZrF, which melts congru- ently at 460°C. A fourth ternary compound with 50n assignment from Pratt & Whitney Aircraft. 6H. A. Friedman, ANP Quar. Prog: Rep. March 10. 1956, ORNL-2061, p 72. PERIOD ENDING JUNE 10, 1956 COMPEENTITI ORNL~LR-DWG t4629 1000 900 Pt \ / \ \ _ 1 800 - 1 | ’ S I [ & / 5 700 / < o w a = 600 ! ) . \ ! \ | ! | / I { \ [ < ! ’\l 4 SNIVINAS/ N ~NE 400 o H— i 5 v/ L u 4 (I A ; "M ' ~ | "_.'5 'T'\ 0 ,l P 3 W = 300 i & RbF 10 20 30 40 50 60 70 80 90 Zrf, ZrFy (mole%) Fig. 2.1.4. The System RbF-Z(F,. the approximate composition 3NaF 3RbF 24eF may exist at high temperatures. Quenching sfudnes have not yet established temperature -limits on its Preliminary quenching studies ‘show compound -~ 3NaF. 3RbF-42rF - melts mcongruently to: NoF-RbF-ZrF ~and hqund The - optical ‘and x-ray data for. the compound are presented in the followmg section. B - stability. that - the. OPTICAL PROPERTIES AND X-RAY PATTERNS FOR 3NuF-3RbF-4ZrF > R.E. Thoma H.lnsley | “Listed below are the identifying characteristics of 3NaF 3RbF 4ZrF o O new compound encounfered “in the phose studies described above, The symbol, _ | d(»&) means the distance between reflecting planes measured in angstroms. The term I/1] refers to the relative intensity of the lines as compared with an arbitrary value of 100 for the strongest line. Under optical properties, N, and N refer to the ordinary and extraordmary lndrces of “refraction of uniaxial crystals, The compound s ~ believed to form an mcomplete Solld solution with . ~ the compound NqF-RbF-?Zr_F_ ¢ - - Optical datu ' Unioxial negatlve ;':' = 10436 Ng = 1f430<.':'_, Xeray data: o dA) S, 677 s 611 15 5.72 15 85 ANP PROJECT PROGRESS REPORT o d(A) 5.37 4.67 4.21 4.11 3.98 3.82 3.72 3.56 3.52 3.44 3.31 3.22 3.12 3.11 3.07 3.01 2.95 2.571 2.482 2.356 2.243 2.164 2.111 2.088 2.060 2,025 2.004 1.976 1.947 1.909 1.861 1.812 1.727 1.638 1.591 /1 30 9 37 10 15 22 30 5 11 100 30 63 100 20 10 10 15 15 33 30 28 18 85 10 10 10 THE SYSTEM NcF-RbF-ZrF‘-U Fa H. A. Friedman H.- Davis? Two areas of the quaternary system NaF-RbF- ZiF ;~UF, offer mixtures which may be svitable fuels for circulating-fuel reactors, Solvents with approximately 10 mole % NaF, 52 mole % RbF, and 38 mole % ZrF, with the addition of 4 mole % UF, give a fuel with a liquidus temperature of approximately 500°C, With the addition of 7 mole % UF , a fuel with a liquidus temperature of ap- proximately 545°C is obtained. These fuyel mixtures have RbF.UF, or RbF.3UF, as the primary phase. No serious segregation of the uranium phase has been evident in the experi- mental work, These mixtures are not optimal choices for the fuel mixture because the uranium phase contains a high concentration of'UF4' and there is o large difference between the temper- atures at which the uranium phase and the primary phase of the solvent separate, Solvents with approximately 35 mole % NaF, 25 mole % RbF, and 40 mole % ZrF4- with the addition of 4 mole % UF4 give a fuel with o liquidus temperature of approximately 480°C. With the addition of 7 mole % UF , a fuel with a liquidus temperature of approximately 510°C is obtained, At both UF, levels the uranium is con- tained entirely in the solid solution 7NaF -6Zr (U)F ,. Mixtures of this general type should have physical properties and corrosion characteristics thot would make them of definite interest as fuels, THE SYSTEM NaF-L!F-Ber R, E, Meadows Quenching studies of the system NaF-LiF-BeF, were continued, and it is now possible to present a diagram (Fig. 2.1.6) of the triangle having LiF, NaF, and NaF:BeF , at the apexes. Work on the other portion of the system is under way, Petro- graphic identification of phases in this system is very difficult, because several of the compounds have nearly the same indices of refraction and are isotropic, or nearly so. As a result reliance was placed on identification by means of x-ray dif- fraction patterns; even this identification is 70n assignment from Pratt & Whitney‘Airc'i-i_;fh‘ - 2NaF- BeF . . PERIOD ENDING JUNE 10, 1956 ORNL-LR-DWG 14630 ZrFy 912 TEMPERATURES ARE IN °C RbF- 2 ZrF, /s ? " 3NoF-4ZrF, /7 403 537 Z RDF-NaF-2ZrF, 510 RbF - ZrF, 7NaF - 6ZrF, 460 385 5RbF- 4ZIF, 505 - 395 438 -\ 400 3NaF . 3RbF - 4ZrFy 475 _ E 630 . =\ 65 620 . 2NoF - ZrF, 2RbF-2rF, - 3NaF-ZrF, 747 NaF 3RbF-2ZrF, RbF 298 - er3 - 790 Fig 2 I 5 The Sysfem NuFaRbF-ZrF - dsfflcult in some cuses, since the s!andord paflernsj i were obtained from sumples whlch are not smgle- ‘ S phose materials, . ;_ - The diagram ‘shows the followmg companblhty' © friangles: " LlF-ZNGF-LiF-25eF2-2NaF-BeF LiF. - © "~ 2NaF. Ber-NaF ‘and 2NaF.LiF+2BeF . -NaF-BeF - ‘In-addifion to the congruently melnng ~ .compounds which- form - the apexes of these . .?compatlblhfy frnangles, the data show a subsolidus compound, NaF.LiF.BeF,, which decomposes to LiF and 2NaoF.LiF 2Bei= above about 240°C (ref. 8), and another subsohfius compound, probably 5NaF L|F~3BeF2, wh:ch decomposes ubove ubout ' - 320°C to 2NaF . LIF 2BeF 2NaF:BeF ., and LiF. . Below - -320°C, g '5NaF-L|F 3BeF ., are the apexes ofa compahb:hty o tnangle. ‘No otfner subsolidus relatlonshlps have - been definitely established, " (Figs 2.1.6) the blnary and quasa-blnary eutectics - LiF, . 2NaFL:F2§ o and ‘In-the diagram . are indicated, and the dotted lines are the approxi- - - mate pnmclry phose boundoraes which meet at the 8w. Jahn, . anorg, u. allgem., Chem. 276, 113—127 (1954). 87 ANP PROJECT PROGRESS REPORT TP TDEN i ORNL-LR-DWG 14631 NGF' Ber EUTECTIC 380 343 EUTECTIC . ' 332 EUTECTIC 2NaF - LiF -28eF, -\ 345 355 - - TERNARY EUTECTIC EUTECTIC 348 340 ~ 2NoF- BeF, NaF -LiF -BeFp ~ 595 - ~a EUTEGTIC TERNARY EUTECTIC — 570 328 : 485 _ ” 7 TERNARY EUTECTIC 7 480 . / / LiF ~~ NoF 845 EUTECTIC 995 649 Fig. 2.1.6. The System NaF-LiF-NaF.BeF,. ternary evutectics shown in each of the three compatibility triangles. There is a limited solid solution region in the ternary system for a high- temperature modification of 2NaF:BeF,, which is not indicated on the diagram. It is not likely that the limits of this region can be accurately de- termined with the experimental methods now being used. The phase which was previously? believed to be a ternary compound and which was tentatively assigned the formula LiF-7NaF-.4BeF, was identified by G, D, White through the use of high- temperature x-ray diffraction as a polymorph of 2NoF-BeF,. This phase, which is stable above 320°C, had not been reported in previous investi- gations of the NaF-BeF_ system!¢:11 and had not been. found in quenched samples of NaF-BeF mixtures examined in this laboratory because |ts inversion is too rapid for it to be quenched in the e M. Bratcher, R. E. Meadows, and R. J. Sfieil, ANP Quar. Prog, Rep. March 10, 1956, ORNL.-2061, p 75. 106, Thilo and F. Liebau, Z. physik. Chem. 199, _ 125..141 (1952). p. M. Roy, R. Roy, and E. F. Osborn, J. Am. Ceram. Soc. 36(6) 185 (1953). 88 - binary system. It is, however, stabilized by the addition of LiF, with which it undoubtedly forms a limited solid solution. Quenching experiments with compositions on the LiF-2NaF.LiF-2BeF, join on both sides of the mixture NoF-LlF-Ber (37-26-37 mole %) show this mixture to have approximately the compo- sition of the eutectic which melts at about 340°C. Petrographic and x-ray diffraction examinations of quenched samples and slowly cooled prepa- rations show that the 2NaF -LiF.2BeF -2NaF.BeF, join is a quasisbinary system wnh a eutectlc that contains 45 mole % NaF, 16.5 mole % LiF, and 38.5 mole % BeF, and melts at about 343°C.. A phase that appecred at a temperoiure below. about 320°C in quenched samples was assigned the formula 5NaF-LiF-3BeF, because the x-ray diffraction pattern of a quenched sample of this composition was not found to contain lines that - could be assigned to any other known phase .in the system. The conclusion that this compound exists only below the solidus temperature is based on the following evidence. First, slowly cooled preparations within the 2NaF.LiF:2BeF,- NaF-BeF ,-2NaF.BeF, triangle contained only the three phases at the apexes, since the reaction rate in the solid state is too slow to permit equilibrium to be obtained. ‘Second, examination of quenched samples in the LiF-2NaF.LiF 2BeF - 2NaF BeF,. triangle showed that the ternary eutectic (mp, 328°C) is above the upper limit of stability of the compound and that the phases just below the solidus temperature are those at the apexes of the triangle, The x-ray diffraction pattern of 5NaF.LiF -3BeF, is related to that given by Jahn12 for low-femper- ature 3NaF-.LiF- 2BeF2 Quenched compositions corresponding to 3NaF.LiF-2BeF, contained 2NaF.LiF-2BeF,, 2NaF-BeF,, and LiF just below the solldus temperature and 5NaF.LiF -3BeF,, 2NaF.LiF. *2BeF,, and LiF below 320°C. No evidence has been found in this laboratory to suggest that 3NaF. LiF-2BeF , exists. The locations of the phase boundaries and ternary eutectics were deduced from the primary and secondary phases observed in quenched samples of the compositions within the compati- bility triangles and the temperatures at which the phases appeared, as well as from previously obtained thermal analysis data, THE SYSTEM NnF-RbF-BeF2 L. M. Bratcher Thermal analysis data were obtained for o number of compositions in the NoF-RbF-BeF system containing 50 mole % BeF,, and some of the slowly cooled melts were examined petro- graphically and by x-ray diffraction. The avmlable . data - show that the 'NaF-3RbF. BeF, join is a quasi-binary system which has a eutectic with - NaF-RbF- BeF the . approxlmate composmon join NaF-2RbF - BeF2 a quosu-bmary system which has a eutectic . with -~ the composmon NaF RbF- BeF2 (43-38-]9 mole %) that melts ot 655 + 5°C. melts mcongruenfly. , mdncate that the - add;tlon ‘of NaF lowers the incongruent . melhng point of RbF-BeF 12y, Jahn, Z, anorg, u. allgem. Chem. 277, 274 (1954). The NaF- RbF. BeF - ~join is’ compllcated because one_or. more. 1ernary ' : compounds exist on or near the join and RbF. Berr “Thermal “anclysis data: None ofr,’ the slowly cooled melrs was completely homo-_ - PERIOD ENDING JUNE 10, 1956 geneous, but the mixtures containing 45 and 50 mole % NoF were predominantly one phase, and therefore there may be a compound with the compo- sition 2NaF-Rb_F-BeF2. THE SYSTEM NGF-KF-L!F-UF4 Ro Ja Sheil 'Data reported previously on the liquidus temper- atures of mixtures formed by adding UF, to NaF- KF-LiF (11.5-42.0-46.5 mole %) indicated that the liquidus temperature varied quite rapidly with changing UFA concentration, at least with mixtures containing approximately 4 mole % UF, (ref. 13). Visuval observations of liquidus temperatures in this system during this quarter gave values of 550, 570, and 595°C for mixtures containing 5.0, 7.5, ond 10 mole % UF respectively. The values obtained by this method for the mixture containing 10 mole % UF, agreed quite well with earlier thermal analysis data. |t now appears that the liquidus temperatures of mixtures in this system containing 0 to 10 mole % UF, are not so strongly dependent upon UF, concentrations as the earlier datal3 had indicated. ALKAL1 FLUORIDE-CeF; SYSTEMS L. M. Bratcher Preliminary data on the NaF-CeF ; and RbF- CeF3 systems were given in the prevrous report, 14 Study of the LiF-CeF, system was initiated recently, and thermal analysis data obtained with five compositions in this system showed that there is a eutectic confaining approximately 19 mole % CeF ‘that melts at 755 £ 5°C. For the NoF-Ce F system, extrapolahon ‘of thermal data obtained - wn‘h mixtures containing 10, 15, and (45-41-14 mole %) that mehs at 640 + soc The-:' 20 mo!e % CeF3 is likewise believed to be indicates that the eutectic which melts at 725 % 10°C contains - ‘approximately 28 mole % CeF3. A compound reported to be present Clin this ‘system has not yet been |denhf|ed The - RbF-CeF, system has been studied more than the oother two systems mentioned, but thermal analysis ~data and studies of slowly cooled melts have so. ~ far failed to give a clear picture of phase relations. |t appears that at least two compounds are formed ' thot p055|bly have the- formulas 3RbF.CeF, and ]3R0 Jo Sheil, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 59. 14 . M. Bratcher, ANP Quar. Prog. Rep. March 10, 89 ANP PROJECT PROGRESS REPORT RbF.CeF,. In addition to liquidus and solidus thermal e&ects on cooling curves, solid transitions were noted at about 425 to 450°C with mixtures in the 10 to 35 mole % CeF, range and at temper- atures ronging from about 500 to 580°C with mixtures containing 33 to 60 mole % CeF,. These transitions probably account for the fact that particle sizes of crystalline phases are so small that microscopic identification is difficult and some phases even give . poor x-ray diffraction ‘patterns, The minimum liquidus temperature of approximately 615°C apparently occurs between 42.5 and 50 mole % CeF,. Study of all three systems mentioned is continuing. THE SYSTEM LIF-CsF L. M. Bratcher The recent availability of pure CsF prompted a re-examination of the LiF-CsF system which is of interest because a compound is formed such as that formed in the LiF-RbF system.1% The results of the earlier studies were not published because of the scatter in the thermal analysis data. While the existence of other: ulkali-hdlide binary compounds is well established, no mention of . alkali-fluoride binary compounds has been found in the literature, The freezing point of the CsF used in the recent investigation was found to be 700 * 5°C, which agrees with recently published values of 705°C (ref. 16) and 703°C (ref, 17). The thermal analysis data shown in Fig. 2.1.7 include some data from the earlier studies. The binary compound, believed to have the compo- sition LiF.CsF, was easily recognized under the microscope because it is birefringent, whereas: both LiF and CsF are isotropic. However, for reasons that are not well understood at present, the mixtures examined to date gave rather poor x-ray diffraction patterns for the compound. The large number of lines observed suggests that the compound is either monoclinic or rhombohedral. 15 . M. Bratcher et als, ANP Quar. Prog. Rep. June 10, 1954, ORNL-1729, p 44. 160, Schmitz-Dumont and E, Schmitz, Z. anorg. Chem. 252, 329 (1944). 174, A, Bredig, He R. Bronstein, and Wme T. Smith, Jre, J. Am. Chem. Soc. 77, 1454 (1955). SOMNPOENTA ORNL-LR-DWG 14632 900 T 800 ~ 700 . 600 TEMPERATURE (°C) 500 400 300 LiF-CsF LiF fo 20 30 40 50 60 70 80 90 - CsF CsF (mole %) Fig. 2.1.7. The System LiF-CsF. 90 THE.SYSTEM,Mng-Can L.. M. Bratcher In the course of an investigation of the system NaF-MgF -Can, thermal data were obtained with two MgF ,-CaF, mixtures which indicated that the literature value for the melting point of the eutectic mixture in this system mlght be too low.?8 There- fore six additional mixtiores were prepared which covered the composition range 37.5 to 50 mole % CaF,. Cooling curves obtained with these mixtures showed that the eutectic mixture con- tained approximately 48.5 mole % CaF . and melted at 985 + 5°C, as compared with the literature value of 940°C. THE SYSTEM RbF-Cqu L. M. Bratcher Preliminary thermal analysis data indicate that there is a eutectic between RbF and a binary compound (probably RbF-CaF ) that contains less than 10 mole % CaF, and has a melting point of 765°C. Cooling curves with mixtures containing 10 to 45 mole % CaF, did not show reproducible liquidus thermal effects, but the thermal effect at 765°C became less marked with increasing C0F2 concentration in this range. Petrographic exami- nation of samples containing 10 to 20% CaF, showed increasing amounts of a cubic compound having a refractive index of approximately 1.410. A 1:1 compound {molar ratio) has been reported!? to be present in the similar system KF-CaF,,. THE SYSTEM LlF-NIF2 L. M. Bratcher Data obtained with three mixtures in the LiF- NiF, system covering the composition range 10 to 30 mole % NiF, were given in the previous report.29 The compound found in this system was tentatively assigned the formula 3LiF. NxF ‘mainly on the basis of petrographic and x-ray diffraction studies of slowly cooled melts. The thermal ~analysis data obtained to. date are dlfhcuh to interpret, but it appears llkely that the compound melts ‘incongruently to NiF., and liquid ot o temperature -only slightly aboye.the 18€, Beck, Metallurgie 5, 504 (1 908). 19p, Silber and ‘M. Ishaque, Compt. rend. 232, 1485 (1951). 20, M. Bratcher, ANP Quar. Prog. Rep., March 10, 1956, ORNL.-2061, p 78. alkali fluoride-ZnF, PERIOD ENDING JUNE 10, 1956 minimum liquidus temperature. Well-crystallized NiF, was found in slowly cooled mixtures con- taining 30 mole % NiF, or more. The first attempt to determine the melting point of NiF2 was unsuccessful, but the melting point was evidently below 1200°C, the maximum temperature to which the material was heated. 'THE SYSTEM UF,-U0, R. J. Sheil Preliminary thermal analysis data and the results of studies of a few slowly cooled mixtures in the UF ,-UO, system were reported previously,?1 Thermal analysis data obtained during this quarter were not very satisfactory because of the previ- ously mentioned undercooling difficulty and a container problem. The use of graphite con- tainers for the UF ,-UQ, mixtures was abandoned when it was dlscoverej thdt mixtures containing 4 wt % UO, or more penetrated both ordinary (C-18) and high-density graphite when heated to a maximum temperature of 1100°C. Mixtures containing 2 wt % UO2 or less did not penetrate ordinary graphite to a noticeable extent when heated to the same temperature. Surface tension measurements of UF ,-UO, mixtures will be made in an effort to account for this interesting phenomenon. The recent thermal analysis studies were conducted in sealed nickel capsules to minimize changes in oxide content of the mixtures while the thermal analysis data were being obtained. This closed apparatus does not permit the use of seeding, which apparently will be required in order to obtain reliable liquidus values in this system, Heating curves and some cooling curves obtained with mixtures containing 4, 6, 8, and 10 wt % UO show evidence of a eutectic which melts at approximafely 910°C. Extrapolation of available liquidus temperature data for these mixtures involves considerable uncertainty, but it appears that the eutectic probably contains 9t0 11 % UO (10 3to 12. 6 mole % U02) THE SYSTEM ZnF,-ZnO L. M. Bratcher H. A. Friedman During the course of earlier investigations of systems,?2 a variation in 21R, ), Sheil, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 71. 22) , M. Bratcher, R. E. Traber, Jr., and C. J. Barton, ANP Quar, Prog. Rep. june 10, 1952, ORNL-1294, p 94. 91 ANP PROJECT PROGRESS REPORT liquidus temperatures was noted in attempts to determine the melting point of pure ZnF,. Cooling curves showed liquidus effects ranging from 885 to 940°C and a second break at or near 845°C. Since ZnO was found in all the somples that . were examined subsequent to the melting-point determinations, it seemed probable that the liquidus temperature was affected by the amount of ZnO present, Confirmation of this postulate was obtained when a cooling curve obtained with oxide-free ZnF,, prepared by NHF.HF treatment of ZnF, previously dried in an HF atmosphere, showed a thermal effect only at 940 + 5°C, Be- cause of the large discrepancies between this value for the melting point of ZnF, and those reported in the literature (872°C by Puschin and Baskow, 23 and 875 + 3°C by Haendler, Patterson, 23N, Puschin and A. Baskow, Z. anorg. Chem. 81, 359 (1913). and Bernard?4), the slowly cooled ZnF, was analyzed chemically and spectrographically. 1t ~ was also carefully examined petrographically and by x-ray diffraction, along with a highly precise determination of cell parameters. = Since no evidence was found of the presence of impurities in more than trace amounts, the melting points for ZnF, reported in the literature are believed to be ~erroneous, possibly because of oxide contami- nation, Thermal analysis data obtained with mixtures formed by adding 5, 10, 15, and 20 mole % ZnO to purified ZnF, show that the eutectic which melts at 850 t 52°C contains approximately 21.5 mole % Zn0O. Only the pure components were found in the slowly cooled mixtures, and thus it appears that solid solutions do not occur, at least at room temperature, 244, M, Haendler, W. L. Patterscn, Jr., and W. J, Bernard, J. Am. Chem. Soc. 74, 3167 {1952). PERIOD ENDING JUNE 10, 1956 2.2, CHEMICAL REACTIONS IN MOLTEN SALTS F. F. Blankenship R. F. Newton ACTIVITY OF CHROMIUM IN ,CHROMIUM;NICKEL ALLOYS M. B. Panish Further measurements were made of the electro- motive forces of electrode concentration cells in in order to.determine the activity of chromium-nickel alloys. The cells used were the same as those described previously,! and they contained as an ‘electrolyte a eutectic mixture of sodium chloride and rubidium chioride with about 0.2 to 0.5 wt % chromous chloride added. ' Activity determinations were made for alloys containing from 11.2 to 53.0 mole % chromium, It was found that there was a marked tendency for the electromotive forces of the cells to drift down- ward because of the reaction Ni + GCl,=NiCl, + C This effect was reduced markedly by packing the lower end of the cell with crushed quartz in order to prevent the transfer of nickel by convection and diffusion. For several cells in which the quartz packing was not used, the equilibrium electromotive force was approximated by extrapolating the steadily drifting electromotive force to zero time. With cells in which the alloy electrode contained over 35 mole % chromium, erratic results were ob- tained after raising and lowering the cell tempera- ture, whereas the electromotive forces obtained should be reproducible. The reasons for this be- havior have not yet been ascertgined, but it is highly probable that the diffusion rate in these elec- trodes is very low and that surface effects play a very important role. It should also be noted that ac- tivity values for the high-activity region will be - approximations because of the low electromotive force produced by cells containing these electrodes vs a pure chromium electrode. If the activity of the chromium in the alloy electrode of such a cell is 0.90, then the elécfrom_otive' force will be about - ~ 0.004.: -The nonreproducibility of the cells in this ~ region is of the same order of magnitude as ihe electromohve forces measured M. B. Panish, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 92, L. G, Overholser G. M, Watson The activities obtained for various chromium- nickel alloys at 750°C are shown in Fig. 221 along with the curve obtained by Grube and Flad? at 1100°C. The two curves are not inconsistent, in that it is quite possible that the differences are due entirely to the difference in the temperatures at which the measurements were made, UNCLASSIFIED ORNL—LR—-DWG 14633 100 | —_ l | i [ | TWO PHASE REGION | 0.75 of AT 750°C A __[—IF - , I’ v\/ > / \QQ’ / e . > Z 0.50 _ . < /" TWO PHASE REGION AT 1100°C 1o 0.25 £ THIS WORK AT 750°C (ELECTROCHEMICAL) o DATA OF GRUBE AND FLAD AT ' 1100°C (2Cr,05+3H, == 3H,0+2Cr) 0 Ni 25 50 75 cr CHROMIUM (at. %) Fige 2.2.1. Activity of Chromium in Nickel- - Chromium Alloys at 750 and 1100°C. It should be noted that the chromium activity in the region near 15% chromium is slightly below the ideal activity which would have been predicted for a perfect solid solution, This is considerably lower than the activity which might have bteen ex- pected from an inspection of the chromium-nickel phase diagram, and it justifies the assumption of near- “‘ideal’ activity for chromium in Inconel. - This assumption is usually made in discussions regarding the chemical equilibria involved in the corrosion of Inconel by molten salts, 26 Geube and M. Flad, Z. Elektrochem. 48, 377 (1942). 93 ANP PROJECT PROGRESS REPORT _REDUCTION OF UF, BY STRUCTURAL METALS Jo D. Redman Further studies were made, by use of the filtration method, of the reduction of UF, by chromium and by iron in reaction mediums that dlffered from those used previously. In the earlier studies, the re- action mediums used were NaF-ZrF , (50-50 mole %e 53-47 mole %4 59-41 mole %3), NaF-LiF-ZrF, (22-55-23 mole %),é and NaF-LiF-KF (11.5-46.5-42 mole %).7 Other alkali fluoride mixtures containing ZrF, have been employed as reaction mediums in the recent studies in order to determine the effects of various alkali fluorides on the interaction of UF, with chromium or iron, The results obtained by using KF-ZrF or LiF-ZrF , (both 52-48 mole %) as reaction medlums are presented in Tables 2.2.1 and 2.2.2. 3J. D. Redman and C. F, Weaver, ANP Quar. Prog, Rep. June 10, 1954, ORNL-1729, p 50; ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 60. 4), D. Redman and C, F. Weaver, ANP Quar. Prog. Repo Ime 10. 195" ORNL‘1896’ P 60- 5). D. Redman, ANP Quar. Prog. Rep, March 10, 1956, ORNL-2061, p 93. 6), D. Redman ond C. F. Weaver, ANP Quar. Prog Rep, Sept. 10, 1955, ORNL-1947, p 74. 75, D. Redman ond C. F, Weaver, ANP Quar, Prog. Rep, March 10, 1955, ORNL-1864, p 56. TABLE 2.2.1., DATA FOR THE REACTION OF UF WITH CHROMIUM IN MOLTEN KF.ZrF (52-48 MOLE %) AT 600 AND 800°C Data for the reaction of UF with chromium at 600 and 800°C in the reaction medium KF- ZrF (52-48 mole %) are given in Table 2.2,1. In these runs approximately 2 g of chromium was reacted with UF, (10.6 wt %, 4.0 mole %) dissolved in approx:mafely 40 g of the KF-ZrF mixture cone- tained in nickel. Similar studies were made with LiF-ZrF , (52-48 mole %) as the reaction medium at a UF, concentrahon of 12.0 wt % (3.9 mole %). these studies chromium from two different sources was employed; the results are given in Table 2.2,2. - As seen from the data in Table 2,2.2 the equi- librium chromium concentrations cre the -same at both temperatures for the two different chromium metals used — hydrogen-fired electrolytic chromium TABLE 2.2.2. DATA FOR THE REACTION OF UF < WITH CHROMIUM IN MOLTEN LIF-ZrF, (5248 MOLE %) AT 600 AND 800°C Conditions of Present in Filtrate Equilibration Temperature Time Total U Total Cr* Total Ni (°C) () (wt%) (ppm) {ppm) 600 3 8.0 1030 35 3 9.4 1060 55 5 9.7 1180 55 5 8.5 1060 235 800 3 8.2 1250 35 3 8.4 1190 25 5 8.1 1150 295 5 8.1 1150 285 16 8.2 1060 30 16 8.2 1140 5 *Blank of 290 ppm of Cr at 800°C. 94 Conditions of Present in Filfrote: Equilibration Temperature Time TotalU Total Gr* Total Ni (°C) (hr) (wt %) (ppm) (ppm) 600 3 9.9 2550 30 3 9.8 3000 75 S 10,0 2980 70 5 9.4 3060 45 5 9.5 2800** 50 5 9.5 2790** 50 12 9.6 2910 30 12 9.6 3060 35 800 3 10.1 3820 55 3 9.2 3830 30 5 B9 4070 35 5 8.6 4040 4; 5 9.2 3810** - 40 5 9.3 3780** 60 12 9.5 -~ 3780 60 | 12 9.5 3830 50 *Blank of 250 ppm of Cr at 800°C. Electfolyfl'c-_chro- mium hydrogen-fired at 1200°C was used in all runs ex- cept those noted. **Blank of 190 ppm of Cr at 800°C. Very pure iodide chromium, not hydrogen-fired, was used in these runs. and . iodide chromium. The iodide chromium was obtdined from Battelle Memorial Institute and con- tained 10 ppm or less of oxygen, This pure chro- mium was not hydrogen-fired, and therefore a com- parison of the blanks obtained for this metal with those found for the electrolytic chromium (used in all - previous studies), which was hydrogen-fired under the usual conditions, should demonstrate the effectiveness - of the hydrogen freatment, The chromium values of 190 and 250 ppm obtained for the unfired iodide chromium and the hydrogen-fired electrolytic chromium, respectively, suggest that the hydrogen-firing is successful (unfired electro- lytic chromium gave a blank of 900 ppm). Evidently the major portion of the blank arises from oxidizing materials (H,0 and HF) present in the LIF-ZI’F4 mixture, The equtllbrlum chromium concentrations and the equilibrium constants calculated from mole frac- tions for the reaction of UF , with chromium in the various solvents are presented in Table 2,2,3 for comparison, The effect of varying the NaF-to-ZrF, ratio in the various NaF-ZrF, mixtures on the chromium concentration has been shown and dis- cussed previously. The values are included in Table 2.2.3 for comparison with the KF-ZrF, and LiF-ZrF, data. It is evident from the volues PERIOD ENDING JUNE 10, 1956 given .in Table 2.2.3 that the particular alkali fluoride used in combination with ZrF , influences the reaction markedly, The different alkali fluorides affect the activity of the UF ; to varying degrees. The activity of the CrF, is also influenced through complexing of the CrF, by both the alkali flucride and the ZrF . Studles will be made shortly with an I‘?I:\F—ZrF4 mixture as reaction medium to com- plete the alkali fluoride series, Studies of the reduction of UF, by iron at 600 and 800°C with LIF-ZrF and KF-ZrF (both 52-48 mole %) as reaction medlums were made at UF concentrations of 12,0 wt % (3.9 mole %) for the Li F-Zrl':i mixture and 10,6 wt % (4.0 mole %) for the KF-ZrF, mixture, The data are presented in Table 2, 2.4 As may be seen from the data the equilibrium iron concentrations are not significantly changed by replacing LiF with KF., |t may be noted that somewhat smaller values result at 800°C than at 600°C for both solvents. These values also are in good agreement with those found when NaF-ZrF, (50-50 mole %), NcF-ZrF (53-47 mole %), and NaF-ZrF (59-41 mole %) were used as the reaction medlums. The iron values obtained by using NaF-LiF-ZrF, (22.55-23 mole %) as the solvent fell in the same range, but in this case the values were slightly larger at 800°C than at 600°C, TABLE 2.2,3, EQUILIBRIUM CONCENTRATIONS AND CONSTANTS FOR THE REACTION | Cr° + 2UF ;=2 2UF + CrF; IN VARIOUS SOLVENTS . _ Temperature - UF, Cr Solvent - {°0) (mole %) ~ {ppm) K.* LiF-ZrF Lo 00 40 2900 7 x 10~4 (52-48 mole %) - 800 40 3900 S 7x 1073 NaF-ZrF, 600 4.1 2250 4x 1074 (50-50 mole o 800 41 2550 5 x 10~4 “NaF-ZrF, 00 4.0 Cowoe o ax et (53-47 mole %) B 800 4.0 2100 - 3x 10=4 © NeF-ZeF, 600 a7 975 lLax10-S (59-41 mole %) - 800 37 10507 T 16 x 1073 KF-ZeF, - 600 3.9 080 0 2.4x 1075 ) (52-43 mole %) 800 3.9 o M0 - 32x 1073 NaF-LiF-ZeF, | 600 25 s 1x 108 (225523 mole %) 800 25 75 4x 10~5 " NaF-LiF-KF _ 600 2.5 1100 -(Il.5-46.5-42 mo|e % 800 2.5 2700 *Kx = XEIF3 XCer/XCr XfiF“, where X is concentration in mole fractions. 95 ANP PROJECT PROGRESS REPORT TABLE 2.2,4. DATA FOR THE REACTION OF UF, WITH IRON IN LIF-ZsF (52-48 MOLE %) AND IN KF-ZrF , (52-48 MOLE %) AT 600 AND 800°C Conditions of . . P . . Equilibration resent in Filtrate Temperature Time Total U Total Fe* Total Ni CO - () wt%) (pem) (ppm) Solvent: LiF-ZrF4 {52-48 mole %) _50(_)',_ . 3 9.1 450 ) 3100 550 35 5 97 460 o 5 9.9 520 65 800 3 9.8 360 15 3 9.6 9 2 5 9.6 400 35 5 9.5 350 35 Solvent: KF-ZrF4 (52-48 mole %) 600 3 7.2 520 1 3 8.0 630 10 5 7.7 490 5 5 7.9 510 10 800 3 8.2 330 65 3 8.1 280 30 5 8.1 310 70 5 8.1 330 35 *Blanks of 100 and 40 ppm of Fe at 800°C for LiF-Zr F4 and KF-ZrF ., respectively, The indifference exhibited by the UF ,-Fe® reaction to changes in the reaction medium is in marked contrast to the behavior noted for the UF ;-Cr re- action. No satisfactory explanation can be offered at this time for these differences. SOME OBSERVATIONS ON MASS TRANSFER OF CHROMIUM BY MOLTEN SALTS -W. R, Grimes When the corrosion of Inconel by NaF-ZrF -UF, mixtures and by NaF-KF-LiF-UF, mixtures is compared, some striking dlfferences are observed. The corrosion, as measured by depth of void forma- 96 tion in the hot region of a system with a temperature gradient, is much worse when the circulated fluid is the alkali fluoride mixture than when it is the ZrF ~bearing fuel. In addition, discrete crystals of nearly pure chromium are found in the cold’ zones of Inconel loops in which ‘the NaF-KF-LiF- UF , mixture has been circulated. Metallic deposits are not usually observed in loops which have cir- culated the NaF-ZrF -UF4 preparations, Depth of void formation does increase with time, however, for both classes of systems, even though the con- centration of chromium compounds in solution - reaches an equilibrium concentration early in the corrosion test and remains essentially constant thereafter. The amount of chromium removed from the loop walls, as estimated from the volume of voids in the hot zone, is considerably larger than can be explained by the amount of chromium com- pound in solution in the melt ofter the test. More- over, the corrosion seems to be nearly independent of the ratio of the.surface area of the loop to the volume of the melt and not strongly dependent on the concentration of uranium in the melt. The processes that occur are undoubtedly very complex, and it is not possible fo explain in quantitative fashion all the effects observed, However, by the use of equilibrium data obtained for the chemical reactions and by making some reasonable assumptions it is possible to rationalize much of the available data and to indicate a pos- sible explanation for the observed similarities and differences in these two fuel classes. For ZrF - bearing mixtures, of which the mixture with 50 mole % NaF, 46 mole % ZrF,, ond 4 mole % UF is typical, the important corrosion reaction is known to be NuF-ZfF4 S ) : 2UF + Cr fi CI‘F2 + 2UF3 The equilibrium constant for this reaction is given by (CrF,) (UF5)? K(a) = r (UF ,)2(Cr) where the quantities in parentheses represent the activities of the materials involved. The equi- librium constant, K('a), cannot be determined by experiment, but the equilibrivm concentration, K('c), which is'defined by the equation 2 o Crerrp Clury) Kiey = ' 2 . “lur o Cien ‘where C U )Eepresents the ethllbnum concentration of the designated component in mole fraction, can be evaluated for pure chromium and has been shown to be constant at a given temperature, For cases in which the CrF, and UF; present are formed only by the reaction described above, C = 2C (UF,) (CeFy) and 3 4C (Cer) Kiey = —0———— - ' C(UF )C(Cr) ‘lfrthe amount of UF, added inifiully is dlways the same and if the fractlon of this UF4 reduced to UI"'3 is small, then Cluey =47 L 4C(CtF2)' K/ & —— (c) ’ ’ € A C(c” , , o\ l_/.'.’; i, o c Kier Cien] 7~ (C'Fz) =- S Co- . i and ) Accordmgly, the CrF2 concenh'ahon ct equuhbrlum,_ if all other factors are constant, depends on the ~.cube “root of the mole fraction. of chromlum in the ~ alloy. Therefore it should be possiblie to calculate,f_j i ~ from experlmental data for pure chromium, the equi~ hbnum concentrations to be expected- from Inconel, - ~ The corrosion reaction for the NaF-KF-LiF-UF ;mlxtures is considerably - more complex. In \‘hls‘ ; - case two consecutlve reactzons uppurently proceed NcFoK F-LIF S , 2UF + Cr CrF + 2UF ’ "3crl='2*'-——--—a' NP KL 2o, +Cr As a result of these reactions abouf 80% of fhe PERIOD ENDING JUNE 10, 1956 chromium compound in solution is trivalent, . The net reaction is {approximately) 14UF, + 5 KL 4oE 4 GF, + 4UF, and , - | o (chz)(chs)HUFz)_H- K(a) = (UF §14(Cr)5 Cieer ) C?cn-'_s) C(IJF:,) K(c) = — . 14 5 S Clen If CrF,, CrF,, and UF, all arise solely from this reaction and if C(UF4) is constant, then C(Cst) = 4C(CtF2) ' Cc ' 4ch 14 CieeF ) 4C1cer ) MC(Cer ) K, = (c} ’ + ~5 A' C(cr) »r ,L 1/19 c (C)A C(Cr) C F = LAt ettt e, (CrF2) T 4w 144 Accordingly, with all other variables held constant, ~~the concentration of CrF, + CrF, is a function of -~ the 5/19 power of the mole fraction of chromium in the alloy, Therefore, from equnllbrwm dota for pure chromlum, correspondmg concentrcmons ccm' be culculoted for lnconel & - ' ' 81y should be emphasized that accuracy of the come - lpllcated equation - s not an essential port of the crgumenf. “The same qualitative conclusions are uppurenf Whether ‘one. uses' ~as the corrosion reaction . oY . er S 2UF +Cr—'"2UF3+CrF2 - T 2 UF, + Cremt BUF, + cu=3 . In feacflon 1, as shown for the ZrF -bearmg ffiel the i Csz concentration varies as (Cr)]/3, and in reaction 2, - the CrF concentration varies as (Cr)l 4, 97 ANP PROJE_CT PROGRESS REPORT Inconel contains 15 wt % Cr or a chromium mole fraction of about 0.16. The assumption that the activity of chromium in Inconel is roughly equal to its mole fraction appears to be rather good. The compounds in the melts, which were treated with pure chromium at 600 and 800°C, are shown in Table 2.2.5, along with corresponding valuves cal- culated for equilibration with Inconel. Experimental values obtained from corrosion experiments with Inconel and these melts are usually slightly lower than these calculated values, presumably because in the experiments equilibrium is dttained with an Inconel surface layer significantly depleted in chromium. - When Inconel powder is used in experi- mental equilibration apparatus, side reactions that involve oxides of the metals complicate the situa- tion; equilibrium concentrations of chromivm come pounds higher than those calculated are often - observed. In general, it appears that the calcu- lations are good only for the idealized case de- scribed. From the data in Table 2.2.5 it is obvious that Inconel exposed to the NaF-KF-LiF-UF , melt will support o higher concentration of CrF 5-CrFgy equilibrium ot 800°C than pure Cr° IS oble fo support at 600°C, Accodingly, chromium is re- moved from lnconel in the hot zone of o loop and deposited as essentially pure chromium in the cold zone. Since no diffusion process is necessary in the cold zone (the chromium can deposit at the sur- face of the Cr° crystais), the rate of attack is con- trolled simply by the rate of diffusion of chromium to the metalesalt interface in the hot zone. The data in Table 2.2,5 also reveal, however, that Inconel exposed to NaF-ZrF ‘-UF « mixtures is in equilibrium with much lower CrF_ concentrations than pure Cr° is in equuhbrwmwfl% when exposed to the fluoride mixture under the same conditions. Accordingly, it is not possible for chromium to dissclve from 800°C Inconel ond to deposit at 600°C as Cr° when NaF-ZrF +UF , mixtures of this general composition are c:rculated Loops of pure Cr® would, of course, mass-trunsfer in this medium; moreover, mass transfer can occur if a sufficiently dilute alloy of chromium can be formed in the cold zone. This suggests that the mass-transfer process takes place in the followmg general way. The molten salt in the hot zone reaches equi- librium with the 800°C Inconel and passes with the dissolved CrF, to the cold zone. In the cold zone, equilibrium is established by deposition of a small amount of the chromium (by reversal of the reaction) in the surface layer of the metal to form an alloy containing slightly more chromium than is normally present in Inconel. If no diffusion of chromium were possible, c true equilibrium would be reached shortly, with the hot surfaces slightly depleted and the colder surfaces slightly en- riched in chromium, The process would then stop (except for the exchange process which has no net effect), with negligible corrosion of the metal. Since diffusion does toke place, however, the process continues. In the hot zcne the con- centration gradient causes a flow of chromium to TABLE 2.2.5. EQUILIBRIUM CONCENTRATIONS OF CHROMIUM FLUORIDES WITH ALKAL! FLUORIDE AND ZrF -BEARING FUEL MIXTURES Chromium Concentration {ppm) In NaF-KF-LiF-UF, In NaF-ZrF -UF , Experimental resuits for melt treated with pute chromjum, (Cr) = 1,0* At 600°C At 800°C Resulits calculated for equilibration of melt with Inconel, (Cr) = 0.16* At 600°C At 800°C 1100 2400 2600 2550 710 1320 1660 | 1400 tConcentration of chromium in mole fraction. 98 the salt-metal interface, and the octivity of chro- mium ot the surfoce is maintained at some level appreciably less than 0.16. In the colder zones the diffusion gradient is away from the salt-metal interface, and the chromium activity is maintained at some activity slightly higher than 0,16, Since diffusion at the lower temperature is slower thon that at the higher temperature, the rate-controlling step is presumably the low-temperature diffusion process. It may be that the intermediatestemperature regions, in which the chemical driving force is less but the diffusion rate is faster, are the most im- portant regions; there is no a priori way of telling. If the dynamicecorrosion and mass-transfer phe- nomena cre examined in this light, it is obvious why neither flow rate nor uranium concenfration of the fuel mixture is an important factor affecting corro- sion attack and mass trensfer in Inconel systems., If it is assumed that most of the corrosion observed invelves mass fransfer to a dilute alloy, the lack of an effect of the surface-to-volume ratio becomes comprehensible. If the relative effect of diffusion rate vs driving force is noted, the reason for the poor correlation of corrosion with temperature drop can be understood; with a ZrF ,-bearing fuel, corro- sion may be worse with a top temperature of 1500°F and a 200°F drop than with a top temperature of 1500°F and o 400°F drop. - The argument points up the fact that, if the low temperature were made sufficiently low, deposition of Cr® might be possible in ZrF (bearing systems; the temperature coefficient of the reaction is not known below 600°C. It is possible that the light frosting of chromium often observed in the cold _dead-leg sump of oId-ster thermul-convechon loops e may have been due to such-an effect, It would also - - be ‘possible to -have alloys’ suffaclently high in o t_'chromwm concentrohon (ond octivny) to cause j?,chromwm deposmon to occur, These considera~ - -tions’ point to. the conscderoblo supericrity of alloys OVer pre mefaIs fO'I' conIdIrlment Of moIIen saIfs.r':!-_f- fa"-.ed by f.ltrohon mefl-.ods. The change |n solu. " bility at 600°C with total quantity of NiF, added ~ may be ascribed to the saturating phase being a - complex compound rather than Nan. . The complex EQUILIBRIUM REDUCTION OF NIF BY H2 IN NcF-ZrF : C M. BIood | Inveshgohon of fhe equshbnum S NIF @ + H2 (g)‘-"-——‘Nl (s) + 2HF (g) 'when Ihe mixture NoF-ZrF (53-47 mole %) is used- -' as the solvent was mmoted. The equipment and experimental fechnlques for this study were very similar to those described previously for the study mixture NaF- ZrF (53-47 mole %) PERIOD ENDING JUNE 10, 1956 of the reduction of FeF, by H, in NaF-ZrF . 911 Equipment changes mcfiuded the subsmunon of nickel as the container material in place of the mild steel used in the previous study. Operational changes consisted in the use of hydregen-helium mixtures of known composition instead of pure hydrogen in the preparation of the equilibrating gaseous streams. This was found to be necessary after the results of preliminary experiments indi- cated the necessity for using small partial pressures of hydrogen and high partial pressures of HF in order to maintain measurable quantities of NiF, in solution. The initial experiments were made in order to ascertain the solubility of NiF, in the solvent The values obtained by fnltrohon of the saturated solution and chemical analysis of the filtrate when o total of 0.17 wt % Ni was added as NIF are summonzed in the following tabulation: Temperature Solubllity of NII-'2 (°C) (ppm) 550 - 400 575 600 -~ 600 980 625 _ 1450 | Solubility values were also obtained by filtration by Redman (see following section on *‘Sclubility _ and Stability of Structural Metal Fluorides in Molten NaF-ZrF "), who showed that the solubility is a function of the quantity of NiF, added. Topol 12 ‘also obtained solubility vulues by measuring the “electromotive forces . of electrolytic cells. The “values for the squblhty at 600°C that were ob- _tained in the different investigations are shown in - Fig. 2.2.2, There appears to be satisfactory corre- " lation of these solubility values with those ob- - 9C, M. Blood and G. M. Watson, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 66. . ‘°C. M. B!ood. ANP Qudf. Progo Repo Dec. 10' 19”0 ORNL-2012, p 85. Ye, m. Blood and G. M. Watson, ANP Quar. Prog. 12 E, Topol ANP Quar. Prog. Rep. March 10, 1956. ORNL-2061, p 89. 99 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG-44634 4000 / 3000 7 CONCENTRATION OF NiF, IN SOLUTION (ppm Ni) 2000 e REDMAN = TOPOL A BLOOD 1000 0 0 2 4 6 8 10 TOTAL NiF, ADDED {wt % Ni) - Fig. 2.2.2. Comparison of Values of Seolubility of NiF, in NoF-ZrF, (53-47 mole %) at 600°C Obtained By Three Investigators. compound ‘is probably, though not necessarily, idenfical with the substance NiF .ZrF, previously synthesized by Sturm. 13 Thermodynamic calculations performed prior to the experimental work indicated that if NiF, ex- hibited activity coefficients similar in magnitude to those previously determined for FeF,, equili- bration at 600°C with H,-HF mixtures containing about 3% HF would allow sufficient NiF, to remain in solution for satisfactory analysis. ly traces of NiF, (<10 ppm) were found after repeated and prolonged equilibration under these conditions and even after the reactor temperature had been re- duced to 550°C and the H,-HF equilibrating mix- ture had been changed to contain 30% HF. Attempts to increase the NiF, concentration in solution by simply raising the HzF concentration of the equili- brating gas were unsuccessful beyond NiF, con- centrations of about 20 ppm Ni. Nickel fluoride concentrations of sufficient mag- nitude for an acceptable degree of precision in the analytical determinations were finally obtained by lowering the partial pressure of hydrogen, in.addi- tion to maintaining a high HF concentration in the equilibrating gas mixtures. This was accomplished by the use of hydrogen-helium mixtures rather than pure hydrogen. The mixtures were prepared by injecting helium gas into a partially empty hydrogen 13g, J. Sturm, ANP Quar, Prog. Rep. Dec. 10, 1955, ORNL-2012, p 91. 100 cylinder. The compositions of the mixtures thus prepared were determined by the proper adaptation of the hydrogen combustion analysis technique to the existing experimental assembly., The experi- mental results are summerized in Table 2,2.6. As is customary, equilibrium was approached from both directions in obtaining the series of measurements described. It is of interest to note that the calculated values of the equilibrium con- stants for the reaction range from 11 to 45 atm at 625°C when the solid and the supercooled liquid, respectively, are used as standard states. The experimental values denote activity coefficients of NiF, a thousand times greater than those ob- tained for FeF,. Detailed numerical calculations TABLE 2,2,6. EQUILIBRIUM CONSTANTS FOR THE REACTION NiF,(d) + H,(g)==Ni(s) + 2HF(g) AT 625°C Nickel Content Hydrogen HF - of Melt Pressure Presswe Kx*x 10-4 (ppm) (atm) (atm) 77 0.0628 0.502 3.06 75 0.0661 0.475 2.68 75 0.0640 0.492 = 296 75 0.0595 0.528 3.67 75 0.0617 0.510 3,30 85 0.0628 0.501 2.76 75 0.0640 0.493 2.97 75 0.0643 0.490 2.92 185 0.0251 0.456 2.74 165 0.0253 0.460 2.98 160 0.0255 0.457 3.00 150 0.0259 0.448 3.03 140 0.0250 0.447 3.22 120 0.0270 0.425 . 3.28 140 . 0.,0270 0.425 2.81 125 0.0269 0.427 3.19 135 0.0274 0.416 2.74 125 0.0276 0.412 2.89 Av 3.01 10.18 *K, = PElF/xNIszfiz‘ where X is mole fractior? and P is pressure in atmospheres. of this reaction will be made ot the conclusion of the experimental work. Experimental work is now in progress at 575°C, Efforts will be made to extend the NiF, concen- tration ranges and to obtain the temperature de- pendence of the equilibrium. Particular attention will be given to the determination of the activity of NiF, in the saturating phase. SOLUBILITY AND STABILITY OF STRUCTURAL METAL FLUORIDES IN MOL TEN NuF-ZrF4 | J. D. Redman The results of studies of the stability and solu- bility of CrF, in NaF-ZrF, (53-47 mole %) at 600°C were presented previously.'®# From these experiments it was concluded that the solubility of CrF, increased as the zirconium-to-chromium ratio decreased, and, accordingly, the solubility of CrF, in this solvent was a function of the amount of CrF, added, It was postulated that the change in so|u2|:ili1y was due to the separation of CrF,.ZeF, as the solid phase, with a resultant increase in the NaF concentration in the melt, Some of these experiments have been rerun in order to obtain more precise values at the larger zirconium- to-chromium ratios, and the values thus obtained, along with some of the earlier ones, are given in Table 2.2.7. . 145, D. Redman, ANP Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012, p 88. "PERIOD ENDING JUNE 10, 1956 Similar studies have been carried out on the solubility of FeF, and NiF, in this solvent, and the data are presented in Tables 2,2,8 and 2.2.9, respectively, The data for these three structural metal fluorides also are plotted on Fig, 2,2.3. An examination of the data presented for 600°C shows that the solubility of CrF, increases quite rapidly as the zirconiumeto-chromium ratio de- creases over the entire range studied. A less pronounced effect is noted for FeF,, and in the case of NiF, @ slight increase is observed for the higher zirconium-to-nickel ratios. At the lower zirconium-to-nickel ratios the solubility is con- stant, These differences in behavior are shown quite clearly in Fig. 2,2.3. It appears that the solid phases present for the GF, and FeF, sys- tems contain ZrF ., probably as MF,.ZrF ,, whereas the solid phase in the NiF, system appears to tie up only very small amounts of ZtF,, It may be noted that at 800°C the solubility of NiF, is inde- pendent of the zirconium-to-nickel ratio, which suggests that the solid phase present is NiF,, There is some question regarding the reliability of the data given in the zirconiumsto-sodium ratio columns in the tables, but for Cer and for FeF, at 600°C it is believed that the decrease in the ratio as the excess of metal fluoride added was _increased is real. The ratio of zirconium to sodium is 0.88 in the starting material, and the consider- ably lower values found for the large additions of TABLE 2.2.7.. SOLUBILITY AND STABILITY OF CeF, IN MOLTEN NaF-ZcF , (53-47 MOLE %) AT 600°C Conditions of . ‘Found in Filtrate - Equilibration _ Ce*t ZrtoGr It ‘Na Zr-to-Na F ot Total Cr (wt %) Rctfip* (wt %) (wt %) Ratio* (wt %) {wt %) Awt %) 1.4 w7 - 415 13,3 0.79 - 45.3 0.66 0.73 1.4 7 2.3 3.4 0.78 44.8 0.65 0.83 40 58 8 127 0.79 45,1 2.1 e 4.0 58 9.5 129 0.78 452 . 20 1.9 64 34 38.4 6 07 8 3.5 3.5 64 34 389 141 e 443 34 36 9.6 21 7.9 . 13.8 069 . 440 .50 52 9.6 2.1 36.3 13.4 0.68 44.2 5.7 5.9 *Ratio calculated from mole fractions. 101 ANP PROJECT PROGRESS REPORT _TABLE 2.2.8. SOLUBILITY AND STABILITY OF FeF, IN MOLTEN NaF-Z¢F , (5347 MOLE %) AT 600 AND 800°C : - Conditions of Equilibration Found in Filtrate Temperature Fe*t Zrto-Fe Zr Na. . Zreto-Na F Fe*t Total Fe {°0. (wt %) Ratio* (wt %) (wt %) Ratio* (wt %) (wt %) (wt %) 600° 0.5 50 415 13.7 0.78 45.7 0.26 0.26 | 0.5 0 419 12.9 0.82 45.3 0.25 0,26 1.0 25 41.6 13.8 0.78 452 . 0.33 . 0.32 1.0 25 4.6 13.6 0.78 45.0 0.35 0.35 5.0 4.8 40.3 1.5 0.76 44.3 1.3 1.4 5.0 4.8 39.8 13.8 0.73 44.4 1.4 1.5 10,0 2.1 37.9 14.0 0.69 45.0 2.9 30 100 2.1 37.5 3.7 0.70 446 3.5 3.5 800 - 6.0 3.9 38.9 10.8 0.91 45.2 6.1 6.0 | 60 3.9 38.5 10.6 0.91 441 4.8 5.8 120 17 34.7 0.0 - 0.8 87 07 1.7 12.0 17 34.5 9.4 0.93 47 N3 1.9 18.0 1.0 32.2 8.8 0.93 41 4 14.2 18.0 1.0 32.1 9.2 0.88 440 128 14.0 *Ratio calculated from mole fractions. TABLE 2.2,9. SOLUBILITY OF NiF, IN MOLTEN NaF-ZrF , (53-47 MOLE %) AT 600 AND 800°C Conditions of Equilibrdfion Found in Filtrate Temperature Ni Zr-to-Ni Zr No Zr-to-Na F Ni (°C) (wt %) Ratio* (wt %) (wt %) Ratio* (wt %) (wt B 600 0.5 43 4.2 13.0 0.80 45.2 0.14 | 0.5 8 418 13.6 0.78 451 0.12 1.5 18 41.4 12.8 0.82 45.5 0.17 1.5 18 41.4 12.5 0.84 45.0 0.19 5.0 5.2 40,4 14.1 0.73 45.3 0.32 5.0 5.2 39.8 14.0 0.72 44,7 0.31 10 2.4 40,0 14.0 0.73 451 0.32 10 2.4 40.8 12.8 0.81 45.3 0.31 800 1.5 18 41.6 12.0 0.88 45.2 1.0 1.5 18 41,5 12.1 0.87 45.3 - 1.1 5.0 5.2 41,6 12.3 0.87 45.6 1.2 5.0 5.2 41,8 11.8 0.89 45.4 1.3 10 2.4 417 12.6 0.84 44,6 1.0 10 2.4 0.3 12.2 0.87 447 13 *Ratio calculated from mole fractions. 102 C PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED ORNL-—-LR—-DWG 14635 6 5 pd Oé, / o Y i 4 7 : . s / e . . 4 > / " 5 . 0090/ 3 / gefz) Q 2 ] m / ~ / A/ _ NiF,; BOO°C OF‘// i A | NiFy; 600°C o et ] 1 5 6 7 8 9 10 0 { 2 3 : 4 METAL FLUORIDE ADDED (wt % M) Fig. 2.2.3. Eflect of Amount of Metal Flucride Added on Solubility of NiF,, FeF,, and Cer NeF-ZrF (53-47 mole %) at 600°C. metal fluoride are odditional evidence for the belief that ZrF is combined with the metal fluo- ride in the solrd phase, mole %) were centinued through the use of the con- cenfration cells discussed previously,15 Cells of the type MIMF 2% ZrF (cy), N_aF-ZrF4| NaF-ZF,, MF xZrF (c )M A comparlson of va Jlues given for Fe""" and total iron shows that Fett tions, The slight devnahons noted are attrrbutoble to expenmentcl errors. The ' ogreement between it is evident that more than 90% of the chromium s present as’ Cr'“' Since the starting materaal does not contain over 95% of the chromium as Crtt, it - - s evrdent that no opprecrable dlsproportlonotron of ".Cer occurs under the condltlons employed SOLUBILITY DETERMNAT'ONS BY MEASURE"’ to ‘measure the temperature of either of the half cells - durectly. T ‘ : . thermocouple placed in g well that was permanently o MENT or= ELECTROMOTIVE FORCES OF CONCENTRATION CELLS ' L. E. Topol Solubility studies of structural metal fluorides and their complexes in molten NaF-ZrF, (53-47 . celis. . as a functlon of temperature, the saturation points * where M was Fe; Nl, o‘r Cr,and x = 0 or 1, were .is stable under these condi= measured at 525 to 750°C in a helium atmosphere, ~ The "containers ‘were alumma, platinum, or the - metal M when M was iron and-nickel; a Zr0, bridge - Cr** ond total chromtum is not so- satrsfactory, but lmpregnoted with NaF-ZrF (53-47 mole %) served as the: electrlcal contact between the two half ‘When the eiectromotwe force was measured were apparent from the points of dlscontmuny in : | the P|°f of yoltage vs temperature. - ln the cell ussembly ernployed it was not feaslble The temperature recorded by a 151, E. Topol, ANP Quar. Prog. Rep, March 10, 1956, ORNL-2061, p 89. 103 ANP PROJECT PROGRESS REPORT fixed midway between the two half cells was - formerly used as the cell temperature. In order to refine the temperature measurements, a calibrated thermocouple, enclosed in a nickel tube, was sub- stituted for an electrode in a dummy cell. In the temperature range of interest (550 to 750°C) the melt temperatures found were 10 + 2°C higher than the melt temperatures at the center between the two crucibles, Thus all the temperatures reported in connection with previous potential measurements -should be revised upward by 10°C, Only the solu- bility data are sufficiently temperature sensitive to be appreciably affected. Application of the temperature correction to the solubility data previously reported!5 gave the following new expressions for the solubility of MF 5. Z¢F , in NaF-ZrF , (53-47 mole %): n3 log N = = 00X 197 (67, for FeF o2eF,, 40.1 x 10° |og N = —m—* + 8.36 ’ fOl‘ Cer'ZI'F4, 28.8 x 103 log N = 'Ts;is_r + 472, for NiF 2 ZcF 16 where N is the mole fruchon of MF -ZrF and T is temperafure in °K. The corrected hects of solution and *‘ideal” melting points are now 36.0, 40.1, and 28.8kcal/mole MF ,«ZrF, and 875, 775, ond 1060°C for the Fe, Cr, ond Ni compounds, respectively. Solubility data obtained by analyses of filtrates from saturated solutions, as described in the pre- ceding sections of this chapter (‘‘Equilibrium Reduction of NiF, by H, in NaF-ZrF '’ and *‘Solu- bility and Stublhty of Structural Metol Fluorides in Molten NaF-ZrF '), did not agree with the elec- trometrically determined solubilities, particularly in the case of NiF,, Since the filtration measure- ments were carried out by adding NiF, rather than NiF,ZrF, as solute (and similarly with FeF, and CrF 2)s the electrometric determinations were re- peated with Nth as solute so that the data could ' léFréh this fnvestigation and xeray and petrographic examinations of quenches of NaF-ZrF, melts, NiF, is now believgd to exist in a ternary compound, rather than as NiFyeZrF,, in saturated solutions. The equation for the solubility was revised to include additional meus'uremeriis, 'qs well as the temperature correction. 104 be compared. X-ray and petrographic examinations of the solidified melt from the half cells should have sufficed to identify either or both of the joins as quasi-binary; in practice, however, both NiF.‘,-ZrF4 and a new phase were found in the slowly cooled half cells. The new phase is thought to be a ternary compound, since it has never been found in the binary systems. Gradient quenches (510 to 540°C) along the two joins revealed only the new phase, with no NiF,ZrF,. Such ambiguous results were dis- appointing, since on a quasi-binary join the solu- bility does not vary with the amount of solute added, as it does along a random join. Cells containing MF, in NaoF-ZrF, (53-47 mole %), with no addition of ZrF,, gave the solubilities of the structural metal solts shown in Table 2,2,10 and Fig. 2.2.4 for NiF, and Table 2.2.11 for FeF,,. The solubilities of the nickel salts along the joins NiF - ZrF -7NaF+6ZrF, and NiF,-7NaF+6ZrF,, as we!l as fhe values for NiF, obtomed by filtration methods, are plotted in Fig. 2 2.4. The discrepancy between the results of the various experiments has not been explained. The electrometrically determined solubilities appear to be virtually the same whether the solute is added as NiF2 or as NiF,.ZrF,, and the same seems to be true for TABLE 2,2,10. SOLUBILITY OF NiF, IN NaF-ZrF , (53-47 MOLE %) AH’;M = 28.8 kcal Ideal melting point = 1060°C Temperature | NiF, Solubility (°C) wt % mole % . 553 - 0.12 0.12 564 0.14 0.15 574 " 0.9 0.20 576 0.20 o 0.21 650 0.79 | 0.83 650 0.93 0.97 655 0.88 0.92 669 0.98 1.0 706 1.97 2,05 725 21 2.29 - SOLUBIITY (mole fraction} % FeF,. The FeF., saturation points were not so 2 2 well defined or reproducible, however, as those for FeF,.ZrF,, especially ot temperatures above 650°C. The potentials of many of the cells flux- tuated greatly and were much lower than the values predicted from the Nernst equation (where complete UNCLASSIFIED 0.0 ORNL-LR-DWG 14636 EMF DATA: . NF 0.05 : © NiFg-2rf4 FILTRATION DATA FOR NiF2: A VARYING TEMPERATURE (BLOOD) A° VARYING AMOUNT OF SOLUTE (REDMAN) 0.02 FROM EMF DATA Q.04 0.005 FROM FiLTRATION DATA ™ 0.004 0.0005 800 750 700 650 600 550 . 500 TEMPERATURE (*C) Fig. .2.2.4. = Solubility. of NiF, and NiF,.ZrF in NaF-ZrF, (53-47 mole %). TABLE 2.2,11. TENTATIVE VALUES FOR THE SOLUBILITY OF Fer IN NaFZrF , (53-47 MOLE %) Temperature FeFy Solubllity : °c oWt % " mole % ' 53 013 o4 561 025 027 %5 . 031 034 8 . 072 . o018 e . 18 . 116 60 w0 s 66 244 263 . 665 . o 323 34y 70 122 775 742 9.9 10.59 PERIOD ENDING JUNE 10, 1956 solution in both half cells was expected). X-ray and petrographic examinations revealed a better- defined temary complex in low-temperature (535°C) quenches of FeF, melts than in high-temperature (672°C) quenches; the FeF, ternary complex gives the same x-ray pattern as that given by the NIF ternary complex, FREE ENERGIES OF FORMATION OF COMPLEX "METAL FLUORIDES MF ,+Z+F , ' L. E. Topol For a saturated solution of MFZ.ZrF4 in a molten electrolyte, the following equilibrium holds: MFZ.ZrF4(crys!)=MF2(so|n) + ZrF y(gotny + AF = 0. In NaF-ZrF, (53-47 mole %) as the solvent, the activities of MFz("ln) and ZrF“uln) are known, - at least approximately, in a few cases, and, since the activity of solid MF,.ZrF, is unity, the free energy of formation of M‘F2-ZrF4 is the only un- known in the following relations as applied to the saturation equilibrium: aManZrF4 = 0 = AFO + RT |n a MFZ'ZUF4 = AF%_ 4 AFS - AF§ MF2 ZrF4 MFz'ZrF,4 -where N is mole fraction and a and y are, respec- - tively, the activity ond the activity coefficient " based on the pure solid as the standard state. The " free ‘energy of complexing of MF -ZrF from solid o 'MF and solid Zrl":4 is o omp _ . : A; o ° . '.AF AFMFQ-ZrF" - AFfiE’? -AF,Z'F‘ - 2.‘3 RT |09 yME NMF#“Z:F ‘ o Hence a knowledge of lhe acflvn‘y of the dissolved - MF, ‘and ZrF , suffices to yield the free energy of | ‘complexing, AF this quantity has been cal- comp' " culated for the proposed family of complexes GF, ZrF FeF,ZrF,, and NiF,-ZrF,, as shown in Toble 2 2 12. Smce for NIF ZrF . AF is comp positive, there is additional evudgnce that this 105 ANP PROJECT PROGRESS REPORT TABLE 2,2.12. FREE ENERGY OF FORMATION OF MF »ZrF Free Energy of , S T Saturation A o Free Energy of o, lemperature ofs ctivity _ . o MPrZFe g Solobilits N e tonty CoPlexing AF o, Formation, AF ) - : {mole fraction) (kcal/mole) (kcal/mole) 700 0.0550 2.20 -8.3 - 520 _ NiF . ZrF, 600 - 0.00331 1900 +0.9 ~510 700 0.0180 1100 +1.6 ~500 CFpZeF, 600 0.0210 115 -8.8 ~545 700 - 0.183 0.87 7.8 -535 solid complex is not stable with respect to solid 4 P The activities of MF, in the saturated solution were obfained by combining solubility and activity coefficient measurements. Henry’s law is expected to hold quite well for dilute solutions in molten electrolytes because the environment of a solute ion is not significantly influenced by the presence of other solute ions unless appreciable concentra- tions are present. This has been demonstrated for FeF, in NaF-ZrF, (53-47 mole %) over o range of concentrations from 0.03 to 0.06 mole % by meas- urements of the activity coefficient based on equi- librium studies at 600 to 800°C.!7 The values of the activity coefficients for FeF, (y at 600°C is 3.28; y at 700°C is 2.20) were combined with electrometrically determined activity coefficient ratios 18 (from bimetallic couples) to give the activity coefficients of CrF, and NiF, shown in Table 2,2,12. Previously published measurements15 showing a twofold change in yg, g, with concen- tration in the range’ 0.06 to 1.5 mole % FeF, are tentatively regarded as in need of reinterpretation. The saturation concentrations listed in Table 2.2.12 were obtained from the equations presented in the preceding section on ““Solubility Determina- tions by Measurement of Electromotive Forces of Concentration Cells." The activities of ZrF , were computed from vapor pressures by using the expression 17¢, M. Blood and G. M. Watson, ANP Quar. Prog. Rep, March 10, 1956, ORNL-2061, p 84, - - 18, E, Topol, ANP Quar. Prog. Rep, Dec. 10, 1955, ORNL-2012, p 97. ‘ 3 106 where p is the vapor pressure of ZrF in NaF-ZrF, (53-47 mole %) and p, is the vapor pressure of solid ZrF . Since 3 logp - 9_243 - M and 12.376 x 103 log p = 13.400 ~ ——r— . 3 Iog_p_= 415 3.124 x 10 o " b T where T is temperature in °K, which gives log £— = —0.578 at 600°C , o The equation for p, was obtained from the work of Sense et al.'? The activity of ZrF , in the satu- rated solution was assumed to be the same as the activity in the pure solvent NaF-ZrF 4 (53-47 mole %). This assumption is probably valid in all cases except that of CrF, ot 700°C; here the solu- tion is too concentrated for much reliance to be placed on the calculation. If the saturating phase were simple NiF, in the NiF, experiments, then 19k, A. Sense et al, Vapor Pressures of the Sodium Fluoride—~Zirconium Fluoride System and Derived In- formation, BMI-1064 (Jan. 9, 1956). aNle' = N should be unity (a pure solid has unit activity). The apparent activity is 9 at 600°C and 20 at 700°C, however, which indicates that the saturating phase is something less soluble than NiF,. The discrepancy amounts to a factor of 10 m the activity coefficient or the solubility and to about 5 kcal/mole in the value of AFNI‘Fz; In computing the AF® for the formation of the complex compound from the elements, the free energy esti- mates given in Table 2,2,13 were employed. 20 201, Brewer et al, p 104=115 in The Chemistry and Metallur of Miscellaneous Materials, Thermod ymamics, NNES 1| 'I 'B, ede by L. L. Quill, McGraw-Hill, New York, 1950. , TABLE 2.2,13. SELECTED VALUES OF FREE ENERGY OF FORMATION AF® {kcal/mole) At 600°C - A+700°C FeF, -137.5 -1339 CeFy ~152.5 -148.9 - NiF, 1279 1248 ZrF, -383.4 ~376.8 -PERIOD ENDING JUNE 10, 1956 An independent check on the differences in the standard free energies of formation of the MF . ZrF complex from the elements is afforded by previously published data obtained for bimetallic couples in saturated half t:ells.z'I To a first approximation, the cell reachon was ’ ’ M+MF‘.ZrF = MFZrF, + M" . 'Thlsreactlon neglects transport and junction poten- tials, which are small if the saturated solutions are sufficiently dilute. From the relation End = -AF for the cell reaction, where E is the electromohve force, n is the number of equivalents transferred, ond J is the number of Faradays passed, an approx- mate value of the difference in the free energy of formation from the elements is found, The com- parison is shown in Table 2.2.14, The values of AF, , in Table 2.2.14 were ob- tained by subtrachng the values of AF® given for these complexes in the last column of Table 2.2.12, White the agreement is good, the effect of the high -solubility of CrF, at 700°C is probably responsible for the larger of the discrepancies. 211, E, Topol, ANP Quar. Prog. Rep. Sept. 10, 1955, ORNL-I947, P 85. TABLE 2,2.14. COMPARISON OF FREE ENERGIES OF FORMATION OF MF,+ZsF, OBTAINED BY TWO METHODS ' ~ Difference Betwee_n AF and AF,, » ' (keal/mole) At 600°C : ~ Ar700°C’ ~End 155 | 6.8 AFcalc ' 15 o ' _ 'IS . =End -19.5 2200 cale -20 , :‘ - =20 . -End 35,1 . =363 AFeulc =35 =35 107 ANP PROJECT PROGRESS REPORT PREPARATION OF SOLUTIONS OF LaFg, BaF,, AND RbF IN NaF.Z¢F (-UF F. L. Ddley : - W, T. Ward ‘A special preparatlon containing LaF, -BaF ,°RbF (2,8-1.0-1.1 mole %) dissolved in a sfandard NaF- ZrF ,-UF 4 (50-46-4 mole %) mixture was requested for use in determining whether the introduction of the simulated fission products would produce sig- nificant viscosity changes, The usual methods of purification and filtration were used, and the equip- ment was modified so that small filtered samples could be dispensed directly into the tubes used for the viscosity measurements, The resulting preparc- tion was studied to obtain some indication of the solubilities of the fission products at 800°C in this solvent. It is believed that the solubility of LaF, may be typical of the solubilities of the rare- earth fluorides, since LaF, presumably forms solid solutions with other rare-earth fluorides. Two “filtered samples of the NaF-ZrF -UF , mix- ture without the additive were obtamed ond two mote samples were obtained cfter the fission. product mixture was added. A small portion of solution "which remained unfiltered in the reactor after withdrawal of all the samples was also re- covered for analysis. All filirations were performed at approximately 800 + 10°C, During the course of the ‘viscosity measurements, the presence of a finely divided solid material was detected in the molten mixture, and a small sample of the solid was obtained by decantation. The solid was tentatively jdentified by x-ray diffraction analysis as free LaF,, with occluded solvent. The resulits of chemical analyses of all the samples taken are presented in Table 2,2,15. The results indicate no greater concentration of odditives in the unfiltered portion (reactor heel) than in the filtered samples, and consequently it appears that the additives were completely dis- solved at 800°C. From visual observations, how- ever, the solution seemed to be essentially satu- rated at this temperature, The saturating phase appeared to be free LaF,, and the probable solu- bility value for this compound at 800°C seems to be about 2,5 wt %. During the preparation of this material no particular effort was made to control the filtration temperature at exactly 800°C, and visual observations could not be made conveniently enough to fix a liquidus temperature; however, the order of magnitude of the solubility of LaF, in NaF-ZrF -UF, (50-46-4 mole %) appears to be about 2 mole % at 800°C. TABLE 2.2,15. CHEMICAL COMPOSITION OF NaF-ZrF ,.UF, SAMPLES BEFORE AND AFTER INTRODUCTION OF LaF4-BaF,-RbF TO SIMULATE FISSION PRODUCTS Filtration temperature, 800°C Major Constituents (wt %) Impurities (ppm) Na Zr U Ba La Rb F Ni Cr Fe NGF-ZI'F4-UF4 - Sample 1 . 10.7 38.8 8.8 0 0 0 41,9 9 50- 65 Sample 2 10.3 38.8 8.7 0 0 42.5 35 40 45 Theoretical 10,41 37.98 8.2 O 0 0 42.99 Na F‘Zde-UF“ plus Lqu-Bon-Rb.F Sample 3 9.2 35.7 8.0 .23 2.62 0.86 41.7 80 50 90 Sample 4 9.5 34.6 7.7 L16 2.5 .00 41,8 100 65 160 Reacfor hee' 8.9 37.8 8.2 lc '8 2.3‘ 0087 4‘-2 25 ) 65 150 Theoretical 9.64 35.17 7.98 1.25 3,40 0.82 41.74 ' Solid detected during 18.1 0.62 2.6 viscosity measurements 108 - Although it appears that the solubilities of the rare earth fluorides in the NaF-ZrF .UF, mixture are sufficiently high to prevent prec:pltutlon upon their formation as :fission products during ART operation, some studies of solubility of individual rare earths and rare-earth mixtures are currently under way. One possible. development from such a study might be the demonstration of a simple means for removal of the rare-earth fluorides from the fuel mixture if the magnitudes and temperature dependences of the solubilities are adequate. PRELIMINARY SOLUBILITY VALUES OF CeF, IN NGF'Z’F4-UF4 C. M. Blood As a start of a systematic determination’ of solu- bilities of the rare-earth fluorides in ART-type fuels, such as NaF-ZrF +UF, (55-40,-57mo!e %), one experiment was performed to determine the approximate solubility of CeF, in a typical solvent, A previously purified mixture of the NaF-ZrFA-UF solvent and approximately 2,5 mole % CeF, were enclosed in a nickel container and heated to 850°C., The mixture was continuously stirred by bubbling hydrogen and helium, alternately, through o dip leg. Two hours prior to filtration, the temperature of the solution was lowered to 800°C. Samples of liquid were then withdrawn through a nicke!l filter tube at 800, 700, and 600°C. A second set of filtrates was obtained in the same manner at 700 and at 600°C after the mixture had been given an odditional “hydrofluorination and hydrogenation trecitrhérih The = . are used and the mixture of salts contains m moles of. AgCl per mole of NoC! N faradays of electricity - are pussed ond the *‘idealized” anode compart- results are summarized in Table 2.2, 16 Smce the CeFa' used in these experlmenfs was - ) nof es.pecmlly pure ond certam dxffu:ulties werei_";;”'r;; , L e o Rl RISt o f_r_‘ment .gains N moles of Ag* by electrode reaction TABLE 22,16, l*APéizdxmrrE sbLU'Bl’Lr‘rv.op"f"" *' '."“1 N‘c; moles of CI= by transfer and loses Nt} L . , : A e gy 0. moles © _~“This may be summcnzed as a gain of N moles of GF N NaF P (U (S5ABSHOLE ) Temperuture to W'M; T MelenCs o ey 0k e0 05 05 0.59 " onode ‘compartment has lost N’ - PERIOD ENDING JUNE 10, 1956 apparent in analysis of the filtrates obtained, these date should be considered as preliminary. |t is anticipated that more relioble values will be . available within the next quarter, CONICEN"I'.RATION CELLS AND TRANSFERENCE NUMBERS IN FUSED SALTS R. F. Newton It is well known that transference numbers are needed for correlating the electromotive forces of concentration cells with activities or other thermo- dynamic properties of the solutions involved, While measurements have been and are being made of the - transfer of ions of fused salts across porous dia- phragms, some -investigators doubt that these data "can be properly interpreted as the ‘'true’’ trans- ference numbers of the ions, since the diaphragm _itselfmay cause a flow of salt across the boundary, 1t is of interest to note, however, that relative transference numbers, which are the values of thermodynamic interest, can be measured in an ordinary Hittorf experiment, At is well known that the transport of water as part of a hydrated ion in aqueous systems cannot be determined without some rather questionable assumptions, but standard transference numbers of aqueous ions are given relative to the water. For example, in a mixture of two salts, such as AgCl and NaCl, the transference number of Ag* may be defined relative to the NaCl. The fictions of the ideal transference numbers, ¢ , 7 , and #g, are useful in this analysis. In an ordinary Hittorf transference experiment in which silver electrodes Na*.and NtA ‘moles of Ag by transfer. ‘ 7 CeF, Solubility St T AgCl- by electrode - reaction and a loss of NtA . moles of- AgCI and of Nt’ moles of NaCl by | frcmsfer. The best iha'r can: Ee deone exper:mentally, S ":however, is to compore ‘the AgCl in a given quantity -of NaCl ot the end with that in the same amount of " NaCl at the start,’ 1, for example, the anode portion i - taken for onalysw at the end of ‘the experiment . ‘contains one mole of NaCl, and the *‘idealized”” Na moles of NaCl, this represents 1 + Nt;, moles of NaCl at the _ start. The AgC! at the start was - m + mNt(, ¢ and 109 ANP PROJECT PROGRESS REPORT the AgCl at the end was (n N + m + mNtj,_ - Neg o, where m is the AgCl in one mole at the start and N is the AgCl from the electrode reaction. If m and N are subtracted from expression 1 and the re- mainder is divided by N, the relative transference of Ag¥ is seen to be (2 fag = thg = Mg o By a similar argument it may be seen that (3) . tE"—']—tAgcr These reluhve transference numbers are all that can be obtained unambiguously from the Hittorf experiment, in which N faradays of electricity are passed, the electrode compartments are analyzed, and the AgCl content of the electrode compartment is compared with that initially present in the same amount of NaCl. They are also the only numbers needed for thermodynamic use. Useful correlations of the relative transference numbers with electric conductance may aisc be made, For these carelations, A (the ionic con- ductance) is defined so that, for pure NaCl, Aya + Mgy = k/C, where k is the specific con- ductance and C is the concentration in equivalents per cubic centimeter, For the NaCl-AgCl mixture, K = Cualna + Cagrag + Coirey CNuxNu K CA gA'Ag ~ > o H - K Cerrel tl, = cl - and, when these values are substituted in Eq. 2, it may be seen that CAg C r - 37 ’ 4 the = Thg = tq Na 110 c AAg = A'Nu = Ag - I\Ag = I\Nd = 7R —, K/CNu ' From Eq. 4 it is evident that the relative fransfer of the silver ion is propertional to the concentra- tion of the silver ion and to the difference in the mobilities of Ag* and of Na*, and hence the relo- tive transfer may be either positive or negative. If it is negative, t,, will be greater than unity. For apphcahon to a concentration cell with transference, consider the cell Ag, AgCl in . . AgCl in NaCl, NeCl; m® mole | . Ag; mB mole AgCl/mole NoCl . - Ag/mole NaCl tT T+ dt” At the anode, for each faraday pass.ed, Ag » Agt(m?) + e, A , A o mole Ag*(m4) is lost by transfer, and A (l - t'Ag) mole C| = is gained, At the cathode the opposite processes occur, with each A replaced by B. In the section with a con- centration gradient between A and B, ‘in each element, such as between the dotted lines, the gain of AgCl is 7, and the loss is £} + th ' which gives a net oss of dt' mole AgCl in the element, and AF::(] —!rAAg)FA- (l—tr ) = FA _ B B =FA-FB+ fBy dF . fBth' If ¢’ is constant this simplifies to TA EB) . AF = (1 -1, ) (FA - FB); however, as shown in Eq. 4, " is not expected to be constant, it PERIOD ENDING JUNE 10, 1956 2.3. PHYSICAL PROPERTIES OF MOLTEN MATERIALS F. F. Blankenship G. M, Watson E. R. Van Artsdalen’ PRESSURE-COMPOSITION-TEMPERATURE RELATIONS FOR THE SYSTEMS KF-ZrF, AND RbF-ZrF, S. Cantor Among the important guiding principles in the search for improved or modified fuel mixtures are predictions regarding the changes in properties that will occur as a result of the substitution of one species of ion for another in an otherwise similar melt. To a very considerable extent these predictions can be based on a determination of how the activity coefficients of the constituents change with composition. One of the easiest activities to measure, and a fairly important one to know, is the activity of ZrF in mixtures with alkali fluorides at composmons such that the vapor is essentrally pure ZeF . Mlxtures con- taining 45 mole % or more ZrF4 produce a vapor which, for practical purposes, is pure ZrF,, and the activity of ZrF, is readily obtained from the total vapor pressure. Accordingly, measurements are being made of the pemnent vapor pressures in alkali fluoride~ZrF, systems. It has been found that the volafll:ty of the ZtF “varies in- versely with the size of the alkali cahon. This is a consequence of the more pronounced com- plexing of ZrF, in the presence of the alkali cations, which have less attraction for fluoride ions. Lithium fluoride represents the case of an ,outsfandmgly strong attraction of ‘a cation for - fluoride ions because of the large charge-fo-radws - ratio “of ihe llfluum ion, Correspondmg1y, the = vapor - pressures from 'the LIF-ZrF4 muxturearef _the _highest; “this 'system is curreritly belng,'_ : ‘measured at: the Battelle Memorlai Tnstitute. - "The - - NaF-ZrFy - mixtures - have ‘been - measured at ~ORNL1=3 qgnd at. Batteile.‘ Vapor pressures for - ’KF-ZrF and RbF-ZrF‘ are currenfly bemg de- ‘R. E. Tfaber, Jr., R, E.. Moore, ond C. J Barfon, _ ‘ 'Z,ANP Quar. Prog. Rep. Dec. 10, 1953, 0RNL-]649, P99 o 2RE, Moore and C."J, Barton, ANP Quar. Prog. Rep - .'_}une 10, ‘1954, ORNL-1729, p 101. - e ' B . E. Moore, ANP Quar. Prog._Rep. Sept 10 1954, S 0RNL"|77]| p129. - K. A. Sense ef al., Vapor Pressures of tbe Sodium Fluoride~Zirconium Fluoride System and Derived Infor- mation, BMI1-1064 (Jan. 9, 1956). : termined by the Rodebush-Dixon method that has been used for other ZrF, mixtures at ORNL. This method depends on t11e *‘valve'’ action of the salt vapor above a liquid held at constant temperature; the valve action is manifested by a differential manometer which registers the re- sistance to flow of inert gas through a region occupied by the refluxing salt vapor. The ‘*valve' becomes suddenly much more definite in effect when the inert gas pressure is lowered to a value corresponding to the salt vapor pressure, An absolute manometer is used to measure the pressure at which this occurs, The System KF-ZrF The vapor pressure experiments initiated pre- viously on the system KF-ZrF, were continved. 5 The Rodebush-Dixon method,® prevnously de- scribed,®*7 was employed. In the present study the temperature determi- nations were carried out potentiometrically by using calibrated platinum—platinum-thodium ther- mocouples. Pure ZrF, was prepared by subliming hafnium-free ZrF, at 720°C under a vacuum, Transiucent crystals were then hand-picked from the sublimate. Spectrochemical analyses for ten possible metallic contaminants showed, in every case, less of the impurity than of the available standard, Potassium fluoride was purified by heating the reagenf-grade product to 50°C above ~the melting point, cooling it slowly, and selechng : clear fragmems of \‘he fused salt, “In the KF-ZrF composmons studied, the as- ' ?sumphon that the ZrF, -is responsible for the ‘Yapor pressure was }usflfled by chemical ‘analyses and x-ray “and petrographlc studies of -sublimed '.;matena? obtalned from the apparatus. after a run was completed The vapor pressures . obtamed for ~ the various mixtures are. summarized in. Table ) ' "2 3 1 The constants A and B are: those expressed ' Ss.- Cantor, ANP. Quar. Prog. Rep. Marcb 10 1 956, L ..tORNL-zom. p 103 T bw. H. Rodenbush and A. L. Dixon, Pbys. Rev. 26, 851 (1925). R E. Moore and C. J. Barton, ANP Quar. Prog. Rep. Sept. 10, 1951, ORNL-1154, p 136. 111 ANP PROJECT PROGRESS REPORT TABLE 2,3,1, SUMMARY OF VAPOR PRESSURES FOR THE SYSTEM KF-ZiF Composition Heat of Yaporization Experimental T:mpemture (mole %) Constants of Zer, _AHU Range (°C) ;:—'7 A B (kcal/mole) 750-900 659 34.1 9.754 8,813 | 40.4 750-900 59.6 40.4 9.142 8,321 38.1 775-900 55,0 45.0 8.403 7,778 35.6 825-950 49.9 50,1 7.722 7,363 33.7 875-950 47.3 52.4 7.679 7,467 34,2 9501000 45.3 54.7 . 7.649 7,721 35.3 975-1100 ' 40.4 59.6 7.612 8,465 38.8 1125-1300 34.9 - 65.1 7.387 9,020 41.3 in the relation B logP(mmHg) A -7 where T is in °K, These constants are obtained by treating the data by the method of least squares. The heat of vaporization, AHv, of ZrF, is obtained by multiplying the constant B by 2.303 (°R) or 4,576. The values listed in Table 2.3.1 are considered to be more reliable than those previously published.® Apparently there was some free Z¢F present in some of the earlier experiments, and for the compositions containing 65.9 to 49,9 mole % ZrF4 a reanalysis by least squares gave the new constants. A plot of the data from which the equation for each mixture was obtained is shown in Fig. 2.3.1. A very odd phenomenon is observed when the apparent heat of vaporization, AH_, is plotted against composition (Fig. 2.3.2). A sharp minimum occurs at 50 mole % ZrF,, and the heat of vapori- zation increases linearly with composition in both directions from the minimum. The change below 50 mole % ZsF, is in the expected direction, since ZrF, becomes more tightly complexed by fluoride ions. The change above 50 mole % was not expected; qualitatively it can be explained by assuming that ZrF, exhibits positive deviations from Raoult’s |uw when dissolved in KZrF_. (This version of Raoult’s law calls for a ‘‘two- particle” depression at the Z¢F, end of the composition range and for zero pressure at the KZrF; end.) In practice, the vapor pressure of supercooled liquid ZrF, is not sufficiently well 112 known for a numerically reliable Raoult law to be established for ZrF, systems at 600 to 800°C. The System RbF-ZF, The vapor pressures of various mixtures in the system RbF-ZrF, were determined in the same manner as in the KF-ZrF, system. Reasonably pure RbF was obtained by selecting clear crystals from a slowly cooled melt. Analyses showed that the other alkali metal ions were present in the following weight percentages: Li, 0.023; Na, 0.005; K, 0.16; Cs, 0.24. The dutu for various mixtures are complled in Table 2.3.2 and shown graphically in Fig. 2.3.3. The constants A and B are from the relation B log P (mmHg) = A - where T is in °K, As in the case of the KF-ZrF, system, the heat of vaporization goes ihrough ct minimum at 50 mole % ZrF (Fig. 2.3.4). The linearity with mole fraction is also similar to that of the. KF-ZrF, system, Thus the effect seems to be a real one in spite of its unnatural appearance. A comparison of the effects of the various alkah fluorides is given in Fig, 2.3. 5, in which pressure is plotted as a function of composition at 912°C, the melting point of pure ZrF,. Moore® has studied two compositions for the system LiF.ZrF , and Sense et al.* of Battelle Memorial lnstltute and Moorel =3 have studled the NoF-ZrF system. -8R, E. Moore, ANP Quar. Prog. Rep. ]une 10, 1955, ORNL-I 896, p 81. wl "PERIOD ENDING JUNE 10, 1956 ORNL~LR~DWG 14637 TEMPERATURE (°C) 1200 1100 | 1000 900 800 200 T T T Y, | % \-a% v.o \\Sc’) 07:.“ » 100 \ .\"»*% X PRESSURE (mm Hg) 10 el — 065 = 070 075 . 080 q/rt"m 0.85 o0s0 095 . 1.00 x 103 ' " ' F:g. 2 3. 1 Vupor Pressures of KF-ZrF Mixtures. | "The value of 905 mm Hg for ZrF at the melflng gpomt was obtamed from the su1: - pressure equahon ' : on,3604 iOgP (mm Hg) = ]2 54]7 - -—-—.F-—-—-- . _ Thrs equuhon represents a least-squares treatmenf ; - systems was measured by the fiodenbush dixon of 33 experimental points obtained with the present apparatus, The results conclusively show that in solutions of ZrF, and alkali fluorides the vapor hmahon voper: i - _ _ VAPOR PRESSURES AND PHASE EQU]L'BRIUM : pressure of ZrF decreases as the ulkoh ion slze _'mcreases. L e R DATA |N Tl‘lE FeClzoKC| SYSTEM | ' C.C. BeuSman ' The vupor—pressure of FeCl in FeCI -KCI. %0n assignment from the Oak Ridge lnsfltute of Nuclear Studies. 113 ANP PROJECT PROGRESS REPORT GO ORNL—LR-—-DWG 14638 a6 a4 42 s O =1 40 38 ‘ \\"--. 36 HEAT OF VAPORIZATION, AH,, CF Zrf, (kcal/mole) . e . -_’_/ 34 i 32 L KF Fig- 203-2. _ 20 Vaporization of Z‘H"4 40 .60 80 ZtF, {mole ?'o-) in KI"’-ZrF4 Mixtures. ZIrfy Effect of Composition on Heat of - method in céncentrcltions from 100 to 45 mole % FeCl, in KCI. Pressure vs composition curves " obta med at 850 and 900°C are shown in Fig. 2.3.6; these are typical of the curves obtained at temper- ‘atures from 700 to 1050°C. The strong devigtion ~ from ideality suggests that the FeCl, is complexed in the liquid at these temperatures. Since the experimental evidence was obtained in phase equi- librium studies of the compound K, FeCl, in the solid state, it is interesting that the same stoi- chiometry may apply to the complex in ‘the liquid. The strong depression of the vapor pressure below 45 mole % precludes further experimental efforts with the' Rodenbush-Dixon technique. The fran- spiration method will be used to complete the experimental curve on the KCl-rich side of the ~ graph., The phase diagram for the FeCI -KCl system, shown in Fig. 2.3.7, was deiermmed by two ex- perimenta} techniques. First, both convenhonol ' cooling and differential cooling curves were ob- tained on small melts, and, second, quenched. samples were prepared by the grodient-quenching techniques used in this laboratory. - Petrographic examination of the quenched and slowly cooled samples revealed two well- defined compounds in the solid state, K:!FeCI4 and KFeCl Both compounds appear to have solid-state transfor- mations, and K,FeCl, melts incongruently. TABLE 2.3.2, SUMMARY OF VAPOR PRESSURES FOR THE SYSTEM RbF--ZrI‘-’4 Experimental Temperature ‘ c‘;mp:s“wn Constants Heat of Vuporizufion | | mole %) of ZrF,, AH Range (°C) A B 4 v ZrF, RbF ‘ r(kca 1/mole) 800-850 81.6 18.4 10.778 9,506 435 725-825 74.6 - 25.4 10,359 9,189 42.1 750-825 70.0 30.0 9.808 8,790 403 . 750-875 . 65.1 34.9 9.143 ° - 8,358. 38.3 . 775-875 . 60.0 40.0 8.789 8,078 . 37.0 775-950 . 548 45.2 8.307 7,902 362 ¢ 850~975 '~ 50.4 49.6 7.555 7,310 33.5 900-1050 45.2 54.8 7.512 7,794 35.7 1000-~1200 40.0 60.0 7.290 8,146 - 373 1125-1275 - 35.0 65.0 7.296 8,790 - .- 40.3. 114 PERIOD ENDING JUNE 10, 1956 ORNL-LR—DWG 13012A TEMPERATURE (°C) 750 850 950 1150 1050 1250 500 o n 200 100 (bH ww) 3YNSSIYH4 20 10 o8 oo 0.60 Yrek) x 10° Fig. 2.3.3. Yapor Pressures of RbF-ZrF , Mixtures. 115 ANP PROJECT PROGRESS REPORT SOMNSEENTTIL ORNL-LR-DWG {13011A 46 H H b / / J A / / / \ 9 H ~n H o /lD/ 2] m Od N .—4/ @ HEAT OF VAPORIZATION, A H,, OF ZrF, (kcal/mote) W H 32 RbF 20 40 60 80 ZrF, Zrf, {mote %) Fig. 2.3.4. Yaporization of ZrF, in RbF-ZeF , SURFACE TENSION AND DENSITY OF NOF-ZI'F4 F. W. Miles Preliminary values of the density and surface tension of NaF-ZrF, (53-47 mole %) were obtained by the maximum-bubtle-pressure method previously described.'? The results are summarized in Table 2.3.3. Difficuities were encountered as a result of plugging of the nickel tube above the capillary tip. In every case the concentration of oxides and oxyfluorides in the plug material appeared to be quite high, and thus it is thought that there was air contamination of the helium stream. The assembly is being examined in an attempt to locate and eliminate the source of 5 wF. W. Miles, ANP Quar. Prog. Rep. March 10, 1956, 116 Effect of Composition on Heat of BN ORNL~LR-DWG 14639 900 o LiF-2rF, {R.E.MOORE) . 4 NoF—2ZrF, (R.E.MOORE) ‘ 800 o NaF~ZrF, (K.A.SENSE, BMI) ‘ 4 KF~ZrF, (S.CANTOR) . // 4 700 * RbF—ZrF, (S.CANTOR) 1 E A M £ 600 - - 7 & / 2 . 'Vg/h 1%/ ' n 500 \ g 87 ) o &7 / | & &, J/{/RbF=2rF, : > /) " 300 S /1 QSQ/" NaF —ZrF,1 1/ , 200 - § / 7 LiF~ZsF, /// #KF"Z'& 100 / / / . 7 %4 0 -—w - MF 10 20 30 40 50 .60 TO B0 90 ZIrfy ZrFy (mole %) Fig.2.3.5. Vapor Pressures of MF-ZI'F4 Mixtures at 912°C. : UNCLASSIFIED ORNL-LR-DWG 14640 240 4 220 / /o 200 : 180 J 160 / 120 140 | /. ] / 60 . .’/ / I TOTAL PRESSURE (mm Hg) 40 7 //Qescrc 20 L - / o _ K 0 20 30 40 50 60 70 80 80 FeCl, FeCly (mole% ) Fig. 2.3.6. Vapor ‘Pressures of FeClz-KCl Mixtures at 850 and 900°C. o TEMPERATURE {°C) PERIOD ENDING JUNE 10, -195‘6 UNCLASSIFIED ORNL-LR~DWG $46414 800 "N — 500 \ / asec 71 - 7 o 400 385°C ~ I i #| 358°C QRO 300 298%C 255°C 200 | K,FeCl, KFeCl, 400 KCt 10 20 30 40 50 60 70 80 90 _ FeCI2 FeCl, (mole %) Fig. 2.3.7. Phase Diagram of the FeCl_-KCl System, TABLE 2.3.3. PRELIMINARY VALUES OF DENSITY AND SURFACE TENSION OF NaF.ZtF, (53-47 MOLE %) OBTAINED BY MAXIMUM-BUBBLE-PRESSURE METHOD Wall Thickness Inside-RadiUsf . . Density . ams o S Temperature Surface Tension f;f Capul!ary(. _ of Cd.pl lor? ey (a/em3) (dynes/cm) o lem) (mils) S : o o 0758 4 o s0 3.049 131 10,0513 9 600 3.014 e 00513 9 | 00 - 3,051 . . 125 9 - 0.0513 700 2916 o120 117 ANP PROJECT PROGRESS REPORT contamination, and changes have been made in the design of the capillary tips. The four measurements listed in Table 2.3.3 were obtained on the same batch of material. Although the melt was repurified, in place, after each determination, the formation of plugs in- dicated contamination during measurement. Further- more, the batch was kept at high temperatures for several weeks, When the reactor was opened, it was observed that some ZrF, had sublimed out of the solution, These factors indicate that the values given in Table 2.3.3 must be considered to be only of a very preliminary nature. |t is encouraging, however, to note that values of the same order of magnitude were obtained with capillary tubes of different wall thicknesses, This behavior seems to indicate contact angles of less than 90 deg. Consequently, the inside radius of the capillary controls the bubble size, and no particular uncertainty exists as to whether the value of the inside or of the outside radius should be used in the calculations. GROUND-GLASS LIGHT DIFFUSER PYREX TUBE z ‘21 PROFILE OF A SESSILE DROP \ Additional measurements will be made on a fresh batch of similar composition. When more reliable values are obtained, the measurements will be extended to other mixtures of interest. SURFACE TENSIONS OF MOLTEN SALTS S. Langer The sessile-drop technique for measuring surface tension, briefly discussed in the previous report,!! was applied to a number of samples of an NaF-ZrF (53-47 mole %) mixture to establish the importance of a number of experimental variables. Since it has been demonstrated that NaF.ZrF, mixtures of this general composition do not wet C-18 graphite, all the experiments described here were conducted with this material as the supporting plaque. ' A schematic drawing of the sessile-drop appa- ratus is shown in Fig. 2.3.8. The sample on its Vg, Langer, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 105. . UNCLASSIFIED ORNL-LR-DWG 14642 FURNACE TUBE SUPPORT SOLENOID-OPERATED SAMPLE-MANIPULATING RODS NICKEL TUBE COOLING-WATER JACKET THERMOCOUPLE WELL REFERENCE SCALE FURNACE SESSILE DROP SAMPLE HOLDER CONSTANT-TEMPERATURE 8LOCK COOLING-WATER JACKET - QUARTZ WINDOW SOLENOQID-QPERATED WINDOW SHIELD Fig. 2.3.8, Diagram of Sessile-Drop Apparatus for Measuring Surface Tensions of Molten Salis, 118 j',-_f";surrounds the drop._, supporting plaque is loaded into the tube through the sample-loading flange. - The central portion of the tube is then baked out under high vacuum to remove volatile materials which might con- taminate the surface of the sample. After the fumace is adjusted to the operating temperature, ‘the sample is pushed to the center of the constant- temperature block * with the solencid-operated sample-manipulating rods. Photographs are taken of the sample as a function of time to ensure the establishment of an equilibrium droplet. A photo- graph of the sessile drop of NaF at 1025°C, taken during calibration work, is shown in Fig. 2,3.9. UNCLASSIFIED] - PHOTO 1822 Fig. 03-9o Sessule Drop of NeF at 1025°C on Gruphlfe P"’q“e':-,ff-.af,-teporfed previously. H_"i'_j_so’nsfactory theoretical interpretations of the re.. The dlmensmns ‘'made under vacuum, of fhe sessnle drop (fl‘le;_r:j‘__- sults . Cor- PERIOD ENDING JUNE 10, 1956 culated by using the tables of Bashforth and Adams!2 and the densities obtained by Miles.!3 The initial ‘experiments indicated no change in surfoce tension when an atmosphere of helium was admitted to the apparatus in which tests had been When tests were conducted in vacuum, however, volatilization of ZrF, caused large changes in composition of the drop; helium atmospheres have been adopted as standard for future tests, Data . typical of those obtained to date for NaF-Z¢F, melts whose initial composition was - 53 mole % NaF and 47 mole % ZrF, are shown in ‘Table 2.3.4. While the data of |11e earlier ex- periments are viewed with some uncertainty, the relative agreement of the surface tension values ' obtained from droplets of widely varying sizes i‘,_,_"--(as evidenced by the spread of sample weights . ondx/z rot.tos), is gratifying. ' "SPECIFIC CONDUCTANCE AND DENSITY 'OF FUSED LITHIUM, POTASSIUM, AND cssnum FLUORIDES“ E. R. Van Artsdalen “Electrical _cond_udance and density have been " measured as o function of temperature for pure, ~fused lithium, potassium, and cesium fluorides, A - ‘newly designed bridge, in conjunction with a S'pei:iei cell eiloy, was used in the measurements. The results fabricated from platinum.rhodium of - thé measurements are expressed by the - f‘”-e'quafions listed in Tables 2.3.5 and 2.3.6. Similar data were also. obtained for several rare-earth = halides. Phofogreph of Reference Scale nnd}.,;_,‘r “These- dofa hove been correlated ‘with those 12F Bashforth and J. C. Adcms, An Attempt to Test the Theories of Capdlary Act:on,,Cambridge Univ. - Press, London, 1883, ; , 13¢, v, Miles, prwate commumcahon, to S. Longer. ”Deto:ls ‘of this work will be published in separofe: ‘ reports and arhcles fro_rrp the Ch__emlsfl'y Dtv_lsion. T 119 It ‘is now possible to. give ‘in _terms of the structure and 1ransport L maximum “diameter, 2x, and " the d;stance frem';j__':"‘%?P’°P°m°5 of the hqu:d 5"“5° : the. line of the maximum dmmefer to the apex, z) " o .-—_ore measured dlrecfly from the photographtc nega-"_-;_f'_'. - tive by using o toolmc:ker s microscope.’ : .]_fi?rechons for amsotroplc shrmkage of the fllm are ~obfaired : by .using the reference - scele “which - ' The surface tensuon is cul-'*'_ ANP PROJECT PROGRESS REPORT TABLE 2,3.4, SURFACE TENSION MEASUREMENTS ON NaF-ZrF, (53-47 MOLE %) | Tempércture Sample Weight Weight Loss Final Composition* Density Surface Tension x/z (°C) (g) (%) {mole % ZeF ) (g/cm’) (dynes/cm) 601 ** 0.870 28.2 36.6 1.71706 3,043 14 603** 0.870 28.2 36.6 1.66592 3.040 141 611 ~ 0.4502 3.1 43.5 1.53637 3.030 141 610 0.4502 3.1 43.5 1.56345 3.031 134 601 0.4502 30 43.5 1.60542 3.043 122 601 0.4502 3.1 43.5 1.57069 3.043 132 601 . 0.4502 3.1 43.5 1.55482 3.043 136 - 601 : . 0.4502 3.1 43.5 1.56994 3.043 131 610 0.5097 3.7 41.4 1.68795 3.031 . - 17 610 0.5097 3.7 41.4 1.65982 3.031 - - 24 602 70,5097 3.7 41.4 1.64464 3.043 o128 " *Coleulated from analysis of material after test. **Tested in vacuum; all other samples tested under helium atmosphere. TABLE 2.3.5, CONSTANTS OF EQUATION FOR SPECIFIC CONDUCTANCE- K, OF FUSED LITHIUM, POTASSIUM, AND CESIUM FLUORIDES K=a+(bx 10~ + (c x 10-5)¢2, ohm™licm=1, where ¢ is in °C Sals b Standard Deviation Applicable Temperature a , @ c (ohm=T.em=1) Range (°C) _ LiF +3.805 +1.004 -3.516 0.008 8471027 KF —3.493 +1.480 —6.608 | 0.009 . 869-1040 CsF - —4,511 +1.642 ~7.632 . 0.009 725921 TABLE 2,3.6, CONSTANTS OF EQUATION FOR DENSITY, p, OF FUSED ' LITHIUM, POTASSIUM, AND CESIUM FLUORIDES p=a—(bx ]0'3)t, g/cma, where t is in °C Slt b Standard Deviation, Experimental Temperature a 7 o a g (g/cms) Range (°C) LiF 2.2243 0.4902 0.0003 | 876-1047 KF 2.4685 0.6515 0.0003 ' 8811037 CsF . 4.5489 1.2806 0.0004 712-912 120 [ . PERIOD ENDING JUNE 10, 1956 2.4. PRODUCTION OF FUELS G. J. Nessle G. M. Watson L. G. Overholser EXPERIMENTAL PREPARATION OF VARIOUS FLUORIDES B. J. Sturm Continued use of structural metal fluorides for research necessitated the preparation of additional quantities of these materials by the several meth- ods developed previously. As in the past, use was made of chemical, x-ray, and petrographic exami- nations to establish the identity ond purn‘y of the materials. Additional CrF; was prepared by hydrofluorina- tion of anhydrous CrC|3, as well as by the thermal decomposition of (NH,), CrF ‘The latter is pre- pared by the mferacnon of hydrated CrF with NH,F-HF, and it may also be reduced by hydrogen to yleld CrF,. Several batches of NiF, were pre- pared by the- hydrofluorination of either hydrated NiF, or hydrated NiCl,. One batch of AgF was prepared by hydrofluormahon of Ag,CO,. Methods described prevmtJSIy were used for the prepara- tion of additional quantities of CeF, and of LaF,. An attempt to remove oxide from a KF-ZrF 4 mnx- ture by hydrofluorination was unsuccessful. Runs at 600°C failed to remove the oxide completely and, while this could be accomplished at 850°C, the loss of ZrF, was sufficient to alter the compo- sition to the point where the material was unsatis- factory for its intended purpose. : A batch of CuF, was prepared for use as a car- rier in certain spectrographlc boron analyses. Ma- terial low in boron and silicon was requested for - this purpose. Anhydrous CuF, was. hydrofluor- inated for 4 hr at 400°C, but this short period of - treatment failed to reduce the boron content to a. sufficiently low level. However, extending the period of hydrofluorlnaflon to 20 hr yleided o ma- _terlal thaf was acceptable. ' A fuel mixture containing Li7 and enrached ura-_f . nium was prepared. for radiation dumage stud|es.‘lgt Since Li7 was available only as the carbonate, 0 prelimmary run was made with normal Li,CO., UF,, KF, NaF; and excess NH,F.HF to esfabhsh‘_”" ' ture containing Li’7 and U233, oxide free, but a sulfur content of 170 ppm was reported, most of which came from the Li7C03. The fuel was subjected to 5 cycles, each of which entailed a 10- to 15-hr hydrogen treatment and 1 hr of hydrofiuorination at 800°C. The product then contained 45 ppm of sulfur and a trace of oxide. Hydrofluorination at 700°C for 6 hr then produced an oxide-free product. LABORATORY-SCALE PURIFICATION OPERATIONS F. L. Daley W. T. Ward With appropriate modzflcahons, the standard hy- _drofluorination-hydrogenation process was used to prepare a number of especially pure materials requested for various purposes. Of these only an NaF-ZrF,-UF, mixture containing additives of Ban, LaF3, and RbF had not previously been pre- pared by the standard procedure. Operational facil- ities to transfer and dispense small samples (200 to 500 g), especially into apparatus for accurate evaluation of physical properties of materials, were constructed and used successfully during this quar- ter. There is evidence that direct transfer of the purified melt into the apporatus has significantly improved the purity of the test material and, ac- cordingly, increased confidence in the test results. Copper-lined stainless steel reactors were used successfully in the purification operations. © PILOT- SCALE PURIFICATION OPERATIONS J P. Blakely C. R. Croft ‘ J. Truitt The pilot-scale purification facility processed 68 '-"-fbatches totaling approximately 1300 Ib of various - fluoride compositions for use in small-scale corro- " sion’ testing, phase equilibrium’ studies, and phys- .,.’,:cal ‘property studies. " The bulk of the material was produced during February when the facility “ was operating on a three-shift, 24-hr-day basis. As a result, the backlog reported last quarter was ,Z-r'.ehmmated and normal operahons were resumed. whether fuswn of this” mtxture would yield “an . : : : oxide-free product After the preliminary run proved - successful the procedure was applied to the mix- The material was 18, J. Sturm, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 100. 121 ANP PROJECT PROGRESS REPORT The use of copper-lined stainless steel reactors in the 5-1b units has proved successful. Therefore, when the nickel reactors now on hand wear out and are discarded, they will be replaced with copper- lined stainless steel vessels. The new copper- lined vessels, in addition to showing promise of having longer lives, ore more versatile than the nickel reactors in that they can be used in the preparation of BeF,-containing mixtures. Although the supply of 50-Ib nickel reactor ves- sels is ot present sufficient, these will also be re- placed with copper-lined stainless steel vessels as they are discarded. PRODUCTION-SCALE OPERATIONS J. P. Blakely Seventeen batches totaling approximately 4230 Ib of fluoride compositions were processed by the production-scale facility. Two new copper-lined stainless steel reactors were received about the first of March and were put into production opera- tion immediately. To date, nine batches have been processed in one reactor and eight in the other with no apparent difficulties. Recent external examina- tion of these reactors show no deterioration or faults, and it may be that the life service of these reactors will be much better than expected. In spite of the temporary shutdown of this facil- ity, the processed fluorides inventory has remained high. Operations were slowed down during the last three weeks of this quarter because of a shortage of storage containers, A survey is now being made to determine the anticipated requirements for proc- essed fluorides during the next six months. A recent communication from Pratt & Whitney Aircraft J. E. Eorgan gave fairly firm commitments for approximately 1700 Ib of processed fluorides per month for the remain- der of the calendar year. The ORNL Aircraft Re- actor Engineering Division has been using an av- erage of about 1000 |b of fluorides per month. Although the new survey is not complete, indica- tions are that Aircraft Reactor Engineering Divi- sion usage will range around 1000 to 1500 Ib per month for the remainder of 1956. The demands will not reach the production capacity of the facility, but ot feast they will make it economically fea- sible to maintain continuous operation, The remaining 7500 Ib of hafnium-bearing zirco- nium fluoride ordered from an outside vendor was received and all the material was accepted as satisfactory. BATCHING AND DISPENS_|NG OPERATIONS J. P. Blckely F. A. Doss The batching ond dispensing facility dispensed 148 batches totaling approximately 4900 ib of proc- essed fluorides in batch sizes ranging from 1 to 250 Ib. The total amount dispensed was about 1000 b less than was dispensed the previous quar- ter. Therefore there was another gain in stock inventory in spite of decreased production. A ma- terial balance for the quarter is given in Table 24.1, The production and use of special compositions is decreasing steadily, but periodic increases can be expected as new materials are developed for testing and investigations. The use of NaF-ZrF, (50-50 mole %) and NaF-ZrF ,-UF , (50-46-4 mole %) or closely related compositions is expected to in- crease to about 3000 ib per month during the first TABLE 2.4.1. MATERIAL BALANCE Material (1b) NOF-ZI'F[UF4 NaF-ZrF4 (50-46-4 mole %) (50-50 mole %) Special Total On hand at beginning of quarter 6,235 1,875 591 8,701 Produced during quarter 3,543 198 1,789 5,530 Total 9,778 2,073 2,380 14,231 Dispensed during quarter 3,828 0 1,078 4,906 On hand at end of quarter 5,950 2,073 1,302 9,325 122 half of fiscal year 1957. This estimate includes Pratt & Whitney requirements. At this rate, full production -should still be able to maintain a high stock inventory and to allow for any necessary shutdown of the production-scale facility. The ef. ficienf handling, maintenance, and dispensing of some 200 storage cans has made it possible to fill all sizes of orders with a minimum of delay to operating sections of the ANP program, - PREPARATION OF ZrF‘ FROM Z.Cl J. E. Eorgan J. P. Blakely The facility for the conversion of hafnium-free Z:Cl, to Z¢F, was operated three times during this quarter. The mechanical operation of the unit is now satisfactery. Handling techniques need to be improved to cut down on raw-material and finished- product losses. Also, the finely divided nature of the product ZrF has presented a dust-loss problem of considerable magnitude. Steps are being taken to eliminate or greatly improve this situation. Large quantities of ZrF, dust were carried into the proc- ess gas exit lines and coused major line-plugging difficulties, A change in piping arrangement im- proved this situation somewhat in the last run. Rough material balances of the three operations indicated a total loss of about 22% of product in the first two runs and 11% in the last run. Approx- imately 75% of this loss was accounted for in ma- terial removed from the exit gas lines. A new dust filter has been ordered for installation on the unit which should cut down the dust loss considerably. Operation of this facility will be postponed until this new filter arrives or there is an urgent. need for more hafnium-free ZrF,. . Analytically, the preduct material was as good or - 'befler than mqterlol previously or presently being used. ' Four samples were taken from each batch as it was being discharged from the unit, each sample - . representing roughly one-fourth of each batch so - that uniformity could be tested. - The analytical - "~ results are given in Table 2.4.2, Although the first" " batch was not so good as the last two, either in - ‘_,umformlfy or quallty, |t was consudered to be us- :able. ‘ ' : o SPECIAL ssnv:css o J P Blckely F. A. Doss Requests for service in filling, 'draining,. and sompling operations were at a normal ond steady level. The greatly accelerated pace set in the PERIOD ENDING JUNE 10, 1956 TABLE 2.4.2. PURITY OF ZrF, PRODUCED FROM Z(Cl, Major | . . - mpuriti Run Somple Conshtuents 'z m;s Zr F Cl Ni Cr Fe 1 1 53,6 450 0.29 90 55 715 2 53.7 44,2 0.40 80 55 595 3 54,3 44,3 0.50 55 45 595 4 542 44.6 0.63 55 45 620 2 1 54,5 45.2 029 65 40 375 2 55.1 44,8 0,33 50 40 400 3 54,6 44,2 034 65 35 375 4 54,6 45.2 0,36 S50 35 375 3 1 54.4 45.4 039 85 50 420 2 54,3 44,3 0.45 70 25 420 3 54,5 44,9 0.43 75 30 410 4 54,8 44.6 039 60 30 400 Theoretical 54,5 45.5 *The hafnium content of the material was less than 100 ppm, the boron content was less than 1 ppm, the sulfur content was about 30 ppm, and the carbon content was 300 ppm. previous quarter was not maintained. Past expe- ‘rience indicates that filling and draining opera- tions will be requested in cycles, depending upon the rate at which tests are started or terminated. Present indications are that another accelerating cycle will occur in the next quarter. Approximately 3000 Ib of processed fluorides and 1500 ib of liquid metals were used to chorge test o equupmenf durmg the quarter in charge sizes rang- " ing from 50 500 Ib. -Since the liquid metal (NaK) “inventory had run below 1500 Ib, another 5000 Ib- “of noneutectic NaK (56% Na-44% K) was ordered ‘and received. It is estimated that this supply - should last six to eight months unless consumption ~is sharply increased. - ' - ~A serious comphcohon arose in’ the d|5posul of NaK during the cold months of this winter. This ‘material, which is normally molten at room temper- atures, could not be disposed of by underwater ~ “injection because it froze in the transfer lines. It - -~ was fortunate that sufficient storage cans were: available to contain the NaK drained from equip- ment until warmer weather arrived and that a com- paratively small amount of NoK was drained from 123 ANP PROJECT PROGRESS REPORT equipment during this quarter. It has been re- quested that electric power be supplied at the disposal quarry so that disposal of both sodium and NaK can proceed regardless of atmospheric temperature. Present rough estimates of the cost of such an installation range from $250 to $1000. Firmer figures are fo be obtained, and, if the range above is correct, adequate power is to be instalied for liquid metal disposal, In-pile loop No. 5 was filled with enriched NaF- 124 ZrF -UF 4(53.5-40-'6.5 mole %). Two new batches for in-pile loops were processed. Dates for filling loops with these batches have not yet been set. Two other small batches of special enriched mix- tures were also processed this quarter. One batch, NaF-ZrF ,~UF, (62.5-12,5-25.0 mole %), was to be used at ORNL for special radiation tests. The other batch, NaF-UF, (66.7-33.3 mole %), was or- dered by Battelle Memorial Institute. ' PERIOD ENDING JUNE 10, 1956 2.5. COMPATIBILITY OF MATERIALS AT HIGH TEMPERATURES F. Kertesz PENETRATION OF GRAPHITE BY MOLTEN FLUORIDES H. J. Buttram G. F. Schenck’ It appears very likely that graphite is the only reasonably effective reactor moderator material that will remain chemically stable upon direct exposure to molten fluoride fuel mixtures. The use of graph- ite to contain such fuel mixtures has been known for a long time to be practical for simple laboratory- scale experimental equipment. The graphite most commonly used for this purpose has the commercial designation C-18, However, little information is available concerning the rate of penetration of reactor-grade graphite by molten fluoride fuel mix- " tures, and accordingly an investigation of this phe- nomenon has been initiated. , Specimens of APC2 graphite 1/ X l/ X !’2 in. were socked at 600°C for various perlods in sealed cap- sules of Inconel containing molten NaF-ZrF or NaF-ZrF4-UF4 mixtures under helium afmospheres. | The specimens were then sechoned and examined under the petrographic microscope. Preliminary observations indicate that the APC graphite was completely penetrated by NaF-ZrF (53-47 mole %) in 1 hr at 600°C but that NaF-ZrF -UF (53.5-40.0- 6.5 mole %) had not detectably penetrated the graphite in 10 hr at 600°C. The rapid penetration of the graphite by the NaF-ZrF, melt is conSIdered. to be surprising, smce such penetraflon is in-con- trast to the stability of C-18 graphite in this and similar fluoride mixtures. Other graphites of higher . density will be examined in these and other fluoride ‘mixtures. If necessary, attempts will be made to - / prevent penetration of the graphite by prior lmpreg- nation with hlgh-melhng-pomf mafenals such as “NaF or CaF EFFECT or ATMOSPHERE ON ANOD!C - ' mssowflon OF NICKEL IN MOLTEN NaOH F A Knox An uf’rempt has been made to study the efflc;ency " of -anodic. dissolution: of nickel in NaOH under - various. atmOSpheres and various condmons of cur- lOrs assignment from Pratt & Whitney Aircraft, Nanona? Carbon Co. code designation, rent density and temperature. The apparatus con- sisted of ‘a crucible of Morganite (recrystallized alumina) containing molten sodium hydroxide into which a nickel anode and a cathode, as well as o nickel thermocouple well, were inserted. This electrolytic cell and its furnace were placed in a large chamber which could be evacuated or filled with the desired atmosphere. A lead storage bat- tery served as the power suppiy, the circuit also contained a voltmeter, ammeter, and ampere-hour meter, as well as rheostats to control the current density applied. - When the assembled electrolytic cell was main- tained at the desired temperature without an applied voltage, attack on the electrodes was negligible. Passage of current resulted in erosion of the anode and subsequent deposition of crystalline nickel on the cathode. The weight gain of the cathode did not correspond precisely to the weight loss of the anode, since the crystalline deposit was not per- fectly adherent; in addition some sodium alumi- nate, formed by the reaction of NaOH with the Al 0, of the crucible, apparently contaminated the cathode deposit. Loss in weight of the anode was accordingly used to evaluate the efficiency of the cell, , The data shown in Table 2.5.1 were obtained at a current density of 2 amp/cm?. At this current density the cell. effmency is very poor under all conditions. The dissolution efficiencyis decreased slightly by substituting hydrogen for helium as the - static cover gas and is reduced considerably by bubblmg hydrogen i'hrough the ‘molten electrolyfe. : ~ TABLE 2.5.1.- EFFICIENCY OF ANODIC - 'SOLUTIONS OF NI_CI_(EL ll_fil NaOH Current denséi.ty: 2 _,a_nfipr/cr'nz' o - " Cell Efficiency (%) . Téir;perutdre : - Under . -With Hydrogen . = ey . U"d" .- Static . - Bubble&rT_hrdfigh o He_llyupj 7_.Hydr'ogen_ - the NaOH 350 0,01 0,006 600 0.5 0.’42-' 800 2,77 2,12 0,66 125 ANP PROJECT PROGRESS REPORT 2.6. ANALYTICAL CHEMISTRY J. C. White DETERMINATION OF BARIUM, LANTHANUM, AND RUBIDIUM IN FLUORIDE FUELS A. S. Meyer, Jr. B. L. McDowell Methods were developed for the determination of barium, lanthanum, and rubidium in NaF-ZrF -UF as part of a program for determining the effect of typical fission products on the physical properties of molten mixtures of fluoride salts. For the barium determination the fluoride salt mixture is dissolved in fuming sulfuric acid, and the barium remains in the insoluble residue, as the sulfate. After dilution of the sulfuric acid solution with water, the barium sulfate, which is contaminated with uranium and zirconium, is separated by filtration and purified by dissolving it in a hot ammoniacal ‘solution of ethylenediaminetetraacetic acid. It is then reprecipitated by acidification of the solution with hydrochloric acid. The reprecipitated barium sulfate is filtered off. The filter paper is charred, and the barium sulfate is then ignited to constant weight at 600°C. Lanthanum is separated from the filtrate from the barium determination by precipitation, as the oxalate, from a neutral solution of ammonium oxa- late, Coprecipitated zirconium and vranium are separated from the original oxalate precipitate by dissolving the precipitate in concentrated nitric acid and carrying out a second oxalate precipita- tion from a neutral solution of ammonium oxalate. The purified oxalate precipitate is converted to the nitrate by digestion with concentrated nitric acid and evaporation to dryness. The residue is taken up in a small amount of water or very dilute nitric acid. Lanthanum is finally precipitated by the addition of solid oxalic acid to the hot, slightly acid solution, The lanthanum oxalate is ignited at 900°C and weighed as La,0,. Ammonium salts in the filtrate from the lanthanum determination are destroyed by digesting the solu- tion with aqua regia. The atkali metals are sepa- rated and converted to the hydroxides by passing the solution through an anion exchange column in the hydroxide form. After all traces of ammonia are removed by evaporating the alkaline solution to dryness, the rubidium is precipitated and weighed as the sparingly soluble tetraphenylboron salt. Potassium, which is present as a trace contaminant 126 in the base fuel, is also precipitated as the tetra- phenylboron salt., The potassium in the tetra- phenylboron precipitate is determined by flame photometry, and the weight of the precipitate is corrected accordingly. ‘ Good precision has been obtained in analyses of replicate samples of fluoride mixtures ‘which contained 1 to 3% of each of the above-mentioned elements. These methods are also being tested for the determination of cerium and the proximate determination of the cerium group of the rare-earth elements in fluoride fuels, DETERMINATION OF NIOBIUM IN FUSED MIXTURES OF FLUORIDE SALTS BY THE THIOCYANATE METHOD A. S. Meyer, Jr. B. L. McDowell R. F. Apple A method for the spectrophotometric determina- tion of niobium as the thiocyanate complex, re- ported by Ward and Marranzino,! was applied to the determination of niobium in NaF-ZrF,-UF . The method is based upon the reaction of niobium(V) with thiocyanate in a mixed solution of 4 M hydro- chloric acid and 0.5 M tartaric acid to produce a complex which exhibits an absorption maximum at 387 mp in an ethyl ether—acetone medium. The addition of acetone to the extract of ethyl ether containing the complex inhibits the polymerization of the thiocyanate ion and stabilizes the soluhon for at least 20 hr. Tartaric acid eliminates the interference of ura- nium, which also forms a complex with thiocyanate. The interference of the red color of the iron(lli)- thiocyanate complex is eliminated by shaking the ether extract with a solution of stannous chloride for 30 sec to reduce the iron(i!l) to iron(ll). The concentration of the reagents in the aqueous phase must be held within rigid limits if repro- ducible absorbance values are to be obtained. In order to maintain fixed concentrations, the fluoride sample is dissolved by fusing it with - potassium pyrosulfate and dissolving the melt in 1 M tartaric acid. A 1- to 10-ml aliquot of the resulting solution is combined with a sufficient 'E. N. Ward and A. P. Murrunzmo, Anal Chem. 27, 1325 (1955). o volume of 1 M tartaric acid solution to yield a total volume of 10 ml.. Then 15 ml-of the hydro- chloric acid—tartaric acid reagent {9 M HCl and 0.5 M tartaric acid) and 15 ml of 20% {w/v) am- monium thiocyanate solution is added, following which the aqueous phase is extracted wnh 5 mi of ethy! ether.. , | Tests of this method have revea!ed thct a Imearr relationship exists between the abscrbance and the concentration of the niobium in the range 0.2 to 2.0 pg/ml. The coefficient of vanohon for this me’rhod bused on standards, is 2%. o ' DETERMINATION OF TRACES OF COPPER IN FLUORIDE FUEL MIXTURES A, S. Meyer, Jro B. L. McDowe” ' Since copper is cons:dered to be compatlble wnth molten fluoride salts containing sulfur’ impurities, the nickel reactors used for the production of fluoride fuel mixtures have been:equppped,wuh ‘copper liners to prolong their useful lifetimes. This modification in production technique neces- sitated the development of a sensitive method for the determination of copper in mixturefi'of fluoride sclts. The spectrophotometric” method .in which ““cuproine’’ (2,2 *.biquinoline) is used as_the chro- mogenic reagent, reported by Gue512 to be Specrflc-- for copper, was adapted for this determination, - - Copper is determined as the copper(l)-cupronne"- corln[:Iex fo;mtehd bl{::u;:b:utnltng an uqueous:(:)ui‘fvatt; - ‘can be separated quanhtahvely from uranium and soiution ot The ple containing coppe! ‘"M zirconium by precipitation with tannin in a slightly a solution of 002% (w/v) cuproine in n-omyl alcohol at pH 6. The copper complex is extracted in the orgdnic phase.’ The complex exhibits its ~ maximum - absorbance at 550 m.® Hydroxylammef.i_j* _' fhydrOCh'or'de was, used 1o :educe the copper. to-';—f'{ffi-??by Volatlhzanon. ‘This was accompllshed by o '/ . the monovalent state, and ‘l'crtarlc ucud was used o - '_':to complex iron cmd uranium, o L . The extraction procedure recommended by Guestz,] . , jf—_i_-was mod:fled for. upphcoflon to. the - defermmahorx' " of traces of - ‘copper in solutaons of hugh ionic. "'fir'.qistrength “The volume of the aqueous phase was- o Vmcrecsed from 30 to 100 ml*in order to ensure,::i.-f Clplfdfed wath tannm ct a pH of Sina hor solutlon. ~.an-.optimum: umount ‘of copper in_the uhquot‘: The "«WTh cioitat fr ite 1 ! and f + - .. extraction’ penod Was increased from 1. 1030 min, e pre 'p' ° e, after | s/ emotfc op igni 4on 7 ° -~ "and- 10 ml of‘an"alcoholic: solut:on of: the reagent;ff'-;.__‘r"3'" e L T 3 J. C. White, Determmauon oj Small Amounts of Tan- - 'was’ used. -Since fhe uqueous-orgamc ratio WAS e S P PR and in NaF-LzF-KF-UF4, ORNL. so iarge, 10 g “of ‘ammonium ‘sulfate was added'_'sz- {"CF -56-1-49 {Jan, 10, 1956). - ; "‘1o reduce the solublllty ‘of fhe olcoho| in fl-:e o 2R, ). Guest, Anal. Chem. 25, 1484 (1953). PERIOD ENDING JUNE 10, 1956 aqueous- volume. The ammonium sulfate served to mask the mass-action effect of ionic constituents in the sample; however, it reduced the effective pH range for quantitative extroction from a range of 4,5 to 7.5 to a range of 6.0 to 6.5. "~ “The abscrbance of the organic extract is a linear function of the concentration of copper up to 7.5 pg/ml, The coefficient of variation of the methed, ~as derived from stondard samples, is 2%. Samples of NoF-ZrF -UF‘ prepared in the copper-lined nickel reactors have been analyzed for copper by ‘this method. The range of the copper concentration ‘was | to 25 ppm, DETERMINATION OF TANTALUM IN NaF.ZrF -UF J P. Young J. R. French The determination® of trace amounts of tantalum in'NaF-LiF-KF has been extended to include the _ determination of tantalum in NaF-ZrF -UF Since z:rcomum dnd uranium interfere w;fh the determ:- . nation of tantalum by the pyrogallol method, 4 it was necessary o develop some means of removing these interferences. The separation of tantalum ,_from uranium in’ NcF LiF-KF-UF,, which was described prev:ously, was accomplished by « ;'fprec:pltahon of the tantalum with cupferron. Zir- ~conium, however, is also precipitated w:fh cup- ,ferron. According to Hillebrand et al.,® tantalum acidic soluhon of oxalate which is half-saturated ‘with ammonium chioride, Since the samples con- tained’ alkall fluorldes, it-was necessary to dis- “solve them ccrefully to prevent ‘the loss of Tqu _:_‘5_‘4'_-cureful dlgeshon of ‘the fluoride sumple in dilute 7 "sulfuric acid.in order to hydroiyze TaFg to Tc:zo_.-,.w ) 5‘The soluhon was then evoporc?ed to dryness, and the residue ‘from this’ ‘evaporation was fused: with f_"i_potcsstum pyrosulfcte. ~The product - of the fusuoni - was ‘dissolved in ammonium-oxalate and then pre- 45 |. Dinnin, Anal. Chem. 25, 1803 (1953). SW. F. Hillebrand et al., El:ed Inorganic Analyszs. : p 598, 2d ed., Wiley, New York, 1953, ] 127 ANP PROJECT PROGRESS REPORT Ta,04, was fused with potassium pyrosulfate, and the tantalum-pyrogallol color was developed, as described in the earlier report.3 The coefficient of variation was 5% for the de- termination of tantalum in NaF-ZrF ,-UF,. In the separation of tantalum from vranium and zirconium, a sample which contained at least 0,5 mg of tanta- lum was taken. DETECTION OF TRACES OF NaK IN AIR A. S. Meyer, Jr, J. P. Young The design of an instrument for the photometric detection of NaK in air was completed. A func- tional description and the operational requn‘ements of the apparatus were given previously.® The electronic components of the instrument have been assembled, and the optical system is now being fabricated. _ Preliminary engineering drawings of the instru- ment for the detection of submicrogram quantities of NaK in air by observation of the sodium reso- nance radiation have been prepared. Experiments are being carried out to devise a method for the introduction of reproducible concentrations of alkali metal oxides into air streams in order to provide synthetic samples for testing these instruments, The alkali metal is injected into the air stream in a jet of helium which is saturated with alkali metal by contact with the molten alkali metal. The temperature dependence of the vapor pressure of the molten metals provides the method for the control of the concentration of the metal in the helium. An apparatus has been assembled to test the stability of sodium aerosols that have been prepared at various mixing temperatures and various flow ratios of helium to air and to determine a quantitative method for transferring the samples of air to the detection instruments. VDET_ERMINATVION OF OXYGEN IN NaK A. S. Meyer, Jr. - G. Goldberg Since it has been established that reasonable precision can be obtained in making the determi- nation of oxygen in alkali metals with the modified Argonne distillation sampler, the sampler was connected to a loop in which NaK was being cir- culated. The loop operates at a temperature of 6a. s. Meyer et al.,, ANP Quar, Prog. Rep. March 10, 1956, ORNL-2061, p 207. 128 1200°F, with a cold trap differential of cbout 600°F. Since no inherent difficulties were en- countered in the operation of the sampler in the loop, a detailed, step-wise procedure’ was written which covers the use of the sampler with both sodium and NaK loops. The analyses made over a five-day period while the cold trap was in oper- ation gave the following concentrations of oxygen, in ppm: 1330, 1265, 575, 310, 265. Since the results obtained with the sampler appear to be satisfactory, additional tests will be conducted in which the results from a plugging indicator will be compared with the results from the sampler. Plans are also being formulated to add a known amount of oxygen to the NaK to make further comparisons of the sampler and the plugging indicator. {n the sampler now being constructed a sight tube is being placed in the distillation cham- ber so that the operations of sampling and distilla- tion may be observed., : EXAMINATION OF COLD TRAPS FROM ALKALI-METAL SYSTEMS A. S. Meyer, Jr. G. Goldberg A program was initiated to carry out the chemical examination of cold traps removed from systems circulating alkali metals, such as heat-exchanger test stands and corrosion-testing loops containing sodium or NaK. A typical request for analysis calls for the determination of oxygen, iron, chro- mium, nickel, and, in the case of NaK, a sodium- potassium ratio. Other constituents which are sometimes requested are uranium, zirconium, and beryllium. The cold traps range from 3 to 5 in, in diameter and from 2 to 6 ft in length. In order to carry out the determination of oxygen successfully, the trap is drained as completely as possible. The residual sodium, or NaK; on the Demister packing within the trap is then reacted with butyl bromide to form the neutral bromide salts of sodium ond potassium; oxides -of these elements do not react with butyl bromide. When the reaction is completed, the organic material is drained, and the trap is cut into two parts, Until the trap is cut open, o blanket of helivm is maintained in it. ' 7.}. C. White, Procedure for the Determination of Oxy- gen in Sodium and NaK by tbe Dzst:l[at:on Metbod' ORNL CF-56-4-31 {April 5, 1956). After the trap has been opened, both sections are washed carefully to dissolve the bromides and oxides, and the washings are combined and filtered. A portion of the filtrate is titrated with hydro- chloric acid, and the concentration of oxygen is calculated from this titration. Both the filtrate and the residue are also analyzed for the corrosion products and any other constituents that were called for in the request for the analysis. The results of a typical analysis are presented in Table 2.6.1. TABLE 2.6.1. ANALYSIS OF NaK DRAINED FROM THE CIRCULATING COLD TRAP OPERATED IN INTERMEDIATE HEAT-EXCHANGER TEST STAND B Determination Water-Insoluble | Total Requested Residue (g} (9) Na 564 K 200 Cr 0.6 0.8 Ni 0.9 1.2 Fe 0.1 0.1 U Not detected Zr Not detected 0, 120 DETERMINATION OF WATER IN HELIUM A. S. Meyer, Jr. G. Goldberg - Traces of moisture in the helium used as a blanket gas in engineering studies can be con-" veniently determined by measuring the dew point - of the gas. However, h|gh and ‘erratic dew-pomt : temperatures ore occasionally measured. < It was =~ " observed. that when -copper tubmg was used for the . ~ connections from the helium to the dew-point meter, dew-pomt temperafures below -145°F correspond- -ing to about 1 ppm-H O ‘were obtumed almost. ~immediately, - Conversely, periods of flushing of - as long as .15 min at a rate of about 10 f3/hr were - required - before comporuble readlngs were - obtained ‘when the gases were passed through 8 2.t length of Tygon tubing. It is therefore recom-r_r o mended that, when possible, connections to the - PERIOD ENDING JUNE 10, 1956 meter be made with metal rather than plastic tubing when dew points below -40°F are to be measured. If plastic tubing is used, the tubing should be adequately purged with test gas before measure- ments are taken, COMPATIBILITY OF BERYLLIUM WITH SOME TYPICAL ORGANIC DEGREASING AGENTS A. S. Meyer, Jr. W. J. Ross Tests have been carried out to determine the compatibility of beryllium with the organic solvents being considered as possible agents fordegreasing the machined reactor components. Suggested de- greasing procedures include 30-min flushing periods with acetone and ethanol at room temperature and trichloroethylene and perchloroethylene at their boiling points. No detectable attack on beryllium was observed even after a 24-hr period of dynamic contact with these reagents. ANP SERVICE LABORATORY W. F. Vaughan Analyses of ten samples were performed for the Wright Air Develooment Center (WADC). The de- terminations made on the WADC fused-fluoride-salt samples included total uranium, trivalent uranium, iron, nickel, and chromium. The hydrogen content of zirconium hydride liners, which had been en- cased in Inconel tubing, was also determined. The butk of the work of the service laboratory was the analysis of fused-fluoride-salt mixtures and alkali metals for the Reactor Chemistry ond Experimental Engineering Groups. A total of 1164 samples was analyzed, which involved 4168 re- ~ported results, an average of 3.6 per sample. The - backleg consists of 48 samples. ‘A breakdown of ,; ;_fhe work follows: Number of Number of - . Samples Reported Results Reacfor Chemisfry, 734 2601 ‘Expenmenfcl L 337 o 1375 VEr_lgmeerlng_' _ S WADC e 50 Miscellaneous 83 142 Tetal 1164 4168 129 Part 3 METALLURGY W. D, Manly 3.1. DYNAMIC CORROSION STUDIES J. H. DeVan FLUORIDE FUEL MIXTURES IN INCONEL FORCED-CIRCULATION LOOPS J. H. DeVun - R, S Crouse " An lnconel forced-curculahon Ioop, 7425-8, in which the cold-leg surface area was: !arger than that in ‘the standard loop, was operated to study the effect of the cold-zone area on mass transfer in the fuel mixture {No. 30) NaF-ZrF -UF (50-46-4 mole %). A tube bundle with a fop and a bottom header and five connecting co;ls of V-in.-dia tubing was used in place of a- standard-s;zed cooling coil " of 4-rn.-d|a tubing.,! ~ With this arrangement the cold-zene area was twice that of o standard loop. The volume of the cold zone was held as nearly as possible the same as that for a standard loop so that the total loop volume would be unchanged. The standord loop (7425-10) operated concurrently with this loop was designed 'G, M. Adamson and Re S¢ Crouse, ANP Q.uar. Prog. Rep. June 10, 1955, ORNL-1896, p 86, Fig-'5o2- as a control loop for this experiment and for 'several previously run experiments in which variables such as wall temperature, temperature gradient, and surface-area-to-volume ratio were studied. Both the loops, 7425-8 and 7425-10, were to operate at a maximum wall temperature of 1600°F and with the other conditions given in Table 3.1.1. However, it was found after operation that both loops had identical discrepancies in the recorders used to measure maximum wall temper- ature, ond thus they actually operated at a maxi- mum wall temperature of 1580°F. While good comparisons can be made between these two tests, the lower wall temperature’ somewhat invalidates the use of loop 7425-10 as a stundord loop for comparison with other test loops. The maximum attack in the lfoop with the larger cold-zone area, 7425-8, was to a depth of 4 mils, which compares quite closely with the attack to a depth of 4.5mils in the standard loop, 7425-10. There were no cold-leg deposits in either loop, and fuel analyses made after the tests showed TABLE 3.1.1. OPERATING CONDITIONS FOR FOUR 'INC_ONEL FORCED-CIRCULATION LOOPS THAT CIRCULATED THE FUEL MIXTURE (No. 30) NuF-ZrF‘-UF‘ (50-46-4 mole %) Loop Number : Operuti_n'g' Cofid itions - ' ' Vo : 7425-8 7425-10 7425-41 7425-43 Operating time, hr 1000 1000 1000 1000 Maximum fluoride mixture =~ 1500 1500 _ 1650 1500 tempera'ture,', °F' o - e R Flmd temperature drop, oF LT 200 200 : | " 200 - 2‘00- - | "‘,Max-mum tube wamempemm, Lm0 .?isso -'f; S 1700 | 'l":?iflevm'ds number L ilo.ooo'f - w,ooo 2950 6,500 L Fludvelecty,fos- o es . es 208 4a1 g Heared surfuce area, m.2 e "_,"262-:‘- o 1‘262'7 i 262 . 262 ~ Cooled surface area, im"’ o sa2 241 a1 U7 Raho of cooled surfoce areo to S o 4a2 71: Sl 2.0 C _A ) ) _: 20 -, s 2_0 " ’ ) L total. |oop volume o RS BRI L e S LT -"/"__\'_V"Raho of heoted surfo::e oreq tol o ].94 24 : 21 _ 2.] o . tofol loop volume L STy RTENEI S , L Maximum depth of aftuck mrls . 4 | 4.5 10 9 133 ANP PROJECT PROGRESS REPORT equivalent chromium contents, approximately 400 ppm. Thus the increase in the cold-zone surface crea produced po apparent effect on corrosion under the conditions of these tests. Some increase in the emount of mass transfer, as evidenced by increased hot-leg attack, was ‘expected with increased cold-leg surface area on - the basis of thermodynamic studies of the reactions of fluoride.salts with pure chromium and with the chromium' in Inconel (see Chap. 2.2, **Chemical Reactions in Molten Salts’'). These studies predict that mass transfer of Inconel in ZrF . ‘bearing fluoride mixtures should be rate-controlled, in part, by the amount of chromium which can diffuse from the circulated fluid into the cold leg. Such a process would be related directly to the cold-zone surface area, | Examination was completed of loop 7425-43, which was operated as a part of a series of tests to study the effect of the bulk fluoride temperature on the corrosion of Inconel. The loop circulated the fuel mixture (No. 30) NaF-ZrF 4 UF, (50-46-4 mole %), and it operated with a mammum wall temperature of 1700°F, a maximum fluid temperature of 1500°F, and a fluid temperature drop of 200°F. The results of operation of this loop are to be compared with those for loop 7425-41, which, as reported previously,? was operated with a similar maximum wall temperature and temperature drop, but with a maximum fluid temperature of 1650°F. Other conditions of operation for these two loops are compared in Table 3.1.1. The maximum attack was found in the region of maximum wall temperature in both loops. As shown in Figs. 3.1.1 and 3.1.2 the types of attack were quite similar, as were the depths of attack, being 9 mils in loop 7425-43 in which the maximum fluid temperature was 1500°F and 10 mils in loop 7425-41 in which the maximum fluid temper- ature was 1650°F. The amounts of chromium in solution in the fluoride mixtures in both loops, ‘as shown in Table 3.1.2, were also comparable, with the amount in the fluid circulated at 1500°F actually being greater. If variations in Reynolds number are neglected, as argued previously,? the insignificance of the effect of the differences in bulk fluoride mixture temperature in these two tests is apparently the result of the similarity in maximum wall temperatures {(1700°F in both loops). 2y, H. DeVan, ANP Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012, p 105. The importance of the wall temperature to.cor-. rosion in Inconel-fluoride fuel systems was discussed previously.? rFig. 3.1.1. Region of Maximum Attack in Inconel Forced-Circulation Loop 7425-43 Which Circulated the Fuel Mixture (No. 30) NaF-ZrF ,-UF, (50-46-4 mole %) for 1000 hr with a Maximum Wall Tempera- ture of 1700°F and a Maximum Fluid Temperature of 1500°F. Etched with modified aqua regia. 250X, Reduced 32.5%. {Secret—with—caption) Fig.-3.'|.2. Region of Maximum Attack in Inconel Forced-Circulation Loop 7425-41 Which Circulated. the Fuel Mixture (No. 30) NaF-Z¢F 4-UI"' 4 (50-46-4 - mole %) for 1000 hr with @ Maximum Wall Tempera- " ture of 1700°F and a Maximum Fluid Temperature of 1650°F. Etched with modified aqua regm. 250X Reduced 32.5%. (Seeretwittrroption) PERIOD ENDING JUNE 10, 1956 TABLE 3.1.2. URANIUM AND IMPURITY ANALYSES OF FUEL MIXTURE (No. 30) NaF-ZrF -UF, (50-46-4 mole %) AFTER CIRCULATION IN INCONEL LOOPS 7425-41 AND 7425-43 Loop No. = - Uranium Content Impurities Found (ppm) Sample-Token o : (wt %) ‘Ni Cr : . Fe | 742541 Before filling 8.61 0 105 45 - After draining 10.3* 40 440 70 7425-43 Before filling 9,20 50 55 70 After draining 9.37 100 . 650 80 ‘*This value is in doubt, ~ SODIUM AND NaK IN INCONEL FORCED-CIRCULATION LOOPS J. H. DeVan R. S, Crouse ‘An Inconel forced-circulation loop completed 1000 hr of operation with sodium which had been specially treated to removed oxide contamination prior to testing. During this test the maximum fluid temperature was 1500°F, and the temper- ature drop was 300°F. Prior to the actual test, ‘sodium was circulated through the loop at’ 1500°F under isothermal conditions to remove oxide films from both the hot and cold legs of the loop. This sodium was then dumped at 1500°F and replaced with a charge that had been cold trapped -and - filtered in a service loop operated at 300°F. A circulatory cold trap maintained at 300°F was also used during the test, trap temperature, An evaluation of mass. transfer in this loop showed that the weight of - deposit was equivalent to that found in loops - operated with normal sodium either wflh or wufhoutf": tnrculafory cold traps. - ~-An experiment .to evaluate 1he effect of a lower' ' cold trap . operating temperature, in -which NaK. - " rather “than sodium- was used, likewise failed _ to show a reduction in mass transfer, Loop 74393, - also constructed ‘of Inconel, was operuted 1000 hr = with the cold trap temperafure held to the minimum " for which Flow through. the trap could be maintained, - o npproxmately 100°F. ~The operahng “conditions, - except for.cold trup temperofure, were identical . to those for the ubove—rnenhoned loops. A visual * * . examination of the loop failed to show. a reduction i deposifs in_comparison ‘with the deposnfs found ~in loop 7439-1, described previously,3 which also circulated NaK and operated with a 300°F cold- SODIUM-BERYLLIUM-INCONEL COMPATIBILITY IN DYNAMIC SYSTEMS J. .H. DeVan R. S, Crouse Previous compatibility studies of Inconel- beryllium-sodium systems, conducted by inserting small beryllium tubes in the hot legs of toroid, thermal-convection, or forced-circulation loops, showed little effect on temperature-gradient mass transfer at temperatures up to 1300°F that could be attributed to the presence of the beryllium. To determine whether increasing the beryllium surface area relative to the surface area of the Inconel would affect the compatibility, two forced-circu- ~lation loops, in which equivalent surface areas of beryllium and Inconel were in contact with the flowing sodium, were operated at 1250°F. Only _one of the loops had an oxide cold trap. The beryllium, in the form of rectangular blocks with - drilled - holes, was canned in Inconel and - inserted,_inrthe hottest section of the loop; the _remaining sections of the loop were constructed of Inconel, ~ A temperature drop of 300°F was maintained between the hot. ‘and cold fluid temper- ' atures in each. test. . After 1000 hr of operation, ‘no increase in the amount of mass transfer was seen in either of these loops as compared with the ‘mass transfer found in similar loops with smaller beryllium inserts, ~ The mass-transferred deposits found were not sufficient in quantity to permit “chemical analysus. They were similar in amount and appearance to deposits seen in Incone! loops ‘without beryllium inserts. The addition of the cold trap had little effect on the test results, 3J0 He DeVan, E. A. Kovacevich, and R« S. Crouse, ANP_ Quar. Prog. Rep, March 10, 1956, ORNL-ZOG'I, ' p 117, 135 ANP PROJECT PROGRESS REPORT THERMAL-CONVECTION LOOP TESTS OF INCONEL CASTINGS - 'J.V H.VDeVon , E. A. Kovacevich In order to evaluate the use of Inconel castings for large, intricate sections required in the ART, three standard Inconel thermal-convection loops with cast Inconel inserts in the hot legs were operated for 500 hr. The castings, whose compo- sitions are given in Table 3.1.3, contained approxi- mately 1.2% manganese, 2% niobium, and 1 to 2% silicon, Inserts 321 and 322, which had the lowest and the highest silicon contents, re- spectively, were tested in loops which circulated the fuel mixture {No. 30) NaF-ZrF 4UF, (50-46-4 mole %) at 1500°F, while the remammg insert, - 323, with an intermediate silicon content, was _tested in a loop which circulated sodium at 1500°F. As shown in Table 3,1.4, very severe corrosion: of the cast specimens was found in the loops which circulated the fuel mixture, The attack occurred not only in the form of the subsurface voids that are typical of the attack of wrought Inconel, but, in addition, very deep intergranular penetrations appeared, which reached in the worst case to a depth of 70 mils, as shown in Fig. 3.1.3, These intergranular penetrations, although a result of rapid attack of grain-boundary con- stituents, were aided considerably by the porosity and the shrinkage cracks present in all the castings in the as-received condition, as shown in Fig. 3.1.4. The casting containing 1% silicon was found to be TABLE 3.1.3. CHEMICAL ANALYSES OF CAST INCONEL INSERTS TESTED IN STANDARD INCONEL THERMAL-CONVECTION LOOPS Cast inconel Insert Chemical Analyses (wt %) No. Ni* Cu Fe Si Mn Cc Cr Nb -8 3N 70.49 0.01 8.20 1.04 1.22 0.23 16.67 2.08 0.005 322 69.79 0.03 8.20 1.93 1.17 0.22 16.51 2.09 0.006 323 70.44 0.02 8.10 1.34 1.6 0.22 16.67 2,00 0.004 *Obtained by difference analysis. TABLE 3.1.4. RESULTS OF METALLOGRAPHIC EXAMINATION OF INCONEL THERMAL-CONVECTION . LOOPS OPERATED WITH CAST INCONEL INSERTS IN THE HOT LEGS Operating time: 500 hr Maximum fluid temperature: 1500°F - Loop ' . Insert " Metallographic Results Circulated Fluid No, ~ Ne. Hot-Leg Attack Cold-Leg Attack 876 o 321 NoF-ZrF4* Cast section, 25 mils Light surface roughening v\lrifh, 7 (50-46-4 mole %). Weld interface, 70 mils a metal deposit present 877 322 NaF-ZrF4-UF4* Cast section, 23 mils Light surface roughening with - (50-46-4 mole %) Weld interface, 25 mils evidence of metal cry;fol; | 878 Control NaF-ZrF -UF ,* 10 mils | " No attack ' (50-46~4 mole %) ' 879 323 Sodium No attack No attack; no deposifs *Fyel mixture No. 30. 136 o/ Fig. 3.1.3. Region of Maximum Attack Near Casting-Weld Interface in-Inconel Casting No. 321 Exposed for 500 hr to the Fuel Mixture (No. 30) NaF.-ZrF -UF4 (50-46-4 mole %) at 1500°F as an Insert in the Hot Leg of a Wrought Inconel Thermal- Convection Loop. Etched with modified aqua regia. 100X. Reduced 32.5%. {Seeret-witireoptien) Fig. 3."-l.'4._7,_:'Pofes'ity in As Recewed |ncene|:- Casfing ,N-°' .321. IOOX » Reduced 32.5%_. : no. befler from the corrosion stondpomt than the.i' ' ;,one contammg 2% suhcon. o S e " Metal depos:ts were “seen metallographlcally o und wsually ‘in the hot “legs of both “loops. 'Spectrogrcphlcally, the . deposits were found to be :predommantly nickel, with aluminum, ‘chromium, “and “iron reported as .major elements. The silicon and mangonese contents™ in’ these deposns were low, Loop 879, which contalned insert 323 and circu- lated sodium, revealed very little corrosion PERIOD ENDING JUNE 10, 1956 (<1 mil) in the insert section. Metallic deposits, which were found by analysis to be predominantly nickel, were visible in the trap area. The amount of deposited material was approximately the same -as that normally found in wrought Inconel loops. It is apparent that Inconel castings of the compo- sitions tested are not svitable for use in contact with the fuel mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %). While the castings showed resistance to sodium, further tests of longer duration will be required to completely evaluate mass-transfer effects. FLUORIDE FUEL MIXTURES IN HASTELLOY THERMAL-CONVYECTION LOOPS J. H. DeVan E. A. Kovacevich Three Hastelloy X (ref. 4) thermal-convection loops, which were constructed from. ¥.in.~dia tubing, were operated 1000 hr, two with sodium at 1500°F and one with the fuel mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %) at 1500°F. Loops 855 and 856, which operated with sodium, had no hot-leg attack and no evidences of metallic deposits in the cold leg. Loop 857, however, which circulated the fuel mixture, showed con- siderable hot-leg ottack to a depth of 27 mils. The cold leg of this loop was attacked to a depth of 1 mil, and there were evidences of metallic crystals and a metallic layer, as shown in Fig. 3.1.5. The metallic crystals in the trap area were analyzed and found to be predomlncntly chromium, Two loops, 872 and 873, constructed of 3/-|n.-d|c ~ Hastelloy W (ref. 5) tubing were also operofed for 1000 hr with sodium and with the fuel mixture (No. 30) NqF-ZrF‘_-UF‘ (50-46-4 mole %), re- - spectively, . at 1500°F, - hot-leg attack; however, loop 872, which operated Neither loop showed with sodium, had scattered metallic crystals in _the cold leg. The hot-leg samples of both loops - were similar metallographically to the as-received samples of each loop. The significant difference in attack on the two alloys, Hastelloys X ‘and W, by the fuel mixture can undoubtedly be explained by the difference ~in the chromium content of the two alloys, 22 4The nominal . composition .of Hastelloy X is 22% Cr-23% Fe~9% Mo-1.5% Co~balance Ni. 5The nominal composition of Hastelloy W is 5% Cr—5% Fe-24% Mo=1% Co-balance Ni. 137 ANP PROJECT PROGRESS REPORT and 5%, respectively, As may be seen in Table 3.1.5, the chromium content of the fuel mixture circulated in the Hastelloy X loop was much higher than that of the fuel mixture circulated in the Hastelloy W loop., This reflects the greater tendency toward chromium = removal -and the resultant void formation in alloys with high chromium contents, such as Hastelloy X. FLUORIDE FUEL MIXTURES IN MONEL " THERMAL-CONVECTION LOOPS J. H. DeVan Three monel thermal-convection loops, 806, 808, and 809, completed 500, 545, and 1027 hr of operation, respectively, with the fuel mixture B URCLASSIFIED B y.95a7 BRI mc{m : B 3L T L Fig. 3.1.5. Metallic Crystals ond Loyer De- posited in Cold Leg of Hastelloy X Thermal- Convection Loop Which Circulcted the Fuel Mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %) for 1000 hr at o Hot-Leg Tempercture of 1500°F. Etched with HCI-H,CrO,. 250X, Reduced 32.5%. s ” or) {(No. 107) NaF-LiF-KF-UF, (11.2-41-45.3-2.5 mole %) circulating at a hot-leg temperature of 1500°F. Loops 808 and 809, which were scheduled to operate 1000 and 1500 hr, respectively, were terminated before completion of the scheduled operating period because of excessive oxidation of the outside surface of the monel tubing. The oxidation did not completely penetrate the tube wall in either loop. Metallographic examination revealed hot-leg attack to a depth of 1 mil, as shown in Fig. 3.1.6, in all these loops. Three other monel loops, 880, 881, and 882, that had been chromium-plated to provide oxidation pro- tection were operated under the same conditions, but, apparently, the platings were imperfect, and two of these three loops, 881 and 882, were Fig. 3.1.6. Region of Maximum Attack of Hot Leg of Monel Thermal-Convection Loop 809 Which Circulated the Fuel Mixture (No. 107) NaF-LiF. KF-UF, (11 .2-41-45.3-2.5 mole %) for 1027 hr at a Hot-Leg Temperature of 1500°F. Etched with HNO,-acetic acid. 250X. Reduced 32.5%. Cocre- ir e eertion) TABLE 3.1.5.- CHEMICAL ANALYSES OF THE FUEL MIXTURE (No. 30) NaF-ZrF -UF4 (50-46-4 mole %) BEFORE AND AFTER CIRCULATION FOR 1000 hr AT 1500°F IN HASTELLOY X AND W THERMAL-CONVECTION LOOPS Loop . Loop Uranium Centent Impurities Found (ppm) Sample Taken - No, Material (wt %) Ni - Cr Fe 857 "~ Hastelloy X Before filling 8.39 260 135 115 After draining 8.68 40 1060 125 873 - Hastelloy W Before filling 8.78 80 60 75 | After draining 8.82 - 85 275 110 138 %) terminated after 1217 and 1339 hr of the 1500 hr scheduled because ' of excessive oxidation and resultont leaks in the heated areas . of the loops, The ! remcmmg Ioop, 880, completec} its scheduled 1000 hr. of operaticn, - Metallographic - unalyses of these Ioops hcve not been completed SPECIAL FLUORIDE FUEL MIXTURES IN INCONEL THERMAL-CONVECTION LOOPS J. H. DeVan - Additional Inconel thermal-convection loops have been operated to evaluate corrosion properties PERIOD ENDING JUNE 10, 1956 of. several special fluoride fuel mixtures, . Tests of 500 hr ‘duration were completed for mixtures. in the MF-ZrF -UF , system, where MF is KF, LiF, NaF, or RbF M:xtures with each of these fluorides and 40 or 46 mole % ZrF , have been tested; each mixture contained 4 mole % UF,. The results obtained for the fuels containing 46 mole % ZrF were reported previously,3 but they are repeated ln Table 3.1.6 for comparison, Lowering the ZrF4 content, as shown in Toble 3.1.6, had no effect on attack by the NaF-containing mixture, but the attack by the KF- and RbF-containing mixtures increased from 1 to 2 mils, The LiF-containing TABLE 3. 6o RESULTS OF METALLOGRAPHIC EXAMINATIONS OF INCONEL THERMAL-CONVECTION LOOPS OPERATED WITH SPECIAL FLUORIDE FUEL MIXTURES AT A HOT-LEG TEMPERATURE OF 1500°F Metallographic Results Loop Fue- 'Mtx‘tur'e“-. Fuel Mixture Comfiosifion OP_::::ng Maximum Hot-Leg . No. Ceode Designation (mole %) (hr) Attack | Cold-Leg Deposits (mils) 845 94 KF-ZrF -UF ;; 50-46-4 500 65 Metallic layer 846 94 KF-ZeF ~UF ,; 50-46-4 500 8 Metallic layer 883 94 KF-ZrF ~UF ; 50-46-4 1500 n Metallic layer 884 94 KF-ZrF -UF ,; 50-46-4 1500 9 Metallic layer 930 WR4 KF-ZrF ~UF ; 56-40-4 500 7 Metallic layer 931 WR4 KF-ZrF 4" UF i 56-40-4 500 , 9 Metallic layer 847 93 _ LiF-ZrF -UF‘; 50-46-4 500 17.5 Possible crystals 848 93 . LIF-ZF, UFG 50—46-4’ 500 19 ‘Possible crystals s WR3 LiF-ZrF CYFG S6-40-4 500 10 Metallic layer and crystals 7 929 WRS LIF-ZiF-UF i 56.40.475 00 10 Metallic layer and crystals ey 95 inF-ZrF‘-Ua-. S0464 . S0 9 Metaltic layer S 840 95 _'fi_RbF-ZrF -UF4;-50-45-4— 8500 - . 9 Metallic lgy"ejl"’ - . 932 . WRS | RI:F-ZrF -ur=4. 56+40-4 < 500 - 10 Metallic layer and crystals 933 WRS RbF-IeF (UFi56-404 500 - 1 Metallic layer and crystals 030 NeRZFaUR s0464 500 105 - Neme 8130 NeF-ZrF CUFiS0-d64- SO0 9 Nome o 926 - WRZ NgE-ZrF_ ~UF i . '40-'4._'*7?".7 500 10 'Metallic layer ond crystuls C 927 - WR2 - NaF-ZrF -UF ; 56-40-4 .. 500 7 Metallic layer and crystals 139 ANP PROJECT PROGRESS REPORT mixture showed quite anomalous behavior in that the attack decreased markedly with the decrease in ZrF ;. These results are being rechecked. "The lbobs' containing r_n'ixtures with 56 mole % of the alkali-metal fluoride, with the exception of the loop operated with the KF-containing mixture, ‘showed traces of metallic crystals in the cold legs. These crystals were most pronounced in the RbF-containing mixturés. Specimens of the crystals found in loop 933, which circulated the fuel mixture RbF-ZrF ,-UF , (56-40-4 mole %), are shown in Fig. 3.1.7. " Two additional Inconel loops, 883 and 884, were operdied under similar conditions to obtain further corrosion data for the KF-ZrF .UF, (50-46-4 mole %) system. An increase in the operating time from 500 to 1500 hr caused an increase in maximum attack of 2 mils. Thin ‘metallic layers were noted in the cold legs after both operating periods. As yet these layers are unidentified, 140 Fig. 3.1.7. Metallic Crystals and Layer De- posited in Cold Leg of Inconel Thermal-Convection Leop 933 Which Circulated the Fuel Mixture (No. WR5) RbF-ZeF, -UF, (56-40-4 mole %) for 500 hr ot a Hot-Leg Temperature of 1500°F, Etched with modified aqua regia. 250X. Reduced 32.5%. (@Eecret-with-capiien; PERIOD ENDING JUNE 10, 1956 3.2. GENERAL CORROSION STUDI ES . E. E. Hoffman BRAZING ALLOYS IN L1QUID METALS AND IN FLUORIDE FUEL MIXTURES D H Jonsen Two Inconel thermal-convection loops with in- serts brazed with 70% Ni-13% Ge-11% Cr~6% Si brazing alloy in the hot-leg section were operated for 500 hr. circulated, and in the other the fuel mixture (No. 44) NoF-ZrF -UF, (53.5-40-6.5 mole %) was cire culated., The hot-leg sections of these loops consisted of seven Inconel segments brazed with the alloy being tested, This type of thermal- convection joop test assembly was illustrated previously.! The hot legs of both loops were !D. H. Jansen, ANP Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012, p 113, Fig. 5.9. In one loop NaK .(56-44 wt %) was maintained at a temperature of 1500°F for the duration of the tests. The cold-leg temperature of the loop that circulated NaK was maintained at 1165°F, and the cold-leg temperature of the loop that circulated the fuel mixture was held at 1150°F. The inner walls of the Inconel segments and two samples from each brazed joint were examined after the tests for evidence of attack. The results of the examinations are summarized in Table 3.2.1. The attack of the fuel mixture on the brazed joints averaged about 5 mils. Typical samples are shown in Fig. 3.2,1. The ottack of the fuel mixture on the Inconel tubing adjacent to the joint ranged from 4.5 mils in the coolest (1450°F) portion of the hot leg to 9 mils in the hottest (1500°F) portion of the hot leg, as shown in Fig. 3.2.1. A slight trace of metallic crystals TABLE 3.2.1. RESULTS OF THERMAL-CONVECTION LOOP TESTS OF INCONEL SEGMENTS BRAZED WITH 70% Ni-13% Ge~11% Cr=6% St BRAZING ALLOY Operating time: 500 hr Brazed Joint Metallographic Notes No. ' Effect of Fuel Mixture Effect of NaK 1* Brazing alloy attacked to o depth of 4.5 mils; adjacent Inconel tube attacked to a maxi- . mum depth of 9 mils 2 Brazing alloy attacked to a depth of 4.5 mils; No attack apparent; large crack through brazing alloy to center of tube wall Attack to o deprth of 0.5 mil spofly, nonunlforrn attack on tube wall to a T maximum depth of 7 mils 3 - _’ Brczlng alloy aflacked to a de'fifh of 6 mils; No attack; large crack, as in joint 1 above - spotfy oflack on tube wu" fo a depth of 4 ' i-:-_:i'--Brazing al loy uttacked to a depfh of 3 mils; Attaék to a depth of 0.5 mil; large crack in - the wall uflacked to a depth of 4 5 mils 5 T Brozing alloy afi'acked to @ depfh of 5 mlls, - - tube woll attacked ‘Io a depth of 4. 5 mils . - 6** Brazing ulloy aflocked to a depth of 5 mils; ' o tube wull oflacked toa depth of 4.5 mils brazing alloy ~ Large crack in brazing allay Attack to'ordepfh'of 1.5 mils; some porosity in - brazing alloy; no large cracks *Brazed joint No. 1 was located at the hottest (1500°F) portion of the hot leg. ** Joint No. 6 was located at the coolest (1450°F) portion of the hot leg. 141 ANP PROJECT PROGRESS REPORT Fig. 3.2.1. Specimens of Inconel Joints Brazed with the 70% Ni~13% Ge-11% Cr—6% Si Brazing Alloy and Exposed for 500 hr to the Fuel Mixture (No. 44) NaF-ZrF ,-UF, (53.5-40-6.5 mole %) in the Hot Leg of an Inconel Thermal-Convection Loop. (@) Joint located in coolest (1450°F) portion of hot leg. (b) Joint located in hottest (1500°F) portion of hot leg. Etched with 10% oxalic acid. 150X. Reduced 19%. {Seerotwith caption) was found in the cold leg of the loop that circu- lated the fuel mixture. Spectrographic analyses of these crystals showed strong lines for chromium and nickel and weak lines for iron. There was no evidence of mass transfer in the loop that circulated NaK. The brazing alloy showed good corrosion resist- ance to NaK, as shown in Fig. 3.2.2, but large cracks which extended to a depth of one-half the tube wall thickness were found in some of the brazed joints, Fig. 3.2,3. It is not known whether these cracks were caused by thermal stresses or shrinkage. Anclysis of the NaK after the test showed 1160 ppm of oxygen. This is admittedly high, but no efforts were made to purify the NaK which was received in container lots. A third Inconel thermal-convection loop was operated for 500 hr in which the hot leg contained Inconel inserts brazed with the 82% Au-18% Ni brazing alloy. The fluid circulated was the fuel mixture (No. 30} NaF-ZrF -UF, (50-46-4 mole %). The hot and cold legs were maintained at tempera- tures of 1500 and 1175°F, respectively. Micro- 142 scopic examination of samples from each brazed joint showed the attack on the alloy to average about 8 mils, with a maximum depth of 10 mils. Tests were also conducted on a series of buttons of Coast Metals brazing alloy No. 52 (89% Ni- 5% Si—-4% B-2% Fe), which were exposed in see- saw apparatus to NaK (56-44 wt %) and to the fuel mixture (No. 44) NaF-ZrF-UF, (53.5-40-6.5 mole %). Since previous tests of this alloy showed depletion of the second phase at the exposed edge,?2 these tests were conducted to determine whether the high-cross-section boron component was removed and, if so, whether the removal was time dependent. The buttons were contained in nickel tubes with hot-zone temperatures of 1500°F for both tests and with cold-zone temperatures of 1100°F in the tests with NaK and 1200°F in the tests with the fuel mixture. In all the tests, the specimens were retained in the hot zone of the test capsule. The duration of the exposures to 2p. H. Jansen, ANP Quar, Prog. Rep. Dec. 10, 1953, ORNL-2012, p 119, Fig. 5.17. : NaK were 100 and 350 hr, and the exposures to the fuel mixture were of 100 and 500 hr duration. The results of the tests are presénted in Table 3.2.2. Boron was found only in very small quan- tities .in the areas depleted of the second phase by exposure to NaK and to the fuel mixture. A microspark traverse on the sample exposed to the. PERIOD ENDING JUNE 10, 1956 fuel mixture for 500 hr showed that the concen- tration of boron was less than 1% from the surface to a depth’of 8 mils. The boron content then rose sharply to its normal value (4%) and stayed there for the remainder of the traverse. Samples of the depleted area, obtained by microdrilling, were analyzed and were found to contain 0.6% boron. Fig. 3.2.2. Specimens of Inconel Joints Brazed with the 70% Ni~13% Ge~11% Cr-6% Si Brazing Alloy and Exposed for 500 hr to NaK at 1500°F in the Hot Leg of an Inconel Thermal-Convection Loop. These two specimens, No. 2 and No. 6, illustrations (z) and (), respectively, were the only samples examined that d:d not have iarge cracks in the brazmg alloy. Etched wuth 10% oxul:c acid. 150X. Reduced 8%. TABLE 3.2.2. RESULTS OF SEESAW CORROS!ON TESTS OF COAST METALS BRAZING ALLOY No. 52 (89% NI-S% Si-4% B-2% Fe) L 7 Durarticzfi"cf_” Bath . Test Weight Loss of Depth of Edge Depleted _ Specimen of Second Phase Ahr) (- {mils) CONaK (S6-44wt®) 100 0.07 X ‘ B - 350 0.23 4 NuF-ZrF -UF o 1000 0.06 - 3 - (53. 5-40-6 5 mole %) ' ' U 500 0.34 | 6 143 ANP PROJECT PROGRESS REPORT It was also found that the silicon content had dropped to about one-third its normal value (5%) in the depleted area. Evidence of the time de- pendence of the amount of depletion is presented in Fig. 3.2.4. None of the buttons tested showed attack to a depth of more than 0.5 mil. " Hardness measurements on the interior of the Fig. 3.2.3. Typical Crack Found in an Inconel Joint Brazed with the 70% Ni=13% Ge=11% Cr~£% Si Brazing Alloy After Exposure for 500 hr to NaK at 1500°F in an Inconel Thermal-Convection Loop. Etched with 10% oxalic acid. 100X. Reduced 29%. button exposed to NaK for 350 hr gave a value of 716 DPH, while measurements at the edge gave a value of 145 DPH, as shown in Fig. 3.2.5. The buttons tested in the fuel mixture gave similar . hardness traverse results. Chemical analysis of the fuel mixture from the 500-hr test and of the NaK used in the 100-hr test showed significant concentrations of boron. NIOBIUM IN STATIC SODIU_M D. H. Jansen Specimens of niobium were tested in . static sodium at 1500°F for a period of 1000 hr to de- termine the suitability of liquid sodium as a pro- tective environment for niobium during high- temperature mechanical property tests. The fests were designed to show whether the niobium would be corroded by the sodium and whether the hard- ness of the niobium would be appreciably altered through the pickup of oxygen from the sodium. The specimens were contained in type 304 stain- less steel and Inconel capsules, and the varicbles such as the volume of the bath, the container size, and the area of the specimen were adjusted to obtain o surface-area-to-volume ratio that would be as close as possible to that found in creep-test equipment. Cold traps were utilized on the bottom of each capsule to reduce the amount of oxygen in the sodium bath. The niobium specimen tested in the type 304 stainless steel capsule showed more evidence of surface roughening than did the specimen tested in the Inconel capsule, Fig. 3.2.6, and the thickness loss was also more than that of the specimen tested in the Inconel capsule, Table 3,2.3. A thin, brittle layer, approximately TABLE 3.2,3. RESULTS OF TESTS OF NIOBIUM IN STATIC SODIUM Exposure time: 1000 hr Temperature of sodium: 1500°F Hardness Thickness Loss Impurity Analysis of Sample (ppm) Sample (VHN) (mils) | " mils H, o, N,y Cc As-received material 130.5 2.4 40 125 300 Specimen tested In type 304 135.0 3.0 3.8 84 250 370 stainless steel tube Specimen tested in Incone! 116.4 1.8 3.4 90 92 1180 tube 144 PERIOD ENDING JUNE 10, 1956 Fig. 3.2.4. Specimens of Coast Metals‘ Brazing Alloy No. 52 (89% Ni-5% Si—4% B=2% Fe) After Ex- posure in Seesaw Apparatus to the Fuel Mixture (No. 44) NaF-ZrF -UF, (53 5.-40-6.5 mole %) for (a) 100 hr and (b) 500 hr. 150)( Reduced 12. 5%._ (See-pat—m-t-h—euphun-}- - Fig. 3.2.5. Specimen of Coast Metals Brazing Alloy No: 52 (89% Ni—5% S$i-4% B—2% Fe) After Ex- posure in Seesaw Apparatus to NaK for 350 he. acid. 150X. Note differences in hardness. Etched with 10% oxalic 145 ANP PROJECT PROGRESS REPORT l UNCLASSIFLED | S Y 18196 Fig. 3.2.6. Specimens of Nicbium After Exposure to Static Sodium at 1500°F in (2) an Inconel Capsule and (b) a Type 304 Stainless Steel Capsule. Etched with 25% HF-25% H,SO,-50% H,0. 250X. Re- duced 17%. 0.3 mil thick, formed on the surface of both speci- mens during the test, This layer, which is very hard (~1180 VHN), is probably a niobium-nickel alloy resulting from dissimilar-metal transfer, but it ‘has not yet been positively identified. Vickers hardness traverses were made on the as-received and as-tested specimens, The hard- ness of the specimen tested in the type 304 stain- less steel capsule increased (130 to 135 VHN) during the test, whereas the hardness of the speci- men tested in the Inconel capsule decreased (130 to 116 VHN). A correlation of the hardness values with the impurity analyses indicates that nitrogen is the major hardening agent, A niobium specimen previously tested in argon at 1500°F for 2000 hr showed a hardness increase from 127 to 168 YHN., The impurities (H,0, Hy, N 50 and 02) of the argon used in the test totaled 130 ppm. THERMENOL IN STATIC SODIUM D, H. Jansen Samples of Thermenol (82% Fe-15% Al-3% Mo) cut from a piece of hot-rolled strip were corrosion tested in static sodium for 100 hr at 1500°F. The specimens were contained in AIS|I 1035 steel and type 430 stainless steel capsules. The specimen tested in the AlSI 1035 steel capsule showed a 146 weight loss of 0.22%, and the specimen tested in the stainless steel capsule showed a weight loss of 0.07%. A slight roughening of the surface occurred on both samples, Fig. 3.2.7. Neither the type 430 stainless steel nor the-AlS| 1035 steel capsule was attacked during the tests. STATIC TESTS OF ALFENOL E. E. Hoffman Specnmens of Alfenol (84% Fe-16% AI), sub- mitted by The Glenn L. Martin Co., were tested under static conditions at 1500°F in the fuel mixture (No. 44) NaF-ZrF «UF, (53, 5-40-6.5 mole %), lead, lithium, and sodlum for 100 hr. Since no Alfenol container tubes were available, Inconel wos used as the container material for all the tests, The Alfenol specimens vused in these tests were 4 -in. cubes, The specimens were placed in l/-m.-OD '0.035-in.-wall Inconel tubes, together wnth sufficient test medium to give 3 in, of liquid bath at the test temperature. The specimen tested in the fuel mixture was covered with black crystals and showed o 30% weight increase. The crystals were analyzed and found to be UF,. The specimen was attacked throughout its thickness along the grain boundaries, Fig. 3.2.8. A PERIOD ENDING JUNE 10, 1956 Flg. 3. 2 7. Specsmens of Thermenol (82% Fe-15% Al-3% Mo) After Exposure to Static Sodium at : 1500°F in (a) a Type 430 Stainless Steel Capsule and (5) an AISI 1035 Steel Capsule. Etched with aqua regia. 250X. Reduced 16%. | Fig. 3 278 Spreclmens of Alfenol (84% Fe-]6% Al) ln" (a) As-Recewed Condmon and (b) After Ex- posure for 100 hr to the Static Fuel Mixture (No. 44) NuF-ZrF UF (53.5-40.6.5 mole %) ot 1500°F. The - grain-boundary attack shown in (b} is typical of the attack found throughout the specimen. (a) Etched L/ with aqua regia. (b} Unetched. 250X. Reduced 13.5%. dSecrétwith caption)m 147 ANP PROJECT PROGRESS REFPORT No weight change data were taken on the speci- men tested in lead becouse porticles of lead ad- hered to the specimen. Metallographic examination revealed attack only in a few scattered areas to a depth of 0.5 mil. The specimen tested in lithium was attacked throughout its thickness along the grain boundaries, which- were apparently aluminum-rich regions. Gentle tapping caused the specimen to break up -inte individual grains, Fig. 3.2.9. The lithium bath was found to contain 0.2 wt % aluminum after the test, The specimen tested in sodium lost only 16 mg (0.009 wt %) during the test and metallographic examination revealed no attack. The sodium bath was found to contain 0.001 wt % aluminum after the test. SODIUM-BERYLLIUM-INCONEL COMPATIBILITY ' IN STATIC SYSTEMS E. E. Hoffman A test was conducted in order to determine the thickness of Inconel that would be consumed by an alloying reaction between Inconel and beryllium in direct contact under: pressure while immersed in sodium at 1300°F. The apparatus used for the tests is shown in Figs., 3.2,10 and 3.2.11. The surfaces of two of the Inconel specimens used in this test, as shown in Fig. 3.2.12, were chro- mium plated to study the effect of chromium in reducing the extent of the alloy formation between nickel (from the Inconel) and beryllium. The test specimens were ‘/4 X yz x 1 in. with the Pz % 1-in. surfaces in contact. Sufficient load was applied through a compression rod and bellows to yield e 500-psi. stress on the specimens, The test assembly was loaded with sodium and held at 1300°F for 1000 hr. The load was applied to the specimens when the test temperature was reached. It was found after the test that the chromiume- plated Inconel specimens could be separated from the adjacent beryllium specimens; however, it was impossible to separate the unplated Inconel speci. mens from the beryllium specimens. The results of metallographic examination of the four specimen interfaces, presented in Table 3.2.4, indicate that Fig. 3.2.9.. Alfenol (84% Fe-|6% Al) Cube That Dlsmtegrafed After 100 hr of Exposure to Static Lithium ot 1500°F. 12X 148 N ~ of BeNi adjacent to the Inconel. UNCLASSIFIED Y2952 LOAD Fig. 3.2. 10., Apparutus for Studymg the Extent of A[ioynng Between Beryllium and Various Meiqls - Under Sh'ess Wl'ule lmmersed in Molten Sodium. a thin’ chrommm p!ute on. lnconel ‘does “not ‘elimis" nate alloying - ‘reactions ‘with beryllium when the two “materials are in contact while lmmersed in hsgh-femperature sodium, but the plate does sub- stantially reduce the extent of alloy formation. Approxlmately 4 mils of Inconel ‘was consumed in the formuhon of the 24.mil nickel-beryllium 'clloy layer which was found where the Inconel and - the beryillum were in direct contact (Fig. 3.2.132). The major porhon of reaction layer was found to be Be,,Nig, with a small percentage The reduction in alloy formation brought about by the 2-mil chro- mium plate is illustrated in Fig. 3.2.13b, A second test is now under way in which the effectiveness of 4- and 6-mil chromium platings will be evaluated. PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED Y1 «17961 Fig. 3.2.11. Enlarged View of Test Specimens ' Shown in Figs. 3.2.10 and 3.2.12. UNCLASSIFIED Y.13508 INCONEL bqRécr-bomAcT BERYLLIUM DIRECT CONTACT INCONEL A-mil CHROMIUM PLATE eERVLLIM 5 nil CHROMIUM PLATE JINCONEL Sodium-Beryllium-Inconel Com- 3.2.12. patibility Test Specimens After Exposure fto Sodium at 1300°F for 1000 h:. Fig. 149 ANP PROJECT PROGRESS REPORT TABLE 3.2.4, RESULTS OF METALLOGRAPHIC EXAMINATION OF THE INTERFACES OF THE SPECIMENS SHOWN IN FIG, 3,213 ~ Test environment: sodium Test duration: 1000 hr Test temperature: 1300°F Stress on specimens: 500 psi Interfccé o ' | Metallographic Notes - A-B:* Inconel vs beryllium - dlrect 10 to 24 mils of alloy formation (Be,Nij plus BeNi) along ‘contact . interface; 4 mils of Inconel consumed in the production of reaction layer B-C: Beryllium vs Inconel = direct 22 to 24 mils of uniform alloy formation along interface contact | C-D: Incone!l with 1-mil chromium 3.5 to 7 mils of alloy formation; 7-mil layer found where plate vs beryllivm plating was thinnest =~ D-E: Berylllum_ vs 2-ml| chromium 2 mils of interaction between specimens along 90% of inter- plate on Inconel ' _ face; 6-mil layer detected at one area where plating appeared to have been defective *See Fig. 3.2.12 for location of interface. FEERE ¥ mcsae:a; BITTTELTTTEL v 1 fi_i; ETT l L T 50 T 1 4 Fig. 3.2.13. Comparison of Results of Direct Contact of Beryllium and Inconel () with Results of Contact Between Beryllium and 2-mil Chromium Plate on Inconel (b). ‘Areas designated A, B, and E may be located on Fig. 3.2.12. (@) As polished. 150X. Reduced 15%. (&) Etched with oqua regia. 1000X. Reduced 15%. 150 - STATIC TESTS_.OF_INCQNEL‘ C_AS'flNGS " R. Carlander ' In thermal-convection loop tests ‘of Inconel cast- ings in the fuel mixture (No. 30) NaF- ZrF -UF, (50-46-4 mole %), much heavier attack of the cast matericl occurred than is normally observed for wrought Inconel (see Chap, 3.1, **Dynamic Cor- rosion Studies’’). Metallographic examination of the as-received material showed the presence of cracks, stringer porosity, and large grains, with the grain boundaries perpendicular to the surface. In order to obtam further information on the effect of the composition of the cast material on its corrosion resistance, static tests of three Inconel castings that differed primarily in silicon content were performed in the fuel mixture (No. 30) NaF- ZrF -UF, (50-46-4 mole %)} in wrought Inconel capsules for 1000 hr at 1500°F. Examination of the tested specimens revealed that the casting with the highest silicon content (1 .93%) was the least attacked, while the casting with the lowest silicon'contenf was the most severely attacked, In all cases, the attack was a combination of subsurface voids and intergranular penetration. The depth of subsurface voids on the Inconel castings was 3 to 4 mils, as compared with 2 to 4 mils on the wrought Inconel capsules, The intergranular - ‘penetration, however, varied from 3 mils on the high-silicon-content casting to 12 mils on the lowssilicon-content casting. This -surface of the casting. lographic examinations of the three cast specimens ~are presented in Table 3.2,5, and the wrought of the lithium on corrosion resistance, PERIOD ENDING JUNE 10, 1956 _difference is attributed to the porosity and to the grain boundaries running perpendicular 1o the The results of metal- Inconel capsule and cast Inconel specimen No. 321 are shown in Fig. 3.2.14. In every case the wrought Inconel container was more corrosion resistant than was the cast Inconel specimen. INCONEL AND STAINLESS STEEL IN NaK CONTAINING LITHIUM R. Carlander Type 316 stainless steel and Inconel were ex- posed to NaK containing lithium in static and in dynamic systems in order to determine the effect In the static tests the specimens were exposed to NaK | (5644 wt %) with lithium additions of 1, 5, 10, 20, and 30 wt % for 100 hr ot 1500°F. No attack occurred in any of the tests, For the dynamic tests, 12-in.-long capsules B were filled to 40% of their volume with NaK plus lithium (5 or 10 wt %) and placed in a tilting- type furnace (1 cpm) at @ hot-zone temperature of 1500°F (cold zone, 1100°F) for 100 hr. No mass transfer occurred in ony of these systems. The ‘Inconel exposed to NaK with 10 wt % lithium _added was attacked to a depth of 1 mil in the hot zone, while the other capsules were unotfocked. TABLE 3.2,5. RESULTS OF METALLOGRAPHIC EXAMINATION OF CAST INCONEL SPECIMENS AFTER EXPOSURE TO STATIC NeF-ZrF‘-UF‘ (50-46-4 mole %) FOR 1000 hr AT 1500° “silicon Depfh of Aflack (mlls) ,-Ccst Inconél _ Mt ’ -hi N f‘.:’ : Cost Specimen Content Wrought lnconel " Cast’ Inconel e :::::elcs' :c:':e:"/ o8 ~ Neo. o Awt %) Capsule Specirnen o Ppecimel 321 o o 1.04 n ‘ . 4 o - 12 ' Subsurface voids 'trp_'ge'pfl'\ of 4 mlls, R L - o - ' o o intergranul'ar'pe'n-e'tr:dtioh' to "dépth' ' _ 7 ’ '_ of 12 rnlls, no crucks cpparent o | 322 o 1.93 B 2 » 3 B ',Subsurfcce volds und Intergrunular - RN o . ‘ penetration to depth of 3 mlls, no . o crucks cpparenf ' : ' - 323 -_: o ,1.734 B : 2 - 8 . 7___-;.Subsurface voids to depth of 3 mi!s, : lnfargrcnulot penetrutlon to depth : of 8 mils; no cracks apparent 151 ANP PROJECT PROGRESS REPORT 8 UNCLASSIFIED § 8883 Fig. 3.2.14. Speéimens of Wrought Inconel Capsule (a) and Cast Inconel Specimen No. 321 (b) After Exposure to the Fuel Mixture (No. 30) NaF-Z¢F ,-UF, (50-46-4 mole %) for 1000 hr at 1500°F Etched' with aqua regia. 250)( Reduced 3%. (€eeret-wfih-up-hvrr) Thermal-convection loop tests were also con- ducted. In these tests the hot-leg temperature was 1500°F (cold leg, ~1250°F), and NaK with 5 wt % lithium added was circulated for 1000 hr. A slight amount of mass transfer, in the form of small adherent crystals containing 71.5% Ni, 4.6% Cr, ond 0.7% Fe, occutred in the Inconel loop, while no mass transfer was found in the stainless steel loop. Portions of the hot and the cold legs of the Inconel and the stainless steel loops are shown in Figs. 3.2,15 and 3.2.16, respectively. The Inconel was attacked to o depth of 1.5 mils in the hot and in the cold legs. The stainless steel was unattacked in the hot leg but was attacked to a depth of 6 mils in the cold leg. Photomicrographs of the cold legs of the Inconel and of the stainless steel loops are shown in Fig. 3.2.17. The addition of 5 wt % lithium to the NaK did not decrease the normal corrosion resistance of Inconel to NaK, but it did decrease 152 that of type 316 stainless steel, The difference in attack between the hot and cold legs of the stainless steel loop may be attributed to the larger amount of carbides present as a network in the cold leg than in the hot leg, where more carbides are in solution because of |ncreased solublhty ' at the higher temperature. ' TRANSFER OF CARBON BETWEEN DISSIMILAR METALS IN CONTACT WITH MOLTEN SODIUM R. Carlander The decarburization of AIS| 1043 sfeel by molten' sodium in Armco iron and type 304 ELC stainless steel containers was demonstrated 'in tests at 1830°F of 100 hr duration, as described previously.3 An additional test has been performed to determine 3E. E. Hoffman, ANP Quar. Prog. Rep Dec. 10, 1955, -~ [ UNCLASSIFIED Y. 18385 HOT ZONE —816°C (1500°F) COLD ZONE-—660°C(1220°F) Fig. 3.2.15. Portions of Hot and Cold Legs of an Inconel Thermal-Convection Loop in Which NaK with 5 wt % Lithium Added Was Circulated for 1000 hr, (Secret with caption) | it - PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED - Y. 18384 ot COLD ZONE —665°C(1230°F) Fig. 3.2.16. Portions of Hot and Cold Legs of a Type 316 Stainless Steel Thermal-Convection Loop in Which NaK with 5 wt % Lithium Added Was Circulated for 1000 hr. (Secret with caption) Fig. 3.2.17. Cold Legs of (a) an Inconel Thermal-Convection Loop and (5) a Type 316 Stainless Steel Thermal-Convection Loop in Which NaK with 5 wt % Lithium Added Was Circulated for 1000 hr ot ¢ Hot-- Leg Temperature of 1500°F and a Cold-leg Temperature of About 1225°F. Etched with aqua regia. - 250X. Reduced 4%. (Secretwithcuprion) 153 ANP PROJECT PROGRESS REPORT the effect of time on the extent of decarburization of the steel specimens. Two AISI 1043 steel specimens and sodium were loaded into evacuated Armco iron and type 304 ELC stainless steel containers and tested at 1830°F for 400 hr. The conditions and results of the tests are given in Table 3.2.6. As expected - on the basis of the previous investigation, exten- sive decarburization of the steel specimens oc- curred in both the iron and the stainless steel containers. The extent of decarburization of the AISl 1043 steel specimens was greater in these 400-hr tests than in the previous 100-hr test, and, in addition, the vapor zone of the stainless steel container was carburized in the 400-hr test and not in the 100-hr test. This discrepancy may be due to the vapor zone sample having been taken closer to the bath zone after the 400-hr test than TABLE 3.2,6. CARBON AND NICKEL ANALYSES OF THE YARIOUS COMPONENTS OF SYSTEMS DESIGNED FOR STUDYING THE DECARBURIZATION OF MILD STEEL BY SODIUM Material Analyzed Welghf Loss in Carbon Content Nicke! Content Test System 400-hr Test o {wt %) (wt %) (g/‘n.2) AISI 1043 steel 3pécimen in Steel specimen Armeco iron contalner As received 0.433 0.008 After 100-hr test 0.121 0.0001 After 400-hr test 0.054 10.0408 iron container As received 0.018 Vapor zone After 100-hr test 0.019 After 400+hr test 0.018 Bath zone After 100-hr test 0.035 After 400-hr test 0.02¢4 AlS| 1043 steel specimen in Steel specimen type 304 ELC stainless As recelved 0.433 0.008 steel container ) S After 100
INCONEL. PIN- %t - I — . - PK-2 PUMP"BRAZE TEST : Fig- 3.4.5.- Detmls of Mlcrometer Mecsurements . B F'g' 3 4_'6' _N"K (PK-Z) P‘"“P Volufe Afle' ~ on Brazed NaK (FK-Z) Pump Yolute, Measurements' RS ruzmg. ‘ ‘made “at four ‘radial “sections (A throud1 D)ot - = ‘ - 90-deg - mtervals at poslhons 1, 2, and 3 (see : V'Tnble 3. 4.2) ' 179 081 TABLE 3.4,3. RESULTS OF MICROMETER MEASUREMENTS ON DIAMETER OF PUMP BARREL Dimensions (in.) Before Additional _Change Due to Position ofo After Welding hange After Rem?vol Additional :Affelz_l B Retiof of " Total Welding Shell Flange of Restraints Changes Anfteqlmg Changes - .Residua. Stresses Chonges A 6.378 6.383 +0.005 6,387 +0.004 6.384 - —0.003 +0.001 +0.006 - 2 6.390 6.391 +0.001 6.392 +0,001 6.393 +0.001 +0.002 +0.003 3 6.378 6.375 ~0.003 6.376 +0.001 6.377 +0.001 +0.002 ~0.001 4 6.378 6.380 +0.002 6.382 +0,002 6.382 0 +0.002 +0,004 B 1 6.384 6.389 +0.005 6.391 +0,002 6.388 ~0.003 ~0.001 +0,004 2 6.390 6.390 0 6.392 +0.002 6.391 ~0,001 +0.,001 © +0,001 3 6.381 6.376 ~0.005 6.380 +0.004 6.381 +0.001 +0,005 0 4 6.382 6.383 +0,001 6,387 +0,004 6.386 ~0.,001 +0,003 +0.004 c1 6.385 6.391 +0.,006 6.392 +0.001 6.391 —0.001 0 +0.006 2 6.386 6.386 0 6.387 +0.001 6.386 ~0.001 0 0 3 6.380 6.375 ~0.005 6.379 +0,004 6.379 0 +0.004 ~0.001 4 6,382 6.385 +0.003 6.385 0 6,385 0 0 +0.003 D1 6.396 6.393 -0.003 6,401 +0.008 6.401 0 +0.008 +0.005 2 6.383 6.382 -0.001 6.384 +0.002 6.385 +0.001 +0.003 +0.002 3 6.380 6.377 -0.001 6.376 ~0.001 6.378 +0.002 +0.001 0 4 6,382 6.381 +0.001 6.383 +0.002 6.385 +0.,002 +0,004 +0.005 E1 6388 6.392 +0,004 6.392 0 6,391 ~0,001 ~0.001 +0.003 2 6387 6.380 ~0.007 6.383 +0.003 6.383 0 +0.003 ~0.004 3 6.380 6.373 ~0.007 6.378 +0.005 6.377 —0.001 +0.004 © ~0.003 4. 6381 6.382 +0.001 6.382 0 o 0 ~+0.001 6.382 LI0dIN SSTJ903d LIDIT0¥d ANV o3 4 ‘,1' FABRICATION OF JOINTS BETWEEN PUMP BARRELS AND THE PRESSURE SHELL OF THE ART P. Patriarca The ability to weld an accurately machined and properly stress-relieved pump barrel to ‘o thick- walled pressure shell without the need for finish machining or subsequent stress relieving would simplify considerably the assembly of the *‘north head” of the ART. A design was suggested to - permit such a procedure, and a test was conducted to evaluate the feasibility of the desngn. : The component ports of the joint used for the test are shown in Fig. 3.4.7. The,’lmer flange 'is shown attached to the pump barrel sleeve, which in turn was welded to the pump barrel. The finer flange is also shown attached to a carbon- steel pipe, which was in tumn welded to_cZ‘é-in.— thick carbon-steel base plate.. This auxiliary weldment was intended to provide a degree of - restraint comparable to that to be expected from the north-head expansion tank. ' It may be noted that the shell flange is quite : large, the intent being to minimize distortion of the flange and thereby require a realistic pro- portion of the distortion during welding to occur in the barrel sleeve and hence in the barrel itself. The restraint provided by the auxiliary weldment shown in Fig. 3.4.8 was intended to supplement the afore-mentioned ,condmen and, hence, to more nedrly simulate the north-head pressure shelf. The Shell'flange was welded into the pump-barrel . ..~ - sleeve by using the weldmg procedure described -~ - in Fig. 3.4.9. " The completed test weldment is ~shown in Flg. 3.4. 10. - Diametrical changes in the © pump’ barrel ‘were determined by micrometer meass . _ urements at the posmons described in Fig.34.11. - .= The restraints were then ‘removed by cutting the carbon _steel 'with -an " oxyacetylene “torch, and . - ~~the barrel cssembly .was subjected to a 6-hr soakf,_ “-at 1500°F prior to further micrometer measurements “ on the barrel diameter, : The results of these . =~ - _,7;,'1‘";d|ometr|ca! measurements are surnmarlzed in Tubleii " °3.4.3. It may be noted that the maximum chonges;"— . are -remarkably’ small considerlng the exient ‘of weldlng involved. PERIOD ENDING JUNE 10, 1956 It is interesting to note that at least partial relief of residual stresses was accomplished by __cutting the barrel assembly from the restraints and soaking at 1500°F. This procedure resulted in additional diametrical changes, which indicate that some distortion will occur in the north head in service that may be unacceptable, ~ Axial changes were also determined by mi- | crometer measurements at the positions described in Fig. 3.4.12. The results of the measurements ‘are summarized in Table 3.4.4. It may be noted that an oxial shift occurred that was significant in magnitude but remarkably small for the amount of welding involved. These results indicate that ~ some distortion is inevitable and must be either - accepted or removed by subsequent machining and - that a stress relief anneal will be necessary to remove effects of residual stresses during oper- ation, | " TABLE 3.4.4., RESULTS OF MICROMETER 'MEASUREMENTS ON AXIS OF PUMP BARREL Dimensions (in.) Position Before After Change Welding Welding Al 2.459 2.447 -0.012 2 2.459 2.468 +0.009 3 2,444 2.468 +0.024 4 2.437 2.431 —0.006 B 1 2.492 2.480 -0.012 2 2.466 2.470 +0.004 3 2416 2.433 +0.017 4 2437 . 2429 - -0.008 .c1 2.524 -2.516 =0.008 """"" o2 2.473 . 2.474 - +0.00 -1 2,380 2394 +0.014 L4 2437 2431 - -0.006 D1 2563 2,561 —=0.002 2 2.475 2573 -0.002 '3 2,351 . 2361 . +0.010 4 2440 . 2,437 -0.003 E1 2594 2594 . 0 _ 2 . 2484 2,478 -0.006 - - 3 2318 - 2,321 +0.003 4 2.442 2438 -0.004 181 81 o UNCLASSIFTED Y.17244 UNCLASSIFIED PHOTO 17242 PUMP BARREL’ AT £! cpesme Fig. 3.4.8. Test Setup Showing Auxiliory Weldment, Fig. .3.4..7.&' Component. Parts for Test Fabrication of a- Joint Between a Pump Barrel and the Pressure Shell of the ART, LI0dIY 5S3JI0Ud LD3F0Ud NV ‘PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED ORNL~LR-DWG 14722 ELECTRODE MATERIAL " INCO 62 INCO 62 © INCO 132 INCO 132 = CURRENT (amp) 130 180 120 {20 ‘.§ o - ¥ WELDING PROCEDURE PASS NUMBER - PROCESS - ELECTRODE SIZE {in.) o " INERT ARC - Y, ‘ .2 " INERT ARC ¥, < © 3= . METAL ARC (NOTE 2) - 42 ' . f1—24 ' METAL ARC (NOTE 3} T .NOTES : , o _ ‘ 1, VARIED FROM 45 TO 55 deg. DEPENDING ON POSITION AROUND PERIPHERY, BECAUSE A * STRAIGHT BEVEL INCLINED 10 deg TO THE HORIZONTAL WAS MACHINED ON THE FLANGE. 2. BUILDUP PASSES INTENDED TO SIMULATE J BEVEL AND MINIMIZE SHRINKAGE. . ~ - 3. FILLER PASSES IN BUILT UP J BEVEL. | | | Fig. 3.'4".97.: “Welding 'Prééeduré and Joint Design for Wélding fh-e'l.'-‘ressure_ Shell"Flcngé.'im'ofhe Pump- o ‘Barrel Sleeve. 183 ANP PROJECT PROGRESS REPORT UNCLASSIFIED PHOTO 17243 Fig. 3.4.10. Completed Test Weldment for Joining Pressure Shell Flange to Pump Barrel Sleeve. INVESTIGATION OF SHRINKAGE OF INCONEL CORE SHELL WELDS A. E. Goldman A series of tests are being carried out to de- termine the weld shrinkage to be expected during fabrication of the Inconel core shell welds. Comparisons of the weld shrinkage formulas given in the literature with preliminary results obtained at ORNL revealed wide discrepancies. It was decided that actual experiments that would yield empirical data would be required. A program of welding Inconel plates under controlled conditions and close observation was therefore carried out. Based upon the results of these tests, the welding of large Inconel hoops was begun, These tests were performed in a manner that as nearly as P. Patriarca 184 possible duplicated the fabrication problems and restrictions of the actual core shells. Each test of the initial progrum consisted i the inert«arc welding of two /-m. Inconel plates, each 6 x 20 in., in accordance wuth the established procedure specifications (PS-1). A 50-deg beve! with a Y% _-in. land was machined on one long edge of each pfafe. A total of nine tests was performed. Each pair of plates was assembled as shown in Fig. 3.4.13. The root gap was fixed at 1/ in, by using four l/s-m tool-steel spacers. The plctes were held ogainst a flat horizontal plate by means of C-clomps. Two large clamps were used to draw the plates tightly against the tool-steel spacers. | The edges of the plates were securely taped to prevent air leakage into the gap, since only the A L " PERIOD ENDING JUNE 10, 1956 UNCLASSIFIED ORNL-LR—-DWG t4723 SHELL FLANGE SLEEVE PUMP BARREL PUMP BARREL - OO SHELL FLANGE SLEEVE ~— T 7T AL AR ANNN e | ysilfl-J | B SIS - N Fig. 3.4.".' rbetails of Micrometer Meusuremefits on .Diulfnefer of Pump Burrel (s?e Table 3.4.3). 185 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-—LR~-DWG 14724 — PUMP BARREL — SHELL FLANGE i — SECTION AA . . i Yy in.-} 13’3 in. SHELL FLANGE SNUONMOUONNNNNNNONNNNANN i = l r'—POST § B g""——PUMP BARREL N o . \ \E iR N , \ N % N \ N AN 186 - Fig'; 3.4.12. Details of Micrometer Measurements on Axis of Pump Barrel (see Table 3.4.4). 7 O UNCLASSIFIED ORNL-LR-DWG 1472% /4 x 6 x 20-in. INCONEL PLATES 2 Y/g-in. TOOL STEEL SPAGERS (4 USED) 100° INCLUDED ANGLE 2 - ' Fig. 3.4.13. Joint Design for Shrinkage Test of Inconel Plate Welds. UNCLASSIFIED . ORNL-LR-DWG 14726 ® BASE PLATE %‘% ROLLER \ ¢ HOLD-DOWN PLATE /] ® — Y~in. TEST PLATES . Ny . PASS SEQUENCE‘V 74— HOLD ~DOWN PLATE . | roLier 3 ———TACK WELD - 0 - SIDE VIEW OF VERTICAL WELDING JIG - . into Vertical Welding Jig. - ‘were again fastened to the flat base, -F‘ig- 3.4.14, '-.As'seinbly.of Shrinkage Test Plates PERIOD ENDING JUNE 10, 1956 torch gas was used to supply backup gas and weld coverage. A 14-in.-|ong tack weld was placed at each end ~of the root gap, and two more lé,--in.-long tack welds were equally placed alorig the root gap. The clamps were removed after the tack welds were made, and the spacers were driven out, Shrinkage measurements were taken, and the plates The root pass was applied affer the tack welds had been wire brushed and the edges had been feathered. Measurements of the root pass shrinkage were then taken. The plates were then assembled in the vertical welding jig as shown in Fig. 3.4.14. The re- maining five weld passes per plate were deposited in accordance with the sequence shown. The completed weldment is shown in the vertical jig in Fig. 3.4.15. Dial-gage readings were taken ot l-min- intervals during the deposition of the UNCLASSIFIED PHOTO 17245 Fig. 3.4.15. Completed Weld in Vertical Welding Jigo ) 187 AN’P' PROJECT PROGRESS REPORT final passes, and micrometer measurements were taken after the final poss had been made. The shrinkage measurements and welding data are presented in Table 3.4.5; the results of tests 106, 107, 108, and 109 were in close agreement, aithough two different welding operators made the welds. : : The welding procedure used for these tests of plates was then used for test welds of hoops. For these tests ¥ -in. Inconel plates, 6 x 138 in., were bent into Lbops approximately 44 in, in diometer. One edge of each hoop had a 50-deg bevel and a ¥ .-in. land. The hoops were placed on the weld-positioner bed in a horizontal plane -in the manner shown in Fig. 3.4.16. Tool-steel spacers, 4 in. long and ¥ in. thick,- shown in Fig. 3.4.17, were placed between the beveled edges at 6-in. intervals to maintain the root gap. Large C-clamps were used to draw the two hoops tightly against the spacers, Smaller C-clamps were halved and tacked to the bottom hoops to aid in the alignment of the upper hoop. Asbestos string was used to seal the gaps between the ~ spacers prior to tacking. The area behind the root gap was sealed with a cover formed from 0.010-in. annealed brass sheet and masking tape, and the enclosed space was purged for 30 min prior to tacking, Alignment was checked con- stantly as the tack welds were placed between the spacers. After removal of the tool-steel spacers, the root pass was deposited. No dressing of the land or feathering of the tack welds was permitted prior to the root-pass deposition. After completion of the root pass, the weld was wire brushed and the five final weld passes were deposited by using the sequence described in Fig. 3.4.14. For these welds, two welders worked 180-deg apart around the hoop while the positioner was slowly rotated. As shown in Fig. 3.4.18, four dial gages were used to record the shrinkages of the final passes. The welders worked on a 15-min-work, 5-min-rest cycle. A typical plot of dial-gage readings is shown in Fig. 3.4,19. The micrometer and dial-gage measurements obtained are summarized in Table 3.4.6. ' TABLE 3.4.5; WELDING CONDITIONS AND RESULTS OF SHRINKAGE MEASUREMENTS ON WELDS OF INCONEL PLATE Total Shrinkage (in.) Dial-Gage Shrinkage* (in.) Maximum Minimum Average Maximum Minimum Average ' Time Welding Current T Wel est Welder o equired Rod Used Used No. No. {min) - (in.) (amp) 100 1 72 - 2135 70 (root) -' o 105-110 101 1 66 23 70 (root) - ' 105-110 102%* 1. 74 180 80 (root) 105-110 106 1 61 198 80 (root) 105-110 107 - 1 60 226 75-80 (root) ) | 105-110 108 2 79 231 80 (root) L 105-110 109 2 64 236 80 (root) 105-110 0.197 0.123 0,126 0.058 0.048 0.053 0.167 0.130 0.146 0.070 0.050 0.060 0.142 0114 0130 0.081 0.079 0.080 0.157 0121 0141 0.092 0,082 0.087 0.148 0.132 0139 0.104 0.096 0.100 0157 0.106 0.138 0103 0.9 0.098 0.154 0.104 0.136 0.102 0.092 0.097 *Measurements made on last five of the six passes;dial-gage readlngs on tests 100 and '|01 are erroneous because the dial-gage actuating arm slipped durmg the test, **Test 102 plates were machined with an 80-deg included angle rather than a 100-deg included ungle. 188 - » *y UNCLASSIFIED PHOTO 17248 Fig. 3.4.16. Inconel Hoops on Weld Positioner Bed. Within the limits of this mveshgohon, the following conclusions can be drawn. - 1. The results of the last four tests on plates indicated that, under controlled conditions, the effect of the welding variables could be minimized . so that a welding. operator could, essentially, ' jduphccte his performance from test to test when welding manually. Also, -for. a given set of con- ~ ditions, o welder could nearly. duphcate another - '~:.welder s performance. ‘ 12, The use -of various lengths of tool-steel . . spacers m ‘rhe fcck weldmg of the plutes caused “‘wide . variations - the resulting - tack-weld ~~shrinkage. - " was found to be mversely propomonol to the ' ~ length | of the spacers.. B 73, The results of the tests on the hoops '“‘_,-' dicated that the inevitable variations in conditions “The. qmount of tack-weld shrinkage and techniques while the operators were pro- gressing around the circumference caused greater PERIOD ENDING JUNE 10, 1956 UHCLASSIFIED ¥-18648 Fig. 3.4.17. Tool-Steel Spacers Used to Main- tain Root Gap While Welding Inconel Hoops. UNCLASSIFIED PHOTO 17247 Fig. 3.4.18. Hoop Welding Assembly Showing ‘Dial Gages Used for Measuring Shrinkage. - variations in shrinkage within a hoop ihan the ~variations observed from hoop to hoop. -4, The over-dll results indicate that for '/-m. Inconel plafe, _inert-arc welded by two welding operators under the conditions utilized for these tests, the transverse shrinkage to be expected will be from 0.111 to 0.138 in., with an average 189 ANP PROJECT PROGRESS REPORT 0.096 DIAL GAGE A 0.080 = - —— DIAL GAGE B ~—— — DIAL GAGE C ~ese-secs- DIAL GAGE D 0.064 0.048 SHRINKAGE (in.) 0.023 0.016 0 20 40 60 80 100 120 - UNGLASSIFIED ORNL-LR~DWG 14727 140 160 180 200 220 240 260 280 300 . " WELDING TIME (min} Fig. 3.4.19. Typical Plot of Dial-Gage Measurements of Slmnkuge Durmg Fmal Fwe Passes of 1he Welding of Inconel Hoops. TABLE 3.4.6. WELDING CONDITIONS AND RESULTS OF SHRINKAGE MEASUREMENTS ON WELDS OF INCONEL HOOPS Te;t Arc Time Rod Used Current Total Shrinkage (in.) Dial-Gage Shrinkage* (in.) " Required - Y. Used No. Amin) (in.) {amp) Maximum Minimum Average Maximum Minimum Average 1 314 1218 80 (root) 0.126 0.111 0.1194 0.095 0.080 0.088 o 110-120 -2 310 1188 80 (root) 0.138 0.115 0.1261 . 0.102 0.088 0.095 110-120 *Measure ments ‘made on -logt five of the six passes; tack shrinkage and root shrinkage not included. \}aide being 0,120 in. The longitudinal shrinkage to be expected will be 0.250 to 0.375 in. for a circumferential length of 138 in. As previously mentioned, the values for the transverse and longitudinal shrinkage for the 44-in. hoops do not correspend to the values obtained by calculation from the formulas given in the literature. Only through the accumulation of empirical data from actual experience can pre- dictions be made for future weld shrinkages: Since each variation in the thickness of the Inconel plate to be welded will create new problems solvable only by more actual test results, ad- 190 ditional tests will be conducted on each of the plate thicknesses of interest. . - EXAMINATION OF NaK-TO-AIR RADIATOR PWA NO. 2 AFTER SERVICE R. J. Gray P. Patriarca A 500-kw high-conductivity-fin radiator, desig- nated PWA HCF radiator No. 2, failed on December 23, 1955, as the result of a leak. - This radiator had been operating in a test rig for a period of - 1199 hr in the temperature range 1000 to 1600°F. For 546 hr of the operating period a temperature differential was imposed on the NaK flowing " o o through the radiater by passing cold air across the fin surfaces. The radiator is shown in Fig. 3.4.20 as it appeared when recelved from 1he test site, The entire radiator was leak checked by pressur- izing it under water and observing it to locate the origin of air bubbles. This procedure revealed the point of failure, which is indicated by the arrow in Fig. 3.4.20. The radiator was then sectioned for further examination, as shown in UNCLASSIFIED Y-17518 p‘”’-A Fig. 3.4.20. NaK-to-Air Radiator PWA No., 2 That Failed in Service. Arrow mdlcotes point of failure, UNCLASSIFIED Y. 17586 Fig. 3.4.21. NaK-to-Air Radiater PWA No, 2 After Sectioning for Metallographic Examination. PERIOD ENDING JUNE 10, 1956 Fig. 3.4.21. The side portions of the support members were removed by using a rubber-bonded masonry wheel in a portable, electric handsaw adjusted . for a shallow cut. Each bank of ‘fins was separated by slicing the support members and the bottom flanged plate, as shown, with a fine- toothed, high-speed, steel hacksaw blade in a portable electric drill equipped with a portable power-saw attachment. The failed area was then carefully removed for metallographic preparation on a standard wet-cutoff machine eqmpped with an abrasive wheel. The area of the failure was again pressurized under water to locate the exact position of the leak before additional preparation for metallo- graphic examination was undertaken. The failure was found to exist in the comer tube on the periphery which faced the side support member, The support member was subsequently removed for unobstructed observation of the emergence of the water bubbles. The point of failure may be seen in Fig. 3.4,22. ' | A UNCLASSIFIED ‘ Y-$7795 s TUBE SHOWING i-FIRE DAMAGE 1 TO FINS im0 i O NaK-to-Air Radiator PWA No. 2 as Viewed from the Support Member Side of the Radictor, " 191 ANP PROJECT PROGRESS REPORT The tube that failed and two adjacent tubes were mounted intact and carefully ground to permit examinations of their longitudinal cross sections as seen against the direction of air flow. The tube that failed is shown in Fig. 3.4.23. The neckdown of the tube wall indicates ‘a tensile fracture similar to that observed? in York radiator Neo. 1. Longitudinal cross sechons of the two ad|acent tubes that were examined are shown in Figs. 3.4.24 ond 3.4.25. Incipient fractures may be seen - in both these tubes. Three tubes were also taken from corresponding positions on the air exit face and prepared for examination in a similar manner. Only one tube. ‘exhibited evidence of incipient fracture, as shown in Fig. 3.4.26. It is concluded that the radiator failed as a result of the initiation of a fracture in a braze ‘alloy fillet by shear forces and the propagation 2P. 'Putriqrco et al., ANP Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012, p 145. UNCLASSIFIED - ¥-17882 " SUPPORT ¥ MEMBER of this fracture through the tube wall by tensde forces or combinations of tensile and shear forces during periods of cyclic operation. Since the incidence of incipient. fractures was associated exclusively with the presence of support members “or plates, it is recommended that these transverse restraints be removed entirely. This can be accomplished, as suggested previously, by using a high-conductivity fin to provide transverse support- at 2< or 4-in. intervals® and modifying the brazing procedure accordingly. . The tensile loading contribution to the York radiator failure was attributed ot the time of examination ‘to the restraining influence of the support member, which extended up the side of the radiator.® In view of this conclusion, the support members of subsequent radiators, including PWA No. 2, were sht as shown in Flg. 3.4.20, 3R. J. Gray and P, Pctrlarca, Metallo rapbzc Ex- amination of ORNL HCF Radiator No. 1 Failures, ORNL CF-55-10-129. ‘ UNCLASSIFIED J§ Y.17881 .. © 2 wehes, Jo jo.04} - F:g. 3.4.23. Longitudinul View of Opposing Walls of the Tube That Failed in Nak-io-Air Radlator PWA No. 2 as Viewed Against the Alr Flow. Note neckdown at fructure. 75X Reduced 22%. ) : 192 L e | ! _ . _ : ' R R mcmsnr = B N mncrunz ERIOD ENDING JUNE 10, 1956 » --Ff:g'z‘:s :. FIRE DAMG cvt:;]\;:gleo S ' - TO FINS - lo.ot | P )§7 . SUPPORT - MEMBER 5 loee] : Fig. 3.4.24. Lengitudinal View of Opposing Wulls of a Tube Aclgucent to the Tube Thet Fclled as Viewed Against fhe Air Flow. Note mcnplent fracfure. 75X. Reduced 34%. ] | | i o | 1 1 . Longitudinal Vlew of Opposing Wulls of a Tube Adlncent to the Tube That Failed as Fig. 3.4.25. VYiewed Against the Air Flow. Reduced 34%. #y - Note fracture within the eutectic structure of the braze metal, 75X. 193 ANP PROJECT PROGRESS REPORT UNCLASSIFIED § & R %%: SUPPORT MEMBER | R Fig. 3.4.26. Corner Tube from Air Exit Face of NaK-to-Air Radiator PWA No. 2 as Yiewed in the Direction of the Air Flow. 75X. Reduced 36%. prior to installation. [t may be noted, however, that the radiator as redesigned retained the bottom flanged plate, a top plate, and the four support members. During the brazing cycle the NaK tubes were brazed to these members, and a relatively rigid condition resulted at each of the five trans- verse sections across the radiator matrix. Local differences in rates of heating ond cooling, par- ticularly between the air inlet face and the remainder of the radiator during blower startup, could therefore bring about the development of tensile loading. This condition could be partially relieved by slicing the support members and plates in @ manner similar to that utilized during dis- section for metallographic examination, as shown in Fig. 3.4.21. ' The development of tensile forces alone, how- ever, cannot be assigned the full responsibility for failure. The incidence of numerous incipient fractures in this radiator has been related to the presence of a support’ member or heavy plate. Over 13,000 tube-to-fin joints have been examined 194 metallographically without the observation of o single incipient fracture, The differences in mass and ‘thermal conductivity of the support members and plates as compared with the high-conductivity “fins could result in significant differences in heating and cooling rates during cyclic operation.. These differences could create lateral forces that could be resbo_'nsib'le for the initiation and propa- gation of fractures in brazed joints between tubes and support members in any portion of the radiator. EXAMINATION OF FUEL-TO-NaK HEAT EXCHANGER AFTER SERVICE G. M. Slaughter Tests of the fuel-to-NaK heat exéhangér, desig-- nated as |HE-3, were terminated as a result of the detection of a leak in a tube bundle after a total of 1794 hr of operation in the temperature range 1100 to 1500°F. There was o temperature differential imposed on the heat exchanger for- 1015 hr of this total time, and 21 thermal cycles were applied over this period. The NaK inlet and NaK outlet headers of the tube bundle that lecked were separated from the heat exchanger to facilitate examination and inspection, Top and bottom views of the inlet header are shown in Figs. 3.4.27 and 3.4,28. The general location of the failure is evident in Fig. 3.4.28, in that a dark reaction product can be distin- guished from the lighter solidified fuel mixture. Forty tubes in the area of the failure were indi- vidually inspected with a dye penetrant and o Borescope, and at least five tubes were found to contain obvicus cracks. The NaK inlet header after dissection with an abrasive cutoff wheel UNCLASSIFIED Yages Fig. 3.4.27. Top of NaK Inlet Header of Fuel-to- NaK Heat Exchanger IHE-3 Showing Tube Welds. ® 9 i e -/ to permit the examination of individual tubes is shown in Fig. 3.4.29. Each tube was numbered for the subsequent investigation as shown in Fig. 3.4.30. S S The frequency and severity of the cracks de- tected in the initidl_inspection were most pro- nounced in the forward rows of tubes, that is, those with short bends.” The distance from the tube bends to the headers was significantly shorter UNCLASSIFIED Y-1m47 Fig. 3.4.28, Bottom of NaK Inlet Header of Fuel-to-NaK Heat Exchanger 1HE-3, Dark reaction product indicates crea of tube failure. The light solidified material is the fuel mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %), . (Secret with caption) PERIOD ENDING JUNE 10, 1956 for the first row of tubes, being 3% in. as com- pared with 6 in, for the last row of tubes. For g given expansion of the 6-ft over-all straight length of the tubes as a result of heating, ¢ substantial degree of strain occurs in these locations, As would be expected, cracking was more pronounced on the tension sides of the tubes. A crack in the tension side of tube 2 could be seen upon visual examination (Fig. 3.4.31). Further evidence of tube distortion at the headers can be seen.in Fig. 3.4.32, which shows the NaK outlet hecder. Several of the tubes of interest were mounted intact in Castolite and polished to the approximate UNCLASSIFIED Y RWEAR o onccnon § Fig. 3.4.29. NaK Inlet Header After Dissection for Further Exomination. UNCLASSIFIED ORNL=-LR-DWG 14728 ® 0060 000 © 0 ©0 006 ® @ 0 ° @ 0.0 00 o. ® O _ DIRECTION OF BENDS Fig. 3.4.30. Diagram of NaK Inlet Header Sfiowing, Identification Numbers of Tubes Removed for Examination. 195 ANP PROJECT PROGRESS REPORT EE UNCLASSIFIED Y-1m315 Fig. 3.4.31. Crack on Tension Side of Tube 2 of NaK Inlet Header. center line for metallographic examination. The tensile side of tube 3, shown in Fig. 3.4.33, illustrates ‘typical severe cracking and corrosion by the fuel mixture which circulated on the outside of the tubes. A similar condition is evident in Fig. 3.4.34, which is a panorama of the tension side of iube 17. The opposite face of tube 17, o pancrama of which is shown in Fig. 3 4.35, does not exhibit so serious ¢ condition, It appears that the corrosion and the stresses combined to form an abnormally unfavorable condition. 196 UNCLASSIFIED T Yame Fig. 3.4.32. Bottom Side of NaK Outlet Header Showing Distortion of Tubes. The extent of the corrosion was investigated by examining tube 93- to ensure that the large degree of attack observed in the previous samples did not resuft from the reaction of the two fluids at the locations of 'rhe failures. The results of the examination (F;g. 3.4.36) indicate that, in general, severe corrosion was prevalent throughout the tubes in the inlet header. The inner tube wall of a typical tube is shown in Fig. 3.4.37; cracks and corrosion emanating from the outer wall may also be seen. A white deposit was found in the cracks in some areas, as shown in Figs. 3.4.38 and 3.4.39. The nature of this deposit, as well as a detailed investigation of the mass transfer and corrosion, will be reported later. | PERIOD ENDING JUNE 10, 1956 § i i i Fig. 3.4.33. Tube 3 of NaK Inlet Header Showing Cracks in Tension Side. Etchant: electrolytic oxalic acid. 100X, 2] Fig. 3.4.34. Panoramc of Tension Side of Tube 17 of NaK Inlet Header Showing Gr-osg Cracks and Cor- g rosion. Etchant: electrolytic oxalic acid. 33X. 197 ANP PROJECT PROGRESS REPORT Fig. 3.4.35. rPcnorcmrm of Compressién Side of Tube 17 Showing"Corrosion and Occasional Cracks. .Etchant: electrolytic oxalic acid. 33X. ’ Fig. 3.4.36. Tension Side of Tube 93 Showing Severe Corrosion, Etchant: electrolytic oxalic acid. 1005(.‘ 198 PERIOD ENDING JUNE 10, 1956 URCLASSI WIEEY Y Fig. 3.4.37. Inner Surface of Tube 17, Cracks and corrosion emanating from outer surface may be seen. Etchant: electrolytic oxalic acid. 200X, Fig. 3.4.38. Cracks and Deposits on Tension Side of Tube 19, Etchant: electrolytic oxalic acid. 100X, 199 ANP PROJECT PROGRESS REPORT . Fig. 3.4.39. Crack and Deposit Evident in Fig. 3.4.38 at a Higher Magnification. Etchant: electrolytic oxalic acid. 500X. 200 o/ -3.5. MECHANICAL D. AI EFFECT OF ENVIRONMENT ON CREEP- RUPTURE PROPERTIES OF HASTELLOY B ‘C R. Kennedy! Revised desugn data obtained from creep tests of solution-annealed Hasfelloy B sheet stock in various envuronments at 1300, 1500, und_'|650°F ore summarized in Figs. 3.5.1, 3.5.2, and 3.5.3. The times to 0.5, 1, 2, 5, and 10% total strain at each temperature in the various environments are the same, and for stresses for which the rupture life is more than 300 hr the effect of environment is shown to be negligible, The creep curves obtained ot 1500°F in air and in argon, shown in Fig. 3.5.4, appear to indicate that the better performance in air than in the other environments at stresses for which the rupture life is less than _IOn assignment from Prott & Whitney Aircraft. PERIOD ENDING JUNE 10, 1956 PROPERTIES STUDIES -Douglas -. ' 300 hr is caused by the ability of air to strengthen Hastelloy B during third-stage creep. - For rupture lives longer than 300 hr there is considerably less third-stage creep because of the aging characteristics of the alloy, and, as seen in Figs. 3.5.1, 3.5.2, and 3.5.3, the effect of environ- “ment diminishes, Design curves produced from limited data for solution-annealed Hastelloy B - sheet tested at 1800°F in argon and in the fuel ‘mixture (No. 30) NaF-ZrF -UF (50-46-4 mole %) are shown in Fig. 3.5.5. At 1800°F Hastelloy B does not age perceptibly and the amount of third- stage creep is great in tests at all stress levels, Thus the fuel mixture strengthens the alloy at all stress levels, Of the environments tested, only those which are ‘‘surface active” (produce a thin, tightly adherent film on the surface of the metal) ‘affect the creep properties of the metal, It can be seen from Figs, 3.5.1, 3.5.2, 3.5.3, " and 3.5.5 that, of the environments tested, only SEORET ORNL-LR-DWG 14729 50,000 <] \\-\ h\$(j). M — P~ \\0,9‘\ , 40,000 \\. ] - \ ‘\\“;/4/4 e N, - \‘-.._____ \\ 7 \ \\ \':GCN '9(,& \\\ \ \ \\ B ;Z."l?é\ N N D N /4/4/ 30,000 \ ) o | o \ 7’5 L ’ '. '- 0.570 \ ._ o ’ _7 ) . 7 : y : 20000 — e T N s L e N N N R \\ NI | | A2 s 1020 50 lOO 200 500 1000 - 2000 - 5000 . 10,000 Lol e e e TlME(hr) - e : Fig. 3.5.1. Design Curves for Hustelloy B Sheet Solution Annealed at 2100°F for 2 hr and Tested in Argon and in the Fuel Mixture (No. 730) NaF.ZeF, UF, (50-46-4 mole %) at 1300°F. 201 ANP PROJECT PROGRESS REPORT — U ORNL-LR~-DWG _14730 20,000 f'g"- « - W L & “ 10,000 9000 - 8000 7000 N 6000 : : ' L 1 2 5 10 20 50 100 200 500 1000 2000_ 5000 {0,000 ' TIME (hr) ) ' ‘Fig. 3.5.2. Design Curves for Hastelloy B Sheet Solution Annealed at 2100°F for 2 hr and Tested - in Yarious Environments ot 1500°F. SEGRENr ORNL-LR-DWG 14734 ) o 14,000 13,000 12,000 14,000 10,000 9000 8000 STRESS (psi} 7000 6000 5000 ~ 4000 . L { 2 5 10 20 50 100 200 500 {000 2000 . 5000 40,000 TIME (hr) Fig. 3.5.3. Design Curves for Hastelloy B Sheet Solution Anneuled at 2100°F for 2 hr und Tesfed . o in Argon and in the Fuel Mixture (No. 30) NaF-ZrF UF , (50-46-4 mole %) at 1650°F. S 202 PERIOD ENDING JUNE 10, 1956 ‘ ! B _ UNCLASSIFIED 60 ORNL-LR-DWG 14732 50 H o W o ELONGATION (%) 20 v 100 200 300 400 500 600 700 800 900 1000 TIME (hr) Fig. 3.5.4. Design Curves for Hastelloy B Sheet Solution Annealed ot 2100°F for 2 hr and Tested in Air and in Argon at 1500°F. - BSOGRE— ORNL-LR-DWG 44733 - - 20,000 - 10,000 ' /RUPTURE IN ARGON so%o /11 ‘ 8000 ©, o, ) ML 7000 1% |2% 5% 5% F R | ."5 6000 03% | IS wp\\!“““i:\fiewflap a PVYVY TN e P ~Zr, ) E ~ o RN ‘--..__- \\:\\\\\‘ \-...:lsio;.qs s o : ' L L ~hao~¢e : N 4000 s = -e o [ .’no/ey - T ] Pl | “-\ B \'::- °) e L TS TR/ IR 3000 - . : N \\ N o SRR NG \\\ . '_.-_“-2_000 _\\\\ i 000 L1 o AL . L , S g - .5 - 10 - 20 . 50. 100 200 500 - 4000 2000 5000 10,000 el ) _ ' . . TIME {hr) T - o - Fig. 3.5.5. Design Curves for Hastelloy B Sheet Solution Annealed at 2100°F for 2 hr and Tested N in Argon and in the Fuel Mixture (No. 30) NaF-ZrF +-UF, (50-46-4 mole %) at 1800°F. ' 203 ANP PROJECT PROGRESS REPORT air and the fuel mixture seem to be surface active, Air, of course, produces an oxide film which is ever present, and the fuel mixture, in effect, creates a very thin surface film by leaching one of the alloy constituents and producing another phase on the surface, as shown in Figs. 3.5.6 and 3,5.7. SHORT-TIME HIGH-TEMPERATURE TENSILE PROPERTIES OF HASTELLOY B C. R. Kennedy The short-time high-temperature tensile proper- ties of solution-annealed Hastelloy B are illustrated in Fig. 3.5.8, which gives the yield and ultimate strengths and the final elongations in the temper- ature range 1000 to 1800°F. As may be seen there is a distinct decrease in the final elongation and in the ultimote strength ot temperatures around 1200°F. This change in properties occurs in the temperature range in which the type of frocture transforms from predominently fransgranular to intergronylar, -It is interesting to note that the change in the yield strength with temperature is relatively smoll. CREEP-RUPTURE PROPERTIES OF HASTELLOY W C. R. Kennedy Creep testing of Haste”oy W is now in progress, ond design data are presented in Figs., 3.5.9, 3.5.10, and 3.5.11 for solution-annecaled sheet tested in argon at 1300, 1500, and 1650°F. Hastelloy W, which has almost the same compo- sition as Hastelloy B, except for the addition of - 5% chromium and the deletion of 3% molybdenum,: has creep properties very similar to those of Hastelloy B. Although Hastelloy W exhibits less of a tendency to age than Hastelioy B, as shown in Figs. 3.5.12 and 3.5.13, a decrease in ductility occurs at 1300°F.. This is also shown in Fig. 3.5.9, in which the absence of a 10% curve indi- cates that the total strain at rupture was less than 10%. Rupture points obtained from tests with the fuel mixture (No. 30) NaF-ZrF ,-UF, (50-46-4 mole %) are also shown in Figs, 3.5.9 and 3,5.10. The times to 0.5, 1, 2, 5, and 10% total strain are identical for the same stress and temperature, and, as shown, only the rupture life is affected by Fig. 3.5.6. Surface Effect on the Unstressed Portion of a Hastelloy B. Specimen After Exposure to the Fuel Mixture (No. 30) NaF-ZrF ,-UF , (50-46-4 mole %) for 1700 he. 1000X. (Secretwith-eaption) 204 1h o PERIOD ENDING JUNE 10, 1956 _Fig. 3.5.7. Surface Effect on the Stressed Portion of a Hastelloy B Spec:men After Exposure to the Fuel Mixture (No. 30) NoF-ZrF, -UF (50-46-4 mole %) for 1700 hr ot a Stress of 13,500 psi. 1000X. (Secret-with—caption) 120,000 | 100,000 80,000 Cy " 'STRESS (psi) . Fig. 3.5.8. UNGLASSIFIED N\ ORNL-LR-DWG 14734 60,000 “Short-Time T Range 1000 tc 1800°F. 60 - : ] lozuvewp L , I & \\ 40 _— ' 20 : - - “ELONGATION - ‘17 - - 7 0 1200 1400 ‘1600 . 1800 - TEMPERATURE("F) - ELONGATION (%) ensile Data for Soluhon-AnneuIed Hastelloy B Tested in the Temperature 205 ANP PROJECT PROGRESS REPORT SECRET ORNL-LR-DWG 44735 50,000 o 40,000 “\\ NS \._.. \ P \ ) . ""--____-.--.-- i \ \P\ e —— @ 30,000 S = -'_'—-——-_--_'--'_i-— "'---_“_-.- .......‘\ \ - e Tt I P 2 \\\ 5% RUPTURE 2 N & \\ "\._\ = : , TN 0 000 ~ 20 0.5% 1% 2% @ RUPTURE POINTS FOR TESTS IN NaF- ZrF4-UF‘ {50-46-4 mole %) 10,000 ‘ v t 2 5 {0 20 50 100 200 500 {000 2000 5000 10,000 TIME (hr) Fig. 3.5.9. Design Curves for Hastelloy W Sheet Solution Annealed ot 2100°F for 2 hr and Tested in Argon at 1300°F. S DGR ORNL-LR-DWG 14736 20,000 18,000 16,000 14,000 12,000 STRESS (psi} 10,000 8000 & RUPTURE POINTS FOR TESTS IN NaF—ZrF, —UF,; (60-46 -4 mole %) 6000 . 1 2 5 10 20 50 100 200 500 {000 2000 5000 10,000 TIME (hr) Fig. 3.5.10. Design Curves for Hastelloy W Sheet Solution Annealed at 2100°F for 2 hr and Tested in Argon at 1500°F. 206 14,000 12,000 10,000 0.5% N% 2% 8000 STRESS (psi) 6000 4000 { 2 5 10 20 5C PERIOD ENDING JUNRE 10, 1956 UNCLASSIFIED ORNL-LR-DWG 14737 RUPTURE 100 200 500 1000 2000 5000 10,000 TIME (hr) Flg. 3.5.11. Design Curves for Hastelloy W Sheet Solution Annealed at 2100°F for 2 hr und Tested in Argen at 1650°F. the environment. The effect of the fuel mixture on the creep properties of Hastelloy W sheet appears to be the same as that shown for Hastelloy B. SHORT-TIME HIGH-TEMPERATURE TENSILE PROPERTIES OF INCONEL - J. R Weir, Jr. The tensule properhes of lnconel sheei huve'fi' been determmed at temperatures from 78 to 2200°F. - The yield point, by 0.2% offset, and the ulhmate ) ~ strength are shown in Fig. 3. 5.14 for both fines __gramed and coarse-gromed material, As ‘may be - the fine-grained material may be the more desirable structural material because of its better tensile properties in the temperature range of interest, CREEP TESTS OF INCONEL IN FUSED SALTS J. R, Weir, Jr. Inconel was creep tested at 1500°F and a stress of 3500 psi in NaF-ZrF (50-50 mole %) in order to compare the severity of attack with that found after similar creeptests in the fuel mixture (No. 30) ‘NoF-ZrF +UF, (50-46-4 mole %), The results are - shown m Flgs. 3.5.16 and 3.5.17." Very little -:v;'surface void formation is seen in the case of the - specvmen tested in the nonuramum-beurmg mixture, " in comparison with the attack by the fuel mixture., _,f‘.-These results are 'Further ev:dence that much of seen,. the fine-grained : material “has the better " the attack by the fuel mixture on lnconel af ]500°F strength properties at temperotures up to-1700°F, - y - ma be attnbuted to the reaction Transuem‘ loads ‘induced by thermal fluctuchons_’:- Y; may, -in some " cases, be of greater concern than ~ the stahc loads that result from pressure differ- " ential. - Therefore, ‘even though the fme-gmmed"_" ,lnconel has less creep resistance in the fused o “salts than the coarse-grdined Inconel (Flg. 3.5.15), 2UF + Cr (m lnconel) —_— 2UF + CrF CREEP TESTS OF WELDED INCONEL “J. R Welr, Jr, o ' Several 0060-m.-fh|ck sheet-type creep speci- mens were machined from welded /-m. Inconel sheet stock and a few creep tests were run in the 207 80¢ 254 UNCLASSIFIED: 5706 4 Fig. 3.5.12, Hastelloy B Sheet After Creep Testing afa - Fig, 3.5.13. Mastelloy W Sheet After Creep Testing at a Stress of 30,000 psi ot 1300°F in Argon; Ruptured in 185hr, - Stress "of-_35.00_’0” psi at 1300°F in Argon; Ruptured in 450 hr. ' S o © - 100X, Reduced 17.5%. o o - IOOX. Reduced 17.5%. N s LA0dIY $STYI0Yd LD3T0dd dNV - » > . 1000 L v o PERIOD ENDING JUNE 10, 1956 UNGLASSIFIED ORNL~LR-DWG 14962 TENSILE STRENGTH| | FINE GRAIN 100,000 | TENSILE g7 \ 80,000 T COARSE GF?EPIGTH \ 5 . \ % N < 60,000 o : \ B \ W \ ' Fl °"'ELD \\ 40,000. \J SRa’ \ CRoPoRy, TION, '\\\ F'INEG LUMIT N RAIN \ \\\ 0.2 20,000 =52 YIELD, Coarge GRAIN N N ‘ N | U= o L : _ ~ o 400 800 1200 1600 2000 2400 2800 TEMPERATURE (°F) Fig.. »3.5.14._ Cbmpurison of Tensile Properties of Fine- and Coarse-Grained Inconel at Tempera- tures Between 78 and 2200°F at a Stress Rate of 0.016 (in./in.)/min, : CEORET 20.000 ORNL-LR-DWG 14963 3 . 10,000 | 8000 6000 4000 " 'STRESS (psi) OOARSE GRAIN,. e fiSO' . 2000 |- b b 1= 1ot | L] {FiNe erA, e b b L 1e500F — Syt 2 08 70 40 20t 5O L 100 200 - 500 {000 2000 5000 {0,000 o Tl ' - ' "TIME(hr),' ST e T Flg. 3 5. 15. Compunson of Stress-Rupfure Properfies of Fine- and Coarse-Gramed Inconel in the Fuel Mixture (No. 30) NaF-ZsF, ~YUF, (50-46-4 mole %) ot 1300, 1500, and 1650°F. 209 ole "Fig. 3.5.16. Inconel After a Creep Test in.the Fuel Fig. 3.5.17. [Inconel After a Creep Test in NaF-ZrF, “Mixture (No, 30) NaF-ZrF -UF , (50-46-4 mole %) at 1500°F - (50-50 mole %) at 1500°F Under a Stress of 3500 psi. 100X. ‘Under a Stress of 3500 psi. 100X. Reduced 18%. (Secret ~ Reduced 18%. ‘ LAOdITY SSTYO0Yd LI3F0dd ANV '\ " fuel mixture (No, 30) NaF-ZrF 1-UF (50-46-4 mole %) and in argon. The .results obtained to date are summarized in Table 3.5.1, The positions of the fractures in"the specimens tested at 1500°F in the fuel mixture are shown in PERIOD ENDING JUNE 10, 1956 psi and No. 3 at 4000 psi. As may be seen, the weld did not deform appreciably, compared with the base metal, in the gage length. Metallographic examination of the specimen tested at 1300°F disclosed that the weld metal was more corrosion Fig. 3.5.18. Specimen No. 2 was tested at 3000 resistant than the base metal. TABLE 3.5.1. RESULTS OF CREEP TESTS OF WELDED INCONEL SHEET IN THE FUEL MIXTURE ‘ (No. 30) NaF-Z¢F (-UF, (50-46-4 mole %) AND IN ARGON Stress - Temperature £ H _ Rupture Life Elongation (psi) (°F) nvironment eat Treatment . (hr) (%) 12,000 1300 Fuel mixture As received 110* _ 21 4,000 ' 1500 Fuel mixture Coarse grained 380 n 4,000 1500 Argon Coarse grained 750%** 10 3,000 1500 Fuel mixture As received 7 820 5 3,000 1500 L Argon B ' As received e 2000** . 4 *Premuture rupture, probably cuused by confuminaiion of the fuel rmxfure. *+Still in test, o ENCLASSIFIED Yalnes) _Fig. 3.5.18. Welded Inconel Specimens After Creep Tests In the Fuel Mixture (No. 30) NaF-ZtF - UF, (50-46-4 mole %) ot 1500°F. Specimen No. 2 was tested at a stress of 3000 psi and specimen No. 3 was tested at a stress of 4000 psi. (Sesretawith-ception) 211 ANP PROJECT PROGRESS REPORT 3.6, CERAMIC RESEARCH L. M. Doney RARE-EARTH-OXIDE COMPACTS FOR ART CONTROL RODS J. A, Griffin L. M. Doney The design of the ART control rods calls for a porous compact of rare-earth oxides. The pores of the compact are to be filled with metallic sodium, and the compacts are to be canned in Inconel. The following process was developed for the fabrication of the compacts. The as-received rare-earth-oxide material (Code 920 from Lindsay Chemical Co.) was pressed -at 4000 psi into compacts, which were calcined at 1325°C for 1 hr, The calcination step was carried out to reduce the shrinkage of the oxide during the final sintering. These compacts “were crushed to pass an 80-mesh screen. The final mixture was compounded from 75% of the 80-mesh calcined -material and 25% of the as- received Lindsay Code 920 oxide. This combi- nation was thoroughly mixed and pressed, with no - binder, at a pressure of 4000 psi. These compacts were then sintered at 1425°C for 55 min in.air. Samples of the sintered compacts. are shown in Fig. 3.6.1. The compacts were oversize in all dimensions so that they could be ground accurately to size and so that no difficulty would be encountered during canning. The grinding was carried out by a commercial ceramic firm. The inside and outside Fig. 3.6.1. Sintered Rare-Earth Compacts Before Being Ground to Desired Dimenéibns. 212 W ¥ s of each compact was ground to the required dimension, and the ends, as well as being ground to the required length, were made flat and parallel. The ground pieces are shown in Fig. 3.6.2. The Lindsay Chemical Co. supplied the foflowing | analysis of the Code 920 rare-earth oxides used in the fabrication of the compacts: Samarium oxide 45.0-49.5% Gadolinium oxide 22.5-27% Neodymium oxide 0.9-4.5% Prasecdymium oxide - 0.9-3.6% Cerium oxide 0-0.9% 3.6.2.Rar PERIOD ENDING JUNE 10, 1956 s e Europium oxide 0.9-1.2% Other rare-earth oxides 8.1-14.8% .Tb.i_olfrure-earfh oxides 90% (minimum) PETROGRAPHIC EXAMINATIONS OF FLUORIDE FUELS ~G. D. White T. N. McVay, Consultant Examinations of fluoride fuel samples with the petrographic microscope were performed at a rate of about 200 samples per month. The materials examined .included quenched samples for equi- librium-diagram determinations, control samples ~from experimental runs, and samples from pro- duction batches. The results of these exami- nations are reported in Part 2, '‘Chemistry.”’ UNCLASSIFIED PHOTO 26348 213 ANP PROJECT PROGRESS REPORT 3.7. NONDESTRUCTIVE TESTING STUDIES R. B. Oliver EDDY-CURRENT TESTING OF SMALL-DIAMETER TUBING J. W, Allen Continued study of the application of the cyclo- graph! to the problem of the inspection of small- diameter tubing has revealed that, in addition to its use as a flaw detector, the cyclograph may also be used to gage adherence to dimensional tolerances. Although changes in diameter and in wall thickness are inseparable in the readout of the instrument, the tolerance limits for both di- mensions may be established by utilizing two standards: (1) the minimum acceptable diameter and wall thickness and (2) the maximum allowable diameter and ‘wall thickness. Dimensional checks made by using this method, augmented with me- chanical measurements, have been used success- fully to determine the dimensional acceptability of approximately 6000 ft of % -in.-OD, 0.025-in.- 1R. B. Oliver, J. W. Allen, and K. Reber, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 164, they have a length of /16» wall and 0.229-in.-0D, 0.025-in.-wall CX-900 Inconel tubing. The 200-kc cyclograph trace of a 0.229-in.-OD, 0.025-in.=wall Inconel tube that has small wall- thickness variations along its length and no per- ceptible diameter variations is shown in Fig.3.7.1. The wall thickness was plofted from Vidigage (ultrasonic-resonance-type thickness gage) and mechanical measurements. It may be seen that, although there is not perfect agreement, there. is a fairly close correspondence between the two types of checks. The disagreement is probably due to minute amounts of eccentricity and inters granular attack on the inside surface of the tube having been detected by the cyclograph. _ The confidence level of the mterpretahons of mdacatnons from this instrument is. _increasing with. continued use. Because the _se_nsmg coil observes at cmy one instant a section of tubing approximately / in. long, it is not possfl:le to detect pin holes, except when they occur in clusters. [n general, defects can be resolved if in. ‘or longer and have UNGLASSIFIED - — ORNL -LR-DWG 13970 - £ 00275 I ~ fi 0.0270 Z . ’ TOLERANCE LIMITS X 0,0265 : r:E ,0.0260‘ f\ /\\ N /\ f\ LN AN f\ N y_ . - UNWTV Y Y YT N \ L0255 ‘ § 00258 | Z 00250 AVERAGE WALL THICKNESS DETERMINED WITH VIDIGAGE OUTSIDE DIAMETER CONSTANT AT 0.230 in, PSS e 1 200-ke¢ CYCLOGRAPH TRACE 0 6 2 18 24 30 36 42 48 54 60 66 72 78 LENGTH (in.) Fig. 3.7.1. Cyclograph Record at 200 kc from a 0.229-in.-OD, 0.025-in.-Wall CX-900 Inconel Tube Compared with Dimensional-Variation Measurements Made Mechanically and with the Vidigage. 214 T S e i " i wy & depths greater than the background dimensional variations in the tube, ULTRASON'C |NSPECT|6N OF TUBING R. W. McClung Approximately 6000 ft of CX-900 Inconel tubing has been inspected by the immersed ultrasound method, Two sizes of tubing 3/16 in. OD, 0.025 in. wall and 0.229 in. OD, 0.025 in. wall, were in- spected and the average rejection by this test was about 2%. Each tube received a double in- spection, with the vltrasound being beamed around the tube in two different directions to improve the chance of detection of unfavorably oriented crack- like defects. This double inspection of the precut tubing reduced the inspection rate to below the original estimate. The current inspection rate is approximately 500 ft in an S-hr day for lengths up to 10 ft, Defects 0.0015 in. deep and I/1 6 in. long on polished, scratch-free tubing can be detected with " PERIOD ENDING JUNE 10, 1956 ‘an estimated confidence of 80 to 90%. If scratches are present they will produce signals comparable to those from the very small defects and effectively increase the minimum detectable defect size. If the defect is appreciably deeper than the scratches and if its length exceeds the maximum dimension of the transducer, it is not difficult to differentiate between defects and scratches. Since the ultra- sonic method is insensitive to dimensions, the inherent dimensional variations that give trouble in an eddy-current inspection. have no effect on ultrasonic inspection. Both the eddy-current and the ultrasonic methods are capable of detecting very small defects, with the eddy-current inspec- tion being confused by dimensional variations and the ultrasound method being confused by scratches. A comparison of the ‘resulfs of the two tests aids in the identification of spurious signals. , A fypical small defect found on the inside of the / 6-in.-0D, 0.025-in.-wall CS$-900 tubing is shown in Fig. 372 This crack is, 0.0015 in. deep (6%) and about ¥ ¢ ine long. Fig. 3.7.2. Small Defect Found by Ultrasonic Inspection on the Inner Surface of a ?'/ -ln.-OD 0.025-in.-Wall Inconel Tube. The defect is 0.0015 in. deep and agpproximately 46 in. long. 215 ULTRASONIC INSPECTION OF PIPE No development work on the ultrasonic method for the inspection of pipe had been planned, be- cause the currently employed contact method was reported to be adequate. Preliminary attempts to employ this method, however, revealed several problems that detracted’ from the reliability of contact inspection. 1t was found to be very diffi- cult to fabricate Lucite shoes that would fit the contour of the pipe and would, at the same time, limit the sound beam to the proper spectrum of incident angles. A sizable part of the sound was propagated as a surface wave and produced inordi- nately large signals from surface scratches. For this reason, defects and scratches could not be separated. Also, the contact method requires that a thin film of oil be maintained between the trans- ducer and the pipe surface, and, thus, any surface roughness, vibration, ovality, or an inadequate supply of oil resulted in a loss of signal and frequent failure to detect defects on the inner surface, Immersed ultrdSound permitted a solution” to these difficulties, and the use of the “B" scan for date- presentahon permitted a high inspection speed with rapid interpretation of "the signals produced in the tube wall. ' A scanning tank of maximum simplicity was designed and fabricated for the ultrasonic inspec- tion of pipe. The tank is 26 ft long, 14 in. wide, and 20 in. deep and is equipped with a variable- speed headstock and chuck and an assembly of 2On assignment to Homogeneous Reactor Project. 216 two 3-in.-dia casters, in liev of a tail stock This assembly contains and rotates the pipe around its own axis. The transducer mounting is aligned with reference to the pipe axis by a pair of Micarta guides, and it is manually translated along the pipe. Very straight pipe can be rotated at speeds of 200 rpm, but much of the pipe is too crooked for this speed. _Even.in the worst case, however, rotational speeds of 60 to 100 rpm are attainable. This equipment is belng used for the inspection of large quantities of pipe in sizes from ?’ to 6 in. IPS for the ETU and the ART. Satisfactory inspection of tubular shopes is de- pendent upon obtaining reference notches of the proper depth (3 to 5% of the wall thickness) on both the inner and outer surfaces. A known, re- producnble notch is easily produced on the outer surface of both pipe and tubing; howe_ver, it is difficult to produce such notches on the inner surface of even large pipe. No really satisfactory notch has yet been produced on the inner surface of small-diameter tubing, and the inside reference " notch ‘is necessary’ to prove that the alignment will reveal defects on.the inner surface.’ Therefore this method is ot being’ used for. the mspectlon . of small-digmeter tibing, Four of the eight nozzle welds on the cell bemg fabricated to house the ART in Building 7503 were inspected by the vltrasound method. The observed indications of cracking in the weld and in the heat affected zone resulted in the removal of all the welds. The cracks revealed were con- firmed by arc gouging and by Magnaflux tests with sufficient frequency to justify the rejection of the welds on the basis of the ultrasound indications. i . Part 4 HEAT TRANSFER AND PHYSICAL PROPERTIES H. F. Poppendiek RADIATION DAMAGE G. W. Keilholtz FUEL RECOVERY AND REPROCESSING H.. K. Jackson _ CRITICAL EXPERIMENTS ADcalhhan | =f. » 4.1. HEAT TRANSFER AND PHYSICAL PROPERTIES H. F. Poppendiek ART FUEL-TO-NaK HEAT EXCHANGER S. |. Cohen J. L. Wantland The ART fuel-to-NaK heat exchanger test system was modified for further studies of heat transfer and friction characteristics. Six 60-deg staggered spacers and six §0-deg inclined spacers were placed alternately on the tube bundle, and trans- verse pressure taps were installed across one spacer of each type. The average transverse pressure drop across each instrumented spacer was plotted in terms of the ratio of the transverse pressure drop to the velocity head times the fluid density over the Reynolds modulus range of 3000 to 8000. For the inclined spacer this term varied from 1.4 to 1.1, ond for the staggered spocer the data fell randomly between 0.01 and 0.03, with no definite trend established. The analytical investi- gation, referred to previously,? on a method of - correlating fluid-friction data for flow parallel to a square array of cylindrical tubes in rectangular cross-section flow channels was completed.3 An experimental study is being made of the fuel-side flvid-friction characteristics of a mockup of the present design of the ART fuel-to-NaK heat exchanger. The dimensions of this apparatus and the spacer configuration will simulate the ART heat exchanger as nearly as possible, except that there will be no headers and the tube bundle will not have the curvature specified for the ART; that is, the friction characteristics for a straight length of tube bundle will be determined. . An experimental heat -exchanger has been de- - signed and is now. bemg fabricated . ‘that will be - similar to the one on whlch tests were conducfed - previously,4 wuth the exception that the tubes will - be .spaced on a triongular array . rather ‘than a square ar_ray. Thls mveshgahon will provnde g 350 ‘Wantland, A’ Metbod of Correlatmg Experz- - ;,- .‘mental Fluid Frzct:on Data for Tube Bundles of . Different Size and Tube Bundle -to Sbell Wall Spacmg, _‘ ORNL- CF-56-4-162 {Aprit 5, 1956). . 4, L. Wantland Thermal Cbaracter:stzcs of tbe ART Fuel-tosNaK Heat Exchanger, ORNL CF=55-12-120 (Dec. 22, ]955)0 data for direct comparison of the heat-transfer and fluid-friction characteristics of the two differ- ent configurations in the transitional Reynolds modulus range. ART HYDRODYNAMICS C. M. Copenha\)er F. E. Lynch G. L. Muller® Sodium Flow in Reflector Cooling System A 5/22-scale mode! of the reflector—core shell cooling annulus and inlet system was designed and fabricated (Fig. 4.1.1) in order to determine quanti- tatively the flow distribution-in the annulus and the various pressure drops in the sodium system. The fluid used in the model was water at temper- atures between 20 and 50°C, Although the model does not incorporate the reflector cooling holes, 12 nozzles were spaced radially around the inlet header to take off fluid at various sectors of the inlet header to mock up flow through the cooling holes. The flow distribution was studied by three methods: (1) the static pressure measured axially across the annulus in conjunction with the tfotal annulus flow rate measured by a rotameter gave " the average velocities through the annulus for the two positions nearest the inlets and the two positions farthest away, (2) pitot-static combi- nation measurements were obtained at positions 1-3 and 3-3 (Fig. 4.1.1), and (3) high-speed motion ‘pictures were made of periodic ink injections. A quantitative measurement of the flow distri- bution was defined as - Vo, + Vy ! ;01@ R < 1+ V, » — — R ' _'where oY Wuntland ANP Quar. Prog. Rep ]une 10. 1955.? S . ?,ORNL-IB%. P 149. L S 2J. L. Wonfland ANP Quar. Prog._Rep. Marcb 10. o _’,1956; ORNL-2061, p 173. . _ 0, = volumemc flow rate in annulus farthest from inlet, -0, = volumetric flow rate in annulus nearesf to ~inlet, : : V- ‘= average llnear velocnty ut station indi- -cated by subscript number. 30n assignment from Pratt & Whitney Aircraft, 219 ANP PROJECT PROGRESS REPORT SCTRET™ ORNL-LR-DWG 14738 PUMP " STATION 4 STATION 3\ PUMP B PUMP A INLET HEADER (PLEXIGLAS) Pg3-5 Fle S, G S 2/ g - ’, 7 2SS 277 7, 7, TC TO : ROTAMETER ' ROTAMETER \\\\\ ORIGINAL GAP = 0.031 in. { x 22 =0.135in.) PRESENT GAP = 0.043 in. ( x =0. 1875 in.) 3.14in. 5 22 5 SHELL TO REPRESENT REFLECTOR (PLEXIGLAS io.91 in.|0.91 in| Pg3 -3 Py3-3 OQUTER CORE SHELL {14 - S-T-41 ALUMINUM ALLOY) 344 in, ROTAMETER ——TO PUMP Fig. 4.1.1. Vertical Section Through Reflector—Core Shell Annulus Medel. v 220 TATION 1 - This relationship gives the ratio ‘of the averages of the velocities at the two stations farthest from the inlets to the averages of the velocities at the two stations nearest the inlets. The relationship for the flow distribution when only one pump was operating was defined as v v, Y_ pump A 4+ ~—n | PUmP B 0 /o V, | operating Va operating Q,, 2 Some of the experimental results are presented in Fig. 4.1.2 in terms of these relationships. The predicted flow distribution in the ART system SR - ORNL-LR-DWG 14739 110 ANNULUS THICKNESS = 0.4875in. [ RATIO OF ANNULUS FLOW TO COOLING - HOLE — FLOW =0.609 A 1.08 A - . ] . N ] - i—r""—’ 1.06 F A& METHOD ¢ } | ONE PUMP OPERATING {‘ METHOD 2 1.04 0 METHOD { TWO PUMPS OPERATING {. METHOD 2 . e e 1.02 o 2 ot = = =] %o 1.00 ‘ Caxi0® 0 2 - 5 a0 ' REYNOLDS MODULUS AT MIDPLANE Fig. - 4.1 .2: Flow Dismbution in _Céné'efiirié Reflecfor-—Core Shell Annulus. | S PERIOD ENDING JUNE 10, 1956 was obtained by extrapolating the data to the Reynolds numbers for the actual system at the mid-plane of the annulus. A summary of the resuylts is given in Table 4.1.1. The minimum flow rate was found in the portions of the annulus nearest the inlets, while the maxi- mum flow rate was observed in the portions farthest away. These conditions cre the reverse of those anticipated, and a possible explanation is that the fluid entering the annulus from the inlet header must abruptly change direction and it probably undergoes high velocity losses, particu- larly near the inlet to the annulus where the momentum is the greatest. After the model inlet header used in these experiments had been fabri- cated, the design of the actual ART sodium system was modified so that uniform distribution of sodium flow would exist in the annulus if there were no eccentricities, The flow in such a concentric annulus will be axial; that is, there will be no spiraling. Studies of the motion pictures of the flow in the model indicated that, for a constant pump pressure, the average velocity fluctuation at any point in the annulus would probably be Iess than £5%. The flow distributions in the annulus for two possible conditions of eccentricity of the core shell were also studied. Buckling of the core shell between the spacers was simulated to produce a local eccentricity. Since the reduction in flow as a function of the reduction in flow area ‘was the prime concern, the buckling of the core shell was simulated by adding pieces of tape, as shown in Fig. 4.1.3. For a locql reduction in annulus thickness of 50% (¢ = 0.5), the flow from _inlet to outlet through that sector of the annulus was 73% of the average flow. - Radial eccentricity was also. simulated to study_ L _the.effect_of a _shght canhng of the core sherll on ] ;. TABLE 4.1.1. SUMMARY OF CONCENTRIC ANNULUS FLOW DISTRIBUTION STUDIES : Annulus width ln. ,VT,ReynoIds modulus at mid-plone | " Ratio of cmnulus flow to cooling hole flow for fwo S condltlons For two pumps operotmg For one pump operating (70% normal flow assumed) | 0.135 e 0,1875 072 x10° 101 x 10° 1025 1.02 1.075 1407 1.045 1.025 1.08 1.06 221 ‘ANP PROJECT PROGRESS REPORT | SEOREP ORNL-ELR-DWG 4740 000000 /-CORESHELL oo 0 TaPE( ' ! CORE SHELL ;‘,:é:a&i___ab\z::mws o ] REFLEGTOR A a, | 1.0 ‘ " o 0.8 — s 0.6 — - - ’/ g / < 04 - / ' REYNOLDS MODULUS AT Py MIDPLANE = 1.4 x10* © ,=0.4875n. T o o 0.2 0.4 - 0.6 0.8 1.0 /% Fig. 4.1.3. Effect of Local Eccentricity on Flow Distribution. the flow distribution in the annulus. displacement of the shell was found to cause a slight spiraling of the flow through the annulus. The configuration studied and the experimental data obtained are shown in Fig. 4.1.4, The data are in good agreement with a simple, derived expression based on parallel and equal system flow resistances. For a mean radial eccentricity (¢, /1) of 0.8 the maximum deviation of the ratio Q /Q3 is about 0.8. The static pressures as a funchon of position in the annulus and in the inlet header were also obtained for the concentric and eccentric cases and are being analyzed. Fuel Flow in Core Further studies of the flow through the 21-in. ART core model with the Pratt & Whitney vortex generators installed in the entrance region were delayed pending the construction of a second test stand for the core-model experiments. Studies of other core configurations that may give stable flow were, however, initiated. In the preliminary study the following three configurations are being considered: a core with an area expansion rate obtained from Nikuradse'’s gamma function, a 222 Radial ontrr™ ORNL-LR-DWG t4984. b % N _ ° 0 [~ REYNOLDS MODULUS AT s MIDPLANE = 0.8 x40*% ,,/’ Qn . =0.435in. . // ! .0 — S ,< — A e @, £ L~ —= | — + 3 0.8 1 0.8 0.9 1.0 W5 an.4.1 4, Effect of Radial Eccenmcuy on Flow Distribution, cylindrical annular core with a large area ex- pansion at the entrance and screens of perforated plates placed in the region of expansion to obtain good flow characteristics, and the present 21-in. core with screens or perforated plates in the expansion region, ART CORE HEAT TRANSFER EXPERIMENT N. D. Greene F. E. Lyf‘lchi G. W. Greene G. L. Muller L. D. Palmer H. F. Poppendiek The instrumentation and operating character- istics of the ART volume-heot-source experiment, described previously,® couples was calibrated. Fifteen complete power runs were made for the swirl entrance system. N. D. Greene et al,, ANP Quar. Prog. Rep. Dec. - 10, 1955, ORNL-2012, p 174. . were checked, and the’ - galvanometer ‘used with -the transient thermo- w *f " At The parameters of the experiment were the follow- ing: Helical Reynolds modu|us 66,000 to 256,000 Prandtl numbe_r ‘ 4105 Axial temperature rise’ S to 10°F Total pr'cwv'er generated ' 0.08 to 0.12 Mw The heat,.bqlan_ces_ obtained for the system were within £5% of being perfect. Two power runs were also made with the swirl entrance system for the simulated case of *‘one pump off.” A photograph of the one-half-scale model used for these experiments is presented in Fig. 4.1.5; it is made almost entirely of Micarta and platinum. The mixing chamber, pump-scroll head, core, power leads, and thermocouple leads for the outer core shell can be identified in the photograph. The panel board containing recording and power control equipment for the experiments is shown in Fig. 4.1.6, and @ view of the fuel annulus after the pump-scroll head was removed at the end of the experiments is shown in Fig. 4.1.7. The platinum—platinum rhodium thermocouples, as well as the platinum electrodes, which are visible in Fig. 4.1.7, were in good condition at the end of the experimental study. A two-dimensional plot of the electric po- tential field for the 24-electrode power circuit is presented in Fig. 4.1.8. Except for the very ends of the system where some flux distortion exists, the axial voltage gradient was within dbou} 5.5% .. of being uniform; consequenfly, the power densfiy was within about 11% of being wniform. In the actual ART system in which the heat sources will - be generated by fission, the volume heat sources will not ‘be uniform; in fact, near the wall they,-f will be from 2 to 3 times as great as they will be near the center of .the channel. The mean, un- - cooled wail und fluid temperature profiles obtained in these experlmenfs ‘with @ uniform volume heat source are .presented in Fig. 4.1.9 in normalized form, .The asymmetries in the outer core shell and island core shell wall temperatures can be ex- plained on'the basis of hydrodynamic flow asymme- tries. For example, the high island core shell ‘wali temperofure in the northern hemisphere exists- because @ separation . region completely encom- - passes the island ‘shell in that region. “The solution for an idealized ART (paraliel-plates system)? is also plotted on Fig. 4.1.9; this predicted uncooled wall temperature profile lies PERIOD ENDING JUNE 10, 1956 Poo.o® s UNCLASSIFIED ‘'PHOTO 25940 x0T ’ ¥ i Fig. 4.1.5. One-Half:Scale Model of ART Core for Yolume-Heat-Source Exp'eriments‘. between the is!and ond outer core shell wall temperature measurements. 7H, F. Poppendiek and L. D. Palmer, Forced Con- vection Heat Transfer Between Paralle! Plates and in Annuli with Volume Heat Sources Within the Fluids, ORNL-1701 (May 11, 1954). 223 ANP PROJECT PROGRESS REPORT 4 a wmm ‘i g3 | : _AT : ‘Ao M.MH UP . - ® £ = o - E = ® o - v v s =2 o vy = < . . - , % £ Q > |- O Yo - - - & o T o fim r 2 : o Ba - e o .Ch um - - S Q o e - B = 0 - g . . -t - 2 - F - £ 224 " ) Figo .401070 Fuel Annblus of the Volume-Heat- Source Apparatus. PERIOD ENDING JUNE 10, 1956 e . ORNL-LR-DWG 1474{ #= WALL OR FLUID TEMPERATURE | "%, = INLET MIXED MEAN FLUID TEMPERATURE r,,, = QUTLET MIXED MEAN FLUID TEMPERATURE ) "2~ 4 « POWER DENSITY 1] _ oA ol i - . - Lt 1 o . [ f———.._;/ 8 ISLAND CORE SHELL b |~ S . » / - < | 2~ 7 0.8 : i - IDEALIZED ART / L . K (THEORETICAL ) ;f”>'\ ! \ - /7| .~ OUTER CORE SHELL ~ 0.6 : Vs /7 | | ' // ,r’ 7 i I - » A 7 MIXED MEAN FLUID ” ’ 7’ £z 0.4 y L. 1’ - 0.2 1 A HELICAL Re = 256,000 ‘ A /,-/ Pr=4.8 ] W =% (UNIFORM) 0 -] | | 0 2 4 6 8 10 12 t4 6 18 AXIAL DISTANCE FROM INLET (in.) Fig. 4.1.9. Temperature Profiles of the Mean, ‘Uncooled Outer Core Shell, Island Core Shell Wall, and Fluid of the One-Half-Scale ART Core Model ‘with a Umform Volume Heat Source. UNCLASSIFIED PHOTO 26304 225 ANP PROJECT PROGRESS REPORT Experimental transient wall and fluid temper- ature data are shown in Fig. 4.1.10 in terms of the total temperature fluctuation divided by the axial temperature rise of the fluid going through the core. The frequencies of the temperature fluctu- ations for this one-half-scale volume-heat-source system vary from about é to 4 cps. The frequency range of temperature fluctuations in the full-scale ART system operating with fluoride fuel should be similar to that in the one-half-scale volume-heat- JA V'I \ | ', At ]\/ LJ \J"' { Afm ISLAND SHELL AT EQUATOR honstii—— § S8 C =] {. WS | QUTER CORE SHELL AT EQUATOR —— A~ /L\ o MID-STREAM BEYOND EXIT 0 10 2.0 3.0 TIME (sec) source system in which.an acid was circulated. The frequencies of the temperature fluctuations were found to be in agreement with the frequencies of the velocity fluctuations observed in the full- scale core model used for flow studies.. The results of a study of the temperature ‘structure in an idealized ART core were presented in the previous report.® The increase in 'the 8H. F. Poppendiek and L., D. Palmer, ANP Quar ' Prog. Rep. March 10, 1956, p 176. SENET ORNL-LR-~DWG 14742 HELICAL Re = 256,000 W=W (UNIFORM) Afg = TOTAL CHANGE IN WALL TEMPERATURE Aty = AXIAL FLUID TEMPERATURE RISE IN GOING THROUGH CORE - LS NN 1 B ot Flg. 4.1.10. Transient Surface and Fluid Temperatures Obta ined fnr ART Core Model wuh a Umform Yolume Heuf Source. 226 » [} » - uncooled wall temperature that would arise if a hyperbolic cosine power density distribution existed rather than a uniform one was determined. The uncooled wall temperature increment, above the fluid temperatures, for this nonuniform power density case was found to be more than twice as great as the corresponding increment for o uniform power density case. The experimental measurements obtained for the uniform power density cose have been modified by this factor and are plotted in Fig. 4.1.11, along with the mathematical prediction for an idealized ART core with a hyperbolic cosine power distribution. As may be seen, temperatures as high as 1800°F might occur if the walls are not cooled properly for this very high Reynolds number case. Elastic thermal stress calculations were also made for the core shell and heat exchanger tubes on the basis of the experimental temperature fluctuations that were observed in this experiment SeCRT™ ORNL-LR-DWG 15045 2.4 2.2 W = POWER DENSITY ANYWHERE IN FUEL CHAMBER W, = POWER DENSITY IN CENTER OF FUEL CHAMBER 20 Ko = RECIPROCAL OF THERMAL DIFFUSION LENGTH IN FUEL [T 7= RADIAL DISTANCE FROM CENTERLINE OF CHANNEL, 48 1™ 7= WALL OR FLUID TEMPERATURE [ fp= INLET MIXED MEAN FLUID TEMPERATURE 1.6 ™ 4, = OUTLET MIXED MEAN FLUID TEMPERATURE 4 ,,/ & . // 1 o RARE: e = ISLAND CORE SHELL — "‘\54"' 1 TE w0 . /:/ f~—] L IDEAUZED ART (THEORETlCAL)-a 77 T _F oun-:a com-: sneu.---_w T [ 0.8 MIXED MEAN FUEL ST | ""”"?4,( A : 17 0.6 A . . /, g /| , _ o4 b 17 AT |/ HELICAL Re = 256,000 v A A - Prw48 T N W = W, cosh K r | LT =t 1 0.2 ——71P 7 - 1 L~ AXIAL D!STANCE FROM lNLET !il'l) Fig. 4111, fefn;pg;u-scre ,mme's of the Mean, ‘Uncooled Outer Core Shell, Island Core Shell Wall, and Fluid of the One-Half-Scale ART Core Model Adjusted for the Nonuniform Yolume Heat Source. o -2 4. 6 -8B 10 12 4 6 18 PERIOD ENDING JUNE 10, 1956 for the uniform—power-density case. For the uniform-power-density system the results indicated that cyclic stresses exist that are similar in magnitude to the endurance limit or fatigue stress of Inconel. It is believed that the stresses in the actual ART system may be higher than those caleulated for the uniform-power-density case. A third set of volume-heat-source experiments is under way. For these experiments entrance vanes that will reduce the rotational velocity component are located in the core throat. THERMAL-CYCLING EXPERIMENT H. W. Hoffman D. P. Gregory? The volume-heat-source experiments with the one-half-scale model of the ART core have verified that the hydrodynamic instabilities that exist in some regions of the ART core, as presently designed, would result in rapid, high-temperature- differential cycling of the Inconel core-shell surfaces. These studies have also indicated that the tube bends at the fuel inlet end of the fuel-to- ‘NaK heat exchanger will be subjected to this temperature cycling. The volume-heat-source experiments indicate that the temperature fluctu- ation at the Inconel surface may be of the order of +60°F. The thermal diffusivity of Inconel is poor, ond therefore these large temperature fluctu- ations will occur, chiefly, in a region close to the metal-fuel interface. For example, the amplitude of a temperature fluctuation with a Y-sec period will be reduced to 40% of its origino? value at a position 35 mils below the metal surface., A ~ possible result of this cyclic thermal expansion ~within the metal will be cracking of the surface ~because of metal fatigue. The possuballtles of accelerated creep and corrosxon * under these " gircumstances also arise. : _ The effect of thermal cycling on materlul strengfh _must be determined experimentally, and a system for accomplishing this study -at reactor temper- - .gtures with a fuel environment is currently being - o consfructed and tested. The ‘apparatus is shown schemohcally in Fig. 4.1.12. -~ NaF+ZrF -UF (50-46-4 mole %) will flow through - the heater under gas pressure and will be subjected ‘to eyclic heating. ‘The surface of the unheated test section will experience the resultmg periodic “The fuel mixture - %0n assignment from Pratt & Whitney Aircraft. 227 ANP PROJECT PROGRESS REPORT SCORES CRNL—LR—DWG 14743 INCONEL TEST SECTION ey ------ ------ EXHAUST - EXHAUST - :-:‘ .:‘. ELECTRODES IN LIQUID TIN B Cswny b _SOLENOID VALVE - . POWER IN . g : SUPPLY \ BELLOWS SUPPLY TANK NO. 1 TANK NO. 2 --------------- 1. FUEL MIXTURE .................................... --------------------------- BEAM SUPPORT WEIGH BEAM MICROFORMER Fig. 4.1.12. System for Thermal Cycling of Inconel Tubing in a Fuel Environment. temperature fluctuations as the fluid moves down- stream. Effects of the thermal cycling will be determined by metallographic examination of the Inconel surface subsequent to the test. . The heater and the test section are fabricated from a single 10-in, length of Inconel tubing (/-ln. OD, 0.035-in. wall). The heater comprises the first 4 in. of this tube and will be supplied with a pulsed current through a device designed to provide a range of pulse frequencies. In order that the stresses imposed on the test section will result only from the thermal cycling of the tube surface, the heater electrodes are floated in pools of molten tin to eliminate tube bending, and 228 a bellows is provided at the outlet end of the test section to take up the over-all thermal expansion, Preliminary tests with NaNO,-NaNO,-KNO (40-7-53 wt %) as the heated fluid were termmatej when oxidation at the electrodes resulted in loss of power. This situation is now being corrected The results of these first experiments for a 11 /-sec power-on — 1V-sec power-off cycle are glven in Fig. 4.1.13. 4T|1e amplitude of the temperature fluctuation at two points in the test section is shown as a function of the heater power., The ““tailing-off’’ of the data at high heater powers is caused by the slow response of the measuring equipment. The cycle perlod will be varled in future experiments, Ty " - TEMPERATURE AMPLITUDE AT OUTER WALL OF TEST SECTION {°F) . \ o SHIELD MOCKUP REA.CI:OR STUDY L. C. Palmer ' A general study of the heat tronsfer and fluid flow characteristics of the proposed Shield Mockup Reactor (see Chap. 5.5, ‘‘Shield Mockup Reactor Design and Construction’’) has been initiated. An anclysis of the temperature structure within the fuel elements and the coolant and a study of the pressure distribution within the core were com- UNCLASSIFIED ORNL~LR—DWG 14744 90 / 80 / iR ® %4 in. FROM TEST SECTION ENTRANCE // o o 3% in. FROM TEST SECTION ENTRANCE ;/ ¢ / TO /ll [ A/ / /] /| £ o o 8 N % T 10 20 30 40 50 60 7O 80 90 {00 HEATER POWER (%) : : 8 i o o o Fig. 4.1,13. Temperature Amplitudes Observed in Preliminary Thermal-Cycling Tests of Inconel Tubing. PERIOD ENDING JUNE 10, 1956 pleted. .The studies are being extended to other regions of the reactor, TEMPERATUR_E smucfuns IN THE REGION BEYOND THE ART REFLECTOR H. W. Hoffman The ‘previous calculations19 on the temperature structure in the region beyond the reflector in the ART were extended to include a recent system modification. The new system, illustrated in Fig. 4.1.14 and designated System C, differs from those previously studied (Systems A and B)!? in that a portion of the boron carbide has been replaced by a stainless-steel-clad copper~boron carbide matrix. This layer, located between the reflector and the remaining B C, will absorb the leakage neutrons from the core and thus prevent -radiation damage of the more brittle boron carbide. A detailed description of the idealized system used in this study has been presented in a separate report.1! Based on these idealizations the system can be analyzed by an iterative procedure to obtain both the temperature rise of the sedium and the temperature profile for the region between the sodium and fuel return streams. The heat deposition rates used in this analysis were tenta- tive upper limits. 12 104, w, Hoffman, J. L. Wantland, and C. M. Copen- haven, ANP Quar, Prog, Rep., March 10, 1956, ORNL- 2061. p 174. Ty, W, Hoffman, Thermal Structure for the Region Beyond the ART Reflector — Supplement I, ORNL CF-56-4-I29 (April 17, 1956). 2H. W, Bertini, personal communication to H, W, Hoffman, SEoREeT. ORNL—LR—DWG 13335A E 4. |8 c D E Fle|H I Jikl] ¢ M I 11 =z o 12 o l s wl |8 |y |k x !5 |z bl o ol sk = alsl & | | 3 1= INCONEL |@la] S o |o|812(8 BORON CARBIDE'. |&|2] 2 ] e |8 a9l wa Qi Wil O T ow | @ z[° &°© |9ziT|=Z zZ| Tl £ | 1 = 8 ge = | v w wn v | X % X% X% X % X Xg X X R A T & SYSTEM C Fig. 4.1.14. Schematic Diagram of Region Beyond the ART Reflector. 229 ANP PROJECT PROGRESS REPORT The recent system modification resulted in a temperature rise of 33°F for the sodium coolant stream and a maximum boron carbide temperature of 1526°F for a sodiuminlet temperature of 1200°F. These values may be compared with the previous results for a system without the copper—boron carbide matrix in which the sodium femperature rise was 24°F and the boron carbide attained a maximum temperature of 1586°F. The sodium temperature rise was found to vary 10% (decreasing from 36 to 33°F) when the sodium inlet temper- ature was changed from 1150 to 1200°F. Typical temperature profiles at the inlet and outlet ends of the system for a sodium inlet temperajure of 1175°F are presented in Fig. 4.1.15, and the temperature profiles for the system without the Cu-B ,C (System B) are alsc shown for comparison. Be, 4. -Barton, Fused Salt Compositions, ORNL CF+55-9-78 (Sept. 16, 1955). NaF-LiF-ZrF (22-55-23 mole %) Beta form (106 to 368°C) Hy = Hygoe The system medification increased the -temper- ature drop across the inner Inconel shell by 30 to 35°F. The temperature gradient in the outer inconel shell was essentially unchanged. - - HEAT CAPACITY W. D. Powers The enthalpies and heat .capacities of two mixtures were determined in the liquid and solid states. For one mixture, NaF-LiF-ZrF 4 (22-55-23 mole %), two distinct discontinuities in the temper- ature-enthalpy relationship were found; one was at 370 to 380°C and the other was at 565 to 585°C. The reported!® liquidus temperature is 570°C. The lower discontinuity is assumed to be a phase transition from the alpha form to the beta form, the beta form being stable below 375°C. The dis- continuity at the higher temperature is the fusion of the alpha form to the liquid at 570°C. Enthalpy - and heat capacity measurements of a second salt mixture, LiF-NaF (60-40 mole %), showed no unusual characteristics. The enthalpy and heat capacity equations for these two salts follow: = ~8.7 + 0.2392T + (7.20 x 10~5)12 c, = 0.2392 + (14.39 x 10=3)T Alpha form (407 to 556°C) H., - H30° T “p Liquid (603 to 897°C) H,. -~ H 30°C €p T At 375°C H, -~ Hg = 28 At 570°C Hyo = Hy = 24 LiF-NaF (60-40 mole %) Solid (112 to 572°C) H - H T 30° c = 642 — 0.009165T + (41.82 x 10~ = ~0.,009165 + (83.64 x 10~3)T ~20.2 + 0.4526T - (5.95 x 10~-5)12? 0.4526 - (11.89 x 10=5)T c = =9.8 + 03191T + (9.94 x 10~5)T2 C, = 0.3191 - (19.87 x 10~3)T 230 .} M Liquid (688 to 896°C) PERIOD ENDING JUNE 10, 1956 Hy = Hygoo = —882 + 0.9249T — (24,62 x 10-5)72 c, = 0.9249 - (49.23 x 10°°)T At 652°C Hyjq = Hyg = 170 In these expressions H = enthalpy in cal/g, = heat capacity in cal/g+°C, = temperature in °C, VISCOSITY AND DENSITY 'S. . Cohen | A study was made of the effect of fission products on the viscosity of a fused fluoride reactor fuel. The increase in viscosity as a result of the presence of dissolved fission-product additives was found to be less than 5%. The formula for the fission-product additives was, however, based on a much higher burnup rate than would actually be encountered in the ART. The formula called for an amount of LaF, (the com- ponent used to representall the rare-earth fluorides formed by fission) that exceeded the solubility limit in the fuel ot temperatures below 800°C. ‘It is felt that the only serious possible problem that - might arise would be the precipitation of solids. This problem is being investigated in solubility | studies on fission products in the reactor fuel (see Chap. 2.3, ‘‘Chemical Reactions in Molten Salts’’). Interest in the fuel mixture NaF-KF-LiF-UF (H 2-4]-45.3-2 5 mole- %) prompted a redetermi- . nation of its viscosity, since the previous measure- ments 14 were made on a relahvely impure sample S - plummet. The ‘density varied from 1.675 g/cm3 “ at 425°C to 1.50 g/cm .at 770°C and may be - expressed as and over o' narrow temperature range. The vise- - f'cosnty of a h:ghly pure sumple prepared for . these Lol measurements varied from 8.1 centipoises at about o 525°C fo 2.7 centiponses at about 725°C and moy— , be represented by the equehon o , ' # - 00292 e4.‘507/'1" ' o :menis are about 10% lower than the prewous ones. V45, |, Cohen, ANP Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 156. Measurements were also made'3 on NaF-LiF-KF- UF, (10.9-44.5-43.5-1.1 mole %). The viscosity vaned from 8.8 centipoises at about 500°C to 1.8 centipoises at about 800°C and may be represented throughout this range by the equation p o= 0.0348 e4265/T where T is in °K, A modest program has been initiated to measure the physical properties of some fused chloride mixtures. Two instruments have been designed to study the viscosities of these materials. One is a special Brookfield viscometer, which has a spindle assembly that contains two large shear surfaces. This increased shear is necessary because of the low viscosity range in which this class of materials falls. The other instrument is an ‘adaptation of the capillary viscometer currently being used for fluorides. The principal modifi- cation is the use of commercial hypodermic needles for the capillary tubes. This affords a number of advantages, such as ease of replocement and interchangeability of needle sizes to cope with different viscosity ranges. ‘A density measurement was made on one chloride mixture, LiCl-KCI-NaCl (56-4]-3 mole %), by using the buoyancy principle with ‘a balance and « p = 1.885 ~ o.oooso:r . ~ where T is in °,_C. These values are in excellent . agreement with values obtained on very similar S : " mixtures by Van Artsdalen and Yaffe.16 o where T is in °K These vclues are: m sahsfoctory-- o T T Ty - ’ A’-_ngreement with the previous data; the new. measure- '-.7155. l. Cohen and T. N. JoneS,r Measurement of the - Viscosity of Compositions 12, 14 and 107, ORNL CF+56-5-33 (May 9, 1956). 16E, R. Van Artsdalen and 1. S. Yaffe, J. Phys, Chem, 59, 118-27 (1955). 231 et 1650 1600 1550 1500 1450 1400 TEMPERATURE (°F) 1350 1300 1250 i2o0 1150 €0 SYSTEM | : INCONEL - CARBIDE SYSTEM C, ExiT NORTHERN HEMISPK - - NLESS STEEL STIAINLESS STEEL COPPER-- YSTEM € STAINLESS STEEL BORON CARBIDE S - INCONEL CARBIDE HELIUM | 50 40 30 . - 20 10 DISTANCE FROM INCONEL—FUEL INTERFACE (ft x 10™%) =LEORES QORNL—LR~-DWG 13336A INCO INCONEL 1¥0d3Y SSIYD0Yd LDIrONd dNV Fig. 4.1.15, Compurison of Temperature Proflle? in Region Beyond the ART Reflector for Various Systems and a Sodium inlet Tempera- ture of 1175°F, ' . ' . (\ o THERMAL CONDUCTIVITY W. D. Powers Two new devices have been built that are being used to determine ‘the conductivities of liquids. in both systems a measured amount of heat flows down through a horizontal sample of known thick- ness and area. The thickness of the sample may be varied so that contact resistances at the surfaces of the sample can be eliminated.’ One device has been fabricated in which the sample is contained within glass. This makes it possible to observe whether any gas bubbles or films are present in the cell at the time measure- ments are being made. Good agreement with the known thermaleconductivity data for water was obtained. Initial experiments with heat-transfer salts have shown the formation of gas bubbles on the surfaces. The effect of this gas on the thermal conductivity is currently being studied.. One - of the difficulties associated . with the measurement of thermal conductivity is in knowing PERIOD ENDING JUNE 10, 1956 -exactly where the heat is flowing. The second device has two heat meters, one directly above and the other directly below the sample. It was assumed that if the heat flow were the same in both meters, the heat flow through the sample would be well established. The original design did not meet this specification and therefore o modification of the system is now being made. ELECTRICAL CONDUCTIVITY N. D. Greene An electrical conductivity cell has been standard- ized at high temperatures with molten potassium chloride, and the cell constant so obtained agreed within experimental etror with those determined by means of standard, aqueous solutions. Further cell standardization with molten NaCl and ZnCIz, as well as refinements of existing measuring techniques, will precede measurements of the conductivities of fluoride melts. 233 ANP PROJECT PROGRESS REPORT 42, RADIATION DAMAGE S U G, W. Keilholtz ' S . : o EXAMINATION OF THE DISASSEMBLED MTR 42.1 and 4.2, 2. The ;ear"beqrmgs,i'Fag: 423, IN.PILE LOOF NO. 3 appeared to be i |n excellent condition, ¢ A. E. Richt The neck of fhe pump, between fhe pump sumpri C. Ellis W. B. Parsley and the heat exchonger, was parhully filled with E. J. Manthes R. N. Ramsey ' oil; however, no- ewdences of fuel or ZtF =vopor .. E. D. Sims deposits were found. in this region. The slmger___ assembly was “removed | cnd ‘examined, and the - . slingers were found to be coated with a black, No. 3 has been completed except for the straight shiny deposnt th. . 24 No ewdences of fuel sections of the fuel line. Disassembly of the fuel- Z F de ' f d'on the slingers. circulating pump. was found to be more difficult or &r -vqpor eposu s were foun on .e 'gers. S e than had been anticipated. The first operation was the removal of the water jacket from the pump bulkhead. The pump was then sectioned by sawing through the bulkhead. During this cutting operation the oil finger below the forward bellows was exposed, and no evidence of oil was seen in the finger, The pump shaft rotated freely and showed no signs of binding after it was freed from the impeller. The pump shaft, forward bellows, bear- ings, and seals were removed from the bearing housing. The forward bellows was partially filled with oil. The bellows, bearings, and seals were cleaned ultrasonically in perchloroethylene to re- move any traces of oil. The face and inside of the forward bellows contained a brittle, amber- o o R colored, amorphous deposit, as shown in Figs. Fig. 4.2.2. Enlorgement of Depesit Shown in Fig. 4.2.1. 2X. Reduced 25.5%, RMG 1380 _ Metallographic examination .of MTR in-pile loop x Flg. 4.2.1. Front Face of Forward Bellows of Fig. 4.2.3. Rear Bearmg from MTR In-Plle Loop B _ Fuel Pump from MTR In.Pile Loop No. 3. 1/2X. No. 3 Fuel Pump. 2X Reduced 14%. (Secref w:ih Lt (Secretwith-captian) caption) | R T u 234 9 » B UNCLASSIFIED RMG 1383 Fig.4.2.4. Slingers from MTR In-Pile Loop No. 3 Fuel Pump. Note shiny, black deposit. 2X. Re- duced 24.5%. (Seere-f—m-fh—ea-p-h-on-)— UNCL ASSIFIED) - RMG 1384 .F‘\ig. .2 5 Impeller (Top) und Shnger (Bofl'om) Halves of Pump Sump from MTR ln-Pile Loop No. 3 Fuel Pump. 1/2X. Reduced 23.5%. (Seeret.with ~ssption} "PERIOD ENDING JUNE 10, 1956 The fuel in the pump appeared to be inhomo- geneous -in that it was made up of two phases. One phase was green and had the normal appear- ance of the ‘fuel mixture used, and the other phase was black. Samples of both phases were obtained for chemical analyses. The two halves of the pump sump, with the fuel still present, are shown in Fig. 4.2.5. The impeller showed no evidence of erosion, Figs. 4.2,6 and 4.2.7. Charcoal samples were obtained from two of the fission-gas adsorption traps in the purge system, A black deposit that was found around the inlet UNCLASSIFIED -RMG 1384 Fi'g'. 4.2.6. Front Face of Impeller from MTR In- Pile Loop No, 3 Fuel Pump, 1/2X. Reduced 15%. Fig. 4.2.7. Side View of |mpe||ef from MTR In- Pile Loop No. 3 Fuel Pump. 2X. Reduced 23,5%. (s Ly trom} 235 ANP PROJECT PROGRESS REPORT hole to one of the traps is shown in Figs. 4.2.8 - and 4,2.9. The walls of the trap on the inlet side also showed a black deposit around the sides for a depth of approximately 1 in., as shown in Fig. 4,2,10. MTR in-pile loop No. 4 has been sectioned at- NRTS, and the active portion is being shipped to Oak Ridge. Disassembly of the loop will begin as soon as the hot cell equipment has been re- paired and checked out. Priorities have been established for the examination of sections that were removed from the ARE, and these exami- UNCLASSIFIED RMG 1388 Fig. 4.2.8. Inlet Side of Fission-Gas Adsorption Trap from MTR In-Pile Loop No. 3. 1/2X. Re- duced 15%. (S=cret-withocgption) Fig- 4,29, Enlargement of Deposit Shown in Flgo 4. 080 2x Reduced 28 5%. 236 previously,? ‘nations will proceed as rapldly as cell tlme_‘ ullows.. _ _ _ 'mvssflénldfis OF MATERIALS REMOVED ~ FROMMTR IN-PILE LOOPS NOS. 3 AND 4 W E. Browmng C C. Bolta R P Shlelds Operctlon of MTR |n-p||e |ooP No. 3 descrlbed. , was terminated beCQUSe ~a plug formed somewhere in the purge system. Therefore a postoperational inspection of the fission-gas adsorption traps was undertaken. A ‘gas sample was obtained by purging the traps for 2 hr with helium and collecting the effluent fission gas in two refrigerated charcoal trops. A radioassay of the gas is to be made. The purged traps were sectioned, and samples of the charcoal adsorber and tubing were collected for radicassay. An ~extensive black deposit was observed around the “inlet to one trap, and a thin brownish film was - present around the inlet to the second trap. These deposits were photographed (Figs. 4.2,8, 4.2.9, and 4.2.10) and then collected for analysis. The resylts of the analyses of the samples are not yet available, but it is suspected that oil or decompo- sition products of oil entered the traps from the fuel-pump bearing housmg ‘ond sump ond caused lOn assignment from Pratt & Whmiey Alrcroft 2. P, Carpenter et al.,, ANP Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012, p 27. Fig. 4.2.10. Inner Walls of Inlet Side of Fission- Gas Adsorption Trap from MTR In-Pile Loop No. 3 Showing Depth of Formation of Black: Deposlt. 1/2X. Reduced 25.5%. (bevret-witircopticnl 1 al . » " a plug when they came in contact with the refr:ger- ated trap inlet. A study of an apparent chemical change brought about in the fuel during operation of MTR in-pile loop No. 3 is also being conducted. Determinations ore being made of the nickel, chromium, and iron content of the fuel, and radiochemical analyses of the fuel are nearly complete. The analysis of the foreign materials taken from the pump is partially completed. The fission-gas adsorption traps from MTR in- pile loop No. 4 have been received, and analyses of copper tubing sections and of acetone washes of the carbon adsorber are under way. CREEP AND STRESS.CORROSION TESTS J. C. Wilson | W. W. Davis A. S, Olson N. E. Hinkle J. C. Zukas A heat transfer test of the stress-corrosion appa- ratus designed® for operation in HB-3 of the LITR was made in the LITR to check the heating to be expected from the fuel, It was not possible to reach the desired temperature of 1500°F with the 2.5 g of the fuel mixture (No. 41) NaF-ZrF -UF, (63-25-12 mole %) used in this test. A bench mockup which reached the same temperature as that attained in the reactor indicated that only about 70 w of heat was generated by fission. About 200 w of heating was expected from this fuel at the depressed flux measured in -HB-3 of the LITR. For the next test a fused salt with more uranium will be used, or the number of con- ductive fins will be reduced. - ' ' Three more’ fube—burst stress-corrosnon speci- mens were exposed to radiation in the LITR, and two more bench tests were completed. The irrodi- ‘ated specimens_ fmled after 164 1260, and’ 285 hr ‘at 1500°F with: ‘a maximum stress of 2000 psi, . - whereas one specimen bench-tested under similar . test’ condmons ‘had not .failed in about. 935 hr. " The other bench test was made at 1450°F, 50 deg ‘_'—-:iower, and the speccmen rupfured ofter only 500 he . under stress, . _ - “Tubing stock’ of the type fo be used in ART' B 'NaK-to-cnr radmtors is- expected to be available " 'soon ‘for “use in tube-burst and stress-corrosion 3J. C. Wilson et al.,, ANP Quar. Prog Rep. Sept. 10, 1955, ORNL-1947, p 165. PERIOD ENDING JUNE 10, 1956 experiments. Components for the MTR tube-burst test apparatus are being fabricated. One of the specimens tested previously in the MTR tensile-creep-test apparatus? ruptured outside the gage length, and therefore the postirradiation elongation measurement was made on this speci- men. The elongation was 1.7% after 760 hr ot 1500°F in helium; this value is higher than was expected. The corresponding bench tests have still not been completed, because experimental difficulties (mainly lecks in the bellows) required that the apparatus be rebuilt. A pneumatic stressing device for bench and in- pile tensile-creep tests was built and is being leok tested. Two pneumatic extensometers for use with this device have been tested. The pneumatic device will provide a powerful, compact means for stressing creep specimens with loads as great as 10,000 Ib, Parts for a stress-relaxation machine are being built, A pneumatic valve operator has been adapted for stressing the specimen, and a pneumatic strain- controlling system is being developed. MTR STRESS-CORROSION APPARATUS J. C. Wilson C. D. Baumann W. E. Brundage The stress-corrosion equipment formerly desig- noted as the alternate stress-corrosion apparatys was modified to change the platinum heater to a simpler, helically wound heater made from molyb- denum wire. . The design, described previously,’ is completed, and fabrication and assembly have been 'parnally completed on a model with which heat fransfer tests w:ll be mcde in the LITR, 'f DUCT’LITY OF NICKEL BASE ALLOYS - J. C. Wllson "T.C. Prtce . An mveshgaflon is . belng made of the factors \?that govern the ductility of mckel-bcse alloys. - As a part of this study, tensxle tests were run on Inconel .and Nuchrome V at strain rates of 0,002 and 0.2 in./min ot temperatures ranging from 500 to 1400°F At. no _temperature was any | obrupt ‘w W. Dovns, N. E. Hmkle, and J C. WI|son, ANP. 'Quar. Prog. Rep. March 10, 1956, DRNL-2061, p 190, Sw. W, Davns, N. E, Hinkle, and J. C, Wilson, ANP Quar. Prog. Rep March 10, 1956, ORNL.2061, Fig. 8.12, p 191. 50n assignment from Prufl & Whutney Aircraft, 237 ANP PROJECT PROGRESS REPORT ‘decrease in ductility observed in either alloy. The ductility of Inconel remained constant between 800 and 1100°F at the slower strain rate and between 800 and 1200°F at the faster strain rate. - At higher temperatures the ductility steadily in- creased. The ductility of Nichrome V decreased steadily as the temperature was increased from 800 to 1400°F. _ Both alloys showed stress fluctuations at both strain rates, the fluctuations being more noticeable at the slower strain rate, but this may have been the result of the inadequacy of the recording system at.-high speeds. In the tests on Inconel the stress fluctuations were present at tempera- tures from 500 to 1200°F and were largest in number and amplitude between 600 and 950°F. The lower temperature limit of these fluctuations has not yet been determined for Nichrome V, but the upper limit seems to coincide with that of Inconel., _ - . Specimens of Monel, Inconel, and Nichrome V have been prepared for testing in the LITR (and later in the MTR) to determine the effect of irradi- ation on ductility. Out-of-plle tests will be made on the Hastelloys. EFFECT OF RADIATION ON STATIC CORROSION OF STRUCTURAL MATERIALS BY FUSED-SALT FUELS H. L. Hemphill The study of the effect of radiation on the cor- rosion of Inconel capsules containing static fused- salt fuels was continued.” Two additional Inconel capsules containing the fuel mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %) were irradiated in the MTR for six weeks each at ]500°F at a power density of approximately 3.5 kw/ cm3. Two other Inconel capsules containing the fuel mixture (No. 44) NaF-ZfF -UF, (53.5-40-6.5 mole %) are being irradiated at 1500°F at a power density of approximately 6 kw/em3. Further mockup tests were run on Hastelloy B capsules under simulated MTR conditions. A method was developed for mounting thermocouples on the capsules without domaging the chromium- nickel plating required to protect the capsule from the atmosphere. Hastelloy B capsules are being prepared for bench tests and for MTR irradiation. W. E. Browning - W, E. Browning and H. L. Hemphill, ANP Quar. Prog. Rep. March 10, 1956, ORNL.2061, p 192. 238 They will be filled with an NoF-KF-LiF-UF, fuel mixture containing approximately 25 wt % enriched vranium and LiF enriched in Li”. LITR VERTICAL IN- PILE LQOP w. E. Browmng D. E. Guss® H. E. Robertson M. F. Osborne R. P. Shields A vertical in-pile loop, which was eSsenhully the same as the one operated prevuously, was installed in the LITR and put into operation; how- ever, the pump stalled when the reactor was started. By reducing the pump temperature it was possible to operate the loop with the reactor at 500 kw; but at higher power levels the temperature drop in the loop was great enough to freeze the fuel in the cold leg. The pump stalled permanently during an effort to increase the pump temperature ‘gradually so that the reactor could be brought to The specifications for the power total power, full power. characteristics of the loop were: 1.5 kw; maximum specific power, 577 w/em’; dilution factor, 7.64. While the cause of the pump failure cannot be precisely determined until the loop has been examined, it is known, from the sounds emanating from the pump and heard through a microphone, and from varigtions in electric power demands of the motor, that contact of moving solid parts overloaded the motor, The parts that had already been fabricated for three additional loops will be changed to eliminate most of the possible couses of pump failure. The changes will include a shorter, more rigid impeller shaft and improved bearings. Also, pump models are being tested to .determine the effect of in- creasing the cleorance around the pump impeller. When these tests are. completed, existing parts will be modified and assembled for bench and in- pile tests. Although failure of the pump prevented operatlon of the loop, the experiment provided a test for the reliability of other components. Thermocouples, heaters, safety circuits, the temperature control system, the tachometer, the lead wires, aond the entire instrumentation system performed satis- tactorily. Piatinum—platinum-rhodium thermo- 80n assignment from United States Air Force. 9G. W. Keilholtz et al., ANP Quar, Prog.. Rep. Sept. 10, 1955, ORNL-1947, p 164. o). . " < couples were used on this loop to eliminate the severe oxidation that occurred in previous loop assemblies where the Chromel-Alumel thermocouple leads passed through heaters.. The change in thermocouple materials necessitated a redetermi- nation of thermocouple corrections. Three conditions contributing to errors in temper- ature measurement were investigated for the platinum—platinum-rhodium thermocouples used. First, the thermoelectric output of the couples might be changed because of incompatibility with Inconel; second, the junction of the thermocouple is cooler than the Inconel surface to which it is attached because of air cooling; third, the outside surface of the Inconel is cooler than the surface in contact with fuel because of the heat flux through the wall. A compatibility test was run in which the plati« num—platinum-rhodium thermocouples were welded to Inconel and heated for 1000 hr ot 900°C. The thermocouples were found to be mechanically sound after this treatment, and their absolute calibration has changed not more than 1,2°C, Tests were also run to determine the temperature correction necessary because of air cooling of the thermocouple kead. A segment of the loop was mocked up in full scale, with electrical heating . TEMPERATURE DIFFERENCE BETWEEN THERMOCOUPLE JUNCTION AND FUEL TUBE SURFACE (°C) PERIOD ENDING JUNE 10, 1956 instead of fission heating, and the thermocouples were mounted on the fuel tube in the same way that they ore mounted on the loop. Surface temper- atures were observed, with an optical pyrometer, through a quartz window ond compared with the thermocouple readings. These tests were similar to those reported previously,? which yielded different results for different thermocouple ma- terials. The differences are no doubt due to bead geometry differences that result from melting-point differences during thermocouple fabrication. The results of these tests are shown for six thermo- couple assemblies in Fig. 4.2.11. For the oper- ating conditions of the LITR vertical in-pile loop the correction amounts to 15 % 10°C, The correction for the differential temperature through the wall of the loop was calculated by vsing theoretical heat tronsfer relationships in the loop and experimental flux data obtained from previous loop operation in the LITR.'! The heat 'Ala FLOW (efm) ‘°w E. Browning, G. W. Keilholtz, and H. L. br!“" ANP Quar. Prog. Rep. March 10, 1955, L-1864, p 146. "G W. Keilholtz et al., ANP Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 164. UNCLASSIFIED CRNL~LR-DWG {4746 66 S 88 n Lo 40 . .Fig. 42,11, Temperature Diflerences Between Six Thermocouple Junctions and Outer Surface of Inconel Fuel Tube as a Result of Air Flow Around the Thermocouple Junction, 239 ANP PROJECT PROGRESS REPORT transfer calculations of Robinson and Weekes!2 were modified to apply to the presently planned position for the loop in the LITR by using certain simplifying assumptions. The tip of the loop is to be 1.5 in. below the center line of the reactor. ‘This is more than 4.5 in. below the position of maximum flux. The flux measured near the fuel tube in the previously operated loop'! was taken as the depressed flux for that loop. 1t was assumed that the flux depression was the same for corre- sponding parts of the present loop. The difference in shape of the flux profile for the two loop po- ‘sitions’ was ignored in calculating the fuel tube wall temperature - differential. Fission-product poison contributions to the macroscopic cross section of the fuel were assumed to be negligible. ' 12M T. Robmson and D W, Weekes, Design Calcu- lations for a Miniature Hng-Temperatw'e In.Pile Circu. 'latmg Fuel Loop. ORNL-1808 (Sept. 19, 1955). The transmission coefficient for neutrons in the fuel was determined from work reported by Holmes, 13 and it was used in conjunction with the flux distri- bution along the loop, as shown in Fig. 42 12 to determine the total loop power. Finally, to find the temperature dn‘ferenhal through the Inconel fuel tube wall, the curves of heat transfer variations along the loop designated Mark VIIi in the work of Robinson and Weekes'? were modified for the higher total power of the loop being investigated. A plot of the calculated temperature differential through the tube wall vs the position on the loop for proposed operating conditions is presented in Fig, 4.2.13. The maxi- mum combined temperature corrections for all three errors totaled 35°C. > 13p, k. Holmes, Problems of Neutron Population in Localized Absorbers, ORNL CF-56-1-141 (Jdn. 27, - 1956). SL bt BN Lldre ORNL-LR-DWG 14747 3 n o = 2 % -J - from Pl‘flff&VWh‘l.fne.i Alrc}uff. By T Robinson, Solid State Semiann. Prog. Rep. Feb, 28, 1954, ORNL-1677, p 27. PERIOD ENDING JUNE 10, 1956 flection of the measured flux in the plane of its maximum value. The computations were carried out on the Reactor Controls Computer according to the technique previously described,® except that the flux function was generated electronicully instead of mechanically. : Calculations were made for the fuel mixture (No. 44) NaF-ZrFl-UF‘ (53.5-40-6.5 mole %) con- “tained in Inconel in the new position and for the fuel mixture (No. 107) NaF.KF-LiF- UF (11.2- 41-45.3-2,5 mole %) contained in Hasfelloy B in both positions. The nhysical property data used for air, for Inconel, and for the zirconium-bearing fuel mixture were the same as those used previ- ously. The data for the alkali-metal fuel and Hastelloy B are given below: Density of fuel,]7 g/cm3 2.09 Specific heat of fuel, '8 wesec/g:°C 2.28 " Yiscosity of fuel, 19 g/cmesec 0.022 Thermal conductivity of fuel,zo 0.035 w/em°C Thermal conductivity of Hasteiloy B, 21 0.113 w/em«°C ' As before, the depression of the thermal-neutron flux because of the presence of the loop was neglected. The results of the calculations are summarized in Table 4.2.1. It is shown that, in spite of the greater length of the irradiated loop in the ‘“‘deep’’ position, the fuel temperature differential is less than that in the position of maximum flux. The principal result of the change in position of the loop is ' a substantidl -'decreGSe in the dilution g R, Mcmn, F. P, Green and R. s Storie, **An - Appendix on Anglog Simulation,’* in Design Calculations . for a Miniature - Hng-’l‘emperature In.Pile Circulating - . Fuel Loop, by M. - ORNL-1808 (Sept. 19, 1955) Robmson and D, F, Weekes, 7 . 17Es!'imctted by method of $. I. Cohen and T N "~Jones, A Summary of Density Measurements on Molten Fluoride Mixtures and a Correlation - Useful for Pre- dicting Densities of Fluoride Mixtures. of Knoun 'i'Composztzons, ORNL 1702 (Moy 14, 1954). 18Eshm«:n'ed on assumption that the alkalismetal fue! " has ‘the same molar heaf capacny as that of the zir- . conium-bearing fuel. : . ‘95 l.: Cohen, personal communi cations. to M, T 2O stimated by W. D, Powers, March 16, 1956. 21 Haynes Stellite Company, Hastelloy High-Strength, Nickel-Base, Corrosion-Resistant Alloys, p 15, Sept. 'I 1951. 241 s v e e ANP PROJECT PROGRESS REPORT TABLE 4.2.1. SOME CALCULATED DESIGN PARAMETERS OF THE VERTICAL IN.PILE LOOP DESIGNED FOR OPERATION IN POSITION C-46 OF THE LITR . Fuel Temperature Maximum Air Fuel Loop Reynolds Numbers Differential Temperature,. Air Flow Position Fuel Air ©C) ©C) Rate (gfm) NaF-ZrF -UF, Shallow®® 1,500 32,000 135 380 4 (53.5-40-6.5 _ 3,000 22,000 65 410 30 mole %) 6,000 20,000 33 510 28 ' 10,000 20,000 33 520 28 Deep® 1,500 74,000 103 250 100 3,000 50,000 56 450 70 6,000 44,000 30 530 60 10,000 42,000 17 550 59 NaF-KF-LiF-UF Shallow? 1,500 19,000 125 450 27 (11.2-41-45.3 3,000 16,000 65 550 23 mole %) 6,000 15,000 K 580 21 ' ‘ 10,000 15,000 20 600 21 Deep® 1,500 32,000 94 380 45 3,000 26,000 50 410 36 6,000 24,000 27 460 33 10,000 23,000 16 480 32 %Tip of loop at position of maximum thermal-neutron flux, bBosed on earlier calculations. 12 “Tip of loop at bottem of active lattice. factor from about 11 in the *“‘shallow’’ position to about 6 in the ‘‘deep’’ position. FAST-NEUTRODON DETECTORS D. Binder A great variety of radiation-damage studies re- quire a knowledge of the fast-neutron flux of the reactor being used for the irradiations, For instance, changes in the mechanical and electrical properties of metals and insulators are dependent on the shape of the neutron-energy spectrum. As the neutron energy increases, the number of atoms displaced increases until a critical energy is reached at which the effect saturates.??2 There- fore, it is essential to know the number of neutrons per unit energy up to about five times the critical energy and the integrated flux above this energy. The energy spectra in two types of reactors from thermal to 100 ev or 1 kev and from 0.7 to 8 Mev 22G, H. Kinchin and R. S. Pease, Reps. Progr. in Phys. 18, 152 (1955). 242 have been investigated by Trice.23:24 For the remaining investigation, that is, the l-kev to 0.7-Mev region, a detector must be developed with which to study the variation of displacements with neutron energy. The development of useful de- tectors is dependent on a knowledge of this vari- ation, and vice versa. The first step therefore is to calculate a known case based on simple as- sumptions. Germanium was selected as the moterml to be studied because it is sensitive to neutrons and the relationship between the solid-state effect and the number of displacements is known. From the work of Cleland et a.2% at low temperatures, it is known that five electron traps are introduced per 23;, , Tnce, Fast Neutron Flux Measurements in E.25 of the Brookbaven Graphite Reactor, ORNL CF- 55-7-130 (July 27, 1955). 24; B, Trice, A Series of Thermal, Epithermal and Fast Neutron Measurements in the MTR, ORNL CF- 55-10-140 (Oct. 28, 1955). 255, W, Cleland, J. H. Crawford, Jf, and J. C. Pigg, Pbys. Rev. 98, 1742 (1955). " o) " n ‘*fast’’ neutron in a graphite reactor and that there are two traps per displacement.: The question is whether it is possible to calculate the number of displacements on the basis of a reasonable graphite reactor spectrum and to achieve approxi- mate agreemenf with the experimental result of 2,5 displacements per fast neutron. (A fast neutron is defined?’ by Cleland et al. to be a neutron with energy greater than that required to displace an atom.) ‘For the calculation the method of Kinchin and Pease?? was modified to include a spectrum which is the sum of 1/E distributions extending to each ‘element of a fission spectrum,2® The agreement between the calculation and the experimental re- sult was within a factor of 4. The spectrum and the variafion of displacements with energy are then roughly correct, or at least there are com- pensating errors, ' In order to test the variation with neutron energy alone, it is plonned to use neutron sources of known energy. The flux from these sources is reduced by factors of 10° to 105 from graphite reactor fluxes, but the sensitivity of pure germanium allows for experiments even ot these low levels. Two sources will be used: a uranium converter plate in a thermal column and an antimony-beryl- lium source. The converter plate gives a pure fission spectrum, and, when used in the slanting animal tunnel of the ORNL Grophlte Reactor, emits 2 x 108 fission neutrons/cmZ.sec near the center of the plute. A first irradiation of 5.8 x 10'3 neutrons/cm? introduced 1.9 x 1075 acceptors/cm? in a germanium crystal. This experimental value agrees to within a factor of 2 wuth the coiculufed | value of 3.4 x '!0I ‘The. result of the firsf expenment mdlcafes fhot'“ ‘the present ossumphons on ‘the variation ‘of d:s—'.:_’. placements with neutron energy are roughlymrrecf. L - This will be checked by &u_’the_r irradiations with fission neutrons and with 30-kev neutrons'fr'on"l:-«-"' ~the antimony-beryllium photoneutron source. More - - adequcte neutron monitoring is to- be used for - ~ future germanium irradiations. This ‘may lead to a _ complete unclerstcmdmg of dlsplocements |n ot . leost one solld-state reaction. S 1'6D Bmder, Solid State Semiann, Prog. Rep Feb, 29, 1956, ORNL-2051, p 48. - PERIOD ENDING JUNE 10, 1956 . VISCOMETER FOR REMOTE MEASUREMENTS OF THE VISCOSITY OF IRRADIATED FUSED-SALT FUELS ' C.C. Webster A viscometer was designed for use in a hot cell which will be capable of measuring the viscosities of irradiated fused-salt fuels at temperatures up to about 750°C in an inert (helium) atmosphere within a shielded cell. The apparatus consists of two vessels, 6 in, long, 1.25 in. OD, mounted in a vertical position, with 2 in. between centers. An Inconel tube 3 ft fong and bent into a U-shape ‘connects the bottoms of the vessels. A small ferromagnetic ball travels with the liquid through the tube, and the position of the ball is detected at two points on each |eg of the U-tube by elec- tronic equipment. Swagelok connections are provided at the top for inserting liquid-level probes and for allowing helium pressure to be applied on the surface of _the liquid in either vessel. The whole apparatus ‘is inserted into a 4-in.-dia vertical furnace with an 18-in. controlled-temperature heated zone. End ~heaters are provided to control the heat loss out the ends of the furnace. EFFECTS OF RADIATION ON ELECTRICAL COMPONENTS J. C. Pigg C. C. Robinson?7 Insulation 28,29 In previous experiments on the effect of _radiation on electrical insulation the insulated _. wire to be irradiated was placed in the reactor as " if it were a sample lead. Care was taken to ensure - ~ that possible feakage paths from the central con- - ‘ductor to ‘the shield were as long .as posmble.' Whenever practicable, the end of the insulation . - that was in the reactor was sealed over at the end ‘of the central conductor. However, the possibility of conductlon by end surfoce contamination and ionization of the air still existed when there was no seal or when the seal was defective. In order to determine whether such leakage was important in the. expemnent:s prev:ously conducted, g 27On uss:gnment from Wright Alr Development Center. 28 C.Pigg and C, C, Robmson, Solid State Semzarm. Prog, Rep. Aug. 30, 1955, ORNL.-1945, p 6. 2%, c. Pigg et al., Communication and Electromcs 22, 717 (Jaonuary 1956). , 243 i e s b ANP PROJECT PROGRESS REPORT two samples were taken from the same spool of Teflon-insulated wire, One sample was the usual length, 25 ft long, and the second was 50 ft long. The two samples were installed at the same time in hole 50-N of the ORNL Graphite Reactor. The 25-ft Teflon-insulated sample was installed in the same manner as in the previous experiments, with no seal at the end. The 50-ft sample was doubled back on itself so that both ends of the wire were outside the reactor. The leakage between the shield ‘and the central conductor was measured to be twice as large for the 50-ft sample as for the 25-ft sample. The photo-emf measurements on the two wires were identical. It can therefore be concluded that the radiation effects described previouslyzs‘ were due to bulk properties rather than to end surface conduction or to air ionization. Barriers A 1IN3BA germanium point-contact rectifier was exposed in a 2 x 10%-r/hr Co®® gamma-ray source, and the forward and reverse currents were measured at l-v bias. The behavior of the current is shown in Fig. 4.2.14. These data are in qualitative agreement with those obtained by Young.so, The increase in reverse current, followed by a return toward its initial value, is an effect not found in . reactor irradiation experiments because it occurs at low total damage. As may be seen from the change in forward current, the effect illustrated in Fig. 4.2.14 would have occurred before the sample 30R, c. Young, Gamma Radiation of Crystal Diodes, ;Jrlglht A)Il‘ Development Center, WCRT-TN-54-255 (Dec. 8, 1954). UNCLASSIFIED ORNL~LR=—DWG 13755 CURRENT AT 1 v BIAS IN38A-B o REVERSE CURRENT, uo ® FORWARD CURRENT, mo X107 {N3BA-A A REVERSE BIAS; NORMALIZED TO COMPARE WITH {N38A—-B TS Ad o 1 2 3 GAMMA EXPOSURE {r) 4 5 6 : T 7{xt0") Fig. 4.2.14, Effect of Gamma Radiation on Conduction Through a Barrier. Germanium point?co;-:tfict rectifier IN38A-B exposed to 2 x 108 t/hr from Cob? source; 1N38A-A exposed to 2.5 x 105 r/hr 1from Cob9 source. 244 i " b reached thermal equilibrium in .an-in-pile experi- ment. Whether or not the effect does result from neutron irradiation has not yet been determined. A series of exposures were also made in a 5 x 10%-r/hr Co®? source, and the voltage-current characteristics of the rectifier were measured after the samples were removed from the source. The change in the characteristic curve can be seen in Fig. 4.2.15. There is little change in the forward curve, which is in agreement with the data in Fig. 4.2,14. The current at l-v reverse bias was normalized to make the preirradiation value coincide with that of the sample exposed in the 2 x 10%r/hr Co®® source. The normalized data are shown in Fig. 4.2.14. UNCLASSIFIED ORNL-LR-DWG 13756 2 —— BEFORE EXPOSURE —— AFTER FIRST EXPOSURE TO # x 40° r —— AFTER SECOND EXPOSURE TO 242 x1C° r —-— AFTER THIRD EXPOSURE TO 7.77 1408 ¢ [: ] M=967 21077 H o REVERSE CURRENT () FORWARD CURRENT {(ma) o B M=a4x1077 2 Me=3ax07 2 =18 %407 0 . — — 0 o oz 0.4'_ 06 08 W0 24 VOLTAGE (v) S Fig. 4 '.2 A5, Effect cf Gclmmu Rcdinhcn on ’-,'-Vohage-Currenr Churucierishc of Germnnium Pointe. : _""Contaci Rectifier IN38A-A at 20°C.; 11\e M vclues_‘ ' ' fj:nre the slopes of the Imes. L - ' f The forwurd chcracterrsflc corresponds to fhe B S ;_'?volfage-current behcvuor of the emitter of a tran- 0 sisters . The “reverse ‘curve . corresponds o fhe:fi” L :collectcr chcrcctertsflc of a transistor. It may be - seenin_Fig.4.2.15 that, ‘olthough the emitter: " characteristic s little” changed,”. rhe collector ~ characteristic changes drastically both in mugmrude : and slope. The magnitude of the current at l-v bias changed by a factor of about 6, and the slope PERIOD ENDING JUNE 10, 1956 changed by a factor of about 5.3. The changes in the characteristics are similar to those caused by neutron irradiation,3! A series of transistors were irradiated in the circuit shown in Fig. 4,2.16 to measure the ampli- fication of the unit during, as well as before and after, irradiation. The series resistances in the emitter and in the collector circuit were changed to match the unit being considered in each experi- ment. Provisions were made to switch from grounded-emitter to grounded-base circuitry so that “both types of circuitry could be studied simul- taneously with the same unit. Exposures were made both in Co%® gammoa-ray sources and in the ORNL Graphite Reactor, The changes in the collector current at a 2-v bias and in the amplification of a Minneapolis- Honeywell H-2 power transistor as a function of gomma irradiation may be seen in Fig. 4.2.17. The collector, base, and emitter voltages were adjusted to the original values before each reading. Hence, the changes in amplification were due to changes ‘in characteristics and not to changes in operating ~ voltages, It may be seen that the collector current and the amplification were essentially the inverse of each other. Whether the change in amplification was due to a change in the slope of the collector characteristic, a change in separation of the collector curves, or a change in impedance match- ing has not yet been determined. The changes in amplification of two Minneapolis- Honeywell - power transistors as a function of irradiation .in hole 51-N of the ORNL Graphite Reactor are shown in Fig. 4.2.18. These curves illustrate the same type of behavior as that ob- o served for- RCA transistor No. 1241.32 The tempo- <~ - rary increase in umpllfucanon observed previously - in RCA transistor No. 1266 was not noted. - Again, '-_.'the flux: Ievel was - such that this region of the ‘eurve would have been traversed before thermal - ethbrwm was obtalned in the m-plle experlmenf. = lnfluence of an Elecmcul Field on Drffusion ' oflntersimul Aioms L When germcnwm is bomborded by hrgh-energy _,,neutrons the tecoils from the neutron-atom col- - *;';.iltswn cnd secondary collrsrons cf fhe orngmal 3!J C. P:gg, Solid State Semzann. Prog. Rep. Aug. ' '31 1953, ORNL-1606, p 81. 3 c. Pigg and J, W, Clelund Solzd State Semzann. ' Prog. Rep. Feb. 10, 1953, ORNL-1506, p 47. 245 ANP PROJECT PROGRESS REPORT UNCLASSIFIED © ORNL-LR-DWG 13757 Q Vs I _ 1.0 ] t i it AW 3 Re - Z1000 |K {0-TURN HELIPOT o @ | L 0-50pa _ Jfi‘éfl: = 0-10v DC T T 5K 10-TURN - . HELIPOT . . b GROUNDED GROUNDED EMITTER BASE 5K 10-TURN :I____ HELIPOT ) = 0-150v DC -L%J HOv-60~ o—4 ] 9) Fig. 4.2,16. Transistor Test Circuit. atom as it travels through the lattice result in the formation of lattice vacencies and interstitial atoms. Since the germanium lottice is fairly open, the interstitial atoms can diffuse readily through the crystal, Because of the high dielectric con- stant of germanium, the interstitial atom is ionized at room temperature, and hence its diffusion should be influenced by the presence of an electric field. Fields of the order of tenths of a volt per micron are available in the barrier region of o diode, and the conductivity of the dicde is extremely sensitive to changes in carrier concentration in the region of the barrier. A 1IN38A point-contact germanium rectifier was exposed for 5 min in hole 51-N of the ORNL Graphite Reactor. After removal from the reactor the rectifier was biased in the forward direction 246 at 0.7 v in order to remove the potential gradient in the borrier region. The measured current was low because of radiation damage. The unit was then placed in an oil bath to maintain a constant - temperature of about 25°C for six days. At the end of the six-day period there had been no observable change in the current passing through the sample at 0.7-v bias. The bias was removed for 2 hr, then restored. When the bias had been restored, it was noted that the current had in- creased by about an order of magnitude. After eight days in the biased condition, the current had not changed. When the bias was removed, an electric'field was established at the barrier to cause the relative interstitial atoms to diffuse from the barrier region, This removal of interstitial atoms from the p-type " -} &) Ll UNCLASSIFIED ORNL—LR-DWG 44749 05 ' | 100 COLLECTOR CURRENT: DC DROP - Y ACROSS 1000 ohms AT 2 volts BIAS 04 & o> P00 O - 80 ho | . 2 03 _ ~ 60 _ b— i b E . /"’ ] E = . -~ t e ~T 1 ~ oz [ - 40 : AMPLIFICATION: AC DROP ACROSS RESISTOR IN COLLECTOR LEG 0.4 20 0 0 0 1 23 456 7 8 910 #1213 14 15(x10%) ' RADIATION DOSAGE (r/hr) Fig. 4.2.17. Effect of Radiation on Collector Cutrent and Amplification of a Minnefi:polisr-Honey- well H-2 Power Transistor kradiated in Gamma- Ray Source of 2,57 x 105 r/hr. PERIOD ENDING JUNE 10, 1956 side of the barrier results in annealing out the type of damage which tends to make the p-type side less p-type and transporting it to the n-type side. Hence both sides of the barrier are annealed toward their initial condition and the forward current increases toward its initial value. RADIATION DAMAGE TO BOR{ON CARBIDE J. G. Morgan The study of radiation effects on B,C, initiated previously,®3 was completed. The apparatus used for the irradiation of samples in the LITR was described in the previous report. The first samples irradioted were hot-pressed, high-density 8,C, 0. Sisman _ supplied by the Norton Company, and slsp-casf B 4C bonded with SiC, supplied by The Carborundum - Company. Four other types of B, C have now been _irradiated. The results of spectrographic analyses 330, Sisman et al., ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 204. UNCLASSIFIED ORNL—LR—DWG {13759 1.0 0.9 0.8 AMPLIFICATION: A-C DROP ACROSS RESISTOR IN COLLECTOR LEG 0.7 06 2 - Lo "2 0.5 & ] . . - E\r 4 1 e 1 - _ " , 204 N(?.A;_fLU* _ltBGx 0 nv — 03— . — _ 02 T . T R b - - ..“'u-. o ‘“*o--,._____ : e L b ‘ 7"“"-3-- — — o NO.G;"-FLJX186x!0‘°nv Sl f Sttt e e S0 b— L : - - ' - : - | e o4tz o3 4 "B 8 7T B 9 .0 . 92 3. 14 .45 16 17(x10% INTEGRATED NEUTRON FLUX (avt,) Flg. 42 18 Chunges in Amplificuiion of Two aneapohs-Honeywe" H-2 Power Trcnsustors as a Function of Integrated Fast-Neutron Exposure. 247 ANP PROJECT PROGRESS REPORT of the six types of B,C are presented in Table 4,2.2, and the resvlts of examinations of 20 irradi- ated samples of the six types are given in Table 4.2,3. The four types of B,C irradiated durlng this quarter are described below‘ 1. The Norton Company supplied low-density, high-purity samples that had been hot-pressed from 325 mesh, and finer, powder and fired at above 2000°C. The boron content of this material was 62,4 + 0.4 wt %. These bodies were hard, strong, and impervious. : 2. The Norton Company also supplied low- density, technical-grade samples that had been hot-pressed from 60-mesh, and finer, material aond fired at above 2000°C. These bodies were hard, strong, and nearly impervious, and they contained 62.4 + 0.4 wt % boron. 3. The Carborundum Company supplied tiles (No. 1) cast from a basic mixture of 90% B,C grains and 10% 200-mesh silicon metal. The cast bodies were fired at 1350°C and refired at 2000°C, They were porous and friable, ond they contained 64.7 + 0.5 wt % boron, 4, The Carborundum Company also supplied tiles (No. 2) cast from a basic mixture of 80% B,C, 10% 200-mesh, and finer, boron-metal powder, and 10% 200-mesh, and finer, silicon-metal powder. The cast bodies were fired at 1350°C and refired at 2000°C. They were porous and friable, and they contained 67.8 % 0.4 wt % boron. - | The samples were exposed for 800 hr in the "LITR in a thermal flux of 2 x 10'3 nv. The sample temperature was less than 200°C, and the total dose received by each sample was 6 x 10'° nvt (thermal). These conditions are to be compored with those expected in the ART in which the temperature of the B,C layer will be 1450°F and . the integrated neutron flux dosage for 500 hr of operation will be 1.2 x 10'? nvt (exposure behind - a 0,030-in. gap in the cermet layer). The average burnup of these samples was 2.9%, and most of the burnup occurred in the first few thousandths of an inch of the surface layer. ' A much greater exposure will be requnred to simulate the ART operating conditions for the copper-B,C cermet layer and the boron steel, and fherefore these materials will be irradiated at the MTR at a temperature of 1600°F. Calculations indicate that these samples should receive total . doses of 1.5 x 1020 nvt for the cermet ond 8 x 1017 nvt for the boron steel, which are equivalent to TABLE 4.2,2. RESULTS OF SPECTROGRAPHIC ANALYSES OF B,C SAMPLES Composition (%) Norton Norton Norton Carborundum Coarborundum Cerborundum Element* High-Density Low-Density Low-Density Sample Cast Tile No, 1 Tile No. 2 Hot-Pressed Technical-Grade High-Purity and Bonded Refired at Refired at Sample Sample Sample with SiC 2000°C 2000°C Al 0.01-0.1 1-10 0.1-1 0.1-1 0.1-1 0.1-1 Ca 0.001-0.01 0.001-0.01 0.001-0.01 0.001-0.01 0.001-0.01 0.001-0.01 Cr ek *x b 0.01-0.1 0.01-0.1 0.01-0.1 Cu *x 0.01-0.1 *h *k >k ' Fe 0.01-0.1 0.1-1 0.1-1 1-10 0.1-1 1-10 Mg -0.0001-0.001 0.00t-0.01 0.0001-0.001 0.001-0.01 0.001-0.01 0.001-0.01 ‘Mn *k 0.01-0.1 >k 0.01--0.1 0.01-0.1 0.01-0.1 Ni *k *x ** 0.1-1 0.01-0.1 0.01-0.1 Ti rx 0.1-1 0.1-1 0.1-1 0.1-1 0.1-1 Zr *x 0.1-1 0.1-1 1-10 1-10 1-10 **Sought but not found. 248 *Limit of detection: Cr, 0.004%; Cuv, 0.0bOI%: Mn, 0.001%; Ni, 0.02%; Ti, 0.04%; Zr, 0.03%. N o ®) PERIOD ENDING JUNE 10, 1956 TABLE 4.2.3. RESULTS OF EXAMINATION OF IRRADIATED B,C SAMPLES Bulk Density Dimensional Specimens Gas ~Appecrance Change ‘ Change fan* Evolution Change (g/cm3) (in./in.) Three Norton Co. high-density 0 £0.4 em? None 0 * 0.01 0 1 0.001 (2.49 g/cms) hot-pressed samples Three Norton Co. low-density 0 * 0.1 emS None 0 1 0 * 0.007 (2.02 g/cma)** technical-grade ' samples Three Norton Co, low-density 0 £0.1 cm® None 0 £0.05 0 *0.001 (2.17_§/cm3) high-purity samples ’ Three Carborundum Co. samples 8.1 cm3/g (av) None, black cast and bonded with S$iC deposit on container Carborundum samples too Two samples of Carborundum Co. 0 0.1 em® None porot.ts and f':i able for tile No. 1 refired to 2000°C | precise density or dimensional measurements Six samples of Carborundum Co. None None tile No. 2 refired to 2000°C 3 *Theoretical production, approximately 1 em”. **Bulk density; sample slightly porous, two weeks aond one week, respectively, in the MTR at a thermal flux at 1.5 x 1074 nv. Irradiations have been initiated on a 7.6% CaB,-92.4% Fe plate clad with stainless steel. Studies of a 10,3% BN-89.7% Ni plate clad with type 304 stainless steel are planned. IRRADJIATIONS OF STRESSED SHIELDING MATERIALS ‘ 4. C. Wilson . W. E. Brundage | W, W Daws B A test apparatus has been demgned for |rrad|ot|on of a 1 wt % boron (Bw)—stamless steel alloy under stress in the LITR ot 1300 and 1600°F, "In these : tests one sample will be irradiated while stressed_ in compression at 500 psi ot each temperoture, and- - one sample will be stressed similarly at the lower - temperature out ‘of .the reactor.” One unstressed - specimen will also be irradiated ot each tempero-"..’;" ture, ‘The irradiation periods will be three to six weeks., The data obmmed will be compared with . available data on” MTR irradiations of boron— stainless steel alloys for Westinghouse (WAPD). - Specimens of an austenitic-stainless-steel-clad copper-B4C cermet (6.6 wt % B,C) are also to be irradiated in the LITR at 1300, 1600, and 1700°F. The samples will be checked for dimensional stability and hardness. THE EFFECT OF RADIATlON ON POLYMERS et 0. Sisman W. W. Parkinson W. C. Sears There has been a continuing program in the Solid State Division to study the effects of high- ‘energy radiation on plastics and elastomers. This - work has heretofore been reported only in the ~ Solid State Division progress reports, but because of increasing interest in this work in connection with allied ANP projects and because there is an increasing - need . for - radaahon-damoge data on _organic- materials, some of the current polymer - work “will now also be presented in the ANP - progress reports. The plastics ancj elostomers have been studied ‘quite extensively, _and some of the chemical re- actions have been identified which. cause the . changes that' have been observed in tfie physical ‘properties of irradiated polymers. A study is currently under way of the reaction products of irradiated polymers. A summary of this work, for 249 ANP PROJECT PROGRESS REPORT which the infrared spectrometer is used, is presented here. The infrared spectra of polystyrene, polyethylene, polybutadiene, GR-S, natural rubber, deproteinized rubber, polyvinyl chloride, and Teflon have been measured before and after irradiation. The con- ditions of irradiation of these materials and some others which were not studied so thoroughly after irradiation are given in Table 4,2.4. The polymer films were cemented on 0 2.9 x 1.1 cm rectangular aluminum wire frame, evacuated for 2 to 5 hr, and then measured to provide a preirradiation spectrum. The films to be irradiated in vacuum were evacuated for three to five days at 0.2 ¢ ond sealed in vacuum in pyrex or quartz tubes prior to irradiation. These sample tubes were packed with aluminum foil in aluminum cans for irradiation. The polymers were irradiated in one or more of the following facilities: (1) a Cob? gamme-ray source, (2) water-cooled hole No. 19 in the ORNL Graphite Reactor, and (3) lattice position C-46 in the LITR, The maximum permis- sible dosage was limited by the formation of open slits in some polymer films (caused by shrinkage and loss in film strength). It was found that the dosage on polystyrene, polyvinyl chloride, mylar, and polymethyl methacrylate could be increased by irradiating unmounted films. The discovery that postirradiation oxidation had occurred in irradiated polymers during exposure to air necessitated opening the evacuated somple tubes in a helium atmosphere. A glove box that TABLE 4.2.4. IRRADIATION DOSAGES OF POLYMERS Irradiation Exposure (X IDB rads) Polymer Cob0 Gomma, .\,]06 o/ he sz,h"z Reactor, LIBTR, Averc'.ge Sample 105 «/hr, ~108 ¢/hr, Thickness In Oxygen In Vacuum in Vacuum in Vacuum (in.) Polystyrene® '} | 2.3, 11 1.5, 16, 23, 31, 35 1000, 1000, 1500 0.0018 Polyalphamethyl styrene® 0.39, 6.9 6.6° 7.0 0,002 Polyethylene® 0.39, 3.9 1.2, 2.3, 6.2 3.4 0.0012¢ Polybutadiene® 3.8, 6.1, 8.9 4.0, 14, 18 0.003 GR-S¢ 0.72, 6.5 3.8, 6.1, 8.9 3.8, 13, 20 0.002 Natural (Hevea) rubber® 1.4, 8.9 4.8,/ 16 0.004 Extracted Hevea® 1.4, 8.9 16, 22 0.004 Polyvinyl chloride® 0.39, 0.72, 3.9 3.9, 5.2, 8.9 12, 40° 0.002 Polymethyl methacrylate® 0.39, 6.9 6.5% 2.0, 6.5 0.0004 Tefion® 0.72 3.8, 6.1 16, 22 | 0.0006 Mylar® 0.39, 6.9 2.3, 6.2 16, 17 0.00025 Nylon€ resin No. 63 0.72 1.4, 6.1 22 C 0.0015 “Supplied by Dow Chemical Company. Sample crumbled; measurement after irradiation impossible. “Commercial film, 9A1s0 0,008 and 0.030 in. exposed to 0.12 X 107 rads. €Supplied by B. F. Goodrich Co. Tube cracked during irradiation, 8Supplied by E. I. du Pont de Nemours & Co., Inc, 250 o) » could be evacuated was used for this operation ond for transferring the specimens to a gas-tight infrared absorption cell for spectral measurement. The absorption cell was then opened in air, and measurements of the spectrum were repeated after various periods of exposure to the atmosphere. As an example of the spectral changes observed in polymers, the spectrum of a polystyrene sample is shown in Fig. 4.2,19. While polystyrene is more \ | UNCLASSFIED ORNL-LA-OWS 10939 ®o T T T T 1 I gr . FEe0 — wl POLYSTYRENE N N BEFORE IRRADIATION ] = 20 |- — — — o i ] 1 to 00 T T T T ] g - ] = 60 - . ] - 100 x 10° RADS =] or MEASURED N HELIM 7] b . o . ! . 00 L ot gw - &0 ] 5 SAME SAMPLE 40 AFTER AR EXPOSURE -1 FOR 1T HOURS - + 20 - - ° 00 y . & 80 - §‘° ; 20 — 0 L ] + 4000 3000 2000 . 800 600 1400 1200 K00 800 s00 Fig.r4.2.l 9. Infrared Spectra of Polyflyrene. ; PERIOD ENDING JUNE 10, 1956 resistant to change by radiation than most polymers are, the changes that may be seen in Fig. 4.2.19 are typical of the alterations of the infrared spectra by irradiation. Dosages of the order of 108 10 1010 rads were required to produce significant changes in the infrared spectra of most polymers. - A large postirradiation effect was observed in the spectra of polystyrene, GR-S, and natural rubber, Upon exposure to air following irradiation in vacuum, oxidation products continued to form for periods of up to 95 days. These reactions were indicated by the growth of strong hydroxyl and carbonyl bonds. The oxidation products formed in postirradiation oxidation were different from those produced by irradiation in oxygen. Irradiation of polystyrene in vacuum to high doses produced a wholesale ‘disruption in which both the aromatic and aliphatic components were equally affected. On the other hand, the aliphatic component of GR-S (a styrene-butadiene copolymer) showed a greater percentage change than did polybutadiene at equal dosages. All polymers showed significant changes in the double-bond regions as a result of irradiation. In polyethylene, RR,C=CH, groups disappeared as trans RCH=CHR, groups formed. In GR-S and -polybutadiene the number of terminal vinyl groups decreased and the unsaturation in the hydrocarbon chains of GR-S ond of natural rubber was de- creased. lrradiation increased the number of trans ‘RCH=CHR, groups in natural rubber, as it did in polyethylene. Conjugated and unconjugated un- saturation was produced in polyvinyl chloride. 251 ANP PROJECT PROGRESS REPORT 4.3. FUEL RECOVERY AND REPROCESSING 7 | H. K. Jackson ' . D. E. Ferguson W. K. Eister H. E. Goeller -VOLAT!LI;I'fY PILOT PLANT DESIGN AND CONSTRUCTION R. P. Milford F. N. Browder The design of the volatility pilot plant for re- covering fused-salt fuels is complete except for the molten-salt sampling device for the fluorinator and -the trap-door closing device for the waste-salt car- rier. - All major equipment items were received, as well as the electrical power and control center and the instrument panelboards. The equipment, with the exception of the ARE fuel hold tank, which is not required until later, is in place; and piping, electrical work, ond instrument installation are proceeding rapidly. It is expected that the plant will be completed by June 30. NICKEL FLUORIDE SLUDGE FORMATION STUDIES G. |. Cathers M. R. Benneit The behavior of NiF, in molten NaF-ZrF, and NaF-ZrF ,-UF, systems was studied to determine whether the presence of this corrosion product would cause the formation of sludges which would interfere with salt transfer in the fluoride-volatility process. In order to determine the solubility of NiF, in the salt mixtures, NiF, was added to mol- ten NaF-ZrF , (50-50 mole %), and the mixture was heated until a clear solution was obtained. it was then cooled until turbidity reappeared. Solubility values estimated by the disappearance of turbidity were in fairly good agreement with solubility values determined electrochemically?! for the solvent NaF- Zrf , (53-47 mole %). Based on the visual deter- minations, the estimates of the solubility of NiF in NaF-ZeF ; (50-50 mole %) were the following: Temperature Solubility of NiF2 (°C) (wt % NiF ) 640 0.7 670 1.0 685 1.3 . E Topol, ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 89. 252 These values show the very definite increase in solubility with temperature when compared with the reported solubility of 0.2 wt % Ni at 600°C. The addition ofas much as 6 wt % NiF, to molten NaF-ZrF4 (50-50 mole %) at 600°C resulted in the formation of a viscous dispersion which was fairly stable, and thus other tests were made to determine the effect of concentration on sedimentation. These tests were made by dry mixing the required amount of salt (~30 g total) and melting it in a %.in.-ID nickel tube. Nitrogen was used initiolly for agita- tion and then as a blanket while the material was kept at 600°C for various times. . The tube was quickly quenched with cold water at the end of the test to fix the NiF, concentration at various heights in the tube. The tube was then cut into '/2-in.-|ong sections, and the salt was analyzed for nickel. Although some settling of NiF, was evi- dent after only 0.5 hr when 2 wt % Ni was added (as NiF,), complete settling had not occurred even after 72 hr (Table 4.3.1). When only 1 wt % Ni was added, settling was more nearly complete at 72 hr; since the NiF, concentration was lower, the in- crease in viscosity ot the bottom was not so great and thus was not so much of a deterrent to settling. In further experiments the sedimentation of NiF in molten NaF-ZrF -UF , (48-48-4 mole %) at 600°C (Table 4.3.2) was similar to that found in the vranium-free system. However, with 2 wt % Ni and with uranium present, the settling was less after 2 hr than in the test with no uranium. Since the solubility of NiF, in NaF-ZrF# is near the lower of the nickel concentrations found in these settling tests, it appears that the solubility of NiF, is ap- proximately the some in uranium-bearing and ura- nium-free mixtures. These experiments have dem- onstrated that NiF, concentrations of up to 2 wt %, more by a factor of 10 than is expected in qircraft reactor fuel reprocessing, would not interfere with salt transfers unless the molten salt were permitted to stand unagitated for long periods of time. DECOMPOSITION OF UF6-3N0F COMPLEX G. |. Cathers R. L. Jolley Some exploratory work was carried out on the decomposition of the UF,.3NaF complex at high " » » ". PERIOD ENDING JUNE 10, 1956 TABLE 4.3.1. SEDIMENTATION OF NiF, IN MOLTEN NuF-ZrF4 (50-50 mole %)-AT 600°C Nicke! Concentration (wt %) ~ Relative Position After : " After - After - After bf Sampling ' S TRETE Initially 0.5 hr 2k 8 hr 72 hr Inttial Nickel Content* — V2 wt % 1 (top) 1.60 0.28 0.20 2 1.74 0.73 0.30 0.22 0.20 3 1.72 1.72 2.20 1.40 0.21 4 178 1.86 2.62 © 3.48 1.23 5 (bottom) 2.13 2.02 3.02 3,54 2.99 Initial Nickel Content* — 1 wt % 1 (top) 0.22 2 0.30 0.16 3 0.31 0.17 4 1.71 0.18 5 (bottom) 3.06 *The nickel was added as Nin. TABLE 4.3.2. SEDIMENTATION OF NiF, IN MOLTEN NaF-ZrF ~UF , (48-48-4 mole %} AT 600°C Nickel Concentration (wt %) Relative Position Initial Ni Content — 2 wt % " Initial Ni Content = 1 wt % Initial Ni Content — 0.5 wt % of Sampling After 2 hr After 48 hr After 2 hr After 48 hr After 48 hr 1 (top) 1.26 2 123 0.25 0.36 0.18 0.24 3 2.12 034 040 0.20 0.33 4 278 275 283 0.47 0.22 281 6.94 308 0.85 55'(b6fiom) ' | ,_’temperatures in an effort to develop a UF 6 desorp-. - tion procedure ‘which avmds UF decomposmon. ' The decomposmon reaction would lead to uranium | “being- held - up on the NaF bed ond wouid thus - necessntofe a subsequenf recovery step. Tests 'were conducted by saturating an NoF bed with UF - at 100°C ond then heating the bed, as @ closeg - sysfem, at the desired temperature for 1'to 4 hr. In - some of the tests the bed was subjected to as much as 55 psia UF pressure during the high-tempera- ture cycle. HoWever, no significant correlations ‘were apparent as to the effect of UF, pressure or length of time of treatment., The residual uranium - content in several tests at 400°C varied in the range 18 to 26%. At 300°C the decomposition ef- fect was much less, a residual ‘uranium content of 2% being produced. Use of excess F, (10 psia) in ‘tests at 400°C also resulted in less decomposition, with residual uranium contents of 8 to 14%. The residual uranium present in the product of all 253 A ANP PROJECT PROGRESS REPORT runs was pentavalent. ~ As a result of dispropor~ tionation of the pentavalent uranium in the analyti- cal procedure, however, the tetravalent uranium content in each case was approximately equal to - the hexavalent content. The decompasition reac- tion is therefore believed to be UF ;-3NaF —> UF ;-xNaF + %F, 254 These results indicate that, if a significont par- tial pressure of UF, is retained in the NaF bed because of a plugged line or cold trap during UF desorption, excessive UF, decomposition will oc- cur when the temperature reaches the 300 to 400°C range. Therefore precautions must be taken to ensure that full sweep-gas flow through the NaF bed is maintained during UF ; desorption. " o PERIOD ENDING JUNE 10, 1956 4.4. CRITICAL EXPERIMENTS A, D, Callihan CRITICAL EXPERIMENTS FOR THE \ COMPACT-CORE REFLECTOR-MODERATED REACTOR E. Demski' J. J. Lynn W. J. Fader! D. E. McCarty D. A. Harvey! E. V. Sandin! D. Sc ofi The study of the NDA2-proposed, sodium-cooled, reflector-moderated reactor with solid fuel ele- ments® has been completed. In the critical as- sembly the fuel region contained 0.004-in.-thick vranium sheets interleaved between aluminum and stainless steel sheets, This fuel region was separated from the beryllium of the island and of the reflector by stainless steel shells. As described previously,* additional uranium, in the form of 0.01-in.-thick disks, was added -in one section of the fuel region to provide excess re- activity for other measurements. An evaluation of the effect of this local nonuniform fuel distribu- tion on the reactivity was made by replacing 2636 g of U235 in 0.004«in.-thick sheets with 2678 g of U235 in 0.01-in.-thick disks in another section of the core. With the other materials unchanged, there was a reactivity loss of only 19 cents, | Evaluation of Stainless Steel Shell The loss in reactivity caused by the stainless steel shells was determined by substituting alumi- num shells of the same dimensions. The exchange of 4.4 kg of atuminum for ll 9 kg of stainless steel resuhed in o gain in reactivity of $4.20, estimated to be equwcient to o 14% decrease in ~the critical ‘mass. The excess reachvn‘y was partly compensafed for by the .removal of some - " of the outer ]cyer of the “/ einethick reflector. - e Removal of a 2/8-|n.-th|ck sectuon that extended" !On nssignmenf from Prctt & Whitney Aircrnft. . o 2Nuclear Developmant Corporoflon of Amer!co. 3CCR-2 A Compact Core Reactor for Azrcraft Pro-g- | pulszon. NY0-3080 (July 30, 1954), Qudrterly Progress " "Report, - ANP Developmeflt- Oct. -1°w"Dec.” 31, 1953, "NDA-20 (Jan, 23, 1956). | 4A. D. Callihan et al., ANP Quar, ng. Rep. March 10, 197536 ORNL-206), p 64; Dec. 10, 1955, RNL 2012, P over the 283/-|n. length of the outer reflector and comprised 70% of the outer cylindrical layer re- sulted in o loss in reactivity of $2.39. Measurements of Gamma-Roy Heating in Beryllivm Capacitive ionization chambers were used to measure the distribution of gamma-ray heating in the beryllium of the island and the reflector of the critical assembly by a method developed at the Knolls Atomic Power Laboratory.> The chambers are constructed of beryllium and have a 10-mil- thick annular cavity 5’16 in. OD and '78 in, deep, The results are expressed as power dissipated as heat in a unit volume per unit reactor power. The reactor power was determined from a calibration based on the intensity of a fission-product gamma ray from an irradiated uranium foil.% A plot of the data obtained from radial traverses 3/4 and 109/‘6 in. from the mid-plane of the reactor is given in Fig. 4,4.1. Three longitudinal traverses, one along the axis of the reactor and the others 4'5’ and 7/ in. from the axis, are plotted in th. 4.4.2, These data have not been corrected for the ionization resulting from the (n,p) reaction in the air-filled chambers and may overestimate ‘the heating adjacent to the fuel by as much as 25%, an estimate based on the work at KAPL. An attempt was made to measure this error by fitling the chambers with CO,, but it was apparent that they were not gas-tight during these experi- ‘__ments._ It may be possible to maoke a correction “to these data by using the results of similar “measurements in ancther assembly. -A layer of beryllium, 2/8 in. thick, was removed from the top of the reactor, and one traverse was ‘repeated with this thinner reflector, = The heating - was observed to be unaffected in the region be- ~tween the fuel and a point 3 in. from the surface - of the modified reflector, - In this outer 3-in. layer " the heating decreased to a value, ‘ot the surface, - about 35% less than it was at the same,d:stonce B from fhe fuel in the thrcker reflector. - ' SC ‘A. Rich and R. E. Slovocek Gamma-Ray Heatmg gd;sas.s')urements in the .S‘IR PPA- 18 KAPL-866 (Jan. , 6S. Snyder, Absolute Determination of Power Produced ;rg a]hslg;nmally Zero Power Reactor, ORNL-2068 {May 255 ANP PROJECT PROGRESS REPORT BTN T ORNL-LR-DWG 44448 20x1077 /- 1.8x 1077 34‘ in. . FROM MIDPLANE 1621077 1451077 L/ 1.2x40°7 est— BERYLLIUM ISLAND —weteest— FUEL REGION —— it \ BERYLLIUM REFLECTOR 1.0x 40°7 8.0x {078 GAMMA-RAY HEATING (w/emS-w) AT | A / 6.0x 4078 e 10%¢ in. FROM MID PLANE (% in. FROM FUEL ANNULUS) N\ 401078 \ 20x 1078 - \\ —-> 0 0 2 4 6 8 2 1 6 18 20 22 DISTANCE FROM REACTOR AXiS {in.) Fig. 4.4.1. Radial Distribution of Gamma-Ray Heating in the Compact-Cu;e Reflector-Moderated- Reactor Critical Assembly, Fa stheufion Leakage | Relative measurements of the fast-neutron leak- age at points on the outer surface of the reflector were made with a Hornyak button, 2 in. in diameter and ‘/4 in. thick, mounted on @ Du Mont 6292 photo- multiplier tube. The button had the same compo- sition, 0.15 g of powdered ZnS in 1,0 g of Lucite, as a similar button described by Hornyak.” By proper choice of a discriminator bias it was possible to detect neutrons with energies above 7W. F. Hornqu, Rev, Sci, Instr. 23, 264 (1952). 256 0.5 Mev against the gamma-ray background of the reactor, : : . S A polar plot of the fast-neutron leakage distribu- tion observed in a fongitudinal traverse over the top of one end of the reactor is presented in Fig. 4.4.3. The observed counting rates, nor- malized to the counting rate at point 19, near the - axis of the assembly, have be’e'rjl“ plotted on the polar radii drawn through points on the reflector surface where measurements were made.. For this traverse, the reflector thickness was approximately ”'/2 in., and the top layer of beryllium wczs,2834 . . (x108) PERIOD ENDING JUNE 10, 1956 seCRET™ ORNL~LR—DWG 14419 20 g ' \5/@1. FROM AXIS 16 _ _ " DISTANCE FROM REACTOR MIDPLANE (in) = '2 ¢ nfi “—.—:._‘:\- AXIAL 3 N g . = - g 10 o ‘ . I X . g \, P ‘\ % 8 \\. " g \. . . Y\ N, N, | . N~. . \‘ ] | 73/,6m FROM AXISN \ \.\ |t CORE REGION —————————— ~ "'-:-.\[’\ | R AR N TS S 0. 0 2 - 4 6 B 10 g2 14 16 18 20 Fig. 4.4.2. Longitudinal Disfribuhon of Gammu-RcyHeafing in the Compact-Core Reflector-Moderuted- Reacfor Crificcl Assembly. 257 ANP PROJECT PROGRESS REPORT FRcRETe ORNL-LR-DWG 14420 10 9 |— ] 8 MEASURED WiTH 2-in.-DIA HORNYAK BUTTON, — TOP BERYLLIUM LAYER: 48%¢ x 28%, in. 7 ] " €6 — o 1 2 3 a4 \ 5 6 7 — 3 4 3 1 5 ) y 5 ) ey ' T ¥ ¥ T T T 5 | | 5 2 I 5 8 > 5 — t 7 - c | o - l BERYLLIUM 5 = REFLECTOR : w ¥ Z 4 b l - 5 - I <+ 10 & 2% in. 4+ i 6 3 Iill + 2 — 3 oy 44 Q [} -:_Lé—_—— EI 3.___ o\. | 7 {2./ | 15 \. ' 9 2 | B FUEL REGION ¢ .’/ 1. © . 10 I T {6 . 47. | i1 — 4147 _ BERYLLIUM ISLAND " 18 AXIS T 19 0 —_ e e e e e e L 20 20 Fig. 4.4.3. Polar Distribution of Fast-Neutron Leakage at Surface of llé-m.-Thlck Reflecfor of Compact-Core Reflector-Moderated-Reactor Critical Assembly. 258 . i & in. long, 18! ]ffi in. wide, and 2/ in. thick. The beryllium layer immediately below it was 34 R long and 24/16 in. wide, The top layer and the corners of the second beryllium layer were subsequently removed, and the resulting reflector was 8% in, thick. The new top layer was ]8”/6 in. wide and 3434 in, long, A second longitudinal traverse was then made. The results are shown in Fig. 4.4.4, in which the top lobe is plotted to a scale one-half that used to plot the end lobe. Two series of lateral traverses were also made at the side of the reactor at several distances from the mid-plane. For reflector thicknesses of 'Hl/2 and 8% in, the outer slab dimensions for each side reflector thickness were the same as those given above. The latitudinal variations of the fast-neutron leckage for the two reflector thicknesses are shown in Figs. 4.4.5 and 4.4.6. The curves show a slight asymmetry about the center of the reactor, which is attributed to the reflection of neutrons by the aluminum and steel structure on which the critical assembly rests. The area under each curve of Figs, 4.4.5 and 4.4.6 is proportional to the counting rate observed at the corresponding distance from the mid-plane in the longitudinal traverses made ot the top of the reactor., By integrating the results of the two longitudinal traverses over the distance from the mid-plane, it was possible to determine that the removal of the 27/ in. layer of beryllium from the top of the reactor mcreased the fast-neutron leak- age there by o factor of 3.7. ' RELATIVE INTENSITY (arbitrary units) o PERIOD ENDING JUNE 10, 1956 SECREY. ORNL-LR-DWG 14424 10 I, 9 [— ] 8 — MEASURED WITH 2-in.-DIA | " HORNYAK BUTTON TOP BERYLLIUM LAYER: 18 X 3434 in. 7 pom —_ NOTE: THE DATA AT THE TOP OF THE REACTOR ARE PLOTTED TO A SCALE ONE-HALF THAT USED FOR PLOT OF THE END I 3 |- €, ' o (2345 6/7g o ! BERYLLIUM ] REFLECTOR o4 I FUEL / REGION 43\' MIDPLANE 7 1 2.'73 in.——-l I-- -BERYLLIUM ISLAND *20 0% AXIS Fig. 4.4.4. Polar Distribution of Fast-Neutron Leakage at Surface of _8"‘/a-in.-Thick Reflector of Compact-Core Reflector-Moderated-Reactor Critical Assembly, 259 ANP PROJECT PROGRESS REPORT ORNL—LR~DWG 14422, 1.5 . . MEASURED WITH 2-in.-DIA HORNYAK BUTTON »'/ \ OUTER BERYLLIUM LAYER: 18'/16 x 28 %4 in. / \ ' - d = AXIAL DISTANCE FROM MIDPLANE OF REACTOR / Cin. ‘ \ S N { < g o // (2) A’i = Ag ? then b - -5 -5 (3) Gpl()\l,E,Q) = sz(Az,E,Q) . The significance of Eq. 3 can be realized if A and G are reduced to conventional units. The relation between distance r (measured in centimeters) and A is given by (4) X = No(Eg)r where N is the nuclear density of the medium. The ‘conventional particle current can be defined as F(7, Efl) given in particles per unit energy at energy E per unit solid angle in direction ( per cm? at position 7. The relation between F and G is then F(r,E,Q) 2.2 Noy(E o) = GIA(),E,Q - 266 By using Eq. 4, Eq. 2 becomes -» Nl-& Py o r ——— ——— = r :.-.——T 2 N2 ]7 p2 1 (6) and by using Eq. 5, Eq. 3 becomes (7) F- - Py o Efi N2 T .1’2 . 2 F GLED) =——F, (s ED) N2 Fo,(r B = 2 PR i | 2 which is the desired transformation.” Proper inte- gration of Eq. 7 gives the flux transformation .as : ’ 2 (8) ‘i’p 2 (-;2 = Pi ., o " o S ruE) =— ¢’p (rp_E) and the _dose rate tronsformati.on as "2 (9 D - [ TS r - ) ps ry = —p—z- )= Dp](r,) . These transformations can be applied to much of the TSF data, but the application is not -entirely general,?2 To make full use of the transformations it will be necessary to obtain additional data at the TSF at several separation distances so that the measurements can be interpolated and applled at any altitude. : ENERGY ABSORPTION RESULTING FROM GAMMA RADIATION INCIDENT ON A MULTIREG[ON SHIELD WITH ' - SLAB GEOMETRY ' S, Auslender3 The code for a Monte Carlo calculation of energy deposition in a multiregion shield with slab geometry“ has been used to obtain the results for 1-Mev gamma rays incident on a slab consisting of regions of fuel, Inconel, sodium, and Inconel again. A diagram of the composite slab is shown in Fig. 5.1.1, which ‘gives the normal thicknesses in 3On assignment from Pratt & Whitney Aircraft. 43, Auslender, ANP Quar, Prog. Rep, March 10, 1956 -) h L & seencr 2-01-059-67 4 -Mev GAMMA RAYS FUEL\!NCONEL SODIUM INCONEL 7 7 NaF~ ZrFy-UF, / / (52.5-42.5- 5.0 mote %) ) % % J NN 2 N\ 379\ \V05767 2149 0488 ™~ {cm) NNV o NN\ Fig. 5.1.1. Geometry of Fuel-lnconel-Sodium- Inconel Slab, ' st (mfp) centimeters ond in mean free paths. The fuel is NoF-ZrF .UF, (52.5-42.5-5 mole %). The per- centages of energy absorption throughout the slab for various angles of incidence of gamma rays are given in Fig, 5.1.2, The percentage of the total energy incident on the slab that is reflected, ab- sorbed, or transmitted for each angle of iricidence is shown in Table 5.1.1, and the percentage of the energy thot is absorbed in each of the four regions of the slab is shown in Table 5.1.2, A plot of the dose buildup factor as e function of the distance through the slab, in mean free paths (mfp), is presented in Fig. 5.1.3. The sharp break in the curves at 2,75 mfp is due to the rapid attenuation of low-energy gamma rays (degraded principally in the sodium) by the final Inconel slab, PERIOD ENDING JUNE 10, 1956 SESREP 2-01-059-88 ISCTROPIC 0.5 ENERSY ABSORBED (% OF TOTAL ENERGY) 0.2 o4 - 005 002 0.01 0 5 10 15 20 25 30 #, NORMAL DISTANCE INTO SLAB {¢m) " Fig. 51.2. Percentage of Total Energy from Incident leMev Gamma Rays That Is Absorbed in a Fuel-lnconel-Sodium-Inconel Slab as o Function of Normal Distance into the Slab, 267 ANP PROJECT PROGRESS REPORT TABLE 5.1,1. PERCENTAGE OF TOTAL ENERGY FROM INCIDENT 1-Mev GAMMA RAYS THAT IS REFLECTED, ABSORBED, OR TRANSMITTED IN A FUEL-INCONEL-SODIUM-INCONEL SLAB 0, Angl.e of Percentage of Total Energy Incidence (deg) ' Reflected Absorbed Transmitted 0 1.151 88.01 10.79 45 2.45 93.36 4.19 60 - 4.10 94,30 1.59 70} - 8.76 90.60 0.635 75!5 ' 13.22 86.05 0.732 TABLE 5.1.2. PERCENTAGE OF TOTAL ENERGY FROM INCIDENT l-Mev GAMMA RAYS THAT IS ABSORBED IN EACH REGION (R;, R,, Ry, AND R,) OF THE FUEL-INCONEL- SODIUM-INCONEL SLAB Percentage of Total sec 0 {0 = Angle Energy Absorbed of incidence) Rl R,y R, R, ] 37.86 29.40 15.04 5.752 V2 48.39 30.97 10.65 3.423 2 ' 60.23 26.10 6.14 1.895 3 69.7 17.32 2913 0.697 4 726 10,52 2,144 0.692 268 €TCRE o 2-01-059-89 : 8 = 70% deg DOSE BUILDUP FACTOR . §=60 =45 deg 8 =0 deg 0 4 2 3 #, NORMAL THICKNESS THROUGH SLAB {mfp) Fig. 5.1.3. Gamma-Ray Dose Bui'ldup Factor as o Function of Normal Thickness Through the Fuel- Inconel-Sodium-Inconel Slab. S i -PERIOD ENDING JUNE 10, 1956 5.2 LID TANK SHIEL DING FAClLlTY R. W. Peelle STUDY OF ADVANCED SHIELDING MATERIALS W. R. Burrus! J. M. Miller W. J. McCool?2 ~ D.R. Otis? - J. Smolen? An extensive mockup experiment was initiated in which combinations of advanced shielding ma- terials, such as lithium hydride, depleted uranium, zirconium, and tungsten, are being investigated. These tests are important to the ANP program because lead and water, the prototype shielding matérials‘ used in most of the previous Lid Tank Shielding Facility (LTSF) mockup tests, are not optimum -aircraft construction moterials. In the initial tests the shielding materials being studied are combinations of lithium hydride and other shielding matericls immersed in transformer oil. The particular configurations used have been those of immediate interest in the GE-ANPD program, but such general interest surrounds the use of these materials that the more reliable data obtained thus far are presented here. This report is preliminary, however, in the sense thct little analysrs work has -been performed. : _]_On assignment from U. S. Air Force, 20n assignment from Pratt & Whitney Aircraft, On assignment from Convair, San Diego. '/4 in~THICK AI PRESSURE PLATE o - Mgein~THICK Al PLATE .~ e |/ Yhe-in~THICK; 21% ENRICHED um sounce vs-mqmcx BORAL SHEET -~ . - - / ~in. -mncx Al SOURCE~PLATE COVER R , f 'fwn -Tmcx ou. LAYER ™ REc:-:ss OF TANK WALL \ Thermal-neutron flux and gamma-rcy tissue dose- rote measurements have been made for conf:gura- tions involving combinations of zirconium, lead, and depleted uranium with lithium hydride and transformer oil. All the measurements were made in the oil along the axis of symmetry of the LTSF source plate. (Fast-neutron dose-rate traverses were also made, but experimental difficulties pre- vent the publication of reliable data at this time.) A typical configuration (No. 69-6a) is shown in Fig. 5.2.1. The various material combinations studied differed only in the region to the right of the line marked ‘'beginning of configuration.” The parameters of all the configurations are given in Tableé 5,2.1, and the known properties of the various materials are given in Table 5.2.2. The lithium hydride is the only material that was - ¢anned in an extraneous material. - All the radiation levels indicated in plots of the data (Figs. 5.2.2 through 5.2,9) are quoted per watt of effective source plate power, based on a total power of 5.5 £ 0.5 w for the 28+in.-dia source plate. A final evaluation of the effective power may ne- cessitate a small change in.the results presented here. All the thermal-neutron fluxes reported are equal to the neutron density times 2200 m/sec. The decreasing magnitude of the slope of the neufron traverses shown in Fig. 5.2.2 at large GEGRETF 2=01~QE7T—-69—272 S TR SN N %S L T e N. ¥ -"§§ 2 sl == T : L N - T 4 1 4 - —_—— NS N s (58 IS 8 / T T \ 'S : A 3 =8 SRHS % Nz ¥ 2p8 152 L X DETECTOR NS N 2Eg =2 08 POSITION NN e iEE Z _ posiTion - NE W "g’g- shTRR T " SOy “Ehg {2 ' - §5 Q;‘ EEa 1R TR £ -] ENER\ B TR n TR S8 NN TS LEOT XIS - - - - — St \ e NVA e Pl c RHNR ' \ N .//'k TE F ol o"‘:“’":’:’o’ooo R NN - TN RSRRRRIRRRS ] U TTBOURCE | .. 07 .+ Fe-BEGINNING OF CONFIGURATION . .~ ...~ - " PATE e g - _ . - - . Lo T~ K . T L. ’ - ,,'- . e : -7, ~ Fig. 5.2.1. Typical Configurution for LTSF Mockup Tests of Advanced Shielding Materials. 269 e e e ANP PROJECT PROGRESS REPORT TABLE 5,21, SUMMARY OF THE CONFIGURATIONS USED FOR LTSF MOCKUP TESTS OF ADVANCED SHIELDING MATERIALS - Configuration " Ne. ' . Composition 69-0 Pure water Transformer oil 69-1 1 #t of LiH in oil 2 £+ of LiH in oil -3 ftof Li_H in oil 69-2 4 in. of Zr in oil : 4 in.of Zr+ 1 £t of LiH in oil - - 4 in..of Zr + 2 ft of LiH in oil 4. in. of Zr + 3 ft of LiH in oil 69-6 * 3 in. of Pb in oil . " 3 in.of Pb+ 1 ft of LiH in oil '3 in. of Pb+2 ft of LiH in oil 3'in. of Pb +3 ft of LiH in oil 69-7 3 in. of U in oil : S 3m. of U+1 ftof LiH monl 3in. ofU+2 ft of LiH in oil 3 in. of U +3 ft of LiH in oil distances from the source is interpreted as an effect of photoneutron production in C1? and deute- rium, although no quantitative analysis has been made. At smaller distances from the source, these data may be used to estimate the macroscopic re- moval cross section of lithium hydride at room temperature in an oil medium. A preliminary value of 0.12 em~! was obtained, which is in good agree- ment with the value expected on the basis of pre- vious removal-cross-section measurements on a lithium-metal slab.4 The observed gomma-ray tissue doses are given in Figs. 5.2.6 through 5.2,9. It can be noted that 270 TABLE 5.22. PHYSICAL PROPERTIES OF THE SHIELDING MATERIALS TESTED _ Material ' Description Density, 0.87 g/cm3 at 20°_C; onalysis, 86.7 wt % C and 12,7 wt % H 5x5x%1 ft slabs encosed in Al cans (/-m.-thu:k wulls), density, obout 0.75 g/crn ; purity, about 95% - Transformer oil Lithium hydride* Zirconjum* 52 x 55 x 2 in. metallic slabs . Lead 55 X 60 X 1.5 in. me-tallic slobrs_ Uranium 52 X 55 X 1.5 in. depleted me- tallic slabs contolnmg 0.24 wt % Y233 ' *Raw material s furnished by GE-ANPD. one slab of lithium hydride behind a heavy shield- ing material reduces the dose at larger distances from the source. This is interpreted as a reduction of secondary gamma-ray production caused by the presence of Li® in the lithium hydride. The data obtained upon subsequent additions of LiH indi- cate, as expected, that LiH does not have so large a macroscopic gamma-ray absorpflon coefficient as that of the oil. It is planned to continue these studies of combi- nations of lithium hydride with other shielding materials. Attempts will also be made to obtain reliable fast-neutron measurements, where pos- sible, for the configurations already studied. 4G. T. Chapman and C. L. Storrs, Effective Neutron g;m?;gé)Cross .S‘ectums for Shielding, ORNL-1843 (Aug. - ;TR s 2-04-057-69- 264 10 BASED ON SOURCE PLATE POWER OF S5 w nv,, THERMAL—~NEUTRON FLUX (neutrons/cm? . sec.w) PURE H,0 PURE OIL CONFIGURATION 69-~0 faft OF LIH CONFIGURATION 69- -4 {2 ft OF LIH 3t OF UH 10~2 5 DISTANCE, FROM SOURCE PLATE: (cm) Fig. 5.2.2 Thermul-Neutron Flux Truverses for Conf:guruhons 69-0 cnd 69-1. n v, THERMAL-NEUTRON FLUX ( neutrons /cm?- sec - w) . 40 60 80 100 1220 440 160 PERIOD ENDING JUNE 10, 1956 oEenTY 2~-01-057—-69-268 in, OF Zr PURE OIL 2 4in. OF Zr AND 1 ft OF 4in. OF Zr AND 2t OF 4in. OF Zr AND 3 ft OF O 20 40 60 - 80 100 120 140 160 23, DISTANCE FROM SOURCE PLATE (cm) “Fig. 5.2.3. Thermal-Neutron Flux Traverses for Configuration 69-2. 271 ANP PROJECT PROGRESS REPORT SEeRp™ 2-04-057-69-271 2 BASED ON SOURCE PLATE POWER OF 55 w v, THERMAL~NEUTRON FLUX (neutrons/cm2-sec:w) . N 0! 3 in. OF Pb AND 4 ft OF LiH 2 3in, OF Pb AND 2 ft OF LiH -2 3in. OF Pb AND 3 ft OF LiH 103 0 20 40 60 80 400 120 440 160 2, DISTANCE FROM SOURCE PLATE {em) o Fig. 5.2.4. Thermal-Neutron Flux Traverses for Configuration 69-6. . ' : 272 omeRET . 2-01- 05769266 BASED ON SOURCE PLATE POWER OF 55 w 3in.OF U 3in.OF U AND 1t OF LiH A Vo, THERMAL~NEUTRON FLUX (neutrons/cm?+sec * w) 3in.OF U AND 2t OF LiH 3in, OF U AND 3 ft OF o 20 40 60 80 100 120 140 160 Zsyr DISTANCE FROM SOURCE PLATE (cm) Fig. 5.2.5. Thermal-Neutron Flux Traverses for | Configuration 69-7. cy c SEONGT 2-01-05T-69-265 BASED ON SOQURCE PLATE POWER OF 5.5 w 3 £ OF LiH 2 ft OF LiH 69-4 69-0, PURE OIL €9-1, ¢ 1t OF LIH 2 69-0, PURE WATER GAMMA-RAY TISSUE DOSE RATE (ergs/g-hr-w) 5 2 w0 20 40 €0 80 400 120 140 160 %, DISTANCE FROM SOURCE PLATE {cm) Fig. 5.2.6. Gamma-Ray Tissue Dose<«Rate Trav- erses for Configurations 69-0 and 69-1. wpreney 2-01-057-69-269 '03 PURE OIL 4 in. OF Ir 4 In. OF Zr ond § f1 OF LIH - 4in. OF Zr and 2 ft OF LIH. = ~4in. OF Zr and 3 ft OF LIH gl " BASED ON SOURCE PLATE-. o ~ -~ POWER OF 5.5w .'\"GAMMA'-R'AY TISSUE DOSE RATE (ergs/g-hr-w) 20 - 40 80 - 80 - 400 . 420 - 440 460 jlso; DISTANCE FROM SOURCE PLATE (cm) S Fig. 5:2.7. Gamma«Ray Tissue Dose 20 2 g \ (F;ONFI(‘JURATlloN 16!) l // 2 16 RN 167 / E DN /] / 2 42 s:"‘ e 7 // "-__-_- § o6 \\ 7~ = . / T & 68a | T~ [=] H O —m 2 = "’\\ o 12 4 16 48 20 22 24 26 28 30 32 34 SPACE POINTS IN CORE Fig. 5 4.3. Compaflson of. quer Distribution in Core - of CFRMR with That ‘in __,.SMC for Various 282 POWER DISTRIBUTION (NORMALIZED) can be obtained in the SMC core power distribution (Fig. 5.4.4) by adjusting the fuel in the five regions from the center out in the proportions of 2:2:3:3:4 and 1:2:3:4:5. These calculations are not final, but they mdlcate that the proper ‘mockup of ‘the CFRMR radlchon can be obtained i in the SMC BESREF 2-01-059-758 4.0 T T T T T T ‘ ; 100 -FUEL-PLATE CORE WITH SPACES ‘BETWEEN FUEL PLATES FILLED WITH IZSZ H20 ;(57 nlu ] CONFIGURATION 273: RATIO OF MASSES OF _ U0z IN FIVE REGIONS OF CORE: 2:2:3:3.4; n o 3.0 o, TOTAL MASS, 10.5kg Y2a3s, W CONFIGURATION 275 RATIO OF MASSES OF UOp:1:2:3:4:5; TOTAL MASSES, O 1.8 kg Y235, '9@ . P —r——: - - S % C2 N 273l 275__4La o st CONFIGURATION 273 o % 18 20 22 24 26 28 30 32 34 36 38 SPACE POINTS IN CORE Fig. 5.4.4, Effect on SMC Power Distribution of Varying the U0, Mass in Five Regions of the Core. Gamma-Ray Sources Another calculation is being carried out at Pruh‘ & Whltney to determine the importance of each region as a gamma-ray source. The reactor is being divided into shells and the gamma-ray intensity from sources in each shell is being determined in a line-of-sight attenuation calculation. This is being done both for the CFRMR and the SMC as a basis of comparison of the two reactors. The remaining region of importance that requires - some further work is the heat exchanger. As pre- viously mentioned the source from the circulation _ of the fuel is not present. Previous LTSF data! indicate that the gamma-ray source resulting from circulation of the fuel contributes approxlmofely 30% of the gamma-ray dose rate outside the reactor shield. It is expected that further anolyms of the TH, Woodsum, ANP Quar, Prog. Rep. Marcb 10, 1956, ORNL-2061, p 237. B o ¥ C. C B e w . fission-product gamma-ray data réported by Zobel and Love? will enable a calculation to be made, 2y, Zobel, T. A. Love, and R. W. Peelle, ANP Quar, Prog. Rep. March 10, 1956, ORNL-2061, p 250. PERIOD ENDING JUNE 10, 1956 by the Monte Carlo method, to determine the gamma- ray dose rate resulting from the circulation of the fuel at least as well as the rest of the radiation from the SMC can be measured.