||\||!|.l||-!;|!| | I|i| Illli,r!l',l!.i:'.l:lT:“|!I'.||'.|I|||'i\|| fl |II|_||'! llllllllfl.\l'.lil' Wl OOCUMERT E flLLrE Gl;i flflfl RY oY 3 4yy5k 03b437L O ORNL=-3124 UC-25 — Metals, Ceramics, and Materials “ s 0 - INOR-8-GRAPHITE-FUSED SALT COMPATIBILITY TEST R. C. Schulze W. H. Cook R. B. Evans lll J. L. Crowley CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON If you wish someone else to see this document, send in name with document and the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S5. ATOMIC ENERGY COMMISSION $1.00 Printed in USA. Priee __~ ., Available froem the Office of Technical Services Department of Commerce Washingten 25, DL C. LEGAL NOTICE This report wos prepored as on oceount of Government sponsored work., Neither the United States, nor the Commission, nor any person acting on behalf of the Commission: A. Mokes ony warranty or reépresentation, expressed or implied, with respect to the accuracy, completenass, or usefulness of the informotion cantained in this report, or that the use of any Infermation, opparatus, methed, or process disclosed in this report moy not infrings privataly awnad rights; or B. Assumes any licbilities with respect to the use of, or for domages resulting from the use of any information, apporatus, method, or process disclosed in this repart. As used in the obove, ""person octing on behalf of the Commission'" includes any employee or contractor. of the Commission, or amployese of such contractoer, 1o the extent thar such employes or controctor of the Commission, er employee of such contracier prepores, disseminates, or provides access to, any information pursuent te his employment or contract with the Commission, or his employment with such centractor. ORNL-312k4 Contract No. W-T405-eng-26 METALLURGY DIVISION INOR-8-GRAPHITE-FUSED SALT COMPATIBILITY TEST R. C. Schulze and W. H. Cook Metallurgy Division R. B. Evans, IIT, Reactor Chemistry Division Je L. Crowley, Reactor Division DATE ISSUED jUN 1 - 1961 OAK RIDGE NATIONAIL LABORATORY Cak Ridge, Tennessee Operated by UNION CARBIDE CORPORATION for the U.5. ATOMIC ENERGY COMMISSION TR 3 Y456 O3L437L O e 5 1 A AR B8 O 0 AL Do b8t i 500+ e i ms INOR-8-GRAPHITE-FUSED SALT COMPATIBILITY TEST R. C. Schulze, R. B. Evans, III,l Je Le Crowley,2 and We. He Cook ABSTRACT For the purpose of evaluating the compatibility of graphite and INOR-8 in a dynamic fluoride fuel medium, INOR-8 Forced Convection Loop No. 9354-5 was operated 8650 hr. The loop operated at a maximum temperature of 1300°F and circulated a fluoride fuel of the system LiF—BeFE—UFM. Post-test examinations of the graphite and loop components revealed no apparent corrosion or carburization problems. INTRODUCTION Design studies have indicated that potentially large gains in the con- version or breeding ratioc of molten fluoride reactor systems can be realized 3 by the incorporation of a graphite moderator and/or reflector. Such a concept requires that graphite, together with other structural materials comprising the reactor vessel, be in direct contact with the molten fluoride fuel mixture. Problems which potentially arise from this arrangement are penetration of the pores of the graphite by the molten salt, carburization of the structural material, and possible reactions between impuritles contalined in the graphite and the molten salt. lReactor Chemistry Division 2Reactor Division 3Report on Fluid Fuel Reactor Task Force, TID-8507, pp. 92-8L (Feb., 1959). Since graphite is inherently porous,u a strong probability exists that this material will be infiltrated by the fuel salt. There are four major reasons why this penetration, if severe, would be detrimental: (1) increased reactor fuel inventory, (2) danger of hot spots in the graphite assembly, (3) fission product retention, and (&) spalling of the graphite because of differential thermal expansion between it and the fused salt. This fourth condition could arise if the salt were allowed to freeze in the pores of the graphite and were subsequently melted. An additional problem resulting from the incorporation of graphite relates to the carburization and resultant embrittlement of the structural material. Presently, the structural material that has shown most promise for use with molten fluoride systems ig INOR-8, whose composition is shown in Table 1. Studies up to this time have shown INOR-8 to be susceptible to carburization when placed in static sodium-graphite systems at temperatures as low as lEOO"F.(ref 5) However, static tests containing INCR-8, graphite, and salt mixture LiF-BeF —UF, (60-37—1 mole %) at temperatures up to 1300°F have given no evidences of INOR-8 carburization in periods as long as 12 000 by, (FeT 657 The third problem area associated with the use of graphite 1s that of reactions between the fuel salt and impurities contained in the graphite. The most serious reaction to be considered is that of uranium-tetrafluoride and oxygen reacting to form uranium dioxide, a product which is relatively insoluble in molten fluoride mixtures of the type considered for present salt reactor concepts. The precipitation of UO if it occurred, would pose 2) a serious hot-spot problem in any stagnant region of the reactor core. Up to the time of the subject experiment, studies concerning these salt-metal-graphite compatibility problems had been limited to static systems. Porosity ranges from 15 to 30% by volume in present commercial grades. 5MSRP Quar. Prog. Rep., Jan. 31, 1959, ORNL-268L, p. T6. 6MSRP Quar. Prog. Rep., July 31, 1959, ORNL-2799, p. 59. 87D. H. Jansen and W. H. Cook, Met. Ann. Prog. Rep., July 1, 1960, ORNL-29808, D.22C. _3_ T+t was felt advisable to complement these studies with a compatibility test in which the flow and temperature conditions of the molten fuel mixture closely simulated those of proposed reactor systems. This report describes the results of such an experiment which was conducted jointly by members of the Metallurgy, Reactor Chemistry, and Reactor Divisions. TEST EQUIPMENT AND METHODS Loop Design A forced-convection loop of the type employed for investigations of the corrosion properties of fused fluoride mixtures was modified to permit the in- corporation of graphite at the point of maximum fluoride temperature. The loop, designated 93545, consisted of a centrifugal pump, two resistance heater sections, the container of graphite, a cooler section, and drain tank which were assembled as shown in Fig. 1. The entire loop and pump bowl were fabricated of INOR-8. The tubing used for the loop was 3/8-in.—o.d. by 0;035—in. wall. The drain tank, which was isolated from the circulating salt mixture during operation, was of Inconel, and tThe pump rotary element wetted parts were fabricated of Hastelloy B. The nominal compositions of Hastelloy B and Tnconel are shown in Table 1. The grephite container, completely constructed of INOR-8, was installed in a horizontal position at the outlet of the second heater leg (Fig. 1). Gas entrapment in the container was prevented by placing the inlet at the lower portion of the end plate and the outlet at the upper portion of the opposite end plate, as shown in Figs. 1 and 2. Figure 2 shows the INOR-O container filled with graphite rods before the cover was attached. The container, fabricated from 0.060~in. sheet, was in the form of a rectangular box ol in. long by 2_1/2 in. square. Orifice plates were placed at the ends to hold the rods in place and to distribute the flow to the spaces between the rods. Baffles were also welded in between the orifice plate and the end of the container to 9id in the distribution of flow. 8J. L. Crowley, W. B. McDonald, and D. L. Clark, "Design and Operation of Forced-Circulation Corrosion Loops with Molten Salts," Paper presented at A.N.S. Meeting, Gatlinburg, Tennessee, June 17, 1959. UNCLASSIFIED ORNL- LR~ DWG 39365R GRAPHITE CONTAINER \SAMFLE LEG FREEZE VALVE DRAIN TANK Fig. 1. Molten Salt Corrosion Loop No. 9354-5 Containing Graphite Specimens. Showing thermocouple and specimen locations. - - Table 1. Nominal Compositions of Various Nickel-Base Alloys Composition (wt %) Alloy N1 Mo Fe Cr C S1 Mn ! \J1 Inconel 72 min -- 6.0~10.0 14,0-17.0 0.15 0.5 1.0 max - INOR-8 bal 15.0-18.0 5.0 max 6.0-8.0 0.4-0.8 0.35 max 0.8 max Hastelloy B bal 26.0-30.0 h,0=7.0 1.0 max 0.05 max 0.03 max 1.0 max Hastelloy W bal 25.0 5.5 5.0 -- - -- UNCLASSIFIED PHOTO 30631 Fig. 2. Graphite Container Before Installation in Loop. -7 - The graphite test specimens consisted of thirty-two 1/2-in.-diam rods and eighteen 3/16-in.-diam rods, each 11 in. long. These specimens were of a low—permeation9 type graphite, National Carbon Grade GT-123-82. Measurements made by the Materials Compatibility Laboratory indicated that the average bulk density of the "as-received" graphite was 1.91 g/cc. This is 84.2% of the theoretical density of graphite.lo’ll Before the graphite rods were installed, they were calipered and welghed by the Reaction Processes Group of the Reactor Chemistry Division. Figures 2 and 3 show the horizontal array in which the graphite rods were stacked. opace was maintained between each of the rods and the sides of the box by means of 0.035-1in.-diam Hastelloy W (nominal composition listed in Table 1) wire spacers wound around the 1/2-in.-diam rods. These spacers were staggered along succegsive layers of rods so that a flow area between the rods was maintained. Operating Procedures Because of the ability of graphite to contain relatively large amounts of sorbed gases, it was necessary to outgas the rods before loop operation was initiated. Outgassing of the graphite was accomplished after the loop was insulated and installed in its facility. A description of the method by which the graphite was outgassed 1s given in Appendix A. Upon completion of the outgassing process, the loop was filled with the salt mixture LiF-BeF —UF) (62—37—1 mole %). This first fill was utilized as a cleaning fluid and circulated approx 12 hr at 1200 to 1250°F. After dumping and refilling with a second salt charge, the loop was placed on the AT con- ditions shown below. Maximum salt-metal interface temperature 1300°F Maximum salt and salt-graphite temperature 1250°F 91n this report, 'permeation! refers to impregnation of the graphite with o fluoride salt and 'permeability! refers to the rate of gas flow through graphite. O The theoretical density of graphite is 2.27-2.28 g/cc. Llvaeneral Properties of Materials," The Reactor Handbook, 3(1), 136, AECD-3647 (March, 1955). - UNCLASSIFIED ORNL-LR-DWG 440884 \ \ . N \ \ A\ Ry ,//, , S %,,,7//%// 3\ A Numbering Scheme for Positioning Graphite Rods. Fig. 3. - 9 - Minimun salt temperature 1100°F AT 200°F Reynolds No. 2200 Flow rate 1 gal/min Pressure on graphite 12.9 psig The calculations made to determine the pressure on the graphite along with other loop statistics are shown in Appendix B. The averages of loop temperatures, which were recorded once per day, are shown in Fig. 4. External heat was applied to the graphite contalner during operation to maintain the temperature at the outlet (TC No. 11) approximately egual to the temperature at the inlet (TC No. 9). The maximum wall temperature of the container, as recorded by TC's No. 7 and No. 8, was maintained at 1300°F. The loop operated under the specified polythermal conditions for a total of 8950 hr. In addition, minor troubles encountered during the course of operation caused the locp to operate 78 hr isothermally. Upon termination, the loop was placed on isothermal operation and the salt was drained through a sampling tube into a pot. A trap was placed in the line between the loop and the pot, in order to obtain a specimen ol the after-test salt for chemical analysis. Along with providing an after-test specimen of the salt, draining the loop facilitated the removal of the graphite rods and loop specimens for examination. A sample of the before-test salt was also submitted for chemical analysis. A chronology of the events that affected the performance of the loop is in Appendix C. EXPERIMENTATL RESULTS AND DISCUSSION Procedure for Ixamination The box containing the graphite rods was removed at the conclusion of loop operation and opened by grinding through top of the container. A portion of the after-test graphite rods was submitted to the Analytical Chemistry Group for determination of any physical or chemical changes, and the remainder of the rods was examined metallographically for any microscopic changes by the Metallurgy Division. TEMPERATURE (°F) 41300 1250 1200 1150 1100 1050 1000 UNCLASSIFIED ORNL-LR-DWG 32366 LUG LUG LUG LUG GRAPHITE 2 o 5 5 CONTAINER COOLER ST ND B n 1STHEATER i 2NPHEATER P | | |1 O o< M < N O W Tqm < m O © © © N o o - O o b b | l @ AVERAGE TEMPERATURE ° e OF CONTAINER WALL o Y Vo r\ / ~ A /,/ \\ '~ ® // \\ /// N p— N ~ // o N 4 Z \\\ // N , ~ /7 \\ & ___ ./ U @ @ A LUG TEMPERATURE ® WALL TEMPERATURE NOTE: DASHED LINE DRAWN THROUGH WALL TEMPERATURES o WHERE LOOP WAS INSULATED AND UNHEATED TO INDICATE FLUID TEMPERATURES 0 100 200 300 400 500 600 CISTANCE FROM PUMP EXIT (in.) Fig. 4. Average Wall Temperatures for INOR-8 (with Graphite Insert) Loop No. 9354-5, - 11 - Specimens of the loop components were also removed from positions indicated in Fig. 1 and were examined metallographically for evidence of carburization and attack by the fluoride mixture. Samples of the galt cir- culated were submitted to the Analytical Chemistry Group for optical microscopy, x ray, and wet chemical analyses. Graphite Analyses Chemical Analysis.- Machined increments of graphite specimens were sub- mitted for chemical analysis. Successive cuttings, 1/32 in. in depth, were taken from two of the larger diameter rods until center portions of less than 3/16-in. diam were left. These portions and "gas-received" impervious graphite "planks! were then ground to -100 mesh in a mortar and pestle, which was thoroughly scoured with Ottawa Sand according to the recommendations of the Analytical Chemistry Division after each grinding. All graphite samples were qubmitted for an analysis of the uranium and beryllium concentrations. Two machine cuttings, 1/32 in. in depth, were taken from four additional rods. These results are given in Table 2 with the beryllium and uranium concentrations graphed as a function of penetration depth in Fig. D. Only a very slight mi- gration of salt to the center of the graphite is noted. Physical Analysis.- Macroscopically, there was no change in the rods. None of the samples was broken or distorted and, except for the bottom layer of rods that was covered with solidified melt, the salt did not adhere to the graphite, as shown in Fig. 6. The weight and dimensional changes observed for the rods after contact with circulating fluorides are listed in Table 3. The dimensional changes for the thirteen 1/2-in.-diam rods corresponded to an average loss of less than 0.5 mil in diameter which approximates the probable error of the mea- surements. Otherwise, there was no evidence of erosion. Weight losses, which ranged from negligible to 0.05% and averaged 0.02%, could be attributed to de- sorption of residual gases from the graphite. No statistically significant differences were noted in the thirteen 1./2-in.-diam rods as compared with the eight 3/16-in.-diam rods for which weight data were available. - 1o - Table 2. Analyses of Machine Cuttings from Graphite Rods Rod Cutting pemn Theoretical” Actual No. No. T Be U/Be U/Be 8 1 30 125 0.573 0.240 0 9 175 0.051 b 10 <1 - 11 1 PP 125 0.176 D 10 110 0.091 14 1 ol 75 0.320 2 28 105 0.267 23 1 17 125 0.136 2 <1 60 0.017 a 5 < 1 - 18 a, 8 < 1 - 1 50 170 0.294 D 15 130 0.115 3 15 125 0.120 b 12 100 0.120 5 10 65 0.154 6 13 105 0.124 7 <1 50 0.020 &8 13 140 0.093 9 5 165 0.030 16 <1 <1 1.000 11 6 105 0.057 10D <1 <1 - Center 100 125 0.800 31 b 5 <1 - 1 20 165 0.121 2 18 140 0.129 3 ol 120 0.199 L 20 85 0.235 5 20 75 0.267 6 20 80 0.250 7 17 55 0.310 8 < 1 80 0.013 9 < 1 95 0.011 10 < 1 <1 1.000 11 < 1 90 0.011 Center 70 170 0.411 b < 1 < 1 - ®Based on chemical analysis at original salt batch, nominally LiF—BeF —UF, (62=37=1 mole %). Samples machined from "as-received" material. URANIUM, BERYLLIUM CONCENTRATION (ppm) UNCLASSIFIED ORNL-LR—DWG—41131A | ROD 31 BERYLLIUM e URANIUM A 200 | ROD 18 oo |3 | «—BERYLLIUM 120 ””,7 A ‘r a0 % URANIUM / 0—..._.__. / 0 D ° —40 8 12 0 /32 /32 X O—~ DEPTH (in.) 4 /32 8 /32 Penetration of an Impervious Graphite by LiF—BeFQ-UFu. 16 732 _E-[_. Fig. Graphite Container After Test. UNCLASSIFIED Y-3033% -1-{-[... of the Graphite Before and After Salt Exposure Table 3. - 15 - Weight and Dimensional Changes Before Exposure After Exposure Net Rod Welght Diam Welght Diam Change Percent No. (&) (in.) (&) (in.) (s) Change Tmpervious Graphite Rods (1/2-in. diam) 1 Lost 3 Lost 6 68.0555 0.498 68.0397 0.496 -0.0158 -0.02 8 68.0571 0.498 68.0438 0.497 -0.0133 -0.02 9 68.5709 0.502 68.5572 0.501 ~0.0137 -0.02 11 68.4152 0.500 68.4096 0.500 -0.0056 -0.01 14 68.7779 0.499 68.7639 0.500 -0.0140 -0.02 16 67.7389 0.496 67.7205 0.495 -0.0184 -0.03 18 68.2650 0.498 68.2517 0.497 -0.0133 =0.02 20 Lost 21 68.5801 0.500 68.5793 0.500 -0.0008 0.00 23 67.9828 0.497 67.9703 0.496 -0.0125 -0.03 26 68.2956 0.499 68.2911 0.500 -0.0045 -0.01 28 68.6806 0.501 68.6666 0.501 -0.0140 -0.02 29 67.8522 0.499 67.8352 0.498 -0.0174 -0.03 31 67.9389 0.499 67.9169 0.498 -0.0220 -0.04 Tmpervious Graphite Rods (3/16-in. diam) 2 9.1095 9.1082 -0.0013 -0.01 L 9.1236 9.1228 -0.0012 -0.01 6 9.4826 9.4810 -0.0016 -0.02 8 9.0329 9.0352 +0.0021 +0.02 10 9.3176 9.3126 ~0.0050 -0.05 12 8.7251 8.7372 +0.0121 +0.1h4 14 9.0932 9.0930 -0.0002 0.00 16 9.5142 9.5098 -0.004L -0.05 18 9.0149 9.010L4 -0,0045 -0.05 R T AR A TR RS SR R e S A e e e e SR AT T e e e e A e T i ST AT e e 1T e, - 16 - Metallographic Examination.- Additional post-test physical examinations were made by the Materials Compatibility Laboratory of the Metallurgy Division. Only a single specimen of this material, a 1/2-in.-diam x ll-in.-long graphite rod, was avallable for establishing the "as-received" characteristics of the graphite rods. Iowever, comparisons of the microstructures of samples from this rod and samples from tested rods indicated that samples were relatively uniform in structure. The typical microstructures of the graphite in the "as-received" and after-test conditions are compared in Fig. 7. In both conditions the graphite characteristically exhibited small, uniform, and widely scattered volds. DBecause of the extremely small size of the voids, it was not possible to identify fuel in them by simple microscopic examination. Both the 3/16-in.-diam and 1/2-in.-diam rods had radial laminations or cracks in sections transverse to the axes of the rods, Fig. 7. The laminations were 0.010 to 0,075 in. long and approx 0.001l in. wide in a 3/16-in.-diam rod and were 0.030 to 0.150 in. long and 0.001 to 0.C02 in. wide in 1/2-in.diam rods. There appeared to be fewer laminations in the 3/16-in.-diam rods than in the 1/2-in.-diam rods. The majority of the laminations began and terminated inside the rods. Only a few extended to the curved surfaces of the rods. The test apparently had no effect on the laminations. Finally, comparisons of the microstructure of the transverse and longi- tudinal sections of tested rods with those of the "as-received" material indi- cated that no attack or erosion had occurred in the tested rods. This observation is in agreement with the weight measurements of the specimens discussed previously. Discussion.- The grade GT-123-82 of graphite was an experimental graphite that was fabricated especially for this particular test. The prime objective was To produce a low permeation graphite. The manufacturer utilized his technology for meking the low permeation graphite but the actual fabrication of grade GT-123-82 was not part of his research for the development of grades of low permeation graphite. The small diameters, 3/16 and 1/2 in., of the graphite rods probably made the reduction of the permeation easier. Therefore, it is not known if the structure of grade GT-123-82 could be duplicated in larger shapes applicable for reactor usage. Fig. T. diam Rods (a) As-Received and (b) After a Qne-Year Exposure to LiF-BeF lUnelassified Low Permeability Graphite: (62-37-1 mole %). As polished. 100X Typical Microstructure of 1/2-in.- o UFy, Unclassified -LT' - 18 - Assuming this particular grade of graphite could be duplicated for the shapes required for use in a reactor, the small amount of salt that permeated the graphite in the loop is not entirely representative of the permeation of the graphite in a reactor, in that the pressure on the graphite in a reactor would be approximately three to four times as great as the 13 psig experienced in the loop test. At thilis time, there is not sufficient data to make an ex- trapolation for the percentage of the bulk volume of (T-123-82 that would be permeated by salt in the reactor. Examination of Loop Components Loop Specimens.- Metallographic examination of specimens from the first and second heater legs indicated negligible attack to have occurred in these sections, based on the absence of surface pitting or subsurface voild forma- tion. The general appearance of the surfaces of these specimens can be seen in Figs. 8 and 9. Similarly, no evidence of attack (i.e.,surface attrition) was found in the specimens removed from the unheated segments of the loop, the tubing connecting the pump to the first heater leg, the unheated bend connecting the two heater legs, and the cooler coil. The condition of the cooler coil surfaces is shown in Fig. 10. Neither the cooler coil nor other cold leg regions of the loop exhibited detectable mass transfer deposits. As evidenced in Figs. 8, 9, and 10, the metallographic appearance of the loop specimens gave no indication of surface carburization, carbide pre- cipitation being no heavier at the exposed surfaces than below. However, an extremely thin film was found to be present on all the specimens examined from the pump exit up to the entrance of the box containing the graphite rods. This film appeared as a continuous second phase which ex- tended below the exposed surfaces of the specimens to thicknesgses ranging up to 1/3 mils. The thickness of this film appeared to increase linearly along the affected length of tubing, increasing in thickness from approx 1/10 mil at the pump outlet to 1/3 mil at the end of the second heater leg. As can be seen in Figs. 8 and 9, there is no evidence of a transition or diffusion zone between the film and the base metal. Unelassgified T-17667 00T .oo8 009 010 211 Fig. 8. Appearance of Specimen Removed from End of the First Heater leg. Etchant: 3 parts HCl, 2 parts H,0, 1 part 10% Chromic acid. 250X Unclassified =) £ T-17467 L=] o s 1 T INCHES 1 1 !_ g g : e ; 007 T Fig. 9. Appearance of Specimen Removed from End of the Second Heater 250X er Leg. Etchant: 3 parts HCl, 2 parts HEO’ 1 part lflfi Chromic acid. 50 Fig. 10. GCeneral Appearance of Coocler Coil. Etchant: 3 parts HC1, 2 parts H,0, 1 part 10% Chromic acid. 250X - 21 - Graphite Container.- Metallographic examination of specimens removed from the entrance and exit of the graphite container revealed no attack and no evidence of carburization. As in the case of the specimens removed from the loop, a very light surface film was found along the interior surfaces of the box (Fig. 11). As a check for carburization, hardness tests were made on two after-test specimens removed from the entrance and exit ends of the box top and on an "as-received" specimen of the box. As shown in Table L, the results of these measurements corroborate the apparent absence of carburization in that the after-test specimens exhibited the same range of hardness as found in the "as-received" specimen. Hastelloy W Spacer Wires.- As described previously,l2 0.035-in.-diam wires circled each of the graphite rods stacked in the container to maintain a fixed flow amulus between each of the rods. Since these spacer wires were in direct contact with graphite, they were examined particularly for evidence of carburization. Metallographic examination of the after-test wires showed them to have a microstructure distinctly different from that of the "as- received” material, as shown in Figs. 12 and 13. Hardness determinations, which utilized a Tukon Hardness Tester, indicated hardnesses from 411 to L35 DPH in the before-test specimens and from 435 to L69 DPH in the after-test specimens. Since Hastelloy W is subject to aging at the temperature range to which it was exposed, a hardness increase for this material would be expected with or without carburization. However, the analyses also showed that the carbon content had increased significantly from a before- test level of 0.028% to an after-test level of 0.055%. Analysis of Corrosion Film.- As noted above, a continuous second phase was observed along the exposed surfaces of various loop components thicknesses up to 1/3 mil. The comparative hardnesses of the film and parent metal in lg"Test Equipment and Methods” p. 7 of this report. iy o e ’ 008 010 Ol 013 Ol4 ] |._-'.r 09 l11. Specimen Removed from Top of Box Containing Graphite L L ol 17 3 | [ % Etchant: 3 parts HCl, 2 parts H_ 0, 1 part 10% Chromic acid. 250K ==2 fo - 23 - Table 4. Hardness Measurements Made on Top of Graphite Container Hardness (DPH) oide Exposed oide Exposed Top of Box to Air to Salt Salt Entrance 201 201 178 177 177 Salt Exit 188 180 178 178 177 As-Received 201 201. 194 194 170 Fig. 12. As-Received Microstructure of Hastelloy W Spacer Wires. Etchant: 3 parts HCl, 2 parts HEO, 1 part 10% Chromic acid. 250X Fig. 13. Typical Microstructure of Hastelloy Spacer Wires After Test. Etchant: 3 parts HCl, 2 parts HEO, 1 part 10% Chromic acid. 500X - 25 - specimen No. 11, on which the thickest layer was found, were determined using a Bergman Microhardness Tester and a l1-g load. On the basis of the 1l-g load, the film was found to have a VHEN of 518, while the Vickers hardness number for the parent metal was 269 (Fig. 1Lk). A check of the base metal using the Tukon Hardness Tester with a 200-g load indicated a VEN of 218. Thus, the absolute magnitudes of the hardness values yielded by the Bergman tester would appear to be slightly high; nevertheless, they demonstrate that the film is approxi- mately twice as hard as the parent metal. Because of the minute thickness of this film, difficulty was encountered in quantitatively determining its chemical composition. However, gualitative analyses of the film were obtained using material which was "scrubbed" from the inside surface of specimen No. 5 (Fig. 8) through the use of a corundum slurry and a glass rod. The specimen was welighed before and after scrubbing +o determine the amount of material removed. The resultant mixture of slurry and metal, together with an unused sample of the slurry as a blank, was then submitted Tor spectrographic analysis. Results of analysis by this technilque are shown in Table 5. Table 5. Analyses Made of Surface Films Composition (wt %) Specimen Ni Cr Fe Mo Other As-Received 71.66 6.99 L.85 15.82 0.68 No. 5 80.0 1.5 14,5 4.0 —- No. 11 67.0 - 0.6 21.6 10.8 A semiquantitative analysis of the film was algso carried out by use of an electron-beam microprobe.l3 This analysis was accomplished by aiming the electron beam directly on the inside diameter surface ol specimen No. 11 (Fig.9) 13MSRP Quar. Prog. Rep., July 31, 1960, ORNL-3014, p. 56, INCHES 02 oca T50X Fig. 1h. Microhardness Measurements Made on Specimen Removed from Second Heater Leg. Etchant: None. 750X. - 27 - > which excited an area of 1.96 x 10 cm2 at the surface to a depth of approx L p. The characteristic x rays emitted by the film were then spectrographically analyzed. The results obtained from this technique are also shown in Table 5. Both types of analyses indicated the film to be composed primarily of nickel. In the sample from specimen No. 5, iron appeared to have increased relative to the base-metal composition while chromium and molybdenum appeared to have decreased. In contrast, the analyses of specimen No. 11, obtained by means of the microprobe, showed an enrichment of molybdenum along with apparent depletion of chromium and iron. At best, these results provide a semiquantitative picture of the chemical composition of the films. Because of inherent errors involved in the analyses "scrubbings" from specimen No. 5, the estimated accuracy of obtained on the the results is within + 20% of each individual cfl.etermination.l)‘L A few of the contributing errors involved are: (1) tramp elements in the abrasive, (2) residual fluorides on the surface of the specimen, and (3) tramp elements in the glass rod used as the scrubber. In the case of the analyses obtained for specimen No. 11, the errors are less easy to resolve. An immediate indi- cation of some error is afforded by the fact that the percentages of nickel, chromium, iron, and molybdenum do not add up to 100%. To assure that the deviation was not associated with residual fluorides remaining on the surface, the specimen was cleaned with an ammonium oxalate solution and then reanalyzed. The ammonium oxalate solution has been found to dissolve fluoride salts quite effectively without disturbing alloys composed predominantly of nickel. The results of this reanalysis, as shown in Table 5, showed no difference from the orlginal analysis. Thus, the failure of nickel, chromium, iron, and molybdenum %o add up to lOO% is believed to have resulted from scattering or absorption losses, which would also introduce some error in the reported relative amounts of these components. IHM. M. Murray, Analytical Chemistry Division, to R. C. Schulze, personal communication, September, 1960. - 28 - Chemical Analyses of Salt.- Chemical analyses of the salt circulated in the loop after test are shown in Table 6. Except for an expected increase in Table 6. Chemical Analyses of Salt Mixture Before and After Operation wt % Theoretical@ ppm Sample Taken U Be U/Be U/Be Fe Cr Ni Before Test 4,87 8.37 0.582 0.767 235 135 5 After Test Loop L.97 9.77 0.509 330 550 25 Sump 4.59 9.55 0.480 460 585 85 aCalculated for LiF-—BeF,-UF) (62=37=1L mole %). chromium concentration, these analyses show little change from the analyses of the concentration of impurities in the before-test salt. As shown in Table 6, chemical analysis of the after-test salt revealed no measurable pick-up of carbon. FExamination of the after-test salt was also made under the petrographic microscope and x-ray diffraction unit to detect the presence of oxide compounds within the salt. These analyses showed the salt to be apparently unaffected by impurities contained in the graphite; however, analysis of the salt by wet analytical methods indicated the presence of 3400 ppm of oxygen. CONCLUSIONS Excluding the possible irradiation effects and keeping in mind the other differences that have been pcinted out for the pump loop and a reactor, there were useful data generated by the pump-loop test. The compatibility of the three-component system, salt-graphite-INOR-8, at slightly above the reactor operating temperature for a relatively extended period of time was demonstrated. The test indicated that: 1. There was no corrosion or erosion of the graphite by the flowing salt. 2, There was very little permeation of the graphite by the flowing salt, and the permeation that occurred was uniform throughout the graphite rods. - 29 - 3. The various INOR-8 loop components exposed to the salt were not carburized. 4. The INOR-8 components exposed to the salt and graphite were negligibly attacked. 5. With the possible exception of oxygen contamination, the salt appeared to have undergone no chemical changes as a result of exposure to the graphite test specimens. ACKNOWLEDGMENT The authors wish to thank J. H. DeVan who made many pertinent and helpful suggestions concerning the experimental work and the preparation of this report. Thanks are due to the following individuals and groups for their contributions: W. H. Duckworth for the operation of the experiment, M. A. Redden for the preparation of loop-component specimens, The Reactor Chemistry Division, especially F. A. Doss and W. K. R. Finnell, The Analytical Chemistry Division, especially C. F. feldman, M. M. Murray, and W. F. Vaughn, The Metallography Group of the Metallurgy Division, and particularly C. E. Zachary. APPENDIX A PROCEDURE FOR OUTGASSING GRAPHITE AND CLEANING LOOP e O R R AR 8 o e, 1 X b SR R - 33 - PROCEDURE FOR OUTGASSING GRAPHITE AND CLEANING LOOP After the entire loop was insulated and installed in its facility, the pump rotary element was removed and the bowl was sealed in preparation for outgassing of the graphite. Outgassing of the graphite was accomplished by first holding a vacuum on the loop for 48 hr at room temperature and then gradually heating both the loop and container to between 1000 and 1100°F while still under vacuum. After 3 reaching this temperature range, a vacuum of less than 5 x 10 ° mn Hg was maintained on the system for 24 hr, following which the system was pressurized with argon and allowed to cool. The cover, used to seal the pump bowl, was removed and the pump rotary element installed while maintaining flow of argon from the opening. The loop was then heated under a positive argon pressure and filled with the salt mixture LiF—BeF,-UF) (62—37—L mole %). This first £i11 was utilized as a cleaning fluid and circulated approx 12 hr at 1200 to 1250°F before being dumped. The loop was then refilled with the operating charge. Sy B R APPENDIX B PRESSURE CALCULATIONS FOR SALT IN GRAPHITE CONTAINER _.3‘7_ PRESSURE CALCULATIONS FOR SALT IN GRAPHITE CONTAINER Using a pump performance curve for the LFB pump and the physical properties of the salt mixture LiF-—BeF ~UF) (62—37—1L mole %), the flow con=- ditions of the test were estimated as follows: Reynolds number, N_ = pdv 2, 180 where p = 123 1b/ft at 1100°F o d = 0.0254 ft v = 4,84 ft/sec L= 0.00692 1b/ft-sec Flow head at discharge of pump, 2 1V 6L htotal = T 3 —g =21.5 ft., where T = fi; = 0.029 1 =50 Tt d = 0.254 ft v = 4.84 ft/sec Flow head at 5&% of length of the loop at the inlet to the graphite container. h(inlet) = 0.54 (21.5) = 11.6 ft Estimated pressure of molten salt at the graphite container (neglecting vertical height difference and assuming a 3-psig cover pressure on the pump suction). Blinlet)P o . P = —~ I - 12.9 psig in the container. Other statistics of the loop and container are as follows: Volume of molten salt No. 130 in circulation 121 in.3 Surface of INOR-8 exposed to salt 1023 in.2 Geometric surface of graphite exposed to salt 683 in.2 Ratio of graphite to INOR-8 surface 0.67 APPENDIX C CHRONOLOGY OF LOOP OPERATION 5-9-58 5.10-58 5-28-58 6-23-58 8-5-58 8-15-58 8-22-58 11-7-58 1-15-59 3-16-59 4-17-59 2-20-59 i e o it i e - 41 - CHRONOLOGY OF LOOP OPERATION The AT cperation was started. The loop was placed on isothermal operation 3 hr for pump motor work. The loop was placed on isothermal operation for 24 hr while alterations were being made to controller. The loop was placed on isothermal operation 18 hr for repairs on pump motor. Power fallure caused cooler to become partially frozen, but the plug was thawed successfully and the loop resumed operation. Only the cooler temperatures dropped below the melting temperature of the salt (approx 850°F). 'The first heater section reached 1500°F momentarily when power was reapplied suddenly. Loop operated 23 hr isothermally. The loop was placed on isothermal operation for 4 hr as precautionary measure during a plant fire. The loop was placed on isothermal operation for 1 hr for pump motor work. The loop was placed on isothermal operation for 2 hr after momentary power outage. The loop was placed on lsothermal operation for 1 hr after momentary power outage. The loop was placed on Isothermal operation for 1 hr to replace broken pump drive belt. The loop was placed on isothermal operation for 1 hr for pump motor work. The loop was terminated after 8 928 hr of operation at AT' conditions. The loop was drained through the sample leg while on isothermal operation. Total Hours AT Operation - 8 850 Total Hours Isothermal Cperation - 78 o e 5L4—58. o8. 93-100. 101. 102. 103. 10L—680. DISTRIBUTION Biology Library 63. Health Physics Library 6l Metallurgy Library 65. Central Research Library 66. Reactor Division Library 67. ORNL Y-12 Technical Library, 68. Document Reference Sectilon 69. Iaboratory Records Department 70. Laboratory Records, ORNL R. C. T1. G. M. Adamson, Jr. T2. D. 5. Billington 73. A. L. Boch Th. E. G. Bohlmann 5. B. S. Borie T6. R. B. Briggs 7. C. E. Center 8. R. A. Charpie T3-81. R. S. Cockreham 82, E. Cohn 83. W. H. Coock 8hL. J. L. Crowley 85. F. L. Culler 86. J. E. Cunningham 87. D. A. Douglas 88. R. B. Evans, III 89. J. H Frye, Jr. 90. W. R. Grimes 91. C. 5. Harrill 92. M. R. Hill 23. A. Hollaender ok, A. S. Housgeholder 95. R. G. Jordan (Y-12) %. W. H. Jordan 97. EXTERNAL DISTRIBUTTION D. E. Baker, GE Hanford D. Cope, AEC, ORO Ersel Evans, GE Hanford J. Simmons, AEC, Washington D. K. Stevens, AEC Washington Given distribution as shown in TID- 1500 (16th ed.) under Metals, Ceramics, and Materials Category (75 copies - 0TS) ORNL-312% Metals, Ceramics, and Materials mtr_l':UO"—|L|POPHQPO:_E.’MEEUM':UU"GELIWH:D_OSM’;U"—IZO TID-1500 (16th ed.) Keim Kelley Tane Livingston . MacPherson Manly McHargue Miller Miller Morgan Murray (K-25) Nelson Patrlarca Phillips Reyling Savage schulze Seagren Shipley Skinner omith Snell Swartout Taylor Weinberg Winters Burr (consultant) Gregg (consultant) Koenig (consultant) Smith (consultant) Smoluchowskl (consultant) L—'FUNOQQUQCD?DH"G pEOPHEEFPEHEBDEEESS . E. Stansbury (consultant) . A. Wilhelm (consultant)