'__5.___;OAK RIDGE NATIONAL I.ABORATORY . . "jf_-_operuted by UNION: CARBlDE CORPORATION - - for the o | U S ATOM!C ENERGY COMMlSSION - e - G L= '35 ~ NOTICE ‘This report contains patentable, preliminary, ‘unverified, or erroneous information. For one or more of these reasons the author or | issuing installation and responsible office - 1-71 have limited its distribution to Governmental 1.7 ] agencies and their contractors as authorized | by AEC Manual Chapter 3202-062. A formal report will be published at a later date when the data is complete enough to warrant publi— . cation. - : : : It is subject e present a final . eprinted or otherwise: gi:‘v‘:n’e::f::m ;l': e ORNL pofent brcnch Legal cnd fnfor- e s -'lnformafion ‘is nof. s o be: -abstracted, - semination without the npprovn! of ihr o mcfion Controi Depurtment. ot i i kR L i s @ e e ey LY LKA g g g ey ke 5 D 'L;dii. »ioflce, Tbis npon was propund os an account of Gov-rnmom sponsoroci work. Noirher the Unmd S!u:u, : nor the Comminlon, nor any person acting on behalf of the Commission: A, Mokes any warranty or representation, ‘expressed or implied, with respect to !ho accuracy,"r'fir completeness, ‘or viefulness of the information contained in this report, or that the use of _ any information, cppurows, molhod or proc“s dm:lond in this rnpofl may net infnnge -privataly owned rights; or - — - B. Assumes any liabilities with respecf |c the use of, or for Jcmogas uwlhng ‘from fl\a use of any informetion, epparatus, nflhod or process disclosed in this report.. As used in the above, *person ueflng on behalf of the Commission® includes any employes or | .. “contractor of the Commission, or smployes of such contractor, to the extent that such employse - |- “or contractor of the Commission, or employes of such contractor prepares, disseminates, or - provides cccess to, any information wsucnt to’ hls ‘employment of contract with fhe Comm:sslon, . or his Cmploymum with such conircflor. o “ w} ORNL-T-500 Copy Contract No. W-7405-eng-26 REACTOR CHEMISTRY DIVISION RADIATION CHEMISTRY OF MSR SYSTEM Written by: W. R. Grimes The experiments described in this document required the cooperative effort of many people in the Reactor, Metals and Ceramics, Operations, Analytical Chemistry, and Reactor Chemistry Divisions of the Oak Ridge National Laboratory. D. B. Trauger, A. R. Olsen, E. M. King, A. Taboada, and F. F. Blankenship have been in respon- sible charge of important segments of this work. While he accepts responsibility for accuracy of the reported material , the author makes no claim to credit for design and operation of the experimental program. DATE ISSUED ‘MAR 13 1963 QAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee - operated by UNION CARBIDE CORPORATION I - for the U. S. ATOMIC ENERGY COMMISSION INTRODUCTION . . . EXPERIMENT ORNL-MTR-47-4 . . . . . . Irradiation of Specimens . . . Post Irradiation Examination . Analysis of Cover Gas . . . Gross Examination of Capsules Examination of the Metal . Examingtion of Salt . . . Condition of the Graphite Conclusions . - - - . - - EXPERIMENT ORNL-MTR-47-5 . . « . . . Behavior Under Irradiation . . Behavior Behavior Behavior Behavior Behavior Conclusions . with Reactor at Full Power. . with Reactor Shut Down. at Intermediate Reactor Power Through Startup and Shutdown. of Capsules After Termination 11 . 11 19 22 22 27 27 28 . 33 . 34 37 41 . 4l . 43 ) -9 RADTATION CHEMISTRY OF MSR SYSTEM INTRODUCTION Compatibility of molten fluoride mixtures, near in composition to that proposed for MSRE, with grfiphite and with INOR-8 has been demonstrated convincingly in many out-of-pile tests over a period of several years. A few in-pile capsule studies of graphite-fluoride-INOR systems with fluoride melts of a rather different composition have been performed in past years; examination of thosé capsules showed no adverse effect of radiation. How- ever, in-pilé testing of this combination of materials under conditions similar to those expected in MSRE has been attempted only recently. Accordingly, no completely realistic experiments have been reported. An assembly ORNL-MTR-3 containing four capsules, was irradiated in MTR during the summer of 1961 to provide assurance as to the non-wetta- bility of grasphite by molten LiF-BeF,-ZrF,-UF,-ThF,; though such assurance was obtained, this experiment yielded some surprising results.ls2,3,% 1In brief, the following anomalous behévior Qas observed. 1. The cover gas within the sealed capsules contained appreciable quantities of CF,. 2. The cover gaé contained the expected quantity éf Kr from all capsules; the Xe was present at the expected }evel in two, but its concenfiration-fias less by 100-fold in the other two capsules. 3. While the-INORfB,-Eothas_capsule walls and as small paftialiy | - immersed teét-speéimens, was eésentially undafiaged, small coupons of molybdenufi'were-thinned to roughly half the original thickness with no observable corrosion deposit. 4. No evidence of wetting of the graphite by the salt was observed. | Ei’ However, the irradiated salt was deep black in colof; it con- tained loose, roughly spherical beads of condensed salt of various colors from clear through blue to black. The difference in behavior of Kr and Xe seemed guite inexplicable; the behavior.of molybdenum, while not of direct conseqfience to the MSRE, was disturbing. The black color of the salt was removed by annealing at temperatures below the liquidus and was shown to be due, largely if not . entirely, to radiation-induced discoloration of crysfalline LiF and 2LiF+BeF,. This color, and the occurrence of the salt beads (wvhich were clearly due to condensation on the relgtively cool capsule walls of mate- rial distilled from the very hot pool of salt) were judged to be trivial. The appearance of CF, in the cover gas, however, remained a most disturbing observation. Thermodynamic data for reactions of fluérides with carbon such as 4UF,; + C = CF, + 4UF;3 suggest that the equilibrium pressure of CF, over such a system should not exceed 10-8 atmospheres. Moreover, out-of-pile controls with identical material , geometry and thermal histories (insofar as possible to obtain same with external heat sources) showed no evidence (> 1 part per million) of CF,. This gas, accordingly, clearly arose from some previously unknown radiation-induced reaction. Such generation might, of course, have serious conseqnences. If CF, were génerated at an appreciable rate in the MSRE core it might well be removed in the gas stripping section of the MSRE pump and be lost to the system. If so, the loss of carbon from the moder- ator graphite would be trivial, but an appreciable loss of fluoride ion g-} L 4] - from the.melt; with the resultant appearance of Ut3 or other reduced species in the fuel, would, at the least, cause frequent shutdowns to reoxidize the fuel; Additional in-pile experiments were, therefore, con- ducted in an attempt to establish the guantity of CF, to be expected under power levels, temperatures and graphite-~fuel geometries more representative of the MSRE. Two assemblies of capsules, designated ORNL-MTR-47-4 and ORNL—MTF—47-5, have been irradiated in the MTR for this program. While the examination of these experiments has not been completed, much informsa- tion has been obtained. The following is a brief statement of progress to date in this experimental program and a review of the conclusions which can presently be drawn from the informagtion. EXPERIMENT ORNL-MTR-47-4 Trradiation of Specimens The irradiation assembly, designated as 47-4 in the following, con- tained six INOR-8 capsules. These were immersed in a common pool of molten sodium which served to transfer £he heat generated during fission through & helium filled annular gap‘to_water circulating in an external jacket. The four large cefisules, as shown in Fig. 1, were l-inch diameter X 2.25-ifiches length-andeonteieea-a core of CGB grephite (l/2—inch diam- eter x ;-inch leegth)_efibmerged about 0.3-inch (at temperature) in about 25 grams of fuel. The two Smalier capsules, see Fig. 2, contained 0.5- inch diameter cylindricai crucibleslof CGB graphite containing about 10 grams of fluoride melt. The large capsules included a thermowell which -permitted measurement'ofitemperature within the submerged graphite speci- men; detailed analysis suggests that this measured'tempereture should approximate closely the temperature at the salt-graphite interface. UNCLASSIFIED ORNL-LR-DWG 67714R Cr=Al THERMOCOUPLE\fl /NICKEL THERMOCOUPLE WELL NICKEL POSITIONING LUGS (2)~ NICKEL FILL L|NE>k < - > )’(MCKEL VENT LINE | INOR-8 CAP ’ A 3 T HELIUM COVER GAS / L (3.5 cm?3) - f \ : ; N [Y=—INOR-8 CAN N PUNCTURE AREA / N_A) A FOR GAS SAMPLING \-_——_—'_: \ N oA \ - N \ MOLTEN SALT FUEL E L \ (25¢g) N _ N E L N N - - N Cr-Al A N THERMOCOUPLE E - — N b B —CGB GRAPHITE 12.4 cm? INTERFACE INOR—8 CENTERING PIN 0 'z { INCH O NICKEL POSITIONING LUG Fig. 1. Assembly 47-4 submerged graphite capsules. - £} HELIUM COVER GAS (2.4 cm3) HELIUM GAS GAP- — UNCLASSIFIED ORNL-LR-DWG 67715 NICKEL POSITIONING LUG PUNCTURE AREA FOR N GAS SAMPLING MOLTEN-SALT FUEL (10 g) —CGB GRAPHITE CRUCIBLE NN N N N N WA -<— INOR- 8 CAN O--— NICKEL POSITIONING LUG Fig. 2. Assembly 47-4 graphite crucible capsules. No such provision for temperature measurement could be incorporated into the smaller capéules. The CGB graphite used in all capsules of 47-4 had a surface area of 0.71 m?/g as determined by the BET method. Other properties of this material follow: Permeability of a 1.5-in.-0D, 6.56 x 10~% cm® of He (STP) 0.5-in.-ID, 1.5-in.-long specimen per second Density (Beckman sir pycnometer) 2.00 g/cm? Bulk density 1.838 g/cm? Bulk volume accessible to air 8.7% Total void volume as percentage of 19.5% bulk volume Spectrographic analysis of the graphite revealed only the usual low levels of trace elements; some of these may have been introduced during machining and handling operations. The UF,; used in preparing the salt mixtures was fully enriched in all cases. The four large (submerged graphite) capsules and one of the small capsules each contained fluoride mixtures of LiF-BelF,-ZrF,-Th¥,-UF,; for which analyzed samples indicated the composition 71.0-22.6-4.7-1.0-0.7 mole %; analysis for Ni, Cr, and Fe in the material showed 25, 24, and 97 parts per million. The other small capsule was loaded with a very similar mixture but with the uranium content raised to 1.4 mole %; the Ni, Cr, Fe analyses were 15, 40, and 123 ppm. The fused salt mixtures were prepared by mixing the proper quantities of pure fluorides, and then, at 750 to 800°C, sparging for 2 hours with H,, 8 hours with a 5:1 mixture of H, and HF', and.48 hours with Hj. The large capsules were filled by transferring the molten fluoride mixture under an atmosphere of He; the level to which they were filled was controlled by blowing excess liquid back through a dip-line adjusted to ) “‘ 0 9 the proper level. Tfi;‘ievel of salt was mofiifored with a television x-ray system. Final closure was made in a helium-filled glove box where the ends of the fill-tubes were crimped and then welded shut. This technique should have affected neither the composition nor the amount of cover gas within Ithe capsule. The smaller capsules were filled, in the glovebox, with ingots of solid fuel mixture. They were closed by inert-gas arc-welding of the end cap. This welding operation, which used argon as the electrode cover gas, raised the capsule temperature gpprecigbly and permitted mixing of some argon with the helium within the capsule before thg seal fias effected. Accordingly, neither the amount of gas nor the He:A ratio within the sealed capsule at time of closure can be certified. | The amount of fuel chosen for each capsule should have yielded a vapor volume of 3.5 cm? in the large, and 2 cm? in the sfiall, capsules at tempergture. The filled and segled capsules were x-rayed and examined to insure that the molten salt had flowed to its proper position in the assembly. Assembly 47-4 wes irradiated thfough three MTR cycles in the period March 15 to June 4, 1962. The'tempeiafure history of the six éapsules (as read from the chromel-alumel thermocofiples'in the four large capsules and as calculated for the two small capsules) is shown in Tsble 1. The maximum measured temperature among’the four instrumented capsules was 1400 * 50CF. The‘mean measured temperatures of the other three instrumented capsules were approximately 1380, 1370, and 1310°F, respectively. The correspond- ing calculated INOR-8 Capsule.walIQto-salt interface temperatures were 1130, 1125, 1115, and 1110°F. The temperature of the surface of the INOR-8 10 Table 1 Temperature History of Fuel Salt in Capsules from 47-4 Time et Temperature (hr) “ Temperature Interval Submerged Graphite Graphite® (°c) Core be Crucible (24, 36, 45, 12, 6°) AN Steady-state operation 0-100 390.6 390.6 100-700 44 .0 ' 12.1 700-750 1203.7 3.8 800-850 11.4 850-900 1456.9 Total 1892.3 1892.3 Nonsteady-state operation 51.7 - 51.7 Total irradiation 1553.4 1553.4 &Calculated temperature history based on temperatures measured in submerged graphite capsules. bGraphite crucible capsule containing fuel with 0.7 mole % UF,. Ceapsule identification number. -} ¥ -) 11 end cap in contact with the fuel salt was calculated to be the same as that.of the graphite-to-salt interface temperature. The area of the end cap wetted by the salt was abofit 12% of the total 5.6 in.? of metal sur- face in contact with the salt. The accumulated time during transient operation ineludes only times for temperature changes of more éhan about 30°C. Of the 121 such temper- ature changes recorded, 60 included decreases to or below the solidus tem- perature of the fuel salt. It is estimated that the fuel freezes within five minutes after shutdown of the MTR, and cooling to below 200°F should occur within half an hour. Exposure data for the capsules in 47-4, and those for out-of-pile controls which were given as similar a thermal history as was practicable, are summarized in Tables 2 and 3. The temperatures were controlled from Capsule 24, and the changes in the retractor position required to hold a constant temperature were not unusual. The thermocouple reading for capsule 45 appeared to drift down- - ward by about 35°C during the 12-week exposure, but this was the only symptom of deviant temperature recdrdings. Post Irradiation Examination | The assembly was rembved.from.the_MTR_and partially disassembled at that loéation for shipment £6 ORNL. The assembly was completely dismantled in ORNL'hqt cells énd the capsules were recovered for complete examination. A1l dismentling operations went smoothly, and no evidence of failure of . any‘capsule was observed. Analeis of Cover Gas The six irradiated and two unirradiated control capsules were et -t e s —p < A G B Table 2. Exposure Data for Large Capsules in Assembly 47-4 Thermal-Neutron Fast-Neutron Temperature (°C) Weight of Urenium Content Flwx® Based (>3 Mev) Flux Power Calculated Capsule Fuel on Co80 Based on Co”® Graphite- INOR-8-to- Density Burnug (g) Mole & Activation Activation to-Salt Selt (w/cm®) (% U23°) ole g (neutrons/cm?-sec) (neutrons/ecm?®-sec) Interface Interface? x 1012 x 1012 3¢ 25.532 0.7 0.993 750 ge 26,303 0.7 1.023 750 12 25,17 0.7 0.979 2.10 2.1 6804 610 67 5.5 24 25.37% 0.7 0.987 2.70 3.2 7602 605 83 7.0 36 24.886 0.7 0.968 2.7 3.3 yal 595 85 7.0 45 25.598 0.7 0.996 3.85 3.8 7od 600 117 9.7 aAverage external neutrqn flux, bEstimated temperatures. CUnirradiated controls. d'I‘hermocouple readings prior to termination of final irradiation cycle. Table 3. Exposure Data for Small Capsules in Assembly 47-4 : Thermal-Neutron Fast-Neutron Weight of Urenium Content Flux® Based (>3 Mev) Flux TemperatuTe . Power Calculated Capsule Fuel e ————— on Co%° * Based om Co”8 Re iona Density Burnug' (g) Mole % Activation Activation (§b) (w/em®) (% U233) : : g (neutrons/cm?-sec) (neutrons/cm?- sec) : x 1013 x 1012 lg ' 9.381 0.7 0.365 . 750 3 6.805 0.7 0.265 750 5 9.829 1.47 0.737 . 895 ' . 4 10.101 1.47 0.758 4.79 5.2 895¢ 260 11.4 6 9.915 0.7 0.386 1.31 1.3 7.5¢ 43 1.3 afiyerage external neutron flux. bUnirradiated controle. thermocouple readings prior to termination of final irradiation cycle. A ") R 13 punctured and the covér gas was recovered for analysis in the interval August 8 to October 23, 1962. These operations, conducted in the hot cells of Bldg. 4501 at ORNL, used a screw-driven puncturing tool sealed with a bellows and by & neoprene O-ring which butted on a flat previously ma- chined into each capsule (see Figures 1 and 2). Gas escaped into a collection system whose volume was calibrated and which had, in each case, been evacuated and checked for leaks. The initial collections were per- formed with a mercury-in-glass Toepler pump and glass éample bulbs in a glass and metal system; when this proved inadequate, as described below, it was réplaced by an all-metal gas collection system without a pump. This metal system was conditioned with elemental fluorine according to wéll-established profiedures before use. The first observations were made on the small capsule (No. 6) which had suffered the least burnup of uranium and on its out-of-pile control (No. 5). These capsules were handled without difficulty with the glass- metal system., After collection of the gas, the void volumes in the cap? sules were determined by pressure drop on expansion of helium from a known volume. Data obtained, including analyses of the gases by the mass spec- trometer, are shown in Table 4.-:The puncturing device leaked slightly, présumably atlthe O;ring séal,.during opening of irradiated Capsule 6. After correction of the analyses for the quantity of air (47%), the total volume of He + A agreed very well with that from the out-of-pile control. The Xe and Kr were.r6coverediat'verj_nearly the expected ratioc, and the absolute volume bf_these_gasesragreed reasonably well with the quantity (0.35 em? total) calculated from the burnup shown in Table 3. The only radiocactive materigls recovered on Toepler pumping from Capsule 6 were Kr and Xe. 14 Table 4 Cover Gas Analysis for Capsules 5 and 6 of Assembly 47-4 Capsule 5 : Capsule 6 Gas collected: 1.3 cm? Gas Collected: 3.2 cm’ Gas volume in capsule: 2.7 cm’® Corrected velue*: 1.7 cm’ , Gas volume in capsule: 2.5 cm? Gas Quantity Quantity Volume Volume % (cm3) g% (cm3) He 15 0.2 23 0.3 1.2 1.2 A 79 1.0 51 0.86 0 0.7 0.01 - - CO + N» 5.3 0.07 - - Kr 0 0 4.2 0.0 0.4 Xe 0 0 19.0 0.32 ' % Corrected for leakage of air (as indicated by mass-spectrometric analysis) at time of sampling. &) " =} ») 15 Cover gas recovery and analysis from all other capsules revealed very different behavior. Capsule No. 24, the first of the large capsules to be opened, showed the cover gas to be quite‘reactive even toward the Hg in the Toepler pump, and truly quantitative analysis of the gas was precluded by the inadequacy of the gas handling system. 1In spite of the considerable but unknown loss of gas by reaction, a large quantity (84 em? at STP) of gas was collected. The product of reaction with the Hg was shown to consist entirely of Hg,F;. Unlike the gas from Capsule 6 this material was quite radioactive; Te activity.(presumably was volatile TeFg) wae primarily responsible. Since the measured free volume in Capsule 24 at ambient temperature was shown to be 4.6 cm?®, a pressure of at least 18 atmospheres existed within the cap- sule at the time of puncturing. Based on mass spectrometry of two 0.1-cm? samples, the gas from Cap- sule 24 was 5% He, 0.1% Xe, 0.4% Kr, 4% (Co + N,), 6% CO,, and 17% CF,. The remainder was O, and SiF, in roughly equal amounts, matching the pro- ducts of the reaction of F, with the glass 810, + 2F; —» SiF, T + 0, T. The krypton yield listed above is in fair agreement with the calculated value, but the xenon yield is far too low. Later samples obtained by long exposuré of the capsule to an evacuated 3-liter container, had 0.3 cem’® of - xenon and less than one-tenth that'amount of krypton. All other capsules in this series were opened into an essentially all- metal, preconditioned:gas systenm, which1permitted much more representative ‘samples of the gas to be obtained. Table 5, for example, shows data ob- tained from Capsule 36, the near duplicate in irradiation conditions of Capsule 24. B L 1T T TS, ] 10 1 1 bS8 om0 1 11 W 0 80 A AR 1.0 O 16 Tfible 5 Analyses of Cover Gas from Irradiated Capsule 36 Volume of gas space in capsule at room temperatuie: 4.6 cm3 Volume of gas space gt operating temperature: 2.6 cm’® Volume of gas removed from capsule: 188 cm? Gas Quentity (vol %) Volume (em?) Fa 83.5 156 CF,, 10.0 | 19 0 2.5 4.7 Xe . 0 0 Kr 0.3 0.56 He 2.8 5.3 CO + Ne 0.1 0.19 002 007 lo3 ") n) ol 17 By contrast, Capsule 3A, which was an unirradiated control for Capsules 24 and 36 and which had been thermally cycled in a manner similar to their in-pile history, yielded 4.5 cm® (STP) of 99.7% He, 0.05% of CO + N, and 0;2% COp. The lack of A and the totai pressure of almost precisely 1 atmosphere was expected because of the speed and simplicity of the final closure of capsules of this type; this analysis undoubtedly represents the gas composition of Capsules 24 and 36 before irradiation. The pertinent data for all irradiatéd capsules in 47-4 are shown as Table 6. Inspection of this information shows that all capsules except No. 6 (previously described) yielded large quantities of F, and consider- able quantities of CF,. The total gas yield from Capsule 45 indicated a pressure of 60 atmospheres at room temperature before opening; had this gas been present at the operating pressure and in the considerably smaller volume then available it must have exceeded 250 atmospheres. It is almost certain that the capsule could not have withstood such a pressure even had the gas been inert. The large volume of gas drastically reduced the concentration of Kr and other minor constituehts, and the resctive nature of the samples required use of-special maés spectrometers whose precision was not of the highest Quality. Both'these*factors_ddversely.affect the precision and accuracy of the analyses.i It appears, however, that most, if not all, the Kr was recovered'from alléémples while the Xe is recovered poorly, if at all, from allrcapsules excépf Nb.'é. Though no positive proof is avail- able, it seems highiy probable that Xenon récovery fails because of forma- tion of .XeF;* (or pérhaps of 'ot’hér fluorides or oxyfluorides) which ma.y—be strongly adsorbed on the fuel or graphite within the system. Quantity and Composition of Gases from 47-4 Capsules Small Capsules Large Capsules Capsule No. 6 36 45 12 Fuel Content (g) 9.915 24..886 24,598 25.174 Uranium Content (g) 0.386 0.968 0.957 0.979 Burnup (%) 1.3 7.0 9.7 5.5 Sampling Date 8/6 9/7 10/16 10/23 Gas Collected (em?® STP) 1.6 188 270 153 Gas Composition (%) P, 0 84.1 88.7 48.1° CF[' 4-3 9-3 8-1. I*l-l-l Xe 9.0 0 0 0 Kr 4.1 0.4 0.8 0.2 He + A 2.6 2.7 2.95 4.0 Other - 3.5 0017 6-6 @Minimum value ;s some gas lost in reaction with container. Corrected from originel analysis which showed SiF, + 0,. " ) ") 19 While it is possible to see that the production of F, (or F, + CF,;) generally increases with quantity of uranium burned, no quantitative cor- relation of yields of either or both together with any of the experimental parameters can be shown. Moreover, the datea does not permit any decisions as to effect of cooling time on yield of these materials. It is worthy of note that the Fp + CF,; produced in the most extreme case (Capsule 45) répresents 3% of the fluorine contained in the salt or 1.6 times the quantity of fluorine in the UF, in that sample. It is clear that these data indicate very large‘losées of F, from the salt. Gross Exafiination of Capsules Visual examination 6f the capsules after removal from the 47-4 assembly showed them to be sound and in good condition. No dimensional éhanges were apparent. All capsules have been sectioned to permif direct observation of the metal, graphite, and salt. Figs. 3 and 4 show typical views (of Capsules 24 and 45) obtained after rough polishing of the sawed specimens. Several features in these views are noteworthy. The fuel salt is black in color, but seems to'shqw_no unusualrtendengy to érack or shatter during the saw- ing process. N§ evidence0fwétting of the graphité by the salt is apparent. Nonwetting of tbe”metal by the salt is 6ccasionally obéerved; this behavior, which isjUnfisual,,seéms-to be due to a layer of "scum" on the salt surface near the_capéule_fiall.' One bilt of curious béhavior, not matched,by the but-of-pile controls, is the oécurrence of'smoothééuffaced voids, which have appérently been gas bubbles during high temperature operation, in the bottom hemi spher- ical section of all the large capsules. o N UNCLASSIFIED 11213 after exposure. 2 -4 MTR~47 2 Capsule 24 Fig- 3. - UNCL ASSIFIED - HCO-1866 [ P - ‘. i ] b 'Fig; 4. Capsule 45 3 MTR-4’7-4, after exposure. 22 A Examination of the Metal | ' ki; Fluorine at elevated temperatures is an extremely corrosive material. | Fig. 5, for example, affords comparison of an untreated specimen of INOR-8 with one after exposure for 22 hours at 1100°F to flofiing helium containing 1% (by volume) of F,. An evaluation of the corrosion of the capsule metal should permit definitive statements as to whether F, was produced and was present during the high temperature operation'of Assembly 47-4. Accord- ingly, a very careful metallographic examination has been made of sections ) from Capsule 24, and selected specimens from other capsules are under study. Especial attention has been paid to metal surfaces exposed in the gas phase as well as at the gas-salt interface and fiithin the molten salt. Typical photomicrographs from Capsule 24 are shown as Figs} 6 and 7. There is no evidence of attack on the INOR-8 at any portion of Cap- sule 24; while definitive data are not yet available for other cafisules, their appearance makes it highly unlikely that evidence of attack will be observed. (No attack would be expected in an Out-of-pile capsule test of ¥ this duration at this general temperature level.) Further, there is evi- dence (see Fig. 7) that wall thickness has not been reduced, so the (w unlikely possibility of uniform corrosion followed by leaching of the corrosion film by the molten salt can also be excluded. It can, apparently, be stated with certainty that none of these capsule walls was exposed to fluorine at high temperature for apprecigble periods of time. Examination of Salt Small samples of salt from the irradiated capsules have been examined carefully with the optical microscope to determine the nature of the crys- talline material. These materials are in general, and as is frequently the i j X004 . X004 o q o] ) | | = o w| | O e L CIHONI GEO'Q ————e——reerrrerrerrrema ) e S3HINI S£0'0 ————————msmnane ] m o w3 O f =3 8 0 <% ? - o> — z . o = St % 23 ' o...wl_ -m + . 0w o g 8 S o o 3 ¥ ‘ 8 | - 5 . m* S o QL HH 9 - o . o By . . et O - 83 - n | . , 0 i : w g o m Q T h L N 2 . . wy | 35 , d“of ~ o g ¥ . _ , o3 &5 , o m . & . ' ¢ % { € i i 1 ok ETCHED Fig. 6. of capsule 24. CONTROL Longitudinal section 24 UNCLASSIFIED R-12958 ETCHED |RRADIATED of INOR-8 from hemispherical portion ah =~ - 25 UNCL.ASSIFIED R-11624 INCHES 0.02 Fig. 7. Transverse section of INOR-8 control material and specimen from hemispherical portion of capsule 24, MTR-47-4. Reduced 8%. 26 O case for rapidly cooled specimens, too fine grained to be positively iden- tified. However, a sampie from near the graphite at the center of Capsule 24 has been shown to contain the normally expected phases, though some are discolored and blackened; these phases included 7LiF-6Th¥, which was shown to contain UF, (green in color) in solid solution. It seems certain, therefore, that the uranium was not markedly reduced at the time the fuel was frozen. | An attempt was made, with Capsule 35;.to remove the salt by melting under an inert atmosphere td provide representative samples for chemical analysis and, especially, for a determination of the reducing power of the fluorine-deficient salt. This melting operation was not successful since only about one-half of the salt was readily removed in this way. That portion which melted and flowed from the capsule was green in color and obviously contained some tetravalent uranium. Since the salt was highly reduced--2% of the fluorine, approximately equivalent to that as UF;, had been lost--melting the material would be expected to result in deposition ¥) of some metal; this metal may have resulted in a high-melting heel of un- certain composition. n A variety of chemical analyses have been performed on materials from these capsules. No evidence of unusual corrosion behavior is observed in any of the analyses. Salt specimens have been examined by gamma ray spectrometry to eval- fiate gross behavior of fission product elements.» In general, the ruthenium activity appears to be concentrated at near the bottom of the capsules; this may indicate its existence largely in the metallic state so that a gravity separation could occur. Cerium as well as zirconium-niobium { J + 27 activity appears, as expected, to be fairly evenly distributed in the fuel. Some cesium, with some Zr-Nb, activity is found on metal surfaces exposed to the gas phase. Condition of the Graphite Graphite specimens from all the capsules‘have been recovered and examined visually with low power microscopes. No evidence of damage can be observed. The surface of the graphite appears identical to unirradiated specimens, the specimen in all cases appears structurally sound, and the saw cuts (see Figs. 3 and 4) appear normal. An autoradiograph of the graphite f}om Capsule 36 indicates quite uniform activity over the specimen with no visible evidence of fuel penetration. No quantitative measurements of physical and structural properties have been completed. Conclusions Of the six capsules in Assembly 47-4, five were found to contain large quantities'of F, and CF,; while the sixth, which had received a con- siderably smaller radiation dose, showed a small quantity of CF, but no F,. This latter capsule alone yielded the expected quantity of xenon while all others retained this element nearly quantitatively. The quantity of F~ released from the fuel melt génerally increased with increase in uranium fissioned (or any of the consequences of this) but no quantitative correla- tion was found. Neither -the. quantity of CF, nor the ratio CF,:F, showed an obvious correlation with fisSiOn"rate; burnup, or time of cooling before examination. | | The appearanée of tfie o§ened'capsules, and"all completed examination of cdmponents'from these, seem to preclude the possibility that the F, was present for any appreciable period during the high temperature operation of 28 this assembly. Accordingly, the generation of F, from the salt at low tem- peratures during the long cooling period must have occurred. Considerable energy is, of course, available from fission product decay to produce radiolytic reacfiions in the frozen salt, and the quantity of energy increases with burnup of uranium in the fuel salt. Capsules as small as those can absorb only a sfiall fraction of the gamma energy re- leased by fission product decay. On the other hand, a substantial fraction (perhaps as much as one-half) of the beta energy would be absorbed. A . yield (G value) of less than 0.04 fluorine atoms per 100 electron volts absorbed would suffice to produce the fluorine ohserved in any capsule. Production of F, with such a low G value seems & reasonsble hypothesis though no observations of fluorine production from salts on bombardment by B irradiation have been published. Attempts to demonstrate F; generation by eiectron bombardment of fuel mixtures at ambient temperatures are under vay. EXPERIMENT ORNL-MTR-47-5 Assembly ORNL-MTR-47-5 was designed, after the observations of CF, in the gas space in capsules from 47-3 referred to sbove, to include six cap- sules of INOR-8; four of these capsules contained graphite specimens such that large variation in salt-graphite interface area and in the ratio of graphite area:metal area in contact with salt was provided. The other two capsules, whose behavior is described below, were generally similar to the large capsules of Assembly 47-4. However, these capsules included addi- tional vapor space above the molten salt, and each was provided with inlet and outlet gas lines extending to manifold systems beyond the reactor. It was possible, accordingly, to sample the cover gas above the melt during ) (n a 29 reactor operation as well as during and after reactor shutdown. The two purged capsuies, referred to as A and B in the following and illustrated in Fig. 8, are 0.050-in. thick INOR-8 vertical cylinders, l-in. diameter x 2.25-in. long‘with hemispherical end caps. Each capsule contains a core of CBG graphite (see preceding éection) l/2-in. diameter x l-in. long submerged to a depth of approximately 0.3 in. in sabout 25 grams of salt. The top cap contains two l/4-in. diameter purge lines, é l/8-in. diameter thermocouple well, and a projéction to keep the graphite sub- merged. The two capsules differ only in the composition (primarily in the uranium concentration) of the fuel salt. Table 7 shofis the nominal compo- sition of the salt mixtures in each capsule. Preparation of the salt, and handling and filling operétions were very similar to those described for Assembly 47-4 above. The six capsules were, as was the case in previous experiments, sus- pended in a tank filled with sodium which served to transfer the thermal energy from the capsules to the tank wall. Heat was removed from the sodium tank through a variable annulus fiiled with helium to cooling water flowing in an external jacket. As in past experiments, no auxiliary heating was provided. The capsule wall afitained temperatures of about 600°C during high level power'operation. The twin manifolds which,comprise the pressure monitoring and gas collection systems were modefately compiex. - A schematic diagram of one of the assemblies is shéwn.as Fié.ié. .The equipment was éarefully calibrated to insufe that thesémplesdrafin as desired into 200 cc metal sampiing bulbs includéd a large and known fraction of the gas contained in the capsule. 30 UNCLASSIFIED ORNL—-LR—DWG 71570OR INOR—-8 THERMOCOUPLE WELL _* INOR-8 INOR-8 COVER GAS COVER GAS PURGE LINE PURGE LINE | ) \\ _ ) \ =3 N\ INOR-8 CAP \ \,—.-_ FOR N / (3 cm”) GAS SAMPLING 7 7 Y N . N—A A— ; = A ) T INTERFACE AR A _;'I:; ::/ MOLTEN-SALT FUEL NT EA - Kl HINRSKY—-- Y - ssad) 2RSSt es e cm AR TR Bl A1 Bk ~p INOR-8 CAN—" [}~ »’0’0’& 1/ / RS HPSRRS 7 '''' AL ey ==~ A - RN XY - - - - / R s / CHROMEL-ALUMEL [}~ — KX Jegied— Y Sen srapiTe THERMOCOUPLE—="FN (%] —— 1/ INTERFACE AREA »,020:0’«! }:0}:: Y (12.5 cm?) qd RN - - RIS =1 LRI INOR—8 CENTERING PIN ) == & & 2% ~—rt :.r/ z’: S > NICKEL POSITIONING LUG Fig. 8. Assembly 47-5 purged capsules. 31 Table 7. Fuel Composition Composition, mol % Fuel | Capsule A Capsule B LiF 6'7.36 67.19 BeF, 27.73 27.96 ZrF, 4,26 .51 UF, - 0.66 0.34 | Specific gravity at 1200°F 2.13 Liquidus 84.2°F Thermsl conductivity at 1200°F 3.21 Btu/hr-ft2.0F/ft Specific heat at 1200°F 0.455 Btu/1bme°F 32 UNCLASSIFIED ORNL-LR-DWG 77838 —— m— X XXX XK X i —P--XXXXX: —b - XXX X i |- | COUPLING —— MOLTEN SALT e SUBMERGED GRAPHITE SAMPLE BOTTLES CARBON TRAPS xxxoooc CAPILLARY TUBING PR PRESSURE RECORDER Fig. 9. ORNL-MTR-47-5 gas collection system. - XX XXX XXX DRIER HELIUM TO MTR OF F-GAS o ® ! 33 Unfortunately, this experimental assembly was designed and con- structed before the caPSUIes-of Assembiy 47-4 were opened. Accordingly, the entire complex was designed to handle CF, but not F,. In the short time between recognitidn of F, in the previous capsules and the insertion of 47-5 into the MTR, a number of materials changes were made to improve the ability to handle fluorine. For example, valves seated with fluorocar- bon plastic were substituted for’less desirable types in the assembly, and sample bulbs of preconditioned nickel were substituted for the original ones of stainless steel. However, some stainless steel tubing, which could not have been replaced without a complete dismantling and reconstruction operation requiring several months, was retained in the assembly. Short lengths of this tubing, which were welded to the INOR-S purge tubes within the sodium tank, attained the temperature of the sodium bath. Stainless steel contains CF, quite well, but it is not at all resistant to fluorine at elevated temperatures. Acco?dingly, evolution of F, with the assembly at high temperature might have been obscured by reaction of F, with the stainless steel. Observations below suggest very strongly that no such effect occurred. Beha#ior Ufider-Irradiation Assembly MTR447-5 was carried'through five MTR qycles during the four and one-half months ending January 23, 1963. Opefating conditions.for this assembly were varied as desired throughout the'test to include conditions anticipated-in'different regions of the MSRE. Temperature and power level ~ in Capsules A and B were'Variedfindependently over g considerable interval - but wete, insofar aS~possible3‘fiéldrconstant during periods represented by accumulagtion of the gas samples. Fuel temperaturés during reactor 34 operation varied from about 190°F to 1500°F with power levels ranging from 3 watts/ce to 80 watts/cc within the fuel. Gas samples were taken by isolating the capsule for a known and pre- determined interval (6 to 96 hours) and then purging the accumulated gases into the sample bulb. Pressure was carefully monitored (to the nearest 0.1 psi) while the capsule was isolated. A total of 59 gas samples were taken from the two capsules during the fifst'four MTR cycles under five distinct sets of conditions within the capsules. These sets of conditions (and numbers of samples) are: 1. Reactor at full power (40 Mw) with fuel molten, (34), 2. Reactor at intermediate power (5 to 20 Mw) with fuel frozen, (8), 3. Reactor shutdown with the fuel at about 90°F, (9), 4. Periods spanning fuel melting at reactor startup, (4), and 5. Periods spamning freezing of fuel following reactor shutdown (4). In addition, after termination of the experiment on January 23, the assembly was removed from MTR, the purge lines were capped with valves, and the assembly was quickly shipped to ORNL. Pressure monitoring of Cap- sules A and B at ambient temperature in the hot cell is under way, and samples of the evolving gas have been analyzed. Behavior with Reactor at Full Power Pressure Monitoring of the capsules with the reactor at 40 Mw and the fuel molten has shown no case of perceptible pressure rise. Moreover, none of the 32 gas samples taken during such conditions (over power levels rang- ing from 3 to 80 w/cc) have shown evidences of F,. [The presence of heated stainless steel in the lines might, as stated earlier, conceivably be responsible.] Analyses of gases collected from Cepsules A and B during the first MTR cycle are shown in Tables 8 and 9. These analyses, which are typical of all that have been collected under these conditions, show Table 8. Analysis of Gas Samples from Capsule A, MTR-47-5 During First MTR Cycle Fuel Composition Fuel Weight 24.822 grams (mole 7 Graphite Weight 5.696 grams LiF 68.0 UF, 0.7 Accumu- lation | _ Power Gas Analysis Time Temp. . Level Percent (by Volume) Parts per Million (hr.) (°F) (W/ee) He N, +CO O» H,0 A CO» H, Kr Xe CF, 29 1050 13 99,2 0.45 0.00 0.04 0.0L 0.11 0.15 24 95