Ty 3 Y456 03bLY4YY9L 5 ORNL-3253 UC-25 - Metals, Ceramics, and Materials CORROSION OF VOLATILITY PILOT PLANT MARK | INOR=-8 HYDROFLUORINATOR AND MARK Il L NICKEL FLUORINATOR AFTER FOURTEEN DISSOLUTION RUNS A. P. Litman (9130171 RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON ou wish someone else to see this yment, send in name with document the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION Printed in USA. Price $0"—?5. Available from the Office of Technical Services Department of Commerce Washington 25, D, C, LEGAL NOTICE This report was prépored as on account of Gevernment sponsored work. Meither the United Stotes, nor the Commission, ner any person octing on behalf of the Commission: A, Makes any warronty of representotion, sxpresssd or implied, with respect to the accurscy, completeness, or usefulness of the informotion contained in this report, or that the use of ony informotion, epparotus, method, or process disclosed in this repart may net infringe privately owned rights; or 8. Assumes any liabilities with respect to the use of, or for domages resulting from the use of any informotion, opparatus, method, or pracess disclosed in this report. As wused in the ocbove, ““person ecting on beholf of the Commission'" includes any employee or contracter of the Commission, or employee of such contracter, ta the extent that such employee or controctor of the Commission, or employee of such contracter prepares, disseminores, er provides access to, any information pursuont to his employment or contract with the Commission, or his employment with such contracror. - ORNL~ 3253 Contract No. W-7405-eng-26 METALLURGY DIVISTION CORROSION OF VOLATILITY PILOT PLANT MARK I INOR-8 HYDROFLUORINATOR AND MARK ITIT L NICKEL FLUORINATOR AFTER FOURTEEN DISSOLUTION RUNS A, P, Litman DATE ISSUED FEB -9 1362 OAK RIDGE NATTONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the U. 5. ATOMIC ENERGY COMMISSION VAT CORROSION OF VOLATILITY PILOT PLANT MARK I INOR-8 HYDROFLUORINATOR AND MARK TTIT L NTCKEL FLUORINATOR AFTER FOURTEEN DISSOLUTION RUNS "A. P. Litman ABSTRACT Process corrosion dccurring in the current Veolatility Pilot Plant hydrofluorinator and fluorinator in operation at the Oak Ridge National Laboratory was evaluated by visual .inspections, chemical analyses, “trans- port studies of Ni, Mo, Cr, and Fe, gamma radiography, wax replication, ultrasonic thickness measurements, -and metallographic studies. The modes, mechanisms, and rates of corrosive attack seem to agree well with previous experimental work. Significant bulk metal losses and moderate pitting attack were noted in both vessels. Maximum attack in the hydrofluorinator, which operated from 650 to 500°C, occurred in the middle vapor region at a calculated corrosion rate of 20 mils/month, based on 765 hr of molten _ fluoride salt residence time, and 0.06 mil/hr, based on 338.5 hr of hydrogen fluoride exposure time. The fluorinator, which operated at about 500°C, sfistained maximum bulk metal losses in the lower vapor region at calculated rates of 28 mils/month, based on 694 hr of salt residence time, and 0.9 mil/hr, based on 30.9 hr of fluorine exposure time. Calculations based on losses in wall thickness indicate that both vessels should be capable of handling approximately 100 additional dissolution runs. These calculations include pitting corrosion in the hydrofluorinator but ignore effects resulting from intergranular or other forms of selective attack which may be present in both vessels. 1. INTRODUCTION The Oak Ridge National Laboratory (ORNL) Fluoride Volatility Process is being developed as a nonaqueous technique for reprocessing zirconium-clad highly enriched uranium fuel elements or homogeneous -2 - fluoride salt mixtures (such as the NaF-ZrF,-UF, Alrcraft Reactor BExper- iment fuel which has been processed, or the LiF-BeF,-ThF,-ZrF,-UF, fuel from the proposed Molten Salt Reactor Experiment). In heterogeneous fuels the zirconium and uranium are converted to their respective tetra-- fluorides in an NaF-ZrF, or NaF-LiF-ZrF, melt, with HF used as the oxi- dant. The UF, is further oxidized, in a differént vessel, to UFg by contact with elemental fluorine. The volatile UFg is purified by an absorption-desorption cycle on NaF pellets and collected in cold traps. The hydrofluorination step is not required for homogeneous fuels. The nature and extent of the corrosion occurring in earlier hydro- fluorination énd fluorination vessels, which must sustain the most cor- rosive environments in the process, have been discussed in other reports, -3 The present Volatility Pilot Plant {VPP) hydrofluorinator is an INOR-8% vessel approximately 17 ft in height and consists of a top right cylinder 24 in. in diameter, a bottom cylinder 5 1/2 in. in diameter, and a conical section connecting the two cylinders. The top cylinder was formed from 3/8—in. rolled and welded plate, the bottom cylinder from 1/4-in, plate, and the conical sectionifrom 1/2-in. plate. The VPP fluorination vessel was constructed entirely from 3/8-in.- thick L (low-carbon) nickel plate., The fluorinator was made by Jjoining two 1l6-in.-diam right cylinders with a 5.5-in.-diam neck, The combined assembly hds a height of 7 ft. Figure 1 is a simplified schematic dia- gram of the current VPP flowsheet showing relative positilons and configu- rations of the Mark I INOR-8 hydrofluorinator and the-Mark ITI L nickel fluorinator. 1A, P, Litman and A. E. Goldman, Corrosion Associated with Fluori- nation in the Oak Ridge National Laboratory Fluoride Volatility Process, ORNL-2832 (June 5, 1961). °p, D. Miller et al., Construction Materials for the Hydrofluorinator of the Fluoride-Volatility Process, BMI-1348 (June 3, 1959). 3A. E. Goldman and A. P. Litman, Corrosion Associated with Hydro- fluorination in the Oak Ridge National Laboratory Fluoride Volatility Process, ORNL-2833 (November 1961). “Nominal INOR-8 composition (wt %): 71 Ni—16 Mo—7 Cr—5 Fe. LIQUID WASTE UNCLASSIFIED ORNL-LR-DWG 61615 Fa DISPOSAL I DESORPTION /o IN. NaF SELLETS OUTLET UFe COLD | F2 TRAPS AQ. KOH SPRAY NaF -LiF-ZrFa ABSORPTION Fa TOWER H2 ' OUTLET UFe NEUTRALIZER MOVABLE COLLECTION { 10% KOH) BED ABSORBER . 400-100%C T} LIQUID WASTE HF + Ho | i FUEL CHARGING | Ho CHUTE F2 FREEZE VALVE 3 HF f— _1» . ’ NaF -LiF-2r Fyq RECYCLE k\ /) YSTEM . S - NaF -LiF -ZrF4-UF 4 \ r , WASTE SALT Zru H ( """"" “: SPENT | = . = ! FUEL L | : L so0o [ ! o ! = 650-500°C 1 ! L .7 | \ LN ] i — HF ) FREEZE ” - [T_:J LINE FLUCRINATOR HYDROFLUOQRINATOR Fig. 1. Simplified Flowsheet — ORNL Fluoride Volatility Process. - - 2., HYDROFLUORINATOR CORROSTION Corrosion in the Volafility Pilot Plant was observed through 14 dissolution runs using both Zircaloy-2 dummy fuel elements and nonirra- diated Zircaloy-2 fuel elements containing 0.2 to 0.4 wt % U. Hydro- fluorination process cycling for the dissolution runs is detailed in Table l.' Special efforts were made to keep temperatures as low as practical by using lower melting NaF-LiF-ZrF, salt baths rather than NaPF- ZrF, melts and by reducing the initial dissolution bath temperature of 650°C to about 500°C as rapidly as possible, Higher temperatures have a strongly adverse effect on Volatility Process corrosion, ™3 2.1 Reaction to Environment After run TU-7, the hydrofluorinator was carefully inspected for determining the extent of process corrosion and for evaluating its‘future usefulness for processing naval reactor fuel elements. The nondestructive inspection procedures included visual examination, chemical analyses of corrosion-dissolution products, transport studies of Ni, Mo, Cr, and Fe, complete gamma radiography of the vessel walls, wax replication of a portion of the interior walls of the hydrofluorinator, and ultrasonic thickness measurements to determine bulk metal losses. There were no surveillance corrosion specimens included in the hydrofluorinator., How- ever, metallographic examination was done on an INOR-8 pipe support clip which had been exposed to the same environments as the salt regibn of the hydrofluorinator vessel proper. 2.1l.1 Visual Examination Following an inspection of the interior walls of the hydrofluorinator by means of the naked eye, the visual inspection techniques consisted in low-magnification viewing with the Omniscope,” an interior-type periscope, and the Questar,® a catadioptric instrument using a combined lens-mirror *Manufactured by Lerma Corp., Northampton, Mass. éManufactured by Questar, Inc., New Hope, Penna. Table 1. VPP Mark 1 INOR+8 Hydrofluorinator Process History - | Wall T erat Salt Composition, Vesse emperature Run No. of Fuel NaF-LiF-ZrF4 During Dissoclution (°C) Molten Salt HF Flow HF Exposure Designation Elements Dissolved (mole %) Vapor Regicn Salt Region Residence Time Rate Time - (hr) (g/min) (hr) Initial Final Moximum Minimum Maximum Minimum _ Salt transfer 27-27-46 27-27-46 435 425 565 535 44.5 0 0 studies. ‘ : T-14 1 43-22-35b 31-17-52 410 N.A, 560 520 87 104 38 T-2 1 38-30-32b 30-27-43 380 N.A. 530 525 . 95 40 41.5 T-3 1 39-39-22b 31-26-43 575 N.A, 625 550 . 57 . 90 25 T-4 1 38-37-25b 30-27-43 590 500 630 495 51.5 90 22,5 T-5 1 ' 38-37-25b 30-27-43 570 460 625 530 62 90 27 T-6 1 38-37-25b 31-30-39 550 390 650 500 59.5 118, 150 26 T-7 2 37-37-26b 27-27-46 390 440 655 525 53 135 23 TU-1€ 2 40-35-257 34.26.40 560 400 670 490 44 92 23 ( + 0.3 wt% U) TU-2 2 37.38-25% 30-27-43 620 450 650 500 35 125 24 (+0.2 wt% U) TU-3 2 33-41-26d 30-31-39 570 390 655 495 32.5 150 19.5 (+0.3 wt % U) _ , ) TU-4 2 42-34-24d 37-29-34 575 440 650 535 36 146 11 TU-5 1 42-35-23d 29-29-42 525 400 650 500 : 36 150 22 TU-6 1 33-37-25d 34-28-38 525 385 650 500 35 121 19 ( + 0-3 wt % U) " TU-7 1 : 39-38-23d 30-30-40 515 405 650 500 37 150 _ 17 (+0.2wt% U) 765 338.5 %Simulated fuel element; made of Zircaloy-2. bBarren salt charge; 60 to 115 kg, ©Zircaloy-2 cladding; Zr-U matrix. Barren salt charge; 95 to 110 kg. -6 - system. In order to examine the vertical walls. of the hydrofluorinator with the Questar, highly polished mirrors and a light source, both held by thin structural members, were lowered into the vessel and the Questar was positioned above the fuel-charging chute.’ The regions of the hydrofluorinator that had been in contact with process vapors were covered with a beige-colored deposit, while the salt- containing area appeared to be black with tints of green and brown. Dark- colored coarse flakes of material were noted on top of the hydrogen fluoride distributor plate (Fig. 2). The flakes had the appearance of dried salts discolored by corrosion-dissolution products. (The LiF-NaF- Zr¥, salts used during dissolution are white in color and develop a greenish cast as UF, is added to the mixture.) The deposits on the dis- tributor plate also appeared to contain metallic particles., Some of the ’defiosits adhering to the vertical walls and interior piping in the salt -region were of appreciable thickness and were loosely adherent, Typical of thése'fegions'is the deposit noted on the thermocouple well and shown in Fig. 3. In order to study the actual walls of the hydrofluorinator, several | attempts were made to remove the corrosion-dissolution deposits. The cleaning schedule used and subsequent observations are given in Table 2. Despite the ambitious cleaning program, about half of the wall deposits remained tenaciously attached to the walls of the hydrofluorinator. Addi- tional low-magnification studies of the interior of the vessel gave indi- cations of pitting attack, especially in the conical section joining the top and bottom right cylinders and in the lower third of the top cylinder. (More information on this pitting attack was acquired later by gamma, radiography and wax replication of the hydrofluorinator and is given in Sec. 2.1.3.) 2.1.2 Chemistry of Corrosion — Transport of Ni, Mo, Cr, and Fe Chromium and iron, constituents of the INOR-8 hydrofluorinator, read- ily react at elevated temperatures with HF gas to form fluorides, probably '7J. B. Ruch designed and operated the Questar mirror-light extension apparatus and was instrumental in producing the photographs in Figs. 2 and 3. R. E. McDonald served as consultant in this work. UNCLASSIFIED PHOTO 540%0 L " 4 . = = y - - .r,'._’. i.* * fan 4 - 5 ¥ b s ‘v“ d‘-:ur .' Sk 2 E - v - | v » Fig. 2. Portion of the HF Distributor Plate Inside the VPP Mark I INOR-8 Hydrofluorinator After Run TU-7 Showing Typical Deposits Occurring from Dissolution and Corrosion. Photograph taken through Questar optical system. Approx 2.5X. UNCLASSIFIED PHOTO 53644 Fig. 3. Portion of the Thermocouple Well in the Lower Salt Region of the VPP Mark I INOR-8 Hydrofluorinstor After Run TU-7 Showing Typical Loosely Adherent Deposits. Photograph taken through the Omniscope optical system, Approx 4X. -9 - Table 2. Cleaning Schedule of VPP Mark I INOR-8 Hydrofluorinator After Run TU-7 Treatment Observations 0.35 M ammonium oxalate, at 60°C for 4 hr 4 high-agitation (by nitrogen sparging) water rinses at 25°C for 8 hr ‘ Nitric acid (5 wt % in water) at 25°C in lower 3 ft of vessel for 3 hr 0.35 M ammonium oxalate at 95 to 100°C for 4 hr Several fiater rinses at 25°C for 6 hr Aluminum nitrate (5 wt % in water) (adjusted to 3.5 pH by potassium hydroxide) at 25°C for 7 hr; high agitation by nitrogen sparging Vapor and salt regions had a dull- gray appearance; heavy deposits still present in bottom of vessel Same as above except some lcoose deposits had been removed from bottom ' Bottom of hydrofluorinator seemed brighter and less congested with loose deposits; crystals similar to metal whiskers noted on the pipe support clips None No significant change“in vessel appearance - About half the wall deposits still present, mostly in the lower portions of the vessel; most of the loose material had fallen to the bottom of the vessel - 10 - CrF, and FeF,. Also, NiF, can be produced during hydrofluorination, but only because of the continuous removal of hydrogen since the free energy of formation (above 490°C) of nickel fluoride by Ni + HF reaction is not 8 Similarly, molybdenum, the other major constituent in INOR-E, favorable. can be forced to react with IF even though a positive free-energy change is involved. Evidence of small but finite dissolution rates of molybdenum metal during hydrofluorination conditions has been reported.® All the oxidation reactions indicated above are part of the initial stages of corrosion in the VPP hydrofluorinator. The corrosion resistance of the hydrofluorinator would be consider- ably enhanced if the corrosion products (fluorides of Ni, Mo, Cr, and Fe) were adherent to the hydrofluorinator wa}ls and impervious to the further passage of HF, ofVlOW'volatility, unaffected by ercsion, and not soluble in nor complexible with the NaF-LiF-ZrF, dissclution bath. However, most of these conditions do not exist and therefore the fluoride films are not protective. Morecver, all or some of the fluorides formed from the INOR-8 + HF reaction can be reduced by the zirconium metal present in the system, by more electropositive elements present (for example, Cr metal should reduce the fluorides of Mo, Ni, and Fe) or by hydrogen gas from dissclution and corrosion reactions. At 600°C, the free energies of formation (relative to HF gas as zero) of MoF,, NiF,, FeF,, CrF,, and ZrF, are +14, +2, -5, -12, and -31 kcal per gram-atom of fluorine, re- _spectively.8 Finely divided particles of Ni, Mo, Fe, and Cr, often 10 are evidence of found in the off-gas stream from the hydroflucrinator, the postulated reduction reactions. The oxidation-reduction cycles described are believed to account for the major source of corrosion during the hydrofluorination stage of the 8A. P. Litwan and R. P. Milford, "Corrosion Associated with the Oak Ridge National Laboratory Fused Salt Fluoride Volatility Process,” presented at Symposium on Fused Salt Corrosion, Fall Meeting of the Electrochemical Society, Detroit, Michigan, October 1-5, 196l. °A. E. Goldman and A. P. Litman, Corrosion Associated with Hydro- flucrination in the Oak Ridge National Laboratory Fluoride Volatility Process, ORNL-2833, pp 45, 49-50 (November 1961). 108, c. Moncrief, Results of Volatility Pilot Plant Dissolution Run T-7, ORNL CF-60-12~-57 (Dec. 22, 1960), i - 11 - Fluoride Volatility Process. Since laboratory experiments are complicated by the numerous possible interactions, most of the knowledge gained to date has been through small-scale process studies and subsequent exami- nation of process development vessels, Chemical analyses of the off-gas stream from the hydrofluorinator, of the waste salt from the VPP, and of the movable-bed absorber have been studied191? as a means of estimating past and future corrosion cf the dissolver vessel. Figure 4 shows the transport paths of Ni, Mo, Cr, and Fe during the 14 dissolution runs, and approximate mean percentages of the total of each element found either as a metal or as a fluoride; Quantitatively, during each run considerably more Ni, Cr, and Fe were found in the waste salt, movable-bed absorber, or off-gas system than had been introduced into the system from the feed salts or the fuel ele- ments. This waé the first positive indication of hydrofluorinator cor- rosion. On the other hand, only about half the molybdenum introduced from the feed salt was found in other regions of the process system.n"'l2 Other work® indicates that the molybdenum may be plating out in the hydrofluorinator and/or possiblyl4 combining as an intermetallic compound with nickel in the INOR-8. The transport and final disposition(s) of molybdenum are being investigated. 2.1.3 Gamma Radiography After run TU-7 the entire shell of the hydrofluorinator was radio- graphed to help confirm or deny the pitting corrosion first noted by , 1lg, ¢. Moncrief, Results of Volatility Pilot Plant Cold Uranium Flowsheet Demonstration Run TU-1, ORNL CF-61-6-62 (June 22, 1961). 12p, . Moncrief, Results of Volatility Pilot Plant Dissolution Run TU~2, ORNL CF-61-6-79 (June 23, 1961). 138, C. Moncrief, Results of Volatility Pilot Plant Dissolfition Run TU-3, ORNL CF-61-7-26 (July 13, 1961). '147, H, DeVan, private communication, Aug. 29, 1961, - 12 - UNCLASSIFIED ORNL-LR-DWG 61591 —= UF, PRODUCT. MOVABLE BED ABSORBER o * ¥ [20% Cr o NI 20% Mo > %Mo % PERCENTAGES . 5% Fe (AS FLUORIDES) OF TOTAL RECOVERED ~1%Cr \ OFF-GAS | 32 : (AS METAL AND - FLUORIDES) WASTE SALT PLUS Ni,Cr,Mo, Fe fi BARREN SALT WASTE SALT CONTAINER * v | ‘ FLUORINATOR 80% Ni L NICKEL FUEL ELEMENT gy (Ze-Sn-U NaF-LiF-ZrF, 80 Cr . + + ° Ni, Fe, Cr) Ni, Mo, Cr, Fe (AS FLUORIDES) HF HYDROFLUORINATOR INOR-8 (Ni-Mo-Cr-Fe) Fig. 4. Transport Paths of Ni, Mo, Cr, and Fe in the VPP Fluoride Volatility Process During l4 Dissolution Runs. Percentages are approximate means of the totals of each element found in the off-gas system, movable- bed absorber,and waste salt. . - 13 - visual examination and to reveal the condition of the weld joints and possible cracking in the vessel walls. Cassettes containing film were wrapped around the exterior walls of the vessel, and an Iri®? gamma source was located inside the hydroflubrinator along the vertical center line. The source strength at the time of use was 17 curies. A 5-mil lead foil directly in front of the film and a 10-mil lead shield behind the film were used to prevent scatter and to intensify the image. - The radiographic film revealed only three regions in the hydro- fluorinator where pitting seemed to be present. One area, which had been exposed to process vapors, was within a 2~-in.-diam circle about 36 in. down from the top of the upper cylinder in the north quadrant. The deepest pit here, based on a 2% sensitivity factor for the film and radicgraphic technique, was about 7.5 mils. The other two areas were in the east gquadrant of the 1/2-in.-thick conical section oflthe hydro- fluorinator and about 6 in. down from the top of the cone. The maximum depth of pitting here seemed to be about 10 mils. | All weld joints appeared to be in good condition on the hydro- fluorination vessel, and no macrocracking was obvious. 2.1l.4 Wax Replication To cross-check the radiographic work and to prove'conclusively the pitting attack, wax replications were made of the interior walls of the vessel, The technique for cobtaining these impressions was proposed by R. P. Milford (Chemical Technology Division) and the early experimental work was done by Smifih;l5 important modifications and the actual wall impressions were made by Crump.l6 Briefly, the replication technique involved lowering into the hydro- fluorinator and horizontally positioning an air cylinder to which was attached a small container filled with dental mclding wax, heating the 15M. 0. Smith, Volatility — Proposed Wax Replication Technique for Measuring Hydrofluorinator Pit Depths, ORNL CF-61-3-59 (Mar. 15, 1961). 16, F. Crump, Jr., Wax Replication Studies of Corrosion Pits in the Volatility Pilot Plant Hydrofluorinator, ORNL CF-61-8-20 (Aug. 9, 1961). - 14 - wax to a suitable temperature, and activating the air cylinder to press the softened wax against the wall of the vessel, Figure 5 shows the replication device in mockup position auring heating. Since the device, as presently designed, requires largé internal clearances, only portions of the top cylinder of the hydrofluorinator could be replicated. Initial replication confirmed the pitting attack in the vapor-phase region observed on the radiographic film. Duplicate impressions, at different times, disclosed that the maximum pitting depth here was 8 mils. Figure 6 shows portions of the replicétions illustrating the configuration of the pit. The height of the replication was measured by a toolmaker's microscope and was cross-checked by an optical comparator and a dial gage with a long lever arm. The excellent results obtained with the initial replication made it desirable to extend the work to a survey of the entire top cylinder. These replications were done in a spiral, clockwise pattern starting at the top of the cylinder and continuing down to the conical section. The results are given in Table 3. While the 8-mil pit previously found proved to be the deepest discontinuity, many other pits were found varying from 0.5 to 5.5 mils in depth. | Replications were also made of portions of circumferential and longi- - tudinal welds in the hydrofluorinatof; portions of the replications are shown in Fig. 7. The impressions indicated the welds to be in good condi- tion with little, if any, evidence of corrosive attack. 2.1.5 Ultrasonic Thickness Measurements A Vidigage,l7 which had an accuracy of approximately il%, was used to determine wall thinning. Readings were taken every 3 surface contour inches, generally vertical, in all four guadrants, and the results were compared with base-line data obtained with the same instrument when the hydrofluorinator was installed in the pilot plant, The losses in wall thickness are plotted in Fig. 8 vs elevation, starting from the top _ 17pn ultrasonic measuring device, manufactured by Branson Instruments, Inc., Stamford, Conn. ‘ ‘™ | UNCLASSIFIED ! PHOTO 55021 CLEVIS a ; WAX CONTAINER WA X HEAT SOQURCE AIR CYLINDER * oLl Fig. 5. Wax Replication Device Used for Obtaining Impressions of Pits in the VPP Hydrofluorinator. In order to insert or remove device through the fuel charging chute, the air cylinder was positioned vertically within the clevis. SR UNCLASSIFIED s PHOTO 54901 Fig. 6. Duplicate Wax Replications of an 8-mil Pit (Encircled) Found in the VPP Hydrofluorinator North Quadrant, About 36 in. Down from the Top of the Upper Cylinder. ») - 17 - Table 3. Depth of Pits in the Vapor Region of the VPP Mark I INOR-8& Hydrofluorinator as Determined by Wax Replication Distance Down from Top Cylinder Pit Depth (in.) Quadrant (mils) Initial Replication 32 North W ~ I_l MO - W I~ H N o0 . - V East Later Replications North 5.5, 5.1, U1 -3 W\t \in , 3.5, 3.1, 2.8 OO O East N |—J HOoOOOOO oyttt Wi - |_‘- . . _b'- . South W W O . Wn 36 3.5, 3.2, - Do \.-l-\ N O West UIUI\.(\JUILN Ut &g W O I~ X OO0OWWO ~MPOOM North 63 66 69 * U\ \in O oo a I8 - UNCLASSIFIED PHOTO 54922 UNCL ASSIFIED PHOTO 54912 Fig. 7. Portions of Wax Replications of Welds in the VPP Hydrofluori- nator. (a) Circumferential weld, southwest quadrant, 48 in. down from top of vessel; (b) longitudinal weld, southeast quadrant, 21 in. down from top of vessel. Approx 4X. ) # - 19 - 24-in.0D T I YT RREFIN T TOP 72 in. —3g-in. WALL WLttt fo i n 16 in. -———'/2-in. WALL 196 in. e = 51/2-in. OD 108 in. b 1, /4 n. WALL BOTTOM CYLINDER - ::E::__—- FV-1000 MATERIAL: INOR-8 SUMMARY OF PROCESS CONDITIONS FOR 14 DISSOLUTION RUNS CYLINDER _ / TEMP (°C) 380-620, VAPOR REGION 490-670, SALT REGION TIME (hr) 765-SALT 338.5-HF SALT COMPOSITION Ncfl'_-l_iF-Zr'E4 (27/43 -17/41-22/52 mole%) PLUS 0.2/0.4 wt% U Fig. 8. o \ CONTOUR SURFACE INCHES FROM TOP UNCLASSIFIED CRNL-LR—-DWG 61589 0.06 mils/hr (HF TIME) 20 mils/month (SALT TIME) /////// - 6 VAPOR REGION AVERAGE LOSS-6 mils SALT REGION VAPOR-SALT INTERFACE REGION AVERAGE LOSS-6 mils o £ ¥ n 0 O 3 w o ! 14 L > a l 75in, 120 in. 0 S 10 15 20 WALL THICKNESS LOSS (mils) ——X INDICATES PITTING ATTACK the VPP Mark I INOR-8 Hydrofluorinator. Wall-Thickness Losses and a Portion of the Pitting Attack on - 20 - of the upper cylinder. ©Since the losses did not seem to vary particu- larly with gquadrant geometry, they were plotted at each elevation. The maximum and minimum loss points in the bottom and top cylinders and in the conical connector were connected to form loss ranges in each geometric region, Average bulk metal losses in the top and bottom cylinders were 6 and 4 mils, respectively, but maximum losses were 18 and 12 mils. The average wall-thickness loss in the connecting cone was 10 mils, while the maximum loss there was 16 mils, Similarly, the average losses in the salt, salt- vapor interface, and vapor regions were, respectively, 4, 6, and 6 mils, 'with corresponding maximum losses of 12, 16, and 18 mils. - The depths of the pits in the upper portion of the hydrofluorinator, as determined by wax replication, are shown in Fig. 8 as extensions of the appropriate wall-thickness loss points, depending on the particular quadrants where each form of corrosion was found. The maximum attack to date occurred in the middle vapor region at a calculated corrosion rate of 20 mils/month, based on 765 hr of molten fluoride salt residence time, and 0.06 mil/hr, based on 338.5 hr of HF exposure time, 2.1.6 Metallography and Chemistry of Corrosion Product Layers Additional information on INOR-8 corrosion in hydrofluorination serv- - ice was obtained by examination of a pipe support clip originally located below the salt level ih the current VPP hydrofluorinator. Metal thickness .losses of 5 mils were noted by micrometer measurement. These correspond to salt-phase hydrofluorinator wall thinning found by ultrasonic thickness measurements, A black, semiadherent scale was present on the clip and was found to be depleted in chromium and iron when compared with the base metal (Fig. 9). The nickel content of the scale was comparable with the nominal INOR-8 analysis, but the molybdenum cbntent was considerably higher. Scale x-ray diffraction patterns could not be matched with known compounds. Since the metallic content of the scale was less than 50% of the total weight, a good possibility exists that the scale is an Ni-Mo-F complex, - 21 - UNCLASSIFIED ORNL-LR-DWG 61655R2 O CORROSION PRODUCT ANALYSES (wt %) Ni Mo Cr Fe 0002 , 69 28 Q.l 0.4(a) i e—66 3l 0.3 0.6(b) < -~ 19 = 5 (c) 7 16 7 5 (d) 0.004 - 00061 - 500 X ETCH: MODIFIED CHROME REGIA Fig. 9. Corrosion Product Layers on INOR-8 as the Result of Hydro- fluorination Service Showing Composition of (a) Surface Scale (Not Shown in Photomicrograph), (b) Spongy-Surface Layer, (c) Subsurface Layer, and (d) Substrate Base Metal. - 22 - Metallographic examination of the clip disclosed a soft (53-63 diamond-point hardness) spongy-surface layer and a subsurface layer of variable hardness (115-335 DPH) as compared with a base INOR-8 value of 210 to 234 DPH (Fig. 10). Millings from the two layers were subjected to spectrochemical analyses, with the results shown in Fig. 9. The spongy- surface layer had approximately the same metallic composition as the scale previously described. The subsurface layer was deficient in chromium, had a slightly higher molybdenum content, and had the same iron content when compared with the base metal. 3. FLUORINATOR CORROSION As indicated in Fig., 1, after the fuel elements were dissolved in the hydrofluorinator, the process salts, containing uranium as UF,, were trans- ferred into the L nickel fluorinator, where the UF, was fluorinated to the process product UFg. Process history for the current VPP fluorinator is shown in Table 4. It should be noted that prior to exposure of the vessel to fluoride salts the unit was fluorine-conditioned at temperatures greater than subsequent operating temperatures, in accord with previous Studies.18 3.1 Reaction to Environment After run TU-7, the Mark III fluorinator was subjected to an ammonium oxalate wash and then to several water rinses to remove process salts. The fluorinator was examined by visual inspection, chemical analysis of wall’ deposits, complete gamma radiography of the fluorinator walls, and ultra- sonic thickness measurements to determine corrosion. 3.1.1 Visual Inspection — Wall Deposit Chemistry The interior of the fluorinator was inspected by means of the unaided eye and the Omniscope,”? All interior areas which had been exposed to IBA. P, Litman and A. E. Goldman, Corrosion Associated with Fluori- nation in the Oak Ridge National Laboratory Fluoride Volatility Process, pp 28-30, ORNL-2832 (June 5, 1961). UNCLASSIFIED PHOTO 54013 (@) ETCH: MODIFIED CHROME REGIA UNCLASSIFIED UNCLASSIFIED Y=a1182 ¥Y-angl e ———— D O (&) (¢) AS POLISHED AS POLISHED UNCLASSIFIED Y=41181 PP LA v DPH e G e "'qr';_‘-_"'- (500!} LOAD) o 335 SRR s R > 0.004 0.002 0.004 . DEFORMATION 0 ’1;7 LINES 4 234 (d) AS POLISHED Fig. 10. Corrosion Product Layers on INOR-8 After Exposure to - drous HF and Fused LiF-NaF-ZrF, at 500-650°C. Typical areas showing (a uniform duplex surface layers, (b) uniform intermediate layer, discontinuous surface layer, and hardness impressions, (c) variable thickness intermediate layer, discontinuous surface layer, and tendency for separation of the product layers, (d) an unusually thick intermediate layer showing brittle nature of layer compared with base metal. Table 4. VPP Mark lIl L Nickel Fluorinator Process History Vessel Wall Temperatures (°C) Salt Molten F2 Flow Run Composition, Without F2a With F, Salt Rate F2 Designation NaF-Li F-ZrF, Residence (liters/min E.xposure (mole %) Vapor Salt Vapor Salt Time (hr) at STP) Time (hr) Region Region Region Region F, conditioning None 630-310 670-560 0 5 psig 53 . (static) Salt transfer 27-27-46 ~375 565-550 49 0 None studies T-1 31-17-52 175-150 555-540 38 0 None T-2 30-27-43 180-150" 570-545 78 0 None T-3 31-26-43 170-160 550-525 94 0 None T-4 30-27-43 340-325 515-485 15 0 None T-5 30-27-43 375-360 580-550 27 0. None T-6 31-30-39 480-475 575-560 15 0 None T-7 27-27-46 360-340 605-580 18 0 None F. sparge tests None 525-500 575-550 520-495 570-535 45 B8 12.0 TU-1 34-26-40 405-375 530-510 360-350 505-500 52 12 2.5 {+ 0.3 wt % U) TU-2 30-27-43 475-450 515-490 460-.440 505-500 40 6 2.5 (+ 0.2 wt % U) TU-3 30-31-39 325-300 575-550 330-320 510-500 38 9 1.75 (+0.3wt%U) TU-4 37-29-34 325-300 535-510 330-315 515-500 50 4, 12 1.25 1.0 (+ 0.4 wt% ) TU-5 29-29-42 390-350 570-525 360-325 510-500 36 4, 16 1.0, 0,5 (+0.3wt%U) TU-6 34-28-38 375-350 580-560 360-340 505-500 34 6, 16 1.0, 0.4 {+ 0.3 wt % U) TU-7 30-30-40 360-340 575-550 360-340 510-500 65 6, 16 1.0, 0.7 {+ 0.2 wt % U) 694 30.9 a . . O . . . O . . . In a few instances, excursions to $80°C maximum in salt region ond 650°C moximum in vapor region were noted. -478— ¢ - 25 - process vapors were covered with relatively thick, light-green deposits. The process vapors were usually F, and UFg plus volatile molybdenum and chromium fluoride corrosion products, probably as MoFg and CrFs. The middle-neck region of the vessel, which joins the top and bottom right cylinders, showed the heaviest deposits. Most of the deposits seemed to be loosely adherent, and where the deposits had flaked off, the substrate metal had a dark, dull appearance. PFigure 11 illustrates these findings in the upper conical section of the fluorinator. At higher magnificatioh evidence of pitting corrosion was found, especially in thé top cylinder of the vessel, _ Samples collected from the deposits in the middle neck and in the upper neck (top-flange region) were ana;yzed by wet chemistry and x-ray diffraction. The latter indicated the deposits to be greater than 90% NiF,. Complete analyseé of the deposits are shown in Table 5, The salt- cortaining region of the fluorinator Wés free of deposits and had a bright, etched appearance. Table 5. Chemical Analyses of Vapor-Region Deposits in the Mark III L Nickel Fluorinator After Run TU-7 and an Ammonium Oxalate Wash Analysis (wt %) Element - Middle Neck “Top-Flange Region Ni 57.3 51.0 Cr 1.0 1.2 Fe 0.1 2.0 Mo 0.03 0.01 Sn 1.0 0.1 Zr _ 0.6 0.9 Na ' 0.9 1.0 Li 0.08 - 0.05 U 14 ppm 22 ppm 3.1.2 Gamma Radiography The technique for obtaining complete gamma radiography of the walls of the VPP fluorinator was similar to that described for the hydrofluori- nator. Evidence of pitting corrosion was noted in thg top cylinder and - 26 - UNCLASSIFIED | PHOTO 55075 SALT SAMPLER LINE 0 2 4 INCHES Fig. 11. Interior of the VPP Mark IIT L Nickel Fluorinator After Run TU-7 Showing Upper Conical Section. Loosely adherent NiF, deposits have flaked off in several regions, revealing darker colored substrate metal., Approx 1/2 size. —% - 27 - cone of the vessel, énd the attack seemed to be especially intensive in the top and bottom third of the cylinder. The pits appeared to be slight- ly deeper than those noted in the hydrofluorinator. With the 2% sensitiv- ity factor for the radiographic technique taken into consideratidn, a first approximation for the maximum pit depths was 10 mils. Initial plans to wax replicate portions of the fluorinator for quan- titative study of the pitting attack were abandoned to avoid the pessibility of metal ignition by reaction of traces of dental wax with fluorine. 3.1;3 Ultrasonic Thickness Measurements - Wall thinning on the VPP fluorinator was determined by Vidigage meas- ‘urements taken every contour surface inch in each quadrant, and the results were compared fiith base~line data. Figure 12 shows the bulk metal losses found. Quadrant geometry did not appear to be significant in analyzing the data; and so all data are presented at each elevation and the loss ranges outlined, | Maximum fluorinator wall thinning of 27 mils occurred in the middle- -neck region which Jjoins the top and bottom 16-in. right cylinders. This ‘area was exposed essentially to process vapors. Average bulk metal losses for the salt, salt-vapor interface, and vapor regions were 5.5, 7, and 8 mils, respéctively, while maximum wall-thickness losses for these regions were two to three times the respective averages. 3.1.4 Nature of Fluorination Corrosion Metallographic and dimensicnal changes on L nickel were folléwed1dur- ing the fluorination process runs by means of control specimens located in the salt and interface regions and in the lower part of the Vapor region of the VPP fluorinator. This work will be reported in detail at a later date. However, to complete this current evaluation of fluorination corro- sibn, a brief summary is included here. The contrcl specimens, l/4-in.-diam rcds 36 in. long, were held in place by use of gas-tight metal connectors attached to short lengths of nickel pipe, which were welded to the bottom head of the fluorinator. At - 28 - UNCLASSIFIED ORNL-LR-DWG 61590 —— ] @ 2 E Qo - & &8 o - T TP | S & V-100 CYLINDER > g ~ MATERIAL: 35-inTK L NICKEL PLATE 0.9 mils/hr 2 (F2 TIME) 7 80in. BOTTOM HEAD BOTTOM CYLINDER -~—— CONTOUR SURFACE INCHES FROM TOP APOR SALT INTERFACE REGION b SALT REGION AVERAGE LOSS-5.5mils 19in AVERAGE LOSS- T mils ] -1 'SUMMARY QF PRQCESS CONDITIONS TEMP (°C) 350-520, VAPOR REGION 500-570, SALT REGION TIME (hr) 690-SALT 30.9-F, 293/4 in, BOT TOM CONE SALT COMPOSITION NoF-LiF-ZrFs (27/37—-17/31—-34/52 mole %) PLUS 0.2/0.4 wt% U [ f 1 BOTTOM 0 5 10 15 20 25 PLATE WALL THICKNESS LOSS (mils) Fig. 12. Wall-Thickness Losses on VPP Mark III L Nickel Fluorinator. - A - 29 . convenient intervals, dimensional and metallographic changes were noted on the rods, which were then replaced by new specimens. Summation of the dimensional losses on those portions of the corro- sion rods exposed to the salt and interface regions compared favorably with losses in the same regions on the fluorinator wall found by ultra- sonic thickness measurements. However, no such correlation was noted for the lower vapor region. Metallographic examination of salt-region samples from the L nickel specimens disclosed spotty loss of grain boundary material and grain boundary widening. Frequently, a few grains had sloughed off due to this attack. A gray, semicontinuous film 0.1 to 0,5 mil thick was on the surface of the samples, and the intergranular attack proceeded below this film to a maximum depth of 1.5 fiils. Similar results were noted for the salt-vapdr interface samples but to a much more moderate degree. Aside from surface roughening and a thick continuous scale, no unusual effects were noted for the vapor-region samples from the specimens. The vapor- induced scale was considerably different in character from the films noted on the salt and interface samples. 4, DISCUSSION AND CONCLUSIONS Corrosion rates on the VPP INOR-8 hydrofluorinator after 14 runs appear to be consistent with previous ORNL developmental work on the dissolution process, with the exception that the region of meximum gttack has shifted from the salt-vapor interface tc the vapor region. Reasons for this shift are not clear, particularly since the vapor-region tempera- tures have been somewhat lower than those in the salt-containing region. One possible reason for the losses in the lower-interface and salt-vapor regions may be the protective nature of the higher concentration molybde- num phase, compared with the base metal, produced at the INOR-8 surface- hydrofluorination interaction region. This higher molybdenum phase 1is believed to have formed as the result of previous process corrosion. Another possibility — at least for the high attack in a localized vapor region — is that barren salt impinged, during transfér, on the upper walls ——y - 30 - of the hydrofluorinator. Rough calculations indicate-that the barren-salt hydrostatic head, plus nitrogen overpressure, is sufficient to cause salt stream impingement on the opposite wall of the hydrofluorinator only a few inches below the elevation of the barren-salt inlet, Based on the 125-mil design corrosion allowance for the hydrofluori- nator, an average of 18 hr per dissolution run, and an admittedly question- able linear extrapolation of the 0.06 mil/hr currefit corrosion rate, the useful life of the VPP Mark I hydrofluorinator seems to be 116 runs. Current corrosion rates for the VPP Mark ITIT L nickel fluorinator and the nature of attack on control specimens in the current vessel compare . closely with corrosion experienced by the Mark I and II vesselst? used in the pilot plant. Also, the geometric location and the maximum losses for the latest vessel, that is, the vapor region, match the Mark T major loss area even though the vessels have different configurations. Based on a corrosion rate of 1 mil/hr based on fluorine time and assuming a corrosion allowance of 175 mils, a fluorine exposure time of 1.5 hr per run, and linear extrapolation, the VPP Mark ITIT fluorinator - should have a useful life of 117 runs to match the predicted life of the hydfofluorinator._ This calculation does not include intergranular or other forms of selective attack. ACKNOWLEDGMENTS The author is indebted to Volatility Pilot Plant and other ORNL per- sonnel for their continued cooperation on this project. Particular thanks “ are due to E. C. Moncrief for providing information on pilot planf run conditions; Inspection Engineering Division for their careful radiography and ultrasonic thickness measurements on the process vessels; G. I. Cathers for assistance in clarifying process chemistry; E. E. Hoffman, R. P. Milford D. A. Douglas, Jr., and J. H. DeVan for their critical review of this report; Graphic Arts personnel; and Metallurgy Reportis Office for typing of the final manuscript. 195, P. Litman and A. E. Goldman, Corrosion Associated with Fluori- nation in the Oak Ridge National Laboratory Fluoride Volatility Process, ORNL~-2832 (June 5, 1961). . » ] 32. 33. 34, 35. 36. 37. 38. 39. 40, 41, 42. 43, by, 45, 46, 47, 48. 49, 50. 51. 52. 53. 54, 55. 56, 57. 58, 59. 60. 61. 62. 63. 64, 6569, 70, DISTRIBUTION Biclogy Library 71. Central Research Library 72, Reactor Division Library 73. OBRNL — Y-12 Technical Library 74. Document Reference Section 75. Laboratory Records Department 76, Laboratory Records, ORNL, RC 77. G. M., Adamson, Jr. 78. E, J. Barber 79. M. R. Bennett 80. H. A. Bernhardt (K-25) 8l. D. S. Billington 82. R. E. Blanco g3. F. F. Blankenship 84—86, A. L. Boch 87. E. G. Bohlmann 88, B. S. Borie 89. J. C. Bresee 90. R. B. Briggs 91. F. N. Browder 92. K. B. Brown 93. W. A. Bush %%, . D. 0. Campbell 95. W. H. Carr, Jr. 96. G. I. Cathers 97. R. E. Clausing 98, W. H. Cook 99. J. A. Cox 100. R. 5. Crouse 101, H. J. Culbert (K-25) 102. F. L. Culler, Jr. 103. J. E. Cunningham - 104, J. H. DeVan 105, D. A. Douglas, dJr, 106. D. E. Ferguson 107. J. H Frye, Jr. 108. J. H. Gillette 109, H. E. Goeller 110. A. T. Gresky 111. W. R. Grimes 112. C. E. Guthrie 113. C. F. Hale (K-25) 114, J. P. Hammond 115, W. 0. Harms 116, C. S. Harrill 117. M. R. Hill 118, E. E. Hoffman 119. - 31 - PEEPONEESDD PR .L|CDOWZHHMF‘EQSQUWW?QWHM?}W%F‘OL|CD‘-E-."-'IHJ>L|’.:U ORNL~3253 Hollaender UC}L—'HU)"UUJEE!L—'.EIbthZQL—'UJEZ Mann MO NOQOQOQY S g H i Horton Householder Jolley Jordan (Y-12) Kegley Kelley Lampton Lane Larson eberman Lewis Lindauer Litman Livingston Long Lotts MacPherson Manly Matherne McHargue McNuse Miles Milforad Miller Miller Moncrief Morgan Murray (K-25) Nelson ‘ Nicol Patriarca Phillips meQOigthE_bdztlj Pierce Pitt Ruch savage Schaffer Seagren Shank apiro (K-25) Skinner Smiley (K-25) Smith Stainker Swartout 120. 121. 122. 123. 124, 125. 126. - 32 - Taboada 127. C. E. Winters . H. Taylor 128, E. L. Youngblood . Ullman ' 129. A. A. Burr (consultant) . Weinberg 130. J. R. Johnson- (consultant) . Whatley 131. C. S. Smith (consultant) . Whitmarsh 132. R. Smoluchowski (consultant) H C. Wilson 133. H. A. Wilhelm (consultant) SORPLEE B EE = EXTERNAL DISTRIBUTION 134. E. L. Anderson, AEC, Washington 135, D. E. Baker, GE Hanford 136. J. E. Bigelow, AEC, OAD 137-138. David F. Cope, ORO 139. P. R. Dowling, AEC, Washington 140, 0. E. Dwyer, BNL . 141. Ersel Evans, GE Hanford 142, F. W. Fink, BMI 143, J. L. Gregg, Cornell University 144. 8. Lawroski, ANL 145. P. D. Miller, BMI 146. W. Seefeldt, ANL 147. J. Simmons, AEC, Washington 148. E. E. Stansbury, University of Tennessee 149, M. J. Steindler, ANL 150. R. K. Steunenberg, ANL 151. Donald X, Stevens, AEC, Washington 152. J. Vanderryn, AEC, CRO 153. R. C. Vogel, ANL : 154—729, Given distribution as shown in TID-4500 (16th ed.) under Metals, Ceramics, and Materials Category (75 copies - OTS) fr