— L ra Contract No. W-7405feng-26 : METALS AND CERAMICS DIVISION A STUDY OF LEAD AND LEAD-SALT CORROSION IN - THERMAL-CONVECTION LOOPS ' @, M, Tolson and A. Taboada , 7 - APRIL 1966 Fa ~OAK RIDGE NATIONAL LABORATORY ~ "0k Ridge, Tennessee | .. - operated by ON. CARBIDE CORPORATION Sowa s for the o U.S, ATOMIC ENERGY COMMISSION \ - CFSTI PRICES ORNL-TM-1437 A, it 4 ’ 'ufi%f b et fi‘&: ’ - -\*‘«—* &ii » A STUDY OF LEAD AND LEAD SALT CORROSION IN THERMAL - CONVECTION LOOPS G.vM. Tolson and A, Taboada ABSTRACT Thermal-convection loop tests of several structural ~alloys were operated using circulating molten lead. Screening tests were run to evaluate Croloy 2 1/4 Cr, ~ carbon steel, AISI-type 410 stainless steel, and No—1% Zr at conditions described in Teble l Two of the "~ test loops contained surge tanks in which fluoride salts, Nb—1% Zr alloy, and graphite were placed in contact with the lead to determine the compatibility of these materials in a direct-cooled lead system. A1l of the steel loops tended to plug in the cold regions because of formation of dendritic crystals of - " iron and chromium. . The hot-leg attack consisted of general surface removal with a few large pits extending to a greater depth. The Nb—1% Zr alloy showed no measurable attack; however, niobium -crystals were found | in the cold leg of a loop which operated 5000 hr at o 1400°F with a AT of 400°F. | INTRODUCTTON Liquid 1ead has béen’proposed as'a coolant-for molten-salt'breeder ? reactors. In one reactor des1gn, lead is in direct contact with salts | _at temperatures up to llOO°F thus eliminating a heat exchanger and | resulting in superlor heat transfer and thermal efficiencies. In another Ndesign, the lead extracts heat from the salt in. a salt-lead heat exchanger. - The materials commonly used in‘reactors at these temperatures, such as’ ’,,300 series stainless steels, Inconels, or Hastelloys, cannot be used in contact with lead because of the high solubility levels of nlckel in 'lead that result in excessive mass transfer.‘ The refractory'metals “1offer very good corros1on resistance but are difficult to fabricate and 7 _are too expensive for a complete reactor system. Carbon steels offer B good ‘corrosion resistance to- liquid lead but are very marginal with _ respect to both strength and oxidation resistance, Since Croloys (steel materials. o . : o 'solution attack but do not adequately'measure rass transfer. In addition, ¥ ‘although the attack rate for Inconel at 1100°F was reported as U‘v with 1 1/4 to 9% Cr and 1/2 to 1% Mo) and 400 series stainless steels have good oxidation resistance, adequate strength, and contain no nickel, these appeared to be the most promising material for this‘appiication. Consequently, the primary effort was expended in evaluating these LITERATURE SURVEY A great portion of the liquid lead corrosion tests'deseribed in the literature involves capsule tests that demonstrate the occurrence of {“ no standard test procedure was used from one test to the hext., Test specimens were contained in a variety of capsule materlals such as graphite, quartz, and the specimen material itself. Inhibitors were frequently used, and a variety of methods were used to prevent oxidation.. The maximum test time was approximately 500 hr. Results in these isothermal tests generally indicated that iron, carbon steels, low-chromium steels, and chromium stainless steels (nickel-free) had good resistance to corrosion by lead at temperatures up to 1400°F. Austenitic stainless steels were geod to 1000°F. Tan- talum, niobium, and molybdenum were not attacked by lead at 1800°F, Nickel and nickel-base alloys had poor resistance to attack by lead, 0.38 mils/year.l Other nonferrous alloys of Cu, Pt, Au, W, Sn, Zn, Mn, | o Zr, and Ti behaved poorly or were not recofimended Carbon demonstrated ‘_ _ _ :r‘ poor resistance to attack by lead at 1800°F but good resistance at 935°F., ‘ In addition to the above, other experimentation has been done - 3 using thermal-convectlon loops to study mass transfer. As in the static work, no standard test design existed, making comparlson'of data diffi- cult, The effect of impurities in lead are not well understood.’ Oxygen 'fi - has been reported to increase corrosion in lead. Some research indicates 1L. R, Kelman, W, D. Wilkinson, and F. I. Yaggee, Resistance of (:)I Materials to Attack by Liquid Metals, ANL-4417 (July 1950). ‘ fy » { -l Y P . h\»‘?é\ N 4},".319—23 (March 1960). ? l\ that many elements may'act_as corrosion inhibitors. These.inhibitors could act to form a protective, film in the case of titanium and zir- conium (ref. 2), to decrease the solubility of the container material in lead as may be the case of nickel, or to remove oxygen from the system in- the case of magnesium. In one of the 1nvest1gations the relative resistance to mass transfer in liquid lead of 24 metals and alloys was measured3 at 1472°F ‘maximm loop temperature and 300°F AT, The tests were run in quartz thermal—cdnvection loops with the alloy being_studled formed-lnto tubes ‘which were inserted into the hot leg and cold lég. In these tests niobium and molybdenum exhlbited no mass transfer after 500 hr of test. Nickel—base alloys and austenitic stainless steels were hlghly susceptible to mass transfer and plugged the loops within 100 hr. Intergranular attack was noted in the hot region.- The pure metals, Fe, Cr, Co, Ti, and Ni, all plugged W1thin 100 hr The .400. serles stainless steel (chromlum) and molybdenum-bearlng alloys showed llttle evidence of mass transfer after approximately 500 hr. There was some evidence of prefer-‘ ential leaching of chromium by'lead.in,thls type alloy, - Several inves- tigators have studiedfithe reslstanoe-of ste61s to corrosion by lead, / bismuth, andxleadAbismutn.alloys-were found to be more corrosive than ' lead,.%he literatureTindicafies-that they all result in the same general behavior.patterns,_"The corrosion of steels in uninhibited lead was reported* to be about 1/40 of that noted in uninhibited bismuth under, comparable conditions (1472°F maximum temperature with 212°F AT). The most sophisticated work with 1ead and lead-bismuth alloys was done by -'e_BNL in conjunetion with the LMFR Program.5 Although only one lead 1oop N ~ e ;=; ST 20, F. Kammerer et al., Trans. ATME 212 20—25 (1958) 3J. V. Cathcart and W. D. Manly3 Corrosion 12 43~48 (February 1956) 43, A, James and J. Troutman, J. Iron Steel I Inst. CLondon) 1%, 5A. J. Romano, C. J Klamut and D. H, Gurinsky, The Investigation - of Container Materials for Bi and Pb Alloys. Part I; Thermal Convection - Loops, BNL-811 (T-313) (July 1963). 4 c - | - was operated; many bismuth and lead-bismuth.loop tests were conducted; iIn'general, the low-alloy steels and low-chromium steels exhibited the best corrosion resistance to bismuth and lead-bismuth; however, titanium and zirconium additions were required to obtain goodlcorrosion resistance in loops that operated above 752°F. The authors concluded that titanium and zirconium inhibited corrosion by forming\ZrN, TiN, and TiC on the walls in the hottest regidns of the loops. No corrbsion occurred in inhibited loops that operated for 10,000 hr at temperatures as high as 1022°F in bismuth and temperatures as high as 1202°F in lead-bismuth eutectic, A single Croloy 2 1/4 steel loop containing‘zifconium—ifihibiteq | L lead was operated for over 27,000 hr with a i022°F hotlleg and approxi- + mately 212°F AT, Although the zirconium additions had been made, the zirconium was not detected by chemical analysis of the solution. . The lead also contained about 250 ppm of magnesium, which had been added to prevent loss of zirconium by oxidation, and it may have héd an inhibiting effect. DESCRIPTION OF TESTS Six uninhibited thermal-convection loop tests were performed at ORNL to explore the compatibility of materials with lead under cofiditions expected in molten salt reactors. ‘One additional loop test was also verformed to evaluate the performance of a newly designed thermal- - convection loop. A description of the loop and the results are presented in the Appendix. The operating conditions for all the loops are summarized - in Teble 1. | Two of the loops (type 410 stainless steel and 2-1/4 Cr steel) had Nb—1% Zr alloy liners in the surge tanks. Molten Salt Reactor Experiment fuel salt® floated on the lead surface in the surge tanks as shown in Fig. 1. These loops also contained a graphite specimen in the surge tank which was exposed to both the salt and the lead at the lead-salt inter- face. The maximum temperature in the loops was about 1200°F and the AT was 300°F, ; | o 67,9% UF,, 7.6% ThF,, 29.2% F1Fs, 37.1% LiF, and 18.2% ZrF,. _- A 'PROBE LINE TO DETERMINE » . LIQUID LEVEL - - B W - el 8 — STOP TO HOLD LINER \ | A - ] | "IN PLACE ) | i Ryl e N Loy LéAD SALT INTERFACE| ~ o - . Fig. 1l.. Section Through the Surge: Tank ‘Used on the 410 Loop. - Notice the salt is floating on the lead and the only metal which : ; B j contacts the salt is the Nb—-1% Zr liner. The graphite specimens were - o .. suspended from the small wire extending 1nto the salt and cannot be - . seen in the picture, - T - S - ; Table 1. Operating Conditions of Thermal-Convection ' Loop Tests Using Lead as a Coolant ' Loop Maximum Material : Temperature AT - Operated - | (°F) -~ (°F) - (hr) Croloy 2 1/4 1210 300 266 AISI-type 410 11210 300 1346 stainless steel o ‘ _ - - Croloy 2 1/4 f 1100 200 5156 ASTM type A-106 1100 200 5064 No-1% Zr clad with 1400 400 385~ type 446 stainless ' ' steel ‘ ‘ | Nb—1% Zr clad with 1400 - 400 5280 type 446 stainless o steel o aLoop containing salt shut down due to instrument malfunction. o _ Two other loops which did not contain salt or graphite operated at 1100°F with a 200°F AT, One of these loops was constructed of 2 1/4 Cr ~ steel, the other of low-carbon steel. A third set of,leop-tests was operated to investigate the compatibility of nidfiium with lead both with ‘and without MSRE salt. These loops were fabricated from niobium clad with type 446 stainless steel. o A cleaning charge of lead was used in all the iron-base loops. The charge was run isothermally at the maximm operating temperature of the loop overnight (ebout 12 hr) and then dumped. The lOops'were then refilled with clean lead and pufi into operation. This was not done with the Nb—1% Zr loop because of the difficulty in attaching a drain line. o loop Showed about 1 mil of3attack, as can be seen in Fig. 3. sion was observed on the Nb—l% Zr liner, as demonstrated by the photo- REsULTs OF TESTS _ Of the two steel loops that contained salt the 2 1/4 Cr steel ?plugged after only 288 hr and the type 410 stainless steel plugged after 1346 hr, ‘As shown in Fig. 2, the plugs were made up of dendritic crystals, which were determined to be iron and chromium by x-ray diffrac- tion and wet chemical analysis. The maximum-depth of attack on the type - 410 stainless steel piping in the hot leg was 2 mils, as determined by The pitting in the hot leg of the 2 1/4 Cr No corro- metallographic examinetion. micrograph shown in Flg. 4. SR Y.58172 Dendritic crystals of Iron and Chromium Which Formed in Fig.' 2. 200x%, Ethe Cold Leg of a 2 1/4 Cr Lead Loop After 266 hr, Unetched Y.58161 Fig. 3. Corroded Section of Hot Leg From 2 1/4 Cr Lead Loop Operated at 1200°F for 266 hr. Unetched. 500x. . Fig. 4. Nb=1% Zr Liner Exposed to Lead and Salt for 1349 hr at 1200°F in 2 1/4 Cr Lead Loop. 200x. Je rate since plugging time is not a function of corrosion rate. After 648 hr of operation, the cold leg of the 2 1/4 Cr Loop, which contained no salt or graphite, began to plug as indicated by In order %o’ determine 1f the hot leg was being selectively attacked the lead was dumped from the loop and decrea51ng cold-leg temperature. the hot leg was radlographed Several areas were noted where the lead had wet the metal and had not drained from the loop, thus indicating ' selective dttack,” The lead that was dumped from the. loop was examined and found to contain crystals of iron and chromium which had‘floated to the top surface during cboling; The loop was then refilled with new lead, restarted, and operated for 5156 hr. tion without plugging does not necessarily mean'a decrease in corrosion ' Post- test metallographic examination of the cold-leg region showed the The increased time of opera- presence of a large amount of dendritic crystals, as revealed in Fig. 5. P - 7Brookhaven previonsly had found that areas of selective attack could be identified in mercury systems in this manner because of the increased wetting action of the liquid metal. Y-62845 - .008 . Figo 5. - Leg of the 2 1/4 ‘Cr Lead Loop Which Ran for 5156 hr at llOO°F with a AT of 200°F , — Dendritlc Crystals of Iron and Chromlum Found in the Cold - 10 The approximate depth of hot-leg attack as determined by the change in wall thickness averaged 5 mils and could have been as deep as 8 mils, The uncertainty in the depth of attack is due to the pitting nature of the attack and the variation in the wall thickness, Analysis of the crystals ih the cold leg showed about the same Fe/Cr/Mo ratio as that existing in the original loop piping alloy. A black film found floating on the lead in the top of the surge tank was identified as MnO by x-ray diffraction and spectrographic analysis, The manganese was probably preferentially leached from the hot leg and then deposited in the surge tank as it scavenged oxygen from the rest of the system materials. The carbon steel loop, constructed of large-diameter pipe, operated for 5000 hr before shutdown. Prior to shutdown, the loop was begimning to show some signs of restricted flow. ‘Subsequent measurements of the wall thickness of the loop piping showed a meximum of 10 and an‘average of 7 mils of attack similar to that found in the 2 1/4 loops (see Fig. 6). V62846 §| 0.045 INCHES N 100X | Y Fig. 6. Section Through Hot Leg of 2 1/4 Cr Lead Loop Which Operated for 5156 hr with a Hot-Leg Temperature of 1100°F and a AT of ' 200°Fo The attack consisted of uniform surface removal. P8 c;;iz » . | - 7 o o 2 o , | I | . | 1ii3t* . . The 1oops constructed of wal% Zr clad with type 446.stainless steel were shut down after only 385 hr of operatlon due to a faulty 'relay in the controlrsystem; The loop that contained salt could not be restarted prdbably due to separation of a hlgh-meltlng constltuent of the salt upon coollng. The 1oop ‘that contained only lead was restarted . | and operated 5280 hr, Posttest metallographic examination showed no signiflcant hot-leg attack, as indlcated by the photograph shownt in Fig. 7. .Some mass transfer crystals were found in this loop, as shown in Fig} 8. The crystals were found to be 90 to- 100% Nb by electron probe as shown in Fig. 9, Nicbium mass transfer has not been reported by any other investlgators., En g, Hallerman and R.'S;'Crouse; personai communication, Nov. 9, 1965, - . ’ . . - . B Y.66685 o I - Fig. 7. Section Through Hot Leg of Nb—l% Zr LOOp Which Operated f "iJ - Over 5000 hr with a Cold-Leg Tbmperature of l400°F and a AT of 400°F. f Unetched. 750x. - , . ay - 12 Fig. 8. Nicbium Crystals Found in Cold Leg of Nb-1% Zr Loop Which Operated Over 5000 hr at 1400°F and a AT of 400°F. Unetched. 250x%. Y-67208 Fig. 9. Optical and Niobium Ly X-Ray Image Taken of the Niobium Crystels Formed in the Cold Leg of a Niobium Thermal Convection Loop Which Operated 500 hr at 1400°F and a AT of 400°F. 250x. (a) Light optics. (b) Niobium Lg X-Ray Image. | ‘ - . _‘ (“h € o> w! 13 - © CONCLUSIONS AND RIECOMMENDATTONS The results of the tests are given in feble 2 along with a Brookhaven loop to use as a comparison. The poor performance of the | ferritic materials described here: suggests that these materials are unsuitable to contain,lead{at the conditions_investigated. This is 'surpriSing-in light of the 27,795-hr Brookhaven loop;_ The 78°F 4if- ference in hot-leg'temperature between our 5156
N - / CLAMSHELL HEATER | 13 in. ‘Iq-ln. CONNAX FITTING 16 34 in. OD X.0.089 In. WALL THICKNESS DUMP TANK—] K THERMAL CONVECTION LOOP _ USED FOR LEAD CORROSION STUDIES ORNL—DWG 65-4744R c‘ L. Va-In. BUTT WELD SWAGELOCK ~ - ' 3-in. : SCHED-40 PIPE i 4Y,-In. LONG 17in. GRAPHITE Nb-1% Zr LINER SLOTS IN GRAPHITE _ ENSURE LEAKAGE SALT ‘I“JL LEAD ) 3 Nb ‘§/|_ozrs l!.':lNER : : ‘» - ; | ENSURE LEAKAGE LOOP PIPING o SECTION THROUGH SURGE TANK OF THERMAL CONVECTION LOOP USED FOR LEAD-SALT STUDIES Fig. 10. Sketch of Prototype .Loop That Will Be Used to Study the Effect of Inhibitors on Lead Salt—Low Alloy Steel Systems. - S 1 In the new loop, better control over the cold-leg temperature is achieved through the use of a cold-leg heater placed at the inlet to the cold leg of the 1oop. Thls heater is controlled by the cold- leg thermo- _couple, A shutdown device is also incorporated which terminates the 1oop when the temperature at the cold-leg heater equals that of the hot 1eg.,_With this design,,the_degree of plugging may be determined by the pouer conéumption of the cold-leg~heater. -The system worked very'well in the prototype test and will be used on future loops. - o~ o 1';—3. 4o 5-6. 7-16. 17, 18, 19. 20, 21, 22. 23, - 2425, 26. 27, 28, 29, 30. 31, 32, 33. i 33, . 7273, - . 75, | 76, . 78-92, INTERNAL DISTRIBUTION Central Research Library 36. W, Reactor Division Library : 37. W. ORNL — Y-12 Technical Library 38, P. " Document Reference Section 3941, M, Laboratory Records Department 42. H, Laboratory Retords, ORNL R.C. 43, P, ORNL: Patent Office . 44. R, G. M. Ademson, Jr, S 45, H. S. E. Beall ‘ ‘ 46. R. E. S. Bettis | 47. W, F. F. Blankenship 48, H, G. E, Boyd S ‘ 49, C. R, B. Briggs 50. R. J. H., Coobs | 51, P. W. H. Cook ; - 52,~ D, R. S. Crouse C , 53, G, J. E. Cunninghem ) 54. 1. J. H, DeVan . _ 55. A. J. R. DiStefano 56, J. D. A. Douglas, Jr. ‘ ~ 57. R. B. Fleischer | 58-67. G. J. H Frye, Jr. " .68, A, R. J. Gray ] 69. J. | ' 70. M, EXTERNAL DISTRIBUTION C, M, Adams, Jr. y Massachusetts Institu‘be of D. .F. Cope, AEC, Oak Ridge Operations Office J. L. Gregg, Bard Hall, Cornell University J. Simmons, AEC, Washington -~ D. K, Stevens, AEC, Washington Research and Development Division, AEC, ORO ‘Division of Technical Information Extension ORNL-TM-1437 R. Grimes 0. Harms N. Haubenreich R, Hill Inouye R. Kasten B. Lindauer G. MacPherson W. McClung B. McDonald F. McDuffie J. McHargue L., Moore Patriarca " Scott M. Slaughter Spiewak Taboada . Tallackson . Thomas . Tolson Weinberg - Weir . Whatley = = & Technology