S ORNL-TM-3352 T ENGINEERING DEVELOPMENT STUDIES " FOR MOLTEN-SALT BREEDER REACTOR PROCESSING NO. 10 L. E.,Mch}ees!e pTIONA s gl AL LABORATORF ey This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Atomic Energy Commission, nor any of their emplayees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights, ' ) salt the induction coil, 'and'theIfluorihator'vessel.a These equatlons predlct a high effi- 01ency for. heatlng molten salt in a fluorlnator for one of the four c011 de51gns tested. SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS - IN A MILD—STEEL FACILITY ’ The new salt feed-and—catch tank was installed 1n the system, and iron oxide was partlally removed from it ‘and from the system by contacting the equlpment with hydrogen.u Reductant was added to the blsmuth about 20 liters of salt (72-16 12 mole % LiF-BeF —ThF ) was charged to the sys- ;-tem to replace salt that had been dlscarded w1th the orlglnal salt feed- ~ and-catch tank, and both phases were circulated through the system in h;order to . complete the removal of ox1des. After the addltlon of a small _'_1amount of 21rcon1um to restore the system 1nventory to about 15 g,rboth _di’jfphases were treated with a 307 HF-—H ‘stream in order to remove oxides in efrom,the salt._;_j(b;_;_,-' ' 2 About 1 g—equlv of reductant, 1n the form of thorlum and Li-Bi "alloy, was added to the treatment vessel for the purpose of establlshlng | & 21rconium dlstrlbutlon coefflolent value in the range of l to 5. After . the blsmuth -and salt had been transferred to thelr respectlve feed tanks, 97 ‘ Zr tracer was added to the salt Mass transfer experiment ZTR 1 was . vi then carrled out using blsmuth and salt flow rates of 216 ml/min and 99 ml/mln, respectlvely Countlng of the flowing-stream samples showed 97Zr tracer had occurred. The reductant that that no transfer of the had been added to eétabllsh the proPer,D _oxidlzed by HF and FeF vessel. 7 value had apparently been - - 5 whlch were present 1n the salt in the treatment DEVELOPMENT OF THE METAL TRANSFER PROCESS: INSPECTION OF EXPERIMENT MTE-2 The equipment used for metal transfer process experimentrMTE—z, com- pleted previously, was disassembled and inspected. Some blistering and cracking of the 20-mil-thick nickel aluminide coating had oecurred‘duringh -the 2370-hr period that the vessel had been held at_650°C.'_Thevessel__ - was sectioned in a manner.euch that the salt and bismuth nhaeeS'could be observed. The salt and bismuth phases appeared to be clean, and the ‘Interfaces were free of contamination. A black material which ‘covered the vessel wailin the fluoride compartment is thouéht_te be a mixture of salt and finely divided bismuth, Deposits containing unusually high concentrations of rare earths were observed on the lip -and overflow spout of the Li Bi alloy contalner | and in the bottom layer of the Th-Bi solutlon._ The total quantltles of rare earths in these dep051ts were only 5 to 10% of the rare-earth 1n—' ventories in the system.- The inside of the carbon steel vessel appeared to be in gobd eondi— tion except for corr051on ‘on items that vere constructed of thin carbon _ steel. The carbon steel pump, which used bismuth check valves, was in ‘good condition at the completion of the experiments (i.e., after being operated to circulate a total of 702 liters of Llcl) " DEVELOPMENT OF THE METAL TRANSFER PROCESS: AGITATOR_TESTS FOR EXPERIMENT. MTE-3 -Equipnent was-censtrueted in order to test the shaft seal design- that is proposed for use in metal transfer experiment MTE-3. The system ~will also allow us to measure the extent to whieh'bismuth'is entrained . &) < ,“' «J) oy vii in salt in a mechanically agitated system, and to evaluate a vapor-. dep031ted tungsten coating as a means for protecting carbon steel from ,corr051on by molten salt and bismuth Several shaft seals were tested for varlous time periods at aglta— tor speeds ranging from 150 to 750 rpm. It was found that the seal life . could be 1ncreased_and-that the 'seal leakage rate was decreased con51d— erably by lubricating'the'seal with mineral oil., A seal des1gn which . allowed satisfactory operation for T1 days was Judged to be acceptable for use 1n experlment MTE—3 Unfiltered salt samples taken during a ThS-hr period in which salt and blsmuth were mechanically ag1tated showed that the concentration of blsmuth in the salt 1ncreased from 8 ppm to 200 ppm as the agitator —speed was 1ncreased from 150 rpm to 750 rpm, During this perlod the ;concentration of nickel in the bismuth 1ncreased from 20 ppm to 1000 wppm which indicated that the bismuth had penetrated the tungsten coating on the interior of the. test vessel. Inspection of the coating after - completion of the test revealed cracks 1n the coating 1n a number of 7"places; Also, 1t appeared that the coating had not been applied over ~the entire surface of the drain 11ne at the base of the test vessel It . was concluded that protection of a vessel from attack by bismuth via a tungsten coating would be’ dlfficult because of the tendency for such a - coating to crack ' DISTRIBUTION OF RADIUM BETWEEN L101 AND L1~B1 SOLUTIONS Data on the distribution of radium.between molten LiCl and Li-Bi solutions contalning 13 to 35 mole % lithium were obtalned prev1ous1y 'Tm»during metal transfer experiment MTE—2 Addltlonal data were obtalned Iby dilutlng a portion of the Li-Bi solution from the experiment (con- .e.taining radium) and equillbrating the resulting solutlon w1th 1icl at s650° All of the radium distribution data could be. correlated well in ;the manner used preV1ously for correlating distributlon data for a large | number of elements 1f radium was assumed to be divalent 1n the LlCl phase.’ It was found that the distribution characterlstics for radium viii are'quite similar to those of the divalent-rare-eafthiand-alkaline-- earth fission products (Sm, Eu “Sr, and Ba), in fact the data for radium and barlum are almost identical. ' DEVELOPMENT OF MECHANICALLY AGITATED SALT-METAL CONTACTORS Efforts involving the»development:of mechanically agitated salt- | metal eontactors'oflthe Lewis cell type‘wefe eontinned.a Préliminany “tests were carried out in contattdrs'of,several sizes and with differ- V:entdagitator configuratiens in order to determine the factors that will - limit the sgitator speedin'stirred—interface'contactors.'-The agitator speed,was fonnd_to be iimited by fihe.transfer of the 1ow—density (water) phase via entrainment in the cifculating highfdensity-(mercury) phase. ‘The limiting-agitater speed was essentially independent'of theisize and shape of the contactor but was strongly dependent on the agitator diam- eter, A test made with a low-meltlng alloy and water, which resulted 1n & density difference of T.l rather than 12.6, indicated that the limiting agitator speed is not highly dependenfi on_the-difference in densities of the two liquid pheses. It is believed that entrainment of salt in . the bismuth.in metal transfer experlment MIE-3 will occur at essentlally the same agltator speed as was observed w1th the mercury—water system (300 rpm), and that the experiment should be operated 1n1t1ally with agltator speeds well below this value. | Two tests were carried out for determining the rate at which blsmuth circulates between the two sides of a compartmented salt-metal contactor contalnlng a captive blsmuth phase._ These tests, in whleh mercury and 'water were used rather than bismuth and saltg'indieated mércurj»flGW' rates of 11.2 and 19. 3liters/minfbr agitaters (speed 195-rpn in each, case) hav1ng straight and canted blades respectlvely. it was'coneluded | that. the bismuth circulation rate in metal transfer experlment MTE—3 will be adequaterand much greater then the mlnlmum_de51redsvalue of 0.5 ~liter/min. o« .‘r'l ' b ) ix ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN 'MOLTEN SALTS AND LIQUID BISMUTH DURING COUNTER- CURRENT FLOW - IN PACKED ‘COLUMNS ' The transfer. of materlalsrbetween a molten salt”and liquid bismuth _results in a conditxon where the fluxes of the transferrlng 1ons are dependent on both concentratlon gradlents and electrlcal potentlal gradl— ents. ThlS greatly compllcates the mass transfer process and makes the design of cont1nuous reductlve extractlon columns dlfflcult. Calculatlons : were completed for both blnary and multlcomponent mass transfer in order to determine the condltlons under which the presence of an electrlc potentlal grad1ent 31gn1f1cantly alters the mass transfer rate, Con- :'dltlons whlch typify molten—salt--bismuth and aqueous—organlc systems were examlned The effect of the electrlc potentlal grad1ent was found to be of ,greater importance when the electrolyte is a molten salt than when it is "an-aqueous solutlon. In the latter case, the nontransferrlng coions redlstrlbute in the electrolyte phase 1n a manner which suppresses the | effect of the electrlc potentlal gradlent., However, it was shown that | 51gn1f1cant errors in calculated mass transfer rate values will result under some operatlng condltlons in both cases. Two cases 1nvolv1ng reductlve extractlon of uranlum from a molten "‘fluor1de salt phase into a 11quid blsmuth phase contalnlng reductant illndlcate that neglect of the effect of an electric potentlal gradlent | probably causes essentlally no error 1n calculated mass ‘transfer rates fI:fOr reductlve extractlon operatlons of 1nterest 1n MSBR proce551ng. ENGINEERING STUDIES OF URANIUM OXIDE PRECIPITATION I'°:; Studles of the chemlstry of protactlnlum and uranlum ox1de preclp- 'f::”pltatlon have 1nd1cated that ox1de prec1p1tatlon may be an attractive .alternatlve process to fluorlnatlon--reductlve extractlon for 1solat1ng o protactlnium and remov1ng uranlum from the fuel salt of an MSBR An _ " expermental facility has been de51gned and equipment is ’oemg installed in order to study the klnetics of uranlum oxide precipitation, to investi- gate the 51ze dlstrlbution and settllng characterlstlcs of oxide precip- '-iitate, and: to gain- experlence Wlth oxide precipltatlon systems. The experimental facility will allow for batch precipitation studies to be made in a b-in.-diam vessel cofitaining'approximately'2 liters of 72—16—12 mole % LiF-Bng-ThFh salt that also containseUFu at an initial ‘concentratlon of about 0.3 mole %. Oxide will be supplied to the pre- cipitator in the form of a water-argon gas mlxture that will be 1ntroduced through a l-in.-diam draft tube to ‘promote contact of the salt and oxide. ~ The salt will be decanted to a receiver vessel after allowing the oxide to settle for a short time, The facility elso includes & system for supplylng hydrogen—HF gas mixtures that will be used for convertlng oxides to fluorldes at the conclusion of an experlment.' The off-gas system 1ncludes eaustlc scrubbers for - ‘removing HF from the preclpltator off-gas | stream in order that addltlonal 1nfbrmatlon relatlve to the extent of prec1p1tat10n of oxldes can be obtained. ' STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS | Salt purlflcatlon studies were continued on the contlnuous reduc— tion of iron fluoride by countercurrent contact of the salt,(72.0-lh.h~ _ 13.6 mole % LiF-Ber-ThFh) with hydrogen‘in & packcd column. Tests carried out to investigate the possibility that iron particleslin the molten salt might be the cause of occa31ona1 high iron analyses showed that iron particles are not present. Sampllng tests w1th varlous sampler designs showed that iron particles, if present, do not remalnsln the salt durlng one pass through the experimental system; that iron analyses bélow 100 ppm sre'uhreliable with sample sizes of 1 g or less; end that salt samples taken in nickel samplers are more subject to iron contamination during removal than samples taken in copper sammlers;!_ccmpariscn of flooding data taken during the countercurrent flow of molten salt and argon indicate . that floodlng occurs at throughput values below those predlcted by the Sherwood correlatlon. &‘j; <) B wl ) 1, - INTRODUCTION . A molten salt breeder reactor (MSBR) will be fueled with a molten | fluoride mlxture that will c1rculate through the blanket and core fregions of the reactor and through the primary heat exchangers We are developlng process1ng methods for use in a close-coupled facility for removing f1s51on products, corrosion products, and fiSSile materials from the molten fluoride mixture. Several operations'assoC1ated with MSER processing are under study, | (1) (3) (4) S8 (7). @ (2) The remainlng parts of this report discuss. experiments conducted in a simulated continuous fluo- ‘rinator for studying 1nduction heating in molten salt, experiments conducted in a mild—steel reductive extrace tion facility to increase our. understanding of the rate at which materials are extracted from molten salt '1nto bismuth in a packed column, ‘the results of 1nspection of equlpment used in experl— mént MIE-2 for demonstrating the metal. transfer proc- -ess for the removal of rare earths from MSER fuel car- r1er salt results of agitator tests carried out for evaluating . various shaft seals for use in metal transfer experi— ment MIE-3, - studies on: the d1str1bution of radium ‘between LlCl ‘and Li-Bi solutions,, development of mechanically agitated salt-metal contactors,: -analys1s of multicomponent mass transfer between ' o - molten salts and liquid bismuth during countercurrent o flow in packed columns,_; des1gn of a fac111ty for conducting englneering studles ' _,related to the precipitation of uranium oxide from il molten fluoride mixtures, and {9) oo f.methods.ci'“ studies of. the purification of salt by continuous o ThlS work was carried out in the Chemical Technology Div131on during the period January through March 1971 2. FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON INDUCTION - HEATING IN A CONTINUOUS FLUORINATOR SIMULATION J. R. Hightower, Jr, - An experlment to demonstrate the usefulness or l&yers of frozen salt for protection agalnst corr051on in a continuous fluorlnator requires an internel heat source that is not subject to corrosion by_the molten salt. High—frequency inductionrheeting has been propOsed'for this purpose, and the estlmated perf‘ormancel of a frozen~wall fluorlnator hav1ng an 1nduc-‘ ' tion coil embedded in the frozen salt near the fluorlnator wall has 1nd1— _cated that such a method may be acceptable,- However, there are uncer- ~ tainties associated with the effect of bubbles in thermolten salt and _ with the amount.of heat that will be generated in the metal walls of the fluorinator. Equipment ha.s.b_een,installed=2 for studying. heat generation' in a simulated frozen-wall.flnorinaiortconteining'provisions for induc- tion heating.. In the simulation a 31 wt. % HNO3 solution,'whieh nas electrical properties similar to those of molten salts, is being used as a substitute for molten salt in e fluorinator vessel.-'We have previously _reported_3 results for the first eight experiments with the earlier indnc— tion heating coil design."During fhis report period, experimentel work 'on:induction heating in the fluorinator simulation was completed. Twenty- one additional runs were carried out in order to test three other induc-~ tion coil designs. Relations were derived for predicting the rates of heet generation in the molten salt, the induction coil, andethe pipe}wall in a fluorinator having frozen-well corrosion protection, 2.1 Modification of Power Genération and Transmission Systems The eleCtrical diagram showing the rf power generation end trans- mission. systems for the induction heating experlment is shown in Fig. 1. The. generator, & Thermonlc Model 1400. oseillator, is rated at 25 kW, operates. at a nominal frequency of 400 kHz, and develops a terminal ~_voltage of about 13,000 V (rms). :The terminals of the generator are- - connected to the primery side of an.oil-filled. rf- step-down transformer, this arrangement reduces the voltage to approximately one-s1xth that o) ORNL DWE. 7i-3813 vy . RE AMMETER. : cch|AL . TRANSMISSION : easLe - ‘ _ WORK COIL _ - - REFLECTED. ; - ' RESONAKCE ACID . GENERATOR @ coiL 3 RESISTANCE "RF CAPACITORS STEP-DOWN ' TRANSFORMER T : FINE TUNING LooP COARSE TUNING CoOIL 3 v o REFLECTED RESONANCE ACID coiL RESISTANCE GQUIvALtUT CIRCUIT Fig. 1. Electrical Diagram for Induction Heating Experiments. » - x of the input5inxorder to not exceed the voltage ratings of. the coaxial transmission cables, .The.coaxial,cdblés (Model T~20-D and Model T-10-D made bylL; C. Miller Company) are connected in series and allow trans- mission of power to a point about 30 £t from the generator. 7During'the current report period, the 10~ft segment (Model T-~10-D) of the cable was replaced with a S-ft-long coaxial conductor, This conductor was madéAof 3/8—1n;ediam copper tubing enclosgd'within 1/2-in.~dism poly- ethylene tubing, fihich'wgs, in turn, placed inside 3/h-in.-diam copper tubing. The new section of the transmission line operated satisfac~ torily. | | o | In order to obtain & high current through the induétion‘cbilland to . -minimize.the current in the coaxial{cable,;additional'1nductances and :capacitanCes,were incorporatedintb:the.ci;guit.with'the induction coil -to'fbrm7a parallel RLC resonant cirCuit.é”Two.capaCitors (General Electric, Cat. NQ;_19F23h),-each:having a capacitance df 0.0105 uF, were - used. These capacitors'were.fated_for a8 meximum current of 196 A at 5h0 kHz -and & maximum voifiage of_SSOO V.. For the induction heating,coils tested, it was possible.to'aéhieve‘conditions néar resonance by%Substi- tuting coils of different sizes,for'the,resonance and coarse-tuning . -coils. (see Fig. 1). Resonance was then approaéhed more closely by adjusting the slide-bar on the fine=tuning loop. It was not possible to obtain precisely resonant conditions because the operating frequency of the generator was affected by each tuning adjustment, waever, with theJcircuit adjusted as close to resonance as possible, we were able‘to drive up to 250 A through one induction coil (Coil III) with a current of only 190 A in the coaxial cable. When>the resonant ciréuit was not ‘used and the coaxial cable was connected directly to the inductidfi coil leads, we could drive only abdut'IGQ A through the induction coil. 2.2 Experimental Results - To date, 29'runs.have.been.mgde with the continuéus fluorinator simulation. to detérmine heat ‘generation rates in the nitric acid, in the pipe surrounding the acid, and in the four induction coils. Each induction coil had & length of 5 ft and an inside diameter of 5.6 in., ) o ) « the relatlon ; ‘and-vasrmade'cf a number of smaller coil sections connected in parallel electricelly, ‘The characteristics of the individual induction coil designs are shown in Table 1. Coil iV'was chosen as the best design because it produced the highest ‘heat generatlon rate in the nitric acid for a glven current 1n the 1nduction 0011 " Table 1. Characteristics of Coils Tested in ' Continuous Fluorinator Simulation — ey - Coil " Length of . No. of S ‘No. of . Conductor = Small Turns in Adjacent - Small - Diameter Section Small Sections Coil Material Sections (in.) - (in.) = Section Wound I Monel"|ter’f'_17~‘,’ - :l/h 3 ' 6-1/4 Opposing II Stainless 18 o 3/8_{ 3 6 Assisting . Steel ' R III Stainless 18 . - 3/8 3 | 6 Opposing ‘ - Steel L R ' IV Copper o W _.'ith - 11-3/4 Opposing Table 2 glves the experlmentally determlned heat generatlon rates Nm:.ln the acld COll and 51mu1ated vessel wall along with the run condi- ’_tlons. Coil currents ranging from 100 to 250 A and oscillator frequen- ‘cies: ranglng from 390 to h26 kHz were used in these experiments,. Heat jgenerat1on rates as hlgh as 1559 W were developed 1n the acid, The experlmentally determlned heat generatlon rates were used to -'*racalculate correctlon factors for use w1th design equatlcns that were der1ved for 1nduction colls having 1deallzed geometrles. The design fequations Whlch deflne these correctlon factors, are. llsted below. The 'ff"rate of heat generation in llquid 1n31de an 1nductlon c01l 1s glven by | | faz N2/ \L | ToT k | P‘Q k 0. 3818( X ) (Pz) .. - .o s i G ey e A . ot et g i e o 4 b £ i, o o = g Table 2. Results of Heat:Generation Measureménts - Average - Total , Heat : Acid Coil Oscillator - Generation Rate Run , Temp. Current Freq. Gas (W) ~ No. Coil (°C) (A) (kHz) ~ Holdup Acid Pipe Coil CFs-1 I 24k - - o 0 22 1ho .2 I 246 130 412 0 Log 167 3 I 26.5 150 h12 0 316 279 1hlk2 b I 25.6 100 L2 0 141 116 - | - 28.1 150 W2 0 378 263 . 29,1 150 - - e 0 . 383 309 5 I 26,7 160 12 0 48 308 6 I Lk6.s 150 412 0 393 282 | 7 I. 28.9 150 ; 412 0 377 273 1356 . 26,6 . 120 lhie 0 235 178 870" 8 ' I s51.k% 150 k2 0 ko2 275 1hk2 9 II 21.9 140 ho2 0 201 128 1179 10 ITI 19.6 165 393 0 195 206 1327 - 20.3 189 390 0 227 . .253 1094 - 11 II 17.k 200 . hoo 0 207. - 304 19k6 12 II 19.3 200 Yoo -0 370 . 313 - 2335 13 II 18.5 160 h22,5 0 > 250 - .201 1253 | 160 4o2,5 0 . | : 1282 - 21.3 200 ho2.5 0 > 66 338 2061 200 ho2,5 0 . . 2083 200 ho2,5 0 -~ - 19k6 1k II 1k.9 150 k02,3 0" 215 166 1150 15 III 15.7 149 402.3 0 370 220 - o 16 III 18,5 150 ho2,8 0 360 318 137k 17 III 18.5 150 Lo2.8 0.130 365 254 1ko8 18 11T 19,3 150 403.8 0.171 325 235 1202 19 III 19.0 150 402,8 0.180 304 188 1099 20 III 19.0 - 1ho ho2.8 0 S ¥~ 186 1030 21 III 20.0 - 1h9 ko2,8 0 382 20k 1195 22- III 22,2 250 L2k, 8 0 1134 626 3606 23 III 20.9 - 250 426.1 0.17 976 688 3366 24 III 24,9 249 - 425,8 0.107 1159 599 - 3457 25 IIT 25.3 252 426.0 0,106 1292 976 3617 26 III 24.8 @ 2l h25.9 0 1243 533 343h 27 III 23.3 2h1 425,3 0,167 1080 - 601 3228 28 IV 29.7 115 k16.6 0o 1559 530 962 29 - 27.2 119 0,16k 560 Iv 42,1 1378 - 962 "-' ‘ where 3 P =eheat generation rate 1n 11quid W a £ : n fi‘average number of turns per meter over length of* c01l m ;, - N #'number of small coil sections, - v ¥ a = radius of fluld zone, m, L = length of coil, m, - (o -1/2 p, = (2nfgu,) " , f = frequency, Hz, _ , o gz = gpecific conductiv1ty of liquid Q 1 l 3 Mo =_magnet1c permeablllty_of,the llquld,_N_A 2, ITOT.= total coil current (rms);-A, | ' k.=vcorrect10n factor, dlmensionless. Equation (l) is based on an approx1mate relatlonl for the rate of heat generatlon in an 1nf1n1tely long cyllnder p051t10ned inside an infinitely long 1nduct10n c01l, and is. valld for (a/pz) < 1 b, | ® The rate at whlch heat is generated in the plpe surroundlng the coil is given by the relation ' R ':' . - -V j : \ Lo Bloor| |%p| L_ | P A p) &p . lwhere Pp'é heat generatlon rate in pipe w, . _gp'= specific conduct1v1ty of plpe, Qfl'mf},, --ap'=1n51de radlus of pipe m, - | : L= 211'f : P, | &,y ) e ‘*ué ='magnetic permeability of pipe, N A : o kp =_correct10n factor dlmen51onless, S L o _ B “'” and ITOT’ n, , L and,f,are Qeflned above, Equation (2) is valid only The rate at which heat is generated in the 1nduction coil is glven by the relatlon Bl | | 2 £1/2 . ek ———b'NT-—-—flpc."Izv"L' o (3) S c ™ 1TN_ [dxl TTOT . — . - - where | P = heat generatibn rate in coil, W, b = inside diameter of coil, n, dc = cpnduCtor diameter, m, - L, = 1ength'qf small coil‘seCtion, m, NT'=—number of turns in small coil Section; N_ = number of small coil. sections.f , pc ='spec1f1c re51st1v1ty of c011 metal Q-m, . , = frequency, kHz, 1 3)L/2, Kl = proportlonallty constant, (fl/kHz m ) ~and ITOT and L are deflned above, The effect. of bubbles in the nitric acid on the heat gefieration rate 'vas,investigated with\coilinI and IV, Eight runs weré‘mgde with air - flow rates up to 2.l6vscfm, which produced bubble volume fractions-in'the' acid as high as 18%. 1In the rapge of bubble #olume fractions examined, the'falue of the correction factor'k defined by Eq. (1), varied approxi- mately llnearly with the bubble volume fraection as shown in Flg. 23 thls variation can be represented by the relatlon | ki: ko(l —'1;0795)’, i B :'7 ' (h) where 7 _ , 'k = correction factor, defined by Eq. (1), kg = cqnstant, | e = bubble volume fraction. The.effECt of bubbles in the liquid on the‘rate of heat'generation'is Vsllght as would be expected if the bubbles remalned near the center of the liquid zone. Values for k., kp,'and K., defined by Eqs. (h), (2), and (3), respectlvely, were determlned for the four induction c01ls tested (see.‘ Table 3). . The values of the constants for coil II are_smaller than the corresponding conSténts for the other coiis; the largést relative varia- ‘tlon occurs in the values of ko, which is proportlonal to the heat genera- tion rate in the 11qu1d. The low heat generatlon rate W1th this 1nduct10ni 10 08 O e o x]f 06 o | i i . cou.m ko :0.178 x cou.nr ko ozsn 04 _ 0.2 - - 0 |- ] 1 1 ] l 1 | | k=ko (1-1.079¢) | C | | ORNL DWG 7i-3846 R I [ ¥ I ! | B 0 002 004 006 008 00 _ 0i2 014 ol o’ «,BUBBLE VOLUME FRACTION‘ Fig. 2. Effect of Bubble Volume Fraction on Correction Coefficient \ for Calculating Heat Generation Rate in Nitrie Acid. 0.20 10 Table 3. Correctlon Factors for Heat Generation Rate Equations ; L . T L o301/ Coil ko K K = (@/xizen) 1 ~ oa3 o6k 1,015x1070 I 0.089 o7 L155x 1070 I 0.178 0.623 - 1.885x 107 o 0261 0.586 215 x1077 coil design is apparently the result of having all of the small coil rsectlons wound in the same dlrectlon since the remalnlng characterlstlcs 'of the coil are similar to those for the other coils, Coil IT would | require tfie largest current in order to produce a given heat generation rate in molten salt; however, it might have a high efflclency for A'heatlng the salt if the diameter of the molten reglon were suff1c1ently large The value of k, for coil I was smaller than that for coil III, -although the two coils have essentially the same design and comparable values were observed for the other‘constants; This varietion in ko values is probably due to changes in experimental technique and operating ¢onditions incorporated after tests with coil I were carried out, The new technique conS1sted in using alcohol-in-glass thermometers to measure the acid temperatures, since the temperatures indicated by thermocouples' used in the runs with coil I were affected to some extent by the rf power generation. Also, the coil designs tested after. coil I resulted in higher heat generation rates in the acid which, in turn, led to smaller relative errors in the measured heat generation rates in the acid, The difference between the k value for coil IV and the k, values for coils I and II is probably due, in large part, to the different spacing of the small coil sections. Coil IV had h—ln.-long small coil sections, with 11. T turns in each section (see Table 1), and the sec- tlons were separated by a space of 2 in. in which turns were not present. vy oy :a'.__ 11 - | Thetsmaller'sections.in the,other*three coils Were.placed closer together so that the coil turns vere'spsced uniformly-over the length of the acid 'column° the total number of turns was”about the same for each coil. In the case of coils I and III th1s spac1ng compressed the magnetlc field between each of the small c01l sectlons, thereby effectlng 8, decrease in the ax1al component of the fleld It is the axial component that pro- v1des the proper eddy currents for heat generatlon. Although any one of the four coils could be used to generate heat __1n the proposed fluorlnator, c011 IV would requlre the lowest coil current to produce the requlred heatlng and, for thls reason, would be “the most de51rab1e. Calculatlons have shown that, for a 5-in,~diam ,molten—salt zone a 5 56—in.-ID c011 made from l/h-ln. n1cke1 tubing (u51ng a c01l IV de31gn in whlch each small . sectlon has 9 5 turns over a 3. 75—1n. length,-with a 2 25—1n. space between small sections), and a 6= 9/16—1n.-ID nickel fluorlnator vessel an eff1c1ency of heatlng the salt (w1th no bubbles) of about h9% would be achieved with a total current of less than 150'A Wlth c01l III the eff1c1ency of heating Ithe salt would be about 58%, but a c01l current of 267 A would be requlred 12 3. SEMICOHTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL FACTLITY 'B. A. Hannaford C. W. Kee n - L. E. MclNeese A We have continued operation of a.facility in which aemicontinuous reductive extraction experiments can be carried.out_in a'mildeSteel sys- tem.h ‘Initial work with theefacility'#as directed toward obtaining data on the hydrodynamice of the countercfirrent flow of molten'ealt'and big=- muth in & 0. 82—1n.—ID 2h-1n.-long column packed with l/h—ln. molybdenum VRaschlg rlngs. We were able to show that floodlng data obtalned with 5 thls column are 1n agreement with predlctlons from a correlatlon based on studles of the countercurrent flow of mercury and aqueous solutlons in backed columns. We have carrled out - several experlments for deter- _lenlng the mass transfér perfbrmance of the packed column in whlch 8 -salt stream contalnlng UFh was countercurrently contacted zlth blsmuth containing reductant over a range of operat1ng condltlons. T It was found that the rate of uranium transfer to the bismuth was controlled by the diffusive resistance in the salt film under'conditions.sueh that the,concentration'of reductant_in the bisfiuth remained higfi,throughOut the column. The extraction dataicould be correlated in terms Of:the‘- height of an overall transfer unit based on the salt phase. In order to measure mass transfer rates under more closely controlled conditions where the controllingiresistance is not in the salt phase, preparations. were begun for experiments in which the rate of exchange of zirconium | isotopes will,be.fieasured'between salt and bismuth phases otherwise,at chemical equilibrinm. Difficulty was encountered at the beginningaof the first run of thls type because of a salt leak in the V1c1n1ty of the _rsalt feed-andecatch vessel 3.1 Replacement of the Salt Feed-and-Catch Tank | A new salt feed-and-catch tank of the initial_designa was fabricated and installed in the system.. Thermal insfilation was removed from ali' transfer ;ines tO'allow their inspection, ahd lines that were more’than C . . - od 13 'moderately oxidized were replaced. 1The'salt transfer 1ine from the . feed tank to the salt jack-leg was rerouted to a. p01nt 11 in. higher hthan in the original design in order to 1mprove control of the salt feed rate and to prevent the backflow of bismuth 1nto the salt feed tank during column upsets.f; After 1ts 1nstallat10n the new salt feed—and—catch tank was stress ' relieved by the same technique used for the original feed-and—catch ' | ;tanks, that is, the rate of heating to the operating temperature (650°C) was maintalned at less than 60°C/hr. The bismuth feed-and-catch tank, ,_which had been allowed to cool -during the time required to fabricate and install the new salt tank was heated to the operating temperature (550°C) at the samelcontrolled rate._ ‘Both the salt and bismuth feed-and-catch © ‘tanks were subjected to}afpressurefprOOf test'at-the'operating tempera- :'ture-after'frOZen bismuth:sealslhad'been'estahlished'in the freeze valves tin order to isolate the feed tanks (rated at 50 p31g) from the receiver ‘1'tanks (rated at 25 p51g) After the pressure tests had been successtlly completed the salt feed—and—catch tank - and the newly 1nstalled salt- :transfer lines were contacted with a hydrogen stream for 13 hr at 600°C in order to remove accumulations of iron oxlde from the 1nternal sur- o faces of the system. : 3 2 Preparation:for Mass-Transfer'Experiment ZTR—l After the system had been treated with' hydrogen for removal of most - .,of the 1ron ox1de, 1t was necessary to (l) add reductant to the bismuth : h-'phase, (2) 1ncrease the salt 1nventory in the system to ‘about 20 liters, "'.; and (3) c1rculate the salt and bismuth phases through ‘the system in 'torder to remove 1mpur1t1es that might have been 1ntroduced during the "maintenance operations.- It was also necessary to 1ncrease the zirconium | glnventory in the system, to remove 1mpurities from the salt phase by "lhhydrofluorination, and finalky, to add 8 suff1c1ent amount of reductant - to the bismuth phase to produce 8 zirconlum distribution ratio of ‘about | 1; These operatlons are discussed in the remainder of this section. A 103—g quantity of thorium was suspended in the bismuth phase in T s 'the treatment vessel in & perforated contalner as described earlier in o order to increase the reductant concentratlon in the blsmuth to about - 0.002 equiv per g-mole of blsmuth During d;ssolutlon of the thorlum, the treatment vessel was held at 650°C, and argon was fed to the draft tube in the vessel at ‘the rate of 2.5 std ft3/hr Only S gdof thorium | remalned undlssolved after a period of U1 hr; analyses of blsmuth sam- ples for uranium and thorium showed that sbout 80% of the thorium had dissolved during the first 2k hr. About 18 liters of salt (72—16-12' mole % L1F-BeF -ThFh) was. then charged to the treatment vessel in order to replace salt that had been discarded when the original salt feed— 'and—catch tank was replaced.. The salt and blsmuth phases‘uererequlll— ,brated'in the‘treatment vessel for about 20 hr before:they vere trans- ferred to their respective feed tanks. Bismuth;andVSaltiwere'then f circulated through the system in order to complete theuremOVal of oxides 'that had'not been removed from the internallsurfaceshof'the system after the previous treatment with hydrogen. In addltlon, column pressure drop measurements were made durlng a period when only salt was flow1ng through the column The observed pressure drop was about 2 1n. H 0 at the salt - flow rate of T0 ml/min, which is in agreement wlth data obtalned soon after the column was installed. It was concluded that the flow charac- teristics.of the colum hed not chenged during runs made to date. The zirconium mass transfer experiments require that a significant quantity of zirconium be present in the salt and bismuth phases to ensure that only a negligible change will ocecur in the zireconium dlstrlbutlon 97 ratio durlng the transfer of ° 'Zr tracer from the salt to the bismuth phase. A 5.2-g quantlty.of Zircaloy-2 was dissolved in the bismuth to. increase the zirconium inventory of'the system to about 15 g. | The salt and bismuth were then contacted in the treatment vessel - with a 30% HF--hydrogen stream hav1ng a flow rate of about 16 std ft3/ hr at 650°C in order to remove oxide (from the salt) that mlght have ‘ accumulated during the prev1ous transfer of the salt and bismuth through the facility. Treatment of the salt and bismuth Wlth the HF—H2 stream " was 1nterrupted after,about 2 hr by a restriction caused,by'deposition‘ of material on the Sintered-Monel_filter in the off—gas_stream from the _ treatment vessel. The filter;fwhich has an external surfaceuarea of ...._.,...AA‘.,..N_V“#AAM. - o -y 15 about 20 1n.2, 1s used for remov1ng partlculates which would otherwise cause restrlctlons in. valve ports in the off-gas system. The filter was removed and replaced byra h—ln.-dlam, 8-in.-long cyllnder?of'com— ':pacted copper meSh | Analyses of the black'solids removed from the Monel filter showed that the material con51sted prlmarlly of carbon, along with ‘substantial amounts of thorlum, uranium, and lithium. After the treatment w1th HF-H,, the salt and bismuth were contacted with argon 23 at about 3 std ft /hr for. a 20—hr period in order to remove HF from the salt. The blsmuth and salt. were subsequently sampled, and a perforated 'basket contalnlng L8 g of thorlum metal was. suspended in the bismuth . phase. After a perlod of 66 hr at a temperature of about 620°C, only »25 g of the thorlum had dlssolved In order to add reductant to the bismuth phase more rapldly, 2h0 g of L1-Bi alloy contalnlng 1.75 wt % lithium was added to,the»treatment vessel. The total quantlty of ' reductant addedfduring'these periods was about 1 g—equiv. The resulting '-21rcon1um dlstrlbutlon ratlo should have been. about 5 1f 1t is assumed that all of. the added reductant was present in the blsmuth as uranlum, ‘z1rcon1um, thorium, and lithlum.- However, the results of exper1ment - ZTR-1, descrlbed ir the follow1ng sectlon ind1cate that thls assumption is not Valld 3. 3 Mass Transfer Experiment ZTR-1 At the conclu51on of a 22—hr equillbratlon perlod which followed __"the flnal addltlon of reductant to the treatment vessel, the salt was “f:transferred to the salt feed tank The transfer of blsmuth from the treatment vessel was only about 50% complete when a failure of the_ ‘-_transfer llne inside the vessel'at the weld which 301ns the molybdenum tublng to. the mlld—steel transfer llne made it necessary to cease this ':“h’0peratlon.' The affected portlon of the transfer 11ne was replaced at a later date however the 1ntended duratlon of experlment ZTR-l was 'h;_reduced in order that the run could be carrled out w1th the smaller ' "famount of blsmuth that was avallable 1n the blsmuth feed tank 96Zr0 that had been irradiated for 12 hr at A 6 T-mg quantlty of 5 | 1k ' -2 -1 a thermal neutron flux of about 2 x lO neutrons cm sec — was -~ 16 -transferred to a 0. TS-ln.-dlam steel capsule after an. 18-hr 'decay perlod to fac111tate addition of the 9TZr tracer to the salt’ phase..‘Perlod;c salt samples taken after immersion of the capsule indicated that‘miiing of the traéér‘with the salt phase‘was COmplete‘after‘é hr;. In'orderrto verlfy that most of the tracer had entered the salt phase the steel o1 ~ addition capsule was counted for “'Zr activ1ty.- The counting results for the bulk salt and for the capsule showed that greater than 99% of the tracer had been transferred to the salt phase. | ,The volumetric flow rates_for bismuth and salt during experiment‘_ :ZTR-l were.2l6 and.99 ml/min respeetively; these values are equivalent l'to abont'90% of the”combined colufin flow capacity'at flOoding. Seven ' sets of flowing stream samples vere taken over a 29-m1n perlod. Countlng o the 2w i transfer.of 91 Zr act1V1t1es in the samples showed that no measurable Zr tracer from the salt to the bismuth phase had_occurred; during the experiment. It was later fofind’that-the lack of transfer was due to an unexpectedly low distribution coefficient for zirconium, which resulted in essentially no zirconiumdbeing present in the bismuth phase. »Wet-chemical analyses of post-run, equilibrated.sampleS‘fOr lithium and uranium'implied a zirconium distribution coefficient Value of about 0.001. A more precise value (o. 023) was obtained by countlng the samples' 97 9TZr activity. It was concluded that essentlally all of the ) reductant that had been added to. the system had been consumed by one or more side reactions, for exemple, the reductlon.of FeF2 1nithe salt phase to metallic.iron, or the reaction of reductant with HF that was _ desorbed from the graphite crucible.’ It'is‘alsozpossible:that a fraction of the 1lithium in the Li-Bi alloy reacted with air or water.vapor during its addition to the treatment vessel, 3.k Variation of Reductant Inventory in the Bismuth Phase in the Treatment VeSsel- ' | Durlng this report perlod we Observed 8 cons1derably greater varia- .dtlon 'in the reductant 1nventory in the blsmuth phase in the treatment vessel than had been expected, consequently, we have begun to give addl- tional attent1on to this subJect. Reductant.can be_removed_from the -l . 1T bismuth phase by a nunber of-side reactions,-including' (l)\reaction of reductant w1th materlals in the salt phase such as FeF HF, or 3 oxygen-contalnlng compounds; and (2) reaction of reduced ietals (uranium, thorium, zirconium) with the graphite_cruclble in the treatment vessel. Information related to the variation of rednctant inventory'in the bis- muth phase, as aell astheinventory~of uranium in the system, will be reported here and in futarejreports covering work in this experimental facility sorthat the phenomens responsible for the observed effects can ' be.identified. Data'on the variation of inventories of reductant, uranium, and | zirconiUm’dnring this report period are summarized in Table 4. At the beginning of the period, the treatment vessel contained_IT.T liters of bismuth and 1:4 liters of salt__(72-16-12~molé % LiF-BeF ~ThF) ). After a 118-day period in which the salt and bismuth were held in the treat- ment vessel, the reductanttinventory’in-the bismuth‘had'decreased from 0.80 g—eduiv to 0.058 g—eqniv; ,Thejaverage rate of decrease in redué- tant concentration during,this period_this period was 0,26'meq/hr.. The addition of 1. 69‘g4equivrof'thorium;metal to the bismuth phase resulted in approx1mately the expected change in the reductant concentration in - the blsmuth based on uranlum and zirconium analyses. FolloW1ng the addition of salt to the system to 1ncrease the salt volume to about 21 5 liters, samples taken from the treatment vessel showed that the "~ reductant 1nventory had decreased sllghtly to 1. 33 g-equiv; 8 decrease ‘was expected because of the probablllty of 1ntroduc1ng small amounts .uof oxxdants during the addltion of salt to the system.: Subsequently, ' ,_athe salt and bismuth phases were circulated through the system (run ’"HRw13) 1n order to remove DdeES whlch may have been 1ntroduced durlng | -ivthe 1nstallat1on of new carbon steel lines and equlpment Analyses of 'blsmuth samples showed that about half of the reductant was removed "from the blsmuth durlng thls operation.r Next, zlrconlum metal (O 23 : g-equlv) was dlssolved in the blsmuth phase in order to achleve the :hd651red 21rconium 1nventory in the system. Followlngrthls addltlon, .the_salt and bismuth‘werercontacted with an HF-H_ mixture. At this 2 point, the salt phase should have contained all of the.uranium and Table 4. Summary of Reductant, Uranium, and Zirconium Inventory Data for Treatment Vessel Salt Phase Bismuth Phase Total ‘ Totsal : ‘ Salt Uranium Zirconium Bi Uranium Zirconium Total Reductant® Combined Phases Wt Inventory Inventory Wt Inventory Inventory (u, Zr, Th, L1) Uranium Zirconium Operation Sequence (g) (g-equiv) (g-equiv) {(g) (g-equiv) (g~equiv) (g-equiv) (g~equiv) (g-equiv) 1. Material remaining 4,760. 0.152 0.0313 171,200 0.601 0.10 0.80 0.753 - 0.13 in treatment vessel . ' at time of salt feed tank failure 2. Following 118-day 4,760 0.7022 0.13% 171,200 0.058 a0 2 N 0.76% 0.13% equilibration period ‘ 3. FPollowing addition of 4,760 0.003° n0% . 171,200 0.757 0.16 1.65 0.76 0.16 - 1.69 g-equiv of Th o ' reductant 4. Following addition of 72,560 = 0.061 0.007% 171,200 0.700 0.24% 1.33 0.760 0.242 67,800 g of LiF-BeF,~ ThF, (72~-16-12 mole %) containing 0.114 equiv of Zr ' 5. Following equil- "~ 63,670 0.337 0.0922 168,250 0.435 ' 0.1522 0.74 0.772 0.242 " ibration run HR-13 ' 6. Following addition of 63,670 0.76Y - 0.478 168,250 ~0 @ no @ nO ‘ 0,762 0.47% 0.23 g~equiv of Zr and . HF-Hz treatment ’ 7 7. Following addition of 63,670 0.76° - 0.47° 168,250 ~0 & o 8 . ap \ '0.763 0.478 1.04 g-equiv of o ' (Th + Li) reductant 8. Following tracer 63,670 0.72 0.62 168,250 <0.003 no 2 . ~0 . 0.72 ' 0.62 experiment ZTR-1 ®Inferred value based on material balance and/or equilibrium considerations. The best value for uranium inventory was taken to be 0.76 g-equiv. bChemicéi analyaié resulted in invefitbry in one phase equal to 0.76 g~equiv t 4Z; this value was taken to be the more accurate measure. 8T » " i ‘Ofi" - - zirconium, and samples showed a uranium concentration in the salt that was within}about»h%'of\the:expected value. We then added to the bismuth . a quantity:of reductant (o. h3 g—eQuiV'of thorium, 0.61 g-equiv of lith- ,1um) theoretlcally sufflclent to produce a zirconlum dlstrlbutlon coef- ficient of about 5, in- the absence of reductantuconsumlng side reactions. 'However analyses of blsmuth and salt before and after tracer experiment - ZTR- 1 showed that the- reductant had been consumed almost entlrely before the experiment was-performed, as discussed earlier. The material balance'for uranium throughout the report period was 'excellent, as shown in Table h the zirconium balance was satisfactory - in view of the greater d1fficulty encountered in’ analy21ng samples for '21rcon1um at low concentratlons. 3.5 . Operation of the'Argon‘Purification'System The argon purlfication'syStem, described earliér,hwasmddified by ~ the addition of a parallel purlflcatlon system (Englehard Deoxo Puri- flers, Models D and C. 1n serles) The purpose of this modification was to evaluate the effectlveness of Englehard un1ts relative to that of the regular purlflcatlon traln, whlch cons1sts of a bed of molecular sieves followed by & bed of uranlum turnings at 650°C The Delph1 trace oxygen analyzer indicated'an oxygen concentration s_of 3.6 ppm 1n the argon stream leaving the Deoxo unlts, as compared j~w1th a value of 1.6 ppm measured in the argon stream leav1ng the regular flr_prpurlflcatlon traln. The comparison was not completely conclu51ve how- . ever, because of the poss1b111ty of sllght air 1nleakage and the pOSSl—- Jblllty of catalyst p01soning 1n the Deoxo unlts The Delph1 analyzer had also shown symptoms 1nd1cative of 31lver cathode p01son1ng, although it had been restored to serv1ce by heatlng the cathode grlds to 800°C ;fln a1r to remove suspected surface contaminatlon. Recurrlng fallure of ¢the Delph1 analyzer mllltated against our obtalnlng a rellable compari- | son of the two. purlflcation systems._ Prlor to the flrst ev1dence of maloperatlon the Delphl analyzer indlcated that the regular purlflcatlon ) system was redu01ng the- oxygen level from about 0. 75 ppm in the inlet argon stream to about 0.2 ppm in the outlet stream ,7 ,20 'MeESufementé of the same argon'stfeamé’Showed water contehts_df' <0.1 ppm, the concentratlons were usually. <0.01 ppu. It was dbservéd o ',‘ | - however, that the water concentration 1ndlcated by each of the Panar | metrlcs probes tendedlto dlm;nlsh over. & perlod of many weeks. This ' . suggested.that'the.calibratioh was éhifting &oWnécéle\with‘tifie, since . o o a new probe installed in the same 1ocat10n would generally 1ndlcate a | 51gn1flcantly hlgher concentratlon (1 €5 2 ppm vs 0.01 ppm) Desplte the dlfflcultles experlenced in measuring the level of oxygen and water in the purified: argon the concentratlons were estab- | jllshed to 11e within limits which were acceptably low. Cn “t 21 b, DEVELOPMENT OF THE METAL TRANSFER PROCESS: | INSPECIION OF EXPERIMENT MIE-2 o E, L,fioungblood L. E;_McNeese | It has been found that rare'earths'distribute.selectively into mol- ten LiCl from bismuth solutions-conteining rare»earths and thorium, and an 1mproved rare—earth removal process ‘based on thlS observatlon has been dev1sed 9 ~Work that will demonstrate all phases of the improved rare-earth removal method Wthh is known as the metal transfer process, is presently under wgy,rpif_u We prevlously . carrled out. an englneering experlment {MTE-1) for . studying the removal of_rere-eerths from 51ng1e-flu1d_MSBR fuel salt by this process?, During'theyexperiment;vspproximately_SO%'of the lanthanum ~ and 25% of neodymium origineliy present“in the'fluQride”salt,were removed at asbout the expected rate. Surprisingly, however, the lanthanum and ‘neodymium removed from the fluoride salt did not accumulate in the Li-Bi | solution used for remov1ng these materlals from LiCl. It is believed that reactlon of 1mpur1ties 1n the system w1th the rare earths caused this unexpected behaV1or._ _ ' A second englneerlng experiment (MTE-2) wes recently completed 11,12 A brlef descrlptlon of the equlpment used for this experlment and the _results of an 1nspectlon carrled out after complefilon of the experiment are presented in the remainder of this section.- h l Descrlption of Equlpment Experlment MTE—E was perfonmed in a vessel constructed of 6~in. ‘ d”,gsched hO carbon steel plpe._ The out51de of the vessel vas spray coated ) with a 20—m11 thlckness of nlckel alumlnlde for protectlon agalnst oxi- hffdatlon. The vessel shown schemstlcally in Fig. 3, was d1V1ded into two _.fcompartments by a partition (constructed of 1/h—1n.—thick carbon steel 7lplate) that extended to w1th1n 1/2 in of the bottom of the vessel. -The ' two compartments were 1nterconnected by 8 2-1n.-deep pool of blsmuth that was saturated with thorlum ~ One compartment contained a 3.6-in.- deep pool of | fluorlde salt- (72-16—12 mole % LiF-BeF -ThFh to which 7 mCi 22 ORNL DWG 70-12503-RI LEVEL ELECTRODN ARGON INLET AND VENT " L =] _—CARBON-STEEL PUMP WITH MOLTEN Bi CHECK VALVES 6-in. CARBON - CARBON-STEEL e = PARTITION ) STEEL PIPE 24 in. ic 72-16-12 MOLE %~ = ——Licl FUEL CARRIER SALT Th-Bi \ Li-Bi Y Fig. 3. Carbon Steel Vessel Used for Metal Transfer Experiment MTE-2. * ny ud 23 of lhTNdand'sufficientLaF were added to produce a 0 3 mole % concen- 3. tration) above the Th—Bl phase The other compartment contained a 4.2 1n.—deep pool of molten Lici above the Th-Bi phase The LiCl compart- ' ment also contained a cup (l.9h in. in diameter, 8.25 in. high) which was initially filled to & depth of ) in. with a 35 at. % Li-Bi solution. The cup was constructed of 0. 031-1n.-th1ck carbon steel sheet metal and was held in place by a holder made of 2-in. sched hO carbon steel plpe. Alumlna spacers were used to electrlcally 1nsulate the cup from the holder. During operatiOn LiCl was circulated through-the cup containing the Li-Bi solution v1a 8 pump constructed of l-l/2—1n.-d1am carbon steel pipe (0.083-in. wall thlckness). The pump used molten bismuth as check _ 9 _ valves. During ther3.3+month'per10d in which the,experlment was in operation, Tb? liters of'LiCl'waS'circulated‘through the cup containing the Li~-Bi solution. Gas-llft sparge tubes were used in both compartments of. the vessel and 1n the cup contalnlng the L1-B1 solutlon to 1mprove contact between the salt and metal phases. The sparge tubes were con- structed of 1/h-in. carbon steel tubing which was placed inside 3/8-in. tubing as shown in Fig. H.' Thermowells, constructed of l/h—ln.—dlam .ecarbon steel tublng, extended 1nto the salt and blsmuth phases for temperature measurements. The lower sectlon of the. vessel was maintained _at the. operatlng temperature (650 to. 660°C) by an 8-kW furnace. The. upper 6 1n of the vessel was wrapped with a coollng coil . through which ewater was c1rculated 1n order to maintain the flange at. about 100° - The equ1pment performed satlsfactorlly durlng operatlon. At the | rcompletlon of the experlment the vessel was cooled to room temperature, '-:;”w1th the salt and blsmuth phases 1n place and- vas. cut apart for 1nspec- .tlon.- h 2 Inspectlon of Equlpment . B The vessel was removed from.the furnace and the exterlor of the vessel, shown in Flg 5, was v1sualLy 1nspected Some bllster1ng and _cracklng of the 20-m11—th1ck nickel alumlnlde coatlng had occurred durlng the 2370—hr‘perrodrthat~the vessel had been_held at-about 650°C. However, 24 ORNL DWG 70-8980RI a3 Q V4-in, CARBON STEEL TUBING 0.035-in. WALL | SIX 0.040-in.-diom HOLES EQUALLY SPACED 3/8-in. CARBON STEEL TUBING 0.025-in. WALL Fig. 4. Gas-Lift Sparge Tube Used for Metal Transfer Experiment MTE-2. 25 at 4y " PHOTO 101874 .y ") . Fig. 5. Photograph ShGW1ng External Air Oxldatlon of the Exterior of the Carbon Steel Vessel Used for Experiment MTE-2. " The exterior of the vessel had been coated Wlth 20 mils of nlckel alumlnlde to retard air oxidation. b 26 the external oxldatlon that had occurred did not cause dlfflculty with the experlment. To'facilitate inspection of its interior, the vessel was cut in such a manner that the lower 11 in. on each 51de of the partition could be removed to expose the salt and metal phases. A view of the fluorlde compartment 1s shown in Fig. 6. The fluorlde salt and Th-Bl phases appeared. to be clean and free from any accumulatlon of materlal at the | -salt—metal interface; ‘hovever, the vapor region of the fluorlde compart- ment was covered with a black powder having the comp051t10n (by weight). of: 2.4% Li, 2.2% Be, 70.2% Bi, 1.1% Th, 0.02% Fe, and 19.9% F. The 'h dep031t was greatest (about 1/8 in. thlck) in the cooler‘portlons of the compartment near the upper flange. Some of the powder had also dlscolored the surfaces of the salt that had contacted the vessel wall The black powder is thought to be a mlxture of salt and metalllc bis- | muth that had‘been entralned 1nto the gas space by the argon sparge. A v1ew of the LiCl compartment (W1th some of the LlCl removed) and the Li-Bi container is shown in Flg. T. The vapor region in the LlCl compartment was covered W1th a white powder consisting of L101 contalnlng 0.6 wt % Bl. This material is belleved to have resulted from vaporiza- tion and from entrainment of LlCl and smaller amounts of bismuth in the ‘argon stream fed to the gas spargers. The LlCl and'Th-Bl phasesrappeared to be c1ean, with no accumulation of impurities at the salt-metal inter- face as had been seen in metal transfer ekperiment MTE—l.el o There,were only two areas in the system where'deposits'containing' unusually high concentrations of rare earths were found. A l/8-in.- “thick 1ayer‘of gray material (shown in Fig.,8) had deposited on the lip and‘overflowZSPOUt,of the Li-Bi container. This»deposithadthe following ‘composition (by weight): 23% LiCl, 59% Bi, 10% L&, and 2% Th. The lan- thanum.contained in the deposit was-eoual to 5 to 10% of the lanthanum - inventory in the-System. -The mechanism by which the material was depo— ] sited on the rim of the lithium-bismuth cup has not been determlned however, 1t may have resulted from the L1-B1 solutlon wetting the con- | tainer wall and subsequently flowing*up 1t.; The lanthanum could then PHOTO 102034) ta Salt and Bismuth _Pharses from the Fluoride Salt Compartment' 6 on Completion of Metal Transfer Exper Fig. t MTE-2. ilmen PHOTO 102035 28 View of LiCl Compartment Following Metal Transfer Experiment MTE-2. Fig. T. Lng owl Experiment Sh contaliner Metal Transfer in -Bi Conta . .\ f , - - o . g o o . , Q ot 4 P oy , o 0 B , , od . . 0 , nc , . o .. | _ : o K : : | , B e S e , ced o o o 34 | | | By 5 <3 ; _ 5 o 5§ | 2 o B oo . n o . 2 | © o _ a W o o ol , g R | | 28 | 30 have been deposited from-the LiCl that was in'contact'with tne metal film. TheklanthanumCOneentration in the bottom layer of'fhe'Th—Bil _.solntion was approximately‘eight'fimes higher than that'observed in | flltered samples taken durlng the run; however thls does not represent a significant fractlon of the total lanthanum in the system The bottom layer of the Th~Bi phase also contained 20 wt % thorlum-and_ls assumed to haxe been a mixture of thorium bismuphide particles and bismuth. The higher lanthanum concentratlon in this material 1s not surprising 3 since it has been shown prev1ously that rare earths distribute prefer- entially to the solld phase via formation of:compounds of. the type ' ThLaBly. -inspeCtion'of the carbon steel vessel interiOrfrevealedflittle evi- dence of corrosion as shown. in Figs: 6 and T. However, some corrosion did occur on the cup that contalned the Li-Bi solution, and on thermo- wells and sparge tubes. All of these 1tems were constructed of thln .cerbon steel..The‘corrosion of the Li—Bi'cop_occnrred mainly at the salt-metal interface. A crack and a 5/8-in.-diem hole had developed in the vicinity of the interface end hed allowed e portion of the Li-Bi solution to run'into the holder. Data from the experiment.indicate : that the hole'had_dereloped sfter about two months of operation; how- ever, it did not cause seriousidifficnlty since the Li-Bi .solution was ~ contained in the holder and did not mix with the other phases in the experiment., The‘lower.portions of the.carbon'steel sperge tubes and ~ thermowells that were in contact with the salt and bismuth phases during the experiment are shown in Fig; 9. The sparge tubes in.both the fluo- ride salt and the LiCl compartments were severely oorroded,>particularly in the srea near the salt—oismuth‘interfeoe. The 3/8—in.¥diam-tubing | was &bsent from the sparge tube that was removed from the fluorlde salt compartment. The sparge tube from the Li-Bi vessel and the thermowells werellessssererely corroded. The corr031on,observed on the carbon steel eomponentslis thought to be due mainly to mass transfer of iron due to a thermsl gradient 1n the bismuth phase.. Iron has a solubility of about 80 ppm in bismuth at 650°C 14 and thermel gradients in the | experlment could cause. 1ron to be dissolved in hot areas and dep051ted 1n cold areas. ot - Fig. 9. Carbon Steel Sparge Tubes and Thermowells Removed After . Completion of Metal Transfer Experiment MTE-2. 32 - The carbon steel pump Wthh used bismnth check valves, was in good ccondltlon at the completlon of the experlment Flgure 10 shows a view - of the lower portlon of the pump after 1t had been sectioned to show the b1smuth check valves. thtle evidence of corr091on could be found by - visual examlnatlon of the pump.' There was no apprec1able loss of bls- o muth from the check valves by its entrainment in the LlCl durlng the ' experlment. The blsmuth 1n the top check valve contained 130 ppm of Li, | less than 50 pPpm cf Th ‘and less than 20 ppm of La after 702 liters of LiCl had been c1rculated through the pump.- While carbon steel is not be1ng considered as a materlal of con- struct1on for an MSBR proce351ng plant it eppears to be suitable for ',use in experlments such asIMTE-2 where a llmited amount of corr051on' is acceptable. S ~ Fig. 10. ILover Portion of the Carbon Steel Pump Used in Metal Transfer Experiment MIE-2. The pump has been sectioned to show the bismuth check va.lves. e T S 3 5. DEVELOPMENT OF THE METAL TRANSFER PROCESS: AGITATOR TESTS FOR EXPERIMENT MTE-3 E. L. Youngblood ~ W. F. Schaffer, Jr. , Mechan1ca1 agitators will be used to promote contact of the salt and metal phases in metal transfer experlment MTE-3, whlch is currently belng designed and.constructed,15 The shaft seals for.the agitators must 'be capable of operating in a dry argon atmosphere and mustrhave a low ' leakage rate in order to prevent-eir and moistureufrom entering the exper- ‘iment. Equipment has been constructed in order to testlthe'shsft sedl . . design that‘iS'proposed for_use in metal transfer experiment MIE-3. .The'j --system will also sllow us to measure the_extent'to which bismuthris ‘entrained in salt in c mechanically'agitated'system and tofevaluate a 'vapor-dep051ted tungsten coatlng as. a means for protecting carbon steel from corrosion by molten salt and blsmuth. 5.1 Description of Equipment " Figure 11 shows the test equipment before'inStallation_of the elec- trical heaters and thermal insulation. The agitator drive assembly (shown in Fig. 12):consistedvof a 1-1/4-in.-diam steinless steel shaft held in position by two ball bearings that were separated by a distance of-fi in. After passing through ‘the ball bearlngs, the shaft dlameter was reduced t0 1 in.; the shaft passed through two Bal—Seals (product of Bal-Seal Englneering Co.) before enterlng the test vessel. The seals were con- structed of graphlte—lmpregnated Teflon and were spring-loaded in order to hold the seallng surface agalnst the shaft. The portlon of the shaft .'that was in contact with the seals was plated with chromium and polished to_a 10- to 12—u1n. surface that would produce satisfactory seallng. The region'between the two seals was pressurized'with-argonpin order to reduce the rate of air 1n1eakage past the seals. A hwin;-long cooling _water jacket was 1ocated below the shaft seal as a means of protectlng the seals from damage by heat from the lower portion of the system (which operates at 650°C). A thermowell was prov1ded for measurlng the temper- ‘ature in the v1c1n1ty of the seals. - - Fig. 11. Agltator Test System Used for Testing the Sha.ft Seal Proposed for Use in Met&l Transfer Experiment MTE-3. - I wp . B ) 37 The agltator used for the test, shown 1n Fig. 13, was machined from a single bar of molybdenum. The agltator shaft was 12 31 in. long and had a diameter of 0.5 in., Separate blade assemblles were located in the ‘salt and bismuth phases. Each assembly had a diameter of l 12 in. and a height of 0.5 in.; each blade was 0.13 in. thlck.\ The upper end of the agitator was threaded to facilitate'its attachment to the drlve unit. The vessel used to conta1n the salt and blsmuth for the agitator test was constructed of 3-in. sched 80 carbon steel pipe (ASTM A 106 Grade B). A standard-plpe cap was used for therbottom_of the vessel. The overall length of_the'veSSel was,20.8;in. jFourvbaffles‘(3 in. long, 1/2 in. wide) were'welded'totthe inside of the'vessel_begihning at a p01nt 2-1/h in. above the bottom of the vessel. A.l/h-in.-sched 40 pipe was attached to the 51de of the vessel to allow sampllng of the salt and ‘bismuth phases, and a l/h-in. pipe was attached to the bottom to allow the salt and blsmuth to draln from the vessel. The lower 10 in. of the ,vessel interior was . coated with tungsten in order to evaluate the effec— tiveness of thls type of coating for reduc1ng corrosion in systems con- C ta1n1ng blsmuth. In applylng the tungsten coatlng, the 1n31de of the vessel was first - plated with nickel (approxlmately 1 mil thlck) by electrodep031tlon. The ‘nickel layer was bonded to the vessel by malntalning the vessel in vacuum at 800°¢c for-h hr. Heaters ‘were then installed on’ the vessel, and the .tungsten coatlng was vapor dep031ted from a H -WF6 m1xture.16 The temper- “ature of the vessel varied from about 400 to 650°C along its length durlng the coatlng operation, consequently,'the coatlng thlckness varled | rt:‘from 0. 00h to° 0. 020 in., w1th the thickest depos1t belng located near a'tthe bottom of the vessel.- Flgure 1h shows & view of the 1nter10r of the fl:vessel after the coatlng had been applled. Examinatlon of the coating _w1th 8 borescope revealed no obv1ous s1gns of cracklng or: bllsterlng. The exterlor of the carbon steel vessel was spray coated wlth 8 20- S mil 1ayer of nlckel aluminide in an effort to. retard air oxldatlon. How- ever, ‘before the n1ckel alumlnlde was applied, half of ‘the vessel was | flrst sprayed with stalnless steel t0 determine whether such a coating would provide improved protection against oxidation. During operation, s, gz e T g i " ey LB S - Fig. 13. . Sl N Molybdenum Agitator. L L5 g # Wi e T A 39 - _InéideJof,Tést Veésel Showing Tungst en Costing. yo the lower section of the vessel was heated with'tubularfelectric heaters. The upper section of the eduipment WaS equipped'with a cooling coil in - - .. order to malntaln the shaft seal temperature at about 50°C. The agitatOr was driven by a 1/h—hp varlable-speed motor that was coupled dlrectly to the drive unit. 5 2 Experlmental Results The shaft seals used in the first test were Bal-Seal No. RBOhA- (SZ)G120 The agitator was anltlally operated ‘at 200 rpm-for 100 hr with . no salt or bismuth in the system. During the first 50-hr period the vessel was held at room temperature durlng the remalnlng 50 hr, temper- atures for the vessel and the seal vere malntained ‘at 650°C. and about _:'50°C respectlvely. Throughout the test the seal leakage rate was deter- o mined by pressurlzlng the region between the two seals with argon and | - measurlng the rate of decrease in pressure when the argon supply was shut .off.. The seal leakage rate durlng the first 100 hr, as measured at 1 atm and ambient temperature, was about 3 cm3/hr. The internal pressure in the reglon between the seals was 15 to 20 p51g inltlally. - ' o - - After the 1n1t1al testlng of the seal 35h9 g of purlfled bismuth and 915 g of fluoride salt (72-16-12 mole % LiF-BeFeuThFh to which 0.3 " mole % LaF ‘had been added) were charged to the system. The salt-bismuth ‘interfacezas located at a point about 1 in. above the lower agitator blades, and’the‘Salt—gas interface was 1ocated'about 1 in.‘aboue:thei ~upper agitator blades. During the f011owihg»one-month period-'the'agi— ‘tator speed was 1ncreased stepwise from 150 to 750 rpm; the salt ‘and V.blsmuth,phases vere maintained at.650°C. The operat;ng tlme.at each speed is sumarized in Table 5. - During most ofrtheAtest period the geal leakage rate remalned constant at about 10 cm /hr however near 'the end of this perlod it 1ncreased to greater than 100 cm3/hr. After a total operatlng perzod of 8h5 hr, the salt and blsmuth were drained from the system and the seals were removed for 1nspect10n._ x The upper seal was found to be badly deter;orated, whlle,the.Teflon | had worn through to the spring in some areas. _The-lower seal, although | Qi') islightly worn, appeared to he-in good condition. Evaluation of various wj 3 L -fually 1ncreased to 0. 4 cms/hr._ A second 1nject10n of 1 cm b1 _Table,5. 'Operating Time During Tests of Bal-Seal No. R30LA-(SZ)G120 Shaft Seals Ag1tator Speed | ~ Operating Time (rpm) () No Salt or Bismuth in System 200 - . 100 - Salt'and Bismuthrin System': 150 50 L2000 - 48 0 18 . Total = 845 seals was continned'withont"the use of'salt'and bismuth'in the system. Short - tests were made at agltator speeds of 100 to 500 rpm ‘using Bal- 'Seals hav1ng a light expander spring, however, these seals leaked exces- sively and were replacedrw;thiBal-Seals (No. R3-6A-(l.OOO)G) having a moderate expander spring*andha'thicker-Cross'section than therseals used ' initially The leakage rate using these seals was satlsfactory, and | Vutestlng was contlnued at amblent temperature for TO days u51ng agitator i‘:fspeeds of 150 to 300 rpm.: Dur1ng the flrst 31 days of operatlon, the L;.'rate gradually 1ncreased from 0.3 cm3/hr to 2 cm3/hr At that tlme, 3 Tdems of mlneral oil was 1njected into the region between seals to {;_determlne whether thls would reduce the seal leakage rate and increase 'hthe seal life After the Oll had been 1nJected the 1eakage rate decreased ;to 0. 05 cm3/hr. During the next 2h days of operatlon, however, it grad- 3 of oil between 'ithe seals resulted in a decrease in the leakage rate to 0 02 cm3/hr and hconflrmed that the use of oil 1s effectlve in reduclng the seal leakage r—rate. 42 After Tl days of operation the seals were-remoVed for inspection. 'Although both seals showed some wear, a sufficient wall thickness remained to allow a eonsiderably longer operatihg time,' BasedLOn'these'test' ‘results, 1t .was determlned that this type of seal is acceptable for use in metal transfer experlment MTE-3. At intervals throughout the period during‘which salt and bismuth were present in the system, unfiltered‘samples-of the salt'phase were _ taken to determine the extent of bismuth entrainment in.the selt. Also, unfiltered bismuth samples were takeh'for‘nickel_and iron analyses in Aorder to determine'whether the’tUngsten coating was intact. Daring the test the bismuth content of the salt increased from 8 ppm- to 101 ppm as shown in Table 6; however, there was no 1ndicat10n that 1arge quan- 'tltles of blsmuth were belng entralned in- the salt. The purlfled blsmuth- charged to the experlment contained 10 ppm of ireon and less than 20 ppm of nickel. During the test the‘concehtration of‘nickel,in the hismuth - increased to about 1000 ppm (shown in Table 6), which indicated that the bismuth had penetrated the tungsten eoating. Thercohcentration of ~iron ia the bismuth phase increased: from 20 ppm to 50 ppm {the approx- inate solubility of iron in bismuth at 650°C during the test. The tungsten. and molybdenum concentratlons in the blsmuth remalned below 20 ppm and o 10 ppm, respectlvely, throughout the test.~ After completion.of the agitastor test, the eqnipment'was disassembled for inspection (see Fig. 15). During operation; the vessel and agitator' --had been maintained at 500 to 650°C for a perlod of 1150 hr, and salt and ‘blsmuth had been present in the system for 1005 hr. Vlsual examlnatlon revealed no evidence of corros1on. " The portion-of the agltator submerged in the’ blsmuth phase had been wet by the bismuth. Also,-droplets of salt and-bismuth could be seen clinging to the agitator shaft at points above | the salt surface;' The upper portions of the shaft and the drive unit were covered W1th~a ‘black material that may have fbrmed as. the result of decom9031tlon of oil from the shaft seals. ' - The tungsten coatlng on the 1n51de of the vessel wes examlned by T Vmembers of the Metals and Ceramies D1v131on. The nickel plate and_tung— ‘sten coating were found to be intact and adherent in two_samfiles taken ( » ) ] ) » n \ Tab1e>6;: Analyses‘of\Salt'and Bismth Samplesah ... Taken During the MTE-3 Agitator Test L » om ( Operafiingg.f | ' Max;.Agitétor» o Bismuth Conc. Timeb . ~ 'Speed ~ in Salt Phase (hr) o rpm) (ppm) 246 30 o u ws . s0 39 @ 0 a; . Nickel Comec. in Bi Phase +(ppm) 500 500 1000 Iron Conc. " in Bi Phase {ppm) *20 , ‘50 ‘aAll samples fiefe unfiltered, - “bTheagitatér.had been operated continuously for at least speed before salt and metal phase samples were taken. 2l hr'at”thé“indicatedfl Fig. 15. Agita.tor After Completion of Test. ~ PHOTO 1099-7) * hs_ from the main vessel. However, cracks in the coating were noted in’ 'several_placés, and ‘there was evidence,that°bismuth'had_penetrated these cracks and had5attacked the nickel"suhstrate;' The cracks are thought to be the result of thermal cycllng since the coefficients of thermal expans1on of the tungsten coatlng and iron vessel are consid- erably dlfferent- In samples taken from the 1/4-in. drain line at the bottom of the vessel the tungsten coat1ng was not in contact ‘with the metal substrate and the nickel layer was absent. The carbon steel was also attacked to a depth. of about 2 mlls in that area. Although the nlckel and tungsten coatlngs were falrly adherent 1n the 1/b-in. sample - 1line attached to the side of the veéssel, numerous cracks were found in thertungsten'cOating It was concluded from examlnatlon of the vessel that complete protectlon of a vessel of this type from exposure to bls- muth by tungstenccoatlng Wlll be difficult because of the ‘tendency of such a coating'to»crack' However although cons1derable dissolution of the nlckel coatlng had occurred attack on the carbon steel base metal was relatlvely mlnor 1n the samples examlned The nlckel aluminlde coatlng on the exterlor of the vessel appeared to be in- good condltion after the test however, the vessel was not held at elevated temperature for a . suff1c1ently long perlod to determlne .whether the use of the. stainless steel coatlng under the nlckel alum1n1de on half of the vessel was beneflcial. '3_N6g- 6. DISTRIBUTION OF RADIUM BETWEEN LiCl AND Li-Bi SOLUTIONS E. L. Youngblood:) | L E. McNeese Redium is present at tracer levels in metal.transfer proCeSS ‘experi- ments as & decay product of thorium Radium.would‘be expected to have distribution characteristics similar to those for dlvalent rare-earth fission products (Sm end Eu) and alkaline-earth fission products (Sr and Ba). Thus, it is of interest to obtain information concerning the behavior | of radium in metal transfer systems. .~ Deta relstive to the distribution of radium between molten LiCl and lithium-bismuth solutions conteining from 13 to 35 mole % lithium, obtained . during metel transfer experiment MIE-2, were reported previously.12 After the completlon of metal trensfer experiment MTE-2, = portion of the Li-Bi solution from the experlment (containing radium) vas diluted'W1th bismuth . end contacted with purified LiCl gt 650°C in order to obtain additional distributiOn data for radium et lower concentrations of lithium in bismuth. " These data ere discussed in the remainder of this section; 6.1 Description of Eqnipment The distribution coefficient measurements were made in 8 12-1n.—hlgh vessel constructed of 2-1/8—in.-diam carbon steel tubing. A thermowell ,and e gas-1lift sparge tube, aiso constructed of carbon steel, were installed infthe vessel for temperature measurement and for contacting the salt and metel phases. The carbon steel vessel containing the LiCl and Li-Bi phases 'was'enclosed'in e heated, 4-in.-diem stainless steel vessel which was 'fmaintainedcunder an argon atmosphere. Semples of the salt and metal phases ‘could be takenby the method described previously.la, Before the LiCl and - the Li-Bi solution containing radium were charged to the system,. the » carbon steel vessel and bismuth were contacted with hydrogenwat 650°C for ~ 12 hr to remove oxide impurities. The LiCl was purified in & separate l'vessel by contact with bisnmmh that had been saturated with thorlum at 650°C. o Y o L by the Ac after a decay period of about 2h hr to ensure that the (half-life, 6.13 hr) was in secular equilibrium with the R ' Inltially, 272 h g of Ll-Bl solution from metal transfer experlment MTE—2 was charged to the system along with h8h 0 g of purified bismuth and 97.0 g of purified LlCl ‘The system‘was then heated to 650°C and maintained at that temperature”durlng the subsequentroperations. Samples ‘vere taken periodically of the LiCl and Li-Bi phases for determination of the radium concentrations;in the phases.' The radium content of the samples was determined by counting the 0.9-MeV gamma radiation emitted 28 228 228 | 6. 2 'Experimental Results _ Samples of the LlCl and Li—Bl phases vere taken 23 hr and 70 hr after the temperature of the system had reached 650° - The concentration of llthlum in the Ll-Bl solution vas. then lowered by the addition of 193 0 g of purified bismuth and additional measurements were made over . & period of 408 hr. Distribution coefficients were calculated from the data obtained. "The results are given in Table T In prev1ous studies of the distribution of materials between fluorides, 19, 20 chlorides or bromides, and bismnth solutions, Ferris et al have determlned that dlstrlbution coefficient data can’ be correlated in terms 1of the lithium concentration 1n the bismuth phase ‘according to the follow1ng relation o A , o _ | R _les.DM;-- f;,n lgDy tlgky, o O) 'where_*. _ L e . f S S | 'D",s_zdistrlbution ratlo for material - - ég,XM(Bl)/XM(salt)";” : 'fo(5i) '="concentration of material M in bismuth phase, mole fractlon, - fi?'xfi(salt}:-? concentration of halide of material M in halide salt mole | e ;_"ffraction, _ : L | _,1n*{=“_valence of material M in. halide salt con D, = distribution coefficlent for 1ith1um,.” T Ll Su o 2constant dependent on materlal M. o Distrlbution coefficient is deflned as the ratlo of the mole fraction of radium in the metal phese to the mole fractlon in the salt phase at 'equilibrium . Teble 7. Data for Radium Distribution Between LiCl and L Lithium-Bismuth Solutions at 650°C EQuilibrafiibfi e - _ Redium Content . ~Lithium Cone. LiCl - Li-Bi 8 - Time - -in Bismuth Phase Phase ' - Phase Distribution {(hr) (mole fraction) (counts min-1 g~1) (counts min~1 g-1) Coefficient. 23 0.05 ~ . 240 . 76k 0.15 70 0.0k 3 19.6 - 803 - 0.12 B 193.0 g of bismuth added to the Li—Bi:phase o - 96 ©0.032 122 . 736 0.08 239 © 0.035 13.0 T84 ' 0.08 08 0.3 1y 82 007 C O » n A - ‘where L9 'Thus, at a glven temperature a plot of the logarithm of the dlstrlbutlon COfolClent for radlum Vs the logarlthm of the mole. fractlon of lithium in the blsmuth phase should give a stralght line hav1ng a slope equal 'to the valence of radlum in the LiCl phase. Flgure 16 shows radium dis- 'trlbutlon data from thls experlment and previously reported data from metal_transfer experiment MEE-E, along with the line hav1ng,a slope of 2-that best fits the data. It is seen that the distribution data can be correlated qulte satisfactorlly by assuming that radium is divalent in- the LlCl phase. It should be noted that only about 30% of the distribution data from }metal transfer experlment MIE-2 is based on LiCl samples taken from the Li-Bi alloy container, while the-remalnderrls based on samples teken from | the main LiCl pool. We ‘believe that the LiCl and Li-Bi phases in the Li-Bi ‘container were essentlally at equlllbrlum at all tlmes ‘however, during __the early stages of the. experlment the main LlCl pool and the Ll—Bl phase would not have been at equilibrlum.W1th respect to the dlstrlbutlon of radlum., For this reason the data based on LiCl samples from the Li-Bi alloy contalner and from the maln LiCl pool during the latter part of ~ the experlment were weighted more heaw1ly in correlatlng the information then date based on LiCl samples taken from the maln 1icl pool early 1n ' the experiment. The distribution coeffmclent data for radium at 650°C 1n the L101—-L1—B1 system.can be represented by the folIOW1ng relation: 1og D'—'2 1og N Li +¢1..-7-57 . - A 6 radlum dlstrlbution coefflcient and M the mole fractlon of 11th1um in the blsmuth phase. A comparlson of the relation summar121ng the radlum dlstrlbutlon data 20,21 -i“'and prev1ously reported dlstrlbutlon data” "’ for Sm, Eu, 'Ba, and Sr is ”7r:jshown in Fig 17 As expected the dlstributlon characterlstlcs for | _[radlum are qulte szmilar to those of the dlvalent rare—earth and alkallne— p;fearth flss1on products, in fact the data for radium and barlum are almost : 1dent1cal. ' ' ' ORNL DWG TI-67RI 10 T T T T TTTT T T T T T T v T AL i 7k * LiCl from Main Pod - 4 LiCt from Li-Bi Cup @ Determined by Dilution of Li-Bi o7 ‘osh 02 RADIUM DISTRIBUTION COEFFICIENT 007 Ll 4 1.1 0051 003 00! 1 L0 gt ) a4 ] Ll 1 b)) ] 1t o1 jo11d 000 0002 0003 00050007 OO 002 003 005 007 Qi 02 03 05 07 10 MOLE FRACTION LITHIUM IN BISMUTH e e e bt a2 ot S S e Sl e A 8 S e ) LA o1 R Fig. 16. Varietion of Radium Distribution Coefficient Between Molten LiCl and Bismuth Phases wrl;h Changes in Concentration of Lithium in the B:Lsmuth. s e e -} - ¥ L) ol : om. WG 73-796. v - ¥ L T Iray . T ¥ T 770 o N T CISTRIBUTION COEFFICENT o e T rrrrr 007 0.03 ek k4 0.03 - 002+ 0.0l . i i3 v afs i i 2 2 13 .1 131 L 1 4 2 2.1 000 002 0003 0005 OC7 G - O@ QO3 005 007 O 02 03 as or 10 ,uos.e rmcfio« LITHIUM IN BISMUTH | ' , Fig. 17. Variation of the D:Lstrlbutlon Coefficients for 8r, Ba, Ra, Eu, and Sm in the LiCl-Bi Alloys with Changes in the Concentra.tion of thhium in the Bismuth Area.. 52 7. DEVELOPMENT OF MECHANICALLY AGITATED SADT—METAL CONTACTORS H 0. Weeren . L. E. McNeese As refiorted previous}y,22 & program.hes been ihitiéted for the'develop- ment of mechanically agitated salt-metal contactors as an alternative to packed column contactors presently under considerationrfor'MSBR processing systems. After brief experimentation-with several centactor types using mercury and aqueous solutlons, it was dec1ded that the Lewis type contac- tor 23 Qh has the greatest potential for achieving acceptable mass transfer rates with minimum dispersion of the salt amd metal phases; This is an 1mportant factor since entralnment of bismuth in processed fuel salt that ‘1s returned to the reactor cannot be tolerated. The Lewis contactor has two agltators ~— one. in the salt phase and one in the-bismuth phase — that are located well away .from the salt-metal interfece. These agltators ~are operated in a manner such that the phases are miXed as vigorously as possible without dlspers1ng one in the other. A review of the 1iterature23 revealed the existence of considerable data concernlng mass transfer coefficients in mqueous-organic. systems in Lewis cell contactors having agitator dlameters of 2 to 4 in. and 1nd1cated that the mass transfer coefficient is strengly.dependent on both the speed ‘and the diameter of.the agitator. However, before Lewis cell cOntacters _ can be designed amd evaluated for salt-metal systems, it will be'necessary to obtaln data for larger contactors, as well as hydrodynamlc and mass transfer rate data for systems having phy31cal propertles that more closely resemble the salt-bismuth system. | | During'this report period, data were obtained on (1) the maximum agltator speed that can be used with a mercury-water system before entrain- ment of'water in the mercury is dbserved, and (2) the rate of circulation of mercury between the two compartments of a stirred-interface contactor 'of the typelbeing considered for the metal transfer process.15 Results from these studies are summarized in the remainder of this section. »y 53 7.1 Studies for Determination of Limiting Agitator Speeds Preliminary tests were carried out in contactors of several sizes ,and:With,different,agitator configurations'in order to determine the factors that will limit the agitator speed in Stirred—interface contactors. - The contactors used in these tests contained two compartments and were of the typefshown'Schematicaliycin Fig. 18. Aqueous solutions and mercury or a low-melting alloy were used to simulate molten salt and bismuth during the studies. It was found that the common factor that limited the agitator.speed.was the transfer of water between the two compartments via entrainment in the circulating metal'phase.__For a given contactor and agitator confignration this‘phenomenon was found to begin at a defi- nite agitator speed below thls speed no entralnment was observed The tests were carried out in compartmented cyllndrlcal contactors having dlameters of 5 5 and 10 1n.,and in & compartmented rectangular contactor measuring 12 X 2l in. In each caSe, the vessel contained no baffles and a single fourAbladed paddle was used 1n the metal phase on one 51de of the contactor. The 11m1t1ng agltator speed was found to be essentlally 1ndependent of the 51ze and shape of the contactor vessel but strongly dependent on the dlameter of the agltator.- Data obtalned durlng these studies are | summarlzed in Fig. 19, where 1t is seen that the allowable agltator speed is 1nverse1y proportlonal to the 1. h3 Pcwer of the agltator diameter. %5 Since the Lew1s correlatlon 1nd1cates that the mass transfer coeffie- | c1ent 1s dependent on the agltator dlameter to the 3. 7 power, it appears ‘that at speeds sllghtly below the llmltlng agltator speed ‘the mess transfer coeff1c1ent will be dependent on the agltator dlameter to the 0 ol power. Thus, 1t should be advantageous to operate a contactor havzng the largest 'vpossible agltator dlameter.s_r" It should.be noted that the datsa shown in Flg 19 are valid only for - the operatlng'condltlons under'whlch they were obtalned. ‘The use of fbaffles, canted agitator blades rather than straight blades, an agitator "~ on each sxde of a contactor, or other changes in. the cell design could make a con51derable dlfference in the llmlting agltator speed. The 54 * ORNL DWG. 73-2545RI BISMUTH PARTITION BETWEEN CELL COMPARTMENTS GA.P FOR BISMUTH FLOW/ - Proposed Contactor Design for Metal Transfer Experiment. Fig. 18. -~} 0‘@0 'BLADE DIAMETER (in) N o1 o CYLINDRICAL CONTACTOR 5.8 IN. DIAM. A CYLINDRICAL CONTACTOR 10 IN. DIAM. + RECTANGULAR CONTACTOR 12 in. x 24 in. ) : L 1 i ) | ‘ 1 0 — 30 50 70 00 200 300 §00 700 - LIMITING AGITATOR SPEED (rpm) Fig.f:l9.' Correlation of Limiting Agitator Speed with Blade Diameter in Several Contactors. : . - | ss 56 following observations relative to this_poifit were noted dfiring the studies: T | R | | (1) The limiting agitator speed was relatively independent of the,#ertical'position of fhe paddle in the mercury phase as long as the paddle was-located well-below_the mércury;water ~ interface. However, when the paddlg was located near the _interfaée;;the'limiting speed was decreased appréci&bly. (2) The limiting agitator speed was apprecisbly fiighef fihéfi the . bismuth-phaSe in each compartment.of:the-confiactor was'agitated than when the bismuth in only one compartment fias'agifiatéd. | For example, the difference-in limifing agitatérfispeeds for a 3-in.-diam agitator blade was sbout 60 rpm. The limiting agitator speed was esséntiallj waffected by the degree of ' agitatiéh of the,a@ueous'fihase. | | | (3) The size of the opening below the'pértitign that separated the two contactor compartments could be increased‘ffom 0.25 in. to 0.75 in.lfiithout appfeCiablj affecting the limiting agitator | épeed; On the other hand, the‘limifiing agitator speed was reduced considersbly as the separation distance was increased above 0.75 in. (L) The use of baffles or the location of the agitator at an off- center pésition increased the limiting agitatbr speed. This éffect‘was more important with small-diameter agitfitors, whefe: the variation in limiting agitator speed was about 30%, then with large-diameter agitators, where the differenée was ohly about 10%. - - (5) When the agitator blades were canted_father than being vertical and when the direction of rotation was such that the sgitator lifted the mercury phase, a considerably higher agitator speed could be axtained without entrainment of_water bétweenlthe two compartments of the cbnfiéétdr. TTherlimitihg7aéitatof speed under these conditions was about;twice that shown in Fig. 19 for both 1.5~ and 3-ifi.—diam agitatoré, but was only about . ax ST - 10% higher'then=the value showhrfor‘a O-in.-diam agitator. The llmltlng agltator speed was found to be affected signif- 1cantly and unpredlctably by baffles. 1n the contactor. A test was carrled out us1ng 8 1ow—melt1ng alloy (Cerrolow 105) and water 1n a heated contactor hav1ng a diameter of 5 5 1n in order to determlne the effect of changes in the propertles of the llquid phases on the contactor performance. The alloy (h2 9-21. 7-8.0-5.0-18.3-4.0 wt % B1-Pb~Sn—Cd—In—Hg) has 8 speclflc gravity of 8.1 and a llquldus tem~ 'peraturehofpabout738°c. At an operating temperature of 60°C the llmltlng ~ agitator speed was essentlally 1dent1cal tovthe llmltlng speed observed 'with"the mercurydwster'syStem. Thus, it appears that the limiting agitator speed is not hlghly dependent on the dlfference in den51t1es of the two 11qu1d phases. Thls observatlon increases our confldence in predicting the performance of a blsmuth-salt system at 650°C by extrapolatlng data obtalned W1th a mercury—water system._ It Was concluded that these data constltute suff1c1ent information for des1gn1ng the salt-metel contactor for experlment MTE-3 and for suggest1ng the 11m1t1ng agltator speed to be used Wlth molten salt and 'blsmuth 1n the contactor at 650°C. Tt 1s believed that entralnment of salt in the blsmuth w1ll oceur | et essentlally the same ‘agitator speed - a8 Was observed with the mercurydwater system (300 rpm) and that experi- ment MI'E-3 should be operated 1n1t1a11y w1th agltator sPeeds well below “this value.. 7 2 Determlnatlon of Mbtal Flow Rate Across Contactor Part1t1on ProPer operation of the salt-metal contactor proposed for use w1th 7 ’,metal transfer experlment MTE—3 w1ll requlre a blsmuth c1rcu1at1on rate °~T;;of 0. 5 11ter/m1n or: hlgher between the two parts of- the contaetor.. Two “,tests were carrled out W1th agltator speeds of 195 rpm - to obtaln data _ :'fiffrelat1ve to thls p01nt.k;,.“' In the flrst test the agltator con51sted of a four4bladed paddle hav1ng vertlcal blades, in the second test the blades were canted at h59 1n a manner such that the mercury was llfted toward the metal surface. 58 During each test, only one side of the contactor cell was agitated}' The experimentalrtechnique'consisted of temporarily preventing'circulationu of mercury between the two halves of the contactor, establlshlng a temper- - ature dlfferentlal between the mercury pools in the two compartments, and allowing resumption of the circulation of metal between the compartments. The rate of change of_the‘temperature_bf.the-mercury‘in one compartment was then measured in order to determine the rate of mercury flow between “the compartments. 'Results of previous tests had shown that'the rate af" conductlve ‘heat transfer across the contactor partltlon and the rate of heat loss to the surroundlngs were negligible as compared with the rate ,of convective heat transfer resulting from c1rculat10n of the mercury phase g between the two. compartments._ ' A mathematlcal analysis was carried out to aid in'interpretation of ‘ the experlmental measurements made for the purpose of determlnlng the ) 'rate of flow of mercury between‘the two contactor compartments. In 'making the analysis,-the following assumptions were'made:' '(l)h Only mercury is present in the'contactor. {2) Equal-quantities of mercury are present in theftWO'compartments; (3) The mercuryrin compartment 1l is initially at temperature TH’. and the temperature in compartment 2 is at temperature TL,' (W) Mercury.circulates between the twofcompartments'ata constant | rate. ' ' | A heat balance on the mercury in compartment 1 ylelds the relatlon ar, S 3 Vpcp-ag—=rpcr—rpcprl,_ S where 'V = volume of mercury in each compartment, cm3, ~ p = density of mercury, g/cm s 'CP = heat capacity of mercury, cal/g-° o Tl = temperature of mercury in compartment 1l at time t °c, T2 = temperature of mercury in. compartment 2 at time t °c, t = time, sec,f F = mercury flow rate5'g/seC. 4 wl ‘relation '_59h_ - 'Since the'rate at Whichrheatjisrexchanged{with the surroundings is negli- gible,ra:heat'balance_on.the mercury in hothncompartmentSVyields the VeC T, S+ VC T = VoC T '+ S (11) S i . . . . ZiceiF . _ - L - o i T 7 ey [gmd Cy * R eeas]s G2 CTer T - N J.%,. =NJd.%=0, (14) L el 1 i Sl 1 i ) ‘ . o 7 i i Cei 'sr Z1/Z | : o c . [c J =Q; » - - (15) er ' ‘ 1) 8 > . 63 Iogagpee— i ) where | J_=_flux of transferrlng component D = dlffus1on coefflclent of transferring component § = thickness of" solvent or electrolyte fllm adjacent to solvent- lelectrolyte 1nterface, | ' | ‘C;= concentrat1on of component 1n solvent or electrolyte, = concentratlon of component i in the solvent phase, C ='concentratlon of component i 1n the solvent phase at the solvent- | 7electrolyte 1nterface, | ' = valence (electrlc charge) of 1on, = electrlc potentlal 1n electrolyte fllm, = Faraday constant gas constant = absolute temperature and o> ¥ w om e N F o o 7=;eqn1l1br1um constant. The subscrlpt i refers to the . 1th transferrlng component the sub- scr1pt r refers to a reference transferrlng component and the subscrlpts e and 8 denote the electrolyte and solvent phases respectlvely., S1m11arly, the rates et whlch components transfer between a molten -’rl:salt and a blsmuth phase are deflned by Eqs. (1) (5), and the relatlon' - "}-j el 1 R’I‘ grad 4, _'-,—_ _ ;.. ST (17) : i €1 ) - /.i ‘ In present1ng the calculated mass transfer rate data for the molten- salt——blsmuth and the aqueous—organic systems, the transfer rates will “be normalized to the transfer rate that would be observed under. the same conditions in the absence of electric potent1al effects. ‘The rates at 5.Whlch components transfer between an electrolyte and a solvent phase in the absence of electric potential effects are’ deflned by Eqs. (1)-(5) and the follow1ng relat1on. Jei = —fDei_grad Cei | S (ls) 8. 2 Calculated Mass Transfer Rates for the Case of Blnary | Exchange w1th Uniform Bulk Concentratlons The extent to which an electric potentlal gradlent alters the rates at Whlch components transfer between electrolyte and solvent phases . having uniform bulk concentrations depends upon (1) the valences of the .transferrlng and nontransferrlng ions in the electrolyte phase, (2) the dlffu51on coefflc1ents of the transferrlng components in the electrolyte'r and solvent phases, (3) the resistance to transfer of components through . the electrolyte and solvent films ad}acent to the electrolyte-solvent | “interface, (4) the equilibrium constants for the transferring components, (5) the relative concentrations of the transferring components in the electrolyte and solvent phases, (6) the concentration of nontransferring ions in the electrolyte phase, and (T) the behavior of the nontransferring ions. Because of the large number of variables inwolved, it is not possi- " ble to portray in & simple manner the complete solution to the set of -equatlons that defines the effect of the electr1c potentlal gradlent on the rates at which components transfer between therelectrolyte and sol- vent phases. Instead' results for selected cases'involviné two trans- »ferrlng components will be given in order to show the 1mportance of ther various factors.‘ In each case, the rate at whlch a component,transfers- between a solvent and an agueous or molten salt phase 1n the presence of _an eleectrice potentlal gradlent Wlll be compared to the rate at which the component would transfer in the absence of an electrlc_potentlal gradient. - In the remainder of this_chapter, the term»relative,flux value‘(RFV)'will:' be used to denote the ratio of the‘flux of a transferring cOmponent-in “ 18 65 the presence of an electric potentlal gradlent to the flux in the absence.' - of an electric potentlal gradient | In all cases 1nvolV1ng blnary exchange w1th uniform bulk concentra- : tions, the valences of the transferrlng and nontransferring ions were Jassumed to be unity. The concentrations of components 1l and 2 in the solvent phase were assumed to be 0.1 and 0. 05 g-mole/cm3, respectiveLy, and the concentrations of . components 1 and 2 in the eleotrolyte phase ‘were assumed to be 0.05 and 0.01 g-mole/cm3. 8 2.1 Effect of Diffusion Coeff1c1ents of Transferring Ions in Electro- ' lyte Phase The effect of the diffusion coefficients of the'transferring'ions in the electrolyte phase during blnary exchange is shown in Fig. 20, ‘where the RFV (for either component) is given as a function of the ratio ~of the diffu51on coefflclent of component 1l in the electrolyte phase to . the diffusion coefficlent of component 2 in the electrolyte phase. In ' obtaining these results, the equlllbrium constant was assumed to be 'unlty, and the thickness of the solvent film was assumed to be negligible (negllglble re51stance to transfer in the solvent fllm) It should be ;noted that the RFV is affected by the relatlve values of the diffusion coefficients for the transferrlng spec1es except in the case where the dlffus1on coeff1c1ents (as well as the valences) are equal The effect of an electrlc potentlal gradient on the mass transfer rate is greater [ in the case of a uniform concentration of coions 1n the electrolyte ~ phase (molten salt solution) than for the - -case of a nonunlform concen- "'-_tratlon of c01ons in’ the electrolyte phase (aqueous solution) because the mobile c01ons in the aqueous electrolyte dlstrlbute across the y 1e1ectrolyte film 1n a manner'whlch suppresses or reduces the effect f-rof the electric potential gradient. If the ratio of the dlffu51on : gcoefflcient fbr component l to that of component 2 is 0 25, ‘neglect of ;-'the effect of an electric potential gradient would result in errors in the calculated mass transfer rate of 23 and 507 for aqueous and molten\ salt electrolyte phases respectively 66 a ORNL OWG 74-677 1or T 1 7 T VIiFETl T 1 T T V1T 7+ N s} i 3 - S [T~ MOLTEN SALT =L ARG = '[C acueous N > | soLuTiOoN 3 - 3 o7 " MO L 0.5 SALT 2 p— L < Jo.s_ — l o 0.1 A ] i 1 P 1 11t | 1 ] | I I | 0.1 0.3 05 07 | 3 S T 10 RATIO OF DIFFUSION COEFFICIENTS OF TRANSFERRING ELECTROLYTE (0, /D,, ) Fig. 20. Effect on Relative Values of Diffusion Coefficients for - Transferring Ions in Electrolyte Phase on Relative Flux Velue. The resistance to mass transfer in the solvent phase was negllgible (zero film thickness in solvent phase). A » 8.2.2 Effect of Ind1v1dual Mass Transfer Coefflclent in Electrolyte _f Phase The effect of the 1nd1v1dual ma.ss transfer coeff1c1ent in the elec—, trolyte phase is shown 1n Fig. 21,where the RFV. (for either component) is glven as a functlon of the ratio of the 1nd1v1dual mass transfer coeffi~- cient for component l in the electrolyte phase to the 1nd1v1dual mass rtransfer coeff1c1ent of component 2 in the electrolyte phase. The mass. transfer coeff1c1ents for components 1 and 2 1n ‘the solvent phase were assumed to be equal to the mass transfer coeff1c1ent of component 2 in the electrolyte phase ' The equlllbrlum constant was assumed to be unity. The error caused by neglect of the effect of an - electrlc potentlal gra- ,dlent is greater 1n the case of the molten salt electrolyte than in the case of an aqueous electrolyte and becomes s1gn1f1cant for the case where - the 1nd1v1dual mass transfer coefflclent for component 1l in the electro- 'lyte phase 1s small relatlve to the. other 1ndiv1dual mass transfer coef- flclent values.z- 8 2. 3 Effect of Ind1v1dual Mass Transfer Coeff1c1ent 1n Solvent Phase The effect of the 1nd1V1dual mass transfer coefficient 1n the sol— vent phase is shown’ in Flg.,22 where the RFV (for either ‘component ) is -given as a functlon of the. ratlo of the 1nd1v1dual mass transfer coef- 'f1c1ent for component 1 or 2 in the solvent phase (assumed to be equal) e to the :|.nd1v1dual mass transfer coeffn.clent for component 2 in the '?electrolyte phase. In obtalnlng these results, 1t was assumed that the 1flequ111br1um constant was unlty and that the ratio of the 1nd1v1dual mass’ fj_yfltransfer coeff1c1ent for component 1 1n the electrolyte phase to that ':for component 2 was equal to 5 When the transfer coeff1C1ent 1n the solvent fllm 1s very hlgh as fd on the rlght s1de of Fig. 22 the transfer rate 1s controlled by the '1‘re51stance in the electrolyte fllm As 1n the earller cases, ‘the effects ' ;of an electrlc potentlal gradlent are more ev1dent in the case of a , - -molten salt electrolyte than 1n the case of an aqueous electrolyte At very high values of thersolvent fllm transfer coeff;clent, mass transfer resistance'is'solelylin.the"electrolyte film, and the influence of an 68 " ORNL DWG 74-68I 10p r T T T T TT] | T T T T T 7.0f | o S N 3 5.0} | | - . - o R 3.0} ' A - . ' - AQUEOUS SOLUTION | MOLTEN SALT RELATIVE FLUX VALUE (J/J,) ° : q 0.1 | 114 oy 1 ! | 1411 o4 03 05 07 | 3 5 7 10 INDIVIDUAL MASS TRANSFER COEFFICIENT (Dei/8 ) - - D/8 Fig. 21. Effect of Individual Mass Transfer Coefficient for Component 1 in the Electrolyte Phase on Relative Flux Value. The A individual mass transfer coefficient for component 2 in the electrolyte phase was assumed to be equal to the individual mass transfer coefficients for components 1 &nd 2 in the solvent phase. : i o e 69 ORNL DWG 74-680RI 1O T r r T T T Tl L i T T T 11Tl T T 1 1T 1rrr1] 7}- ] 51 - 3 - 3| | - | MOLTEN SALT ] : /—’7 AQUEOUS SOLUTION = / 24 a ' . _ > _ | _ 307 . - | - - w Sk - > poen Z — 5 wous- - o 0.l i L 1 L1 o1y i 1 11111 0.1 0.3 0.5 o - 3.0 50 70 10 D. /5 30 56 70 100 INDIVIDUAL MASS TRANSFER COEFFICIENT RATIO (—‘—-‘-’8 ) De Fig. 22. Effect of Individual Mass Transfer Coefficient for Component 1 or 2 in Solvent Phase on Relative Flux Value. The individual mass transfer coefficients in the solvent phase were assumed to be equal. The ratio of the mass transfer coefficient in the electrolyte phase for component 1 to that for component 2 was assumed to be 5. 10 electric-potentlal gradient.approaches_an asymptotic value. As the sol- -vent phase transfer coefficient decreases, the.effects of electric fields decrease since the transfer‘rate.is no longer_controlled by the electro- lyte film. Eventually, for very low values of the solvent film transfer coefficient (left side ofrFig 22), the transfer rate is controlled by the film resistance in the solvent and is not 51gn1flcantly affected by electr1c fields. 8. 2. 4 Effect of Equilibrium Constant The effect of the equ111br1um constant on the RFV is shown in Fig. - 23, In obtaining these results, it was assumed that the individual mass transfer coefficients in the solvent phase were . equal to the individual | mass transfer coefficient for component 2 in the electrolyte phase. It was also assumed>that the individual mass transfer coefficient.for com- h ponent 2 in the electrolyte phase was five times that for.component 1. ) Changes in the equilibrium constant result in changes in the interfacial concentrationS'in both phases. For very small values for the equilibrium “constant, the equilibrium concentrations of the transferring species on the solvent side of the interfacerbecome small relative to,the‘concen— trationston the electrolyte side of the interface. 1In this case, the. resistance to transfer between the phases is principally in the solvent film;tflneelectric potential gradient has essentially'no‘effect,;as shown in the left-hand portion of Fig. 23. For high values of the equilibrium constant, the primary resistance to mass transfer is in the electrolyte film and the effect of the electric potentiallgradient becomes important,, '8.2.5- Effect of Concentrationfof Nontransferring Ions in' the Electrolyte Phase The effect of the concentration of nontransferring ions in_the elec- trolyte phasefdnring,binary exchange is shownfin Fig. 2k, where the RFV - is given_as a function of the fraction‘of the electrolyte concentration that cannot exchange between the phases. In maklng these calculat1ons, the bulk concentratlons of the transferring ions in the electrolyte and ' solvent prhases remained constant at the values given earlier, and the i, i_‘ 'fl\ ORNL DWG 74-675 T.0F 5.0} 3.0 (474.) o o 1 RELATIVE FLUX VALUE 0 < | ] MOLTEN SALTY | i i AQUEOUS SOLUTION 1 1 i i - 0. 0. - 0.3 0.5 o.7 1.0 3.0 50 720 10 EQUILIBRIUM CONSTANT (Q) 30 50 70 100 Fig. 23. Effect of Equilibrium Constant on Relative Flux Value. The individual mass transfer coefficients in the solvent phase were assumed to be equal to the individual mass transfer coefficient for component 2 in the electrolyte phase. The individuel mass transfer coefficient in the electrolyte phase for component 2 was assumed to be five times that for component 1. ' L T2 ~ ORNL DWG 74-679R — — | ol > | - | N | =< o g | ™ z o > W RELATIVE FLUX VALUE (J/J,) | J | l | | | 2 4 .6 .8 1,0 Ce(Trans) /ce('l_'qiul) g0 o Fig. 24. Effect of Concentration of Nontransferring Ion on Relative Flux Value. » 73 bulk concentratlon of a nontransferrlng component 1n the electrolyte ‘phase was varled in a manner whlch.produced nontransferrlng ion fraction values ranging from O_to_0.5. The valence of the nontransferrlng ion was +1. As”in'the case of the coion, -the value of the diffusion coef- ficient of the nontransferrlng ion is un1mportant The nontransferring ion behaves as a supporting electrolyte (as does the c01on) and suppresses the effect of the electric potentlal gradient. ‘As shown in Fig. 24, the RFV value decreases steadlly as the con- centratlon of the nontransferrlng 1on (relatlve to the total electrolyte ,v concentrat1on) is increased. As in the earlier cases, the effect of the electric potential gradient'is more prononnced inrthe case where the electrolyte is a molten salt solutlon than in the case where the electro- lyte is an aqueous solutlon. 8 3 Calculated Mass Transfer Rates in an Extractlon ' Column Hav1ng Nonunlform Bulk Concentrations ~ The rates at Wthh components transfer between the solvent and electrolyte phases in a packed column 1nvolve many independent varlables, and no attempt w1ll be made to show the effect of each of these varla— bles. Instead, results wxll be presented for two operatlng conditions. The first is & 51mple blnary exchange process in whlch the electrolyte that enters one end of the column contalns component l but not compo-— . nent 2, whlle the solvent that enters the opposite end of the column conta1ns component 2 but not component 1. Results - are presented for the -cases of an aqueous electrolyte and a molten salt electrolyte, and for the case where there is- no effect of an electrlc potential gradient. The second operat1ng condltlon con31dered corresponds to an actual -”reductlve extractlon operatlon 1nvolv1ng a molten fluorlde mlxture and | llquld blsmuth contalnlng reductant. ‘In thls case, four cat1ons are o+ o b 3+ ) + b+ con51dered (Ll 5 -BeTy Th' and U= However, only Li. ,_Th- , and o U3+ are allowed to transfer between the electrolyte and solvent phases. fResults are presented for the case where the electrolyte is-a molten salt and for the case where there is no effect of an electrlc potentlal . gradient. Th 7'8;3.1 Binary Exchange - - Table 8 shows the operatlng cond1t10ns for the case of blnary -‘exchange in a packed column - The valences of the transferrlng ions and the coion ‘were assumed to be unity. Thesequlllbrlum constant for the ' transferrlng ions was assumed to be unity. - It was assumed that the _ column was of such length that the total 1nterfac1al area between the 'psolvent and electrolyte phases was’ 2000 cm2¢ Table 8. ‘Operating Condltlons for the Case of Blnary Exchange in a Packed Column ' Quantity . Phase | . , Electrolyte . -~ - Bolvent Flow rate, cm3/sec . - - 50 . 50 Film thickness, cm - 0 2x10t . 1x107 Inlet concentrations, g—mole/cm3 | o Component 1 S ' 2 x 1073 -0 Component 2 | S | 0 | : 1l x 10-3 ~Diffusion coefflclents, cm /sec | | | - Component 1 - "1x107° o 2x 10-5_ Component 2 | o 1x 10-6 | b 1072 In making the calculations, the column was divided into 50 axial increments of equal length and the electrolyte film in a‘given inerement of column length was divided into 20 1ncrements of equal thickness. 'Calculated values for the concentratlon proflles for the two. transferrlng components are shown in Figs. 25-27 for the cases where the electrolyte, phase consists of an aqueous solution or a molten-salt solution and for the case where the electric potential'gradient has no effect. The con- centratlons of the transferrlng jons. have been normallzed to their maxl- mum - values to facllltate presentatlon of the results. It is apparent from Flg.- T that an electric potentlal gradlent has .the effect of decre351ng the rate at which materlals transfer between ®) €| 75 S ORNL DW6G 74-672 AELECTROLYTE PHASE OSOLVENT PHASE o S 9---1---1 “ofijfi'rfi[l1rl l_]_‘ll1]lj‘f Tlfll - - 0.8}~ -~ L1 z/L Coi T o b i b g L v oo vl v 0 11 ¥ ‘0 02 04 _ 06 08 10O | T T e HZI/CMAX o - Fig. 25. Concentration Profiles for the Case of Binary Exchange - in o Packed Column Where the Electrolyte Phase Is en Aqueous Solution. 76 - ORNL DWG 74-676 O ELECTROLYTE PHASE A SOLVENT PHASE 9--=)-=-1i PP 0T 1T T T T T T T T T TR 0.8¢ 4 0.6 Z/7L o4} N\ 0.2} [ - o‘llllllJliLllj_lt!ll'lll X D~ 0.2 04 06 0.8 10 CLZ)/CMAX : Fig. 26. Concentration Profiles for the Case of Binary Exchange in a Packed Column Where the Electrolyte Phase Is a Molten-Salt Solution. ¥ 77 ORNL DWG 74-658 O CASE1 A CASE Il + NO ELECTRIC 9-ee|-—I1 ITITIIIT—ll‘r1||"‘,T‘ NO ELECTRIC POTENTIAL- - GRADIENT . - -l 4 Ny N — AQUEOQUS 7 SOLUTION - MOLTEN-SALT SOLUTION - . JIL_I'IIJIIQ_'-_—ll‘."JJl-, | 08 L0 ~ Fig. 2T7. Compariséfi'of Concéntratidn Profiles for Component 2 for ,the Case of Binary Exchange in & Packed'Column Results are shown for the cases where the electrolyte is an aqueous solution or-a ‘molten-salt solutlon, and for the case where_the effects of the electrlc potential T gradient are not. 51gnificant.= 78 the éleétrolyte and solvent phases, and that this effécfi is-mbfe pré— nounced when the electrolyte is a molten-salt solution. For the latter' case, the caiculated Value for the extent to which-component_2 transfers from thersblvent‘to the electfolyte'is seen to be in error by about 24% if the effect of the electric potential gradient is not considered. For vthe.éase where the éledtrolyte is.an a@uedué solution, thegcofresponding error is about 1h4%. 8.3.2 Multicomponent Exchange Table 9 gives the operating conditions for tworcases involving multicomponent exchange in a packed column. The conditions_for these cases correépond'to redfiCtive extracfiion involving the transfer of uranium,.thorium,'and'lithium between a molten fluoride salt ?hase (71.7- 16-12-0.3 mole % LiF-BeF -ThFh-UF ) and liquid bismuth containing reduc- 2 3 2+ 3+ ' + + tant. Although only four cations (Li , Be” , U , and Thh ) are con- sidered here, calculational methods and computer programs which‘have the capability of considering up to ten cations have been developed. - Two elements (Be and F), assumed to be present in the éalt phase as Be? and F , do not transfer to the bismuth phase. In each of the cases, the solvent phase (bismuth) that enters the column contains no? uranium. In one case, the solvent contains lithium (1 mole %) but no thofium; in the other, the solvent contains thorium (1 mole %) but no lithium. The column length is such that the total interfacial area between the molten salt and bismuth phases is 2000 bm3. The calculated concentration prdfiles for the two cases are shown in Figs. 28 and 29. In both cases, the salt and bismuth-phases were -essentialiy in equilibrium with respeét to 1ithium and thorifim'negr:the bismuth inlet. There were no Significant differences between the results obtained when the effect of an electric potential gradient fias considered and when it was neglected. This result is not surprising;~however, since the low concentratibn of feductant (Li and'Th) in the solvent'phase dic- tates that the mass transfer rates are controlled primarily by the resis- tance to transfer in the solvent phase. Also, uranium is & minor compo- nent in the electrolyte (molten salt) phase, and the coions act Asra o, 1 19 o ORNL DWG 74-659 OSOLVENT PHASE AELECTROLYTE PHASE 1O R, 2l . ]—'i"ljrl"_lrrl,]illi]ll'l’lli1 0.9 0.8 0.7 0.6 0.4 0.3 'llll'r"'lllrflllll'lll’f-l‘lllllllll’llff 0.2 " _ll......ll..-."-n—-n-nIl!!.l_..-l-.-..-'-"u.'"i't B, . N * b 0.l ~Js !_! TIT‘I'IITTII - B0 [ - 0 9o 0 a2 iy flb g IJJ_I e b g oo - 0.2 o.-4, 0.6 . 08 —_-‘lfo - o C(Z)/CMAX ' S O Fig. 28 Concentration Prof:.les for U, Th, Li, and Be in the Molten Sa.lt and Bismuth Phases in a Packed Column. The bismuth phase entenng the column contained only lithium at the concentration of 1l mole %. 80 | "ORNL DWG 74-660 DELECTROLYTE PHASE ASOLVENT PHASE - 9--|--2-11 '.ofl111llllfilill|1fi‘lil]—t T 3 L S . | Li T | | Th 0.8}- | [ : | Be I v Li ™ 0.6} J = . ‘N i o4 0.2 o-l . a a9 9 4 4 1 2 4+ . 41 0o 0.2 04 0.6 0.8 .0 C (Z)/C MAX Fig. 29. Concentration Profiles of U, Th, Li, and Be in the Molten Selt and Bismuth Phases in a Packed Column, The bismuth phase entering the column conteined only thorium at the concentration of 1 mole %. b L 81 | Table 9. Operatlng Conditions for the Cases of : Multlcomponent Exchange 1n a Packed Column - . Quantity - o thase ' S s Electrolyte o Solvent -Flow rate;“cm3/Sec | - f; _ ' 7. 50 L B SO Film thickness, em C1x1073 \ 1x10° Inlet concentrations, g—nole/cm3_ | ‘_7 57 R | o 3.745 x 107° 4.6 x 1072 Be - 8.637x103 0 ~Th - o 6.627 x 10 -3 ; 4.6 x 10778 U . B 560 x 10 b 0 i Diffusion coefficlents,ncmg/sec | . | Li B o o 3 x_lO-S‘ 3 x 110-IS Be _7.,nr | - - - —_— | | —— ™ C1x107 C1x1077 v S exao 2 2x107 a'Only;lit'hium,(or thorium) was present in the inlet solvent. supporting electrolyte whlch suppresses ‘the effect of an electric poten- tial grad1ent 8 h Summary A mathematical analy51s has been carrled out and computer programs Eihave been developed whlch allow determinatlon of the 1mportance of the ';electrlc potentlal gradlent that 1s generated in an electrolyte during 'f'the transfer of materials between an electrolyte and a solvent phase. S The effect of the electric potentlal gradlent was found to be of greater 711mportance when the electrolyte is a molten salt’ than when 1t is an . -td"aqueous solutlon., In the latter case, the nontransferring colons redis- tribute in the,electrolyteiphase in a manner whlch-suppresses the effect of the electric potential gradient.-'However, it was shown that 82 751gn1f1cant errors in calculated mass transfer rate values will result - | &"; ,under some operatlng condltlons in both cases.- Two cases 1nvolv1ng reductive extractlon of uranium from a molten fluorlde salt phase into a liquid blsmuth phase contalnlng reductant o l‘ _': v 1nd1cate ‘that neglect of the effect of an electrlc potentlal gradlent probably causes essentlally no errortln calculated mass transferrates- for reductive extraction operations of interest in MSBR processing.- o 83 9. ENGINEERING STUDIES OF URANIUM OXIDE PRECIPITATION M. J Bell* L. E. McNeese. : -Studies‘Of the chemistry of protactinium and'uranium oxide pre- cipitation by Baes et al. have indlcated that ox1de precipltatlon may be an attractive alternatlve process to fluorlnatlon-—reductlve extrac- : tion for isolatlon of protactinlum and removal of uranium from the fuel salt of an MSBR‘26 et Calculations made by Bell and McNeese 1nd1cate that protactlnlum removal tlmes of 3 to 5 days can be obtalned for a reasonable range of flowsheet varlables, and that uranium removal effi- ciencies of greater than 99% can be reallzed with relatlvely few equi- ~librium stages. 28,29 Englneerlng studles of uranlum oxide precipitation are planned in order to- study the kinetics of uranium: oxide precipitation, to 1nvestlgate the size distrlbutlon and settling characteristics of the oxlde preclpltate, and to gain experlence with oxide precipitation systems. During thls report period, an experimental facility was designed and 1nstallat10n of equipment was inltlated The facility is described ‘1n the remalnder of thls section. .9.l"bescriptiOn of.Facility [j A schematic diagram of the major components of the proposed oxide prec1p1tatlon fac111ty 1s shown in Flg. 30 ‘This faclllty will allow _,batch prec1p1tat10n studies to ‘be made in a vessel contalnlng approx- ~imately 2 liters of 72-16-12 mole. % LiF~BeF -ThFh salt thet has an f‘rlnltlal UFh concentratlon of about 0. 3 mole %. 0x1de Wlll be- supplled ,'to the prec1p1tator in the form of a water-argon gas mlxture that will “ be 1ntroduced through a ‘draft tube to promote contact of the salt and ) oxide. The method proposed for separatlon of the salt and oxide phases dmils decantatxon of the salt to a receiver vessel after the oxlde ‘has been | ;allowed to settle fbr a short perlod. The fa0111ty also 1nc1udes a ‘f,system for supplylng hydrogen—HF gas mlxtures that will be used for 'convertlng the ox1des to fluorldes at the conclusmon of an experlment. - ¥Present address: Dlrectorate of- Regulatory Standards Washlngton, " DC 205ks Ha=HF SUPPLY ORNL DWG 72-64 VENT MF £ I REMOVAL SYSTEM | M > | B _ - [ | | o SUPPLY ND PRECIPITATOR VESSEL Fig. 30. Conceptual Design of a Sihgle-Stage Uranium Oxide Precipitation Facility. "8 85 The’ off-gas system 1nc1udes caustlc scrubbers for remov1ng HF from the gas streams in order that 1nformat10n relatlve to the extent of oxide prec1p1tat10n can be obtalned by collectlng the HF evolved by the pre- . c1p1tatlon process. Ball valves on both the feed and prec1p1tator ' vessels prov1de a8 means for obtalnlng samples of salt and ox1de 9.2 Precipltator Vessel De51gn In order to obtain 1nformatlon applicable to the design of the ”prec1p1tator vessel the experlmental system was simulated u51ng sand and water 1n1t1ally, and then iron powder and a glycerolewater solut1on, in place-of oxlde_and salt, a1r was employed_as the sparge gas. A 1- in.-diam draft tube was placedrin:a;h;in.—diam glass pipe, and the ~effects of the distance between the bOttom of the draft tube and the bottom of the vessel, as well as the shape of the vessel bottom, on the degree of mixing of the sollds vere determlned. It was observed ._that only the solids located 1mmed1ately beneath the draft tube were 'dfentralned in the circulatlng 11qu1d and that the system performed most feffectlvely when the bottom of the draft tube was positioned w1th1n a _'few mllllmeters of the bottom of the vessel. When the draft tube was raised as llttle as. l/h 1n. from the bottom, m1xing performance became 'very poor It was also observed that, if the solids were allowed to 'settle around the bottom of the draft tube, dlfflculty was sometlmes - 'experlenced in resumlng c1rculat10n of the solld material The preclpltator vessel that was de51gned as’ the result of these - B observatlons is described in Table 10 and shown' 1n F1g 31. The lower " part of the vessel, 1n wh1ch the salt 1s contalned, is a sectlon of '_h-in._sched Lo 1ow-carbon nlckel plpe. The bottom of the vessel is 'machlned from a h-ln.-dlam 1ow-carbon nickel bar and 1s tapered at . h5° *:gln order to d1rect the sollds under a draft tube that is used to promote contact of the salt and oxlde. The draft tube is constructed of a __f;sectlon of 1-in. sched hO n1cke1 plpe that 1s welded to the gas inlet . fdtube. The 1nlet tube also supports e baffle that is used to prevent ilarge quant1t1es of - salt from belng carrled 1nto the upper part of the vessel. The upper part of the vessel is constructed of & section of 86 h-in.-diem x 1/8-in. plate Table 10. Materials Used for Construction of the Precipitator Vessel - Part No. Item Material 1 ,'Sein. sched-ho X 6-in, pipe o Nickel 2 1/b-in. plate rolled to fit L' Nickel 3 h-.i‘n. sched 40 x 12-1/2-in. pipé "L Nickel L h-in,~diam x l—3/h-1n. circular bar machined "L" Nickel as shown _ - 5 6-5/8—1n.-d1am x 1/2—in‘ plate "L" -Nickel -6 - 1-ih; sched 40 x711-1/2vin._pipe - "L" Nickel T 1/2-in. 20 BWG tubing " _Nickel 8 3/8-in. 20 BWG tubing MM Nickel 9 1-1/4-in. diam x 1/2-in. circular bar ot Nickei’ ‘machined as shown ' 'lO' Calrod heaters - | 11 Insulatiofi—-Q-in. minimum thickness __Fiberfrak 12 5-in.-diam x-1/8-in. plate "L" Nickel 13 5-1/2-in.-diam x 1/8-in. plate "L”'Nickelj 1k "L" Nickel 87 ORNL DWG 72-6853RI [:m-____;_____) ':”’—ofiER&?s cAsiNLETQf”'i - LINE o _ N Y ’ 7 o n 1. .. e e P \ , | , o . SALT } ~ TRANSFER - LINE DRAF T~ RINCN | LY - fFig.”31.J SChematibifiiéérémfiof-the Precipitator Vessélffor a'Single- Stege Urenium Oxide Precipitation Facility. The numbers correspond to the materials list shown in Table 10. - ' 88 6-in. sched .40 ldw-carbpn nickel pipe and serves as & deentrainment' section. Nickel plates welded to the dréft tube in'this_section of the vessel serve as heat shields and flow diverters. The gqs.inlet'tube is’ used to intrddfice either argon;watér mixtures or hydrogen-HF mixtures ihto the vessel and, in additidn, functions as one of the-sampling noz-- éles. The upper end of this tube is fltted with a l/2—1n. nlckel-plated Jball valve through which samples of the salt and oxide are obtalned The vessel is heated with tubular resistance heaters and insulated with © 3 in. of,Fiberfrax;. Nozzles on the uppéer flange of the vessel are uséd forlthermowells,'an Off-gaé line,. and ser?icé lines for'infroducing salt to the vessel. The'nozzleé,ére’déscribed in more detail in Table 11. The feed-and-receiver tank was fabricated from a 19-in.-long section of hein. sched L0 low-carbon_nickel pipe and has heads made from 1/2-in. nickel plate. It also is heated with tubular resistance,heaters.and'is insulated wlth 3 in, of Flberfrax. The equipment is béing ifistalled in a hood in Bldg. 35h1. ' ' - Table'll}j N§zz1e Schéduleffor Prééipitator Vessél . ?:::;;iSepvice ';Numbef of‘N6zz1es ' T Description\ ‘ ‘;‘Materiél Cas and sample ialet A_ Therfipwe11s . | Sait'tranéfer_-" 1/2-in. tube x 1/2-in. pipe ;f1/2-in.vffibe'x‘l/2-in; pipe 1/2-in. tube x 1/2-in. pipe 3/8-in. tubing Monel _Monel Monel LM Nickel 68 90 10. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS R. B. Lindsuer . We have previously described equipment for studying the puriflca-' 30 Initial work with this system. was directed at the measurement of the flooding rates in a 1.25-in.- tion of salt by continuous methods. diam, T-ft-long column packed with 1/4-in. nickel Raschig rings. Flooding data were obtained during the countereurrent flow of molten salt (66-34 mole % LiF—BeFa) andrhydregen or argon.Bl The objective of the present work is to study the continuous reduction of iron fluo- ride in molten salt by countercurrent contact of the salt with hydrogen in a packed column. We have previously reported on two iron fluoride ~ reduction runs that were carried out at 700°C with salt having the composition 66-3h mole % LiF-BeF (ref. 32) and nine runs that were carried out at T00°C with salt hav1ng the comp031tion 72 O—lh 4L-13.6 mole % L1F-BeF -ThFh 33 Tron analyses from samples'taken during the 1ast sefiés of nine iron fluoride reduction runs have shown considér&blescatter_(Fig.'3é),_ Also shown in this figure are the efialyses for ten samples takep.after the last reduction run when the soluble iron (fluoride) concentration ‘should have remained constant. Suspected reesons for the inconsis- tencies are: presence of iron particles suspended in the salt, con- ‘tamination of the sample during ahalysis, and analytical difficulties. at_iow i:on eoncentretions{ Several tests were carried out during this report period in order to resolve these questions; the_results and con- clusions are discussed in the remeinderpoffthis chapter. 10.1 Tests rbr'Determining the Presence of Suspended Iron Particles in Salt Samples ~ Tests were made to investigate the possibility that iron particles, | suspended in the molten ssalt, might:be‘the cause of the occasional high' iron analyses such as those after runs b, 10, and 11 reported previously. 30 _‘It is belleved that any metal particles which were suspended in the salt would deposit on metsl specimens suspended in the,salt.-iTherefore,‘ . O ORNL DWG 74-66) _BATCH SAMPLES AFTER RUN - NICKEL FILTER SAMPLER FLOWING-STREAM SAMPLES DURING RUN : COPPER FILTER SAMPLER ' COPPER DIP SAMPLER (NO FILTER) +%xDO | T 4007 W QL e L 200 o -IOC)-" ~ IRON CONCENTRATION (ppm) = - Qo » ,, 1 4 g i 1 3 4 5 6 T 8 9 10 N0 |53z e Ll Reoucnou RUN NO. = | SAMPLER TEST NoO. e ‘ ‘Fig. 32. Iron Ana.lyses Dur1ng Recent Iron Fluoride Reduction Runs and Sampler Tests.- 16 92 specimens of nickel were suspended in the salt in the feed tank at | varying}depths for 5 to 15 min. Leaching these specimené with HC1 ~dissolved a negligible amount Of iron. To further subStantiate the ; absence of iron partlcles in the salt, four dip (unflltered) samples were taken. The iron concentratlons in these samples ranged from 65 to 75 ppm, which is equivalent to the concentratlon in the filtered samples having the lowest iron content. Such resultsxshow conclu51vely that iron particles are not suspended in the salt. ' 10.2 Tests of Various Sampler Types _ The samplers used durlng the iron fluoride reduction runs carrled out to date have consisted of l/h-ln.—OD nickel samplers to which a porous metal- filter is attached. The sample size is about 1lg. Removal of the salt sample from these samplers (via a steel tublng cutter fol— ‘ lowed by crushing) is somewhat difficult and prOV1des a possible source of iron contamination. Therefore, following the last reductlon,run, both filtered and nonfiitered'samples-were taken in cOppér:samplers, some of which allow removal of a larger {k-g) sample. The éopper tubing 'is,more easily crushed, and samples are less likely to be cdntaminated- with iron during this ‘operation. ,Ayeragés of the iron concentr#tions for the samples taken after run 11 are shown in Table 12. . Table 12. Average Reported IrOn.Concentratiofis. for Various Sampler Types : _ , - Average Iron Conc.. . Number Type of Sampler , (ppm) ~ of Samples 1/4-in.-diem Ni (filtered) 262 2 1/U4-in.-diem Cu (filtered) | 109 2 3/8-in.-diam Cu (filtered) 102 3 3/8-in.-diam Cu'(unfiltered) R | 68 o - e folloving‘conclusions-have been”dravn from these dataf The - (1) 1f partlcles of 1ron are. produced dur1ng the reduction - of iron fluoride in the column, the iron does not remain - suspended in the salt for one pass through the system _'(column, piping, recelver tank). (2) Iron analyses below 100 ppm are unrellable W1th sample - sizes of 1 gor less. | (3) Samples taken with nickel samplers are more subject to - contamination during removal than are samples from copper samplers. . 10 3 Correlation of Floodlng Data Typlcal flow data obtalned in the present 1. 25~1n.-d1am T-ft-long packed column are compared 1n Flg 33 w1th a floodlng ve1001ty corre- latlon developed by Sherwood 3h The follow1ng nomenclature is used in the flgure T GLzé mass veloc1ty of salt lb/hr-ft2 G, =:mass veloc1ty of gas, Ib/hr ft2 a, = ft /ft of packlng, My %,saltfviscosity,;cP E = packlng v01d fractlon, - &, ='standard grav1tatlonal acceleratlon ft lb/lb hr2, .oL-=:gas density, lb/ft3 _pvfi#.salt density, lb/ft .'717 d-”' t”i,The llquld deentralnment section at the top of the column had 1nsuff1— C1ent capac1ty to permlt flooding of the column w1thout danger of plug-' '”laglng the column 8a5 outlet W1th salt and these data represent near- '_pmaxlmum flow rates at. whlch the column was stlll operable ~However, 'dthe operatlng condltions are believed to be sufflclently close to :floodlng condltions to determlne ‘that actual floodlng condltlons would _ fall 51gnif1cantly below values predicted by the Sherwood correlatlon. ORNL DWG 74-673R! A LiF-BeF, AND ARGON ' 1 lllllll LiF-BeF, AND HYDROGEN LiF~-BeF2-ThFg AND ARGON . LiF-BeF2-ThF4 AND HYDROGEN 1 11 v a1l ! ! ¥ 1 PP 1 1 L1 10'*_ X O + 50'2:- N . o'f_. - Q._I [ 3’. Q> [~ o - S| w N -0 B o] o 1073— -4} ‘00" Fig. 33. Correlation. 10-! Comparison of Actual Column Data with Sherwood Flooding + L £ O '.w* . 4 C 95 11. REFERENCES L. B. McNeese Engineering,Development pp. 16- ~29. 5, - 10. ll;l 1e. . : 'Df13,_ BV L. E. McNeese, Eng;neering Development Reactor Processing No. 8, ORNL-TM-3258 L.'E;.McNeeSe,nEngineering Development - Reactor Processing No.. 9, ORNL-TM-3259 L. E. McNeese,]EngineeringrDevelopment ~ Reactor Processing No. 1, ORNL-TM-3053 L. E. McNeese, Engineering Development Reactor Processing No. 6, ORNL-TM-31k1 L. E. McNeese, Eng1neering:Deve10pment Reactor Proce551ng No. 7 0RNL—TM—3257 . 'L E. McNeese Englneerlng Development | Reactor Proce531nngo. 8 0RNL—TM~3258 L. Bd McNeese, Eng1neer1ng,Development | Reactor Proce551ng No. l 'ORNL-TM-3053 'L E McNeese, Englneerlng,Development Reactor Proce551ng No. 5 ORNL-TM-3lh0 L. E McNeese Englneerlng Development Reactor Proce351ng No 7, 0RNL—TM—3257 L. E. McNeese, Eng;neerlng Development '-Reactor Proce551ng No. 8, 0RNL—TM—3258 L. E McNeese, Eng1neer1ng Development *Reactor Process1ng No: 9, ORNL~TM~3259 Studies for Molten-Salt .Breeder Reactor Proce551ng No 7, ORNL-TM—BQST (February.l972), Studies for Molten-Salt Breeder (Mey 1972), pp. 2k-30. Studies'for=Molten-Salt Breeder (December 1972), pp. 30-32. Studies for Molten-Salt Breeder (November 1970), pp. 1-1kh. ]Studies'for Molten-Salt Breeder (December 1971), pp. T3-T9. Studies for Molten-Salt Breeder (February 1972), pp. 52-58: Studies for Molten-Salt Breeder (May 1972), pp. 64-89. StudieS'for Molten—Salt Breeder | (November 1970), p. 6. Studies”for Molten-Selt Breeder (Oetober 1971), pp. 2—15 Studles for Molten-Salt Breeder (February 1972), pp. 29-46. Studies for Molten-Salt Breeder (May 1972}, pp. 56-63. Studles for Moltenesalt Breeder (December 1972), pp 167-95 rF J Smlth . Less-Common Metals ;2, 73 79 (1972) _J A. Lane “H. G MacPherson and F Maslan, Fluld Fuel Reactors, p. "l 728 Addlson-Wesley Publlshlng Co 5 Inc., Read1ng, Mass s 1958 s R Reactor Processing No 9, 0RNL-TM—3259 16{_ __17’ L E McNeese, Englneerlng,hevelopment Studles for Molten-Salt Breeder (December 1972) pP- 196-20h J. 1. Federer ORNL personal communlcatlon March 1971 J. R. Dlstefano and J 1. Federer, ORNL, personal communlcatlon, February 1972. et e e e < o2 T R B T WS b e e ok e e b b ke i - 18. 19. 20. 21, . Feb. 28, 1970, ORNL-4548, P 171 22. 23. 2k, 25, 26. 27, - 28. 29. 30. 31. ~ e & 33, 3k. | ORNL-h622, p. 20L, L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 8, ORNI-TM-3258 (May 1972), p. 61. . L. M. Ferrls et al., MSR Prqgram Semiann. Progr; Rept. Feb. 28, 1970, ORNL-hshB p. 289. , | | , L. M. Ferrls et al., MSR Program Semlann Prqgr Rept. Aug 31 1970 D. M. Moulton and J. H. Shaffer MSR Program Semiann. Progr. Rept. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 9, ORNIL-TM-3259 (December 1972)5 pp. 205-15. J. B. Lewis, Chem. Eng. Sci. 3, 24B8-59 (195L). . J. B. Lewis, Chem. Eng. Sci. 3, 260-78 (1954). L.”E.rMcNeese, Engineering-Develqpment,Studies.for Molten-Salt Breeder Reactor Processing No. 9, ORNL-TM-3259 (December 1972), pp. 238-50. R. G. Ross, C. E. Bamberger, and C F. Baes, Jr., MSR Program Semlann Progr. Rept. Aug. 31, 1970, ORNI-4622, pp. 92-95. C. E. Bamberger and C. F. Baes, Jr., J. Nucl. Mater. 35, 177 (1970). M. J. Bell and L. E. McNeese, MSR Program Semlann. Progr Rept Aug. 31, 1970, ORNL-L622, pp. 202—8 : M. J. Bell and L. E. McNeese, MSR Prqgram.Semlann. Progr. Rept Feb., 28, 1271,_0RNL—&676l PP- 237-&0. L. E. McNeese,-Engineerigg_Development Studies for Molten-Salt .Breeder Reactor Processing No. 6, ORNL-TM-31L41 (December 1971), pp. 59-T2.. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. 46-52. L. E. McHNeese, Engineering Development Studies'for‘MoltenQSalt Bfeeder Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 97-106. L. E.VMeNeese, Engineering Development Studies for Moltenéselt Breeder Reactor Processing No. 9, ORNL—TM+3259.(December‘1972), pp. 256-57. A. S. Foust et al., Pr1nc1ples of Unit Operations, p. 270, Wiley, New York, 1960. ' o o : %] £ sy a5 - - - » O 1O\ &+ 11. 12. 13. 1k, 15. 16. 17. 18. 19. 20. 21. 22. 23. 2k, - 25. 26. ) V 27- ' 284 ‘ 29. 30. 31. 33. 3k, 35. 36. 37, 9T h 99. Cc. F. C. E. M. R. E. S. R. E. J. -0-- E. G. Baes, Jr. ‘Bamberger Bennett Bettis Blanco - ‘Blomeke Bohlmann J. Braunstein M_o' Ao ‘R. B. H. R. ‘R. E. K. B. Bredig Brigegs Bronstein Brooksbank ‘Brown J. Brynestad D. W. . W' Lo Ho D.A w. HO R. M. Jd. L. F. L. J. M.V J. H. J. R. W. P. J., R. D.‘ .E.' ‘L. M. L. 0. J. C. W. R. Rl _H., - S. Cantor _ Cardwell Carter Cochran, Jr. Cook ‘Counce Crowley Culler Dale ‘ DeVan DiStefano Eatherly Engel Ferguson - Ferris Gllpatrlck Griess ' Grimes ‘Guymon B. A, J., R. ‘B. F. ‘R. W. Hannaford Hightower, Jr.~_:‘ Hiteh Horton o1 INTERNAL DISTRIBUTION o 39 L0, L], - h2, 43, L, - 45-63. 6L, 65. 66. 67. 68. 69. - 70. T1. ~T2.. 73, . 75. 76. 7. 78. 79. 80. 81, - 82, 83. 8k. 85. 86-87 88. 89-91. 92. 9. R "J. S. Watson " R. G. Wymer W. R. Huntley C. W. Kee A. D. Kelnmers J . Ao I{.lein W. R. Laing R. E. MacPherson ‘H. E. McCoy L. E. McNeese A, P. Malinauskas A. S. Meyer R. L. Moore J. P. Nichols H. Postma M. W. Rosenthsl A. D. Ryor H. C. Savage W. ¥, Schaffer, Jr. 'C..D. Scott M. J. Skinner .F. J. Smith G. P. Smith I. Spiewak - Martha Stewart 0. K. Tallent - L. M. Toth D. B. Trauger W. E. Unger e E. L. Youngblood ‘Central Research Library - Document Referénce Section Laboratory Records Laboratory Records, RC Y-12 Document Reference ~ Section ORNL Patent Offlce N EXTERNAL DISTRIBUTION | - 37830 T. A. Nemzek, Dlrector, USAEC .RRD, Washlngton 205h5 ,D' F. Cope USAEC RDT Slte Offlce, ORNL Oak Rldge Tennessee. ;Deslonde R. de301sb1anc, Ebasco SerV1ces, 2 Rector Street New " York, New York 10006 o 37830 iK- 0. Laughon, USAEC RRD Wfiéhlngton, DC 205&5 . o ';C L. Matthews USAEC RDT Slte Offlce ORNL, Oak Rldge Tennessee b A g 11 St 2 s b e i e+ 1T 98 100. D. J. Royer, Continentel 0il Co., Ponca City, Oklahoma TL601 101. A. M. Weinberg, Institute for Energy Anelysis, Post Office | ; Box 117, O=ak Ridge, Tennessee 37830 . | 102. E. A. Womack, USAEC, RRD, Washington, DC 205h5 103-10k. TIC, Oak Ridge, Tennessee -105. ‘Research and Technical Support D1v151on AEC ORO