A major purpose of theTechnical Information Center is toprovide the broadest disseminationpossible of information, contained inDOE’S Research and DevelopmentReports to business, industry, theacademic community, and federal,state and local governments.

Although portions of this reportare not reproducide, it is beingmade available in microfiche tofacilitate the availability of thoseparts of the document which arelegible.

2

LOS Alamos f4al,onSl LabOrSIO(y Is opermld by ma LMIIWOIYV of C91Morn@formaUnd.dSlsfosDoswrmmOf EnargyUndarconl~aclW-7405-ENG-39

TITLE Cryogenic Hydrogen Isotope Distillation For the Fusion Fuel Cycle

AUTHOR(S)’

SUBMITTEDTO

LA-UR--85-982

DE85 009635Robert H. Sherman

Materials Science and Technology DivisionLos Alarms National LaboratoryLos Alarm, New Mexico 87545

Anerican Nuclear Society National Topical Meeting, Tritium Technologyin Fission, Fusion and Isotopic ApplicationsDayton, Ohio April 30-MY 2, 1985

DISCLAIMER

Thin mpm wanpreparedu an●cauru O( work spamorodby an ?Boncyof the United SmIOSOovornmam, Nolthortho Unltcd SIatcaOovornmonlnor any ●goncythored, nrx any of tholremploy=, makesmy warrmrty,oaprossor Impllml,or ●sumea any Iogal Ilablllty or roi~nsl.blllty for tho wxuracy, wmplotonou,or umfulna.sof any irrformullon,●ppmratu,prcwlucl,orpromMdlsclosai,or roprosontothat ho UM wouldnot Infrlngc prlvatolyownmirl@t~, Rofor-oncohemln to msyIpeclflc mmmtirclalprrduct, procaq or Mtwb by Irado rmmo,lrdem~rk,mnnufaclutor,or otharwlM d~ not na~rlly ~nstltuto O: Imply~1.end~r~menl,-m.Iuondatlon,or fnvrrrlngby Ihc United WIM Oovornmentor ●y ●gencyIhoroof.Tho viewsand oplnlonsof mhom oxpti horoln do not nccamrlly stato or mffoct Ihoso of thoUrdmdSMos (lovornmontor ●ny ~~ncy thereof

_—..—z

‘- ~

bclsAuarinmLos Alamos National LaboratoryLos Alamos,New Mexico 87545

I 01(M N(1 RIR114,<!M() ;O/ 9.,n!

Dayton Paper 3/12/ 198s Add Abstract 3/18/ l’9B5.

CRY~NIC NY~ ISUTOPE DISTILLATI~F~ T- FUSI~ FUEL CYCLE

by

Robert H. ShermanMaterials Chemistry Group

Materials Science and Technology DivisionLos Aiamos National LaboratoryLos Alamas, New Mexico 8754!5

ABSTRMT

Cryogenic distillation is an attractive method far the hydrogenisotopic separations required in fusien fuel cycles. The theoreticaland practical aspects of designing and constructing such systems ●rewe: 1 e~tabl ished. Practical considerations in the design of systemsare presented and applied to the Isotope Separation System (1SS) atthe Tritium Systems Test Rssembly (TSTA), as well as systems of dis-tillation columns that might be used for a reactor such as the TokamakFusion Core Experiment (TFCX) and the r~covery of breeding blanketproduct.

INTRODUCTION

There are a number of different methods which may be used to=eparate the mixtures of D~~ DT, and T= (together with a small admixt-ure of t-k, H13P and HT) which exit as unburned fuel from a fusionreactor chamber. Of the various methods, distillation iz the oldest1:nawn t.a map and alsm probablv the I?e=t under~tood. Other method~which have been used nr prclposed include thermal dlffusiun. palladiumdiffusion, gas! chromatoqraphvs differential hydridinq~ and laser acti-vated separation proces%es.

Distillatic)n has several advantage= over the other separationm~thod~. because the separation factor~ are relatively larqe. thrauqh-put is hiqh, startup times are relatively 5hQrt, and there is laraede~] qn fle):ibllltv, G ~y=tem uf ccllumn5 ger(erally’ Gperates at asteadv s.t.ate arid there{nre c,:)rltrc!l1= rf=latlvel~’ easv. Most lnGtr!.t-

Klt-ilt~tl(T,rl h-s FJPE’1”1 WF1 1 dEk’elcIPecj arid i~ ~\,,~Jlatllp aE Qff-the-~hel f:?(.JOI:. bJtlli, ttl..-l t +F(. ldr ae tell{per k,+.1,.lrc rltddlerlts tc the II.?*I tempcl’-

dt.(lr 2 ~rivllc:rllll~:rl[,c?{ :tle cc{lL\rrlilE. d~(rlIIU ~-,~e~ d~] C!fI Lller e 15 rlcl terrlLer –~!~(rk, .:(.1 lr,(l, dIILl }IPI II.. ~~w~:r rrIOc h5r1!P.11 =(F -.~J15 ,:II-eFCt. Llr ,rl t-t,c~ \y f.~lll,i [Ir..,+.l,AtIII(1 r.r1...:::l.(r p: a 1’ F,’ :qpc,r L JI1’il’.kl ,, (,-.11.,~ +tr T@sL’hGr”c< +,-,If

tI,,../ 1; C{rL, 1“1(:,, ],.!t J(’ L,rtll,:(.lrE) ar acilt2-l I!.: WJ :!III, f IIE: E;+. L&lIl. b’3’.Ef1,1r~!fl,,l,~p.c1~.]]!, cc!rl”: Lt-ui- ! ec! F!’.arocl=ll I 5 r.,t (.-ILIFJ FErlC(rcft1clrl5 ,,’gtr.wlk are c-~l-

retid’~ J rl uper - d J CIII. t)lL151 lt 15 ~lclsslble to ~ltlli:e th~~ exl=tlr)q LQIII-rrlerclal P):perI erlc~ Irl th~ select] m of CC31Uml-Ipdr-ameter~ and seller-alc~e%lun.

TII+ II I., 7~.~r rl! ! WII,:!LI (r.! rlll:tl]l,?t] (l’ !,(:. ~ ,1( .1-,1,] ,Ylllt:. (.1{ 1 ~.(.~f {.”,[-Of.E!?L,+!r’til1011 11: ttIEJ re!ldtl~tlv ldrde Jll+erltclr# (.14trl(l(lrrlwlIILII 15 held(IL, ,3‘:-~111fc~f.tl+ I I (T1 C-III t Ill:’ : L’ 1(1(1111 F!?(LI11,f:l!.fll(.l WI”(1 15 (2I’ d’s tlL1ll

1 f t 1’1$.llcluld

l’(.’ 1,(,1 1*1’ . I (, W(:I’ P L.UUI Y w~llerlt 5 fc~r tll~ r~frluerat.or rleressarv tc

t

1

~J@#? lWWlllllONOfllilSl)OGUMWISUllllMl~

)+$

/A,(2 I

Davton Paper

maintain the ca. ~~ ~ tem~eratLlre5 are slgni+lcant.

While the detailed slmuldtion Of a multicornponent distillationcolumn is a complicated mathematical problem? present generationdlaital comp~~ters permit a rapid solutiun of the ●quations requirad todescribe the system. Dynamical simulations are possible in additionto steady state solutions which ● re the basis of system d~mign. Com-puter studies have also been made of the start-up conditions. s

The design concepts devmloped at the Tritlum Systmms TestAssembly (TSTA) have been ●pplied ta other tusion energy ●pplicationsuch as the separations required for the principal fuel stream% ●t theTokamak Fusion Core Experiment (TFCX), International Tokamak Reactor(INTOR) ~ and the Technology Development Facility (TDF). AdditionallyPpreliminary distillation system desi qns have been developed forseveral different tritium breeding blanket concepts to recover thetritium for use in the main fuel cycle. The earliest design studieswere directed at the isotope separations needed at the Intense NeutronSource (INS), which was never constructed.

GEIIERA PRIIWIPLES

There are several basic principles to be considered in the designof an isotope

1)

2)

separation system based on distillation.

The order of volatility of the hydrogenic molecules isH2~ HD~ HT, D2, DT, and T2P with H2 being the mostvolatile.In aeneral ~ a calumn can be designed to provide aspecified separation between any two 5pecies9 calledthe light and heavy I[eyp which are usual]v adjacent toeach other an the v@latili ty scale. For preliminary

design! species more volatile than the light hey can becc~n=idered to be recovered 100’Z in the distillate (ortop) product while species less vulatile than the heavyLev c an be considered to be recovered 1<)0% in the

reboiler (or bottom) product.A minimum of one column i= required for el?.c.’pur eproduct.Simple sinale columns can be used satisfacturilv over asurprlslnalv w~de range of uperatlnq conditions and i{~,<~t~(l) m+ sever .s1 lnd] vldua] (:olL!mrls rn!sv be rer:nn-flqured, s.hc(uld sv~tem Ctararrlwt[+l-E c:hdnqe. ‘The maralnal(:c1s-t Of dddirld an addlt,lona] coll,~fnrltci d s.xstern iE c?~ttlP CIrdF+r c!+ i(~;; ~f the total ~c,~.t.Utll]!? lt I+ pcII:.~It,]G tc refllcf,’e;~ddltlorlal gtrearr~=. ??selected lot,a+.lcjns al~na a cc]umrl. tc, du SCI w1ll result

lrl .4 cc,lumfl wh]ch ~% dlfflcult t,u control and CIf a

ger~erallv inflexible design.Similarlv. a column !nav ha~e anv number of feed’s. butttlev are specl fled at. t.hr+o’:per)~~ O{ svstem fle~’]b] 1 i-tv.R Ilc,lqht eq~~]vn]c~rlt to s tlleuretlc~i staqe (HE”TF-’) C14 :(

c rrl Illl?-fvhe ue~-d fc,r ~c~l~lmtrs p~~led w]th stalrlle=~steel 13d(:1 lr)cr such as) “l+c~ll-p,nl “.= Th]s 15 5Llltdbl~

Dayton Paper 3/12/1983 Add Abstract 3/18/ i9eS Paqe 3

9)

10)

11)

TSTR

At TSTA.constructed.

for columns from 1 cm l.d. to about lo cm id.A superflcla] vapor velocltv r2f CS.5 to 5 in/sec wiliresul t in adequate flow without flooding of the packedsection. This value is COmpL!tfSd based on the unpacked●rea.Holdup of liquid on the packing ●mounts to approximate-ly 9.3% of the superficial volume.rritiurn has a heat of decay of 0.326 W/g. This mani -fests itself as a composition dependent heat leak tothe column or reboiler.The undesirable species! HT, may be removed by promot-ing the equilibrium

HT + D= == HD + DT (Eq. 1)

after which the HD may be removed as an innocuousdistillat= product and the DT recovered for recycl inginto the fuel stream.

four cryogenic distillation col.~mns have been designed?and installed to effect the 5eParation5 required. ‘-4

TSTR personnel were responsible for the process design and specifica-tions. Arthur D. Little, Inc. (Cambridqe~ I’IRIwas. selected as thevendor responsible for the engineering design, process control, andconstruction. The col~!mns are approximately 2.5 cm in diameter~ 3-4m~ters lcnoq and filled with stainless steel packing. They are oper-ated in two cascaded with no internal return paths. The nominaldesign feed is assumed tn be an equilibrium mixture of 25?4 Dz~ 50% DT.and TC./Ld . T> ( with 1% protium) at a rate of approximately Z63 qram-moles/day. The product streams of HD waste (containing only 2 x 10-7

mole fraction HT), 99.99+% D= for fuelinq neutral beams. llT for reac-tor fuel inq. and 9~+X Tz to fuel a tritium neutral beam (if one isused) or for experiments requiring pure tritiurn. In all likelihood. aCo]urnn to prOdLICe pure Tx would not be needed in an operating reactor.but at TSTA it is required In order to ~jeve]op the design and opera-tlno expertise and to produc= pure t:ritiurn for experiments.

Figure 1 15 a flaw Echamal,lc uf the TSTASV=tenl

1 E.ot@pP 5.eparatlnrl(1ss) a{~d the ~l”I~Cl~.,1 d~s]on speclfluatlorls are gI ‘F’en I n

TatI]p 1. (’.:!l’.lfTlri JJ:IL dezl rim+-j $c, pro,idp a 5.eL,ilrdtlLlll W1 (11 D7 as. theIlf:lhlt i I.(,, ar,d EIT +s Lr((?‘I~;,u le~. ThP reboller craduci s~reaiII 1s

t.!,ere+(~t-c’ uti. <j-”:. N: i,nd ::,;. 1.s. Thl~ stream IS used as the feed fur~r:(l{lfllrl ~ LJt1 (Tk, C,c,r.t,:1te= dE a twcl carrlclclrlerlt{-c.lumri ar’d prc,>lde~ a

c’]r:!JIL~t~ stre~m I I D? ~ncl ~ ret,clller :tre~(l, uf 19.

Tt,p (:1~1~-. ~,t L C!tl (J+ the =vstem of Cnlurnrl~. 2 and 4 15 much mar e

Cmmple}: . ThesE~ iw~ cc)lumn% must provide a stream of pure 11= and atiaste ~.tream 0+ HD tu remnve anv pratl~~rnwhl~h f(]L(rldlt!~ Wav Into ther~a~t~r ~~,,~tem, Tfhe molecule HT 1 ~es, tnldwav between H13 and n= 1r,~olatlllt~,/ tirldl= ~(rld~slrablc~, v~t half clf anv crutlum In the Fvsterr

el)ds Ll~ ~S HT. It cannot bp discarded with the H12 waste beta’.lse ltcontain% trltlum wh]ch would con~.t~t~~te a hazard and there is tee, muchtcl ccl]le~t fcr st.oraue or burls] . Tlier@fure, lt should be treated a=

Davton Paper 3/12/ 198s fidd Abstract 3/10/ 190S Pagm 4

the heavv kev In Column 2. In Column 4. however. the separation i%between D= which then is removed as the top or distillat~ strmam ●ndDT which is extracted as the rmboiler stream. Any HT would necessari-lV appear in the distillate and contaminate the D= which is to be usedfor neutral beam injectors. Tritium in this Stream is ●lso highlyundesirable. The solution to this dilemma is to promote the prm+iouslymentioned chemical reaction (1) which i% one of several which occurwhen the three hydrogen isotopes ● re mixed. This reaction i6 cata-lyzed by smal 1 catalytic converters between Columns 1 ●nd 2, ●ndbet ween Columns 2 and 4. The separation in Column 1 depletes DT inthis loop. In additionp a large stream of D= is rmcirculmted throughthe Column 2-4 loop. Both of thmse steps, tog=ther with the separa-tion in Columns 2 and 4 tend ta drive the equilibrium r-action to theright side of the ●quation. Thus the HT molecule is ●ffectivelyremoved from the system allowing the discard of H ● m HD and ther=tention of T as DT.

These columns have been operated successfully ●nd several nf theimportant design parameters verified by special experiments. Earlyexperiments were carried out to establish the HETP which was found tobe 5 cm. Sev~ral runs have been made with the system of cGlumns whichcombined operator training with production of 50 m= of D-depleted Hz(containing <3 ppm HD) and H-depleted D= (containing < 2 ppm HD) +orresearch purpose=. Measurements of liquid holdup on the packing &ndwalls and of column pressure drop have been reported recently.=

Rt thi5 time only preliminary results are available +or feedswhich include H. D? and T. The results are very encouraging and tendto verify ti~at the multi-column system will perform as de~igned.

Several minor design error5 were made in the 1SS whichq while notaffecting nveral 1 perform~nce, have prmved to be inconvenient, To

save & trivial cost? several intercolumn 5treams were not instrumentedwith flowmeter5. but instead a calculated flow wa5 indicated. Thi5 isfine for steady statr conditions. however the actual flows are sorelynlis.sed durinq 5tartL\p when conden5atlon renders calculated f 1OHSmeanlnales5. The mis5inq flow transducer-s are now being instal l~d.cGlumn pre55urez are mea=ured bv MKS Baratron capacitance manometers.TheEe normally WCIr} quite satlsfactorll~, however we chose tc LISP on~0+ t.h~ mc~del~ wlthc,ut temperature cclmpen~atic,n and as a result there15 arl electric-al uncertainty 0+ Up to ~pvpral torr mainly due tc+,jflr

zeroand.’c,r t.cm~,eratur= Pffects...- ‘51rlc~ we are deslrclus @f malntalrl–

Jrlc: ~rlt?r CC!lLI 17111 PrPi59J f-.2 (jr (j~l,z clf the? mt-der of TO torr t)~l~ cre=erits +~C-,r~r~,l ~irc!kllpfll. The CIbVIrI_(E gc!l~(hlcrl bc,(~lclhE tc, ln~tdl] the s.llahk-1,. mot” e ~::~~erls 1 VP, t~mc,el-ature cmntrcll led s,tirls,mr~.

uavton Paper 3/12/19esi Add Abstract 3/18/198S Paqc 5

pluq lntercolurnn flow lines. During operat~on. flow tends to be inand out of the surqe volume rather than directlv from one column tcanother if there are any pressure variations in the interconnectedcolumns. The problem is being addressed by valving the tank out ofthe system with air-operated valves which will fail open. ●nd a bypassvalve. Each of the columns is individuality provided with a rupturedi sk which exhausts into the surge volume. A bleed valve is beinginstalled to allow the tank to be brought up to operating pressuregradually since if it were initially evacuated and then an orl/offvalve EiCtUated the pressure swing would totally disrupt the operatingdistillation system.

As delivered? we had specified that there were to be no hardshutoff kalves between columns. This has proved to be a great disad-vantage when repairs or modifications were necessarv because of therelatively 1arge vol ume of the columns and their interconnectionthrough long lines. With the installation of the individual rupturedisks! we are also in=,talling shutoff valves on the reboiler~ f●ed ~and distillate Iir=s of each column. These hand-operated valves canbe shut for maintenance and lea}: ~:hecking! but would normally remainopen with the handles removed. In addition, a micruswitch is attachedto each valve! and al 1 the switc))es are connected in series. Thus ifany One valve is closed! a potentially unsafe condition is indicated

In the control room.

The TSTGl isotope separation syskem is a provides a good exampleof the flexibility inherent in a system designed with individualcolumns that are externally interconnected. The question arose of howto recover In case there were to be an accidental release of ca. 100grams of tritium irito the TSTA building. Th~ Emergency TritiurnCleanup SVstem (ETC) would be activated and ca. 85 grams of tritiurnwould be recnvered as HTO in various concentratims. With additionalequipment the water could be electrolyzed to H~,. HD! and HT. After areplumbing of the faur colurnn~ of the 1SS9 the S-/stem COL!ld be used toseparate the HT. wh]ch could then be catalyied according to the reac–tion

~Hl” == H= + T>. (Eq. 2)

TFCX

—— -------- - w. .“. ● TUd rqw o

Column 1 is deslqnecl. as in TSTfi9 to provide a principal separa-tion between Dz and DT. The distillate product? which is mostly D~PIS equilibrated to promote the HT dissociation reaction. Column 2 isdesigned to provide a separation between HD and HT! thus any inqrowthof protium into the TFCX system is discharged as innocuous HD waste asthe distillate product. The reboiler product contains mostly Da ●swell as any HT and DT present in the column feed. Thar Column 2reboiler stream is split into two streams with the bulk being r~cycl~dto the appropriate location in Column 1 as a principal feed.

The reboi]er stream from Column 1 is mostly DT. To ranove excamsD= the stream is passed over a small room-temperature catalyst bmd toestablish isotapic ●quilibrium. This reactor is simply ●bout 10 cm=of finely divided palladium supported on an inert ceramic substrate.This stream is the feed to Column 39 which separates DT and T=. Thedistillate stream containing all of the Dz and most of the DT isrecirculated to same location as the external feed to Column 1,an

afterequilibration step to shift the species towards ● larger fraction

of Tz. The reboiler stream is combined with a small quantity of thereboi Ier stream from Column 2 to provide an essentially ●quimolar D-Tmixture for reactor fueling.

Table II is a summary of the important design parameters for theTFCX 1SS =ystem. An important consideration was minimizing the totaltritium inventory. At first glance it would seem that Column 1 mightbe unnece55ary. However the recycle stream to Column 1 serves toincrea5e the flows in the top section of Column 1 ●nd thus increasethe required diameter. Otherwise? the column would have a desi gndiameter nf only a few millimeters. clearlv almost impossible toconstruct and control .

For this. design. criteria for D= and T= product puritv were setby per50nnel from the TFCX team at Frir)ceton University. It is po5-sible. however, to significantly reduce the complexity of the svstemand the intientory of tritium. If it ~~ assumed that reasonably pureII= i5 still required for fuelinq neutral beams and that thegoal i~

actualan appruv]mately equimmlar mixture of D and T in the torus.

th~rl a sv~itefilof onl~ two columns Should suffice. Such a svstem ]sski,:,wn In FlqLlre .1 and the CJesiqn speclfication~ are alven in Table II.It i= p~l~5ible tQ prp~(are d wide range nf torus fueling mixtures b~vi-r ~,lr,c) t.tl[? flclwE cl+ the I-)7 stream fr@rr) the rebcller c+ Column 1ttlp

and~T’T> <tt-eafllffurll the reL,oller cI+ Column Z. EIV utlllzlrlu such a

twn-cnlufl,n ]~~,t@pp ~epdratlon :’/~tpfy, rather than a three-cclurfln~;~tp(l), tt~~’ tr]t.l~ln)]r,:e]-ltclrvVJQUld b~ reduced +rom ‘71:1,OIn~(II cl tc) ~n~ ,J-, -& , 1:),: )(”1 ~1 =lrlr.e t.klegre,,t.~sl. pc,rtl[)r-~clf ttle lrlverltorv 15 I n ttle 1 owerpcrt~mn of t.17GI pact ]nu and ttle Fetlclller n+ the third cmlumn.

BLANKET PRODUCT RECOVERY

would be ‘He due to the decav of the trltlum.oresent The resultant~eed 5tream would contain Only ? small amount of elemental trltlum asmost of the tritium would be in the form of HT of DT and hence a hiqhdegree of separation is necessary.

The first consideration was. a choice of eluent gas. If thechoice was Dz then the species in the feed stream to the distillationsystem would be basically D= and DT ●long with ‘He, whereas if theelU@nt were H~. then the species would be HZ and HT togeuher with ~He.The relative volatility for the H=/HT pair is significantly hiaherthan that for the Dz/DT pair. Hence the decision to utilize normalhydrogen (protium) as tne eluent.

A three column system wa5 designed to effect the separationrequired. The system is shown schematically in Figure 4. The impor-tant parameter5 are given in Table 111. The first column is designed

with HD as the light key and HT as the heavy key. The small amount ofHT in the distillate may be reduced by passing the gas over a 5mal 1catalyst bed and then into Column 2. In the second column a smal 1stream of ca. 1% is removed as the distillate. Because helium is onlyslightly soluble in liquid hydrogen and the withdrawl from the reboil-ers is from the liquid phase, any helium will be removed from the5ystem through the top product stream. If there were no flow ofhydrogen from the top of column 29 the helium conc~r,tration in theupper portion of the distillation column and the condenser WOU1 dbecome very high and inhibit the 5eparation. The bull: of the feed isremoved as a reboiler product. which i5 recycled to the blanket re-covery system. finy HT is thus recirculated and never need appear as awa5te product.

The reboi ler product of Column 1 is enriched from CS. 1% HT to.:a. B5% HT. ~t this point there is e55entially no T=. Therefore the5tream is sent over a catalyst bed to effect the equilibrium reac-tion given previously by Equation 2. The feed tn Column 3 th”:n con–tain5 Cs. 29”A Tz. In column Z the design separation i5 between DTand Tz, The D= and DT are recvcled into column 1 as the top productand the bc,ttom pr@duct of 9~.8Y T= i~ recovered f@r introduction intothe reactor fuellna s.vstem.

COMPUTER SIf’lULATIOFdS

ques) . In same cases, band matr~cles mav be used . How=vmr, thematricles are often 11l-conditioned and unless the computer can retaina 1arae number of significant fiqures! ● rrors will creep into thecalculations and propaqate. While an S-bit microprocessor with 64K ofmemor v has been used for preliminary design studie~first

which computeapproximations of flows and feed locations? a large mainframe

computer such as a CDC-76QCI or Cray-1 is necessary for the detailedplate-by-plat~ calculations of composition, flow, ●nd temperature.Several studi=5 have appeared which would indicate that ● large numberof smal 1 columns might be necessary for the separations required inthe fu5ion energy processes. These studies were unrealistic in thatthey were limited by the size of computer memory ●vail able or theamount of computer time alloted.

Currently available computer design simulations uf multi-compo-nent di~tillation columns coupled with the body of commercial engi -neering design and fabrication expertise will permit theconstruction of

design ●ndefficient cryogenic distillation systems for the

hydrogen isotopic separations which will be required for fusion energyreactor sy5tems0 Distillatioll systems designed with several simplecolumns provide a flexibility of configuration and operation to permitadaptation to altered feed or product specifications.

The author would like to acknowledge valuable discussions withWilliam H. Denton~ formerly with the Atomic Energy Research Establish-ment ~ Harwell~ U.K. and John R. Eartl it. Prc)fessor Donald N. Hanson,

University 0+ Californiaq Berkeley? California was most help+u? in th)eearlv staaes of the development of the computer simulations useci here.Masahiro Kino5hita of the Japan ~tomic Enerqv Research Institute,Tol;ai-mura, Japan? has been mo5t oracious in 5haring his computersimulati~ns both while an exchange scientist at Los Alamos and sub -5equentl v.

Masahlro kinashita. John R. Eartl it. and Robert H. Sherman? “AStart-up Analys15 of Four Interlined Distillation Columns forHydrogen Isotope Separation”. Fusior~ TeChnOloq~~ in press.

Reliance Glass Works? Inc.! 17 Gateway Road ~ Eensenville~ IL60106

J. R. Bartlit, W. H. Denton, and R. H. Sherman, “Hydrogen IsotopeDistillation for the Tritium Sy5tems Test Assembly”p Proceedingsof the Third Topical Meeting on the Technology of ControlledNuclear Fusion? Santa Fe, Nllp CONF-7805C)e, p778–783(197e)

John R. @artlit~ William H. Denton, and Robert H. Sherman?“Hydrogen Isotope Separation”g U.S Patent 4P353~971. Oct. 1Q82.

Robert H. Shermang John R. Bartlit, and D. Kirk Veirs? “Experi-mental Results from Hydrogen/Deuterium Distillations at the

Tritium Systems Test Rssembly”q Fusion Technoloa~ , 6962S-428(1984)

Donald N. Hanson, Private communication

William R. Wilkes. Private Communication

Masahiro Kinoshita~ “an Efficient Simulation Procedure EspeciallyDe’/eloped for Hydrogen I sot~pe Distillation Column5”!Fusion TechnoloQ y, 6~S74-E&3(19S4)

F’. C. ~o~~er5~ “Cryogenic Hydraaen Data Fertinent to MaqenticFusi on Energy”, UCRL-52b28~ Lawrenc~ Liverm@re National

Laboratory (1979)

Glbert E. Sherwood and F. Tlar}: Souers. “Thermodynamics n+ LiquidHvdr@gen Solutions”, Fusion Technoloqy~ 5.TSC)–555(19E14;

HD Wast@ Stream+ ------- ------------------- ___ -_>

I

,1

+ --- +

#:

+----+ n ,I I D2 Stream

+ ----->: EQ :----------->; 2 : +----- ---- >I + ---- + : ,I :,I 1 ,1 I :I + --- + +---+

! : : :II II + ---- + I :I + ---–--–>: E(2 :---?; 4 :

+ --- + + ---- + II

1 I 1I : I

1 1+ --- +

----- --- ?“;1 : I

Feed : : IStream : : + ----- -— --- ----- ----- + —------- _:

+—–-+ 1 DT Stream# #

: +--- +I m ,,,

1I I ,I bII :3;+------ --___ —___ ----- __ -...! ,,}

: I

+-–-+,1I

+-— ----- -_-_——-__-—- ----- ______ ,>

T2 Stream

Fioure 1.

lSCITCIFE SEF’6K’ATION SYSTEM FOR TSTA

HD Waste Stream+ ------ ---->,I

:+ --- +

: I

+----+ I ,I+ ----- ----- >.: E(3 :---------------------->; 2 :,I + ----- +

II

II

i

I + ------ +

+ --- + + --—-_——- : PUMP :’:-----, 11 I +–-----+1 : <-----+1

: 1I

i 1I

+ --- +1,

v+- . ---— --- - ----- - —- >

D2 Stream

: 1 : +-------+ +-----+\l--- - -- - - .,. , :<--! PC : <–----–--------–--: E(J !~:----+

Feed : : +------+ + ---- + I

Stream +---+ ##I + --- +

II

I 1I I

I + ---- + :II

+------------ ,>. EQ :-------------------->. : 3 :

+ ,---— + 1I

II i

+ -—- +

,I

I

+–––––--– –,>

T2 Stream

II

I

Figure 2.

IS~Toi’E SEPARATION SYSTEM F@13 TFCX

HD Waste Stream+ ------ --- >I

8

+ --- +

B #I I

+ 1---- + i I

+ - --- - -- - - - > : EQ : ----------------------:> : ~ :I + # I---- + II 1I I: +---+

: 11 + -- - - -- + v

+ --- + + - - - ----- : PUMP : <----------------+------”-->I I #I + ----- - + D2 Streamt :< -.---+

:11--------:> : #I

Feed ! :Stream +---+

II#

+------------------------------------------------ >DT + T2 Stream

Figure 3.

ALTERNATE ISOTOPE SEPARATION SYSTEM FOR TFCX

H2 + l-k Waste

II

I

+ --- +I 0I

+ ---- + I II

+ ----- ----- >.: EQ ; _--_--______-____--_--> : ~;,I + ---- +

II I

I) : I

II

I + --- +I I

+---+ 11I II + ------ -- >II : H2 Recycle Stream: 1 : +------+ + ---- +

-_-_____:. ; 1<--: PUMP : <-----------------: ECl :<----+Feed : : +------+ + ---- + :Stream +---+ I

i ~---+#I 1:I 1 1

I II1 + ----- +

-1:dl+ \;---- - - -- .- - - ,, E(J : --------------------> I #

+ ---- + I II+ --- +

,I

i

+ --------—>.

T2 Str”e~m

Figure 4.

ISOTOPE SEPARATION SYSTEM FOR ANL EILfiNP:ET PRODUCT RECOVERY

IKE 1CDLM PMWIERS FM lSlt ISOWE WMfiTlti SYSTER

Colitm Cohlal Col!m calm1 1 ! 4

. . . . . . . . . . . .. ----— . . . . . . ..- .. --—...

c of st~s, PacktiSrctiom 91 t! u 91t 34 RKtif~inq StOqrV SJ 54 10 419 Df Intwadiato Staqu o 0 0 00 d Stripplrq Stiqw n 2b 41 34lm, is. 2,0 2,0 2.0 2,0tfdt~d pKhhQ UM W09 Um UoOi

13,30 1s.33 10.5 13.51,12 0.7s9 0. 9?s 1.495

000 w m w

Fred:F1OM me, tol*s/tK 4,4(-314lb,oolt frirtion 1.4(-4)HD, sOlr friction 100(-2)HT,molt fraction 9,3(-3)h, molr frac{ion 0,240Dl, adlt friction 0,4B312 !Dolr frwtion 0,249

Dislillttr ProductsFIw rdtp, oolcs/sec 1.1(-3)

1%1 ool~ fraction 5,3(-4)IID, fiol P Irutirm 0.041HT, molt fr~ction O,OJbDa, molp friction o,91eDT, SOIF frwlion 4,?(-3)12, molt frdct)on 6.1(-E)

Rebel 1Fr ProductSF]ouRatt,to]ts!m 3,1(-31Ht, molt IrtctionHD,oolcfract]on 4,7(-9)Ml! sol? frtction 1,7(-61Da, molt friction 0.017DT, self fr~ctlon (1,At?

la ,BOIQfriction (r#JJ5

lnmtorlmHz, qraas O.lbbh, Qrtns 11,s1., qrm 24,4

ill0,990.000050,019492

4,4[-3)1,2(-4)0,020l, V;-4)0m9b90,0113,2(-s)

0,7(-5)4.3(-1)0,9941,5(-7)‘*3(-8)

4,3(-3)

2,0(-4!1,01-4)0,9B90,0! 13,3(-5)

3,392?,0O*IO

25,32 24.1327.bl 49,0920.49 47,b2

1X97 22,s9 24,47 23.0124,97 23,9427,2b 10,1,2s.02 17,9920,9b 192,0D2 M01O*9S0,W5o,2sb419

f, ?bEl120,99s0.10.b9SS20

3.1 (-31s309(-9)1,3(-b)0,0170,b490.335

2.2(-3)

5,2(-9)l,E(-b)0,014(I,92BO,(M

9,5(-4)

2.4(4)0,0110.997

0,0001*IO

41,0

9,bbD2M0. W7S0.00s0.97b3M

4,3(-s)2,9(-0)3.0(-4)lc9(-b)0.9890.O11

3,3(-s)

flJ2(-Tl2,9(-0)3,1(-4)I,b(-b)100005,7(-51

1.0(-4)

0,3270,4721.4(-31

0,01109,5i,u

.-.●

●

-------- -------- . . . . . . ... . . . . . . .Rcctlfving stqrs MC ]Katrd In ttw top w! Ion of the Colum$trlppinq stqrc artIwattd In thr bottm wction of thr colunRolw ratio of !iwid rctwn~nq frm cmdwwr to Mtrri#l sithdrom as dlstlllatrfractlm of light ky c~rnt rccowrd in ttw distillateFrtction of hmvy icy coqm.mt r~avwti in tlw distlllttcTbr tiw in pumthcm rqmwnts th? mpmrnt of 10, i.r., 1.1(-6) I 1,1 I 10*Entrlm ldt blavk wc ( 10-10

T&u IIcmLm PMmlEM Fa n TFCI Imlm SMNIM SYSTEM

O,s-n 0.247UO m

23*9J 22*2I24.b3 2s.812*4a 1.0?

I.u 1.094.3 Em.o12 KD

Fttd:Flm ~dh, M] FS/SK

h, mole fractionKD,~1~ frtctlonHT,KIlrfr~rtionD?,KIlrfr~ctionD7,solefr~cttonT2,BOIFfr~ctlan

Dlstillit~Product!FIOUrJtr,tolrs/trc%, oo!r fr~ctlonHO,aolrfrictionHl,molefr~rtlon02,sol?frtct]cmDl,to’zIrtc!lonT3,~o!rfrtttlon

Rebc! ler PrDducts

FIOM Rite, 00] PS/5tC

Hx, Qolc lr~tt]onHD, aolr fr~ct]onHI, BOIP frtttlon

Dz, B91r Irict]on

Dl, aolr frution

11 ,aolr frtctlon

91

O.?YO*M10,?blbJ5

5,7(-5)?1.4(-4)0.01!?,3(-3)0,24H0, 40!0,24B

3,3(-4)1.0(-6)

2.2(”3)j, J(-fJ

2,Wh1,9(-4)J,2(-q)

]I(lf-il

e,:l.~)

!,7(-$)

0,lbOO,bls0,2!7

0.002haI1,5

0,3712,2B1,22

HT

O.w0.001o.m47J

3,3(-4)3.8(-b)3,5[-3)2.4(-b)0,995

1#5(-3)

4.0(-7)

1.1(-6}

1.1(-3)0.999b. 9(-7)1,0(-5)

2.9(-5)

1,:(-5)

214(-b)0, 9VQ

0.0015

b.0(-1)

0,s7

4*J

0,00?

0.307

w24.3125.20

2.042.14

21.31112

0, W45o,#l0,7217bb

1,0(-4)9

0,2320,4?2

o.27k

7,5(-5)

0,322o#b7R3,%(-4)

I,q(-s@

1.3(-6)9,7(-3)0,?90

0,000],q?,7

Tw Coluul Srstrt

Calm calm

1 2.- . . ..-. . . . . . ,------- .

19 lblb 43

48 0

13 112,0 2.0

13012 UO1212.U 12.b70.J)5 6.307

MO m

2L93 22.2a

24.8J 230014.05 1.09J,97 I.w

~o.e m. o

02010.9950,0010.07b971

5.7(-S]

1.4(-4)1,1(-2)9.3(-s)0. 24B0)404

oi?4e

S, I(-3)2,5(-5)119(-3)1,7(-1)o.99b9,0(-s)

4,4(-5)

0?2(-8)2,5(-6)0,03bO.b!l0, J2J

0,002J,b2il

HBH’T

O.w0.0010.W47J

3,1(-414.2!-b)3,7(-313.0(-b)0,?9sIme(-s)0,3(-7)

I.1(-bl1,1(-3)o#w9

bi9(-7)1.0(-5)

I,?(-!II*

3,5(-5)2i4(-bl0,W80,0015btO(-7t

0,574,J

0,00s

. . . . . . . . . . . . . . . . . . . . . . . .- —.... --● ktliyinqft~ uc la~tcdInthatq uctim 04ttwCOIUUIb Str]ppinqst~gm wr Iadsd In ttwbottm wct]m of the COIUUIm Holw rttlo of lIWId rrtwnlnq Iro# cmdmwr to ~twltl wthdraw M dl~tlll~tr● Fr~ctim d l]qht h~ c~mwd rtiovwrd in ths distillate● Fr&ctim of Ikmvy k8y c~wwnt rHovwd in tW distillatee Tht ndtr in pwmt~ r~rwnts tht cIpmmt of 10, i.e., 1,1(-6) ■ 1,1 I 10+Q Entrim ltft blink &r@ ( 10-1*

TMJ Ill~~ PMAKIEkSFMMMMEl PUOWTMCMRYSV’SM

ColUul co] urn ColwlI 2 1

------ ---..-— --. —-— .

9 of Stqw, Packd ~tim W II $a6 of RKtlfyln9 Shqrr 11 10 14# 01 lnt@r*lttc Stqm 12 0 0# of Stflpplng st*s* ? I 24WV, In. 2.0 2.0 2.0MI I-P,i ?ick]nq UooS tww 43012

Frd :

Flou Ratr, S91Q3/5CC

IA, mlr fr~ctionKD, Mle fr~ctionHI, mle fractionk, sole fr~ction01, oole fraction

Tz ,oole friction

D]stlll~tr Products

Flw ritr, mlrs/sec%, Mle Ir#ctionHD, mle fraction

HT, oole fraction

k, nole Fr~ct)on01, mle Fract]on

11, nole trictlon

Rebollrr ProductsFlou Rate, Boles/seeHa, ~olt {ractlonliD, SOIP frictionHT, sole fraction

Da, colt fractionDl, tolt fraction

12 ,oo]e fraction

b. n 1.03 b.n4,n 1.40 0.?57

w no m20.b9 20,%n. 13

U0,5W?. J

1.MIt2HI

0. WIB8970.N72010,9W209

22, s320.5b 2%22SI.70 28.9053.11 slow?%00 10.00it2 HTW 12@.01~ 0, W9blb500b2(-b) b.5d7(-5)’0,0100000 o.7b7?5b

0.233

O.wo2.0(-7)9.9(-3)

●

1.3(-9)

3.!(-5)

0,231

I.OM

2.6(-B)1.1(-4)

4,0(-?)

0,0591,1(-5!04855

3i6(-b)o,oe-b

,2s!!.OM1,3(-8)1,3(-6)

2,3(-3)1,000

si2(-El)

O,??q1.0001,s(-011.3(-b)

A4.V0.0000,001

4,0(-3)

0, 2b5!.0(-5)0.443

1,1(-5!0,292

2.0(-3)

0.3741.4(-5)0.b2b

4,4(-B)2,7(-5)

l,?(-1~

5,0(-4)

3.9(-5)0,9W

1,20,000

13,1