Embed Size (px)
DESCRIPTION
Molten-salt breeder reactors ( MSBR's ) are being developed by the Oak Ridge National Laboratory for generating low-cost power while extending the nation's resources of fissionable fuel.
Citation preview
M OLTEN-SALT R EACTORS-H ISTORY,
STATUS, AND POTENTIAL
M. W. ROSENTHAL, P. R. KASTEN, and R. B. BRIGGS
Oak Ridge National Laboratory, Oak Ri.dge, Tennessee 37830
Received August 4, 1969
Revised October 10, 1969
Molten-salt breeder reactors ( MSBR's ) arebeing deuelo\ed by the Oak Ridge National Lab-oratory for generating low-cost power whileextending the nntion's resource s of fis sionable
fuel. The ftui,d fuel in these reactoTS, consistingof UFa and ThF4 dissolued in fluordes of beryl-lium and lithi,um, is circalated throu'gh a reactorcore moderated by gra\hite. Technology deuelop-ments oaer the past 20 yea?'s haue culmirnted inthe successfut operation of the ?-MW(th) Molten-Satt Reactor Experiment (MSRE), and haue in-di,cated that operatioyt with q molten fuel ispractical, that the salt is stable under reactorconditions, and that corrosion is uery low. Pro-cessing of the MSRE fuel has demonstrated the
MSR proc es sing a,s s ociated utith high'performanceconaerters. New fuel processing methods underdeaelopment should permit MSR's to operate as
economical breeders. These feahtres, combinedwithhighthermalefficiency(44%)andlot,tsfrimarysystem PressuT€, giue MSR cotzuerters and'
breeders potentially faaorable e c on o m i c, fuelutilization, and safety characteristics . Further,these reactors can be initiatty fueted ouith 233(J,
"u(J, or ptutonium. The cortstructiort cost of on
MSBR power ptant is esti.mated to be about the
same as that of tight-water rea,ctors. This could
tend to power costs 'n 0.5 to 7.0 mill/hwh less thmt
those for light-water reactors. Achieuement ofeconomic molten-salt breeder reactor s requiresthe construction and, operation of seueral reactorsof increasing size artd, their a,ssociated frocessingplants.
THE HISTORY OF MOLTEN.SALT REACTORS
Investigation of molten- salt reactors started inthe late 1940's as part of the United States' pro-
KEYWORDS: molten-solt re-ocfors, breeder reocfors, de-si gn, operotion, performonce,economics, power reocfors
gram to develop a nuclear powered airplane. Aliquid fuel appeared to offer several advantages,so experiments to establish the f easibility ofmolten- salt f uels were begun in 1947 on "theinitiative of V. P. Calkins, Kermit Anderson, andE. S. Bettis. At the enthusiastic urging of Bettisand on the recommendation of R. Grimes,R. C. Briant adopted molten fluoride salts in 1950
as the main line effort of the Oak Ridge NationalLaboratory's Aircraft Nuclear Propulsionr pro-gram. " The fluorides appeared particularly ap-propriate because they have high solubility foruranium, are among the most stable of chemicalcompounds, have very lo'w vapor pressure even atred heat, have reasonably good heat transferproperties, are not damaged by radiation, do notreact violently with air or water, and are inert tosome common structural metals.
A small reactor, the Aircraft Reactor Experi-ment, was built at Oak Ridge to investigate the use
of molten fluoride fuels for aircraft propulsionreactors and particularly to study the nuclearstability of the circulating fuel system. The AREfuel salt was a mixture of NaF, ZtFa, and UF+, themoderator was BeO, and all the piping was In-conel. In 1954 the ARE was operated successfullyf or I days at steady- state outlet temperaturesranging up to 15B0oF and at powers up to 2.5 IUW(tfr;. No mechanical or chemical problems wereencountered, and the reactor was f ound to be
stable and self -regulating.2That molten- salt reactors might be attractive
f or civilian power applications was recognizedfrom the beginning of the ANP progr&ffi, and in1956 H. G. MacPherson formed a group to studythe technical characteristics, nuclear per-formance, and economics of molten-salt convert-ers and breeders.t After considering a number ofconcepts over a period of several years, Mac-Pherson and his associates concluded that $r&ph-ite- moderated thermal reactors operating on a
NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 19?O r07
Rosenthal et al. MSR-HISTORY, STATUS, AND POTENTIAL
thorium fuel cycle would be the best molten- saltsystems for producing economic power. The tho-rium fuel cycle with recycle of 233u was found togive better performance in a molten-salt thermalreactor than a uranium fuel cycle in which "tu isthe fertile material and plutonium is produced andrecycled. Homogeneous reactors in which theentire core is liquid salt were rejected becausethe limited moderation by the salt constituents didnot appear to make as good a thermal reactor asone moderated by graphite, and intermediate spec-trum reactors did not appear to have high enoughbreeding ratios to compensate f or their higherinventory of fuel. Studies of fast spectrum molten-salt reactorsn'u indicated that gooo breeding ratioscould be obtained, but very high power densitieswould be required to avoid excessive fissile in-ventories. Adequate power densities appeared dif -ficult to achieve without going to novel anduntested heat removal methods.
TWotypes of graphite-moderated reactors wereconsidered by MacPherson's group-single-fluidreactors in which thorium and uranium are con-tained in the same salt, and two-fluid reactors inwhich a f ertile salt containing thorium is keptseparate from the fissile salt which contains ur-anium. The two-fluid reactor had the advantagethat it would operate as a breeder; however, thesingle-fluid reactor appeared simpler and seemedto offer low power costs, even though the breedingratio would be below 1.0 using the technology ofthat time. The fluoride volatitity process,o whichcould remove uranium from fluoride salts, hadalready been demonstrated in the recovery ofuranium from the ARE fuel and thus was availablefor partial processing of salts from either type ofreactor.
The results of the ORNL studies were con-sidered by a IJ.S. Atomic Energy Commission taskforce that made a comparative evaruation of fluid-fuel reactors early in 19bg. one conclusion of thetask force? was that the molten- salt reactor, al-though limited in potential breeding gain, had ..thehighest probability of achieving technical f easi-bility. "
By 1960, more complete conceptual designs ofmolten-salt reactors had emerged. Although em-phasis was placed on the two-fluid concept be-cause of its better nuclear perf ormance,t thesingle-fluid reactor was also studied.e ORNLconcluded that either route would lead to low-power-cost reactors, and that proceeding to thebreeder either directly or via the converter wouldachieve reactors with good fuel conservation char-acteristics. since many of the features of civilianpower reactors would diff er from those of theARE, and the ARE had been operated only a shortperiod, another reactor experiment was needed to
investigate some of the technology f or powerreactors.
The design of the Molten-Salt Reactor E>cperi-ment was begun in 1960. A single-fluid reactorwas selected that in its engineering features re-sembled a converter, but the f uel salt did notcontain thorium and thus was similar to the fuelsalt for a two-fluid breeder. The MSRE fuel saltis a mixture of uranium, lithium-7, beryllium, andzirconium fluorides. Unclad graphite serves asthe moderator (ttre salt does not wet graphite andwill not penetrate into its pores if the pore sizesare small). AtI other parts of the system thatcontact salt are made from the nickel-base alloy,INOR-B (also called Hastelloy-N), which was spe-cially developed in the aircraft program for usewith molten fluorides. The maximum power is^-8000 kw, ffid the heat is rejected to the atmos-phere.
construction of the MSRE began in LgGz, andthe reactor was f irst critical in 1go 5. Sustainedoperation at fuII power began in December 1966.successf uI completion of a six- month run inMarch of 1968 brought to a close the first phaseof operation during which the initial objectiveswere achieved. The molten fluoride fuel was usedfor many months at temperatures ) 1200T withoutcoruosive attack on the metal and graphite partsof the system. The reactor equipment operatedreliably and the radioactive liquids and gaseswere contained safely. The fuel was completelystable. Xenon was removed rapidly from the salt.when necessary, radioactive equipment was re-paired or replaced in reasonable time and withoutoverexposing maintenance personnel.
The second phase of MSRE operation began inAugust 1968 when a small processing facilityattached to the reactor was used to remove theoriginal uranium by treating the fuel salt withfluorine gas. A charge of 233u fuel was added tothe same carrier salt, and on october z the MSREwas made critical on tttu. six days later thepower was taken to 100 kw by Glenn T. seaborg,Chairman of the IJ. S. Atomic Energy Commission,bringing to power the first reactor to operate on,.ru.
During the years when the MSRE was beingbuilt and brought into operation, most of the de-velopment work on molten- salt reactors was insupport of the MSRE. However, basic chemistrystudies of molten fluoride salts continued through-out this period. one discoveryto during this timewas that the lithium fluoride and beryllium flu-oride in a fuel salt can be separated from rareearths by vacuum distillation at temperaturesnear 1000'C. This was a significant discovery,since it provided an ine>rpensive, oD-site methodf or recovering these valuable materials. As a
APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 19?O108 NUCLEAR
consequence, the study eff ort looking at f uturereactors focused on a two-fluid breeder in whichthe fuel salt would be fluorinated to recover the
uranium and distilled to separate the carrier saltfrom fission products. The blanket salt would be
processed by f luorination alone, since f ew fis-sion products would be generated in the blanket ifthe uranium concentration were kept low. Graph-ite tubes would be used in the core to keep the fueland fertile streams from mixing.
Analyses of these two-f luid systems showedthat breeding ratios in the range of 1.07 to 1- 08
could be obtained, which, along with low fuel in-ventories, would lead to good fuel utilization. Inaddition, the fuel cycle cost appeared to be quiteIow. Consequently, the development eff ort f orfuture reactors was aimed mainly at the featuresof two-fluid breeders. A review of the technologyassociated with such reactors was published in196?.11 The major disadvantage of this two-fluidsystem was recognized as being that the graphitehad to serve as a piping material in the corewhere it was exposed to very high neutron f luxes.
In late 196?, new experimental information and
an advance in core design caused the molten- saltprogram at ORNL to change from the two-fluidbreeder to a single-f luid breeder. Part of theinformation influencing this change concerned the
behavior of graphite at higher radiation exposuresthan had been achieved previously, and the otherpart related to a development in chemical proces-sing.
The irradiation datalz,13 showed that the kindof graphite planned for use in an MSBR changes
dimensions more rapidty than had been antici-pated. This made it necessary to lower the corepower density for the graphite to have an accept-able service life, and to plan on replacement of thecore at fairly f requent intervals. tn Moreover'complexities in the assembly of the core seemedto require that the entire core and reactor vesselbe replaced whenever a graphite element reachedits radiation limit or developed a leak. Undersuch circumstances, many years of operation of aprototype reactor would be required to proveconvincingty that the two-fluid core is practicable.
At about the time the problems associated withIong graphite exposure became evid.ent,4 chemicalprocessing development occurredtu that greatlyimproved the prospect for a single-fluid breeder.To obtain good breeding performance in a single-fluid reactor, ttt Pa (27.4- day half- Iif e) must
^beheld up outside the core until it decays to "t U.
The processing development that showed promiseof accomptishing this was a laboratory demon-stration of the chemical steps in a liquid-liquidextraction process for removing protactinium and
uranium from molten fluoride salts. The tech-
Rosenthal et aI. MSR-HISTORY, STATUS, AND POTENTIAL
nique is to exchange thorium and lithium dissolved
in molten bismuth for the constituents to be re-moved from the salt. The process has similari-ties to one being developed at Argonne NationalLaboratory for processing fast reactor fuels and
involves technology explored at Brookhaven Na-
tional Laboratory for removing fission products
from a liquid bismuth fuel. Additional data have
confirmed the early results and have shown thatthe uranium can be selectively stripped f rom the
salt into bismuth, the protactinium can be trappedin salt in a decay tank, and the uranium can be
transferred back to the salt by electrolysis forreturn to the reactor. Calculations indicated thatthe extraction and electrolysis could be carriedout rapidly and continuously, and that the processequipment would be relatively small.
Laboratory experiments also indicated thatrare earths might be extracted f rom salt f romwhich the uranium had previously been removed.Unfortunately, the chemical potentials that de-termine which constituents transfer to the bis-muth are relatively close f or thorium and thehigher cross-section rare earths, and the separa-tion is more difficult than it is for uranium andprotactinium. The extraction process may still be
workabte, however, because with no processingthe rare earths have a much smaller effect on
breeding ratio than does protactinium and so need
be removed only relatively slowly (50- to 80-daycycle time). Several other processes for rareearth removal are under investigation, andwhether liquid metal extraction will be the mostattractive is still uncertain.
The advance in core design that was importantin the switch to the single-fluid breeder was therecognition that a f ertile 'sblanket" can be
achieved with a salt that contains uranium as wellas thorium. The blanket is obtained by increasingthe volume f raction of salt and reducing thevolume fraction of graphite in the outer part of thereactor. This makes the outer region under-moderated and increases the capture of neutronsthere by the thorium. With this arrangement,most of the neutrons are generated at some dis-tance from the reactor boundary, and captures inthe blanket reduce the neutron leakage to anacceptable level. H. G. MacPherson had proposedthis scheme several years before, but it was onlystudied thoroughly after the discovery that pro-tactinium could be removed from the salt. Opti-mrzation calculations then showed that properselection of dimensions and volume f ractionscould keep the inventory of uranium in the outerregion from being excessive, while producing a
blanket region.As a result of the developments cited above,
design studies of single-fluid breeders were pur-
NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 1970 109
Rosenthal et al. MSR-HISTORY, srATus, AND POTENTIAL
sued. The studies indicated that the fuel utiliza-tion in single-fluid, two- region molten- salt reactorscan be almost as good as in two-ftuid reactors,and with the present limitations on graphite life,the economics probably can be better. conse-quently, in 1968 oRNL's Morten-salt ReactorProgram was directed toward the development ofa single-fluid breeder reactor. The immediateeffect was to bring the reactor itseU to a moreadvanced state, since in many respects the single-fluid concept is a scaled-up MSRE. At the sametime, however, a number of new features wereintroduced into the processing concept by adoptionof the extraction processes.
The characteristics of single-fluid reactorsand the work that is required to develop commer-cial- size plants are discussed in the sectionswhich follow.
ECONOiIIC AND NUCLEAR PERFORMANCEOF MOLTEN.SALT BREEDER REACTORS
Molten- salt breeder reactors off er attractionsas power producers because of potentially favor-able economic, fuel utilization, and safety charac-
ETtrEN.UfiAN}U*I
R€AETOR
teristics. The avoidance of fuel fabrication, theease of processing, and the low fissile inventoryshould result in low f uel cycle costs. Capitalcosts benefit from high thermal efficietrcy, lowprimary system pressure, and low pumping re-quirements. Some inherent safety features assistthe designer in providing a safe plant.
The above characteristics can more than com-pensate for the extra costs of handling radioactivefluids and maintaining radioactive equipment, andthat MSBR's can have low power costs. since thereprocessing can be done in a small on-site plant,the attainment of low costs does not await thedevelopment of a large fuel reprocessing industry.The use of bare graphite in the core and the rapidremoval of fission products, when combined withlow fuel inventory, result in good fuel utilization.In addition, molten-salt reactors can be started upeconomically on "t u, " uIJ, or plutonium, attohence, can use the fuel that gives the lowest costat a given time.
A simplified flow diagram of a single-ftuidtwo-region MSBR having these characteristics isshown in Fig. L, and some of the features of suchreactors are given in Table I. An example of the
GRA:F,HI TE [Jf ODE RAISRCOOLANT $ALT
:E
:lE{a6.a5(t
=
$ATT CONTAININS. FUf,I-&ND,FjERT}L-H MATERIAI-
',,TURB6':,.,,snfiiennron
STEAfuI
Fig. 1. Simplified flow diagram of a single-fluid two-region
NUCLEAR APPLICATIONS &
molten-salt breeder reactor.
TECHNOIOGY VOL. 8 FEBRUARY 19?O1t0
Fuel-fertile salt, mo1e Vo I zz tLiF, 16 BeFz,
I 12 ThFe , 0.3 uFr
| *"ntng point, 930'F
Moderator I CttPhite (bare)
Salt volume fractions, Vo I Cote, 13; blanket, 40
Core temperatures, oF I tntet, 1050; outlet, 1300
Reactor power, Mw(e) I 1000-2000
Steam system I f SOO Psia, 1000'F,
| ++Vo net cYcle efficiencY
Breeding ratio I f .OS- r'0?
Specific fissile fuet inventory,a I t.O-t'SkglMw(e)
Doubling time (compound interest),al 15-25
year
Fuel cycle cost,a mitls/kWh I 0'6-0'?(including graphite replacement)
TABLE I
Characteristics of One-Fluid, Two-RegionMolten-SaIt Breeder Reactors
aThe lower values are associated with the higher reactor powers-
fuel cycle cost breakdown for a 1000 UW(e) MSBR
is shown in Table II; to permit comparison withthe fuel cycle costs of solid fuel reactors' the
capital and operating costs of the processing plantand the cost of replacing the graphite have been
included in the table.We have also estimated the capital cost of
1000 nrrw(e) molten-salt reactor plants and com-pared them with cost estimates of light-waterreactors made on the same bases. The estimatesassume that a molten-salt industry has advanced
to the point where development costs have largelybeen absorbed, and the manufacture of materialsand the construction and licensing of plants have
become routine. Under these conditions we f ind
that the capitat costs of molten- salt and light-water reactors should be about the same. Whilethe molten- salt reactor has some f eatures which
add cost, such as the off - gas system and the
TABLE II
Fuel Cycle Cost Breakdown for a Single-FluidMotten-Salt Breeder Reactor*
*At 1070 per year inventory charge on material, B.T%o,p^e1
year fixed charge rate on processing plant, $13/g'ooIJ,"fitt.z/s
"uu, $
ONLY LIGHT WATE
CONVERTER REACTORSBU I LT THROUGHENTIRE PERIOD
€or.(._lo(n.h
I GHT I{ATER BREEDERSSTART I 976. ONLYL I GHT I.JATER BREEDERS
ILT AFTER I99
oo/,
IoFe
@ooat
=l3l
oF
Iu)y+
@oca)
=1sl
Rosenthal et al. MSR-HISTORY, STATUS, AND POTENTIAL
4.5
4.0
3.5
3.0
2.5
?.0
1.5
1.0
0.5
0
-- FBR t5 5 .35
SIGN
.07
.31
.07
.07
.63BREEDERS STARTBREEDERS BUILT
cessing schemes, but low power costs can be ob-tained with smaller breeding ratios usingprocessing already demonstrated at the MSRE,namely, xenon stripping and fluoride volatility.The MSRE processing plant, if run in a semicon-tinuous fashion, is large enough to process a1000 MW(e) reactor on a three-year cycle. Ifused with a single-f luid reactor to recover theuranium, and with the fission products and otherconstituents of the salt being discarded , a con-verter having a breeding ratio between 0.9 and 0.gwould be obtained. tB rhe accompanying fuel cyclecost, including graphite repracement cost and thefixed charges and operating. cost of the processingplant, would be ,., 0. B mills/kwh. Thus , a reactorwhich is a scaled-upr higher power-density MsRE,using an MSRE-type processing plant, is capableof producing power at attractive costs. with theaddition of extractive processing, such a reactor,or one very similar, becomes a high-performancebreeder.
SAFETY
safety is an important factor to be consideredin evaluating civilian power reactors, but it cannotbe divorced from economics. Most reactors can
APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 19?O
Altb=Ge6Y=} )->=<, o
tJ.J llt' l<
= >.-\ hl< Z, (lJF >.<- d,
(5 (J vr <9Z,f<Vl=t.{ L.l- t<J >r$- Oco cJ u-,
= t&l (7) U-lo o-J d,O (/t t-t dt
(t,F^A=, @ vr rtrtJ-, O (U +)=
c-l urtu.J= L od. 5(J,.< an O= C tt'lcro (u t-'ur +l d. (Ud.
+, (FO
=!- ou
=o cJH-c Q)d'=, ttr g)
d(F .u (u
= O tY''-(F
U.J .t1 C) 'r
l.< O -Cl (llF ._ .t_ C\d- tlat)J_ .tlJ'r' O l-
=E o-o=
\-? \--t tF(J
REFERENCE MSBR
MSR ?O I .3FBR IO 3MSR l5 1.0MSR l0 0.7FBR 53I 982. oNLYAFTER I 998.
DE
1970 I 980 I 990 2000 2ol o 2o2o 2030YEARS
Fig. 2. Uranium ore requirements for a growing nuclear power economy.
compete with fast breeder reactors in limiting theresource requirements.
To demonstrate further the importance of in-ventory, the peak ore requirements obtained fromcurves like those shown in Fig. 2 have beencross -plotted in Fig. 3 as a f unction of specif icinventory and doubling time. A point to note isthat values plotted in Fig. B depend on the assump-tions made about the growth of nuclear power. Itwas assumed that the US nuclear power capabilitywould reach 140 000 MW(e) in 1980, 930 000 l\rlw(e)in 2000, and expand at a rate of 100 000 Mw(e) peryear afterwards. If nuclear power grew expo-nentially after 2000, the ore requirements wouldnot reach a peak unless the doubling time of thereactor considered was less than the doublingtime of the long term growth of the nuclear powereconomy.
MOLTEN.SALT CONVERTER REACTORS
Nearly all the present ORNL effort on molten-salt reactors is devoted to developing a breeder,since it has the lowest potential power cost aswell as good fuel utiLrzation characteristics.Achieving good breeding performance is dependentupon the successful development of advanced pro-
112 NUCLEAR
@
O(Y)
=FRAJ
roo
F4z,u.J
=u-ld.l-l ?=vOu.te.
U2d.
oV'lrld. I
DOUBL ]
TIMI35Y
NG
eor_
/I
l5 Yeor,-at-l
./ / -/l0 Yeor!
4 5 Yeor
01234567SPECIFIc INVENT0RY [kg/MW(e)]
Fig. 3. Maximum uranium ore requirements.
probably be made adequately safe by conservativedesign features and redundant engineered safe-guards that add to the cost of power. In thisregard, molten-salt reactors have some inherentfeatures that may result in cost savings relativeto other systems: the primary system operates atIow pressure with fuel salt that is more than1000'F below the boiling point, some fission prod-ucts are removed from the reactor continuously,iodine and strontium form stable compounds in thesalt, and the salts do not react rapidly with wateror air. The need for excess reactivity is reducedby the continuous fuel processing, and a promptnegative temperature coefficient is associatedwith heating of the fuel salt.
A safety disadvantage of molten- salt reactorsis the accumulation of fission products in the pri-mary system, the off-gas system, the fuel storagetanks, and the processing plant, which requiresprovisions to ensure that the fission products wiIIbe contained and their decay heat will be removedunder all conceivable circumstances. Partiallyoffsetting this is the ability to drain the fuel intoa tank having always ready, redundant coolingsystems.
Preliminary evaluation of conceptual designssuggests that the safety characteristics of molten-salt reactors wiII be a net asset, but a thoroughevaluation of fairly detailed designs wiII be re-quired before the economic importance of this canbe assessed.
Rosenthal et al. MSR-HISTORY, STATUS, AND POTENTIAL
REMAINING TECHNOLOGY DEVELOPMENTS
While molten-salt reactors offer many attrac-tive featur€s, there is much to be done beforefull-scale economic molten-salt breeders can be
built. What is seen as the remaining primary re-search and development in diff erent areas issummatized below.
Physi cs
No crucial physics problems are seen in thedevelopment of molten-salt reactors, but in viewof the relatively small breeding gain of an MSBR,particular care must be taken to ensure that thenuclear calculations are accurate. The combineduncertainty in breeding ratio due to uncertaintiesin cross sections of all the reactor materials is- t0.016, of which - 0.012 results f rom uncertaintyin the 4 value of "t LJ.
To reduce the uncertainty, the capture-to-f ission cross- section ratio of 233U is being de-termined from changes in isotopic ratios in theMSRE fuel. Additional measurements of the ab-sorption cross sections of the salt constituent-Li, B€, and F-could reduce the uncertaintyfurther. tattice experiments with concentrationsand geometries appropriate f or the core of anMSBR should be made in facilities such as thePhysical Constants Tests Reactor and the HighTemperature Lattice Test Reactor at BattelleNorthwest Laboratories, with emphasis on verify-ing the temperature coeff icients of reactivity,including Doppler coefficients. These experimentswould also provide an integral check on theadequacy of cross-section values and on compu-tational models.
Chemistry
An extensive program of molten-salt chemicalresearch over the past 20 years has established asound basis for understanding the chemistry of thepertinent f luoride salts. Additional work is de-sirable, however, in several areas.' A betterunderstanding is needed of the behavior of thenoble metal fission products in MSBR systems.Alternative processes f or rare earth removalshould be investigated, and chemical informationon the protactinium removal process needs to be
extended to a wider range of conditions. Thephase behavior of PuFs in molten salts ought to bestudied more thoroughly to provide a firmerchemical basis for the use of plutonium. Furthermeasurements are needed of the physical andchemical properties of molten salts, and particu-Iarty of the sodium fluoride/sodium fluoroboratesalt mixtures which are the presently proposedsecondary coolants for MSBR's.
NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 1970 il3
Rosenthal et aI. MSR-HISTORY, STATUS, AND POTENTIAL
Analytical Chemistry
Adequate control of the chemistry of the saltsin the reactor systems depends on development ofrapid, reliable, and economical analytical tech-niques for the precise determination of uranium,thorium, and other constituents of the f uel salt,the detection of bismuth in fuel salt, the measure-ment of the concentration of trivalent uranium infuel salt, and the analysis of constituents of fueland coolant system cover gases. At present, suchmeasurements are obtained through sampling andsubsequent analysis, but promising electrochem-ical and spectrophotometric methods f or con-tinuous, or almost continuous, analyses are underdevelopment. These methods need to be developedto the point where dependable instruments can bedemonstrated.
Fuel Processing
Obtaining good breeding in a single-fluid reac-tor requires rapid on- site reprocessing to keepprotactinium out of the high neutron flux as wellas to remove f ission products. Investigation ofmost of the steps needed to effect rapid repro-cessing has not proceeded past preliminaryIaboratory tests. Thus, development of the liquid-bismuth extraction system will require an ex-tensive program of engineering experiments.Materials problems, including the inherentty dif-f icult coruosion problem of the electrolyzer f ortransferring uranium from bismuth to salt andthorium and lithium from salt to bismuth, willhave to be overcome.
Nonradioactive engineering experiments shouldbe sufficient to develop and prove out componentsandto demonstrate the practicality of salt-bismuthcontractors and of the electroly zer. Removal ofrare earths also can be tested using nonradio-active material, but experiments using significantconcentrations of protactinium will have to beperformed in facilities that have provisions forhigh q activity. Demonstration of a processingsystem with high fission-product heating rateswill only be possible on a reactor.
Although a workable fluorination process hasalready been demonstrated, advantages would ac-crue from using a layer of frozen salt to protectmetal surfaces from corrosion. Since fluorinationcan have an auxiliary role in single-fluid breederprocessing as well as the main role in processinga converter, further work should be done on this.Additional work is needed on vacuum distillation,which appears applicable to recovery of lithiumand beryllium, and possibly thorium, from wastestreams.
Materials
Graphites that would be acceptable for thecores of early molten-salt reactors can probablybe obtained commercially in the near future, butproduction techniques must be scaled up, andcommercially produced materials must be testedfor reactor use. Damage from high-energy neu-trons eventually leads to deterioration of graphiteat temperatures of interest to molten- salt reac-tors. However, there is some e4perimental evi-dence that the resistance of graphite to radiationdamage can be significantly increased, and im-proved grades need to be developed and tested.Graphite development work centers on use of theHigh- Flux Isotope Reactor, in which a dose of4 x L0" nvt (nyear.
Helium sparging witl remove 135Xe from moltensalts, and reduction of xenon poisoning by a factorof 6 has been demonstrated in the MSRE. Toachieve the factor of 10 reduction desired for abreeder, howev€r, will require that the perme-ability of the graphite, as measured with helium,be reduced to '-" 10-8 cm' /sec. Pyrolytic carbonimpregnation of the surface has been shown to becapable of producing graphite of such low perme-ability, but much more testing under imadiation isrequired, and the process must be scaled up foruse with reactor-size graphite bars.
The Hastelloy- N used in the past for containingmolten salts is embrittled by helium produced inthe metal when it is irradiated. Although thepresent material has been satisfactory f or theMSRE, &tr improved alloy is needed f or powerreactors. A modified form of Hastelloy-N con-taining small amounts of titanium or hafniumappears to meet this requirement. Further testingunder a variety of irradiation conditions must becompleted to demonstrate fully that the desiredimprovement has been accomplished. Large heatsmust be obtained, and the modified material mustbe subjected to enough testing to have it approvedfor pressure vessel use by the ASME. Thechanges in composition from standard Hastelloy-Nare not great, however, and full requalifications isprobably not required. For the longer term, creeprupture tests at low stress conditions and extend-ing over many years should be made to verify theextrapolation of data f rom the requalificationtests. Resistance of the modified alloy to corro-sion by fuel and coolants salts-particularly thefluoroborate-must be demonstrated at all reactoroperating conditions.
In a reductive extraction fuel processing sys-tem, materials will be exposed to both salt andbismuth at temperatures between 500 and 700'C.Candidate materials will have to be tested, and
APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 1970114 NUCLEAR
methods of protection, such as deposition of filmsor frozen salt layers on the wall, may have to bedeveloped.
Mol ten-Salt lrradiation Experiments
MSRE operation has shown the compatibility ofgraphite with fissioning fuel at low power densi-ties and low burnup, and a variety of in-pilecapsules and loops have provided additional infor-mation. However, the experience to date has notincluded exposures at the power densities ofbreeder reactors. In-pile loop experiments thusneed to be performed using appropriate materialsat the peak power densities of high-performancebreeders.
Engi neering Development
Development of large components is probablythe most costly of all the technology needs of theMSR. Most of the components built so far havebeen of MSRE size or smaller. While the MSREhas demonstrated the workability of small com-ponents, much larger equipment will be needed forMSBR's.
Vertical shaft pumps with overhung impellerssimilar to the MSRE pumps would be used, and thebasib requirement is to increase the size. Heatexchangers for MSBR's must be of very highperformance to obtain a low fuel salt inventoryand must be capable of replacement. The develop-ment of steam generators heated by molten saltinvolves an area in which there has been almostno experience, and the high melting point of thesalt will probably cause the design to be consider-ably different from that of other steam generatorsbuilt to date. The control rods and drives forlarge MSBR's wiII be different from those on theMSRE, because air cooling will not suffice at thehigher power densities. Rods and drives that per-mit insertion of control elements directly into thefuel salt will be needed.
Xenon is effectively stripped from the fuel saltin the MSRE pump bowl, but for a large, high-performance MSBR, it will be necessary to injectbubbles into the salt under controlled conditionsand to remove them effectively. Reliable methodsfor disposing of afterheat in the f uel salt afterreactor shutdown must be demonstrated f or avariety of reactor conditions.
The ability to maintain the reactor primarysystem and the processing system of an MSBRpower plant must be assured. New techniquesneed to be developed that go beyond those used onthe MSRE. The ability to cut and weld pipingremotely would be valuable, and existing automaticwelding equipment should be adapted for this useand demonstrated.
Rosenthal et aI. MSR-HISTORY, STATUS, AND POTENTIAL
A number of test facitities will be needed foruse in the development and testing of componentsand systems.
These and other developments must be accom-plished on an appropriate scale before the nextmolten- salt reactor is completed. Where they aresuccessful, only scale-up wiII be required for thelarger reactors that follow.
Reactor Development
Achievement of economic molten- salt breederreactors requires the construction and operationof several reactors of increasing size and theirassociated p r o c e s s i n g plants. Construction ofthese reactors wiII establish the industrial capa-bility to manufacture molten- salt reactors andwill test and demonstrate the technology whichIeads toward dependable reactor operation and lowpower costs. The next reactor could be a molten-salt breeder experiment which has all the featuresand systems of a full- scale po\I/er breeder andoperates at power densities and other conditionsextreme enough to provide a thorough test offuture reactor conditions. Alternatively or con-currently, a converter reactor could be built whichuses technology largely demonstrated by theMSRE except for equipment scale-up to higherperformance and larger size.
coNcLUsroNs
The objective of developing breeder reactors isto obtain a source of low- cost energ'y f or our-selves and for future generations. Three eventsof the past year make it seem likely that molten-salt reactors can accomplish this:
1. the very successful operation of the MSREon both 235U and 233U
2. the processing of a complete charge of fuelat the MSRE
3. the advances in chemical processing andcore arrangement that make it possible todesign a single-fluid breeder.
The f irst event gives assurance that molten salttechnology can be used dependably f or powerreactors. The second event means that a con-verter reactor having a breeding ratio of 0.8 to0.9 and attractively low power costs can be basedon a demonstrated processing system. And thethird means that molten- salt breeders, &s well asconverters, can essentially be scaled-up versionsof the MSRE.
While adoption of the single-fluid conceptbrings the breeder reactor nearer to fruition, italso increases the development needed for the
NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 1970 115
Rosenthal et al. MSR-HISTORY, srATUS, AND POTENTIAL
9. L. G. ALEXANDER, \ry. L. CARTER, C. W. CRA-VEN, D. B. JANNEY, T. W. KERLIN, and R. VANWINKLE, "Molten-Salt C onve rte r Reactor DesignStudy and Power Cost Estimated for a 1000 MweStation," ORNL-TM-1060, Oak Ridge National Lab-oratory (September 1965).
10. M. J. KELLY , "Removal of Rare Earth FissionProducts from Molten-Salt Reactor Fuels by Distilla-tion,t' Trans. Ayn. Nucl. Soc., 8, 170 (196b).
11. The status of molten-salt reactor development wassurveyed in a series of ORNL reports issued in 1967 .They are:
R. B. BRIGGS, ' 'summ ary of the objectives, theDesign, and a Program of Development of Molten-SaltBreeder Reactors,r, ORNL-TM-I9b1 , Oak Ridge NationalLaboratory (June 12, LTGT).
w. R. GRIMES, "chemical Research and Develop-ment for Molten-salt Breeder Reactors,rr ORNL-TM-1853, oak Ridge National Laboratory (June Lg6T).
H. E. MccoY and J. R. WEIR, "Materials'Devetop-ment for Molten Salt Breeder Reactors,r, ORNL-TM_1854, oak Ridge National Laboratory (June Lg67).
DUNLAP scoTT and A. G. GRINDELL,',componentsand Systems Development for Molten-Salt Breeder Re-actors," ORNL-TM-1955, oak Ridge National Labora-tory (June 3 0, L|GT ).
A. M. PBRRY, "Physics program for Molten saltBreeder Reactors," ORNL-TM-185Z, Oak Ridge Na-tional Laboratory (June L}GT).
PAUL R. KASTEN, "safety Program for Morten-saltBreeder Reactors,', ORNL-TM-I8b9, Oak Ridge Na-tional Laboratory (June 9, L}GT).
J. R. TALLACKSON, R. L. MOORE, and S. J.DITTO, "Instrumentation and controls Development forMolten-salt Breeder Reactors," oRNL-TM-1gb6, oakRidge National Laboratory (May 22, Lg6T).
'W. L. CARTER and M. E. WHATLEY, ,,Fuel andBlanket Processing D e ve I op m e nt for Molten-SaltBreeder Reactors,t' ORNL-TM-1852, Oak Ridge Na-tional Laboratory (June L967).
6. G. I. CATHERS, ,.Uranium Recovery from SpentFuel by Dissolution in Fused Salt and Fluorination,r, ROBERT BLUMBERG, ,,Maintenance Development!Lc1. sc.i. Enf..., 2t 768 (195?); w. H. CARR, "Volatility for Molten-salt Breeder Reactors," oRNL-TM-1859,Processing of the ARE F\rel," chem. Eng. sjtnp. series, oak Ridge National Laboratory (June 1962).56, 28 (1960).
breeder processing plant. Scale-up towards com-mercial-size breeder reactors, however, need notawait the demonstration of breeder processing.Since the converter and breeder are basically thesame except for the processing plant, the scale-upcan begin by building converters, with expectationthat they will evolve into single-fluid breeders asthe processing is improved. This f lexibility-generally a result of using a fluid fuel-extendsalso to the freedom to start up the reactors with" uIJ, "t fJ, or Pu, depending on which is mosteconomic at the time.
Much remains to be done bef ore economicmolten- salt breeder reactors can be achieved, butthe basic virtues of the molten- salt system andthe success of the MSRE offer promise that thebreeder development program will meet withearly success.
ACI(NOWLEDGTIENTS
This research was sponsored by the u. s. AtomicEnergy commission under contract with the UnionCarbide Corporation.
REFERENCES
1. H. G. MACPHERSON, "Molten-salt Reactors,',fuoc. Intern. conf. constructiue (Jses At. Energy,Washington, DC, Nouember jgGB, American NuclearSociety (March 1 96 9).
2. ''The Aircraft Reactor Experiment,r , Nucl. Sci.Eng., ,2, 7 97 (1 957).
3' J. A. LANE, H. G. MACPHBRSON, and F. MASLAN,Fluid Fuel Reactors, pp. 56T ff, Addison-wesley, Read-ing, Mass. (1958).
4, L. G. ALEXANDBR et &I., Oak Ridge NationalLaboratory, Personal communication (1 96 2-L963).
5. P. R. KASTEN, "The MosEL Reactor concept,r,Third u.N. Intern. conf. peaceful uses of At. Energy,Geneua, A/Conf za/p/sga (1964).
7. us-AEC TASK FORCE, "Report of the Fluid FuelReactors Task Forc e ," TID-8502 (Febru ary 1 959),
8. L. G. ALEXANDER, L. CARTER, R. H. CHAP_MAN, B. KINYON, J. 'W. MILLER, and R. VANWINKLE, " Thorium Breeder Reactor Evaluation-partI, Fuel Yield and Fuel Cycle Costs in Five ThermalBreeders," ORNL-CF-61-B-9, Oak Ridge National Lab-oratory (September 1 gG1).
L2. R. W'. HENSON, A. S. PERKS, and J. H. W. SIM-MONS, "Lattice Parameter and Dimensional changes inGraphite Irradiated Between 300 and 13b0'c," AERE R5489, United Kingdom Atomic Energy Research Estab-lishment, Harwell (June L}GT).
13. J. w. HELM , " Long Term Radiation Effects onGraphite,tt Paper MI 77 , 8th Biennial Conference onCarbon, Buffalo, New York (June Lg67).
APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 19?Olt6 NUCLEAR
L4. P. R. KASTEN, E. S. BETTIS, W. H. COOK'W. P. BATHERLY, D. K. HOLMES, R. J. KEDL,C. R. KENNEDY, S. S. KIRSLIS, H. E. McCOY, A. M.PERRY, R. C. ROBERTSON, D. SCOTT, and R. A.STREHLOW, "Graphite Behavior and Its Effects on
MSBR Performance,t' Nucl. Eng. Desigrt, 9r 2 (Feb-ruary 1969).
15, R. GRIMES, "Molten-Salt Reactor Chemistry,ttNucl. Appl. Tech., 8, 137 (1970).
16. JACKSON & MORBLAND and S. M. STOELLERASSOCIATES, "Current Status and Future Technical and
Rosenthal et al. MSR-HISTORY, STATUS, AND POTENTIAL
Economic Potential of Light Water Reactofs,t' WASH-1082, IJ.S. Atomic Energy Commission (March L968).
L7. "Performance Potential of Large Seed-BlanketBreeder Reactof s," WAPD-LSBR-300, Bettis AtomicPower Laboratory, Westinghouse Electric Corporation(February 1965).
18. A. M. PERRY and H. F. BAUMAN, " ReactorPhysics and Fuel-Cycte Analys€s,t' Ntrcl. Appl. Tech.,8, 208 (1970) .
NUCLEAR APPLICATIONS & TECHNOLOGY VOL. 8 FEBRUARY 1970 117
