An Analysis of Misconceptions About Thorium-Based Fuel Cycles As a Technical Fix to the Nuclear Proliferation Problem

An analysis of misconceptions about thorium-based fuel cycles as a technical fix to the nuclear proliferation problem.

This paper analyzes several pre- rich to bomb-usable levels than natu- of the plutonium economy are taken valent misconceptions about nuclear ral uranium. Such enrichment can be as read. Nothing in this article fuel cycles that breed fissile done at a modest scale with readily should be construed to imply that: uranium-233 from thorium. Its main available centrifuges. L e the proliferation risks of conclusions are: a Several other features of mixed plutonium fuel cycles can be made a Uranium-233, despite the thorium-uranium cycles are rather tolerable, or gamma radioactivity of associated unattractive from a safeguards point 9 any commitment (even an ex- isotopes, is a rather attractive mate- of view. ploratory one) to a plutonium econ- rial for making fission bombs, and is Thus thorium cycles of any kind omy need be made for many de- a credible material for subnational as cades, perhaps even a century or two, well as national groups to use for this are not a technical fix for proliferation (national or subnational) and, any improvement in existing purpose. safeguards (such as those sought by e "Pure" thorium cycles, which in though probably more safeguardable than plutonium cycles, are less so the Carter administration1 is not effect merely substitute uranium-233 worth making only because it does for plutonium, would take many than once-through uranium cycles that entail no reprocessing. not solve the whole problem, or decades and much uranium to any form of nuclear power is a While thorium cycles have some establish, and offer no significant necessary or desirable. safeguards advantage over plu- potential technical advantages, including flexibility, they cannot pro- All these fallacies are addressed tonium cycles. elsewhere [2]. While some of my ar- vide major savings in nuclear fuel re- e "Denatured" thorium-uranium guments have been previously ad- cycles, which dilute the uranium-233 sources compared to simpler ways of saving neutrons and uranium. vanced by advocates of plutonium with inert uranium-238 to a level not fuel cycles, presumably seeking to directly usable in bombs, are not an Thus while advocates of nuclear suggest that their option is accept- effective safeguard even against power may find thorium cycles able because the thorium alternative subnational bomb-making. This is worth exploring, such cycles do not to it is nearly as bad, I shall suggest because three factors-higher initial differ fundamentally from uranium that neither is acceptable. Thus any fissile content, larger mass dif- cycles in any of the respects- out-of-context quotation of this ference between isotopes, and including safeguards and fuel paper in support of plutonium cycles smaller amount of material needed resources-that are relevant to the would be perverse and misleading. for a bomb-make such "de- broader nuclear debate, and should Since proposals were widely pub- natured" uranium-233/238 fuel one not be euphorically embraced as if lished in 1976-1977, notably by or two hundred times easier to en- they did.-A.B.L. Taylor and Feiveson 131, for a "denatured" [4] thorium cycle which If nuclear fission is to be used as an FO~~-MITREreports. Those who be- might be less proliferative (espe- energy source, whether it should use lieve plutonium is nasty and uranium cially to subnational groups) than a fuel cycles that require plutonium to may be scarce call attention to fuel- plutonium-uranium fuel cycle [5], be extracted and re-used turns on efficient, non-plutonium fuel cycles. some people [6] have jumped to the two independent questions [I]. The It is this last, perhaps most interest- conclusion that thorium cycles in first is whether plutonium is bearable ing, group that has recently recalled general and denatured thorium cy- and the second is whether a lot of attention to thorium-based fuel cycles in particular offer a technical fix uranium is available (geologically cles which appear to many nuclear to the proliferation problem. This is and politically) at relatively low advocates to merit study on several incorrect. It is also certainly not prices. Few if any analysts answer grounds, potentially greater resist- what such authors as Taylor, Feive- yes to both questions; and for any ance to the proliferation of nuclear son, Cochran, and von Hippel had in who do, policy choices are un- weapons being high among them. mind: they proposed denatured constrained. Those who believe It is this cluster of proposals, and thorium only as proof of the exis- plutonium is bearable and uranium some prevalent misinterpretations of tence of fuel cycles which might, on scarce build fast breeder reactors. them (especially common in Britain closer study, prove to be inherently Those who believe plutonium is and Canada), that this article briefly less proliferative than uranium/ nasty and uranium plentiful write discusses. The arguments of critics plutonium cycles. Since at least one The concern over weapons proliferation problems has lately led to serious international consideration of alternative nuclear fuel cycles. Among these. cycles involving the breeding of the uranium-233 isotope from thorium, especially with provisions for its "denaturing" with ordinary uranium, are generally believed to show great promise for a nuclear energy future without weapons proliferation. In this article the author dashes cold water on the most optimistic of such expectations.-Ed .

such fuel cycle could be envisaged, 15.8 grams per cubic centimeter). ting use of a gun-type rather than an they argued, it is important not to (Corresponding figures for the alpha implosive assembly-a further op- rush ahead with a plutonium eco- phase of plutonium, at around 19.8 tion which may be easier than an im- nomy before potentially less pro- grams per cubic centimeter, are 11 plosive design. Both chemically and liferative options are carefully kilograms for nearly pure metallurgically ui'anium-233 is con- examined. plutonium-239 and 13 kilograms for siderably more tractable than The early stages of that closer "reactor-grade" plutonium pro- plutonium: it is not very pyrophoric examination have made it clear that duced at a burnup around 30 thermal and has one solid phase, not six. Its thorium fuel cycles, including de- gigawatt days per metric ton of inhalation toxicity is considered to natured ones, do not qualitatively uranium. Of course all these quan- be about 70 times lower than that of alter the prolifefation problem tities can be reduced greatly-more plutonium-239, several hundred (either national or sbbnational) and than fivefold-by reflection and im- times lower than that of reactor- in some respects offer more avenues plosion in a nuclear explosive.) grade plutonium. for proliferation than once-through Broadly speaking [7],for fast neut- The only significant disadvantage uranium cycles. The reactor-physics rons the relative neutron yield is of uranium-233 is that after produc- and safeguards considerations sup- tion, various daughters of uranium- porting this conclusion are some- 232 build up in it (peaking about 10 what elementary, but the scarcity of years after extraction), notably systematic analyses of them has and the fission cross-section thallium-208. Some uranium-232 lately generated much misplaced en- daughters emit intense and penetrat- thusiasm [6]: hence this summary, ing gamma rays, of which thallium- These conclusions are relatively in- 208 typically contributes more than Thorium and Uranium-233. There sensitive to contamination of the half the total with its 2.6- and 0.6- are two basic kinds of thorium cy- uranium-233 with traces of other megavolt emissions. Commercial cles, both of which may breed fissile uranium isotopes. handling of uranium-233, especially uranium-233 from fertile thorium- Because uranium-233 is about if it were more than a few days old 232, mainly via the intermediate three times as reactive as uranium- (whereas aging for about a year is product protactinium-233, which has 235 (whose bare-sphere critical mass desirable for full decay of the a 27-day half-life. The first kind is the at 93.5 percent enrichment is about protactinium-233 arent of the "pure" thorium cycle, a self- 50 kilograms), one might expect that uranium-2331, woul 1require heavy sustaining one relying on previously a third as much of it would be re- shielding-of order twice as much as bred uranium-233 for fission neut- quired for a bomb. In fact the re- the gamma- and neutron-emitting re- rons to breed more uranium-233 in ciprocal of this ratio, the "substitu- cycle plutonium requires-to protect fuel consisting of uranium-233 mixed tion factor," is not 3 but typically 5 workers from overexposure. But if with thorium-232. Burnup of this fuel to 10 because a physically smaller relatively small amounts of produces various isotopes, up to and core can be compressed more in a uranium-233 were being dealt with including small amounts of convergent implosion. (An in- for relatively short periods outside plutonium formed by successive compressible solid spherical shell, if the restrictions of prudent health- neutron captures. imploded, provides a pressure vary- physics practice, shielding could be From a weapons physics point of ing as the inverse fourth power of its greatly reduced or even dispensed view, uranium-233 is similar and in radius @I.) with. some respects perhaps superior to Ordinarily uranium-233 has a sub- A kilogram of uranium-233 pro- plutonium-239, the classical mate- stantial neutron background because duced at 46 thermal gigawatt days per rial. They are roughly equal in re- of (a,n) reactions with light-element metric ton of heavy metal burnup activity, as shown by the nearly impurities. With a significant chemi- and 32 thermal megawatts per metric identical bare-sphere critical masses cal effort, which is not normal prac- ton of heavy metal specific power, [7] of uranium-233 (16 kilograms at a tice, these impurities can be greatly for example (corresponding to about density around 18.8 grams per cubic reduced so that the neutron back- 140 milligrams of uranium-232 per centimeter) and the delta phase of ground becomes lower than in kilogram of uranium), then aged 90 plutonium-239 (17 kilograms at about weapons-grade plutonium, permit- days. would contain [9] about 82 mil-

February 1979 The Bulletin 17 Amory Lovins, a former Oxford don and Regents' Lecturer at Berkeley, is the British Representative of Friends of the Earth, Inc., and a consuitant physicist active in energy po1icy in about 15 countries.

licuries of thallium-208. Total un- thorium cycle is one attempt, on pa- need to have at least seven atoms of shielded gamma dose from that per, to reduce those problems to a uranium-238 in fresh fuel per atom of kilogram at a distance of 1 meter more tractable level. uranium-233-is nearly 40 kilograms would be of order one-quarter per electric gigawatt-year sent out: 5 roentgen per hour. An American Denatured Thorium Cycles. The to 7 times [I21 less than with a Physical Society working group [ 101 denatured thorium cycle-actually a light-water reactor, not zero. cites a uranium-232 content of order mixed thorium-uranium or thorium- Second, "secured" facilities are 200 to 1,000 milligrams per kilogram uranium-plutonium cycle-relies on to be centrally sited in at most a [11], implying maximum dose rates the possibility of denaturing few places under international con- {after roughly hundred-day aging) of uranium-233, that is, making it into a trol. Irradiated fuel from national re- order 1 roentgen per hour from 1 form not directly usable- for actors is then shipped to these cen- kilogram at 1 meter. Such dose rates, weapons, by mixing it with at least ters and reprocessed. All fissile ma- though substantial, are hardly a mat- seven times as much non-fissile terials, including both plutonium and ter of major concern to people likely uranium-238 from which it cannot be uranium-233, are recoverable, and in to put the materid to illicit use in the chemically separated as plutonium many analyses are assumed to be re- first place. It is also possible that could. This denatured fuel flows be- covered, then used in various ways supervisors of such use might ne- tween two different kinds of reactors to breed more uranium-233 for de- glect to tell workers of the hazard. in different places. naturing and shipment back to the Because uranium-233 is an attrac- First, "national" reactors-most national reactors as fresh, isotopi- tive weapons material and because or all of the total reactor population- cally denatured fuel. This breeding chemical processing by methods are built in relatively dispersed sites might be done in a thorium-232- thoroughly described in open litera- which could be internationally ad- blanketed breeder (for example, a ture can readily separate uranium- ministered but are more likely to be liquid-metal fast breeder reactor) 233 from thorium-232, the safeguards under national control (often of that burns pluto~um-239and some implications of a pure thorium cycle non-nucjear-weapons states). These uranium-233 in its core. (The good differ only in detail from those of a reactors are breeders or advanced neutron yieid of the resulting plutonium-239 cycle. The former of- converters with good neutron eco- uranium-233 in thermal converters fers no significant safeguards advan- nomy$ such as heavy-water, might make the asymptotic ratio of tage over the latter-not even the spectral-shift> or high-temperature national converters to secured ease of detecting uranium-233 by gas-cooled reactors* They are fueled breeders as large as 3 to 4.) The prin- remote sensors such as door with a mixture of uranium-233, ciple, however, is that no monitors, since modern techniques uranium-238, thorium-232, and plutonium-239 goes off-site: the only can also (we are assured) sensitively perhaps uranium-235. {Liquid-metal fissile material leaving the "se- detect the high-burnup plutonium fast breeder reactors appear unsuit- cured" facilities is denatured fuel. present in plutonium-based fuel cy- able for this role: with normal de- ~n some other proposa1s. any ex- cles. signs their fuel would have to be too cess plutonium-239 is burned on the Making enough uranium-233 to set enriched in fissile isotopes for de- secured sites in special burner re- up a self-sustaining pure thorium naturing to work. and if the core actors that lack fertile inventories. cycle also requires that a mixed were redesigned to permit denatur- This yields energy from the thoriumluranium-235 cycle first be ing, it would make too much plutonium but wastes neutrons: the operated for a very long time- plutonium-239 and not enough system is not a net fissile producer. typically several decades at least. uranium-233 for the cyck to work In a further variation due to von Such mixed cycles breed surplus properly. j Hippel, the pIutonium is never even uranium-233 as a by-product. The Regardless of the type of national extracted, but is discarded (under neutrons needed to do this and to reactor used, the thorium-232 would international controls) still mixed sustain fission in the mixed cycle be largely in a blanket (according to with the fission products. This must be derived from either highly most analyses) and the denatured makes immediate safeguards easier enriched uranium-235 or fissile uranium-233luranium-238 (lura- (the plutonium-bearing waste must plutonium-both weapons materials. nium-235) fuel mainly in a central still be protected from eventual re- Thus the persistent initial phase of core. The thorium-232 and covery} and is not such a great eco- any sort of thorium cycle inevitably uranium-238 respectively produce nomic loss because the high entails the presence, in some form, uranium-233 and plutonium-239 of plutonium-236 and -238 content of- of weapons materials, incurring in which part is burned in the reactor fers a disincentive to plutonium some degree the safeguards prob- and part discharged in irradiated extraction and reuse [13]. In a lems that the thorium cycle was fuel. The minitnwn net production of plutonium-throwaway cycle-which alleged to prevent. The denatured plutonium-2394onstrained by the can still be 5 to 6 times as uranium- Muitinational reprocessing plants would legitimize reprocessing and dissserninate its technology, somewhat like a hospitai spreading disease, for thorium cycles unlike uranium cycles require reprocessing. efficient as a once-through natural uranium with such cen- denatured thorium-cycle fuel can be pressurized-water reactor- trifuges would require patience and achieved not only with homemade reactivity in fresh fue1 going to na- tens or preferably hundreds of units centrifuges designed for this purpose tional reactors is supplemented as because each has a relatively small by persons moderately skilled in the needed, at least in the early stages, mass flow.) But in interpreting this art but also with modified versions of by makeup uranium-235. (This might incident it must be recalled that: commercially available medical ul- be added as a feedstock at about 60 e denatured thorium-cycle fresh tracentrifuges. (These may have percent enrichment-whose critical fuel is not 0.7 percent enriched in the large volumes, speeds over 75,000 mass is somewhat over twice that of fissile isotope as natural uranium is, revolutions per minute, and acceler- fully enriched uranium-235-and but rather about 12 percent-so that ations over half a ,million gravities.) then diluted with uranium-238 to a 85 percent of the separative work With patience, strategic quantities of denatured level. Thus the total needed to attain high enrichment has uranium-233 could be accumulated number of uranium-238 atoms in de- already been done. (For example, at with a few centrifuges (though a natured fuel would be at least 7 times constant 0.25 percent tails assay. it government would presumably use the number of uranium-233 atoms can be calculated that 6.8 times less hundreds and make more uranium- plus 4 tin~esthe number of uraniuth- separative work is needed to enrich 233). Since the holdup in each cen- 235 atoms.) The secured facilities to a kilogram of, say. 93.5 percent trifuge would be typically of order wo~ildthen contain no reactors at all, product starting with 12-percent- grams, the hard-gamma dosc even but only reprocessing, enrichment, enriched feed than with 0.71- for brief contact maintenance would and fuel fabrication plants. percent-enriched feed.) probably be acceptable to those con- Regardless of the disposition of e the mass difference between cerned (or at least to their the plutonium within the .'secured" uranium-233 and uranium-238 is supewisorsj-though in the long run sites, the important feature of these two-thirds larger than the mass dif- the concomitant enrichment in uranium-thorium cycles is that the ference between uranium-235 and uranium-232 may increase the dose. denatured fuel that leaves those sites uranium-238. In gas-diffusion en- It could be argued fairly persua- and forms an item of commerce is richment of hexahorides this means sively that conversion to uranium not directly usable for weapons nor, the theoretical separation factor per hexafluoride, centrifugal enrich- it is alleged, readily convertible to stage would be 1.0072 (that is, en- ment, and reconversion would be weapons material as mixed richment from 0.12 to 0.935 in 286 somewhat more difficult than extract- urani~tm-plutoniumoxides, carbides, ideal stages} rather than 1.0043 (in ing plutonium from fresh mixed- or nitrides would be. This allegation 478 stages), since diffusion velocities oxide fuel. But though different kinds rests on the assumption that isotopic vary as the square root of molecular of sophistication are required and the enrichment is not available to weight. With centrifugal enrichment, tasks are not strictly comparable, amateurs as chemical separation of however, maximum separation fac- they do not seem to differ qualita- fresh fuels presumably is. This as- tor varies exponentia1Iy with, and tively in difficulty, especially for sumption seems shaky, because the separative power of a particular high-burnup recycle plutonium. Nor "the technology to carry out such centrifuge varies as the square of, do the times required to perform enrichment on non-economical, differences in n~olecularweight. both tasks necessarily differ by as non-commercial scale is available in Hafele confirms [I71 that uranium- much as an order of magnitude. the open literature. This requires 233 is three times (about .!i2/3>)easier Denatused fuel, then, is not is- further evaluation" [14]. The rele- to enrich than uranium-235-and revocably denatured even for vant physical principles merit dis- separation factors for uranium-235 amateurs-let alone for govern- cussion here. can be of order 1.2 to 1.5 or more per ments. But the rest of the denatured Wolf Hafele has stated [I51 that stage. thorium cycle also lends itself rather '.the construction of a set of crude e as noted earlier, about 5 to 10 well to misuse. The national reactors centrifuges does not require unusual times less urani~im-233product is have, of necessity, an excellent skills as long as efficiency and com- required than uranium-235. neutron economy. CANDU or similar mercial competition are not the These three considerations reactors are likely in this role and, point." He added [16] that, to show suggest that denatured thorium- with their on-load refueling and this, a European centrifuge-builder cycle fuel is 100 to 200 times (= 7 x 3 slight excess of refueling-machine once made in an American univer- x 5-10) easier to enrich to bomb us- capacity, are well suited to clandes- sity machine shop (in about three able levels than natural uranium- tine introduction and brief irradia- weeks and with one technician) a two orders of magnitude more than tion of adequate amounts of fertile crude centrifuge capable of re- the difference between low-enriched materials (thorium-232 or uranium- spectably enriching natural uranium. and natural uranium. This factor im- 238) in a few channels for later clan- (In practice, making a bomb from plies that significant enrichment of destine reprocessing, perhaps in hot

February 1979 The Bulletin 19 It is important that thorium cycles not be euphorically and falsely advertised as a solution to the nuclear proliferation problem ...

cells. In fairness, one must add that and disseminate its technology, (and undemonstrated) plutonium-239 it may be easier to safeguard a re- somewhat like a hospital spreading cycles. actor than a reprocessing plant; but disease, for thorium cycles, unlike If one assumes that the world will the possibility of abuse remains, so uranium cycles, require reprocess- depend essentially forever on nu- that fully effective international in- ing. clear fission power on an enormous spection is still essential [18]. (If all these problems of the multi- scale, this may be important- The spent fuel itself, rich in both especially if one considers proneness plutonium-239 and uranium-233, and national centers were solved, one could also envisage a heavily spiked to proliferation less important. If, especially any blanket fuel, must be however, one assumes more modest safeguarded against diversion and plutonium fuel cycle which would then have broadly similar problems. rates of growth in nuclear capacity clandestine reprocessing-just what or electrical demand, or views nu- one is worried about with ordinary From a safeguards point of view the two cycles would appear similar-at clear power as a parenthesis-even thermal reactors, with or without one of a century or two-then breed- plutonium recycle or breeding. least as regards small-scale bomb production; the plutonium cycle ing gain is unnecessary, since (Most nuclear power countries can efficient converters can stretch today do laboratory-scale reprocess- would probably offer more opportu- nities for rapid large-scale bomb known uranium reserves far enough ing, suitable for one or a few bombs, [5]. production.) to suffice In particular, over the and could build a bomb-per-day next 75 to 100 years, advanced con- crude reprocessing plant in a year or It might be thought that centraliza- verters can save as much uranium as two for a cost one or two orders of tion of reprocessing and ofthe use of plutonium-239 breeders. The ad- magnitude less than the cost of a pure fissile materials might make vanced converters' relative in- single power reactor.) Prompt re- them easier to guard and assay; but it sensitivity to the price of uranium processing of irradiated thorium- can equally be argued that this cen- expands reserves, and anxieties over cycle fuel-though it must cope with tralization makes the "secured" site security of supply can be cheaply re- the high initial radioactivity-can a more vaiuabIe prize and a more lieved by stockpiling [21]. also readily extract pure vulnerable site to both accident and The mixed thorium/uranium-235 protactinium-233, which would then malice. It certainly makes the cycle has high initial requirements decay to isotopically pure uranium- farflung net of dependent national for uranium-235 and enrichment or 233 with negligible gamma back- reactors more vulnerable to all sorts for plutonium-239. These large first- ground [19]. And as with non- of surprises. core requirements can produce a denatured fuel cycles, governments dynamic materials-flow problem if can disguise their own diversions as Thorium and Nudear FueI Re- one tries to set up such a fuel cycle terrorist thefts, and can in dozens of sources. Recent calcujations- quickly. This is analogous to the effective but legal ways deceive preliminary but probably fairly problem of rapid construction of fast international inspectors or make it accurate-suggest [20] that no com- breeders; a plutonium shortage for impossible for them to do their job bination of national and multina- initial cores [22] is liable to arise un- properly. tional reactors that includes ad- less many thermal reactors are also Further, the "secured" sites vanced converters can have a total built and run in tandem with the would indeed have to be secured- conversion ratio greater than 1.0 breeders for many decades to pro- not only on paper-from both em- (that is, be a net producer of fissile duce their plutonium. The combined bezzlement and seizure of their ex- material)-except that a system population of fast and th~rnIa1re- tremely large fissile inventories. using a one-to-one converter like a actors is then likely to run short of Though both embezzlement and sei- molten-salt reactor might in principle uranium-235 (assuming this to be in zure can be made more difficult by have a low breeding gain. The Gen- short supply to start with). Staaing technical means, it appears to be im- eral Atomic Company also maintains with breeder cores of low specific possible in principle to deny them that a carefully optimized system plutonium inventory also means low altogether to a determined adver- using high-temperature gas-cooled breeding gain, thus uranium short- sary, The formidable political prob- reactors can have a significant breed- age, and substituting highly enriched lems of finding acceptable sites for ing gain [20]. Even breeder-breeder uranium for scarce plutonium still multinational centers and protecting combinations would have a small does not take the pressure off of them from instability and expropna- fissile gain, typically of order two to uranium supplies. tion would have to be overcome: three percent per year, and so would The reason this problem cannot be most analysts who have carefully take many decades to come to equi- evaded points up an important fea- studied this problem consider it in- librium. They thus appear incapable ture of thorium fuel cycles-a fea- soluble. Multinational reprocessing of yielding the rapid gains in fissile ture so simple it often goes un- plants would legitimize reprocessing inventow hoped for from advanced recognized. Using thorium does not . . . a problem th to reject all forms of nuclear power in favor of fission-free energy strategies.

magically expand the supply of nu- generous military subventions basis for public policy. Yet both con- clear fuel, only of fertile material, of throughout their establishment, are tain an important lesson. Advocates which we already have a great deal now having to pay their own way, of nuclear power should learn that in the form of uranium-238. No fer- and are finding the burden of fuei- there is no need to rush into the tile material, whether it be uranium- cycle investments unsustainable. plutonium economy [27], and that a 238 or thorium-232, can be bred into Thorium cycles would have the same wide range of alternative fuel cycles a fissile material save by neutrons. problem from the start: the visibility [28] may offer attractive substitutes These7the coin of the nuclear realm, of fu11 fuel-cycle costs could be a for it later. It is equally important must be derived. ultimately from most ineluctable deterrent. that thorium cycles not be euphon- hranium-235 (either directly or via Among the criteria that Cochran cally and falsely advertised as a solu- bred fissile matei-ials such as and others proposecl [5] as a minimal tion to the proliferation problem-a plutonium-239) [23]. It is therefore and partial basis for acceptability of problem that is the most compelling the supply of uranium-235 in the commercial fuel cycles is that their reason to reject all forms of nuclear ground that poses nuclear fuel con- power in favor of fission-free energy straints. The only sense in which "deve1opment and commercial utili- stixtegies 1291. using thorium helps with the nuclear zation of the technology by a non- fuel problem is that thorium-232 can nuclear weapons state leaves that be bred with thermal neutrons (and 1. This formu!ation is due to Lew KOW- state no closer to a nuclear weapons arski. I am indebted for discussions and cor- can then fuel efficient thermal con- capability than would be the case if respondence to him+ Tom Cochran. Floyd verters), whereas uranium-238 can all its nuclear power were derived Culler, Warren Donnelly, Peter Fortescue, Sir be most efficiently bred with fast from low-enriched uranium fueled Brian Flowers. David Hafemeister, Walt Pat- neutrons. The former process may reactors operating in a once-through terson, Ted Taylor, and Frank von Hippel. many of whom suggested significant im- be more convenient, though the lat- fuel cycle . . . and with verified provements in an earlier draft. 1 alone am re- ter has higher neutronic spent fuel storage in secured inter- sponsible For the results. efficiency-a property not wholly national facilities' ' [25]. 2. See, for example, references 3 and 5; A. unlinked io its relevance to bombs. B. Lovins, Sofr Energy Paths: Toward a And the fad-breeding ability of Durable Peace (New York: Harper & Row, It appears very doubtful that de- 1979), especially Chap. 11; International uranium-238 is irrelevant if, as ap- natured thorium cycles could meet Technology Project, Institute of International pears to be the case, the available this criterion. Certainly they cannot Studies, University of CaIifornia, Berkeiey, uranium-235 is ample [5] when used greatiy improve the present Non-Prolgeration and Nuclear VVas~eMan- in thermal converters. safeguards situation; and for reasons agement, Contract AC6AC72S (Washington, D.C.: U.S. Arms Control & Disarmament People who wish to ascribe magi- similar to those set out above, Theo- Agency, 1977); A. Wohlstetter et al., Moving cal properties to thorium cycles dore Taylor has recently described Toward Lge in A Nuciear Armed Cro~txd?, should also remember that the above the notion that they can as ''a snare ACDAIPAB-263 {Los Angeles, Ca.: Pan description of a denatured thorium and a delusion'' [26]. A more mature Heuristics, 1977); WohlstetterandT. B. Coch- cycle is theoretical. No such cycle ran, respective Proofs of Evidence for the assessment by the technical commu- summer 1977 Windscaie inquiry (U.K.). has operated anywhere, though nity ofthe grounds for his skepticism 3. H. A. Feiveson and T. B. Taylor, "Se- there has been minor use and re- will probably lead many to share it. curity Implications of Alternative Fission FU- processing of thorium fuef at pilot And he and 1 would both emphasize tures:'* Bit//eiin 32: I0 (December 1976), scale. Dissolution and extraction are that the safeguards/proliferation 14-18, 46-8. 4. Of course, uranium cycles can be isotop- somewhat harder than for low- problem has no iechnicd fix. It is ically denatured too-as is the conventional burnup uranium dioxide fuel [24]. It partly, even predominantly, a puliti- once-through !ow-enriched uranium cycle. is possible that actual behavior of a cal problem. Facile assumptions [6] 5. T. B. Cochran. R. E. Train, F, von Hip- thorium cycle might depart sig- about technical fixes or near-fixes pel, and R. H. Williams, '.Proliferation Re- nificantly from the above results of sistant Nuclear Power Technologies: Pre- will not solve the political problem ferred Alternatives to the PIutonium preliminary calculations, or that un- and might well make it worse by giv- Breeder,'' including additional statements by expected technical problems might ing a false sense of security. T. B, Cochran and R. E. Train, part of the arise as they have done persistently Thorium cycles have had the mis- report of the LMFBR Review Steering Commit- with moderate-burnup plutonium fortune to be attacked both by tee (R. Thorne! Chairman}, report to the Pre- sident, April 6, 1977. cycles. It does seem rather late in the pIutonium advocates (for competing 6. See, for example, Sir Fred Hoy!e, En- day to start with a whole new set of with their favorite technology) and ergy or Exiincrion? (London: Heinemann fuel physics and chemistry. Of by nuclear critics (for having the po- Educational, 19771, pp, 50-1. course, just the same could be said tential, in the hands of the un- 7. H. C. Paxton et al., "Critical Di- of high-burnup plutonium cycles- mensions of Systems Containing U235,fi239, informed, to give nuclear power an and U233,''TID-7028 (Los Alamos Scientific which, unlike their low-burnup undeserved new lease on life) [I]. Laboratory, June 1944): R. W. Selden, "Re- predecessors that have rested on Neither basis for attack is a sound actor Plutonium and Nuclear Explosives"

February f979 The Bulletin 21 (Lawrence Livermore Laboratory, Nov. 1, tion factor, respectively, of 3.0 kilograms of 36, S 153). 1976); Reactor Physics Constants, ANL-5800 separative work per year and 1.10 for 25. Verified fresh-fuel storage would also (2d ed.), Argonne National Laboratory, 1963. uranium-235 (8.3 kilograms of separative be required to prevent diversion lo clandes- 8. E. Teller, in S. A. Blumberg and G. work per year and 1-17 for uranium-233). Gas tine enrichment facilities or production re- Owens, Energy and Conjlkt: T%e Age and centrifuges with better performance than this actors. Times of Edward Teller (New York: htnam, and costing around $10,000 or less are de- 26. T. B. Taylor to Lovins, Sept. 30, 1977. 19761, Appendix. scribed in the open literature. 27. S. M. Keeney el al,, Nuclear Power: 9. M. J. Bell and J. P. Nichols, "Penetrat- 18. APS Study Group (reference 10, p. S 1ss~ie.sand Choices (Cambridge, Mass.: Bal- ing Radiation Dose Rates and Shield Re- 1021 states: "It must be remembered that with linger, 1977): H.A. Feiveson et al., "The quirements in Fabrication of Fuels Containing any denatured fuel cycle, international inspec- Plutonium Economy: Why We Should Wait 233U and High Exposure Plutonium," in tion is still essentid in guaranteeing that fresh and Why We Can Wait," Bulletin 32:iO (Dec. Sol-Gel Processes and Reactor Fitel Cycles, fuel and/or natural uranium are not being di- 1976), 10 ff; Feiveson, "Proliferation Re- CONF-700502 (Oak Ridge National Labora- verted to small enrichment facilities, that resistant Fiicl Cycles," Anniiaf Re~iwof En- tory, 1970), pp. 74-84. For dose vs. aging, see actors are not being misused to produce ergy-. 3 (1978), 357-94. See also references 2 and reference SO. p. S 154, figure 11. plutonium> and that small non-safeguarded At. The BNL paper cited in reference 17 states separation Ptcilities are not operated in addi- 28. See reference 27, Cochran [S], APS on page 23 that for uranium-233 dioxide with 1 tion to those at international centers. Such in- Study Graup [lo]. Unfortunately the discus- gram of uranium-232 per kilogram, a 100- spection goes beyond present IAEA sion of alternative fuel cycles may be tem- roentgen dose will be accumulated in 3 days, 5 agreements but must become part of any fu- porarily derailed by sophistical discussions of hours, and 2 hours respectively, if an individ- ture political arrangements which in- plutonium cycles based on breeder reactors ual is 1 foot from 30 kilograms of uranium stitutionaiize denatured fuel cycles." that do not breed. dioxide aged 0, 10, and 20 days since chemical The International Atomic Energy Agency If the original goal of high breeding gain is separation. report of June 8, 1977 to the IAEA Board of to be abandoned (see. for example, W. Mar- 10. American Physical Society, Study Governors, GUVil842. calls attention in its shall, "Nuclear Power and the Proliferation Group on Nitdear Fziel Cycles and Wa.ae Annex to the serious and, so far, un- issue" (London: United Kingdom Atomic En- Management,Re~*je~~sofModern Physics 50: 1 surmounted difficulties of effectively ergy Authority, Feb. 19781, fast reactors be- (Jan. 19781, Part 11, p. S 154. safeguarding both reprocessing plants and come unnecessary; and if it is felt desirable to 11. See reference 10, p. S. 171, citing T. Pig- on-load rcfnciing reactors. incinerate stbcks of plutonium (and perhaps ford and C. Yang, "Thorium Fuel Cycles,'. 19. APS Study Group (reference 10, p. S other actinides), that are already separated, UCB-NE3227 (Berkeley: University of 154) referred in its prepublication preprint this can be done in low-powcr-density thermal California. June 1977). The authors calcu- (VIII. F 1) to "a relatively simple chemical reactors with a mn-fertile matrix. Spent fuel recycled uranium from a high-temperature separation." H. C. Ott states (Po)\,er En- not yet reprocessed can be disposed of irrc- gas-cooled reactor as about 500 milligrams gineering, Nov. 1977. p. 28) that the minimum trievably (one hopes) in the manner intended per kilogram. but predict as high as 9 grams protactinium-233 assay-that characteristic of for vitrified high-level wastes. ,45 for per kilogram for a uranium-thorium light a pressurized water rcactor-is 3.8 percent of plutonium not yet generated in thermal re- water reactor fuel cycle. the fissile uranium assay. actors, the opportunity of generating it in 12. APS Study Group (reference 10, p. S 20. F. L. Culler, Jr., to Lovins, May 1977; order to be able to build non-breeding fast re- 155) says a factor of about seven in idealized the calculations, mainly by Argonne National actors in which to incinerate it can simply be fuel cycles, but notes that about three is pos- Laboratory, are to be published in connection foregone. sible in "&re practical designs." with supporting research for the international L. Kowarski has aptiy remarked that the 13. APS study Group (reference 10, p. S Fuel Cycle Evaluation. See also C. L, Rickard Marshall proposal reminds him of the entre- 171). The plutonium-236 and-238 are formed (General Atomic Co.), "The Thorinm Fuel preneur whose herd of cows gttve milk copi- largely from neptunium-237. Cycle," AIF Fuel Cycle Conference, Yew ousiy while there was a glut of it and Europe York, March 5-8, 1978. was already sinking under a mountain of pow- 14. APS Study Group (reference 10, p. S 21. F. von Hippel and R. H. Williams have dered miik. He said, "Don't worry: with 102. S 156). calculated that each ten years' worth of high-pressure salesmanship 1 shall dispose of 15. W. Eafeie, "Non-Proliferation of Nu- uranium stockpile bought at $40 per pound of all my rniik anyway.'. On being told, how- clear Weapons'' (LaxenburgsAustria: Inter- u3ou incurs a carrying charge (at 16 percent ever* that the milk was full of cholesterol and national Institute for Applied Systems per annumj that raises the delivered price of might hence be dahgerous to drink, he said. Analysis, Dcc. 1976). p. 10. nuclear electricity by an amount ranging from "That is more serious. But I have a solution: 1 16- W. HiLfele to Lovins- Feb. 16, 1977. 0.9 percent (once-through PWR) to 0.2 per- shall simply feed the milk lo thc cows. They 17. W. Hgfele, in Arg~irnenrein der Ener- cent (heavy water reactor with uranium-233 can be trained to drink vast quantities of it!" giediskussion: i. Schnelle Briifer: Pro lind and-235 recycle and throwaway plutonium) 29. See, for example, A. B. Lovins [2]; for Contra edited by H. Matthofer (Necker- (H. A. Feiveson, '.Proliferation Resistant a compendium of critiques and responses, see Verlag, Villingen, Oct. 19771, pp. 44-45. Fuel Cycies." Annual Review of Energy 3 U.S. Senate, Select Committee on Small BUS- For a discussion of separation factors for (1978). 391, Tabie 8 1. iness and Committee on Interior & Insular Af- uranium-233 vs uranium-235, see Brookhaven 22. L. Grainger, Energy Policy 4 (Dec. fairs, Alternative Long-Range Energy Sfrule- National Laboratory, "Preliminary l976), 322-9: D. Merrick, Nature 264 ( 1976) gies. 2 vols. (Washington, D.C., U.S. Gov- Safeguards Analysis of Denatured Thorium 596-8. ernment Printing Office, 19771, condensed by Cycles," file #5.8.13 (Upton, N.Y.: BNL 23. Surplus neutrons might in principle be H. Nash in The Energy Cut~troversy:.SO$ Technical Support Organization, Nov. 10, available from sources other than pure fission Path Questions and Ansb$-crs (San Francisco, 1976). An addendum (July 14, 1977) calculates cycles-for example, spallation reactions in Ca.: Friends of the Earth. 1979); Lovins, that five kilograms per year of 90 percent accelerators, or thermonuclear or near- "Re-Examining the Nature of the ECE En- uranium-233 can be separated (at I percent thermonuclear reactors-but such sources are ergy Problem," ECE (XXXIII)/2/I.G- tails assay) from 49 kilograms of uranium per neglected here because of their uncertain or (Geneva: U.N. Economic Commission for year that is 10% uranium-233 and 90 percent unfavorable feasibility, lead times, energy Europe. Feb. 20, 1978): Lovins, ,*Soft Energy uranium-238 with an ideal cascade of 43 cen- balance, and economics. Technologies," Annual Review of Energy 3 trifuges, each having a capacity and a separa- 24. APS Study Group (reference 10, pp. .S (1978) 477-5 17.