T he Economics of Reprocessing vs . D irect D isposal of S pent Nuclear F uel

Final Report 8/12/1999-7/30/2003

Matthew Bunn Steve Fetter John P. Holdren

Bob van der Zwaan

December 2003

DE-FG26-99FT4028

P roje ct on M anaging the A tom

B elfer C enter f or S cience and Intern ational A ffairs J oh n F. K ennedy S chool of G overnment

H arvard U niversity

79 J ohn F. K ennedy S tr eet C ambri dge, M assachuse tts 02138

© 2003 President and Fellows of Harvard University Printed in the United States of Am erica

This report was prepared as an account of work sponsored by an agency of the United States Governm e nt. Neither the United States Governm e nt nor any agency thereof, nor any of their em ployees, m akes any warranty, express or im plied, or assum es any legal liability or responsibility for the accuracy, com p leteness, or usefulness of any inform ation, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific com m ercial product, process, or service by trade nam e, tradem ark, m anufacturer, or otherwise does not necessarily im ply its endorsem ent, recom m endation, or favoring by the United States Governm e nt or any agency thereof. The vi ews and opinions of authors expressed herein do not necessarily state or reflect those of the United States Governm e nt or any agency thereof.

The authors of this report invite liberal use of the inform ation provided in it for educational purposes, requiring only that the reproduced m aterial clearly state: Reproduced from Matthew Bunn, Steve Fetter, John Holdren, and Bob van der Zwaan, The Economics of Reprocessing vs. Direct Disposal of Spent Nuclear Fuel (Cam bridge, Mass.: Project on Managing the Atom , Harvard University, 2003).

Project on Managing the Atom

Belfer Center for Science and International Affairs

John F. Kennedy School of Governm e nt Harvard University

79 JFK Street

Cam bridge, MA 02138

W eb: http://www.ksg.harvard.edu/bcsia/atom

Abstract

This report assesses the econom ics of reprocessing versus direct disposal of spent nuclear fuel. The breakeven uranium price at which reprocessing spent nuclear fuel from existing light-water reactors (LW R s) and r ecycling the resulting plutonium and uranium in LW Rs would becom e econom ic is assessed, using central estim ates of the costs of different elem ents of the nuclear fuel cycle (and other fuel cycle input param e ters), for a wide range of range of potential reprocessing prices. Sensitivity analysis is perf orm e d, showing that the conclusions reached are robust across a wide range of input param e ters. The contribution of direct disposal or reproce ssing and recycling to electricity cost is also assessed. The choice of particular central estim ates and ranges for the input param e ters of the fuel cycle m odel is justified through a re view of the relevant literature. The im pact of different fuel cycle approaches on the volum e needed for geologic repositories is briefly discussed, as are the issues surrounding the possibility of perform i ng separations and transm utation on spent nuclear fuel to reduce the need for additional repositories. A sim ilar analysis is then perform e d of the breakeven uranium price at which deploying fast-neutron breeder reactors would becom e com p etitive com p ared with a once-through fuel cycle in LW Rs, for a range of possible differences in capital cost between LW Rs and fast-neutron reactors. Sensitivity analysis is again provided, as are an analysis of the contribution to electricity cost, and a justification of the choices of central estim ates and ranges for the input param e ters. The equations used in the econom ic m odel are derived

and explained in an appendix. Another appe ndix assesses the quantities of uranium likely to be recoverable worldwide in the future at a range of different possible future prices.

iii

Table of Contents

1.1. W h at Is Reprocessing?

1.2. Data and Sources

1.3. Cost vs. Price

1.4. Currency Conversion

1.5. Cost of Money, Discount Rate, and Taxes

1.6. Real vs. Nom i nal Dollars

1.7. Plan of the Report

2. Direct D i sposal vs. Reprocessing and Rec y cling in Thermal Reactors 13

2.1. How to Com p are Costs of Different Fuel Cycles

2.2. Calculating Breakeven Prices

2.3. Breakeven Price Sensitivity Analysis

2.4. Contribution to the Cost of Electricity

2.5. Com ponent Costs of the Fuel Cycle

2.5.1. Uranium Prices

2.5.2. Reprocessing Costs and Prices

2.5.3. Costs of Disposal of Spent Fuel and Reprocessing W a stes

2.5.4. Costs and Prices for Mixed Oxide Fuel Fabrication and Use

2.5.5. Costs of Interim Storage of Spent Fuel

2.5.6. Enrichm e nt Prices

2.5.7. Low Enriched Uranium Fuel Fabrication Prices

2.5.8. Prem ium s for Handling Reprocessed Uranium

2.5.9. Conversion Prices

2.5.10. Non-Price Factors: Fuel Burnup, Discount Rate

Sidebar: Volumes of Wastes From Direct Disposal and Reprocessing Sidebar: Reprocessing to Reduce the Need for Additional Repositories

3. Direct Disposal vs. Recyclin g in Fast-Neutron Reactors 67

3.1. Plutonium Breeding and Recycling in Fast Reactors

3.2. Breakeven Uranium Price for Recycling in Fast Reactors

3.3. Cost of Electricity for Fast Reactors and Once-Through System s

3.4. Cost Param e ters and Variations

3.4.1. Difference in Capital Cost

3.4.2. Reactor Ownership and Financing Arrangem e nts

3.4.3. Reprocessing Costs

v

3.4.4. Core and Blanket Fuel Fabrication Costs

3.4.5. Geological Disposal of Reprocessing W a ste

3.4.6. Breeding Ratio

3.4.7. Depleted Uranium Price

Sidebar: Thermal Neutron and Fast-Neutron Reactors Sidebar: Characteristics of the Model Fast Reactor

4. Conclusions 87

Appendix A. Fuel Cycl e Cost Calculations 89

A.1. Direct Disposal vs. Reprocessing and Recycle in LW Rs

A.1.1. Direct Disposal

A.1.2. Reprocessing-Recycle

A.1.2.1. Value of Recovered Plutonium

A.1.2.2. Value of Recovered Uranium

A.1.3. Uranium Breakeven Price

A.2. Direct Disposal vs. Recycling in Fast-Neutron Reactors

A.2.1. Capital Cost

A.2.1.1. Interest During Construction

A.2.2.2. Fixed Charge Rate

A.2.2. Operations and Maintenance Cost

A.2.3. Fuel Cost

A.2.3.1. LW R Fuel

A.2.3.2. LMR Fuel

A.2.4. Breakeven Uranium Price

Appendix B. World Uranium Resources 105

B.1. Introduction

B.2. Fallacy of the Traditional Econom ic Resource Model

B.3. Estim ates of Uranium Resources

B.4. Uranium From Seawater

B.5. Uranium Consum ption

vi

List of Figures and Tables

Figure 2.1. Breakeven uranium price as a function of the cost of reprocessing 18

Table 2.1. Estim ates of fuel cycle costs (2003 dollars) and other param e ters 19

Table 2.2. Breakeven p r ices of selected param e ters 20

Figure 2.2. Sensitiv ity o f the uraniu m breakeve n price 21

Figure 2.3. Additiona l c o st of elec tricity f o r th e reprocess i ng-recycle option 22

Figure 2.4. Uranium prices, 1972-2000 24

Table 2.3. Notional cost reduction f o r disposal of reprocessing wastes 42

Figure 3.1. Breakeven uranium price for governm e nt-owned reactors 69

Figure 3.2. Breakeven u r anium price f o r utility-o wned reacto rs 70

Figure 3.3. Breakeven uranium price for private venture ownership 70

Table 3.1. Sensitiv ity a n alysis f o r the breakev e n uranium price 71

Table 3.2. Breakeven p r ice of selected param e ters 73

Figure 3.4. Difference in the cost of electricity between an FR with recycling and an LW R with direct disposal 74

Table A.1. Isotopic composition of fresh and spent LEU 92

Table A.2. Isotopic composition of fresh MOX fuel 93

Table A.3. Optim um tails assay 95

Table A.4. Fixed charge rates 101

Table B.1. Typical u r an ium concentrations 106

Table B.2. Exponential uranium resource estim ates 113

vii

Acknowledgements

This report was prepared with the support of the U.S. Departm e nt of Energy (DOE), under Award No. DE-FG26-99FT40281. However, any opinions, findings, conclusions, or recom m e ndations expressed herein are those of the authors and do not necessarily reflect the views of DOE. Additional funding was provided by the John D. and Catherine T. MacArthur Foundation. The authors would like to thank Laure Mougeot for extensive research assistance in the early phases of the project, and Brian Torpy, Annaliis Abrego, and Anthony W i er for additional research assistance. W e are grateful to Professors Mike Driscoll and Richard Lester of the Massachusetts Institute of Technology (MIT) for discussions of the f i ner points of cost levelization in the presence of corporate incom e taxes; to Marvin Miller of MIT for discussions of uranium resources; to Nigel Mote of International Nuclear Consultants and Geoff Varley and Dan Collier of the Nuclear Assurance Corporation, for discussions of the costs of som e elem ents of the nuclear fuel cycle; and to David W a de of Argonne National Laboratory for data and discussi ons relating to recent analyses of recycling in fast-neutron reactors. W e are also gratef ul to Driscoll, Chaim Braun, Jor-Chan Choi, and Per Peterson for useful com m e nts on an earlier draft. All responsibility for rem a ining errors and m i sjudgm ents, of course, is our own.

viii

Executiv e Summary

For decades, there has been an intense debate over the best approach to m a naging spent fuel from nuclear power reactors—whether it is better to dispose of it directly in geologic repositories, or reprocess it to recover and recycle the plutonium and uranium , disposing only of the wastes from repro cessing and recycling. The relative costs of reprocessing vs. not reprocessing are one im portant elem ent of these debates. Econom ics is not the only or even the principal factor affecting decisions concerning reprocessing today. But econom ics is not unim portant, par ticularly in a nuclear industry facing an increasingly com p etitive environm ent. At a m i nim u m , if reprocessing is being done to achieve objectives other than econom ic ones, it is worthwhile to know how m u ch one is paying to achieve those other objectives.

W h ile som e analysts have argued in recent years that the costs of reprocessing and direct disposal are sim ilar, and that reprocessing will soon be the m o re cost-effective approach as uranium prices increase, the data and analyses presented in this report dem onstrate that the m a rgin between the cost of reprocessing and recycling and that of direct disposal is wide, and is likely to persist for m a ny decades to com e .

In particular:

 At a reprocessing price of $1000 per kilogram of heavy m e tal (kgHM), and with our other central estim ates for the key fuel cycle param e ters, reprocessing and recycling plutonium in existing light-water reactors (LW R s) will be m o re expensive than direct disposal of spent fuel until the uranium price reaches over $360 per kilogram of uranium (kgU)—a price that is not likely to be seen for m a ny decades, if then.

 At a uranium price of $40/kgU (com parable to current prices), reprocessing and recycling at a reprocessing price of $1000/ kgHM would increase the cost of nuclear electricity by 1.3 m ills/kW h. Since the total back-end cost for the direct disposal is in the range of 1.5 m ills/kgW h, this represents m o re than an 80% increase in the costs attributable to spent fuel m a nagem e nt (aft er taking account of appropriate credits or charges for recovered plutonium and uranium from reprocessing).

 These figures for breakeven uranium price and contribution to the cost of electricity are conservative, because, to ensure that our conclusions were robust, we have assum e d:

 A central estim ate of reprocessing cost, $1000/kgHM, which is substantially below the cost that would pertain in privately f i nanced f acilities with identical costs and capacities to the large com m e rcial facilities now in operation.

 A central estim ate of plutonium fuel fabrication cost, $1500/kgHM, which is significantly below the price actually offered to m o st utilities in the 1980s and 1990s.

 Zero charges for storage of separated plutonium or rem oval of am ericium .

 Zero additional security, licensing, or shut-down expenses for the use of plutonium fuels in existing reactors.

 A full charge for 40 years of interim storage in dry casks for all fuel going to direct disposal, and no interim storage charge for fuel going to reprocessing—

ix

even though m o st new reactors are built with storage capacity for their lifetim e fuel generation, so few additional costs for interim storage need be incurred.

 Geological disposal of spent MOX fuel at the sam e cost as disposal of spent LEU fuel .

 Reprocessing and recycling plutonium in fast-neutron reactors (FRs) with an additional capital cost, com p ared to new LW Rs, of $200/kW e installed will not be econom ically com p etitive with a once-through cycle in LW Rs until the price of uranium reaches som e $340/kgU, given our central estim ates of the other param e ters. Even if the capital cost of new FRs could be reduced to equal that of new LW Rs, recycling in FRs would not be econom ic until the uranium price reached som e

$140/kgU.

 At a uranium price of $40/kgU, electricity from a plutonium -recycling FR with an additional capital cost of $200/kW e , and with our central estim ates of the other param e ters, would cost m o re than 7 m ills/kW h m o re than electricity f r om a once- through LW R. Even if the additional capital cost could be elim inated, the extra electricity cost would be over 2 m ills/kW h.

 As with reprocessing and recycling in LW Rs, these figures on breakeven uranium price and extra electricity cost for FRs are conservative, as we have assum e d:

 Zero cost for providing start-up plutonium for the FRs.

 Zero additional cost for reprocessing higher-plutonium -content FR fuel.

 Zero additional cost for m a nufacturing higher-plutonium -content FR fuel.

 Zero additional operations and m a intenance costs for FRs, com p ared to LW Rs.

 Costs for the far m o re com p lex chem ical separations processes and m o re difficult fuel fabrication processes needed for m o re com p lete separation and transm utation of nuclear wastes would be substantially higher than those estim ated here for traditional reprocessing. Therefore the extra electricity cost, were these approaches to be pursued, would be even higher. Argum ents for separations and transm utation to lim it the need for additional repositories rest on a num ber of critical assum p tions that m a y or m a y not be borne out in practice.

 W o rld resources of uranium likely to be econom ically recoverable in future decades at prices far below the breakeven price am ount to tens of m illions of tons, probably enough to fuel a rapidly-growing nuclear en terprise using a once-through fuel cycle for a century or m o re.

In this report, we have focused only on the econom ic issues, and have not exam ined other issues in the broader debate over reprocessing. Nevertheless, given (a) the costs outlined above; (b) the significant proliferation concerns that have been raised (particularly with respect to those reprocessing approaches that result in fully separated plutonium suitable f o r use in nuclear explosives); and (c) the availability of saf e , proven, low-cost dry cask storage technology that will allow spent fuel to be stored for m a ny decades, the burden of proof clearly rests on thos e in favor of investing in reprocessing in the near term .

x

1 . Introduction

For decades, there has been an intense debate over the best approach to m a naging spent fuel from nuclear power reactors—whether it is better to dispose of it directly in geologic repositories, or reprocess it to recover and recycle the plutonium and uranium , disposing only of the wastes from reprocessi ng and recycling. These debates have becom e even m o re salient in recent years, as increasing accum u lations of both spent nuclear fuel and separated plutonium from reprocessing generate increasing concern worldwide. Countries that have chosen to reprocess are f acing high costs and rising political controversies, while m a ny of those that have chosen not to reprocess are f acing signif i cant political obstacles to providing adequate storage space for spent fuel. No country in the world has yet opened a perm anent repository for either spent nuclear fuel or the high-level wastes from reprocessing. In several countries, proposals to separate a nd transm ute not only plutonium and uranium , but other long-lived radioactive m a terials in spent fuel as well, have gained increasing attention in recent years.

The relative cost of reprocessing vs. direct-disposal is an im portant elem ent of these debates. Econom ics, of course, is not the only or even the principal factor affecting decisions concerning reprocessing today—the inertia of fuel-cycle plans and contracts initiated long ago, hopes that plutonium recycling will contri bute to energy security, lack of adequate storage space for spent fuel, environm ental concerns, and other factors also play critical roles. 1 But econom ics is not unim portant, particularly in a nuclear industry facing an increasingly com p etitive environm ent, where the difference between producing electricity at slightly higher or lower cost than com p etitors is the dif f e rence between bankruptcy and profit, and where fuel-cycle costs are am ong the few costs reactor operators can readily control. At a m i nim u m , if reprocessing is being done to achieve objectives other than econom ic ones, it is worthwhile to know how m u ch one is paying to achieve those other objectives.

There is general agreem ent in recent studies that with today’s low uranium and enrichm e nt prices, reprocessing and recycling is m o re expensive than direct disposal of spent fuel . 2 The only argum ent is over the m a gnitude of the difference and how long it is likely to

1 For a useful (t hough now som e what dat e d) overvi e w of reprocessi ng and recy cl i ng of pl ut oni um i n count ri es around t h e worl d, wi t h project i ons for t h e fut u re and so m e suggest i ons for pol i c i e s t o address t h e rel e vant issues, see David Albright, Fr ans Berkhout, and W illiam W a lker, Pl ut oni um and Hi ghl y Enri ched Urani um 1996: World Inventories, Capabilities, and Policies (Oxford, UK: Oxford Uni v ersi t y Press for t h e St ockhol m International Peace Research Institute, 1997).

2 The st udi es on t h i s t opi c are t oo num erous t o l i s t here. Offi ci al st udi es are of part i c ul ar i n t e rest . For exam pl e, a recent study for the French governm e nt com p ared a scenario in which all of the low-enriched uranium fuel produced in French reactors was repro cessed to a hypothetical scenario in which reprocessing and recycling had never been introduced, and found that not reprocessing would have saved tens of billions of dolla rs com p ared to t h e al l -reprocessi ng case, and woul d have reduced t o t a l el ect ri ci t y generat i on cost s by m o re t h an 5 percent . See Jean-M i c hel C h arpi n, B e njam i n Dessus, and R e né Pel l a t , Economi c Forecast St udy of t h e N u cl ear Pow e r Opt i on (Paris, France: Office of the Prim e Minister, July 2000, available as of Decem ber 16, 2003 at http://fire.pppl.gov/eu_fr_fission_plan.pdf ), Appendi x 1. The 1994 st udy by t h e Nucl ear Energy Agency of t h e Organi zat i on for Econom i c C ooperat i on and Devel opm ent , The Economi c s of t h e N u cl ear Fuel C ycl e (Paris, France: OEC D / N EA, 1994), whi l e fi ndi ng a t o t a l fuel cy cl e cost onl y about 14% great er for t h e reprocessi ng

2 E conomics of R eproce ssing vs. D irect D isposal of S pe nt N uclear F uel

persist. Advocates of reprocessing often argue that the extra cost of reprocessing is sm all today, and will soon disappear as uranium supplies becom e scarce and their price rises. 3 The data and analyses presented in this report, by contrast, dem onstrate that the m a rgin between the cost of reprocessing and recycling and that of direct disposal is wide, and is likely to persist for m a ny decades to com e .

These issues are increasingly im portant, as a num ber of countries face m a jor decisions about future m a nagem e nt of their spent fuel. In the United States in particular, the Bush adm i nistration has supported developm ent of new reprocessing approaches that, it is argued, m i ght be m o re proliferation-resistant than previous ones, while m i nim i zing nuclear wastes; the Departm e nt of Energy plans to spend several hundred m illion dollars over the next several years on research and developm ent related to reprocessing in the Advanced Fuel Cycle Initiative, 4 although the idea of building a large (1500 m e tric tons of heavy m e tal per year) aqueous reprocessing plant in the United States around the m i ddle of the next decade has apparently been abandoned. 5

1.1. What Is Reprocessing?

Reprocessing does not elim inate any of the radioactive m a terial in spent fuel—it m e rely divides that m a terial into several cate gories (plutonium , uranium , and various types of

opt i on, found a back-end cost t w i ce as hi gh for t h e reprocessi ng opt i on as for t h e di rect di sposal opt i on. In 2003, a m a jor study from the Massachusetts Institute of Technology (MIT) (i n which one of the present authors (Holdren) participated) cam e to conclusi ons quite sim ilar to those we reach in this study. See John Deutch and Ernest J. M oni z, co-chai r s, The Future of Nuclear Power: An Int e rdi s ci pl i nary MIT St udy (C am bri dge, M A : Massachusetts Institute of Technology, 2003, available as of Decem ber 16, 2003 at h ttp ://web .m it.ed u / n u c learp o w er ). The M I T st udy present s i t s resul t s by consi d eri ng t h e cost of reprocessi ng as part of t h e cost of prepari ng pl ut oni um fuel (and t h e pl utonium fuel therefore appears several tim es as expensive as urani u m fuel of equi val e nt energy val u e), whi l e we present our resul t s wi t h reprocessi ng count ed as part of t h e cost of wast e m a nagem e nt – but t h i s di fference i n present a t i on does not affect t h e cont ri but i on of reprocessi ng and recy cl i ng t o t o t a l el ectricity cost. (The MIT study’s central estim ate of the increase in el ect ri ci t y pri ce resul t i ng from use of reprocessi ng rat h er t h an once-t h rough fuel cy cl es i s hi gher t h an t h e one i n this study, prim arily because they do not assign an extra cost for several decades of dry cask storage of spent fuel for t h e once-t h rough cy cl e, as we do.) The RAND corporat i on al so produced a com m onl y ci t e d st udy of t h i s subject i n t h e earl y 1990s: see B r i a n G. C how and Kennet h A. Sol o m on, Li mi t i ng t h e Spread of Weapons- Usable Fissile Materials (Sant a M oni ca, CA: RAND, 1993). A useful sum m a ry of st at em ent s and st udi es on t h i s subject from t h e m i d-1990s and before can be found i n Ingo Hensi ng and W a l t e r Schul z, An Economi c Comparison of Different Disposal Met hods Used by Nuclear Power Plants: A Cost Simulation of Alternative St rat e gi es From t h e German Poi n t of Vi ew Energiewirtschafliche s Institute (EW I), University of Cologne (Ol e nbourg-Vourl a g,1995), whi c h al so fi nds si gni fi cant l y hi gher cost s for t h e reprocessi ng fuel cy cl e.

3 For a typical version of the argum ent, see Jam e s Lake (president of the Am erican Nuclear Society), “Outdated Thi nki ng i s Hol d i ng Us B ack,” The Washi ngt on Post , M a y 12, 2001, whi c h assert s t h at “t he econom i c t r ade-off i s approxi m a t e l y equal ” t oday , and t h at for t h e fut u re, reprocessi ng offers “si gni fi cant advant ages i n sust ai ni ng low-cost nuclear fuel supplies.”

4 See, for exam ple, U.S. Departm e nt of Energy (DOE), FY 2004 Detailed Budget Justifications—Office of N u cl ear Energy, Sci e nce, and Technol ogy (W ashington, D.C.: DOE, February 2003; available as of Decem ber 16, 2003 at ht t p : / / www.m b e.doe.gov/ budget / 04budget / c ont ent / e s/ nucl ear.pdf ), p. 18 and p. 45; see al so DOE, Offi ce of Nucl ear Energy , Sci e nce, and Technol ogy , Report to Congress on Advan ced Fuel Cycle Initiative: The Fut u re Pat h f o r Advanced Spent Fuel Treat m ent and Transmut a t i on Research (W ashi ngt on, DC : DOE, January, 2003, available as of Decem ber 16, 2003 at ht t p : / / www.nucl ear.gov/ report s / A FCI_CongRpt 2003.pdf ). 5 Ernest J. M oni z, present a t i on at t h e Second M o scow Int e rnat i onal C onference on Nonprol i f erat i on, Sept em ber 20, 2003, sum m a ri zi ng Deut ch and M oni z, co-chai r s, The Future of Nuclear Power, op. ci t .

I ntroduction 3

radioactive wastes). In a current-technology reprocessing plant, the spent fuel from nuclear reactors is chopped into pieces and dissolved in boiling nitric acid. The uranium and plutonium in the spent fuel are extracted from this nitric acid solution using organic solvents (typically tributyl phosphate). Since this extraction is accom p lished by m a nipulating the chem ical reduction-oxidation (redox) states of the plutonium and uranium ions in solution, this process (the only one that has been operated at com m e rcial scale) is called the Plutonium -Uranium Redox Extraction (Purex) process. 6

The result is that the original spent f u el is transf orm e d into reprocessed uranium (representing approxim a tely 95% of the m a ss of the original fuel m a terial), plutonium (roughly 1%), and a nitric acid solution containing the intensely radioactive fission products and other isotopes that m a ke up the rem a ining 4% or so of the original spent fuel—a solution known as high-level waste (HLW ). In addition, a variety of low-level and interm ediate-level wastes (LLW and ILW , som e of which are referred to in the U.S. system as transuranic wastes, or TRU) also result from the process. During the processing operation, a sm all portion of the radioactivity is released into the atm o sphere or into liquid wastes from the reprocessing plant—releases which have been the focus of considerable controversies regarding the operation of existing plants.

The liquid HLW from reprocessing m u st eventually be solidified (usually by m i xing it with m o lten glass, which is then hardened, a process known as vitrification), and is then slated for disposal in a geologic repository, the sam e destination as is planned for spent fuel in countries where spent fuel is not reprocessed, such as the United States. Despite occasional claim s to the contrary, 7 in traditional reprocessing m a ny of the long-lived isotopes that pose particularly serious threats to the environm ent and hum an health rem a in in the HLW . Hence a repository would have to be designed to contain the m a terial for m a ny m illennia, whether the m a terial disposed of was spent fuel or HLW from reprocessing. The ILW from reprocessing also requires isolation in a geologic repository, because of its plutonium content. Substantially m odif i ed approaches—currently expected to have still higher costs—would be needed to separate out and recycle the other long-lived isotopes from spent fuel, for possible transm utation in a react or or in an accelerator-reactor system . (Such concepts for separations and transm utation are discussed in Chapter 3.)

In principle, both the uranium and plutonium separated from spent fuel during reprocessing can be m a de into new reactor fuel and recycled. In practice, this is done for only a sm all fraction of the uranium recovered from reprocessing today, because freshly m i ned uranium is cheap enough that the uranium recovered from reprocessing (which is less

6 For a useful descri pt i on, see M . B e nedi ct , T.H. Pi gford, and H.W . Levi , Nuclear Chemical Engineering, 2 nd Ed. (New York, NY: McGraw-Hill, 1981).

7 For a t y pi cal cl ai m t h at aft e r reprocessi ng i t i s onl y necessary t o di spose of “short e r-l i v ed fi ssi on product s ,” whi c h can be hel d i n st orage desi gned t o l a st “for a few hundred y ears,” see Lake, “Out dat e d Thi nki ng i s Holding Us Back,” op. cit. In reality, expected dos es from a nuclear repository over hundreds of thousands of y ears are dom i n at ed by l ong-l i v ed fi ssi on product s such as t echnet i u m and i odi ne, whi c h are not rem oved i n t r adi t i onal reprocessi ng approaches. For a di scussi on of t h e effect of vari ous reprocessi ng approaches on repository requirem e nts a nd perform ance, see U.S. National Res earch Council, Com m ittee on Separations Technol ogy and Transm ut at i on Sy st em s, N u cl ear Wast es: Technol ogi es f o r Separat i ons and Transmut a t i o n (W ashi ngt on DC : Nat i onal Academ y Press, 1996), Appendi x G, “Effect s on R e posi t o ry ,” pp. 315-353.

4 E conomics of R eproce ssing vs. D irect D isposal of S pe nt N uclear F uel

desirable because of various isotopes created during irradiation in the reactor, including U- 234 and U-236) is not com p etitive for use in fresh fuel. So nearly all of the uranium recovered from reprocessing every year sim p ly rem a ins in storage.

Sim ilarly, a substantial fraction of the plutonium recovered each year from reprocessing also rem a ins in storage. The fabrication of uranium - plutonium m i xed-oxide (MOX) fuel from this plutonium and its use in reactors has not kept pace with the continued separation of additional plutonium through m o re reprocessing. The result is that today, there are m o re than 200 m e tric tonnes of separated civilian plutonium in storage around the world. 8 This separated plutonium , while “reactor-grade,” is usable in nuclear weapons (by any state or group capable of m a king a nuclear weapon with weapon-grade plutonium ), 9 and the current world stock is enough for tens of thousands of nuclear weapons. For this reason, this growing accum u lation—which will soon exceed the am ount of separated plutonium in all of the world’s nuclear weapon stockpiles com b ined—has been the subject of considerable controversy. Unf o rtunately, as this report will describe in detail, f a bricating f u el f r om this plutonium is m o re expensive than m a king fuel from freshly m i ned uranium . The use of MOX f r om reprocessed plutonium has also been the subject of substantial political controversy, focused particularly on safety concerns. These econom ic and political factors continue to delay the use of MOX fuel in a num ber of countries, and alternatives such as im m obilizing separated plutonium as waste have not yet been adopted. As a result it is not yet clear how or when the large world stockpile of separated plutonium will ultim ately be reduced. 10

Originally, no one intended that the plutonium recovered from reprocessing light- water reactor (LW R ) spent fuel would be recycled as fuel in LW Rs. Rather, the nuclear power industry expected that there would be a rapid transition to fast-neutron reactors, which would use this plutonium as their start-up fuel . Com m e rcialization of fast-neutron “breeder” reactors (so-called because they can be configured to produce m o re plutonium from uranium than they consum e in their fuel) has been delayed decades longer than originally expected, however. Therefore, in addition to com p aring once-through use of uranium to reprocessing and recycling in light-water reactors, this study will also com p are once-through use of uranium fuel in light-water reactors to repro cessing spent fuel and using the plutonium (and, perhaps, other actinides) in future fast-neutron reactors.

1.2. Data and Sources

For som e industries, reasonably good data on costs and prices are readily available. Data on uranium prices, enrichm e nt prices, and conversion prices, for exam ple, are widely

8 See, for exam pl e, Davi d Al bri ght and M a rk Gorwi t z , “Tracki ng C i vi l Pl ut oni um Invent ori e s: End of 1999” (W ashington, DC: Institute for Scien ce and International Security, Octobe r 2000, available at http://www.isis- onl i n e.org/ publ i cat i ons/ puwat ch/ puwat ch2000.ht m l ).

9 For an aut hori t a t i v e di scussi on, see U.S. Depart m e nt of Energy , Offi ce of Arm s C ont rol and Nonprol i f erat i on, Final Nonproliferation and Ar ms Control Assessment of Weapons-Usabl e Fissile Material Storage and Excess Pl ut oni um Di sposi t i on Al t e rnat i ves , W a shi ngt on DC : DOE/ NN-0007, January 1997, pp. 37-39.

10 See fo r ex am p l e, Kev i n O’Neill, ed ., Addressing Excess Stocks of Civil and Military Plutonium: Proceedings of the December 10, 2001 Conference (W ashington, DC: Institute for Scien ce and International Security, 2002, available as of Decem ber 16, 2003 at ht t p : / / www.i s i s -onl i n e.org/ publ i cat i ons/ 2001ci v i l pu/ 2001ci v i l put oc.ht m l ).

I ntroduction 5

available from several reliable sources. This is not the case, however, for the reprocessing and MOX fuel fabrication industries. Dom i na ted by a sm all num ber of state-owned firm s, these industries have m a intained strict secrecy over both their costs and their contract prices, in an effort to m a intain a variety of com m e rcial advantages.

Data are nevertheless available from several sources, which we have com b ined in preparing this report. First, som e official da ta on costs associated with specific existing or proposed plants are available, and we have relied on these data where possible. Second, a variety of national or international studies over the years have provided cost estim ates based on data provided by the industry, and we have relied heavily on these figures as well. Third, a variety of reported costs have found their wa y into the nuclear industry trade press, and where particular figures could be confirm e d fr om other sources, we have also m a de use of these. Finally, we have had the opportunity to review generic, representative cost data prepared by the Nuclear Assurance Corporation (NAC), a m a jor nuclear industry consulting firm , for the Departm e nt of Energy, and these figures have been helpful in confirm i ng the estim ates available from other sources. 11

Fortunately, the sensitivity analysis provided in this report dem onstrates that our conclusions are robust over a broad range of variations in the input param e ters. Even changing the cost of reprocessing or of MOX fuel fabrication by a factor of two, for exam ple, would not m a ke the reprocessing fuel cycle m o re cost-effective than the once-through cycle under current conditions. Thus differences in estim ates from different sources should not have any substantial effect on our conclusions.

Estim ating the costs of disposal of spent fuel or high-level nuclear waste poses an even m o re knotty problem . No repository for spent fuel or high-level waste has yet been com p leted or operated anywhere in the world. Hard data on real costs are therefore nonexistent, and cost estim ates inherently uncer tain. Different countries are planning quite dif f e rent types of repositories with a wide range of capacities, and hence their estim ated unit costs (per ton of spent f u el or of solidif ied high-level waste) vary signif i cantly; the quality and detail of the available estim ates also varies. W h en com p aring direct disposal of spent fuel to disposal of the wastes that would result from reprocessing it, one is com p aring approaches that generate different volum es of waste, different physical and chem ical form s of waste, different rates of heat generati on from the wastes, and different degrees (and lif etim es) of the wastes’ radiotoxicity. There is only a very m odest literature analyzing how these different waste characteristics m i ght affect repository cost. W e have done our best with the literature available—f o cusing prim arily on the projected cost f o r the U.S. repository, which is the one for which the m o st detailed and consistent cost inform ation is available— but this is clearly an area for additional research. Fortunately, here, too, even very broad variations in assum p tions about the relative cost of disposing of spent fuel vs. disposing of reprocessing wastes do not change the basic conclusions of this study.

11 Geo ff Varley an d Dan Co llier, Fuel Cycle Cost Data (At l a nt a, GA: NAC , Oct ober 1999). Thi s report was prepared on cont ract t o t h e U.S. Depart m e nt of Energy , and whi l e i t i n cl udes no propri e t a ry i n form at i on on cost s or pri ces at exi s t i ng reprocessi ng or pl ut oni um fuel fabri cat i on pl ant s , t h e com p i l a t i on of avai l a bl e dat a i t does cont ai n i s propri e t a ry t o NAC ; hence t h e report i s not publ i c l y avai l a bl e.

6 E conomics of R eproce ssing vs. D irect D isposal of S pe nt N uclear F uel

1.3. Cost vs. Price

In the classic m odel of a fully com p etitive m a rket, the cost of providing a good or service and its m a rket price are very closely related. The price is sim p ly the long-run m a rginal cost of providing the good or service plus a rate of profit sim ilar to what could be m a de by taking sim ilar risks elsewhere in the econom y. If the m a rket price rises to a higher level, this will create opportunities for unusually high profits that will lead m o re producers to enter the m a rket, and the resulting com p etition will drive the price back down to the com p etitive level. If the m a rket price is below m a rginal cost plus a com p etitive prof it, producers will choose to produce other, m o re profitable goods and services, and the decline in supply will drive the price back up again.

Few nuclear m a rkets, however, m a tch this classic com p etitive equilibrium m odel particularly well. The uranium and enrichm e nt m a rkets m a tch the m odel at least slightly better than som e of the other nuclear m a rkets, in that com p etition in these m a rkets is sufficiently intense that no producer can afford to charge greatly m o re than its costs for very long. Hence, in this report we will rely on price data for these elem ents of the nuclear fuel cycle, rather than attem p ting to separately assess the underlying costs of providing these goods and services.

The reprocessing and MOX fabrication indus tries, however, have been dom inated by an oligopoly of only two or three firm s, which have set prices that m a y in som e cases be quite different from real long-run m a rginal costs. Originally, they were able to set reprocessing prices at levels above the full capital plus operating cost of reprocessing, because custom er utilities faced governm e nt requirem e nts to reprocess and had no other choice. Today, by contrast, with the capital costs of the reprocessing plants already paid for, they are able to set prices at levels that reflect only operating cost, future capital and decom m i ssioning costs, and profit—and are therefore below a realistic estim ate of the f u ll cost of providing the reprocessing service. (These paid-of f plants will not last f o rever, however, and if reprocessing were to continue, pri ces would have to rise to levels that would pay for both capital and operating costs for new plants to replace the existing facilities.) Thus, the fact that one of these services is being offered at a particular price does not in itself dem onstrate that its full cost m u st be at that price or below—a com m on m i sconception.