Nuclear Control Institute
1000 Connecticut Ave, NW, Suite 804
Washington, DC 20036
Presented at the U.S. Department of Energy Plutonium Stabilization and Immobilization Workshop
December 14, 1994
The concept of geologic disposal of wastes containing plutonium has come under increasing criticism by some analysts from a non-proliferation standpoint, who argue that repositories will become attractive sources of large amounts of weapons-usable fissile material in the future.1 This argument has two main components:
1. The plutonium emplaced in repositories will be recoverable assuming sufficient time, effort and financing, even after the repository has been made "irretrievable." Repositories therefore will have to be safeguarded for hundreds of thousands of years, until the plutonium has decayed to a very low concentration. However, it is impossible to guarantee that safeguards can be maintained over such a long time frame. Furthermore, the ability to employ material accountancy, a cornerstone of International Atomic Energy Agency (IAEA) safeguards, will effectively be lost following closure of the repository.
2. The attractiveness of repository spent fuel to a proliferant will increase with time in several ways. First, the penetrating radiation barrier that renders spent fuel extremely hazardous to handle will decay to a very low level after a few centuries of cooling, so that the material can be reprocessed for much lower cost than spent fuel of more recent vintage. Second, the time and effort necessary to recover the fuel from the repository will decrease as mining technologies improve. Third, the isotopic quality of the plutonium in commercial spent fuel will approach (although never quite reach) that of "weapons-grade" with time.
While this issue has been raised most often by promoters of reprocessing, such as an American Nuclear Society (ANS) special panel on the "Protection and Management of Plutonium,2 it has also been noted in less partisan analyses. The National Academy of Sciences (NAS) recommended that "follow-on studies should continue on the longer-range questions of whether and how the residual security risks of ... plutonium should eventually be reduced beyond the spent fuel standard," though it also emphasized that this endeavor should not distract attention from the more immediate problem of reducing the risks posed by separated plutonium.3
Some observers have argued that this problem can only be avoided by foregoing geologic disposal of plutonium entirely and maintaining it in retrievable storage, while redirecting resources toward the development of technologies that can achieve the near-total "destruction" of plutonium.4,5 Others argue that disposal in deep boreholes, rather than in mined repositories, can provide a more effective impediment to future recovery.6
However, replacing the current plan for geologic disposal of spent fuel with one of these even less well-established approaches would represent a drastic response to a highly uncertain future risk. Furthermore, some of the proposed cures would be worse than the "disease"; for instance, the plutonium "destruction" schemes require the large-scale separation of plutonium from spent fuel and would result in a dramatic increase in proliferation risks in the near-term. Clearly, some objective means are needed for evaluating whether the prospect of future "plutonium mines" really poses an unacceptable long-term risk and, if so, for determining how current plans for geologic disposal of spent fuel must be modified to achieve a reasonable level of security.
This definition is conceptually similar to the idea of the "spent fuel standard," introduced by the NAS to evaluate disposition options for separated plutonium.7 It can be used as a basis for rejection of costly spent fuel disposition options which would render plutonium recovery far more difficult for a proliferant than new production, and therefore would achieve only a marginal increase in overall security. Similarly, it can identify options that, if pursued, would leave spent fuel relatively vulnerable to diversion.
In defining the MPS, one must determine which alternative source of fissile material is the most appropriate for comparison to repository mining. This depends to some extent on the future of commercial nuclear power, as the following discussion illustrates.
The standard that a repository must meet can be less stringent if one assumes that a nation will continue to operate nuclear reactors and other nuclear fuel cycle facilities. Such a nation will always have a ready supply of spent fuel available, either in retrievable storage or in the reactor cores. (If the nation operates commercial reprocessing plants as well, it will also possess stockpiles of separated plutonium.) Under these circumstances, it is apparent that spent fuel in a sealed geologic repository would be relatively unattractive with respect to both state- sponsored and sub-national diversion, assuming that repositories were safeguarded at a level consistent with other stages of the fuel cycle.
On the other hand, the relative attractiveness of spent fuel in a repository would be greatest in the context of a "nuclear-free" future, in which nuclear power had been phased out and neither operable reactors nor retrievable spent fuel storage facilities existed. In this case, the only means of acquiring spent fuel other than mining the repository would be the construction and operation of production reactors and associated front-end facilities (e.g. uranium mining and fuel fabrication) from scratch. The assumption of a nuclear-free future is therefore the most conservative one and is adopted here.
The most difficult part of this analysis, as in other aspects of repository performance assessment, is the task of making reasonably credible observations while minimizing speculation about the future, such as technological advances, political systems and human behavior. Use of a relative standard (e.g. the MPS) may be more sensible than an absolute standard (the IAEA's "practicable irrecoverability" criterion, discussed below), because any assumptions one makes about the future must apply equally to both pathways. For instance, one cannot expect that technological improvements that may reduce the difficulty of the mining route will occur without commensurate technical advances that may be useful for the new production route, such as the development of advanced materials.
Because of the need to avoid excessive speculation about the future, this analysis must necessarily be limited in scope. However, it can be used to provide insight into some questions, including:
1) from a non-proliferation standpoint, how long after reactor discharge should spent fuel be kept in monitored retrievable storage?
2) is it indeed necessary to maintain long-term safeguards on spent fuel, and if so, does this constitute a "fatal flaw" of the geologic repository concept?
3) under what circumstances will spent fuel repositories pose unacceptably high proliferation risks? Are there ways of mitigating those risks?
i) A "minimum acquisition" plutonium program, based on a gas-graphite production reactor rated at 30 MW-thermal (MWt) and capable of producing around 8 kg of plutonium per year. This program is estimated to require a capital cost of $120 - $300 million, of which $35 - $100 million is the construction cost of the reactor.8 The average cost of the reprocessing component is $15 - $40 million, or about 12% of the total. The time for construction of this project is estimated to be 3-4 years, with a crew of 100.
ii) An "intermediate acquisition" program capable of producing around 100 kg of plutonium per year. One program, based on a 400 MWt reactor, was estimated to require a capital investment in the range of $400 million to $1 billion (for the reactor alone), with a cost overrun of up to 100% possible in the event of delays. The construction time for this reactor is estimated to be 5-7 years, requiring a staff of 200-300.9 The cost of the reprocessing plant in this case was not given; scaling from the previous example (40% of the reactor cost) yields a value of $160 - $400 million.
Actual construction times could be considerably shorter than these estimates, as is apparent from the original historical example of "new production." The 250 MWt B- Reactor at Hanford, the first plutonium production reactor of the Manhattan Project, was completed in about a year; 15 months later, two other reactors had begun operation as well. Together, these three reactors were producing plutonium at a rate of about 100 kg per year within two years after the start of construction.10 Thousands of workers, however, were employed in these projects.
iii) A "maximum acquisition" plutonium program, in which the desired production rate is limited only by the resources available. For example, during the Cold War, multiple 2150 MWt reactors were constructed at the Savannah River Site, each capable of producing about 600 kg of plutonium per year. Scaling the previous cost estimates with a factor of 0.7 implies that the cost of each reactor of this size to be between $1.3 billion and $3.2 billion. The average time between start of construction and startup of these reactors was under three years.11
Large underground mining operations today typically require capital investments on the order of a few hundred million to well over one billion dollars (with the higher figure reflecting isolated sites and/or difficult climates),12 and development times of 2-5 years before production can begin.13
Excavation of a spent fuel repository would require a similar level of investment. For example, the development costs of the original (1988) design of the Yucca Mountain repository include $200 million for preparation of the site, $320 million for constructing the shafts and ramps, initial excavations at the repository level, and underground service systems, and $510 million for construction of surface facilities, for a total of over $1 billion (all costs in undiscounted 1992 dollars).14 (Lest one think that these values are artifacts of fiscal inefficiencies of the U.S. radioactive waste program, it should be noted that the estimated underground costs in the U.S., when normalized to the volume of rock excavated, are the lowest among six countries surveyed. 15)
One may argue that these costs are based on rigorous standards for occupational and environmental safety which would probably not be observed by those seeking to acquire fissile material rapidly. However, there are limits to the extent to which one can skimp on mine safety and still guarantee a given level of productivity. A catastrophic accident could lead to extensive loss of personnel and set back the project by several years. Even if the cost were reduced by a factor of two or three, it would still exceed a few hundred million dollars.
Although the theoretical maximum advance rates of modern drilling methods are quite high, the minimum time necessary to gain access to a repository will probably be determined by the significant effort involved in the on-site assembly and preparation of equipment. For example, the tunnel boring machine (TBM) being used to drill the Exploratory Studies Facility (ESF) at the Yucca Mountain site was brought there in 52 separate truckloads. One engineer has estimated that one to two years would be needed to assemble a large TBM and auxiliary equipment, and to prepare a platform to launch the drilling.16
Furthermore, these costs and advance rates do not reflect the unusually harsh conditions that may be experienced when mining a filled spent fuel repository, as opposed to a conventional mineral deposit. Even after a few hundred years, when the gamma radiation levels have dropped substantially, the thermal conditions may remain severe. For instance, in the "extended hot, dry" scenario being considered for a repository at the Yucca Mountain site, waste package surface temperatures would remain above 100C for several thousand years, and near-field rock temperatures would remain above 80C for 10,000 years.17 The maximum ambient temperature that could be tolerated by workers is about 35C. Temperatures in the deepest mines today reach 55C, and therefore extensive ventilation and cooling systems must be employed, capable of discharging several megawatts of heat. The additional infrastructure necessary will increase both the capital outlay and the detectability of the operation by providing both visual and thermal signatures (as discussed below).18
The above analysis suggests that for the "minimum" and "intermediate" acquisition scenarios, both the spent fuel mining route and the new production route would require capital investments in the range of several hundred million to over one billion dollars, and a development time of 1-5 years. In these cases, therefore, the mining route does not appear to have a decisive advantage with respect to these criteria.
The rate of removal of spent fuel from a repository, which will be on the order of the rate of emplacement, is controlled by the inventory of each waste canister and the spacing between canisters. Because of the high concentration of plutonium in commercial spent fuel and the limited space available for emplacment, production rates from a large plutonium mine could be quite high. For example, the rate of loading of the planned Yucca Mountain repository is currently anticipated to be 3000 tHM of spent fuel a year. Assuming equal rates of retrieval and emplacement, this mine could produce 25 tonnes of plutonium per year. More than forty large Savannah River-type reactors would have to be built to attain a level of production equivalent to the mining route, at a much greater capital cost. Therefore, current repository designs would fail to meet the MPS under scenarios in which a nation in a "nuclear-free world" suddenly wants to begin accumulating nuclear weapons at a rate of thousands per year. However, it remains to be determined whether this scenario is sufficiently credible to warrant concern.
However, when compared to new production, this advantage is not as overwhelming as it may first appear. The burnup of spent fuel from dedicated gas-graphite weapons- grade plutonium production reactors is typically quite low, below around 800 MWD/t.19 This is around fifty times smaller than the typical burnup of commercial fuel.
For low and intermediate production rates, low-burnup fuel can be processed in small, rudimentary reprocessing cells known as "caves," provided the operators are willing to accept high but not debilitating radiation doses. These cells, which date from the 1950s, utilize primitive, mechanically operated remote-handling devices known as ball-joint manipulators.20 The difference in cost between a small cave and a glovebox facility would probably not be great enough to influence the decisions made by a proliferant group.
The highest whole-body radiation dose that employees could tolerate without a significant loss of efficiency, taking into account the effect of fractionated dose, is around 3 Gy/yr, or about 15 mGy/hr for a standard work-week. This level of exposure is an order of magnitude below that at which acute prodromal symptoms would appear.21 However, an operator in this environment will receive just under the LD10 dose annually, which means that on average, ten percent of those exposed will succumb to radiation sickness each year. (Whether or not this is an acceptable loss depends on the total number of available personnel).
After two years of cooling, each tonne of uranium gas-graphite reactor fuel, irradiated to a burnup of 800 MWD/t, will contain about 720 grams of weapons-grade plutonium and about 4000 curies (Ci) of hard (greater than 0.4 MeV) gamma activity. In order to produce 100 kg of plutonium per year, about 0.5 tonnes of fuel would have to be reprocessed per day (assuming 80% plant availability). This quantity of fuel could be subdivided into five batches per day, each containing about 400 Ci of hard gamma activity. The unshielded exposure from a point source of this strength at a distance of 50 cm is over 5 Gy/hr; to reduce this to the target value of 1.5 mGy/hr would require attenuation by a factor of 3300. This can be achieved with 8 cm of lead or 15 cm of dense lead glass. These dimensions are well within the constraints of process cells of the type described above.
At higher production rates, small, locally-shielded caves would no longer be feasible for reprocessing production-reactor fuel, and larger plants, utilizing remote operation and maintenance, would have to be employed. Thus the comparative attractiveness of repository-grade spent fuel is maximized under these circumstances. However, at high production rates, the cost of a facility for reprocessing aged spent fuel would increase as well, because of the need to provide greater containment of alpha particles and to mitigate the increased risk of a criticality accident.
On the other hand, the IAEA intends to terminate safeguards on vitrified high-level wastes (VHLW) with a fissile content of less than 2.5 kg per cubic meter (a concentration of about 0.1 weight-percent for a typical HLW glass of density of 2.5 g/cm3). Adoption of this threshold also implies that safeguards can eventually be terminated on spent fuel as well, since the concentration of plutonium will decrease with time. The plutonium concentration in a commercial PWR spent fuel assembly upon discharge (including the internal void volume) is approximately 25 kg/m3; the concentration will fall below the IAEA threshold after decreasing by a factor of ten, which will occur after approximately 80,000 years (in the absence of geochemical concentration mechanisms). (However, the proposed 2.5 kg/m2 threshold has an unclear technical basis and may have to be revised.)
Techniques that could be employed in a repository safeguards program include seismic monitors to detect unauthorized drilling, satellite reconaissance to observe hard- to-conceal surface activities, and random on-site inspections.24 These methods are neither technically challenging, intrusive or resource-intensive. In addition to visual cues, repository mining may also have a thermal signature, as ventilation and cooling systems may have to be employed in order to permit tolerable working conditions. The ground resolution available from commercial satellite reconaissance data (5-10 m visual, 120 m infrared) is more than adequate to detect repository mining activities.
It appears that the necessity of maintaining long-term safeguards on spent fuel repositories is inconsistent with one of the fundamental principles of geologic disposal; namely, that a repository, after it is sealed, should not require active monitoring for any purpose. However, one can argue, using the logic of the material production standard, that this issue does not compromise the validity of the concept of spent fuel disposal.
Recently, in view of the events in Iraq and North Korea, the IAEA has acknowledged the inadequacy of the current safeguards system, which focuses primarily on inhibiting the diversion of weapons-usable fissile materials from declared fuel cycle operations, in the absence of additional mechanisms for verifying the absence of clandestine material production.25 In an attempt to rectify this situation, the IAEA Board of Governors is considering an array of expanded verification activities, known as the 93+2 program. This program would include enhanced access to all sites where nuclear materials are present, greater authority to inspect undeclared sites and the use of supplemental techniques such as environmental monitoring. Although the sharing of national satellite reconaissance data is not currently contemplated, it may very well become part of a future verification regime.
The threat posed by mining of a geologic repository will be maximized in the context of a world in which there are no operating nuclear reactors or retrievable spent fuel storage facilities, as discussed above. However, without an effective mechanism for verifying the absence of clandestine fissile material production, the notion of a "nuclear-free" world is not meaningful from a non-proliferation standpoint. If we cannot guarantee that safeguards on spent fuel repositories will remain in place indefinitely, then we also must assume that safeguards on clandestine production also cannot be guaranteed. In such an environment, it is far from clear that spent fuel repositories would pose the greatest proliferation risks.
On the other hand, if one assumes that a means of detecting clandestine production is in place, then such techniques could be extended quite simply to safeguard geologic repositories. For example, assume that a satellite system for the detection of thermal output from production reactors is deployed. This system could equally well be applied to the monitoring of repository sites for the presence of unauthorized mining activities. Furthermore, the task of monitoring a series of known sites would be far more straightforward than the task of verifying the absence of clandestine activities, which could occur anywhere (and in particular, in industrial areas where the visual and thermal signals could be camouflaged). This suggests that in the presence of a monitoring regime, a covert proliferation program would be more likely to escape detection if the clandestine production route, rather than the repository mining route, were employed.
Backfilling all drifts and shafts in a repository would ensure that regaining access would be as difficult and costly as originally excavating the site. However, in one design being proposed for the U.S. repository at Yucca Mountain, shielded canisters would be brought into repository drifts and simply left sitting on rail cars; no backfilling of the drifts would take place (the access shafts would, however, be backfilled). It is clear that this approach would permit rapid retrieval of spent fuel, once the repository horizon were reached. Eventually, the drift walls would collapse and bury the canister; but a densely packed backfill would provide a more reliable barrier in the near-term.
Canister inventory and spacing
To lower the rate of plutonium removal possible from a repository, the spent fuel inventory of disposal canisters could be reduced and the canisters placed farther apart. However, in order to do this effectively, spent fuel would have to be distributed at a very low emplacement density in a repository, substantially raising the cost of geologic disposal. Also, the thermal loading of such a repository would be very low, eliminating the deterrent effect of high temperature.
The risk involved in underground mining increases with the depth of the operation. The casualty rates in the very deep South African gold mines (over 2,300 m deep) are very high. Locating repositories at these depths would increase the cost of both retrievability and emplacement.
Composition of plutonium glass
It is apparent that if one takes the long-term proliferation problem seriously, then weapons-grade plutonium (W-Pu) vitrified with radioactive wastes does not meet the spent fuel standard. This follows from the fact that Pu-239 decays to U-235, another fissile isotope. The uranium in spent fuel, however, will remain low- enriched even after all the Pu-239 has decayed. In WPu-glass, however, the plutonium content of the glass will eventually become highly-enriched uranium with a few percent U-236 admixture. This can be easily corrected by denaturing the product, e.g. adding 5 weight-percent U-238 to the glass, assuming a WPu loading of 1 weight-percent (this may need to be adjusted to accommodate differential leaching that may occur). This will ensure that the U-235 concentration never exceeds the HEU threshold of 20%. Borosilicate glass compositions can easily be adjusted to accept a 5% U-238 loading.
1. Swahn, J., "The Long-Term Nuclear Explosives Predicament," Technical Peace Research Group, Institute of Physical Resource Theory, Chalmers University of Technology, G"teborg, Sweden, 1992. Back to document
2. American Nuclear Society, "Protection and Management of Plutonium," Special Panel Report, American Nuclear Society, LaGrange Park, Illinois, 1995, p. 12. Back to document
3. Committee on International Security and Arms Control, National Academy of Sciences, Management and Disposition of Excess Weapons Plutonium: Reactor-Related Options, National Academy Press, Washington, D.C., 1995, p. 15. Back to document
4. Jensen, R. et al., "Accelerator-Based Conversion (ABC) of Reactor and Weapons Plutonium," in the Proceedings of the International Conference and Technology Exposition on Future Nuclear Systems: Global '93, Seattle, Washington, September 12-17, 1993, p. 833. Back to document
5. "Burying Spent Fuel is Not the Best Approach to Nonproliferation, Seaborg tells ANS," Spent Fuel 2, November 6, 1995, p. 2. Back to document
6. Swahn (1992), op. cit., p. 166. Back to document
7. Committee on International Security and Arms Control, National Academy of Sciences, Management and Disposition of Excess Weapons Plutonium, National Academy Press, Washington, D.C., 1994. Back to document
8. U.S. Congress, Office of Technology Assessment (OTA), Technologies Underlying Weapons of Mass Destruction, OTA-BP-ISC-115, U.S. Government Printing Office, Washington D.C., December 1993, p. 156. Back to document
9. Ibid, p. 158. Back to document
10. Cochran, T., Arkin, W., Norris, R. and Hoenig, M., Nuclear Weapons Databook, Volume II: U.S. Nuclear Warhead Production, Ballinger, Cambridge, MA, 1987, p. 59-64. Back to document
11. Ibid, p. 61. Back to document
12. Peters, W., Exploration and Mining Geology, 1987, p. 262. Back to document
13. Hartman, H., Introductory Mining Engineering, Wiley, New York, 1987. Back to document
14. Nuclear Energy Agency/Organisation for Economic Cooperation and Development (NEA/OECD), The Cost of High-Level Waste Disposal: An Analysis of Factors Affecting Cost Estimates, OECD, Paris, 1993, p. 136. Back to document
15. Ibid, p. 53. Back to document
16. Robert Saunders, Morrison-Knudsen Company (Las Vegas, Nevada), private communication, October 1995. Back to document
17. TRW Environmental Safety Systems (Las Vegas, Nevada), "Strategy for Waste Containment and Isolation for the Yucca Mountain Site," Preliminary Review Draft, October 1995. Back to document
18. Sidney Whittaker, Atomic Energy of Canada Ltd (AECL), private communication, November 1995. Back to document
19. Albright, D., "North Korean Plutonium Production," Sci. Global Sec. 5 (1994), p. 63. Back to document
20. Goertz, R., Ferguson, K. and Doe, W., "Mechanical Handling of Radioactive Materials," in the Nuclear Engineering Handbook (J. Etheringon, ed.), McGraw-Hill, 1958, p. 7-128. Back to document
21. United Nations Scientic Committee on the Effects of Ionizing Radiation (UNSCEAR), Sources and Effects of Ionizing Radiation, United Nations, New York, 1988, p. 576. Back to document
22. International Atomic Energy Agency, "Advisory Group Meeting on Safeguards Related to Final Disposal of Nuclear Material in Waste and Spent Fuel," Secretariat Working Paper, STR-243, IAEA, Vienna, 1988. Back to document
23. Seneviratne, G., "IAEA Developing Safeguards Conditions for Vitrified Waste, Repositories," NuclearFuel, September 12, 1994, p. 9. Back to document
24. International Nuclear Fuel Cycle Evaluation (INFCE), "Safeguards for Geologic Repositories," INFCE/DEP/WG.7/18, September 1979. Back to document
25. Jennekens, J.; Parsick, R.; von Baeckmann, A., "Strengthening the International Safeguards System," IAEA Bulletin 34 (1992), p. 6. Back to document
NCI Home Page