Over the past 20 years, there has been little or no net growth in installed nuclear capacity in much of the world with the exception of Asia, where there has been some limited new nuclear construction. Many energy analysts now expect, however, a dramatic nuclear renaissance, provoked in part by anxieties over global warming and claims that nuclear power can play a substantial role in easing these concerns.
Yet, nuclear power will have to expand fivefold or more worldwide to make even a modest contribution to greenhouse gas reductions. Such an expansion does not appear feasible in the next quarter-century or longer, and in the interim, there are alternative paths to low-carbon-emitting electricity that appear equally or more promising than nuclear power.
At the same time, the Bush administration and some advocates of this nuclear renaissance are using another environmental concern—the growing quantities of spent fuel—to justify a sharp change in U.S. policy toward spent fuel reprocessing. The administration’s Global Nuclear Energy Partnership (GNEP) aims at deploying reprocessing on a large scale, thus ending a more than 30-year moratorium in the United States, and coupling this reprocessing with deployment of a fleet of fast reactors designed to burn the plutonium and other transuranics separated in the reprocessing operation. The technologies envisioned, however, have not yet been proven, and even if they were to be developed, they will do little to manage the spent fuel problem while costing an enormous amount of money and risking increased weapons proliferation.
Rather than rushing headlong into promoting nuclear power as a means of curbing global warming, policymakers should take the ample time they have to assess the practical feasibility of implementing other alternatives on a large scale. Until this is done, there is no reason that nuclear power should be privileged by government policies.
Projections
Nuclear power plants often take decades to build. Despite the hoopla over a nuclear renaissance, there is little evidence of a vast surge in construction before 2030, the farthest point in time where the projections at least roughly can be based on actual plans. At the end of 2005, 443 nuclear plants, with an installed capacity of 367 gigawatts-electric (GWe), were in operation in 31 countries. These units provided about 2,700 terawatt hours in 2005, approximately 15 percent of electricity generation worldwide.[1] Six countries—the United States, France, Japan, Germany, Russia, and South Korea—accounted for 75 percent of the capacity. Projections from international, governmental, and private groups vary somewhat but generally fall into the range of 400-600 GWe of installed nuclear power in 2030. For example, the International Energy Agency forecasts a worldwide capacity in 2030 of 416-519 GWe. The International Atomic Energy Agency (IAEA) projects a nuclear capacity in 2030 of 414-679 GWe.[2] The U.S. Energy Information Administration estimates 2030 capacity at 438 GWe.[3] The trade journal NUKEM, which relied on a “number of informational sources” for its projections, projects a capacity in 2030 of 535 GWe.
For those advocating or expecting a nuclear renaissance, it is the period after 2030 that is of the greatest interest. A 2003 MIT interdisciplinary study presented one scenario of what a nuclear system might look like by mid-century.[4]
The MIT study imagined a worldwide nuclear capacity in 2050 of 1,000-1,500 GWe, respectively termed low and high scenarios. The authors point out that the projection “is certainly not a prediction of rapid growth in nuclear power. Rather, it is an attempt to understand what the distribution of nuclear deployment would be if robust growth were realized.”
It is unlikely that the more robust projections of the expansion of nuclear power will come to pass. Investors in developed countries continue to shy away from nuclear power technology because of the continuing high financial risk they see in the sector. Few developing countries have the infrastructure or incentive to support a large nuclear capacity. It is doubtful that nuclear capacity in 2050 could reach even the low MIT scenario of 1,000 GWe of installed nuclear power, much less the high scenario of 1,500 GWe.
If nuclear power were to grow to 1,000-1,500 GWe by mid-century, it is noteworthy that plants would have to be built in several developing countries that have no or negligible nuclear power today. These countries include Algeria, Armenia, Azerbaijan, Belarus, Georgia, Indonesia, Iran, Malaysia, Mexico, North Korea, Pakistan, the Philippines, Poland, Romania, Slovakia, Thailand, Turkey, Turkmenistan, Uzbekistan, Venezuela, and Vietnam. A level of 1,000-1,500 GWe of global production is also the level to which nuclear power would have to rise if it were to play a significant role in combating global warming.
Nuclear Power and Global Warming
Nuclear power’s environmental appeal is that it produces less carbon than coal or natural gas, today’s predominant fuels for generating electricity. Each kilowatt-hour produced by nuclear energy “saves” 186 grams of carbon compared to an equivalent amount of energy produced by a modern coal-fired electric plant (if those plants do not use carbon capture and storage) and 93 grams of carbon compared to electricity produced using natural gas.
One gigawatt of nuclear capacity operating at an average capacity factor of 85 percent will therefore save about 1.5 million metric tons of carbon annually compared to coal and 0.75 million metric tons of carbon compared to gas-turbine-generated electricity. (Carbon dioxide emissions are measured here in the quantity of contained carbon molecules. The emissions of carbon dioxide would be 3.66 times greater than the carbon.)[5]
Total global carbon emissions today are approximately 7 gigatons (billion metric tons) per year and are projected to reach 14 gigatons per year in 50 years under business as usual. These emissions would be lessoned by 1 gigaton per year if nuclear energy were used instead of coal to produce about 700 gigawatts of electricity. An additional 700 gigawatts of electricity produced from nuclear power would roughly triple the current nuclear capacity to about 1,000 GWe. Nuclear power, therefore, can play a role in addressing global warming. For it to do so, however, it would have to grow substantially, to a level that will not be feasible for the next several decades.
Moreover, nuclear power is not the only low-carbon generation alternative possible, nor is nuclear power generation expansion the most effective way to reduce carbon emissions in the near future. First, improvements in end-use efficiency would have the greatest and the most immediate impact. For example, the International Energy Agency in its World Energy Outlook 2006 developed an “alternative scenario” in which governments adopted an array of actions to reduce greenhouse gases. In this scenario, the study attributed to end-use efficiency advances two-thirds of the reductions in carbon dioxide.[6] Other studies of greenhouse gas emissions have reached similar conclusions.[7]
Secondly, efficiency improvements in the power sector itself could have substantial impact. Thus, under business as usual, the World Energy Outlook 2006 estimates that coal electricity will grow from 6,900 terawatt hours in 2004 to 14,700 terawatt hours in 2030.[8] New coal-fired plants have efficiencies up to 46 percent, and by 2030 the efficiencies could reach 50 percent or higher.[9] Today, the world average is less than 30 percent. Investments that would drive the average world efficiency of coal plants in 2030 to, say, 42 percent from 30 percent would save roughly the same amount of carbon as would building 800 nuclear plants.
China provides an especially vivid illustration of the salience of efficiency improvements. The average efficiency of its coal plants today is 23 percent. If this could be raised to 42 percent by 2030, that would save the same amount of carbon as 350 GWe of nuclear power if it replaced the current inefficient plants. How dramatic this is may be seen in the World Energy Outlook projections of capacity. In 2004, China had 6 GWe of nuclear power and 307 GWe of coal energy. These amounts were projected to grow to 31 GWe of nuclear power and 1041 GWe of coal power in 2030.[10]
Finally, it appears from various studies that several low-carbon electricity supply options could be available after 2020-2030, including advanced nuclear reactors, integrated gasification combined-cycle coal plants with carbon capture and storage, and wind farms, at roughly competitive prices if states through taxes, cap and trade schemes, or other schemes place a charge of $100 per ton of carbon emitted to the atmosphere. How suitable these alternatives are to a large-scale expansion is at present unclear. By 2030, however, experts will have a far-better understanding of the practical capacity available worldwide for carbon storage, for example, and of the practicality of large-scale wind farms with compressed air storage.
We should also by 2030 have a clearer idea of the prospect of some advanced nuclear technologies, which could have attractive proliferation-resistance characteristics if nuclear power were to expand beyond 1,500 GWe. These include concepts such as pebble-bed gas reactors; molten-salt thorium reactors; and small, sealed, lead-cooled reactors. Above all, it may be hoped that by 2030, we will understand better whether a safeguards system could be put into place to allow a true nuclear renaissance. Until that time, there is little reason for states to provide disproportionate subsidies to nuclear power.
Management of Spent Fuel
Moreover, if nuclear power were to increase in future decades at the level advocates support, it would exacerbate the already difficult problem of dealing with spent nuclear fuel. Worldwide, about 10,000 metric tons of spent fuel is discharged from reactors each year. This includes roughly 6,500 metric tons from 325 GWe of light-water reactors (LWRs) and 3,500 metric tons from 42 GWe of heavy-water and gas-cooled reactors. The total plutonium contained in the spent fuel is approximately 75 metric tons. By 2030, spent fuel discharges will rise to about 11,000-13,000 tons annually, containing more than 100,000 kilograms of plutonium.
To manage this material, two spent fuel strategies are being used or are under consideration by countries. First, reprocessing of the spent fuel, with the separated plutonium either stored indefinitely for possible future use in fast breeder or burner reactors or recycled as mixed oxide fuel (MOX) in LWRs.[11] Second, interim storage of the spent fuel with the object either of ultimate direct disposal in geological repositories or of making a later decision on ultimate disposal (reprocessing or direct emplacement in a repository). Naturally, with no repository yet built and none likely to be built for at least a decade or longer, the two disposition routes under the interim storage strategy are for the moment indistinguishable.
Countries adopting the first strategy, at least partially, include France, Japan, India, Russia, and the United Kingdom, all of which have their own reprocessing plants. Belgium, Germany, the Netherlands, and Switzerland in the past had sent their spent fuel to France or the United Kingdom for reprocessing but have now decided to adopt instead a strategy of interim storage before disposal and have cancelled all reprocessing contracts abroad. Japan in the past also sent much of its spent fuel to Europe for reprocessing but in 2006 opened its own reprocessing plant at Rokkasho and will in the future reprocess on its own territory.
Countries following the second strategy of no reprocessing include Finland, South Korea, Sweden, Taiwan, and the United States (at least to date).
Whichever strategy is adopted, it appears increasingly likely that most countries will have to dispose of their spent fuel domestically. Hopes have faded that some countries would be willing to store spent fuel sent by other countries. Russia, the one country that had most strongly raised the possibility of hosting spent fuel from abroad, has taken the position that it would not do so other than for spent fuel derived from reactors built by the Soviet Union or Russia and using Russian fuel.
At present, approximately 2,500 metric tons of spent fuel is reprocessed annually, representing about one-quarter of the total annual spent fuel discharge worldwide. The total amount of plutonium separated through reprocessing is about 25 metric tons per year. Roughly one-third of this plutonium is being recycled into MOX, with most of the other two-thirds added to existing stockpiles at the reprocessing plants in Japan, Russia, and the United Kingdom. In 2005, 32 LWRs in Belgium, France, Germany, and Switzerland used MOX, with an additional 18 reactors in these countries licensed to use the fuel. Japan hopes to use MOX in one-third of its reactors by 2010.
The MOX disposition strategies now being pursued by France and Germany and envisioned for other countries have several environmental and economic limitations.[12] Economically, MOX provides only 30 percent or less of the fuel in the reactors in which it is employed. Furthermore, six spent uranium-oxide fuel assemblies must be reprocessed to obtain the plutonium for the equivalent of one MOX assembly.
MOX also provides fewer environmental benefits than it might seem because spent MOX fuel is not reprocessed.[13] Moreover, for the same cooling-off period, the space in a repository required to store MOX is three or more times greater than the space for uranium-oxide fuel. In addition, the depleted uranium obtained from the reprocessed uranium-oxide spent fuel generally is not recycled.
The alternative to reprocessing is dry-cask storage. The U.S. Nuclear Regulatory Commission has deemed such storage to be safe and secure for many decades, and there is now considerable industry experience with dry casks. Indeed, virtually every currently operating reactor in the United States either already had dry-cask storage as of the end of 2004 or has now such storage under construction or planned.
In comparing the costs of the two alternative routes, two flows of material should be kept in mind. First, in the dry-cask alternative, all of the spent uranium fuel is kept in the reactor pools for about 20 years and then put into dry-cask storage. Second, with reprocessing and MOX recycling, the separated high-level wastes are stored at the reprocessing plant, and the spent MOX fuel is stored indefinitely at the reactor (Germany) or reprocessing plant (France).
The MIT study assumed that the costs of storage under the two alternatives would be similar. This assumption is reasonable because the high-level wastes contain all the fission products and all the transuranics other than plutonium. The costs of the dry-cask storage are anyhow quite low, approximately $100-200 per kilogram of heavy metal.
Given little difference in storage costs, cost issues boil down to whether it is cheaper to reprocess spent fuel and fabricate MOX or purchase fresh fuel. The answer is clear: reprocessing and MOX fabrication costs would far exceed those from uranium fuel, even if the real costs of uranium and enrichment services were several times their current levels. (The spot prices for uranium, which have risen rapidly over the past year due to limitations on current uranium mine production and rising short-term demand, are not a good indicator of the real costs of mining and processing.) At a reprocessing cost of $1,000 per kilogram of heavy metal from LWR spent fuel and a MOX fabrication cost of $1,500 per kilogram, one kilogram of MOX will cost roughly 2 cents per kilowatt-hour more than an equivalent kilogram of uranium fuel.
As noted, the route of dry-cask storage does not preclude a decision decades hence to reprocess. In the longer run, advocates of reprocessing believe that rising uranium prices and limited waste repository space in a growing nuclear economy will drive countries to pursue closed fuel cycles based on fast reactors. Neither argument is compelling.
With respect to uranium scarcity, several recent analyses show that unless the real cost of uranium rose to well more than $300 per kilogram, the LWR, which makes up 88 percent of today’s nuclear power generating capacity, will remain less expensive than any fast reactor so far proposed.[14]
There appears to be ample uranium. In the 1,500 GWe scenario, the annual uranium consumption in 2050 would be roughly 270,000 metric tons, with a cumulative uranium requirement of about 8 million metric tons, assuming still the use of LWRs on once-through fuel cycles. This would still be well below the 14.9 million metric tons estimated by the Organization for Economic Co-operation and Development and the IAEA to be available at less than $130 per kilogram. It also supports the MIT study conclusion that there is sufficient uranium globally to support a nuclear system growing to 1,500 GWe capacity by 2050.[15] Moreover, several recent studies have estimated much greater amounts of uranium to be available at higher prices, more than 60 million tons at $200 per kilogram.[16]
The Bush administration has advanced a second argument for reprocessing and fast reactors with GNEP: that reprocessing could drastically reduce the costs of waste disposal if the separated plutonium and other transuranics are subsequently forged into fuel and burned in fast reactors. For example, GNEP has argued that, by burning the long-lived transuranics, the capacity of Yucca Mountain can be effectively expanded and the need for another repository indefinitely postponed even if nuclear power in the United States grows substantially.
The implications of such partition and transmutation have been examined in detail elsewhere.[17] A report from the International Panel on Fissile Materials advances several reasons to question the practicality and wisdom of the GNEP scheme, including its cost, proliferation risks, and waste disposal rationale. On the latter, it is striking that, in order to save space at the repository, the separated fission products first would have to be stored above ground for hundreds of years after reprocessing before their emplacement in Yucca Mountain.
Moreover, it is not even clear that the GNEP scheme is the simplest or most direct way to increase the capacity of Yucca Mountain. A recent EPRI study suggested that the capacity could be 260,000-570,000 tons, far greater than the 63,000-ton legal limit currently stipulated or the 105,000-120,000-ton technical limit often suggested.[18]
However one looks at the nuclear future, it appears illogical for countries to undertake reprocessing today. The once-through fuel cycle has the inestimable proliferation-resistance advantage that no nuclear-explosive material appears anywhere. Although not absolutely precluding diversion paths to support country proliferation, it does provide a nearly intractable barrier to substate acquisition of fissile material from the power-reactor fuel cycle. By contrast, reprocessing puts separated weapons-usable material into the civilian fuel cycle. This danger is widely recognized even by the advocates of reprocessing who, while sending the message that reprocessing is essential for the future of nuclear energy, argue that not all countries participating in that nuclear future can be trusted with reprocessing.
In sum, it does not appear feasible in the next quarter century for nuclear power to make a significant contribution to greenhouse gas reductions. In addition, there are alternative paths to low-carbon emitting electricity that appear equally or more promising than nuclear power and should be seriously considered. The practicality of these alternatives will come into clearer focus in the next few decades. In the meantime, there is no reason for policymakers to favor nuclear power.
If policymakers choose to invest in nuclear power, there is no economic rationale for reprocessing or for the recycling of plutonium in LWRs, and strong security reasons to avoid doing so. The absence of such rationale becomes still more evident if nuclear growth over the next several decades is relatively slow.
Harold A. Feiveson is a senior research scientist at Princeton University and a member of Princeton’s Program on Science and Global Security of the Woodrow Wilson School of Public and International Affairs. He is the editor and a principal author of the book, The Nuclear Turning Point: A Blueprint for Deep Cuts and De-alerting of Nuclear Weapons (1999). He is also co-founder and editor of the international journal, Science & Global Security.
ENDNOTES
1. International Energy Agency, World Energy Outlook 2006, p. 347, table 13.1; NUKEM, “Data Feature: 2005/2006 World Nuclear Electricity Generating Capacity,” December 2006.
2. International Atomic Energy Agency, “Energy, Electricity and Nuclear Power Estimates for the Period to 2030,” July 2006.
3. Energy Information Administration, U.S. Department of Energy, International Energy Outlook 2006, table F5.
4. Massachusetts Institute of Technology, “The Future of Nuclear Power: An Interdisciplinary MIT Study, 2003.” The study was co-chaired by John Deutch and Ernie Moniz.
5. S. Pacala and R. Socolow, “Stabilization Wedges: Solving the Climate Problem for the Next 50 Years With Current Technologies,” Science, August 13, 2004.
6. “At the point of use, the largest contributor to avoided carbon dioxide emissions is improved end-use efficiency, accounting for nearly two-thirds of total savings.” International Energy Agency, World Energy Outlook 2006, p. 190. “The electricity saved in the residential and commercial sectors combined accounts for two-thirds of all the electricity savings in the Alternative Policy Scenario. By 2030, the savings in these two sectors avoid the need to build 412 gigawatts of new capacity.” Ibid., p. 213.
7. United Kingdom Sustainable Development Commission, “The Role of Nuclear Power in a Low Carbon Economy,” 2006.
8. International Energy Agency, World Energy Outlook 2006, p. 493.
11. Fast reactors utilize fast neutrons produced by fission. The typical commercial light-water reactor (LWR) uses a moderator to slow the neutrons to thermal speeds. Fast reactors, which have so far not been developed commercially, could in principle be used as breeders to produce more plutonium than they burn up or as burners to burn down plutonium and other transuranics as fuel. Absent their development, some of the countries separating plutonium have turned to using MOX fuels in LWRs. These fuels combine plutonium with standard or depleted uranium.
12. Jean-Michel Charpin et al., “Economic Forecast Study of the Nuclear Power Option,” July 2000. This study, whose principal author was the French planning commissioner, sought to compare systematically the back-end fuel cycle strategies for France.
13. Spent MOX fuel is hotter than spent uranium-oxide fuel, and the percentage of fissile isotopes in the plutonium in spent MOX fuel is less than in the plutonium in spent low-enriched uranium fuel and thus less attractive as a source for a new MOX assembly.
14. Erich Schneider and William Sailor, “Nuclear Fission,” Science & Global Security, Vol. 14, Nos. 2-3 (2006), pp. 194-196; Matthew Bunn et al., “The Economics of Reprocessing Versus Direct Disposal of Spent Nuclear Fuel,” Nuclear Technology, Vol. 150 (2005). It is very likely that, up until at least 2030, the LWR will remain the dominant reactor type, based on so-called Generation III or III+ designs, reactors with evolutionary improvements on existing LWRs. These reactor types include the European Pressurized Water Reactor (PWR) being marketed by Areva, the Westinghouse AP600 and AP1000 PWRs, and an Advanced Boiling Water Reactor being developed by General Electric. One non-LWR under consideration in China and South Africa that could conceivably be deployed before 2030 is the gas-cooled pebble-bed modular reactor. Generation IV reactors, including fast-neutron reactors, are the subject of research and development today but are unlikely to be deployed until well after 2030. See John Aheane, “Advanced Nuclear Reactors—Their Use in Future Energy Supply,” InterAcademy Council, 2005.
15. See International Energy Agency, World Energy Outlook 2006, table 13.12.
16. Son Kim and Jae Edmonds, “Nuclear Energy in a Carbon-Constrained World,” PNWD-SA 7184, November 1, 2005, figure 3-2; Erich Schneider and William Sailor, “Long Term Uranium Supply Issues and Estimates,” LAUR-05-8879, 2005; See Schneider and Sailor, “Nuclear Fission,” pp. 194-196.
17. Frank von Hippel, “Managing Spent Fuel in the United States: The Illogic of Reprocessing,” International Panel on Fissile Materials, January 2007.
18. J. Kessler, “Room at the Mountain,” EPRI Technical Update, May 2006. The implications of the uniqueness in the Yucca Mountain case of its limited geographic extent were pointed out to the author by Steve Fetter in a private communication.