Reprocessing: Poised for growth, or on death’s door?

Some observers believe that plutonium reprocessing is on the verge of an expansion—while others argue that the end of the practice is in sight. The risk of nuclear proliferation has always been the chief objection to reprocessing, but proponents argue that today, with uranium enrichment technology more easily available, reprocessing no longer represents an efficient route toward nuclear weapons. Supporters also tout the energy security that reprocessing could provide to nations without indigenous uranium sources and the reductions in high-level nuclear waste that reprocessing might achieve. Opponents counter that reprocessing offers only marginal benefits in waste reduction and in any event makes little economic sense. Taking into account issues ranging from proliferation to nuclear waste to cost, how should nations approach plutonium reprocessing?

Round 1

Reprocessing in China: A long, risky journey

Since 1983, a closed fuel cycle has been an official element of China's nuclear energy policy. According to proponents, plutonium reprocessing and breeder reactors will allow full utilization of China’s uranium resources, drastically reduce the volume of radioactive waste that must be stored in an underground repository, and establish a way to dispense with the spent fuel accumulating in China’s reactor pools. But Beijing's attempts to develop commercially viable reprocessing facilities and breeder reactors have been afflicted with technological difficulties, serious delays, and cost overruns. At this point—especially taking into account China's ample uranium resources and its easy access to additional resources abroad—it appears very doubtful that reprocessing and fast reactors are the proper way forward for China's nuclear energy sector.

Not according to plan. In 1986, China's State Council approved construction of a pilot civilian reprocessing plant at the Jiuquan nuclear complex in Gansu province. Construction of the plant, designed to produce 50 metric tons of heavy metal a year, started in 1998 and finished in 2005. But the construction process encountered difficulties, delays, and higher-than-expected costs. Finally a hot test was conducted in 2010—24 years after the project's approval. Even then, after only 10 days of operation and the separation of less than 14 kilograms of plutonium, new problems were identified. As of the end of February, 2015, reprocessing had not resumed. Indications are that the plant's annual reprocessing capacity upon resumption could be much lower than the 50 metric tons of heavy metal originally planned.

Separately, the China National Nuclear Corporation (CNNC) has since 2007 been negotiating with France's Areva on the purchase of a commercial reprocessing plant capable of producing 800 tons of heavy metal a year. A series of agreements has been signed, but price remains a sticking point. And Chinese experts are not unanimous about whether China should import a commercial reprocessing plant at all. Some would like to fast-track the deal, while others believe that China should prioritize indigenous technology in order to maintain independence. Indeed, even amid its negotiations with Areva, CNNC began to plan a medium-scale demonstration reprocessing plant, using the pilot plant as a basis. The proposal has not been approved by the government, but in any event the future of the Areva deal is by no means clear.

Parallel with development of the pilot reprocessing plant, China has been working to establish commercially viable plutonium breeder reactors. According to a plan in place until 2013, development of breeder reactors was to be a three-stage process. The first stage was to complete a project known as the China Experimental Fast Reactor. The second stage would involve building, by about 2020, a few demonstration fast reactors. Finally, commercialized fast reactors would be deployed around 2030. Progress always ran far behind schedule.

The China Experimental Fast Reactor is a sodium-cooled experimental fast reactor using technology developed for Russia’s BN-600 reactor. The project, with a planned capacity of 20 megawatts, was approved in 1995. Construction began in 2000. As with the pilot reprocessing plant, the experimental fast reactor encountered many difficulties during construction. Capital cost estimates had to adjusted twice, with each estimate double the previous one. The reactor went critical in July 2010 and, by July 2011, 40 percent of its full power was incorporated into the grid. The reactor, however, was online for only 26 hours during the remainder of 2011, and it produced the equivalent of just one full power-hour. Not until December 2014 did the reactor manage to operate at full capacity for 72 hours. So 19 years passed between project approval and operation at full capacity.

As for the second stage of the pre-2013 plan, CNNC in 2009 signed an agreement with Russia’s Rosatom to jointly construct two copies of Russia’s BN-800 fast neutron reactor in China. But Beijing has not officially approved the project. As with the French reprocessing plant, Chinese experts complain that Russia is demanding too high a price. It is not clear when or if the project will go forward. Instead, CNNC in 2013 began focusing on the development of the indigenous 600-megawatt China Fast Reactor (CFR-600). The start of construction is envisioned for 2017, with operations to commence in 2023—but the government has not approved the project yet.

Experts from CNNC have also, since 2013, urged the development of China's first commercial fast reactor—a 1,000-megawatt reactor based on experience gained from the CFR-600. But CNNC expert Gu Zhongmao—an advocate of the closed fuel cycle—said at a recent workshop on nuclear energy in East Asia that “China needs at least another 20 to 30 years of effort before commercialization of fast reactor energy systems, and there are so many uncertainties ahead. It is beyond our ability to draw a clear picture 20 years ahead.”

Why rush? Should China continue pursuing its plans for fast breeder reactors and commercialized reprocessing? Good reasons exist for avoiding this course of action. First, because most of China's power reactors are newly built, Beijing will face little pressure over the next two decades to reduce its spent fuel burden. And spent fuel can be stored safely, at low cost, in dry casks—or disposed of safely in a deep geological repository.

Second, China faces no shortage of uranium resources for the foreseeable future. The nation's identified resources more than tripled between 2003 and 2012, to 265,500 metric tons from 77,000 metric tons. China's potential uranium reserves amount to more than 2 million tons. Beijing in recent times has also secured huge overseas uranium resources—about three times as large as its own identified uranium reserves. More such reserves could easily be added.

In any event, the cost of uranium accounts for only a small percentage of the cost of power that reactors generate. Simply put, the cost of uranium will not increase in the foreseeable future to levels that would justify the cost of reprocessing and breeder reactors. To the extent that China is concerned about potential disruptions in its uranium supply, it could easily and inexpensively establish a “strategic” uranium stockpile.

China should carefully examine the experiences of nations that have launched large reprocessing programs and built demonstration breeder reactors in the expectation that the commercialization of these reactors would follow. Commercialization did not follow in those countries—but huge expenses were incurred for cleaning up reprocessing sites and disposing of separated plutonium. For China, there is no urgent need to go down this risky road.

Plutonium recycling is much more expensive, and much less safe and secure, than operating light water reactors with a once-through fuel cycle. As for nuclear waste, dry cask storage is a safe, flexible, and low-cost option that can postpone for decades the need either to reprocess spent fuel or to dispose of it directly—allowing time for technology to develop. China has no convincing rationale for rushing to build commercial-scale reprocessing facilities or plutonium breeder reactors.


For sustainable nuclear energy, a closed fuel cycle

Reprocessing and recycling uranium and plutonium have been a matter of intense debate in many countries over the last few decades. A number of nations—including the United States, the United Kingdom, France, India, Japan, Russia, and others—have developed technologies for reprocessing and recycling. A select group of nations, including India, have followed a sustained policy in favor of reprocessing and recycling. They have pursued these policies not only out of concern over limited uranium resources, but also because they see reprocessing and recycling as the best path toward making nuclear energy sustainable over the long term.

Indeed, for countries such as India and China, large-scale expansion of nuclear energy is not sustainable without reprocessing and recycling. These countries' uranium resources are limited. Thorium seems unlikely to become a valuable resource for energy production in the near term. It is impractical to manage large volumes of high-level waste in repositories for long periods of time. Elsewhere, France has also had a sustained program for reprocessing and recycling, and Russia is pursuing pyroprocessing to produce plutonium-uranium mixed oxide fuel for its BOR-60 reactor.

Many critics of reprocessing believe that the world's available resources of uranium are adequate—and that closing the fuel cycle is therefore not an urgent necessity. This is a short-sighted argument. The quantity of uranium available on Earth—either in the planet's crust or in the waters of the oceans—is finite, and fission energy is therefore not renewable. The only debate about this question can concern how much uranium is available and how long it will last. The "Red Book" of uranium resources, produced by the International Atomic Energy Agency and the Nuclear Energy Agency, indicates that the world's available uranium as of 2013 amounted to about 7.6 million metric tons—enough to last for about 150 years at current rates of consumption. Consumption, though, is bound to increase as an increasing number of countries turn to nuclear energy to meet their energy requirements.

Putting aside additional uranium resources that may be identified in the future, and also putting aside nuclear energy's future growth rate, one must conclude that uranium-based fission energy cannot in any event last for more than a few centuries. This is not much time—when measured against the length of time that humankind is likely to exist. We, the authors, believe that the current generation bears a responsibility toward future generations not to deplete the world's uranium resources. This means that uranium cannot be discarded as waste after only 1 percent of its energy is utilized—as happens today. Rather, reprocessing and recycling must be pursued so that 75 percent of uranium (if not more) is used to produce fission energy. Reprocessing and recycling have the potential, compared to the once-through use of uranium, to increase by a factor of at least 50 the amount of time during which humankind can derive fission energy from uranium resources.

Several countries—India, France, Russia, and China in particular—have concluded that fast reactors (which can exploit for electricity generation the plutonium and depleted uranium generated in thermal reactors) will be an important element of their future nuclear power programs. And of the six concepts for innovative nuclear energy systems being developed by the Generation IV International Forum (a collective undertaking of 13 governments), four are based on fast reactors. But arguments against reprocessing and recycling continue to be made.

These include—in addition to the argument that uranium resources will remain adequate for the foreseeable future—the supposed immaturity of the technology, its perceived high cost, and proliferation considerations. But the technology surrounding fabrication and recycling of reactor fuel is in fact mature, and has already been developed on a reasonably large scale (though only in a handful of countries). France and India—through extensive research and development, stable government policy, and successive implementation of improved plants—have demonstrated the safety, maturity, and acceptable cost of reprocessing and recycling of plutonium. Meanwhile, resistance to proliferation can easily be built into the design of the fuel cycle. It is possible to envisage separation schemes in which uranium and plutonium from irradiated fuel are recovered together—and no pure plutonium product, which might cause proliferation concerns, is produced. Pyroprocessing, for example, which has undergone considerable research in the United States and Russia, achieves less decontamination of fission products than the aqueous processes based on the PUREX method (for plutonium uranium extraction), which have been the mainstay of the nuclear industry to this point. This provides inherent proliferation resistance. Proliferation concerns no longer constitute a convincing rationale for adopting a once-through fuel cycle.

The waste dimension. Management of nuclear waste is the single dimension of nuclear energy on which the public focuses most. The once-through fuel cycle presents two serious problems in this regard. First, because the once-through cycle involves disposing of uranium and plutonium as waste after only one use, a larger volume of waste results. This creates a greater need for repositories and an impractically long time period over which they need to be monitored—which the public is never likely to welcome. Second, a significant number of countries are establishing nuclear power sectors today, but most nations are not endowed with sites geologically appropriate for disposal of irradiated fuel. Who, then, will bear the burden of the spent fuel that these new nuclear power facilities generate?

Such issues must be taken into account when the cost of various approaches to the fuel cycle is calculated. Usually, recycling's cost is compared with the cost of the once-though cycle under the assumption that uranium will remain at its current cost. But even assuming that uranium's price stays stable for centuries—unlikely—one still must account for the once-through cycle's higher waste management costs. One must also account for the extra environmental harm associated with the greater mining requirements of the once-through cycle—if uranium and plutonium are recycled, less uranium needs to be mined per unit of energy produced. This sort of holistic approach to cost shows that reprocessing is not prohibitively expensive, as it is often portrayed, but in fact is cost-competitive.


How I changed my mind on reprocessing

This very personal story starts in the late 1960s, when I worked as a postdoctoral engineer on a short research project involving the Phénix fast breeder reactor at Cadarache, the research facility that was the center of France’s research into safety for fast breeder reactors. I, like nearly every other nuclear engineer at the time, was convinced that Glenn Seaborg, the Nobel Prize–winning chemist, was right: Breeder reactors and plutonium recycling would provide the world with unlimited cheap electricity. The consequences for the environment would only be positive. When my time at Cadarache came to an end, therefore, I decided to work in the country that was the apparent leader in breeders and reprocessing, and I moved on to General Electric’s Breeder Reactor Development Operation in Sunnyvale, California.

In those days, uranium was becoming more expensive. According to many, it was also becoming scarce. Therefore plutonium’s value was high—at least for military uses, and theoretically for civilian ones as well. When France and Germany decided in 1970 to undertake a large project demonstrating the commercial potential of fast breeder reactors, via the 1200-megawatt Superphénix reactor, I felt the need to participate in this European project. So I moved on again, but this time I took a position on the operator’s side, with the largest German utility, RWE. During the pre-project phase, RWE gave me the opportunity to learn project management at the SNR-300 reactor in Kalkar, Germany, which was then under construction.

RWE also sent me to the Phénix fast breeder reactor in France to follow its connection to the grid; this was part of the preparation for the Superphénix project. I fondly remember the view from the control room of the Phénix reactor (yes, it had a large window). I could see the Rhône and the many fighter planes that would take off from the airfield on the other side of the river. It was a wonderful period, full of optimism. Optimism seemed particularly justified in 1974 when André Giraud, head of France’s Atomic Energy Commission and later minister of industry, announced that international construction of the Superphénix reactor would soon begin. By 1990, construction of two additional 1500-megawatt fast breeder reactors would be completed.

But soon my optimism started to crumble somewhat. I was trying to get a clearer picture of the economics involved in breeder reactors and plutonium reprocessing by doing simple comparisons of the expenses involved in loop breeder reactors, pool breeder reactors, and light water reactors. I took into account the materials that go into constructing them, their ability to vary electricity output as necessary, and so on. I came to the conclusion that fast breeder reactors would always end up with considerably higher capital costs than light water reactors—at least 30 to 50 percent higher. I reluctantly concluded that commercial fast breeder reactors would not succeed in my generation.

Unless, of course, changes in fuel cycle costs came to the rescue. And indeed, in the late 1960s, the United Kingdom’s Atomic Energy Authority had begun offering reprocessing at unbeatable prices—just $15 per kilogram of heavy metal. Eurochemic, a joint European reprocessing facility, offered low prices as well. So did WAK, a small German reprocessing facility that, despite its affordable prices, still had difficulties filling its order book as of 1971. Uranium prices, on the other hand, were on the increase for several years beginning in about 1975, making things difficult for firms such as Westinghouse, which sold fuel reloads at fixed prices.

Then, in 1976, the German government mandated reprocessing as the only legal approach to the back end of the fuel cycle. This forced the nation’s utilities to pursue reprocessing. But WAK was the only reprocessing facility in the whole country. So fulfilling the government’s mandate meant relying on foreign contracts, and BNFL—a successor to Britain’s Atomic Energy Authority—lacked capacity after a fire at its reception facility. I moved back to Germany to negotiate reprocessing contracts with COGEMA, predecessor of the French nuclear energy firm Areva, and to identify storage capacity for spent fuel until COGEMA could begin receiving fuel. In parallel we developed dry cask storage as an alternative.

But by this time—1978—COGEMA would offer reprocessing only under cost-plus conditions, and it expected a hefty profit margin of 25 percent. Meanwhile, big away-from-reactor storage pools with aircraft and sabotage resistance would not be online for some time. They were also more expensive than the alternative—dry cask storage. By 1979 RWE had scrapped away-from-reactor pool projects in favor of casks.

In 1982 and 1983 COGEMA came online with big pools that offered relief to German utilities. But the future of reprocessing became bleak nevertheless. When COGEMA quoted the owners of the Superphénix reactor unexpectedly high prices for reprocessing, the owners decided to build a large on-site storage pool for their spent fuel. That was the real end of fast breeder reactors in France—not the political shutdown decision in 1998. The cost realities had sunk in. Breeders without a commercial fuel cycle just don’t make sense.

The German utilities realized in 1989 that reprocessing would entail unacceptable costs, even without taking into account the expense of fabricating plutonium-uranium mixed oxide fuel. And the high burn-up rates of uranium fuel made reprocessing even less economical. At the same time, it was becoming clear that uranium was not scarce at all. In fact, it was now available again at decent prices. Also, reprocessing projects in the United Kingdom and Japan never experienced the same production success as in France, and economic prospects for reprocessing became very dark.

Today, plutonium is no longer a high-value asset on a utility’s balance sheet. It is at best a zero-value item. But except in countries that might successfully develop cheap breeder reactors and an affordable fuel cycle (China, India, and Russia still maintain such hopes), it’s more likely to be a negative-value item. One fears that privately owned utilities will become hostages of their own plutonium. For national utilities, meanwhile, it is the government—or, more precisely, the taxpayer—who will become the hostage.

Is it possible, someday, that Seaborg will be proven right after all? Well, I concluded more than 30 years ago that commercial fast breeder reactors would not become a reality in my generation. The same still appears to be true for the young generation today. And this conclusion does not even take into account the proliferation threat that plutonium poses.


Round 2

Long-term ideal versus short-term reality

Baldev Raj and P.R. Vasudeva Rao have argued that reprocessing and fast breeder reactors are necessary for the long-term sustainability of nuclear power. Indeed, fast breeder reactors' potential to produce more fuel than they consume has held an attraction since the advent of nuclear power, especially to those who envision a time when uranium will no longer be available cheaply. Unfortunately, several decades of experience have shown that plutonium recycling systems are much more costly and much less reliable than water-cooled reactors. If establishing sustainable nuclear power means successfully managing important issues such as nuclear safety and proliferation resistance, while also achieving economic competitiveness, minimizing production of radioactive waste, and using natural resources wisely, breeder reactors and plutonium recycling still have far to go before they can meaningfully contribute.

Raj and Rao wrote in Round One that for countries such as India and China, with limited uranium resources, "large-scale expansion of nuclear energy is not sustainable without reprocessing and recycling." But a country’s own limited uranium resources do not necessarily constrain its nuclear power development. In fact, the global distribution of uranium resources can generally be characterized this way: Countries with more nuclear energy have less uranium, and countries with more uranium have less nuclear energy. Trade in uranium naturally constitutes a global market.

In Round Two, Raj and Rao went on to write that "discussions about long-term, sustainable sources of energy cannot be based solely on today’s economics. Uranium prices will go up when supplies begin to run out." But past predictions that uranium prices would steadily rise have been proven wrong. Even when demand has increased, uranium prices have remained relatively low. This is unsurprising in a way—prices for most minerals have decreased in constant dollars over the past century even as extraction has increased. For uranium, increased exploration and advances in technology have caused known resources to increase faster than uranium has been depleted. Known uranium resources are a dynamic economic concept, and global resources will surely prove to be greater in the long run than the amount currently reported in the Red Book.

Raj and Rao also argue for plutonium recycling on the grounds that it produces lower volumes of nuclear waste than the once-through cycle produces. But plutonium reprocessing and recycling still produce high-level waste, long-lived intermediate waste, and low-level waste. All of these waste streams must eventually be buried, so reprocessing does not eliminate the need for repositories. Moreover, a geological repository's capacity is determined by the waste's decay heat, not the physical volume of waste. Thus, at the geological repository for high-level waste that China envisions establishing in Gansu Province, capacity would merely double if all transuranic elements were separated from nuclear waste—the same increase that could be achieved by waiting 100 years before burying the waste. China, therefore, instead of building an expensive reprocessing plant, could opt for relatively low-cost dry cask storage. In the end, the capacity of geological repositories is increased to only a tiny extent by reprocessing plutonium and recycling it once via mixed-oxide fuel.

Regarding the proliferation risks that breeders and reprocessing might introduce, Raj and Rao have written that "resistance to proliferation can easily be built into the design of the fuel cycle" and that pyroprocessing "provides inherent proliferation resistance." It is true that pyroprocessing does not produce pure plutonium, as traditional PUREX reprocessing does. But the end product of pyroprocessing is much less radioactive than is spent fuel itself. It would be a relatively straightforward process to separate plutonium after pyroprocessing is complete—easier than to separate plutonium directly from spent fuel. Raj and Rao also portray plutonium recycling—as opposed to separation—as a "nonproliferation measure." But plutonium recycling and breeders require that plutonium be separated in the first place, allowing it to be put to military uses. Indeed, India's 1974 "peaceful" nuclear explosion used plutonium purportedly separated for use in the country's breeder program. And even if governments have no desire to proliferate, separated plutonium is much more vulnerable to theft or misuse than is spent fuel.

Finally, Raj and Rao contend that "the current generation bears a responsibility toward future generations not to deplete the world's uranium resources." But if the current generation is unable to guarantee the safe and secure operation of nuclear power plants today, what's the point of maximizing uranium resources for future generations? In particular, what's the point of doing so through breeder reactors and plutonium reprocessing—problematic technologies that introduce additional security risks?


Not whether, but when

In Round One, Hui Zhang wrote that China shouldn't rush into development of commercial-scale reprocessing and breeder reactors. In Round Two, Janberg generalized Zhang's point by posing the following question to the authors of this essay: "Why rush into breeders and reprocessing?" But we, the authors, haven't argued that great urgency surrounds plutonium reprocessing or breeder reactors. Rather, our argument is that widespread use of these technologies is inevitable over a long timeline (assuming that the world’s uranium resources are to be used effectively). Neither Zhang nor Janberg has addressed the long-term sustainability of nuclear power so far in this roundtable. Both authors seem to think it satisfactory if fission energy is produced for only a few more decades.

Janberg and Zhang have both built their arguments against breeders and reprocessing partly on the basis of economics—uranium prices in particular. But discussions about long-term, sustainable sources of energy cannot be based solely on today’s economics. Uranium prices will go up when supplies begin to run out; it's only a question of time. But then, perhaps only people in countries starved of natural resources can appreciate what Homi Jehangir Bhabha, father of India's nuclear program, meant when he said that "no power is costlier than no power." That is, no means of generating energy poses greater burdens than the lack of energy poses.

Janberg likewise depends on short-term rather than long-term economics when he describes reaching his conclusion that "fast breeder reactors [will] always end up with considerably higher capital costs than light water reactors—at least 30 to 50 percent higher." Janberg's conclusion is hurried. It is derived from comparing a mature technology (light water reactors) to technologies that haven't yet been demonstrated on a comparable scale (breeders and reprocessing). So when Janberg presents his belief that "commercial fast breeder reactors [will] not succeed in my generation," we can only reply that our concern is not the current generation. It's future generations.

Zhang, meanwhile, wrote in Round One that "Plutonium recycling is much more expensive … than operating light water reactors with a once-through fuel cycle." But operational experience with reprocessing plants around the world is inadequate to make such a pronouncement. Indeed, France has pursued a commercial-scale reprocessing and recycling program with good success.

Waste and weapons. Janberg and Zhang also give short shrift to plutonium reprocessing and breeder reactors when it comes to reducing volumes of nuclear waste. But what then is to be done with the nuclear waste produced by the once-through fuel cycle? Neither author mentions Yucca Mountain—the nuclear waste repository in the United States that has faced intense political opposition during decades of planning and is still not close to becoming operational. If all countries with nuclear power sectors adopt a once-through fuel cycle, many Yucca Mountains will be required for waste disposal. This hardly seems feasible. Do all nations have space available for such facilities? Who will bear the expense of monitoring and ensuring these facilities' safety for centuries?

A final point: Janberg wrote in Round Two that perhaps "proliferation doesn't seem so important a concern" in India, and that India's 1974 nuclear detonation was the "ignition point" for nuclear weapons programs in many countries. It's somewhat unfortunate that Janberg chose to discuss proliferation in these terms. We, the authors, are participating in this roundtable as experts in fast reactors and the fuel cycle—we are not participating to present "the Indian position." But since Janberg has raised these issues, we will point out that India is recognized as a responsible nuclear power with an immaculate nonproliferation record. In any event, in discussions of proliferation one must distinguish between reprocessing and recycling. Could reprocessing lead to proliferation? That is a country-specific issue. But recycling, on the other hand, is a nonproliferation measure. Can a safer place be imagined for plutonium than the core of a reactor?


New era, same arguments

Life offers wonderful surprises, even when you reach old age. In my roundtable colleague Hui Zhang, I have discovered a Chinese “twin” in spirit! In fact, I would address to our colleagues Baldev Raj and P.R. Vasudeva Rao the question at the heart of Zhang's Round One essay: Why rush into breeders and reprocessing? Why advance the same arguments that were advanced in the 1960s and 1970s—when, over the decades, these arguments have proven unsound in multiple countries?

Raj and Rao base much of their Round One essay on the idea that global uranium resources are inadequate. But uranium ore, in constant dollars, is cheaper today than in the past. True, prices increased just before Fukushima, but afterwards fell so low that many mines couldn't cover their costs. Production had to be reduced and mining operations had no incentive for further exploration. Until mines can invest again, supplies will be short and prices will rise—but short-term fluctuations aside, the uranium industry today finds it challenging just to survive.

Proven, easily accessible reserves of uranium will be adequate to meet demand for decades to come. But Raj and Rao choose to analyze fuel supply over a very long time scale—writing "one must conclude that uranium-based fission energy cannot … last for more than a few centuries." Even if this is true, it expresses an incredible lack of trust in market forces and human ingenuity. History has shown time and again that if human beings face an urgent need, a path toward meeting that need will be discovered. Or existing methods will be improved so that extra time is gained for research and development.

Indeed, many energy technologies under development today may reduce the need for fission energy in the future. I won't argue that photovoltaic cells, for example, are the answer—but their efficiency is steadily improving and their capital costs are decreasing, while nuclear reactors are growing more expensive (for new reactors at Finland's Olkiluoto plant and France's Flamanville facility, costs have more than tripled). For that matter, do Raj and Rao believe that humanity won't deliver adequate electricity storage systems within the next, say, 15 to 20 years? What pessimism for researchers to display!

In any event, if plutonium is to be the answer to a possible uranium shortage in the future, significant stockpiles of plutonium already exist. The United Kingdom has on hand more than 110 metric tons of civilian plutonium—and no reactors in which the fuel is used. Japan has a stockpile of about 47 metric tons (stored in Japan and elsewhere). France is home to more than 20 metric tons of plutonium. These stockpiles constitute a very meaningful buffer against future shortages of nuclear fuel.

Raj and Rao also argue for reprocessing on the basis of its purported benefits in nuclear waste reduction, maintaining that a once-through cycle results in a larger volume of waste than does the closed fuel cycle. They're correct—if one focuses only on spent fuel and has in mind the heat released in disposal mines. But reprocessing also involves discharges of liquid and gaseous wastes into the environment. Then there are the vitrified wastes that reprocessing entails; the hulls and structures of spent-fuel assemblies that must be disposed of; and the solidified residues of the reprocessing process itself. For the sake of completeness, one should also mention the waste that ultimately results from dismantling a reprocessing facility. These waste streams make the once-through cycle clearly superior in terms of waste volumes. Pyroprocessing will not alter this reality in any significant way.

Finally, there is the proliferation issue. Raj and Rao write that "Proliferation concerns no longer constitute a convincing rationale for adopting a once-through fuel cycle." Perhaps for India, which hasn't signed the Nuclear Non-Proliferation Treaty, proliferation doesn't seem so important a concern. But it has been a grave concern for many other countries since India's "peaceful" nuclear detonation in 1974. That detonation, which used plutonium reprocessed from spent reactor fuel, was the ignition point for nuclear programs in Pakistan, North Korea, Libya, Iraq, and perhaps others countries not yet recognized. I concede that nowadays it may be easier for some nations to produce fissile material through uranium enrichment. But reprocessing, even if it doesn't achieve full separation of uranium and plutonium from spent fuel, still provides the easiest route to a "dirty bomb"—and a quick route too, complicating surveillance efforts. I believe that the proliferation risks associated with reprocessing still exist—and are in fact increasing.


Round 3

Pay more, risk more, get little

In Round Three, Baldev Raj and P.R. Vasudeva Rao claim to have identified "several technical inaccuracies" in my Round Two essay. I consider their claims misleading.

For example, I wrote that "high-level waste, long-lived intermediate waste, and low-level waste [from plutonium reprocessing and recycling] … must eventually be buried, so reprocessing does not eliminate the need for repositories." Raj and Rao respond that "only high-level waste requires burial in deep geological repositories. Intermediate waste can be buried in shallower repositories." But long-lived intermediate-level waste produced through plutonium reprocessing and recycling indeed must be buried in a deep repository (while shallow burial is adequate for short-lived radioactive waste). My key point, however, was that reprocessing does not eliminate the need for repositories. Raj and Rao essentially ignore this idea.

Raj and Rao also mischaracterize my argument that the capacity at China's envisioned repository for high-level waste in Gansu Province would merely double if all transuranic elements were separated from nuclear waste—and that the same increase could be achieved by waiting 100 years before burying the waste. Raj and Rao dispute my statements by comparing the toxicity of reprocessed waste to the toxicity of waste that has not been reprocessed. But my focus in Round Two was on reductions in the volume of the waste that China envisions placing in its Gansu repository. I made no claim about toxicity.

All in all, regarding volumes of nuclear waste, I agree with my roundtable colleague Klaus Janberg, who wrote in Round Two that a thorough accounting of all waste streams involved in reprocessing and breeder reactors makes the once-through cycle "clearly superior."

Easy choice. Lately, advocates for fast neutron reactors have been arguing that breeders and reprocessing can reduce the long-term hazards associated with burial of high-level waste. But these long-term benefits are offset by short-term risks and costs. For example, breeder advocates argue that the risks surrounding leakage in geological repositories could be reduced if all the long-lived isotopes of plutonium and other transuranics contained in spent fuel were transmuted (or fissioned), thus significantly reducing the doses of radioactivity that could escape due to any leakage. But studies show that long-lived fission and activation products in spent fuel—not isotopes that could be fissioned through breeders and reprocessing—dominate the radioactivity doses that leakage could release. Plutonium, in fact, is quite insoluble in deep underground water. So, reprocessing delivers no obvious long-term benefits in reducing leaked doses of radioactivity—but it does involve routine releases of long-lived radioactive gases from spent fuel. Reprocessing also increases the risk that tanks for high-level liquid waste might explode. (In a similar vein, advocates for fast neutron reactors argue that reprocessing, by reducing the need to mine uranium, can reduce human radiation exposure. But any such benefit is canceled out because plutonium reprocessing and recycling themselves expose workers and the public to radiation. In short, the net effects may well be negative.)

Meanwhile, all reprocessing and fast neutron reactor programs currently under consideration significantly increase the economic costs of nuclear energy. This means that nuclear decision makers must choose between achieving rather insignificant reductions in the long-term hazards associated with nuclear waste—and achieving short-term gains in the areas of safety, security, human health, and the environment. The choice seems rather clear-cut. The US National Academy of Sciences concluded in 1996, based on a review of the costs and benefits of reprocessing and fast neutron reactor programs, that "none of the dose reductions seem large enough to warrant the expense and additional operational risk of transmutation." That assessment remains valid today.

Finally, Raj and Rao assert that Janberg and I have not contested their argument that "today's generation bears an obligation to future generations not to deplete the world's uranium resources." Instead, Raj and Rao write that Janberg and I "have focused on economic questions, waste issues, and also on India's nuclear weapons program… ." But why shouldn't Janberg and I discuss issues—such as economics, waste, and nuclear proliferation—that must be satisfactorily addressed if sustainable systems for nuclear power are to be established? And even if one devotes as much attention to the future availability of uranium as Raj and Rao would like, does reprocessing make sense if uranium remains cheaply available, as seems likely, for a long time to come?


Closing the fuel cycle: Majority viewpoint

Klaus Janberg and Hui Zhang have not contested the central argument that we have advanced in favor of reprocessing and breeder reactors: that today's generation bears an obligation to future generations not to deplete the world's uranium resources. Instead, Janberg and Zhang have focused on economic questions, waste issues, and also on India's nuclear weapons program—which comes out of context, as this roundtable's purpose is to discuss a subject, not any nation's position on that subject.

In Zhang's Round Two essay, we also identify several technical inaccuracies. For example, Zhang argues that "high-level waste, long-lived intermediate waste, and low-level waste [from plutonium reprocessing and recycling] … must eventually be buried, so reprocessing does not eliminate the need for repositories." This is far from accurate. The fact is that only high-level waste requires burial in deep geological repositories. Intermediate waste can be buried in shallower repositories—which are easier to establish. And low-level waste can be diluted and dispersed, with no need for repositories at all. Indeed, this represents the international approach to waste management (not merely India’s).

Zhang also suggests that the once-through cycle requires no more space in geological repositories than the closed fuel cycle demands—as long as waste is stored for 100 years in dry casks. Again, this is far from correct. Numerous papers in the literature clearly illustrate that, via the once-through fuel cycle, the toxicity of high-level waste is reduced to levels equivalent to that of natural uranium only after tens of thousands of years. But when transuranics are removed through reprocessing, toxicity levels equivalent to that of natural uranium are achieved after just 300 years. This is a fundamental truth based on the nature of radioactive decay. It cannot be wished away. In any event, Zhang and Janberg's emphasis on dry cask storage is misguided. Is there any way to guarantee the safety of spent fuel stored in casks for hundreds or thousands of years? As we wrote in Round Two, the safest place for plutonium is inside a reactor. Dry casks can't compare.

Now or later. Another important advantage of reprocessing, to which Zhang and Janberg have given no attention, is its ability to reduce the need for uranium mining. Of all the stages in the nuclear fuel cycle, including ultimate waste management, mining is responsible for delivering the highest doses of radioactivity to human populations. The 2015 International Congress on Advances in Nuclear Power Plants, recently conducted in Nice, France, included presentations and discussions about nuclear systems powered by depleted uranium and plutonium (derived through reprocessing) and their potential to reduce the need for mining uranium, if not obviate it completely.

Zhang wrote in Round Two that "breeder reactors and plutonium recycling still have far to go before they can meaningfully contribute" to the establishment of sustainable nuclear power. We are pleased to see that Zhang believes breeders and reprocessing might meaningfully contribute at some point, even if the time isn't now. But that, actually, is the focus of our argument—that the wide-scale adoption of breeder reactors and plutonium processing is a question of when, not whether. The global need to achieve energy sustainability has decided the "whether." The "when" depends on individual nations' policies, technological maturity, and energy requirements.

To our knowledge, no nation has ever taken the view that the closed fuel cycle will never be required. In fact, only a very few countries favor the once-through cycle, and some of these nations intend to phase out nuclear power entirely. The balance of nations hold the view that the closed fuel cycle is not now economical in their countries, or is not now required. But they reserve the option to adopt the closed fuel cycle in the long run. The importance of fast reactors and plutonium reprocessing has often been acknowledged in international programs such as the International Congress on Advances in Nuclear Power Plants and the Generation IV International Forum. The reasoning behind this majority position is that if nuclear power is to be sustainable—where the parameters of sustainability include economics, safety, and waste management—the closed fuel cycle and fast reactors are indispensable.


An indefensible waste

In Round Two, Baldev Raj and P.R. Vasudeva Rao continued to claim that plutonium reprocessing and breeder reactors offer significant benefits in waste reduction. They also downplayed the economic challenges and proliferation risks associated with breeders and reprocessing. Their views on these issues do not hold up to scrutiny.

Regarding waste, Raj and Rao discussed the failure so far to open the Yucca Mountain nuclear waste repository in the United States. "If all countries with nuclear power sectors adopt a once-through fuel cycle," they wrote, "many Yucca Mountains will be required for waste disposal," which they called "hardly feasible." So doesn't it make sense, their argument runs, to reduce waste volumes through reprocessing? But the fact is that no geological repository for high-level waste is operating anywhere in the world. So what would it matter if waste volumes were reduced through reprocessing? Then again, as I argued in Round Two, a full accounting of waste streams shows that the once-through fuel cycle produces lower waste volumes than breeders and reprocessing.

In any event, a better approach to the waste problem is dry cask storage. A typical nuclear power plant, operating for 50 years, produces waste that, stored in dry casks, could fit in a plant's empty turbine hall. Moreover, this method of storage does not produce the many operational and dismantling wastes that are involved in reprocessing.

Regarding economics, Raj and Rao wrote that "operational experience with reprocessing plants around the world is inadequate" to conclude they are more expensive than light water reactors and the once-through cycle. "France," they wrote, "has pursued a commercial-scale reprocessing and recycling program with good success." I will agree that UP3, France's reprocessing facility for foreign customers, has been a surprising technical success. But commercially, the story is very different. I can testify to this, having been personally involved with the facility to some extent. At one time, France was able to sign reprocessing customers to contracts under cost-plus terms—the customers assumed all risks, from licensing on through to dismantling—because the customers' governments compelled them to sign. These contracts, which never would have been concluded without government pressure, are now expired or expiring. New customers in sufficient numbers are not in the offing, and the French national electricity utility shows no great enthusiasm to fill the order void. And meanwhile, reprocessing facilities in other countries—West Valley in the United States and Sellafield in the United Kingdom—were total failures.

When it comes to proliferation, Raj and Rao argued that plutonium recycling is a "nonproliferation measure"—but, as Hui Zhang has already pointed out, plutonium must be separated before it can be recycled, making it much more susceptible to misuse than if it were still contained in spent fuel. Again, dry cask storage is a better option. Spent fuel's high radioactivity protects it against theft for more than 150 years. Isn't a century and a half long enough to determine whether plutonium recycling is really necessary?

I am convinced that allocating large sums of money to fast breeder reactors over the next 30 years—a time period during which breeders and reprocessing cannot reach economic maturity—is an indefensible waste. Better alternatives for investment exist. I will not claim that Germany's Energiewende, or energy transformation, is an exemplary alternative. In fact, I consider it very wasteful. But investment should be directed toward the promising energy alternatives that do exist—not toward breeder reactors and plutonium reprocessing.



Receive Email