3 November 2016

A fast reactor at any cost: The perverse pursuit of breeder reactors in India

M. V. Ramana

M. V. Ramana

Ramana has a doctorate in theoretical physics and is currently appointed jointly with the Nuclear Futures Laboratory...


In October 2016, Japan’s science and technology ministry announced that it was going to start decommissioning its Monju fast breeder reactor in 2020. Monju had a troubled history, and the decision to shut it down was in line with decisions in the United States and Western European countries to cancel their breeder reactor programs—which had come to be seen as unnecessary, unsafe, or uneconomical. (When something as expensive as a nuclear reactor is shut down, countries typically do not explicitly use words like “unsafe” or “unnecessary” or “uneconomical” in their official statements. Nevertheless, one can read between the lines of the documents issued by their public relations departments.)

In contrast, in October 2016, the chairman of India’s Atomic Energy Commission announced that the country’s Prototype Fast Breeder Reactor (PFBR) will be commissioned next year, with six more breeder reactors planned. Projections for the country’s nuclear capacity produced by India’s Department of Atomic Energy (DAE) call for constructing literally hundreds of breeder reactors by mid-century. For a variety of reasons, these projections will not materialize, making the pursuit of breeder reactors wasteful.

But first, some history. The DAE’s fascination with breeder reactors goes back to the 1950s. The founders of India’s atomic energy program, in particular physicist Homi J. Bhabha, did what most people in those roles did around that time: portray nuclear energy as the inevitable choice for providing electricity to millions of Indians and others around the world. At the first major United Nations-sponsored meeting in Geneva in 1955, for example, Bhabha argued for “the absolute necessity of finding some new sources of energy, if the light of our civilization is not to be extinguished, because we have burnt our fuel reserves. It is in this context that we turn to atomic energy for a solution… For the full industrialisation of the under-developed countries, for the continuation of our civilization and its further development, atomic energy is not merely an aid; it is an absolute necessity.” Consequently, Bhabha proposed that India expand its production of atomic energy rapidly.

There was a problem though. India had a relatively small amount of good quality uranium ore that could be mined economically, at least based on what was known at that time. But it was known that the country did have large reserves of thorium, a radioactive element that was considered a “great potential source of energy.” But despite all the praises one often hears about it, thorium has a major shortcoming: It cannot be used to fuel a nuclear reactor directly but has to first be converted into the chain-reacting element uranium-233, through a series of nuclear reactions. To produce uranium-233 in large quantities, Bhabha proposed a three-step plan that involved starting with the more readily available uranium ore. The first stage of this three-phase strategy involves the use of uranium fuel in heavy water reactors, followed by reprocessing the irradiated spent fuel to extract the plutonium. In the second stage, the plutonium is used to provide the startup cores of fast breeder reactors, and these cores would then be surrounded by “blankets” of either depleted or natural uranium to produce more plutonium. If the blanket were thorium, it would produce chain-reacting uranium-233. Finally, the third stage would involve breeder reactors using uranium-233 in their cores and thorium in their blankets. Breeder reactors, therefore, formed the basis of two of the three stages.

Bhabha was hardly alone in thinking of breeders. The first breeder reactor concept was developed by Leό Szilárd in 1943, who was responding to concerns, shared by colleagues who were engaged in developing the first nuclear bomb, that uranium would be scarce. The idea of a phased program involving uranium and thorium had also been proposed in October 1954 by François Perrin, the head of the French Atomic Energy Commission, who argued that France will “have to use for power production both primary reactors [using natural or slightly enriched uranium] and secondary breeder reactors [fast neutron plutonium reactors] … in the slightly more distant future … this second type of reactor … may be replaced by slow neutron breeders using thorium and uranium-233. We have considered this last possibility very seriously since the discovery of large deposits of thorium ores in Madagascar.” (At that time, Madagascar was a French colony, achieving independence only in 1960.)

That was then. In the more than 60 years that have passed since the adoption of the three-phase plan, we have learned a lot about breeder reactors. Three of the important lessons are that fast breeder reactors are costly to build and operate; they have special safety problems; and they have severe reliability problems, including persistent sodium leaks.

These problems were observed in countries around the world, and have not been solved despite spending over $100 billion (in 2007 dollars) on breeder reactor research and development, and on constructing prototypes.

India’s own experience with breeders so far consists of one, small, pilot-scale fast breeder reactor, whose operating history has been patchy. The budget for the Fast Breeder Test Reactor (FBTR) was approved by the Department of Atomic Energy in 1971, with an anticipated commissioning date of 1976. But it was October 1985 before the reactor finally attained criticality, and a further eight years (i.e., 1993) elapsed before its steam generator began operating. The final cost was more than triple the initial cost estimate. But the reactor’s troubles were just beginning.

The FBTR’s operations have been marred by several accidents of varying intensity. Dealing with even relatively minor accidents has been complicated, and the associated delays have been long. As of 2013, the FBTR had operated for only 49,000 hours in 26 years, or barely 21 percent of the maximum possible operating time. Although the FBTR was originally designed to generate 13.2 megawatts of electricity, the most it has achieved is 4.2 megawatts. But rather than realizing that the FBTR’s performance was typical of breeders elsewhere and learning the appropriate lesson—that they are unreliable and susceptible to shutdowns—the DAE terms this history as demonstrating a “successful operation of FBTR” and describes the “development of Fast Breeder Reactor technology” as “one of the many salient successes” of the Indian nuclear power program.

Even before the Fast Breeder Test Reactor had been constructed, India’s Department of Atomic Energy embarked on designing a much larger reactor, the previously mentioned Prototype Fast Breeder Reactor, or PFBR. Designed to generate 500 megawatts of electricity, the PFBR would be nearly 120 times larger than its testbed cousin, the FBTR. The difficulties of such scaling-up are apparent when one considers the French experience in building the 1,240 megawatt Superphenix breeder reactor; that reactor was designed on the basis of experience with both a test and a 250-megawatt demonstration reactor and still proved a complete failure. Nonetheless, the DAE pressed on.

Full steam ahead. Work on designing the PFBR started in 1981, and nearly a decade later, the trade journal Nucleonics Week reported that the Indian government had “recently approved the reactor’s preliminary design and … awarded construction permits” and that the reactor would be on line by the year 2000.

That was not to be. After multiple delays, construction of the PFBR finally started in 2004; then, the reactor was projected to become critical in 2010. The following year, the director announced that the project “will be completed 18 months ahead of schedule.”

The saga since then has involved a series of delays, followed by promises of imminent project completion. The current promise is for a 2017 commissioning date. Regardless of whether that happens, the PFBR has already taken more than twice as long to construct as initially projected. Alongside the lengthy delay comes a cost increase of nearly 63 percent—so far.

Even at the original cost estimate, and assuming high prices for uranium ($200 per kilogram) and heavy water (around $600 per kilogram), my former colleague J. Y. Suchitra, an economist, and I showed several years ago that electricity from the PFBR will be about 80 percent more expensive in comparison with electricity from nuclear power plants based on the heavy water that the DAE itself is building. These assumptions were intended to make the PFBR look economically more attractive than it really will be. A lower uranium price will make electricity from heavy water reactors cheaper. On the global market, current spot prices of uranium are around $50 per kilogram and declining; they have not exceeded $100 per kilogram for many years. Likewise, the heavy water cost assumed was quite high; the United States recently purchased heavy water from Iran at a cost of $269 per kilogram instead of the $600 per kilogram assumed figure.

The calculation also assumed that breeder reactors operate extremely reliably, with a load factor of 80 percent. (Load factors are the ratio of the actual amount of electrical energy generated by a reactor to what it should have produced if it had operated at its design level continuously.) No breeder reactor has achieved an 80 percent load factor; by comparison, in the real world the UK’s Prototype Fast Reactor and France’s Phenix had load factors of 26.9 percent and 40.5 percent respectively.

Consequently, even with very optimistic assumptions about the cost and performance of India’s Prototype Fast Breeder Reactor, and the deliberate choice of high costs for the inputs used in heavy water reactors, the PFBR cannot compete with nuclear electricity from the others kinds of reactors that India’s Department of Atomic Energy builds. With more realistic values and after accounting for the significant construction cost escalation, electricity from the Prototype Fast Breeder Reactor could be 200 percent more expensive than that from heavy water reactors.

But such arguments don’t resonate with DAE officials. As one unnamed official told sociologist Catherine Mei Ling Wong, “India has no option … we have very modest resources of uranium. Suppose tomorrow, the import of uranium is banned … then you will have to live with this modest uranium. So … you have to have a fast reactor at any cost. There, economics is of secondary importance.” This argument is misleading because India’s uranium resource base is not a single fixed number. The resource base increases with continued exploration for new deposits, as well as technological improvements in uranium extraction. In addition, as with any other mineral, at higher prices it becomes economic to mine lower quality and less accessible ores. In other words, if the price offered for uranium is higher, the amount of uranium available will be larger, at least for the foreseeable future.

One must keep these factors in mind when making economic comparisons between breeder reactors and heavy water reactors. Even for the earlier set of assumptions, without the dramatic cost increase of the PFBR factored in, breeders become competitive only when uranium prices exceeded $1,375 per kilogram—a truly astronomical figure, given the current spot price of $50 per kilogram. Significantly larger quantities of uranium will become available at such a price. In other words, the pursuit of breeder reactors will not be economically justified even when uranium becomes really, really scarce—which is not going to happen for decades, perhaps even centuries, given that nuclear power globally is not growing all that much.

The DAE, of course, claims that future breeder reactors will be cheaper. But that decline in costs will likely come with a greater risk of severe accidents. This is because the PFBR, and other breeder reactors, are susceptible to a special kind of accident called a core disassembly accident. In these reactors, the core where the nuclear reactions take place is not in its most reactive—or energy producing—configuration. An accident involving the fuel moving around within the core, (when some of it melts, for example) could lead to more energy production, which leads to more core melting, and so on, potentially leading to a large, explosive energy release that might rupture the reactor vessel and disperse radioactive material into the environment. The PFBR, in particular, has not been designed with a containment structure that is capable of withstanding such an accident. Making breeder reactors cheaper could well increase the likelihood and impact of such core disassembly accidents.

What of the DAE’s projections of large numbers of breeder reactors to be constructed by mid-century? It turns out that the methodology used by the DAE in its projections suffers from a fundamental error, and the DAE’s calculations have not accounted properly for the future availability of plutonium that will be necessary to construct the many, many breeder reactors the DAE proposes to build. What the DAE has omitted in its calculations is the lag period between the time a certain amount of plutonium is committed to a breeder reactor and when it reappears (along with additional plutonium) for refueling the same reactor, thus contributing to the start-up fuel for a new breeder reactor. A careful calculation that takes into account the constraints flowing from plutonium availability leads to drastically lower projections. The projections could be even lower if one takes into account the potential delays because of infrastructural and manufacturing problems. The bottom line: Even if all was going well, the breeder reactor strategy will simply not fulfill the DAE’s hopes of supplying a significant fraction of India’s electricity.

Ulterior motives? For all the praises it sings of breeder reactors, there is one reason for its attraction to the PFBR that the DAE does not talk much about, except indirectly. Consider this interview by the Indian Express, a national newspaper, with Anil Kakodkar, then-secretary of the DAE, about the US-India nuclear deal: “Both from the point of view of maintaining long-term energy security and for maintaining the minimum credible deterrent, the fast breeder programme just cannot be put on the civilian list. This would amount to getting shackled and India certainly cannot compromise one [security] for the other.” (There is some code language here. “Minimum credible deterrent” is a euphemism for India’s nuclear weapons arsenal. “Put on the civilian list” means that the International Atomic Energy Agency will not safeguard the reactor, and so it is possible for fissile materials from the reactor to be diverted to making nuclear weapons.)

What this points to is the possibility that breeder reactors like the PFBR can be used as a way to quietly increase the Department of Atomic Energy’s weapons-grade plutonium production capacity several-fold. But as mentioned earlier, this is not a reason that the DAE likes to publicly admit. Nevertheless, the significance of keeping the PFBR outside of safeguards has not been lost, especially on Pakistan.

Breeder reactors have always underpinned the DAE’s claims about generating large quantities of electricity. That promise has been an important source of its political power. For this reason, India’s DAE is unlikely to abandon its commitment to breeder reactors. But given the troubled history of breeder reactors, both in India and elsewhere, the more appropriate strategy to follow would be to simply abandon the three-phase strategy. The DAE’s reliance on a technology shown to be unreliable suggests that the organization is incapable of learning the appropriate lessons from its past and makes it more likely that nuclear power will never become a major source of electricity in India.