DRAFT

Attempts to Reduce the Proliferation Risks of Nuclear Power:
An Overview of Whats Old and Whats New

Marvin Miller
Security Studies Program & Department of Nuclear Engineering
Massachusetts Institute of Technology

Prepared for a Conference on Nuclear Power and the Spread of Nuclear Weapons: Can We Have One Without the Other?,
sponsored by the Nuclear Control Institute, Washington, DC, April 9, 2001

Introduction

At the same time that Amory Lovins, was posing the question, Can we have nuclear power without proliferation? in his book, Soft Energy Paths, [1] a prominent nuclear scientist was wrestling with the same issue in a series of lectures on the energy problem delivered at the Technion in Israel. Commenting on opposition to nuclear power on the grounds that it would give both sub-national groups and nations access to weapons-useable materials, he argued that the sub-national threat could be handled by increased security, but added: [2]

I wish I could be as optimistic and positive about the remaining objection: as nuclear reactors spread among nations their production will enable almost every country to acquire nuclear weapons. This statement, most unfortunately, is true. I believe that eventually nuclear proliferation is unavoidable unless we find better solutions to international problems than are now on the horizon.

Edward Teller was neither the first nor the best-known nuclear scientist who was concerned that nuclear power would facilitate the acquisition of nuclear weapons. Similar sentiments were expressed by Enrico Fermi: It is not certain that the public will accept an energy source that produces vast amounts of radioactivity as well as fissile material that might be used by terrorists. [3] And there is the oft-quoted statement in the Acheson-Lilienthal Report of 1946 that stressed the inadequacy of international inspections to prevent proliferation: [4]

There is no prospect of security against atomic warfare in a system of international agreements to outlaw such weapons controlled only by a system, which relies in inspections and similar police-like methods. The reasons supporting this conclusion are not merely technical, but primarily the insuperable political, social, and organizational problems involved in enforcing agreements between nations each free to develop atomic energy but only pledged not to use bombsSo long as intrinsically dangerous activities [i.e., production and use of weapons-useable materials such as plutonium and highly-enriched uranium] may be carried out by nations, rivalries are inevitable and fears are engendered that place so great a pressure upon a system of international enforcement by police methods that no degree of ingenuity or technical competence could possibly hope to cope with them.

However, in the euphoria generated by Atoms for Peace, most people reassured themselves that the type of safeguards system that had been judged to be inadequate in the Acheson-Lilienthal Report could nevertheless minimize proliferation risks. Proliferation might be a problem down the road, but it was difficult to stand in the way of a technology that would make the deserts bloom and would be too cheap to meter. Developing as well as developed countries were eager to avail themselves of the benefits of this new energy source. Publicly, this meant peaceful use, i.e., power, desalination, and production of special isotopes for medicine and agriculture. So nuclear technology flowed out of countries like the US and the Soviet Union, and foreign students and scientists flowed in, eager to learn the tricks of the nuclear trade. However, it was clear that the some of the same technologies, materials, and manpower could be applied to making weapons, and that safeguards could not prevent their diversion to such use.

The wake-up call on the linkage between the peaceful and military atom was the Indian test in 1974. Much has been written about the Indian nuclear program, [5] and others will address it here. So I confine myself here to the observation that Homi Bhabhas strong stance that India would never accept colonialism in the nuclear sphere, coupled with the eagerness of the nuclear suppliers to sell their wares, provided India with the opportunity to produce plutonium and then use it in a peaceful nuclear explosive.

The ensuing efforts to minimize the risk that civilian nuclear activities could be used as a cover for a weapons program encompassed both national and international initiatives. Since the US was the key player in these efforts, I turn next to the reassessment of non-proliferation policy in the US that started at the end of the Ford administration and was vigorously pursued by the incoming Carter administration.

The plutonium economy: problem or solution?

As many attendees at this meeting will recall, the debate about the connection between nuclear power and nuclear weapons was particularly heated during the Carter administration. The concern of the Carterites stemmed from the 1974 Indian test and the prospect that a rapid spread of nuclear power after the 1973 oil crisis would provide the rationale for the acquisition of the materials and technologies that had been deemed dangerous in the Acheson-Lilienthal Report: highly-enriched uranium and plutonium, as well as the uranium enrichment and reprocessing technologies which can produce these materials from natural or low-enriched uranium and irradiated reactor fuel, respectively.

The counter-argument that international safeguards at enrichment and reprocessing plants would be able to detect and hence deter the production of significant quantities of weapons-useable materials was met with skepticism that such methods could detect such actions in a timely manner, that is, before their use in weapons. It was also pointed out that safeguards can only be effective when applied, and non-NPT states were not legally bound to accept safeguards on their indigenous facilities. Moreover, NPT states can legally withdraw from the treaty on three months notice, and there are no sanctions specified for violations of treaty commitments.

In sum, the only proliferation-resistant fuel cycles were those in which neither highly-enriched uranium nor plutonium were used in separated form. And any reprocessing or enrichment should preferably take place under international or multinational control. The Carter administration sought to implement these views by both domestic legislation and international persuasion, with a focus on eliminating commercial use of plutonium. However, although the International Nuclear Fuel Cycle Evaluation (INFCE) program, organized at the behest of the Carter administration, was supportive of the need to raise the level of proliferation consciousness, there was strong resistance in Western Europe and Japan to the attempt to delegitimize their use of plutonium in the nuclear fuel cycle.

In brief, the Europeans, Japanese, and their supporters elsewhere argued that: (1) Nuclear power was essential, and there wasnt enough uranium to support a large nuclear enterprise if only once-through fuel cycles were utilized; and (2) The proliferation risks involved in plutonium use were exaggerated since:

        It is difficult to make reliable nuclear weapons using reactor-grade plutonium; and

        Once-through fuel cycles are not a panacea since it easy to extract plutonium from power reactor spent fuel in a quick and dirty reprocessing plant.

Both of these contentions were discussed at great length at the time, and the former continues to cause controversy despite the efforts of knowledgeable individuals such as the late Carson Mark, [6] Richard Garwin, [7] and John Kammerdiener [8] to shed light on the subject. Part of the problem is that key aspects of weapons design, including how the obstacles involved in using reactor-grade plutonium, need to and can be overcome, are still classified. Another consideration is that the stakes are high, which tempts some people to pay selective attention to the facts. In lieu of an extended discussion, I quote here the latest unclassified guidance on the subject from the US Department of Energy: [9]

The degree to which these obstacles [i.e., higher probability of predetonation in some designs, as well as increased heating and radiation compared to weapons-grade plutonium] can be overcome depends on the sophistication of the state or group attempting to produce a nuclear weapon. At the lowest level of sophistication, a potential proliferating state or sub national group using designs and technologies no more sophisticated than those used in first-generation nuclear weapons could build a nuclear weapon from reactor-grade plutonium that would have an assured, reliable yield of one or a few kilotons (and a probable yield significantly higher than that. At the other end of the spectrum, advanced nuclear weapon states such as the United States and Russia, using modern designs, could produce weapons from reactor-grade plutonium having reliable explosive yields, weight, and other characteristics generally comparable to those of weapons made from weapons-grade plutoniumProliferating states using designs of intermediate sophistication could produce weapons with assured yields substantially higher than the kiloton-range possible with a simple, first-generation nuclear device.

But even if reactor-grade plutonium can be used in weapons, it might still be possible to make access to it more difficult by technical or institutional means. Examples of the former are modifications of the standard Purex reprocessing flowsheet so that plutonium would not be separated from uranium or only partially decontaminated from fission products. Such schemes were assessed during the Carter administration and found not to offer significant nonproliferation advantages. [10]

The institutional schemes were more ambitious; their essence was to restrict weapons-useable materials and the technologies that produce them to international or multinational energy centers. The centers would take back spent fuel from national reactors and process it, along with inputs of natural uranium (and thorium in some schemes), to produce fresh fuel for both the national reactors and reactors located within the center. Many variants on this theme were studied during the Carter administration; an example using U-Pu and U-Th fueled PWRs is shown in Figure 1.

Unsurprisingly, none of these schemes made it past the paper stage. The technical, economic, and institutional difficulties involved in setting up and operating such centers are considerable. [11] But beyond this is a more fundamental problem: whom would sign up on the national reactor side of the proposed energy divide? Such schemes would be seen by the non-nuclear weapons states as another means by which the weapons states seek to maintain their weapons monopoly. Are countries today more willing than they were in the past when the perceived benefits of nuclear power were much greater - to embrace such discriminatory arrangements? I return to this point below.

In retrospect, the fears of the Carterites that the spread of nuclear power would lead to a nuclear-armed crowd did not materialize. Because of a number of factors, e.g., concerns about reactor safety, especially after the Chernobyl accident, lack of progress in disposing of spent fuel, and the availability of cheap natural gas as feed for efficient, modular gas turbines, nuclear power has hardly spread beyond the countries that had already implemented it when Carter took place. During the Reagan administration the proliferation issue moved to the back burner, and its focus also shifted from the fuel cycle to the efforts of Pakistan to develop a nuclear weapons capability via unsafeguarded dedicated facilities. As a result of the Chernobyl accident, technical innovation in nuclear power for the past 15 years has focused largely on designing safer reactors. However, attempts to increase the proliferation resistance of the fuel cycle never disappeared entirely. For example, the proponents of the Integral Fast Reactor (IFR) claimed that it not only was safer and produced less long-lived waste requiring geologic disposal than a standard fast breeder reactor, but that its integral design and pyroprocessing technology also made it more proliferation-resistant. [12] And the major selling point of the thorium-uranium LWR fuel invented by Alvin Radkowsky is the claim of increased proliferation resistance compared with an LWR fueled with low-enriched uranium. [13] How valid are these claims, and can newer designs do even better?

Will a second nuclear era be more proliferation resistant?

The mood of proponents of nuclear power in the US is decidedly more bullish these days. There is an electricity crisis in California, and nuclear power has powerful friends not only in Congress but also in the new Administration. Within the nuclear R&D community there is much discussion of and some funding for new reactor designs that are advertised as simpler, safer, and also more proliferation resistant. But are such designs technically feasible, economically viable, and politically acceptable? I comment briefly on these dimensions in the following.

Technical feasibility

Some of the new designs involve evolutionary, if still undemonstrated, extensions of current technology; others have important novel features and/or require the integration of several advanced technologies. Examples of the former are the High Temperature Gas Reactor (HTGR), the Radkowsky thorium-uranium fueled LWR, and the IRIS (International Reactor Innovative and Secure) LWR variant of the Secure Transportable Autonomous Reactor (STAR) concept. Examples of the latter are the Encapsulated Nuclear Heat Source (ENHS) variant of the STAR concept and the Accelerator Transmutation of Waste (ATW) concept proposed by the Los Alamos National Laboratory.

A common proliferation-resistant feature of the reactor designs are high burnup, long life cores on the order of 100,000 MWD/MT and ten years, respectively without fuel shuffling or refueling. This reduces the need for frequent core access in off-load designs, and also increases the fraction of the even plutonium isotopes in the discharged fuel. In the Radkowsky design, the high burnup also increases the amount of U-232 produced along with U-233 in the thorium-fueled blanket.

The development of such cores requires considerable R&D. Naval reactor experience is probably relevant, but technology transfer is hampered by classification restrictions. [14] And how much the degraded isotopics contribute to proliferation resistance is debatable. As the previously cited DOE guidance on reactor-grade plutonium implies, there are nuclear weapons designs, which are predetonation proof; i.e., they work with any isotopic mixture of plutonium. While such designs are certainly advanced compared to the Trinity bomb, they were developed many years ago, and many states now have the capability required to implement them. And while the excess heat from Pu-238 in reactor-grade plutonium and the radiation from the U-232 daughters in thorium cycles must also be accounted for in weapons design and production, neither is considered to be significant obstacles. [15]

The other side of the proliferation coin with regard to high burnup is the increased enrichment required at the front end, on the order of 10% for a burnup of 100,000 MWD/MT. As is well known, the use of such material instead of natural uranium as feed for a uranium enrichment plant significantly reduces the separative work required to produce weapons-grade uranium, and hence the size of the plant. Less well known is that U-233 is excellent weapons material, and is easier to enrich to weapons-grade than U-235, particularly using centrifuges.

Economics

In the good (?), old days, the development of civilian nuclear power was heavily subsidized by governments on the basis that it was important for both national prestige and security; the fact that it could also serve as a convenient conduit and cover for a nuclear weapons programs was also widely appreciated. In such circumstances, it was easy to fudge the numbers with regard to the real cost of nuclear electricity. Now, in an era of increasing deregulation, nuclear power must meet the test of the marketplace. An important consideration is whether increasing the proliferation resistance of the fuel cycle will be a significant burden in this regard. Or conversely, whether new technologies designed to make nuclear power more competitive will make it less proliferation-resistant. Several examples come to mind; I begin with the rise and fall of laser isotope separation technology in the US.

In the 1960s, it occurred to scientists in several countries, including the US, France, and Israel, that it might be possible to use the then newly-developed tunable dye laser technology to selectively excite and then separate (enrich) isotopes of interest, e.g., the U-235 in natural uranium. Two means to this end were suggested, one using molecules and the other using atoms as the working material. These processes became known as Molecular Laser Isotope Separation (MLIS) and Atomic Vapor Laser Isotope Separation (AVLIS), respectively. For uranium enrichment, the former used uranium hexafluoride, , and the latter used uranium vapor. In the US, government sponsored programs to develop MLIS and AVLIS for uranium enrichment started in the early 1970s at the Los Alamos and Livermore National Laboratories, respectively. A private consortium, Jersey Nuclear Atomic Isotopes (JNAI), also pursued AVLIS for uranium enrichment independently.

By the mid-1970s, in the expectation of a major expansion of nuclear power in the wake of the 1973 oil crisis, the US Department of Energy (DOE) was supporting R&D of four uranium enrichment processes, MLIS, AVLIS, the Plasma Separation Process (PSP), and the Advanced Gas Centrifuge (AGC), as well as a major upgrading of our large Gaseous Diffusion capacity. However, the expansion did not materialize, and DOE began to winnow the advanced options. In a 1982 process selection, it chose AVLIS over MLIS and PSP for further development, and AVLIS again emerged the winner over AGC in a 1986 process selection.

One consideration in the process selections was proliferation resistance, and, with my MIT colleagues, Manson Benedict and George Rathjens, I was asked to make proliferation assessments of the competing technologies. As might be expected, proponents advertised that their processes would produce Low Enriched Uranium (LEU) at very low cost, but could only be configured to produce Highly Enriched Uranium (HEU) with great difficulty. We concluded that a process which made LEU could also be modified to make HEU with comparable difficulty, but we did not dig deeply enough into the issue of whether it could make LEU as advertised. The wake-up call, at least for some of us, came during the 1986 process selection, which the nuclear pioneer Karl Cohen characterized as a contest between a technology that didnt work (AGC), and one that didnt exist (AVLIS). Amidst all the hoopla, many people, including the undersigned, hadnt appreciated that AVLIS as an integrated system existed mostly on paper; the actual separative work that had been performed was miniscule. However, DOE, and after privatization of the enrichment program, the United States Enrichment Corporation (USEC), continued their its support of AVLIS until they pulled the plug in 1999.

The lesson here is that technology not only has to work on paper or in the lab; it must also be economic on a commercial scale. Similar pitfalls may lie ahead. For example, advocates of nuclear power using once-through fuel cycles point to seawater as a source of sufficient uranium to support thousands of gigawatts for hundreds of years. This on the basis of extraction costs on the order of $200/kg U estimated from recent experiments that have recovered gram quantities of uranium from the sea off the coast of Japan. The argument is that further development will reduce these costs substantially, to say, $100/kg U. Since uranium at current costs of about $25/kg U accounts for approximately 5% of the busbar cost of nuclear electricity, an increase to $100/kg U would only increase the busbar cost by 15%. While this may appear tolerable, the worlds nuclear vendors are unanimously of the opinion that nuclear power costs must be reduced by about 30% to be competitive with combined cycle gas turbines. And of course, projected costs on the order of $100/kg U are far from assured; in fact, the latest estimate is about $1,200/kg U. [16] This shouldnt be taken as a criticism of continued R&D in this area. Quite the reverse: such efforts should be increased to provide more realistic estimates of recovery costs.

In sum, it is unclear whether todays safer, simpler, modular, and more proliferation resistant paper reactors and fuel cycles will turn out to be economically competitive. What does seem clear, at least to me, is that while greater proliferation resistance is desirable, it isnt sufficient. The public is still concerned about safety and waste disposal, and rightly so. Moreover, innovation is not restricted to the nuclear fuel cycle; future reactors will face stiff competition from both non-nuclear electricity sources as well as opportunities for reducing energy demand via greater efficiency in end-use.

Political Acceptability

Last, but certainly not least, is the issue of the political acceptability of changes designed to increase the proliferation resistance of the nuclear fuel cycle. I have been talking mostly about new reactor designs as a means of increasing the intrinsic barriers to proliferation. However, the view that nuclear power generated, e.g., by LEU-fueled once-through cycles, represents an acceptable proliferation risk depends on the viability of both the intrinsic and the extrinsic barriers to proliferation, including safeguards, physical security, and export controls. Thus, going beyond what might be called the once-through cycle standard for proliferation resistance logically involves upgrading extrinsic as well as intrinsic barriers. This is an ongoing process, e.g., the recently negotiated additional protocol to the NPT model safeguards agreement designed to make a states nuclear activities more transparent. Whether such changes are sufficient is another matter. Those who think not but want to retain nuclear power as a global, long-term energy option commonly invoke international or multinational energy centers or parks containing all sensitive nuclear activities as the ultimate solution to the proliferation problem.

I am doubtful. As previously noted, [17] such arrangements would require that countries outside the fence restrict their peaceful nuclear activities. In a world in which the non-nuclear weapons states parties to the NPT are increasingly unhappy with the progress towards nuclear disarmament called for in Article VI of the treaty, such new restrictions would run against the tide, not to speak of Article IV of the treaty. Alternatively, one could imagine a non-discriminatory regime in which all countries would accept international control of their nuclear energy programs, as proposed in the Baruch Plan of 1946, which was based in turn on the findings of the Acheson-Lilienthal Report. But the Baruch Plan also required the destruction of all nuclear weapons; indeed it is hard to imagine a viable nuclear weapons free world without international control of peaceful nuclear activities. This seems as visionary today as it did in 1946, but it is probably the only way, short of the abandonment of nuclear power, to break the power/proliferation linkage.

Figure 1. LWR fuel cycle for international safeguards, national reactors fueled with thorium and denatured uranium.

Source: Report to the American Physical Society by the study group on nuclear fuel cycles and waste management, Reviews of Modern Physics, Volume 50, Number 1, Part 2, January 1978, p. S156.