• No results found


Research reactor programs can be used, overtly or covertly, to assist in the manufacture of nuclear w e a p o n s in several ways. The most direct link is the use of research reactors for plutonium production. Another link is that H E U research reactor fuel can be diverted for w e a p o n s production. Another possibility is that H E U can be extracted from spent research reactor fuel. The operation of research reactors can also provide s o m e justification for the development of enrichment and/or reprocessing facilities, which can facilitate w e a p o n s programs. M o r e generally, research reactor programs involve the development of a nuclear infrastructure which lowers the barriers to w e a p o n s development.

It w o u l d be completely impossible to produce plutonium-239 weapons from s o m e research reactors, such as zero-power test facilities. It would be extremely difficult to use m a n y other research reactors to produce plutonium weapons because of insufficient production volumes or insufficient plutonium purity - for example m a n y of the low-power reactors in developing countries would not be suitable.

N o r are research reactors fuelled with H E U suitable for plutonium production.37 Similarly, the potential to use research reactor programs in support of HEU weapons development should not be overstressed. The fuel stockpiles and

throughput of m a n y HEU-fuelled research reactors are too small to be of m u c h concern. A n d it would be difficult or impossible to use research reactors fuelled with L E U or natural uranium in support of an H E U weapons program (though these reactors are m o r e suitable than HEU-fuelled reactors for plutonium


Despite the above qualifications, research reactors can be and have been used in support of nuclear weapons programs. Even low-power reactors can be of concern.

For example the Iraqi IRT-2000 research reactor, which originally operated at 2 M W but w a s later upgraded to 5 M W , could have produced sufficient plutonium for one w e a p o n over a period of several years. This risk, albeit small, w a s

amplified by the fact that I A E A inspections of the reactor were infrequent because of the low-risk status of the reactor. (Snyder, 1985.)

In general terms, the most useful research reactors for covert weapons programs are m e d i u m to high-power reactors fuelled with natural uranium or very lightly

enriched uranium (thus producing significant quantities of plutonium-239), or m e d i u m to high-power reactors which use significant quantities of H E U fuel (which can be diverted before irradiation, or H E U can be extracted from spent fuel).39 M a n y other research reactors can be used for weapons-related research, or more generally to develop nuclear expertise. It is notable that research reactors operate in roughly twice the n u m b e r of countries as power reactors: in 1996, 59 countries operated research reactors and 32 operated power reactors ( A N S T O , 1996F; 1997).

2Z However, as discussed in section 2.3, it is possible to "blanket" an H E U reactor core with natural uranium and thus generate significant volumes of Pu-239.

2S Definitions vary, but L E U is generally considered to be < 2 0 % uranium-235 by weight, with H E U containing 2 0 + % uranium-235. Some categorisations also include medium-enriched uranium (MEU), with 20-50% uranium-235. A weapon could be produced with M E U , but it would be a cumbersome process and a crude device; H E U is far more suitable.

22 For general discussions on research reactors and nuclear weapons, see Wohlstetter et al., 1979,

Sales of research reactors are an important part of the broader picture of dubious nuclear trade. The major capitalist powers have sold research reactors to a n u m b e r of countries with weapons ambitions: recipient countries include India (sales and support from Canada, the U S , the U K , etc.), Pakistan (US, France), Algeria

(France), South Africa (US), Iran (US), Iraq (France), Israel (France), Argentina (US, West G e r m a n y ) , South Korea (US), Taiwan (Canada, U S ) , and Brazil (US). (There is no evidence of serious pursuit of a nuclear weapons program in Chile, but research reactor sales and support to Chile by the U S , the U K , and Spain could also be questioned.) The Soviet Union has also been involved in s o m e questionable research reactor sales, including sales to Libya, Egypt, and North Korea. China has a long history of dubious nuclear sales; with respect to research reactors, China has sold two small research reactors to Iran, it supplied Algeria with the 10-15 M W Es Salam research reactor, and it is assisting Pakistan with the 50-70 M W research reactor at Khusab. Several other countries have developed reactor industries -such as Argentina and India - with further potential for domestic use of research reactors in support of weapons programs or exports of research reactors to

countries with weapons ambitions. Thus India has used research reactors in support of domestic nuclear weapons development and has reportedly been negotiating the sale of a research reactor to Iran. Most research reactors in

Argentina have been indigenously designed and built, and for s o m e years there were plans to build a 70 M W research reactor in Argentina which would not be subject to any international safeguards and would have enabled plutonium weapons production without violation of I A E A safeguards agreements. While Argentina has struggled to establish a nuclear export industry, s o m e sales have been m a d e including the sale of a small research reactor to Algeria. (Poneman, 1985; Watford, 1993; Spector et al., 1995; Carnesale, 1981; de la Court et al., 1982.) The most direct use of research reactors for nuclear weapons development is

extraction of plutonium-239 from irradiated research reactor fuel. The two most important examples of this are India and Israel. In both cases, research reactors have been used in conjunction with reprocessing facilities to produce substantial volumes of plutonium for nuclear weapons. In Iraq, IAEA-safeguarded reactors have been used to produce small quantities of plutonium, and larger volumes would probably have been produced and separated if not for the bombing of Iraq's research reactors o n four occasions from 1979 to 1991. In Romania, experimental plutonium extraction, using spent research reactor fuel, m a y have taken place in support of the covert weapons program prior to Ceausescu's overthrow in 1989.

Similarly, there m a y have been covert plutonium separation in North Korea, probably involving irradiated fuel from the 5 M W ( e ) "Experimental Power

Reactor" at the Y o n g b y o n site. (It is a m o o t point whether this Experimental Power Reactor counts as a research reactor, a power reactor, or a dedicated plutonium-producing w e a p o n s reactor - suffice it here to note that the term research reactor can be a misnomer.) In Pakistan, one of the two operating research reactors, PARR-I, m a y have been used to produce tritium for advanced nuclear w e a p o n s despite being subject to I A E A safeguards, and the 50-70 M W research reactor under construction at Khusab m a y provide Pakistan with its first supply of unsafeguarded spent fuel. In Algeria, there m a y have been plans to produce plutonium for weapons in the 15 M W Es Salam research reactor,

although that it less likely n o w that the reactor is under I A E A safeguards and Algeria has acceded to the N P T . In Taiwan, it w a s suspected that the Canadian-supplied T R R research reactor w a s being used in conjunction with a small reprocessing plant for weapons development; under pressure from the U S , the reprocessing plant w a s dismantled in 1977 and the T R R reactor w a s shut d o w n in 1987 although there are n o w plans to restart the reactor.40

Tied in with plutonium production is the question of reprocessing facilities for plutonium extraction. The longstanding view that reprocessing is a legitimate part of the nuclear fuel cycle - and perhaps a necessary step in the longer term - has legitimated the establishment of reprocessing facilities in a n u m b e r of countries and has assisted in a n u m b e r of covert weapons programs. The five declared weapons states all have substantial facilities for separation of weapons-grade plutonium. A n u m b e r of other countries - including India, Israel, Iraq, and Pakistan - have sought help from advanced supplier states to develop

reprocessing facilities. North Korea apparently succeeded in constructing a reprocessing facility without foreign assistance. In Argentina and Brazil,

construction of reprocessing facilities w a s suspended. A n u m b e r of other countries have expended s o m e effort towards the establishment of reprocessing facilities, and in s o m e cases, such as Taiwan and South Korea, this m a y have been

associated with w e a p o n s ambitions (notwithstanding the substantial nuclear power programs in those countries). (Camilleri, 1984; Spector et al., 1995.)

In most of the cases listed above, nuclear power programs have provided the major rationale for developing reprocessing facilities, with research reactors being of less importance. That said, reprocessing facilities have certainly been used in several countries in support of covert weapons programs.

4Q The main source for this information on plutonium production is Spector et al., 1995. See also the references listed in the survey of weapons programs in section 2.3. of mis thesis. The comments on

The use of hot cells - lead-shielded radiochemical laboratories with remote handling equipment for examining and processing radioactive materials - is more closely related to research reactors. Hot cells can, if adequately equipped, be used to extract plutonium from spent fuel. The simpler and cheaper the facilities, the lower the volume and the lower the purity (and thus the weapons-useability) of the plutonium. H o t cells are "dual-use" facilities: they can be used for

radioisotope processing, and numerous other non-military purposes, as well as for plutonium separation. Thus for example there has been a dispute as to whether one of the facilities at the North Korean Yongbyon site is a "radiochemical

laboratory" or a plutonium separation facility. Unsafeguarded hot cells, supplied by Italy, have been used in Iraq for plutonium separation. In Argentina, hot cells operated from 1969-1972 and m a y have been used to extract spent fuel. A small volume of plutonium w a s separated from hot cells in Romania. In Brazil,

laboratory-scale reprocessing facilities were completed but are not k n o w n to have operated. A hot cell facility w a s built in Pakistan, with French and Belgian

assistance, and might have been used for plutonium separation even though larger-scale facilities were also built. H a d Algeria's covert weapons program

proceeded, the existing hot cells could have been used for plutonium separation as well as for radioisotope processing. (Spector et al., 1995; Cronin, 1985; Snyder, 1985;

de la Court et al., 1982.)

Civil nuclear programs - involving power and/or research reactors - are also implicated in the development of H E U bombs. There are three methods of using the cover of a civil nuclear program for H E U weapons production. O n e is

diversion of imported H E U . A second possibility is extraction of H E U from spent reactor fuel. A third, less direct connection is that civil nuclear programs provide justification for the development of enrichment facilities. Generally a nuclear p o w e r program is a far m o r e plausible rationale for the pursuit of a domestic enrichment capability than a research reactor program, because of the cost and complexity of enrichment facilities. In other cases enrichment is seen as a m e a n s of adding value to uranium exports. Justifications for the development of

enrichment facilities cannot neatly be separated from each other. In Australia, for example, enrichment research w a s pursued for numerous reasons - adding value to uranium exports, doubts about the ongoing availability of H E U fuel for H I F A R , the possibility of enriched-uranium power reactors being introduced, and the research m a y also have been pursued to keep open the nuclear weapons option (see chapter 3.4).

Historically there has been more interest in, and concern about, the production of plutonium for covert weapons programs; plutonium w a s easier and cheaper to

produce than H E U . H o w e v e r improvements in enrichment technology have altered this balance somewhat. Another consideration is that weapons can be m a d e from a simpler design using H E U .

Most of the countries which have pursued covert weapons programs have built or purchased uranium enrichment facilities or have pursued research into uranium enrichment. (Indeed most of these countries have also put s o m e effort into

plutonium production and separation, thus doubling their options.) In some cases, such as South Africa and Pakistan, a nuclear power program has provided legitimacy for the development of enrichment plants which have been used to produce H E U bombs. In other cases, such as Argentina and Brazil, a nuclear power program has provided legitimacy for the development of enrichment technology but the w o r k has not progressed beyond the research stage, or enrichment facilities exist but have not been used to produce H E U . In a number of other countries, research reactor programs have been implicated in covert H E U weapons

programs. O n e of the strategies pursued in Iraq w a s diversion of imported H E U fuel supplied for research reactors. In 1980, Iraq announced that I A E A inspections would be temporarily suspended because of the circumstances of the Iran-Iraq war, and 26 pounds of H E U were removed from the core of the low-power T a m m u z II reactor and stored in an underground canal. Something similar happened during the 1991 Gulf W a r . So too domestic enrichment work w a s in progress before the bombing and/or dismantling of m u c h of Iraq's nuclear infrastructure in the early

1990s. A centrifuge enrichment program m a y have been pursued in Iran - to the extent that there w a s any attempt to justify this covert research in relation to the civil nuclear program, it could have been justified for production of fuel for research reactors or for the planned nuclear power program. Israel has

concentrated on plutonium weapons production but has m a d e s o m e progress in the development of enrichment technology. To the extent that there is any

pretence that the Israeli nuclear program is a non-military program, it would be possible that enrichment work could be justified for the production of H E U for the IRR-1 HEU-fuelled reactor; alternatively the pretence of a nuclear power program in the longer term m a y be relevant. (Spector et al, 1995; Spence, 1984; Bellany, 1972; Holdren, 1983; Falk, 1983, ch.9.)

There are other cases where imported HEU research reactor fuel has raised concerns even if there has not been any diversion for weapons production so far as is publicly k n o w n . The supply of H E U to Libya by the Soviet Union (and n o w Russia) has been contentious. If there w a s a more serious pursuit of nuclear

weapons in Sweden, diversion of H E U fuel, supplied by the U S for the R 2 research reactor, might have been a feasible option, safeguards notwithstanding. Just before

pulling out of Vietnam, the U S removed 12-13 kg of 2 0 % enriched uranium from the Dalat research reactor, which it had supplied. Supply of H E U research reactor fuel from the U S has been suspended a number of times over the years because of concerns about the potential for diversion: there is no evidence of diversion of H E U in South Africa, but the U S supplied 104 kgs of H E U research reactor fuel before supply w a s cut off in the mid 1970s; supply to Mexico w a s cut off for some months in 1978; supply to Israel w a s suspended in 1981; and supply to Romania w a s cut off from 1989. There are a number of other examples of supply of H E U

research reactor fuel, or H E U targets for radioisotope production, being suspended or refused by the U S . These instances reflect ongoing concern, dating from the 1970s, about the international trade in H E U and the weapons implications of the trade. (Jaster, 1988; Harby, 1988; Spector et al, 1995; de la Court et al., 1982.)

Leaving aside specific examples, it is widely acknowledged that research reactors are important in the H E U economy. The level of uranium enrichment for power

reactors rarely exceeds 3-5% uranium-235, and this fuel is far short of the level of enrichment necessary for weapons production. M a n y research reactors, by

contrast, are fuelled with H E U . In the 1950s and 1960s, low-power research reactors were built around the world using L E U fuel. L E U fuel w a s chosen in part because it is not suitable for weapons manufacture. However L E U fuels gave w a y to H E U , which can be used for longer in the reactor core, and can generate a higher

neutron flux which is preferable for purposes such as fundamental research and materials testing. In addition, the use of H E U fuel facilitates the generation of high neutron fluxes and this facilitates radioisotope production; this w a s another

reason for the use of H E U fuels. H E U became readily available and w a s used not only for high-power research reactors but also for low-power reactors for which L E U would have been sufficient if not ideal. (Muranaka, 1983.) The U S has been the main supplier of H E U , since it had a near monopoly of enrichment facilities for m a n y years, and has exported over 25 000 kg of H E U - most of this (85%) w a s sold to the 12 Euratom countries, but a total of 51 countries have received H E U for use in research reactors from the U S . (Takats et al., 1993.)

The use of HEU fuel in research reactors has provided impetus for the production of and trade in H E U with implications for weapons proliferation. The weapons

implications gave rise to the Reduced Enrichment for Research and Test Reactor (RERTR) Program, a U S initiative which arose out of the 1978 Nuclear N o n -Proliferation Act. The primary aim of the R E R T R program is the conversion of HEU-fuelled reactors to enable use of L E U fuels. The implications of the H E U ->

L E U reactor conversion program for radioisotope production will be taken u p in chapter seven; for the m o m e n t m y concern is with weapons proliferation. A

considerable proportion of HEUfuelled research reactors in the U S , or U S

-supplied research reactors in other countries, have been converted to L E U . S o m e in-principle agreement to support the program has also been given by Russia and China. These signs are promising, but it is likely that the connection between HEU-fuelled research reactors and w e a p o n s proliferation will be an issue for the foreseeable future. A significant n u m b e r of reactors still use H E U fuel. S o m e reactor operators are refusing to convert reactors to L E U fuel - the technical logistics of reactor conversion are neither simple nor fully developed and the effect of conversion o n reactor performance is a contested issue. There is little or no likelihood that w e a p o n s ambitions are influencing decisions in those countries where s o m e reactor operators are refusing conversion (e.g. Belgium, the

Netherlands, South Africa, the U S ) . Nevertheless reactor conversion raises a familiar dilemma facing all attempts to deal with proliferation concerns with technical fixes - countries wishing to acquire a nuclear w e a p o n s capability using materials gained for or from research reactors are precisely those countries the least likely to be interested in the acquisition of or conversion to proliferation-resistant technologies. Another limitation of the reactor conversion program is that conversion to L E U will increase plutonium production.

As well as the potential for research reactors to be used for nuclear weapons production via the plutonium or H E U routes, research reactors can be used for weapons-related research - perhaps the most striking example is the 19 M W

Purnima research reactor in India, which w a s essential for theoretical calculations relating to nuclear explosions and thus played an important role in the Indian nuclear w e a p o n s program including the 1974 test explosion (Reiss, 1988, ch.7).

There w o u l d be m a n y other examples of research being used for weapons-related research. This issue is greatly confused by the overlap between civil and military nuclear technologies - for example materials testing research is often ambiguous in its potential applications.

More generally, research reactor programs, as with nuclear power programs, require the establishment of a nuclear infrastructure, involving various nuclear fuel cycle technologies, technical expertise, the establishment of nuclear trade links, and so on. Development of this infrastructure can facilitate the later pursuit of nuclear w e a p o n s even if the original intention w a s only to pursue non-military nuclear development. Sometimes this has occurred through the intermediary of nuclear power: research reactor programs have been developed as a forerunner and/or an adjunct to nuclear power, and the p o w e r program is then entangled in a covert w e a p o n s program. Thus in a n u m b e r of countries - e.g. South Africa, Pakistan, Argentina, Brazil - p o w e r programs have been used as cover for covert

w e a p o n s development, with little or n o direct involvement of research reactors in the w e a p o n s program, yet in all these countries research reactors played an

important role in the development of the nuclear infrastructure.41

The indirect links between research reactors and covert weapons programs are complex and little is to be gained by commenting and speculating on all the cases where research reactors have, or m a y have been, indirectly involved in covert weapons programs. A n u m b e r of points can be illustrated using the example of South Africa. There is n o evidence that research reactors were used to produce plutonium for w e a p o n s in South Africa, nor that H E U research reactor fuel w a s diverted. Yet the two research reactors, Safari I and II, were probably of indirect value to the w e a p o n s program. They might have been used for weapons-related research - for example neutron generation and control research, or materials testing (Jaster, 1985). The operation of research reactors w a s certainly important in the development of a nuclear infrastructure in South Africa, and that

development w a s certainly important for the w e a p o n s program. The termination of U S supply of H E U research reactor fuel gave s o m e impetus to, and justification for, the pursuit of a domestic enrichment capability, which w a s crucial to the

weapons program. Lastly, as in Argentina, the supply of a research reactor (Safari I) to South Africa facilitated the later indigenous design and construction of a

research reactor (Safari II) which w a s not under international safeguards (Jaster, 1985).

Since much of this thesis is concerned with the medical radioisotope industry, it is worth noting the connections between radioisotope production and covert

weapons programs. O n e of the direct links between radioisotope production and weapons proliferation is plutonium extraction using hot cells. Another close link is that the use of H E U reactor fuel facilitates production of high specific activity radioisotopes42 - this w a s one of the reasons for the historical trend towards using H E U fuel, and it m a y be a reason s o m e reactor operators have refused conversion to L E U fuel m o r e recently (see chapter 7.7).

Another set of links between radioisotope production and covert weapons programs involves enrichment facilities. In Iraq, it w a s discovered in 1991 that large calutrons - electromagnetic isotope separation devices, also k n o w n as high-current m a s s spectrometers - were being used for uranium enrichment.

^ 1 A related issue is whether it makes m o r e sense to pursue a covert weapons program under cover of a nuclear p o w e r program or a research reactor program. This debate is taken u p in The Bulletin of the Atomic Scientists b y Fainberg (1983) and Holdren (1983; 1983B).

^2 Specific activity refers to the ratio of the desired radioisotope to contaminants (which can include isotopes of the desired product).