by Wilfried Krull

Research reactors under threat

29 September 2000

The programme to convert research reactors to use low-enriched uranium fuel has ground to a halt; what is more, long term disposal of the fuel may pose problems that will shut some reactors for good.

In the years between 1978 and 1980, huge international efforts were made under the auspices of the United Nations to evaluate the proliferation risks posed by the nuclear fuel cycle. Measures that would reduce the proliferation risk were considered and enumerated in a series of reports. The International Fuel Cycle Evaluation conference concluded – with the agreement of all participants – that “INFCE has strengthened the view that effective measures can and should be taken to minimise the danger of the proliferation of nuclear weapons”.

This conclusion is the source of the programme on Reduced Enrichment of Research and Test Reactors (RERTR), by which research reactors were to be converted to use low enriched uranium fuel.

In its original form, INFCE was very clear that all existing research reactors in all countries were part of the core conversion programme, even including, at the upper end, the USA’s 250MW ATR.

“The term ‘research reactor’ is used here for thermal neutron reactors which are designed, built and used as neutron ... sources for fundamental research, material irradiations, isotope production, fuel element and reactor safety tests, training etc. Over 150 research reactors of significant power (between 10kW and 250MW) are in operation with highly enriched uranium in more than 35 countries.

The INFCE called for reduction to 20% enrichment, while acknowledging that at first some intermediate enrichment (eg to 45%) might be necessary:

“Enrichment reduction preferably to 20% or less which is internationally recognised to be a fully adequate isotope barrier to weapons usability of U-235.

“For some (currently operating) research reactors it may be desirable to use an intermediate enrichment fuel, at least until new fuels with higher uranium concentration have been successfully developed and demonstrated. For these reactors it may be advantageous to adopt a standard value of 45% for the intermediate enrichment range.

“Therefore, although it may not be technically possible in some research reactors, decreasing the enrichment from the 90% range as far as reasonable toward 20% would be a worthwhile improvement in proliferation resistance of research reactor fuels.”

By 1984, a US Federal Register Note on limiting the use of HEU in US reactors had stated: “The applicant or licensee must use HEU fuel of enrichment as close to 20% as is available and acceptable to the Commission (NRC)”. This was repeated in a Federal Regulatory Note in 1988. But these Federal Regulatory Notes have never been applied to a reactor in the USA, even though they could have been applied since 1985. Why?


U3Si2 fuel has been qualified with a density of up to 4.8gU/cm3 since 1985. This means that up to 95% of all research reactors should be able to convert to <20% enrichment, and the remainder could operate at 30% enrichment. Some 20 years after the INCFE demands were made, all the countries that are still operating reactors above these levels are violating the INCFE conclusions; in the USA they are flouting their own regulations.

The worst offenders on this list are the USA and Russia. But others are with them: Argentina, Australia, Belgium, Chile, China, Czech Republic, France, Germany, Greece, Hungary, India, Israel, Japan, Kazakhstan, Korea DPR, Libya, Mexico, Netherlands, Poland, Romania, Russia, South Africa, Turkey, Ukraine, USA, Uzbekistan, Yugoslavia, and the Euratom community. There are no signs that this is likely to change.

The INFCE communique recognized that there are specific applications requiring high flux reactor operation that, at that time, could only be met with high enrichment fuel. This statement has been used to justify the continued use of HEU – for example by the designers of FRM-II. On the other hand, the statement demanded at least the conversion to intermediate enrichment.

The possible use of HEU at that time for a few research reactors must be seen in the light of the estimated potential for the development of high density fuel at that time. The potential was believed, on a long term development basis, to be only around 3gU/cm3. But fuels of 4.8gU/cm3 were tested in full core tests within five years after INFCE: any reactor that has started up since 1985 has had the option of using high density fuel.

In assessing the practical feasibility of utilising lower enriched fuel in existing research reactors, the criteria agreed by INFCE are: safety margins and fuel reliability should not be lower than the current design based HEU; and that loss in reactor performance (flux per kW) or increases in operating costs should be marginal. What is marginal? Unfortunately this has not been fixed. Assessments vary from 0% through 5% (FRM-II), and to 20% or more (ANL). Many operators converting their reactors have accepted a figure of 15%.

Because U3Si2 fuel of up to 4.8gU/cm3 has been developed and qualified since 1985 up to 95% of research reactors (even in Russia and countries supplied with Russian fuel) can be converted. Other reactors said to have unique purposes can use enrichment of 30% or less. This has been demonstrated even for the FRM-II with its advanced design and extremely small core.

  Fuel development for higher density fuels to meet the conversion demands was of high priority. UAlx and U3O8 have been in use at lower densities, but 3.08gU/cm3 was achieved for U3O8 in 1983 and is achievable for UAlx today. High density UO2 fuel was considered (mainly Caramel), and other forms, such as U3Si-Al, were believed to be of interest, although without giving an estimate of achievable densities. But there was a clear distinction between standard fuels (UAlx, U3O8), higher density fuel (U02) and advanced fuels (U3Si, U3Si2).

“Research reactor fuel types with uranium densities in the meat of 1.4gU/cm3 are currently commercially available. About 2g/cm3 may be possible after development and testing and about 3gU/cm3 may be obtained on a longer time scale.”

This definition in this context is very important as the INFCE document says:

“In the longer term, it was estimated that, if advanced fuel types are developed and demonstrated, 90% or more of all research reactors would have the potential for conversion to less than 20% enrichment.”

It was believed that with the progress of development and using 3gU/cc some 90% of research reactors would be convertable to 20% enrichment. But 3gU/cc has been qualified since 1983, so where are the 90% converted research reactors now?

The INFCE document also says:

“In some cases the use of an intermediate enrichment value as a temporary stage on the way to the 20% limit may increase the effort needed in implementing new fuel types.”

The historical development of the RERTR programme clearly demonstrates that this sentence contains the kernel of the truth, and gives the reason why RERTR is on the way to becoming a flop.

Development has been fast:

•1982/1983: UAlx qualified up to 2.2gU/cc.

•1983/1984:.U3O8 qualified up to 3.1gU/cc.

•1986:U3Si2 qualified up to 4.8g U/cc.

  In 1986 U3SI2 was successfully demonstrated in a whole core conversion and prototype fuel elements have been successfully irradiated in many research reactors.

With higher achievable densities, and accepting the “marginal” penalties (of 15%) all research reactors are convertable and should be converted in agreement with the INCFE conclusions.

France has begun to develop higher density U-Mo fuel (up to 9gU/cc), including reprocessing investigations. That programme is in a far better shape than US dvelopment. The French activities are being driven by the need to have fuel for the 100MW JHR (criticality planned for 2010).

This new U-Mo fuel still has to be qualified. It will convince the international research reactor community when:

•A report from a high level member of NRC is presented at an international meeting describing how U-Mo-fuel can be qualified for US research reactors.

•The US RERTR programme fixes a US research reactor to be converted to U-Mo.

•The French promoters of the U-Mo-fuel promise to convert Osiris immediately after the fuel has been qualified.


There is an important issue of discrimination against reactors that have been converted.

For example, the 5MW FRG-1 reactor was converted in 1991 to LEU. Conversion normally involves thermal neutron flux reduction of around 15% (for U-Mo fuel this reduction will be around 20%). The cost of operating this reactor is around $14 million for a year of 250 full power days. Neglecting all other costs, FRG-1 has effectively been spending $2.1 million per year since 1991 on non-proliferation. Operators whose reactors have not been converted reactors are enjoying a competitive and financial advantage, and non-conversion – especially of US reactors – is therefore an important discrimination issue.

Compare the investment made at FRG-1 with other programmes to reduce HEU. Since 1982 development of fuel above 4.8gU/cc has been promised. No progress was made, however, and a renewed promise was made in August 1996 when the US DOE announced a five-year programme totalling $20 million. By the end of 2000 only 60% of the money will have been spent and so far no schedule for development has been presented.

The US DOE has spent spome $16 million over six years on dveloping U-Mo fuel. When this is compared with the penalties a reactor operator has to pay when it converts to LEU it demonstrates clearly that the UMo programme is a peanut programme.


The main parts of the fuel cycle are cost for uranium fabrication, shipment of fresh fuel, shipment of spent fuel and reprocessing or direct disposal. In the INFCE communique it was assumed that spent fuel would be reprocessed, as had been done in the past by the US DOE and others:

“No special reprocessing problems are anticipated for high-concentration current-type elements. However, for very-high-concentration fuels of new types, some development effort may be needed if these fuels cannot be treated with conventional reprocessing methods.”

Within the RERTR programme reprocessing was a key issue from the early beginnings in 1980. Reprocessing of UAlx, U3O8 and UO2 fuel was well established, even for very high burnup fuel. Reprocessing of UxSiy fuel had not been demonstrated, but following an RERTR demand the practicability of reprocessing was proven in the US and published in 1983.

In recent years the situation has changed:

•The AEA plant at Dounreay has closed.

•At Cogema’s Cap la Hague plant only UAlx fuel can be reprocessed. Products including waste have to be taken back.

•The import of spent fuel to Mayak in Russia has ben prohibited by law since 19 December 1991.

•The Savannah River and Idaho National Engineering sites in the USA may take back spent fuel of US origin under special conditions until 13 May 2006.

After 2006, when the US ‘take back’ programme ends, the situation will be:

•No reprocessing or take back programme for Triga fuel and silicide fuel if the US is not willing to prolong the acceptance of US origin fuel. Triga and silicide fuel is clearly a dead end of development.

•Reprocessing of UAlx is only possible at Cap la Hague.

•Reprocessing of U-Mo may be possible at Cap la Hague.

•There will be no reprocessing options for operators who cannot take back the resulting high level waste.

On the other hand, direct disposal in repositories for HEU and LEU (20 %) fuel may not be acceptable. The US is being convinced that HEU and even LEU research reactor fuel may not be directly finally stored in a repositoryfor criticality reasons – it may be necessary for spent fuel to be reprocessed or “melted and diluted” for safety reasons (see DOE/EIS-0279 D, below).

Taking this into account, after 2006 countries with US-origin fuel and no ‘take back’ possibility must terminate the operation of their research reactors after 2006.

Silicide fuel and Triga fuel is a dead end of development and should not be used after 2006. Countries that can take back waste will probably have two options:

•For LEU, mothballing reactors and waiting for U-Mo fuel to be qualified which can be reprocessed at Cap la Hague (optimistically at the end of this decade) and accepting additional penalties for that fuel. This is the LEU version.

•For HEU, using UAlx fuel

A long-term solution for the back end is key to the RERTR programme.


The INFCE conclusions do not refer to new reactors, but conclusions may be drawn.

At a minimum the same demands must be made as for research reactors current in 1980 when the RERTR programme began – incurring the same penalties (eg 15-20 % reduction in performance or neutron flux) if necessary.

It is clear that the use of HEU violates INFCE conclusions. Enrichment of 45% was tolerable in the 1980s when UAlx or U3O8 fuel had to be used. Now that U3Si2 fuel is qualified up to 4.8gU/cc only 20% enrichment is in agreement with INFCE.

Germany and the US are violating the conclusions of INFCE in their designs for the FRM-II and the ANS, respectively.

As the US has halted the ANS project, the construction of FRM-II (20MW, 93% enrichment, assumed criticality 2001) to is the only new reactor violating the RERTR conclusions, despite the fact that the reactor could be redesigned with an LEU core that would provide the same neutron flux for the users. Its 40-year contract with Russia, which will supply HEU for the reactor, means that Russia is also in violation of the agreement.


After 20 years of RERTR fuel development, though underfunded, has had good results. Although new fuels are available, the rate of conversion to LEU has been slow. The result is that operators who have acted on RERTR are discriminated against: it is time for penalties to be placed on those not complying.

After 2006, the most pressing problem will be the lack of disposal facilities for spent fuel. At that time, unless France’s U-Mo programme – for fuel and reprocessing – is qualified, reactors will have to shut down. Otherwise the operators will be sitting forever on their spent fuel.

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.