The FRM-II research reactor was designed as a high-performance neutron source. It is optimised as a multi-purpose reactor, primarily for beam tube experiments in fundamental research but also for various technical, medical and industrial applications. The general features of this novel design were established in the 1980s.

The core provides a high thermal neutron flux in a large volume outside the reactor core accessible for experiments at the moderate thermal power of 20MW. The total power was fixed at a low value in order to minimise the radioactive inventory and environmental impact. These requirements have been met by the compact fuel element of FRM-II which uses silicide fuel of relatively high uranium densities (3.0 and 1.5g/cm3) of highly enriched uranium (HEU, 93% U-235).


The compact core of the FRM-II consists of only one cylindrical fuel element (24.3cm outer diameter, 70cm fuel zone height), which contains 113 fuel plates that are cooled by light water and placed in the center of a large heavy water moderator tank. The element contains 8.1kg of HEU (with 93% U-235) in the plates which each include uranium-silicide dispersion fuel (U3Si2/Al) between two Al cladding layers. These fuel plates are 1.36mm thick and welded to the inner and outer core tubes; they have involute geometry to provide cooling channels of constant width (2.2mm). Because of the extremely strong gradient of the thermal neutron flux between the highly absorbing fuel zone and the virtually absorptionless D2O moderator, it is necessary to reduce the power density peaking effects at the outer edge of the fuel zone by grading the fuel density in each plate: the density of uranium is 3.0g/cm3 inside and only 1.5g/cm3 outside of a ‘grading radius’ of 10.56cm.

The unperturbed maximum of the thermal neutron flux in the D2O moderator is as high as 8.0x1014cm-2s-1 at 20MW and for a reactor cycle length of at least 52 full-power days. A horizontal cut-through of the inner part of the moderator tank with the fuel element is shown in Figure 1.

Figure 1. Perspective view from above down the core midplane; one can see the inner part of the moderator tank with the fuel element in the centre. The control rod (not shown) moves inside the fuel element. The five shutdown rods (AS-1 through AS-5) are shown in their shutdown positions. One further recognises some of the 10 horizontal and two inclined beam tubes as well as some vertical irradiation facilities.

As can be seen from the figure, the fuel element is placed in the central core tube which separates the H2O cooling circuit from the D2O moderator. The control rod, which also serves as one of the two fast shutdown systems, moves in the free inner space of the fuel element (not shown). The other of the two fast shutdown systems consists of five shutdown rods which are arranged in the D2O tank as close as possible to the core tube with a safety margin of 1cm. They are shown in Figure 1 in their shutdown position (under operation they are withdrawn). One further recognises some of the 10 horizontal and two inclined beam tubes as well as some vertical irradiation facilities.


The construction of the new neutron source began in 1996 next to the old reactor, sometimes known as the ‘atomic egg’. In spite of a delay of two years to the operating licence due to an extensive review by the Federal Ministry of Environment and Reactor Safety the startup phase began in 2003 and on 2 March 2004 the first criticality was achieved. Full power of 20MW was reached for the first time in August 2004. 52 full power days –– the time projected for one fuel cycle – was reached on 21 October 2004. The first fuel element had demonstrated that its cycle length was equal to the minimum predicted in calculations some 16 years before. The start of routine operation of the FRM-II is scheduled for next month.


At present many of the low- or medium-flux reactors and all of the high-flux reactors in the world use HEU fuel. The International Nuclear Fuel Cycle Evaluation conference of 59 states and six international organisations stated during the years 1978–80 with respect to a conversion from HEU to LEU (low enriched uranium) that: “the agreed criteria are that safety margins and fuel reliability should not be lower than for the current design based on HEU and that neither any loss in reactor performance, eg flux per unit power, nor any increase in operating costs should be more than marginal.” It is widely understood that the term ‘marginal’ means about 5%.

In cases where the discussion of nonproliferation has led to the conversion of a research reactor from HEU to LEU, the following principal has to be pursued: to reach comparable criticality conditions when reducing enrichment by compensation with uranium density the fissible mass must be at least increased by a certain degree. The reason for this overcompensation of enrichment by density is the increased concentration of the isotope U-238, which acts as a ‘neutron poison’. The FRM-II fuel element actually contains 7.5kg U-235 and 0.5kg U-238. Would it contain 20% enriched U at the same U-235 mass, the U-238 mass would be 38kg. similarly, that is why the production of plutonium by the U-238 isotope is drastically lower in the first case.


In the 1980s, enrichment studies for a new FRM were performed by Technische Universität München (TUM). The maximum uranium density with the fuel U3Si2 was assumed to be not higher than 4.8g/cm3. In comparison to an LEU variant, the FRM-II HEU element achieved a maximum flux level that was 35% higher. In 1995-96, nearly one decade after the final TUM design of the element and at beginning of the construction phase for FRM-II, there were presented alternative designs for FRM-II by the Reduced Enrichment for Research and Test Reactors (RERTR) group at Argonne National Laboratory in the USA. It was shown that a conversion to LEU was impossible without severe geometry changes and an increase of the reactor power to 32MW, even with unrealistic uranium metal densities of 19g/cm3.


In 1998 the new ‘red-green’ German government expressed its wish to take the FRM-II into operation with a fuel element of reduced enrichment and so far the scientific aims have been met. Hearings with experts about the feasibility of the conversion have been organised by the Federal Ministry of Science and Education. In October 2001 an ‘Agreement between the Federal Republic of Germany and the Free State of Bavaria on the conversion of FRM-II’ was written down and finally signed after the final licence was granted in May 2003. The compromise found in this agreement is part of the final nuclear licence – development of a new medium-enriched (MEU, with not more than 50% U-235) fuel element before the end of the year 2010. It is understood that the conversion should neither reduce the reactor safety nor degrade the neutron flux and reactor cycle time more than marginally.

In November 2001, TUM established an international working group to study the possibility of a new fuel with then unqualified densities and reduced enrichment. The group consisted of representatives of TUM (FRM-II), the fuel element manufacturer, Cerca, and the constructor of FRM-II, Framatome-ANP (formerly Siemens-KWU). In its kick-off meeting in July 2003, the group declared to accept the technological challenge of the development of a fuel with densities in the order of 8g/cm3 within the very short time limit of 7.5 years.

The project was broken down into three 2½-year phases:

  • Phase 1: The search for the fuel type and test irradiation.
  • Phase 2: Further test irradiations, the final decision on the fuel type and design of the fuel element.
  • Phase 3: Fabrication of the fuel element and licensing.

In Phase 1 the high density fuel shall be chosen or at least preselected. This first 2½-year phase lasts from July 2003 to December 2005 and includes the following tasks:

  • Calculations for the high density fuel(s) for FRM-II geometry.
  • Irradiation of full-size MEU fuel plates.
  • Participation in international research programmes.

Any scenario to convert the FRM-II has to keep the reactor power constant at 20MW. Without changing the whole D2O moderator tank with all its installations and its shutdown rods the core geometry must remain unchanged. This is necessary not only to avoid a complete new licensing procedure but involve also a reactor shutdown period of many years and costs estimated at more than €150 million.

For an enrichment of 50%, an advanced high density fuel is required when maintaining the dimensions of the HEU compact core fuel element. The new uranium-molybdenum dispersion fuel (UMo-Al) presently under research worldwide promises realistic uranium densities of up to 8.0g/cm3. TUM did actual neutronics calculations for this fuel with the same procedures that led to the HEU design of FRM-II (criticality, burn-up, etc). The result was that the density in the fuel must be at least 7.75g/cm3 with an enrichment of 50% U-235. When taking into account other aspects like the less fortunate power density distribution with UMo, the minium density can hardly be below 8.0g/cm3. The core would then contain 10.8kg U-235 instead of 7.5kg in the actual HEU case. The total U mass is 22kg instead of 8kg now. All efforts to flatten the power distribution will reduce criticality.

One key point is the maximum of the fission density (FD) in the fuel. Because of the very heterogeneous power and fission density in the plates, maxima are reached only in very small areas of the plate. This is different from irradiation tests, where the plates are much more homogeneously irradiated. The relevant maximum of the value in FRM-II must be assigned to the inner part of the plate and possibly to the plate ends. The result of these preliminary calculations is that the maximum of fission density in the ‘meat’ is up to 2.0×1021cm-3 in a narrow area at the inner and outer border of the high density region (see Figure 2). Test irradiations show whether the new fuel withstands the mentioned fission densities.

Still the thermal flux is depressed by about 8.0% when compared to the actual HEU fuel, ie the aim of only marginal consequences for the scientific use has not yet been met.


For Phase 1 it was decided to plan an irradiation of four full-size test plates. Six plates had already been produced by CERCA with UMo/Al dispersion fuel (8% wt Mo), ready for irradiation.

The uranium densities were 8.0g/cm3 for four plates and 7.0g/cm3 for the two other plates. Two of the 8.0g/cm3 plates contain the additive of a possible considerable ‘diffusion blocker’. Those blockers are presently the object of research and development worldwide. We used a 2% silicon additive in the Al matrix. Including a safety margin an FD of 2.3×1021cm-3 shall be reached by the test irradiations. According to the CEA’s calculations, this will be reached in a remarkable area of the plates within four cycles of the materials test reactor Osiris at CEA-Saclay.

In case any of the 8.0g/cm3 plates would fail during the irradiation, they could be replaced by the two 7.0g/cm3 plates in the test facility.

Great attention was given to make conditions during the irradiations as close as possible to those in the future FRM-II core. Specifically the maximum temperatures in the ‘meat’ should be comparable. To do so, Osiris needs an extension of its irradiation licence for heat flux values in the order of 280 to 300W/cm2. The decision from the French safety authority is expected in March 2005. The irradiation of the plates will start immediately after the authorisation is granted.

The alternative of monolithic UMo fuel is presently under study at the RERTR programme. One of the main challenges with this very high density fuel is the fabrication of full-size plates. So far only two test irradiations of small mini-plates have been performed.

CEA, CERCA and TUM agree in a ‘memorandum of understanding’ about their cooperation in feasibility studies of the fabrication of and the irradiation behaviour of full-scale monolithic fuel plates.

Author Info:

Based on a paper presented at the ‘Reduced Enrichment for research and Test Reactors (RERTR)’ meeting, held on 7-12 November 2004 in Vienna, Austria. Anton Röhrmoser, Winfried Petry, Klaus Böning, Nico Wieschalla, Zentrale Wissenschaftliche Einrichtung FRM-II, Technische Universität München, D-85747 Garching bei München, Germany