Neutronic and nuclear post-test analysis of MEGAPIE

The MEGAwatt PIlot Experiment (MEGAPIE) project was started in 2000 to design, build and operate a liquid metal spallation neutron target at the power level of 1 MW. The project is an important step in the roadmap towards the demonstration of the Accelerator-Driven System (ADS) concept and for high power molten metal targets in general. In an ADS the spallation target is placed inside a sub-critical reactor core. The role of the spallation target is to provide the extra neutrons needed by the sub-critical core to keep the reactor working. Since an ADS is a fast neutron system, there is no moderation and the spallation neutron spectrum is therefore a typical fast spectrum. For a sub-critical core with k{sub eff} = 0.95, a strong neutron source is needed, and in the roadmap an accelerator current higher than 10 mA is indicated as baseline parameter for the experimental ADS. The choice of the accelerator current and energy depends primarily on the number of neutrons that need to be generated, and that are used to drive the reactor. With the 590 MeV cyclotron delivering a continuous beam on target with a current up to 1.8 mA, SINQ was chosen for the MEGAPIE experiment as the most powerful spallation neutron source in the world, with a proton beam power on target that can reach 1 MW. Up to MEGAPIE all SINQ targets were based on a bundle of heavy material rods (full zircaloy, steel rods filled with Pb, zircaloy rods filled with Pb) cooled by a flow of heavy water. For the MEGAPIE target a loop of about 82 litres of lead-bismuth eutectic (LBE) circulates enclosed by a steel structure. The target is about 5 m long and the LBE is made circulating by means of a main electromagnetic pump. The neutronic performance was deduced from flux measurements done at different positions and distances from the spallation target, because the neutron yield (number of neutrons per incoming proton) cannot be directly measured. The presence of the heavy water moderator in the SINQ facility changes the spectrum, from a fast one to a prevalently thermal one, in most of the measurement points (with the exception of measurements performed near the centre of the target). The neutronic performance of a liquid target is compared to the standard solid targets used in SINQ. In the MEGAPIE experiment the neutron flux is measured in the close proximity of the spallation zone by means of innovative micro fission chambers which give a current proportional to the neutron yield. Coupled with very detailed Monte Carlo simulations, these integral measurements provide accurate data on the neutron generation. Spallation residues accumulation or temperature influence the neutron balance and the neutron energy spectrum. Overall, the results obtained with the 3 codes FLUKA 2006.3b, MCNPX 2.5.0 and SNT are consistent. The comparison was performed for the LBE, where the results compare well, and for the structure of the target for which the discrepancies are larger. The reason is related to the different origin of the activation: residual nuclei in LBE are mainly due to spallation reactions, while target structure activation is mainly due to low-energy neutron capture. The latter is sensitive to the simulated thermalization process and to the capture cross sections data used. By comparing measurements and calculations of the neutron flux, differences of 20% were found for thermal fluxes. For epithermal flux the 'background' of neutrons with E 1 MeV) a disagreement of a factor 2-3 (depending on the chamber position) was found. It seems that the calculation of the fission rates is not correct due to the inherent difficulty of reproducing the mixed neutron spectrum, with strong thermal, epithermal and fast components at the detector locations. MEGAPIE has a neutronic performance higher than the solid targets of SINQ. The performance change between the two different solid targets and MEGAPIE has been correctly reproduced. To achieve a good accuracy in the calculation of the neutronic performance of an ADS system, an accurate definition of the geometrical model taking into account the influence of structural materials is of primary importance. The results depend also on the beam profile used in the simulations, at least for the flux calculations close to the target interaction point. Radioactive nuclides produced in liquid metal targets are transported into hot cells, pumps or close to electronics with radiation sensitive components. Besides the considerable amount of decay {gamma} activity in the irradiated liquid metal, a significant amount of the Delayed Neutron (DN) precursor activity accumulates in the target fluid. The transit time of a liquid metal target being as short as a few seconds, DNs may contribute significantly to the activation and dose rates. The importance of the DN issues in liquid metal targets is confirmed. Another problem is the gas production and release in an ADS target, the proton beam generating a large amount of gas by spallation reactions. A large amount of Po isotopes, volatile at relatively high temperatures, are produced in the LBE. The gas production was measured by {gamma} spectroscopy. The release rates of noble gases in MEGAPIE are at the % level after 1-2 days of operation, while the release becomes almost complete weeks after the beginning of operation. Pressure increase in the cover gas could be reproduced with calculations within a factor of 2. The effect of the impurities in the radionuclide inventory of the LBE, using the actual chemical composition of the LBE used in MEGAPIE, is minimal.