Current designs for nuclear fusion reactors call for the use of hydrogen isotopes deuterium and tritium as the fuel for the energy producing fusion reactions. Deuterium is plentiful in nature, but tritium undergoes radioactive decay with a half-life of 12 yrs and does not occur naturally. The use of tritium must be carefully controlled due to its cost and radioactivity and there will be strict limits on the tritium inventory in fusion experiments or in a reactor. These concerns led the IAEA Nuclear Data Section to organize a coordinated research project (CRP) on “Tritium Inventory in Fusion Reactors”. During the years 2002–2006 this CRP brought together specialists in fusion materials and plasmamaterial interaction for exchange of information and coordination of research activities on interaction of tritium with plasma-facing materials. This volume of Atomic and Plasma-Material Interaction Data for Fusion is an output of that CRP. The focal point of the present international fusion energy research programme is the ITER proof-ofprinciple experimental reactor now under construction at Cadarache, France, in collaboration among

This work is focused on the material migration into gaps between tiles and into castellation grooves on plasma-facing components from JET: water-cooled Mk-I divertors and belt limiter blocks. The essential results are summarised by the following: (i) co-deposition occurs up to a few cm deep in the gaps between the Mk-I tiles; (ii) fuel inventory in the the CFC tiles gaps exceeds that on plasma-facing surfaces by up to a factor of 2; (iii) in gaps between the beryllium tiles from the inner divertor corner and in belt limiter the fuel content reaches 30% of that on plasma-facing surfaces, whereas in the grooves of castellation in Be the fuel content is less than 3.0 % of that found on top surface; (iv) fuel inventory in the castellation of the Be divertor and limiter tiles is strongly associated with co-deposition of carbon. Implications of these results for a next-step device are addressed and the transport mechanism into the gaps is briefly discussed. The results presented here suggest that in a machine with non-carbon walls in the main chamber (as foreseen for ITER) the material transport and subsequent fuel inventory in the castellation would be reduced. 1.