Introduction. The amount of tritium buried in the first wall of a fusion reactor could turn out to be the decisive factor for the choice of wall material. In the case of copper, the common material for beam dumps, we have found a large discrepancy between data published from surface analysis groups and data from neutron generators on the amount of implanted hydrogen [1]. In neutron generators with high particle energies and fluxes, trapping appears to be dominant and implanted hydrogen concentrations of up to 40 at% are reported [2,3]. Well controlled low flux experiments done for surface analysis yield concentrations below 1 at% and the amount of implanted hydrogen is dominated by diffusion and surface recombination [4]. To investigate if similar discrepancies are found in Beryllium we analysed a test section with Nuclear Reaction Analysis after an exposure test with Deuterium beams. In this exposure surface temperature and power density are representative of a first wall component. However the particle energy is higher than for first wall components and the flux is correspondingly lower. The test section is an actively cooled Beryllium Monoblock (Brush Wellman grade S65C) which had been used before for several power handling tests. Part of the surface had been above liquidus in these tests with a melt depth of less than 1 mm. Aim of this experiment was to determine the quantity if implanted deuterium with Nuclear Reaction Analysis. The experiment was done in two steps: 1.The test section was exposed to a power density of 4 - 5 MW/m2 and analysed some days after exposure. 2.The test section was exposed to high power densities of 20 MW/m2 which brings the surface into melting and was analysed after exposure.