Evidence was presented which showed that the H emission rate during exposure to an external atomic D source had the same activation energy (0.6 to 0.8eV) as that for diffusion under the same conditions. This result was similar to that in the layer diffusion case. This result, as well as H exchange with the sample during deuteration (and results which indicated a narrow H emission distribution), could be explained in terms of an exchange mechanism in which mobile D exchanged places with bound H; with no significant activation barrier. It was also shown, by an analysis of H evolution curves, that the H chemical potential was pinned as H was lost from the sample. This result was consistent with the negative-U model for H bonding and H clustering.

W.Jackson: Journal of Non-Crystalline Solids, 1993, 164-166[1], 263-8