Some important features of desorption kinetics from MgH2 can be found in the next paper: http://www.sciencedirect.com/science/article/pii/S0925838813007627
Decomposition kinetics of MgH2 have the next characteristics:
1) long incubation (induction) period
2) slow hydrogen diffusion in hydride phase (and very slow rates of associative desorption from hydride surface due to it's non-metallic nature)
3) relatively fast diffusion in metal phase
4) decomposition starts after first metallic nuclei formed on the surface of MgH2
So, MgH2 decomosition kinetics can be described in terms of nucleation-and-growth model.
So, if during the synthesis of MgH2 external surface of MgH2 particles is pure hydride, and don't have areas of metal-phase Mg (or other metallic particles/ impurities) on the surface, then at the same temperature (in isothermal decomposition) or heating rate (in TDS experiments with linear heating) desorption will start later than for MgH2 of same morphology, but with metallic nuclei on the surface.
One of the methods to enhance decomposition kinetics is the way of preliminary formation of metall-phase areas on the surface. It helps to reduce incubation time.
It can be made by "thermoactivation" - preliminary heating until decomposition starts and metal nuclei on the surface are formed, and subsequent fast cooling. Next run with thermoactivated sample will show, that incubation period is drastically reduced.
It can be made also via well-known ball-milling with inetr additives or without any. Durind collisions local heating and desorption occures, and hydride-free areas are formed on the surface. But ball-milling also cahange morphology because of hydride particles crushing. And ball-milling with chemical-active additives can lead to another mechanism of activation - formation of catalytic active particles on the surface of hydride, or formation of new compounds on the surface of hydride particles. So, it seems to me, that thermoactivation is the only method to affect decomposition kinetics of MgH2 without change in morphology or chemical composition. I also made some photoactivation experiments with MgH2 (preirradiation to form metallic nuclei on the surface). This method works well with AlH3 (see my papers for more information). But in case of MgH2 in my experiments photoactivation didn't work. I have some assumptions, why. But it's another question, that require discussion on relations between photochemical reactions and electronic structure of crystalls.