.
As long as the bond-breaking dominates the rate equation, the reverse rate is
negligible because there is simply not enough free hydrogen. Thus the
degradation within this initial stress phase is only proportional to the stress time
.
Once the interface reaction reaches equilibrium, the previously released
hydrogen species diffuse towards the oxide. Under the assumption of a well
passivated interface with only a few initial dangling bonds 
and atomic
hydrogen 
as diffusion species the diffusion-limited stress regime can be
approximated by a power-law of the form 
. The mathematics behind this
diffusion-limited process can be found in Appendix C.
However, not all measurement results are consistent with the predictions of
the reaction-diffusion theory using atomic hydrogen. Quite to the contrary, the
extracted exponents were found to depend on the measurement delay time; the
exponents of 
were obtained for measurement data with large delay time
only [7, 59]. Shorter delay times on the other hand yielded exponents of around
. Chakravarthi et al. now interpreted this weaker time dependence by
introducing instant dimerization of the released atomic hydrogen at the interface
via 
and subsequent diffusion of the created 
[60].
This theoretical assumption yields a smaller exponent of approximately 
because of the larger kinetic exponent (
for 
instead of 
for
, cf. Appendix C). When performing even faster measurements with
delay times around a micro-second, e.g. OTF-measurements done by
Reisinger et al., exponents of 
were obtained [12] which does not
correspond to the stress behavior predicted by the reaction-diffusion
theory.