“…The diffusion data show that for T = 950°C and 1150°C, this flux is more than large enough to ensure the transportation of vacancies from system 1 to system 2 and thus can ensure condition ii) as stated above. The thermal diffusion to pre-existing pores, which has been proposed as a alternative vacancy sink [16], was shown to be much more difficult since the distance to be travelled by the vacancies was in this case much larger than the raft thickness [44]. As speculated in the literature [16,30], we may assume that the density of sinks, that is the density of system 2, climbing a[100] dislocations, may be the key point.…”