The presence of a high dose
of aluminium in the sample helped to retain a higher percentage of
bismuth in the sample for the 1000 C anneals. However, despite
this it is clear the vast majority of the implanted bismuth has either
diffused to the surface or segregated out even for very short anneal
times.
The simple explanation for the lack of bismuth related emission
from the sample is therefore due to the difficulty in introducing
sufficient bismuth centres into the silica in a situation where
there is a need for thermal annealing to remove radiation damage.
It has been reported recently [11] that the presence of nanoparticles
prevents the diffusion of bismuth in cooling alloys. We propose
that a similar process may be happening in our samples
where aluminium oxide formed during implantation is preventing
diffusion.