Annealing free-standing membranes
I first reported the crystallization of free-standing amorphous silicon membranes in this post in January. The last experiment was done inside the susceptor (top and bottom piece). This week, I annealed a set of amorphous 15 nm and 30 nm with the susceptor top off. This allowed for the direct heating of the films. One set was annealed with the protective oxide still present while the other set did not have the protective oxide. The membranes that had the protective oxide during the anneal was treated with BOE after the RTP and before imaging.
In general, the films that were annealed without the protective oxide did not form any pores. At temperatures > 800 C tears start to form indicative of a large stress during the crystallization process. What is interesting is that in the 30 nm no oxide case, you can actually see an intermediate film thinner than 30 nm. This might be more evidence that the pore formation process starts on one side of the silicon and progresses through the amorphous volume.
Films annealed with the protective oxide show pores at temperatures > 700 C in both the 15 nm and 30 nm thick samples. There does not appear to be any pattern in pore size as the temperature is increased. The crystal grain size does increase with higher annealing temps. This is especially apparent in the 30 nm film.
From these results, we can conclude that the strain caused by the bulk silicon layer is a necessary for pore formation. By annealing with the protective oxide, a small amount of strain is induced in the silicon layer which is why there are a few sparse pores in these samples.