Non-uniformity, of a different kind
Today I looked at a wafer (391) that had been annealed in the RTP on top of a 6″ wafer substrate. The idea is that the six inch substrate would create a “shadow” around the 4″ wafer which would improve thermal uniformity by blocking the multiple reflections from the lamps inside the chamber. Here is what the temperature profile looked like:
One of the parameters we can control is the amount of relative power that is emitted from the front bank of lamps and the back bank. After several attempts at trying to get the two thermocouple readings to match, I settled for a 30% output in the back and %100 in the front. You can see that the steady-state temperature differs by about 10 C. As a comparison, here is the RTP profile for wafer 393 (no 6″ substrate used).
The two readings are almost on top of each other. From my last post, there is a radial symmetry to the pore morphology in wafer 393. Wafer 391, however, seems to have a linear (kind-of) dependence. See pictures below (apologies for the poor quality).
(0,0) – 35kx
(2,0) – 35kx
(4,0) – 35kx
(0,-2) – 35kx
(0,-5) – 35kx
We should keep in mind that I’m only telling part of the story, since I have only imaged two of the four quadrants. I’ll need to look at the other two “arms” of TEM samples to confirm this new non-uniformity.
It will be interesting to see what the other arms look like. If the radial symmetry is really broken, this is an indication that thermal non-uniformity is the cause of our problems, and we can look into ways to fix it.
Annealing on top of a 6″ wafer is very non-ideal, considering how it warps. Seeing any breakage of the symmetry is significant.