Uniformity results Oct 28 &29

John Shott shott at snf.stanford.edu
Tue Oct 29 17:17:49 PST 2002

Nancy et al:

Nancy, thanks for collecting and posting this information ... 

Yes, this doesn't look good ... in fact, I would say that it is rather unusual
to see an oxidation with that much variation!

There are two disturbing things (over and above the fact that the average
overall thickness isn't as great as it should be ...):

1. The top to bottom uniformity on each wafer really stinks.  In fact, I don't
think that this is easily explained by cantilever position in the same way
that it would be in a LPCVD situation.  To me, particularly given the fact
that these two runs are just after a tube pull, this is a symptom of a
temperature uniformity problem.  The first thing that I would ask, would be:
was all thermal blocking/glass wool installed properly ... I'm thinking that
maybe we are drawing air out the vestibule that is being drawn along the
length of the tube by leakage near one side of the tube. If cool air is
flowing along the one side of the tube (and not along the other), this could
affect the top to bottom uniformity. Also, are the spike thermocouples
positioned properly and in good thermal contact?

2. We seem to see a pretty significant increase in thickness near the torch
end of the tube.  This would lead me to think that we don't have a set of
calibration numbers that are properly set for the "torch on" condition.  Can
we double check this?

Lastly, I'm a bit surprised that when we saw these problem yesterday ...
particularly a top to bottom uniformity problem ... that a more comprehensive
search wasn't launched at that point.  As I understand it, the conjecture was
that a door interlock problem was at fault ... can anyone explain to me how a
door interlock can affect top to bottom uniformity?  I think that we probably
wasted a day re-running an experiment that was bound to produce bad results
again today ....



More information about the tylan4-pcs mailing list