Update on Lam Particle Problem
mcvittie at cis.Stanford.EDU
Mon Mar 19 22:39:39 PST 2001
To follow up on the particle problem, I looked at the open chamber during
the wet clean, I looked at a wafer with particles on it from the chamber
before the clean and I looked at the log book. The particles are coming
for the deposition on the outer ceramic ring on the top of the chamber.
This is a common place for deposition for the HBr process, however we have
not seen flaking like this until very recently. Cesar commented that he
had only seen particles like this in cases after a chamber had been
exposed to air. The particles on the wafer were pretty gross and
consistent with the flaking found in the chamber. I did not see a
particle problem on Friday when I last used the system. Particles were not
reported until Sunday after a series of long runs up to 30 min. It looks
like the particle problem is related to the long etch runs although others
have previously run such long runs with no reported particle problems.
There is another problem that has been occurring lately which may be
related to the current particle problem. For some weeks wafers have been
sticking to the wafer clamp. This has resulted is wafers getting stuck in
the chamber and has resulted in the chamber having to be opened up to air
to get the wafers out. When this has occurred, I believe that no wet
chamber cleaning was done. It is possible, that these air exposures
weakened the bond between the deposition and the chamber surfaces and as
the deposition got thicker the film stress caused the flaking. If this is
the case, the solution is that a wet clean be done every time the chamber
is opened and in-situ plasma cleaning be used to keep the deposition thick
Several users have voiced concern about long etches being done in the Lam.
My answer is that 10 um deep etches are not new for the Lam. However, I am
going to work with the user doing the long etches to make sure the chamber
deposition thickness is kept down.
Finally, I would like to point out that Stanford is somewhat unique in
using nearly pure HBr for etching. HBr has a representation of being a
pretty dirty process which most engineers do not like to run in
production. One solution to our problem would be to try and develop a
Cl2/HBr process which meets our needs.
James P. McVittie Senior Research Scientist
Allen Center for Integrated Systems jmcvittie at stanford.edu
Stanford University Tel: (650) 725-3640
Rm. 336, 330 Serra Mall Fax: (650) 723-4659
Stanford, CA 94305-4075
More information about the lampoly