Dear X. Yan:

I think you have two problems here. First the control of the thickness,
and the second on stoichometry. During the strike of a  plasma and
initial heating of the chamber walls there will outgassing of both
material from past etches ("seasoning") and also any atmosphere that
leaks in, which includes water. Though you can pump way down, as soon as
heat is applied materials will degass, plus there is the effect of
plasma bombardment. So you will have stoichometric problem. Your
amorphous layer will not be Si, but instead SiOxZy, with Z being other
stuff.

Also, I don't think you can strike a plasma and then shut it down in 1.5
seconds with real process control.

My speculative thought is that I would have a prior step of Argon plasma
for 15 to 20 seconds to degass the chamber. Then I would dilute the
silicon containing gas with argon to reduce considerably the deposition
rate, so that the deposition is 10 or 15 seconds. You could increase the
argon flow a lot, and reduce the silicon containing gas to a low level.
You don't want to go so low that the Mass Flow Controller doesn't really
control. For MFCs you want to be between 15 to 85% ideally, but you can
go down to 10% with okay results. At 5% you will run into risks that you
really don't what you are flowing or you might not get flow. For
prototyping it could work, just realize you might really be flowing 2 or
3% and if they recalibrate the MFCs you will have to reset up your
process.

Lowering the pressure might slow down the process a lot.

However, if you still can't get low enough, you have to go to a diluted
silicon gas bottle and that is a cost problem for a lab, though not a
production facility.

I would consider sputting silicon, but that isn't done very often and I
am not knowledgable and that might be a set of even worse problems.

Best of luck

Ed


-----Original Message-----
From: mems-talk-bounces@memsnet.org
[mailto:mems-talk-bounces@memsnet.org] On Behalf Of memser
Sent: Sunday, November 25, 2007 5:23 AM
To: mems-talk@memsnet.org
Subject: [mems-talk] Precision of film deposition 10 A

Hello all,
     I designed an optical modulator based on a film structure. But the
requirement of precision of film deposition (PECVD, Oxford plasma system
100)  is approximately 10 A.  It means a deposition time of 1.5 second
for PECVD amorphous silicon. I don't know whether this requirment is too
difficult to achieve. Maybe subtle changes in chamber of plasma system
will lead a disabled device, although we measure the film deposition
rate every time before fabriction.

Best regards!

X.Yan
memser@tom.com