Nate:

I have a few suggestions that might help.  You don't mention why the
profile is not meeting your requirements but based on the information
you provided, you are probably seeing an undercut profile.  Also, you
don't provide any parameters for the conditions that you have tried but
presumably you are running at a few hundred millitorr and at or near
room temperature, or higher, for the system that you described.

In order to get a straight vertical profile, with the combination of
SF6/O2, you might consider trying to operate at the lowest pressure and
lowest temperature that you can safely achieve in your system.  The safe
operating ranges are typically provided by the manufacturer.  The use of
backside helium cooling will also help to keep the wafers cooler during
processing and should be set to 5-10 Torr is possible.  If an
electrostatic clamp is used, you might find that you need to optimize
the clamping voltage setpoint to get optimal clamping and cooling.  If
you are processing small pieces of wafers, then you could investigate
bonding of the pieces to a larger substrate using eutectic alloys or
vacuum grease, if allowable in your system.

Once you have determined the operating limits in temperature and
pressure, you might consider starting with an oxygen rich process, say
10:1/O2:SF6 to determine if you can obtain sufficient passivation of
your sidewall to produce a non-undercut profile.  If you can achieve a
non-undercut, angled profile at 10:1, then with further reduction in the
O2/SF6 ratio you should be able to find a mix that produces the vertical
profile angle.  If you do see an undercut at 10:1, then you may want to
try increasing the O2/SF6 ratio further.  As you introduce more O2 to
the mix, your silicon and SiN etch rates will drop but this is a
tradeoff that you may be forced to make.  If you are unable to get a
positive profile, then you could try reducing the bias power level.
This will also result in further reduction in etch rate but may provide
the profile performance that you are targeting.

If you cannot find a set of conditions that produces the non-undercut
profile by reducing the pressure, temperature, and power, then it is
unlikely that this chemistry will work for you.

Regards,
Robert Ditizio


-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of
Nate Lawrence
Sent: Monday, August 23, 2010 7:55 AM
To: [email protected]
Subject: [mems-talk] Anisotropic RIE Etching of Si and SiN

Hi,

I'm trying to find a good RIE recipe for anisotropically etching Si
(crystalline and poly-Si) and SiN while maintaining straight and as
smooth
as possible sidewall profile. I'm using a metal etch mask (have tried
both
Cr and Ni) and would like some selectivity over SiO2, though high
selectivity is not crucial. I would like to be able to etch vertically
as
much as possible, my current goal is 2um. I have a plasma therm RIE
system
with the following gases

Channel1 Ar, CH4, He
Channel2 O2, H2
Channel3 SF6, CBrF3, CHF3, CF4
Channel4 Cl2

All combinations of gases I have tried with O2 seem to give poor
sidewall
profile. My latest attempt was CF4/H2 with the hope of forming some
polymer
passivation on the sidewalls though this the etch rate of crystalline Si
seems too slow in this process and my mask is not holding up to as deep
an
etch as I want (also no selectivity to SiO2).

If anyone has any suggestions I would greatly appreciate it! I have read
that etching with Cl or Br may improve the process and give selectivity
to
SiO2 though I have no experience myself with those gases.

Thanks!

Nate Lawrence
PhD Candidate
Department of Electrical and Computer Engineering,
Boston University