Hi Maryla,

I think you have an issue with stress. LPCVD nitride has tensile
stress, and thermal oxide has compressive stress. Ideally, you'd like
to know the stresses in each film and then adjust the thickness so that
they compensate each other. Since your oxide layer is much, much
thicker than your nitride layer, the resulting membrane has a lot of
compressive stress, so it should at least buckle and possibly rupture
after release.

Randy Grow
Ph.D. Candidate
Dept. of Applied Physics
Stanford University
Stanford, CA 94305-4090

> Message: 4
> Date: Mon, 10 Jan 2005 22:20:36 +1100 (EST)
> From: rmk111@rsphysse.anu.edu.au
> Subject: [mems-talk] Si02/Si3N4 membranes
> To: MEMS-talk@memsnet.org
> Message-ID: <63939.210.9.143.35.1105356036.squirrel@210.9.143.35>
> Content-Type: text/plain;charset=iso-8859-1
>
> Hi all
> I was wondering if someone could help me.
> I am trying to etch SiO2/Si3N4 membranes in 100 Si wafer.  The
> thickness
> of the Si3N4 and thermal oxide layers are  100nm
> and  2.5micron, respectively.
> The Si3N4 layer is removed by the  RIE ,  while  SiO2 in BHF and
> then etched in KOH to produce the window membrane. Unfortunately,  each
> time I repeat my process the membranes brake.
>  Could someone  advise me why this is happening and suggest solution to
> this problem.
> Also, what are  the best conditions for fabrication of the SiO2
> membranes?
> Thank you very much in advance.
>
> Maryla