I'm guessing that your nitride was initially not too stressy, either
compressive or tensile, as PECVD nitride can be either.
PECVD silicon nitride is often written as SixNyHz because it can contain
well 10-25% hydrogen.
Upon heating, some of the hydrogen is "outgassed," leaving a more tensile
film behind on one side of the wafer.
The thermal oxide on the other side of the wafer is compressive (at room
temperature).
The two added together produce a curved wafer, which must be really curved
in your case if you see it with the eye.
    --Kirt Williams

----- Original Message -----
From: "Duan" 
To: "General MEMS discussion" 
Sent: Thursday, October 06, 2005 7:50 AM
Subject: [mems-talk] silicon nitride stress problems

> Hello all,
>  I have encountered some problems with the stress in thin films.
> What I have done is that, I first got 1 micron PECVD silicon nitride on a
6
> inch silicon wafer(400 um thick), then I put it in the oxidation furnace
at
> 1100 degree to get 0.4 um oxide film. After this, the wafer was bent very
> much, which was flat before oxidation. I don't know what the reason is. So
> could someone kind enough to tell me why and possiblly tell me the way to
> get rid of that. Any comments or literatures are quite appreciated