Hi junjun wu, I also found that the etch rate decreases as the feasure size
decreases when I did XeF2 etching. However, the etch rate decreases very fast as
the feaure size decreases if the feature size is under certain value. On the
other hand, the etch rate decreases slow as the feaure size decreases if the
feature size is over certain value. I think it is maybe due to that XeF2 is more
easy to reach the silicon surface and SiF4 is more easy to pump out as the
feature size increase.

It's difficult to obtain uniform etch rate on different feature size on the same
surface using XeF2. But you can carefully arrange the etch windows to get a
uniform etching.


2009-02-25

xudehui0108



On 2009-02-24  01:55:31  junjun wu wrote:

Hi Edward, no plasma is used in the XeF2 process. As you and Mehmet pointed
out in his comment, this seems to be a universal effect for dry etching,
which I think might have something to do with how fast the resulting SiF4
can be removed from the surface. I wonder if there are any ways to alleviate
it, such as dilution of the XeF2 vapor, temperature control, or in stead of
static XeF2 soaking using a constant flow over wafer surface ... Thanks.
Junjun