according to my experience, when use HF solution to etch the SiO2 layer, you
should monitor the process frequently. you want to take out the wafer as soon as
the SiO2 is etched, since this process is istropic, which means the etchant can
attack the SiO2 under your photoresist as well, and that's why the resist tends
to peel off and make some saw-shape edge when u leave the wafer in the solution
for too long. also, thick resist is easier to peel off than thin layer. (some
one also mentioned that 49% HF solution tends to peel off ur resist)
  I feel a little bit strange that you need 30 min to etch 1.5um SiO2, I
remember that it took me about 10 minutes to etch 2um SiO2 when I was using BHF.

  But since you have RIE, why dont you just use RIE for this process? I think
that can make your life easier and safer. :)

  Guocheng Shao

"Xiaoguang \"Leo\" Liu"  wrote:
  Dear all

For one of my projects, I need to pattern thick thermal oxide of
1.5um. I used 49%HF:DI = 1:6 as the etchant and 20um AZ9260 as a mask
(the reason of using this thick resist is that I need to do a through
wafer Deep RIE after patterning the SiO2).

The etching of the oxide takes more than half an hour in the etchant.
My problem is that the photoresist can not stand the HF solution and
start to peel off (especially the smaller features) after about 20min.
HF also attacks the photoresist reducing the overall thickness of the
photoresist, which can impact my later DRIE process.

Could anybody illuminate me with a better method to etch the oxide
layer without destroying the photoresist?