Hi Leonardo,

The problem you're facing points to a classic discussion on the CF-based
plasma etching process of silicon/silica.  Without knowing additional
details (substrate, power mode (ICP?)  etc.), I would suggest trying the
following variations of parameters: reduction of O2 (this will enhance
the polymerization/sidewall passivation, hence producing less undercut)
and lower pressure (50 mT seems high if a more vertical etch is
intended).  As a start, I would eliminate the use of O2, and use a lower
(but high enough to sustain a stable plasma) pressure, say, ~10 - 20mT
(one side effect of a reduced pressure is the higher V/P ratio,
therefore a lower selectivity).

Good luck.
Trent

On 2009/12/3 23:05, Leonardo Lesser-Rojas wrote:
> I am using a CHF3/02 RIE process in an Oxford 80Plus etching reactor
> that uses ZEP520 e-beam resist as a mask to etch  shallow features on
> SiO2 (40-50nm deep) and that achieves almost anisotropic profiles (a
> bit barrel like but ok).  However the slight undercut (10/15 nm) will
> make the very thin features on the mask (6-15 nm or so resist stripes
> between exposed areas ) to be removed away during the etching process,
> I guess the undercuts though small, overlap under the stripe.  What
> can be done to improve that?  Power reduction, gas reduction, or
> pressure reduction?  The current conditions are Power=200W, P50 mtorr
> CHF3/O2 50/2 sccm.  I don't have much samples to study the behavior
> and must try first shots...any suggestions are highly appreciated.
>
> Best regards
>
> Leonardo Lesser