Xin, regarding Al etching - RIE is really the only way to go.  I have
several years experience with wet etching sputtered aluminum layers of
varying thicknesses, and I can assure you that control of a wet etch
on deep submicron (smaller than 0.5um) features is exceedingly
difficult, and I have found it difficult to control below 0.75um even
with heavy dilution.  For your features, you definitely want a dry
etch chemistry.
--I have found that the following is very good for extremely small
features (I don't know what etch system you have) - the following
proportions of gases:  40:12:5 BCl3:Cl2:CHCL3 (or similar) at the
lowest pressure your RIE will support with approximately 300W RF
power, you can try lower (that was on an older dual-chamber
Plasmatherm RIE system).

PMMA hardening is best dealt with in an oxygen plasma.  The resulting
alumina that forms can be easily removed by a very quick exposure to
that same plasma - in an ideal situation, you would want to switch
your sample under vacuum to an O2 plasma immediately after finishing
the Cl2 etch of the aluminum, and then switch back to the aluminum
etch for about 10 seconds when you were done.

> From: Xin Heng 
> To: General MEMS discussion 
> Date: Sat, 12 Mar 2005 12:33:38 -0800 (PST)
> Subject: [mems-talk] etch sub-micron holes in PMMA patterned Aluminum
> Hi,
>  We want to etch holes of a couple hundred nanometers in diameter in an
> e-beam evaporated Aluminum film. The substrate is clean Glass and the
submicron
> hole pattern is defined by PMMA and e beam lithography.
>  Are there any wet chemstry or dry chemistry ways to leave a good result.
>
>  p.s. we tried RIE but PMMA is hard to be stripped off with Acetone.