--Thank you Jie, but I need undercutting more than 10µm :(

--And thanks again, Ed, I did have an SEM pics of sample no.1, but I could not
see clearly whether the polymer is removed and then the structures got stuck
into the substrate, or the polymer is not removed at all.

I will try using NMP and other metal as well for the devices.
The processes I had are (as u wrote):

1. Spin polymer
2. Deposit SiNx/SiOx
3. Deposit Al
4. Photomask pattern.
5. Etch Al (wet)
6. Etch SiNx/SiOx
7. Strip photoresist (wet first, then plasma)
8. Oxygen plasma undercut

About your old-patent, do you mind if I take a look at that´stupid´process? (no
idea why u called it like that though)

regards, Onny


> From: esebesta@tx.rr.com
> To: mems-talk@memsnet.org
> Date: Fri, 30 Jan 2009 10:24:20 -0600
> Subject: [mems-talk] hard ashing polymer (if Al deposited): possible
isotropic etch by ICP?
>
> Dear Onny,
>
>           The Al is over the SiNx so it can't be leaving a residue on
> the polymer and so it shouldn't have an effect. Though if you want to
> try a different metal, give it a try. It would likely give some insight
> whether the test is successful or not. I patented a stupid process years
> ago with Al over polyimide and got it to life off quite easily in an old
> Tegal etcher.
>
> You say, "But the cantilevers in sample no. 1 seemed stuck to the
> substrate." I would SEM them and see what is actually happening. It
> might be that the Al/SiNx ribbon is flexing downward or falling
> downward. Or doing polymer removal the stip is curling down at the
> edges. (rather fanciful speculation I know.) A SEM could be very
> informative.
>
> Again, try NMP it will clean off easily resist burnt on hot plates,
> labels, attack clean room gloves and most everything.
>
> I am assuming your process is:
>
> 1. Spin polymer
> 2. Deposit SiNx
> 3. Deposit Al
> 4. Photomask pattern.
> 5. Etch Al (Plasma or wet?)
> 6. Etch SiNx
> 7. Strip photoresist? (wet or plasma?)
> 8. Oxygen plasma undercut.
>
> Ed