help request for image reversal and lift off with AZ
5214
tony
2004-02-19
Zhang:
Just curious: How do you know it is vertical profile? You checked
under SEM? Anyway, I used AZ5214 IR process to get 0.8um lines on MJB3. I
used these conditions:
100C(hot plate) 60s
exp 11mW for 5s
delay 2mins
105C 60s PEB
flood exp 60s at 14mW
dev 12s AZ1:1.
Development is pain in the neck: even add 3 secs more and pattern will
peel off. You have to minimise developement. Room Temperature has to be
around 21C all time, sometimes my lab goes to 29C and it mess up it. Trick
is develope in short time steps, check it under MJB3. If not clear, add
short time again.
I can see double edges under very high mag microscope: I believe it is
undercut.
I used oxygen plasma short time to remove 3000A PR after develope. This
is crucial for metal adhesion.
Hope this can help.
----- Original Message -----
From:
To:
Sent: Wednesday, February 18, 2004 9:01 PM
Subject: [mems-talk] help request for image reversal and lift off with AZ
5214
> Hello, everyone
>
> I am working on AZ 5214 and trying to get reversed image with nice
undercut
> profile. So for, I can only get vertical profile, which caused a big
problem
> for lift-off process.
>
> The substrate I am using is highly doped silicon deposited with SiO2
(200nm).
> The recipe I tried is as follows:
> prebake 90-100 degree 45-60 sec
> image exposure 2-5 sec at 10mW/cm^2 (Karl suss MJB3)
> PEB 115 degree 20-60 sec
> Flood exposure 15-40 sec at 10mW/cm^2
> Develpment 20-30 sec in AZ 400K (1:4)
>
>
> When I tried to increase exposure dose and PEB time, some inversed
> patterns(<10um) were gone and develpment became very diffult.The profile
is
> overcut. When the exposure dose is low and PEB time short, I can get
small
> features in the inversed patterns, but the small gaps (the parts with
> photoresist)between the inversed patterns are develped away easily . At
this
> time, the profile is vertical because of overdevelpment.
>
>
> Does anyone has experience with AZ 5214? Please give me some suggestion
about
> this problem.
>
>
> Thanks,
>
>
>
>
>
>
>
>
>
>
>