Hi Alex,

I tried the solvent/developer surface treatment thing with S1813 a while
ago with absolutely no luck. Now I use LOR3A and it works well. I
believe the main difference between LORA and LORB is the
dissolution/undercut rate. LORB undercuts faster, so if you have big
features or need a huge overhang, LORB is good. LORA undercuts slower,
so it's better for finer geometries (if something undercuts too much
then the top layer just floats off). But then because it undercuts
slower, you have to trade off possibly overdeveloping your 1813 vs not
getting enough undercut for good lift off. You tune the dissolution rate
with the bake temperature, though, so it's something you can optimize.
It's worked out well for me lifting off 100-150nm thick metal lines 10um
wide. Also, I used 300MIF as my developer, which I believe is a little
stronger than MF319, so you might have to develop a little bit longer
than I did (1 min). A vs B depends on your process, obviously, but I can
attest that A has worked well for me.

-Joe Grogan

Alex Mellnik wrote:
> Hello,
>
> I've used NR9 for a few years now and have had good results with it since
> developing a robust recipe for it.  However, in the past year I've
> noticed
> that NR9 seems to leave some sort of invisible residue behind on many
> materials.  If the area of interest is Au for example, I can generally
> get
> ohmic contact with Ti/Au contacts patterned with NR9, but the second
> layer
> of metal often peels off during wirebonding.  When I try to make ohmic
> contacts to highly doped p-type Ge with Ti/Au patterned with NR9 the
> contacts are only ohmic for large areas, which is not the case when I
> pattern the contacts with S1813.
>
> Normally I would just ash for 1 min in oxygen plasma prior to deposition,
> but here I can't do this because it would either damage the surface of
> the
> Ge or an isolation layer in our device of hard-baked SU8 2000.5 .  I
> spoke
> with the Futurrex engineer about this, but all he could suggest was to
> adjust the bake times, which didn't make any difference.  Does anyone
> have
> any other ideas I could try?
>
> In the meantime we have been using S1813 to do the deposition, but it's a
> hassle to do liftoff since our contacts are usually 2kA or more thick.
> I'm planning to buy a LOR resist to do two-step processing, and I was
> thinking about LOR 3A since I use the metal-ion free developer MF 319.
> However, I noticed that a lot of the people on this site use LOR B with
> that same developer, even though MicroChem seems to not recommend it.  Is
> there any particular reason to use one or the other?  Should I try the
> trick with hardening the top layer with toluene first?
>
> Thanks,
>
> Alex