DIY photolithography

February 24, 2008 11:02 AM

DIY photolithography: has anyone ever tried to etch their own high-resolution PCB layouts or electrode patterns?

I am interested in resolutions down to a few microns, so I don't think inkjet solutions will work. I have access to a spin-coater and should be able to order the required resist and etching solution. I am primarily looking for information on resist selection, mask fabrication, alignment, and exposure.

posted by Krrrlson to science & nature (6 comments total)

If you need res down to a micron, you might need to get friendly with a local printer and have them burn film.
posted by nathan_teske at 11:07 AM on February 24

1 micron with photo-lithography: no way. 10 microns might be doable but seems hard. I know this group, in a field I'm familiar with, made their own small-feature PCBs for Science and Great Justice; a nice email to one of the grad. students may yield some helpful advice.
posted by fatllama at 1:37 PM on February 24

When I said "a few microns" I meant more on the order of 10 microns than 1 micron. Mind you, with the right equipment 1 is more than feasible.
posted by Krrrlson at 4:21 PM on February 24

Generally, people use a contact aligner for what you're talking about. Do you have access to a cleanroom? If not, you're certainly in for a challenge to get 10um resolution. You need good/flat contact between your mask and your sample and a well-characterized UV source. Then, it's a matter of developing a process that works. You might search the literature for resist/exposure/development conditions to get a good starting point. If you have multiple mask layers (I'd guess you don't), alignment will really require an aligner.

Most people order their masks from a foundry after designing them on a CAD program such as LEdit. Making one's own mask with reasonable feature sizes requires one of a few options. I assume you already know this, but even for a bare-bones solution, you're looking at 100s to 1000s of dollars in cost.

An extremely interesting alternative (which MAY work for you) is Shrinky-Dink lithography. Extremely clever researchers at UC Merced used the thermoplastic children's toy to create microfluidic channels. Here is a popular-press article from the American Chemical Society. Email me if you need help sourcing the academic paper, which includes recipes.

Also, for fatllama's benefit- 10um is pretty easy in a contact aligner. 1um is difficult, but possible. With a stepper (projection lithography), submicron features are trivial.

With very advanced optical lithography using all sorts of "tricks," ~50nm is actually possible, but there, you're talking about millions of dollars in investment.
posted by JMOZ at 6:12 AM on February 25 [1 favorite]

Cool, thanks.
posted by fatllama at 10:05 AM on February 25

You can't use a copper layer that's thicker than your intended lateral resolution, or the undercut will destroy it for you regardless of the exposure resolution. Again, there is special equipment to work around it, but nothing that's cheaply available.
posted by springload at 1:55 PM on February 26