Sciencemadness Discussion Board

Patterning Glass for Microfluidics

krfkeith - 30-3-2013 at 22:16

I've been interested in playing around and experimenting with microfluidics as of lately. I still have a lot of questions about that subject, but those I will have to research myself.

After much deliberation, I decided that regular old soda-lime glass is the way to go. In the few references I have found to DIY microfluidics, the preferred material seems to be PDMS. While attractive, it's extremely high price (sometimes as much as $50 for a few grams) make it unsuitable for my purposes. Glass has the advantage of being very cheap and easily to pattern. Well, sort of.

Chemically speaking, etching glass is rather easy. The first choice would seemingly be HF. However, I would prefer to avoid concentrated HF for obvious reasons. There are two main options instead. The first are the multiple varieties of etching cream that are widely available in hobby stores. The second is hardware store rust removal, which contains low quantities (~1-2%) of HF, which should be suitable if I give it enough time. The only problem is that HF tends to leave the surface extremely rough. The most common solution I have seen in the literature is to add relatively small quantities of HCl, which is supposed to remedy the frosted, uneven surface HF normally leaves behind. HCl is obviously easily acquired.

Now here's the problem: how exactly do I selectively etch small channels? I could obviously just draw on some resist. As I understand it, wax is commercially used for artistic glass etching. The problem here is that if I want any assemblage of repeatability and precision, hand drawn patterns will never fly. Unfortunately, I cannot think of any reliable way to deposit wax. I have found a few references to modified inkjet printers which can directly print wax, but that is beyond the scope of this project. My next thought was polyethylene, which is of course extremely inert and resistant to HF and HCl. There again, however lies the problem of how to reliably pattern.

Now, I know there are commercial resists such as SU-8 which probably resist attack, but those are extremely expensive and hard to come by, so I would prefer to avoid them if possible.

By this point I think it is clear that there is just no easily obtainable material with both good resistance to HF and HCl, as well as the ability to be photographically patterned. I say good resistance, because a common them I noticed in a literature review was the problem of pinhole defects, even with commercial resists, which was countered by sputtered layers of chromium as a second layer of protection. I don't think I need to say that that is definitely beyond the reach of an amateur, hobbyist such as myself.

The solution I had in mind was to use a "soft" resist which would be in turn used to pattern a "hard" resist, which would finally be used to pattern the actual glass. Wax or PE would seem to work. One idea I had was to dissolve PE in a commonly available solvent such as xylene, spin-coat (I have a home made spin coater) it onto the glass wafer, and then bake it. Molten wax could possibly be spun coat, but the viscosity might pose a problem. Any suggestions in this area would be greatly appreciated. Anyhow, as for the soft mask, I remembered back when I was dabbling in alternative photographic processes, the so-called gum dichromate process. The process is extremely simple. Basically, a solution of gum arabic and a dichromate (I think potassium is preferred) is used to coat a plate. The plate is then exposed. The dichromate in the exposed sections hardens the gum rendering it insoluble in water. The plate is then washed under water to remove the unexposed gum, and the image is revealed. Interestingly enough, many materials besides gum arabic can be used such as gelatin, agar, and PVA. In any case, the main problem is finding some way to remove the exposed soft mask after it has been used to pattern the hard mask beneath. Furthermore, the solvent used to pattern the hard mask must not affect the soft mask. As far as I know, gum arabic should not be affected by non-polar solvents like xylene, but I am not as sure about the others (my suspicion is that they wouldn't be since they are water soluble). These two issues combined present a problem. As a result, I have two questions for everyone here:

1) What exactly is going on, chemically speaking, with the exposed material (gum, PVA, etc.) that makes it insoluble in water after exposed to light?


2) How can I remove it without damaging the mask beneath it?

[Edited on 31-3-2013 by krfkeith]

[Edited on 31-3-2013 by krfkeith]

[Edited on 31-3-2013 by krfkeith]

watson.fawkes - 31-3-2013 at 06:31

Quote: Originally posted by krfkeith  
how exactly do I selectively etch small channels? I could obviously just draw on some resist. As I understand it, wax is commercially used for artistic glass etching. The problem here is that if I want any assemblage of repeatability and precision, hand drawn patterns will never fly.
Regardless of whatever specific method you choose, I'll suggest that it's going to include silk screening.

You should be able to silk screen molten wax directly. It will take some fiddling to get a process that work and doesn't burn your hands constantly. To clean the screen, you can steam off the wax as they do with lost wax molds in jewelry work.

If you want a two stage process, screen on a thick layer of water soluble paint (maybe even tempera) in the positive. Then spin cast wax on top of everything. Now using a sliding knife edge to scrape off the top layer of wax; you'll need to build a jig for this. Finally dissolve out the paint with water. You'll be left with a wax negative.

itb - 6-4-2013 at 23:03

I think it depends how tiny you want your channels and what your trying to make them do...silkscreening would only be good if your channels aren't very small or for an even surface...you will most certainly need some form of resist which possibly a pcb resist might work I've never tried it nor do i know if hf eats it. Then the next step would be some form of pattern on a transparent surface to shine on your resist to polymerize it. If you want really fine lines you'll need to add some optics most likely plus have to make sure your resist can accomadate the size of lines you want. Don't forget you can always look into synthesizing hf resist if your budget is tight or looking for fun. I've never done this though but looking forward to in the future : )

Also google diy microfluidics other people have interesting ways of getting these things done.

[Edited on 7-4-2013 by itb]

[Edited on 7-4-2013 by itb]

[Edited on 7-4-2013 by itb]

12AX7 - 7-4-2013 at 11:51

Regular PCB etching processes should work. If you do photoresist, you'll have to check if the stuff is affected by HF. I suspect most aren't. Otherwise, typical things include permanent markers and regular printer toner (preferably printed on a plastic transfer sheet, but even regular paper has been known to give passable results). Iron on the toner (go carefully so as not to stress the glass), remove the transfer sheet (peel off, or if paper, moisten and rub off; sulfuric acid may help to decompose the fibers). The glass may need to be roughened to help adhesion, and needless to say, it has to be clean (no fingerprints, etc.).

Tim