Looking into the viability behind using a cheap 3D resin printer (photon s), silicon wafers, and dry film photoresist to make microfluidic devices through maskless photolithography. I've seen a similar process done to make PCBs so I think theoretically there's a good shot of it working. Lets find out.
I currently have a good method to make chips using channels around 400μm wide (shameless plug to my instructable below) so the goal for this project is to fabricate channels that are ~100μm tall and ~100μm deep.
I selected a 40mm silicon wafer to be the substrate for my first attempt as they are the cheapest of my collection and I have a bunch of them. For the photoresist layer I will be using the cheapest roll of dry film photoresist that amazon has to offer. [Note: turns out this was a really bad idea as this roll came with a ton of defects.]
After cutting a square of resist I removed one side of the film (a little piece of tape helps a lot here) and got to work trying to get it to lay flat on the wafer. This process is very similar to putting a screen protector on a phone, and seeing as how I have been doing that since the OG nintendo DS I thought I would have no problems.
This was a faulty assumption.
Laminating the wafer:
It turns out that getting that film down without any bubbles is very difficult. I tried plastic cards, blades, finger tips and vacuum chambers and still could not get all the air cleared out. Part of me wonders if I need to dehydrate the chips first? I'll try that next time for sure.
Once the film was on as flatly as I could get it, I set my trusty cricut heat press to 100c and set it on top of the resist+silicon for 2 minutes. This resulted in what appears to be pretty solid adhesion. I'll have to do some tests to see what affect this on the final outcome once I have the process down. I found it helpful to have a kimwipe underneath and on top of the stack to prevent the film from sticking to the press or the build plate.
A quick word on the cricut easy press:
It's great. I got it for like $50 on sale at a Michaels and its been so useful for so many different projects. You program in the desired temperature, set a time, and you're good to go. It heats up quick and is really comfortable to use. The ceramic heating element is really flat and consistent too.
I fanboy over this thing so hard because I spent literal years looking for somewhat decent hot plates and heating presses and this thing blows them all out of the water. Cannot recommend enough.
Exposing the resist.
Dry film photoresist works by hardening when exposed to UV light. Usually a mask (can be from an inkjet printer, mylar, steel etc.) is placed on top of the resist and is then placed in a box with a UV light source. The UV light cures the photoresist in all the areas that the mask isn't covering which makes it resistant to the developing agent. This method works very well, but masks can be really expensive to produce and it is tough to make revisions once made. Inkjet printers work well enough for larger channel geometries, but we are trying to get below 100 microns wide here so "well enough" won't cut it.
Enter the 3D resin printer.
A resin printer is basically just a vat of goop sitting on top of a small, high resolution monochrome UV LCD screen that displays an image and hardens the resin in that particular spot. Remove the vat of goop and it seems like a very good candidate for maskless photolithography.
I whipped up a quick test pattern in Fusion 360 to try and get a baseline. It was just a simple set of .8 - .1mm x 20 mm channels, and then matrices of pillars of various widths and separation distances. No real hard thought went into this design, I just figured they would be good markers to see what we're working with.
To trick the resin printer into working for this project I took the test pattern and extruded it to be 1mm thick, then set the layer size in the slicer to 1mm. This results in the LCD displaying the exposure image for whatever the single layer time is set to. I started off with a 1 minute exposure time, which proved to not be enough. Eventually I ended up around 3 minutes...
Decided to start this off by just jumping in and seeing how it goes.
My rough plan of action/goal is as follows: Create sub 150μm diameter microfluidic channels using maskless photolithography. To do this I will use dry film photoresist, silicon wafers, an iron, and a 3D resin printer.
My first steps are:
Laminate the silicon wafers with the dry film photoresist (DFR) using a programmable iron
use the high resolution LCD screen from a cheap resin printer (photon S) to project an image onto the DFR, hardening the film in that area.
Use sodium carbonate to wash away any unexposed DFR
Pour polydimethylsiloxane (PDMS) over the silicon wafer
Slap that piece of PDMS onto another slab of PDMS and boom, we've got ourselves a microfluidic chip.
In theory, anyway.
I will be doing this both at MakeIt Labs in Nashua NH and at my apartment. The goal is to find a good recipe that can be followed by anyone with access to a resin printer and allow them to make their own chips.
Throughout this project I'll upload logs as I go and probably photo dump a bunch. Some logs will be lab reports, others will be thoughts and musings on material choice, fluid handling etc. Feel free to comment or PM with any thoughts!