How and Why We Reinvented the Spin Coater
Posted on Thu, Oct 23, 2014 by Jack Stephenson
Spin coating is probably the most commonly used wet deposition technique for thin films, and we've been using it for as long as we can remember. It's the standard starting point when trying to make highly uniform and very thin layers of everything, from photoresist to nanomaterials. However, the majority of spin coaters have come from the silicon processing industry; where processing is carried out on whole wafers at a time. This means that the vast majority of spin coaters have not been designed with small scale substrates in mind - crucial for most research environments.
Don't get us wrong, many of the spin coaters that we've used in the past have been great and very reliable. But they've often been too large, difficult to maintain and/or plumb in and not designed for the sort of usage that is normal in universities and research labs. There are a few small scale analogue spin coaters available but when precision matters the idea of turning a dial to control the speed seems a little inaccurate.
One of the biggest problems we noticed was caused by the way that substrates are held onto the chuck. In most cases this is done by use of a vacuum. However, this causes multiple problems. In a research and teaching environment we often found that new users inevitably ended up getting solutions and solvents down the vacuum chuck which is lethal for the spin coater bearings; at one point I was stripping down the vacuum bearing in one of the old spin coaters in our glovebox every three months or so. Solvent is also not good for the vacuum pumps, and between the four spin coaters in different labs it always seemed like there was at least one pump that needed servicing.
However, most of all we found that when working on thin glass substrates of 0.7 mm or below (as is often the case for coverslips) we often had a slight warping of the substrate which affected the uniformity of the films, and therefore produced less uniform data. In a discipline where accurate statistics are important this was a problem.
About five years or so ago we therefore tried out a new design for the spin coater chuck. We started to design our own chucks to retro-fit on top of our existing spin coaters. We did this by creating a small recess the size of our standard substrate so that we could just place it in without requiring the vacuum line. I remember saying to my boss at the time - when persuading him to sign the purchase order - that this was a gamble as I had no idea whether it would work, but if it did then it would solve multiple problems in one go.
At that point we had no confidence that the substrate would stay in the recess when being spun, especially at high speed. Although our basic physics told us that resolution of the applied forces (gravity, centripetal, acceleration) should mean that there were no upwards lifts to make the substrate jump out of the recess, we were worried that turbulence could affect it. We designed the chuck such that the substrate would be as flush to the surface as possible but ultimately just hoped for the best. However, we gave it a try and were very surprised. No matter what we did - high speed, low speed, fast ramp, slow ramp, even shaking the spin coater - the substrates didn't budge.
That was five years ago and after those first initial tests using the new chuck design we found that we had better film uniformity and it rapidly became our favourite chuck. After we'd used it for a couple of months without any problems we rolled it out to all four spin coaters. Again, no problems and after several more months we began to be more adventurous.
We wanted to do a scale-up project on 10 cm substrates and needed to spin at 6000 rpm. You would have thought that without a vacuum to hold it in place it wouldn't be possible, but we had no issues at all. Then we wanted to try spinning flexible PET substrates, and thought surely it wouldn't work... but it did. Silicon wafer fragments of arbitrary size and shape? No issue as long as they have one straight edge. The culmination of our spin coating experiments was when we started spin coating onto Parafilm. It started as just a bit of fun, but we found that it really does work! See for yourself in the videos below; you'll be amazed.
After 3 years of using vacuum-free chucks on standard spin coaters, we realised we had simultaneously improved the film quality, increased the versatility of substrates we could spin onto, and reduced the service interval for our vacuum pumps. As such, when we it came to buying another spin coater for a new lab, we looked at what was available on the market and realised we could make one that was simpler, better and cheaper.
This is where our spin coater project began. We spent a year developing the perfect electronics and control systems using a DC brushless servo motor, rare-earth metal magnets and military specification bearings in order to provide a small but compact power unit with smooth acceleration and low vibration.
We used an 8 pole hall sensor in conjunction with new generation of 32bit microcontrollers (ARM Coretex M3) with real-time system clocks to give timing accuracy down to as low as 11 ns. Then we built a rugged control system based on standard proportional-integral-differential (PID) feedback but with corrections for digital noise and hunting. We tested it in our own labs and beta tested with some of our customers. All this resulted in a high accuracy, digitally controlled spin coater that is small and cheap enough for everyone to own one - that's why we call it the personal spin coater.
Put simply, we built the Ossila Spin Coater from a combination of love, fun, experience and requirement. After five years in the making and with many satisfied customers, we couldn't be happier with our creation!
Want to learn more? Watch us spin coat flexible substrates, parafilm and silicon substrate fragments, or view our Spin Coating Guide for technical information on spin coating for nanotechnology.