Dissolved O2 Sensing - omg, it could work for cheap

1 Apr 2011

Hi kids -

Sean here. I'm the dude trying to make a dissolved O2 (DO) sensor from scratch! You may be asking yourself why you care about dissolved oxygen in the first place...the short version is that dissolved oxygen is an important factor to know when determining if a biologic system will live or die. Only a tiny bit of oxygen, and you wind up in a system that you readily see in ponds: most organisms survive based on photosynthesis alone (like algae). A lot of oxygen? You get a system like the ocean, dominated by giant killer fishies and more advanced animals. In fermentation of beer, you need to suppress the amount of oxygen your yeast sees, or else it will make carbon dioxide only, not tasty tasty ethanol. To this hacker...dissolved oxygen is a really interesting parameter.

So, why don't people monitor this all the time? Dissolved oxygen probes are *expensive*. They typically involve platinum catalysts and silver systems (details available here: http://en.wikipedia.org/wiki/Clark_electrode). This, combined with low demand, put commercial DO probes near $200 at a bare minimum...and that doesn't even include a meter to read the signal! If that price could get slashed by an order of magnitude, I have no doubt that all sorts of interesting correlations could be derived by hackers. Is this even possible?

Enter the optode. Instead of using a typical electrode system where oxygen concentration is related to a voltage or current from a chemical reaction, otpodes work by relating oxygen concentration to fluroescence of a fancy pants material (in this case, its a crazy Ruthenium based complex). The main advantage to using an optode is that it removes the need to build a sensitive analog circuit to amplify a tiny current. Also, it reduces the number of expensive materials needed from 2 to one. Handy! The hardest part in building this sensor is trying to determine a way to expose the system to the catalyst without degrading the catalyst or exposing the system to a ruthenium ion (this could allow some unexpeted zany reactions to occur).

I'm a chemistry dude by day, so I had some ideas in the back of my head to solve these problems. Rather than using the bare catalyst, I thought of encapsulating the catalyst inside a set of cheap, clear, food approved polymers that were also oxygen permeable. Mylar is oxygen permeable and food approved, so it seemed to make an ideal outer layer. Cost? $0.85/foot...yes! The downer with mylar is that it doesn't fuse to itself very easily. However, vinyl fuses to other polymers really really well and is also optically clear, and $1.25/yard...double yes!

Before blowing any expensive catalyst, I decided to demo the fusing process using the polymers and a solvent that the Ru complex would be dissolved in. Great successes were had! Here's the lowdown:

Here's a drop of acetone (available at any home improvement store) on top of a ~2"x2" square of mylar. A similar sized piece of vinyl was cut and placed on top. This spreads out the acetone in a beautiful, thin layer (if you've ever put a cover slip on top of a microscope slide, you know exactly what I'm talking about). Next, this assembly was sandwiched between 2 aluminum plates, and heated with a heat gun:

After doing this for about 5 minutes (aluminum plates should be too hot to touch), the assembly was removed from the vice. Don't have some aluminum plates and/or a vice? Unfortunately, you'll need to borrow one for this step. The results below show what happens if you don't happen to use a compression based technique:

Far left: No compression at all. Mylar shrinks and buckles a bunch from the applied heat. Middle: one aluminum plate on top...so minimal pressure. Most of the severe bubbles are gone, but you can still see a lot of bubbles that didn't get their way out of the film layer. Far right: beautiful bilayer that is essentially bubble free, and awaiting some catalyst sandwiching!

So, in short, good news everybody! Looks like the film creation for this sensor will work out well...now to build the sensing array and do all the math...more on that next post!