This is a pretty wild paper, if you’re like most of us in the human race and not following developments in membrane separation technology. If we think about such things at all, we imagine a sort of high-end filtration device that lets through small stuff and blocks out the larger gunk. That’s accurate as far as it goes – you may recall that there was a difficulty earlier during the pandemic in sourcing some of these membrane filters that are used for purification during the manufacture of vaccines and monoclonal antibodies. These typically pass small oligos and proteins while retaining larger cellular debris from the production process, not to mention whole viruses or bacteria.

For synthetic organic chemists, filtration comes in as a way to remove floating solid particles from solutions, which is an even cruder and more easily visualized use of membranes. We often have fine particulate stuff in our solutions from reagents, catalysts, solid-support carriers, bulk drying agents and the like, and a pass through a membrane filter provides a gratifyingly clear solution and an equally gratifying stain of retained schmutz on the surface of the filter. This is routinely done directly in the sample vial for things that you’re sending into an HPLC/MS machine – they make the vials so that as you close them, your sample is pushed through a tiny disk of filter material. HPLC machines themselves have filters built into them, naturally, but the sort of gorp that bench chemists are willing to put in for analysis will clog those pretty rapidly, annoyingly shutting down the whole system for maintenance. So it’s a lot easier to take care of the problem on an individual sample basis, as those of us who did LC/MS before the filter vials became commonly available well recall. There were always a couple of people in any hallway who tried to skip past the “please filter your samples!” exhortations and who merrily put vials of cloudy brown stuff into the queue. It didn’t take many of those to bring things to a halt for everyone.

The really humungous use (to use the technical term of the art) for membrane filtration is with water. You’ll have heard of desalination plants that process salt water, and many people (if they think about it at all) assume that this must be done by distillation – boiling off the pure water and leaving a salt residue. There are indeed flash-distillation desalination plants, but the leading technology is “reverse osmosis”, which involves forcing the salty water through membrane at high pressure. There are several other membrane-based desalination methods, but RO is the biggest, and it generally uses less energy than distillation. You have to prepare the salt water carefully, because a lot of things can foul the membranes (either quickly or slowly over time), but in most places it’s still the way to go. A lot of money and effort has been put into improving the technology, and that’s still ongoing.

But the paper linked in the first paragraph is the cutting edge of membrane filtration. The authors report a very interesting membrane made from a polyaryltriazole material, highly crosslinked with phenol groups bridging through the polymer. The material is cast into membrane form from a polar solvent like NMP before being thermally crosslinked, and it comes out quite mechanically robust. And it’s full of nonpolar pores that will pass small aliphatic hydrocarbons, but not large ones or polyaromatic ones. In other words, you can put dilute crude oil on one side of the membrane and, by applying pressure, drain off what is more or less gasoline from the other. The pressure is about half what you’d use for desalination, so it’s nothing extreme.

This has a ways to go before commercialization, if indeed it really makes it that far. The paper isn’t just blorping thick dark crude onto a membrane and shoving it through – they’re using very light grades of crude oil and diluting them in solvent for now. But it’s a very promising start. The authors show that different membranes produced by somewhat different levels of crosslinking, etc., show varying degrees of permeability – it might be easiest to use one of the less-stringent ones to get most all the asphaltic material out on the first pass and then send that material through a tighter membrane, for example. Eventually you might set up a whole cascade of these things, allowing you to draw off various fractions as you wish and creating chemical feedstocks with far less effort and energy than we use today. Producing these membranes through polymer chemistry means (as usual) that there are a lot of changes that will produce a lot of different materials, and these folks certainly aren’t the only group working on such ideas. But this is (from what I can see) the most impressive result so far.