An article was recently released in Science detailing a system that is capable of testing and optimizing reactions. “Reconfigurable system for automated optimization of diverse chemical reactions” is a system with interchangeable modules that currently can perform a certain task: hot reactor, cold reactor, packed bed reactor, visible light photo-reactor, liquid-liquid separation, and a blank. With five pumps, and the potential to add a liquid at each module, the system offers a lot of flexibility. The system will also automate the experimentation process, you don’t need to know anything about the system or the kinetics or previous successes, the system will simply go through a variety of permutations of experimental parameters and automatically optimize the reactions. The method is called SNOBFIT, which seems to work by having a user defined search area and partitioning it into smaller chunks, and using functions built from previous results to determine what set of parameters to test next. But according to the article it also tests these functions by not only pushing in the direction the functions suggest an improving yield, but also probing complete opposite directions to test the validity of the functions and whether the system is converging on a local maximum or a global maximum. The article then goes on to describe seven reactions which were tested and optimized on the system. The system allowed them find the conditions for very high yields and short reaction times. In the case of a photoredox reaction, the system decreased reaction time from three hours to only three minutes. The overall system is very cool and the researchers have shown it is easily extensible for a variety of chemistries. However, there are some things that seem like limitations. The heated reactors have an upper temperature limit of 120°C and the system does not have any sort of pressurization capabilities. Additionally, it seems the system doesn’t tolerate gases, all of the conditions test were ones where the reactants would be liquid at atmospheric pressure. That can severely hinder the application of this technology to oxidations, where the reactants are either a gas (oxygen or air) or evolve a gas (like peroxides). Regardless, the system seems like a significant advancement in automating experimentation.
In Chemical and Engineering News, “Should plastics be a source of energy?” starts off in New Jersey with a massive white smokestack in Rahway. I’ve actually seen it a few times and wondered what it’s for. Turns out it’s a facility that turns garbage into energy by burning it, reducing the amount of waste that goes into landfills and capturing recyclable metals from entering a landfill in the first place. It continues describing the massive amounts of plastic wastes generated and asks “what can we do with this?” It used to be that China would take plastic waste for cheap, pull out the valuable and easily recycled polyethylene terephthalate (PET), high-density polyethylene (PE), and polypropylene (PP) and dump the rest. The way plastics were sent over were in big shipments of everything. American recycling is inherently poor because we don’t separate things out. It all goes into one bin and it’s shipped over to someone else. If they don’t do a good job, or send it out to someone else to deal it with like China, a lot ends up in landfills or the easily recyclable materials are missed because of contaminants. The article points out how places, especially the West Coast, can’t handle the plastics coming in because their infrastructure was never set up to actually deal with the problem, just push it off to someone else. A deceptively simple solution is to burn it, but plastic is more energy dense than standard trash so the energy quota for the plants are quickly saturated. The article says that another issue is landfills are cheap in the US, so there is not as much of a reason to pursue burning the trash. A lot of criticism comes from idea burning plastic moves the problem from the ground into the air by releasing far more green house gases. It also means new plastic has to be made instead of recycling the material that was sent to burn. Proponents of incineration say that burning plastic is cleaner than coal, and the industry has cleaned up a lot from where it once was.
Another option in the article is to pyrolyze the plastics back to their monomers. Plastics are made of up long chains of repeating units, such as the polypropylene monomer below. Pyrolosis heats the plastics to a high temperature without burning them, forcing the long chains to break apart into smaller pieces. Once small enough to be fluid, these short chains or monomers can be reused to make new plastics. This is especially good as the plastics will essentially be like new, without actually using new materials. Not only that, it makes it possible to reuse plastics where erstwhile could only be sent to a dump. Styrene has been a particular success story, and others are trying to make diesel and naphtha replacements.
The final proposal the article discussed is using plastics to power cement kilns, since they apparently use an absurd amount of energy to reach balmy temperatures of 1400°C. Seems like the first suggested solution but more focused. It’s an interesting article worth the read, and I think I’ll go look for some plastic pyrolosis papers for my next installment.
Jakub's Personal Blog 
