If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Computational Chemistry

Robot runs almost 700 chemistry experiments in 8 days

Proof-of-concept work improves the yield of a hydrogen-making catalyst 6-fold

by Sam Lemonick
July 9, 2020 | A version of this story appeared in Volume 98, Issue 27

Credit: Andrew I. Cooper
A mobile robot worked night and day to optimize a hydrogen-producing reaction in a lightly-modified lab.

Imagine a chemist who spent 90% of their day working in the lab, repeating the same experiments. Benjamin Burger, Andrew I. Cooper, and colleagues at the University of Liverpool built one, a robot that is designed to operate in a mostly standard lab with equipment made for human scientists (Nature 2020, DOI: 10.1038/s41586-020-2442-2).

Industry has used highly customized or fully bespoke robotic systems and academic scientists have described similar automation concepts. Burger and Cooper took a different approach, adding specialized grippers and sample trays to a robot designed to work in warehouses or factories so the robot could operate liquid and solid chemical dispensers, a gas chromatograph, and other lab instruments. Their robot optimized a photocatalytic process that generates hydrogen gas, a potential green fuel, from water. The machine performed 688 experiments over 8 days, adjusting the concentrations of 10 chemicals using artificial intelligence, to arrive at a mixture that produced 6 times more hydrogen than the starting conditions produced.

Optimizing this kind of system “can easily take more than a year of practical, very tedious work,” says Yann Pellegrin of the University of Nantes, who studies hydrogen-generating photocatalysts and was not involved in the research. The Liverpool group estimates their robot is 1,000 times faster than a human experimenter. But Pellegrin also points out that this type of catalytic system is probably not viable because it consumes one of the component chemicals. Burger and his colleagues agree and call this project “a proof of concept.”

Automation experts like Nicola L. B. Pohl of Indiana University say a robot that can work in spaces largely designed for humans could make it easier for some chemists to automate some tasks. Burger and Cooper hope to start selling this system. One near-term improvement is to give the robot more computing power and chemistry knowledge so it can make better decisions as it experiments.



This article has been sent to the following recipient:

Chemistry matters. Join us to get the news you need.