Robot scientists have solved the biggest challenge in chemistry
Prepare to meet your new lab mate.
Coronavirus has highlighted a fact well-known in the scientific community, but not often recognized by the public at large: the scientific process itself is often time-consuming and tedious.
Speeding up the scientific process without sacrificing accuracy has been a dream scientists have chased for years. Now, researchers have designed a new autonomous robot chemist that can complete tasks up to 1,000 times faster than human scientists while simultaneously freeing them up to make important new discoveries.
What is it?-- This robot chemist may not have the same number of Ph.D.'s under its belt as its fellow scientists, but in many ways, it is designed to be just as human -- minus a few key cosmetic features.
Using laser scanning and touch feedback to navigate its surroundings instead of vision, this robot has no need for a human-like face. Likewise, instead of manipulating vials with precarious human arms or moving around the lab on unstable human-like legs, this robot can autonomously wheel itself around a lab and complete science experiments using an incredibly sensitive and precise single arm.
And with human-like proportions, this robot is designed to fit into pre-existing lab space and jump right into scientific exploration without needing specialized modifications. With only half-a-day of human set-up, this robot can work autonomously on science experiments for over 20-hours straight.
No Ph. D needed for this robot chemist to start conducting important research in the lab along side human scientists.
Who made this? -- Introduced in a study published Wednesday in the journal Nature, this bot is designed by researchers from the University of Liverpool in the U.K. Lead researcher on the study and director of the university's Material Innovation Factory, Andrew Cooper, tells Inverse that this new robotic companion to chemistry labs stands apart from other types of scientific robotics in one crucial way.
"There are quite a few instruments in chemistry that people refer to as 'robots.' These are not new things," says Cooper. "But almost all automated systems up until now are built to do a particular thing; they're basically hardwired. This is a different idea [because] we've automated the researcher, [meaning] we built a robot that uses instruments like a human."
Because this robot is not restricted to a single task, Cooper says it can actually be much more useful and flexible in a lab setting.
What did it do? -- In this study, the robot was tasked with helping scientists discover a new photocatalyst -- a chemical that when exposed to sunlight would separate into hydrogen and water. These catalysts are incredibly important for clean energy production because they allow hydrogen to be collected without fossil fuel use. But finding an optimized catalyst is like looking for a needle in a haystack.
Because the experiment space of this problem is so vast and would take human researchers months to explore, the scientists used their robot to do the same work in a little over a week. With parameters (such as what chemicals would be used and what questions would be asked) set in place by human researchers, the robot autonomously chose between 98 million different possible experiments and autonomously conducted them in the lab to discover the optimized chemical substance for this catalyst. And, after completing initial experiments, the robot was also able to refine its discovery using five hypotheses to narrow in even more on the most optimized formulae.
At the end of the process, the robot had discovered a new catalyst that was six-times more reactive than those previously discovered.
With human-like dimensions, this robot chemist can autonomously maneuver chemistry labs with ease.
What this means for robots replacing scientists -- As for whether or not robot scientists will be replacing human scientists anytime soon, Cooper says that these robots are designed to be helpful, collaborative partners.
"We don't see it as replacing jobs in research," says Cooper. "We're not replacing the scientist, it's something to assist the scientist... That will be more powerful than trying to do everything with a robot."
What's next -- In order to better facilitate collaboration and communication between these robots and their human counterparts, Cooper says they're working now to include features like voice recognition and increased intelligence in the robots. This would help them work better with scientists in their own labs while also even enabling cross-ocean collaboration with robot scientists in labs around the world. Cooper says to expect these advances as early as in the next 18-months.
And, while not anticipated when this research originally started, Cooper says that these robots can help scientists continue social distancing in the coming months without significant loss of research efforts.
"If you can set-up an experiment on Monday morning and then you can stay out of the building for the rest of the week -- and still make progress -- that's enormously powerful," says Cooper. "I think this idea looks even better than it did before the pandemic."
Abstract: Technologies such as batteries, biomaterials and heterogeneous catalysts have functions that are defined by mixtures of molecular and mesoscale components. As yet, this multi-length-scale complexity cannot be fully captured by atomistic simulations, and the design of such materials from first principles is still rare. Likewise, experimental complexity scales exponentially with the number of variables, restricting most searches to narrow areas of materials space. Robots can assist in experimental searches but their widespread adoption in materials research is challenging because of the diversity of sample types, operations, instruments and measurements required. Here we use a mobile robot to search for improved photocatalysts for hydrogen production from water. The robot operated autonomously over eight days, performing 688 experiments within a ten-variable experimental space, driven by a batched Bayesian search algorithm. This autonomous search identified photocatalyst mixtures that were six times more active than the initial formulations, selecting beneficial components and deselecting negative ones. Our strategy uses a dexterous free-roaming robot, automating the researcher rather than the instruments. This modular approach could be deployed in conventional laboratories for a range of research problems beyond photocatalysis.
