Robot chemist discovers a new catalyst
Researchers at the University of Liverpool have built an intelligent mobile robot scientist that is claimed to be the first of its kind, with the ability to work almost 24 hours a day and to carry out experiments by itself.
Robots have been used before in chemistry research, but they are typically hardwired to a specific experiment. In contrast, the 1.75 m-tall robot developed by the Liverpool team is mobile and can roam around the laboratory, performing a wide range of different tasks. It can also work with equipment designed for human operation thanks to its human-like dimensions and physical reach, relying on an articulating arm to interact with lab equipment and a special gripper that allows it to pick things up and operate machinery.
“What we’ve done, unusually, is to use a mobile system,” said Professor Andrew Cooper, who led the project. “That has advantages because you can really drop it into an existing lab and, without much modification, use equipment which is already there. To some extent, if a human can use the equipment, the robot can use the equipment.”
Unlike a human, however, the 400 kg robot has infinite patience, can think in 10 dimensions and works for 21.5 hours each day, pausing only to recharge its battery. In place of a vision system it uses lasers to map its 3D environment (lidar), coupled with a touch feedback system for positioning.
“The robot taps a cube which is located in a fixed space, and that then provides a six-position reference point,” Prof Cooper explained. So the robot is programmed with the coordinates of the equipment and instruments that it needs to use, and each of these stations is marked with its own cube.
“The biggest challenge was to make the system robust,” said PhD student Dr Benjamin Burger, who built and programmed the robot. “To work autonomously over multiple days, making thousands of delicate manipulations, the failure rate for each task needs to be very low. But once this is done, the robot makes far fewer mistakes than a human operator.”
The robot was designed to help the team in their work developing photocatalysts, which are materials that help produce hydrogen from water using light. Writing in the journal Nature, the researchers recounted how the robot conducted 688 experiments over eight days, working for 172 out of 192 hours.
“The robot weighs out the solid catalyst; it dispenses the liquids; it shines the light onto the samples; it measures the output, which is the hydrogen; and based on this output, the hydrogen emission from the experiment, it then makes a decision about what to do next,” Prof Cooper said.
The robot’s brain used a search algorithm to navigate a 10-dimensional space of more than 98 million candidate experiments, deciding on the best experiment to do next based on the outcomes of the previous ones. By doing this, it autonomously discovered a catalyst that is six times more active, with no additional guidance from the research team.
“If you explore too much then you end up doing 98 million experiments, and that’s just not possible,” Prof Cooper said. “But if you exploit too much, if you optimise too greedily or too quickly, you don’t necessarily find the best solution.
“So the robot uses an optimisation algorithm, using a method called Bayesian optimisation, to navigate that chemical space.”
While the robot won’t be replacing human jobs any time soon, as it still needs scientists to devise the experiments, the researchers say their technology could be used to tackle problems of a scale and complexity that are currently beyond our grasp. For example, autonomous robots could find materials for clean energy production or new drug formulations by searching vast, unexplored chemical spaces.
“Our strategy here was to automate the researcher, rather than the instruments,” Prof Cooper said. “This creates a level of flexibility that will change both the way we work and the problems we can tackle. This is not just another machine in the lab: it’s a new superpowered team member, and it frees up time for the human researchers to think creatively.”
The researchers have also received a lot of interest from laboratories that are looking to recommence operation while still maintaining a level of social distancing. Prof Cooper noted, “If you have a technology where you can load the system up, and then that takes three hours and then it works for eight days or two weeks by itself, then you can see possibilities there. So I’d be delighted if we can find some way to contribute to that.”
The scientists argue for the 'carbohydrate-insulin model', which explains obesity as a...
Researchers reveal more than 50 distinct cancer, immune and connective cell types and states...
A probable cause of Alzheimer's disease is the leakage from blood into the brain of...