A robot that might typically be found in a car assembly line can be modified to operate in a chemistry laboratory alongside humans, reports a paper published this week in Nature. The robot is linked to machine-learning algorithms to enable it to select which experiments to carry out when given hypotheses to test.
Automated chemistry setups are becoming more common in both academic and industrial laboratories. These are being combined with in-line analysis and decision-making to enable a degree of autonomy. However, chemistry robots are custom-made and require specialized interfaces with laboratory equipment and analytical instruments or dedicated instrumentation that only the robot will use.
Andrew Cooper and colleagues describe a modified robot that uses the same standard analytical instruments as would a human chemist, automating the researcher rather than the instruments. The robot uses a combination of laser scanning coupled with touch feedback for positioning, rather than a vision system. Therefore, it can operate in complete darkness, an advantage for carrying out light-sensitive photochemical reactions. The robot has human-like dimensions and can operate in a conventional, unmodified laboratory. Unlike many automated systems that can dispense only liquids, this robot dispenses both solids and liquids with high accuracy and repeatability, broadening its utility in materials research.
The authors programmed the robot to explore various hypotheses to improve the performance of a polymeric photocatalyst. It optimized the reaction conditions within two to three days, instead of the several months that a human would be expected to require. The authors suggest that this robot could be used in conventional laboratories to solve a range of research problems beyond photocatalysis.
Technology: Soft robots catch more flies with magnetsCommunications Materials
Technology: Self-driving cars drive more safely with new algorithmNature Machine Intelligence
Ecology: Fast-growing trees die young and could affect carbon storageNature Communications
Epidemiology: US COVID-19 cases may be substantially underestimatedNature Communications