Chemical engineering has been a laborious process: marking up the whiteboard to lay out a hypothesis, then putting on the lab coat, grabbing the test tubes, and hitting the workbench for long sessions of “wet chemistry” to test the idea.
Now artificial intelligence can do most of that wet work instead.
IBM’s new RoboRXN chemistry research software combines artificial intelligence, cloud computing, and automated processing to test new combinations of substances to see how they blend and whether a particular combination can meet a specific goal or need.
A chemist specifies the compound he wants to create and the RoboRXN figures out the ingredients, catalysts, and reactions needed to create it.
The AI program also will predict the result of chemical combinations an engineer proposes.
IBM’s AI-based offering is a cousin of the robot chemist created by scientists at the University of Liverpool. The mobile robotic arm performed 688 physical experiments, working with ten variables, in eight days to identify the most effective catalysts to pull hydrogen from water.
The resulting compounds were found to be six times more effective than those the scientists had thought of.
TRENDPOST: Artificial intelligence will continue to automate processes of discovery, redefining… and perhaps eventually eliminating career fields in engineering, science, and technology.
