Searching for new sustainable energy materials will be easier because self-driving labs are working on it. Argonne's Polybot was developed for the same goal.
What Is Argonne's Polybot?
New materials are needed to be used as components for sustainable energy. In particular, materials with high radiation tolerance supporting quantum are preferred for nuclear fusion and quantum computing. Also, the materials should be safe, practical, and sustainable.
Since many things have to be considered, searching for these materials has been very challenging. It involves testing tons of hypothesized materials, IEEE Spectrum reported.
MIT professor of nuclear science Mingda Li said the suitable materials are very few versus the hypothesized materials and can be likened to a drop of water in an ocean.
Fortunately, there are already tools that can help us find those materials fast - like the self-driving labs. These laboratory systems use a combination of robotics and machine learning to run independent experiments. Polybot, a newly-minted self-driving lab at Argonne National Laboratory in Lemont, Illinois, is among these self-driving labs.
Polybot comprises various components, including computers running machine learning software and three robots. The latter have distinct tasks - one is assigned to run chemical reactions, another refines the products or reaction, and the last one is equipped with wheels and a robotic arm to transport samples between stations.
The robots are programmed using Python scripts and perform manual tasks in an experiment like loading samples and collecting data.
The machine learning software will analyze the sample and suggest changes for the next experiment by adjusting temperature, quantity of reagents, and duration of reactions. Polybot could perform all the tasks without human intervention making it a fully self-driving lab with a closed-loop system since June.
Benefits of Self-Driving Labs Like Polybot
Argonne scientist Jie Xu planned Polybot in 2019. She wanted it to be a "universally applicable and reconfigurable resource." They used Plybot to study a plastic's conductivity to create sustainable versions of polymers.
Xu said self-driving labs like Polybot could make the job faster because they can exhaust various ways to synthesize the target electronic polymer by doing half-million experiments in a shorter time. The expert added that human researchers could only generate ten molecules in two years.
According to Xu, self-driving labs can synthesize new materials from two directions. Unlike humans, they use robotics to perform synthesis and analysis, which can run continuously. They also use machine learning to prioritize parameters, making automatic adjustments with better results for the next experiments.
Xu said that prioritization is important, and repeating the same experiment with new parameters - such as temperature and quantity of reagents would be exhausting.
Also, self-driving labs can generate a large amount of experimental data, which is useful for machine learning algorithms. Training a machine learning algorithm must be fed with tons of data, so it will work well and produce better results. For this reason, some labs pool their data with other researchers.
Henry Chan, Xu's colleague at Argonne, added that they wanted to improve Polybot's capabilities, so it could perform more tasks other than optimizing experiments. He wanted it to be used to discover and create new materials like polymers.
RELATED ARTICLE : How Do 7 to 9 Hours of Sleep Help Improve Mental Health
Check out more news and information on Tech & Innovation in Science Times.
