Skip to content
Using AI to Accelerate Autonomous Material Discovery and Optimization

Using AI to Accelerate Autonomous Material Discovery and Optimization

In today's rapidly progressing world, researchers are developing innovative methods to accelerate discovery by merging automated experiments and artificial intelligence with high-performance computing. A standout tool from Oak Ridge National Laboratory showcases precisely how far technologies have reached. It proves that AI can significantly influence materials synthesis and managing associated experiments independently, without any need for human supervision.

This autonomous material synthesis tool employs pulsed laser deposition (PLD) to deposit a thin layer of substance onto a base material. Following the deposition, the AI starts its analysis, examining how the newly created material's quality links to various synthesis conditions. These conditions include temperature, pressure, and the energy emitted during the PLD process. Based on its analysis, the AI suggests a new set of conditions believed to improve the quality of the final product and then autonomously controls the PLD equipment to perform the next experiment.

Sumner Harris, the leader of the study from Oak Ridge National Laboratory's Center for Nanophase Materials Sciences, remarked, "We integrated computer control of all processes into the system and incorporated some hardware innovations to allow AI to drive experimentation." Harris added that such automation speeds up their work tenfold, while AI's capability to understand significant parameter spaces results in needing far fewer samples.

Researchers are optimistic that incorporating automation and AI into this process widely can result in rapid advancements in science. These tools can continue their strides in the world of scientific discovery, making significant strides to move humanity forward in areas like materials sciences, artificial intelligence, and high-performance computing.

Disclaimer: The above article was written with the assistance of AI. The original sources can be found on ScienceDaily.