AI uncovers 9,389 reactions on copper catalyst for turning CO2 into fuel

Researchers at the Indian Institute of Science used machine learning and quantum simulations to expand a copper-catalyst reaction network from 152 to 9,389 reactions. The larger model predicted about 40 times more CO2 conversion and correctly identified methanol as the main product. The findings were published in Nature Communications.

• Copper-catalyst reaction network expanded from 152 to 9,389 reactions
• Expanded model predicted roughly 40x more CO2 conversion
• Small network wrongly named formic acid as the main product
• Hydrogen can transfer as intact H2 molecules, not only atoms