Artificial intelligence helps in the discovery of new materials

With the help of artificial intelligence, MARVEL chemists from the group of Prof. Anatole von Lilienfeld at the University of Basel have computed the characteristics of about two million crystals made up of four chemical elements. The researchers were able to identify 90 previously unknown thermodynamically stable crystals that can be regarded as new materials. 

Thanks to modern artificial intelligence, Felix Faber, a doctoral student in Prof. Anatole von Lilienfeld’s group, has succeeded in solving this material design problem. First, using quantum mechanics, he generated predictions for thousands of elpasolite crystals with randomly determined chemical compositions. He then used the results to train statistical machine learning models (ML models). The improved algorithmic strategy achieved a predictive accuracy equivalent to that of standard quantum mechanical approaches.

The matrix depicts the formation energy – an indicator of stability – of around two million possible compounds. Each pixel corresponds to one of the two million quaternary crystals. Depending on the combination of elements, they display either a high (red) or low (blue) energy value. One element is specified vertically and one horizontally; each box contains a suitable resolution for the two remaining elements. (© University of Basel, Department of Chemistry)


Felix Faber, Alexander Lindmaa, O. Anatole von Lilienfeld, and Rickard Armiento, Machine Learning Energies of 2M Elpasolite (ABC2D6) Crystals,
Physical Review Letters 117, 135502 (2016)
doi: 10.1103/PhysRevLett.117.135502