Prediction of new Ti-N phases using machine learned interatomic potential

Key Points

• New titanium-nitrogen (Ti-N) phases were accurately predicted using a machine learning approach, enhancing material characterization.
• The model shows high accuracy with a predictive capability of 92% in discovering new Ti-N phases across diverse conditions.
• Analysis employing a machine learned interatomic potential algorithm allows for efficient exploration of atomic interactions in materials.
• These findings highlight the potential for machine learning to revolutionize the discovery process in the field of material science.