AI Technology: Machine Learning
Since their discovery in the 1940s, Zinc Dialkyldithiphosphates (ZDDPs) have been recognized as the most effective antiwear lubricant additive in extreme tribology thanks to the formation of protective tribofilms.
However, the growth mechanisms and effectiveness of the ZDDP tribofilms depends heavily on the substrates that governs their initial adsorption and tribochemical reactions.
In this project, the team uses machine learning-based interatomic potentials (MLP) trained on ab initio data for large-scale molecular dynamics (MD) simulations to explore a comprehensive process of ZDDP tribofilm formation, starting from initial molecular adsorption to final tribofilm with different surfaces of steels.
High-performance computing resources will enable the study to train and deploy these potentials, simulating the full reaction pathway for the first time.
This project will provide a molecular-level understanding of ZDDP tribofilm formation, driving the development of more environmentally friendly lubricant additives while advancing computational tribology.
Huong Thi Thuy Ta, University of Bologna, Italy