Comparing fine-tuning strategies for machine learning force fields in lithium-ion diffusion

Machine learning force fields (MLFFs) are transforming materials science and engineering by enabling the study of complex phenomena, such as those critical to battery operation. In this work, we explore the predictive capabilities of pre-trained and fine-tuned MACE MLFF and compare different fine-tuning strategies to predict interstitial lithium diffusivity in LiF, a key component in the solid electrolyte interphase in Li-ion batteries. Our results demonstrate that the MACE-MPA-0 foundational model achieves comparable accuracy to the well-trained DeePMD in predicting key diffusion properties based on large-scale molecular dynamics simulations, while requiring minimal or no training data. For instance, MACE-MPA-0 predicts an activation energy Ea of 0.22 eV, the fine-tuned model with only 300 data points predicts Ea = 0.20 eV, both of which show good agreement with the DeePMD model reference value of Ea = 0.24 eV. In this work, we provide a solid test case where fine-tuning approaches, whether using data generated for DeePMD or data produced by the foundational MACE model itself, yield similar robust performance to the DeePMD potential trained with over 40,000 actively learned data, albeit requiring only a fraction of the training data.