Scientific Calendar Event



Description

Andrea Grisafi
(Sorbonne Université - CNRS)


Abstract:
Machine-learning methods are playing a prominent role in the atomistic simulation of materials. When it comes to electrochemical systems, however, the effective implementation of these approaches presents unique challenges. Notably, the presence of a heterogeneous interface between an electrode and an electrolyte solution under an applied electric bias requires predicting complex long-range interactions that drive the dynamics of the system at the nanoscale. In this context, I will introduce a machine-learning method suitable to treat the electron density of a system [1] and show how it can be used to predict the nonlocal polarization of gold electrodes that are embedded in concentrated ionic solutions [2]. I will continue discussing the application of this method to simulate the electrostatic properties of model ionic capacitors under an applied voltage, matching the accuracy of density-functional QM/MM methods while reaching time scales of several nanoseconds [3].
 
1.Andrea Grisafi, Alberto Fabrizio, Benjamin Meyer, David M. Wilkins, Clemence Corminboeuf, Michele Ceriotti, Transferable machine learning model of the electron density, ACS Central Science 5, 57-64 (2019) [https://doi.org/10.1021/acscentsci.8b00551
2.Andrea Grisafi, Augustin Bussy, Mathieu Salanne and Rodolphe Vuilleumier, Predicting the charge density response in metal electrodes, Physical Review Materials 7, 125403 (2023) [https://doi.org/10.1103/PhysRevMaterials.7.125403]
3.Andrea Grisafi and Mathieu Salanne, Accelerating QM/MM simulations of electrochemical interfaces through machine learning of electronic charge densities, The Journal of Chemical Physics 161, 024109 (2024) [https://doi.org/10.1063/5.0218379]

 
Go to day