Recorded 30 October 2025. Jing Yang of the Max Planck Institute for Iron Research presents "Molecular dynamics simulation for electrochemical systems: challenges of the field fluctuation" at IPAM's Boundary Conditions for Atomistic Simulations in Macroscopic Electrochemical Cells Workshop. Abstract: Ab initio molecular dynamics (AIMD) serves as powerful method for studying the electrochemical processes at solid/liquid interfaces. To drive these surface reactions, a macroscopic electric field needs to be applied to the system, which is a major source of computational complexity. For these calculations, there exists multiple methods with different treatments of the counter charges and the boundary conditions, some of which are not necessarily included in standard DFT codes. In addition, the AIMD calculations are often prohibitively expensive for observing the reaction process in real time. Machine learning force fields (MLFFs) offer the promise of a much cheaper alternative, but the accuracy of such potentials for describing long-range electrostatic interaction is still questionable. In this talk, we first discuss the implementation for calculating electrified surface under arbitrary field by utilizing the newly released VASP-python interface. This development allows one to perform AIMD with applied electric field using the standard VASP code with great control and flexibility. Next, we discuss some of the recent developments of machine learning potentials with long-range electrostatic interaction. We show that ML models that are trained on local charge distributions, such as Hirshfeld charge or Wannier centers, do not necessarily reproduce the macroscopic electrostatics. We propose a model which is instead trained on the macro dipole moment and show that such a model can predict the macroscopic electric field for realistic electrochemical systems. Learn more online at: https://www.ipam.ucla.edu/programs/workshops/workshop-iii-boundary-conditions-for-atomistic-simulations-in-macroscopic-electrochemical-cells/