Water at Interfaces\nInvestigated by Neural Network-Based\nMolecular Dynamics Simulation

Interfacial\nwater plays a crucial role in various biological, chemical,\nphysical, and material processes. Sum-frequency vibrational spectroscopy\n(SFVS) is a powerful method for studying molecular orientation, structure,\nand dynamics at interfaces. By employing molecular dynamics simulations\nbased on artificial neural network potentials, we explore water at\ninterfaces, taking into account non-Condon effects from the electric\nfields of surrounding water molecules and neglecting a rigorous treatment\nof quantum nuclear motion. Our simulations are not able to replicate\nthe shoulder of the free OH peak. The influence of the electric dipole\ndominates SFVS, while that of the electric quadrupole is relatively\nsmall but still important and nonignorable. This influence contributes\nto the absence of the low-frequency positive peak of OH stretch vibrations.\nThis research represents a significant advancement in the understanding\nof the properties of water at interfaces and the mechanics of SFVS.