Understanding the response of interfacial water to electric fields is critical for elucidating its unique vibrational dynamics. Here, we employ a mixed quantum/classical approach coupled with neural network-based molecular dynamics (MD) simulations to compute the in-plane transverse electric field at hydrogen sites of interfacial water. Our results reveal a bidirectional field distribution (positive and negative values), where a positive field blueshifts the bending mode frequency, while a negative field fails to induce a redshift in sum-frequency vibrational spectroscopy (SFVS). By integrating ab initio MD simulations with quantum chemistry calculations, we reproduce experimental SFVS and explain the absence of a redshifted bending peak. Crucially, we identify the electric quadrupole contribution as the dominant factor governing the bending mode response. The proposed ab initio framework maintains high accuracy while reducing computational costs significantly, offering a robust tool for studying the interfacial spectroscopy of water.
Zheng et al. (Mon,) studied this question.