Infrared spectrum of water under external electric fields remains unresolved

Determine, from first principles, the infrared absorption spectrum of bulk liquid water under finite external electric fields, characterizing how applied fields affect spectral features and intensities across relevant vibrational modes in the liquid phase.

Background

Accurate theoretical prediction of the infrared spectrum of liquid water has historically been challenging due to the intricate hydrogen-bond network and the coupling of intra- and intermolecular motions. This difficulty is exacerbated under external electric fields, where field-induced changes in structure and dynamics modify spectral features. The paper introduces a machine-learning interatomic potential framework (CACE-LR with LES) that infers Born effective charges from energy and force data, enabling electric-field-driven simulations and IR spectrum prediction without directly training on charges or polarization.

The authors emphasize that despite progress, the theoretical description of water's IR response—particularly in the presence of external fields—remains not fully resolved, motivating scalable, accurate approaches capable of capturing electrical response in bulk systems.