Nonlinear Effects of Hydrophobic Confinement on the Electronic Structure and Dielectric Response of Water

Fundamental studies of the dielectrics of confined water are critical to understand the ion transport across biological and synthetic nanochannels. The relevance of these fundamental studies, however, surmounts the difficulty of probing water's dielectric constant as a function of a fine variation in confinement. In this work, we explore the computational efficiency of machine learning potentials to derive the confinement effects on the dielectric constant, polarization, and dipole moment of water. Our simulations predict an enhancement of the axial dielectric constant of water under extreme confinement, arising from either the formation of ferroelectric structures of ordered water or larger dipole fluctuations facilitated by the disruption of water's H-bond network. Our study highlights the impact of hydrophobic nanoconfinement on the dielectric constant and on the ionic and electronic structure of water molecules, pointing to the importance of geometric flexibility and electronic polarizability to properly model confinement effects on water.