[Paper Review] Propensity of water self-ions at air(oil)-water interfaces revealed by deep potential molecular dynamics with enhanced sampling

Pengchao Zhang, Xuefei Xu|arXiv (Cornell University)|Apr 10, 2024

Spectroscopy and Quantum Chemical Studies4 citations

TL;DR

This study uses deep potential molecular dynamics with enhanced sampling to resolve the long-standing controversy over water self-ion (H₃O⁺ and OH⁻) distribution at air- and oil-water interfaces. It reveals a stable interfacial double-layer structure where H₃O⁺ resides in the topmost layer and OH⁻ is enriched in the deeper interfacial layer due to its stronger interfacial stabilization free energy, reconciling conflicting macroscopic and microscopic observations.

ABSTRACT

The preference of water self-ions (hydronium and hydroxide) towards air/oil-water interfaces is one of the hottest topics in water research due to its importance for understanding properties, phenomena, and reactions of interfaces. In this work, we performed enhanced-sampling molecular dynamics simulations based on state-of-the-art neural network potentials with approximate M06-2X accuracy to investigate the propensity of hydronium and hydroxide ions at air/oil(decane)-water interfaces, which can simultaneously describe well the water autoionization process forming these ions, recombination of ions, and ionic distribution along the normal distance to the interface by employing a set of appropriate Voronoi collective variables. A stable ionic double-layer distribution is observed near the air-water interface, while the distribution is different at oil-water interfaces, where hydronium tends to be repelled from the interface into the bulk water, whereas hydroxide, with an interfacial stabilization free energy of -0.6 kcal/mol, is enriched in the interfacial layer. Through simulations of oil droplets in water, we further reveal that the interfacial propensity of hydroxide ions is caused by the positive charge distribution of the oil-water interface contributed by hydrogens of the dangling OH bonds of interfacial water layer and the outmost layer decane molecules laying flat on the droplet. The present results may aid in understanding the acid-base nature of water interfaces with wide applications.

Motivation & Objective

• To resolve the conflicting experimental and simulation results on the interfacial propensity of water self-ions (H₃O⁺ and OH⁻) at air- and oil-water interfaces.
• To investigate the role of water autoionization and ionic recombination in shaping interfacial ion distributions.
• To validate the stability and accuracy of the double-layer ionic distribution model under realistic reaction conditions.
• To compare interfacial ion behavior between air-water and oil-water interfaces, particularly regarding ion stabilization free energy differences.
• To reconcile macroscopic measurements (e.g., negative zeta potential) with microscopic observations by accounting for depth-dependent probing in experiments.

Proposed method

• Employed deep potential molecular dynamics (DPMD) with neural network potentials trained to M06-2X functional accuracy for high-accuracy, efficient simulation of water and ions.
• Applied on-the-fly probability enhanced sampling (OPES) with Voronoi collective variables to efficiently sample rare events such as water autoionization and ionic recombination.
• Used a set of collective variables based on Voronoi tessellation to track ionic formation and recombination dynamics in real space.
• Calculated free energy profiles for water autoionization and interfacial stabilization using umbrella sampling and free energy perturbation techniques.
• Performed simulations on both air-water and oil-water interface systems to compare ion distribution and stabilization energies.
• Validated the model by reproducing experimental autoionization free energy (155.5 kcal/mol) and ensuring force prediction accuracy within 0.3 kcal/mol per atom.

Experimental results

Research questions

• RQ1 What is the true interfacial distribution of H₃O⁺ and OH⁻ ions at air-water and oil-water interfaces, considering their dynamic formation and recombination?
• RQ2 How does the interfacial stabilization free energy of H₃O⁺ compare to that of OH⁻, and what drives the observed double-layer structure?
• RQ3 Why do macroscopic experiments (e.g., zeta potential) show negative surface charge while some microscopic techniques detect H₃O⁺ at the surface?
• RQ4 How does the presence of oil instead of air affect the interfacial propensity of OH⁻ and H₃O⁺, and what molecular factors contribute to this difference?
• RQ5 Can enhanced sampling with deep potential models accurately reproduce the free energy of water autoionization and its interfacial effects?

Key findings

• The study confirms a stable interfacial double-layer distribution of H₃O⁺ and OH⁻ at both air-water and oil-water interfaces, with H₃O⁺ localized in the topmost interfacial layer and OH⁻ enriched in the deeper interfacial layer.
• OH⁻ exhibits a significantly more negative interfacial stabilization free energy (by ~0.3 kcal/mol) than H₃O⁺, explaining its higher interfacial concentration and the observed negative zeta potential.
• The free energy of water autoionization in bulk water was accurately reproduced (155.5 kcal/mol), validating the deep potential model and sampling methodology.
• The oil-water interface stabilizes OH⁻ more strongly than the air-water interface due to greater asymmetry and amphipathic character in OH⁻, consistent with more frequent detection of negative zeta potential in oil-in-water systems.
• The double-layer model reconciles conflicting experimental and simulation results by showing that different techniques probe different interfacial depths: surface-sensitive methods detect H₃O⁺, while bulk-averaged or subsurface probes detect OH⁻.
• High-level heterodyne-detected vSFG experiments may corroborate the findings, as they also suggest H₃O⁺ at the outermost surface and OH⁻ in the subsurface layer.