Anomalous corrections to the Kelvin equation at complete wetting

TL;DR

This paper investigates the corrections to the Kelvin equation in a confined Ising model, revealing anomalous scaling behaviors at complete wetting due to thin wetting layers, supported by numerical and simplified models.

Contribution

It uncovers unexpected 1/L^{4/3} corrections at complete wetting, contrasting with the predicted 1/L^{5/3} scaling, and explains this via the influence of thin wetting layers.

Findings

At partial wetting, corrections follow a 1/L^2 scaling.

At complete wetting, a 1/L^{4/3} correction dominates for intermediate system sizes.

A crossover to 1/L^{5/3} correction occurs for larger systems, confirmed by a simplified solid-on-solid model.

Abstract

We consider an Ising model confined in an $L \times \infty$ geometry with identical surface fields at the boundaries. According to the Kelvin equation the bulk coexistence field scales as 1/L for large L; thermodynamics and scaling arguments predict higher order corrections of the type $1/L^2$ and $1/L^{5/3}$ at partial and complete wetting respectively. Our numerical results, obtained by density-matrix renormalization techniques for systems of widths up to L = 144, are in agreement with a $1/L^2$ correction in the partial wet regime. However at complete wetting we find a large range of surface fields and temperatures with a correction to scaling of type $1/L^{4/3}$. We show that this term is generated by a {\it thin} wetting layer whose free energy is dominated by the contacts with the wall. For $L$ sufficiently large we expect a crossover to a $1/L^{5/3}$ correction as predicted by the theory for a {\it thick} wetting layer. This crossover is indeed found in a solid-on-solid model which provides a simplified description of the wetting layer and allows the study of much larger systems than those available in a density-matrix renormalization calculation.