The paper tackles a problem in surface science: when a liquid droplet sits on a rough surface, the angle at which the liquid meets the solid (the contact angle) is not a single fixed value but rather a range. The droplet resists moving until external forces push the contact angle outside this range, a phenomenon called contact angle hysteresis. On a rough surface with a repeating microscopic texture, the angle measured at the large scale can differ substantially from what the local chemistry of the surface would predict. The challenge is that this range of angles also depends on direction, meaning a droplet may be much easier to slide one way than another. The authors develop a mathematical framework to precisely define and compute this directional range of angles for surfaces with a doubly periodic pattern, meaning the roughness repeats in two independent directions like a wallpaper pattern.
Computing this range is numerically difficult because two very different length scales are involved simultaneously. The microscopic roughness determines where the contact line pins and sticks, but the apparent angle only becomes meaningful when you zoom out to the macroscopic view of the droplet surface. Resolving both at once on a single numerical grid is expensive and problematic. To handle this, the authors introduce a two-scale alternating method inspired by a classical technique for splitting overlapping computational domains. Near the contact line where pinning happens, they use a diffusion-based scheme well suited to capturing sharp geometric features. Far from the contact line, they solve a simpler linearized problem describing a nearly flat liquid surface. These two solvers are applied in alternation, and the authors prove that a suitable measure of surface energy decreases with each iteration, giving confidence the method converges correctly.
When they test the method on a representative rough surface, the results reveal strongly anisotropic hysteresis, meaning the range of contact angles varies dramatically depending on which direction the contact line runs across the surface. The width of this range changes by more than a factor of three depending on orientation and shifts sharply near diagonal directions of the surface texture. This anisotropy has a direct physical consequence: it constrains the possible shapes of stationary droplets. A droplet will naturally orient itself so its edges align with the directions where pinning is strongest, and the calculations predict a roughly square droplet shape with sides running diagonally relative to the underlying surface pattern. This connects microscopic surface geometry to observable macroscopic droplet behavior in a quantitatively precise way.