In a recent paper, Sharifi-Mood et al. study colloidal particles trapped at a liquid interface with opposite principal curvatures c1 = −c2. In the theory part, they claim that the trapping energy vanishes at second order in ∆c = c1 − c2, which would invalidate our previous result [Phys. Rev. E, 2006, 74, 041402]. Here we show that this claim arises from an improper treatment of the outer boundary condition on the deformation field. For both pinned and moving contact lines, we find that the outer boundary is irrelevant, which confirms our previous work. More generally, we show that the trapping energy is determined by the deformation close to the particle and does not depend on the far-field.
PACS numbers:Colloidal particles trapped at a curved liquid interface are subject to capillary forces that do not depend on their mass or charge but on geometrical parameters only. In Ref.[1], Sharifi-Mood et al. provide an interesting analysis of the role of contact line pinning. Regarding the trapping energy, however, these authors assert that it vanishes at second order, contrary to previous work, and they state that "the origin of the discrepancy is an inappropriate treatment of the contour integral" in [2]. The present comment intends to refute this claim of [1], to unambiguously determine the trapping energy, and to clarify the role of the far-field.Previous works [2-4] rely on the assumption that curvature-induced forces arise from the interface close to the particle and that the far-field is irrelevant. Thus the profile of the bare interface is taken in small-gradient approximation, h 0 = ∆c 4 cos(2ϕ)r 2 , which is valid only at distances shorter than the curvature radius R c = 1/∆c. Adding a particle modifies the profile as h = h 0 +η, where the deformation field η = ∆c 12 cos 2ϕ r 4 0is determined from the contact angle at the particle surface. By the same token, the trapping energy comprises only near-field contributions, and is given by the boundary term along the contact line of radius r 0 ,Since a similar line integral of η∇η is cancelled by the area change due to displacement of the contact line on the particle surface, one obtains the curvature-dependent energy E = E in [2], which was confirmed in [3,4]. Sharifi-Mood et al. attempt to go beyond the nearfield approach and to evaluate the term arising at the outer boundary,where, in the simplest geometry, R out (ϕ) is the distance of the container wall from the particle. Using the above expressions h 0 ∝ r 2 and η ∝ r −2 , and letting R out → ∞, these authors find in (24) of [1] the relation E out = −E in . This leads them to the conclusion that the trapping energy vanishes at second order, E = E in + E out = O(∆c 4 ). Yet this argument is flawed by the fact that E out is calculated with the near-field deformation (1) which is not correct at R out . (Moreover, h 0 ∝ r 2 is valid at distances within the curvature radius only [5].) In the following we evaluate E out with the the correct deformation field η, which satisfies appropriate conditions at the outer bou...