Index Estimates for CMC and Minimal Surfaces with Capillary Boundary
Luca Seemungal · 2026
Read Index Estimates for CMC and Minimal Surfaces with Capillary Boundary by Luca Seemungal (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-18783.txt.
Introduction
Seemungal bounds the index of CMC surfaces with capillary boundary. The index is capped by a linear mix of genus, boundary count, and branching order. A second bound uses a Willmore-type energy from area, curvature, and contact angle. The key tool compares second variations of area and energy at a branched conformal map. An appendix derives second variation formulas for area, volume, and wetting terms. Those formulas also cover variations away from critical points and non-admissible cases.
Catalog record
| CREATOR | Luca Seemungal |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 111828 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.18783
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Index Estimates for CMC and Minimal Surfaces with Capillary Boundary Luca Seemungal Abstract We prove that the index of a cmc surface with capillary boundary is bounded from above linearly by its genus, number of boundary components, and branching order, and also by some Willmore-type energy involving the area, mean curvature, contact angle, and ambient curvature. The main auxiliary theorem of more general interest is a comparison of the second variations of area and energy at a branched conformal map with boundary. In the appendix we derive the various second variation formulae for area, enclosed-volume, and wetting functionals away from critical points and for non-admissible variations, the purpose of which is to rather comprehensively fill a gap in the literature. 1 Introduction Let \Sigma be a compact Riemann surface with boundary, and let M^{3} be a Riemannian three-manifold with boundary. It is well-known that, for boundaried (branched) conformal maps 1 1 1 We call a map u boundaried if u(\partial\Sigma)\subset\partial M . We call a map branched if \mathrm{d}u vanishes on a discrete subset of the domain and these singularities have the structure of branch points, as in [ gulliver-osserman-royden-1973 ] . u:\Sigma\to M , the area \mathcal{A}(u) and Dirichlet energy \mathcal{E}(u) agree up to zero’th and first order: u\text{ conformal}\implies\mathcal{A}(u)=\mathcal{E}(u)\text{ and }\mathrm{d}\mathcal{A}(u)\equiv\mathrm{d}\mathcal{E}(u). One is therefore behoven to study the second order relationship between them. Indeed, the idea of comparing the area and energy goes at least as far back as the work of J. Douglas (see, for instance, [ douglas-1939 , Sec. 5.3] ), but it was not until early in this millennium that N. Ejiri and M. Micallef performed the first detailed comparison of area and energy [ ejiri-micallef-2008 ] , where they computed the second order deficit \mathrm{d}^{2}{\left(\mathcal{E}-\mathcal{A}\right)}\!\left(u\right)\left[v,v\right] for a minimal surface u (a conformal map which is also harmonic), in terms of a certain \overline{\partial} -problem dependent on v . That such a comparison might be achievable is suggested by the following heuristic, firstly and elegantly exposited by Ejiri and Micallef, and only a brief outline of which we give here. As is well known, the area is invariant the group of diffeomorphisms of the domain, but the energy is only invariant under the subgroup of conformal diffeomorphisms. Starting therefore at a conformal map, one expects on the one hand that variations in the target that induce conformal diffeomorphisms will keep area and energy the same, whilst on the other hand that variations which do not induce conformal diffeomorphisms will incur some kind of conformal deficit between the area and the energy. Of course, that some variations cannot induce conformal diffeomorphisms is revealed by the dimension of the moduli space of conformal structures on the domain. Ejiri and Micallef formalised this heuristic and used the comparison to prove index bounds for branched minimal surfaces. For minimal surfaces, the second variation of area is given by q=\left|A\right|^{2}+\mathop{\mathrm{Ric}}(\nu,\nu) , where A is the second fundamental form, \mathop{\mathrm{Ric}} is the ambient Ricci curvature, and \nu is the normal to the minimal surface. Unfortunately, the fact is that, given L=-\Delta-q a Schrödinger operator on a two-dimensional domain 2 2 2 This is in contrast to higher dimensional domains, where one does have the Cwikel–Lieb–Rosenbljum inequality, see for instance [ li-yau-1983 ] . \Omega , it is not possible (as simple examples of q show) to have a general inequality of the form (here i_{L} is the number of negative eigenvalues of L ) i_{L}\leq C\int_{\Omega}\!\left|q_{+}\right|; such an inequality would give a bound of the desirous form i_{\mathcal{A}}\leq C(M)\mathcal{A}(u)+Cg , where C(M) depends on M and g is the genus of the surface. Fortuitously, there are by now established methods to bound the index of the energy. For the sake of exposition, we infidelitiously reduce the work of Ejiri–Micallef down to two steps. Firstly, the area–energy comparison \mathrm{d}^{2}{\left(\mathcal{E}-\mathcal{A}\right)}\!\left(u\right)\left[v,v\right] gives the comparison of area and energy indices i_{\mathcal{E}}\leq i_{\mathcal{A}}\leq i_{\mathcal{E}}+r , where r (simply put) is the dimension of the Teichmüller space, and therefore a topological-branching term. Secondly, one bounds i_{\cal{E}} by relevant geometric quantities (area in the compact case) using heat-kernel methods [ cheng-tysk-1994 ] . The upshot is that one gets the desired index bounds of the form (for, say, compact minimal surfaces) i_{\mathcal{A}}\leq C\mathcal{A}(u)+r , where C is a constant depending only on the ambient space M and r is some topological term. Let us also mention that V. Lima has carried out this procedure for free-boundary minimal surfaces [ lima-2022 ] , computing firstly the second variation comparison \mathrm{d}^{2}{\left(\mathcal{E}-\mathcal{A}\right)}\!\left(u\right)\left[v,v\right] at a free-boundary minimal surface, and secondly applying these heat-kernel methods to deduce the bounds. Most of that which I have hitherto described has assumed the surface to be minimal (which, one recalls, is a conformal and harmonic map). A brief glance at the heuristic however reveals that it should apply to arbitrary conformal maps. Indeed, together with B. Sharp, we computed [ seemungal-sharp-2026 ] the second variation comparison \mathrm{d}^{2}{\left(\mathcal{E}-\mathcal{A}\right)}\!\left(u\right)\left[v,v\right] at an arbitrary conformal map, but without boundary. Yet another brief glance reveals that it should also apply to arbitrary conformal maps with boundary, which is our main auxiliary theorem below. The proof we give is different to that in [ seemungal-sharp-2026 ] , in addition to dealing with multitude of terms appearing on the boundary. Theorem (Theorem 5 in main body) . Let M be a Riemannian three-manifold (possibly with boundary), let \Sigma be a Riemann surface (possibly with boundary), and suppose that u:\Sigma\to M is a boundaried branched conformal immersion. Then, for any variation \Phi:I\times\Sigma\to M of u such that \Phi(I,\partial\Sigma)\subset\partial M , we have \frac{\mathrm{d}{{}^{2}}}{\mathrm{d}{t^{2}}}\Big|_{t=0}{\left(\mathcal{E}-\mathcal{A}\right)\left(\Phi(t)\right)}=4\int_{\Sigma}\!\left|\mu\right|^{2}\,\mathrm{d}\Sigma, where, denoting v:=\Phi^{\prime}(0)\in\Gamma_{\partial M}(u^{*}TM) and splitting v=s+\sigma into s:=v^{\perp} and \sigma:=v^{\top} , \mu=\left(\nabla^{\top}_{\partial_{z}}\sigma^{\left(0,1\right)}-2e^{-2\lambda}\left\langle s,A(u_{z},u_{z})\right\rangle u_{\overline{z}}\right)\otimes\mathrm{d}z is the infinitesimal conformal deficit of v . Remark . (i) For an elucidation of the so-called ‘infinitesimal conformal deficit’ \mu , see Section 3 . Suffice it to say for now that \mu should be viewed as a PDE which asks, given a normal variation s , to find a tangential variation \sigma which reparametrises the domain so as to stay conformal (at the infinitesimal level, of course). (ii) What we have written above is a comparison of the two derivatives of \mathcal{E} and \mathcal{A} , which (since u is not necessarily a critical point of \mathcal{E} nor \mathcal{A} ) is not the same as the comparison of the “second variation” or Hessian of \mathcal{E} and \mathcal{A} . Indeed, it is not clear from the outset that the left-hand side is even bilinear. Nevertheless, the right-hand side is bilinear, so the non-bilinearities in the second derivatives must cancel precisely (as one can see in the proof). For this reason, if one wished to write the comparison as one of the Hessians of \mathcal{E} and \mathcal{A} over some infinite-dimensional space of variations, then one could readily deduce this from what we have here, even though this is not necessary in this paper. (iii) Weaker forms of this area–energy comparison for minimal surfaces has appeared in various works, (for instance, [ cheng-zhou-2023 , gao-zhu-2024 ] ), where the computations are much simplified by somewhat weaker theorem statements (i.e., the area–energy comparison is an inequality rather than an equality) and with assumptions about the target or the domain