On the existence of area-minimizing surfaces with free boundary (Q1174587): Difference between revisions

If \(N\) is a complete Riemannian manifold of dimension \(n\geq 3\) and \(M\subset N\) an embedded submanifold of dimension \(m\geq 2\) (e.g. \(M=\partial N\)) one is naturally led to the following free boundary problem. Given a free homotopy class of closed curves on \(M\) find a minimal surface with free boundary belonging to this class. Here, a two- dimensional minimal surface \(S\subset N\) is said to have a free boundary on \(M\) if \(\partial S\subset M\), and \(S\) meets \(M\) orthogonally along \(\partial S\). To attack this problem, the author uses the standard procedure of minimizing the Dirichlet integral among parametric surfaces. Of course, the basic difficulty is to control the boundary values of a minimizing sequence to ensure that the homotopy condition is fulfilled in the limit. In contrast to Courant's classical approach via linking numbers (related to homology) the author uses so-called \(L\)-minimizers, i.e. the Dirichlet integral is minimized among surfaces whose boundary length does not exceed \(L\). As \(L\to \infty\) these \(L\)-minimizers may only develop singularities at finitely many points along their boundary. These lead to splitting of surfaces and decompositions of boundary homotopy classes. In this way it is possible to derive a-priori-Hölder-estimates for \(L\)- minimizers (independent of \(L\)). With this approach the author is able to prove the following result: If a free homotopy class satisfies the ``Douglas condition'', then there exists an area-minimizing surface with free boundary in it.