The Kato square root problem for mixed boundary conditions (Q2253147)

Let \(-\nabla\cdot\mu\nabla\) be an elliptic differential operator in divergence form with bounded complex coefficients over a domain \(\Omega,\) subject to Dirichlet boundary conditions on some closed subset \(D\subset\partial\Omega\) and natural boundary conditions on \(\partial\Omega\setminus D\) in the sense of the form method. Let \(A\) be the maximal accretive realization of \(-\nabla\cdot\mu\nabla\) on \(L^2(\Omega)\). The \textit{Kato Square Root Problem} for \(A,\) as posed by \textit{J. L. Lions} [J. Math. Soc. Japan 14, 233--241 (1962; Zbl 0108.11202)], amounts to identify the domain of the maximal accretive square root of \(A\) as the domain of the corresponding form, that is, the subspace of the first order Sobolev space \(H^1(\Omega)\) whose elements vanish on \(D.\) Considering the negative Laplacian subject to mixed boundary conditions on \(\partial\Omega,\) the authors prove that under very general geometric assumptions there exists an \(\alpha>1/2\) such that the domain of \((-\Delta)^\alpha\) is a Sobolev space of optimal order \(2\alpha.\) In combination with a reduction theorem recently obtained by the authors, this solves the Kato Square Root Problem for \(-\nabla\cdot\mu\nabla\) under the same geometric assumptions.