Regularity of the free boundary in the biharmonic obstacle problem (Q2338494)

In the paper under review, the author uses a flatness improvement argument to study the regularity of the free boundaries for biharmonic obstacle problem with zero obstacle. More precisely, given a bounded domain \(\Omega\subset \mathbb{R}^n\), let \(u\) be the unique minimizer of the functional \[ \int_{\Omega}(\Delta u(x))^2 dx, \] among the admissible set \[ \{u\in W^{2,2}(\Omega), \ u\geq 0, \ u=g>0, \ \partial_\nu u = f \text{ on } \partial\Omega\} \] for some given boundary data \(g\) and \(f\). Assume that \(0\in \Gamma_u:=\partial\{u>0\}\) is a free boundary point and \(B_2\) is compactly contained in \(\Omega\). The author proves the following improvement of flatness result: if \(u\) is almost one-dimensional, i.e. \(\|\nabla'u\|_{W^{2,2}(B_2)}\leq \epsilon\), where \(\nabla'=\nabla-e_n \partial_{x_n}\), if \(\{u>0\}\) is a non-tangentially accessible domain and if \(u=0\) in \(B_2\cap \{x_n<-t\}\) for some \(t\in (0,2)\), then \(\Gamma_u\cap B_1\) is a \(C^{1,\alpha}\) graph for some \(\alpha\in (0,1)\). As a consequence, \(u \in C^{2,1}(B_1)\), which is the best possible regularity due to the model solution \((x_n)_+^{3}\). The author also provides a counterexample, showing that without the flatness assumption \(C^{2,\frac{1}{2}}\) is the best possible regularity in dimension \(n\geq 2\). To prove the result the author adopts the linearization argument by \textit{J. Andersson} [``Almost everywhere regularity for the free boundary of the normalized \(p\)-harmonic obstacle problem \(p>2\)'', Preprint, \url{arXiv:1611.04397}; Invent. Math. 204, No. 1, 1--82 (2016; Zbl 1339.35345)].