Removable sets for continuous solutions of quasilinear elliptic equations with nonlinear source or absorption terms (Q681678)

The authors investigate the removability of certain singularities of weak solutions of quasilinear elliptic equations in an open set \(\Omega\subset \mathbb{R}^n\) (\(n\geq 2\)) of the form \[ (1) \;\;-\text{div}\mathcal{A}(x,\nabla u)=\mathcal{B}(x,u,\nabla u), \;\;\;\text{and} \;\;\;(2) \;\;\text{div}\mathcal{A}(x,\nabla u)=\mathcal{B}(x,u,\nabla u). \] Here, \(\mathcal{A}:\Omega\times \mathbb{R}^n\rightarrow \mathbb{R}^n\) is a Carathéodory function satisfying the structure conditions (\(\|\cdot\|\) stands for the Euclidean norm in \(\mathbb{R}^n\)) \((a)\) \ \(\displaystyle{\inf_{(x,\xi) \in \Omega \times (\mathbb{R}^n\setminus\{0\})}\frac{\langle \mathcal{A}(x,\xi),\xi\rangle}{\|\xi\|^p}>0}\), \(\displaystyle{\sup_{(x,\xi) \in \Omega \times (\mathbb{R}^n\setminus\{0\})}\frac{\|\mathcal{A}(x,\xi)\|}{\|\xi\|^{p-1}}< \infty}\), \ \ \(p\in (1,n)\), \((b)\) \ \(\big\langle\mathcal{A}(x,\xi_1)-\mathcal{A}(x,\xi_2),\xi_1-\xi_2\big\rangle>0\), for all \(x\in \Omega\) and \(\xi_1,\xi_2\in \mathbb{R}^n\) with \(\xi_1\neq \xi_2\), \((c)\) \ \(\mathcal{A}(x,\lambda \xi)=\lambda|\lambda|^{p-2}\mathcal{A}(x, \xi)\), for all \((x,\xi,\lambda)\in \Omega\times\mathbb{R}^n\times (\mathbb{R}\setminus\{0\})\), and \(\mathcal{B}:\Omega\times \mathbb{R}\times \mathbb{R}^n\rightarrow \mathbb{R}\) is a Borel measurable function such that \(\mathcal{B}(x,\cdot,\xi)\) is nonnegative in \([0,+\infty)\) for all \( (x,\xi)\in \Omega\times \mathbb{R}^n\). The main results are the following removable singularities theorems for equations \((1)\) and \((2)\): Let \(m\in [0,n-p)\) and let \(E\subset \Omega\) be a compact set such that for some constants \(r_0>0\) and \(C_0\geq 1\), \[ \big|\big\{y\in \mathbb{R}^n: d(y,E)<r,\; d(x,y)<R \big\}\big|\leq C_0r^{n-m}R^m, \quad\text{for all}\quad x\in E\text{ and } r,R\in (0,r_0). \] Assume that \(\mathcal{B}\) satisfies the growth condition \[ \sup_{(x,t,\xi)\in \Omega\times \mathbb{R}\times \mathbb{R}^n}\frac{|\mathcal{B}(x,t,\xi)|}{1+|t|^q}<\infty,\quad\text{for some}\quad q\in \left(0,\frac{(n-m)(p-1)}{n-m-p}\right). \] Then, every nonnegative continuous weak solution \(u\) of equation \((1)\) in the open set \(\Omega\setminus E\) which satisfies \[ u(x)=o\Big(d(x,E)^{\frac{p-n+m}{p-1}}\Big) \;\text{as} \;x\rightarrow y, \quad\text{for all}\quad y\in E, \] can be continuously extended to the whole \(\Omega\). The same conclusion is obtained for equation \((2)\) provided that \(\mathcal{B}\) satisfies the growth condition \(\mathcal{B}(x,\cdot,\xi)\in L_{\text{loc}}^\infty(\mathbb{R})\), for all \((x,\xi) \in \Omega\times \mathbb{R}^n\), and \(E\) satisfies the additional condition \[ \frac{1}{C_0}r^m\leq \mathcal{H}^m(\{y\in E: d(x,y)<r\})\leq C_0r^m, \quad\text{for all}\quad x\in E\text{ and }r\in (0,r_0), \] where \(\mathcal{H}^m\) denotes the \(m\)-dimensional Hausdorff measure. One of the main ingredients of the proofs is a Wolff potential estimate established by Kipel\(\ddot{\text{a}}\)inen and Malý.