Multi-bump analysis for Trudinger-Moser nonlinearities I. Quantification and location of concentration points (Q2216753)

In this interesting paper the authors study the blow-up behavior of sequences of solutions of certain elliptic PDEs which contain a nonlinearity with Trudinger-Moser growth, in dimension two. The authors answer questions posed in [the first author, Duke Math. J. 132, No. 2, 217--269 (2006; Zbl 1281.35045); \textit{M. del Pino} et al., J. Funct. Anal. 258, No. 2, 421--457 (2010; Zbl 1191.35138); \textit{A. Malchiodi} and \textit{L. Martinazzi}, J. Eur. Math. Soc. (JEMS) 16, No. 5, 893--908 (2014; Zbl 1304.49011); \textit{L. Martinazzi}, Calc. Var. Partial Differ. Equ. 36, No. 4, 493--506 (2009; Zbl 1180.35211)]. In particular, the authors prove that the blow-up phenomenon is very restrictive because of the strong interaction between bubbles. Let \(\Omega\subset\mathbb{R}^2\) be a smooth bounded domain. In [the first author, Duke Math. J. 132, No. 2, 217--269 (2006; Zbl 1281.35045)] it was proved that if \((f_{\varepsilon})_{\varepsilon>0}\) is a sequence of functions of uniform critical growth in \(\Omega\), and \((u_{\varepsilon})_{\varepsilon>0}\) is a sequence of solutions of \[\Delta u_{\varepsilon}=f_{\varepsilon}(x,u_{\varepsilon}(x)), \] with \(||\nabla u_{\varepsilon}||_2^2\rightarrow\beta\) as \(\varepsilon\rightarrow 0\), for some \(\beta\in\mathbb{R}\), then there exists a solution \(u_0\in C^0(\overline{\Omega})\) of \[ \Delta u_0=f_0(x,u_0(x)) \text{ in } \Omega, \qquad u_0=0 \text{ on } \partial \Omega, \] and there exists \(\tilde{N}\in\mathbb{N}\) such that \[ ||\nabla u_{\varepsilon}||_2^2=||\nabla u_0||_2^2+4\pi\tilde{N}+o(1). \] If \(\tilde{N}=0\), the convergence of \(u_{\varepsilon}\) to \(u_0\) is strong in \(H^1_0(\Omega)\), and holds in \(C^0(\overline{\Omega})\). In this paper the authors discuss whether it is possible to have both \(u_0\neq 0\) and \(\tilde{N}\ge 1\). Moreover, they analyze if the concentration points appearing when \(\tilde{N}\ge 1\) are isolated or not. More specifically, the authors consider a sequence \((\lambda_{\varepsilon})_{\varepsilon>0}\) of positive real numbers, a sequence \((f_{\varepsilon})_{\varepsilon}\) of smooth functions in \(\overline{\Omega}\) satisfying \(f_{\varepsilon}\rightarrow f_0\) in \(C^1(\overline{\Omega})\) and \(||\nabla^2f_{\varepsilon}||_{L^{\infty}(\Omega)}=O(1)\), where \(f_0>0\) in \(\overline{\Omega}\). Then, they consider a sequence \((u_{\varepsilon})_{\varepsilon>0}\) of smooth positive solutions of \[ \Delta u_{\varepsilon}=\lambda_{\varepsilon}f_{\varepsilon}u_{\varepsilon}e^{u_{\varepsilon}^2} \text{ in } \Omega, \qquad u_{\varepsilon}=0 \text{ on } \partial\Omega. \] The authors show that if blow-up occurs, then the weak limit has to be zero, so lack of compactness can only happen for \(\beta=4\pi N\), with \(N\ge 1\). The authors also obtain a characterization of the location of concentration points, confirming the conjecture of \textit{M. del Pino} et al. [J. Funct. Anal. 258, No. 2, 421--457 (2010; Zbl 1191.35138)].