An inverse stability result for non-compactly supported potentials by one arbitrary lateral Neumann observation (Q264477): Difference between revisions

Consider a Schrödinger equation \(-i\partial_t u-\Delta u+p(x)u=0\), where \((x,t)\in\Omega\times(0,T)\), \(\Omega=\omega\times\mathbb{R}\) an infinite straight cylinder with bounded cross section \(\omega\subset\mathbb{R}^{n-1}\) and boundary \(\Gamma=\partial\omega\times\mathbb{R}\). Fix initial data \(u_0:\Omega\to\mathbb{R}\). Fix a nonempty open subset \(S_*\subset\partial\omega\), and set \(\Gamma_*=S_*\times\mathbb{R}\). The main result of the paper states that the potential \(p\) is uniquely determined by the trace of the normal derivative \(\partial_\nu u\) at \(\Gamma_*\times(0,T)\) of the solution \(u(x,t)\) which satisfies \(u(\cdot,0)=u_0\) and zero Dirichlet boundary conditions, \(u=0\) at \(\Gamma\times(0,T)\). Moreover, a logarithmic stability estimate is shown. The initial value is assumed non-degenerate in the sense that \(|u_0(x',x_n)|(1+x_n^2)^d\geq\kappa\) holds with some constants \(d\) and \(\kappa>0\); here \(x_n\) denotes the coordinate of the cylinder axis. Moreover, the following restrictive hypothesis is put on the class of potentials \(p\): There is a potential \(p_0\) such that all allowed potentials \(p\) are equal to \(p_0\) in a neighbourhood of the boundary. The authors argue that these assumptions are acceptable from a practical viewpoint for the inverse problem of recovering the impurity potential \(p\) in a nanotube \(\Omega\). The proof of the uniqueness and stability theorem employs the technique of Carleman estimates.