On a theorem of Deift and Hempel (Q1104487)

The paper presents in strikingly simple proof of a previous result in \textit{P. A. Deift} and \textit{R. Hempel} [Commun. Math. Phys. 103, 461-490 (1986; Zbl 0594.34022)] on eigenvalues of a Schrödinger operator \(H_ 0+\lambda W\) in spectral gaps of \(H_ 0\). More precisely, the following theorem has been proven: Theorem. Let \(V,W\in L^{\infty}({\mathbb{R}}^{\nu})\) be real-valued, \(\nu\in {\mathbb{N}}\), supp(W) compact, \(W_-(x)\geq 1\) for all x in some open ball of radius \(\epsilon >0\). (Here \(W_{\mp}:=(| W|_{\mp}W)/2)\). Define \(H_ 0=-\Delta +V\), \(H=H_ 0+\lambda W\), \(\lambda\geq 0\) on \(H^ 2({\mathbb{R}}^{\nu})\). Let \((a,b)\subset {\mathbb{R}}\setminus \sigma (H_ 0)\) be a spectral gap of \(H_ 0\) and assume \(E_ 0\in (a,b)\). Then there exists a sequence of positive numbers \(\lambda_ n\uparrow \infty\) such that \(E_ 0\in \sigma_ p(H_{\lambda_ n})\), \(n\in {\mathbb{N}}.\) The proof (in contrast to that of Deift-Hempel (loc. cit.)) is based on the use of a self-adjoint Birman-Schwinger kernel. While the condition supp(W) compact has not been used there, the present theorem avoids the notion of ``exceptional levels'' introduced in that paper. In the meantime generalizations of the results of Deift-Hempel and the above theorem appeared in a preprint by \textit{S. Alama}, \textit{P. Deift} and \textit{R. Hempel} [Eigenvalue branches of the Schrödinger operator \(H-\lambda W\) in a gap of \(\sigma(H)\), New York Univ. (1988)].