Local time decay of high energy scattering states for the Schrödinger equation (Q793213)

The main result of this paper is an estimate for the time decay of the time evolution of certain scattering solutions for the Schrödinger equation. \(-id\phi(t)/dt=H\phi(t),\) \(t\in {\mathbb{R}}\) where H is the self- adjoint closure of \(-\Delta +V\) on \(C_ 0^{\infty}({\mathbb{R}}^ n)\), \(V=V_ s+V_ 1\) being the sum of a short range part and a long range part of the potential, i.e. real-valued functions satisfying the conditions \(<x>^ NV_ s(-\Delta +1)^{-\delta}\) is bounded for \(N\in {\mathbb{N}}\) and suitable \(\delta\in [0,frac{1}{2})\) with \(\quad<x>^{2N}V_ s(x)\to 0,\quad | x| \to \infty;\) and \(V_ 1\in C^{(N+2+[n/2])}({\mathbb{R}}^ n)\) such that \(| D^{\alpha}V(x)| \leq C_{\alpha}<x>^{-| \alpha | - \epsilon}\) for \(\epsilon>0\) and any multi-index \(\alpha\) with \(| \alpha | \leq N+1.\) More precisely it is shown that the time evolution operator \(e^{-itH}\) considered as a mapping between the weighted spaces \(L^ 2_ s\to L^ 2_{-s}\quad(L^ 2:=L^ 2({\mathbb{R}}^ n,<x>^{2s}dx)\) decays like \(t^{-s+\eta}\), where \(\eta\) depends on the ''smoothness'' of the long- range part of the potential \(V_{\ell}\) and the ''localization'' weight (i.e. \(\eta =S/N)\), provided one cuts off energies below a suitable high level. Note that this estimate depends not on the dimension n of the configuration space, in contrast to low-energy estimates. The result is obtained by a suitable transformation of some estimates of the power of the resolvent of H near the real axis. These estimates are obtained by using two new techniques. One is the use of a modified complex scaling (i.e. the use of the modified generator \(A=frac{1}{2}(x- {\tilde \sigma}(p)+{\tilde \sigma}(p)\cdot x)\) (instead of the usual dilation generator \(frac{1}{2}(x\cdot p+p\cdot x)\) for defining the dilation family \(\theta \to e^{-\theta A}He^{\theta A}\) which turns the essential spectrum of H into a parabola shaped line if \(\theta\) runs from 0 to a suitable \(\theta_ 0>0\). (\({\tilde \sigma}\) being a suitable bounded function.) The other technique is the use of a finite Taylor approximation of the dilation family for solving certain differential inequalities, an idea which goes back to E. Mourre.