Finite energy subspace for time-periodic Schr\"odinger operators

TL;DR

This paper establishes the existence of channel wave operators for N-body Schrödinger operators with time-periodic short-range potentials and introduces a finite energy subspace concept, proving their equivalence with wave operator subspaces for two-body systems.

Contribution

It defines a finite energy subspace for time-periodic Schrödinger operators and proves its equivalence with the wave operator subspace in two-body systems, extending understanding of asymptotic completeness.

Findings

All channel wave operators exist for N-body systems.

For N=2, asymptotic completeness is proven using a simplified method.

The finite energy subspace coincides with the wave operator subspace for N=2.

Abstract

For $N$-body Schr\"o\-dinger operators with time-periodic short-range pair-potentials we show by a time-dependent commutator method that all channel wave operators exist. For $N=2$ we prove asymptotic completeness by a simplified version of the method, recovering Yajima's completeness result \cite{Yaj1} proven by a stationary method. We propose a definition of a \emph{finite energy subspace}, intuitively consisting of states with `finite asymptotic energy'. This geometric notion is used to characterize the \emph{wave operator subspace} given as the direct sum of the ranges of the channel wave operators. Thus our main result states that the two subspaces coincide. In turn they coincide for $N= 2$ with the orthogonal subspace of the pure point subspace of the monodromy operator (by asymptotic completeness). For $N\geq 3$ asymptotic completeness for time-periodic short-range potentials remains an open problem. The main result of the paper may in this case potentially serve as an intermediate step for proving (or possibly disproving) asymptotic completeness. Another potential ingredient could be a key intermediate result of the paper, asserting that the states orthogonal to the pure point subspace obey a (sharp) \emph{minimal velocity bound}. Thus it remains an open problem possibly to derive a good maximal velocity bound for $N\geq 3$. Our results apply to $N$-body systems of particles in a time-periodic external electric field with time-mean equal to zero, for example the AC-Stark model.