Complete positivity violation of the reduced dynamics in higher-order quantum adiabatic elimination

TL;DR

This paper investigates the limitations of quantum adiabatic elimination, revealing that higher-order corrections can violate complete positivity in reduced dynamics, challenging previous assumptions and conjectures.

Contribution

It demonstrates that higher-order terms in quantum adiabatic elimination can break complete positivity, refuting a conjecture that parametrization could always restore it.

Findings

Higher-order contributions can violate complete positivity.

Complete positivity cannot always be restored by parametrization.

Violations occur even without truncating the asymptotic expansion.

Abstract

This paper discusses quantum adiabatic elimination, which is a model reduction technique for a composite Lindblad system consisting of a fast decaying sub-system coupled to another sub-system with a much slower timescale. Such a system features an invariant manifold that is close to the slow sub-system. This invariant manifold is reached subsequent to the decay of the fast degrees of freedom, after which the slow dynamics follow on it. By parametrizing invariant manifold, the slow dynamics can be simulated via a reduced model. To find the evolution of the reduced state, we perform the asymptotic expansion with respect to the timescale separation. So far, the second-order expansion has mostly been considered. It has then been revealed that the second-order expansion of the reduced dynamics is generally given by a Lindblad equation, which ensures complete positivity of the time evolution. In this paper, we present two examples where complete positivity of the reduced dynamics is violated with higher-order contributions. In the first example, the violation is detected for the evolution of the partial trace without truncation of the asymptotic expansion. The partial trace is not the only way to parametrize the slow dynamics. Concerning this non-uniqueness, it was conjectured in [R. Azouit, F. Chittaro, A. Sarlette, and P. Rouchon, Quantum Sci. Technol. 2, 044011 (2017)] that there exists a parameter choice ensuring complete positivity. With the second example, however, we refute this conjecture by showing that complete positivity cannot be restored in any choice of parametrization. We discuss these results in terms of unavoidable correlations, in the initial states on the invariant slow manifold, between the fast and the slow degrees of freedom.