Almost complete solution for the NP-hard separability problem of Bell diagonal qutrits

TL;DR

This paper presents a nearly complete solution with 95% success probability for distinguishing separable from entangled Bell diagonal qutrit states with positive partial transposition, using novel geometric classification methods.

Contribution

It introduces a geometric approach to classify Bell diagonal qutrit states, achieving high accuracy in identifying separable and bound entangled states, and quantifies their relative volumes.

Findings

81.0% of PPT states are separable

13.9% are bound entangled

5.1% remain unclassified

Abstract

With a probability of success of $95 \%$ we solve the separability problem for Bell diagonal qutrit states with positive partial transposition (PPT). The separability problem, i.e. distinguishing separable and entangled states, generally lacks an efficient solution due to the existence of bound entangled states. In contrast to free entangled states that can be used for entanglement distillation via local operations and classical communication, these states cannot be detected by the Peres-Horodecki criterion or PPT criterion. We analyze a large family of bipartite qutrit states that can be separable, free entangled or bound entangled. Leveraging a geometrical representation of these states in Euclidean space, novel methods are presented that allow the classification of separable and bound entangled Bell diagonal states in an efficient way. Moreover, the classification allows the precise determination of relative volumes of the classes of separable, free and bound entangled states. In detail, out of all Bell diagonal PPT states $81.0 \%\pm0.1\%$ are determined to be separable while $13.9\pm0.1\%$ are bound entangled and only $5.1\pm0.1\%$ remain unclassified. Moreover, our applied criteria are compared for their effectiveness and relation as detectors of bound entanglement, which reveals that not a single criterion is capable to detect all bound entangled states.