Two-copy distillability of one-copy-undistillable negative-partial-transpose states in every dimension

A central problem in quantum information theory is whether negative-partial-transpose (NPT) states that cannot be distilled into pure entanglement using a single copy (one-copy-undistillable) can become distillable when multiple copies are available. The authors investigate this within the canonical two-parameter family of states introduced by DiVincenzo et al., which serves as a symmetry-reduced testbed for the NPT-distillability problem.

The study focuses on a distinguished state (point C) within the canonical family. The researchers employ a uniform equal-norm tight-frame construction to derive explicit Schmidt-rank-two certificates. By analyzing the expectation values of these certificates across two copies, they determine whether the state violates the conditions for undistillability. They further extend this analysis to three-copy witnesses to explore the surrounding parameter space and establish inner bounds for the distillable region.

The authors demonstrate that point C is two-copy distillable in every dimension d >= 3, despite being one-copy undistillable. This result serves as a counterexample to the long-standing conjecture that the entire one-copy-undistillable region of the canonical family remains undistillable for any finite number of copies. The study reveals that the canonical family contains states with opposite two-copy behavior, separated by a region where finite-copy distillability remains an open question. The authors also provide analytic inner bounds for the distillable region, showing that the two-copy distillability extends to an open neighborhood around the counterexample.

This work provides a definitive answer to a fundamental question regarding the limits of entanglement distillation. By proving that one-copy-undistillable states can become distillable with just two copies, the paper clarifies the complex geometry of NPT states and highlights that finite-copy distillability does not necessarily follow the same patterns as one-copy distillability. The findings offer a concrete, dimension-independent counterexample that refines our understanding of bound entanglement and provides a new benchmark for future studies in quantum information theory.