The Zero-error Entanglement Cost is Highly Non-Additive

TL;DR

This paper investigates the non-additivity of the zero-error entanglement cost, revealing states with constant Schmidt number across multiple copies and exploring how quantum operations can increase this measure.

Contribution

It demonstrates the high non-multiplicativity of the Schmidt number and provides exact calculations of the regularized zero-error entanglement cost, introducing new classes of quantum operations.

Findings

Existence of states with constant Schmidt number across multiple copies

Exact computation of the regularized zero-error entanglement cost for certain states

Quantum operations that preserve pure state Schmidt number but can increase mixed state Schmidt number arbitrarily

Abstract

The Schmidt number is an entanglement measure whose logarithm quantifies the zero-error entanglement cost of generating a given quantum state using local operations and classical communication (LOCC). %However, the Schmidt number is a notoriously difficult quantity to compute, and its relationship to other entanglement measures is largely unknown. In this paper we show that the Schmidt number is highly non-multiplicative in the sense that for any integer $n$, there exists states whose Schmidt number remains constant when taking $n$ copies of the given state. These states also provide a rare instance in which the regularized zero-error entanglement cost can be computed exactly. We then explore the question of increasing the Schmidt number by quantum operations. We describe a class of bipartite quantum operations that preserve the Schmidt number for pure state transformations, and yet they can increase the Schmidt number by an arbitrarily large amount when generating mixed states. Our results are obtained by making connections to the resource theory of quantum coherence and generalizing the class of dephasing-covariant incoherent operations (DIO) to the bipartite setting.