arXiv:2610.15664
Separability criteria and lower bounds of entanglement via realignment
Yuhang Xie, Yanjun Chu, Shao-Ming Fei
Abstract
The detection and quantification of quantum entanglement are central problems in quantum information science. Based on the realignment of density matrices and the vectorization of reduced density matrices, we construct a family of enhanced separability criteria for bipartite and multipartite quantum systems. Our criteria recover the previously proposed realignment-based criteria as special cases and detect more entanglement than the earlier ones. Moreover, we derive tighter lower bounds for the concurrence and the convex-roof extended negativity, criteria for detecting the genuine tripartite entanglement, and lower bounds on genuine-multipartite-entanglement concurrence. The advantages of our results are demonstrated with detailed examples.
AI Summary
The paper addresses entanglement detection and computable lower bounds on convex-roof entanglement measures using extensions of the realignment criterion. It introduces a block matrix $\mathcal Q_{P,Q,\mu,\nu}(\rho)$ containing $\mathcal R(\rho)$, vectorized reduced states, and an auxiliary block $PQ^T$ of arbitrary rank, with $\nu^T\mu=1$. Rank-one choices recover several existing realignment-based constructions. For separable bipartite states, it establishes $\|\mathcal Q(\rho)\|_{\mathrm{tr}}\leq\mathfrak M$, where $\mathfrak M=\|(P^TP+\nu\nu^T)^{1/2}(Q^TQ+\mu\mu^T)^{1/2}\|_{\mathrm{tr}}$. Writing $\Delta=\|\mathcal Q(\rho)\|_{\mathrm{tr}}-\mathfrak M$ and $d=\min(d_A,d_B)$ gives $C(\rho)\geq\sqrt{2/[d(d-1)]}\,\Delta$ and normalized CREN $N(\rho)\geq\Delta/(d-1)$. The proofs combine product-state matrix factorizations, trace-norm convexity, Schmidt-decomposition estimates, and convex-roof arguments. For three equal-dimensional subsystems, the average block-matrix norm over the three bipartitions is at most $\mathfrak M+2(d-1)/3$ for biseparable states; its excess also lower-bounds GME concurrence after division by $\sqrt{d(d-1)}$. A separate tensor-product construction provides full-separability criteria for general multipartite systems. Examples compare selected parameter settings: one bipartite detection threshold changes from $x=0.223267$ to $0.223231$, while a noisy three-qutrit W-state GME threshold changes from $q=0.805132$ to $0.804647$. Parameters are selected numerically. Analytic optimal choices and a parameter-independent optimized criterion remain open; measurement-based implementations and extensions to finer multipartite separability classes are proposed as future directions.
Error found: The general multipartite map in Eq. (13) depends on the chosen tensor-product decomposition, and an allowed rescaling of factors makes a fully separable product state violate Theorem 5's stated bound. The numerical examples demonstrate improvements over selected earlier parameter choices, not strict improvements over the earlier parameterized families after optimization. Theorem 1's proof incorrectly applies ordinary-transpose Gram identities to complex matrices, although Hermitian adjoints and a conjugated factor repair the proof without changing the theorem.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 17 pages, 3 figures