A Note on Lower Bounds of Concurrence for Arbitrary Dimensional Bipartite Quantum States
Zhi-Bo Chen, Shao-Ming Fei · 2025
Read A Note on Lower Bounds of Concurrence for Arbitrary Dimensional Bipartite Quantum States by Zhi-Bo Chen, Shao-Ming Fei (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-16291.txt.
Introduction
Zhi-Bo Chen and Shao-Ming Fei give new lower bounds for a measure of quantum entanglement called concurrence. Entanglement links two parts of a quantum system in ways that have no classic match. Concurrence turns that link into a number, but the exact value is hard to get for large systems. Bounds that can be computed from data are thus prized.
Catalog record
| CREATOR | Zhi-Bo Chen, Shao-Ming Fei |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 15 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-16291
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Note on Lower Bounds of Concurrence for Arbitrary Dimensional Bipartite Quantum States Zhi-Bo Chen, Shao-Ming Fei Abstract Quantification of quantum entanglement plays a crucial role in the study of quantum information tasks. We present analytical lower bounds for both concurrence and 2-concurrence based on the correlation matrices of bipartite quantum states. Compared with other related lower bounds, our approach provides a better estimation of the entanglement, particularly for states with large purity. Keywords: concurrence, entanglement measure, q-concurrence Quantum entanglement is a remarkable property of quantum systems, serving as fundamental physical resources in quantum information processing, including quantum cryptography, quantum teleportation and quantum computation QCQI . The resource theory of quantum entanglement has been well established with the LOCC (local operation and classical communication) IE as the free operations and the separable states as the free states. To quantify the entanglement, many entanglement measures have been presented. However, most entanglement measures have no explicit analytical formulas for general quantum mixed states, due to the optimization involved in the definition with convex roof extension. An important entanglement measure is the concurrence Intro Concurrence ; af . For a bipartite pure state \ket{\psi}\in\mathbb{C}^{d_{1}}\otimes\mathbb{C}^{d_{2}} in systems A and B , the concurrence of \ket{\psi} is defined by C(\ket{\psi}):=\sqrt{2(1-Tr({\rho_{A}}^{2}))} , where \rho_{A}=Tr_{B}\ket{\psi}\bra{\psi} is the reduced state of subsystem A obtained by tracing over the subsystem B . The concurrence is then extended to mixed states \rho by the convex roof extension, C(\rho)=\underset{\{P_{i},\ \ket{\psi}\}}{min}\sum_{i}P_{i}C(\ket{\psi_{i}}), where the minimization goes over all possible pure state decompositions of \rho=\sum_{i}P_{i}\ket{\psi_{i}}\bra{\psi_{i}} QCQI . Generally it is a challenge to calculate exactly the concurrence for general mixed states. Many lower bounds of concurrence have been presented Concurrence PRL ; Concurrence PRA ; Concurrence old ; Concurrence KF ; Concurrence 1 ; Concurrence QIP ; Concurrence 2 ; Concurrence 3 . Besides, the lower bounds of q -concurrence q-Concurrence have been also studied, where the q -concurrence of a pure state is defined by C_{q}(\ket{\psi})=1-Tr({\rho_{A}}^{q}) , q\geq 2 . The q -concurrence of mixed states is defined similarly by the convex roof extension. In this paper, we present new lower bounds of concurrence and 2-concurrence based on the correlation matrices of bipartite quantum states. For an arbitrary bipartite quantum state \rho\in\mathbb{C}^{d_{1}}\otimes\mathbb{C}^{d_{2}} , \rho has a generalized Bloch representation KF1 ; KF2 , \displaystyle\rho=\frac{1}{d_{1}d_{2}}\left(I_{d_{1}d_{2}}+\sum_{i,j\ \mbox{\scriptsize{are not all 0} }}u_{ij}\ \mathaccent 869{\lambda_{i}}^{(1)}\otimes\mathaccent 869{\lambda_{j}}^{(2)}\right), (1) where \mathaccent 869{\lambda_{0}}^{(1)}=\sqrt{d_{1}-1}I_{d_{1}} , \mathaccent 869{\lambda_{0}}^{(2)}=\sqrt{d_{2}-1}I_{d_{2}} , \mathaccent 869{\lambda_{i}}^{(1)}=\sqrt{\frac{d_{1}(d_{1}-1)}{2}}\lambda_{i}^{(1)} , \mathaccent 869{\lambda_{j}}^{(2)}=\sqrt{\frac{d_{2}(d_{2}-1)}{2}}\lambda_{j}^{(2)} , \left\{\lambda_{i}^{(1)}\right\}_{1\leq i\leq{d_{1}}^{2}-1} and \left\{\lambda_{j}^{(2)}\right\}_{1\leq i\leq{d_{2}}^{2}-1} are generators of SU(d_{1}) and SU(d_{2}) , respectively, satisfying Tr(\lambda_{i}\lambda_{j})=2\delta_{ij} and Tr(\lambda_{i})=0 , u_{ij}=\frac{1}{(d_{1}-1)(d_{2}-1)}Tr\left(\rho\cdot\mathaccent 869{\lambda_{i}}^{(1)}\otimes\mathaccent 869{\lambda_{j}}^{(2)}\right) . Eq.( 1 ) can be further rewritten as \displaystyle\rho=\frac{1}{d_{1}d_{2}}\left(I_{d_{1}d_{2}}+\sum_{i=1}^{{d_{1}}^{2}-1}r_{i}\ \mathaccent 869{\lambda_{i}}^{(1)}\otimes\mathaccent 869{\lambda_{0}}^{(2)}+\sum_{j=1}^{{d_{2}}^{2}-1}s_{i}\ \mathaccent 869{\lambda_{0}}^{(1)}\otimes\mathaccent 869{\lambda_{j}}^{(2)}+\sum_{i,j\neq 0}t_{ij}\ \mathaccent 869{\lambda_{i}}^{(1)}\otimes\mathaccent 869{\lambda_{j}}^{(2)}\right). (2) Denote R=(r_{i}) and S=(s_{j}) the real vectors, and T=(t_{ij}) the ({d_{1}}^{2}-1)\times({d_{2}}^{2}-1) correlation matrix. Set \bm{u}=(u_{ij}) a real vector, called the vector representation of \rho . Then \bm{u}=(R,S,vec(T)) is the vector representation of \rho , where vec(T) is vectorization of T . For a matrix A=a_{ij} , the Frobenius norm \left\|A\right\|_{F}=\sqrt{Tr(AA^{\dagger})} . It is verified that {\left\|A\right\|_{F}}^{2}=\sum_{i,j}|a_{ij}|^{2}=|vec(A)|^{2} . For the vector representation \bm{u}=(R,S,vec(T)) , we have \displaystyle|\bm{u}|^{2}=\sum_{i,j\ \mbox{\scriptsize{are not all 0} }}{u_{ij}}^{2}=|R|^{2}+|S|^{2}+{\left\|T\right\|_{F}}^{2}. (3) Furthermore, it is straightforward to verify that \left\|T\right\|_{F}=\sqrt{\left\|TT^{\dagger}\right\|_{tr}} , where \left\|A\right\|_{tr}=Tr(\sqrt{AA^{\dagger}}) is trace norm, also known as Ky-Fan norm. It is easily verified that \left\|T\right\|_{F}\leq\left\|T\right\|_{tr} . Based on \left\|T\right\|_{tr} , necessary separability conditions and lower bounds of concurrence Concurrence KF have been derived. A quantum state is entangled when \left\|T\right\|_{tr} is sufficiently large KF1 ; KF2 ; KF3 ; KF4 . Interestingly, the norm \left\|T\right\|_{F} can also gives rise to lower bounds of concurrence. We first prove that the 2-norm of the vector representation of \rho and the purity of \rho are equivalent. Lemma 1 For bipartite state \rho\in\mathbb{C}^{d_{1}}\otimes\mathbb{C}^{d_{2}} given by ( 2 ), we have \begin{array}[]{rcl}Tr(\rho^{2})&=&\displaystyle\frac{1}{d_{1}d_{2}}\left(1+(d_{1}-1)(d_{2}-1)|\bm{u}|^{2}\right),\\[8.53581pt] Tr(\rho_{A}^{2})&=&\displaystyle\frac{1}{d_{1}}\left(1+(d_{1}-1)(d_{2}-1)\cdot|R|^{2}\right),\\[8.53581pt] Tr(\rho_{B}^{2})&=&\displaystyle\frac{1}{d_{2}}\left(1+(d_{1}-1)(d_{2}-1)\cdot|S|^{2}\right).\end{array} Proof By ( 1 ) we have \displaystyle Tr(\rho^{2}) \displaystyle=\frac{1}{(d_{1}d_{2})^{2}}\left(d_{1}d_{2}+d_{1}d_{2}(d_{1}-1)(d_{2}-1)\sum_{i,j}\sum_{k,l}u_{ij}u_{kl}\delta_{ik}\delta_{jl}\right) \displaystyle=\frac{1}{(d_{1}d_{2})^{2}}\left(d_{1}d_{2}+d_{1}d_{2}(d_{1}-1)(d_{2}-1)\sum_{i,j}{u_{ij}}^{2}\right) \displaystyle=\frac{1}{d_{1}d_{2}}\left(1+(d_{1}-1)(d_{2}-1)|\bm{u}|^{2}\right). By ( 2 ) we have \rho_{A}=\frac{1}{d_{1}}\left(I_{d_{1}}+\sqrt{d_{2}-1}\sum_{i=1}^{{d_{1}}^{2}-1}r_{i}\ \mathaccent 869{\lambda_{i}}^{(1)}\right), then \displaystyle Tr({\rho_{A}}^{2}) \displaystyle=\frac{1}{{d_{1}}^{2}}\left(Tr({I_{d_{1}}}^{2})+(d_{2}-1)\sum_{i=1}^{{d_{1}}^{2}-1}\sum_{j=1}^{{d_{1}}^{2}-1}r_{i}r_{j}Tr\left(\mathaccent 869{\lambda_{i}}^{(1)}\mathaccent 869{\lambda_{j}}^{(1)}\right)+2\sum_{i=1}^{{d_{1}}^{2}-1}Tr(\mathaccent 869{\lambda_{i}}^{(1)})\right) \displaystyle=\frac{1}{d_{1}}\left(1+(d_{1}-1)(d_{2}-1)\sum_{i=1}^{{d_{1}}^{2}-1}{r_{i}}^{2}\right) \displaystyle=\frac{1}{d_{1}}\left(1+(d_{1}-1)(d_{2}-1)\cdot|S|^{2}\right). Tr({\rho_{B}}^{2}) can be similarly derived. \Box Based on the generalized Bloch representation, the concurrence of a pure state is given by the Frobenius norm of the correlation matrix. Theorem 1 For any bipartite pure state \ket{\psi}\in\mathbb{C}^{d_{1}}\otimes\mathbb{C}^{d_{2}} , we have \displaystyle C(\ket{\psi}) \displaystyle=\sqrt{\frac{2(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|T\right\|_{F}}^{2}-1\right)}, \displaystyle C_{2}(\ket{\psi}) \displaystyle=\frac{(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|T\right\|_{F}}^{2}-1\right), where T is correlation matrix of \ket{\psi}\bra{\psi} . Proof For a pure state \rho=\ket{\psi}\bra{\psi} , from Lemma 1 and Eq.( 3 ) we have \displaystyle|R|^{2}+|S|^{2}+{\left\|T\right\|_{F}}^{2}=\frac{d_{1}d_{2}-1}{(d_{1}-1)(d_{2}-1)}. (4) Since Tr({\rho_{A}}^{2})=Tr({\rho_{B}}^{2}) for pure state QCQI , by Lemma 1 we have |R|^{2}=\frac{d_{1}Tr({\rho_{A}}^{2})-1}{(d_{1}-1)(d_{2}-1)},~\ |S|^{2}=\frac{d_{2}Tr({\rho_{A}}^{2})-1}{(d_{1}-1)(d_{2}-1)}. From Eq.( 4 ) we have \displaystyle{\left\|T\right\|_{F}}^{2}-1 \displaystyle=\frac{d_{1}d_{2}-1}{(d_{1}-1)(d_{2}-1)}-|R|^{2}-|S|^{2}-1 \displaystyle=\frac{d_{1}d_{2}-1-(d_{1}Tr({\rho_{A}}^{2})-1)-(d_{2}Tr({\rho_{A}}^{2})-1)-(d_{1}-1)(d_{2}-1)}{(d_{1}-1)(d_{2}-1)} \displaystyle=\frac{d_{1}+d_{2}}{(d_{1}-1)(d_{2}-1)}\left(1-Tr({\rho_{A}}^{2})\right). Therefore, 1-Tr({\rho_{A}}^{2})=\frac{(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|T\right\|_{F}}^{2}-1\right) , which proves the theorem. \Box For general bipartite mixed states, we have the following conclusion. Theorem 2 For any bipartite state \rho\in\mathbb{C}^{d_{1}}\otimes\mathbb{C}^{d_{2}} , we have C(\rho)\geq\frac{\sqrt{2}(\sqrt{K+1}+1)}{K}\left(\left\|T\right\|_{F}-1\right), C_{2}(\rho)\geq\frac{(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|T\right\|_{F}}^{2}-1\right), where T is the correlation matrix of \rho , K=\frac{d_{1}+d_{2}}{(d_{1}-1)(d_{2}-1)} . Proof Since \sqrt{2}>C(\ket{\psi})\geq 0 , by Theorem 1 we have 2>\frac{2(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|T\right\|_{F}}^{2}-1\right)\geq 0 . Hence, the correlation matrix of a pure state always satisfies \displaystyle\sqrt{\frac{d_{1}+d_{2}}{(d_{1}-1)(d_{2}-1)}+1}>\left\|T\right\|_{F}\geq 1. (5) It is easy to verify that if a\geq x\geq 1 , then \sqrt{x^{2}-1}\geq\sqrt{\frac{a+1}{a-1}}(x-1) . Take a=\sqrt{K+1} , where K=\frac{d_{1}+d_{2}}{(d_{1}-1)(d_{2}-1)} , by ( 5 ) we have \displaystyle\sqrt{{\left\|T\right\|_{F}}^{2}-1}\geq\sqrt{\frac{\sqrt{K+1}+1}{\sqrt{K+1}-1}}(\left\|T\right\|_{F}-1). From Theorem 1 we have \displaystyle C(\ket{\psi})=\sqrt{\frac{2}{K}\left({\left\|T\right\|_{F}}^{2}-1\right)}\geq\sqrt{\frac{2}{K}\cdot\frac{\sqrt{K+1}+1}{\sqrt{K+1}-1}}(\left\|T\right\|_{F}-1). (6) Next consider the optimal decomposition \rho=\sum_{i}P_{i}\ket{\varphi_{i}}\bra{\varphi_{i}} for the concurrence. Using ( 6 ) we have \displaystyle C(\rho)=\sum_{i}P_{i}C(\ket{\varphi_{i}}) \displaystyle\geq\sqrt{\frac{2(\sqrt{K+1}+1)}{K(\sqrt{K+1}-1)}}\sum_{i}P_{i}\left(\left\|T_{i}\right\|_{F}-1\right) \displaystyle\geq\sqrt{\frac{2(\sqrt{K+1}+1)}{K(\sqrt{K+1}-1)}}\left(\left\|T\right\|_{F}-1\right)=\frac{\sqrt{2}(\sqrt{K+1}+1)}{K}\left(\left\|T\right\|_{F}-1\right), where T_{i} is correlation matrix of the state \ket{\varphi_{i}}\bra{\varphi_{i}} . For the second inequality in Theorem 2, consider the optimal decomposition \rho=\sum_{i}P_{i}\ket{\psi_{i}}\bra{\psi_{i}} for the 2-concurrence. Using Theorem 1 we have \displaystyle C_{2}(\rho)=\sum_{i}P_{i}C_{2}(\ket{\psi_{i}})=\frac{(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\sum_{i}P_{i}\left({\left\|T_{i}\right\|_{F}}^{2}-1\right), where T_{i} is correlation matrix of the state \ket{\psi_{i}}\bra{\psi_{i}} . Since f(A):={\left\|A\right\|_{F}}^{2} is convex function, we have C_{2}(\rho)\geq\frac{(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|\sum_{i}P_{i}T_{i}\right\|_{F}}^{2}-1\right)=\frac{(d_{1}-1)(d_{2}-1)}{d_{1}+d_{2}}\left({\left\|T\right\|_{F}}^{2}-1\right). \Box The lower bound of the 2-concurrence in Theorem 2 is exact for pure states. One would expect that Theorem 2 gives a better lower bound for states of large purity. Interestingly, this is also true for the lower bound of concurrence in Theorem 2. Example 1 Consider the 4\times 4 state \rho_{x}=x\ket{\phi}\bra{\phi}+\frac{1-x}{16}I_{16} , where \ket{\phi}=\frac{1}{\sqrt{2}}\left(\ket{00}+\ket{33}\right) and 0<x<1 . For this state we have {\left\|T\right\|_{F}}^{2}=\frac{13}{9}x^{2} . By Theorem 2 we obtain C(\rho_{x})\geq\frac{9\sqrt{2}}{8}\left(\frac{\sqrt{17}}{3}+1\right)\left(\frac{\sqrt{13}}{3}x-1\right),~~~~~C_{2}(\rho_{x})\geq\frac{9}{8}\left(\frac{13}{9}x^{2}-1\right). In Concurrence PRL a lower bound of concurrence has been presented for m\times n ( m\leq n ) states, C(\rho)\geq\sqrt{\frac{2}{m(m-1)}}\left(max\left\{\left\|\rho^{T_{A}}\right\|_{tr},\ \left\|\mathcal{R}(\rho)\right\|_{tr}\right\}-1\right)\ , where T_{A} is partial transposition with respect to the first subsystem PPT , \mathcal{R} is realignment operation R . For \rho=\sum_{ijkl}\rho_{ij,kl}\ket{ij}\bra{kl} , \rho^{T_{A}}=\sum_{ijkl}\rho_{ij,kl}\ket{kj}\bra{il} and \mathcal{R}(\rho)=\sum_{ijkl}\rho_{ij,kl}\ket{ik}\bra{jl} . For the state \rho_{x} , one has C(\rho_{x})\geq\frac{1}{\sqrt{6}}\left(\frac{7+9x}{8}-1\right) . When x>0.914 , the lower bound of C(\rho_{x}) given in Theorem 2 is significantly better than the one given in Concurrence PRL , as shown in Fig.1. Figure 1: Red solid line (N) is for the lower bound of concurrence from Theorem 2 of this paper, green dashed line (O) is for that from Concurrence PRL . In q-Concurrence the authors presented a lower bound of 2-concurrence for m\times n ( m\leq n ) states, C_{2}(\rho)\geq\frac{1}{m(m-1)}\left(max\left\{\left\|\rho^{T_{A}}\right\|_{tr},\ \left\|\mathcal{R}(\rho)\right\|_{tr}\right\}-1\right)^{2}. For the state \rho_{x} , one has C_{2}(\rho_{x})\geq\frac{1}{12}\left(\frac{7+9x}{8}-1\right)^{2} . When x>0.854 , the lower bound of C_{2}(\rho_{x}) given in Theorem 2 is significantly better than the one given in q-Concurrence , see Fig.2. Figure 2: Red solid line (N) is for the lower bound of 2-concurrence from Theorem 2 of this paper, green dashed line (O) is for that from the Theorem 1 in q-Concurrence . Example 1 seems to imply that