---
title: Bipartite Concurrence
url: https://www.emergentmind.com/topics/bipartite-concurrence
type: topic
---

# Bipartite Concurrence

Bipartite concurrence is a central quantitative measure of entanglement for bipartite quantum systems of arbitrary finite local dimension. It is defined for both pure and mixed quantum states and serves as a convex-roof entanglement monotone, capturing the degree of nonseparability of a state. Originally motivated by the need for computable certification of entanglement beyond separability criteria—such as the positive partial transpose (PPT) or realignment criteria—bipartite concurrence has become a fundamental tool in quantum information theory, with critical analytical lower bounds playing a key role in entanglement detection and quantification. This article organizes the theory and methodology underlying bipartite concurrence, including its standard definition, the construction of analytical lower bounds via positive maps, comparative performance with other entanglement criteria, explicit examples, and limitations and open problems [1104.1000].

## 1. Definition and Convex-Roof Extension

For a bipartite quantum system with Hilbert space \(\mathcal{H}_A \otimes \mathcal{H}_B\) and local dimension \(N\), the concurrence for a pure state \(|\psi\rangle\) with Schmidt decomposition
\[
|\psi\rangle = \sum_{i=1}^N \alpha_i\,|e_i^A\rangle \otimes |e_i^B\rangle,\quad \sum_i \alpha_i^2 = 1,
\]
is defined as
\[
C(|\psi\rangle) = \sqrt{2\left(1 - \mathrm{Tr} \rho_A^2\right)} = 2 \sqrt{\sum_{i<j} \alpha_i^2 \alpha_j^2},
\]
where \(\rho_A = \mathrm{Tr}_B |\psi\rangle \langle \psi|\) is the reduced density operator. For a mixed state \(\rho\) on \(\mathcal{H}_A \otimes \mathcal{H}_B\), the concurrence is defined via the convex-roof procedure:
\[
C(\rho) = \min_{\{p_k, |\psi_k\rangle\}} \sum_k p_k\,C(|\psi_k\rangle), \quad \rho = \sum_k p_k |\psi_k\rangle \langle\psi_k|,
\]
where the minimum is taken over all pure-state decompositions of \(\rho\) [1104.1000].

## 2. Analytical Lower Bound via Positive Maps

A central result in the theory of bipartite concurrence is the construction of explicit analytic lower bounds that can be evaluated efficiently and, in certain regimes, outperform those derived from other separability criteria. The key ingredient in this approach is the use of a positive (but not completely positive) linear map \(\Phi: M_N \to M_N\). The map \(\Phi\) is defined by its action on the entries \(a_{ij}\) of any \(N \times N\) matrix \(A\):
- For \(i \neq j\):
  \[
  [\Phi(A)]_{ij} = -a_{ij}
  \]
- For the diagonal elements:
  \[
  [\Phi(A)]_{ii} = (N-2)a_{ii} + a_{i',i'}, \qquad i' = (i \ \mathrm{mod}\ N) + 1
  \]

This map is positive but fails to be completely positive—it witnesses entanglement [1104.1000].

For any density operator \(\rho\), the lower bound is constructed via the convex functional:
\[
f(\rho) = \|(I_N \otimes \Phi)\,\rho\|_1 - (N-1),
\]
where \(\|\cdot\|_1\) is the trace norm. For all pure states \(|\psi\rangle\) with Schmidt coefficients \(\{\alpha_i\}\), it is shown that
\[
f(|\psi\rangle\langle\psi|) \le 2 \sum_{i<j} \alpha_i \alpha_j
\]
[1104.1000].

Invoking a sharp inequality from Breuer's framework, the principal analytic lower bound for arbitrary mixed or pure states becomes:
\[
C(\rho) \ge \sqrt{\frac{2}{N(N-1)} f(\rho)} = \sqrt{\frac{2}{N(N-1)}\left(\|(I \otimes \Phi)\rho\|_1 - (N-1)\right)},
\]
or, introducing \(\mathcal{B}(\rho)\),
\[
\mathcal{B}(\rho) = \sqrt{\frac{2}{N(N-1)}\left(\|(I \otimes \Phi)\rho\|_1 - (N-1)\right)},
\]
where the bound is stated as
\[
C(\rho) \ge \max\{ 0, \mathcal{B}(\rho) \}
\]
[1104.1000].

## 3. Comparison with PPT, Realignment, and Other Bounds

Analytic lower bounds for concurrence based on positive maps, PPT, realignment, and entanglement witnesses can be written in a unified form:
\[
C(\rho) \ge \sqrt{\frac{2}{N(N-1)} f_{\mathrm{old}}(\rho)},
\]
where:
- PPT criterion: \(f_{\mathrm{ppt}}(\rho) = \|\rho^{T_A}\|_1 - 1\)
- Realignment criterion: \(f_r(\rho) = \|\tilde{\rho}\|_1 - 1\), with \(\tilde{\rho}\) the realigned matrix
- Breuer's witness criterion: \(f_W(\rho) = -\mathrm{Tr}(W \rho)\), where \(W\) is an entanglement witness

The positive-map \(\Phi\) bound strictly improves upon these in certain regimes, particularly on classes of states where standard separability criteria are less sensitive [1104.1000].

| Bound Type             | Formula for \(f(\rho)\)                          | Offset   |
|------------------------|---------------------------------------------------|----------|
| Positive map \(\Phi\)  | \(\|(I \otimes \Phi)\rho\|_1\)                   | \(N-1\)  |
| PPT                    | \(\|\rho^{T_A}\|_1\)                             | \(1\)    |
| Realignment            | \(\|\tilde{\rho}\|_1\)                           | \(1\)    |
| Breuer's witness       | \(-\mathrm{Tr}(W \rho)\)                         | \(0\)    |

## 4. Explicit Example: Four-Qubit “Hou-State”

Consider the specific \(4\times4\) density matrix family (“Hou-state”):
\[
\rho = \frac{1}{4}\mathrm{diag}(q_1, q_4, q_3, q_2, q_2, q_1, q_4, q_3,\ldots)
+ \frac{q_1}{4}\sum_{i \neq j \in \{1,6,11,16\}} E_{ij},
\]
with \(q_i \ge 0\) and \(\sum_i q_i = 1\). The eigenvalues of \((I_4 \otimes \Phi)\rho\) can be computed in closed form. The positive-map bound yields:
\[
C(\rho) \ge \sqrt{\frac{1}{6}\left( \|(I_4 \otimes \Phi)\rho\|_1 - 3 \right)} = \frac{1}{4\sqrt{6}}|q_1 - q_4|
\]
[1104.1000].

For comparison:
- PPT bound: \(C_{\text{PPT}}(\rho) \approx \frac{1}{2\sqrt{6}}(2q_1 + |q_1-q_3| + \dots - 1)\)
- Realignment bound: \(C_r(\rho) \approx \sqrt{\frac{1}{6}(3q_1 + \frac{1}{4}\sqrt{(q_1-q_2+q_3-q_4)^2} + \dots - 3)}\)
- Breuer's witness: \(C_W(\rho) = -\frac{1}{2\sqrt{6}}(q_2 + 2q_3 + q_4) \leq 0\)

Further specialization to \(q_2 = 0.5\), \(q_4 = 0.01\), \(q_1 = (1-q_2-q_4)\sin^2\theta\), \(q_3 = (1-q_2-q_4)\cos^2\theta\), numerical evaluation reveals:
- The positive-map bound is positive for \(\theta > 0.143\)
- The PPT bound for \(\theta > 0.390\)
- The realignment bound for \(\theta > 0.613\)
- The witness never detects entanglement in this family

Thus, the positive map bound demonstrates concrete superiority in detection ability for specific state families [1104.1000].

## 5. Relation to Pure-State Tightness and Limiting Cases

For pure states, the positive-map lower bound becomes exact (i.e., saturates the true concurrence) if
\[
\|(I \otimes \Phi)|\psi\rangle \langle \psi|\|_1 - (N-1) = 2 \sum_{i<j} \alpha_i \alpha_j
\]
holds for the particular Schmidt vector. Not all pure states realize this equality—identifying which do is an open problem. In general, as with all convex-roof lower bounds constructed from separability criteria, the gap between the bound and the true concurrence remains nontrivial except for these saturating cases [1104.1000].

## 6. Limitations, Parameter-Tuning, and Open Research Directions

While the positive-map approach outperforms PPT, realignment, and witness-based lower bounds for significant families of states, several limitations persist:
- The bound is generally not tight for arbitrary mixed states.
- The positive map \(\Phi\) is not unique; optimization over alternative positive maps or convex combinations thereof may further sharpen the lower bound.
- Characterization of the full class of pure and mixed states for which the bound is tight remains incomplete.
- Efficient computation of \(\|(I \otimes \Phi)\rho\|_1\) may be challenging in large dimensions, although analytically tractable for highly structured states.

Open questions include the systematic construction and classification of positive maps delivering the optimal analytic bound for arbitrary dimensions, and exploration of extremal points in the convex hull of positive but not completely positive maps [1104.1000].

---

In summary, bipartite concurrence serves both as a quantitative entanglement monotone and as a diagnostic criterion, with analytic lower bounds derived via positive maps supplying a practical and, in certain nontrivial cases, maximally sensitive tool for entanglement detection and certification. The positive map framework underlying these bounds is both conceptually and practically superior to PPT, realignment, and witness-based criteria for targeted families of quantum states and remains an active area of research for further optimization and generalization [1104.1000].

Source: https://www.emergentmind.com/topics/bipartite-concurrence