Quantifying mixed-state entanglement via partial transpose and realignment moments

Technical Summary: Quantifying Mixed-State Entanglement via Partial Transpose and Realignment Moments

Problem Statement
Quantifying entanglement in mixed quantum many-body systems remains a notoriously difficult problem, particularly in the context of Noisy Intermediate-Scale Quantum (NISQ) devices where experimental noise renders states mixed. While pure-state entanglement is well-characterized by measures like the entanglement (Rényi) entropy, extending these to mixed states faces significant hurdles. The standard metric, the Partial Transpose (PT) negativity (or logarithmic negativity), requires full diagonalization of the partial transpose, which is exponentially costly. Furthermore, experimental measurement of PT negativity typically necessitates full-state tomography. Previous efforts focused on binary entanglement detection using low-order moments or positive maps, but these lack the quantitative power to characterize the degree of entanglement or entanglement dimensionality in large, noisy systems.

Methodology
The authors introduce families of quantitative entanglement witnesses constructed from the moments of the partial transpose and the realignment matrix. These are defined as:

1. $p_\alpha$-negativity ($E_\alpha$): Derived from the PT moments $p_\alpha = \text{tr}((\rho^\Gamma)^\alpha)$. Specifically, $E_\alpha(\rho) = \frac{1}{2-\alpha}[\ln(\tilde{p}_\alpha) + (\alpha-1)S_2(\rho)]$, where $\tilde{p}_\alpha = \text{tr}(|\rho^\Gamma|^\alpha)$ and $S_2$ is the 2-Rényi entropy.
2. $r_\alpha$-negativity ($C_\alpha$): Derived from the realignment moments $r_\alpha = \text{tr}((R_\rho^\dagger R_\rho)^{\alpha/2})$, where $R_\rho$ is the realignment matrix.

The authors focus on the case $\alpha=4$ ($E_4$ and $C_4$), demonstrating that these quantities can be efficiently measured on quantum hardware using SWAP tests or Bell measurements on a constant number of copies (2 or 4 copies) without full-state tomography. The framework is extended to symmetry-resolved (SR) witnesses by decomposing the density matrix into charge sectors (e.g., under global $U(1)$ or $\mathbb{Z}_2$ symmetries), which tightens the bounds and improves detection capabilities.

The paper also develops efficient classical algorithms to compute these moments for Matrix Product States (MPS) and Matrix Product Operators (MPO), with computational costs scaling polynomially with system size and bond dimension.

Key Contributions and Results

1. Quantitative Witnessing and Bounds:

   • The $p_\alpha$-negativity and $r_\alpha$-negativity serve as quantitative witnesses. For $\alpha \ge 1$, they provide rigorous lower bounds on the PT negativity ($E$) and the CCNR negativity ($C$), respectively.
   • Crucially, they provide lower bounds on the robustness of entanglement and the Schmidt number (entanglement dimensionality).
   • For pure states, these witnesses collapse to the standard entanglement Rényi entropies, effectively generalizing them to mixed states.

2. Efficient Entanglement Testing:

   • The authors present an efficient algorithm to distinguish whether a mixed state has low entanglement ($O(\log n)$) or high entanglement ($\omega(\log n)$), provided the state has bounded 2-Rényi entropy ($S_2 = O(\log n)$). This generalizes previous results known only for pure states.
   • Efficient algorithms are provided to test the Schmidt rank (using 2-copy measurements) and the operator Schmidt rank (using 4-copy measurements).

3. Circuit Depth Certification:

   • The witnesses enable the robust certification of the minimal circuit depth required to prepare a state, even in the presence of noise, assuming the final state is weakly mixed ($S_2 = O(\log n)$). This extends previous noiseless certification protocols.

4. Pseudoentanglement and Cryptography:

   • The paper establishes rigorous limits on pseudoentanglement (the ability of low-entanglement states to mimic high-entanglement states).
   • It is shown that for mixed states with low entropy ($S_2 = O(\log n)$), the pseudoentanglement gap is fundamentally limited to $f(n) = \Theta(n)$ vs. $g(n) = \omega(\log n)$.
   • The results imply that entropy is a necessary resource for securely hiding entanglement information; to achieve the maximal gap ($g(n)=0$), the state must be highly mixed ($S_2 = \omega(\log n)$).

5. Entanglement Phase Diagrams:

   • The authors demonstrate that the entanglement phase diagram of Haar random states (comprising PPT, Maximally Entangled, and Entanglement Saturation phases) can be fully established solely by computing the $p_4$-negativity, bypassing the need for analytic continuation of all PT moments.
   • This result extends to any state 4-design. Surprisingly, stabilizer states (which do not form a 4-design) are shown to share the exact same entanglement phase diagram as Haar random states.
   • The $p_4$-negativity effectively captures volume-law entanglement in Haar random states even under exponentially strong global depolarizing noise.

6. Many-Body Applications:

   • Numerical results on the Transverse Field Ising Model (TFIM) and XXZ model show that the witnesses, particularly the symmetry-resolved versions, effectively detect entanglement in ground states subject to noise and finite temperature.
   • The SR witnesses are shown to be more robust than non-SR witnesses, detecting entanglement in regimes where standard witnesses fail (e.g., disconnected partitions or near critical points).

Significance
The paper claims to open new avenues for quantifying entanglement in large and noisy quantum systems. By introducing $p_\alpha$ and $r_\alpha$ negativities, the authors provide a practical, experimentally accessible alternative to the PT negativity that retains quantitative power. The work bridges the gap between pure-state entanglement theory (via Rényi entropies) and mixed-state physics, offering tools to characterize entanglement dimensionality, certify circuit depth, and understand the complexity of pseudoentanglement. The ability to compute these witnesses efficiently via MPS/MPO algorithms and measure them via SWAP tests makes them directly implementable on current and near-future quantum hardware, addressing the critical need for benchmarking NISQ devices.