Susceptibility of qudit-qudit entanglement to quantum noise: insights from the negativity rank

Technical Summary: Susceptibility of Qudit-Qudit Entanglement to Quantum Noise: Insights from the Negativity Rank

Problem Statement
Quantum entanglement is a fundamental resource for quantum computation and many-body physics, yet it is fragile against environmental interactions that introduce quantum noise. While entanglement entropy suffices for pure states, detecting and quantifying entanglement in mixed ensembles is computationally intractable for general qudit-qudit systems. Existing methods often rely on Positive but Not Completely Positive (PnCP) maps, such as the partial transpose (PT), to witness entanglement. However, the relationship between the structural properties of entanglement—specifically the "negativity rank" (the number of negative eigenvalues in the partially transposed density matrix, $\rho^\Gamma$)—and the robustness of that entanglement against various noise channels remains underexplored. This paper investigates how the negativity rank governs the susceptibility of qudit-qudit entanglement to depolarizing noise and statewise mixtures.

Methodology
The authors employ a combination of analytic derivations, spectral analysis, and constructive proofs to explore the operational implications of the negativity rank ($n_-$).

1. Depolarizing Noise Analysis: The study models global depolarizing noise as a linear map shrinking the Bloch vector. By analyzing the recursive evolution of the negativity spectrum under this channel, the authors derive the decay rate of negativity as a function of $n_-$.
2. Geometric and Convex Analysis: The paper utilizes the geometry of the Bloch sphere and the convexity of the Absolutely Positive Partial Transpose (APPT) subspace. It identifies extremal points and derives boundaries for the APPT subspace, particularly for qutrit-qutrit systems.
3. Spectral and Inertia Analysis: To determine robustness against mixing, the authors analyze the eigenvalues of partially transposed density matrices. They employ the interlacing theorem and inertia analysis (specifically regarding rank-1 updates) to determine how many separable or entangled states can be mixed into a pure NPT (Negative Partial Transpose) state before the ensemble becomes PPT.
4. Constructive Proofs: The paper constructs specific subspaces and mixing strategies (e.g., using generalized Bell bases) to demonstrate tight bounds on the number of states that can destroy entanglement.

Key Contributions and Results

• Negativity Decay Rate: The authors establish that the rate of negativity decay under global depolarizing noise is governed by the negativity rank $n_-$. Higher negativity ranks lead to more rapid decay. Specifically, for a state with $n_-$ negative eigenvalues, the negativity decreases by $n_-/d^2$ per unit of noise probability $p$ (until the rank changes).
• APPT Boundary Derivation: Leveraging the decay rate, the paper provides an intuitive derivation of the APPT boundary for ensembles radially connected to pure states on the Bloch surface. It confirms that the boundary is determined by the state with the minimal negativity rank ($n_-=1$), corresponding to the embedded Bell state.
• Maximum Purity of Qutrit-Qutrit APPT: The paper analytically derives the maximum purity for a qutrit-qutrit APPT ensemble, finding $P_{max}^{APPT} = 17/121$. This result:
  • Confirms previous numerical evaluations.
  • Resolves Conjecture 6.7 from Ref. [56] for the qutrit-qutrit case.
  • Coincides with the maximum purity of an Absolutely Separable (ASEP) ensemble in this dimension, offering evidence (though not proof) that ASEP and APPT subspaces may be equivalent for $d=3$.
  • Provides a computationally amenable constraint for semidefinite programs (SDPs) searching for APPT entangled states.
• Maximal NPT Subspaces: The paper reviews and details the proof that the maximum negativity rank for a $d_A \times d_B$ system is $(d_A-1)(d_B-1)$. It constructs a specific maximal NPT subspace $S$ and discusses how noise channels can increase the negativity rank (unlike the negativity itself, which is monotonic under CPTP maps), noting that while global depolarization decreases rank, specific local channels can increase it.
• Robustness to Statewise Mixture:
  • Separable Mixtures: The authors prove a tight bound on the number of arbitrary separable states that can be mixed into an entangled pure state $|\psi_0\rangle$ while retaining NPT classification. The maximum number $N$ is exactly equal to the negativity rank of the initial state: $N = n_- = s(s-1)/2$, where $s$ is the Schmidt rank. This extends previous lower bounds by one.
  • Entangled Mixtures: For mixtures of entangled states, the paper shows that $N = s-1$ entangled states are sufficient to eliminate entanglement (rendering the mixture PPT), and this bound is tight.

Significance and Claims
The paper claims that the negativity rank is a critical structural parameter that dictates the resilience of entanglement against specific noise mechanisms.

• Operational Insight: The negativity rank provides a direct link between the spectral properties of a state and its practical durability in noisy environments. High-rank states are more fragile to global depolarization but more robust against the addition of separable noise.
• Classification of Entanglement: By deriving the maximum purity of APPT states and analyzing the boundaries of ASEP and APPT subspaces, the work contributes to the ongoing effort to classify mixed-state entanglement, particularly in dimensions where PPT is not sufficient for separability.
• Practical Implications: The results offer guidance for quantum simulation and computation. For instance, in lattice gauge theory simulations where negativity rank increases toward the continuum limit, understanding these decay rates informs decisions on noise mitigation. Furthermore, the derived purity bounds serve as constraints for numerical searches of APPT entangled states, potentially helping to distinguish between ASEP and APPT subspaces in higher dimensions.

The authors maintain a modest tone regarding the equivalence of ASEP and APPT subspaces in higher dimensions, noting that while their results support the possibility for qutrit-qutrit systems, a definitive proof remains elusive. Similarly, they clarify that while the negativity rank can be increased by specific noise channels, the negativity measure itself remains monotonic under CPTP maps.