Absorption capacity of separable noise: Bell-mixing thresholds on separability and teleportation

Technical Summary: Absorption Capacity of Separable Noise

Problem Statement
The paper addresses the dynamics of entanglement and quantum teleportation utility in two-qubit systems subject to noise. Specifically, it investigates the transitions that occur along a "Bell-mixing line," defined as a convex combination of a fixed Bell reference state $|\Phi^+\rangle$ and a separable noise state $\sigma$:
$$\rho_\lambda = \lambda |\Phi^+\rangle\langle\Phi^+| + (1-\lambda)\sigma, \quad \lambda \in [0, 1].$$
Two critical operational thresholds are identified along this line:

1. The Separability Threshold ($\lambda^*$): The smallest $\lambda$ for which $\rho_\lambda$ becomes entangled (equivalent to the state losing Positive Partial Transpose, or PPT, on two qubits).
2. The Teleportation Threshold ($\lambda_F$): The smallest $\lambda$ for which the maximal teleportation fidelity $F(\rho_\lambda)$ exceeds the classical bound of $2/3$.

While these thresholds are well-known for specific cases (e.g., Werner states), the paper seeks a unified framework to determine these values for general separable noise states $\sigma$ and to track them as $\sigma$ evolves under local noise channels.

Methodology
The authors introduce a "fixed-reference" perspective, shifting the focus from how much noise destroys a fixed entangled state to how much Bell reference a fixed separable background can "absorb" before a transition occurs. This is formalized through two absorption capacities:

1. Entanglement Absorption Capacity ($C_{abs}(\sigma)$): Defined as the supremum of a mixing parameter $\eta$ such that the partial transpose of the mixture remains positive:
   $$C_{abs}(\sigma) := \sup \{ \eta \ge 0 : \sigma^{TB} + \eta Q \succeq 0 \},$$
   where $Q = (|\Phi^+\rangle\langle\Phi^+|)^{TB}$ and $\sigma^{TB}$ is the partial transpose of the noise state.
2. Fidelity Absorption Capacity ($C_F(\sigma)$): Defined similarly, but with the stopping condition that the fully entangled fraction $f(\rho_\eta)$ remains at or below $1/2$ (equivalent to $F \le 2/3$):
   $$C_F(\sigma) := \sup \{ \eta \ge 0 : f(\rho_\eta) \le 1/2 \}.$$

The paper establishes a Möbius transform linking these capacities to the operational thresholds:
$$\lambda^*(\sigma) = \frac{C_{abs}(\sigma)}{1 + C_{abs}(\sigma)}, \quad \lambda_F(\sigma) = \frac{C_F(\sigma)}{1 + C_F(\sigma)}.$$

The methodology proceeds by deriving closed-form expressions for these capacities for two structured families of separable noise:

• Product Noise ($\sigma = A \otimes B$): Independent local decoherence.
• Separable X States: States with non-zero entries only on the diagonal and anti-diagonal in the computational basis.

Additionally, the paper utilizes decomposition certificates, leveraging the concavity of the capacities to provide lower bounds for general separable states by decomposing them into solved components (product or X states).

Key Contributions and Results

1. Closed-Form Thresholds for Product Noise:

   • The entanglement absorption capacity depends only on the local marginal purities:
     $$C_{abs}(A \otimes B) = \sqrt{(1 - \text{Tr} A^2)(1 - \text{Tr} B^2)}.$$
     This implies the separability threshold $\lambda^*$ is determined solely by local impurities and is independent of the specific maximally entangled reference frame.
   • The fidelity absorption capacity $C_F$ depends on both local purities and the orientation of the product state relative to the Bell reference (specifically the alignment of Bloch vectors). Consequently, $\lambda_F$ can differ from $\lambda^*$, creating a regime where the state is entangled but not yet useful for teleportation.
   • For arbitrary pure-state references $|\psi\rangle$ with concurrence $C_\psi$, the capacity rescales universally: $C_{abs}(A \otimes B; \psi) = C_{abs}(A \otimes B) / C_\psi$.

2. Closed-Form Thresholds for X States:

   • The paper derives explicit formulas for $C_{abs}$ and $C_F$ for separable X states in all four Bell frames ($\Phi^\pm, \Psi^\pm$).
   • For the $|\Phi^+\rangle$ frame, $C_{abs}$ depends on the anti-diagonal coherence $u$ and the populations $b, c$ of the $\Psi$ block:
     $$C_{abs}(\sigma_X; \Phi^+) = -2 \text{Re}(u) + 2\sqrt{bc - (\text{Im}(u))^2}.$$
   • The teleportation capacity $C_F$ involves a similar structure but depends on the sum of populations $(b+c)$.
   • A key result is the explicit characterization of the gap $\lambda_F - \lambda^*$. This gap vanishes (i.e., entanglement and teleportation utility appear simultaneously) if and only if the middle populations are equal ($b=c$). Otherwise, a "useless entanglement" interval exists.

3. Decoherence Trajectories:

   • The framework is applied to local amplitude-damping and dephasing channels, which preserve the X-state structure.
   • By substituting time-dependent entries into the closed-form formulas, the paper generates explicit threshold trajectories $\lambda^*(t)$ and $\lambda_F(t)$.
   • The results show that while dephasing leads to monotone threshold changes, amplitude damping can cause non-monotonic behavior (thresholds rising and falling) due to transient redistribution of diagonal populations.

4. Decomposition Certificates:

   • The paper proves that both capacities are concave under convex decomposition. This allows for the calculation of certified lower bounds on thresholds for general separable noise by decomposing them into product or X components, avoiding the need for full numerical optimization in some cases.

Significance and Claims
The paper claims to provide a unified analytical framework for locating both the entanglement and teleportation transitions along a Bell-mixing line. Its primary significance lies in:

• Analytic Control: Moving beyond numerical checks of PPT or fidelity at specific points to provide closed-form formulas that can be bounded, differentiated, and tracked along channels.
• Directional Specificity: Explicitly treating the noise state as the input and the reference as fixed, distinguishing these capacities from direction-independent entanglement measures.
• Operational Insight: Clearly identifying the parameter regime where a state is entangled but fails to provide a quantum advantage in teleportation (the "useless entanglement" interval), and quantifying how noise orientation and type (product vs. X) influence this gap.
• Extension to Pure References: Providing a simple rescaling law for product noise when the reference is an arbitrary pure state, dependent only on the reference's concurrence.

The authors note that for general separable states without special structure, the exact thresholds require solving a small Semidefinite Program (SDP), but the derived closed forms for product and X states, combined with decomposition certificates, offer a practical and analytically tractable approach for many physically relevant noise models.