Communication with Quantum Catalysts

1. Problem Statement

Quantum communication systems are inherently limited by noise arising from imperfect devices and environmental decoherence. This noise degrades the quality of quantum channels and entangled states, reducing the capacity to transmit both quantum and classical information.

• Traditional Approach: "Exact" quantum catalysts (states that facilitate transformations without changing) have been proposed to enhance communication. However, recent theoretical work suggests that for certain noisy channels (e.g., dephasing channels), exact catalysts yield a zero catalytic channel capacity, rendering them ineffective for specific tasks.
• The Gap: There is a need to explore whether allowing the catalyst to undergo slight, controlled alterations (a concept borrowed from chemical "deactivation" or "embezzling") can overcome these limitations and enhance transmission efficiency for both quantum and classical information.

2. Methodology

The authors introduce and analyze "embezzling catalysts"—quantum states that facilitate information transmission while undergoing negligible but non-zero changes. They employ two primary mathematical frameworks to construct and analyze these protocols:

• Convex-Split Lemma: A tool from single-shot information theory used to construct catalysts. The authors utilize this to create a bipartite state $\tau$ that approximates a target state by mixing a noisy channel's Choi state with a full-rank state.
• Embezzling States: Specifically, the van Dam-Hayden embezzling states ($\tau^E$), defined as a superposition of basis states with specific coefficients ($\frac{1}{\sqrt{j}}$), which allow for the extraction of entanglement resources with minimal disturbance to the catalyst.
• Distinguishability Measures: The performance is quantified using the Purified Distance ($P$) and Uhlmann Fidelity ($F_U$) to measure how close the output state is to the ideal target state (e.g., a maximally entangled state) and how much the catalyst itself changes.
• Dimensionality Reduction: The authors investigate strategies to minimize the dimensionality of the catalysts (the number of copies required) to make the protocols practical. They compare using maximally mixed states versus randomly selected full-rank states as the base for constructing catalysts.

3. Key Contributions

A. Non-Zero Catalytic Channel Capacity

The paper proves that while exact catalysts may fail (yielding zero capacity) for certain noisy channels (like two copies of a dephasing channel with noise parameter $p < 0.817$), embezzling catalysts can achieve a strictly non-zero catalytic channel capacity ($Q_c^\epsilon(N) > 0$) for any noisy quantum channel, provided a small error threshold $\epsilon$ is allowed.

B. Catalytic Superdense Coding

The authors extend the concept of catalysis to classical information transmission via Superdense Coding.

• They demonstrate that standard correlated catalysts (like Duan's state) cannot be used here because superdense coding protocols forbid direct classical communication between parties (only Local Operations, LO, are allowed).
• However, embezzling-state-assisted protocols work because they rely solely on local operations. This allows the enhancement of superdense coding capacity even with noisy entangled states.

C. Dimensionality Reduction Strategies

A significant practical contribution is the method for reducing the dimensionality of the required catalysts.

• The authors show that using randomly selected full-rank states to construct the catalyst (via the convex-split lemma) requires significantly fewer copies (lower dimension) compared to using the standard maximally mixed state ($I/d^2$).
• Numerical experiments indicate that selecting a finite number of random states (e.g., 200 for dephasing channels) yields near-optimal performance, drastically reducing the resource overhead.

D. Long-Distance Entanglement Distribution

The paper applies these findings to long-distance entanglement distribution.

• Conventional methods and correlated-catalyst methods fail beyond a specific distance limit (e.g., $l > (2 \ln 3)/\alpha$ for depolarizing channels).
• Embezzling catalysts allow for entanglement distribution across distances exceeding these theoretical limits, effectively extending the range of quantum networks.

4. Key Results

• Theorem 1 (Non-Zero Capacity): For any noisy channel $N$ and error $\epsilon > 0$, there exist finite-dimensional quantum catalysts such that the single-shot $\epsilon$-error catalytic channel capacity is non-zero.
• Theorem 2 & 3 (Protocol Construction):
  • Using the Convex-Split Lemma, a catalyst $\tau^{CS}$ can be constructed to achieve capacity $\log m$ if the number of copies $n$ satisfies specific bounds related to the max-relative entropy.
  • Using Embezzling States ($\tau^E$), the protocol achieves the same capacity with a Schmidt rank $M$ dependent on the target dimension and error tolerance.
• Performance Comparison:
  • Embezzling-state protocols generally outperform convex-split-lemma protocols in terms of the fidelity achieved for a given catalyst dimension.
  • However, embezzling-state protocols induce a slightly larger change (consumption) in the catalyst state compared to convex-split protocols, though this change remains within acceptable bounds.
• Numerical Findings:
  • For dephasing channels, using 200 randomly selected full-rank states reduces the required catalyst dimension by a significant margin compared to using the maximally mixed state.
  • In superdense coding, as the dimension of the embezzling state increases, the capacity approaches the theoretical upper bound of $2 \log d$.

5. Significance

• Overcoming Zero-Capacity Limits: This work fundamentally shifts the understanding of quantum catalysis by proving that "imperfect" (embezzling) catalysts are superior to "perfect" (exact) ones in noisy environments, unlocking communication capabilities previously thought impossible.
• Practicality: By demonstrating methods to reduce the dimensionality of catalysts, the paper addresses a major barrier to implementation: the difficulty of generating and manipulating high-dimensional entangled states.
• Unified Framework: It unifies the enhancement of both quantum (entanglement distribution, channel capacity) and classical (superdense coding) information transmission under a single catalytic framework.
• Future Directions: The paper opens new avenues for research into universal catalysts and their application in other quantum tasks like teleportation (as hinted in their accompanying paper), potentially revolutionizing the design of future quantum networks and repeaters.

In summary, the paper establishes that controlled imperfection (embezzling) in quantum catalysts is a powerful resource that enables reliable, high-capacity communication over noisy channels where traditional exact catalysts fail, while offering concrete strategies to minimize the physical resources required for implementation.