Observation of Non-Markovian Evolution of Tripartite Quantum Steering

This paper presents the first experimental observation of non-Markovian evolution in tripartite quantum steering using GHZ-type mixed states, demonstrating the unique asymmetric steering structures and information backflow effects that distinguish multipartite systems from bipartite ones.

The Gist

The Big Picture: Quantum Memory in a Noisy World

Imagine you are trying to send a secret message to three friends (Alice, Bob, and Charlie) using a special kind of "quantum magic" called quantum steering. This magic allows one person to instantly influence the state of the others just by making a measurement, even if they are far apart.

However, in the real world, everything is "noisy." Think of this noise like a crowded, chaotic room where your secret message gets muddled and lost. Usually, once the message is lost to the noise, it's gone forever. This is called a Markovian process (like dropping a glass; once it shatters, it doesn't put itself back together).

But this paper explores a different scenario called Non-Markovian evolution. In this world, the noise has a "memory." It's as if the room remembers where the glass shards fell and, after a moment, pushes them back together, reassembling the glass. This "memory effect" allows information that seemed lost to flow back into the system, reviving the quantum magic.

What the Scientists Did

The researchers at Hangzhou Normal University wanted to see if this "memory effect" works when you have three people involved (a tripartite system), rather than just two. They used photons (particles of light) to act as Alice, Bob, and Charlie.

1. The Setup: They created a special mixed state of light (a "GHZ-type" state) where the three photons were deeply connected.
2. The Noise: They introduced a "decoherence" effect (the noise) to one of the photons (Alice) using quartz plates. This was designed to break the connection between the three, making the "steering" disappear.
3. The Memory: Then, they applied a second operation to "fix" the noise. Because of the memory effect, the lost connection didn't just stay gone; it came back.

The Key Findings: A Complex Dance of Control

The most exciting part of their discovery is how the "steering" behaves differently depending on who is looking at whom. In simple two-person relationships, steering is often one-way or mutual. But with three people, it gets complicated and asymmetric.

Think of it like a game of "Who can control the group?"

• Phase 1 (The Death): As the noise increased, the group lost their ability to steer each other. First, they could only steer in a specific way (where two people control the third). Then, they lost that ability too and became completely "unsteerable" (the magic was gone).
• Phase 2 (The Revival): Thanks to the memory effect, the magic came back. But it didn't return all at once.
  • First, the group regained the ability where Bob and Charlie could steer Alice, but Alice couldn't steer them back.
  • Later, the full connection was restored, and everyone could steer everyone.

The "Asymmetric" Surprise:
The paper highlights that this revival isn't the same for everyone.

• If you look at the group from Alice's perspective, the rules for when the magic returns are different.
• If you look from Bob's or Charlie's perspective, the rules are identical to each other but different from Alice's.

It's like a dance where the lead dancer (Alice) has a different rhythm than the two backup dancers (Bob and Charlie). The backup dancers move in sync with each other, but the lead dancer has a unique, more complex pattern of losing and regaining control.

Why This Matters (According to the Paper)

The paper claims this is the first time scientists have experimentally watched this specific "death and revival" of steering in a three-person quantum system.

• Hierarchy: It proves that in a group of three, the relationships aren't just simple pairs; they have a complex hierarchy. Some people can steer others in ways that don't happen in two-person groups.
• Directionality: The "memory" of the environment affects the group differently depending on who is interacting with the noise.
• Resource Protection: This shows that in a noisy, real-world environment, we might be able to use these memory effects to protect and recover useful quantum resources (like the ability to steer) that we thought were lost.

Summary Analogy

Imagine a group of three friends holding a rubber band together.

1. The Noise: A strong wind (environment) blows, stretching the band until it snaps. The friends can no longer feel each other's pull.
2. The Memory: The wind suddenly stops and reverses direction, pulling the rubber band back together.
3. The Result: The band doesn't just snap back to normal immediately. First, only two friends can feel the pull of the third. Then, the third friend feels the pull of the other two. Finally, everyone feels everyone again.

The paper shows that in a group of three, the way the rubber band snaps back is lopsided and complex, depending on which friend is standing where. This gives scientists a new map for understanding how to keep quantum connections alive in the messy, noisy real world.

Technical Summary

1. Problem Statement

• Context: Open quantum systems inevitably interact with their environments, leading to decoherence and information loss. While Markovian processes describe irreversible loss, non-Markovian processes involve memory effects where information flows back from the environment to the system.
• Gap: While non-Markovian dynamics (specifically the "death and revival" of quantum correlations) have been studied in bipartite systems (entanglement and steering), they remain largely unexplored in multipartite quantum steering.
• Specific Challenge: Tripartite systems exhibit complex hierarchical and directional structures (asymmetry) that do not exist in bipartite systems. Understanding how these structures evolve under non-Markovian noise is crucial for scalable quantum networks and asymmetric quantum information processing (e.g., one-sided device-independent protocols).

2. Methodology

The researchers employed a photonic experimental platform to simulate and observe the non-Markovian evolution of tripartite quantum steering.

• System State: They utilized Greenberger-Horne-Zeilinger (GHZ)-type mixed states encoded in the path and polarization degrees of freedom of photons. The state is defined as:
  $$\rho_{ABC}(\eta, \theta) = \eta |\Phi(\theta)\rangle\langle\Phi(\theta)| + (1-\eta)\mathbb{1}_A \otimes \rho_{BC}^\theta/2$$
  where $|\Phi(\theta)\rangle = \cos\theta|000\rangle + \sin\theta|111\rangle$. The parameter $\eta$ controls the noise admixture.
• Non-Markovian Simulation:
  • Environment: Introduced via the photon's frequency degree of freedom.
  • Interaction: The system (Alice, Bob, Charlie) interacts with the environment. Specifically, subsystem Alice (A) undergoes time-dependent interactions ($U_1$ and $U_2$) with the environment, while Bob and Charlie share a different environmental frequency.
  • Mechanism:
    • Dephasing ($U_1$): Implemented using quartz plates (QPs) at $0^\circ$ to induce decoherence, causing the steering to decay (death).
    • Recovery ($U_2$): Implemented using QPs at $90^\circ$ to introduce a compensatory phase, simulating the memory effect that allows information backflow (revival).
• Scenarios Tested:
  • 1SDI (One-sided Device-Independent): Alice's device is untrusted; Bob and Charlie's are trusted.
  • 2SDI (Two-sided Device-Independent): Alice and Bob's devices are untrusted; Charlie's is trusted (and permutations thereof).
• Measurement:
  • Standard three-qubit quantum state tomography was performed to reconstruct the density matrix ($\rho_{ex}$).
  • Witnesses:
    • Entanglement Witness (EW): To verify entanglement survival.
    • Steering Witness (SW): To quantify steerability in both 1SDI and 2SDI scenarios.
  • Non-Markovianity Quantification: The degree of non-Markovianity ($M$) was calculated by fitting the experimental death and revival curves to theoretical models, specifically looking at the distinguishability growth of quantum states.

3. Key Contributions

1. First Experimental Observation: This is the first experimental demonstration of non-Markovian evolution (death and revival) specifically for tripartite quantum steering.
2. Revealing Asymmetry: The study experimentally demonstrated the intricate asymmetric steering structure unique to tripartite systems. Unlike bipartite systems where steering is often one-way, tripartite systems show complex directional dependencies based on which subsystems are trusted/untrusted.
3. Hierarchical Dynamics: The work maps the transition between different steering regimes:
   • Steerable in both 1SDI and 2SDI $\to$ Steerable only in 2SDI $\to$ Unsteerable (Death).
   • Unsteerable $\to$ Steerable only in 2SDI $\to$ Steerable in both (Revival).
4. Quantification of Memory Effects: The authors provided an operational method to quantify the degree of non-Markovianity ($M$) by observing the full dynamical evolution of steering correlations, rather than just static snapshots.

4. Key Results

• Death and Revival: The experiment successfully observed the complete cycle of steering death and revival.
  • As the effective thickness of the quartz plates ($L$) increased (simulating time), the system transitioned from being steerable in both scenarios to unsteerable.
  • Upon applying the compensatory operation ($U_2$), the steering was revived, confirming the presence of strong memory effects.
• Non-Markovianity Degree: The fitting of the experimental data yielded a non-Markovianity degree of $M \approx 1$ (specifically $0.99999(1)$ for entanglement and $1(1)$ for steering), indicating an ideal non-Markovian process where information backflow is nearly perfect.
• Asymmetric Revival Patterns:
  • When analyzing different bipartitions (A|BC, B|AC, C|AB), the researchers found that the critical thickness ($L(\lambda)$) required for the transition from unsteerable to steerable differed depending on the direction of steering.
  • For example, in the (A|BC) configuration, the revival thresholds were distinct from (B|AC) and (C|AB), highlighting the directional nature of tripartite steering.
• Entanglement vs. Steering Hierarchy: The results confirmed the existence of states that are entangled but unsteerable, visually demonstrating the hierarchy where entanglement is a broader resource than steering.

5. Significance

• Foundational Insight: The study provides foundational insights into the hierarchical and directional structures of multipartite quantum correlations in open systems, moving beyond the simpler bipartite models.
• Resource Protection: It highlights the potential of non-Markovian environments to protect and recover quantum resources (steering) in noisy conditions, which is vital for real-world quantum technologies.
• Quantum Networks: The findings are directly applicable to asymmetric quantum information processing and quantum networks, where different nodes may have varying levels of trust (device independence) and where memory effects can be engineered to maintain connectivity.
• Experimental Framework: The paper establishes a versatile experimental platform for controlling system-environment interactions, offering a blueprint for future studies on complex open quantum systems and quantum resource dynamics.