⚛️ quantum physics

Generation of Quantum Entanglement in Autonomous Thermal Machines: Effects of Non-Markovianity, Hilbert Space Structure, and Quantum Coherence

This paper theoretically demonstrates that a quantum autonomous thermal machine composed of two qubits can generate entanglement in an external system exclusively through a specific thermodynamic cycle (Cycle A) that leverages non-Markovian memory effects, Hilbert space structure, and quantum coherence, with parameters compatible with superconducting qubit platforms.

Original authors: Achraf Khoudiri, Khadija El Anouz, Abderrahim El Allati

Published 2026-03-24

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Achraf Khoudiri, Khadija El Anouz, Abderrahim El Allati

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a tiny, self-contained factory made of two little switches (qubits) that are constantly being heated by a hot stove on one side and cooled by a freezer on the other. This factory is the Quantum Autonomous Thermal Machine (QATM).

Now, imagine you have two other switches sitting nearby, completely disconnected from the stove and freezer. These are your External System.

The big question this paper asks is: Can our little factory, just by doing its own work, magically "glue" the two disconnected switches together so they start acting as a single, synchronized unit? In physics, this "gluing" is called entanglement, a spooky connection where two particles share a fate, no matter how far apart they are.

Here is the story of how the researchers found the answer, explained through simple analogies.

1. The Setup: A Factory with Two Modes

The factory (the QATM) has two switches, let's call them M1 and M2.

M1 is near a cold reservoir (freezer).
M2 is near a hot reservoir (stove).

Because of the temperature difference, energy flows through the factory. The researchers discovered that this flow can happen in two distinct "dance routines" or cycles:

Cycle A (The "Memory" Dance): In this mode, the factory absorbs heat from the outside world in a specific way. It acts like a sponge that soaks up energy but also holds onto a "memory" of what happened a moment ago.
Cycle B (The "Forgetful" Dance): In this mode, the factory dumps heat out. It acts more like a standard machine that forgets everything instantly, letting information leak away into the environment.

2. The Secret Ingredient: Non-Markovianity (The Echo)

In the quantum world, there is a concept called Markovianity. Think of it like a conversation in a noisy room where you only hear what someone says right now. You have no memory of what they said five seconds ago.

The opposite is Non-Markovianity. This is like being in a canyon where your voice echoes back to you. You hear what you said a moment ago, and it mixes with what you are saying now. This "echo" is a memory effect.

The paper shows that Cycle A creates a strong echo. The factory and the outside switches talk to each other, exchange information, and then the information bounces back. Cycle B is quiet; the information just disappears into the noise.

3. The Result: Entanglement Only Happens in Cycle A

The researchers ran simulations to see if the two outside switches (S1 and S2) would become entangled.

In Cycle B (The Forgetful Dance): The outside switches barely talk to each other. They get "decohered" (scrambled) by the environment. No entanglement forms. It's like trying to whisper a secret to a friend while a hurricane is blowing; the message is lost.
In Cycle A (The Memory Dance): Because the factory is acting like a "structured reservoir" (a smart sponge with an echo), it protects the outside switches. The "echo" of information bouncing back and forth allows the two outside switches to synchronize. Entanglement is born.

4. The Role of "Virtual Temperature"

You might wonder, how does the factory know which dance to do? It depends on the Virtual Temperature.

Think of the factory not just as two switches, but as a single "super-switch" with a made-up temperature.

If the real temperatures of the stove and freezer are set just right, this "super-switch" gets a negative virtual temperature.
This sounds weird (like having less than absolute zero!), but in quantum mechanics, it means the system is "inverted" (more excited than calm). This inversion is the key that unlocks Cycle A.
If the temperatures aren't set right, the virtual temperature stays positive, and the system falls into Cycle B, where no magic happens.

5. Coherence: The Glue

The paper also talks about Coherence. Imagine the outside switches are two dancers.

Local Coherence: Each dancer knows their own steps.
Correlation Coherence: The dancers are watching each other and moving in sync.

The researchers found that in Cycle A, the factory helps convert the dancers' individual steps (local coherence) into a synchronized routine (correlation coherence). This synchronization is the fuel that creates entanglement. In Cycle B, the dancers lose their rhythm and stop watching each other.

The Big Picture

This paper is a blueprint for building better quantum computers. It tells us:

Don't just build a machine; build a machine with a memory. If your quantum device forgets everything instantly (Markovian), it will fail to create the strong connections (entanglement) needed for powerful computing.
Temperature is a tool. By carefully tuning the heat and cold, you can switch a machine between a "forgetful" mode and a "memory-rich" mode.
The "Echo" is good. In the noisy quantum world, having information bounce back (non-Markovianity) isn't a bug; it's a feature that can be used to create and protect quantum magic.

In short: The authors found a way to turn a simple heat engine into a "quantum glue factory," but only if they tune the heat just right to make the machine remember its past. When it remembers, it can link distant particles together. When it forgets, the link breaks.

1. Problem Statement

The paper addresses the challenge of generating and controlling quantum entanglement in open quantum systems using Quantum Autonomous Thermal Machines (QATMs). While previous studies have explored entanglement generation in steady-state regimes or focused on the second law of thermodynamics, there is a need to understand:

How non-Markovian dynamics (memory effects) arising from the interaction between a thermal machine and an external system influence entanglement.
The role of Hilbert space structure and virtual temperatures in defining thermodynamic cycles that can act as resources for entanglement.
Whether specific thermodynamic cycles can suppress decoherence and enhance quantum correlations compared to others, particularly in the context of the second law of quantum thermodynamics (entropy production).

2. Methodology

The authors propose a theoretical model involving a QATM interacting with an external quantum system.

System Architecture:
- QATM ( $M$ ): Composed of two qubits ( $M_1, M_2$ ), each coupled to independent Markovian bosonic reservoirs ( $R_1, R_2$ ) at different temperatures ( $T_{R1} < T_{R2}$ ).
- External System ( $S$ ): Composed of two qubits ( $S_1, S_2$ ) initially prepared in superposition states.
- Interaction: The QATM and the external system are coupled via an energy-conserving interaction Hamiltonian ( $\hat{H}_{MS}$ ) that facilitates transitions between specific states (e.g., $|01\rangle \leftrightarrow |10\rangle$ ) without external time-dependent driving.
Theoretical Framework:
- Dynamics: The system evolves under a Lindblad master equation in the weak-coupling limit. The reduced dynamics of the subsystems are analyzed to observe non-Markovian effects.
- Thermodynamic Cycles: Based on the Hilbert space structure and energy constraints ( $E_{M2} - E_{M1} = E_{S2} - E_{S1}$ $E_{M 2} - E_{M 1} = E_{S 2} - E_{S 1}$ ), two distinct thermodynamic cycles are defined:
  - Cycle A: Heat flows from the external system to the machine ( $Q_M \geq 0$ ). This corresponds to a negative virtual temperature ( $T_M < 0$ ) and occurs when $T_{M1} \leq \frac{E_{M1}}{E_{M2}}T_{M2}$ .
  - Cycle B: Heat flows from the machine to the external system ( $Q_M \leq 0$ ). This corresponds to a positive virtual temperature ( $T_M > 0$ ) and occurs when $T_{M1} \geq \frac{E_{M1}}{E_{M2}}T_{M2}$ .
- Metrics:
  - Non-Markovianity: Quantified using the Breuer-Laine-Piilo measure (based on trace distance backflow) and the entropy production rate ( $\dot{\Sigma}$ ). Negative $\dot{\Sigma}$ indicates memory effects.
  - Entanglement: Measured via Concurrence ( $C$ ).
  - Correlations: Measured via Mutual Information ( $I$ ) and Coherence Correlations ( $\Delta C_S$ ).

3. Key Contributions

Definition of Thermodynamic Cycles via Virtual Temperature: The authors rigorously define two operational cycles (A and B) for a QATM based on the ratio of qubit temperatures and energy spacings, linking them to the concept of virtual temperature (including negative temperatures).
Identification of Non-Markovianity as a Resource: The study demonstrates that even with Markovian reservoirs, the exchange of correlations between the QATM and the external system induces non-Markovian memory effects in the reduced dynamics of the subsystems.
Cycle-Dependent Entanglement Generation: A critical finding is that entanglement is generated only when the machine operates in Cycle A. Cycle B, despite generating similar total correlations (mutual information), fails to generate entanglement.
Role of Coherence: The paper establishes that coherence correlation (the difference between global coherence and the sum of local coherences) is the specific resource driving entanglement generation. In Cycle A, local coherences are preserved longer, allowing them to convert into entanglement.

4. Key Results

Entropy Production and Irreversibility:
- Entropy production ( $\Sigma$ ) is positive for both cycles, satisfying the second law.
- However, the entropy production rate ( $\dot{\Sigma}$ ) becomes negative during evolution, signaling non-Markovian dynamics.
- Cycle A exhibits more frequent and larger negative values of $\dot{\Sigma}$ compared to Cycle B, indicating stronger memory effects and lower irreversibility.
Entanglement vs. Total Correlations:
- Cycle A: Generates significant concurrence (entanglement) between $S_1$ and $S_2$ . The mutual information is high, and the system benefits from the "structured reservoir" effect of the QATM.
- Cycle B: Generates negligible concurrence (approaching zero), even though the mutual information (total classical + quantum correlations) remains comparable to Cycle A. This highlights that total correlations do not guarantee entanglement; the type of cycle matters.
Coherence Dynamics:
- In Cycle A, local coherences decay non-monotonically (due to information backflow), preserving the quantum resource necessary for entanglement.
- In Cycle B, local coherences decay rapidly and monotonically to zero due to stronger decoherence and irreversibility.
- The coherence correlation ( $\Delta C_S$ ) is positive and persistent in Cycle A, directly linking the preservation of coherence to the generation of entanglement.
Experimental Feasibility:
- The parameters used (energy spacings in the GHz range, coupling strengths in the MHz range, and time scales in microseconds) are consistent with current superconducting qubit platforms.

5. Significance

Bridging Thermodynamics and Information: The work provides a concrete mechanism showing how thermodynamic constraints (temperature gradients, virtual temperatures) can be engineered to control quantum information resources (entanglement).
Non-Markovianity as a Control Knob: It challenges the view of non-Markovianity solely as a nuisance, showing that in autonomous thermal machines, memory effects induced by cycle selection can be harnessed to protect and generate entanglement.
Cycle Selection Strategy: The results suggest that for quantum technologies requiring entanglement (e.g., quantum networks or computing), operating a thermal machine in a specific thermodynamic cycle (Cycle A) is crucial, as the alternative cycle (B) leads to rapid decoherence and loss of entanglement despite similar heat flows.
Experimental Relevance: By mapping the theoretical model to superconducting qubit parameters, the paper offers a viable roadmap for experimental verification of non-Markovian entanglement generation in autonomous thermal devices.

In conclusion, the paper demonstrates that temperature differences, Hilbert space structure, and coherence act as quantum resources. Specifically, Cycle A of a QATM acts as a structured, non-Markovian reservoir that mitigates decoherence, allowing local coherences to evolve into robust entanglement, whereas Cycle B fails to do so due to higher irreversibility.