Strong Mpemba Effect Through a Reentrant Phase Transition

This paper demonstrates that strong direct and inverse Mpemba effects arise in the antiferromagnetic Ising model during temperature quenches across a reentrant phase transition, driven by the selective excitation of the slowest staggered relaxation mode in the paramagnetic phase.

The Gist

Imagine you have two cups of coffee. One is scalding hot, and the other is just warm. You put both in a refrigerator to cool them down. Common sense tells you the warm cup should reach the fridge's temperature first. But what if the hot cup actually got cold faster?

This counter-intuitive phenomenon is called the Mpemba effect. While it sounds like a magic trick, a new paper by Blom and colleagues explains how it happens in a very specific, theoretical world of tiny magnetic particles (spins).

Here is the story of their discovery, broken down into simple concepts.

The Setting: A Magnetic Dance Floor

Imagine a giant dance floor filled with dancers (the "spins"). These dancers have a rule: they want to hold hands with their neighbors, but they want to hold hands with someone facing the opposite direction (like a checkerboard pattern). This is called "antiferromagnetic" order.

However, there is also a loud DJ (a magnetic field) shouting, "Everyone face the same way!" The dancers are stuck in a tug-of-war between their desire to be opposite (the checkerboard) and the DJ's command to be uniform.

The "Reentrant" Twist

Usually, if you turn up the heat (temperature), the dancers get too jittery to hold hands at all, and they just spin around randomly. This is the "disordered" state. If you cool them down, they settle into their checkerboard pattern.

But in this specific setup, the authors found a weird "reentrant" behavior. Imagine cooling the dancers down:

1. Hot: They are chaotic and random.
2. Medium: They calm down and form the perfect checkerboard pattern.
3. Very Cold: Suddenly, they get confused again and break the pattern, returning to a chaotic state!

It's like a party where people start dancing wildly, then slow down to a synchronized dance, and then, as the music gets too slow, they start dancing wildly again. This "return to chaos" at low temperatures is the reentrant phase transition.

The Race: Who Cools Down Faster?

The researchers set up a race between two groups of dancers:

• Group A (The "Hot" Start): Starts in the chaotic state (high temperature), then the temperature is suddenly dropped to a cold, chaotic state.
• Group B (The "Warm" Start): Starts in the synchronized checkerboard state (medium temperature), then the temperature is dropped to that same cold, chaotic state.

The Result: Even though Group B started closer to the final destination (both are chaotic, but Group B was already "calm" in a different way), Group A (the hot start) actually arrived first.

This is the Mpemba effect: The system that started "farther away" from the goal finished the race sooner.

Why Does This Happen? The "Slow Lane" Analogy

To understand why, imagine the relaxation process (cooling down) is like a car driving home. The car has two gears:

1. Fast Gear: Moves quickly but only covers short distances.
2. Slow Gear: Moves very slowly and gets stuck in traffic.

The researchers discovered that the "Slow Gear" in this system is a specific type of movement called the "staggered mode." This is the specific wobble required to break the checkerboard pattern.

• Group B (The Warm Start): Because they started in the checkerboard pattern, they were already "wearing" the Slow Gear. When they tried to cool down, they got stuck in traffic. They had to slowly unwind their pattern before they could relax.
• Group A (The Hot Start): Because they started in the chaotic state, they had zero of that specific "checkerboard wobble." They didn't have to use the Slow Gear at all. They skipped the traffic jam entirely and zoomed home using only the Fast Gear.

Because Group A didn't have to deal with the slow, sticky part of the process, they won the race, even though they started further away.

The Key Ingredient: The Shape of the Map

The paper proves that this effect only happens because of that weird "reentrant" map (where cooling leads to order, then back to chaos).

If you change the rules of the dance floor (specifically, how many neighbors each dancer has), the "reentrant" map disappears. The path becomes a straight line: Hot $\to$ Cold. In that straight-line world, the Mpemba effect vanishes. The "hot" cup cools slower than the "warm" cup, just like normal physics predicts.

The Bottom Line

This paper doesn't tell us how to cool water faster in your kitchen. Instead, it provides a mathematical proof that in complex systems with competing forces, starting from a "farther" state can sometimes be an advantage if that starting point avoids a specific, slow bottleneck that the "closer" state gets stuck in.

They showed that the shape of the system's equilibrium (the "map" of order and chaos) dictates whether this strange racing effect can happen. If the map has a "reentrant" loop, the Mpemba effect is possible; if the map is a straight line, it is not.

Technical Summary

1. Problem Statement

The paper investigates the Mpemba effect, a counter-intuitive phenomenon where a system initially farther from thermal equilibrium relaxes faster than one initiated closer to it. While observed in various systems (spin glasses, ferromagnets, fluids), the specific mechanisms driving the strong and inverse Mpemba effects in antiferromagnetic (AF) systems near reentrant phase transitions remain underexplored.

The authors focus on the antiferromagnetic Ising model in a uniform magnetic field. In this system, nearest-neighbor antiferromagnetic interactions compete with the external field's tendency to align spins parallel. This competition creates a phase boundary between antiferromagnetic (Néel ordered) and paramagnetic (disordered) phases. The central question is how the topology of this phase boundary—specifically the presence of reentrance (where cooling leads to order, then disorder, then order again, or vice versa)—influences non-equilibrium relaxation dynamics.

2. Methodology

The authors employ a combination of statistical mechanics and dynamical systems theory:

• Model System: The Ising model on a Bethe lattice (a tree-like structure) with coordination number $z$ in a uniform magnetic field $H$. The Hamiltonian is $H(\sigma) = -J \sum_{\langle ij \rangle} \sigma_i \sigma_j - H \sum_i \sigma_i$, with $J < 0$ (antiferromagnetic).
• Dynamics: Single-spin-flip Glauber dynamics governed by a master equation satisfying detailed balance.
• Approximation Scheme: The Bethe-Guggenheim (BG) pair approximation. This allows the reduction of the many-body master equation into a closed set of three autonomous ordinary differential equations (ODEs) for the macroscopic order parameters:
  1. Total magnetization ($m$)
  2. Staggered magnetization ($s$) (order parameter for AF order)
  3. Nearest-neighbor correlation ($q$)
• Thermodynamic Analysis: Derivation of the BG free energy functional $F(x; T)$, proving it acts as a Lyapunov function for the dynamics (ensuring monotonic relaxation to equilibrium).
• Spectral Analysis: Linearization of the dynamical equations around the final equilibrium state to analyze the eigenvalues ($\lambda_k$) and eigenvectors ($v_k$) of the Jacobian matrix. This identifies the "slowest relaxation mode" governing long-time behavior.
• Protocol: Simulating temperature quenches from an initial temperature $T_i$ to a final temperature $T_f$. The study specifically compares quenches starting from the paramagnetic phase versus the antiferromagnetic phase, both terminating in the paramagnetic phase.

3. Key Contributions

1. Mechanism for Strong Mpemba Effect: The paper identifies that the strong Mpemba effect (where the initially hotter system relaxes strictly faster) arises from the selective excitation of the slowest relaxation mode.
2. Role of Reentrance: It establishes a direct causal link between reentrant phase transitions and anomalous relaxation. The authors demonstrate that reentrance is a necessary condition for the strong effect in this model; when the coordination number $z$ is low (monotonic phase boundary), the effect vanishes.
3. Spectral Origin: They provide a rigorous spectral explanation:
   • In the paramagnetic phase, the slowest relaxation mode is purely staggered (aligned with the $s$-axis).
   • A quench starting from the paramagnetic phase has zero overlap with this slow mode ($a_{slow} = 0$), leading to fast relaxation (dominated by faster modes).
   • A quench starting from the antiferromagnetic phase has a non-zero staggered component, exciting the slow mode, resulting in a slow-relaxation tail.
4. Coordination Number Dependence: The study quantifies the critical coordination number $z^* \approx 5.14$. For $z < z^*$, the phase boundary is monotonic, and no reentrant Mpemba effect occurs. For $z > z^*$, the boundary becomes non-monotonic (reentrant), enabling the effect.

4. Key Results

• Phase Diagram Topology:
  • For low $z$ (e.g., $z=3$), the critical field $H_c(T)$ decreases monotonically with temperature.
  • For high $z$ (e.g., $z=7$), $H_c(T)$ becomes non-monotonic. There exists a temperature window where the system is antiferromagnetic, bounded by paramagnetic regions at both higher and lower temperatures.
• Relaxation Dynamics (Cooling Quench):
  • Consider a quench to a final paramagnetic state $T_f$.
  • Path A (Hot $\to$ Cold): Start at $T_{i1}$ (paramagnetic). The system has $s(T_{i1}) = 0$. Overlap with the slow mode is zero. Relaxation is fast.
  • Path B (Warm $\to$ Cold): Start at $T_{i2}$ (antiferromagnetic, closer to $T_f$). The system has $s(T_{i2}) \neq 0$. Overlap with the slow mode is maximal. Relaxation is initially slow due to the "staggered tail."
  • Result: Despite Path B starting closer to equilibrium, Path A overtakes it at finite time. This is the Strong Mpemba Effect.
• Inverse Mpemba Effect:
  • By reversing the protocol (heating quench from a cold paramagnetic state to a hot paramagnetic state, passing through the AF phase), the system starting closer to the final state (but with non-zero staggered order) relaxes slower than the one starting farther away (with zero staggered order). This demonstrates the Inverse Mpemba Effect.
• Spectral Evidence:
  • The eigenvalue spectrum shows a clear separation between the slowest eigenvalue $\lambda_{slow}$ (associated with staggered fluctuations) and faster modes.
  • In the paramagnetic phase, the left eigenvector of the slow mode is $w_{slow} = (0, 1, 0)^T$, confirming it probes only the staggered magnetization $s$.

5. Significance

• Unification of Equilibrium and Non-Equilibrium: The work bridges equilibrium phase behavior (reentrance) with non-equilibrium relaxation phenomena. It proves that the topology of the equilibrium phase diagram dictates the spectral properties of the relaxation dynamics.
• Resolution of "Strong" vs. "Weak" Effects: It clarifies the conditions for the strong Mpemba effect (zero overlap with the slow mode) versus the generic effect, providing a quantitative framework (Mpemba index) applicable to spin systems.
• Theoretical Framework: By using the Bethe-Guggenheim approximation, the authors provide an analytically tractable model that captures complex behaviors (like reentrance) often missed by simple mean-field theory (which predicts reentrance for all $z$) but is more accurate than exact solutions for high-connectivity lattices.
• Future Directions: The paper highlights the need for spatially extended dynamical field theories to incorporate staggered magnetization as an independent slow field, moving beyond mean-field descriptions to understand pattern formation and spatially resolved relaxation in antiferromagnets.

In summary, the paper demonstrates that reentrant phase transitions act as a natural mechanism for generating strong Mpemba effects in antiferromagnetic systems by creating a scenario where the initial state's projection onto the system's slowest relaxation mode is qualitatively different depending on the starting phase, regardless of the initial distance from equilibrium.