Phase Ordering in a few O(n) Symmetric Models: Slow Growth, Mpemba Effect and Experimental Relevance

Through Monte Carlo simulations of the three-dimensional nonconserved XY and Ising models, this study reveals anomalously slow phase ordering growth at zero temperature and demonstrates a robust Mpemba effect where systems quenched from higher initial temperatures reach equilibrium faster, with findings that hold across different initial magnetization distributions and offer significant experimental relevance.

The Gist

Imagine you have a large crowd of people in a room, all spinning around randomly like dizzy dancers. This represents a "hot" state where everything is chaotic. Now, imagine you suddenly turn off the music and tell everyone to stop spinning and stand still, facing the same direction. This is what physicists call "cooling down" or a "quench."

Usually, you would expect the people who started out spinning the fastest (the hottest) to take the longest time to stop and get organized. However, this paper reports a surprising discovery: sometimes, the people who were spinning the fastest actually get organized faster than those who were spinning slowly.

This counter-intuitive phenomenon is called the Mpemba Effect. You might know it from the old saying that "hot water freezes faster than cold water." While that specific claim is debated in real life, this paper shows that a similar "hot beats cold" race happens in the microscopic world of magnets and spins.

Here is a breakdown of what the researchers found, using simple analogies:

1. The Two Types of "Dancers"

The researchers studied two different models of how these spins behave, which they call the Ising model and the XY model.

• The Ising Model: Imagine people who can only face either North or South. They are like binary switches.
• The XY Model: Imagine people who can face any direction on a flat circle (North, East, South, West, or anywhere in between). They have more freedom to move.

The researchers simulated these systems in 3D (like a cube of people) and 2D (like a flat sheet of paper).

2. The "Slow Motion" Mystery

When they cooled the 3D XY model down to absolute zero (the coldest possible temperature), they expected the "dance floor" to organize at a standard speed. In physics, there is a rule of thumb that says the size of the organized groups should grow at a specific rate (like a car driving at a steady speed).

However, they found that at absolute zero, the 3D XY model was extremely slow. It was like the dancers were stuck in mud, moving at only about 30% of the expected speed.

• Why? In this 3D world, the "mistakes" in the dance (called defects) aren't just flat lines; they are long, tangled strings or ropes that weave through the 3D space. Untangling these 3D ropes takes a lot of time and effort, causing the system to crawl.

3. The Mpemba Race: Hot Starts Win

The main experiment involved starting the "dance" from different temperatures:

• Group A: Started very hot (spinning wildly).
• Group B: Started just above the freezing point (spinning moderately).

They were all cooled down to the same final temperature. The researchers expected Group B to finish first because they started closer to the goal. Instead, Group A (the hot starters) finished first.

The Analogy: Imagine two runners. Runner A starts at the top of a steep hill, running wildly. Runner B starts halfway down, jogging calmly. You expect Runner B to reach the bottom first. But in this experiment, Runner A's wild momentum and the way they scrambled at the start actually helped them clear the obstacles faster than Runner B, who got stuck in a "traffic jam" of indecision.

4. The Dimensionality Twist (2D vs. 3D)

This is where it gets really interesting. The researchers found that this "hot wins" effect depends heavily on whether the system is flat (2D) or a solid block (3D).

• In 3D (The Real World): The "hot wins" effect happened naturally, even when the starting group had a mix of all kinds of spins. The system didn't need any special rules to make this happen. This suggests the effect is robust and could be seen in real-world experiments.
• In 2D (Flat World): The effect disappeared unless they forced a very specific rule: they had to make sure the starting crowd had zero net direction (equal numbers facing North and South). If they let the crowd start with any random mix, the "hot wins" effect vanished.

Why the difference? In 2D, the "mistakes" are just points. In 3D, they are long lines. The researchers argue that the way the crowd fluctuates (wiggles and changes) near the critical point is much wilder in 2D than in 3D. In 3D, the wild fluctuations of the "hot" start actually help the system find the right path faster, whereas in 2D, those fluctuations just cause chaos that slows things down.

5. Why This Matters

The paper emphasizes that previous studies often forced the starting conditions to be perfectly balanced (zero magnetization) to see this effect. That's like forcing a race to start with everyone standing perfectly still.

This study is special because they let the starting crowds be messy and random, just like they would be in a real experiment. They found that even with this messiness, the "hot starts" still won in 3D. This makes the result much more relevant to real-world physics and potential experiments, suggesting that the Mpemba effect is a genuine feature of how magnetic materials order themselves, not just a trick of the math.

In summary: The paper shows that in 3D magnetic systems, starting "hotter" can actually help a system organize itself faster than starting "cooler," a phenomenon that survives even when the starting conditions are messy and realistic. However, this trick only works in 3D; in a flat 2D world, you need very specific conditions to see it.

Technical Summary

Technical Summary: Phase Ordering in a few O(n) Symmetric Models

Problem Statement
This study investigates the kinetics of phase ordering in nonconserved $O(n)$ symmetric models, specifically the three-dimensional (3D) XY and Ising models. The research addresses two primary phenomena: the anomalous growth of the characteristic length scale during phase ordering at zero temperature ($T_f = 0$) and the existence of the Mpemba effect (ME)—where a system starting from a higher initial temperature ($T_s$) reaches equilibrium faster than one starting from a lower $T_s$—in magnetic transitions. A critical focus is placed on the experimental relevance of these findings by avoiding artificial constraints on the initial order parameter (magnetization), which have been used in previous theoretical studies but are difficult to realize experimentally.

Methodology
The authors employ Monte Carlo (MC) simulations on simple cubic lattices with linear dimensions $L=256$ (and $L=64$ for specific visualizations).

• Models: The study utilizes the 3D XY model (two-component unit vectors) and the 3D Ising model (scalar spins $\pm 1$). For comparison, results from the 2D Ising model are also presented.
• Protocol: Systems are quenched from a paramagnetic phase (starting temperatures $T_s > T_c$) to final ferromagnetic temperatures ($T_f$). Simulations cover quenches to $T_f = 0$ and finite temperatures ($T_f = 0.5T_c$ and $0.6T_c$).
• Initial Conditions: Unlike prior studies that constrained the initial magnetization to zero, this work samples initial configurations from the full distribution of order-parameter values at each $T_s$, reflecting realistic experimental conditions where instantaneous magnetization fluctuates.
• Dynamics: The simulations use the Metropolis algorithm with nonconserved dynamics (spin flips for Ising; random trial rotations for XY).
• Observables:
  • Order Parameter: Magnetization distributions are analyzed to confirm critical fluctuations.
  • Correlation: The two-point equal-time correlation function $C(r, t)$ is calculated to determine the characteristic domain length $\ell(t)$. For the XY model, $\ell(t)$ is derived from the decay of $C(r, t)$ to 0.1; for the Ising model, it is the first moment of the domain size distribution.
  • Energy: Potential energy $E$ is tracked to monitor relaxation.
• Statistics: Results are averaged over thousands of independent initial configurations (e.g., 1450 for XY at $T_f=0$).

Key Results

1. Anomalous Growth Exponent in 3D XY Model:
   For quenches to $T_f = 0$, the characteristic length scale $\ell(t)$ in the 3D XY model grows as $\ell \sim t^\alpha$ with an exponent $\alpha \approx 0.15$. This is significantly slower than the theoretical expectation of $\alpha = 1/2$ for nonconserved dynamics. This slow growth persists in the long-time limit within the accessible simulation window, resembling the behavior observed in the 3D Ising model at intermediate times. The growth is driven by the annihilation of vortex-antivortex pairs, which form line defects in 3D geometry. In contrast, quenches to $T_f = 0.5T_c$ exhibit the expected $\alpha \approx 1/2$ behavior.

2. Observation of the Mpemba Effect:
   The study confirms the Mpemba effect in both 3D models. Systems initialized at higher $T_s$ (closer to or above $T_c$) reach the final equilibrium state faster than those initialized at lower $T_s$, despite starting with higher potential energy. This "overtaking" behavior is observed in both domain growth ($\ell(t)$) and energy relaxation ($E(t)$).

   • Dimensionality Dependence: In the 3D Ising and XY models, the effect is robust even when initial configurations are drawn from the full magnetization distribution.
   • 2D Ising Contrast: In the 2D Ising model, the ME is not observed when using the full distribution of initial magnetizations. It only appears when the initial magnetization is artificially restricted to $m \approx 0$. The authors attribute this difference to the broader critical fluctuations in 2D (higher susceptibility exponent $\gamma$) compared to 3D.

3. Scaling and Fluctuations:
   The data for $T_f = 0$ in the XY model exhibits dynamic scaling (self-similar growth), collapsing onto a master function. The study explicitly demonstrates that critical fluctuations in the initial state are significant and that the ME arises from differences in these fluctuations rather than metastability or specific initial constraints.

Significance and Claims
The paper claims that its results are of "much experimental relevance" because the observation of the Mpemba effect does not rely on the artificial restriction of the initial magnetization to zero, a condition that is difficult to enforce in physical experiments. By utilizing the full distribution of order parameters, the study suggests that the ME is a generic feature of para- to ferromagnetic transitions in 3D systems, driven by critical fluctuations in the initial state.

The authors modestly note that while their results suggest the ME can arise without metastability, the issue requires further investigation given that freezing phenomena are observed in Ising models at $T=0$. They also highlight that the slow growth exponent ($\alpha \approx 0.15$) in the 3D XY model at $T_f=0$ remains an open question regarding whether a crossover to $\alpha=1/2$ occurs at timescales beyond current computational capabilities.

The work connects the ME to the dimensionality of the system and the nature of critical fluctuations, providing a unified picture for magnetic transitions that contrasts with previous findings in lower dimensions or constrained systems.