Mpemba effect in a two-dimensional bistable potential

This paper presents an exactly solvable two-dimensional model of the Mpemba effect in an overdamped Langevin system within a radially symmetric bistable potential, analytically demonstrating how specific initial conditions can lead to faster relaxation to equilibrium through non-monotonic mode amplitudes and a derived crossing condition for the Kullback-Leibler divergence.

The Gist

The Big Idea: The "Hot Water" Mystery

You've probably heard the old saying: "Hot water freezes faster than cold water." While this sounds impossible (like trying to run a marathon faster by starting with a head start), it's a real phenomenon called the Mpemba Effect.

Usually, if you have two cups of water, one hot and one cold, and you put them in a freezer, the cold one should reach the freezing point first. But sometimes, the hot one catches up and wins the race. Scientists have been arguing about why this happens for decades.

This paper is a mathematical "proof of concept." The authors built a perfect, simplified model of a particle (like a tiny ball) moving in a specific landscape to show exactly how and when this "hotter wins" race can happen.

The Setting: A Two-Valley Landscape

Imagine the particle is a hiker trying to get to the bottom of a valley.

• The Landscape: The authors created a map with two valleys (a "bistable potential").
  • Valley A: A deep, deep hole far away from the center. This is the "goal" (the final equilibrium state).
  • Valley B: A smaller, shallower dip right near the center (the origin).
• The Rules: The hiker is moving through thick mud (overdamped). They are being jostled by random bumps (heat/temperature).

The Race: Starting from Different Heights

The experiment involves two hikers starting in different spots:

1. Hiker Cold: Starts in a small dip near the center, but they are "cold" (calm, low energy).
2. Hiker Hot: Starts in the same small dip, but they are "hot" (jittery, high energy).

Both hikers are told to run to the Deep Valley (Valley A).

The Intuition: You'd think the Cold hiker would win because they are already closer to the goal. The Hot hiker is jittery and might run around in circles, wasting time.

The Surprise: The paper shows that under specific conditions, the Hot hiker actually arrives first.

How Did They Solve It? (The Magic Map)

Solving the math for a particle moving in mud with two valleys is usually a nightmare. It's like trying to predict the path of a drunk person in a storm.

The authors used a clever trick:

1. The Shape: They designed the landscape using a very specific, "piecewise" shape (like a smooth curve made of different mathematical Lego blocks).
2. The Translation: They translated the messy "mud and bumps" problem into a clean "quantum mechanics" problem (specifically, a Schrödinger equation). This is like translating a chaotic street fight into a chess game where you can calculate every move perfectly.
3. The Result: Because they translated it, they could solve it exactly. They didn't have to guess or use a computer to approximate; they found the precise mathematical answer.

The Secret Sauce: The "Effective Wall"

Here is the most interesting part of their discovery.

In a 1D world (a straight line), if you have two valleys, the hot hiker usually loses unless you put a wall in front of them to force them to turn around.

But in this 2D world (a flat plane), the center point (where the hiker starts) acts like a ghost wall.

• Think of the center as the tip of a cone. You can't go through the tip; you have to go around it.
• This "ghost wall" changes the rules of the game. It forces the hot hiker to explore the landscape in a way that, surprisingly, helps them find the exit faster than the calm, cold hiker who gets stuck in the local dip.

The "Crossing" Moment

The paper proves that for the Mpemba effect to happen, two things must align:

1. The Amplitude: The "Hot" starting temperature must be tuned just right so that the hiker's initial jitteriness helps them escape the small dip efficiently.
2. The Crossing: As time goes on, the "Hot" hiker's path must cross the "Cold" hiker's path. The paper shows that if the deep valley is far enough away and the shallow dip is close enough, this crossing will happen.

Why Does This Matter?

Before this paper, we mostly understood this effect in simple, one-dimensional lines or with artificial walls. This paper is a breakthrough because:

• It works in 2D (like real life, where things move in all directions).
• It needs no artificial walls. The geometry of the space itself creates the effect.
• It provides a perfectly solvable model. Scientists can now use this exact math to predict when the Mpemba effect will happen in other systems, from cooling coffee to quantum computers.

The Takeaway

Imagine you are trying to find a lost key in a dark room.

• The Cold approach: You move slowly and carefully, checking every inch near where you think you dropped it.
• The Hot approach: You are frantic, shaking the room, throwing things around.

Usually, the careful person wins. But if the room has a specific shape (like a cone or a funnel) and the key is in a specific spot, the frantic shaking might accidentally kick the key into the light faster than the careful search.

This paper proves that mathematically, in a 2D world, the "frantic shaking" (high temperature) can indeed be the winning strategy to reach equilibrium faster.

Technical Summary

1. Problem Statement

The Mpemba effect is a counter-intuitive phenomenon where a system initially at a higher temperature ($T_{ini}$) reaches thermal equilibrium faster than an identical system starting at a lower temperature after being quenched to a cold bath. While observed in various systems (colloids, granular fluids, spin glasses), the theoretical understanding has been limited by a lack of exactly solvable models in smooth, higher-dimensional potentials.

Previous analytical studies were largely restricted to:

• One-dimensional (1D) systems.
• Discrete-level models or piecewise linear potentials.
• Systems requiring hard confining walls to observe the effect in asymmetric potentials.

The authors address the open question: Can the Mpemba effect occur in a smooth, two-dimensional (2D) bistable potential without artificial confining walls? Furthermore, they aim to clarify the role of dimensionality and geometry in nonequilibrium relaxation.

2. Methodology

The authors construct an exactly solvable model for an overdamped Langevin particle in a 2D radially symmetric bistable potential.

• Potential Construction:
  They define a piecewise potential $V(r)$ (where $r = |\mathbf{r}|$) composed of three regions:

  1. Inner region ($0 \le r < r_-$): Quadratic (harmonic).
  2. Middle region ($r_- \le r < r_+$): Quadratic plus logarithmic (anti-harmonic/barrier).
  3. Outer region ($r \ge r_+$): Quadratic plus logarithmic (second well).
     The parameters are tuned to ensure $V(r)$ and its first derivative $V'(r)$ are continuous at the matching radii $r_\pm$. This specific "quadratic-logarithmic" form allows for exact analytical solutions.

• Mapping to Schrödinger Equation:
  The Fokker-Planck equation governing the probability distribution $P(\mathbf{r}, t)$ is mapped to a Schrödinger-type eigenvalue problem via a similarity transformation:
  $$P(\mathbf{r}, t) = e^{-\beta V(\mathbf{r})/2} \psi(\mathbf{r}, t)$$
  This transforms the non-Hermitian Fokker-Planck operator into a Hermitian Hamiltonian $H = -T\nabla^2 + V_S(r)$, where $V_S(r)$ is an effective potential.

• Analytical Solution:
  The eigenvalue problem is solved in each region using confluent hypergeometric functions (Kummer's $M$ and Tricomi's $U$ functions). The global eigenvalues and eigenmodes are determined by enforcing continuity of the wavefunction and its derivative at the matching radii $r_\pm$.

• Relaxation Analysis:
  The time evolution of the system is expanded in terms of eigenmodes. The authors focus on the coefficients $a_m$ of the eigenmodes, particularly the slowest relaxation modes ($m=2, 3$), to analyze the relaxation dynamics from an initial equilibrium state at $\beta_{ini}$ to a final bath at $\beta$.

3. Key Contributions

A. Exact Solvability in 2D

The paper provides the first analytically tractable 2D model of the Mpemba effect in a smooth potential. Unlike 1D models which often require hard walls to trap particles and induce the effect, this 2D model relies on the natural geometry of the radial coordinate.

B. The "Effective Wall" at the Origin

A crucial theoretical insight is the identification of an effective wall at $r=0$ in 2D systems.

• In 2D, the equilibrium measure includes a Jacobian factor $r$ (i.e., $P_{eq} \propto r e^{-\beta V(r)}$).
• This creates an effective potential $V_{eff}(r) = V(r) - T \ln r$, which diverges at $r=0$, preventing the particle from crossing the origin.
• This geometric constraint mimics the role of a hard wall in 1D, allowing the Mpemba effect to emerge without physical boundaries.

C. Rigorous Crossing Condition (KL Divergence)

The authors move beyond simple observable crossings (like mean position) and derive a sufficient condition for the Mpemba effect based on the Kullback-Leibler (KL) divergence, a monotone measure of distance to equilibrium.

• They derive a condition involving the amplitudes of the two slowest modes ($a_2$ and $a_3$):
  $$\frac{\Delta_3}{\Delta_2} < -1$$
  where $\Delta_n = (a_n^{(A)})^2 - (a_n^{(B)})^2$ for two different initial temperatures.
• This condition ensures that the initially hotter system (with a larger initial KL distance) crosses the trajectory of the colder system due to a specific suppression of the slowest mode amplitude.

4. Key Results

• Existence of Mpemba Effect:
  For a potential where the global minimum is far from the origin ($V(\alpha) < V(0)$), the authors demonstrate that the coefficient of the slowest mode, $a_2(\beta_{ini})$, exhibits a non-monotonic dependence on the initial inverse temperature $\beta_{ini}$. Specifically, $a_2$ has a peak at a critical $\beta^*_{ini}$.

  • Systems starting near this peak have a suppressed slowest mode, allowing them to relax faster than systems starting at lower temperatures.
  • Numerical verification confirms a crossing of the KL divergence trajectories, validating the Mpemba effect.

• Role of Potential Depth ($V(\alpha)$):
  The authors analyze three cases based on the relative depth of the two wells:

  1. $V(\alpha) < 0$ (Global min at outer well): Mpemba effect observed.
  2. $V(\alpha) = 0$ (Degenerate wells): A peak in $a_2$ exists, but the crossing condition is violated. Neither Mpemba nor inverse Mpemba effects are observed.
  3. $V(\alpha) > 0$ (Global min at origin): $a_2$ decreases monotonically with $\beta_{ini}$. No Mpemba effect is possible.

• Dimensionality Contrast:
  The study confirms that while 1D asymmetric potentials require a hard wall to observe the effect, 2D systems naturally possess the necessary confinement via the radial Jacobian, making the effect robust in unbounded 2D space.

5. Significance

• Theoretical Advancement: This work bridges the gap between discrete/1D models and realistic continuous 2D systems. It proves that the Mpemba effect is not an artifact of hard boundaries but a fundamental consequence of the interplay between potential landscape geometry and relaxation modes.
• Analytical Benchmark: By providing exact eigenvalues and eigenmodes, the model serves as a rigorous benchmark for testing numerical simulations and approximate theories of nonequilibrium relaxation.
• Generalization of Criteria: The derivation of the KL divergence crossing condition offers a robust, observable-independent criterion for identifying the Mpemba effect in complex systems, moving beyond specific "temperature-like" observables.
• Geometric Insight: The identification of the "effective wall" at the origin in 2D provides a new perspective on how dimensionality influences stochastic dynamics and metastability.

In summary, Hayakawa and Takada successfully demonstrate that the Mpemba effect is a generic feature of 2D bistable systems with smooth potentials, driven by the non-monotonic projection of initial states onto the slowest relaxation eigenmodes, and solvable without artificial confinement.