Thermal relaxation asymmetry persists under inertial… — Plain-Language Explanation

✨

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

The Big Picture: The "Hot vs. Cold" Race

Imagine you have two identical cups of coffee.

Cup A is boiling hot (90°C).
Cup B is ice cold (10°C).
The room temperature is a cozy 50°C.

You put both cups on the table at the same time. Common sense tells us they will both eventually reach 50°C. But here is the surprising discovery the scientists made: The hot cup cools down to 50°C slower than the cold cup heats up to 50°C.

Even though they start the same distance away from the target temperature (one is 40 degrees above, the other 40 degrees below), the cold cup "rushes" to catch up, while the hot cup "strolls" along. This is called Thermal Relaxation Asymmetry.

The Old Story vs. The New Discovery

The Old Story (Overdamped):
Scientists already knew this happened for things that move through thick syrup (like honey). In this "thick" world, friction is so strong that objects don't have momentum; they stop instantly when you stop pushing them. They proved that in this sticky world, heating is faster than cooling.

The New Story (Underdamped with Inertia):
This paper asks: What if the object isn't in syrup, but in air?
In air, things have inertia (momentum). If you push a ball, it keeps rolling even after you let go. It overshoots, bounces back, and wobbles before settling. This is "underdamped" motion.

The big question was: Does the "hot vs. cold" speed difference still happen when things have momentum and bounce around?

The Answer: Yes! The authors proved mathematically that even when objects are bouncing, spinning, and carrying momentum, the cold object still heats up faster than the hot object cools down.

The Metaphor: The Bouncing Ball in a Valley

To understand why this happens, imagine a ball in a valley (a potential energy landscape).

The Setup: The bottom of the valley is the "target temperature."
The Cold Ball: You drop the ball from a high cliff on the left side. It falls down, gains speed, and smashes into the bottom. Because it has inertia, it doesn't stop immediately; it shoots up the other side, slows down, and rolls back. It takes a chaotic, energetic path to settle.
The Hot Ball: You drop the ball from a high cliff on the right side. It also falls, smashes, and bounces.

The Twist:
The paper shows that the "path" the cold ball takes to get to the bottom is more efficient than the hot ball's path.

Think of it like a dance. When the system is cold, the "dance moves" (the way the ball moves and spins) align perfectly to rush toward the target.
When the system is hot, the "dance moves" are a bit more chaotic and inefficient. The momentum causes the ball to overshoot the target more aggressively, wasting energy in the process of trying to settle.

The authors call this "Thermal Kinematics." It's like saying the cold system has a "head start" in its velocity, allowing it to overshoot the target quickly and settle faster, whereas the hot system gets stuck in a loop of overshooting and correcting.

The "Ghost" of Velocity

One of the most interesting parts of the paper is what happens when we try to ignore the speed (velocity) of the ball and only look at its position (where it is).

Usually, in physics, if you have a very heavy ball in thick syrup, we say "forget about the speed, it's zero." We just look at the position.

The Paper's Surprise: Even if you pretend the speed doesn't exist (the "overdamped" limit), the memory of that speed still matters.
The Analogy: Imagine you are trying to calculate how much energy it took to move a car. If you only look at the distance traveled but ignore the fact that the car was speeding up and slowing down, your math will be wrong. The "ghost" of the velocity contributes to the energy budget even if you can't see the speed anymore.

The authors found that if you ignore the velocity, you miss a hidden chunk of energy that affects how fast the system relaxes. It's like trying to explain a car crash without mentioning that the car was moving; you can describe the damage, but you can't explain why it happened the way it did.

Why Does This Matter?

It's Universal: This isn't just about coffee cups or balls. It applies to tiny particles, biological molecules, and even complex systems driven by external forces (like a motor).
No "Local Equilibrium": Usually, we assume that if a system is relaxing, it passes through a series of calm, balanced states. This paper shows that it doesn't. The system is always in a state of "chaotic flux," mixing position and speed in complex ways that never look like a calm, balanced state until the very end.
Better Engineering: If we want to build tiny heat engines (machines that run on temperature differences) or cool down computer chips faster, we need to understand this asymmetry. Knowing that "cold heats up faster" helps us design better protocols to control temperature.

Summary in One Sentence

Even when objects are bouncing and spinning with momentum (inertia), nature still has a bias: cold things heat up faster than hot things cool down, and this happens because the "dance" of position and speed is more efficient for heating than for cooling.

1. Problem Statement

The paper addresses the phenomenon of thermal relaxation asymmetry, specifically the observation that systems heated from a lower temperature ( $T_c$ ) to an ambient temperature ( $T$ ) relax faster than systems cooled from a higher temperature ( $T_h$ ) to the same $T$ , provided the initial conditions are thermodynamically equidistant (TE).

Context: Previous work established this asymmetry for overdamped dynamics (where inertia is neglected, e.g., Ornstein-Uhlenbeck processes in position space).
The Gap: It remained an open question whether this asymmetry persists when inertial effects are included (i.e., in underdamped dynamics where both position $r$ and velocity $v$ are dynamic variables).
Specific Challenge: In underdamped systems, the coupling between position and velocity creates complex phase-space dynamics. Furthermore, the definition of "temperature quenches" and the behavior of the overdamped limit (where velocity relaxes instantaneously) introduce ambiguities regarding free energy contributions from velocity degrees of freedom.

2. Methodology

The authors employ a rigorous algebraic approach based on stochastic thermodynamics and linear response theory for Gaussian processes.

Framework: They model the system using Underdamped Langevin Dynamics (Ornstein-Uhlenbeck Process in phase space). The state vector is $x = (r, v)^T$ .
Quantification of Relaxation: They utilize the Excess Free Energy (EFE), defined as the Kullback-Leibler (KL) divergence between the time-dependent probability density $\rho(x,t)$ and the steady-state density $\rho_s(x)$ :
$D(t) = k_B T \int dx \, \rho(x,t) \ln \left[ \frac{\rho(x,t)}{\rho_s(x)} \right]$
For equilibrium relaxation, $D(t)$ corresponds to the total entropy production. For non-equilibrium steady states (NESS), it represents the non-adiabatic part of entropy production.
Thermodynamically Equidistant (TE) Conditions: The initial conditions for heating ( $T_c$ ) and cooling ( $T_h$ ) are chosen such that the initial excess free energies are equal: $D_c(0) = D_h(0)$ .
Mathematical Strategy:
1. Formulate the dynamics as a linear Stochastic Differential Equation (SDE): $dx_t = -Ax_t dt + \sigma dW_t$ .
2. Express the KL divergence $D(t)$ analytically in terms of the covariance matrix $\Sigma(t)$ , leveraging the fact that the process remains Gaussian.
3. Define the asymmetry metric $\Delta D(t) = D_h(t) - D_c(t)$ .
4. Prove that $\Delta D(t) > 0$ for all $t > 0$ using matrix analysis, specifically analyzing the eigenvalues of a transformed covariance matrix $X(t)$ .
5. Investigate the overdamped limit ( $m/\gamma \to 0$ ) and marginalized dynamics (projecting onto position or velocity alone).

3. Key Contributions

A. Generalization to Underdamped Dynamics (Theorem 1)

The authors prove that the thermal relaxation asymmetry holds for the entire range from overdamped to underdamped dynamics.

Result: For any linearly driven system approaching an equilibrium or NESS with a positive-definite steady-state covariance, the heating process ( $T_c \to T$ ) is strictly faster than the cooling process ( $T_h \to T$ ) in terms of excess free energy decay.
Significance: This confirms that the asymmetry is a fundamental feature of non-equilibrium relaxation in phase space, not an artifact of neglecting inertia. It highlights that position-velocity coupling does not restore symmetry but rather underscores the intricate, non-local-equilibrium nature of the relaxation path.

B. Resolution of the Overdamped Limit Ambiguity

The paper clarifies a subtle issue in the transition from underdamped to overdamped descriptions.

Finding: In the standard overdamped literature, the free energy contribution from the relaxation of velocity degrees of freedom (from variance $\propto T_i$ to $\propto T$ ) is often implicitly assumed to vanish or occur instantaneously before the quench.
Insight: The authors show that this contribution does not trivially vanish. The "excess free energy" in the overdamped limit depends on the precise interpretation of the temperature quench. If the velocity relaxation is considered part of the process, it contributes to the initial free energy landscape. However, because TE conditions are defined based on the total initial free energy, the asymmetry result remains robust across both interpretations, provided the time scales are consistent.

C. Marginalization and Projections (Theorem 2)

The authors investigate whether the asymmetry persists if one only observes a subset of variables (e.g., only position $r$ or only velocity $v$ ).

Result: Even when the phase-space density is marginalized (projected) onto position or velocity subspaces, the thermal relaxation asymmetry ( $\Delta D_{marg}(t) \geq 0$ ) persists.
Mechanism: This is proven by showing that the eigenvalues of the projected covariance matrices remain within the interval $(0, 1]$ , ensuring the algebraic inequality holds.

4. Key Results and Proofs

Algebraic Proof: The core of the proof relies on showing that for the matrix $X(t) = e^{-At}\Sigma_s e^{-A^T t} \Sigma_s^{-1}$ $X (t) = e^{- A t} Σ_{s} e^{- A^{T} t} Σ_{s}^{- 1}$ , the eigenvalues $x_j(t)$ $x_{j} (t)$ satisfy $0 < x_j(t) \leq 1$ $0 < x_{j} (t) \leq 1$ .
- The condition $x_j(t) \leq 1$ is derived using the log-norm inequality and the positive semi-definiteness of the diffusion matrix.
- The strict inequality $x_j(t) < 1$ for at least one $j$ at any $t > 0$ ensures that $\Delta D(t) > 0$ .
Phase-Space Geometry: The study reveals that during relaxation, the probability density in phase space does not pass through intermediate states of defined temperature. Instead, it exhibits transient correlations between position and velocity (e.g., $r$ - $v$ correlations) that are absent in the steady state.
Driven Systems: The results apply to systems linearly driven into Non-Equilibrium Steady States (NESS), including those with rotational drift (e.g., harmonic potentials with a magnetic-field-like term $\omega$ ), provided the system remains confined (eigenvalues of the drift matrix have positive real parts).

5. Significance and Implications

Universality of Asymmetry: The finding establishes that the "heating is faster than cooling" rule is robust against the inclusion of inertia. This suggests the asymmetry is a deep property of stochastic thermodynamics driven by the geometry of the probability space, rather than a specific feature of overdamped diffusion.
Non-Local Equilibrium: The persistence of asymmetry in underdamped systems reinforces the concept that far-from-equilibrium relaxation does not proceed through a sequence of local equilibria. The coupling of position and velocity creates complex trajectories that cannot be described by simple effective temperatures.
Refining Overdamped Models: The work provides a necessary correction to how free energy and temperature quenches are interpreted in overdamped models. It suggests that for accurate thermodynamic accounting (e.g., in Brownian heat engines), one must explicitly consider the velocity degrees of freedom, even if the dynamics are effectively overdamped.
Future Directions: The authors propose that this phase-space framework could be used to generalize thermodynamic uncertainty relations (TURs) and speed limits to underdamped systems. They also identify the geometry of level sets in phase space (rotations during heating vs. cooling) as a promising area for further "thermal kinematics" research.

In summary, the paper provides a rigorous algebraic proof that thermal relaxation asymmetry is a fundamental feature of linear stochastic systems, persisting regardless of inertial effects, and clarifies the thermodynamic subtleties involved in the overdamped limit.

Thermal relaxation asymmetry persists under inertial effects