Asymptotic Behavior of Thermal Nonequilibrium Steady States for a Driven Chain of Anharmonic Oscillators

Luc Rey-Bellet, Lawrence E. ThomasCommunications 2000 · DOI · arXiv

Introduction

In classical equilibrium statistical mechanics, the Boltzmann-Gibbs distribution serves as a fundamental "cheat code" for describing the state of a system. If a system is in thermal equilibrium at a uniform temperature $T$, the probability of finding it in a state with energy $H$ is proportional to $\exp(-H/T)$. However, this elegant framework collapses when a system is driven out of equilibrium—for instance, a wire connected to a heat source at one end and a heat sink at the other.

In such Non-Equilibrium Steady States (NESS), a constant flow of energy prevents the system from reaching a standard equilibrium distribution. This paper, published in 2000, provides a rigorous mathematical treatment of NESS for a specific, physically realistic model: a chain of $n$ anharmonic (nonlinear) oscillators coupled to two heat reservoirs at different temperatures ($T_L$ and $T_R$). The authors establish a variational principle that describes the steady-state probability distribution as the average temperature approaches zero, effectively finding a non-equilibrium analogue to the Boltzmann distribution.

The Model and Methodology

Stochastic Dynamics

The authors model the system as a Hamiltonian chain of $n$ particles. To simulate the thermal environment, the chain is coupled at its boundaries to two heat reservoirs, which are modeled by wave equations. This complex setup is mathematically reduced to a system of Markovian Stochastic Differential Equations (SDEs) on an extended phase space.

Extending Large Deviation Theory

The primary analytical tool used is Freidlin-Wentzell large deviation theory, which describes how small random perturbations influence dynamical systems. A significant challenge in this specific model is that the noise is "degenerate" or "thin"; it only enters the system at the first and last oscillators. Standard theory typically requires noise to hit every variable directly.

To overcome this, the authors employ tools from geometric control theory and the concept of hypoellipticity. They prove that the nonlinearity (anharmonicity) of the oscillator chain allows the thermal fluctuations from the boundaries to "spread" throughout the entire chain. This ensures that the boundary temperatures influence the internal state of every oscillator, allowing for the extension of large deviation results to this degenerate case.

The Variational Principle and the Action Function

The central insight of the paper is the definition of an "action" function, $W(x)$. This function represents the minimum "cost"—in terms of probability—required for the system to move from a stable equilibrium to a specific state $x$ due to thermal fluctuations.

The authors demonstrate that the steady-state distribution $\mu$ behaves asymptotically like $\exp(-W/T)$ to leading order as the mean temperature $T$ approaches zero. This $W$ is determined by a variational principle: it is the infimum of a rate function over all possible paths the system could take through its phase space.

Key Results

1. Rigorous Low-Temperature Limits

The authors provide a formal proof (Theorems 1.4 and 3.3) for the existence of the low-temperature limit for the invariant measure. Specifically, they show: $$\lim_{\epsilon \to 0} \epsilon \log \mu_\epsilon(D) = -\inf_{x \in D} W(x)$$ where $\epsilon$ represents the mean temperature of the reservoirs. This confirms that the "action" $W$ dictates the most likely states of the system in the cold limit.

2. Recovery of Equilibrium

A critical test for any non-equilibrium theory is its consistency with known equilibrium states. The authors show that if the temperature difference between the two reservoirs is zero ($T_L = T_R$), their formula for $W$ collapses perfectly back into the standard Boltzmann-Gibbs distribution.

3. The Role of Nonlinearity and Entropy

The paper highlights that anharmonicity is not just a detail but a requirement; it is the mechanism that allows the system to reach a steady state and ensures the internal oscillators "feel" the temperature gradient. Furthermore, they relate the probabilistic "cost" of a path in their framework to the physical entropy production along that path, bridging the gap between stochastic analysis and thermodynamics.

Discussion and Limitations

While this work provides a robust mathematical foundation for NESS, it is bounded by several theoretical constraints:

  • Asymptotic Scope: The results are strictly valid as $T \to 0$. At higher temperatures, the system may explore more of the phase space in ways not captured by the "skeleton" of the distribution provided by $W$.
  • Convexity Requirements: The proofs assume the 2-body potential of the oscillators is strictly convex. This excludes systems with more complex phase transitions or specific types of molecular bonds that do not follow this potential shape.
  • Computational Complexity: Although the variational principle is mathematically explicit, calculating $W$ for a chain with a large number of oscillators ($n \gg 1$) is computationally intensive, limiting its immediate application in numerical simulations of large-scale materials.

In conclusion, this paper successfully extends the reach of large deviation theory to non-equilibrium physics, providing a rigorous path toward understanding how heat and energy flow through microscopic structures at the limit of absolute zero.

Summary updated

BibTeX
@inproceedings{reybellet2000asymptotic,
  author = {Luc Rey-Bellet and Lawrence E. Thomas},
  title = {{Asymptotic Behavior of Thermal Nonequilibrium Steady States for a Driven Chain of Anharmonic Oscillators}},
  booktitle = {Communications},
  year = {2000},
  doi = {10.1007/s002200000285},
}
Figure