[Paper Review] Diffusive scaling in energy ginzburg-Landau dynamics

Carlangelo Liverani, Stefano Olla+1 more|arXiv (Cornell University)|Sep 21, 2015

Advanced Thermodynamics and Statistical Mechanics5 references3 citations

TL;DR

This paper establishes that energy fluctuations in a stochastic Ginzburg-Landau dynamics model, derived from weakly coupled anharmonic oscillators, converge to the solution of the linearized heat equation under diffusive space-time scaling. Using a non-gradient approach adapted to positive energy states and linearly growing potentials, the authors prove hydrodynamic limit convergence under a spectral gap condition, which is verified for stochastic perturbations but remains open for deterministic cases.

ABSTRACT

Ginzburg-Landau energy models arise as autonomous sto-chastic dynamics for the energies in coupled systems after a weak coupling limit (cf. [3, 6]). We prove here that, under certain conditions, the energy fluctuations of these stochastic dynamics are driven by the heat equation, under a diffusive space time scaling.

Motivation & Objective

• To establish the hydrodynamic limit of energy fluctuations in a stochastic Ginzburg-Landau dynamics model derived from weakly coupled systems.
• To show that under diffusive space-time scaling, the energy fluctuations converge to the solution of the linearized heat equation.
• To adapt Varadhan’s non-gradient method to the Ginzburg-Landau dynamics with positive energy constraints and linearly growing potentials.
• To verify the spectral gap condition required for the convergence, specifically for the stochastic perturbation case.
• To clarify the two-step derivation of the heat equation: first weak-coupling limit, then diffusive hydrodynamic limit.

Proposed method

• Formalizes the energy Ginzburg-Landau dynamics via stochastic differential equations with energy exchange currents driven by local energy variances.
• Defines the generator of the dynamics as a sum of local generators involving drift and diffusion coefficients related to the energy potential U and the diffusion function γ².
• Uses reversible Gibbs measures with potential U(ℰ) ∼ log ℰ for small and large ℰ, ensuring proper stationary distributions.
• Applies the non-gradient method of Varadhan by introducing a corrector function to handle non-gradient fluctuations.
• Employs spectral gap estimates on finite systems to control the variance of energy fluctuations and ensure convergence.
• Uses Boltzmann-Gibbs principle to replace local functions by their conditional expectations, reducing the fluctuation problem to a martingale estimate.

Experimental results

Research questions

• RQ1 Does the energy fluctuation process in the weakly coupled Ginzburg-Landau dynamics converge to the linearized heat equation under diffusive scaling?
• RQ2 Can the non-gradient method be adapted to Ginzburg-Landau dynamics with positive energy states and logarithmic potentials?
• RQ3 Is the spectral gap of the finite dynamics generator bounded uniformly in system size, enabling the hydrodynamic limit?
• RQ4 What is the role of the microcanonical conditional expectation in the Boltzmann-Gibbs principle for this model?
• RQ5 How does the convergence depend on the choice of the diffusion coefficient γ² and the potential U(ℰ)?

Key findings

• The energy fluctuations of the Ginzburg-Landau dynamics converge to the solution of the linearized heat equation under diffusive scaling, under a spectral gap condition.
• The spectral gap condition is verified for the stochastic perturbation case studied in [6], ensuring the convergence result holds in that setting.
• The Boltzmann-Gibbs principle is applied successfully by showing that the conditional expectation of the fluctuation function is small in L² norm, uniformly in system size.
• The variance of the energy fluctuation process is bounded by Ck³, and the conditional expectation of the corrector function is O(k⁻¹), ensuring convergence.
• The proof relies on the fact that the microcanonical conditional expectation of the function φ₀ is O((ℰ̄ₖ − ē)² + k⁻¹), which leads to uniform L² bounds.
• The key technical step is the estimate of the second-order term in the Boltzmann-Gibbs expansion, which vanishes as k → ∞ due to the spectral gap and decay in the sum.