Equilibrium fluctuations for \(\nabla\varphi\) interface model (Q1872221)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Equilibrium fluctuations for \(\nabla\varphi\) interface model
scientific article

    Statements

    Equilibrium fluctuations for \(\nabla\varphi\) interface model (English)
    0 references
    0 references
    0 references
    0 references
    6 May 2003
    0 references
    The subject of the paper are the space-time fluctuations of the shape of an interface (for instance a phase boundary), which is flat in equilibrium. The deviation from the flat interface (the height) is modeled by a massless scalar field with reversible Langevin dynamics. Neighboring heights are connected through an elastic potential. For a strictly convex elastic potential it is proved that on a large space-time scale these fluctuations are governed by an infinite-dimensional Ornstein-Uhlenbeck process. Its effective diffusion type covariance matrix is characterized through a variational formula. The paper uses the following techniques: 1. An infinite-dimensional version of an identity of Helffer and Sjöstrand is proved, expressing static correlations as the solution of suitable elliptic partial differential equations. Rewriting them in terms of diffusion processes, a random walk in a dynamic environment is obtained. 2. This representation translates the original fluctuation problem into a problem involving an invariance principle for a certain random walk in a dynamic random environment. The convergence of this random walk to a Brownian motion in \(R^d\) with a specific diffusion matrix is proved. 3. Applying Nash inequality and a generalization of it, it is shown that the dynamic covariance has a limit as \(\varepsilon \to 0.\) 4. Finally, the convergence of the processes is established, using the convergence of the covariance.
    0 references
    0 references
    fluctuations
    0 references
    interface
    0 references
    stochastic process
    0 references
    covariance matrix
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references