Limit theorems for basic states of the Anderson model (Q1320689)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Limit theorems for basic states of the Anderson model
scientific article

    Statements

    Limit theorems for basic states of the Anderson model (English)
    0 references
    0 references
    18 May 1994
    0 references
    Let \(V=\{x=(x^ 1,\dots,x^ \nu) \in \mathbb{Z}^ \nu:| x^ i | \leq n\), \(i=1, \dots, \nu\}\), and let \(N=(2n+1)^ \nu\) be the number of points in \(V\), \(\nu \geq 1\). Let us consider the following Hamiltonian in \(L^ 2(V)\): \[ H_ V=\kappa \Delta_ V+\xi (x),\quad x \in V, \tag{1} \] where \(\Delta_ V\) is the discrete Laplacian on \(V\) with zero Dirichlet boundary conditions (the restriction of the operator \(\Delta \varphi (x)=\sum_{| y-x |=1} \varphi (y)\), \(x \in \mathbb{Z}^ \nu\), to \(V)\), \(| x |=| x^ 1 |+\cdots+| x^ \nu |\), \(\xi (x)\), \(x \in \mathbb{Z}^ \nu\), is a random field of independent identically distributed variables, \(\kappa\) is a positive constant. Let \(\lambda_{1,N}>\cdots>\lambda_{N,N}\) be (random) eigenvalues of the operator (1). This paper presents limit theorems for the first (normalized) eigenvalues \(\lambda_{1,N},\dots,\lambda_{i,N}\) under the following assumption on the distribution density \(p(t):=(d/dt) P(\xi (0)<t)\): \[ p(t)=\bigl( A \alpha+o(1) \bigr) t^{\alpha-\beta-1} \exp (- At^ \alpha) \quad (t \to \infty), \tag{2} \] where \(\alpha>0\), \(A>0\), \(- \infty<\beta<\infty\). In the theory of Lifshits tails, this class of distributions is of special importance due to existence of bifurcations with respect to \(\alpha\) in the Anderson model (1).
    0 references
    0 references
    0 references
    0 references
    0 references
    discrete Laplacian
    0 references
    Dirichlet boundary conditions
    0 references
    limit theorems
    0 references
    Lifshits tails
    0 references
    Anderson model
    0 references