Abstract:The paper continues the development of the rigorous supersymmetric transfer
matrix approach to the random band matrices started in J Stat Phys 164:1233 --
1260, 2016; Commun Math Phys 351:1009 -- 1044, 2017. We consider random
Hermitian block band matrices consisting of $W\times W$ random Gaussian blocks
(parametrized by $j,k \in\Lambda=[1,n]^d\cap \mathbb{Z}^d$) with a fixed
entry's variance $J_{jk}=\delta_{j,k}W^{-1}+\beta\Delta_{j,k}W^{-2}$, $\beta>0$
in each block. Taking the limit $W\to\infty$ with fixed $n$ and $\beta$, we
derive the sigma-model approximation of the second correlation function similar
to Efetov's one. Then, considering the limit $\beta, n\to\infty$, we prove that
in the dimension $d=1$ the behaviour of the sigma-model approximation in the
bulk of the spectrum, as $\beta\gg n$, is determined by the classical Wigner --
Dyson statistics.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the universal behavior of one - dimensional random band matrices (1D Random Band Matrices, 1D RBM) in the spectral central region. Specifically, the author studied the behavior of 1D RBM in the spectral central region and, through the σ - model approximation method, proved that when the bandwidth \(W\) and the system size \(n\) tend to infinity and \(\beta \gg n\), the second correlation function of 1D RBM is consistent with the local spectral statistics of the Gaussian unitary ensemble (GUE).
### Background of the Paper
Random band matrices (RBM) are a model used to describe quantum systems with random quantum transition amplitudes in large boxes. These matrices can be regarded as a transition model from Wigner matrices to random Schrödinger operators. Wigner matrices have no spatial structure, while random Schrödinger operators contain a deterministic Laplacian operator and a random diagonal potential term.
### Research Questions
The main research questions in the paper are:
1. **Universality**: In 1D RBM, does the local spectral statistics in the spectral central region show universality?
2. **σ - model approximation**: How to study the second correlation function of 1D RBM through the σ - model approximation method?
### Main Results
The author solved the above problems through the following steps:
1. **σ - model approximation**: The author derived the σ - model approximation expression of the second correlation function of 1D RBM.
2. **Universality proof**: By analyzing the σ - model approximation, the author proved that when \(\beta \gg n\), the behavior of the second correlation function of 1D RBM in the spectral central region is consistent with the local spectral statistics of GUE.
### Formula Summary
The key formulas involved in the paper include:
- **Density function**:
\[
\rho(E)=\frac{1}{2\pi}\sqrt{4 - E^{2}}, \quad E\in[- 2,2]
\]
- **Second correlation function**:
\[
R_{2}(\lambda_{1},\lambda_{2})=\frac{1}{(2i\pi N)^{2}}\lim_{\epsilon\rightarrow0}\mathbb{E}\left\{\Im\text{Tr}\left(H - \lambda_{1}-i\epsilon\right)^{-1}\Im\text{Tr}\left(H - \lambda_{2}-i\epsilon\right)^{-1}\right\}
\]
- **σ - model approximation**:
\[
R_{2}^{(n\beta)}(E,\epsilon,\xi)=C_{E,\xi}\int\exp\left\{\frac{\tilde{\beta}}{4}\sum\text{Str}Q_{j}Q_{j - 1}-\frac{c_{0}}{2|\Lambda|}\sum\text{Str}Q_{j}\Lambda_{\xi,\epsilon}\right\}dQ
\]
where:
\[
\tilde{\beta}=(2\pi\rho(E))^{2}\beta, \quad U_{j}\in\dot{U}(2), \quad S_{j}\in\dot{U}(1,1)
\]
\[
C_{E,\xi}=e^{E(\xi_{1}+\xi_{2}-\xi'_{1}-\xi'_{2})/2\rho(E)}
\]
\[
Q_{j}=\begin{pmatrix}U_{j}^*&0\\0&S_{j}^{-1}\end{pmatrix}\begin{pmatrix}I + 2\hat{\rho}_{j}\hat{\tau}_{j}L&2\hat{\rho}_{j}\\2\hat{\tau}
\]