B. Eynard
Abstract:Exact integral expressions of the skew orthogonal polynomials involved in Orthogonal (beta=1) and Symplectic (beta=4) random matrix ensembles are obtained: the (even rank) skew orthogonal polynomials are average characteristic polynomials of random matrices. From there, asymptotics of the skew orthogonal polynomials are derived.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the asymptotic behavior of skew orthogonal polynomials in orthogonal ($\beta = 1$) and symplectic ($\beta = 4$) random matrix ensembles. Specifically, the author aims to derive the exact integral expressions of these skew orthogonal polynomials and extract their asymptotic properties from these expressions.
### Main problem background
Random Matrix Theory (RMT) has many applications in mathematics and physics, especially in the fields of disordered solid - state physics, quantum chromodynamics (QCD), statistical physics, etc. In these fields, researchers are interested in the spectral characteristics of matrices (such as Hamiltonians, transfer matrices, S - matrices, Dirac operators, etc.). For various reasons (such as disorder, random impurities, quantum fluctuations, chaos or non - integrability), these matrices can be regarded as random matrices. Research shows that the spectra of large random matrices have universal properties, such as short - range and long - range two - point correlation functions, universal conductance fluctuations in mesoscopic conductors, etc.
### Core problem of the paper
To understand and prove these universal properties, one possible method is through the "orthogonal polynomial" method. For different matrix ensembles, different types of orthogonal or skew orthogonal polynomials are involved. This paper focuses in particular on O(N) and Sp(2N) - invariant matrix ensembles, which appear in physical systems with time - reversal symmetry and/or half - integer spin and broken rotational symmetry. These problems involve families of skew orthogonal polynomials.
### Solution
The main contributions of the author include:
1. **Exact integral expressions**: Derived the exact integral expressions of skew orthogonal polynomials, showing that the even - order skew orthogonal polynomials are the average values of the characteristic polynomials of random matrices.
2. **Asymptotic behavior**: Derived the asymptotic behavior of skew orthogonal polynomials from the exact expressions, especially in the large $n$ limit and in the case where $n - n_F\sim O(1)$ close to the Fermi level.
### Formula summary
For different ensembles ($\beta = 1$ and $\beta = 4$), the exact integral expressions of skew orthogonal polynomials are as follows:
- **Orthogonal ensemble $\beta = 1$**:
\[
P^{(1)}_{2n}(x)=\frac{1}{Z^{(1)}_{2n}}\int_{M\in E^{(1)}_{2n}}dM\det(x - M)e^{-N\text{tr}V(M)}
\]
\[
P^{(1)}_{2n + 1}(x)=\frac{1}{Z^{(1)}_{2n}}\int_{M\in E^{(1)}_{2n}}dM(x+\text{tr}M + c_n)\det(x - M)e^{-N\text{tr}V(M)}
\]
- **Symplectic ensemble $\beta = 4$**:
\[
P^{(4)}_{2n}(x)=\frac{1}{Z^{(4)}_n}\int_{M\in E^{(4)}_n}dM\det(x - M)e^{-\frac{N}{2}\text{tr}V(M)}
\]
\[
P^{(4)}_{2n+1}(x)=\frac{1}{Z^{(4)}_n}\int_{M\in E^{(4)}_n}dM(x+\text{tr}M + c_n)\det(x - M)e^{-\frac{N}{2}\text{tr}V(M)}
\]
### Asymptotic results
For large $N$ and $n$, the asymptotic behavior of skew orthogonal polynomials can be described by introducing analytic functions (such as the resolvent function $W(z)$). The final results include:
- **Orthogonal ensemble $\beta = 1$**:
\[
P^{(1)}_{2n}(x)e^{-N