Optical and Hall conductivities of a thermally disordered two-dimensional spin-density wave: two-particle response in the pseudogap regime of electron-doped high-$T_c$ superconductors

TL;DR

This paper calculates the optical and Hall conductivities in a two-dimensional metal with thermally disordered antiferromagnetism, revealing behaviors consistent with electron-doped cuprates in the pseudogap regime.

Contribution

It extends the Lee-Rice-Anderson approximation to compute two-particle response functions in disordered antiferromagnetic states, incorporating vertex corrections and impurity scattering.

Findings

Conductivities satisfy the $f$-sum rule.

Results align qualitatively with experimental data on electron-doped cuprates.

Long-range order behaviors emerge for large but finite spin correlation length.

Abstract

We calculate the longitudinal ($\sigma_{xx}$) and Hall ($\sigma_{xy}$) optical conductivities for two-dimensional metals with thermally disordered antiferromagnetism using a generalization of an approximation introduced by Lee, Rice and Anderson for the self energy. The conductivities are calculated from the Kubo formula, with current vertex function treated in a conserving approximation satisfying the Ward identity. In order to obtain a finite DC limit, we introduce phenomenologically impurity scattering, with relaxation time $\tau$. $\sigma_{xx}(\Omega)$ satisfies the $f$-sum rule. For the infinitely peaked spin correlation function, $\chi(\mathbf{q})\propto \delta(\mathbf{q}-\mathbf{Q})$, we recover the expressions for the conductivities in the mean-field theory of the ordered state. When the spin correlation length $\xi$ is large but finite, both $\sigma_{xx}$ and $\sigma_{xy}$ show behaviors characteristic of the state with long-range order. The calculation runs into difficulty for $\Omega\lesssim 1/\tau$. The difficulties are traced to an inaccurate treatment of the very low energy density of states within the Lee-Rice-Anderson approximation. The results for $\sigma_{xx}(\Omega)$ and $\sigma_{xy}(\Omega)$ are qualitatively consistent with data on electron-doped cuprates when $\Omega>1/\tau$.