Random Magnetic Field and the Dirac Fermi Surface

TL;DR

This paper analyzes a 2D Dirac fermion at finite density under a quenched random magnetic field, mapping it to 1D chiral fermions, and computes disorder-averaged observables along an exactly solvable fixed line.

Contribution

It introduces an exactly solvable fixed line for a disordered 2D Dirac fermion system, enabling direct calculation of observables without replica or supersymmetry methods.

Findings

Longitudinal dc conductivity is nonuniversal.

Conductivity varies continuously along the fixed line.

The low-energy theory maps to coupled 1D chiral fermions.

Abstract

We study a single 2d Dirac fermion at finite density, subject to a quenched random magnetic field. At low energies and sufficiently weak disorder, the theory maps onto an infinite collection of 1d chiral fermions (associated to each point on the Fermi surface) coupled by a random vector potential. This low-energy theory exhibits an exactly solvable random fixed line, along which we directly compute various disorder-averaged observables without the need for the usual replica, supersymmetry, or Keldysh techniques. We find the longitudinal dc conductivity in the collisionless $\hbar \omega/k_B T \rightarrow \infty$ limit to be nonuniversal and to vary continuously along the fixed line.