Giant Negative Magnetoresistance Driven by Spin-Orbit Coupling at the ${\mathrm{LaAlO}}_3/{\mathrm{SrTiO}}_3$ Interface

The ${\mathrm{LaAlO}}_3/{\mathrm{SrTiO}}_3$ interface hosts a two-dimensional electron system that is unusually sensitive to the application of an in-plane magnetic field. Low-temperature experiments have revealed a giant negative magnetoresistance (dropping by 70%), attributed to a magnetic-field induced transition between interacting phases of conduction electrons with Kondo-screened magnetic impurities. Here we report on experiments over a broad temperature range, showing the persistence of the magnetoresistance up to the 20 K range—indicative of a single-particle mechanism. Motivated by a striking correspondence between the temperature and carrier density dependence of our magnetoresistance measurements we propose an alternative explanation. Working in the framework of semiclassical Boltzmann transport theory we demonstrate that the combination of spin-orbit coupling and scattering from finite-range impurities can explain the observed magnitude of the negative magnetoresistance, as well as the temperature and electron density dependence.

Figure 1

(a) Measured magnetoresistance at $T=1.4\mathrm{K}$ for different gate voltages (left panel) and at $V_G=50\mathrm{V}$ for various temperatures (right panel). Inset: Schematic drawing of the device in a Hall bar geometry (in-plane field perpendicular to current direction), showing the source $S$, drain $D$, longitudinal voltage $V^{xx}$, transverse voltage $V^{xy}$ and gate voltage $V_G$. (b) Magnetoresistance calculated from the Boltzmann equation, at fixed $T=1.4\mathrm{K}$ (left panel) and at fixed $n=2.2\times 10^{13}{\mathrm{cm}}^{-2}$ (right panel).

Figure 2

(a) Dispersion relation for the mobile electrons at the ${\mathrm{LaAlO}}_3/{\mathrm{SrTiO}}_3$ interface, calculated from the model Hamiltonian (2) for $n=2.2\times 10^{13}{\mathrm{cm}}^{-2}$ at $B=0$ (solid line) and $B=12\mathrm{T}$ (dashed line). Colors indicate the orbital character of the bands. (b) Corresponding Fermi surfaces when the chemical potential is located at the “sweet spot” above the Lifshitz point where the system becomes very sensitive to changes in carrier density and magnetic field.

Figure 3

Energy-dependent density of states (a) and temperature-dependent chemical potential (b), calculated from the three-band Hamiltonian (2). Both quantities are shown for the sweet-spot carrier density $n=2.2\times 10^{13}{\mathrm{cm}}^{-2}$, at $B=0$ (solid line) and $B=12\mathrm{T}$ (dashed line).

Figure 4

Measured (a) and calculated (b) magnetoresistance at 1.4 K for different densities or gate voltages as a function of the rescaled magnetic field $B/B^{\star }$. The characteristic field $B^{\star }$ is chosen such that the rescaled curves all pass through the point with $\mathrm{MR}=-0.05$.