Joule heating of dilute 2D holes in a GaAs quantum well

TL;DR

This study measures Joule heating effects in a 2D hole gas within a GaAs quantum well, revealing how heating power density relates to temperature and scattering, and comparing results with theoretical predictions.

Contribution

It provides the first detailed measurements of Joule heating in dilute 2D holes in GaAs, showing the power-temperature relationship across the metal-insulator transition.

Findings

P ~ T^2 in insulating phase at low T

P ~ T^4 at higher T in insulating phase

P ~ T^5 or T^6 in metallic phase

Abstract

We present measurements of the Joule heating of a 2D hole gas (2DHG) formed in a 30nm GaAs quantum well. The hole density is in the range (4.6-18.9)*10^9cm^-2 and exhibits an apparent metal-to-insulator transition (MIT) with a critical density 6*10^9 cm^-2. In the limit of zero heating power density P, the GaAs lattice is within 2 mK of the 6 mK base temperature of our dilution refrigerator determined by He-3 melting curve thermometry. Throughout the range of heating power densities used (1 to 10^6 fW/cm^2), the temperature rise of the lattice is estimated to be negligible compared to the temperature rise of the hole gas. We argue that the hole scattering rate is only a function of the hole temperature, with little dependence on the lattice or impurity temperatures in the relevant temperature range below 150 mK. We have therefore made measurements of the hole resistivity at negligible heating power density (P<5fW/cm^2) as a function of measured lattice temperature in the range 6 to 150 mK. We then use the hole resistivity measured in a cold lattice to estimate the temperature of the 2D hole gas as a function of P. In the low hole density insulating phase, the heating power density P(T) that heats the hole gas to a temperature T exhibits a dependence P ~ T^2 for T<30mK, gradually changing to P ~ T^4 for higher temperatures. On the metallic side of the MIT, P is proportional to T^5 or T^6 with a magnitude roughly 4 times less than the power density required to heat the insulating phase to the same temperature. Our measurements are within a factor of two of the available quantitative theoretical predictions for the hole energy loss rate as a function of temperature.