Conductivity of a 2DEG in Si/SiGe heterostructure near metal- insulator transition: role of the short and long range scattering potential

TL;DR

This study investigates the metal-insulator transition in a 2D electron gas within Si/SiGe heterostructures, examining the roles of short and long-range scattering potentials and comparing experimental results with existing theories.

Contribution

It provides new insights into the scattering mechanisms affecting conductivity near the MIT, highlighting limitations of current theories in describing the full scattering potential.

Findings

Observation of a zero-field MIT in Si/SiGe heterostructures.

Discrepancy between experimental magnetoresistance and existing theories.

Importance of considering both short and long-range scattering potentials.

Abstract

We report the observation of a metal-insulator transition (MIT) in a two- dimensional electron gas (2DEG) in a Si/SiGe heterostructure at zero magnetic field. On going through the MIT we observe the corresponding evolution of the magnetic field induced transition between the insulating phase and the quantum Hall (QH) liquid state in the QH regime. Similar to the previous reports for a GaAs sample, we find that the critical magnetic field needed to produce the transition becomes zero at the critical electron density corresponding to the zero field MIT. The temperature dependence of the conductivity in a metallic-like state at zero field is compared with the theory of the interaction corrections at intermediate and ballistic regimes $k_{B}T\tau/\hbar\geq1$. The theory yields a good fit for the linear part of the curve. However the slope of that part of $\sigma_{xx}(T)$ is about two times smaller than that reported in other 2D systems with similar values of $r_s$. At the same time, the recent theory of magnetoresistance due to electron-electron interaction in the case of arbitrary $k_{B}T\tau/\hbar$, smooth disorder and classically strong fields does not seem to be quite adequate for the description of the parabolic magnetoresistance observed in our samples. We attribute these results to the fact that neither of these theories deals with the whole scattering potential in a sample but leaves either its long range or its short range component out of consideration.