Reliable modeling of weak antilocalization for accurate spin-lifetime extraction

We examine models for the magneto-conductivity correction in 2DEGs with both Rashba and Dresselhaus spin-orbit coupling (SOC) for their applicability to experimental data fitting. In particular, we compare the Landau-quantized Cooperon approach, which is mostly only numerically treatable, and the quasi-classical approximation that was recently employed to obtain an explicit solution for arbitrary Rashba and Dresselhaus SOC [PRL 112, 156601 (2019)]. It is found that the quasi-classical approximation yields significantly different results even to lowest order in the magnetic field and appears unsuitable for reliable parameter fitting. The discrepancy emerges when a sum over Landau levels is replaced by an integral over wave vectors. Substantial improvement is achieved by supplementing the quasi-classical model with the first two corrections given by the Euler-MacLaurin formula. Corresponding modifications are, however, only feasible in special SOC parameter configurations where the mixing of Landau bands is negligible and a closed-form solution that accounts for Landau quantization is also available. Such a scenario appears in a parameter regime where a persistent spin helix emerges and a transition between weak anti- and weak localization takes place. Combining recent findings, we derive a generalized closed-form expression for the magneto-conductivity correction applicable to generic 2DEGs that are grown along a crystal direction with at least two growth-direction Miller indices equal in modulus. The result is a function of spin lifetimes of the long-lived spin textures and is valid close to the persistent-spin-helix regime. The accuracy of the derived formula is validated by comparing with results from numerical diagonalization of the multiband Cooperon as well as a recently established Monte-Carlo-based real-space simulation in exemplary (001)-, (113)-, and (110)-2DEGs.
Comment: 18 pages, 7 figures