Magnetotransport Spectroscopy of Strongly Rashba-Split Hole Subbands Reveals Many-Body Interactions

Abstract

We report the results of magnetotransport experiments carried out on low-disorder 2D hole gases (2DHG) in the strongly correlated liquid regime, hosted in dopant-free (100) GaAs/AlGaAs single heterojunctions. Over a wide range of 2DHG densities (from 0.7 $\times$ 10$^{15}$/m$^2$ to $2 \times 10^{15}$/m$^2$), Fourier analysis of low-field (B < 1 T) Shubnikov-de Haas oscillations reveals two spin-orbit-split heavy-hole (HH) subbands with distinct effective masses contributing to transport. Surprisingly, the lighter-mass HH subband exhibits a parabolic dispersion with Fermi wavevector below the anticrossing between the heavy-hole and light-hole subbands, while the heavier HH subband is non-parabolic throughout. Quantitative comparison with numerical calculations based on the Luttinger model reveals that both effective masses are enhanced by a common factor ($\approx$ 2.3), which we attribute to many-body interactions. This common scaling factor has a very weak dependence on the 2DHG density, likely due to band hybridization. Our measured hole masses are compared with published cyclotron resonance and magnetotransport values. We propose a cohesive framework reconciling the long-standing three-way discrepancy between Luttinger theory, magnetotransport, and cyclotron resonance measurements of density-dependent effective masses in partially spin-orbit-polarized heavy-hole systems in GaAs.

Figures (4)

• Figure 1: (a) Diagram of the dispersion relations for spin-orbit polarized LH and HH subbands companionPRB2025, illustrating the anticrossing of HH$-$/LH$+$ in the presence of LH/HH hybridization. (b) Shubnikov-de Haas (SdH) oscillations at a fixed 2DHG density. (c) Comparison between carrier densities extracted from SdH oscillations and the Hall effect. The quantities $p_1$ (squares), $p_2$ (triangles), and $p_1 + p_2$ (circles) were obtained from the Fourier transform analysis of SdH oscillations. Cross symbols represent $p_{\text{2d}}$ determined from the low-field Hall effect. Solid (dashed) lines are linear least-squares fits to the FT (Hall) data. (d) Spin-orbit polarization $\Delta p/p$ versus 2DHG density.
• Figure 2: Fit values of: (a) the quantum scattering times ($\tau_q$) and (b) the effective massess ($m_1, m_2$), obtained from the Fourier analysis of SdH oscillations at different 2DHG densities in sample J (full symbols) and sample H (empty symbols), for HH$-$ (squares) and HH$+$ (diamonds).
• Figure 3: Comparison of 2DHG effective mass measurements from cyclotron resonance (CR) experiments and our SdH experiments in (100) GaAs single heterojunctions. Each filled (empty) triangle symbol represents a separate modulation-doped wafer from Ref. lu2008cyclotron (Ref. zhu2007density), measured at T = 4 K. Circles represent the weighted averages $(p_1m_1+p_2m_2)/(p_1+p_2)$ of the low-temperature data in sample J, to emulate T = 4 K (see main text). For all data, error bars are similar in size or smaller than the symbols.
• Figure 4: Comparison of $m_1$ and $m_2$ values obtained from our magnetotransport results in sample J (full diamonds and squares), published SdH studies (empty diamonds and squares) stormer1983energyhabib2004spingrbic2004, and Luttinger theory (empty circles) companionPRB2025. The inset demonstrates that all experimental $m_1$ (squares) and $m_2$ (diamonds) are larger than Luttinger theory predictions by approximately the same scaling factor ($\approx$ 2.3). All dashed and dotted lines are polynomial fits to the data.