Spin-valley 0.7 anomaly in bilayer graphene/WSe$_2$ quantum point contacts

Abstract

We report a well-resolved 0.7 conductance anomaly at $G = 0.7\times(2e^2/h)$ in bilayer graphene/WSe$_2$ quantum point contacts. Proximity-enhanced spin-orbit coupling splits the four-fold ground state of bilayer graphene into well-separated spin-valley locked Kramers doublets. The anomaly emerges between these opposite spin-valley states. Despite fundamentally different band structure and wavefunction characteristics, the temperature and bias phenomenology closely mirror GaAs systems. In contrast, the parallel magnetic field response differs significantly, confirming the central role of valley degrees of freedom. This opens new pathways to study valley-exchange correlation physics in regimes inaccessible to conventional semiconductors.

Figures (8)

• Figure 1: (a) Top-view SEM image and schematic cross-section of the BLG/WSe2 heterostructure. Split gates (SGs) electrostatically confine the 1D channel, while the channel gate (CG) controls the carrier density. (b) Transconductance $dG/dV_\mathrm{CG}$ versus $V_\mathrm{CG}$ and $B_\perp$ showing individually resolved transport modes enabled by proximity-enhanced SOC. The legend indicates the evolution of the first subband states $(\mathrm{K}^+\uparrow, \mathrm{K}^-\downarrow)$. (c) Conductance $G$ versus $V_\mathrm{CG}$ for varying perpendicular ($B_\perp$) and (d) parallel magnetic field ($B_\parallel$). Curves are offset horizontally for clarity. The $2e^2/h$ plateau exhibits a pronounced anomaly at roughly $0.7 \times 2e^2/h$ (black arrow) at zero magnetic field. Increasing $B_\perp$ leads to an evolution of the shoulder into a fully quantized $e^2/h$ plateau through valley--Zeeman splitting, while $B_\parallel$ leaves the 0.7 feature unchanged because of spin--valley locking.
• Figure 2: (a) Differential conductance $G$ versus source-drain bias $V_\mathrm{SD}$ for varying $V_\mathrm{CG}$ with a prominent finite-bias plateau appearing at $0.7 \times 2e^2/h$ and a zero-bias anomaly below the $2e^2/h$ plateau. (b) Temperature dependence of the zero-bias anomaly in conductance $G$ versus $V_\mathrm{SD}$, showing reduced amplitude and increased width at higher temperatures.
• Figure 3: (a), (b) Temperature-dependent conductance $G$ versus $V_\mathrm{CG}$ for different split gate voltages $V_\mathrm{SG}$. The conductance shows characteristic suppression at elevated temperatures typical of the 0.7 anomaly. (c), (d) Kondo temperatures $T_K$ extracted from the temperature-dependent fits in (a) and (b) versus $V_\mathrm{CG}$ for both $V_\mathrm{SG}$ values. (e) Subband spacing $\Delta E_{1,2}$ and (f) electrostatic potential curvature $\hbar\omega_{x}$ as functions of $V_\mathrm{SG}$.
• Figure A.1: (a) Temperature-dependent conductance $G$ versus $V_\mathrm{CG}$ for varying split gate voltages $V_\mathrm{SG}$. All conductance traces show suppression at elevated temperatures typical of the 0.7 anomaly. (b) Respective extracted Kondo temperatures for all SG configurations.
• Figure A.2: (a)$dG/dV_\mathrm{CG}$ versus $B_\perp$ and channel gate voltage $V_\mathrm{CG}$ at 10mK with additional anomaly at $G=1.2G_0$ (marked by the red circle). (b) Respective conductance evolution in $B_\perp$. (c) Temperature dependence of the conductance showing $G$ suppression at elevated temperatures. (d) Differential conductance versus source-drain bias $V_\mathrm{SD}$ showing both a zero-bias anomaly and finite-bias plateaus at $1.2G_0$.