Few-Mode and Anisotropic Quantum Transport in InSb Nanoribbons Using an All-van der Waals Material-Based Gate

TL;DR

This work demonstrates all-vdW gating of non-vdW InSb nanoribbons using an hBN dielectric and a few-layer graphite gate, achieving low gate hysteresis and high-quality quantum transport. Through systematic transport measurements, including magneto-spectroscopy and bias spectroscopy, the authors observe ballistic, few-mode conductance and anisotropic g-factor behavior arising from the NR cross-section, with quantized plateaus persisting to relatively low magnetic fields and shorter channels than typical InSb nanowire devices. The study highlights the effectiveness of all-vdW gates in reducing disorder and enabling clean quantum transport in non-vdW materials, with implications for spintronics and topological superconductivity in quantum devices. The supplemental devices reinforce the main conclusions and illustrate that contact quality is a key bottleneck, guiding future improvements in contact engineering alongside all-vdW gating strategies.

Abstract

High-quality electrostatic gating is a fundamental ingredient for successful semiconducting device physics, and a key element of realizing clean quantum transport. Inspired by the widespread improvement of transport quality when two-dimensional van der Waals (vdW) materials are gated exclusively by other vdW materials, we have developed a method for gating non-vdW materials with an all-vdW gate stack, consisting of a hexagonal boron nitride dielectric layer and a few-layer graphite gate electrode. We demonstrate this gating approach on MOVPE-grown InSb nanoribbons (NRs), a novel variant of the InSb nanowire, with a flattened cross-section. In our all-vdW gated NR devices we observe conductance features that are reproducible and have low- to near-zero gate hysteresis. We also report quantized conductance, which persists to lower magnetic fields and longer channel lengths than typical InSb nanowire devices reported to date. Additionally, we observe level splitting that is highly anisotropic in an applied magnetic field, which we attribute to the ribbon cross-section. The performance of our devices is consistent with the reduced disorder expected from the all-vdW gating scheme, and marks the first report of ballistic, few-modes quantum transport in a non-vdW material with an all-vdW gate. Our results establish all-vdW gating as a promising approach for high-quality gating of non-vdW materials for quantum transport, which is in principle applicable generically, beyond InSb systems. In addition, the work showcases the specific potential of all-vdW gate/InSb NR devices for enabling clean quantum devices that may be relevant for spintronics and topological superconductivity studies.

Figures (9)

• Figure S1: Hysteresis in the 1.5 $\mu$m channel as a function of wait time between the setting of gate voltage, $V_g$, and the measurement of current through the device. a) Linear conductance, $G=I_\mathrm{meas}/V_\mathrm{bias}$, for forward and backwards sweeps of $V_g$ with wait times of 0.1, 1.5, and 7.5 s. The impact of wait time on the observed hysteresis is negligible, and the traces lie almost entirely on top of each other. b-d) Details of the hysteresis loops in the neighborhood of the threshold voltage for each wait time. Threshold voltages for each sweep are indicated with horizontal dashed lines, and the corresponding normalized near-pinchoff hysteresis value is given, using the same definition as in the main text.
• Figure S2: Forward pinch-off and fitting for mobility extraction in the 1.5 $\mu$m channel. A fixed bias of 5 mV is used, and conductance is calculated as G $=I_\mathrm{meas}/V_\mathrm{bias}$. No series resistance is subtracted, as it is a variable to be fitted.
• Figure S3: Differential conductance, $G=\mathrm{d}I_\mathrm{meas}/\mathrm{d}V_\mathrm{bias}$, colormap as a function of $V_\mathrm{bias}$ and $V_g$ in the 1.5 $\mu$m channel, with zero applied magnetic field. No evidence of quantized conductance is present, as this channel is in the diffusive transport regime, but the effects of a somewhat non-Ohmic contact are visible at low biases throughout the whole $V_g$ range. Here a series resistance of 12.2 k$\Omega$ has been subtracted, as determined in the mobility fitting.
• Figure S4: Out-of-plane field sweep, 200 nm channel. a) Linear conductance at $V_\mathrm{bias}=1$ mV as a function of $V_g$ and out-of-plane magnetic field, $B_z$. Both the 0.5 and 1.0 $G_0$ plateaus appear to persist and continuously evolve through the full range of field values. b) Linecuts from a) at various field values, offset in $V_g$ by -60 mV (1.55 T) and -120 mV (0.3 T) for clarity. In all traces, plateau-like features with intermediate, non-half-integer values of $G$ are observed in addition to the quantized plateaus. The exact nature of these features is unclear, and may vary with field strength.
• Figure S5: Differential conductance, $G=\mathrm{d}I_\mathrm{meas}/\mathrm{d}V_\mathrm{bias}$, colormap as a function of $V_\mathrm{bias}$ and $V_g$ in the 400 nm channel for different angles of the external 2.8 T magnetic field. a) Field angle of $\theta=0$, fully in-plane with the NR. b) Field angle of $\theta=45\degree$, halfway between in-plane and out-of-plane. In both panels, the edge of the 0.5 $G_0$ plateau has been identified by hand and indicated with black dashed lines. Both field angles give a lever arm of $\sim 110$ meV/V, the same as was found for the $\theta=90\degree$ (fully out-of plane) case, justifying the use of plateau width as a proxy for $g$-factor in the main text.