Radial etching of strongly confined crystal-phase defined quantum dots

TL;DR

The paper addresses the need for strong confinement in crystal-phase defined InAs nanowire quantum dots to enhance charging energies and enable spin–orbit qubit studies. It advances a hybrid fabrication approach that combines axial WZ barriers around a ZB QD with isotropic radial etching, supplemented by 3D finite-element, density-gradient simulations to quantify electrostatics. The authors report $E_C$ up to about $33$ meV for etched QDs with $d_{NW}\approx 22$ nm and observe that $E_C$ and the gate lever arm $\alpha_G$ increase as diameter shrinks, while stray capacitances eventually cap further gains. The findings illuminate the balance between geometric confinement and parasitic capacitances, offering a path toward hard-wall NW QDs suitable for robust spin–orbit qubit platforms and guiding surface-passivation strategies to mitigate scattering. Overall, the work presents a practical framework for enhancing electronic confinement in crystal-phase-defined nanowire QDs and underscores the role of WZ/ZB interfaces in spin–orbit physics.

Abstract

We realize strongly confined quantum dots (QDs) in InAs nanowires (NWs) by combining epitaxial crystal-phase control with chemical wet etching. A strong axial confinement is first introduced by growing closely spaced wurtzite (WZ) tunnel barriers in NWs to enclose a zinc blende (ZB) QD. The NW cross-section is then reduced by isotropic etching to obtain very small QDs, with a maximum observed charging energy > 30 meV. Using low-temperature electrical characterization and finite-element method simulations, we study how charging energies and the onset of electron filling scale with QD diameter. For extremely small diameters, we identify a regime where stray capacitances become non-negligible, limiting further increase in charging energy by diameter reduction alone. This approach to increasing confinement is particularly relevant for understanding the strong spin-orbit interaction observed in crystal-phase QDs, possibly linked to polarization charges at the WZ/ZB interfaces. Small diameter QDs allow considerably weaker interfering electric fields when studied, but the QDs cannot be realized with epitaxial growth alone due to a loss of crystal phase control.

Figures (6)

• Figure 1: a) SEM images before and after etching in a diluted citric acid (C$_6$H$_8$O$_7$) and hydrogen peroxide (H$_2$O$_2$) mixture for 15 s. Etching procedure and mixture details are found in the main text. b) ECCI and SEM micrographs of a representative NW hosting two QDs, with crystal-phase dependent contrast revealing the zinc blende (ZB) and wurtzite (WZ) segments. c) ECCI and SEM micrographs of an etched QD segment (left, corresponding to device D2) and an unetched QD (right). The ECCI methodology reveals clear contrast differences between the crystal phases for segments with well-defined crystal facets. However, the small interaction volume in etched segments prevented extraction of their dimensions in this work. d) Higher resolution ECCI micrograph of the left QD in b, with extracted dimensions. ZB rotational twin planes reverse the contrast, which is observed in the left ZB segment.
• Figure 2: a, b) Stability diagrams of two etched QDs near depletion at $B=0$. c,d) Magnetic field dependence of the 1e charge state, where the field is applied along the direction shown in e,f. The $V_{SD}$ sweep in c is measured along the dashed line in a, while d is along the dashed line in b. e,f) SEM images of the two devices, where the thinned-down QDs are connected to electrodes. The sidegates visible in e are not used. g) Gate voltage at which the first electron occupies the QD, i.e., $V_G(N=1)$ versus NW diameter. Confinement and screening effects result in a positive shift with a reduced cross-section. h) Schematic showing that the subbands in both the WZ and ZB phases gain energy from the radial confinement, delaying the loss of axial confinement for thin diameters.
• Figure 3: a-c) Simulated charging energies, capacitances, and gate lever arms as a function of radial extent of QDs. Legend in a refers to QD dimensions. Solid markers correspond to experimental results, where numbering corresponds to the devices in Table \ref{['table:1']}. d) Schematics showing parameters for the simulation. A NW with varying diameter, barrier, and dot length is suspended a distance away from a back-gate. The barriers are treated as dielectrics, and the capacitances are extracted for each electrode (black). e) Charging energy as a function of gate bias for a 20 nm (blue), a 30 nm (green), and a 40 nm (orange) thick QD, with $w_B=40$ nm and $w_{QD}=12$ nm. The color maps show the electron distribution at certain charge configurations. Each colormap is normalized to itself, with the maximum value $n_{max}$ included next to each plot.
• Figure S1: SEM images used for extraction of QD dimensions.
• Figure S2: Charge stability diagrams of devices used to extract parameters in this work. Units in S.