⇦Chinese Physics Letters, 2022, Vol. 39, No. 5, Article code 058101Express Letter In Situ Epitaxy of Pure Phase Ultra-Thin InAs-Al Nanowires for Quantum Devices Dong Pan (潘东)1,2†, Huading Song (宋化鼎)3,4†, Shan Zhang (张珊)3†, Lei Liu (刘磊)1, Lianjun Wen (文炼均)1, Dunyuan Liao (廖敦渊)1, Ran Zhuo (卓然)1, Zhichuan Wang (王志川)5, Zitong Zhang (张梓桐)3, Shuai Yang (杨帅)3,4, Jianghua Ying (应江华)3,4, Wentao Miao (苗文韬)3, Runan Shang (尚汝南)4, Hao Zhang (张浩)3,4,6*, and Jianhua Zhao (赵建华)1,2* Affiliations 1State Key Laboratory of Superlattices and Microstructures, Institute of Semiconductors, Chinese Academy of Sciences, P. O. Box 912, Beijing 100083, China 2Center of Materials Science and Optoelectronics Engineering, and CAS Center for Excellence in Topological Quantum Computation, University of Chinese Academy of Sciences, Beijing 100190, China 3State Key Laboratory of Low Dimensional Quantum Physics, Department of Physics, Tsinghua University, Beijing 100084, China 4Beijing Academy of Quantum Information Sciences, Beijing 100193, China 5Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China 6Frontier Science Center for Quantum Information, Beijing 100084, China Received 18 March 2022; accepted 28 March 2022; published online 1 April 2022 †D. Pan, H. D. Song, and S. Zhang contributed equally to this work.
*Corresponding authors. Email: jhzhao@semi.ac.cn; hzquantum@mail.tsinghua.edu.cn Citation Text: Pan D, Song H D, Zhang S et al. 2022 Chin. Phys. Lett. 39 058101    Abstract We demonstrate the in situ growth of ultra-thin InAs nanowires with an epitaxial Al film by molecular-beam epitaxy. Our InAs nanowire diameter ($\sim $30 nm) is much thinner than before ($\sim $100 nm). The ultra-thin InAs nanowires are pure phase crystals for various different growth directions. Transmission electron microscopy confirms an atomically abrupt and uniform interface between the Al shell and the InAs wire. Quantum transport study on these devices resolves a hard induced superconducting gap and 2$e$-periodic Coulomb blockade at zero magnetic field, a necessary step for future Majorana experiments. By reducing wire diameter, our work presents a promising route for reaching fewer sub-band regime in Majorana nanowire devices.
cpl-39-5-058101-fig1.png
cpl-39-5-058101-fig2.png
cpl-39-5-058101-fig3.png
cpl-39-5-058101-fig4.png
cpl-39-5-058101-fig5.png
 DOI:10.1088/0256-307X/39/5/058101 © 2022 Chinese Physics Society Article Text Semiconductor nanowires proximately coupled to superconductors have attracted increasing interest in recent years due to the prospect of hosting topologically protected Majorana zero modes (MZMs), which can be used in fault-tolerant quantum computing.[1–9] Among them, narrow band gap semiconductor nanowires (InAs and InSb) coupled to a superconductor (Al, Sn and Pb) are of particular interest due to the strong spin-orbit coupling and large $g$-factor in the semiconductor.[10–18] Many experimental signatures for possible MZMs in structures based on these hybrid nanowires have been reported.[10–14] These experimental studies and related theoretical work have demonstrated that an atomic scale uniform and defect-free crystal interface between the semiconductor and the superconductor is a crucial ingredient for high quality experiment.[15,19–21] This interface quality mainly depends on the quality of the semiconductor and the superconductor as well as how they are combined. The superconductor layer fabricated by conventional (ex situ evaporation/sputter) methods often shows undesired low-energy states within the proximity-induced superconducting gap, the soft gap problem. The soft gap is a source of decoherence and thus detrimental to future topological quantum information processing. This problem was solved by Krogstrup et al. using in situ deposition of Al shells on the sidewalls of InAs nanowires by molecular-beam epitaxy (MBE).[21] To date, high-quality superconductor layers (such as Al and Pb) have been successfully grown on InAs, InSb and InAsSb nanowires.[21–26] See also related networks[27,28] using this low temperature epitaxial technique. Another factor is related to the quality of the semiconductor nanowires. In particular, InAs nanowires usually exhibit random mixtures of wurtzite (WZ) and zinc-blende (ZB) crystal structures, which deteriorates the electrical transport properties due to electron scattering at stacking faults or twin planes.[29–31] For example, this type of scattering related disorder can easily create unwanted quantum dots which lead to the formation of Andreev bound states, the most popular alternative (topologically trivial) explanation of current zero bias peaks.[32–34] Theoretical and experimental studies demonstrated that controlling the diameter down to a small value (e.g., below 50 nm for Ag catalyzed InAs growth) is one of the effective methods to obtain pure phase InAs nanowires.[35–38] Nevertheless, to the best of our knowledge, the InAs-Al nanowires reported have InAs diameters $\sim $40–150 nm[18,21–23,39] (typically $\sim $100 nm). The epitaxy of ultra-thin InAs-Al nanowires (with pure phase InAs, diameter below 40 nm) for quantum devices has not been experimentally demonstrated yet. Besides the material quality, another motivation for the pursuit of thin diameter is aiming to reduce the number of occupied sub-bands in the wire. Previous experiments likely have multiple sub-band occupation. The precise occupation number is unknown. In a multi-band system, each spin-resolved sub-band could host a pair of MZMs, leading to multiple MZMs overlapping in the wire's ends. The coupling between these spatially overlapped MZMs (non-topological) can be significantly suppressed by the smooth potential variation in the system. In transport experiment, they could mimic the topological Majorana signature, e.g., a quantized zero bias conductance peak, and thus are dubbed as quasi-MZMs with a topologically trivial origin.[40,41] Another disadvantage of multi-band is that the tunnel coupling between the probing lead and the proximitized wire is dominated by the lower sub-bands. However, it is the top sub-band which hosts topological MZMs. The small tunnel coupling between the top sub-band and the probe could significantly suppress the MZM signal (e.g., the MZM peak width and the peak height with the presence of thermal averaging). To reduce the number of occupied sub-bands, electrostatic gating, however, may have a limited effect due to the superconductor screening based on recent Schrödinger–Poisson simulations.[42–44] Therefore, one possible solution is to reduce the wire diameter, which can increase the sub-band energy spacing. In this Letter, we present the epitaxy of InAs-Al nanowires with the InAs wire diameter significantly reduced. The InAs nanowires with thin diameters are pure phase crystals for various different growth directions. This diameter ($\sim $30 nm) is only one-third of those used in literature (typically $\sim $100 nm). Our nanowire diameter is comparable to the width of quantum well formed at the InAs-Al interface due to band bending,[42] therefore may enhance the sub-band energy spacing. This large sub-band spacing may help to reduce the number of occupied sub-bands or even reach the single sub-band regime if the diameter is continuously decreased, a true one-dimensional electron system. For the Al deposition,[21] we found that the substrate temperature and Al shell thickness are crucial parameters. By optimizing these parameters, we can obtain continuous and smooth Al shells on the ultra-thin InAs nanowires. The transmission electron microscope (TEM) study confirms that our InAs-Al interface is atomically abrupt and uniform. As a direct consequence, these hybrid-wires show hard induced superconducting gap and 2$e$-periodic Coulomb blockade in the electron transport study of corresponding quantum devices. Large zero bias peaks, reaching the order of 2$e^{2}/h$, are observed on these wires, reported in a separate work.[45] Our results on ultra-thin hybrid nanowires pave the way for realization of fewer or ultimately single sub-band MZMs. In Situ Epitaxy of Al Half Shells on the Ultra-Thin InAs Nanowires. Figures 1(a)–1(c) illustrate the diameter dependent crystal structure and crystal quality of InAs nanowires. For thick diameters, twin defects or mixture of ZB and WZ are commonly observed in InAs nanowires. By reducing the wire diameter (below 50 nm), one can achieve a pure crystal phase which forms the base of this work.[38] We grow InAs nanowires with thin diameter using Ag as seed particles. Due to the small sizes of the seed particles, the wire diameters ranging from $\sim $20 to 30 nm can be obtained (see the Supplementary Material for growth details). From the TEM study, we confirm the pure phase crystal structure of these ultra-thin InAs nanowires: along the $\langle 111\rangle$ direction, the InAs has a pure WZ crystal structure; along the non-$\langle 111\rangle$ directions, the InAs has either a pure WZ or pure ZB crystal structure (for more details of the thin InAs nanowires, please see Refs. [46–52]). To obtain a uniform semiconductor-superconductor interface, Al shells [Figs. 1(d) and 1(e)] were deposited after the growth of the InAs, with an in situ epitaxial growth manner. We find that the substrate temperature during Al growth plays a key role on the Al shell morphology: a low substrate temperature is a prerequisite in enabling the deposition of continuous Al layers on the nanowire facets. Otherwise, with high growth temperature ($\sim$$1\,^{\circ}\!$C), the Al shells form discontinuous islands and look like ‘pearls on a string’ on the side walls of all the InAs nanowires [see Figs. S1(a)–1(e) in the Supplementary Material]. Cooling down the substrate to about $-10\,^{\circ}\!$C can form continuous Al shells [see Figs. S1(f)–1(j) in the Supplementary Material]. However, the Al shells have a rough and faceted outer surface, suggesting that the $-10\,^{\circ}\!$C temperature is still not low enough. Moreover, the thickness of Al shell measured from high-resolution TEM images is $\sim $30 nm [see Figs. S2 and S3 in the Supplementary Material], which is beyond the critical thickness of the Al layer for smooth shell growth. Based on these feedbacks, we further lower the substrate temperature during Al growth and reduce the Al film thickness (less than a critical thickness $\sim $15 nm in our case). We show in Figs. 1(f)–1(l) our ultra-thin InAs-Al nanowires grown under the optimum conditions, the growth temperature and time of the Al shells are about $-40\,^{\circ}\!$C and 3 min, respectively. It is evident from the scanning electron microscopy (SEM) images that continuous and smooth Al half-shells are formed on the facets of all the InAs nanowires (also see Fig. S4 in the Supplementary Material). Figures 1(f) & 1(j), 1(k), and 1(l) are from three separate growth rounds with different Al shell thicknesses. Notably, the total diameters of the InAs-Al nanowires are small, ranging from $\sim $29 nm to $\sim $45 nm. We also find that some InAs-Al nanowires have a natural shadowed region that separates the Al shell into two islands, which can be used to estimate the thickness of the Al shells. As can be seen from Fig. 1(j), the thicknesses of the shadowed region (pure InAs nanowire) and the segment covered with Al are $\sim $18 nm and $\sim $35 nm, respectively. Thus, the Al thickness (including native oxide layer) is $\sim $17 nm, matching our expectations based on growth rate and time. Figures 1(k) and 1(l) show thinner Al films (down to $\sim $10 nm and 6 nm in thickness), also consistent with our growth condition. In Fig. S5, we show InAs-Al nanowires with various different growth directions, continuous and uniform Al shells can be observed.
cpl-39-5-058101-fig1.png
Fig. 1. In situ epitaxy of Al half shells on the ultra-thin InAs nanowires. (a)–(c) Schematic illustration of the diameter dependent InAs nanowire crystal structures. As the InAs diameter becomes thinner, the crystal structure can be tuned from a ZB crystal phase with twin defects to a mixture of the WZ and ZB phases, and finally to a pure WZ crystal phase with no stacking faults or twin defects. The yellow lines in (a) and (b) denote the twin defects and stacking faults. (d) Schematic illustration of in situ epitaxy of Al half shells on an ultra-thin InAs nanowire. The nanowire is pure phase and can be either WZ or ZB for various growth directions. (e) Schematic cross-section of InAs-Al. (f)–(j) SEM images of several typical ultra-thin InAs-Al nanowires with various wire diameters. The Al shell thickness is $\sim $17 nm. The missing of Al on part of the InAs nanowire in (j) was due to the shadow effect of another wire in front of it during the Al growth. (k)–(l) Two more shadow InAs-Al wires (from different growths) with the Al shell thickness being $\sim $10 nm and 6 nm, respectively. All the SEM images were taken at a tilt angle of 25$^{\circ}$.
Crystal Structure and Chemical Composition of the Ultra-Thin InAs-Al Nanowires. We next perform TEM and energy dispersive spectrum (EDS) measurements on our ultra-thin wires to evaluate the InAs-Al interface quality, which is crucial for MZM experiments. Figure 2(a) is a typical low-magnification TEM image showing a continuous and smooth half Al shell on the facet of the InAs nanowires, consistent with the SEM results in Fig. 1. Figures 2(b) and 2(c) are high-resolution TEM images of the InAs-Al nanowire. We can see that the Al shell is well epitaxially grown on the InAs nanowire. Meanwhile, axial and radial rotations of the Al are not observed. The InAs nanowire has a pure ZB crystal structure and the Al has a face-centered cubic crystal structure. As shown in Fig. 2(c), the Al shell forms a sharp and uniform interface to the InAs nanowire. The thicknesses of the InAs nanowire and Al shell are $\sim $19 nm and $\sim $15 nm, respectively. According to the fast Fourier transform (FFT) of the Al shell [Fig. 2(d)], the growth direction of the nanowire is $\langle \bar{1}0\bar{1} \rangle $. The interface of the ultra-thin InAs-Al nanowire has also been identified by high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and EDS measurements. Figures 2(e) is an HAADF-STEM image taken from the section of the ultra-thin InAs-Al nanowire. This again shows that the ultra-thin InAs-Al nanowire has a sharp and uniform interface. The false-color EDS elemental maps of In [Fig. 2(f)], As [Fig. 2(g)], Al [Fig. 2(h)] and their overlay [Fig. 2(i)] taken in the middle region of the InAs-Al nanowire further confirm that a sharp and uniform interface has formed between the InAs nanowire and the Al layer. Detailed TEM observation of a dozen such InAs-Al nanowires (Section 3 in the Supplementary Material), all revealed that the InAs-Al interface is sharp, uniform, and oxide-free. Continuous, uniform and thin half Al shells can be well epitaxially grown on the facets of InAs nanowires with all the growth directions we have measured (see Figs. S8 and S9 in the Supplementary Material).
cpl-39-5-058101-fig2.png
Fig. 2. Crystal structure and chemical composition of the ultra-thin InAs-Al nanowires. (a) Low-magnification TEM image of an InAs-Al nanowire. (b) and (c) High-resolution TEM images of the InAs-Al nanowire. (d) The FFT of the Al segment viewed along the $[\bar{1}11]$ axis. (e) HAADF-STEM image taken from the middle section of the InAs-Al nanowire. The white rectangle in (a) highlights the region where HAADF-STEM image was recorded. (f)–(h) False-color EDS elemental maps of In (yellow), As (green), and Al (red) taken in the middle region of the InAs-Al nanowire, respectively. (i) False-color overlay EDS elemental maps of In (yellow), As (green), and Al (red).
Hard Superconducting Gap. We now turn to the quantum transport study of our ultra-thin InAs-Al devices. Figure 3(a) shows a normal-nanowire-superconductor (N-NW-S) device (device A). The global back gate ($V_{\rm BG}$) tunes the electron density in the InAs nanowire (grey), as well as the tunnel barrier (the wire segment with Al shell etched away). Details of device fabrication and measurement circuit can be found in the methods. The InAs wire diameter in this device is $\sim $34 nm. Figure 3(b) shows the differential conductance $dI/dV$ as a function of the source-drain bias voltage $V$ and $V_{\rm BG}$. Sweeping $V_{\rm BG}$ more negative, the overall conductance decreases to zero, tuning the barrier segment of the nanowire from open to pinched off. The features of two horizontal white lines (around $\mathrm{\pm 0.24\,mV}$) indicate the superconducting gap edge. In the tunneling regime (near pinched off), $dI/dV$ reveals the density of states in the proximitized nanowire part, as shown in Figs. 3(c) and 3(d). The sub-gap conductance is suppressed to zero, suggesting a hard induced superconducting gap.[15,16,21,53,54] This hard gap is a direct result of the high-quality atomic-abrupt InAs-Al interface as demonstrated in Fig. 2.[20] The suppression ratio of the sub-gap conductance compared with the normal state conductance (outside gap) is around 50, comparable with the values reported in literature.[15,16,21,53,54] The superconducting gap remains hard with sub-gap conductance sticking to zero after applying a magnetic field $B$, as shown in Fig. 3(e). The $B$-direction is along the wire axis. The gap closes around 1.1 Tesla, due to orbital effect of the bulk Al film with an estimated film thickness of 15 nm. Large Zero Bias Peak. At a different $V_{\rm BG}$, the $dI/dV$ spectroscopy of device A reveals an Andreev bound state indicated by the two symmetric sub-gap peaks at $B = 0$ as shown in Fig. 4(a). We note several charge rearrangements between the measurements of Figs. 3 and 4, which reset the device condition. The two peaks in Fig. 4(a) Zeeman split in $B$-field. The inner peaks merge at zero bias, forming a zero bias peak (ZBP). The effective $g$-factor is estimated to be $\sim $7.6 based on the Zeeman splitting dispersion in $B$.[27,55] This ZBP remains non-split against $B$-sweep from 0.28 T to 1.1 T. This $B$ range of $\sim $0.8 T translates into a Zeeman energy scale (1/2$g\mu _{\scriptscriptstyle{\rm B}}B$) of $\sim $0.18 meV, comparable with the superconducting gap and much larger than the ZBP width. A ZBP sticking at zero bias in $B$ sweep was initially regarded as a possible sign of MZM in hybrid nanowire devices.[10,56,57] However, a more likely alternative explanation with a topologically trivial origin based on Andreev bound states was quickly proposed.[32–34,58] Figure 4(b) shows the zero-bias line-cut where the ZBP-height first reaches 80% of 2$e^{2}/h$, and then decreases by further increasing $B$ field. We note that our reported $dI/dV$, in a two-terminal device design, is purely the conductance of the device (including unknown contact resistance) with the outer measurement circuit resistance (e.g., fridge filters) subtracted based on independent calibration (see the Supplementary Material for details).
cpl-39-5-058101-fig3.png
Fig. 3. Hard superconducting gap. (a) False-color SEM image of an N-NW-S device (device A). Scale bar is 1 µm. Part of the Al film (red) on the ultra-thin InAs nanowire (grey) was selectively etched. The nanowire was then contacted by normal metal (yellow, 10 nm Ti and 70 nm Au). The substrate is p-doped Si, acting as a global back gate, covered by 285-nm-thick SiO$_{2}$ (gate dielectric). Fridge base temperature is $\sim $20 mK for all the measurements with various fridge filters. (b) Differential conductance $dI/dV$ of device A as a function of bias voltage $V$ and back gate voltage $V_{\rm BG}$ resolving the superconducting gap of $\varDelta \sim 0.24$ meV. (c) and (d) Vertical line-cuts (linear and log-scale) at $V_{\rm BG}$ of $-1.258$ V labeled by the black bar in panel (b), resolving a hard superconducting gap. (e) $B$ dependence of the gap with $B$ direction aligned with the wire axis. Lower panel shows the horizontal line-cuts within the gap (0 bias, black curve) and outside the gap ($V = 0.4$ mV, red curve).
cpl-39-5-058101-fig4.png
Fig. 4. Large zero bias peak. (a) Behavior of $dI/dV$ of device A shows a large zero bias peak (ZBP) at a $V_{\rm BG}$ of $-1.808$ V. (b) Zero-bias line-cut shows the peak height exceeding 80% of 2$e^{2}/h$. (c) and (d) The $dI/dV$ curve (vertical line-cut) at 0 T and 0.39 T, respectively. (e) Another ZBP remains non-split against the $B$-sweep at a different $V_{\rm BG}$ of $-1.88$ V. (f) $V_{\rm BG}$ sweep of the two ZBPs from (a) and (e) at $B = 0.5$ T. The blue and red arrows roughly correspond to the gate voltages of (a) and (e), respectively, with a small gate sweep hysteresis.
Figures 4(c) and 4(d) show the $dI/dV$ curves at 0 T and 0.39 T, respectively. Our ZBP height is large but still not reaching 2$e^{2}/h$, which is of the similar order with the 2DEG defined InAs-Al nanowires[59] and significantly larger than the ZBP-height reported on similar InAs-Al devices in the literature.[56] In an independent measurement of a separate device from the same nanowire batches (sharing identical or similar growth conditions), our ZBP height can reach 2$e^{2}/h$.[45] Though we speculate that thinner wires with single sub-band occupation and improved crystalline properties could enhance ZBP-height,[60,61] more experiments are needed to establish such connection,[45] a subject of our future study. Figure 4(e) shows another ZBP at a different $V_{\rm BG}$, which also remains non-split over a sizable $B$ range. This ZBP height, however, is much less than the one in Fig. 4(a). Figure 4(f) shows the $V_{\rm BG}$ sweep at 0.5 T, which reveals the two ZBPs in Fig. 4(a) and Fig. 4(e) as two level-crossing points, similar to the typical Andreev bound state peaks reported in the literature.[58] This level-crossing feature of ZBP rules out its topological origin[55] since MZM induced ZBPs should remain non-split over a sizable range in both $B$ scan and gate scans. The full phase diagram (in $B$ and $V_{\rm BG}$ space) of this trivial ZBP of Figs. 4(a) and 4(e) is shown in Figs. S10 and S11, demonstrating that the large ZBP, robust in $B$ scan [Fig. 4(a)], is probably due to fine-tuned level repulsion. The formation of these trivial sub-gap states is less likely due to crystalline defects, where our TEM analysis has confirmed its pure-phase and single-crystal nature. We do note that other types of disorder (e.g., smooth potential due to electro-gates, contact preparation and nanowire oxides) still remain and are likely the cause of these trivial states, which is a subject for our future device improvement.
cpl-39-5-058101-fig5.png
Fig. 5. The 2$e$-periodic Coulomb blockade and $2e$–$1e$ transition of an island device. (a) False-color SEM image of an InAs-Al island device (device B). Scale bar is 1 µm. The Al island (red) was defined by wet chemical etching. The normal metal contact (yellow) and side gates (cyan) are Ti/Au (10 nm/70 nm). The plunger gate voltage $V_{\rm PG}$ tunes electron number on the island, while the lower two tunnel gates tune the tunnel coupling between the island and the contacts. (b) Characteristics of $dI/dV$ of device B as a function of $V$ and $V_{\rm PG}$, revealing regular Coulomb-blockade diamonds. The two tunnel gates and back gate were kept grounded for the measurement. (c) Horizontal line-cuts at $V = 0$ mV (black curve) and $V = 0.3$ mV (red curve). (d) Vertical line-cuts at the Coulomb valley (red curve) and Coulomb peak degeneracy point (black curve) indicated by the corresponding arrow in (b). (e) Behavior of $dI/dV$ (at $V = 0$) as a function of $V_{\rm PG}$ and $B$ (direction along the nanowire), revealing the $2e$–$1e$ transition, with the vertical line-cuts at different $B$ values shown in (f).
The 2${e}$-Periodic Coulomb blockade and 2${e}$–$1{e}$ Transition of an Island Device. Next, we study a different device with the Al shell floated instead of grounded as shown in Fig. 5(a). The InAs wire diameter in this device is $\sim $34 nm. The nanowire segment with the floated Al island tunnel couples to the source and drain normal metal contacts through the InAs nanowire, forming a gate-tunable hybrid superconducting island device with a finite Coulomb charging energy. The plunger gate ($V_{\rm PG}$) tunes the electron number on the island, leading to Coulomb blockade diamonds in the $dI/dV$ spectroscopy as shown in Fig. 5(b). The hard induced superconducting gap ensures negligible quasi-particle poisoning. Therefore, each Coulomb diamond in Fig. 5(b) corresponds to 2$e$ (Cooper pair) instead of 1$e$ (quasi-particle) charge transport. This 2$e$ feature is also reflected by comparing the Coulomb oscillation period at zero bias (2$e$) and high bias (1$e$) in Fig. 5(c). Based on the Coulomb diamond size ($V \sim 0.22$ mV), we estimate the charging energy of $E_{\rm C}=e^{2}/2C \sim 28\,µ$eV, smaller than the superconducting gap and thus fulfilling a necessary condition for 2$e$. Figure 5(d) shows the $dI/dV$ curve at the Coulomb blockade valley (red) and the degeneracy point (black). Negative differential conductance (NDC) is observed at the degeneracy point, which is the known features for these hybrid islands. Based on the onset bias of NDC and $V_{\mathrm{NDC}}=2(E_{\rm O}-E_{\rm C})/e \sim 68\,µ\mathrm{V}$,[62] we estimate the energy of the lowest sub-gap state $E_{\rm O} \sim 62\,µ$eV. In Fig. S12, we show another island device (with InAs wire diameter of $\sim $28 nm), where the extracted parameters are in similar ranges. Figure 5(e) shows the 2$e$-periodic Coulomb peaks (at $V = 0$ mV) splitting into two 1$e$-periodic peaks in a $B$ sweep (direction along the nanowire), see Fig. 5(f) for typical line-cuts. This $2e$–$1e$ transition is likely due to the closing of the superconducting gap: when the gap or $E_{\rm O}$ is smaller (lower) than $E_{\rm C}$, odd parity ground state starts to be available, and the 2$e$-peak splits. The $B$-field induced $2e$–$1e$ transition and its further peak-spacing-oscillation pattern is another possible MZM signature in these devices,[63] with also alternative explanations proposed based on trivial origins.[64,65] In summary, we have developed an MBE growth technique for the epitaxy of high-quality Al layers on the pure phase, ultra-thin InAs nanowires for various different growth directions. A sharp and uniform interface is identified. As a result, electron transport reveals a hard induced superconducting gap and 2$e$-Coulomb blockade in related quantum devices. Since disorder is currently the biggest challenge for MZM nanowire experiments,[66,67] our result eliminates crystalline twin defects and stacking faults as a source of disorder in our system. We do note that thinner diameter leads to larger surface-volume ratio therefore making the device more sensitive to surface disorder. Decreasing other types of disorder by device improvement is in progress, which may open possibilities for observation of clean signatures of MZMs in the future. Acknowledgments. This work was supported by the National Natural Science Foundation of China (Grant Nos. 92065106, 61974138, 12104053, and 11704364), the Beijing Natural Science Foundation (Grant No. 1192017), and Tsinghua University Initiative Scientific Research Program. D.P. also acknowledges the support from Youth Innovation Promotion Association, Chinese Academy of Sciences (Grant No. Y2021043). H.D.S acknowledges China Postdoctoral Science Foundation (Grant Nos. 2020M670173 and 2020T130058). References Majorana Fermions and a Topological Phase Transition in Semiconductor-Superconductor HeterostructuresHelical Liquids and Majorana Bound States in Quantum WiresTopological quantum computationPaired states of fermions in two dimensions with breaking of parity and time-reversal symmetries and the fractional quantum Hall effectUnpaired Majorana fermions in quantum wiresSuperconducting Proximity Effect and Majorana Fermions at the Surface of a Topological InsulatorMajorana zero modes in superconductor–semiconductor heterostructuresNext steps of quantum transport in Majorana nanowire devicesFrom Andreev to Majorana bound states in hybrid superconductor–semiconductor nanowiresSignatures of Majorana Fermions in Hybrid Superconductor-Semiconductor Nanowire DevicesAnomalous Zero-Bias Conductance Peak in a Nb–InSb Nanowire–Nb Hybrid DeviceZero-bias peaks and splitting in an Al–InAs nanowire topological superconductor as a signature of Majorana fermionsThe fractional a.c. Josephson effect in a semiconductor–superconductor nanowire as a signature of Majorana particlesSuperconductor-nanowire devices from tunneling to the multichannel regime: Zero-bias oscillations and magnetoconductance crossoverHard gap in epitaxial semiconductor–superconductor nanowiresShadow-wall lithography of ballistic superconductor–semiconductor quantum devicesParity-preserving and magnetic field–resilient superconductivity in InSb nanowires with Sn shellsEpitaxial Pb on InAs nanowires for quantum devicesSuperconducting proximity effect in semiconductor nanowiresSoft Superconducting Gap in Semiconductor Majorana NanowiresEpitaxy of semiconductor–superconductor nanowiresRobust Epitaxial Al Coating of Reclined InAs NanowiresShadow Epitaxy for In Situ Growth of Generic Semiconductor/Superconductor HybridsZero-bias peaks at zero magnetic field in ferromagnetic hybrid nanowiresEngineering hybrid epitaxial InAsSb/Al nanowires for stronger topological protectionHighly Transparent Gatable Superconducting Shadow JunctionsSelective-Area-Grown Semiconductor-Superconductor Hybrids: A Basis for Topological NetworksSelective-area chemical beam epitaxy of in-plane InAs one-dimensional channels grown on InP(001), InP(111)B, and InP(011) surfacesElectron transmission through silicon stacking faultsBallistic electron transmission through interfacesSingle-electron transport in InAs nanowire quantum dots formed by crystal phase engineeringTransport spectroscopy of N S nanowire junctions with Majorana fermionsNear-zero-energy end states in topologically trivial spin-orbit coupled superconducting nanowires with a smooth confinementAndreev bound states versus Majorana bound states in quantum dot-nanowire-superconductor hybrid structures: Trivial versus topological zero-bias conductance peaksAn Empirical Potential Approach to Wurtzite-Zinc-Blende Polytypism in Group III-V Semiconductor NanowiresControlled polytypic and twin-plane superlattices in iii–v nanowiresMethod for Suppression of Stacking Faults in Wurtzite III−V NanowiresControlled Synthesis of Phase-Pure InAs Nanowires on Si(111) by Diminishing the Diameter to 10 nmMBE growth of Al/InAs and Nb/InAs superconducting hybrid nanowire structuresTwo-terminal charge tunneling: Disentangling Majorana zero modes from partially separated Andreev bound states in semiconductor-superconductor heterostructuresReproducing topological properties with quasi-Majorana statesEffects of Gate-Induced Electric Fields on Semiconductor Majorana NanowiresHybridization at Superconductor-Semiconductor InterfacesEffective theory approach to the Schrödinger-Poisson problem in semiconductor Majorana devicesLarge zero bias peaks and dips in a four-terminal thin InAs-Al nanowire deviceElectrical characteristics of field-effect transistors based on indium arsenide nanowire thinner than 10 nmSuspended InAs nanowire gate-all-around field-effect transistorsPhase-coherent transport and spin relaxation in InAs nanowires grown by molecule beam epitaxyContact properties of field-effect transistors based on indium arsenide nanowires thinner than 16 nmCrystal Phase- and Orientation-Dependent Electrical Transport Properties of InAs NanowiresMeasurements of the spin-orbit interaction and Landé g factor in a pure-phase InAs nanowire double quantum dot in the Pauli spin-blockade regimeCrossover from Coulomb blockade to ballistic transport in InAs nanowire devicesHard Superconducting Gap in InSb NanowiresBallistic superconductivity in semiconductor nanowiresElectric field tunable superconductor-semiconductor coupling in Majorana nanowiresMajorana bound state in a coupled quantum-dot hybrid-nanowire systemBallistic Majorana nanowire devicesSpin-resolved Andreev levels and parity crossings in hybrid superconductor–semiconductor nanostructuresScaling of Majorana Zero-Bias Conductance PeaksEnhanced Zero-Bias Majorana Peak in the Differential Tunneling Conductance of Disordered Multisubband Quantum-Wire/Superconductor JunctionsTowards a realistic transport modeling in a superconducting nanowire with Majorana fermionsCoulomb blockade of two-electron tunnelingExponential protection of zero modes in Majorana islandsConductance of a superconducting Coulomb-blockaded Majorana nanowireDecays of Majorana or Andreev Oscillations Induced by Steplike Spin-Orbit CouplingLarge zero-bias peaks in InSb-Al hybrid semiconductor-superconductor nanowire devicesDisorder-induced zero-bias peaks in Majorana nanowires
[1] Lutchyn R M, Sau J D, and Das S S 2010 Phys. Rev. Lett. 105 077001
[2] Oreg Y, Refael G, and von Oppen F 2010 Phys. Rev. Lett. 105 177002
[3] Freedman M H, Kitaev A, Larsen M J, and Wang Z H 2003 Bull. Amer. Math. Soc. 40 31
[4] Read N and Green D 2000 Phys. Rev. B 61 10267
[5] Kitaev A Y 2001 Phys.-Usp. 44 131
[6] Fu L and Kane C L 2008 Phys. Rev. Lett. 100 096407
[7] Lutchyn R M, Bakkers E P, Kouwenhoven L P, Krogstrup P, Marcus C M, and Oreg Y 2018 Nat. Rev. Mater. 3 52
[8] Zhang H, Liu D E, Wimmer M, and Kouwenhoven L P 2019 Nat. Commun. 10 5128
[9] Prada E, San-Jose P, de Moor M W, Geresdi A, Lee E J, Klinovaja J, Loss D, Nygård J, Aguado R, and Kouwenhoven L P 2020 Nat. Rev. Phys. 2 575
[10] Mourik V, Zuo K, Frolov S M, Plissard S R, Bakkers E P A M, and Kouwenhoven L P 2012 Science 336 1003
[11] Deng M T, Yu C L, Huang G Y, Larsson M, Caroff P, and Xu H Q 2012 Nano Lett. 12 6414
[12] Das A, Ronen Y, Most Y, Oreg Y, Heiblum M, and Shtrikman H 2012 Nat. Phys. 8 887
[13] Rokhinson L P, Liu X, and Furdyna J K 2012 Nat. Phys. 8 795
[14] Churchill H O H, Fatemi V, Grove-Rasmussen K, Deng M T, Caroff P, Xu H Q, and Marcus C M 2013 Phys. Rev. B 87 241401
[15] Chang W, Albrecht S M, Jespersen T S, Kuemmeth F, Krogstrup P, Nygård J, and Marcus C M 2015 Nat. Nanotechnol. 10 232
[16] Heedt S, Quintero-Pérez M, Borsoi F, Fursina A, van Loo N, Mazur G P, Nowak M P, Ammerlaan M, Li K, Korneychuk S, Shen J, van de An Y P M, Badawy G, Gazibegovic S, de Jong N, Aseev P, van Hoogdalem K, Bakkers E P A M, and Kouwenhoven L P 2021 Nat. Commun. 12 4914
[17] Pendharkar M, Zhang B, Wu H, Zarassi A, Zhang P, Dempsey C P, Lee J S, Harrington S D, Badawy G, Gazibegovic S, Op H V R L M, Rossi M, Jung J, Chen A H, Verheijen M A, Hocevar M, Bakkers E P A M, Palmstrøm C J, and Frolov S M 2021 Science 372 508
[18] Kanne T, Marnauza M, Olsteins D, Carrad D J, Sestoft J E, de Bruijckere J, Zeng L, Johnson E, Olsson E, Grove-Rasmussen K, and Nygård J 2021 Nat. Nanotechnol. 16 776
[19] Stanescu T D and Das S S 2013 Phys. Rev. B 87 180504(R)
[20] Takei S, Fregoso B M, Hui H, Lobos A M, and Das S S 2013 Phys. Rev. Lett. 110 186803
[21] Krogstrup P, Ziino N L B, Chang W, Albrecht S M, Madsen M H, Johnson E, Nygård J, Marcus C M, and Jespersen T S 2015 Nat. Mater. 14 400
[22] Kang J H, Grivnin A, Bor E, Reiner J, Avraham N, Ronen Y, Cohen Y, Kacman P, Shtrikman H, and Beidenkopf H 2017 Nano Lett. 17 7520
[23] Carrad D J, Bjergfelt M, Kanne T, Aagesen M, Krizek F, Fiordaliso E M, Johnson E, Nygård J, and Jespersen T S 2020 Adv. Mater. 32 1908411
[24] Vaitiekėnas S, Liu Y, Krogstrup P, and Marcus C M 2021 Nat. Phys. 17 43
[25] Sestoft J E, Kanne T, Gejl A N, von Soosten M, Yodh J S, Sherman D, Tarasinski B, Wimmer M, Johnson E, Deng M T, Nygå R J, Jespersen T S, Marcus C M, and Krogstrup P 2018 Phys. Rev. Mater. 2 044202
[26] Khan S A, Lampadaris C, Cui A, Stampfer L, Liu Y, Pauka S J, Cachaza M E, Fiordaliso E M, Kang J H, Korneychuk S, Mutas T, Sestoft J E, Krizek F, Tanta R, Cassidy M C, Jespersen T S, and Krogstrup P 2020 ACS Nano 14 14605
[27] Vaitiekėnas S, Whiticar A M, Deng M T, Krizek F, Sestoft J E, Palmstrøm C J, Martí-Sánchez S, Arbiol J, Krogstrup P, Casparis L, and Marcus C M 2018 Phys. Rev. Lett. 121 147701
[28] Lee J S, Choi S, Pendharkar M, Pennachio D J, Markman B, Seas M, Koelling S, Verheijen M A, Casparis L, Petersson K D, Petkovic I, Schaller V, Rodwell M J W, Marcus C M, Krogstrup P, Kouwenhoven L P, Bakkers E P A M, and Palmstrøm C J 2019 Phys. Rev. Mater. 3 084606
[29] Stiles M D and Hamann D R 1990 Phys. Rev. B 41 5280
[30] Stiles M D and Hamann D R 1988 Phys. Rev. B 38 2021
[31] Nilsson M, Namazi L, Lehmann S, Leijnse M, Dick K A, and Thelander C 2016 Phys. Rev. B 93 195422
[32] Prada E, San-Jose P, and Aguado R 2012 Phys. Rev. B 86 180503
[33] Kells G, Meidan D, and Brouwer P W 2012 Phys. Rev. B 86 100503
[34] Liu C X, Sau J D, Stanescu T D, and Das S S 2017 Phys. Rev. B 96 075161
[35] Akiyama T, Sano K, Nakamura K, and Ito T 2006 Jpn. J. Appl. Phys. 45 L275
[36] Caroff P, Dick K A, Johansson J, Messing M E, Deppert K, and Samuelson L 2009 Nat. Nanotechnol. 4 50
[37] Shtrikman H, Popovitz-Biro R, Kretinin A, Houben L, Heiblum M, Bukała M, Galicka M, Buczko R, and Kacman P 2009 Nano Lett. 9 1506
[38] Pan D, Fu M Q, Yu X Z, Wang X L, Zhu L J, Nie S H, Wang S L, Chen Q, Xiong P, von Molnár S, and Zhao J H 2014 Nano Lett. 14 1214
[39] Güsken N A, Rieger T, Zellekens P, Bennemann B, Neumann E, Lepsa M I, Schäpers T, and Grützmacher D 2017 Nanoscale 9 16735
[40] Moore C, Stanescu T D, and Tewari S 2018 Phys. Rev. B 97 165302
[41] Vuik A, Nijholt B, Akhmerov A, and Wimmer M 2019 SciPost Phys. 7 61
[42] Antipov A E, Bargerbos A, Winkler G W, Bauer B, Rossi E, and Lutchyn R M 2018 Phys. Rev. X 8 031041
[43] Mikkelsen A E G, Kotetes P, Krogstrup P, and Flensberg K 2018 Phys. Rev. X 8 031040
[44] Woods B D, Stanescu T D, and Das S S 2018 Phys. Rev. B 98 035428
[45] Song H D, Zhang Z T, Pan D, Liu D H, Wang Z Y, Cao Z Y, Liu L, Wen L J, Liao D Y, Zhuo R, Liu D, Shang R N, Zhao J H, and Zhang H 2021 arXiv:2107.08282 [cond-mat.mes-hall]
[46] Fu M Q, Pan D, Yang Y, Shi T, Zhang Z, Zhao J H, Xu H Q, and Chen Q 2014 Appl. Phys. Lett. 105 143101
[47] Li Q, Huang S Y, Pan D, Wang J Y, Zhao J H, and Xu H Q 2014 Appl. Phys. Lett. 105 113106
[48] Wang L B, Guo J K, Kang N, Pan D, Li S, Fan D, Zhao J H, and Xu H Q 2015 Appl. Phys. Lett. 106 173105
[49] Shi T, Fu M Q, Pan D, Guo Y, Zhao J H, and Chen Q 2015 Nanotechnology 26 175202
[50] Fu M Q, Tang Z Q, Li X, Ning Z Y, Pan D, Zhao J H, Wei X L, and Chen Q 2016 Nano Lett. 16 2478
[51] Wang J Y, Huang S Y, Lei Z J, Pan D, Zhao J H, and Xu H Q 2016 Appl. Phys. Lett. 109 53106
[52] Wang L B, Pan D, Huang G Y, Zhao J, Kang N, and Xu H Q 2019 Nanotechnology 30 124001
[53] Gül Ö, Zhang H, de Vries F K, van Veen J, Zuo K, Mourik V, Conesa-Boj S, Nowak M P, van Woerkom D J, Quintero-Pérez M, Cassidy M C, Geresdi A, Koelling S, Car D, Plissard S R, Bakkers E P A M, and Kouwenhoven L P 2017 Nano Lett. 17 2690
[54] Zhang H, Gül Ö, Conesa-Boj S, Nowak M P, Wimmer M, Zuo K, Mourik V, de Vries F K, van Veen J, de Moor M W, Bommer J D S, van Woerkom D J, Car D, Plissard S R, Bakkers E P A M, Quintero-Pérez M, Cassidy M C, Koelling S, Goswami S, Watanabe K, Taniguchi T, and Kouwenhoven L P 2017 Nat. Commun. 8 16025
[55] de Moor M W A, Bommer J D S, Xu D, Winkler G W, Antipov A E, Bargerbos A, Wang G, van Loo N, Op H V R L M, Gazibegovic S, Car D, Logan J A, Pendharkar M, Lee J S, Bakkers E P A M, Palmstrøm C J, Lutchyn R M, Kouwenhoven L P, and Zhang H 2018 New J. Phys. 20 103049
[56] Deng M T, Vaitiekėnas S, Hansen E B, Danon J, Leijnse M, Flensberg K, Nygård J, Krogstrup P, and Marcus C M 2016 Science 354 1557
[57] Gül Ö, Zhang H, Bommer J D S, de Moor M W A, Car D, Plissard S R, Bakkers E P A M, Geresdi A, Watanabe K, Taniguchi T, and Kouwenhoven L P 2018 Nat. Nanotechnol. 13 192
[58] Lee E J, Jiang X, Houzet M, Aguado R, Lieber C M, and de Franceschi S 2014 Nat. Nanotechnol. 9 79
[59] Nichele F, Drachmann A C C, Whiticar A M, O'Farrell E C T, Suominen H J, Fornieri A, Wang T, Gardner G C, Thomas C, Hatke A T, Krogstrup P, Manfra M J, Flensberg K, and Marcus C M 2017 Phys. Rev. Lett. 119 136803
[60] Pientka F, Kells G, Romito A, Brouwer P W, and von Oppen F 2012 Phys. Rev. Lett. 109 227006
[61] Rainis D, Trifunovic L, Klinovaja J, and Loss D 2013 Phys. Rev. B 87 024515
[62] Hekking F W J, Glazman L I, Matveev K A, and Shekhter R I 1993 Phys. Rev. Lett. 70 4138
[63] Albrecht S M, Higginbotham A P, Madsen M, Kuemmeth F, Jespersen T S, Nygård J, Krogstrup P, and Marcus C M 2016 Nature 531 206
[64] Chiu C K, Sau J D, and Das S S 2017 Phys. Rev. B 96 054504
[65] Cao Z, Zhang H, Lü H F, He W X, Lu H Z, and Xie X C 2019 Phys. Rev. Lett. 122 147701
[66] Zhang H, de Moor M W A, Bommer J D S, Xu D, Wang G Z, van Loo N, Liu C X, Gazibegovic S, Logan J A, Car D, Op H V R L M, van Veldhoven P J, Koelling S, Verheijen M A, Pendharkar M, Pennachio D J, Shojaei B, Lee J S, Palmstrøm C J, Bakkers E P A M, Das S S, and Kouwenhoven L P 2021 arXiv:2101.11456 [cond-mat.mes-hall]
[67] Das S S and Pan H N 2021 Phys. Rev. B 103 195158