Quantum transport signatures of Majorana zero modes in proximitized nanowires

Introduction

Majorana zero modes in one-dimensional topological superconductors are predicted to appear as spatially separated zero-energy boundary states with non-Abelian exchange properties. The conceptual foundation goes back to the Kitaev chain [1], while semiconductor-superconductor nanowires with spin-orbit coupling, Zeeman splitting, and induced superconductivity provide a practical solid-state route [2,3]. In tunnelling spectroscopy, an isolated Majorana mode is expected to produce resonant Andreev reflection and an ideal zero-temperature conductance of 2e^2/h under appropriate coupling conditions [4].

Experiments beginning in 2012 reported zero-bias conductance peaks in hybrid nanowire devices and established the field of Majorana nanowire transport [5,6,7]. Device quality improved substantially with epitaxial semiconductor-superconductor interfaces and hard induced gaps [8,9]. However, the interpretation of zero-bias peaks remains difficult. Local Andreev bound states, quantum-dot levels, disorder, smooth confinement, finite temperature, and dissipation can all produce peaks that appear stable over finite magnetic-field and gate ranges [10,11,12]. Several reviews and perspective articles therefore argue for multi-terminal and multi-probe evidence standards rather than relying on local tunnelling alone [13,14].

The present study follows that philosophy. We do not claim a topological demonstration. Instead, we ask which combination of transport measurements most effectively rejects trivial alternatives in a realistic device set. The article reports a fictional but technically bounded screening workflow for gate-defined InAs-Al nanowires, combining local conductance, nonlocal conductance, Coulomb-blockade island spectroscopy, and disorder-aware simulation.

Device fabrication and measurement protocol

The devices were fabricated from molecular-beam-epitaxy grown InAs nanowires with two facets covered by approximately 8 nm of epitaxial Al. Individual wires were transferred to oxidised Si substrates with local bottom gates, Ti/Au normal contacts, and selectively etched Al segments defining tunnel barriers, proximitized wire sections, and superconducting islands. The proximitized segment lengths ranged from 0.9 to 1.8 micrometres. Devices were measured in a dilution refrigerator with base electron temperature calibrated between 38 and 55 mK using Coulomb-peak thermometry and superconducting gap-edge fits.

Local tunnelling conductance was measured using a 3-5 micro V rms lock-in excitation at 17 Hz. For devices with contacts at both wire ends, simultaneous two-terminal conductance was measured with independent low-frequency excitations, allowing extraction of local conductance at each end and nonlocal conductance through the proximitized segment. Magnetic field was applied nominally along the nanowire axis, with transverse misalignment below 3 degrees estimated from orbital suppression of the induced gap. Gate sweeps were repeated after thermal cycling for four devices to estimate reproducibility.

A subset of six devices was configured as floating superconducting islands by depleting both end barriers. Coulomb-blockade peak spacings were measured as a function of axial field and plunger gate. We use island parity evolution as a supporting diagnostic because overlapping Majorana modes can produce characteristic even-odd peak-spacing changes, although quasiparticle poisoning and subgap states can complicate the interpretation [15].

Simulation and classification criteria

Transport simulations used a one-dimensional Bogoliubov-de Gennes tight-binding model with Rashba spin-orbit coupling, induced pairing, orbital-free Zeeman splitting, smooth electrostatic confinement, and random onsite disorder. The induced gap, effective g factor, and spin-orbit energy were drawn from ranges consistent with the measured gap closing fields and normal-state subband spacings. Disorder was represented by Gaussian-correlated potentials with correlation lengths from 30 to 120 nm. The simulations were not fitted device by device; they were used to compare how often different physical mechanisms reproduce the same qualitative signatures.

A local zero-bias peak was labelled stable if its centre stayed within 8 micro eV of zero bias over at least 200 mT and if its height exceeded three times the local background conductance. A nonlocal gap-reopening signature required correlated changes in conductance at both ends and a finite-bias gap edge that closed and reopened with field. An island parity signature required a reproducible reduction of even-odd spacing contrast over a field range overlapping the local zero-bias regime. These thresholds are conservative but not definitive. They are intended to reduce false positives, not to prove topology.

We classified each device into four evidence tiers: no stable subgap state, local-only zero-bias stability, local plus nonlocal consistency, and local plus nonlocal plus island consistency. This tiered scheme follows theoretical work showing that end-to-end correlations are more discriminating than local spectra when distinguishing Majorana bound states from ordinary Andreev bound states [16,17].

Results

All devices showed hard induced gaps at zero magnetic field, with subgap conductance suppression between factors of 22 and 83 relative to normal-state conductance. The induced gap extracted from tunnelling spectra was 185 +/- 24 micro eV. Five of the 16 devices developed zero-bias peaks that satisfied the local-stability criterion. The median field window for those peaks was 260 mT, beginning near 0.55 T and ending before the parent Al gap fully collapsed. Peak heights varied strongly with tunnel-barrier setting and temperature. After thermal broadening correction, the largest peak height was 1.55 +/- 0.18 e^2/h, below the ideal resonant value.

Local peaks did not always correlate between wire ends. In two devices, the left-end peak remained pinned near zero while the right-end spectrum showed only dispersing subgap states. In one device, both ends showed near-zero peaks but in non-overlapping gate windows. Two devices showed simultaneous local peaks at both ends and a nonlocal conductance feature consistent with gap closing and reopening. The nonlocal signal was small, typically below 0.015 e^2/h, and required averaging over repeated sweeps to distinguish from charge rearrangement noise.

Coulomb-blockade measurements provided an additional filter. Of the two devices with local and nonlocal consistency, one showed a reproducible reduction in even-odd peak-spacing contrast over the same field range as the zero-bias peaks. The second showed irregular peak motion and intermittent quasiparticle poisoning, preventing a stable parity assignment. No device showed all signatures with quantized local conductance. We therefore classify one device as tier 3, one as tier 2, three as local-only, and the remaining eleven as non-candidates under the present protocol.

The disorder ensemble reproduced several local-only observations. Smooth confinement with weak disorder generated stable zero-bias peaks in 9% of simulated parameter sets without a topological invariant change, in agreement with previous demonstrations of trivial zero-bias mechanisms [11,12,18]. Simultaneous local stability and nonlocal gap reopening occurred in only 1.7% of trivial simulations, while adding island parity consistency reduced the false-positive fraction below 1% in our ensemble. These percentages depend on the assumed disorder model and should not be treated as universal.

Discussion

The results reinforce a cautious interpretation of Majorana nanowire transport. Local zero-bias peaks remain valuable because they identify devices and gate regimes worth studying, but they are not sufficient evidence for a topological phase. The strongest candidates in this dataset are not the devices with the tallest local peaks; they are the devices where local spectra, nonlocal response, and island behaviour are mutually consistent. This agrees with the broader field movement toward nonlocal diagnostics and structured evidence protocols [10,13,14].

The absence of quantized 2e^2/h conductance is not by itself fatal. Finite temperature, tunnel-rate asymmetry, dissipation, soft residual subgap states, and multi-subband coupling can all reduce the observed peak height. Conversely, nearly quantized conductance can be produced by non-Majorana states under some conditions [12]. Peak height should therefore be read together with its dependence on barrier transparency, temperature, end-to-end correlations, and island parity. In our devices, the conductance data are consistent with strongly broadened end states but do not establish ideal resonant Andreev reflection.

The main limitation is statistical. Sixteen devices are enough to evaluate a screening workflow but not enough to estimate a fabrication yield for topological devices. The one-dimensional simulations also omit electrostatic self-consistency, orbital magnetic effects, faceted geometry, and microscopic interface disorder. These omissions matter because nanowire devices are sensitive to local potential variations and contact details. Future work should combine multi-terminal transport with self-consistent electrostatic modelling and deliberately varied disorder profiles.

A practical outcome is that device screening should be sequential. Local spectroscopy can triage devices quickly. Candidate regimes should then be tested by nonlocal conductance with independent contacts, followed by island measurements where fabrication permits. A device that fails the nonlocal test should not be promoted on the basis of local peak stability alone. This is a stricter standard than much early work used, but it is better matched to what is now known about false positives.

Conclusion

We examined quantum transport signatures in 16 proximitized InAs-Al nanowire devices using local tunnelling, nonlocal conductance, Coulomb-blockade island spectroscopy, and disorder-aware modelling. Five devices showed stable local zero-bias peaks, but only two also showed nonlocal consistency, and only one additionally showed reproducible island parity evolution. The study therefore supports a tiered evidence standard rather than a single-signature interpretation.

The data are compatible with Majorana-like end-state formation in a small subset of devices but do not constitute a topological demonstration. The strongest conclusion is methodological: local zero-bias peaks should trigger further nonlocal and island tests, not serve as standalone evidence.

Data and code availability

Processed conductance maps, Coulomb-blockade peak tables, electron-temperature fits, and simulation input files are included in the supplementary archive. Raw lock-in time traces are available as compressed HDF5 files. The tight-binding simulations were run with Python 3.10, NumPy 1.23, SciPy 1.9, and Kwant 1.4.

References

1. Kitaev, A. Y. Unpaired Majorana fermions in quantum wires. Phys. Usp. 44, 131-136 (2001).
2. Lutchyn, R. M., Sau, J. D. & Das Sarma, S. Majorana fermions and a topological phase transition in semiconductor-superconductor heterostructures. Phys. Rev. Lett. 105, 077001 (2010).
3. Oreg, Y., Refael, G. & von Oppen, F. Helical liquids and Majorana bound states in quantum wires. Phys. Rev. Lett. 105, 177002 (2010).
4. Law, K. T., Lee, P. A. & Ng, T. K. Majorana fermion induced resonant Andreev reflection. Phys. Rev. Lett. 103, 237001 (2009).
5. Mourik, V. et al. Signatures of Majorana fermions in hybrid superconductor-semiconductor nanowire devices. Science 336, 1003-1007 (2012).
6. Das, A. et al. Zero-bias peaks and splitting in an Al-InAs nanowire topological superconductor as a signature of Majorana fermions. Nat. Phys. 8, 887-895 (2012).
7. Deng, M. T. et al. Anomalous zero-bias conductance peak in a Nb-InSb nanowire-Nb hybrid device. Nano Lett. 12, 6414-6419 (2012).
8. Krogstrup, P. et al. Epitaxy of semiconductor-superconductor nanowires. Nat. Mater. 14, 400-406 (2015).
9. Chang, W. et al. Hard gap in epitaxial semiconductor-superconductor nanowires. Nat. Nanotechnol. 10, 232-236 (2015).
10. Albrecht, S. M. et al. Exponential protection of zero modes in Majorana islands. Nature 531, 206-209 (2016).
11. Zhang, H., Liu, D. E., Wimmer, M. & Kouwenhoven, L. P. Next steps of quantum transport in Majorana nanowire devices. Nat. Commun. 10, 5128 (2019).
12. Pan, H. & Das Sarma, S. Physical mechanisms for zero-bias conductance peaks in Majorana nanowires. Phys. Rev. Res. 2, 013377 (2020).
13. Yu, P. et al. Non-Majorana states yield nearly quantized conductance in proximatized nanowires. Nat. Phys. 17, 482-488 (2021).
14. Lai, Y.-H., Sau, J. D. & Das Sarma, S. Presence versus absence of end-to-end nonlocal conductance correlations in Majorana nanowires: Majorana bound states versus Andreev bound states. Phys. Rev. B 100, 045302 (2019).
15. Poschl, A. et al. Nonlocal conductance spectroscopy of Andreev bound states in gate-defined InAs/Al nanowires. Phys. Rev. B 106, L241301 (2022).
16. Prada, E. et al. From Andreev to Majorana bound states in hybrid superconductor-semiconductor nanowires. Nat. Rev. Phys. 2, 575-594 (2020).
17. Liu, J., Potter, A. C., Law, K. T. & Lee, P. A. Zero-bias peaks in the tunneling conductance of spin-orbit-coupled superconducting wires with and without Majorana end-states. Phys. Rev. Lett. 109, 267002 (2012).
18. Das Sarma, S. & Pan, H. Disorder-induced zero-bias peaks in Majorana nanowires. Phys. Rev. B 103, 195158 (2021).