Two signals emerge in the spectra
A laboratory study of SrSn3, a kagome superconductor, points to two superconducting signals: one associated with the surface and another with the bulk, or interior. The distinction emerged when researchers varied RN, the tunnel-junction resistance, which is equivalent to changing the tip-sample distance. The specimens were epitaxial SrSn3 films grown on Si(111) by molecular beam epitaxy and transferred in situ to an STM for measurement.
At large RN, the spectra emphasized the surface gap. At small RN, they instead showed a larger bulk-attributed gap. The latter reached about 1.87 meV when measured from one peak to the other, while a Dynes model fit to the conductance curve gave 1.4 meV with a reported uncertainty of 0.1 meV. This resistance sweep was the study's main way of separating the two signals.
The surface-associated gap followed the BCS form used in the analysis to describe how a superconducting gap changes with temperature. That fit gave a zero-temperature gap of about 0.68 meV, a transition temperature of about 4.2 K and a reduced gap ratio of about 3.8.
Thickness and field reinforce the split
Thickness comparisons added a second distinction. The surface gap remained nearly unchanged, while the bulk gap decreased by about 31 percent as thickness fell from roughly 78 nm to roughly 5 nm. The crossover resistance also decreased in the thinner films.
Field-dependent spectra gave a field limit of about 0.82 T for the surface channel and about 0.25 T for the bulk channel. The paper presents these values as estimates based on spectral evolution and extrapolation.
The vortex peak remains unresolved
The clearest contrast appeared in vortex-core spectra, measurements taken at the centres of vortices. In surface-dominated measurements, vortex cores showed pronounced zero-bias conductance peaks, an enhanced signal at zero energy. In bulk-dominated measurements, the spectra instead showed pronounced zero-bias suppression. The peak weight decreased in thinner films.
The zero-bias peaks showed no discernible spatial splitting. Their peak position and linewidth, or breadth, remained essentially unchanged away from the vortex centre. Supplemental spatial grid spectra found no discernible low-energy variation at any RN and showed uniform evolution as RN changed.
That pattern is notable, but it is not a standalone identification of a Majorana state. The authors conclude that a non-split zero-bias conductance peak alone is insufficient evidence for a Majorana zero mode in a multichannel superconductor. The absence of spatial splitting may disfavor some clean-limit interpretations of the vortex state, but it does not exclude dissipation linked to the tunnel junction or an intrinsic splitting below experimental resolution.
A measured result with a clear boundary
Taken together, the resistance-dependent spectra, thickness trend, field response and vortex data form the basis for the authors' interpretation of coexisting surface and bulk channels. The authors describe this as a spectroscopic separation, with the channel assignments following how the signals respond when RN, thickness and field are varied.
The work is an arXiv preprint. Its data statement says supporting data are available from the corresponding authors upon reasonable request.
Paper data and sources
Original title: Distinct Surface and Bulk Superconductivity in the Kagome Superconductor SrSn$_3$
Authors: Qun Zhu, Yong-Wei Wang, Ji-Hai Zhang et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text