Fe(Te,Se) zero-energy states: spin-resolved spectroscopy separates Majorana from Yu-Shiba-Rusinov

Why Fe(Te,Se) became a Majorana platform — and why zero-bias peaks were never decisive

Fe(Te,Se) earned its status as an intrinsic topological superconductor from theory and spectroscopy. Xu et al. predicted from ab initio calculations and an effective eight-band model that extended s-wave Fe-based superconductors such as Fe1+ySe0.5Te0.5 host a metallic topologically nontrivial band structure and can exhibit a normal-topological-normal superconductivity transition on the (001) surface, with a Majorana zero mode trapped at the end of a magnetic vortex line [9]. Song et al. later used first-principles calculations to argue that iron atoms in crystalline domain walls of FeSe0.45Te0.55 spontaneously form ferromagnetic order, imposing a π phase difference between the separated surface superconducting regimes and giving rise to one-dimensional Majorana modes trapped by the wall [3]. Reviews of the field summarized how Majorana zero modes and dispersing Majorana modes manifest in scanning tunneling spectroscopy — the former as a pronounced conductance peak precisely at zero energy, the latter as a constant or slowly varying density of states — and catalogued FeTe0.55Se0.45 and (Li0.84Fe0.16)OHFeSe among the leading platforms [6].

That same literature also flagged the ambiguity. The 2023 review by Li et al. noted that zero-energy conductance peaks are the expected signature of Majorana zero modes but did not claim they are unique to them [6]. The anchor paper by Wang et al. now supplies the missing discrimination: it reports near-zero-energy localized states at multiple structural defects on Fe(Te,Se) that could be misidentified as Majorana zero modes without additional verification, and uses spin-polarized scanning tunneling spectroscopy to assign them to topologically trivial Yu-Shiba-Rusinov states [1]. The distinction is not semantic. A zero-bias peak is a spectroscopic endpoint; a Majorana zero mode is a topological object with specific spin, spatial, and field-response properties. The new work tests those properties directly.

What spin-resolved spectroscopy changed: step edges, wrinkles, and line-defect bends

The anchor study examined three defect classes on freshly cleaved FeTe0.55Se0.45 (Tc = 14.5 K) at 1.1 K in ultrahigh vacuum, using a bulk Cr tip for spin-polarized measurements and magnetic fields up to 3 T [1]. At step edges, spatially resolved spectra showed two particle-hole symmetric peaks on one path and a zero-bias peak on another; the zero-bias peak split within a few nanometers and the split peaks vanished within about 8 nm of the edge [1]. The authors modeled this as two interacting Yu-Shiba-Rusinov bound states with different coupling strengths, plus temperature broadening, rather than a Majorana mode [1]. At Type I wrinkles, zero-energy density of states was strongly localized at some positions and absent at others along the same wrinkle — an observation the authors state clearly rules out a topological dispersing mode — and line spectra across the wrinkle showed two peaks symmetric about zero, compatible with a Yu-Shiba-Rusinov bound state [1].

The Type II line-defect bend is the most instructive case because it had been the strongest prior claim. Earlier work interpreted a flat in-gap line shape at a one-dimensional defect as a dispersing Majorana mode caused by a π phase shift of the lattice, mimicking the Fu-Kane model [1]. The anchor paper extracted the phase shift with a more refined method and found it near 0.5π for the first Bragg peak and almost no shift for the second — much less than π [1]. Spin-polarized spectra at three atomic-precision positions showed a density-of-states difference within ±0.8 mV whose sign reversed between positive and negative energies, similar to the spin polarization of Yu-Shiba-Rusinov bound states, and the crossing point of the +1 T and −1 T spectra was not at zero energy for two of the three positions [1]. The authors model this as an ensemble of irregularly coupled magnetic impurities below the surface, most plausibly excess Fe introduced during synthesis, and note that surface Fe adatoms have been reported repeatedly for this stoichiometry [1].

The harder case: zero-energy bound states in regions with no surface defects

The anchor paper also examined a region with no adatom or line defect visible at atomic resolution, where the zero-energy dI/dV map nonetheless showed a donut-shaped enhancement with two local maxima labeled point 1 and point 2 [1]. Both points showed zero-bias peaks with the same full width at half maximum near 0.5 mV, comparable to zero-bias peaks observed in vortex cores at lower temperature [1]. Spatial and magnetic-field-dependent spectroscopy then separated them: the peak showed no clear energy shift along two directions but a finite shift to positive energies along a third, and the peak at point 1 exhibited a magnetic response in spin-polarized spectra while point 2 did not [1]. Because a Majorana bound state should not shift in peak energy and is spin-polarized, the authors conclude that neither point can be explained by a Majorana bound state [1].

The point 2 result is presented with appropriate caution. The authors note that its non-dispersing zero-bias peak could in principle indicate a Majorana bound state, but the absence of an obvious magnetic response argues against a spin-polarized Majorana mode; they also acknowledge that when the defect is farther from the surface, spin sensitivity is reduced, so other explanations for point 2 may exist [1]. This is a finding with an explicit open edge, not a closed verdict. It also connects to a broader theoretical point: Pathak et al. showed analytically for an extended Kitaev chain that even within a single topological phase, Majorana zero modes can exhibit qualitatively distinct decay behaviors — monotonic, oscillatory, or perfectly localized — and that boundary-origin modes need not have their maximum probability at the edge, sometimes peaking at interior lattice sites [2]. Spatial profile alone, in other words, is not a reliable topological fingerprint, which strengthens the case for the multi-observable approach the anchor paper advocates.

How this compares with precursor claims and competing explanations

The most direct precursor is the domain-wall Majorana proposal. Song et al. argued that ferromagnetic order of iron atoms in the domain wall imposes a π phase difference between the separated surface superconducting regimes, producing one-dimensional Majorana modes trapped by the wall, and proposed a FeSe0.45Te0.55–ferromagnet–FeSe0.45Te0.55 junction to create, fuse, and braid Majorana zero modes [3]. The anchor paper directly addresses the phase-shift premise of that class of claim, reporting a measured shift near 0.5π rather than π at the Type II line-defect bend and citing prior work that reached the same conclusion with a refined phase-extraction method [1]. The disagreement is methodological, not merely interpretive: the two results rest on different phase-extraction procedures applied to the same kind of lattice-shift measurement, and the anchor paper's spin-resolved spectra provide an independent, non-phase-based argument for a trivial origin [1].

A second precursor line comes from engineered platforms. Chen et al. deposited Bi islands on Fe(Te,Se) and reported robust zero-energy modes appearing everywhere on some islands, attributing them to proximity-induced topological superconductivity on the Bi islands with strong spin-orbit coupling [7]. That claim rests on a different heterostructure and a different proximity mechanism, so it is not directly contradicted by the anchor paper's surface-defect results; it does, however, share the same evidentiary vulnerability — a zero-energy mode assigned to a topological origin without spin-resolved verification. On the theory side, Hein et al. analyze Yu-Shiba-Rusinov state formation in Ising superconductors with spin-orbit coupling and an in-plane magnetic field, treating magnetic impurities as local probes of the superconducting condensate [4]. That framework is directly relevant to the anchor paper's interpretation, because it shows how spin-orbit coupling and magnetic field reshape Yu-Shiba-Rusinov states — precisely the ingredients the anchor paper invokes when it models coupled magnetic impurities below the surface [1][4].

What remains uncertain, and what would settle it

The anchor paper's conclusions are bounded by its method. All results come from surface scanning tunneling spectroscopy on cleaved Fe(Te,Se) at 1.1 K with fields up to 3 T; the study does not probe bulk states, does not perform device-level braiding or interferometry, and does not establish whether any topological phase exists elsewhere in the same crystals [1]. The authors themselves flag the point 2 ambiguity and the reduced spin sensitivity for deeper defects [1]. The theoretical literature reinforces why this matters: Rivière et al. show that in interacting Majorana X- and Y-junctions, exact zero modes arise from incommensurate short-range correlations and manifest as level crossings between in-gap states, with the number and character of in-gap states depending on junction geometry and the sign and strength of the coupling [5]. Zero modes in realistic geometries are therefore not a single universal signature, and a spectroscopic zero-bias peak cannot be mapped onto a topological invariant without additional structure.

The practical consequence for the field is a revised checklist rather than a rejection of Fe(Te,Se). The anchor paper's own summary emphasizes using multiple methods to examine the nature of zero-bias peaks [1], and the broader platform literature already treats Fe(Te,Se) as one of several candidate hosts alongside nanowire, atomic-chain, and oxide-dislocation approaches [6][8][10][11][12]. What would move the question forward is exactly what the anchor paper could not supply: spin-resolved measurements extended to vortex cores and to engineered junctions, combined with transport or interferometric tests that probe non-Abelian exchange rather than spectral shape alone. Until then, the defensible reading is that Fe(Te,Se) remains a plausible topological superconductor, but the specific near-zero-energy states reported at its common surface defects are better explained by trivial Yu-Shiba-Rusinov physics [1].

About These Sources

This research page is built on 12 peer-reviewed studies — published from 2016 to 2026, 4 from 2024 or later — selected as the most relevant from 12 studies that passed quality screening, drawn from 57 papers retrieved from a database of over 500 million.

Sources used in this answer