FeTe1-xSex: Tuning, Structure & Superconductivity

• FeTe1-xSex is an iron-chalcogenide superconductor characterized by a tetragonal PbO-type lattice and tunable structural phases via Se substitution.
• Increasing Se content induces a monotonic lattice contraction that shifts the material from antiferromagnetic order to bulk superconductivity with distinct transition temperatures.
• Interstitial Fe, annealing processes, and nanoscale chemical inhomogeneity critically influence its magnetic, superconducting, and topological surface states.

FeTe$_{1-x}$Se$_x$ (FTS), also written Fe(Te,Se), is an iron-chalcogenide “11” superconductor in which isovalent Te$\rightarrow$Se substitution continuously tunes a structurally simple Fe–Ch layered lattice (Ch = Te, Se) across antiferromagnetic, superconducting, nematic, and topological regimes. Across bulk crystals, thin films, and nanostructures, the system is consistently described as tetragonal PbO-type or anti-PbO type with space group P4/nmm at room temperature, while its low-temperature ground state depends strongly on Se content, interstitial Fe, annealing history, strain, and local chemical inhomogeneity (Maheshwari et al., 2019).

1. Crystal chemistry and average structure

At room temperature, FTS crystallizes in the tetragonal P4/nmm structure over broad composition ranges. Single-crystal studies on $0 \le x \le 0.50$ report a tetragonal PbO-type phase with Fe at $(3/4, 1/4, 0)$ and chalcogen atoms at $(1/4, 1/4, z)$, while aberration-corrected STEM describes the same average structure in anti-PbO language, with lattice Fe on the 2a Wyckoff site and chalcogens on 2c (Maheshwari et al., 2019). In thin films spanning the full substitution series $0 \le x \le 1$, all films likewise retain a tetragonal structure at room temperature, enabling an “unabridged” phase diagram unobscured by the phase separation that affects bulk targets in the interval $0.6 \le x \le 0.8$ (Zhuang et al., 2014).

The dominant structural trend with increasing Se content is monotonic lattice contraction, especially along $c$. In single crystals, representative Rietveld values evolve from $a = 3.8263(2)$ Å, $_x$0 Å, $_x$1 Å$_x$2 at $_x$3 to $_x$4 Å, $_x$5 Å, $_x$6 Å$_x$7 at $_x$8, with the chalcogen height $_x$9 decreasing from approximately $\rightarrow$0 Å to $\rightarrow$1 Å (Maheshwari et al., 2019). In the underdoped regime $\rightarrow$2 to $\rightarrow$3, refinement gives $\rightarrow$4 Å, $\rightarrow$5 Å for FeTe and $\rightarrow$6 Å, $\rightarrow$7 Å for $\rightarrow$8, directly corroborating the shift of $\rightarrow$9 reflections toward higher $0 \le x \le 0.50$0 with Se substitution (Maheshwari et al., 2016). The paper explicitly relates this to Bragg’s law, $0 \le x \le 0.50$1, together with $0 \le x \le 0.50$2 for $0 \le x \le 0.50$3 peaks.

Morphologically, single crystals are slab-like and strongly layered, with only $0 \le x \le 0.50$4 reflections visible from plate surfaces and SEM images showing plate morphology with clear cleavage parallel to the layers (Maheshwari et al., 2016). This layered microstructure is consistent with the Fe–Ch stacking motif emphasized in STM studies, where an Fe layer is sandwiched between two chalcogen layers and the topmost Te and Se atoms can be directly distinguished in real space (Zhao et al., 2021).

Average crystallography, however, is not the whole structural story. STEM resolves excess interstitial Fe, denoted Fe$0 \le x \le 0.50$5 or Fe(2), on the 2c site in the chalcogen layer, with Fe$0 \le x \le 0.50$6 concentration decreasing from approximately $0 \le x \le 0.50$7 in Fe$0 \le x \le 0.50$8Te to $0 \le x \le 0.50$9 in Fe$(3/4, 1/4, 0)$0Te$(3/4, 1/4, 0)$1Se$(3/4, 1/4, 0)$2 and $(3/4, 1/4, 0)$3 in Fe$(3/4, 1/4, 0)$4Te$(3/4, 1/4, 0)$5Se$(3/4, 1/4, 0)$6 (Ma et al., 2024). This local occupancy is central to the magnetic and superconducting phenomenology discussed below.

2. Magneto-structural parent state and the low-Se regime

The Te-rich end member is a first-order magneto-structural system rather than a superconductor. In transport studies on underdoped single crystals, FeTe shows a sharp resistive step near $(3/4, 1/4, 0)$7 K that shifts to lower temperature with Se substitution, reaching around $(3/4, 1/4, 0)$8 K at $(3/4, 1/4, 0)$9 (Maheshwari et al., 2016). In a broader single-crystal series, the same transition is tracked as $(1/4, 1/4, z)$0 K for $(1/4, 1/4, z)$1, $(1/4, 1/4, z)$2 K for $(1/4, 1/4, z)$3, $(1/4, 1/4, z)$4 K for $(1/4, 1/4, z)$5, $(1/4, 1/4, z)$6 K for $(1/4, 1/4, z)$7, and $(1/4, 1/4, z)$8 K for $(1/4, 1/4, z)$9, with no resistive signature for $0 \le x \le 1$0 (Maheshwari et al., 2019).

Neutron powder diffraction resolves the structural component of this transition. For $0 \le x \le 1$1, the high-temperature tetragonal P4/nmm phase transforms on cooling to monoclinic P2$0 \le x \le 1$2/m, and the structural transition temperature $0 \le x \le 1$3 remains coincident with the AFM ordering temperature $0 \le x \le 1$4 within experimental uncertainty: $0 \le x \le 1$5 K at $0 \le x \le 1$6, $0 \le x \le 1$7 K at $0 \le x \le 1$8, $0 \le x \le 1$9 K at $0.6 \le x \le 0.8$0, and $0.6 \le x \le 0.8$1 K at $0.6 \le x \le 0.8$2 (Martinelli et al., 2010). The associated magnetic order is commensurate with propagation vector $0.6 \le x \le 0.8$3, equivalent to the bi-collinear or double-stripe order commonly summarized in the 1-Fe Brillouin zone as $0.6 \le x \le 0.8$4 (Dong et al., 2010).

The first-order character is visible in resistivity hysteresis. In low-Se single crystals the hysteresis width $0.6 \le x \le 0.8$5 narrows systematically from approximately $0.6 \le x \le 0.8$6 K in FeTe to $0.6 \le x \le 0.8$7 K at $0.6 \le x \le 0.8$8, $0.6 \le x \le 0.8$9 K at $c$0, and $c$1 K at $c$2, indicating that Se weakens the first-order magneto-structural transition (Maheshwari et al., 2016). Mössbauer spectroscopy provides a complementary local-probe view: at 5 K the average hyperfine field decreases from $c$3 T at $c$4 to $c$5 T at $c$6, $c$7 T at $c$8, and $c$9 T at $a = 3.8263(2)$0, while no hyperfine field related to magnetic ordering is resolved at $a = 3.8263(2)$1 (Maheshwari et al., 2019).

A recurrent theme in this regime is coexistence rather than abrupt replacement. Superconductivity appears for $a = 3.8263(2)$2 in neutron work, so that $a = 3.8263(2)$3 hosts superconductivity together with long-range AFM and monoclinicity (Martinelli et al., 2010). Transport and Mössbauer studies on single crystals likewise identify a coexistence regime around $a = 3.8263(2)$4 to $a = 3.8263(2)$5 (Maheshwari et al., 2019). This immediately qualifies a common oversimplification: in FTS, AFM and superconductivity are not generically separated by a clean nonmagnetic interval.

3. Superconductivity, phase diagrams, and sample dependence

The superconducting phase diagram of FTS is highly sensitive to interstitial Fe, annealing protocol, and sample form. In as-grown single crystals examined by magnetic susceptibility, long-range AFM is fully suppressed only for $a = 3.8263(2)$6, weak superconductivity appears when $a = 3.8263(2)$7, and bulk superconductivity emerges only beginning at $a = 3.8263(2)$8 (Kawasaki et al., 2011). By contrast, air-annealed crystals with partially removed excess Fe yield a revised phase diagram in which bulk superconductivity appears as low as $a = 3.8263(2)$9 and coexists with AFM over $_x$00 (Dong et al., 2010). Oxygen annealing at 300 °C for 2 h under 1 atm O$_x$01 produces yet another boundary set, with complete suppression of magnetic order and bulk superconductivity for $_x$02 (Kawasaki et al., 2011).

Single-crystal transport studies on $_x$03 self-flux samples place the onset of superconductivity near $_x$04, with $_x$05 K at $_x$06, $_x$07 K at $_x$08, $_x$09 K at $_x$10, $_x$11 K at $_x$12, and $_x$13–$_x$14 K at $_x$15, where $_x$16 K (Maheshwari et al., 2019). In these crystals, room-temperature Raman spectra show the allowed phonons Eg(Te/Se) near $_x$17 cm$_x$18, A$_x$19(Te/Se) near $_x$20 cm$_x$21, and B$_x$22(Fe) near $_x$23 cm$_x$24, all hardening slightly with Se content, which is consistent with lattice contraction (Maheshwari et al., 2019).

Thin films shift the superconducting optimum. In single-phased FeSe$_x$25Te$_x$26 films grown over the full $_x$27 range, the AFM kink of FeTe near $_x$28 K disappears by $_x$29, superconductivity extends from approximately $_x$30 to $_x$31, and the highest $_x$32 K occurs in the newly identified interval $_x$33 rather than at the canonical bulk composition near $_x$34 (Zhuang et al., 2014). This thin-film result shows that the apparent bulk optimum at $_x$35 is not universal and can be masked by phase separation in bulk specimens.

The superconducting state is also robust against magnetic field. For FeTe$_x$36Se$_x$37 single crystals, a Ginzburg–Landau analysis gives $_x$38 T, $_x$39 T, and $_x$40 T for the 90%, 50%, and 10% resistivity criteria, respectively, while the thermally activated flux-flow barrier $_x$41 decreases from about $_x$42 meV at $_x$43 T to $_x$44 meV at $_x$45 T (Maheshwari et al., 2019). In thin films, WHH analysis yields $_x$46 T for $_x$47, with $_x$48 nm (Zhuang et al., 2014).

Another strong form of tuning is Fe-site substitution. In FeTe$_x$49Se$_x$50 crystals grown by Bridgman’s method, only Co, Ni, and Cu substitute into the Fe sublattice, and superconductivity is fully suppressed by roughly $_x$51 at% Co, $_x$52 at% Ni, and $_x$53–$_x$54 at% Cu (Gawryluk et al., 2010). This establishes that the superconducting state is extremely sensitive to disorder within the magnetic Fe network.

4. Interstitial Fe, annealing, and nanoscale inhomogeneity

Interstitial Fe is one of the defining non-idealities of FTS. In Fe$_x$55Te$_x$56Se$_x$57, excess Fe is magnetic, acts as a pair breaker, and contributes to carrier localization; removing or compensating it substantially reshapes the phase diagram (Dong et al., 2010). Air annealing of an $_x$58 crystal at 270 °C for 2 h reduces the Fe content from Fe$_x$59 to nearly Fe$_x$60, while preserving the Se/Te ratio, and converts a non-bulk-superconducting sample into a bulk superconductor (Dong et al., 2010). Oxygen annealing introduces approximately $_x$61 O per formula unit in Fe$_x$62Te$_x$63Se$_x$64, a value comparable to the typical interstitial-Fe content $_x$65, and is interpreted as electronic compensation of the electrons donated by interstitial Fe (Kawasaki et al., 2011).

At the atomic scale, STEM directly images Fe$_x$66 ordering in Fe$_x$67Te. Selected-area electron diffraction reveals a superstructure with wave vector $_x$68 along [010], corresponding to a $_x$69 supercell, while FFTs along [001] show satellite reflections characterized by $_x$70 with domain-dependent orientation (Ma et al., 2024). These interstitial atoms form what the study terms an iron polycomplex: Fe$_x$71 interacts with neighboring lattice Fe, contracts local Fe–Fe distances, distorts the FeCh$_x$72 tetrahedra, and lowers the local anion height. Se substitution suppresses both Fe$_x$73 concentration and Fe$_x$74 ordering, and no superlattice spots are observed in the $_x$75 and $_x$76 samples (Ma et al., 2024).

Chemical disorder persists even when Fe$_x$77 is reduced. HAADF-STEM shows random Te/Se substitution and Te-rich or Se-rich domains of characteristic size around $_x$78 nm in Fe$_x$79Te$_x$80Se$_x$81 and Fe$_x$82Te$_x$83Se$_x$84 (Ma et al., 2024). The same paper notes $_x$85 nm in FTS, implying that pairing samples several local chemical environments.

Nanostructure transport makes the inhomogeneity operational. In Fe(Te$_x$86Se$_x$87) flakes, superconductivity is progressively suppressed as thickness is reduced, and zero resistance disappears at a critical thickness $_x$88 nm (Yue et al., 2016). At $_x$89 nm the resistivity shows only a partial superconducting drop, whereas at $_x$90 nm and $_x$91 nm no zero-resistance state survives. The study explicitly connects this threshold to the $_x$92–$_x$93 nm scale of Te/Se fluctuations reported by STEM/EELS and interprets the transport as direct evidence for nanoscale inhomogeneous, percolative superconductivity (Yue et al., 2016). This suggests that in FTS the superconducting path is often an emergent network rather than a spatially uniform condensate.

5. Electronic coherence, nematicity, and strain

ARPES shows that superconductivity in FTS is tied more closely to electronic coherence than to large Fermi-surface reconstruction. Over $_x$94, the Fermi-surface topology is nearly unchanged, yet the spectral line shape evolves strongly: broad, incoherent spectra with small quasiparticle weight at low $_x$95 sharpen progressively with Se substitution, and a well-defined quasiparticle peak appears around $_x$96, where bulk superconductivity is realized (Ieki et al., 2014). In FeTe, by contrast, the normal-state spectra are highly incoherent, although a sharp quasiparticle appears below $_x$97 in the AFM-reconstructed state. This disconnect between nearly fixed Fermi-surface geometry and rapidly changing spectral coherence argues against a purely nesting-based description of the superconducting onset.

Near the composition $_x$98, spectroscopic-imaging STM resolves an additional electronic instability: nematicity. At $_x$99 the system is tetragonal with no detectable electronic nematicity; at $\rightarrow$00 it breaks into micrometer-scale puddles, some with pronounced unidirectional nematic electronic modulations and some without; and at $\rightarrow$01 nematic modulations are pervasive (Zhao et al., 2021). The anisotropy is visible in QPI as an arc-like scattering vector oriented along the $\rightarrow$02 axis, with representative magnitude near the Fermi level of $\rightarrow$03 Å$\rightarrow$04.

The superconducting gap remains comparatively stable across these nematic textures. Local spectroscopy finds $\rightarrow$05 meV across $\rightarrow$06, $\rightarrow$07, and $\rightarrow$08, but in strongly nematic regions near $\rightarrow$09 the coherence peaks are drastically suppressed or vanish and the gap fills with in-gap spectral weight (Zhao et al., 2021). The relative coherence peak height,

$\rightarrow$10

anti-correlates with anisotropic strain with coefficient approximately $\rightarrow$11, while the local QPI amplitude correlates with strain with coefficient approximately $\rightarrow$12. The relevant strain scale is only about $\rightarrow$13 in $\rightarrow$14.

These results establish a distinction between gap magnitude and superconducting coherence. Static nematic order, plausibly interpreted as pinned nematic fluctuations near a critical composition, degrades coherence or superfluid density without measurably shifting $\rightarrow$15 (Zhao et al., 2021). A plausible implication is that multiple low-energy phenomena in FTS—superconductivity, nematicity, and topology—must be interpreted in a strongly local framework rather than solely through spatially averaged bulk phase diagrams.

6. Topology, correlated surface states, and time-reversal symmetry breaking

FTS is also a correlated topological material. In the band-inversion picture, non-trivial $\rightarrow$16 topology arises along $\rightarrow$17–$\rightarrow$18 when an odd-parity band derived from hybridized Fe $\rightarrow$19 and chalcogen $\rightarrow$20 states crosses an even-parity Fe $\rightarrow$21 band, yielding a Dirac-like topological surface state at $\rightarrow$22 (Kim et al., 23 Jul 2025). A recent ARPES study directly locates the lower topological boundary between $\rightarrow$23 and $\rightarrow$24: no observable $\rightarrow$25-centered topological surface state appears at $\rightarrow$26, whereas clear Dirac-cone-like surface states are present at $\rightarrow$27, $\rightarrow$28, and $\rightarrow$29 (Kim et al., 23 Jul 2025).

The same work shows that topology is fragile in practice under strong correlations. At $\rightarrow$30, the surface state is well defined at $\rightarrow$31 K, broadens rapidly between $\rightarrow$32 and $\rightarrow$33 K, and is nearly indistinguishable above $\rightarrow$34 K, where the MDC-derived mean free path falls below $\rightarrow$35 Å (Kim et al., 23 Jul 2025). Yet the bulk band inversion itself remains essentially unchanged up to $\rightarrow$36 K at $\rightarrow$37. This is interpreted as a consequence of an orbital-selective Mott phase: the bulk $\rightarrow$38 invariant survives, but OSMP-related self-energy effects destroy the coherence of the topological surface quasiparticles.

A complementary DFT and model study emphasizes that the topological scale is also highly sensitive to Se concentration and local Se/Te arrangement. The SOC gap $\rightarrow$39 at the avoided crossing between the chalcogen $\rightarrow$40 and Fe $\rightarrow$41 bands is zero at $\rightarrow$42 and $\rightarrow$43, follows a dome-like composition dependence, and reaches its largest reported value of $\rightarrow$44 meV near the center of the series (Wang et al., 2023). At fixed $\rightarrow$45, explicit 2$\rightarrow$462 supercell configurations yield local SOC gaps spanning $\rightarrow$47–$\rightarrow$48 meV, and one configuration is topologically trivial with $\rightarrow$49 (Wang et al., 2023). This provides a microscopic route to the patchy occurrence of Majorana-like vortex signatures reported in the broader literature.

A further complication is time-reversal symmetry breaking (TRSB) at the surface. Sagnac SMOKE and ARPES measurements on FeTe$\rightarrow$50Se$\rightarrow$51 and FeTe$\rightarrow$52Se$\rightarrow$53 detect spontaneous polar Kerr signals and a Dirac gap in the topological surface state, while bulk AC susceptibility shows no bulk ferromagnetic transition (Farhang et al., 2022). In FeTe$\rightarrow$54Se$\rightarrow$55, spontaneous Kerr angles up to $\rightarrow$56 nrad appear below $\rightarrow$57 K, just below $\rightarrow$58 K, and field training with $\rightarrow$59 T yields remanent $\rightarrow$60 nrad (Farhang et al., 2022). In FeTe$\rightarrow$61Se$\rightarrow$62, a non-superconducting Type A sample shows spontaneous Kerr signals up to $\rightarrow$63 nrad, while a superconducting Type B sample shows signals up to $\rightarrow$64 nrad. Because TRSB appears even in the non-superconducting sample, the paper attributes it to an intertwined surface ferromagnetic order rather than a TRSB superconducting order parameter.

Taken together, these results correct another common misconception: neither the existence of a non-trivial bulk invariant nor the observation of low-energy surface spectral weight guarantees a coherent, purely topological surface channel. In FTS, local Se/Te distribution, correlation-driven decoherence, anisotropic strain, and surface ferromagnetism all modulate the surface state that underlies Majorana-based interpretations (Zhao et al., 2021).