betaFeedback

WorksPhysics and Astronomy

Quasi-one-dimensional topological band structure and van Hove singularities in monolayer TaIrTe4_4 from laser μμ-ARPES

Abstract

Recent transport experiments reported a quantum spin Hall insulator phase in monolayer 1T-TaIrTe4_4 gated away from charge neutrality. This phase is not predicted by band structure calculations and has been attributed to an electronic instability induced by strong correlations at a putative van Hove singularity. Here, we investigate the electronic structure of exfoliated monolayer 1T-TaIrTe4_4 using micro-focus laser angle resolved photoemission. We find a strongly anisotropic band structure susceptible to density wave instabilities. Our data further reveal a saddle point singularity in the density of states. However, we find that the saddle point lies at a density for which transport experiments found no anomalies. Moreover, the quasiparticle line widths suggest weak electron correlations only. This points to a secondary role of van Hove singularities and electron correlations in the transport phase diagram of monolayer 1T-TaIrTe4_4.

Main

The interplay of correlations and topology can give rise to novel phases of matter. [1, 2, 3] While this prospect was recognized early on, the field long remained experimentally challenging. However, recent advances in two-dimensional moiré materials brought about remarkable progress. The archetypal manifestation of the synergy between band topology and strong correlations is the fractional quantum anomalous Hall (FQAH) effect in Chern insulators – 2D topological insulators with broken time-reversal symmetry. Evidence for the FQAH effect was first reported from optical measurements of twisted bilayer MoTe2_2. [4] Shortly after, the direct transport observation of fractional Hall plateaus unambiguously confirmed the FQAH effect in MoTe2_2 and multilayer graphene. [5, 6] Time reversal invariant quantum spin Hall insulators, [7] the first type of topological insulator predicted in 2006 and realized in 2007 in HgTe quantum wells, [8, 9] do not show a charge Hall effect but may also host fractional states manifest in fractional edge conduction. [8, 10, 11] Recent studies further provide evidence for topological Kondo insulators [1, 12] and anomalous Hall crystals, topologically non-trivial correlations. related insulators at fractional filling that spontaneously break translational symmetry and time-reversal symmetry. [13] Topological Mott insulators have been discussed theoretically, [14, 15] but proved elusive in experiment, in part because their signatures are unconventional and may be difficult to detect. [16]

1T-TaIrTe41\mathrm{T}\text{-}\mathrm{TaIrTe}_{4} provides a promising material platform in this context. In its bulk form, 1T-TaIrTe41\mathrm{T}\text{-}\mathrm{TaIrTe}_{4} was identified as a non-centrosymmetric type-II Weyl semimetal candidate. [17, 18] Upon dimensional reduction, it was predicted that monolayer and bilayer 1T-TaIrTe41\mathrm{T}\text{-}\mathrm{TaIrTe}_{4} realize quantum spin Hall insulating phases at charge neutrality. [19, 20, 21] Intriguingly though, recent experiments by Tang et al. revealed two distinct insulating states with quantized edge conduction in monolayer 1T-TaIrTe41\mathrm{T}\text{-}\mathrm{TaIrTe}_{4}. This adds a new aspect to the rich physics of interacting topological insulators. [22] The first insulating state, observed at charge neutrality, is consistent with density functional theory (DFT) predictions [19, 20, 21] and is likely a single particle topological band insulator. The second insulating state is more puzzling. It was observed in gated devices at a carrier density ∼6⋅1012 cm−2\sim6 \cdot10^{12}\ \mathrm{cm}^{-2}, where the chemical potential lies a few 10 meV in the conduction band, and is separated from the topological band insulator at charge neutrality by a narrow, weakly metallic region. These characteristics are incompatible with a single particle band structure picture and strongly suggest a many-body nature of the insulating gap.

Tang et al. interpret the finite-density insulating state in monolayer 1T-TaIrTe41\mathrm{T}\text{-}\mathrm{TaIrTe}_{4} as a topological charge density wave (CDW) insulator and propose that the instability is induced by enhanced electron correlations when a van Hove singularity (VHS) predicted by DFT approaches the Fermi level. [22] Raman, nonlinear Hall, and nonlinear optical measurements in few-layer TaIrTe4\mathrm{TaIrTe}_{4} have equally been interpreted as evidence for a low-temperature one-dimensional CDW phase. [23] However, several questions remain open. To date, neither a density-wave reconstruction nor van Hove singularities have been directly established in monolayer 1T-TaIrTe41\mathrm{T}\text{-}\mathrm{TaIrTe}_{4}. Also, the robustness of an interpretation based on DFT calculations remains largely untested. A recent synchrotron nano-focus angle resolved photoemission (ARPES) study found a band structure in qualitative agreement with DFT on large energy scales but could not resolve the relevant low-energy band structure with the predicted VHS. [24]

Moreover, Tang et al. find the many-body QSHI state at a much lower density than expected from the nesting vector predicted by DFT. [22, 25] Finally, a systematic study on a large number of devices revealed significant variations with some devices showing only a single QSHI phase and a third category of devices showing two topologically trivial insulating states. [25]

Here, we use laser μ\mu-ARPES to map the full low-energy quasiparticle dispersion in ML 1T-TaIrTe4_4. Together with DFT calculations, our data unambiguously confirm a topologically non-trivial band structure with gapped Dirac points at 4 equivalent momenta on the Brillouin zone boundary. We further demonstrate the presence of VHSs with significant density of states originating from shallow saddle points connecting the Dirac points with the global valence band maximum at Γ\Gamma.

Figure 1a illustrates the sample design used for this study. We exfoliate 1T-TaIrTe4_4 in a N2_2-filled glove box. MLs are identified from the optical contrast and encapsulated between a graphite bottom layer and ML graphene using polycarbonate (PC) stamps. The stacks are placed on Au coated Si/SiO2_2 wafers and rinsed in chloroform prior to transferring the samples under protected atmosphere to the vacuum system for μ\mu-ARPES, where they are located in ARPES spatial maps (Figure 1b). All measurements were performed at T=6T = 6 K, a photon energy of 6.01 eV and an energy and momentum resolution of ∼6\sim6 meV / 0.002 Å−1^{-1}. DFT calculations in the generalized gradient approximation with on-site Hubbard corrections (GGA+UU) were performed using the Quantum ESPRESSO package. [26, 27] Additional details of the experimental and theoretical methods are given in Supporting Information.

Figure 1. Anisotropy of crystal structure and electronic band structure. (a,b) Optical micrograph and ARPES spatial mapping of an encapsulated ML 1T-TaIrTe4_4 sample measured for this study. (c) Structure of ML 1T-TaIrTe4_4 with the primitive unit cell and lattice vectors indicated. The Brillouin zone is shown on the right. (d) ARPES dispersion plots along Γ\GammaX and Γ\GammaY. DFT calculations with and without SOC and DFT+UU calculations (for U=2.5U = 2.5 eV) are overlaid in dashed and solid lines, respectively. All calculations are shifted in energy to match the experimental Fermi wave vector along Γ\GammaX.

ML 1T-TaIrTe4_4 has an orthorhombic structure with alternating zigzag chains of Ta and Ir atoms in edge sharing, distorted Te octahedra (Figure 1c). The in-plane lattice constants measured in the bulk are a=3.770a = 3.770 Å (along the chains) and b=12.421b = 12.421 Å (perpendicular to the chains). [28, 29] Importantly, the structure has a glide mirror symmetry, which protects Dirac band crossings and allows the Dirac points to move away from time-reversal invariant momenta.[30] The product of the glide-mirror with a 2-fold screw rotation makes the structure inversion symmetric.[30, 31] The low-energy states of 1T-TaIrTe4_4 arise from hybridization of Ta 5d5d and Te 5p5p orbitals,[22] suggesting a strongly anisotropic electronic structure. Our μ\mu-ARPES data of ML 1T-TaIrTe4_4 shown in Figure 1d directly confirm this expectation. We observe a strong dispersion along the Ta chains (Γ\GammaX) and suppressed dispersion in the orthogonal Γ\GammaY direction. Notably, our DFT calculation shown in the same figure underestimates the quasiparticle dispersion along Γ\GammaX by nearly a factor of 2, opposite to the expectations for a correlated material with enhanced quasiparticle mass (see Figure 1d). Investigating the origin of this discrepancy, we notice that 1T-TaIrTe4_4 has a light and heavy valence band, found in our ARPES data at EFE_F and E−EF≈0.3E - E_F \approx0.3 eV, respectively. Without SOC, these

bands cross at 2 degeneracy points along ΓX\Gamma X at E−EF∼−0.13E - E_F \sim-0.13 eV (Figure 1d). Adding SOC lifts the degeneracies by hybridizing the two bands. This renders their dispersion highly sensitive to their relative energy position. Here, we use DFT+UU as a pragmatic way to tune the latter. Empirically, we find that adding U≈2.5U \approx2.5 eV on Ta and Ir sites correctly reproduces the experimental valence band positions at Γ\Gamma and at the same time improves the overall description of the experimental quasiparticle dispersion. For the remainder of this paper, we will refer to this DFT+UU calculation. The systematic evolution of the band structure with UU is shown in Supporting Information, Figure S1.

Figure S1. DFT+UU calculations showing a systematic evolution of the electronic band structure with on-site Coulomb correction UU. Fermi levels are chosen to reproduce the experimental Fermi wave vectors along ΓX. The ARPES data reproduced from Figure 1,Figure 2 is shown for comparison.

In Figure 2, we focus on the low-energy quasiparticle band structure. Consistent with Figure 1, we find that our samples are degenerately doped with the chemical potential lying ∼7\sim7 meV below the valence band maximum. The Fermi surface (top panel in Figure 2a, also shown to scale in the full Brillouin zone) consists of a very small elliptical hole pocket at Γ\Gamma with a carrier density of 7⋅1011 cm−27 \cdot10^{11}\ \mathrm{cm}^{-2} (0.00160.0016 holes per Ta). At present, it is not clear if the hole doping arises from defects in our crystals or is induced by charge transfer to the graphite and graphene encapsulation layers. With increasing energy, the ARPES intensity first evolves into an unusual hashtag like structure before developing the characteristic warped lines of a quasi-1D band structure. The series of dispersion plots in Figure 2b reveals that the peculiar hashtag contours arise from the evolution of the Γ\Gamma valley into an extended kk-space area with nearly flat states at an energy ∼−60\sim-60 meV.

Figure 2. Gapped Dirac cones and low-energy quasiparticle dispersion. (a) Stack of constant energy contours. For scale, the spectral weight at EFE_F is also shown in the full Brillouin zone. (b) Dispersion plots along kxk_x for equidistant kyk_y values ranging from Γ\Gamma to Y. (c) Dispersion plot along the edge of the BZ. (d) DFT+UU band dispersion showing the gapping of the Dirac cones with the inclusion of SOC.

Figure 2c shows an expanded view of the band map along the Brillouin zone boundary. The abrupt change in dispersion near the Λ\Lambda points is a fingerprint of an avoided crossing between a light hole band and a heavier electron-like band. Our DFT+UU calculations (Figure 2d) readily identify these bands as Ta dd states with two symmetry-protected Dirac crossings along SYS. The inclusion of SOC lifts the degeneracy, resulting in topological band gaps at Λ1,2\Lambda_{1,2}, consistent with earlier DFT studies. [19, 20, 21] Adding correlations in DFT+UU affects band gaps and reshapes the band structure. The gap at Λ\Lambda increases from 69 meV at U=0U = 0 to 108 meV at U=2.5U = 2.5 eV and moves to Y for U≥4U \ge4 eV (see Supporting Information, Figure S1). Importantly though, we find that the quantum spin Hall state of 1T-TaIrTe4_4 is robust in DFT+UU up to at least U=5U = 5 eV. The valence band maximum lies at Γ\Gamma in all our calculations. At U=2.5U = 2.5 eV, the valence band valley splitting EΓ−EΛE_{\Gamma}-E_{\Lambda} is 59 meV, close to the experimental value of ≈57\approx57 meV obtained from EΛ=−50E_{\Lambda} = -50 meV (Figure 2c) and EΓ≈7E_{\Gamma} \approx7 meV estimated in Figure 1. The local conduction band minima at Γ\Gamma and Λ\Lambda are not detected in our experiments. The calculations show them nearly degenerate with the direct and indirect gap both increasing slowly with UU from 33 meV at U=0U = 0 to 49 meV at U=2.5U = 2.5 eV, slightly larger than the gaps of 5–30 meV deduced from transport experiments. [22, 25]

The band structure of 1T-TaIrTe4_4 has two key-properties that favor many-body instabilities. It is quasi-1D and in addition shows extended flat regions. In Figure 3 we analyze these properties quantitatively. To this end, we first determine in Figure 3a the full 2D low-energy quasiparticle band structure E(kx,ky)E(k_x,k_y) from an analysis of energy- and momentum distri bution curves (EDCs, MDCs). Figure 3b shows the quasiparticle density of states (DOS) obtained from a histogram of the E(kx,ky)E(k_x,k_y) values. The pronounced peak at −52-52 meV can be attributed to a VHS at a filling of 6.2⋅1012 cm−26.2 \cdot10^{12}\ \mathrm{cm}^{-2}. We directly confirm the presence of VHSs in Figure 3c. Dispersion plots extracted along the intersecting kk-space paths marked green and violet in Figure 3a show a shallow saddle point with opposite curvature along k1,2k_{1,2}, implying a VHS with diverging DOS. At the same time, it is evident from Figure 3c that a minute deformation of the bands could transform the saddle into a ridge and move the VHS out to the Λ1,2\Lambda_{1,2} points, or could even merge two singularities at the Y point, as we observed it in DFT+UU for U≥4U \ge4 eV (see Supporting Information, Figure S1). Supporting Figure S2 further compares the experimental quasiparticle DOS with DFT and DFT+UU.

Figure 3. Van Hove singularity. (a) ARPES quasiparticle dispersion obtained from the analysis of EDCs and MDCs. (b) Experimental quasiparticle DOS. (c) Dispersion along the two kk-space paths intersecting at the VHS indicated in panel a. (d) Static Lindhard susceptibility calculated from the experimental quasiparticle dispersion.

Figure S2. DFT+UU DOS for U=0U = 0 and U=2.5U = 2.5 eV. Fermi levels are chosen to match the slight hole doping observed in experiment. The experimental quasiparticle DOS from Figure 3 is shown for comparison.

To connect our findings to transport data on gated devices, we calculate the static susceptibility χ0(q)\chi^{0}(\mathbf{q}) from the experimental quasiparticle dispersion, assuming a rigid band shift of 52 meV to move the VHS to the Fermi level. The most prominent feature at Q∗=0.031 (2π/a)\mathbf{Q}^{*} = 0.031\ (2\pi/a) corresponds to nesting between two VHSs and closely resembles the structure in the bare DFT susceptibility at Q∗=0.052 (2π/a)\mathbf{Q}^{*} = 0.052\ (2\pi/a) identified by Tang et al. as origin of the putative CDW instability. However, several arguments call the relevance of VHSs and nesting at Q∗\mathbf{Q}^{*} into question. We first note that the periodicity of CDWs is often determined by a strongly momentum dependent electron-phonon interaction and rarely coincides with nesting vectors. [32] Further, a CDW insulator with the experimental periodicity Q∗\mathbf{Q}^{*} requires at a density n=Q∗a/π=0.062/u.c.n = Q^{*}a/\pi= 0.062/\mathrm{u.c.} (1.3⋅1013 cm−2)(1.3 \cdot10^{13}\ \mathrm{cm}^{-2}). Although lower than the DFT value, this remains beyond the range investigated by Tang et al. [22] On the other hand, our ARPES data show that the van Hove filling occurs at a lower density of ≈6.2⋅1012 cm−2\approx6.2 \cdot10^{12}\ \mathrm{cm}^{-2} where transport data is available. [22, 25] If the system is tuned to this density, one expects anomalous transport from scattering with ’hot’ van Hove carriers leading to an increased resistivity and deviations from the canonical T2T^{2} behavior, [33] as it was observed in moiré systems or in Sr2_2RuO4_4. [34, 35] However, thus far, no signs of such anomalies have been reported in 1T-TaIrTe4_4 gated to van Hove filling. This all points to a secondary role of VHSs in 1T-TaIrTe4_4. Instead, the properties of ML 1T-TaIrTe4_4 may be dominated by its unusual electronic structure 1T-TaIrTe4_4 with extended regions with nearly flat bands. Placing the chemical potential near or inside a flat band naturally leads to a large susceptibility at small Q\mathbf{Q}. Indeed, calculations using the experimental quasiparticle dispersion show that such small-Q\mathbf{Q} features may exceed the susceptibility peak from nesting of the VHS (Supporting Information, Figure S3). In this situation, the precise wave vector of an instability is likely determined by the momentum dependence of electron-phonon coupling. At present, it remains open to what degree the above arguments apply to the conduction band where Tang et al. found the many-body QSHI state. [22] Consistent with Ref., [22] our DFT calculations for U=0U = 0 show a nearly particle-hole symmetric band structure with a single singularity in both the conduction and valence band DOS. However, for U=2.5U = 2.5 eV, we find a splitting of more than 20 meV in the conduction band singularity, while the valence band singularity stays intact (Supporting Information, Figure S2). It remains to be understood if this qualitative change underlies the different behavior of 1T-TaIrTe4_4 under electron and hole doping observed in experiment. [22]

The combination of flat bands and substantial Coulomb repulsion in the Ta and Ir d-shell suggest a potentially important role of electron correlations in 1T-TaIrTe4_4. The strength of correlations is often assessed from the renormalization of the quasiparticle Fermi velocity over the bare dispersion. However, this requires an accurate bare Hamiltonian – typically obtained from DFT in the local density approximation. [36] We presently cannot give such a Hamiltonian for ML 1T-TTaIrTe4\mathrm{TaIrTe}_{4}. The primary effect of the Hubbard correction UU in our DFT+UU calculations is to tune the relative energy position of bands, and with it their hybridization. However, our calculations still underestimate the quasiparticle dispersion for U=2.5U = 2.5 eV which reproduces the relative energy position of the valence bands at Γ\Gamma (Figure 4a).

Figure 4. Many-body interactions. (a) Comparison of the quasiparticle dispersion along ΓX\Gamma X from three devices (S1, S2, S3) with the DFT+UU bands. (b) Quasiparticle linewidth Γ=∂E/∂k Δk\Gamma= \partial E/\partial k \,\Delta k calculated assuming a constant dispersion ∂E/∂k=2.08\partial E/\partial k = 2.08 eVÅ. Here, Δk\Delta k is the full width at half maximum of Lorentzian MDC fits. The inset shows fits of MDCs at representative energies marked by vertical arrows of the same color. Red lines are the background of the fit.

The evolution of the quasiparticle linewidth provides an alternative, albeit less quantitative, measure of interactions. In strongly correlated materials, the linewidth typically exceeds the excitation energy beyond a few 10 meV from the Fermi level where defined quasiparticles gradually evolve into incoherent excitations.[37, 38] Our dispersion maps show no sign of such a crossover. For a quantitative analysis, we use Lorentzian fits of MDCs along ΓX\Gamma X and calculate quasiparticle linewidths Γ\Gamma assuming a constant group velocity. This reveals an exceptionally slow variation with energy. As shown in Figure 4b, Γ\Gamma increases barely more than 30 meV over the full energy range of ∼200\sim200 meV. This differs markedly from typical correlated electron systems[39, 40] and is more comparable to nearly free electron metals, pointing to weak correlations.[41] We note, however, that correlation effects can be highly sensitive to the band filling. Intriguingly, our analysis in Figure 4b shows a small but statistically significant linewidth anomaly at the energy of the VHS, pointing to possibly relevant electron correlations around this filling.

In summary, we reported a comprehensive mapping of the low-energy quasiparticle dispersion in ML 1T-TTaIrTe4\mathrm{TaIrTe}_{4}. Our data provide strong evidence for a topological band inversion and directly resolve low-lying van Hove singularities. This provides a solid foundation for the interpretation of complementary transport experiments. Our findings broadly support the notion of a topological CDW insulator in gated ML 1T-TTaIrTe4\mathrm{TaIrTe}_{4} put forward by Tang et al. 22 However, they also reveal a mismatch between the density of the many-body QSHI state in ML 1T-TTaIrTe4\mathrm{TaIrTe}_{4} and the wave vector of the most prominent feature in the single particle susceptibility. This highlights the need for structural studies and for future ARPES experiments on gate-tuned devices to probe the finite-density insulating regime directly and search for spectral signatures of density wave order and enhanced many-body effects.

Acknowledgments

We thank I. Gutiérrez-Lezama for help with the development of the device fabrication. The experimental work was supported by the Swiss National Science Foundation (SNSF) through grants 200020-215548 and 200021-204065. A.C. and M.G. acknowledge financial support from the EMPEROR project, CUP E93C24001040001, funded by the European Union—NextGeneration EU (M4C2INV1.3) through the National Quantum Science and Technology Institute (Spoke 5).

Supporting Information

Crystal growth

Single crystals of 1T1T-TaIrTe4\mathrm{TaIrTe}_{4} were grown using the self-flux technique from the corresponding elements. Powders of Ta (99.95 %, Thermo Scientific), Ir (99.95 %, Alfa Aesar) and Te (ground from lumps, 99.999 %, Thermo Scientific) were mixed together in molar ratios of 1:1:30 inside an alumina crucible. This was placed inside an open quartz ampoule, and a ceramic sieve followed by 2–3 cm of compressed quartz wool were placed on top. Then, the ampoule was sealed off under an inert atmosphere of 300 mbar Ar and placed vertically in a muffle furnace. The temperature program consisted of heating to 1000∘C1000^\circ\mathrm{C} with a ramp of +30∘C+30^\circ\mathrm{C}/h, holding the maximum temperature for 48 h before cooling down with −2∘C-2^\circ\mathrm{C}/h to 600∘C600^\circ\mathrm{C}. After a few hours at this temperature, the ampoule was taken out of the furnace, flipped around, and excess Te-flux was removed through hot-centrifugation at ca. 2000 rpm. Needle-shaped crystals of the title compounds with a length of several mm could be isolated.

The elemental composition of the grown 1T1T-TaIrTe4\mathrm{TaIrTe}_{4} crystals was confirmed using energy dispersive X-ray spectroscopy (EDS) on a JEOL JSM-7600F scanning electron microscope, equipped with an X-MaxN^{\mathrm{N}} 80 detector (Oxford Instruments), using an accelerating voltage of 20 kV. In total, 90 point-analyses were performed on 14 crystals originating from six batches. On average, a composition of Ta1.00(3)Ir1.00(3)Te3.95(9)\mathrm{Ta}_{1.00(3)}\mathrm{Ir}_{1.00(3)}\mathrm{Te}_{3.95(9)} was obtained, which lies within one standard deviation from the nominal composition. The crystal structure of 1T1T-TaIrTe4\mathrm{TaIrTe}_{4} was confirmed with both single-crystal (on a Oxford Diffraction SuperNova diffractometer from Agilent/Rigaku) and powder (Rigaku SmartLab) X-ray diffraction.

Sample preparation

Mechanical exfoliation using scotch tape and polydimethylsiloxane (PDMS) was employed to obtain monolayers of 1T1T-TaIrTe4\mathrm{TaIrTe}_{4} from the bulk crystals. Single crystals of TaIrTe4\mathrm{TaIrTe}_{4} were first placed on scotch tape and exfoliated 3 to 4 times to obtain thin flakes, which were then transferred onto the PDMS. The PDMS with flakes was subsequently brought into contact with a 290 nm SiO2/Si\mathrm{SiO}_2/\mathrm{Si} substrate and heated to 100∘C100^{\circ}\mathrm{C} for 1 minute. After heating, the PDMS was slowly peeled off from the SiO2\mathrm{SiO}_2 substrate at room temperature, transferring the flakes on the substrate.

The thickness of TaIrTe4\mathrm{TaIrTe}_4 flakes on the SiO2/Si\mathrm{SiO}_2/\mathrm{Si} substrate was initially estimated using optical contrast analysis. Monolayer TaIrTe4\mathrm{TaIrTe}_4 exhibited normalized contrast values of approximately 7.5%, 11%, and 2.5% in the red, green, and blue channels, respectively. These thickness assignments were later confirmed by ARPES measurements.

For ARPES measurements, the monolayer TaIrTe4\mathrm{TaIrTe}_4 flake was encapsulated in between monolayer graphene and bulk graphite. The top graphene layer serves as a capping layer, while the bottom graphite flake provides a flat and conductive substrate. The heterostructure was assembled using a PDMS/polycarbonate (PC) stamp by dry transfer technique. In this process, the topmost flake (graphene) was first picked up, followed by the monolayer TaIrTe4\mathrm{TaIrTe}_4 and the graphite flake. During pickup, the substrate temperature was initially set to 90∘C90^{\circ}\mathrm{C} upon contact of the stamp and ramped to 120∘C120^{\circ}\mathrm{C} before cooling down. After the pick up, the heterostructure along with the PC film, was released on a gold -coated SiO2/Si\mathrm{SiO}_2/\mathrm{Si} substrate at 180∘C180^{\circ}\mathrm{C}. Finally, the sample was immersed in chloroform overnight to dissolve the PC layer and to remove residual polymer contamination.

The fabricated sample was transferred to the load-lock chamber of the ARPES system under a protected environment to prevent contamination and oxidation from ambient exposure. All processing steps, including exfoliation, heterostructure assembly, and sample mounting, were carried out inside a nitrogen-filled glovebox with impurity levels maintained below 0.1 ppm for both H2O\mathrm{H}_2\mathrm{O} and O2\mathrm{O}_2.

DFT calculations

Density functional theory (DFT) calculations were carried out using the Quantum ESPRESSO package, [26, 27] employing the PBE[42] exchange−correlation functional together with pseudopo tentials obtained from the PseudoDojo [43, 44] library. The kinetic energy cutoffs for the wavefunctions and charge density were chosen as 90 Ry and 360 Ry, respectively, for the DFT simulations of the 1T-TaIrTe4_4 two-dimensional monolayer. A Γ-centered Monkhorst−Pack (12×4×1)(12 \times4 \times1) k-point grid was used to sample the Brillouin zone (BZ). A cold smearing of 0.02 Ry was employed, with the input files generated using the Materials Cloud Quantum ESPRESSO input generator. To eliminate spurious electrostatic interactions between periodically repeated images along the out-of-plane direction, a Coulomb cutoff[45, 46] scheme was employed, ensuring appropriate boundary conditions. The electronic band structure was calculated using the experimental crystal structure, with spin–orbit coupling incorporated through fully relativistic pseudopotentials. To achieve better agreement between the experimental and calculated electronic band structures, the GGA+UU approach was employed by applying on-site Hubbard corrections to the Ir-dd and Ta-dd orbitals. Several combinations of UU values were examined, including (0, 0), (0, 1), (1, 0), (0, 2), (2, 0), (1, 2), (2, 1), (2, 2), and (2.5, 2.5) eV for the Ta and Ir atoms, respectively. These show that adding UU on the Ta atoms dominates the effects on the low-energy band structure. For simplicity, we thus

focus on equal UU values on Ta and Ir. Figure S1 shows the systematic evolution of the band structure with UU. The best agreement with experiment is found for U=2.5U = 2.5 eV on both Ir-dd and Ta-dd orbitals.

In Figure S2 we further compare the experimental quasiparticle DOS with DFT+UU for U=0U = 0 and U=2.5U = 2.5 eV.

Susceptibility calculation

The charge susceptibility for monolayer 1T-TaIrTe4_4 shown in Figure 3d of the main text was computed from the static Lindhard function

χ0(Q,ω=0,T)=1Ω∑kf(Ek)−f(Ek+Q)ω+Ek−Ek+Q+iδ(1)\chi_0(\mathbf{Q},\omega= 0,T) = \frac{1}{\Omega}\sum_{\mathbf{k}}\frac{f(E_{\mathbf{k}})-f(E_{\mathbf{k}+\mathbf{Q}})}{\omega+ E_{\mathbf{k}} - E_{\mathbf{k}+\mathbf{Q}} + i\delta} \tag*{(1)}

using the experimental band dispersion EkE_{\mathbf{k}} shown in Figure 3a. Here, Ω\Omega is the system volume, f(Ek)f(E_{\mathbf{k}}) is the Fermi function and δ\delta is a broadening parameter set to 5 meV. The temperature was fixed to T=4.2T = 4.2 K. The calculation in Figure 3a is performed for a rigid band shift by μ=−52\mu= -52 meV to place the VHSs at the chemical potential. Figure S3 shows the systematic evolution of ξ0\xi_0 with band filling. The susceptibility is strongly enhanced near van Hove filling where it develops a well defined peak from nesting of the VHSs around kx=0.05k_x=0.05 Å−1^{-1}. In addition, the calculations show a broad local maxima at lower momentum if the chemical potential is placed within the nearly flat region of the valence band around the van Hove filling.

Figure S3. Line profiles of the charge susceptibility χ0(Qx,Qy=0)\chi_0(Q_x,Q_y=0) for different band fillings.

References

  1. [1]Dzero, M.; Sun, K.; Galitski, V.; Coleman, P. Topological Kondo Insulators. Phys. Rev. Lett. 2010, 104, 106408.
  2. [2]Neupert, T.; Santos, L.; Chamon, C.; Mudry, C. Fractional Quantum Hall States at Zero Magnetic Field. Phys. Rev. Lett. 2011, 106, 236804.pubmed.ncbi.nlm.nih.gov/21770534
  3. [3]Witczak-Krempa, W.; Chen, G.; Kim, Y. B.; Balents, L. Correlated Quantum Phenomena in the Strong Spin-Orbit Regime. Annual Review of Condensed Matter Physics 2014, 5, 57–82.
  4. [4]Cai, J. et al. Signatures of fractional quantum anomalous Hall states in twisted MoTe₂. Nature 2023, 622, 63–68.DOI
  5. [5]Lu, Z.; Han, T.; Yao, Y.; Reddy, A. P.; Yang, J.; Seo, J.; Watanabe, K.; Taniguchi, T.; Fu, L.; Ju, L. Fractional quantum anomalous Hall effect in multilayer graphene. Nature 2024, 626, 759–764.DOI
  6. [6]Park, H. et al. Observation of fractionally quantized anomalous Hall effect. Nature 2023, 622, 74–79.DOI
  7. [7]Kane, C. L.; Mele, E. J. Z₂ Topological Order and the Quantum Spin Hall Effect. Phys. Rev. Lett. 2005, 95, 146802.
  8. [8]Bernevig, B. A.; Zhang, S.-C. Quantum Spin Hall Effect. Phys. Rev. Lett. 2006, 96, 106802.pubmed.ncbi.nlm.nih.gov/16605772
  9. [9]König, M.; Wiedmann, S.; Brüne, C.; Roth, A.; Buhmann, H.; Molenkamp, L. W.; Qi, X.-L.; Zhang, S.-C. Quantum Spin Hall Insulator State in HgTe Quantum Wells. Science 2007, 318, 766–770.pubmed.ncbi.nlm.nih.gov/17885096
  10. [10]Levin, M.; Stern, A. Fractional Topological Insulators. Phys. Rev. Lett. 2009, 103, 196803.
  11. [11]Kang, K.; Shen, B.; Qiu, Y.; Zeng, Y.; Xia, Z.; Watanabe, K.; Taniguchi, T.; Shan, J.; Mak, K. F. Evidence of the fractional quantum spin Hall effect in moiré MoTe₂. Nature 2024, 628, 522–526.DOI
  12. [12]Han, Z.; Xia, Y.; Xia, Z.; Zhao, W.; Zhang, Y.; Watanabe, K.; Taniguchi, T.; Shan, J.; Mak, K. F. Topological Kondo insulator in MoTe₂/WSe₂ moiré bilayers. Nature Physics 2026, 22, 396–401.DOI
  13. [13]Dong, J.; Wang, T.; Wang, T.; Soejima, T.; Zaletel, M. P.; Vishwanath, A.; Parker, D. E. Anomalous Hall Crystals in Rhombohedral Multilayer Graphene. I. Interaction-Driven Chern Bands and Fractional Quantum Hall States at Zero Magnetic Field. Phys. Rev. Lett. 2024, 133, 206503.DOI
  14. [14]Raghu, S.; Qi, X.-L.; Honerkamp, C.; Zhang, S.-C. Topological Mott Insulators. Phys. Rev. Lett. 2008, 100, 156401.pubmed.ncbi.nlm.nih.gov/18518132
  15. [15]Pesin, D.; Balents, L. Mott physics and band topology in materials with strong spin-orbit interaction. Nature Physics 2010, 6, 376–381.DOI
  16. [16]Wagner, N.; Crippa, L.; Amaricci, A.; Hansmann, P.; Klett, M.; König, E. J.; Schäfer, T.; Sante, D. D.; Cano, J.; Millis, A. J.; Georges, A.; Sangiovanni, G. Mott insulators with boundary zeros. Nature Communications 2023, 14, 7531.DOI
  17. [17]Koepernik, K.; Kasinathan, D.; Efremov, D. V.; Khim, S.; Borisenko, S.; Büchner, B.; van den Brink, J. TaIrTe₄: A ternary type-II Weyl semimetal. Phys. Rev. B 2016, 93, 201101.
  18. [18]Haubold, E. et al. Experimental realization of type-II Weyl state in noncentrosymmetric TaIrTe₄. Phys. Rev. B 2017, 95, 241108.
  19. [19]Liu, J.; Wang, H.; Fang, C.; Fu, L.; Qian, X. van der Waals Stacking-Induced Topological Phase Transition in Layered Ternary Transition Metal Chalcogenides. Nano Letters 2017, 17, 467–475.DOI
  20. [20]Marrazzo, A.; Gibertini, M.; Campi, D.; Mounet, N.; Marzari, N. Relative Abundance of Z2 Topological Order in Exfoliable Two-Dimensional Insulators. Nano Letters 2019, 19, 8431–8440.DOI
  21. [21]Guo, P.-J.; Lu, X.-Q.; Ji, W.; Liu, K.; Lu, Z.-Y. Quantum spin Hall effect in monolayer and bilayer TaIrTe₄. Phys. Rev. B 2020, 102, 041109.DOI
  22. [22]Tang, J. et al. Dual quantum spin Hall insulator by density-tuned correlations in TaIrTe₄. Nature 2024, 628, 515–521.DOI
  23. [23]Jiang, H.; Xi, T.; Li, J.; He, Y.; Ma, H.; Mao, Y.; Taniguchi, T.; Watanabe, K.; Rhodes, D. A.; Zhang, Y.; Xiao, J.; Wang, Y. Probing interplay of topological properties and electron correlation in TaIrTe₄ via nonlinear Hall effect. Nature Communications 2025, 16, 6351.
  24. [24]Ekahana, S. A. et al. Visualization of Tunable Electronic Structure of Monolayer TaIrTe₄. 2026; https://arxiv.org/abs/2601.11504.
  25. [25]Li, J.; Tang, J.; Primeau, L.; Ding, T. S.; Soni, R.; Qian, T.; Watanabe, K.; Taniguchi, T.; Ni, N.; Maestro, A. D.; Ma, Q.; Zhang, Y. Interaction-Driven Topological Transitions in Monolayer TaIrTe₄. 2025; https://arxiv.org/abs/2506.18412.
  26. [26]Giannozzi, P. et al. QUANTUM ESPRESSO: a modular and open-source software project for quantum simulations of materials. Journal of Physics: Condensed Matter 2009, 21, 395502.DOI
  27. [27]Giannozzi, P. et al. Advanced capabilities for materials modelling with Quantum ESPRESSO. Journal of Physics: Condensed Matter 2017, 29, 465901.
  28. [28]Mar, A.; Jobic, S.; Ibers, J. A. Metal-metal vs tellurium-tellurium bonding in WTe₂ and its ternary variants TaIrTe₄ and NbIrTe₄. Journal of the American Chemical Society 1992, 114, 8963–8971.DOI
  29. [29]Xing, Y.; Shao, Z.; Ge, J.; Luo, J.; Wang, J.; Zhu, Z.; Liu, J.; Wang, Y.; Zhao, Z.; Yan, J.; Mandrus, D.; Yan, B.; Liu, X.-J.; Pan, M.; Wang, J. Surface superconductivity in the type II Weyl semimetal TaIrTe₄. Natl Sci Rev 2020, 7, 579–587.DOI
  30. [30]Muechler, L.; Alexandradinata, A.; Neupert, T.; Car, R. Topological Nonsymmorphic Metals from Band Inversion. Phys. Rev. X 2016, 6, 041069.DOI
  31. [31]Zhang, X.; Mao, N.; Janson, O.; van den Brink, J.; Ray, R. Layer dependent topological phases and transitions in TaRhTe₄: From monolayer and bilayer to bulk. Phys. Rev. Mater. 2024, 8, 094201.DOI
  32. [32]Johannes, M. D.; Mazin, I. I. Fermi surface nesting and the origin of charge density waves in metals. Phys. Rev. B 2008, 77, 165135.DOI
  33. [33]Mousatov, C. H.; Berg, E.; Hartnoll, S. A. Theory of the strange metal Sr₃Ru₂O₇. Proceedings of the National Academy of Sciences 2020, 117, 2852–2857.
  34. [34]Wei, L.; Xu, Q.; He, Y.; Li, Q.; Huang, Y.; Zhu, W.; Watanabe, K.; Taniguchi, T.; Claassen, M.; Rhodes, D. A.; Kennes, D. M.; Xian, L.; Rubio, A.; Wang, L. Linear resistivity at van Hove singularities in twisted bilayer WSe₂. Proceedings of the National Academy of Sciences 2024, 121, e2321665121.DOI
  35. [35]Barber, M. E.; Gibbs, A. S.; Maeno, Y.; Mackenzie, A. P.; Hicks, C. W. Resistivity in the Vicinity of a van Hove Singularity: Sr₂RuO₄ under Uniaxial Pressure. Phys. Rev. Lett. 2018, 120, 076602.
  36. [36]Tamai, A. et al. High-Resolution Photoemission on Sr₂RuO₄ Reveals Correlation-Enhanced Effective Spin-Orbit Coupling and Dominantly Local Self-Energies. Phys. Rev. X 2019, 9, 021048.DOI
  37. [37]Tamai, A.; Allan, M. P.; Mercure, J. F.; Meevasana, W.; Dunkel, R.; Lu, D. H.; Perry, R. S.; Mackenzie, A. P.; Singh, D. J.; Shen, Z.-X.; Baumberger, F. Fermi Surface and van Hove Singularities in the Itinerant Metamagnet Sr₃Ru₂O₇. Phys. Rev. Lett. 2008, 101, 026407.DOI
  38. [38]Mravlje, J.; Aichhorn, M.; Miyake, T.; Haule, K.; Kotliar, G.; Georges, A. Coherence-Incoherence Crossover and the Mass-Renormalization Puzzles in Sr₂RuO₄. Phys. Rev. Lett. 2011, 106, 096401.DOI
  39. [39]Koralek, J. D.; Douglas, J. F.; Plumb, N. C.; Sun, Z.; Fedorov, A. V.; Murnane, M. M.; Kapteyn, H. C.; Cundiff, S. T.; Aiura, Y.; Oka, K.; Eisaki, H.; Dessau, D. S. Laser Based Angle-Resolved Photoemission, the Sudden Approximation, and Quasiparticle-Like Spectral Peaks in Bi₂Sr₂CaCu₂O₈₊δ. Phys. Rev. Lett. 2006, 96, 017005.DOI
  40. [40]Smit, S. et al. Momentum-dependent scaling exponents of nodal self-energies measured in strange metal cuprates and modelled using semi-holography. Nature Communications 2024, 15, 4581.
  41. [41]Potorochin, D. V.; Kurleto, R.; Clark, O. J.; Rienks, E. D. L.; Sánchez-Barriga, J.; Roth, F.; Voroshнин, V.; Fedorov, A.; Eberhardt, W.; Büchner, B.; Fink, J. Lifetime of quasiparticles in the nearly free electron metal sodium. Phys. Rev. B 2022, 106, 125138.
  42. [42]Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Physical Review Letters 1996, 77, 3865–3868.pubmed.ncbi.nlm.nih.gov/10062328
  43. [43]Hamann, D. R. Optimized Norm-Conserving Vanderbilt Pseudopotentials. Physical Review B 2013, 88, 085117.DOI
  44. [44]van Setten, M. J.; Giantomassi, M.; Bousquet, E.; Verstraete, M. J.; Hamann, D. R.; Gonze, X.; Rignanese, G.-M. The PseudoDojo: Training and Grading a 85 Element Optimized Norm-Conserving Pseudopotential Table. Computer Physics Communications 2018, 226, 39–54.DOI
  45. [45]Rozzi, C. A.; Varsano, D.; Marini, A.; Gross, E. K. U.; Rubio, A. Exact Coulomb Cutoff Technique for Supercell Calculations. Physical Review B 2006, 73, 205119.DOI
  46. [46]Sohier, T.; Calandra, M.; Mauri, F. Density Functional Perturbation Theory for Gated Two-Dimensional Heterostructures: Theoretical Developments and Application to Flexural Phonons in Graphene. Physical Review B 2017, 96, 075448.DOI

Paper details

vlnc:W_pjqgvje4
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents