Unveiling the orbital texture of 1T-TiTe₂ using intrinsic linear dichroism in multidimensional photoemission spectroscopy

Angle-resolved photoemission spectroscopy (ARPES) provides momentum-resolved electronic band structures but measured intensities are modulated by transition dipole matrix elements, complicating direct access to the spectral function. The layered semimetal 1T-TiTe₂ has long served as a benchmark for many-body effects and quasiparticle lifetimes in ARPES, yet the role of matrix elements in this material has been less explored. Matrix elements contain both extrinsic (geometry, polarization, orientation) and intrinsic (initial-state orbital symmetry) information; disentangling them yields access to orbital texture, chirality, and quantum geometric properties. Orbital textures in TMDCs are sensitive to structural phases like charge density waves and excitonic order. This study investigates how sample rotation-induced modulation of ARPES intensity can isolate an intrinsic linear dichroism in photoelectron angular distributions (ILDAD) in inversion-symmetric 1T-TiTe₂, enabling direct probing of momentum-dependent orbital textures.

Prior ARPES work established 1T-TiTe₂ as a model semimetal for observing Fermi-liquid behavior, many-body spectral features, and photohole lifetimes. However, matrix element effects were shown to significantly modulate ARPES intensities and sometimes mimic interaction signatures. Recent approaches propose either computational removal of matrix elements or, conversely, leveraging them to extract wavefunction information including orbital character, pseudospin textures, and Berry curvature. Orbital texture interplays with charge density waves and excitonic phases in TMDCs, with single-layer TiTe₂ exhibiting thickness- and substrate-dependent CDW and pseudogap behavior. A related differential ARPES methodology (TRDAD) in non-centrosymmetric 2H-TMDCs linked crystal rotations to time-reversal-like operations, isolating intrinsic dichroism tied to orbital pseudospin. Building on this, the present work generalizes to ILDAD for centrosymmetric 1T-TiTe₂ via mirror-related crystal orientations.

Experiment: A tabletop femtosecond XUV source (center ~21.7 eV, ~110 meV FWHM) based on 500 kHz OPCPA-driven high-harmonic generation supplies p-polarized pulses incident in the x–z plane (~65°). Photoelectrons are recorded with a time-of-flight momentum microscope (METIS 1000), enabling simultaneous 3D acquisition I(E, kx, ky) without rotating the sample. Bulk 1T-TiTe₂ crystals are cleaved in UHV (~5×10⁻¹¹ mbar). Due to the XUV inelastic mean free path, the signal is dominated by the topmost trilayer. Data processing uses an open-source workflow to generate calibrated 3D intensity hypervolumes with artifact and symmetry-distortion corrections. Dichroism extraction: Two measurements are performed for crystal orientations related by a 60° azimuthal rotation (R60), which corresponds to a layer mirror operation (z → −z) in 1T polytype. From the multidimensional data, left-right linear dichroism asymmetry ALDAD(E, kx, ky) = [I(E, kx, ky) − I(E, −kx, ky)] / [I(E, kx, ky) + I(E, −kx, ky)] is computed for both orientations, and the intrinsic component iLDAD is obtained as the antisymmetric part under R60. Thresholding ensures only momentum regions with sufficient symmetrized intensity are considered. Theory: (i) One-step photoemission calculations within the fully relativistic KKR method (SPR-KKR) using LDA+SOC, TRLEED final states, proper experimental geometry, and lifetime broadening simulate ARPES intensities and iLDAD, including photon-energy dependence. (ii) DFT (PBE) bulk and monolayer band structures are computed with QUANTUM ESPRESSO; projective Wannierization (WANNIER90) yields an 11-band tight-binding (TB) model (Ti d and Te p). A TB + plane-wave (PW) final-state model evaluates matrix elements in dipole gauge, enabling orbital-resolved analysis by selectively including Ti d orbitals (dz2, dxz, dyz, dxy, dx2−y2). Layer-resolved KKR confirms top-layer dominance near EF and M/M′. Photon-energy scans of iLDAD are computed theoretically to assess final-state effects.

• Strong momentum-dependent linear dichroism is observed in 1T-TiTe₂ ARPES at the Fermi level, with iLDAD up to approximately ±97% around Ti-3d electron pockets at M/M′ and weaker (up to ±35%) but symmetry-consistent signals around Te-5p hole pockets at Γ. Adjacent M and M′ pockets exhibit opposite iLDAD signs. • The dichroic patterns rotate with the sample by 60°, and the antisymmetric component under R60 isolates an intrinsic contribution, indicating orbital-origin rather than extrinsic geometry effects. • One-step KKR simulations reproduce the measured iLDAD at M/M′ for photon energies of 18.7–21.7 eV; Γ-region features show stronger photon-energy dependence and occasional sign reversals, attributed to the 3D character of Te-5p states and final-state nuances. • Theoretical photon-energy scans show iLDAD near M/M′ is robust against photon-energy variations, while Γ-region iLDAD varies strongly, supporting that the dominant dichroism at M/M′ is intrinsic and not governed by final-state effects. • TB + PW modeling identifies the main orbital contributors at M/M′ as Ti dz2, dxz, and dyz. Coherent superpositions of these orbitals are essential to reproduce the measured iLDAD symmetry and magnitude; incoherent sums fail to capture the effect. • Orbital-projected band structures reveal a momentum-dependent orbital texture: along different symmetry cuts, the relative weights of dz2, dxz, and dyz vary, yielding a threefold orbital symmetry despite sixfold band symmetry. • Real-space analysis shows the initial states form momentum-dependent hybrid orbitals that are effectively tilted d-like states. Enhanced (suppressed) photoemission occurs when the hybrid orbital lobes align (anti-align) with the photoelectron momentum, explaining the pocket-by-pocket iLDAD pattern and its inversion under 60° rotation (mirror).

By rotating the sample and extracting the antisymmetric component of the left-right linear dichroism, the study cleanly separates intrinsic orbital information from extrinsic geometric effects in ARPES of 1T-TiTe₂. The excellent agreement between experiment, one-step KKR simulations, and TB + PW modeling demonstrates that the strong dichroism around M/M′ arises from momentum-dependent hybridization and orientation (tilt) of Ti dz2, dxz, and dyz orbitals. The robustness of iLDAD at M/M′ across photon energies and the theoretical identification of coherent orbital superpositions as the source of the effect together rule out final-state effects as the primary origin. Thus, ILDAD provides direct, momentum-resolved access to orbital texture beyond conventional band mapping or standard linear dichroism (s ↔ p polarization), enabling differentiation of regions with equivalent eigenvalues but distinct wavefunctions.

The work introduces and applies intrinsic linear dichroism in photoelectron angular distributions (iLDAD) to unveil the orbital texture of 1T-TiTe₂. Multidimensional ARPES combined with a 60° crystal rotation cleanly isolates intrinsic dichroism, which is reproduced by one-step KKR and TB + PW models and traced to momentum-dependent hybridized Ti d orbital orientations. This establishes a route to go beyond eigenvalue mapping and access wavefunction-level information in solids via matrix element effects. Future directions include temperature-dependent iLDAD across charge-density-wave transitions (e.g., monolayer TiTe₂), probing putative excitonic insulator behavior via expected changes in orbital admixture, strain-dependent studies of topological phase transitions, investigations of nematicity-related gap anisotropy, and ultrafast pump-probe iLDAD to track non-equilibrium orbital texture dynamics, topological switching, and photoinduced orbital order.

• Exact photon-energy scans of iLDAD were not experimentally feasible with the present HHG source; photon-energy dependence was explored theoretically. • Discrepancies at Γ between experiment and KKR at fixed photon energy likely reflect limitations in DFT-based final-state energies (TRLEED shifts) and the 3D character of Te-5p states. • Minor up/down asymmetries in measured dichroism indicate small experimental imperfections (alignment, orientation, imaging). • TB + PW employs a plane-wave final-state approximation and simplified Wannier radial parts; while effective near M/M′ with Ti d dominance, this may be less accurate for states with strong Te p or complex final-state character. • Layer sensitivity is inferred (and supported by layer-resolved KKR) to be dominated by the top trilayer at XUV energies; deeper-layer contributions could still modulate details.