Angle-resolved photoemission and inverse photoemission provide direct experimental access to the occupied and unoccupied electronic structure E(k) of solid surfaces, respectively. The additional quantum number “spin” carries valuable information about exchange interaction in magnetically ordered systems and about spin–orbit interaction, which becomes especially important in systems with heavy elements. The experiments aim at determining the intrinsic spin polarization of the electronic states under investigation—but is this what is measured? While in ferromagnets the magnetization direction serves as the spin quantization axis, the situation is more complex in spin–orbit-influenced systems. First, spin–orbit interaction leads to k-dependent spin textures of the electronic states, called spin-momentum locking. Second, the experimental measurement of spin polarization by photoemission and inverse photoemission contains additional, even k-independent spin effects—depending on the orbital composition of the state under investigation in combination with the choice of experimental parameters. Is the electron spin polarization in or out of control? On the one hand, examples show that it is not straightforward to determine the spin polarization of electronic states. On the other hand, spin-resolved measurements performed with deliberately chosen geometries can provide comprehensive information about the orbital symmetries of the involved states. The (110) surface of tungsten serves as a textbook example for “controlling” the electron spin polarization.
Phys. Rev. Materials 1, 064604 (2017).
The metal/semiconductor hybrid system Tl/Si(111)-(1×1) exhibits a unique Tl-derived surface state with remarkable properties. It lies within the silicon band gap and forms spin-momentum-locked valleys close to the Fermi energy at the K and K′ points. These valleys are completely spin polarized with opposite spin orientation at K and K′ and show a giant spin splitting of more than 0.5 eV. We present a detailed preparation study of the surface system and demonstrate that the electronic valleys are extremely robust, surviving exposure to 100 L hydrogen and 500 L oxygen. We investigate the influence of additional Tl atoms on the spin-polarized valleys. By combining photoemission and inverse photoemission, we prove the existence of fully spin-polarized valleys crossing the Fermi level. Moreover, these metallic valleys carry opposite Berry curvature at K and K′, very similar to WSe2, promising a large spin Hall effect. Thus, Tl/Si(111)-(1×1) possesses all necessary key properties for spintronic applications.
Phys. Rev. B 95, 115401 (2017).
The optical control of spin currents in topological surface states opens new perspectives in (opto-) spintronics. To understand these processes, a profound knowledge about the dispersion and the spin polarization of both the occupied and the unoccupied electronic states is required. We present a joint experimental and theoretical study on the unoccupied electronic states of the topological insulator Bi2Se3. We discuss spin- and angle-resolved inverse-photoemission results in comparison with calculations for both the intrinsic band structure and, within the one-step model of (inverse) photoemission, the expected spectral intensities. This allows us to unravel the intrinsic spin texture of the unoccupied bands at the surface of Bi2Se3.
Phys. Rev. B 95, 085416 (2017).
We report on joint experimental and theoretical investigations of the unoccupied surface electronic structure of W(110). The spin-resolved inverse-photoemission experiments reveal a number of bands influenced by spin-orbit interaction and an image-potential state. The bands disperse differently within the two nonequivalent mirror planes of the surface, which is explained by their origin and their localization within the surface region. Surprisingly, the image-potential state also exhibits anisotropic dispersion, although it is strongly located within the surface barrier. The experimental findings are confirmed by first-principles electronic-structure calculations.