Experimental verification of a spin effect in photoemission: Polarized electrons due to phase-shift differences in the normal emission from Pt(100) by unpolarized radiation

Abstract

A theoretical prediction of Tamura and Feder has been experimentally verified: Photoelectrons from the fourfold-symmetric surface of a centrosymmetric crystal, Pt(100), can be polarized even if the incident radiation is unpolarized and the electrons are emitted normal to the surface. For 21.2and 16.9-eV photon energies, a spin-polarization component P~p erpendicular to the reaction plane is found. The degree of polarization is up to 15% and does not change when the crystal is rotated about its surface normal. This supports strongly the prediction that the effect is due to phase-shift differences.This work reports on the experimental verification of an effect in photoemission that produces spin-polarized electrons without making use of magnetically ordered electron spins (as in photoemission from magnetic materials') or of optical spin orientation by excitation with circularly polarized light. It was predicted in a recent theoretical work by Tamura and Feder and yields spinpolarized electrons with unpolarized light even in the normal photoemission from the fourfold-symmetrical Pt(100) surface, for which other earlier reported spinpolarization effects with linearly and unpolarized radiation from nonmagnetic crystals are forbidden by symmetry (see below}. The effect has its origin in a broken symmetry due to off-normal light incidence in combination with hybridization [b,656 hybrids for (100) surfaces] and yields spin-polarized electrons due to phase-shift differences. The spin-polarization vector P is perpendicular to the reaction plane of the incident radiation and the emitted electrons. The effect is only predicted for the "one-step model" of photoemission, while it is absent in the "three-step model. " It is expected to occur for practically each crystal surface, which makes it widely applicable. In photoemission from ferromagnetic samples, for example, it competes with exchange-induced effects and might influence the interpretation of spinresolved photoemission spectra and conclusions concerning ferromagnetic properties. There have been previous reports in the angle-resolved photoemission from nonmagnetic materials with linearly and unpolarized radiation but they all differ fundamentally from the effect reported in this work: In Ref. 5 spinpolarized electrons arise from spin-dependent photoelectron diffraction or phase matching conditions at the solid vacuum interface and are only obtained when the electron emission occurs a+normal Spin-polarized e. lectron emission with linearly and unpolarized light was also observed in the photoionization of free unpolarized atoms and molecules where it is a consequence of a quantummechanical interference between different photoelectron partial waves. It was, however, until now not clear whether a counterpart of this effect for solids can exist at all. Even for normal photoemission with linearly and unpolarized light from nonmagnetic crystals spin-polarized electrons were found when the sample is noncentrosymmetric.Phase shifts do not, however, play any role in this case and...

Widespread spin polarization effects in photoemission from topological insulators

“…SMEs in the solid state are not confined to core levels, and have been predicted [42][43][44][45] and subsequently observed [46][47][48][49][50][51] in various related forms in the valence bands of Pt and Au single crystals. As seen in various Rashba-Bychkov spin-split surface states [52][53][54], TRS in these systems requires that any intrinsic spin polarization of quasiparticles, P qp , must be antisymmetric with respect to k = 0, where k is the quasiparticle in-plane crystal momentum.…”

“…This ensures zero net spin imbalance integrated over all k-space and that P qp ( k = 0) = 0. In each of these cases [46][47][48][49][50][51], however, normal emission ( k = 0) photoelectrons from valence bands were found to be significantly spin polarized (∼ 10 -20%) dependent on the photon energy and polarization. These effects were explained in terms of SMEs.…”

Spin Polarization and Attosecond Time Delay in Photoemission from Spin Degenerate States of Solids

“…The spin polarization is produced by the phase shift between the two interfering transitions T 1;2 , which in the case of atomic photoionization correspond to the two final partial waves with orbital quantum number l AE 1. In solids one should also distinguish between probed in-plane and out-of-plane orbitals [47], as well as mixed spatial symmetries of the considered state in the double group symmetry representation [8,20], both in the initial and final states. In any case, the interfering transitions are different photoemission channels which do not correspond to different photoelectrons, but they together build up the photoelectron wave function.…”

Photoelectron spin-flipping and texture manipulation in a topological insulator

“…6, [17][18][19][20] Angle-resolved photoemission spectroscopy (ARPES) directly maps the dispersions and Fermi surfaces of such electronic states in energy-momentum space. Spinresolved ARPES also measures the spin polarization of the corresponding photoelectrons.…”