Photoemission from commensurate and incommensurate alkali metal monolayers on Be(0001)
Algdal, Jonathan; Breitholtz, Marcus; Wallden, Lars; Lindgren, S.-A.; Balasubramanian, Thiagarajan; Hellsing, Bo; Chis, Vasile Sweden
As may be expected from the relative sizes of the atoms LEED shows that a full K monolayer on Be(0001) has 2x2 order while for Na it is incommensurate forming a close packed layer aligned with the substrate and around 1.65x1.65 order. The electronic structure revealed by photoemission is similar for the two systems and characterized by a quantum well state (QWS) band with quadratic dispersion extending 0.55 eV (Na) and 0.40 eV (K) below EF with the band minimum at the center of the Brillouin zone. Model potential (Na) and first principles calculations (K) show that this QWS has one node in the film and that most of its charge is in the outermost region of the monolayer. There is no bottom nodeless state with a maximum in the overlayer quantum well but instead the Be(0001) surface state (SS) 2.8 eV below EF at Γ obtains an extended tail outside the substrate and is downshifted to 3.7 eV for Na and 3.4 eV for K. The Be(0001) SS at the M-point, 1.9 eV below EF, obtains a smaller shift (0.3 eV for Na and 0.2 eV for K). The reason is that this state deposits little charge in the outermost substrate layer.
While the electronic structure of the two systems is quite similar and well understood the photoemission spectra show qualitative differences, which we ascribe to the difference with respect to structure. For the commensurate K overlayer one finds strong emission lines corresponding to Be states folded either to or from the smaller zone defined by the 2x2 overlayer. As examples the SS at M in the Be(0001) zone gives a prominent emission line in spectra recorded along the surface normal and the Γ SS band is repeated in the second 2x2 zone. For the incommensurate Na overlayer these diffraction effects are less important. This reflects how the charge is shared between substrate and overlayer for the states, the QWS residing mainly in the overlayer and the SS mainly in the substrate. The QWS band shows the lateral period defined by the incommensurate overlayer, and the Be derived states the period of the substrate.
The interpretation of one emission line remains obscure. This is observed for both Na and K spectra recorded near the surface normal and may be a loss satellite to the QWS line separated from it by around 0.75 eV.