Electronic structure of pure and alkali-metal-intercalatedVSe2

Abstract

The valence bands of the layered compound ${\mathrm{VSe}}_2$ and the related intercalation compounds ${\mathrm{Na}}_x{\mathrm{VSe}}_2,$ ${\mathrm{K}}_x{\mathrm{VSe}}_2,$ and ${\mathrm{Cs}}_x{\mathrm{VSe}}_2$ have been investigated by means of angle-resolved photoelectron spectroscopy, and compared to self-consistent linear augmented plane-wave (LAPW) band calculations. The intercalation compounds were prepared in situ by deposition of Na, K, and Cs on ${\mathrm{VSe}}_2$ cleavage surfaces. The intercalation was monitored by core-level spectroscopy, and although K was found to intercalate more slowly than Na and Cs, estimated alkali concentrations of $x=0.2\mathrm{}–0.3$ were reached for all three alkali metals. Additional depositions mainly seemed to increase the intercalation depth. Good agreement between LAPW calculations and valence-band spectra was found, in particular for the dispersion along the layers. Normal-emission spectra, obtained at different photon energies, indicated vanishing perpendicular dispersion, but in spectra measured under variation of the emission angle some band-edge signatures were seen, which suggests that some perpendicular dispersion remains, in accordance with the LAPW calculations. The lack of dispersion in the normal-emission spectra could be due to intercalation induced structural transformations, leading to stacking disorder. Also correlation effects may contribute. The rigid-band model is found inadequate, except as a crude approximation, for describing the changes during the initial phase of intercalation. It might be used to describe the continued intercalation, however, under condition that the intercalation modified bands are used. The need for studies that probe both electronic and crystallographic structure (including defects) is stressed.