Cesium adsorption on (2×4)-GaAs(001) was studied by photoemission and low -energy electron diffraction. The different Cs-induced changes of As 3d and Ga 3d core level spectra show that charge transfer is almost complete for Ga surface sites, but is negligible to surface As at a coverage Θ Cs < 0.3 ML. The situation becomes opposite for Θ Cs >0.3 ML, at which transfer occurs to As but no longer to Ga. Charge transfer to As atoms leads to surface disordering and destabilization and induces surface conversion from As-rich to Ga-rich (4x2)-GaAs(001) surface after annealing at a reduced temperature of 450°C.1/09/09 2 Adsorption of alkali metals on semiconductor surfaces has been investigated in most detail for silicon 1,2 and for the (110) cleavage face of GaAs, [3][4][5] for which the nature of chemical bonding, the amount of charge transfer, and the origin of the surface dipole have been essentially clarified. On the other hand, despite of technological applications, the fundamental aspects of adsorption on GaAs(001), including the difference in the adsorption mechanisms at the empty gallium-related and at the occupied arsenic-related dangling bonds is still far from being known. For the (2×4) As-rich surface, the early stages of Cs adsorption, characterized using both X-ray diffraction and theoretical calculations, 6 give rise to adsorption near both As dimers and Ga dangling bonds with a significant charge transfer only in the latter case.Here, we report investigation of the microscopic nature of the bonding between Cs and (2x4)-GaAs(001) using photoemission spectroscopy. The investigations were performed at the SU3 beamline of the SuperAco storage ring (Orsay, France), in a UHV chamber with a base pressure in the 10 -11 mbar range. The photoemission spectra were obtained using a hemispherical electron analyzer, at an overall energy resolution of the order of 0.15 eV at photon energy of 100 eV. Clean reconstructed surfaces of epitaxial p-GaAs(001) were obtained by using a treatment by HCl in isopropyl alcohol under dry nitrogen atmosphere, followed by introduction into ultra-high vacuum (UHV) and annealing to 450°C. 7 Low energy electron diffraction (LEED) investigations and measurements of the work function changes were performed in an independent setup by a retarding field method using a LEED gun.
8Several successive exposures to cesium (20, 30, and 50 minutes) were performed at RT using thoroughly outgassed dispensers in a vacuum not exceeding 10 -10 mbar. Using a previousAuger investigation, 9 we estimate that a deposition time of 50 minutes corresponds to a cesium coverage Θ Cs of about 50% of the saturation concentration, the latter concentration being taken as one monolayer (ML). The Cs 4d core level spectra (CLs), not shown here, are .50-1.55 were used. 11 For As 3d at Θ Cs <0.3ML , we have used two surface components.There is first a contribution S 1 , shifted by 0.62 eV to the higher BE side. This contribution is known to be caused by excess arsenic under the form of As dimers which are bonded t...