Photoemission studies of the interaction of oxygen with GaAs(110)

Abstract

Room-temperature adsorption of oxygen on cleaved GaAs(110) surfaces has been studied with photoemission measurements using synchrotron radiation ($h\nu =20-120\mathrm{}$ eV). Particular effort has been spent to obtain the proper photoemission spectrum that represents the density of valence states (DOVS) of the oxygen bonded to GaAs(110). Simple theoretical interpretation, substantiated with comparison made between the DOVS of O-GaAs(110) and the experimental DOVS's of ${\mathrm{Ga}}_2$ ${\mathrm{O}}_3$ and ${\mathrm{As}}_2$ ${\mathrm{O}}_3$, of the DOVS of O-GaAs(110) is given. This interpretation of the DOVS has led to the proposal of a new adsorption model involving both nonbridging oxygen (As=O) and bridging oxygen (Ga-O-As). Evidence extractable from other measurements for the new model is also discussed. Another form of oxygen which saturates at relatively low oxygen coverage (≤0.02 monolayer) has also been observed, and is suggested to adsorb at defect sites of the GaAs(110) in the form of Ga-O-Ga units. The role of oxygen adsorption at defect sites in the initial steps of adsorption, particularly the dissociation of oxygen molecules, is discussed. Such discussion is further substantiated be detailed studies of the oxygen adsorption on sputter-disordered GaAs(110) surfaces. A two-step adsorption process has been found to occur on sputter-disordered surfaces. On either ordered or disordered surfaces, direct formation of ${\mathrm{Ga}}_2$ ${\mathrm{O}}_3$ by room-temperature exposure to unexcited oxygen is highly unlikely. The formation of ${\mathrm{As}}_2$ ${\mathrm{O}}_3$ by room-temperature adsorption of oxygen on GaAs(110) is also definitely ruled out by thermal annealing experiments. On the other hand, the results of the thermal annealing experiments are consistent with the existence of the As=O bond.