Arsenic for antimony exchange on GaSb, its impacts on surface morphology, and interface structure

We quantify the rates and total amounts of the arsenic for antimony exchang e on both the Sb-terminated and Ga (or In)-terminated GaSb (001) surfaces u sing in situ real time line-of-sight mass spectrometry (LOS-MS) during mole cular beam epitaxy. On the Sb-terminated GaSb (001) surface, an As for Sb e xchange is observed to occur at all values of incident As-2 flux considered . At high substrate temperature, three-dimensional (3D) nanometer-sized clu sters from as a consequence of As/Sb exchange and lattice mismatch strain b etween GaAs and GaSb. The 3D clusters are found to have lateral dimensions of similar to 10-30 nm and heights of 1-3 nm by atomic force microscopy (AF M). By contrast, at lower substrate temperatures a two-dimensional surface morphology is maintained, and AFM reveals an array of atomically flat terra ces. On the surface terminated by one monolayer (ML) of Ga or In, there exi sts a critical As-2 flux below which the As/Sb exchange is greatly diminish ed. The net amounts of Sb leaving the surface during one period of InAs/GaS b type-II superlattice growth are measured in real time by LOS-MS and estim ated to be in the range of 0-0.4 ML for the various conditions used. By sup plying only an As-2 beam to a GaSb surface covered by InAs, the Sb riding o ver the InAs layer is replaced by arsenic and the total amount of such Sb i s measured. The amount of Sb riding on the InAs can be as large as 0.8 ML f or the first 1 ML of InAs and it gradually decreases to zero as the number of InAs monolayer increases. X-ray diffraction data show that all the InAs/ GaSb superlattices coherently match with the GaSb substrate in the growth p lane. The average lattice constant along the growth direction reduces with decreasing Sb mole fraction shown by the increased Sb desorption signal. Us ing the information on As/Sb exchange and Sb riding on the InAs surface, we predict an average lattice constant along the growth direction to be consi stent with the measured one to within 2x10(-4). (C) 1999 American Institute of Physics. [S0021-8979(99)05013-6].