No abstract

No abstract

No abstract

Compositionally graded films of SiGe/Si(100) and GaInAs/GaAs were grown under different conditions in order to investigate the different modes of strain relaxation associated with the compositional grading. We show that, when the growth conditions are very clean and the gradient is shallow enough (about 1% misfit per half micron), very good, relaxed films are obtained. This coincides with the introduction of large numbers of dislocations in the substrate itself, which is counter-intuitive at first since the substrate is under negligible strain. We show that this introduction of dislocations is the result of the activation of novel Frank–Read-like sources located in the graded region, and is directly correlated to the lack of other low energy nucleation sites for dislocations. We detail the conditions of growth necessary for this phenomenon to occur, and show that it operates both for the SiGe/Si system and the GaInAs/GaAs system. Pure, relaxed Ge films have been grown in this manner on Si(100), with a defect density as low as 106/cm2.

By comparing the C 1s photoemission intensities from a clean diamond {111} surface to those obtained from a fluorine-covered surface, we conclude that a single t, '111) layer of atoms participates in the diamond 2X 1 surface reconstruction. The data interpretation presented does not rely on assuming an electron mean free path, but, in fact provides an independent measurement of its value. For clean diamond (111) a surface core-level peak is observed at 0.80+0.05 eV lower binding energy than the bulk-C 1s peak at 285.0 eV below E~. After fluorine exposure a single chemically shifted carbon peak is observed at 1.85+0.05 eV higher binding energy than the bulk peak, indicating the presence of CF units at the surface.

No abstract

We have employed the techniques of photoelectron spectroscopy {with tunable synchrotron radiation 30& h v& 100 eV) and inverse photoemission spectroscopy at h v=9.7 eV to study respectively the occupied and unoccupied electronic states of n-type SrTi03 doped with 1 at. /o Nb. Doped samples have the advantage that charging effects are avoided and hence a rather accurate determination of peak positions is possible. The total (occupied and empty) experimental density of states agrees with the calculated density of states of Pertosa and Michel-Calendini when their band-gap energy is readjusted. %'e have no evidence for intrinsic surface states either in the band gap or in the conduction band, as was calculated by Wolfram et al. The admixture of Ti 3d states in the 0 2p valence states is small, as can be concluded from the Ti 3p~3 d resonance behavior.

No abstract