Electronic structure of some SiC and Be surfaces

Sammanfattning: Silicon carbide (SiC) is a wide-bandgap semiconductor whose properties make it a good candidate for high-power, high-voltage, high-temperature and highfrequency devices. Beryllium (Be) is a metal with some unusual electronic properties.In this thesis synchrotron radiation and photoemission have been utilized as tools to extract information from both the core levels and valence band states of different SiC and Be surfaces.Several earlier SiC investigations have dealt with the silicon face and not so many with the carbon face. In paper I the carbon faced 6H-SiC(0001̅ ) surface is studied. The C 1s and Si 2p core levels are characterized for different photonenergies and after annealing to different temperatures. A 3 x 3 reconstruction developed after annealing to 950° C. After heating to 1200° C, spectra showed evidence of surface graphitization.In paper II, an investigation of the SiO2 /SiC-interface using high energy photoemission spectroscopy is reported. A low defect density in the interface is crucial for the manufacturing of power MOSFETs. One limiting factor for the formation of a high quality oxide is thought to be a carbon containing by-product at the interface. In this investigation, no carbon could be detected at the interface, but instead a 3 Å layer of graphite like carbon was found on top of the oxide.Paper III and the first part of paper IV deals with the surface state on the √3 x √3 reconstructed surfaces of 6H-SiC(0001) and 3C-SiC(111) respectively. In paper III the sensitivity of the surface state to oxygen and Na exposures was investigated and the symmetry and dispersion of the surface state was determined in both paper III and IV.Theoreticians have proposed a Mott-Hubbard model to explain the semiconducting behavior of this reconstruction, but this would also give a surface state with the periodicity of the √3 x √3 Surface Brillouin Zone. This periodicity has not been observed for either the 6H-SiC(0001) or the 3C-SiC(111) crystals.In the second part of paper IV, an investigation similar to the one in paper I has been done, but on the 3C-SiC(111) and 3C-SiC(1̅1̅1̅ ) surfaces. Similar results as for the 6H and 4H polytypes were found for both the Si-face and the C-face.The mean values of the surface core level shifts (SCLS) were lower for 3C than for 4H and 6H, but when taking the error bars into account the shifts cannot be said to be different. For both the 6H-SiC(0001̅ )  and the 3C-SiC(1̅1̅1̅ ) the C 1s spectra for the 3 x 3 reconstruction indicate the possibility of a quite complex atomic structure.  Paper V and VI deal with two different Be surfaces. Paper V reports a fourthlayer core level shift on Be(0001), which is unmatched by any other metal. In paper VI experimental evidence for a model for the assignment of the SCLSs on Be(101̅ 0) earlier proposed by theory is reported. The most intuitive way is to assign the largest shift with the first layer, the second largest shift with the second layer and so forth. This was also supported by the intensity relations of the components. Other groups have later proposed that the component with the largest shift is from the second layer, the second largest shift is from the first layer and that the third largest shift is from the third layer. In paper VI the intensity ratios calculated using a multiple scattering method show that the component with the largest shift is connected to the second layer, the second largest shift to layers three and four and the smallest shift to the first layer.