Structural and electrical characterisation of Si:B/Si1-xGex/Si(001) heterostructures

In this work the problems associated with the fabrication of high germanium content (x≥0.2) Si1-xGex heterostructures have been examined. An optimum growth temperature of 550-600 deg. C was determined for solid source molecular beam epitaxy of Si0.7Ge0.3 heterostructures. At lower growth temperatures, the poor 2DHG carrier mobility was attributed to poor material quality, ie. large vacancy concentration. At higher growth temperatures, strain induced roughening was seen to degrade the 2DHG mobility. One method of reducing the alloy layer roughening was to use surfactant mediated growth. While initial work showed that segregation of germanium on silicon was suppressed, using low energy SIMS, we also observed a dramatic loss of crystallinity and a high level of boron contamination. The crystallinity was recovered by growing under a lower residual hydrogen pressure. The boron nitride RF cell of the atomic hydrogen source was considered the most likely source of the high boron contamination and was replaced with a quartz one. To reduce the detrimental roughening effect in high germanium content heterostructures, we also investigated the alternative of using graded germanium concentration profiles in the heterostructure channel. By this method the total effective strain was reduced to promote two dimensional growth while maintaining a high peak germanium content. Analysis of linearly graded profile channels showed the structural advantage of normal graded profiles over inversely graded profiles. Electrically the opposite was observed due to the high effective field produced in the alloy layer reducing the 9K 2DHG mobility for normal modulation doped heterostructures. For the first time, the Hall-and-Strip method has been successfully applied to modulation doped heterostructures to permit direct measurement of the 2DHG carrier mobility at room temperature. The successful application of the Hall-and-Strip method was found to be conditional upon the removal of an unintentional boron doping layer at the substrate/epilayer interface, inherent to solid source MBE growth, which was achieved using careful substrate preparation. Hall-and-Strip analysis of modulation doped Si0.8Ge0.2, showed the 2DHG Hall scattering, factor was well below unity at room temperature (∼0.3). Using the experimentally determined Hall scattering factor, a room temperature 2DHG carrier drift mobility of ∼400 cm2/Vs was determined for a sheet carrier concentration of 3.3x1011cm-2. Similarly, a minimum room temperature drift mobility of 300 cm2/Vs (ns=6.6x1011cm-2) was determined in a modulation doped Si0.7Ge0.3 heterostructure. Unfortunately, copper contamination of this wafer prevented an accurate drift mobility from being determined. However, we may expect a much higher drift mobility in uncontaminated material. (author)