Chemical State Characterization of Layers Produced by Ion Beam Synthesis: The Silicon-Germanium-Oxygen System.

This thesis describes the work using RBS and XPS analytical techniques in conjunction with thermodynamic analysis to study buried oxide layers in a Si0.5Ge0.5 alloy produced by ion beam synthesis. The motivation for this research was to demonstrate the importance of characterizing the chemical state of multi-element or multi-valent compounds by a method which allowed thermodynamic interpretation of the equilibrium reached in ion beam processed systems. The Si-Ge-O system was so chosen because it is of current technological interest to synthesize buried oxide layers in SiGe alloy and theoretical importance to understand its oxidation mechanism. This system acts as a good working example of what can be achieved in both multi-element and multi-valent compound systems and the results obtained from this study may have wide-range applicability. To synthesize the buried oxide layer in a SiGe alloy, three high doses of oxygen ions (0.6, 1.2 and 1.8x10 18O+cm-2) with an energy of 200keV were implanted into samples of a Si0.5Ge0.5 alloy. The alloy was grown by molecular beam epitaxy (MBE) on a n-type (100) silicon substrate. Selected samples were, subsequently, annealed at different temperatures of 800°C, 900°C and 1000°C for one hour. The sample composition and crystal quality, corresponding to each of the above treatments were first inspected using RBS and the channelling technique. The elemental distribution and chemical bonding of Si and Ge associated with their oxides in the sample were then examined in detail by XPS depth profiling and spectrum synthesis. It has been observed that the formation of silicon and germanium oxide is highly dependent on the dose of oxygen. The implanted oxygen reacts preferentially with silicon to form SiO2 and this leads to a rejection of germanium from the buried oxide layer and its segregation in the Si/SiO2 interfaces. GeO2 can be formed only in the buried layer when the applied oxygen dose is higher than the value for stoichiometric SiO2 in the alloy. Based on these results, the changes of Si, Ge and O activities with the concentration of oxygen were calculated using the Thermo-Calc program, and chemical-potential depth profiles for the three components in the system throughout the oxygen projected range from all the as-implanted samples were determined. In light of the thermodynamic analysis, oxygen implanted into the SiGe alloy results in an increase of the chemical potentials of oxygen and germanium in the buried oxide layer, and a significant decrease in that of silicon. Transport during implantation and in subsequent annealing then follows the gradient of chemical potential: silicon atoms diffuse in the buried layer to form SiO2 by the replacement of Ge in Ge oxides, and the released germanium atoms diffuse out to accumulate at each side of this layer. Increasing annealing temperature promotes this mutual diffusion process, which favours formation of more SiO2 and rejection of more germanium from the oxide, leading to the system reaching a more stable thermodynamic state. However, it has been found that in all the cases there is some Ge and its oxides trapped in the SiO2 so that after annealing the ultimate thermodynamic equilibrium is not achieved in the system. The reason why Ge has a 'snowplough' or 'pile-up' behaviour in the oxidation of the SiGe alloy is attributed to the fact that the driving force given by its gradient of chemical potential is so strong that Ge has to diffuse in this direction even if it is against its gradient of concentration.