|
Index |
Doping Issues in Wide Band-Gap SemiconductorsExeter, United Kingdom21-23 March 2001 |
Next |
Most studies on dopant issues have focused on their properties in bulk. However, recent experiments indicate that dopants may significantly affect the growth morphology (and thus the material quality) and also the bulk concentration. Examples are the anti-surfactant behavior of Si on GaN(0001) surfaces (which leads to the formation of quantum dot like structures) or the reduction in the oxygen concentration when going from the GaN(000-1) to the (0001) surface.
In order to get some insight into the mechanisms causing these effects we have studied the adsorption and incorporation of dopants on GaN surfaces employing density-functional theory. In the present talk we will focus mainly on two dopant impurities - oxygen and silicon - which both are well known to act as shallow donors in GaN. We have thereto calculated the surface energy, atomic geometry and electronic structure for a large number of possible surface reconstructions where we have varied the ratio between Ga, N and impurity atoms. Based on these results we construct a phase diagram which gives all stable surfaces as a function of the chemical potentials.
Using these results we find a qualitatively different behavior for oxygen and silicon. Oxygen has a strong tendency to segregate to the surface: the oxygen concentration on the surface will be therefore significantly higher than in bulk. Also, the incorporation of oxygen on GaN surfaces is highly stable against the formation of the solubility limiting phase Ga2O3. In contrast, silicon is more stable in the bulk than on the surface and almost all silicon induced surface structures are unstable against the formation of Si3N4. These results give immediate insight into recent experiments. In particular, we identify the conditions under which Si acts as an anti-surfactant, we discuss how the doping efficiency of silicon can be increased, and we explain why the oxygen concentration on (0001) is lower than on (000-1).
This work was supported in part by the DFG (Schwerpunktprojekt Gruppe-III-Nitride).