Surface morphology of growing epitaxial layers: simulations using kinetic Monte Carlo method
MetadataShow full item record
AbstractThis thesis attempts to study the process of surface roughening during the deposition of a semiconductor material on another, a process commonly known as epitaxy, using kinetic Monte Carlo modeling. The surface profile and its morphology change during epitaxial deposition, and information about the atomistic processes involved is important for fabrication of semiconductor nano-devices with desired characteristic features. A multi-particle kinetic Monte Carlo model is used to simulate deposition of Si on Si(00l) (homoepitaxy) in (1+1)-dimensions and Ge on Si(00l) (heteroepitaxy) in (2+ 1 )-dimensions. The simulation model is based on a discrete description of atoms so that the unit length scale coincides with the atomic diameter. The first set of results corresponds to (1 + 1 )-dimensional simulations of Si layers deposited on initially flat Si(00l) and initially structured Si(00l) substrate. The driving force for the motion of diffusing atoms ( or adatoms) is the tendency of an adatom to maximize the number of bonds it makes with its neighbors. Various roughness parameters were evaluated and analyzed to give insight into the process. Results obtained from one-dimensional simulations show the evolution of surface roughening as Si is deposited on Si(00l) substrate. We also performed annealing simulations of flat and stepped Si(00l) deposited at different temperatures by turning off the deposition rate in our kinetic Monte Carlo model. Results are presented for annealing simulations at different temperatures showing that surface roughness can be decreased and smoother surfaces can be obtained by annealing at high temperatures. Annealing simulations at low temperatures do not show observable smoothing, thereby indicating that adatoms are m a stable configuration at low temperatures, and that the probability for their movement is very small. The last set of results belongs to (2+ 1 )-dimensional simulations to study the heteroepitaxial growth of thin films . The aim of these simulations is to study the atomistic processes leading to the formation of heterostructures on semiconductor surfaces during heteroepitaxial growth. Apart from taking into consideration the number of bonds an adatom makes with its neighbors, adatom-adatom interaction is also considered in this model, which is also a departure from ( 1 + 1 )-dimensional simulations. The interaction between the adatoms is modeled using the famous Lau and Kohn expression for the interaction of surface atoms. Formation of heterostructures is not observed and it can be argued that the amount of material deposited is not sufficient enough to observe any such clustering. The results obtained do not give much insight into the processes occurring at the surface due to limitations imposed by large computational time. The usefulness of the simulations in (2+ 1)-dimensions lie in the fact that they represent successful implementation of extension of our one-dimensional code.
Bibliography: p. 100-107