Preliminary Study Of The Computer Simulation Of Semiconductor Thin Film Growth

The early stages of the deposition of thin films onto substrates consists in the formation and growth of clusters that eventually connect each other to give a continuous film. This process is of great importance for the properties of the final film, since the average grain size of poly crystalline films is affected by the size distribution of the clusters before the formation of a continuous phase. The grain size , in turn , strongly affects the mechanical and electrical characteristics of the film: for example, the larger the grains, the higher the resistance to electromigration and therefore the longer the life of very large scale integration(VLSI) interconnections.At the opening of the dissertation, the growth mechanism of thin films is discussed.Based on the investigation of molecular dynamics and statistics physics of thin films growth, a program for the simulation of the nucleation and growth of thin island films on substrate is presented. The program is based on the description of the phenomenon in terms of adsorption, diffusion and reevaporation of single atoms on the substrate surface, which is represented as a triangular lattice of adsorption sites. A method of taking into account the interaction with neighboring particles in the evaluation of the diffusion activation barrier is proposed. The simulation is devoted to investigating the role of diffusion and reevaporation in the determination of the particular microstructure of the clusters. The effects of the substrate temperature on the morphology of the simulated island of films are also presented.Secondly, a program of simulating the growth of SiGe thin films is presented as an example of the above model. As the change of growth rate with temperature and the content of Ge with the ratio of gas flow, the simulation results indicate the similar growth tendency with our experimental data.At last, a simple model is also proposed , which is based on the regular solution theory and explains the Ge enrichment in the surface or subsurface of Si].xGexexpitaxial thin films. The model using the Monte Carlo method can shows fine agreement with the experimental data.