Silicon dioxide thin films are the basic components for microelectronics and other micro technologies. As the size of the transistors and relevant micro devices shrinks for faster and smaller chips and other microsystems, growth of high quality and ultrathin silicon dioxide films becomes more important. In this study, we grew ultrathin silicon dioxide films by using a rapid thermal processing furnace, and characterized the film using a high resolution SIMS (secondary ion mass spectrometry). We found the first direct evidence for a transitional layer between the pure silicon substrate and the pure silicon dioxide. We then established a unified model to predicting the growth rate based on experimental observation. The growth of silicon dioxide has been the focus of research for the last three decades. The prevalent view is based on the seminal paper by Deal and Grove1. It has been known for a long time, however, that the growth rate for ultrathin films is faster than that would be predicted from the kinetics of the Deal-Grove model. To overcome this shortcoming of the Deal-Grove model, a term has to be added, which could enhance the growth rate in the thin-film regime yet still approach the Deal-Grove behavior asymptotically in the thick-film regime2. Physical characterization by using Auger and X-ray photoelectric spectroscopy found3 that the oxide has suboxide states of Si+3, Si+2 and Si+1. We grew silicon dioxide films in a furnace-based rapid thermal processing system (Axcelis’s Summit -200TM). The sample was then analyzed with Accurel’s Atomika SIMS. Figure 1 shows the profile of silicon and oxygen for a sample grown at 1100oC for 40 seconds. It clearly demonstrates that there is a transitional layer with a thickness of 2 nm. Based on the experiments, a growth process consisting of three major steps is proposed. First, oxygen diffuses through oxide. Then some oxygen reacts with fresh silicon at the SiOx/Si interface. The rest penetrates into the silicon substrate while reacting with oxygen in a layer (Fig.2). Other assumptions of the model include: 1) Oxidation reactions take place in a hybrid layer, in addition to at the interface of oxide/silicon, 2) The concentration of silicon in the hybrid layer is approximately constant. This can be justified by the fact that, the concentration of silicon is much larger than the oxygen concentration, 3) The process is quasi-steady, 4) The system being studied has a uniform temperature even though oxidation is a highly exothermic reaction, and 5) All the oxygen diffused through the oxide layer will be consumed to produce the oxide. Figure 2 shows the schematic arrangement of interest to this analysis, which is the hybrid layer in the middle, where oxygen diffuses and reacts with silicon. In Fig. (3a), calculation data (lines) are compared to experimental data for oxidation at 1 atm O2. Fig (3b) compares experimental data from ref. (3) and calculation for oxidation at 0.2 atm O2.
Journal: TechConnect Briefs
Volume: 2, Technical Proceedings of the 2003 Nanotechnology Conference and Trade Show, Volume 2
Published: February 23, 2003
Pages: 40 - 43
Industry sector: Sensors, MEMS, Electronics