LAUSR

Abstract The development of functional relationships between the observed deposition rate and the experimental conditions is an important step toward understanding and optimizing low-pressure chemical vapor deposition (LPCVD) or low-pressure chemical vapor infiltration (LPCVI). In the field of ceramic matrix composites (CMCs), methyltrichlorosilane (CH3SiCl3, MTS) is the most widely used source gas system for SiC, because stoichiometric SiC deposit can be facilitated at 900 °C–1300 °C. However, the reliability and accuracy of existing numerical models for these processing conditions are rarely reported. In this study, a comprehensive transport model was coupled with gas-phase and surface kinetics. The resulting gas-phase kinetics was confirmed via the measured concentration of gaseous species. The relationship between deposition rate and 24 gaseous species has been effectively evaluated by combining the special superiority of the novel extreme machine learning method and the conventional sticking coefficient method. Surface kinetics were then proposed and shown to reproduce the experimental results. The proposed simulation strategy can be used for different material systems.