Characterization of low pressure chemically vapor deposited silicon nitride using experimental design

Abstract

During the low pressure chemical vapor deposition of Si 3N 4, significant interactions occur between the temperature, time, and gas flows, and thus a designed set of experiments has been used to establish the optimum deposition conditions. The conditions employed in this study were dictated by a full factorial matrix of the independent variables over 20 deposition runs. This procedure allowed us to vary the flow rates of NH 3 and SiH 2Cl 2 as well as the temperature and time of deposition. In this way we were able to determine the effects of these variables and their interactions on the electrical, physical, and chemical properties of the Si 3N 4. Deposition conditions can then be chosen to optimize film properties. Two sets of samples were prepared: Si 3N 4 was deposited either directly on silicon substrates or on silicon covered with 25 nm of SiO 2; the nitride films deposited on SiO 2 were subjected to a wet oxidation to simulate the structure in working devices. Following the formation of polysilicon electrodes on both types of samples, capacitance-voltage measurements were performed to obtain data on fixed charge in the insulating layers and on interface state density. The properties investigated were thickness and composition of the Si 3N 4; profiles of the silicon, nitrogen, oxygen, and chlorine content were obtained by secondary ion mass spectrometry and Rutherford backscattering analysis. We discovered that a significant amount of chlorine was incorporated in the nitride layers during deposition, that this chlorine was able to diffuse through an underlying SiO 2 layer and to accumulate at the SiSiO 2 interface, and that the chlorine had a strong effect on the amount of fixed charge in the structure. The deposition rate increased with the ratio of SiH 2Cl 2 to NH 3 flow, as did the amount of chlorine in the nitride, while the magnitude of the negative fixed charge in the dielectric layer decreased. Following an induction period that depended on temperature and gas flow rates, the deposition rate was thermally activated and constant.