No abstract

Gas phase transport in Si epitaxial growth has been investigated by means of in situ mass spectrometry measurements. of reaction gas concentration profiles. Five species, SIC14, SiHC13, SiHfC12, SIC12, and HC1 are observed in the reaction gas. Not only SIC14 (source gas) but the other Si chlorides as well contribute to the surface growth reaction. The primary reaction species shifts from SIC14 to the other Si chlorides with decreasing gas velocity and with increasing temperature and position along the susceptor.Gas phase reactions and transport phenomena in Si epitaxial growth have been studied extensively by a number of authors and reaction models deduced from an equilibrium viewpoint (1-5). However, the actual reaction is not in equilibrium, but rather quasiequilibrium (6), and the Si epitaxial reaction is difficult to describe by a theoretical model, It is necessary to observe the reaction gas species by in situ measurements in a reactor. Recently such measurements have been made and reaction models proposed. Ban and Gilbert observed the reaction gas profiles in a reactor by using mass spectrometry and reported that SIC14 and only a small amount of SIC12 were found (7, 8). A reaction mechanism was proposed in which SiCI~ was transported to the substrate surface by diffusion and the source gas decomposed into solid Si and HCI byproduct. The following heterogenous reaction occurred at the surface SiC14(g) + 2H2(g) --> Si(s) -t-4HCI(g) Sedgwick and Smith observed reaction gas concentration profiles using laser Haman spectroscopy in Si epitaxy with SiH2C12 as source material (9). The SiHfCI2 decomposed into SIC12 in the gas phase, and then SIC12 was transported to the surface where the epitaxial layer grew SiHfC12---> SiCI# + H2 (gas phase) SiCI2 + Hf--> Si + 2HC1 (substrate surface) Nishizawa and Nihira observed reaction gas profiles by using infrared absorption spectroscopy and found SIC14, SiHC!~, SiHfC~2, and HC1 species (10). They proposed the following models SiHClz SiCla § H2 ~ SIC1,2 ~__-Si % SiH2C12 + HC1Duchemin also investigated Si epitaxial reactions extensively by observing reaction gas profiles in a vertical type reactor (11).Recently, Pollard and Newman have reported a model in which multicomponent mass transfer was considered (12). However, they used calculated gas concentration profiles for the calculation of growth rates. We have observed the reaction gas concentration profiles by in situ measurements using mass spectrometry while changing the growth parameters. It is apparent that the three models mentioned above are likely to describe only a part of the actual reaction. In this paper, it is proposed that all the Si chlorides, SiCIly, SiHC13, SiH2C12, and SIC12, participate in the surface reaction, and the primary surface reaction species which contributes to the growth rate is suggested to shift with changing gas velocity, substrate temperature, and position along the susceptor.

Impurities at the interfaces of HF-treated Si (100) substrates and Si films prepared by low-temperature (900 °C) chemical vapor deposition using SiH2 Cl2 were measured by secondary-ion mass spectroscopy. Si substrates were prepared by 49% HF, 5% HF, and 0.05% HF acid solutions and loaded into the growth chamber without a water rinse. Carbon, fluorine, oxygen and chlorine were detected at the interfaces for 49% HF and 0.05% HF treated substrates, but they were not detected for 5% HF-treated samples. Desorption of the contaminants appeared to be sensitively related to a difference in chemical states of the HF-treated surfaces.

HF‐normaltreated Si (100) substrates were used for low‐temperature (900°C) epitaxy. A conventional low‐pressure chemical vapor deposition (LPCVD) system was used for Si deposition. The stacking fault density (SFD) in the epitaxial layer drastically increased with the duration of the ultrapure water rinse time after HF dipping. Oxygen, fluorine, and carbon contaminants after HF treatment were measured by x‐ray photoelectron spectroscopy. The chemical bonding state of the residual carbon was related to the HF concentrations. For the 5% HF treatment, which corresponded to the optimum epitaxial condition, mainly C‒O was present.

No abstract