In this paper, we present a process which provides low thermal budget removal of carbon, oxygen, and chlorine from the silicon surface by annealing the substrate in a vacuum ambient with water and oxygen partial pressures less than io-Torr. Following the in situ clean, silicon was deposited epitaxially, and carbon, oxygen, and chlorine levels at the epitaxy/substrate interface were quantified using secondary ion mass spectroscopy. It was demonstrated that these contaminants were removed from the silicon surface with a 10 s anneal at temperatures as low as 75 0°C. Both Si02 patterned and bare silicon wafers were employed, and the patterns on the surface did not show any effect on the in situ clean efficiency. The impact of chlorine as a contaminant was also studied, and no effect on the in situ cleaning efficiency was observed. Surface roughness was quantified by atomic force microscopy, which revealed that surface roughness on Si<100> did not increase after in situ cleaning in vacuum. A defect analysis was performed using optical and scanning electron microscopy. A correlation was obtained between the defect density and the carbon levels at the epitaxy/substrate interface. However; electrical device characterization did not show any correlation between different in situ clean temperatures and the n-channel metal-oxidation-silicon field effect transistor leakage current or transconductance.) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 141.214.17.222 Downloaded on 2015-03-09 to IP ABSTRACT Single-crystal SiC etching in an inductively coupled plasma (ICP) chamber using a gas mixture of 02/CF4 has been investigated for the first time. The dependence of etch rate on JCP power, substrate dc bias, and oxygen percentage was studied. It is found that etch rate increases with ICP power and substrate dc bias, and an etch rate exceeding 200 nm/mm can be obtained at -100 V substrate dc bias. The etch rate is as high as 20 nm/mm, even when there is no substrate bias. Clean and smooth surfaces can be obtained readily. Trenches with different depths were etched and their profiles were examined by scanning electron microscopy.

The robustness of a low thermal budget surface preparation method for selective silicon epitaxial growth has been investigated. After the HF dip, the stability of hydrogen passivation on Si(100) in deionized water and air has been studied. No significant increase was observed in oxygen and carbon coverage for deionized water rinse times varying from 10 to 1000 s. On wafers exposed to air for up to 10,000 s, carbon coverage on Si(100) stayed at the same level, whereas the oxygen coverage increased steadily. An in situ clean at 800°C for 10 s reduced the interfacial oxygen below the secondary ion mass spectroscopy detection levels on wafers that had been contaminated by exposure to air for up to 1000 s. In situ cleaning was studied in ambients with different partial pressures of intentionally introduced oxygen and nitrogen backgrounds. Oxygen was removed from Si(100) during the in situ clean for nitrogen partial pressures up to 1×10−6 normalTorr . When the oxygen partial pressure is sufficiently high false(1×10−6 normalTorrfalse) , oxide removal was not complete after in situ cleaning. There was no observable increase in the surface roughness for samples annealed in oxygen partial pressure up to 1×10−5 normalTorr . Hydrogen passivation was removed from the substrates and the surfaces were exposed to vacuum at room temperature for different times. After 10,000 s, the oxygen coverage was less than 2% of a monolayer. The carbon contamination on the surface was instantaneous and no additional carbon accumulation on the surface was observed up to 10,000 s. There was no apparent increase in the defect density for these wait times. © 1999 The Electrochemical Society. All rights reserved.