The formation of titanium silicide grown on the arsenic or boron heavily doped silicon substrate with cleaning method was studied using X-ray diffraction, auger electron spectroscopy, transmission electron microscopy, and secondary ion mass spectroscopy. Experiments were divided cleaning method in two types: HF cleaning and in-situ RF cleaning method. The growth of titanium silicide was enhanced when the silicon substrate was cleaned in-situ RF cleaning method. In the first place, titanium film was deposited to confirm the existence of native oxide(< 20Å) after HF cleaning at room temperature. To remove native oxide, in-situ RF cleaning before titanium deposition was carried out. Cross-sectional transmission electron microscopy revealed that the native oxide was formed at the interface between silicon and as-sputtered titanium film during room temperature sputtering. Both cleaning condition samples were found to have amorphous TiSix layer when titanium was deposited on undoped-silicon at 400℃ temperature. Amorphous TiSix layer for HF cleaning sample was found to have larger thickness(20%) than that of in-situ RF cleaning condition.
Our research would like to attention on the relation between dopant and cleaning methods. In case of arsenic redistribution, sheet resistance of titanium silicide was reduced below ~ 3.7 Ω/㎠ which was about 10 times less than the conventional HF cleaning method. Transmission electron microscopy micrographs showed as follows. In-situ RF cleaning condition sample showed that had 550 C54-$TiSi_2$ thickness and smooth interface. But HF cleaning sample had discontinuous and local TiSi2 growth and uncompleted C54-$TiSi_2$ transformation. Depth profile of arsenic showed that anomalous arsenic distribution of arsenic in $TiSi_2$ film when silicidation using Ti 270 is carried out. High concentration of arsenic inside $TiSi_2$ is attributed to the native oxide formed by HF cleaning. It was seemed that native oxide suppressed the diffusion of arsenic. In-situ RF cleaned sample had low arsenic concentration in $TiSi_2$ layer after 850℃, 30sec RTA. It was speculated that titanium silicide was well formed in low arsenic concentration in $TiSi_2$. High concentration of arsenic in $TiSi_2$ film showed that native oxide in HF cleaning sample filled the role of barrier of diffusion of arsenic. Hence, in-situ RF cleaning should be chosen to form uniformed titanium silicide as well as to reduce sheet resistance. These results were similar to samples using the boron-implanted silicon substrate. It has been studied that boron rarely affected $TiSi_2$ formation. But HF cleaned samples seemed that the native oxide made boron diffuse faster. Depth profile of boron in HF cleaning sample showed that boron concentration in $TiSi_2$ was higher than in-situ RF cleaned sample. Transmission electron microscopy micrographs revealed that HF cleaning sample had rougher interface and smaller $TiSi_2$ thickness than in-situ RF cleaning sample. But both samples showed no difference in sheet resistance and X-ray diffraction. In our research, the important point to note is native oxide makes interface rougher than that of in-situ RF cleaning and impedes diffusion of dopant or silicon.
