AVS 56th International Symposium & Exhibition | |
Thin Film | Tuesday Sessions |
Session TF2-TuA |
Session: | ALD/CVD: Oxides and Barriers |
Presenter: | R. Methaapanon, Stanford University |
Correspondent: | Click to Email |
In this work, TiO2 ALD is carried out in an integrated ALD reactor/UHV chamber that allows for X-ray photoelectron spectroscopy (XPS) analysis after each precursor pulse without vacuum break. Titanium tetrachloride (TiCl4) and water (H2O) are selected as precursors due to their molecular simplicity and broad operating temperature range that result in several achievable TiO2 phases. The initial growth at 100oC on two substrates − chemical oxide on silicon prepared by piranha treatment, and hydrogen-terminated silicon prepared by HF etch − is compared. The intensities and binding energies of characteristic peaks from the XPS spectra are used to analyze the elemental compositions and chemical state of each species as the deposition progresses. TiO2 growth on both SiO2 and H-Si surfaces exhibits linear behavior, as normally achieved by ALD, but the TiO2 growth rate is lower on hydrogen-terminated surface than on silicon dioxide surface. Interestingly, no incubation period is observed on either surface.
The chemical shifts of the Si 2p, O 1s and Ti 2p XPS peaks after TiO2 deposition on the SiO2 substrate suggest bond formation between titanium and silicon-bound oxygen at the interface. The data also suggest that some chlorine is trapped at the SiO2/TiO2 interface and that the titanium oxide right at the interface is sub-stoichiometric. The results on the hydrogen-terminated Si surface show different interfacial properties. There is no detectable amount of oxidized silicon species on hydrogen-terminated silicon after deposition under vacuum. Together with the results of ex situ studies, it can be concluded that interfacial silicon dioxide grows after air exposure, not during ALD reactions. The absence of silicon oxide and a shift in the Si 2p binding energy in the as-grown samples suggest the possibility of an ALD mechanism which involves direct bonding between titanium and silicon on the surface. The differences between the two substrates will be discussed.