Penetration of fluorine into the silicon lattice during exposure to F atoms, F₂, and Xe F₂: Implications for spontaneous etching reactions

The purpose of this article is to present evidence about the quantity and distribution of fluorine in silicon after and during spontaneous etching with F atoms, F2, and Xe F2. X-ray photoelectron spectroscopy spectra were analyzed using the method developed by Sven Tougaard designated in this paper as "ST." It is found that fluorine penetrates deeply into the silicon lattice during the etching reaction. It is shown that the surface concentration of fluorine (first 10 Å) is relatively independent of whether F atoms of Xe F2 were used and is also relatively independent of doping level. In contrast, the amount of fluorine in the silicon lattice does depend upon these parameters. It is suggested that the fluorine in the lattice exists primarily as negative ions. Fluorine on the surface (top 10-20 Å) is probably Si Fx (x=1-3). The authors data are consistent with the data and interpretation of Lo [J. Vac. Sci. Technol. A 11, 2054 (1993)] who found the surface concentration of F at saturation to be ∼1.7× 1015 F cm2 (1.7 ML, 1.1× 1015 F /ML). They found that Si F3 was the dominant species on the surface at saturation. The concentration of F- ions in the lattice is so large [∼ 1020 to 4 × 1021 cm3] that they probably neutralize holes and donor sites and also are likely to dope the system with deep lying acceptor states which lead to an increased concentration of holes at the valence band maximum. A p-n junction may be created. In addition, the total quantity of fluorine in the lattice is estimated. These new values are derived from previously published work using a more reliable calibration. They are about 2.5 times higher than the original estimate. It will also be shown that the etch properties are influenced by the negative ions in the silicon lattice. The ions adjust their concentration and depth distribution to fit the etch conditions. The time constant for this adjustment is undetermined but is many minutes at room temperature and moderate fluxes [see Fig. 15 of Ref. 3(a)].