Mobilization of sodium in SiO2films by ion bombardment

Abstract

As part of an investigation of the mobilization and movement of ions in Si${\mathrm{O}}_2$ films resulting from bombardment by ions in the 5-2000 eV energy range, we have used radioactive ${}_{}{}^{22}{}_{}{}^{}\mathrm{Na}$ to follow the movement of sodium. The sodium (1-at.% ${}_{}{}^{22}{}_{}{}^{}\mathrm{Na}$) was initially on the surface of the Si${\mathrm{O}}_2$ at a concentration of ∼2 × ${10}^{13}$ sodium atoms/${\mathrm{cm}}^2$. We have shown (a) that ion bombardment of Si${\mathrm{O}}_2$ films on Si by many ions of interest in device processing or testing (${\mathrm{Ar}}^{+}$, ${\mathrm{N}}^{+}$, ${\mathrm{N}}_2^{+}$, ${\mathrm{Ne}}^{+}$, ${\mathrm{He}}^{+}$) causes sodium drift away from the oxide surface to the Si${\mathrm{O}}_2$/Si interface, with a small percentage remaining in the bulk, (b) that nearly all the impurity sodium on the outside surface of an oxide film can be drifted into the oxide as a result of ion bombardment, given high enough dose [∼ ${10}^{16}$ ${\mathrm{cm}}^{-2\mathrm{}}$ for the ion in (a)], (c) that the transfer of sodium away from the oxide surface at a given ion dose is energy dependent, peaking for ${\mathrm{Ar}}^{+}$ ions around 500 eV, but still remaining measurable down to 5 eV and up to 2 keV, (d) that ion implantation of ${\mathrm{B}}^{+}$ and ${\mathrm{P}}^{+}$ or ${\mathrm{Si}}^{+}$ ions at 2 keV does not cause sodium movement in Si${\mathrm{O}}_2$, (e) that the amount of sodium moved through an oxide film by bombardment increases linearly with ion dose over the dose range 1 × ${10}^{12}$ to 1 × ${10}^{16}$ ${\mathrm{cm}}^{-2\mathrm{}}$ for ${\mathrm{Ar}}^{+}$ ions, (f) that the drift of sodium away from the oxide surface to the Si${\mathrm{O}}_2$/Si interface is independent of dose rate, (g) that the flatband shifts in bias-temperature stressing of metal-oxide-semiconductor samples subjected to ion bombardment correlate well with the amount of sodium transferred to the Si${\mathrm{O}}_2$/Si interface seen in the tracer experiments, (h) that the sodium movement results observed corroborate the hypothesis that the dominant charge-transfer and drift process is related to the ion-neutralization processes at the oxide surface and, (i) therefore, all experiments or processes in which ions are impacted on insulator films should be carefully examined for harmful effects due to the sodium-mobilization effects described here. These include backsputtering, ion microprobe analysis, and ion scattering spectrometry.