Spectroscopic studies of oxygen speciation in potassium silicate glasses and melts

Abstract

To resolve discrepant results between O1s XPS and 29Si and 17O NMR techniques, O1s XPS spectra were collected and fitted for 15 potassium silicate glasses containing 10−37 mol% K2O. The effects of melting on the compositions of the glasses are documented and quantified, as are uncertainties associated with fitting the O1s spectra. These spectra are well resolved into two symmetric peaks: a narrow peak due to NBO (Si−O−K) plus “free oxide” (O2– or K−O−K) and a broader peak due to BO (Si−O−Si). Values of mol% BO (uncertainties of ±1%) are obtained from the computed peak areas and indicate 2.2 (±0.8) mol% O2– for glasses containing 32 (±1) mol% K2O. Reassignment of Q species for previously published 29Si NMR spectra leads to BO and O2– values in good agreement with these O1s XPS spectra. A recent 17O NMR study on potassium silicate glasses concluded that there is <1 mol% O2– in these glasses based mainly on the assumption that the 17O chemical shift for O2– should be similar in the glasses and the oxide in crystalline K2O. This assumption is questionable. Free oxide in these glasses should be highly reactive toward gases such as CO2 (via CO2 + O2– → CO32–) and it seems likely that the reactive species in silicate glasses is O2– rather than NBO as often assumed. The small amounts of CO2 found in reacted glasses are qualitatively consistent with the abundance of O2– obtained from these O1s XPS spectra of silicate glass.