DISCRETE BOND MODEL (DBM) OF SODIUM-SILICATE GLASSES DERIVED FROM XPS, RAMAN AND NMR MEASUREMENTS

In sodium silicate glasses, the fraction of differently bound Si speci es Q[i] (i = 0-4), depending on the number i of bridging oxygens bound to the quarternary silicon, is a function of stoichiometry and the Na /Si ratio. Sodium silicate glasses were investigated by high resolutio n X-ray photoelectron spectroscopy. To explain the differences in chem ical shifts and linewidths of the 0 Is signal of the bridging and the non-bridging oxygen as a function of alkali concentration, an extended glass model was developed. This new model takes into account the infl uence of the alkali concentration on the Q[i] distribution and on the appearance and concentrations of differently bound bridging oxygens as well as non-bridging oxygens. In principle, four different bonds for the non-bridging oxygens (Na-O-Q[i], i = 0-3) and seven different bond s for the bridging oxygens (Q[i]-O-Q[j], i, j = 1-4, \i - j\ < 2), lin king the different Q[i] species could exist. According to this 'discre te bond model', for each glass composition the number of these possibl e bridging and non-bridging oxygen species is limited to a maximum of two for each of them. An approximation of the measured X-ray photoelec tron 0 Is signal by a superposition of 0 Is signals as calculated on t he basis of this model allows an excellent reproduction of the experim ental results if the glasses are non-phase-separated. In the case of p hase separation, the measured spectrum shows only Q[i]-O-Q[i] bonds (w ith i = 1-4) identical to the bonding behaviour in the corresponding w ell-known crystalline phases. The chemical shifts of the non-bridging oxygens relative to the bridging oxygens (as well as the absolute bind ing energies), as measured by X-ray photoelectron spectroscopy, can be explained by changes in the relative concentrations and intensities o f the different oxygen bonds as a function of glass composition. The e nergy differences between two energetically neighbouring bridging oxyg ens (i.e., Q[i]-O-O[i] and Q[i]-O-Q[j] with \i - j\ = 1) are equal (wi thin the experimental error) for all bridging oxygen bonds in sodium s ilicate glasses and can be correlated with the cation field strengths of the modifying cations.