Leaching of borosilicate glasses. I. Experiments

The corrosion of sodium borosilicate glasses xNa(2)O-yB(2)O(3)-(1-x-y)SiO2 is studied at constant temperature and pH as a function of the glass composition for 0.10 < x approximate to y < 0.25. Two series of experiments are reported with different glass surface area to solution volume ratios (S/V = 100 and 2000 m(-1)). The dissolution kinetics are monitored by following the concentrations of the three cations (Si, B and Na) in solution as a function of time. The altered glasses are characterized by NMR and SAXS. For all glasses, the concentrations of released cations reach a stationary limit. The time required to reach the stationary state decreases from 40 days to a few hours when the content in boron and sodium oxides is increased from 10 to 25 mol.%. The final leached fraction of silicon, LF(Si), is controlled by an apparent silica saturation concentration, which is nearly independent of the initial glass composition. On the contrary, the final leached fraction of boron and sodium, LF(B) and LF(Na), which are always close to each other, show a strong compositional dependence. The LF(B) and LF(Na) are low and nearly equal the LF(Si) for the glasses with boron (and sodium) oxide content less than 15%. This is the mark of a nearly congruent dissolution for these glasses. When the boron (and sodium) oxide content increases from 15% to 20%, there is a sharp increase of LF(B) and LF(Na) in comparison to LF(Si). Above 20%, the LF(B) and LF(Na) saturate to unity, which means that all the soluble cations have been removed, whereas the dissolved fraction of silicon may be less than 10% (depending on the S/V ratio). Si-29 NMR shows a considerable reconstruction of the altered glasses. SAXS puts in evidence the formation of a porous network with a typical pore diameter of 5 nm. The results are qualitatively explained through the competition between (i) the extraction of the soluble boron and sodium, which initiates the formation of a porous network at the nanometer scale, and (ii) the recondensation of silica, which makes the surface layer more polymerized and more resistant than the original glass at the atomic scale. (C) 2004 Published by Elsevier B.V.