Insights into the structure of alkali borate glasses

In the light of extensive new glass transition temperature (T_g) and density (ρ{variant}) data obtained for the ternary glass system Li₂O(LiCl)₂B₂O₃ we have re-examined the well accepted structural models developed for borate glasses by Krogh-Moe and co-workers. We find that, in contrast to the conclusions of NMR, Raman and IR spectroscopy, glasses in the binary and in the ternary systems with common O/B ratios do not exhibit identical boron-oxygen networks. Rather, we find that as LiCl in the ternary is increased the network is systematically weakened and expanded to accommodate the large Cl^- anion. On the basis of these observations we question the ability of spectroscopic techniques such as Raman and NMR to confirm the uniqueness of structural models based on combinations of intermediate-range units. We propose a tentative model which is qualitatively in agreement with the observed changes in T_g and ρ{variant} upon addition of LiCl, and explains how the large Cl^- ions can be accommodated without large increases in energy due to strain and electrostatic effects while maintaining BO₄/BO₃ constant. Recent transport results on some fast ion conducting glasses in this system are also discussed in terms of the proposed model.