Ionic Conduction and Bond Valence in Glasses

Doctoral thesis, 2008

Ion conducting glasses are highly interesting for several different applications, among them, chemical sensors, smart windows with variable optical properties and as electrolytes in batteries. The last example can take full advantage of the ease with which a thin film (< 1 μm) of g lass electrolyte can be sputtered onto a substrate and thereby provide power for micro-scale electronics, e.g., a chip on so called smart cards. How ever, despite intense research there is still no fully satisfactory model for the atomic mechanics governing the ion conduction in these materials. In many promising materials (e.g. nitrated phosphate films) we have not even access to good models of their structure. The aim of this thesis is to shed some light on both the structure and ionic properties of ionic oxide glasses. The structure is determined by a combination of neutron and X - ray diffraction experiments together with the reverse Monte Carlo technique, with which we can create structural models of the investigated g lasses. These structural models can then discussed in terms of pair correlations, topology and possible structural inhomogeneities. Further more, the application of the bond valence method to the structures yields vivid representations of the ionic path ways through the structures. From these, not only estimations of the activation energy and conductivity can be extracted, but also important information on coordination of mobile ions and pathway topography can be gained. These techniques have here been used for investigations of two different phenomena. Firstly, when a network modified oxide glass, such as M2O-B2O3, is doped by a metalhalide salt, MX, a strong increase in conductivity can be seen. Earlier structural studies have shown how salt-rich channels through the oxide network is formed. Our investigations show the distinctly different pathways in differently doped systems. Further more, we show that ordering of the glass beyond the nearest neighbors play little role in ion conduction. Secondly, the ever elusive mixed mobile ion effect (M M IE) has been placed under scrutiny. The strong M M IE in a mixed Li/Rb meta-phosphate has been investigated in term s of the pathway topography and the blocking effect at room temperature and near the glass transition. A study of the weak MMIE in a Ag/Na meta-phosphate system reveals the similarities in pathways for the two types of ions. In this system we find no blocking effect and propose the existence of cooperative Ag-Na pairs in the pathways.