Network former mixing effects in ion-conducting lithium borotellurite glasses: Structure/property correlations in the system (Li2O)y[2(TeO2)x(B2O3)1 − x]1 − y

Ternary lithium borotellurite glasses of composition(Li2O)y[(2TeO2)x-(B2O3)1 − x](1 − y) (y = 0.33 and 0.40; and 0 ≤ x ≤ 1) reveal a negative network former mixing effect, implying that their electrical conductivities and ionic mobilities are lower than those expected by linear interpolation between their binary endmembers. For the y = 0.33 system, the room temperature ionic conductivity decreases from ~ 10− 8 (Ω·cm)− 1 to 10− 11 (Ω·cm)−1, while the activation energy increases from ~ 0.7 to 1.0 eV with increasing tellurium content. The structural foundations of the network former mixing effect have been probed by multinuclear solid state NMR spectroscopy. Static 125Te NMR spectra can be modeled by considering inhomogeneous broadening due to the chemical shift anisotropy effects, subject to Gaussian distributions of principal tensor values. The lineshape parameters are only weakly dependent on glass composition. 11B magic angle spinning (MAS)-NMR studies reveal that the fraction of four-coordinate boron species, B(4), increases moderately with increasing tellurium content. Overall, the data suggest that the concentration of heteroatomic connectivities is relatively low, suggesting chemical segregation and/or incipient phase separation into binary lithium tellurite and lithium borate units at the atomic- or the nanoscale. Static 7Li NMR spectra at low temperatures are governed by the strength of homonuclear 7Li7Li and heteronuclear 7Li11B/7Li10B magnetic dipole-dipole interactions. Considering the second moments for the dipolar magnetic interactions in the lithium-diborate model compound, the second moments of the 7Li NMR lineshapes could be successfully modeled by a random spatial distribution of the Li+ ions. Phenomenological modelling of the motional narrowing effects observed with increasing temperature results in a consistent description of the negative network former mixing effects in terms of activation energies and dynamic heterogeneities. The negative network former mixing effects is stronger in the system with lower lithium ion content (y = 0.33).