Modelling configurational entropy of silicate melts

The Adam-Gibbs theory provides a robust connection between the transport or relaxation properties of melts and their thermochemical properties. In its expanded form: log eta = A + B/T[S-c(Tg) + Cp-c in (T/T-g)] the equation has adjustable unknown parameters A, B and Sc (Tg) which can be estimated from experimental estimates of configurational heat capacity (Cp-c), glass transition temperature (Tg) and viscosity (eta). Here, we use recently published datasets for anhydrous and hydrous silicate melts and glasses (N similar to 50) for which there are measurements of log eta and calorimetric measurements of Cp-c and Tg. Our fitting strategy follows the approach developed by previous workers with the sole exception that we assume all silicate melts converge to a common, but unknown, high temperature limit to melt viscosity (e.g., A = log eta(infinity)). Our optimal value forA is -3.51 +/- 0.25. A consequence of a common, high-temperature limit to silicate melt viscosity is that the corresponding model values of glass transition temperature (Tg(12)), melt fragility (m), and the ratio Cp-c/S-c are constrained to lie on a single plane approximated as: Cp-c/S-c = Tg(12)/243399 - m/15.518 + 0.996 thereby establishing a quantitative connection between calorimetric and rheological measurements. Lastly, we show a good correspondence between values of Tg(12) and fragility (m) from this Adam-Gibbs based model of melt viscosity and values predicted by the GRD viscosity model for multicomponent silicate melts (cf. Giordano et al., 2008). (C) 2016 Elsevier B.V. All rights reserved.