Ca-Fe-Mg olivines: phase relations and a solution model

Abstract

Reversed phase equilibrium experiments in the system (Ca, Mg, Fe)₂SiO₄ provide four tielines at P≪1 bar and 1 kbar and 800° C–1,100° C. These tielines have been used to model the solution properties of the olivine quadrilateral following the methods described by Davidson et al. (1981) for quadrilateral clinopyroxenes. The discrepancy between the calculated phase relations and the experimentally determined tielines is within the uncertainty of the experiments. The solution properties of quadrilateral olivines can be described by a non-convergent site-disorder model that allows for complete partitioning of Ca on the M2 site, highly disordered Fe-Mg cation distributions and limited miscibility between high-Ca and low-Ca olivines. The ternary data presented in this paper together with binary solution models for the joins Fo-Mo and Fa-Kst have been used to evaluate two solution parameters:

$$\begin{gathered} F^0 \equiv 2(\mu _{{\rm M}o}^0 - \mu _{{\rm K}st}^0 ) + \mu _{Fa}^0 - \mu _{Fo}^0 = 12.660 (1.6) kJ, \hfill \\ \Delta G_*^0 \equiv \mu _{{\rm M}gFe}^0 + \mu _{FeMg}^0 - \mu _{Fo}^0 - \mu _{Fa}^0 = 7.030 (3.9) kJ. \hfill \\ \end{gathered} $$

.

Ternary phase quilibrium data for olivines tightly constrain the value of F⁰, but not that for ΔG ⁰_* which describes nonideality in Fe-Mg mixing. From this analysis, we infer a function for the apparent standard state energy of Kst:

$$\begin{gathered} \mu _{{\rm K}st}^0 = - 102.79 \pm 0.8 - (T - 298)(0.137026) \hfill \\ + (T - 298 - T1n(T/298))(0.155519) \hfill \\ + (T - 298)^2 (2.8242E - 05)/2 \hfill \\ + (T - 298)^2 (2.9665E + 03)/(2T(298)^2 ) kJ \hfill \\ \end{gathered} $$

where T is in Kelvins and the 298 K value is relative to oxides.