To constrain the mixing properties of (Ca,Fe,Mg)
2SiO
4 olivine, Fe-Mg exchange experiments have been carried out on the assemblages of low-Ca olivine+clinopyroxene, low-Ca olivine+high-Ca olivine+clinopyroxene and high-Ca olivine+clinopyroxene with bulk Fe/(Fe+Mg) in the range of 0.4 to 0.6, mainly about 0.5, at pressures from 1 atm to 2.0 GPa and temperatures from 1100°C to 1300°C, mainly at 1 atm and 1150°C. The phase relations of (Ca,Fe,Mg)
2SiO
4 olivine have been also determined at 1 atm and 1150°C. Experiments under 1 atm were performed in an electric furnace equipped with CO
2/H
2 control device. The oxygen fugacity around the sample space in the furnace was kept to that realized with the iron-wüstite buffer. High-pressure experiments were carried out using a piston-cylinder apparatus. The Fe-Mg distribution coefficient between olivine and clinopyroxene,
KD [=(
XMgCpx XFeOl)/(
XFeCpx XMgOl)], increases with bulk Ca/(Ca+Fe+Mg) content, but insensitive to pressure. The solvus between low-Ca and high-Ca olivines becomes narrow with pressure and Fe/(Fe+Mg) of the system, but the pressure effect on the solvus is very small.
Thermodynamic analyses were made on the present and previous data to refine the compositional dependence on
KD of olivine-clinopyroxene pair and the solvus of quadrilateral (Ca,Fe,Mg)
2SiO
4 olivine using the Fe-Mg mixing data, and the free energy changes of the Fe-Mg exchange reactions between these phases and in olivine. The symmetric and asymmetric two-site regular solution models were applied to (Ca,Fe,Mg)
2SiO
4 olivines in terms of the Fe-Mg intracrystalline exchange between M1 and M2 sites and the excess energies of Ca-Fe-Mg mixing. In the present model, it is assumed that substitution of Ca is limited to M2 sites alone whereas Fe and Mg substitute on both M1 and M2 sites. The mixing parameters of the asymmetric Ca-Mg and symmetric Fe-Mg and Ca-Fe models are evaluated as
WCaMg=36.31±0.19,
WMgCa=32.73±0.18,
WCaFe=21.34±0.06,
WCaFeMg=6.12±0.98, δ
1=2.24±1.30 and δ
2=−8.45±0.94 (unit in kJ), whereas the symmetric Ca-Fe-Mg mixing model (
WCaFeMg=0) resulted as
WCaMg=
WMgCa=34.20±0.18,
WCaFe=23.75±0.58, δ
1=−1.50±2.35 and δ
2=−10.20±1.37. The present results are consistent with the previous analyses of miscibility gaps in Ca-Fe-Mg olivines. The Ca-Mg mixing is slightly less non-ideal than the previous estimates on the forsterite-monticellite solvus. Considerations about the Fe-Mg intracrystalline exchange show a slight preference of Fe in M1 site in Ca-free olivines, and indicate preferential order of Fe to M2 site with increase of Ca contents of olivine.
View full abstract