HIGH-PRESSURE PHASE-TRANSFORMATIONS IN A NATURAL CRYSTALLINE DIOPSIDEAND SYNTHETIC CAMGSI2O6 GLASS

A phase transformation study has been carried out on a natural crystal line diopside and a synthetic CaMgSi2O6 glass at pressures of up to 34 .5 GPa and 30 GPa, respectively, at approximately 1000-degrees-C in a diamond-anvil cell in conjunction with a YAG laser-heating system. On the basis of X-ray diffraction data obtained from samples under in sit u pressure conditions and those quenched and unloaded to the ambient c onditions, we found that the crystalline diopside breaks down into Mg2 SiO4 (spinel) + SiO2 (stishovite) + CaSiO3 (perovskite) at approximate ly 17 GPa; at higher pressures, while the CaSiO3 (perovskite) remains stable, the Mg2SiO4 (spinel) and SiO2 (Stishovite) recombine to form M gSiO3 (ilmenite) at approximately 22 GPa which, in turn, transforms in to an orthorhombic MgSiO3 (perovskite) at pressures above 24 GPa. We h ave also found that the CaMgSi2O6 glass is transformed directly into c ubic (Ca,Mg)SiO3-perovskite at pressures higher than 13 GPa. These exp erimental results demonstrate that a different starting material (crys talline or glassy) used in a high pressure and temperature study indee d leads to different high pressure phase(s), thus reconciling the exis ting controversy on the phase transformation in diopside. On the basis of the pressure-volume data obtained for the cubic (Ca,Mg)SiO3 perovs kite, it is also suggested that the cubic (Ca,Mg)SiO3 perovskite, alth ough kinetically favorable to form from the CaMgSi2O6 glass, is not a thermodynamically stable phase with respect to the assemblage of the o rthorhombic MgSiO3 and the cubic CaSiO3 perovskites at pressures betwe en 10 and 30 GPa. Another interesting finding in this study is that in the case of the natural diopside, stishovite is present with MgSiO3 ( ilmenite) + CaSiO3 (perovskite) at 22-24 GPa, and with MgSiO3 (perovsk ite) + CaSiO3 (perovskite) at pressures above 24 GPa. In view of the p revious results on pure synthetic diopside in which no stishovite was observed at pressures above 22 GPa, this intriguing result is most lik ely due to the trace amount of other cations (i.e. Fe, Mn, Na, Al, and Ti) present in the sample. The mechanism of precipitating the stishov ite phase from ilmenite or perovskite(s), or both, is yet unknown, but the results suggest positively that MgSiO3 (ilmenite or perovskite) o r CaSiO3 perovskite, or both, are non-stoichiometric and deficient in silica.