The evidence for the occurrence of two successive transitions in Al2O3 from the analysis of Hugoniot data

We calculated the Hugoniots of Al2O3 to 400 GPa using a thermoelastic model developed previously and compared the results with measurements. The P-H-rho(H) data and U-u data for crystalline corundum by McQueen and Marsh agree well with our results calculated by the elastic-plastic model to 130GPa, while Marsh's later data lie in between ours obtained by the elastic-plastic model and those obtained by the elastic- isotropic model to 130GPa. Above 130GPa, the P-H-rho(H) data of the two sets of measurements begin to depart towards the higher-density side. On the other hand, the Hugoniot data of Mashimo et al to 104GPa agree well with ours obtained by the elastic -isotropic model, suggesting rapid relaxation of shear stress. The first two experiments (flash-gap method) revealed no phase transitions, but the third one (inclined-mirror method) showed a phase transition at 79.3 GPa. The discrepancy can be largely attributed to the measurement method. Incorporating the theoretical values of the elastic quantities and cohesive energy in the static lattice at zero pressure for the high-pressure Rh2O3(II) and Pbnm-perovskite phases predicted by Thomson et al in our model, we have calculated the Hugoniots of Al2O3 by using the three-phase model. Comparison of our results with Erskine's Hugoniot data to 340GPa suggests that Al2O3 undergoes a further transition at similar to250GPa in agreement with the theoretical static-lattice value (223 15 GPa). The calculated temperatures are similar to773 K and similar to3427K for the respective transitions. The unexpectedly low slope of the U-u curve observed by Erskine can be attributed to the fact that the occurrence of the transitions was not considered in his analysis. The calculated Hugoniots of porous Al2O3 with porosity up to 14% are also in good agreement with experiments.