| Abstract: | The olivine-spinel phase transformation in Mg2GeO4 does not occur by a martensitic mechanism. The evidence, from samples transformed in a Griggs-type solid medium deformation apparatus, are: 1. (1) lack of microstructural features in the olivine phase which can be specifically associated with a martensitic mechanism 2. (2) the orientation relationship between the two phases that is predicted by the martensitic mechanism does not occur nor is there any apparent consistency of relative orientations 3. (3) application of a differential stress to the transforming sample resulted in an anisotropic growth rate for the spinel phase indicating that growth was externally controlled rather than crystallographically controlled. Anisotropic growth of the spinel phase results in elongation of the residual olivine phase grains in the plane normal to the direction of maximum principal compressive stress. A velocity ratio of 1.7−0.7+5.4 has been determined for the growth rate of the spinel from measurements on residual olivine grains. The interphase grain boundary in samples transformed under stress has cusp-shaped fingers of spinel with a blunt end separated by thin spikes of olivine. Samples transformed isostatically do not exhibit this feature providing further confirmation of anisotropic growth of the spinel. The preferred growth of the spinel is consistent with a theory of phase transformation under nonhydrostatic stress. The predicted spinel finger shape based on this theory is generally consistent with observed shapes except for the blunt end. The discrepancy may be due to surface energy which has not been considered here, or to local deviations of the applied macroscopic stress. |
