Structure and properties of crystalline inclusions trapped in minerals

This thesis deals with the characterization of mineral inclusions by means of various non-destructive techniques. Two type of inclusions are analyzed: inclusions in diamonds and inclusions in metamorphic rocks, in particular the host-inclusion pair quartz in garnet. The work on inclusion in diamonds focuses on rare inclusions of magnetic minerals, such as iron oxides, and employs a multi-analytical approach: X-ray diffraction, Raman spectroscopy, magnetometry, and X-ray tomography. Magnetic properties can help in the identification of the composition of the inclusions and thus of the environment in which the host-inclusion system grew. X-ray tomography was employed to locate the inclusions in the samples, asses the presence of fractures, and support the identification of the phases. The characterization of the quartz-in-garnet pair is focused on the determination of the stress state of the inclusions and the influence of such stress on the structure and properties of the system. It is shown how to characterize the crystal structure of inclusions in-situ by means of X-ray diffraction. A thorough characterization of the polarized Raman scattering of quartz as a function of pressure and temperature was performed to guide the interpretation of the results from inclusions. The high-pressure Raman experiment verifies the validity of the approach based on the phonon-mode Grüneisen tensor to calculate the pressure in the inclusions, also for the E modes. Furthermore, it points out that strong multiphonon interactions can contribute to the stability of alpha-quartz at ambient conditions and provides new pressure calibrations. The heating experiment performed in situ by Raman spectroscopy shows that the quartz inclusion in garnet does not undergo the alpha-beta phase transition. The comparison of the free and trapped crystal data clearly shows a different response to heating and the applied models cannot reproduce the experimental data. It can be shown that the disagreement between the data and the prediction is due to the elastic anisotropy of quartz, suggesting that at a new model taking into account elastic anisotropy is needed.