Theoretical Approaches to Structure and Spectroscopy of Earth Materials

Abstract

The characterization of complex materials in terms of their structure, electronic, magnetic, vibrational, thermodynamic or other physical and chemical properties is often a challenging task that requires the combination of a number of complementary techniques. Experimental approaches such as diffraction or spectroscopic methods usually provide fingerprint information about the material under investigation. The interpretation of measured data is either done by reference to analogue materials or by constructing a theoretical (e.g., structural) model that fits the experimental data. For the latter, computational methods have become very powerful in recent years. For example, Rietveld refinement of powder diffraction data or curve fitting of various spectra is now done on a routine basis. The continuous improvement in hardware performance resulting in a huge and progressive increase of computing power by a factor of ~1000 per decade, as well as advanced algorithms and codes provide the basis for predictive modeling of material properties ab initio , i.e., from first principles using quantum chemical methods such as density functional theory (DFT). DFT enthalpy predictions for the major lower mantle minerals, MgSiO3 perovskite and post-perovskite, periclase (MgO) and CaSiO3 perovskite at zero temperature, over the relevant pressure range from the transition zone to the core-mantle boundary can be made in a few hours on an office PC. Free energy calculations for these phases at finite temperatures using lattice vibrational modes in the (quasi-)harmonic approximation require at most a couple of days. More realistic compositions of these mantle minerals with Fe substituting some of the Mg atoms in a solid solution are computationally more demanding but have also become accessible. The same is true for structural investigations of disordered phases, such as glasses, melts and fluids. Both first-principles and classical molecular dynamics simulations are useful methods for a statistically significant sampling of disordered structures. …