Stategy for Locating Potential Sites for Hydrogen in Minerals

Neutron diffraction is ideally suited for determining structural positions of hydrogen in minerals and usually the sample is deuterated in order to reduce the incoherent scattering from the hydrogen. In many cases, however, difficulties in location of hydrogen in minerals by neutron diffraction may arise because the there may be too little hydrogen in the sample to detect, the sample cannot be deuterated, there may be deterioration in the quality of diffraction data collected in pressure cells, etc. We propose a strategy that will aid in both the determination of potential docking sites of hydrogen and in the determination of the crystallographic orientation of O-H bonds. The strategy is based on determining the Laplacian of the electron density, -∇ ^2ρ (r), of the mineral. As demonstrated by Bader et al. (1984), a mapping of the Laplacian of the electron density distribution yields a series of concentric shells centered at the nucleus of each atom defining where the electron density distribution, ρ, is alternately locally concentrated and locally depleted, a distribution that reflects the shell structure of the atom. The region where the distribution is positive is called the valence-shell charge concentration (VSCC) of the atom (Bader, 1990). When two atoms combine and a bond is formed, the VSCC of each atom is distorted to one degree or another with the concomitant formation of maxima and minima in the VSCC of each atom. The maxima define domains where Ÿƒ is locally concentrated and the minima define domains where ρ is locally depleted. It has been found that the number, the location and the relative sizes of the maxima provide a faithful representation of the bonded and non-bonded electron pairs of the Lewis model. The non-bonded electron pairs correspond with sites of potential electrophilic attack. Gibbs et al. (2002) has extended this approach to minerals and predicted potential sites for hydrogen in coesite. The theoretical results agree very well with Koch-Mueller et al.­Ýs (2001) infrared spectroscopic study of H-doped coesite. We present further results here for other high-pressure silicates, including stishovite, periclase, wadsleyite, ringwoodite, akimotoite, and MgSiO₃ perovskite, and compare the results with available experimental data. In general, the correspondence between the theoretical predictions and experimental data is good. We propose that the combined theoreticial and experimental approach provides a powerful tool for understanding the role of hydrogen in minerals.