Molecular dynamics simulations of silicate and borate glasses and melts : structure, diffusion dynamics and vibrational properties

Molecular dynamics simulations of silicate and borate glasses and melts: Structure, diffusion dynamics and vibrational properties. In this work computer simulations of the model glass formers SiO2 and B2O3 are presented, using the techniques of classical molecular dynamics (MD) simulations and quantum mechanical calculations, based on density functional theory (DFT). The latter limits the system size to about 100−200 atoms. SiO2 and B2O3 are the two most important network formers for industrial applications of oxide glasses. Glass samples are generated by means of a quench from the melt with classical MD simulations and a subsequent structural relaxation with DFT forces. In addition, full ab initio quenches are carried out with a significantly faster cooling rate. In principle, the structural properties are in good agreement with experimental results from neutron and X-ray scattering, in all cases. A special focus is on the study of vibrational properties, as they give access to low-temperature thermodynamic properties. The vibrational spectra are calculated by the so-called ”frozen phonon” method. In all cases, the DFT curves show an acceptable agreement with experimental results of inelastic neutron scattering. In case of the model glass former B2O3, a new classical interaction potential is parametrized, based on the liquid trajectory of an ab initio MD simulation at 2300 K. In this course, a structural fitting routine is used. The inclusion of 3-body angular interactions leads to a significantly improved agreement of the liquid properties of the classical MD and ab initio MD simulations. However, the generated glass structures, in all cases, show a significantly lower fraction of 3-membered planar boroxol rings as predicted by experimental results (f=60%-80%). The largest boroxol ring fraction of f=15±5% is observed in the full ab initio quenches from 2300 K. In case of SiO2, the glass structures after the quantum mechanical relaxation are the basis for calculations of the linear thermal expansion coefficient αL(T), employing the quasi-harmonic approximation. The striking observation is a change change of sign of αL(T) going along with a temperature range of negative αL(T) at low temperatures, which is in good agreement with experimental results.

Lead 210 and silicate profiles from sediments of the equatorial Pacific and South Atlantic

Particle mixing rates have been determined for 5 South Atlantic/Antarctic and 3 equatorial Pacific deep-sea cores using excess 210Pb and 32Si measurements. Radionuclide profiles from these siliceous, calcareous, and clay-rich sediments have been evaluated using a steady state vertical advection diffusion model. In Antarctic siliceous sediments210Pb mixing coefficients (0.04-0.16 cm**2/y) are in reasonable agreement with the 32Si mixing coefficient (0.2 or 0.4 cm**2/y, depending on 32Si half-life). In an equatorial Pacific sediment core, however, the 210Pb mixing coefficient (0.22 cm**2/y) is 3-7 times greater than the 32Si mixing coefficient (0.03 or 0.07 cm**2/y). The difference in 210Pb and 32Si mixing rates in the Pacific sediments results from: (1) non-steady state mixing and differences in characteristic time and depth scales of the two radionuclides, (2) preferential mixing of fine-grained clay particles containing most of the 210Pb activity relative to coarser particles (large radiolaria) containing the 32Si activity, or (3) the supply of 222Rn from the bottom of manganese nodules which increases the measured excess 210Pb activity (relative to 226Ra) at depth and artificially increases the 210Pb mixing coefficient. Based on 32Si data and pore water silica profiles, dissolution of biogenic silica in the sediment column appears to have a minor effect on the 32Si profile in the mixed layer. Deep-sea particle mixing rates reported in this study and the literature do not correlate with sediment type, sediment accumulation rate, or surface productivity. Based on differences in mixing rate among three Antarctic cores collected within 50 km of each other, local variability in the intensity of deep-sea mixing appears to be as important as regional differences in sediment properties.