V23F-2188
Structure of Amorphous Titania Nanoparticles
Ultrafine (2 - 3 nm) titania (TiO2) nanoparticles show only diffuse scattering by both conventional powder x-ray diffraction and electron diffraction. We used synchrotron wide-angle x-ray scattering (WAXS) to probe the atomic correlations in this amorphous material. The atomic pair-distribution function (PDF) derived from Fourier transform of the WAXS data was used for reverse Monte Carlo (RMC) simulations of the atomic structure of the small nanoparticles. Molecular dynamics simulations were used to generate input structures for the RMC. X-ray absorption spectroscopy (XAS) simulations were used to screen candidate structures obtained from the RMC. The structure model that best describes both the WAXS and XAS data consists of particles with a highly distorted shell and a small strained anatase-like crystalline core. The average coordination number of Ti is 5.3 and the Ti-O bond length peaks at 1.940 Å. Relative to bulk titania, the reduction of the coordination number is primarily due to the truncation of the Ti-O octahedra at the titania nanoparticle surface, and the shortening of the Ti-O bond length is due to bond contraction in the distorted shell. Core-shell structures in ultrafine nanoparticles may be common in many materials (e.g. ZnS).
V23F-2189
Nanoscale Phase Stability Reversal During the Nucleation and Growth of Titanium Oxide Minerals
Fine-grained titanium oxide minerals are important in soils, where they affect a variety of geochemical processes. They are also industrially important as catalysts, pigments, food additives, and dielectrics. Recent research has indicated an apparent reversal of thermodynamic stability between TiO2 phases at the nanoscale thought to be caused by an increased contribution of surface energy to the total free energy. Time-resolved X-ray diffraction (XRD) experiments in which titanium oxides crystallize from aqueous TiCl4 solutions confirm that anatase, a metastable phase, is always the first phase to nucleate under our range of initial conditions. Rutile peaks are observed only minutes after the first appearance of anatase, after which anatase abundance slowly decreases while rutile continues to form. Whole pattern refinement of diffraction data reveals that lattice constants of both phases increase throughout the crystallization process. In addition, transmission electron microscope (TEM) observations and kinetic modeling indicate that anatase does not undergo a solid-state transformation to the rutile structure as once thought. Instead, anatase appears to re-dissolve and then feed the growth of already nucleated rutile nanocrystals. Density functional theory (DFT) calculations were employed to model 1, 2, and 3 nm particles of both mineral phases. The total surface energies calculated from these models did yield lower values for anatase than for rutile by 8-13 kJ/mol depending on particle size, indicating that surface free energy is sufficient to account for stability reversal. However, these whole-particle surface energies were much higher than the sum of energies of each particle's constituent crystallographic surfaces. We attribute the excess energy to defects associated with the edges and corners of nanoparticles, which are not present on a 2-D periodic surface. This previously unreported edge and corner energy may play a dominant role in the stability reversal of nanocrystalline titanium oxides, as well as other mineral systems susceptible to reversals in phase stability at the nanoscale.
V23F-2190
Molecular Simulation of Copper Complexation by a Model of Dissolved Organic Matter
The coadsorption of cations and organic matter to mineral surfaces has been proposed as a mechanism to stabilize soil organic carbon and explain enhanced cation adsorption in soils. Spectroscopic investigations of coadsorption in copper (Cu)-organic ligand-mineral systems have provided conclusive evidence that ternary surface complexes do occur in instances where the simple organic ligand possesses a minimal number of homogeneous binding sites. Although the results of many experimental and spectroscopic investigations of ternary systems with dissolved organic matter (DOM) suggest that ternary surface complexes may be present, unambiguous conclusions cannot be drawn due to the nature of DOM and the heterogeneity of DOM binding sites. Molecular simulations of ternary surface complexes are powerful tools that can be used in the interpretation of spectroscopic data from these experimental systems. Molecular simulations of Cu-DOM coadsorption at mineral surfaces require a model of Cu-DOM that adequately reproduces the physical-chemical behavior of natural Cu-DOM. The Cu-saturated Schulten DOM molecule, whose charge from deprotonated carboxyl groups was balanced by Cu2+ ions, was modeled using the COMPASS forcefield and energy minimization and molecular dynamics algorithms. Compared to previous simulations of Na- and Ca-saturated Schulten DOM, the Cu-DOM was more compact and spherical, as demonstrated by the decreases in the surface area, surface area: volume ratio, porosity, and radius of gyration and by an increase in the volume. The Cu-DOM complex is more polar than Ca-DOM and less polar than Na-DOM, as indicated by the dipole moment and extent of hydrogen bonding. The Cu2+ ion coordination environment contains an average of four oxygen atoms at 2.04 Å from the metal center. These first shell atoms are primarily water and carboxylate oxygen, although oxygen atoms from carbonyl, phenolic, and esters functionalities as well as nitrogen groups are also present. The DOM model reproduces the physical properties and reactivity of natural organic matter with Cu and is therefore an appropriate model for further simulations of Cu-DOM-mineral ternary surface complexes.
V23F-2191
Slow diffusional dynamics at the mineral-water interface: Nanoscale view from the computational molecular modeling perspective
Molecular-scale knowledge of the thermodynamic, structural, and transport properties of water and ions at mineral interfaces is crucial for quantitative understanding and prediction of many geochemical and environmental processes. At mineral interfaces, individual water molecules and hydrated ions simultaneously participate in several dynamic processes characterized by different, but equally important time- and length- scales. Most of these processes are very difficult to investigate experimentally and require a broad range of sophisticated analytical techniques, but many of them can be effectively studied by molecular dynamics computer simulations in a single MD run. On a relatively short time scale (~1-100 ps), the interfacial dynamics is dominated by the molecular librational and re-orientational motions. The librations (hindered rotations) of surface hydroxyls also occur on this time scale. These motions are responsible for the reformation and breaking of individual H-bonds, and the strength of these bonds can be directly correlated with the frequencies of intra-molecular O–H vibrations at an even shorter sub-ps time scale. However, the diffusional processes related to reformation of the entire interfacial hydrogen bonding network, surface adsorption of H2O molecules and ions are characterized by a much longer time scale (~0.1-10 ns). We quantify the slow diffusional motion of H2O molecules on the surfaces of quartz and tobermorite using van Hove self-correlation functions, GS(r,t). The results clearly show a distinct characteristic time, τ ~0.8 ns for the hopping diffusion of water molecules on the surface of tobermorite, in excellent qualitative and quantitative agreement with 1H NMR field cycling relaxometry results for surface-associated water in tobermorite (Korb et al., 2007a,b; Kalinichev et al., 2007). Similar hopping diffusional dynamics is several times faster on the ideal (001) surface of quartz, but in reality can be significantly slowed down by the atomic- scale roughness of the surface. These results are also consistent with recent neutron spectroscopic data for water dynamics on rutile and cassiterite surfaces (Mamontov et al., 2007, 2008).
V23F-2192
Diffusion in Argillaceous Media: Bridging the Molecular and Laboratory Scales
The diffusion coefficients of water and solutes in clay-rich media are important parameters in contaminant hydrology, sediment and aquitard pore water geochemistry, and CO2 and high-level radioactive waste geological storage. In the present study, we report long (> 10 ns) molecular dynamics (MD) simulations of water and solutes in Na-montmorillonite interlayer nanopores. We show that the diffusion coefficients of water and solutes are strongly affected by the time-scale of measurement for time-scales of 10-12 to 10-9 s. We demonstrate for the first time that the nanopore-scale diffusion coefficients of water and solutes (sodium, strontium, cesium) in Na-montmorillonite interlayers obtained from sufficiently long MD simulations are consistent with centimeter-scale laboratory data on the apparent diffusion coefficients of water and solutes in smectite clay plugs, i.e., MD simulations of individual nanopores can provide quantitative estimates of laboratory-scale transport and cation-exchange in nanoporous media.
V23F-2193
The Hydration Sheath
Scale is intimately related to the nature of the dielectric behavior of water at an interface. The dielectric properties of interfacial water were studied above two different oxides, goethite and quartz. We examined double interface systems, (oxide|water|vacuum) including monolayer/bilayer coverage, and single interface systems (bulk water|oxide). This study employed classical molecular dynamics in order to determine correlation effects of individual waters; which, in conjunction with classical physics of interfacial systems, provided us with useful estimates of both the magnitude of dielectric shielding present at the interface and the geometric extent of perturbation from bulk water behavior at an interface. These results compare favorably with existing experimental measurements of both monolayer dielectric constants and hydration behavior around a solute. This hydration sheath is shown both here and in experimental studies to be limited to approximately one monolayer's thickness in a bulk water system. Overall, the calculations provide a consistent picture of dielectric screening, one that can play a significant role in the behavior of the electric double layer.
V23F-2194
An Automated Method of Scanning Probe Microscopy (SPM) Data Analysis and Reactive Site Tracking for Mineral-Water Interface Reactions Observed at the Nanometer Scale
Developing a method for bridging the gap between macroscopic and microscopic measurements of reaction kinetics at the mineral-water interface has important implications in geological and chemical fields. Investigating these reactions on the nanometer scale with SPM is often limited by image analysis and data extraction due to the large quantity of data usually obtained in SPM experiments. Here we present a computer algorithm for automated analysis of mineral-water interface reactions. This algorithm automates the analysis of sequential SPM images by identifying the kinetically active surface sites (i.e., step edges), and by tracking the displacement of these sites from image to image. The step edge positions in each image are readily identified and tracked through time by a standard edge detection algorithm followed by statistical analysis on the Hough Transform of the edge-mapped image. By quantifying this displacement as a function of time, the rate of step edge displacement is determined. Furthermore, the total edge length, also determined from analysis of the Hough Transform, combined with the computed step speed, yields the surface area normalized rate of the reaction. The algorithm was applied to a study of the spiral growth of the calcite(104) surface from supersaturated solutions, yielding results almost 20 times faster than performing this analysis by hand, with results being statistically similar for both analysis methods. This advance in analysis of kinetic data from SPM images will facilitate the building of experimental databases on the microscopic kinetics of mineral-water interface reactions.
V23F-2195
AFM Observations of Weathering and Microbiological Alterations on the Surface of Calcite Buried in Arctic Soil (Spitsbergen)
This study focused on the direct observation of chemical weathering and biological activity on mineral surfaces in the newly forming arctic soil of West Spitsbergen . Chemical weathering and soil forming processes associated with glaciers may affect several geochemical cycles including global carbon cycle and as a result have negative feedbacks on the global climate. Study areas are the foreland of the Werenskiold glacier, continuously retreating by several meters a year. Several samples of freshly cleaved calcite had been buried in the soils for one year. Samples were analyzed with the use of Atomic Force Microscopy (AFM). Results of AFM investigation show changes observed on a calcite samples located respectively about 2500 meters (sample calcite 1) and 100 m (sample calcite 2) from the glacier front as compared to a control sample calcite 0, that has never been exposed to glacier environment. Samples calcite 1 and calcite 2 were recovered from Spitsbergen after 1 year. Compared to the control sample calcite 0, which displays sharp edges and smooth surfaces, both field-treated samples calcite 1 and calcite 2 display rounded edges, irregular surfaces, numerous dissolution features and rounded pitches associated with bacterial activities. The observations suggest that both samples calcite 1 and 2 undergo intensive and rapid chemical and biological weathering when exposed to relatively unsaturated with respect to calcite glacial meltwaters. Several types of analyses have been applied to various regions and lines on the calcite surface. Selected regions on the calcite surface included (a) the entire area of the observed surface (b) top step region roughness, and (c) bottom step region roughness. Selected line parameters have been calculated along: (a) three randomly selected parallel lines, (b) top step line roughness, and (c) bottom step line roughness. Both surface area roughness and line roughness are calculated as the mean deviation of the height. Significant differences have been observed between the samples in calculated roughness parameters, with increase of these parameters ranging from 28% to 241% in calcite 1 located 2500 m from the glacier front and from 100% to 486% in calcite 2 located 100m from the glacier front, as compared to calcite 0. Roughness of the entire surface area for the control calcite 0 was 5.73nm, which increased by 28% in calcite 1 and by 100% in calcite 2. Top step edge roughness increases from calcite 0 (1.82 nm) by 241% to 6.2 nm in calcite 1 and by 486% to 8.84 nm in calcite 2. Bottom step edge roughness increases from 2.08 in calcite 0 by 98% to 4.52nm in calcite 1 and by 149% to 5.67 nm in calcite 2. Line roughness for calcite 0 is 3.26, which increased by 102% in calcite 1 and by 217% in calcite 2. Top step line roughness increases from 2.04 nm in calcite 0 by 191 % to 5.93 nm in calcite 1 and by 276% to 7.67nm in calcite 2. Bottom step line roughness increases from 2.28 nm in calcite 0 by 198% to 4.52 nm in calcite 1and by 249% to 5.67 nm in calcite 2. In calcite 0
V23F-2196
Relative Effects of Nonredox vs. Redox on Fe2+/Fe3+ Equilibrium Isotopic Fractionation in Aqueous Solutions
Previous studies have shown both theoretically and experimentally that nonredox effects (such as bond partner, bond length, and coordination number) may be as significant to equilibrium iron isotope fractionation as the effects of oxidation state [1, 2]. Since 56Fe/54Fe isotope fractionation in the geological record is often taken as an indicator of environmental redox conditions, it is important to understand the influence of nonredox factors (i.e., important ligands present in the environment as found in the solution chemistry of an aqueous solution such as a lake or the ocean) on this isotopic signal. To this end, we explored the relative effects of nonredox vs. redox effects on the iron isotope signature with both theoretical models and experiments, using the aqueous iron chloride system as an easily modeled proxy for potential iron bond-partners found in nature (e.g., sulfides, siderophores, small organic molecules, etc.). We developed ab initio models for the ferric and ferrous chloride complexes using both Unrestricted Hartree Fock and Density Functional Theory. Our experiments consist of a series of low pH solutions of both ferrous and ferric chlorides in varying ratios, combined with an equal amount of diethyl ether. We take advantage of the unique solubility of the Fe(III)Cl4- complex in ether to create a spectator phase against which variations in 56Fe/54Fe partitioning in the aqueous solution can be quantified. Extrapolation to Fe(II)/Fetotal=1 allows us to calculate the fractionation between the dominant ferric and ferrous complexes at a given chlorinity. We ran three series of experiments, extending the range of chlorinities examined to 5M. As the chlorinity of the solution increases, the dominant ferric and ferrous chloride species change, thus altering the nonredox effects of the solution, enabling us to monitor the changing fractionation between the ferric and ferrous species. The difference in δ56Fe(ferric) - δ56Fe(ferrous) is 3.3‰, 2.7‰, and 2.3‰ at [Cl-]=1.5, 2.5, and 5.0M respectively. Our first-principles models predict a slightly smaller sensitivity (~0.2‰/M [Cl-]) on chlorinity. Thus, the nonredox factor of chlorinity decreases the fractionation by ~0.3‰/M [Cl-]. References: [1] Hill & Schauble, GCA 72 (2008) 1939-1958 [2] Hill et al., submitted.
V23F-2197
Nickel Sorption to Bacteriogenic Manganese Oxides: Insights from X-ray Absorption Spectroscopy and Density Functional Theory
Bacteriogenic Mn oxides are ubiquitous, highly reactive minerals with a remarkable capacity to scavenge metals due to their nanoparticulate dimensions and abundant structural defects. These minerals are commonly deposited in a matrix of bacterial cells and extracellular polymeric substances, forming geosymbiotic systems whose reactivity with contaminant metals is not fully characterized. In the current study, a synergistic experimental-computational approach was used to study the mechanism of Ni adsorption at varying loadings and at pH 6-8 using the Mn oxide produced by Pseudomonas putida GB-1. Extended X-ray absorption fine structure (EXAFS) spectra showed two dominant coordination environments: Ni bound as a triple corner sharing (TCS) complex at octahedral vacancy sites and Ni incorporated into the octahedral sheet. The proportion of adsorbed and incorporated Ni varied as a function of surface coverage and pH, with the latter form of Ni being favored at higher loadings and decreased proton activity. These two coordination environments, although consistent with data published for Ni sorbed by synthetic MnO2(s), did not describe fully all of our EXAFS spectra, leading us to consider the binding of Ni at particle edges or via a non-specific sorption mechanism. In parallel to the spectral analysis, density functional theory (DFT) calculations were performed to test different adsorbate-adsorbent configurations and the pH dependence of the adsorption mechanism. Geometry optimized structures for Ni sorbed above vacancies (i.e., TCS) or incorporated into the Mn oxide structure were in excellent agreement with corresponding structural parameters obtained from EXAFS analysis. The calculated energy barriers for the transition from adsorbed TCS to incorporated Ni were consistent with the hypothesis that the TCS complex is a precursor for Ni incorporation and that incorporation is favored by decreased proton activity. The combined perspectives obtained from these two powerful techniques provide a molecular-scale picture of Ni speciation that is integral to predictive models of contaminant remediation.
V23F-2198
First-principles study on elastic constants of C-S-H type minerals
Calcium-Silicate-Hydrate (C-S-H) is the mineral binding phase of all Portland concrete materials, and the principle source of their strength and stiffness. Despite decades of research, the elastic properties of C-S-H mineral crystals are unknown. Here we investigate two natural analogs of C-S-H, tobermorite and jennite, and characterize their mechanical properties by first-principles calculations. First, we calculate their lattice parameters and elastic constants. Second, we show that in contrast to previous suggestions, for natural tobermorite 11 Å , the mechanically weakest directions are two inclined regions that form a hinge mechanism. By studying bond length changes under deformation in tobermorite 14 Å and jennite, we show that water molecules play a major structural role in defining their elastic properties. Averaged elastic moduli obtained by first-principles calculations of tobermorite 14 Å and jennite compare well with corresponding nanoindentation experiment on C-S-H.
V23F-2199
Speciation of Aqueous Silica at High pH Using Raman Spectroscopy
This study presents Raman spectra of silica solutions taken at ambient conditions with varying pH (11.4-14.2) and SiO2 concentrations (0.005-5.5 molal). These results are the foundation of a comprehensive set of hydrothermal diamond anvil cell experiments aimed at exploring the effects of temperature, pressure, and concentration on silica polymerization. Dissolved silica in basic aqueous solution plays an important role in the manufacture of detergents, adhesives, and the formation of zeolites. The nature of dissolved silica is also of central importance in understanding lithospheric fluid chemistry. At lower crustal to upper mantle temperatures and pressures, silica is one of the most soluble major rock-forming oxides, and will dominate the aqueous chemistry at these conditions, especially at conditions approaching the upper critical end point in the SiO2-H2O system. The concentration of dissolved silica will not be limited to the quartz saturation surface, however, and it would therefore be useful to understand what the independent effect of silica concentration is on polymerization. The pH of lithospheric fluids is not as high as in this study, but high pH solutions dissolve large amounts of silica, and thereby allow concentrated silica solutions to be studied at ambient conditions. The effect of pH on silica polymerization is therefore also important to characterize. We collected Raman spectra of silica solutions at ambient conditions with varying pH and silica concentrations, using first-principles calculations to interpret the spectra. Total silica concentration was varied from 0.1 to 5.5 molal while keeping the total K/Si ratio constant at three different K/Si ratios from 1 to 1.4. Raman spectra were also taken at a 1.6 K/Si ratio for silica concentrations ranging from 0.005 to 5.5 molal. The spectra show increasing polymerization with increasing silica concentration, and decreasing polymerization with increasing K/Si ratio. The polymer (475-650 cm-1) and deprotonation (850-1125 cm-1) peak areas scale roughly linearly with concentration, nearly independent of the K/Si ratio. The monomer (735-800 cm-1) peak area scaling, however, depends greatly on the K/Si ratio. Moreover, in the dilute solutions (0.005-0.1 molal), the monomer peak area scaling implies that the solution is starting to become polymerized even at concentrations as low as 0.05 molal. Polymerization of aqueous silica plays an important role in lithospheric fluid chemistry, and could afford silicate-melt-like or crystal-like sites into which otherwise insoluble elements such as titanium could substitute, leading to enhanced mobility for those elements.
V23F-2200
The Structural Role of Aluminium in Silicate Melts
The multiple structural roles which Al3+ in tetrahedral co-ordination assumes in silicate melts as a function of the number, charge and radius of the charge balancing cations as well as the Al/Si ratio leads to a complex variation in the physical and thermodynamic properties of the melt, as well as in the flow mechanism of the melt. Literature data show that there is a change in viscosity trend as a function of composition at the condition when there are no longer enough charge balancing cations available for the Al3+ in tetrahedral co- ordination. This has been observed previously for Li+-, Na+- and Ca2+-bearing melts, but not for Mg2+-bearing aluminosilicate melts. Here we show that the configurational heat capacity (Cpconf) of aluminosilicate melts also shows changes in trend as a function of the cation/Al; with a minimum in Cpconf occurring in melts in the Na2O-Al2O3-SiO2 series at Na=Al(atoms), and a maximum in the CaO- Al2O3-SiO2 series at Ca=2Al(atoms) (the mol.% SiO2 is held constant in these melts). There is however, a minimum in Cpconf in the CaO-Al2O3-SiO2 series at the condition Ca=Al(atoms), suggesting that a further modification of the melt structure occurs when there are no longer enough Ca2+ such that each Al3+ can have its own charge-balancing cation, and the Al3+ must begin to share the calciums. Investigation of the metaluminous CaO-Al2O3 - CaO- SiO2 system (CaO>Al2O3) supports this conclusion as there is a maximum in both Cpconf and viscosity at the Ca=Al(atoms) condition. Cpconf is a measure of the energy required for the melt structure to rearrange to be in equilibrium with temperature. There is, however, no clear correlation between the magnitude of the viscosity or Tg12 (temperature at which viscosity is 1012 Pa s) and the Cpconf data for the different series of compositions. There does however appear to be a good correlation between fragility and the magnitude of Cpconf; as expected from the Adam-Gibbs description of viscosity as a function of configurational entropy. Raman spectroscopy of the glasses has also been used to investigate the changes in structure as a function of Na/Al, Ca/Al and Si/Al. The spectra show systematic changes as a function of composition; but the overlap of the Raman bands for the different Al and Si Q species (e.g. T-O stretching vibration) does not allow any quantitative interpretation of the spectra.
V23F-2201
Moderate pressure phase diagram of methane by Molecular Dynamics simulations
By using classical and ab initio Molecular Dynamics simulations we have investigated the phase diagram of methane up to ~ 25 Gpa. The melting line of phase I (fcc) was computed in a range of pressure corresponding to the Earth's crust conditions by using classical potentials and three different approaches -free energy calculations, phase coexistence method and integration over the coexistence line. The three techniques consistently give a phase boundary in good agreement with known experimental values. The solid phases in a range of temperature between 100K and 300K were investigated using a metadynamics technique, our results providing a possible assignments of structure and explanation of existing, controversial experiments.
V23F-2202
The Role of Biomolecules in Cation Desolvation During Calcification: A Molecular Dynamics Study
The calcifying organisms exhibit a remarkable level of control over the mineralization of their skeletons into calcium carbonate and phosphate phases. Processes of skeletal formation are largely directed by an organic matrix rich in acidic proteins, particularly those high in aspartate content. A number of studies have implicated these Asp-rich biomolecules in the selection of crystal morphology and polymorph. Moreover, we have found that Asp-rich molecules enhance calcite growth kinetics and hypothesized that these compounds favor growth by partially desolvating the cation to lower the energy barrier to mineralization (Elhadj et al., 2005, PNAS). Indeed, ion desolvation is recognized as an influential interfacial process although little is known about mineral-ion-biomolecule interactions. In this study, we test the idea that biomolecules may modulate the barriers to ion desolvation during biomineralization by conducting molecular dynamics (MD) simulations that examine the effect of acidic peptides on the Ca2+ hydration environment. Each experimental system consists of a periodic water box containing one calcium ion and one of three biomolecules: Asp, (Asp)2, or Asp-Leu. MD runs were performed with the LAMMPS software using the TIP3P model of water, CHARMM22 force fields, and Åqvist ion-water potentials. Systems were equilibrated for 20ps before 120ps of data collection at 300K. Calculated radial distribution functions were then used to determine the average first-shell Ca2+ coordination number. We show that the hydration state of Ca2+ is strongly related to its proximity to the amino acid or peptide. For large separation distances, the Ca2+ solvation sphere is unperturbed by the acidic molecule and contains eight first-shell waters. At smaller separation distances, a decrease in average hydration number is seen for all three experimental molecules with minimum hydration numbers around 5.5. The Ca-O distance for first-shell water oxygen atoms is unaffected by the presence of Asp-containing molecules and remains at 2.41 Å for all simulations. Our results suggest that one role of Asp-rich matrix molecules during calcification may be to facilitate nucleation or growth by promoting Ca2+ desolvation.