CONTRACTOR: PACIFIC NORTHWEST NATIONAL LABORATORY

MS K2-25

Richland, Washington 99352

CONTRACT: D-AC06-76RLO-1830

CATEGORY: Geochemistry

PERSON IN CHARGE: A.R. Felmy

A. Surface Structure and Chemistry of Carbonate Minerals (D. R. Baer [509-375-2375; Fax 509-375-5965; E-mail dr_baer@pnl.gov], J. E. Amonette [509-372-6125; Fax 509-372-6328; E-mail je_amonette@pnl.gov], and J. P. LaFemina [509-375-6895; Fax 509-375-4486; E-mail jp_lafemina@pnl.gov])

Objectives: The purpose of this program is to develop a fundamental, microscopic understanding of the structure and chemistry of carbonate surfaces, including the interactions between adsorbates and mineral surfaces.

Project Description: This project involves an interdisciplinary theoretical and experimental effort designed to gain a fundamental, molecular level understanding of carbonate mineral surface structure and chemistry. Carbonate minerals are particularly important in the global CO₂ cycle and in subsurface contaminant migration processes. The availability of large single-crystals allows fundamental measurements to be made on well-defined surfaces. By linking experimental studies of geochemical reactions on single-crystal surfaces with first-principles quantum-mechanical model calculations to describe the surface and interfacial structure and chemistry, a systematic study of the factors controlling the surface chemistry of carbonate minerals can be made. In particular, the effects of substitutional impurities and other point chemical defects on the structure and geochemical reactivity of carbonate mineral surfaces and interfaces can be isolated and quantified. Moreover, this improved microscopic understanding will eventually provide insights into the behavior of these materials in natural systems.

The approach to meeting program goals involves three interdependent areas of effort: development of ab

initio and kinetic Monte Carlo (KMC) models for the structure and chemistry of the calcite cleavage surface, vacuum studies of the structure and chemistry of the cleavage surface, and comparison of surfaces in vacuum with those in model geochemical environments.

Results: High resolution atomic force microscopy (AFM) measurements of the CaCO₃ (101-4) surface in different aqueous solutions indicate a slight relaxation of the uppermost surface carbonate oxygen atoms (Liang et al., 1996a). These measurements imply that the surface maintains a 1x1 structure in which the carbonate group within the surface unit cell slightly reorients itself. In previous studies some evidence of a 2x1 reconstruction had been reported (Stipp et al., 1994). A series of ab initio density-functional computations has been completed that shows no evidence for a 2x1 reconstruction and suggests only a slight (0.1) relaxation of the outermost oxygen atoms and a small reorientation of outer carbonate groups. These calculations found no evidence (cleaved or water-exposed surfaces) of an asymmetrical relaxation as observed by the AFM in aqueous solutions.

AFM observations of the motion of steps during dissolution (Liang et al., 1996b) have been extended to more geologically relevant conditions that include impurities, as well as various solutions with different degrees of saturation. These measurements show that increasing amounts of Mn in solution decrease dissolution and minimize the anisotropy in step velocities

observed in more pure conditions. As might be expected, increasing levels of bicarbonate in solution slow step velocities during dissolution. In addition, the anisotropy of step velocity is changed such that the steps that dissolve the fastest in "pure" conditions nearly stop while the "slow" moving steps are relatively unaffected. The effects are also sufficiently anisotropic so that the shape of pits changes from being rhombohedral to being partially rounded.

In earlier work, a KMC model of dissolution was developed that, after parameterizing to AFM results, was able to reproduce quantitatively all of the available AFM data on shallow pit growth. Fitting to experiment in this way allowed individual elementary microscopic rates involved in the dissolution to be determined for the five types of step edge sites. It also allowed pit growth to be simulated for very small pits in the early stage of growth, a regime that is extremely difficult to observe experimentally. More recent work has clarified the relationship existing between the terrace-ledge-kink (TLK) and KMC models of dissolution. Despite its simplicity, the TLK model provides a

description of the dissolution of calcite that is semiquantitatively correct when tested against the more sophisticated KMC model. This indicates that the dominant microscopic processes occurring during dissolution in the real system, as incorporated into the TLK model, have been correctly identified.

The KMC model is now expanded to study the surface reactivity when closer to equilibrium. In this case, as well as dissolution, growth from solution also occurs and produces observable changes in pit morphology. Again, matching the KMC behavior to AFM data is being used to provide quantitative atomistic information. This approach is also being adopted to study how the elementary rates are affected by the presence of impurities.

References:

Liang. Y., A. S. Lea, D. R Baer, M. H. Engelhard, Surface Science 351, 172 (1996a).

Liang, Y., D. R. Baer, J. M. McCoy, J. P. LaFemina, Journal of Vacuum Science and Technology A14, 1368 (1996b).

B. Theoretical Characterization of the Physics and Chemistry of Soil Minerals (Anthony C. Hess [509-375-2052; Fax 509-375-6631; E-mail ac_hess@pnl.gov] and Maureen I. McCarthy)

Objectives: This program develops and uses solid state quantum mechanical and classical mechanical methods to investigate the atomic scale properties of aqueous mineral interfaces that affect the transport and speciation of geochemically important systems.

Project Description: This research program investigates the microscopic properties of minerals and mineral interfaces that affect the macroscopic transport of contaminants through the subsurface. Our principal focus is on the mineral/water interface with the goal of identifying critical or controlling aspects of atomic scale phenomena that influence the behavior of complex geochemical systems on longer length and time scales. Our strategy is based on an integrated theoretical approach that combines methods from ab initio quantum mechanics and classical mechanics. Ab initio and first principle quantum mechanical methods implemented on massively parallel computer architectures are used to investigate such phenomena as adsorption, dissocia

tive chemisorption, diffusion and desorption on the internal and external surfaces of oxide, metal oxide, and aluminosilicate minerals. Molecular dynamics and molecular mechanics techniques are employed in conjunction with the quantum mechanical calculations to study interfacial dynamics and ensemble effects.

This research program is jointly supported by OBES/Geosciences and OBES/Chemical Sciences.

Results: Our recent work has focused on completing the implementation of a new periodic first principles Gaussian basis density function program (GAPSS) and in using molecular dynamics potentials, developed from our previous quantum mechanical studies using periodic Hartree-Fock theory to investigate mineral water interfaces.

The new program, GAPSS, as implemented on large-scale parallel computer systems, can now study significantly larger systems to higher degrees of accuracy than could previously be achieved. To establish

the accuracy and reliability of this new approach, we have tested its predictions against the available theoretical and experimental data for a variety of bulk oxides and semiconductors, including selected external surfaces of these materials. In addition, studies involving the physisorption of several molecular species (H₂O, CO, HCl, etc.) on the surface of MgO, a-Al₂O₃, and ZnO have been completed and compared to our previous quantum mechanical calculations of these systems. Having successfully completed the initial validation

phase, GAPSS is currently being used to investigate the interaction of molecular water with complex defect structures on MgO and a-Al₂O₃.

Studies have also been completed using molecular dynamics methods that are increasing our understanding of the solvation and near surface geometric structure of ionic species at the water/MgO(001) interface. Comparisons of experimental and theoretical XAFS spectra are currently being used to understand the geometric structure of this aqueous mineral interface.

C. Structure and Reactivity of Ferric Oxide and Oxyhydroxide Surfaces (James R. Rustad [509-372-6313; Fax 509-372-6328; E-mail jr_rustad@pnl.gov] and Andrew R. Felmy [509-372-6296; Fax 509-372-6328; E-mail ar_felmy@pnl.gov])

Objectives: The objectives of this program are to (1) develop the capability to develop large-scale molecular models of hydroxylated ferric oxide and oxyhydroxide surfaces, (2) use these models to better understand the relationship between surface structure and reactivity for this class of minerals, and (3) use this knowledge to advance the predictive capability of thermodynamic models for adsorption.

Project Description: Ferric oxides have high specific surface areas and high affinities for oxyanions and heavy metals and actively respond to changes in redox conditions in natural environments. These minerals are therefore important in a variety of low-temperature geochemical processes, particularly those in which adsorption and dissolution couple with fluctuations in redox potential. For many solutes, measurement of sorption density versus aqueous concentration suggests the presence of a heterogeneous array of surface sites having a range of affinities for the probing solute. Crystallographic differences in the arrangement of surface oxide sites are a fundamental aspect of this heterogeniety. In this project, the effects of crystallographic heterogeneity on adsorption are evaluated using large-scale computational molecular models. These results are then used to produce a more robust thermodynamic description of adsorption at the mineral-water interface.

Results: During FY 1996: (1) We have performed molecular-dynamics simulations of the goethite-water interface in an effort to assess the role of solvent in controlling surface speciation. Water molecules (125)

were placed in a thin film between two goethite (110) surfaces. The hydroxyls associated with the dissociatively-adsorbed water remained bound to the FeOH₂ groups. Thus, the reaction mechanism for the uptake of protons by the goethite surface appears to be FeOH₂-OH + H⁺ -> FeOH₂⁺ + H₂O. (2) The infrared spectrum of hydroxylated goethite was computed and shown to be in reasonable agreement with the experimental results. In particular, it was shown that several surface species were not infrared active. The results explain the existence of only two peaks in the surface vibrational spectrum, despite the presence of five distinct surface species. (3) Studies of nonhydroxylated hematite (001) and (012) surfaces were undertaken to provide a benchmark of our classical simulation methods with quantum mechanical calculations. Our results are in excellent agreement with available LDA, Hartree-Fock, and tight-binding calculations. (4) Hydroxylation of the hematite surface strongly affected the manner of surface relaxation through surface hydrogen bonding. The coordination of the surface iron ions was strongly distorted for surfaces in which the face-sharing octahedra are exposed. (5) The shape of the surface charge vs solution pH curve for goethite was well-represented in the dilute region by the model derived from the molecular statics methods; however, prediction of surface titration data at higher background electrolyte calculation required introduction of specific pair formation constants between the background electrolyte ions and the surface.

References:

Rustad, J. R., A. R. Felmy, and B. P. Hay, Molecular statics calculations of proton binding to goethite surfaces: A new approach to estimation of stability constants for multisite surface complexation models, Geochim. et Cosmochim. Acta, 60, 1563-1576, 1996.

Rustad, J. R., A. R. Felmy, and B. P. Hay, Molecular statics calculations for iron oxide and oxyhydroxide minerals: Toward a flexible model of the reactive mineral-water interface, Geochim. et Cosmochim. Acta, 60, 1553-1562, 1996.

Lockheed Martin

Albuquerque, New Mexico 87185

CONTRACT: DE-AC04-94AL85000

CATEGORY: Geophysics and Earth Dynamics

PERSON IN CHARGE: M. C. Walck

A. Micromechanics of Failure in Brittle Geomaterials (Joanne T. Fredrich [505-844-2096; Fax 505-844-7354; E-mail fredrich@sandia.gov] and Teng-fong Wong [State University of New York at Stony Brook])

Objectives: The objective of this project is to provide a fundamental understanding of the effects of grain boundary structure and cementation, damage state, and load path on the deformation and failure mode of brittle porous and nonporous geologic materials by measurement of mechanical behavior under high pressure and deviatoric stress, quantitative microstructural characterization of pristine and deformed samples, and theoretical analysis.

Project Description: Knowledge of the failure behavior of rocks is important for several energy-related applications, including oil and gas exploration and production, underground disposal of nuclear waste, and drilling technology. The experimental investigation will provide a detailed understanding of the micromechanical processes associated with the brittle failure of geomaterials and includes triaxial tests following various load paths that are defined by the ratio K of the change in the radial confining (horizontal) stress to the change in the axial (overburden) stress. Tests are conducted to various stages of failure and include measurement of strain and acoustic emission. The micromechanical failure process is further elucidated and characterized quantitatively using light microscopy, laser scanning confocal microscopy (LSCM), and scanning electron microscopy. Work focuses on porous carbonate and siliciclastic rocks, although related experiments are also being performed on low-porosity crystalline rocks in order to study completely the ef

fect of certain parameters. The results of the laboratory tests and microstructural studies are used to guide analyses using fracture mechanics and continuum plasticity theories.

Results: Application of a high resolution three-dimensional imaging technique developed previously under this project to various engineering materials was demonstrated. Thermal Protection System (TPS) materials possess a complicated, porous microstructure that directly influences their thermal, mechanical, and chemical properties. This is especially true of rigid, fibrous insulations, such as the LI, FRCI, and AETB families of TPS tile insulations used on the Space Shuttle. Conventional imaging techniques, such as optical microscopy or scanning electron microscopy, cannot be used to quantify the microstructures of these porous materials (~80%), which is crucial for TPS modeling, development, and optimization. We determined that the LSCM technique can be used to image the microstructure of these fibrous insulations and also to quantify various aspects of the microgeometry of both the solid matrix and void space. As a result of the successful preliminary results, Fredrich is now conducting a collaborative project with Dr. J. Marschall of NASA Ames Research Center with funding from NASA Ames Research Center. In a second application, the LSCM technique was applied to image growth defects in SiC crystals. The growth defects are of interest since they are detrimental to the performance of

semiconductor devices fabricated from SiC. The LSCM technique was used to reveal the three-dimensional structure of superscrew dislocations that can be used to constrain theories of their development. This work was conducted in collaboration with Dr. W. M. Vetter and Prof. M. Dudley of SUNY Stony Brook. Preliminary results were reported in MRS Symp. Proc. V. 406.

Analysis of the experimental test series conducted on Westerly granite with differing initial damage states was conducted. The experimental data suggest two regimes of crack growth in brittle geomaterials subjected to all-compressive loading: a low pressure regime (<25 MPa), where the failure process is sensitive to the preexisting crack population, and a higher pressure regime, where microcrack growth leading to brittle

failure is unrelated to the preexisting microcracks. We are applying Gupta and coworkers' analysis of stress singularities at grain triple junctions due to elastic anisotropy to interpret the experimental results.

A triaxial test series to investigate the effect of ductile grain-boundary-lining second phases on brittle failure in compression was conducted. The experimental data reveal a transition in crack propagation behavior at elevated pressures where shear localization is inhibited. The data suggest that the second phase enhances the damage tolerance of the material by causing grain-boundary delamination with no loss in load-bearing capacity. Detailed microscopy studies and micromechanical analyses of crack propagation at an interface are underway.

B. Laboratory and Theoretical Analyses of Transport Paths in Single Natural Fractures (S. R. Brown [505-844-0774; Fax 505-844-7354; E-mail srbrown@sandia.gov], N. G. W. Cook, L. R. Myer, and G. Yang [University of California at Berkeley and Lawrence Berkeley National Laboratory])

Objectives: Fluid flow in fractured rock is an important phenomenon to understand in connection with energy production and containment or disposal of wastes. The objective of this project is to address several outstanding questions of the effects of void topology on flow and transport in single fractures by quantitative, visual observations and measurements of single- and two-phase flow.

Project Description: We have developed a method for obtaining precise replicas of real fracture surfaces using transparent epoxy resins. These replicas are being used to examine flow in the fracture void space using digitized optical imaging and nuclear magnetic resonance imaging (NMRI). In parallel work, we are examining single- and two-phase flow in irregular fracture aperture distributions using graph theory, effective medium theory, and percolation theory to analyze the topology of the conducting network of void space. This work emphasizes that the topology of the network is of at least as much importance as the conductance of the individual elements in determining fluid flow. The topological characteristics of the preferred paths at different scales will be analyzed to study the effects of scale on flow and dispersion in fractures. Numerical

predictions based on graph theory will be compared with experimental observations of multiphase flow.

Results: Field specimens of several natural rock joints were collected. Two samples were chosen for detailed study. Matched two-dimensional images of the surface topography of each fracture surface were measured using laser profilometry, allowing computation of the aperture distribution. Silicon rubber molds were used to construct epoxy replicas of both specimens. Clear and dyed fluids were injected into the fracture pore space. The Lambert-Beer Law for light attenuation of dyes was used to determine the aperture distribution of the replicas. Dye was injected into clear fluid to observe flow channels. NMRI was used for quantitative measurements of flow velocity. Both NMRI and video imaging techniques show distinct and strong channeling of fluid flow at the sub-millimeter to centimeter scale. Both replicas have one large dominant flow channel. The aperture distributions derived from NMRI and video imaging are very similar; however, the correlation between direct aperture measures and those reconstructed from surface profilometry was poor. The single large channel does not appear in the reconstructed aperture. Differences between actual and reconstructed

sample assembly and differences in the scale of the measurements are likely reasons for this discrepancy.

A new effective medium theory (EMT) was developed and documented in collaboration with J. B. Walsh (MIT). This model departs from standard theories in that it includes short-range spatial correlations in aperture size. Adding aperture correlation to EMT improves substantially on the accuracy of flow predictions from statistical aperture data as compared to most earlier methods.

In collaboration with R. L. Bruhn (U. of Utah), a model has been developed and documented for the for

mation of voids and veins along geologic faults. This model includes progressive dilation of the fault during multiple slip events and the elastic deformation of the surfaces normal to the fault plane (closure) as new void space develops. The model predicts vein geometries that are qualitatively similar to those observed in fault-controlled mineral deposits. The model shows that new void space opened by slip along a typical rough fault cannot easily be closed by elastic processes due to 10 km burial. This emphasizes the high capacity of faults to transmit fluids at depth.

C. Shear Strain Localization and Fracture Evolution in Rocks (W. R. Wawersik [505-844-4342; Fax 505-844-7354; E-mail wrwawer@sandia.gov], D. J. Holcomb, and W. A. Olsson)

Objectives: This research seeks an improved understanding of the mechanism of the formation of faults and fractures in rock and the prediction of their causative stresses, location, orientation, thickness, and spacing. Specifically, the research examines the applicability of a theory by Rudnicki and Rice that describes faulting as a constitutive instability leading to a localization of shear deformation from a homogeneous pattern of deformation.

Project Description: An experimental program is undertaken that entails four parts: (1) A systematic evaluation of the Rudnicki-Rice constitutive parameters under axisymmetric and truly multiaxial stress states. Several rock types are used to cover a range of porosities, including low-porosity Tennessee marble and intermediate porosity Gosford sandstone. (2) Experiments to measure and compare the actual and predicted development of strain localization in axisymmetric and plane-strain compression. (3) Multiaxial stress tests on pressurized thin-walled cylinders with superimposed torsion to investigate special phenomena leading to strain localization at relatively small deformations. (4) Post test petrographic observations concerning the details of shear banding and the potentially accelerating effects of imperfections in experimental boundary conditions. The foregoing research is integrated with a parallel theoretical study by Rudnicki.

Results: Previous work had shown that a vertex forms at the load point on the yield surface for Tennes

see marble. To investigate this phenomenon for a porous rock, a zigzag stress history was applied to both Gosford and Berea sandstones. Initial data analysis indicates that vertices also form on the yield surface for these rocks, suggesting that strain localization can occur earlier than smooth-yield-surface predictions. A collaboration was begun with a computational group at Sandia with the goal of incorporating the constitutive properties measured in this project into codes suitable for modeling large-scale geological structures. There is increasing interest in the petroleum industry in developing this capability. In addition to incorporating general criteria for localization, realistic modeling requires the inclusion of nonnormality, hardening, and pressure dependence; both the form and the values of these quantities were measured for the rocks studied under the current project.

A larger, true triaxial apparatus was brought into operation during the year, capable of deforming samples as large as 10 by 10 by 20 cm. Initial applications of the apparatus to Gosford sandstone showed the effects of stress state clearly. At the same mean stress, plane strain deformation resulted in failure with dilation, while conventional triaxial testing led to failure with compaction. The associated fault angles differed by 10 degrees in accord with theory. The apparatus' ability to apply stress states that are not axially symmetric is crucial to testing theories of localization.

CATEGORY: Geochemistry

PERSON IN CHARGE: M. C. Walck

A. Cation Diffusion Rates in Selected Minerals (Diana K. Fisler, Randall T. Cygan [505-844-7216; Fax 505-844-7216; E-mail rtcygan@sandia.gov], and H.R. Westrich)

Objectives: Determine experimental cation diffusion coefficients for enstatite and carbonate minerals at temperatures less than 1000°C for evaluating disequilibrium behavior in geological, nuclear waste, energy, and materials concerns.

Project Description: Evaluation and modeling of geochemical processes related to nuclear waste, energy, and materials problems will require the accurate determination of cation diffusion data in silicate minerals and carbonates. A new technique for the preparation of diffusion couples using thin film technology was developed in an effort to evaluate the relatively slow diffusion of Mg²⁺ and Ca²⁺ in phases such as pyroxene and calcite, where diffusion rates are on the order of 10^-22 to 10^-16 m²/s in the temperature range of 700° to 1000°C. Resistive evaporation of enriched stable isotopes onto polished mineral surfaces is used to create a thin film-mineral diffusion couple. Diffusion couples are annealed in a controlled oxygen fugacity furnace for periods up to three months in order to provide a diffusive penetration depth of approximately 0.2 microns. Depth profiles of the tracer isotope are obtained using an ion microprobe and are then fit to appropriate diffusion models to obtain precise and reproducible diffusion coefficients.

Results: Work has been completed on the enstatite phase of this project. Samples of enstatite with compositions ranging from Mg_0.99Fe_0.01SiO₃ to Mg_0.83Fe_0.17SiO₃ show no dependence on iron content within the resolution of this technique. Samples of the composition Mg_0.91Fe_0.09SiO₃ and Mg_0.99Fe_0.01SiO₃ were annealed at oxygen fugacities ranging from the quartz-fayalite-magnetite to the iron-wustite buffers for determining the oxygen-fugacity dependence of the diffusion

coefficient and show small or no dependence on oxygen fugacity in contrast to assumptions based on comparisons to the diffusion of cations in olivine. In addition, extensive transmission electron microscopy on the enstatite/oxide interface reveals the character and thickness (400Å) of the thin film. Ten to 15 samples each of calcite, dolomite, rhodochrosite, and magnesite were obtained and prepared to provide surfaces parallel to cleavage and coated with ²⁵Mg-enriched oxide. Five samples each of calcite and dolomite were prepared and coated with ⁴⁴Ca-enriched oxide. Preliminary experiments on the diffusion of ²⁵Mg in calcite suggest that diffusion coefficients are on the order of 10^-23 m²/s for anneals performed at 500°C and 600°C. An ion microprobe profile analysis for a ⁴⁴Ca diffusion experiment is shown in the accompanying figure. These results suggest a similar value of 10^-23 m²/s for the ⁴⁴Ca self-diffusion coefficient for an anneal completed at 650°C. Further experiments are in progress to confirm the results for Mg and Ca diffusion in calcite and to extend the measurements to lower temperatures.

B. An Investigation of Organic Anion-Mineral Surface Interactions During Diagenesis (Patrick V. Brady [505-844-7146; Fax 505-844-7354; E-mail pvbrady@sandia.gov], Randall T. Cygan, and Henry R. Westrich)

Objectives: Mineral surface-organic acid interactions affect organic anion budgets and often control the dissolution and growth of aluminosilicate minerals during soil formation and diagenesis. Determination of temperature-dependent adsorption of oxalate, benzoate, and salicylate onto enstatite, albite, aluminum oxide, and kaolinite as functions of organic anion type and concentration will lead to improved understanding of soil weathering characteristics and diagenetic porosity evolution.

Project Description: The adsorption of organic anions to mineral surfaces in soils and deep basins can be understood if temperature-dependent adsorption isotherms of carboxylate and phenolate groups onto aluminosilicate surfaces are first known. Moreover, the catalytic role of adsorbed organic anions on mineral dissolution and porosity evolution in soils and during diagenesis may be reliably estimated if the mechanistic link between anion adsorption and reaction rate is quantitatively established. These hypotheses are being examined by wet-chemical measurements of the temperature-dependent adsorption of oxalate, salicylate, and benzoate onto quartz (SiO₂), corundum (Al₂O₃), enstatite (MgSiO₃), albite (NaAlSi₃O₈), and kaolinite (Al₂Si₂O₅[OH] ₄). To complement this experimental program, a variety of computer-based techniques are being used to examine the molecular interactions at the mineral-solution interface, including ionic modeling, Monte Carlo docking and molecular dynamic simulations. Linking observed anion-surface interactions with an atomistic evaluation

of reaction mechanisms and pathways will lead to a set of general rules for predicting the extent of organically mediated phase changes during diagenesis. At the same time, field measurements of soil weathering in the presence and absence of organic acids are being used to determine linkages among soil biota, the silicate-carbonate cycle, and global climate.

Results: (1) Mineral dissolution at pH 3, enstatite dissolution at 25° and 60°C varies from 10^-14.7 to 10^-14.0 moles/cm²s, respectively. At similar temperatures and near neutral pH, dissolution rates range from 10^-15.5 to 10^-14.9 moles/cm²s. At 25°C albite dissolution from pH 3 to 5 is approximately 10^-16.5 moles/cm²s. Oxalate at 100 µmol levels depresses the dissolution of both enstatite and albite at 25°C .

(2) Temperature-dependent sorption albite at 25°C and kaolinite at 60°C both strongly sorb salicylate under acidic conditions (pH < 3). Only mild sorption of benzoate occurs on albite, Al₂O₃, and kaolinite at 25° and 60°C.

(3) Digital imaging of field weathering: We have completed digital imaging of abiotic and lichen-controlled weathering of plagioclase and olivine as a function of temperature and precipitation on Hualalai Volcano in Hawaii. The activation energies for abiotic weathering are very close to those measured in the lab. The lichen-controlled activation energy for plagioclase dissolution is 13 kcal/mol. Precipitation effects on runoff are fit with a power law and are found to be much greater for lichen-controlled weathering.

C. Heterogeneous Nucleation and Growth Kinetics of Clays (Kathryn L. Nagy [505-844-5337; Fax 505-855-7354; E-mail klnagy@sandia.gov] and Randall T. Cygan)

Objectives: Clay nucleation and growth often occur heterogeneously and/or epitaxially on detrital minerals. Kinetics of these processes and role of nucleating substrates are unknown. The goal is to quantify clay mineral growth by examining the reaction at the surface,

assess reactive surface area, and provide kinetic data for accurate modeling of weathering and diagenesis.

Project Description: Experiments are performed in which the growth surface (single crystal or organic substrates) is characterized before and after reaction

by various techniques used by the surface physics community to investigate thin films. These include atomic force microscopy (AFM), ion beam analyses (Rutherford backscattering and elastic recoil detection), rotating anode and synchrotron X-ray reflectivity, and X-ray diffraction (XRD). Comparison is made with results from standard powder experiments in which rates are quantified from solution chemistry changes. Clay growth occurs under controlled solution composition and temperature conditions that mimic nature. Nucleation site densities and nucleated crystal morphologies are monitored to acquire information on reactive surface areas, a parameter considered typically by bulk measurements such as gas adsorption. Molecular modeling of the bonding interactions between substrate and overgrowth as well as overgrowth morphology provides a fundamental basis for interpreting the experimental results.

Results: In the first year, an AFM was acquired and set up for application to measurement of clay growth. Initial batch and flow-through experiments on brucite growth on muscovite were conducted at 80°C, in pH 9 solutions. Growth was characterized by Tapping Mode™ AFM, ion beam analyses, and rotating

anode XRD. Results from the ion beam and low-angle X-ray reflectivity analyses show that a layer 100- 300 Å thick forms at supersaturations seven times higher than the equilibrium ion activity product at 1 to 5 days. Tapping Mode™ AFM images show a distinction in morphology and growth mechanism of the brucite islands as a function of saturation state. At a supersaturation of 7, islands are spaced about 0.5 µm apart, have trigonal outlines with convex edges 0.3 - 0.5 µm in length, and are one unit cell high. In contrast, at a supersaturation of 15, islands are spaced every 0.05 - 0.1 µm, exhibit no obvious shape, and are 4-5 unit cells high. At lower supersaturation, monolayers appear to form sequentially perpendicular to the basal plane, while, at higher supersaturation, individual islands grow upwards before spreading laterally. Calculated rates based on the brucite film thickness are similar in magnitude to published dissolution rates at comparable temperatures and in comparable solutions. Brucite dissolves seven orders of magnitude faster than other sheet-structured minerals, including gibbsite, kaolinite, and muscovite. Therefore, it is promising to obtain growth rates that to first order match dissolution rates.

CATEGORY: Hydrology

PERSON IN CHARGE: M. C. Walck

A. Laboratory Investigation of Constitutive Property Upscaling (Vincent C. Tidwell [505-848-0574; Fax 505-848-0558; E-mail vctidwe@sandia.gov], John L. Wilson [New Mexico Institute of Mining and Technology])

Objective: The basic objective of this research program is to enhance fundamental understanding of the processes and media characteristics that govern permeability upscaling. Specifically, we will address the following questions:

• Does permeability upscale in a predictable and quantifiable manner?

• What characteristics of the geologic medium influence upscaling behavior?

• What are the appropriate measures of permeability upscaling (i.e., first two statistical moments, fractal dimension, other)?

Project Description: Laboratory investigation of permeability upscaling is accomplished through the use of a specially adapted mini-permeameter that we have termed the multi-support permeameter (MSP). The MSP allows rapid, precise, nondestructive measurement of gas permeability over a range of discrete sample

supports (i.e., sample volumes). Measurements are made at different sample supports by simply varying the size of the permeameter tip seal. In this way, measurements spanning five orders of magnitude on a per volume basis are made subject to consistent boundary conditions and flow geometry. Experiments progress by collecting thousands of measurements on multiple faces of meter-scale blocks of rock with each of five different tip seals (0.31, 0.62, 1.27, 2.54, and 5.08 cm inner diameter[ID]) plus a single large-scale (15.24 cm ID) measurement designed to integrate over the entire sampling domain. This process is repeated on multiple rock samples each of which is carefully selected according to its depositional and diagenetic history. The acquired data is used to evaluate the performance of theoretical upscaling models, to bound their application with respect to limiting assumptions, and to explore alternative measures and models of upscaling behavior.

Results: In this project's first year, efforts have focused on evaluation of the MSP. Modifications to tip seal design have been made to achieve improved control over tip seal geometry under compression. Evaluation of seal quality as a function of compression time and pressure has also been accomplished. In addition, effects of gas slippage, head loss, nonsteady-

state flow, and non-Darcy flow on permeability calculations have been assessed. A series of experiments were then performed on synthetic, "homogeneous" plaster and mortar blocks to demonstrate that measurements made at different sample supports are free from bias induced by the measurement technique. A suite of exhaustive measurements was also collected from natural rock samples to demonstrate that measurement error is low and consistent across the different tip seals.

Upon completion of the system evaluation, extensive upscaling data were collected from two fluvial sandstones, a 0.3 by 0.3 by 0.3 m block of Berea Sandstone (14,000 permeability measurements) and a 1.0 by 1.0 by 1.0 m block of Massillon Sandstone (75,000 permeability measurements). It should be noted that such 3-D, spatially exhaustive, multi-support data sets do not exist elsewhere. Reduction of this data has revealed strong trends in the mean, variance, and correlation length scale as a function of sample support. Considering the care that has been taken to insure consistency in the multi-support measurements, we are convinced that the measured trends are not simply artifacts of the sampling program but are diagnostic of permeability upscaling. Analysis of this data and comparison to theoretical upscaling models are ongoing.

B. Two-Phase Immiscible Fluid Flow in Fractured Rock: The Physics of Two-Phase Flow Processes in Single Fractures (Robert J. Glass [505-848 0556; Fax 505-848-0558; E-mail rjglass@nwer.sandia.gov], Harihar Rajaram [University of Colorado, Boulder], Michael J. Nicholl [Oklahoma State University])

Objectives: The objective is to develop a quantitative understanding of the critical processes controlling two-phase flow and transport in fractures based on detailed physical experiments and high resolution numerical simulations. This understanding may subsequently be abstracted for use in conceptual models applied at large-scale to applied problems in petroleum extraction and the isolation of hazardous or radioactive waste.

Project Description: Under two-phase, immiscible fluid-flow conditions, phase geometry within the fracture (i.e., the geometry saturated with each phase) ultimately controls the permeability to each phase, fluid

pressure/saturation relations, and solute dispersion within each phase. Phase geometry is a function of both the aperture field and the two-phase flow processes themselves. Capillary, gravitational, and viscous forces in combination with boundary and initial conditions have all been demonstrated to play roles in the formation of fracture phase-geometry. Phase "fingering," separate from the single phase concept of flow channelization, occurs where any of these forces are dominant (capillary fingering, gravity-driven fingering, viscous fingering), emphasizing the importance of understanding the two-phase displacement processes themselves.

In this collaborative project between Sandia National Laboratories and the University of Colorado, systematic physical experimentation is coupled with concurrent numerical simulation to explore the interplay among capillary, gravitational, and viscous forces in the control of phase structure and, thus, flow and transport in rough-walled fractures. Understanding of capillary, viscous, and gravity fingering is explored in order to identify mechanisms that may be used to damp or enhance fingered flow. In addition, a conceptual understanding is developed for the phase invasion processes controlling replacement of a fully entrapped, immobile phase as it dissolves into the flowing phase.

Results: Experimental study of two-phase flow in fractures has been limited by the inability to measure fracture aperture fields, phase occupancy geometry, and solute concentration fields within a given experiment at sufficient resolution. At Sandia National Laboratories, high resolution (time, space, concentration, aperture) full-field light transmission techniques to make such measurements in transparent fractures constructed out of analog materials have been developed and evaluated. Aperture and solute concentration fields measured in a series of fractures that increase roughness systematically are being used to test and develop capabilities for large-scale modeling of single phase flow, solute transport (diffusive to advective dominated), and

phase dissolution in fractures containing entrapped phase structures (see University of Colorado contribution for this project).

A modified invasion percolation (MIP) model has been developed to simulate slow-flow phase invasion processes; MIP incorporates both gravity and capillary forces in the absence of viscous forces (i.e., applicable to small capillary number flows). Capillary forces are modeled by the Laplace-Young equation that relates pressure drops across fluid/fluid interfaces to its curvature and surface tension. Interfacial curvature includes both curvature between the rough walls (determined by the local aperture, contact angle, and local plane convergence/divergence angle) and "in-plane" curvature as one would see looking normal to the plane of the fracture. When only aperture-derived curvature is included, the model does not predict the experimental results; however, inclusion of in-plane curvature allows good simulation of past experimental results for gas dissolution, gravity-driven fingers, horizontal invasion, and gravity stabilized rises and drains. A series of simulations using measured aperture fields (0.01 mm resolution) form an analog rough-walled fracture (isotropic, correlation length 0.7 mm) is being conducted to define the set of physical experiments required for a full test of the MIP model.

C. Multicomponent Convection in Porous Media and Fractures (Robert J. Glass [505-848-0556; Fax 505-848-0558; E-mail rjglass@nwer.sandia.gov], Harlan W. Stockman, and Scott W. Tyler [Desert Research Institute, Reno, Nevada])

Objectives: Objectives are to understand the physical processes controlling multicomponent convection in natural porous and fractured media and to develop quantitative relationships between system parameters (permeability, porosity, and solute concentrations) and the magnitude of convective mass transport in subsurface hydrologic environments.

Project Description: Multicomponent convection is a transport phenomenon that can occur in fluid bodies and fluid-filled porous media that contain two or more components (heat, solutes, etc.) that influence the local fluid density. If two or more of the components individually influence local fluid density, differ in dif

fusivity, and have opposing individual density gradients, then the globally stable system can develop local buoyancy-driven instabilities in the form of thin fingers or larger convection cells. These instabilities may in turn drive larger scale motion within the porous media or fluid body. Multicomponent (double-diffusive for two component systems) convection may contribute significantly to mass and heat transport in porous and fractured systems; leading to fluxes in excess of two orders-of-magnitude higher than would be predicted in an equivalent, density-stable system.

In this collaborative project between Sandia National Laboratories and the Desert Research Institute,

we combine systematic physical experimentation with concurrent numerical simulation to explore multicomponent convection in porous media and fractures. Sandia National Laboratories' high resolution, full-field energy transmission techniques are used to study the onset and development of multicomponent convection in simulated porous media (Hele-Shaw cells), sands, and fracture analogs. Boundary and initial conditions, as well as the permeability field are varied to evaluate the conditions under which multicomponent convection occurs in natural systems. Numerical modeling efforts focus on application of lattice gas automata techniques to consider two- and three-component systems.

Results: This project was initiated in February 1996. We have analyzed 13 experiments that investigate the factors controlling the transition from stable diffusive transport to unstable multicomponent convection. Each experiment began by layering a sucrose solution over a sodium chloride solution in a Hele-Shaw cell such that the fluid was density stable, although double-diffusively unstable. The concentration of the

sucrose solution was systematically varied between experiments to test the analytical theory predicting the stability boundary. Using light transmission techniques, the sodium chloride concentration field was tracked using a nonreactive dye. Analysis of images taken throughout an experiment show mass conservation better than 2%, in spite of the evolution of highly contorted concentration fields.

Experiments conducted to date suggest that, for sufficiently large component Rayleigh numbers, system stability can be predicted without detailed knowledge of variability within the permeability or solute concentration field. If found to hold, this will be an important result, as uncertainty with respect to the permeability and solute concentration fields is often extreme within the subsurface environments. The experiments also demonstrate that solute fluxes can be quantitatively determined and that these fluxes are well in excess of that predicted if the solutes are transported by diffusion alone.