GRANTEE: STANFORD UNIVERSITY

Department of Geological and Environmental Sciences

Stanford, California 94305-2115

GRANT: DE-FG03-93ER14347-A000

TITLE: Cation Chemisorption at Oxide Surfaces and Oxide-Water Interfaces: X-Ray Spectroscopic Studies and Modeling

PERSONS IN CHARGE: Gordon E. Brown, Jr., and George A. Parks (415-723-9168; Fax 415-725-0979; E-mail gordon@pangea.stanford. edu)

Objectives: This project concerns chemical interactions between metal ions in aqueous solution and oxide surfaces representative of those found in the Earth's crust. These "sorption" reactions partition the metal between fluid and solid phases and must be understood at a molecular level to develop quantitative geochemical understanding of mineral surfaces and the macroscopic models required to predict the fate of contaminants in the environment. Our objectives are to (1) characterize sorption reactions by determining composition, molecular-scale structure, and bonding of the surface complexes produced using direct sorption measurements and synchrotron-based X-ray absorption fine structure (XAFS), X-ray photoelectron (XPS), and UV/Vis/IR spectroscopies, (2) investigate how these properties are affected by the solid surface and aqueous phase composition, and (3) develop molecular-level and macroscopic models of sorption processes.

Project Description and Results: (1) Grazing-Incidence XAFS Studies of Co(II) and Pb(II) Sorption Complexes on Single Crystal -Al₂O₃: Using an apparatus we developed for fluorescence XAFS of adsorbates on wet single crystal surfaces in grazing-incidence geometry, we examined sorption complexes of Co(II) and Pb(II) on -Al₂O₃ (0001) and (1-102) surfaces and of Co(II) on TiO₂ (rutile) (110) and (001) surfaces. As an example of our results, we found that Pb(II) forms outer-sphere (OS) complexes on -Al₂O₃ (0001) but inner-sphere (IS) complexes on -Al₂O₃ (1-102). In contrast, Co(II) forms about equal concentrations of IS complexes on both surfaces of -Al₂O₃, adsorbing

dominantly to tridentate sites on the (0001) surface and dominantly to tetradentate sites on the (1-102) surface. These differences are explained using a bond-valence model that accounts for structural differences on the alumina surfaces and differences in coordination chemistry of Co(II) and Pb(II) in solution. These results provide the first direct structural evidence that different surfaces of an oxide can adsorb cations in different modes.

(2) Fluorescence-yield XAFS Studies of Pb(II) and Co(II) Sorption Complexes on Powdered -Al₂O₃ and Iron Oxides: We have used XAFS to study Co(II) and Pb(II) sorbed onto high surface area -Al₂O₃ and iron oxides. On -Al₂O₃, Pb(II) sorbs preferentially to edges of Al(O,OH)₆ octahedra as mononuclear bidentate IS complexes at pH 6-7 and sorption densities of 0.5 to 5.2 mmoles/m². Pb(II) sorbs on goethite and hematite as mononuclear bidentate complexes to edges of FeO₆ octahedra. Our bond-valence model and XAFS results suggest that Pb(II) sorbs primarily at unprotonated surface sites. Hydrolysis of Pb(II) appears to be the primary source of proton release during Pb(II) sorption. Absence of dimeric Pb(II) complexes on iron oxides and their presence on -Al₂O₃ may be related to differences in Al(O,OH)₆ and FeO₆ octahedral edge-lengths and bond-valence sums at the surface oxygens to which Pb(II) bonds. XAFS study of aqueous Co(II) sorption on alumina powders (pH 8, 2; [Co]_T = 100 mM - 12.6 mM) provides strong evidence for a new type of mixed Co-Al hydroxide precipitate at Co 2. Under these

solution conditions, the Co concentration is below the solubility of solid Co(OH)₂. The co-precipitates are less soluble than either Al(OH)₃ or Co(OH)₂ and apparently form when Al dissolves from Al₂O₃.

(3) Copper(II) and 2,2'-bipyridine Co-adsorption on Oxide Surfaces: We have used XAFS and other methods to determine how 2,2'-bipyridine (bipy) influences sorption of Cu(II) on amorphous SiO₂ and 2O₃. Cu sorbs as IS complexes on both oxides. Complexes are multinuclear, possibly dimeric, and monodentate on am-SiO₂ but monomeric and probably monodentate on alumina. Under the conditions studied, bipy inhibits Cu(II) sorption on 2O₃. On am-SiO₂, bipy enhances Cu(II) sorption when 2, but inhibits sorption when 2. With solution Cu:bipy of 1:2, Cu:bipy in the sorption complex was 1:1 on 2O₃ but 1:2 on am-SiO₂, indicating that one bipy dissociates from the complex during sorption on 2O₃. FTIR showed that all sorbed bipy was bound to Cu on both solids. Comparison of EXAFS and XANES spectra of the Cu/bipy₂/2O₃ sample with spectra from the Cu(bipy)₁ and Cu(bipy)₂ solutions indicates that the surface species on 2O₃ is predominantly Cu(bipy)₁, in agreement with the Cu:bipy ratio derived from uptake data. XAFS analysis suggests IS binding of Cu. Comparison of the EXAFS of Cu(II)/bipy₂/am-SiO₂ with those of crystalline Cu(bipy)₂ compounds shows that the surface species is Cu(bipy)₂.

(4) Macroscopic Sorption Modeling: Using a surface complexation, electrical triple layer sorption model (SCM-TLM) in the speciation codes HYDRAQL and

FITEQL, we have found that the structural and compositional information derived from XAFS constrains the selection of sorption reactions. For sorption of Co(II) on 2O₃, many sorption reactions are capable of simulating uptake data alone. In this system, as for Co(II) on -Al₂O₃, XAFS-derived near-neighbor Co stoichiometry requires multinuclear surface complexes, however, and thus rules out strictly mononuclear models. Our recent observation of mixed (Co,Al) hydroxide precipitates in undersaturated solutions reinforces assumptions that precipitation is an important sorption reaction. A small set of inner-sphere, bidentate, mononuclear sorption reactions, dimeric and monomeric multinuclear complexes, and Co(OH)₂ precipitation approximate uptake and XAFS data well.

(5) Other XAFS Studies of Metal-Ion Sorption Complexes: U(VI) sorbed onto kaolinite was studied using XAFS spectroscopy to develop a molecular-level understanding of sorption of U(VI) by clays in soils and sediments. A parallel XAFS study of U in crystalline solids demonstrated the accuracy of structural information derived from U XAFS, including the second coordination shell around U. Sorption complexes on kaolinite contained the uranyl moiety bonded in IS mode, surrounded by five equatorial oxygens. At pH 6-7 in air, mononuclear U species dominate. At pH 7-7.5 in air, small multinuclear U complexes dominate, with numbers of U atoms similar to those of dominant species in the corresponding aqueous solution. With CO₂ absent and pH 7-8, multinuclear complexes also dominate.

GRANTEE: STANFORD UNIVERSITY

Department of Geological and Environmental Sciences

Stanford, California 94305-2115

GRANT: DE-FG03-94ER14462

TITLE: Development of Fracture Networks and Clusters: Their Role in Channelized Flow in Reservoirs and Aquifers

PERSONS IN CHARGE: D. D. Pollard (415-723-4679; Fax 415-725-0979; E-mail dpollard@pangea.stanford.edu) and A. Aydin (415-725-8708; Fax 415-725-0979; E-mail aydin@pangea.stanford.edu)

Objectives: The objectives of this project are to describe and document the geometry of opening-mode fracture networks and clusters in sedimentary rock, understand the mechanics of their development in relationship to faults and folds, and develop a sound methodology for prediction of spatial variations of permeability in fractured aquifers and reservoirs using an integrated program of field mapping, laboratory simulation, and theoretical analysis.

Project Description: The principal elements of the project are to (1) investigate the variations in orientation and distribution of fracture sets near normal faults and characterize their geometric and fluid flow properties, (2) investigate the mechanisms of fracture clustering and construct conceptual and mechanical models for the impact of fracture clusters on the development of faults and folds in sedimentary rocks, (3) design and carry out laboratory experiments using a brittle-coating technique to investigate the development of fracture networks in folded layered materials, (4) design and carry out laboratory experiments on three-dimensional fracture propagation in blocks of PMMA under mixed-mode loading to understand the relationships between loading and complex fracture geometries, and (5) develop a computer code for predicting the connectivity and spatial density of fractures in a reservoir or aquifer using wellbore data.

Fracture Clusters in the Perturbed Stress Field Near Normal Faults

Results: Over the past year, we focused on the characterization of joints that form approximately orthogonal to normal faults. Classic Andersonian theory invokes a stress regime that produces joints parallel to the strike of faults; however, this theory fails to incorporate the influence of the fault on the stress field. As a fault grows, the surrounding stress field is perturbed so that joints may form with orientations different from Andersonian predictions.

At Arches National Park, Utah, field evidence, such as joint densities and age relationships, supports the hypothesis that joint growth was driven by the stresses induced by faulting. The joints are nearly perpendicular to the faults and are steeply dipping. The spacing of the joints is on the order of the thickness of the unit in which they are contained (Moab Member of the Entrada Formation), and the joints frequently occur as clusters of several closely-spaced joints.

Using a boundary element method computer program and principles of linear elastic fracture mechanics, the stress perturbation around a normal fault does indeed predict joint growth at high angles to faults in the vicinity of the fault tips, where stress concentrations are highest. Three-dimensional rendering of the stress field provides an informative demonstration of the spatial variation in the stress field and associated joints around normal faults.

Fault Development from Localized Shearing of Joint Zones

Results: Faults with offsets ranging from ~1cm to ~150m were mapped in the Valley of Fire State Park, southern Nevada. These maps include the geometric and physical attributes of the associated fractures and provide a large data base for conceptual models of faults and fractures in sandstone and their hydraulic properties. Fracture localization, produced by shearing of preexisting joint zones, results in faults with a highly damaged and fragmented core. The maps enable identification of fragmentation mechanisms and characteristic fracture geometries associated with accumulated slip. Fragmentation occurs at irregularities in the trace of joints subjected to shear, at sheared joint intersections, and in the span between overlapping echelon sheared joint segments.

Two classes of faults produced from sheared joint zones have been identified. One type is associated with right stepping joint zones sheared in a left sense or left stepping joints sheared in a right sense. In these cases fragmentation occurs by failure of intervening rock in compression. The other fault type is associated with right stepping joint zones sheared in a right sense or left stepping joints sheared in a left sense. Fragmentation in these cases is produced by tensile failure of rock spans between sheared joints.

Numerical Modeling of Multilayered Flexures with Fractures

Results: Two mechanisms that produce joint clusters in folds are bending and bedding plane slip. Therefore, fractures are likely to be associated spatially with regions of greater curvature and with larger slip gradients on bedding plane faults. Bedding planes were mapped through the thickness of the aeolian Navajo Formation of East Kaibab Monocline, Utah, and the occurrence of joints related to bedding plane slip and fold curvature were documented. Slip along bedding planes, as evidenced by joint clusters oblique to bedding, develops in the center of the formation. Joints perpendicular to bedding and parallel to the fold axis form near the synclinal hinge.

Numerical experiments used the boundary element method to examine bending of a layer flexed to match the Navajo Formation at Hackberry Canyon. Both uniform and observed distributions of frictional layers were modeled. The numerical models include inelastic frictional slip and the development of opening-mode fractures from points of stress concentration. Slip along horizontal frictional interfaces develops in the center of the layer and opening-mode fractures related to curvature form within the anticlinal and synclinal hinges of the fold. Thus, the first-order numerical results match field observations.

The same numerical code was used to investigate frictional bedding plane slip in the initial stages of fault-related folding. Results show how fold shape depends on dip of the underlying fault; asymmetric anticlines are produced under contraction boundary conditions and asymmetric synclines are produced under extension. At depths as shallow as 1 km there is no significant difference between fold amplitudes in the hanging wall and the footwall. The presence of frictional bedding planes near fault tips encourages development of fault flats from fault ramps.

Fracture Characterization and Prediction from Borehole Data

Results: A computer code, Pred2/3D, for predicting ("rebuilding") subsurface fracture networks using borehole image data provides 2D and 3D relative fracture spatial density, connectivity, and prediction uncertainty maps in the range from tens to hundreds of feet from the wellbore. The underlying physical basis for the prediction comes from laboratory and numerical modeling of the evolution of fracture sets. Reservoir engineers can use the information provided by this computer program to find the best drilling direction, predict the effectiveness of a new well, or assess the future value of an aged well.

The code has been tested using 2D data from experimentally produced fracture networks in a layered brittle material and 3D data from two Mobil horizontal boreholes, two Arco horizontal boreholes in chalk, one North Sea horizontal borehole in chalk, and some USGS testing wells in the Mirror Lake area of central New Hampshire.

GRANTEE: STANFORD UNIVERSITY

Department of Geological and Environmental Sciences

Stanford, California 94305-2115

GRANT: DE-FG03-93ER14366

TITLE: Experimental Investigation of Kinetics and Rheology During Diagenesis

PERSONS IN CHARGE: J. G. Liou (415-723-2716; Fax 415-725-2199; E-mail liou@pangea.stanford.edu) and B. R. Hacker (415-725-0045; Fax 415-725-2199; E-mail hacker@pangea.stanford.edu)

Objectives: Objectives are to determine the dehydration rate of laumontite and its effect on the frictional rheology of laumontite.

Project Description: Two processes of enormous economic consequence occur within the upper to middle crust: (1) the formation, migration, entrapment, and degradation of hydrocarbons and (2) hazardous seismicity. Substantial scientific evidence suggests that both these processes are influenced by devolatilization reactions during diagenesis; however, surprisingly few laboratory studies have been conducted on materials actively undergoing low-grade metamorphism or diagenesis. Because of this, there exists no suitable basis for understanding the rates at which devolatilization occurs and what effects this process has on deformation at shallow to moderate depths in the crust.

We are conducting a coordinated deformation and kinetic study of an important devolatilization reaction, the breakdown of laumontite. Laumontite is a common zeolite whose equilibrium phase relations and room-temperature frictional behavior are well understood. Besides serving as a model system for more complicated rocks, laumontite is an important mineral in its own right, particularly for hydrocarbon fields in sandstones and for fault zones in the crust. Hydrostatic experiments are being conducted to investigate the kinetics and mechanism of laumontite dehydration, and triaxial deformation experiments will enable characterization of the effect of differential stress on the reaction and the effect of synkinematic dehydration on the

mechanical behavior of rock. We anticipate results of significant import for hydrocarbon exploration and recovery and for understanding the strength and seismic potential of crustal fault zones.

Results: We have determined the rate and mechanism of the laumontite &lig; wairakite + H₂O reaction in experiments as long as three months at P_H2O = 100 MPa and temperatures of 350°_450°C. At 350°_400°C, nucleation occurred on the smallest laumontite fragments in the starting material. Growth proceeded by the dissolution of large laumontite grains, transport within the fluid, and reprecipitation of euhedral-subhedral wairakite. At 425° and 450°C, each sample contains two product phases, wairakite and an unidentified plagioclase-like phase. The plagioclase-like silicate is stabilized by the uptake of Na and formed early as ~10 mu-wide skeletal grains along laumontite grain boundaries. The wairakite grains subsequently nucleated and grew into large laumontite grains.

Nucleation rates varied from >10⁷ wairakite grains per square meter of laumontite surface per second at 425°C to >10⁸ m^-2 s^-1 at 450°C; at lower temperatures the volumetric nucleation rate is >10¹¹_10¹² m^-3 s^-1. Growth rates varied from 1.5x10^-11 m/s at 350°C to 2.1x10^-10 m/s at 450°C; the low-temperature data can be fit with an apparent activation energy of 72 ± 13 kJ/mole and a pre-exponential "interface jump distance" of ~1 x 10^-18 m. These rates are comparable to those predicted by Walther and Wood to characterize interface-controlled reactions in silicates under H₂O-

saturated conditions. Extrapolation of our data indicates that the rate-limiting step in the transformation of laumontite to wairakite under natural conditions cannot be interface-controlled growth under H₂O-saturated conditions and must instead be nucleation rate, heating rate, rate of fluid transport, or fluid activity. We have completed nine friction experiments on laumontite powder at temperatures of 23°, 350°, and 400°C. Sliding rates have been varied between 0.01 and 1.0 mm/sec at a controlled H₂O pressure of 100 MPa and a fixed effective pressure of 100 MPa. All experiments so far are baseline measurements of the variations in laumontite friction with increasing temperature in the absence of transformation. At room temperature the strength of the laumontite gouge follows Byerlee's law, as expected, but at 350°_425°C the coefficient of friction is the highest ever measured for a silicate, m = 0.95 &lig; >1. All samples strain harden, so the steady-state value may be higher. Pressure-stepping tests reveal that the strength is sensitive to confining pressure, suggesting

that the deformation is dominantly brittle. This is confirmed by microscopic observation of the samples, which contain Riedel shears spaced 300_400 mu apart and are made up of submicron-sized particles. Transmission-electron microscopy in progress is designed to reveal the deformation/transformation mechanism responsible for this unusual mechanical behavior.

Published Abstracts

Jové, C.F., and Hacker, B.R., 1994, Kinetics of laumontite breakdown during diagenesis: preliminary results. Eos, Transactions American Geophysical Union, v. 75, p. 703.

Hacker, B.R., Blanpied, M.L., Lockner, D.A., and Jové, C.F., 1995, Dehydration and friction: laumontite &lig; wairakite + H₂O. Geological Society of America Abstracts with Programs, v. 27.

Jové, C.F., and Hacker, B.R., 1995, Experimental diagenesis of laumontite breakdown. Geological Society of America Abstracts with Programs, v. 27.

GRANTEE: STANFORD UNIVERSITY

Geophysics Department

Stanford, California 94305-2215

GRANT: DE-FG03-95ER14535/A000

TITLE: Crosshole Seismic Attenuation Tomography and Attenuation Logging in Boreholes

PERSON IN CHARGE: Jerry M. Harris (415-723-0496; Fax 415-725-2032; E-mail harris@pangea.stanford.edu)

Objectives: The objectives of this research are to (1) develop methods of attenuation estimation for crosswell seismic and sonic log data and (2) interpret the attenuation estimates, along with velocity data, for in situ reservoir properties.

Project Description: Attenuation imaging is possible using the frequency-shift dispersion method, where the downshift in the centroid frequency of a seismic wavelet is used as projection data in tomographic inversion. This is analogous to using travel time for velocity imaging. Frequency shift data are less sensitive to effects that severely contaminate amplitudes which might otherwise be used to estimate attenuation. If attenuation is caused by friction between pore fluids and the pore walls, then attenuation measurements may be useful as indirect measurements of fluid permeability. Travel time tomography algorithms will be adapted to invert the frequency-shift data for local attenuation, and, finally, the attenuation images will be interpreted, cooperatively with other data, for in situ permeability near and between boreholes. The project makes extensive use of synthetic modeling of waves in viscoelastic and poroelastic media and real field data from the crosshole and sonic logging geometries.

Results: A model that relates the integrated attenuation and the frequency shifts has been found. For a constant-Q model and a signal spectrum that can be fit

to a Gaussian distribution, the integrated attenuation equals the frequency-shift divided by the variance of the radiated spectrum. To verify this model, synthetic tests were needed. Since the frequency shift method uses information on dispersion, purely numerical modeling methods (e.g., finite differences) could not be used. We developed a semi-analytical modeling method and the corresponding computer codes to simulate the complete viscoelastic wave field in radially layered media. This forward modeling method is efficient and accurate and has no grid dispersion. Complete crosshole seismic surveys are simulated and used for synthetic attenuation tomography. Inversion results reveal that the frequency shift method is capable of imaging the attenuation distribution. Initial tests on field data also show that the attenuation tomograms exhibit good correlation with the geological structures.

We also investigated the effects of scattering from thin layers on the frequency shifts. We found that geometrical spreading caused by refraction around complicated boreholes is highly frequency-dependent (i.e., dispersive) at high logging frequencies. The apparent spreading factor (1/z^p with p<1, p=1, or p>1 for damaged, simple, and flushed boreholes, respectively) alters the frequency shifts in a way that significantly affects the estimation of intrinsic attenuation from the logging geometry.

GRANTEE: STANFORD UNIVERSITY

Geophysics Department

Stanford, California 94305-2215

GRANT: DE-FG03-90ER14152

TITLE: Induced Seismicity

PERSON IN CHARGE: P. Segall (415-725-7241; Fax 415-725-7344; E-mail segall@pangea.stanford.edu)

Objectives: The objective of this project is to develop a fundamental understanding of seismicity associated with energy production.

Project Description: Earthquakes are known to be associated with oil, gas, and geothermal energy production. The goals of the project are to develop physical models that predict when seismicity is likely to occur and to determine to what extent these earthquakes can be used to infer conditions within energy reservoirs. Early work focused on earthquakes induced by oil and gas extraction, which we were able to demonstrate are caused by poroelastic stressing.

Our present work is focused on two problems: (1) earthquakes within geothermal fields, such as The Geysers in northern California and (2) fundamental physics of earthquake nucleation, including effects of dilatancy, rate and state-dependent friction, and shear heating. The former has involved modeling thermoelastic and poroelastic effects of geothermal production and water reinjection, as well as Global Positioning System (GPS) measurements of deformation within and around The Geysers. The latter has involved the development of constitutive laws for dilatancy based on laboratory data and theoretical and numerical analyses of the stability of frictional sliding under conditions appropriate for crustal faulting.

Results: We have developed a simple model of the stresses induced by the injection of cold water into a highly permeable fault zone within a geothermal reservoir. Pore pressures within the fault zone depend on the mass flow rate of injected fluid, duration of injection, and product of permeability and fault zone thick

ness (kh). Maximum pore pressures are on the order of 2 to 3 MPa above ambient for injection rates of 10 kg/s, kh = 1011 m³ (typical values for The Geysers) and injection times of 10 years or less. Advective cooling of rock adjacent to the fault zone induces thermoelastic stresses. The thermoelastic reduction in compressive stress normal to the fault zone is expected to weaken the fault against frictional sliding. For a temperature contrast between the injected water and initial rock temperatures of 200°C (appropriate for The Geysers), the maximum reduction in effective normal stress across the fracture is on the order of 10s of MPa. The thermal stress decreases with increasing distance from the borehole, eventually becoming negative (compressive). The result that thermoelastic stresses dominate pore pressure changes is true even when rapid volume changes due to heating and boiling of the injectate are taken into account.

Observations of surface subsidence and horizontal strain can help constrain the distribution of strains within the reservoir due to changes in temperature and pore-pressure. In 1994 and 1995, 40 geodetic monuments within The Geysers geothermal field were reoccupied with GPS receivers. Precision is sub-centimeter in the horizontal coordinates and 1-2 cm in the vertical. Displacements between 1994 and 1995 are significant at the 95% confidence level and indicate subsidence within the reservoir as well as significant horizontal deformation.

We have developed constitutive laws for dilatant fault gouges that include rate-dependent effects. Using these results together with laboratory-derived fric

tion laws, we determine conditions under which dilatant faults can exhibit stick-slip instabilities. Under isothermal conditions, dilatancy is stabilizing while pore-fluid exchange with the surrounding rocks is de

stabilizing. For adiabatic conditions a shear heating instability competes with frictional weakening to cause stick-slip instability. For the shear heating instability both dilatancy and pore-fluid exchange are stabilizing.

GRANTEE: STANFORD UNIVERSITY

Geophysics Department

Stanford, California 94305-2215

GRANT: DE-FG03-86ER13601.A004

TITLE: Porous Reservoir Rocks with Fluids:

Acoustic and Reservoir Transport Properties

PERSON IN CHARGE: Amos Nur (415-723-9526; Fax 415-723-1188; E-mail nur@pangea.stanford.edu)

Objectives: The objectives of the research are to (1) relate seismic characteristics of hydrocarbon reservoirs and groundwater aquifers to rock, soil, and pore-fluid properties and (2) apply the results to the design and interpretation of in situ seismic measurements to obtain subsurface fluid saturation, flow description, and flow monitoring. The research program involves laboratory measurements, theoretical modeling, and in situ field studies. It spans several projects.

Effect of Gas in the Pore Space

Project Description: Conduct a high resolution shallow seismic refraction and reflection experiment with repeated profiling of every 20 minutes to 1.5 hour to investigate the seismic response of water table excursions (and the air-saturated layer above the water table) in beach sand due to ocean tide.

Results: It has been found that a low velocity layer exists right above the fully saturated sand layer and right below the near-surface dry sand. This low velocity layer is caused by the relatively high density of the almost saturated sand, where only small amounts of gas are present, and its low compressibility, which is close to that of the dry sand. Thus, small gas bubbles trapped in the flooded sand strongly affect the system's seismic response. The history of water level variation is a key control of the seismic response. The results can be used to map water table from which both subsurface flow patterns and relative permeability can ultimately be obtained.

Effect of Liquid in the Pore Space

Project Description: Investigate the seismic and sonic effects of the topology of fluid distribution in porous rocks, as applied to various processes of pore fluid replacement during drilling, well completion, hydrocarbon recovery, and also geological fluid migration.

Results: Whereas Gassmann's equation can be used for perfectly homogeneous rock systems, even slight lithologic inhomogeneity (e.g., clay content spatial variation) may lead to significant saturation inhomogeneity (or patchiness) where Gassmann's equation is not directly applicable. When saturation is patchy, newly derived modified equations have to be used to predict saturation from seismic and to calculate rock-frame properties from well logs. Accounting for saturation inhomogeneity is especially important in the interpretation of repeated surface and crosswell seismic data and sonic well logs for monitoring fluid recovery in two or three phase reservoirs.

Effect of Solids in the Pore Space

Project Description: Investigate the sensitivity of seismic velocities to the amount and location of the solid phase (such as diagenetic cement or gas hydrates) in the pore space.

Results: In diagenesis of typical sandstones and carbonates, cement first precipitates at grain contacts so that the stiffness and hence velocity increase rap

idly with even small decrease in porosity. Permeability, on the other hand, is not very sensitive to filling the grain contacts (at high porosity) and shows only small decrease with decreasing porosity. In the reverse situation, where a solid phase is placed in larger pores and away from the grain contacts, the opposite is true; the velocity is only weakly affected by the decreasing

porosity but permeability decreases rapidly. This latter situation appears to be the case in the formation of gas hydrates under Earth's ocean floor and is likely to be responsible for the trapping of large pools of free methane under impermeable hydrated layers associated with bottom simulating reflectors.

GRANTEE: TEMPLE UNIVERSITY

Department of Chemistry

Philadelphia, Pennsylvania

GRANT: DE-FG03-ER

TITLE: The Surface Chemistry of Pyrite: An Interdisciplinary Approach

PERSONS IN CHARGE: Daniel R. Strongin (215-204-7119; Fax 215-204-1532) and

Martin A. A. Schoonen (Department of Earth and Space Sciences, SUNY, Stony Brook; 516-344-3124; E-mail schoonen@sbmpo4.ess.sunysb.edu)

Objectives: A research program has been initiated to study the surface reactivity of pyrite, the most abundant metal sulfide on Earth. A unique aspect of this research program is that it integrates experimental observations made using low-temperature aqueous techniques and modern surface science probes. Low-temperature aqueous techniques, such as electrophoresis, provide insight into the overall acid-base behavior, charge development, and sorption onto the surface. Modern surface probes, such as high resolution electron loss spectroscopy, photoelectron spectroscopy, low energy electron diffraction, and temperature programmed desorption, provide molecular level information on the geometric and electronic structure of both the chemisorbed molecule and adsorbing surface. With these probes it is possible, for example, to determine the nature of surface groups that are responsible for the acid-base reactions and charge development observed in electrophoresis experiments.

Project Description: In the first stage of this project, the focus will be on understanding the charge development on pyrite and the interaction of its surface with water. Electrophoresis will be used to determine the charge development of pyrite in solutions with known concentrations of inorganic and organic species. Modern surface science probes will be used to interrogate the molecular structure and thermal chemistry of adsorbed species on pyrite. These surface science studies encompass two types of experimental studies. The objective of the first type of surface science study is to investigate the reactivity of an atomically clean "as

grown" surface of pyrite, in the ultra high vacuum (UHV) environment, toward a variety of molecular reactants. Naturally occurring samples of pyrite will be cleaned using a technique developed in our research group (Chaturvedi et al., 1996, Am. Min. 81, 261-264). This cleaning technique allows us to study the reactivity of atomically clean cubic and octahedral surfaces of pyrite and avoids the more extensive structural and chemical changes associated with cleaving, cutting, and polishing samples. The interaction of the clean surfaces with selected molecules (e.g., H₂O, H₂S, CH₃OH) will be investigated in UHV using the array of electron spectroscopic and thermal desorption techniques. The goal of these experiments will be to develop an atomic level view of the structure and reaction chemistry of the chemisorbed species and pyrite surface. The second type of study investigates the surface of pyrite after exposure to solutions of known composition. These experiments provide us with a bridge between the electrophoresis experiments and the UHV studies. To conduct this type of experiment, a transfer cell has been designed that allows exposure of a well-defined pyrite crystal to a solution at room temperature and pressure, followed by transfer of the sample into the UHV environment for analysis with surface science probes. With this transfer cell it is possible to investigate the change in surface composition as a function of solution composition. Our choice of solution compositions for these studies will be guided by the results of the electrophoresis experiments.

Results: The combined results of this interdisciplinary study will provide a detailed picture of the type of functional groups at the pyrite surface, its charge development, the acid-base behavior, and the reactivity of these surfaces. Because of the ubiquity of pyrite

in the natural environment and its importance in industrial applications, the results of this research will be of interest to geochemists, environmental chemists, and material scientists.

GRANTEE: UNIVERSITY OF TENNESSEE

Institute for Rare Isotope Measurements

10521 Research Drive, #300

Knoxville, Tennessee 37932

GRANT: DE-FG05-95ER14497

TITLE: Development of Laser-Based Resonance Ionization Techniques for ⁸¹Kr and ⁸⁵Kr in the Geosciences, II

PERSONS IN CHARGE: N. Thonnard (423-974-9700; Fax 423-974-8289; E-mail nthonnar@utk.edu), T.C. Labotka, and L.D. McKay

Objectives: Objectives are to (1) bring into operation a new analytical methodology for ⁸¹Kr and ⁸⁵Kr, (2) identify performance limitations and implement improvements in reproducibility, accuracy, throughput, and sample size, and (3) initiate research in the geosciences.

Project Description: The ⁸¹Kr and ⁸⁵Kr noble gas radioisotopes, with abundances and concentrations of 10^-12 and 10^-22 in modern water, could contribute to our understanding of processes in the environment, including dating of polar ice and very old groundwater, ocean circulation, and modern water flow patterns. Cosmogenic ⁸¹Kr (2.1x10⁵ year half-life) can date events in the 50,000- to 1,000,000-year time period, while anthropogenic ⁸⁵Kr (10.8-year half-life) is useful in tracing events during the last 50 years. Neither isotope is accessible by accelerator mass spectrometry. A few measurements of ⁸⁵Kr from 200-liter water samples, using decay counting, have been demonstrated. The only ⁸¹Kr measurements from natural samples to date are a handful of results from old groundwater and polar ice, using the laser-based analytical technique under development here. The chemical inertness of the noble gases can simplify interpretation of results. When fully operational, this technique should permit ⁸⁵Kr measurements using only 1- to 5-liter samples and ⁸¹Kr measurements from 10- to 20-liter samples. The technique presently consists of a multi-step process starting with (1) degassing of the sample, (2) separating Kr from the remainder of the gas, (3) a first isotopic enrichment, reducing interfering isotopes by 10⁵, (4)

second isotopic enrichment of 10³, and (5) detecting the rare krypton isotope in a time-of-flight mass spectrometer utilizing resonance ionization. As there are only a few thousand analyte atoms in the sample, the sensitivity, element selectivity, and immunity to interferences of resonance ionization are required. A detection limit of 100 ⁸⁵Kr atoms had been demonstrated earlier in the final mass spectrometer. Required still is characterization of the efficiency, accuracy, and blank level of each step; stabilization and automation of operating parameters in steps (1), (2), and (3); complete redesign of step (4); and improvement to the laser and data acquisition systems in step (5).

Results: Most of the effort this past year has been to steps (3), (4), and (5). The first enrichment system (3) uses a 15-inch velocity filter and plasma ion source in a closed, gas recirculating system. The ion-source controller has been rebuilt, a Hall-Effect probe installed, an isotope dilution calibration system designed, and the ion path completely realigned. These changes should significantly enhance accuracy, stability, reliability, and safety. Gold coating of the rods in the quadrupole-based second enrichment system (4) and redesign of the collector and detection hardware should breathe a little more life into this obsolescent system until the design of the magnetic sector-based replacement is finished. The resonance ionization time-of-flight mass spectrometer (5) has a significantly improved data acquisition system, a completely new sample introduction system, and enhanced vacuum system.