GRANTEE: NORTHWESTERN UNIVERSITY

Department of Civil Engineering

Evanston, Illinois 60208-3109

GRANT: DE-FG02-93ER14344

TITLE: Shear Strain Localization and Fracture Evolution in Rocks

PERSON IN CHARGE: J.W. Rudnicki (847-491-3411; Fax 847-491-4011; E-mail jwrudn@nwu.edu)

Objectives: To obtain an improved understanding of the occurrence, development, and evolution of zones of shear localization (faults) in rocks and their relation to the macroscopic constitutive description, especially that governing multiaxial response, and microscale mechanisms of deformation.

Project Description: Because of the significance of fractures to energy production, waste disposal, and mineral technologies, prediction of their causative stresses, location, orientation, thickness, and spacing is important. This project examines the applicability of a theory of localization that describes faulting as an instability of the constitutive description of homogeneous deformation. Because the predictions depend strongly on the constitutive parameters governing abrupt changes in the pattern of deformation, theoretical work is being done to develop a more realistic constitutive model based on the growth and interaction of microcracks and resulting increase in overall compliance of the solid. This constitutive relation is calibrated by comparison with axisymmetric compression tests and then used to predict the response in more complex experiments (compression-torsion) with abrupt changes in the pattern of loading. Comparison of numerical studies with experiments addresses the effects of realistic geometries and boundary conditions.

Results: Abrupt changes in the pattern of deformation, such as occur in localization and in the combined compression versus torsion tests (Olsson, Mech. Mat., 1995), cause unloading of open cracks in some orientations. Consequently, the response to abrupt changes in the pattern of deformation depends strongly on the unloading response. A phenomenological description of unloading and reloading has been implemented in a microcrack model. Implementation is guided by understanding gained from laboratory observations and from a detailed study of the unloading response predicted by a model in which opening and extension of tensile wing cracks is driven by sliding on an oblique crack with Coulomb friction (Jeyakumaran and Rudnicki, GRL, 1995). Comparison of the predictions with observations suggests that there is rapid partial rehealing of small microcracks or some other source of resistance to crack reopening.

Experimental results on simulated fault gouge (e.g., Marone and Kilgore, Nature, 1993) have suggested that the deformation response of the fault depends on processes of shear localization within the gouge. A simple theoretical model suggests that the development of localized deformation in hydraulically isolated fault zones is strongly affected by coupling between internal fluid flow and slight variations in porosity.

GRANTEE: UNIVERSITY OF NOTRE DAME

Department of Civil Engineering and Geological Sciences

Notre Dame, Indiana 46556-0767

GRANT: DE-F602-93ER14391

TITLE: Energy Partitioning of Seismic Waves in Fractured Rocks

PERSON IN CHARGE: Laura J. Pyrak-Nolte (219-631-8377; Fax 219-631-9236; E-mail pyrak-nolte @nd.edu)

Objectives: The primary objectives of the proposed research are to investigate through numerical and laboratory investigations: (1) partitioning of seismic energy between body waves, guided waves, and scattered waves produced by sources of finite size that are transmitted, reflected, and channeled along single and multiple fractures, (2) techniques for exciting these waves, (3) effects of spatial variations in the mechanical properties along the fracture on seismic wave propagation, and (4) effects of a finite fracture on seismic waves.

Project Description: Rock masses contain fractures and discontinuities on all length scales that affect the mechanical stability of a rock mass and the flow of fluids through a rock mass. A goal of site characterization for waste isolation or mineral exploration is to detect and characterize the hydraulic and mechanical properties of fractures using seismic techniques. Seismic data are often difficult to interpret because of wave conversions that occur at interfaces. Converted wave modes can arise when seismic waves are propagated through a fractured rock mass and these waves inhibit direct interpretation of the received signals. Because the existence of interface waves has not previously been taken into account in seismic data interpretation, it is important to understand how non-welded interfaces, such as fractures and joints, give rise to interface waves. In this research project, the partitioning of seismic wave energy into body waves and interface waves caused by fractures (non-welded contacts) is studied through laboratory experiments and numerical analysis.

Results: The specific results for the third year of this project include (1) experiment evidence of compressional-mode interface wave that co-propagates with the bulk compressional wave, (2) generation of Rayleigh-mode interface waves for sources at oblique angles of incidence to the fracture, and (3) determination of the effect of asperity spacing on Bragg scattering and on the limits of applicability of the displacement discontinuity theory for interface waves. The compressional-mode interface wave propagating along a fracture was observed in a fracture in limestone. This compressional-mode interface wave is localized to the region surrounding the fracture and has particle motions containing both transverse and longitudinal displacements. As stress across the fracture increases, the wave energy shifts to higher frequencies and to earlier arrival times and is accompanied by a decrease in the transverse displacement and an increase in the longitudinal displacement of the particle motion.

Previously, interface waves were generated by straddling the fracture with seismic transducers. The ability to generate interface waves for off-fracture placement of the source was confirmed through measurements of particle motion of the interface waves and the behavior of the waves under stress. For off-fracture generation, interface waves were only observed when the shear wave was incident at or greater than the critical angle for shear-wave to compressional-wave conversion.

GRANTEE: UNIVERSITY OF OKLAHOMA

School of Geology and Geophysics

Norman, Oklahoma 73019

GRANT: DE-FG05-913414209

TITLE: A Study of Hydrocarbon Migration Events: Development and Application of New Methods for Constraining the Time of Migration and an Assessment of Rock-Fluid Interactions

PERSONS IN CHARGE: R. D. Elmore (405-325-3253; Fax 405-325-3140; E-mail delmore@uoknor.edu) and M. H. Engel

Objectives: The objective of the research is to test and refine a paleomagnetic method for dating hydrocarbon migration and maturation of organic matter. The specific objectives include field tests of the dating method and laboratory simulation experiments to better constrain the mechanisms for the precipitation of authigenic magnetite under a variety of geologic conditions.

Project Description: Investigations of fluid migration are commonly hindered by a lack of temporal control. There is little doubt that the ability to constrain the time of oil migration would be of significant benefit for exploration. The paleomagnetic dating method is based on a genetic connection between hydrocarbons/organic matter and precipitation of authigenic magnetite. Isolation of the magnetization carried by the magnetite and comparison of the corresponding pole position to the apparent polar wander path allow the timing of diagenetic events to be determined. The research involves paleomagnetic field tests of the method and laboratory simulation studies to constrain appropriate chemical and physical conditions for magnetite authigenesis.

Results: Field and laboratory studies completed this past year provide further evidence for chemical remanent magnetizations (CRMs) in sedimentary rocks that resulted from fluid migration and from maturation of organic-rich sediments. For example, paleomagnetic and rock magnetic analyses of bitumen-impregnated limestones from the Kimmeridgian-Portlandian

Asphaltkalk deposits of the Hils Syncline near Holzen (northern Germany) contain a late Cretaceous/early Tertiary age CRM that resides in magnetite. A paleomagnetic fold test indicates that the CRM was acquired after folding. The stable CRM was not observed in nonbitumen-impregnated limestones from the same localities, indicating a connection between hydrocarbon migration and the precipitation of authigenic magnetite. The age of the CRM is consistent with that suggested for petroleum generation (late Cretaceous) in this region. A second case study consisted of Mississippian age limestones in southern Indiana that contain hydrocarbons seeping from large stylolites. The stylolites are conduits for hydrocarbon migration. The limestones contain a modern magnetization and a southeasterly and shallow Late Paleozoic CRM, both residing in magnetite. Magnetic intensity of the modern component decreases away from the stylolites and is interpreted to result from recent hydrocarbon migration. The occurrence and intensity of the Late Paleozoic CRM are independent of the stylolites. Finally, we have completed a study of organic-rich, Jurassic age limestones (Blue Lias) that occur in a fault zone along the southern margin of the Bristol Channel Basin at Kilve, West Somerset, England. The Bristol Channel Basin is one of several fault-bounded sedimentary basins in the Celtic Sea/Irish Sea region in which the timing of deformation, maturation of organic matter, and hydrocarbon migration is uncertain. Deformation at Kilve consists of numerous normal faults

and associated folds. Fluid flow within the fault zone is indicated by widespread calcite veining associated with the faults. The veins contain hydrocarbons. The veins and limestones both contain a northerly and down component that resides in magnetite. This magnetization is a CRM that was acquired when the vein calcite precipitated. A revised pole position suggests remanence acquisition in the late Cretaceous/early Tertiary. Fold tests on several folds associated with the faults suggest the presence of a pervasive CRM in the limestones that is synfolding. The timing of alteration in the limestones, however, is difficult to determine because of rotations on the faults and the problem with determining paleohorizontal. Preliminary analyses of the same limestones on the northern margin of the Bristol Channel Basin near Penarth (Wales) indicate the presence of hydrocarbons and a magnetization (CRM) similar to that in the rocks on the southern margin of the basin.

This past year our test for a relationship between maturation of organic matter and magnetite authigenesis focused on limestones adjacent to a dike on the Isle of Skye, Scotland. Paleomagnetic results indicate a southerly and negative magnetization in a 0.9-meter-thick Tertiary dike and in the organic-rich limestones within

a meter of the dike. With increasing distance from the dike, the limestone does not contain a stable magnetization. The dike emplacement has clearly caused a magnetization in the surrounding limestone. Immediately adjacent to the dike, the magnetization has high maximum unblocking temperatures and is thermochemical in origin. These unblocking temperatures decrease to approximately 300°C at 110 cm. Magnetic susceptibility and NRM intensities are higher around the dike compared to background levels away from the dike. This suggests the addition of magnetite, perhaps by chemical processes. The magnetic susceptibility decreases to background levels within the first 20 cm away from the dike but then increases near 25 cm and stays above background out to 110 cm. Interestingly, it has been reported (Bishop and Abbot, 1993, Geochim. Cosmochim. Acta 57, 3661-3668) that the oil generation window begins at a distance of about 25 cm from the dike. These preliminary results suggest a possible connection between the magnetization and the maturation of hydrocarbons adjacent to the dike. We are currently performing rock magnetic measurements and conducting petrographic studies to determine whether a CRM is present in the 25-110 cm interval.

GRANTEE: OREGON STATE UNIVERSITY

College of Oceanic and Atmospheric Sciences

Ocean Administration Building 104

Corvallis, Oregon 97331

GRANT: DE-FG03-96ER14595

TITLE: Multi-Station Magnetotellurics

PERSON IN CHARGE: Gary D. Egbert (503-737-2947; Fax 503-737-2064; E-mail egbert@oce. orst.edu)

Objectives: The principal objective of this project is to develop and test new methods for collecting and processing remote reference magnetotelluric data in areas with significant cultural noise, with special emphasis on the "dead band" (approximately 0.1-10.0 Hz) where signal levels are low and noise levels are high.

Project Description: The project has three aspects: (1) In the first year, data were collected in a series of three station MT arrays that span conditions from very noisy (just south of San Jose, CA) to very remote sites. Our emphasis was on collecting a large volume of dead band data at a small number of sites. Multiple remote sites at varying distances and in various combinations of noisy and quiet were included to allow us to characterize spatial and temporal properties of signal and noise and to test different approaches to remote reference data acquisition. (2) New approaches to remote reference data processing are being developed. We are adapting multivariate and robust statistical methods to the special problems of MT data processing in the dead band. (3) Using the multistation MT (MSMT) data from our experiment and other compilations of remote reference MT data, we are testing and comparing various strategies for collecting and processing remote reference MT data.

Results: Efforts over the past year have focused on making the multiple station processing program (MULTMT) useful for routine MT data processing. Initial documentation for MULTMT has been prepared, the user interface has been improved, and formats of output files have been standardized so that commonly available plotting packages can be used for displaying results. A preliminary set of programs and documentation has been made available to interested MT researchers via MTNet (http://www.cg.NRCan.gc.ca/mtnet/mtnet.html). A manuscript describing the new processing methods has been submitted for publication.

We are also continuing efforts to test and further develop multiple station processing methods. Efforts in this direction have included collaborative work with H.F. Morrison and students at U.C. Berkeley, applying multivariate processing methods to data from two experiments: a natural source induced polarization (IP) array experiment conducted at Battle Mt., Nevada, and an ongoing electromagnetic earthquake precursor monitoring experiment being conducted at two sites on the San Andreas Fault in California. We are also extending previously developed methods, in particular focusing on improving error estimates and on more rigorous statistical tests for the presence of coherent noise in MT array data.

GRANTEE: PENNSYLVANIA STATE UNIVERSITY

College of Earth and Mineral Sciences

University Park, Pennsylvania 16802

GRANT: DE-FG02-95ER14547

TITLE: Dissolution Rates and Surface Chemistry of Feldspar Glass and Crystal

PERSONS IN CHARGE: Susan L. Brantley (814-863-1739; Fax 814-863-7823; E-mail brantley@essc.psu.edu) and Carlo G. Pantano (814-863-2071;

Fax 814-865-0016; E-mail pantano@ems.psu.edu)

Objectives: This project aims to develop dissolution/surface hydration models for feldspar that can be used to relate laboratory data to geoenvironmental field systems.

Project Description: The key issue in relating experimental feldspar dissolution rates to dissolution of feldspar in soils, aquifers, and watersheds concerns the nature of the surface. In this program, the flow conditions, form of the feldspar, and solution composition in laboratory experiments are being varied and correlated with direct measurement of the dissolution rate and the surface layer characteristics. These relationships will provide a road map that can be used to evaluate the more complex dissolution behavior of minerals in the natural environment. The hypothesis is that the surface is the link. This requires the ability to characterize surfaces of naturally weathered minerals with a precision and reliability that can be used to back-calculate how those surfaces formed. Thus, the laboratory samples include glassy and crystalline feldspars in both granular and bulk forms. The dissolution rates are measured in static and in flow at neutral to basic pH as a function of dissolved aluminum. The surface analytical methods rely on XPS and depth-profiling methods that can be applied to both laboratory and natural specimens. The ability to distinguish leaching and surface adsorption/precipitation is critical. These findings will provide the basis for interpretation of the surface analyses of natural feldspars so that the factors controlling dissolution in the field can be determined.

Results: In the first year of the project, the following tasks were completed: design of the flow-through dissolution experiments, establishment of a protocol for surface analysis of feldspars, and initial comparison of natural and laboratory-dissolved feldspars.

The completed static experiments were designed to investigate dissolution of feldspar under conditions of near neutral pH and close to equilibrium. These experiments were run at pH 7.3 to 8.3 with albite and potassium feldspar. The crystals were allowed to dissolve for up to 2200 hours. After 1700 hours of dissolution, a few samples were removed and were analyzed using X-ray photoelectron spectroscopy to determine Al/Si and Na/Si ratios of the surface layers. For all experiments, solution chemistry gave no indication of ongoing dissolution (the rate was too small to be measured because of the small surface area/water volume ratio of the experiment). In addition, no evidence for dissolution was observed under SEM; however, XPS analysis of the investigated surfaces showed a clear change in the Na/Si and Al/Si surface ratios. The observed decrease in the Na/Si ratio of the surface may be related to either Na leaching or to Si precipitation on the albite surface. Similarly, the decrease in Al/Si ratio of the near equilibrium experiment could be related to either aluminum leaching or Si precipitation. Because the atomic percent Si in the sample dissolved in solution with 37.1 ppm Si was increased above that of the unleached starting crystal, the preliminary conclusion is that a Si back-reaction has occurred. Such a back-reaction may

explain the observed decrease in Al/Si ratio; however, some Al and Na leaching must also have occurred to explain the observed ratios.

In contrast, polished samples of Quebec albite that were buried 15 cm deep in a Pennsylvania soil show an opposite trend: increasing Al/Si ratio. An increasing Al/Si ratio was also found on the Norwegian moraine feldspar samples. It appears that natural samples mani

fest either (1) increased Al surface content or (2) decreased Si content. XPS analysis of Cape Cod feldspars shows a similarly high Al/Si ratio on the surface. Further investigation of this consistent discrepancy between naturally dissolved feldspars (high Al/Si surface ratio) and laboratory-dissolved feldspars (low Al/Si surface ratio) is necessary in order to better understand controls on natural feldspar weathering.

GRANTEE: PENNSYLVANIA STATE UNIVERSITY

Ore Deposits Research Section

University Park, Pennsylvania 16802

GRANT: DE-FG02-96ER14634

TITLE: Zeolite Thermodynamics and Kinetics

PERSONS IN CHARGE: H. L. Barnes (814-865-7573; Fax 814-863-2001; E-mail barnes@geosc.psu.edu) and Rick T. Wilkin (814-865-3565; Fax 814-863-2001; E-mail rwilkin@geosc.psu.edu)

Objectives: The goal of this project is to obtain data from hydrothermal experiments pertaining to the basic thermochemical properties and kinetic factors that govern dissolution and crystallization of zeolite phases, especially clinoptilolite, mordenite, and analcime. These data are needed to understand chemical and physical processes that occur in shallow crustal environments during the low-grade metamorphism of volcanic and sedimentary rocks.

Project Description: The project includes both closed- and open-system experimental investigations on high-purity natural zeolites. Closed-system experiments involve measuring zeolite solubilities at temperatures to about 300°C and pressures to about 0.20 kbar. Reversible solubility measurements are made on analcime and cation-exchanged varieties of clinoptilolite and mordenite so that thermodynamic data are acquired for Na-, K-, Ca-, and Mg-end-member compositions with a fixed water content and Si/Al ratio. The solubil

ity measurements provide a basis for relating measurements of dissolution and precipitation reaction rates to measured or specified departures from metastable equilibrium conditions. Kinetics of congruent zeolite-water reactions are tracked using a hydrothermal flow-through system. The flow-through system is advantageous because reaction rates can be measured under conditions of fixed fluid composition, flow rate, temperature, and pressure. In addition to providing basic thermodynamic and kinetic parameters, the experimental data permit kinetic evaluations of zeolite transformations, for example, clinoptilolite-analcime-albite. Although the mechanisms of zeolite transformations remain equivocal, such reaction paths define the transition from zeolite to greenschist metamorphic regimes and are characterized by the release of water and silica and large, negative molar volume changes.

Results: This project has only recently been initiated.

GRANTEE: PRINCETON UNIVERSITY

Department of Geological and Geophysical Sciences

Princeton, New Jersey 08544

GRANT: DE-FG02-85ER13437

TITLE: Thermodynamics of Minerals Stable Near the Earth's Surface

PERSON IN CHARGE: A. Navrotsky (609-258-4674; Fax 609-258-1274; E-mail alex@geo.princeton.edu)

Objectives: The goals of the project are to increase both the data base and the fundamental understanding of the thermodynamics of volatile-bearing mineral phases (amphiboles, micas, clays, zeolites, carbonates) important to surficial, sedimentary, and shallow crustal processes.

Project Description: Using high temperature solution calorimetry, this research determines the enthalpies of formation of hydrous minerals and carbonates. Systematics in energetics of ionic substitutions are sought in order to predict the thermodynamics of complex multicomponent minerals. Mixing properties of mica, amphibole, clay, zeolite, and carbonate solid solutions are also studied.

Results: Carbonates: Study of a largely disordered series of synthetic ankerites (Mg, Fe) Ca(CO₃)₂ and of more ordered natural samples confirm significant differences in enthalpy between the two groups of materials. These are consistent with an enthalpy of disordering of MgCa(CO₃)₂ of 20-25 kJ/mol and

an enthalpy of disordering of FeCa(CO₃)₂ of about 10 kJ/mol. The rather large enthalpy of disordering of dolomite is supported both by the systematics of solid solution energetics derived by Davies and Navrotsky (1982) and by a calorimetric study of a metastable apparently totally disordered CaMg(CO₃)₂ sample made from aqueous solution. The latter sample is energetically less stable than a mixture of huntite CaMg₃(CO₃)₄ and Ca-rich dolomite to which it decomposes on heating. The system SrCO₃-CaCO₃ has been studied by calorimetry.

Zeolites: Work on the Ca-zeolites (laumontite, leonhardite, yugawaralite, warakite) and their ion-exchanged forms has been published. The energetics of ion exchange and hydration in other zeolites, including chabazites, is under investigation. The energetics of the Si = Al + Na substitution is similar in faujasites and aluminosilicate glasses.

Clays: A calorimetric study of kaolinite, dickite, smectites, and illites has commenced.

GRANTEE: PURDUE UNIVERSITY

Department of Earth and Atmospheric Sciences

West Lafayette, Indiana 47907

GRANT: DE-FG02-93ER14365

TITLE: Rupturing and Ground Deformation During the 28 June 1992 Landers, California, Earthquake

PERSON IN CHARGE: Arvid M. Johnson (317-494-3259; E-mail gotesson@omni.cc.purdue.edu)

Objectives: This project has the overall objective of understanding the form and significance of surface rupture produced by earthquake sequences. Specific objectives are to describe fracturing, strains, and other manifestations of broad belts of ground rupture during earthquake sequences (Loma Prieta, Landers, and Northridge); to mechanically analyze those structures, where appropriate; and to relate the observations and analyses to earthquake processes. The specific project funded here is our research on the Landers earthquake.

Project Description: Our project has been to document, via detailed maps, the ground rupture and deformation in earthquake areas and to relate them to the broader processes of earthquake sequences. At Landers in 1992-1994, we mapped about 10 km of rupture zones along three faults at scales of 1:400 or more detailed, including rupture belts associated with growth of a tectonic ridge, Tortoise Hill ridge, along the Emerson fault zone and rupture belts along the Kickapoo and Homestead Valley fault zones. In 1995 we used photogrammetry, precision land surveying, and leveling to determine deformations at different scales in the vicinity of the rupture zones and across the tectonic ridge. We resurveyed a network of 46 benchmarks of horizontal control points established during the 1970s by Southern California Edison that extended across the Emerson fault zone. The network includes a grid of benchmarks set every 1/4 mile and is supplemented by 30 wing points set for elevation control. We also used photogrammetry to survey a ladder of quadrilaterals

across Tortoise Hill ridge. The surveys show, for the first time, details of the determination of horizontal strains and differential vertical displacements over a broad area across a fault zone and within the fault zone and included a growing ridge.

In the 1994 Northridge earthquake sequence, we were presented with a different opportunity to study ground deformation during an earthquake. Since the Northridge earthquake sequence occurred in a highly-developed metropolitan area, land survey data and therefore detailed deformation data are available that do not exist in the more rural settings in 1989 at Loma Prieta and in 1992 at Landers. In 1995 and 1996, we have been analyzing hundreds of City of Los Angeles survey records on relative positions of street-intersection monuments in parts of the San Fernando Valley so that we could describe the strains and identify areas or belts of elevated strains. We have been relating the survey data to damage to streets, sidewalks, sewers, and houses.

Results: Our results have been more exciting than we could have predicted. Our mapping and analysis of ground rupture in the epicentral area of the Loma Prieta earthquake sequence and our analysis of strains and changes in uplift of ground at Northridge have together documented, for the first time, that many faults or shear zones with different kinematic signatures move coactively with the fault responsible for the main shock during an earthquake sequence and cause much of the damage to man-made structures during an earthquake. At Landers we have rediscovered a phenomenon ap

parently described earlier only by G.K. Gilbert in the 1906 earthquake, that intense ground rupture along the main earthquake-fault that intersects the ground surface occurs across a broad zone, 50 to 200 m wide. Any structures within this zone can be severely damaged. At Landers we have documented, for the first time, actual growth of a tectonic ridge. Although such growth was suspected based on previous studies, nobody had observed it. Our detailed surveys at Landers documented growth of a ridge by one meter as the fault zone slipped three meters. Furthermore, we obtained new insight into the origin of ridges along strike-slip fault zones. Tortoise Hill ridge is along a relatively straight segment of the Emerson fault zone, so it cannot be a result of a left-step or bend on a right-lateral fault zone (or vice versa), which is the current explanation for such phenomena. We suggest that Tortoise Hill ridge, like analogous ridges that occur in landslides, grew as a result of localized dilation of material within a belt of shear zones along the Emerson fault zone.

We suspect that many earthquakes occur because the seismogenic structure, within which the main shock occurs, is growing rapidly, as happened for example for the anticline at Kettleman Hills. The most far-reaching result of our research is that, at Northridge, we have identified a specific seismogenic structure; it is a heart structure, consisting of a basin and a high on either side. The highs are the Santa Monica Mountains on one side and the Santa Susana Mountains on the other, and the basin is the San Fernando Valley. Movement along a limb of the dish fault responsible for the heart structure produced the main shock and coactive movement along many smaller faults within this structure and was apparently responsible for much of the damage during the Northridge earthquake sequence.

This year we have written drafts of two papers on Tortoise Hill ridge and two on coactive faults at Northridge, which are in review as U.S. Geological Survey open-file reports. We will publish shorter versions in referred journals.

GRANTEE: RENSSELAER POLYTECHNIC INSTITUTE

Department of Earth and Environmental Sciences

Troy, New York 12180-3590

GRANT: DE-FG02-95ER144532

TITLE: Transport Phenomena in Fluid-Bearing Rocks

PERSON IN CHARGE: E. B. Watson (518-276 8838; Fax 518-276-8627; E-mail watsoe@rpi.edu)

the presence of aqueous fluid. After recovery from the piston-cylinder apparatus, the samples are mounted for permeability characterization at room conditions using conventional gas-flow techniques.

Results: During the past year, the investigation of silica solubility and diffusion in H₂O at 0.5-1.5 GPa and 530°-880°C was brought to completion. At 1.0 GPa, the diffusivity is given (in m²/s) by

D_SiO2 = (4.58x10^-5)exp(-6684/T)

corresponding to an activation energy of 55.6 kJ/mole. This equation is consistent with diffusion behavior predicted on the basis of the Stokes-Einstein equation, which relates the diffusivity to the size of the diffusant and the viscosity of the diffusion medium. Interestingly, the equation above specifies aqueous SiO₂ diffusivities that approach the rate of heat transport in rocks, suggesting that diffusion in fluid-bearing rocks is a highly effective means of geochemical transport even in the absence of fluid advection.

Several additional findings resulted from completion of this part of the project: (1) the solubility data for SiO₂ in H₂O obtained in the present study are in substantial agreement with previously-published values, (2) the effect of pressure on diffusion of aqueous SiO₂ is relatively small in the 0.5-1.5 GPa range, with D_SiO2 decreasing by no more than 0.5 log units over a 1-GPa increase in pressure, and (3) Soret (thermal diffusion) effects are immeasurably small in silica-saturated H₂O at 1 GPa with a temperature gradient of ~10°/mm.

Objectives: The objective of this project is to shed light on chemical transport in the Earth through (1) development and implementation of a technique for measuring mineral solubilities and diffusivities of dissolved mineral components in aqueous fluids at extreme P-T conditions (up to 3 GPa and 1200°C) and (2) characterization of the grain-scale permeability of fluid-bearing rocks under conditions of chemical and mechanical equilibrium.

Project Description: Deep in the Earth, fluid-assisted geochemical transport is controlled mainly by: 1) the solubilities of rock components in the fluid of interest, 2) diffusion characteristics of the dissolved solutes, and 3) the permeability of the rock to fluid flow. (Under circumstances where P_fluid>>P_total and an open fracture system is not sustainable, the permeability of interest is that dictated by the equilibrium microstructure of the rock.) All three of these properties are poorly constrained. Solubility data are scarce at pressures in excess of 1 GPa, and information concerning solute diffusion and rock permeability is virtually nonexistent. The Rensselaer project involves the development and implementation of techniques to characterize these key properties.

The principal methodology for the solubility and diffusion measurements is experimentation at high pressures (0.5-3.0 GPa) and temperatures (500°-900°C) in a conventional solid-media, piston-cylinder apparatus, using noble-metal capsules (or cells) developed for the purpose. In the permeability study, the piston-cylinder apparatus is used to fabricate metal-jacketed "rock" samples exhibiting near-equilibrium microstructure in

An investigation of the solubility of CaF₂ in water at 1 GPa and 500°-900°C was undertaken during the past year as a prelude to diffusion experiments like those completed for SiO₂. The goal is to see whether diffusion in H₂O of simple ions like Ca²⁺ and F^- is still faster than SiO₂, which presumably diffuses as a relatively large anionic complex (Si[OH]₄). Exploratory runs involving solubility and transport of TiO₂, ZrSiO₄, clinopyroxene, and basaltic glass components were also made.

The grain-scale permeability side of the project was initiated by synthesis (at 1 GPa and 850°C) of a series of porous CaCO₃ rocks (i.e., marbles) characterized by near-equilibrium microstructure and a grain size of

100+ microns. Water present during the syntheses resulted in interconnected, fluid-filled porosities ranging from ~1% to 13%. The Ni-jacketed cylindrical samples were recovered from the high-pressure apparatus and sawn open at both ends, allowing their permeabilities to be characterized by the transient-pulse, gas-flow technique. The resulting permeabilities vary in proportion to ³.

A series of porous synthetic quartzites has also been synthesized; in these samples, the fluid geometry is characterized by a lower dihedral angle than that for the CaCO₃ system. Porosity measurements on these samples are in progress.

GRANTEE: RICE UNIVERSITY

Geology and Geophysics

Houston, Texas 77251-1892

GRANT: DE-FG03-95ER14552

TITLE: Transition Metal Catalysis in the Generation of

Petroleum and Natural Gas

PERSON IN CHARGE: Frank D. Mango (713-497-0384; Fax 713-497-4874)

Objectives: It is proposed that the light hydrocarbons in petroleum, including natural gas, are formed catalytically through the condensation of n-alkenes and hydrogen in fine-grained carbonaceous sedimentary rocks. The transition metals are suggested as the catalytic agents. The objective of this research is to test this hypothesis and to explore the catalytic properties of transition metals under realistic geologic conditions.

Project Description: Various natural sources of transition metals, including the asphaltene fraction of petroleum and source rocks, are being analyzed for catalytic activity. Pure transition metal complexes (metal porphyrins, acetylacetonates, oxides, and sulfides) are also under study. Reactions are being conducted in gas manifold systems under steady-state and batch reactor conditions and products analyzed by high-resolution gas chromatography.

Results: Petroleum source rocks are catalytic in the conversion of hydrogen and olefins into natural gas, marking the first time that natural gas has been generated in the laboratory under mild conditions (Mango et al., Nature, 368, 535, 1994). The evidence supports transition metals as the active agents (Mango, Ad. Org. Geochem., in review). Various metal compounds in the pure state show the same levels of catalytic activity as sedimentary rocks, and the products are identical. The metal oxides (V, Co, Fe, and Ni) are particularly active and remarkably robust. Activity is unaffected by typical poisons, like air, water, CO, or CO₂, and kinetic studies suggest high levels of catalytic activity at all stages of catagenesis. Nickel oxide promotes the formation of n-alkanes in addition to natural gas (NG),

demonstrating the full range of the hypothetical catalytic process:

[NiO*]

n-C_x⁼ + H₂ &lig; NG + n-C₅ + n-C₆ + ... n-C_x-1

The cyclohexanes and cyclopentanes are a challenge in any theory on the origin of light hydrocarbons. They are major components of oil, yet their origin has remained obscure. Unlike the higher polycyclic biomarkers, they are without biological parents and it is unlikely they could be degradation products. It was suggested that they might be catalytic products, formed from kerogen-derived dienes (Mango, Geochim. Cosmochim. Acta., 56, 553, 1992). Our results with NiO support this view. Under the same conditions giving natural gas, NiO promotes the cyclization of dienes to cyclohexanes and cyclopentanes. For example, 1,6 heptadiene gives methylcyclohexane and the various isomers of dimethylcyclopentane seen in petroleum:

[NiO]

C₇H₁₂ + H₂ &; +

We have now demonstrated the primary tenets of the original hypothesis, namely that the light hydrocarbons in petroleum - n-alkanes, iso-alkanes, cycloalkanes, and natural gas - can be formed through the catalytic action of transition metals in sedimentary rocks. These results support the view that transition

metal catalysis is probably a major pathway through which oil and gas is formed in the earth. This catalytic process is, to our knowledge, largely unprecedented,

and attention is now directed to its scope and mechanistic details.