CONTRACTOR: LAWRENCE LIVERMORE NATIONAL LABORATORY

University of California

Livermore, California 94550

CONTRACT: W-6405-ENG-48

CATEGORY: Geophysics and Earth Dynamics

PERSON IN CHARGE: F. J. Ryerson

A. Effects of Heterogeneity on the Fracture of Rock (S. C. Blair [510-422-6467; Fax 510-423-1057; E-mail blair5@llnl.gov], joint research with L.R. Myer [Lawrence Berkeley Laboratory; 510-486-6456])

Objectives: The objectives of this research are to understand how microscale (or grain scale) heterogeneity affects macroscopic mechanical behavior of rocks, study the process of progressive fracture of rock in compression, and evaluate the role of crack interaction in rock deformation and fracture. Moreover, this work also aims to investigate the relationship of static to dynamic moduli (in collaboration with LBL) and to characterize the role of heterogeneity at a variety of scales in fracture and the scaling properties of rocks.

Project Description: This project is concerned with simulation of rock deformation and fracture at grain and larger scales. During the initial phase of the project, a 2-D field-theory model for rock fracture was developed and then used to determine how heterogeneity in different microscale parameters affects behavior in the simulated compression tests and how macroscopic stress strain behavior is related to the formation of cracks. During the past year, a new 3-D model for deformation and fracture has been developed. This is a lattice-spring model in which elastic properties of individual volume elements (voxels) in a material are assigned separately. Stresses are transmitted by linear springs that connect the nodes surrounding each voxel.

Results: Results from the 2-D simulations show that the model exhibits an implicit size effect (inverse power law) that closely matches laboratory and field data for the dependence of sample strength on sample size and that local heterogeneity in geometry and strength may lead to crack interactions along macroscopic surfaces that are similar to those caused by shear localization and that are often observed in real experiments. The model was also used to relate macroscopic stress strain behavior in compression to the formation of microcracks and to strain energy due to cracks. Two patterns of cracking were found, including spatially uncorrelated, noninteracting cracks that have little effect on macroscopic deformation and spatially correlated interacting cracks associated with extension or linking of cracks and fractures. These cause strain softening and large changes in crack strain energy. The 3-D model has been used with images of bone microstructure obtained from X-ray tomography to estimate deformation and determine likely models of failure. Results indicate that this technique can be used to efficiently calculate the elastic properties of heterogeneous materials, incorporating on the order of 106-107 occupied voxels.

B. The Role of Carbon and Temperature in Determining Electrical Conductivity of Basins, Crust, and Mantle (A. G. Duba [LLNL; 510-422-7306; Fax 510-423-1057; E-mail alduba@llnl.gov] joint research with T. J. Shankland [505-667-4907; Fax 505-667-8487; E-mail shankland@lanl.gov], and E.A. Mathez [American Museum of Natural History; 212-769-5379; Fax 212-769-5339; E-mail mathez@amnh.org])

Objectives: The intent of this work is to comprehend the electrical conduction mechanisms in carbon-bearing rocks and in mantle minerals for the purpose of relating electrical conductivity measured in the field to formation conditions and existing state of crustal rocks and to temperatures in the mantle.

Project Description: Electrical conductivity depends strongly on temperature (T) and on the presence of other phases, such as carbon, fluids, or ore minerals at the lower temperatures of the crust and basins. Thus, one research approach is to measure of mantle minerals as functions of temperature, orientation, oxygen fugacity (fO2), and iron content. These data supply the best models for "electrogeotherms" yet available. Another approach is to document textures of carbon in crustal rocks from basins and metamorphic zones and relate them to rock conductivity. In this case texture of carbon distribution is mapped with electron microscopy in the same samples used for conductivity measurement.

Results: We have documented that the electrical conductivity of a water-saturated schist, collected from a surface outcrop near the Denali Fault Zone in the Yukon-Tanana terrane of south-central Alaska, increases slightly with pressure to about 200 MPa. Thus, the accepted hypothesis that electrical conductivity of saturated crustal rocks decreases with pressure is not necessarily true. Detailed petrographic examination of one sample, a quartz-mica-garnet-schist, revealed the presence of a stringer of carbonaceous material generally less than 10 µm thick within one of the muscovite layers. The stringer extends for about 2 cm along the foliation and is probably responsible for the anomalous conductivity change with pressure. The carbon

aceous stringer together with its host muscovite layer is deformed and broken around a rotated garnet porphyroclast. We interpret this to indicate that the carbonaceous material formed by fluid deposition in a fracture formed within the muscovite layer, possibly during the main phase of metamorphism and deformation. The mica and carbon stringer were then deformed by a non-coaxial deformation responsible for rotation of the garnet porphyroclasts. The deformation was accommodated by plastic deformation of quartz, indicating that it occurred in the ductile regime under conditions at least equivalent greenschist facies metamorphism. This result demonstrates that the carbonaceous stringer was present at depth. Brittle deformation on the microscopic scale is observed in the rock and interpreted to have been caused by subsequent unloading due to uplift. The brittle deformation broke the connectivity of the carbon stringer, explaining in part why the rock does not exhibit anomalously high conductivity at 0.1 MPa (1 atm) pressure. The observations indicate that carbonaceous material may exert a primary control on crustal electrical conductivity because it may be present as interconnected arrays in grain boundaries or microfractures or in megascopic, through-going fractures. We are currently pursuing laboratory measurements on the effect of carbon deposition on the electrical conductivity of rocks during dilatancy associated with fracture. Preliminary indications are that there is an increase in conductivity as carbon is precipitated on newly formed crack surfaces as a rock fails in the laboratory. In addition, we have documented a small effect of intracrystalline hydrogen on the electrical conductivity of olivine.

CATEGORY: Geochemistry

PERSON IN CHARGE: F. J. Ryerson

A. Thermodynamic and Transport Properties of Aqueous Geochemical Systems (Joseph A. Rard [510-422-6872; Fax 510-422-0208; E-mail rard1@llnl.gov] and Donald G. Miller [510-422-8074; Fax 510-422-6363; E-mail dmiller@llnl.gov])

Objectives: The objectives are to (1) measure precise and accurate osmotic/activity coefficients, solubilities, densities, and mutual (Fick's law) diffusion coefficients for aqueous brine salts and their mixtures and osmotic/activity coefficients for acidic sulfate mixtures, (2) develop reliable methods to estimate such properties for multicomponent solutions from binary solution properties, and (3) calculate generalized transport coefficients.

Project Description: The general techniques of classical thermodynamics and of linear irreversible thermodynamics are used to understand and model equilibrium and transport processes in brines and other aqueous electrolyte mixtures relevant to energy programs. Properties being measured are osmotic/activity coefficients and solubilities by the isopiestic method, densities by pycnometry and vibrating densimetry, and diffusion coefficients by Rayleigh and Gouy interferometry. One goal is to measure highly accurate data for systems involving geochemical brines, radioactive waste isolation, and chemical pollutants. A second goal is to develop estimation methods for accurate predictions of these properties for aqueous electrolyte mixtures of arbitrary complexity, using the accurate new data as test systems. Transport data are being analyzed as Onsager transport coefficients and osmotic/activity coefficients are being analyzed using extended Pitzer's equations.

Results: Diffusion experiments were performed previously at two different solute ratios of mixtures of NaCl and Na2SO4 at 0.5 mol/dm3 and 25°C using Gouy interferometry. Two additional solute ratios were studied as were the limiting binary solutions NaCl and Na2SO4, all at 0.5 mol/dm3, using Rayleigh interferometry. The Rayleigh experiments were done with the Gosting diffusiometer, which is now automated for "real

time" data acquisition using a computer-controlled photodiode array. Similar experiments are planned at 1.0 mol/dm3. Isopiestic experiments were performed for aqueous H2SO4 + MgSO4 mixtures at 25°C and are now complete for the acid-rich half of this system at three molality fractions z of H2SO4. A total of 195 data points have been measured, covering the total molality (sum of the molalities of H2SO4 and MgSO4) and water activity ranges of mT between 0.12548 and 12.050 mol/kg and aw between 0.9958 and 0.2758 at z 6/7, of MT between 0.13563 and 11.011 mol/kg and aw between 0.9958 and 0.3312 at z 5/7, and of mT between 0.14741 and 7.2060 and aw between 0.9958 and 0.5582 at z 4/7. The highest molalities for all three z values extend well into the supersaturated region. Stock solutions of H2SO4 + MgSO4 mixtures were prepared with z 3/7, 2/7, and 1/7, and one series of experiments was started. Another of our isopiestic chambers is being prepared for a second series at these composition fractions. Once the experiments have been performed for the MgSO4-rich region, the results will be modeled with extended versions of Pitzer's equations and will yield reliable parameters for the interaction of Mg2+ with HSO4-.

During this period three journal articles were published and three more submitted. One published paper reported isopiestic results for aqueous H2SO4 solutions at low molalities where literature data were discrepant. The other two analyzed various possible mixing rules for estimating ternary solution densities and electrical conductances based on the properties of their constituent binary solutions. Papers submitted for publication include NaCl + Na2SO4 diffusion coefficients at 0.5 mol/dm3, the effect of different-sized concentration differences on the diffusion coefficients of NaCl + KCl, and isopiestic results for acid-rich H2SO4 + MgSO4 mixtures.

B. Experimental Determination of Mineralogical Controls on UThPb Redistribution: Implications for Crust/Mantle Differentiation (H.F. Shaw [510-423-4645; Fax 510-423-1057; E-mail shaw4@llnl.gov] and F.J. Ryerson [510- 422-6170; Fax 510-422-1002; E-mail ryerson@s91.es.llnl.gov])

Objectives: The objective of this work is to determine mineral/aqueous fluid and mineral/silicate liquid partition coefficients for a suite of trace elements (U, Th, Hf, Zr, Nb, Ta, Sr, Ba, Rb, and Pb) under conditions relevant to fluid metasomatism and partial melting in the upper mantle. The results of the project will provide important constraints on the petrogenetic interpretation of trace element, U, Th, Pb, and U-series disequilibrium data obtained on igneous rocks, particularly those formed in subduction-zone environments.

Project Description: Along with the formation of the Earth's core, the differentiation of the crust and mantle represents the major chemical fractionation process occurring on the Earth. The nature of this process has been constrained by a wide variety of trace element and isotopic analyses of crust- and mantle-derived samples. Effective utilization of these data requires a quantitative understanding of the fractionation of the elements of interest between minerals and both silicate melts and aqueous fluids. For many elements, however, the relevant mineral/melt partition coefficients are poorly known and there is an almost complete lack of data for the partitioning of trace elements between minerals and aqueous fluids. The experimental data generated in this project will provide quantitative information of the partitioning of trace elements of geologic interest between minerals and melts and aqueous fluids, with emphasis on the partitioning of U, Th, Pb, and the high field strength elements (Zr, Hf, Nb, Ta). Partition coefficients for elements of interest are being obtained as a function of O2, T, P, and fluid or melt composition. Experimental charges are produced using standard and newly developed techniques that utilize one-atmosphere gas-mixing furnaces and high-pressure piston-cylinder devices. The trace element composition of the charges is being measured primarily by quantitative ion microprobe techniques, supplemented by electron micro-probe analyses and solid-source mass spectrometry.

Results: The partitioning of U, Th, Pb, Ba, Sr, Nb, and Ta between aqueous fluids (with and without added Cl- and CO32-) and rutile, clinopyroxene, orthopyroxene,

pyrope, olivine, and pargasitic amphibole under upper mantle conditions has been studied. Using these data, together with literature data for Sr and Pb isotopic compositions of Pacific island arc basalts (IABs), the calculated composition of a fluid in equilibrium with the average Pacific IAB source is 87Sr/86Sr = 0 .7036, 207Pb/204Pb = 15.57, Sr ~500 ppm and Pb ~30 ppm. The Subarc fluid composition is consistent with a mixture of fluids from both altered MORB (~96wt%) and sediment (~4wt%), with ~96% of the Sr from the former and ~70% of the Pb from the latter. Although the mass fraction of sediment-derived fluid in the Subarc fluid may be minor, sedimentary input for Ba (0-50%), Th (53- 88%), U (61-83%), in addition to Pb, can be significant. Our results are consistent with, and provide additional support for models of, slab input to IAB sources. Preliminary results indicate that metasomatism by water-rich fluids and silicate melts will produce similar but distinguishable trace element signatures in IABs. Experiments were conducted to determine partition coefficients for Ti, Rb, Ba, Sr, Zr, Nb, Ta, Hf, Pb, U, and Th between pargasitic amphibole and a synthetic hydrous silicate liquid. Rb, Ba, Nb, and Ta are dramatically less compatible in pyroxene than in amphibole, while other elements, such as Th, U, Hf, and Zr, have similar compatibilities. Because of these differences, liquids produced by small degrees of partial melting of amphibole-bearing mantle sources, or by high-level fractionation of amphibole, should have distinctively lower Th-normalized Rb, Ba, Nb, and Ta concentrations than melts from amphibole-free systems. Using mineral-melt partition coefficients determined in this project, the isotopic evolution of the uranium-series nuclides 238U, 230Th, 226Ra, and 231Pa during partial melting in an ascending column of mantle was modeled. The observed [230Th] and [231Pa] excesses in e-MORB and n-MORB can be generated by initiating melting in the garnet stability field at varying depths, and these excesses can be preserved during equilibrium transport of the melt at geologically reasonable rates through the overlying spinel herzolite. The mineral/melt partition

ing of Li, Be, and B between olivine, orthopyroxene, clinopyroxene, amphibole, and synthetic basaltic melt has been determined at 1 atmosphere and 15kb. Calculations using these data indicate that the B/Li and Be/Li ratios in mantle-derived melts will be fractionated with respect to their source but that the B/Be ratio in

the melt will faithfully record the value in the source throughout melting. This implies that the systematic variation in the B/Be ratio of arc basalts as a function of distance from the trench is a primary feature of the mantle source region and not a result of the melting process.

C. Uranium, Thorium, Lead, and Oxygen Diffusion in Rock-Forming Minerals: Implications for Reactive Transport (F.J. Ryerson [510-422-6170; Fax 510-422-1002; E-mail ryerson@s91.es.llnl.gov] and K.D. McKeegan [University of California at Los Angeles; 310-825-3580; Fax 310-825-2779; E-mail kdm@argon.ess.ucla.edu])

Objectives: The spatial distributions of isotopes of elements, such as uranium, thorium, lead, and oxygen, observed on a microscale can be utilized to constrain thermal histories of crustal rocks and the extent of their interactions with fluids. There are two requirements necessary for application of this idea to real geologic systems: first, microanalytical techniques must be employed to quantitatively measure the isotopic or elemental heterogeneities, and, second, fundamental diffusion data must be experimentally determined in order to know the rates at which equilibrium between fluid (either melts or aqueous fluids) and host rocks can be approached as a function of various external conditions.

Project Description: The diffusion coefficients for uranium, thorium, lead, and oxygen will be determined in a number of different minerals under a variety of external conditions. The experimental diffusion runs are performed at the Lawrence Livermore National Laboratory and the analytical phase of the work is done at UCLA, using the Cameca ims 1270 ion microprobe. Techniques for the in situ measurement of either experimentally induced or naturally occurring variations of isotopic and elemental concentrations are being developed for the UCLA ion microprobe.

Results: Silicon and oxygen diffusion have been measured in a grossular, spessartine, and pyrope almandine garnets at 800°C, 1 GPa, and oxygen diffusion measured at 800°C, 1 atm. The 1 GPa runs employ an overgrowth technique in which a single crystal of

garnet is annealed in an H218O fluid-containing powdered 30SiO2; 1 atm runs employ an 18O-enriched gas reservoir. The garnet overgrowth formed in the 1 GPa runs (typically 300 nm thick) decreases the tendency for surface dissolution under hydrothermal conditions, and our experiments yield well-behaved diffusion profiles. "Water" concentrations were determined by IR and ranged from 0 ppm (Tanzania pyrope-almandine) to 2800 ppm (Jeffrey Mine grossular). At 800°C, 1 atm (air and NNO), the garnets showed no loss of water when annealed up to 120 hr. At 1 atm, oxygen diffusion in all of the studied garnets is ~10-24m2 s-1, independent of both water content and bulk composition. Oxygen diffusion in grossular and spessartine both increases to ~10-21m2 s-1 at 1 GPa under hydrous conditions; unfortunately, the short diffusion profiles preclude IR analysis for water in the diffused region. Oxygen diffusion is inversely correlated with silicon concentration (approximated from the sum of 28Si and 30Si intensities) and suggests that the decreased rates of oxygen diffusion are related to a decrease in the hydro-garnet component through the exchange of SiO44- = O4H44-. At 1.0 GPa, silicon diffusion coefficients are equal to oxygen diffusion coefficients for both grossular and spessartine. Closure temperatures for oxygen diffusion in a 1mm diameter garnet are always in excess of 1000°C for geologically reasonable cooling rates. Sluggish oxygen diffusion in garnet confirms that oxygen isotopic compositions recorded during the growth will be faithfully retained.

D. An Experimental Investigation of Mechanisms Controlling Glass Dissolution (Susan A. Carroll [510 423-5694; Fax 510-422-0208; E-mail carroll6@llnl.gov] and William L. Bourcier [510-423-3745; Fax 510-422-0208; E- mail billb@llnl.gov])

Objectives: The objective of this project is to identify the underlying molecular mechanisms responsible for dissolution of glass and to utilize this understanding in the development of quantitative models for predicting glass dissolution rates in nature.

Project Description: This project uses a combination of conventional glass dissolution experiments, potentiometric surface titrations, and NMR characterization of glass and its solution interface to determine the mechanisms controlling glass dissolution. The dissolution behaviors of three simple glass compositions (SiO2, Na2O4•SiO2, and NaAlSi3O8) will be used to develop a mechanistic model based on three principal observations: (1) the pH dependence of dissolution rates, (2) the saturation effect, and (3) the effect of absorbed alkali cations on dissolution rates. The results will be relevant to a number of problems, including the stability of radioactive waste glasses, weathering of volcanic glasses, and obsidian hydration age dating, among others.

Results: In this first year of the project, we are focusing on the mechanisms controlling simple SiO2 glass dissolution as a function of solution pH, Al, Si, Na, and Cs concentrations. We have completed the experimental measurements of silica glass dissolution rates at 25° and 70°C in pH 2, 4, 6, 8, 10, and 12 solutions and in similar buffer solutions at pH 4 and 10 doped with NaCl and CsCl. We have also reacted high

surface area (270 m2/g) silica gels in constant pH 4 and 10 solutions (pH-stat experiments) also doped with either CsCl or NaCl . These samples are to be used in NMR studies of Si, Cs, and Na. Their high surface areas are optimum for enabling the NMR probe to detect adsorbed alkali cations on the silica surface. We have made NaAlSi3O8 (albite) glass for our next set of dissolution experiments that are currently in progress. Samples of the reacted glass and silica gels are currently being analyzed at the NMR lab at UC Davis. A sample probe suitable for making measurements on the reacted glass samples has been installed and tested. The NMR work is in progress. Our results show that the dissolution rate of silica glass as a function of pH is independent of pH below pH 8 but increases with increasing pH above 8. This trend is similar to that observed for quartz; however, the dissolution rate for silica glass is higher than for quartz, as expected. At 25°C, silica glass dissolves about two times faster with each ten-fold increase in Na concentration at pH 10. Under the same conditions, Cs does not affect silica glass dissolution rates. This difference in behavior of the two ions of quite different sizes is the focus of the NMR work. We anticipate that the NMR will be able to determine how Na and Cs are coordinated at the glass-solution interface, and this information we anticipate will help to explain the different effects the two cations have on glass dissolution.

E. Mineral Dissolution and Precipitation Kinetics: A Combined Atomic-Scale and Macro-Scale Investigation (Kevin G. Knauss [510-422-1372; Fax 510-422-0209; E-mail knauss@s19.es.llnl.gov] and Carrick M. Eggleston [307-766- 6769; Fax 307-766-6679; E-mail carrick @uwyo.edu])

Objectives: Our objectives are to build and test a contact atomic force microscope (AFM) capable of operation at up to 150°C and 6 atm pressure, to apply this AFM to direct, in situ, and real-time observations of step dynamics during dissolution and growth of oxide and silicate minerals at elevated temperature and pressure, and to use rate and stoichiometric data from parallel macroscopic dissolution and growth experi

ments to interpret mineral rates, using a combined microscopic Burton-Cabrera-Frank and macroscopic surface-complexation model.

Project Description: This project combines atomic-scale and macro-scale approaches to the study of mineral-fluid interaction in order to significantly improve our understanding of, and ability to predict the course of, mineral dissolution and precipitation pro

cesses. We are building a high temperature flow-through fluid cell for the AFM. This will allow atomic-scale kinetic experiments under geologically relevant conditions for important oxide and aluminosilicate minerals. Identical conditions will be investigated, using macroscopic wet-chemical rate experiments, including conditions both near and far from equilibrium. We will measure rates of dissolution and precipitation, determine activation energies, measure rates of step motion across surfaces (including anisotropy), and investigate step-step interactions that affect rate. We will then be able to address many still-open questions

concerning the exact forms for rate laws near and far from equilibrium, the microscopic interpretation of these rate laws in terms of dissolution and precipitation mechanisms operating under various conditions, and the question of what exactly the "active area of interaction" and "active sites" are on mineral surfaces.

Results: This project was only recently initiated. We are still in the design stages for the proposed high temperature and pressure AFM. Several important design modifications have been made so that we can use existing commercial control electronics, software, and, in some cases, hardware.

CATEGORY: Energy Resource Recognition, Evaluation, and Utilization

PERSON IN CHARGE: F. J. Ryerson

A. Linear and Nonlinear Mechanics of Rocks (J. G. Berryman [510-423-2905; Fax 510-422-1002; E-mail berryman@s123.es.llnl.gov], P. A. Berge [510-423-4829; Fax 510-423-1057; E-mail berge@s44.es.llnl.gov], and D. Elata [510-423-8281; Fax 510-422-1002; E-mail elata@llnl.gov])

Objectives: Our major objective is to understand factors affecting physical properties of rocks in order to improve our ability to predict rock behavior from knowledge of rock components. One new tool developed to accomplish this objective is the recent discovery of exact results in poroelasticity and thermoelasticity for two component composite rocks. This project exploits these as well as other new results, with the expectation that new insight into the linear and nonlinear mechanics of rocks will result. Such insight may prove important for understanding earthquake source mechanisms and for oil field engineering practices related to drilling and pumping. Also, such information is important for interpretation of both seismic and electrical field data.

Project Description: Modeling of idealized two-mineral component rocks has been one of the main thrusts of the effort. Recent advances show that it is possible to compute all the compressibilities (jacketed, unjacketed, jacketed pore, and unjacketed pore) exactly

for certain models. Although very general results on effective-stress rules for various physical properties of rocks have already been published, more explicit applications to examples of well consolidated and poorly consolidated rocks have now been studied and will soon be reported. The new approach based on exact results offers promise of analytical and/or numerical modeling capability from linear to semilinear to fully nonlinear deformation of rocks, including rocks containing cracks, within the same basic theoretical framework. These types of results are of interest in the oil and gas industry, as they play a significant role in interpretation of AVO (amplitude versus offset) data used as direct hydrocarbon indicators. The same basic framework can also be employed to treat reservoir characterization problems, especially regarding the effects of changing stress on matrix and fracture permeability in double-porosity models used for reservoir pumpdown studies. In addition to linear and nonlinear elastic materials, mixtures of viscoelastic materials have also been studied.

Results: We have developed a unified approach for deriving effective medium theories and demonstrated the range of applicability and relationships between implicit and explicit schemes. A paper on this work was recently published in the journal Mechanics of Materials. A new theory of the elastic behavior of granular materials and/or cracked materials under uniaxial stress has been developed and its implications continue to be explored. In collaboration with Stanford researchers, we have developed a new approximate analytical solution describing the pressure dependence of contact stiffnesses for coated spheres, which can be used to describe the elastic behavior of cemented sandstones. These results have also been submitted for publication in the journal Mechanics of Materials. We also show that theoretical models for granular rocks must be modified to avoid violating thermodynamic constraints. A paper on these results has been submitted to the ASME Journal of Applied Mechanics. In collaboration with Professor Graeme Milton of the University of Utah, we have developed a method for obtaining rigorous bounds on the shear modulus of viscoelastic

composites, such as rocks that contain mixtures of two viscoelastic constituents (for example, quarts and a viscous fluid, or quartz and clay). This work extends and completes the previous work of Gibiansky (Princeton) and Milton on the bounds for the bulk modulus of viscoelastic composites and furthermore shows how well various realizable theoretical models produce results consistent with the bounds. A paper describing this work has been submitted for publication to the Proceedings of the Royal Society of London. In collaboration with researchers at the University of Wisconsin, we have developed methods to determine and in some cases drastically reduce the number of elastic coefficients required to describe the behavior of a double-porosity system in the presence of changing pore pressure for applications to reservoir pumpdown. The first part of this work has been accepted for publication in the Journal of Geophysical Research. This work continues both at the theoretical level and also at the level of developing new computer simulation techniques to make use of our results for reservoir engineering and resource management.

B. Velocity Analysis, Parameter Estimation, and Constraints on Lithology for Transversely Isotropic Sediments (P.A. Berge [510-423-4829; Fax 510- 423-1057; E-mail berge@s44.es.llnl.gov], J. G. Berryman [510-423-2905; Fax 510- 422-1002; E-mail berryman@s123.es.llnl.gov], D. Elata [510-423-8281; Fax 510- 422-1002; E-mail elata@llnl.gov], I. Tsvankin [Colorado School of Mines] [303-273- 3060; Fax 303-273-3478; E-mail ilya@dix.mines.edu], K. Larner [Colorado School of Mines] [303-273-3428; Fax 303-273-3478; E-mail klarner@dix.mines.edu], F. Muir [Stanford University] [415-723-9390; Fax 415-723-1188; E-mail francis@pangea.stanford.edu])

Objectives: Our major objective is to obtain constraints on lithology, using the anisotropy parameters recovered from seismic data, in order to improve analysis of seismic reflection data collected in areas where the geology is complicated by anisotropy and heterogeneity.

Project Description: The influence of anisotropy leads to significant distortions in seismic reflection data processing and errors in interpretation that can result in drilling in the wrong place or to the wrong depth and errors in data analysis that may turn a play into a non-play or vice versa. Theoretical constraints on the elastic stiffnesses in a transversely isotropic medium

and algorithms newly-developed at the Colorado School of Mines for processing seismic reflection data exhibiting transverse isotropy are being combined with rock physics analysis to determine how constraints on anisotropy translate into constraints on lithology and improved interpretation of seismic reflection data. Expected benefits of this project are improved processing and interpretation of seismic reflection data in the oil exploration industry and increased understanding of the connections between seismic properties and other physical properties of rocks and sediments exhibiting anisotropy.

Results: Investigators from the Colorado School of Mines have developed new algorithms and codes for processing seismic reflection data in vertically inhomogeneous, transversely isotropic media and have applied these codes to a Chevron data set to produce estimates of some of the anisotropy parameters. These results have been published in Geophysics and in The Leading Edge and were presented at the 7th International Workshop on Seismic Anisotropy. Colorado School of Mines researchers have also developed algorithms for processing seismic reflection data that exhibit transverse isotropy with a nonvertical symmetry axis. These results have been submitted to Geophysical Prospecting, to the EAGE 58th Annual Meeting, and to the 1996 annual meeting of the Society of Exploration Geophysicists. Results from research at Stanford include the development of a set of mathematical tools for describing the elastic constants of anisotropic materials using a minimum number of parameters, where these parameters have lithologic significance. These results were presented at the 7th In

ternational Workshop on Seismic Anisotropy. LLNL investigators have found some links between constraints on anisotropy parameters in certain models of anisotropic media and constraints on lithology. These results were presented at the 7th International Workshop on Seismic Anisotropy. LLNL researchers collaborating with Schlumberger researchers have also developed models to describe stress-induced anisotropy in granular media. Some of these results have been presented at the 7th International Workshop on Seismic Anisotropy and submitted for publication in Phys. Rev. E. LLNL and Stanford researchers are currently developing and analyzing additional theoretical models of rocks to determine how anisotropy parameters are related to rock properties, especially for rocks containing fluids, and how these relationships constrain both lithology and fluid content. The Stanford/LLNL group is also determining what information from seismic reflection data is necessary and sufficient for correlating anisotropy parameters and lithology. Work on fluid substitution in anisotropic sediments is in progress.

C. Compositional Kinetic Modeling of Oil and Gas Formation (Alan K. Burnham [510-443-8779; 510-423-7914; E-mail burnham1@llnl.gov])

Objectives: This work develops and tests models of petroleum generation, migration, and thermal stability. These models are used to reduce petroleum exploration risk and costs through integrated basin analysis, which combines many aspects of geology, geophysics, geochemistry, and hydrology to determine where and when oil is generated, migrates, and accumulates.

Project Description: Oil and gas generation kinetics and oil destruction kinetics have been measured by a variety of techniques, including isothermal hydrous pyrolysis, temperature-programmed pyrolysis using various detectors, and sealed capillary tube pyrolysis, and kinetic models are developed to predict oil and gas generation and composition and their expulsion from the source rock. The final remaining experiments are directed towards understanding the kinetics of oil cracking to better understand the floor for oil survival and formation mechanisms of natural gas.

Results: During the past year, four papers were published on earlier work on a new kinetic model for well-preserved algal kerogens, a test of the conventional parallel reaction model for marine and terrestrial kerogens using hydrous pyrolysis residues, experiments and a kinetic model for high-pressure cracking of hexadecane, and a basin analysis study of the Maracaibo basin. New experimental work expanded our use of isotopically labeled hydrocarbons to study the cracking of hydrocarbons in crude oil, including hexadecane cracking at 310°C, the fate of alkene intermediates, and the rate of dealkylation of benzene and cyclohexane. The experiments are now essentially complete, but since data analysis is currently in progress, there are only a few conclusions that can be given at this time. Earlier results indicated that both the rate and mechanism of alkane cracking are different in pure alkanes and in crude oils. Cracking of oils doped with 1,2-13C-labeled dodecene indicated that alkanes are readily saturated

and much less likely to undergo addition reactions in crude oils as compared to alkane mixtures. This supports the idea that there are sufficient hydrogen donors in crude oils to quench free radical chain reactions and slow the overall cracking rate of alkanes. Of the alkyl radicals and alkenes that are not saturated by hydrogen donors, 13C NMR measurements show that addition to aromatic rings rather than alkylation of alkenes is the predominant pathway. The cracking rate of 1,2-13C-labeled hexadecane doped in a North Sea crude oil was measured at 310°C over nine months in order to more accurately extrapolate to geologic time frames. No

cracking products were observed, which means that the rate must be at least 150 times slower than at 350, which indicates an activation energy of 90 kcal/mol. This suggests that all free-radical propagation reactions have been quenched and the effective activation energy approaches that of initiation reactions. If thermal cracking were the only destruction mechanism, petroleum could survive for nearly a million years at 250°C and more than a billion years at 200°C, in agreement with some recent geological observations.