@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Earth, Ocean and Atmospheric Sciences, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Haywood, Jennifer"@en ; dcterms:issued "2011-04-26T15:08:56Z"@en, "2011"@en ; vivo:relatedDegree "Master of Science - MSc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Carbonates and shales are common in fold and thrust belts worldwide: carbonates typically comprise the hanging wall of fault zones and the shale forms the footwall. Generally, a cataclasite is developed in both the carbonate and shale materials, demonstrating that strain is accommodated in both rock types. Despite the wide occurrence of carbonate and shale cataclasites, little is known about the rheological behavior of these composites. The results of two suites of triaxial frictional sliding experiments designed to analyze the effects of composition, temperature, pore fluid pressure and forcing block composition on gouge strength and stability are presented. Experiments were conducted at 70 MPa effective confining pressure and displacement rates varied between 1 to 100 μm s⁻¹. Gouge material was created from quartz-bearing phyllosilicate-rich shale combined in various volumetric proportions with reagent grade calcite powder with an average grain size of ~5 μm. Experiments were performed on each endmember composition as well as 75%, 50% and 25% mixtures of shale and calcite. At room temperature (T), saturated conditions strain localization in the composite gouges causes significant weakening relative to the strong carbonate endmember. At 150°C and 15 MPa pore fluid pressure (Pf), the carbonate gouge undergoes significant strain weakening followed by the evolution to stick-slip sliding. Microstructure analysis indicates that deformation in the shale endmember gouge is distributed across the gouge zone. In the composite gouges, the fine grained carbonate facilitates phyllosilicate rotation and strain localization. In the carbonate gouge, strain localizes in R₁ and shear zone boundary parallel Y shears. Results show that in the absence of elevated T and Pf, the carbonate hanging wall cataclasite is strong relative to the underlying shale footwall cataclasite. At these conditions strain is most likely to localize in the shale or shale-rich composites. Elevated T and Pf promote strain localization and seismic faulting in the carbonate cataclasite. The coefficient of friction values determined for shale carbonate composites are less than the μ = .85 value predicted by Byerlee for rocks deformed at less than 200 MPa normal stress and should replace Byerlee’s value in numerical thrust sheet models."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/33956?expand=metadata"@en ; skos:note "CARBONATE AND SHALE FAULT GOUGE: EXPERIMENTAL INSIGHTS INTO THE ROLE OF GOUGE COMPOSITION, TEMPERATURE AND PORE FLUID PRESSURE ON THE MECHANICAL STRENGTH OF FAULTS by Jennifer Haywood B.A., The Colorado College, 2006 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Geological Sciences) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2011 © Jennifer Haywood, 2011 ii Abstract Carbonates and shales are common in fold and thrust belts worldwide: carbonates typically comprise the hanging wall of fault zones and the shale forms the footwall. Generally, a cataclasite is developed in both the carbonate and shale materials, demonstrating that strain is accommodated in both rock types. Despite the wide occurrence of carbonate and shale cataclasites, little is known about the rheological behavior of these composites. The results of two suites of triaxial frictional sliding experiments designed to analyze the effects of composition, temperature, pore fluid pressure and forcing block composition on gouge strength and stability are presented. Experiments were conducted at 70 MPa effective confining pressure and displacement rates varied between 1 to 100 µm s-1. Gouge material was created from quartz-bearing phyllosilicate-rich shale combined in various volumetric proportions with reagent grade calcite powder with an average grain size of ~5 µm. Experiments were performed on each endmember composition as well as 75%, 50% and 25% mixtures of shale and calcite. At room temperature (T), saturated conditions strain localization in the composite gouges causes significant weakening relative to the strong carbonate endmember. At 150°C and 15 MPa pore fluid pressure (Pf), the carbonate gouge undergoes significant strain weakening followed by the evolution to stick-slip sliding. Microstructure analysis indicates that deformation in the shale endmember gouge is distributed across the gouge zone. In the composite gouges, the fine grained carbonate facilitates phyllosilicate rotation and strain localization. In the carbonate gouge, strain localizes in R1 and shear zone boundary parallel Y shears. iii Results show that in the absence of elevated T and Pf, the carbonate hanging wall cataclasite is strong relative to the underlying shale footwall cataclasite. At these conditions strain is most likely to localize in the shale or shale-rich composites. Elevated T and Pf promote strain localization and seismic faulting in the carbonate cataclasite. The coefficient of friction values determined for shale carbonate composites are less than the µ = .85 value predicted by Byerlee for rocks deformed at less than 200 MPa normal stress and should replace Byerlee’s value in numerical thrust sheet models. iv Preface This thesis is written as a modified version of two papers written for publication in peer reviewed journals. Part I of the thesis “Room Temperature, Velocity Stepping Sliding Friction Experiments” reports the results of room temperature, velocity-stepping frictional sliding experiments conducted on carbonate and shale composites. These rock deformation experiments were designed to study the effects of gouge composition and forcing block porosity on the strength, stability, and microstructural evolution of calcite and shale gouge zones. The results are applied to natural fault zones. This chapter has been submitted for publication, entitled Carbonate and shale fault gouge: Experimental insights into the role of gouge composition on the mechanical strength of faults (submitted January 13, 2011) Authors: Haywood, J. and Kennedy, L. A.. I performed the experiments, the microstructural analyses and XRD data collection. Kennedy supervised the project. Part II “Temperature and Pore Fluid Pressure Controls” reports the results of experiments performed on shale and carbonate composites (identical to those used in chapter one) at elevated temperature and pore fluid pressure to investigate how the strain is partitioned between these cataclasites as temperature and pore fluid pressure conditions in the fault zone fluctuate, and to evaluate the deformation mechanisms that are active under elevated T and Pp. This section will be revised and submitted as a paper: Temperature-fluid pressure controls on the mechanical evolution and shale-carbonate fault gouge. Authors: Haywood, J., Kennedy, L. A., and Faulkner D. R.. I performed the experiments, the microstructural analyses and XRD data. Collection. Kennedy supervised the project. Experiments were conducted at Faulkner’s lab at the University of Liverpool. v Table of Contents Abstract ....................................................................................................................................ii
 Preface .....................................................................................................................................iv
 Table of Contents ....................................................................................................................v
 List of Tables.........................................................................................................................viii
 List of Figures .........................................................................................................................ix
 List of Symbols .......................................................................................................................xi
 Acknowledgements................................................................................................................xv
 1.
 Introduction ......................................................................................................................1
 1.1.
 Introduction ..............................................................................................................................1
 1.2.
 Coefficient of Friction and the Rate and State Properties of Friction ......................................2
 1.3.
 Thesis Layout ...........................................................................................................................9
 2.
 PART I: Room Temperature, Velocity Stepping Sliding Friction Experiments......11
 2.1.
 Introduction ............................................................................................................................11
 2.2.
 Methods ..................................................................................................................................13
 2.2.1.
 Experimental Conditions.................................................................................................13
 2.2.2.
 Large Sample Rig (LSR).................................................................................................17
 2.3.
 Results ....................................................................................................................................19
 2.3.1.
 Mechanical Data .............................................................................................................19
 2.3.1.1.
 Effects of Compositional Variation .........................................................................19
 2.3.1.2.
 Effects of Forcing Block Porosity ...........................................................................21
 2.3.1.3.
 Effects of Pore Fluid ................................................................................................24
 2.3.2.
 Microstructure Analysis ..................................................................................................27
 2.3.2.1.
 Compositional Variations and Microstructure Development ..................................27
 2.3.2.2.
 Strain and Microstructure Development..................................................................35
 2.3.2.3.
 Effects of Forcing Block Porosity ...........................................................................39
 2.4.
 Discussion...............................................................................................................................43
 2.4.1.
 Comparison to Previous Experimental Data ...................................................................43
 2.4.1.1.
 Microscale Deformation Mechanisms and Foliation Development ........................45
 2.4.2.
 Fault Core Width.............................................................................................................46
 vi 2.4.3.
 Implications for Natural Thrust Faults ............................................................................48
 2.5.
 Conclusions ............................................................................................................................50
 3.
 Part II: Temperature and Pore Fluid Pressure Controls ...........................................52
 3.1.
 Introduction ............................................................................................................................52
 3.2.
 Methodology...........................................................................................................................58
 3.2.1.
 Sample Description and Preparation...............................................................................58
 3.2.2.
 Experimental Procedure ..................................................................................................59
 3.2.3.
 Petrographic Observations ..............................................................................................64
 3.3.
 Results ....................................................................................................................................65
 3.3.1.
 Frictional Strength and Sliding Behavior of Gouge Mixtures ........................................65
 3.3.2.
 Effects of Pf and T on Mechanical Strength....................................................................70
 3.3.3.
 Description of Gouge Textures .......................................................................................77
 3.3.1.
 Orientation of Riedel Shear Localization........................................................................86
 3.4.
 Discussion...............................................................................................................................88
 3.4.1.
 Effect of Increasing Carbonate Content on Frictional Strength......................................88
 3.4.2.
 Effects of T and Pf on Carbonate Gouge Strength and Stability.....................................97
 3.4.3.
 Implications for Natural Fault Zones ............................................................................100
 3.4.3.1.
 Strain Partitioning in Thrust Faults........................................................................100
 3.4.3.2.
 Implications for Thrust Belt Models......................................................................103
 3.5.
 Conclusions ..........................................................................................................................106
 4.
 Discussion......................................................................................................................108
 4.1.
 Comparison of the Two Experimental Suites and Conclusions ...........................................108
 4.1.1.
 Comparison Between Relative Strengths of Shale and Composite Gouges .................108
 4.1.2.
 Effects of the Copper Jacket on Frictional Strength .....................................................111
 4.1.3.
 Effects of Pore Fluid Pressure.......................................................................................111
 4.1.4.
 Microstructure ...............................................................................................................113
 4.1.5.
 Ideas for Future Work ...................................................................................................116
 5.
 Conclusion.....................................................................................................................118
 Works Cited .........................................................................................................................119
 Appendices ...........................................................................................................................136
 Appendix A
 Grain Size Distribution ...........................................................................................137
 Appendix B
 XRD Analysis .........................................................................................................141
 vii Appendix C
 Data Acquisition .....................................................................................................149
 Appendix D
 Data Processing.......................................................................................................152
 Calculations ................................................................................................................................153
 Correction for Change in Surface Area During Sliding .............................................................158
 Matlab Code ...............................................................................................................................160
 LSR Stiffness Calibration Procedure..........................................................................................173
 Appendix E
 Replicate Experiments.............................................................................................182
 Appendix F
 Compilation of Previous Experimental Data...........................................................184
 viii List of Tables Table 2.1: Starting Material.....................................................................................................16
 Table 2.2: Experimental Data..................................................................................................23
 Table 3.1: Experimental Conditions........................................................................................54
 Table 3.2: Experimental Data..................................................................................................66
 Table A-1: Grain Size Distribution .......................................................................................140
 Table B-1: Endmember Composition....................................................................................145
 Table B-2: UBC Composite Composition.............................................................................146
 Table B-3: Liverpool Composite Compositions ...................................................................147
 Table C-1: Raw UBC Output File .........................................................................................150
 Table C-2: Raw Liverpool Output Data File.........................................................................151
 Table D-1: Calibration Experiment Conditions ....................................................................175
 Table F-1: Compilation of Previous Work............................................................................185
 ix List of Figures Figure 1.1: Rate and State Friction Graph.................................................................................5
 Figure 1.2: Fault Zone Figure (after [Tullis et al., 2007]).........................................................8
 Figure 2.1: Sample Prep ..........................................................................................................15
 Figure 2.2: LSR .......................................................................................................................18
 Figure 2.3: Mechanical Data ...................................................................................................22
 Figure 2.4: Dry Graphs............................................................................................................26
 Figure 2.5: Riedel Shear Zone Terminology...........................................................................29
 Figure 2.6: Microstructure I ....................................................................................................30
 Figure 2.7: Microstructure II ...................................................................................................33
 Figure 2.8: Strain Percent Graph.............................................................................................37
 Figure 2.9: Strain Percent Photos ............................................................................................38
 Figure 2.10: Porosity I.............................................................................................................40
 Figure 2.11: Porosity II ...........................................................................................................41
 Figure 3.1: Sample Assembly .................................................................................................60
 Figure 3.2: Rig Diagram (from [Mitchell, 2007]) ...................................................................63
 Figure 3.3: Mechanical Data ...................................................................................................68
 Figure 3.4: Mechanical Data II................................................................................................69
 Figure 3.5: Coefficient of Friction vs. Carbonate Content......................................................71
 Figure 3.6: Effects of Pf on Mohr-Coulomb Failure ...............................................................73
 Figure 3.7: Strain Hardening Rate...........................................................................................75
 Figure 3.9: Microstructure.......................................................................................................80
 Figure 3.10: Regimes ..............................................................................................................84
 Figure 3.11: Carbonate ............................................................................................................85
 Figure 3.12: Riedel Shear Localization Angle (after Tembe et al. 2010) ...............................87
 Figure 3.13: Comparison with Other Experimental Data........................................................92
 Figure 3.14: Framework Model...............................................................................................95
 Figure 3.15: Critical Taper (after [Davis et al., 1983]) .........................................................105
 Figure 4.1: Carbonate Content vs. Coefficient of Friction....................................................109
 Figure 4.2: Comparison to Previous Data .............................................................................110
 x Figure A-1: Volume Percent .................................................................................................138
 Figure A-2: Cumulative Volume Percent..............................................................................139
 Figure B-1: Shale - UBC .......................................................................................................142
 Figure B-2: Carbonate ...........................................................................................................143
 Figure B-3: Shale - Liverpool ...............................................................................................144
 Figure B-4: Ternary Diagram................................................................................................148
 Figure E-1: Replicate Experiments .......................................................................................183
 xi List of Symbols* (in order of appearance in text) Chapter 1 μ coefficient of friction τ shear stress σn normal stress μ0 initial steady state coefficient of friction V0 initial sliding velocity V new sliding velocity a experimentally determined constant, direct effect Dc critical sliding distance b experimentally determined constant θ state variable t time Chapter 2 ϕ porosity μ coefficient of friction XRD X-Ray Diffraction LVDT linear variable differential transformer SEM scanning electron microscope Pc confining pressure Pe effective confining pressure γ shear strain T temperature Chapter 3 μ coefficient of friction *
selected
 xii T temperature CKH ceramic knuckle band heater SEM scanning electron microscope Pf pore fluid pressure Pc confining pressure Pe effective confining pressure γ shear strain ϕ angle of internal friction α angle between Riedel shear and shear zone boundary μi coefficient of internal friction Chapter 4 UBC University of British Columbia U of L University of Liverpool Pf pore fluid pressure Chapter 5 T temperature Pf pore fluid pressure Appendices UBC University of British Columbia U of L University of Liverpool XRD X-Ray Diffraction Phyll phyllosilicate Qtz quartz DCDT direct current displacement transducer LVDT linear variable differential transformer Flbs applied force (i.e. axial load in lbs ft s-2) Fkg applied force (i.e. axial load in kg m s-2) xiii r radius of core (in m) A surface area of the top of the core (in m2) σdiff (kg) differential stress (in kg m-1 s-2) σdiff differential stress (in MPa) Pc confining pressure (in MPa) angle between core axis and saw cut r radius of core 
 σ1app raw applied axial stress θ angle between axial load (σ1) and the pole to the saw cut surface σ1app applied axial stress corrected for change in surface area din axial displacement (in inches) dmm axial displacement (in mm) kkg/cm rig stiffness (in kg/cm) kkg/mm rig stiffness (in kg/mm) dR deformation in the rig (in mm) dcorr axial displacement corrected for deformation in the rig (in mm) γ shear strain d displacement (mm) Φ
 angle of saw cut to axial load TT precompaction gouge zone thickness (mm) l axial displacement As initial area of sliding surface Asr area of sliding surface after axial displacement l ! \" xiv k stiffness CESL Center for Experimental Studies of the Lithosphere E Young’s Modulus A area of the top of the calibration rod L length of the calibration rod krod calibration rod stiffness krig rig stiffness ksystem combined stiffness of the rig and calibration rod xv Acknowledgements Funding for this project was provided by an NSERC Discovery Grant to Lori Kennedy. The Yukon Geologic Survey provided logistical support to collect the shale gouge material. The Egil H. Lorntzsen Scholarship provided tuition funding for the first year of study. I would first like to thank Lori Kennedy for help and guidance from the very first experimental design through conducting an experiment to writing and compiling the final thesis. I truly appreciate your flexibility when things didn’t work as planned, and I am honored that you trusted me and allowed me to have significant input into the direction of my project. Thank you also for providing me with opportunities to make connections and collaborate with your colleagues from around the world. My experiences with the Rock Deformation Lab in Liverpool and the Yukon Geological Survey have provided me with valuable connections to the larger scientific community. Thanks to Kelly Russell for welcoming me into your lab, you provided me with a great group of people with whom I interacted and discussed geology with every day. I am humbled by your unfailing support, and your availability at short notice to discuss anything from MATLAB coding and error propagation to things slightly outside your areas of expertise, like the mechanical behavior of fault gouges. Also, thanks to Terry Gordon for helping me write MATLAB code and for always being ready to help sort through complicated equations, whether for my thesis research or just a class homework assignment. Go raibh mile maith agaibh as bhur gcunamh! Joern Unger and Doug Polson, thanks for keeping the drill press and LSR running, and for helping out with little projects along the way. xvi I would also like to thank Dan Faulkner, Julia Behnsen, Pete Armitage, Tom Mitchell and the rest of the Rock Deformation Lab at the University of Liverpool. Thanks for letting me use your lab, for helping me whenever pieces broke, and for showing me some of the really cool things that can be done with fancy equipment! Rhodri Jerrett - thanks for being an amazing host/cook/tour guide/ historian/housemate/ friend in Liverpool. I could not have done it without you. Lucy, you are everything a post-doc should be. You provide support, advice, a bouncing board for ideas, general life support and guidance - always with a sense of humour. I know that the entire VPL appreciates what you do. Betsy, you’ve seen me through a lot of phases of life and development as both a person and a geologist. Thanks for always being there to listen. R-E, Chanone, Shelley and the rest of the VPL, thanks for the continued entertainment, for keeping things fun, and for teaching me a little bit about volcanoes along the way. Other people I’d like to thank: Liz Hearn, Mark Jellinek, Steve Israel, Mati Raudsepp, Jenny Lai and the rest of the XRD/SEM facility people, Teresa Woodley, Cary Thomson and the rest of the administrative staff, Ken Hickey & his lab, plus the rest of the UBC faculty and grad students who have helped me and kept me engaged in fields of geology that are completely unrelated to my own research. 1 1. Introduction 1.1. Introduction Carbonates and shales are common in fold and thrust belts worldwide: carbonates typically comprise the hanging wall of fault zones and shale forms the footwall [Burkhard, 1990; Erickson, 1994; Kennedy and Logan, 1998; Vrolijk and van der Pluijm, 1999; Wojtal and Mitra, 1986]. Generally, a cataclasite is developed in both the carbonate and shale materials. The limestone cataclasite is typically depleted in carbonate relative to the hanging wall carbonate, and the shale cataclasite is enriched in carbonate relative to the footwall shale [Erickson, 1994; Kennedy and Logan, 1998]. Both cataclasites can be foliated and the degree of strain partitioning between the two cataclasites is unknown. In other upper-crustal fault zones, phyllosilicates such as illite, chlorite and white micas commonly form in gouge zones as replacement products of feldspars [Jefferies et al., 2006]. Calcite easily goes into solution and commonly precipitates in fractures within gouge zones, chemically enriching fault gouge material in calcite [Peyaud et al., 2006]. Furthermore, active fault zone material recovered during the San Andreas Fault Observatory at Depth (SAFOD) drilling project [Solum et al., 2006; Tembe et al., 2006], and the Nojima Fault Zone in Japan [Ohtani et al., 2000] contained both phyllosilicates and calcite. The detailed studies of the Carboneras Fault, Spain an exhumed crustal-scale, phyllosilicate and carbonate-rich strike slip fault zone highlight the importance of these materials in fault zone rheology. However, despite numerous and diverse examples of phyllosilicate and carbonate- rich fault zones little is known about the rheological behavior of phyllosilicate and carbonate gouge composites [Faulkner et al., 2003]. Mechanical grain-size reduction during faulting of intact rock produces fine-grained cataclasite and gouge material [Engelder, 1974] and the frictional properties of both 2 unconsolidated and lithified gouge material profoundly affect whether seismic faulting or aseismic slip dominates fault movement [Brace and Byerlee, 1966; Dieterich, 1978; Dieterich, 1979; Marone, 1998; Ruina, 1983]. Laboratory-based sliding friction experiments and the resulting friction constitutive laws are used to examine the effects of the frictional characteristics of fault zone materials on the stability and style of sliding in fault zones [Marone, 1998]. The purpose of my research is to analyze the effects of composition, temperature, pore fluid pressure and forcing block composition on the strength and stability of carbonate, shale and carbonate/shale composite fault gouges. The aim of this research is to assess how fluctuations in temperature and pore fluid pressure affect strain accommodation in fault zones containing both shale and carbonate minerals. Microstructural analysis seeks to correlate fabric development with the strength and stability of these gouges. Although carbonate-on- shale thrust faults are the primary target of this study, the result will be applicable to other shallow fault zones that contain both phyllosilicates and carbonates. Below, a review of friction and the rate and state properties of friction is provided as background information. 1.2. Coefficient of Friction and the Rate and State Properties of Friction Laboratory-based rock friction experiments are widely used to examine the strength and stability of fault zone materials. Early models of earthquake nucleation and propagation invoked fracturing of intact rock as a way to relieve stored elastic strain in the crust [Reid, 1911]. However, fracturing of intact rock, especially at high confining pressures, creates stress drops much bigger than those recorded in natural earthquakes [Chinnery, 1964]. 3 Furthermore, earthquakes typically nucleate in preexisting fault zones. As a result, Brace and Byerlee [1966] suggested that the frictional properties of fault zone materials, rather than the strength of intact rock may control the stability of a fault zone. Specifically, they suggested that stick-slip behavior, a phenomenon observed in the laboratory in which there is a cyclical increase in stress followed by a stress drop, may be the equivalent of shallow crustal earthquakes. Their model, in which the frictional properties of rocks determine whether or not seismic faulting will occur, is widely accepted and provides the basis for rock friction studies. The Coulomb coefficient of friction (µ) relates the amount of stress perpendicular to a fracture to the amount of stress parallel to a fracture and is a measure of the resistance of a material to sliding. It is described by the equation: Equation 1.1 in which τ is the shear stress and σn is the normal stress. Byerlee [1978] suggested that for a wide variety of rock types, the shear stress required for sliding in the upper crust can be approximated by the equation τ=.85σn for normal stresses less than 200 MPa, and by the equation τ = 50 + 0.6σn for normal stresses greater than 200 MPa. Although this is a simple model, for a wide variety of rock types, these coefficient of friction values are appropriate [Byerlee, 1978]. However, they are not representative values for many phyllosilicates [Byerlee, 1978]. This is significant since fault zones often contain large amounts of phyllosilicates (i.e.: San Andreas Fault [Carpenter et al., 2009; Logan et al., 1981; Morrow et al., 2007; Tembe et al., 2006], Chelungpu Fault, Taiwan [Mizoguchi et al., 2008], Neodani 4 Fault, Japan [Tsutsumi et al., 2004], Nojima Fault, Japan[Ohtani et al., 2000], Carboneras Fault, Spain [Faulkner et al., 2003]). The coefficient of friction is not constant and depends on the grain size, the amount of slip, sliding velocity, sliding history, the time of static contact between surfaces, the area of the sliding surface and the presence of pore fluids [Paterson and Wong, 2005]. A change in sliding velocity results in a change in the value of the coefficient of friction, meaning that friction is dependant on sliding velocity. The following equation known as the Dieterich- Ruina Law [Dieterich, 1979; Ruina, 1983] is used to describe the dependence of friction on sliding velocity and the amount of static contact time (also known as Rate and State Friction): Equation 1.2 In this equation, µ0, is the steady state coefficient of friction for a material sliding at a constant initial sliding velocity, V0. When the sliding velocity is increased to a new, faster (V/V0 = e) sliding velocity, V, there is an immediate increase in the coefficient of friction (Figure 1.1), known as the instantaneous rate dependence of friction, represented by the variable a. Over a characteristic sliding distance, Dc, the coefficient decays to a new steady state coefficient of friction, µ. The variable b describes the magnitude of the decrease in coefficient of friction from the maximum immediately following the velocity step to the new steady state coefficient of friction. 5 Figure 1.1: Rate and State Friction Graph 1 mm/s 1 mm/s10 mm/s a b Dc .54 .59 .58 .55 .56 .57 9.5 10 10.5 11 11.5 Displacement Co e! ci en t o f F ri ct io n Figure 1.1 (modified from [Mair and Marone, 1999]): Schematic showing the rate and state properties of friction. As the sliding velocity increases (by a magnitude of e) there is an immediate increase in the coefficient of friction. The increase, a, is the ‘direct effect’. Over a characteristic sliding distance, Dc, the coefficient of friction decays to a new steady-state coefficient of friction. The magnitude of the decrease in the coefficient of friction from the maximum value following the velocity step to the new steady-state coefficient of friction is described by the variable b. a and b are both material properties, and the relationship between them describes the rate dependence of the gouge material. If (a - b) is positive the gouge is velocity strengthening, meaning that the strength of the gouge increases as sliding velocity increases. If (a - b) is negative, the gouge is velocity weakening, meaning the strength of the gouge decreases as sliding velocity increases. The gouge in this figure displays velocity weakening behavior since (a – b) < 0. 6 In the Dieterich model, the state variable, θ, represents the average age of asperity contacts, and changes as a function of time according to the equation [Dieterich, 1979]: Equation 1.3 θ is an empirically derived value and the microstructural mechanics of state variable evolution are poorly understood, but the state variable is thought to represent an average contact lifetime. A good summary of rate and state friction is available in Scholz [1998]. If the ‘new’ steady state coefficient of friction (following an increase in sliding velocity) is less than the previous steady state coefficient of friction the material is velocity weakening. In this case, (a-b) < 0. Conversely, if the coefficient of friction increases with increasing sliding velocity the material is velocity strengthening. In this case, (a-b) > 0. Velocity weakening behavior can result in a frictional instability, while velocity strengthening materials promote steady-state sliding. In the laboratory, velocity weakening behavior can result in stick-slip sliding. Velocity strengthening materials are inherently stable, whereas velocity weakening materials are inherently unstable. An earthquake can only nucleate in velocity weakening portions of a fault zone. Velocity strengthening portions of a fault zone deform primarily by stable, aseismic creep [Scholz, 1998] (Figure 1.2). The earthquake can then propagate some distance into velocity strengthening portions of the fault, but ultimately the earthquake dissipates. 7 Figure 1.2 (following page): The distribution of velocity weakening and velocity strengthening materials in a fault zone varies with depth and along fault zone strike. Changes along strike can be due to changes in lithology. Earthquakes that nucleate in velocity weakening portions of the fault zone can propagate into velocity strengthening portions of the fault zone, but ultimately dissipate. The presence of unconsolidated gouge in the upper several km of the crust may define the upper limit of the seismogenic zone, as dilation in the unconsolidated gouge causes velocity strengthening behavior (e.g. [Marone, 1998; Marone et al., 1990]). 8 Figure 1.2: Fault Zone Figure (after [Tullis et al., 2007]) Unstable frictional slip, (a - b) < 0 Stable frictional slip, (a - b) > 0 10 km 3 km Stable or Unstable Slip? ?? Nucleation Heterogeneities Damage Zone Slip Zone Ultracataclasite Zone Gouge Zone DEPTH Propagation Stable frictional slip, (a - b) > 0 9 1.3. Thesis Layout This thesis is written as a modified version of two papers written for publication in peer reviewed journals. Part I of the thesis “Room Temperature, Velocity Stepping Sliding Friction Experiments” reports the results of room temperature, velocity-stepping frictional sliding experiments conducted on carbonate and shale composites. These rock deformation experiments were designed to study the effects of gouge composition and forcing block porosity on the strength, stability, and microstructural evolution of calcite and shale gouge zones. The results are applied to natural fault zones. This chapter has been submitted for puplication, entitled Carbonate and Shale Fault Gouge: Experimental insights into the role of gouge composition on the mechanical strength of faults (submitted January 13, 2011) Authors: Haywood, J. and Kennedy, L. A.. I performed the experiments, the microstructural analyses and XRD data analysis. Kennedy supervised the project. Part II “Temperature and Pore Fluid Pressure Controls” reports the results of experiments performed on shale and carbonate composites (identical to those used in chapter one) at elevated temperature and pore fluid pressure to investigate how the strain is partitioned between these cataclasites as temperature and pore fluid pressure conditions in the fault zone fluctuate, and to evaluate the deformation mechanisms that are active under elevated T and Pp. This section will be revised and submitted for publication as Temperature-fluid pressure controls on the mechanical evolution and shale-carbonate fault gouge. Authors: Haywood, J., Kennedy, L. A., and Faulkner D. R.. I performed the experiments, the microstructural analyses and XRD data analysis. Kennedy supervised the project. Experiments were conducted at Faulkner’s lab at the University of Liverpool. 10 The discussion chapter will compare the results of Part I and II and the results from these experiments will be discussed in the context of natural fault zones. 11 2. PART I: Room Temperature, Velocity Stepping Sliding Friction Experiments1 2.1. Introduction In this section, the results of room temperature, velocity-stepping frictional sliding experiments conducted on carbonate and shale composites are presented. Specifically, the effects of gouge composition and forcing block porosity on the strength, stability, and microstructural evolution of carbonate and shale gouge zones are examined. Velocity-stepping frictional sliding experiments are designed to examine the strength and stability of a fault gouge. In these experiments the sliding velocity is systematically varied to examine the velocity dependence (i.e. stability) of a fault gouge. Steady-state sliding, the equilibrium state at which continuous movement along a fracture surface occurs without an increase in the applied load [Paterson, 1978], is represented by a horizontal line on a stress-strain plot, and is indicative of stable sliding. Jerky sliding in which there is a cyclical increase in stress followed by an abrupt stress drop is called stick-slip behavior [Paterson, 1978]. Stick-slip behavior is the result of a frictional instability and may be the laboratory equivalent of shallow crustal earthquakes [Bowden and Tabor, 1950; Brace and Byerlee, 1966; Bridgman, 1936]. In addition, laboratory experiments are capable of reproducing microstructures found in natural fault zones, thus providing constraints on the conditions required to produce those fabrics [Logan and Rauenzahn, 1987; Logan et al., 1992; Rutter et al., 1986]. 1
A
version
of
this
chapter
has
been
submitted
for
publication.
Haywood,
J.,
Kennedy,
K.
A.,
In
Review.
Carbonate
and
Shale
Fault
Gouge:
Experimental
insights
into
the
role
of
composition
on
the
mechanical
strength
of
faults.

 12 Previous experiments on carbonate gouges show that these materials are unstable when deformed under a wide variety of conditions [Kawamoto and Shimamoto, 1998; Logan et al., 1992; Shimamoto, 1977; Shimamoto and Logan, 1981]. Kawamoto and Shimamoto [1998] showed that calcite can display violent stick-slip behavior at temperatures up to 700°C. However, stable sliding has been observed at low confining pressures (< 50 - 100 MPa [Logan et al., 1981; Logan et al., 1992; Shimamoto and Logan, 1981]). Pure dolomite gouge also displays violent stick-slip at confining pressures above 60 – 90 MPa, and stable sliding below these confining pressures [Shimamoto and Logan, 1981]. Phyllosilicates typically display stable sliding and strain hardening behavior at low confining pressure and temperature (montmorillonite [Ikari et al., 2011; Logan and Rauenzahn, 1987; Morrow et al., 1992; Tembe et al., 2010], illite [Morrow et al., 1992; Saffer and Marone, 2003; Tembe et al., 2010], muscovite [Mariani et al., 2006; Scruggs and Tullis, 1998; van Diggelen et al., 2009], chlorite [Shimamoto, 1977], kaolinite [Bos and Spiers, 2000]), however they can display stick-slip behavior at high confining pressures (illite, talc, chlorite [Byerlee and Summers, 1973]), and slow sliding rates (illite, montmorillonite [Moore et al., 1986]). Various studies on composite gouges show that mixing a stable mineral endmember with an unstable endmember can either stabilize an unstable endmember or unstablize a stable endmember: Logan and Rauenzahn [1987] showed that the addition of just 5% montmorillonite was enough to suppress stick-slip behavior in quartz resulting in stable slip, while Shimamoto and Logan [1981] showed that adding just 10% anhydrite to a previously stable quartz gouge was enough to induce stick-slip behavior; the stick-slip amplitude increases with increasing anhydrite content. Niemeijer and Spiers [2005] showed that halite 13 and muscovite composite gouges display velocity weakening behavior at high sliding velocities despite the fact the both endmember gouges show no velocity dependence, suggesting that mixtures of these gouge materials may actually be more unstable than either endmember gouge. Experiments on composite gouges containing phyllosilicates (kaolinite/halite [Bos and Spiers, 2000; 2002; Bos et al., 2000]) suggest that the presence of phyllosilicates in gouge mixtures can cause extreme weakening. Barnhoorn et al. ([2005]: calcite/anhydrite) and Delle Piane et al. ([2009]:muscovite/calcite aggregates) suggest that heterogenous phase distribution may be required for strain localization and weakening. Several studies have been conducted on natural gouge material containing both carbonate and phyllosilicates [Carpenter et al., 2009; Collettini et al., 2009a; Lockner et al., 2000; Smith and Faulkner, 2010; Tembe et al., 2006], however the focus of these experiments has been on characterizing a specific fault zone rather than the effect of compositional variations on frictional characteristics. 2.2. Methods 2.2.1. Experimental Conditions Room temperature, triaxial, velocity-stepping frictional sliding experiments were conducted on 25.4 mm diameter by 50 mm length cores containing a 1 mm gouge layer spread along a polished 35° angle saw cut (Figure 2.1). Gouge material was combined 1:1 weight percent with water to create a slurry which was then spread on the saw cut. Prepared samples were jacketed in polyolefin heatshrink tubing. Shale gouge material was created from quartz-bearing phyllosilicate-rich shale (X-ray Diffraction analysis using the Rietveld Method [Rietveld, 1967; 1969] indicates: 31% quartz, 14 39% illite, 18% clinochlore, 11% feldspar) combined in various volumetric proportions with reagent grade calcite powder (80% calcite, 20% dolomite (Appendix B for Rietveld analyses)) with 98% of grains less than or equal to 4.37 µm diameter. Rietveld analysis of the carbonate powder using a 10% CaF2 spike shows the all the carbonate is crystalline. Shale gouge material was powdered to pass through a 147 µm mesh sieve. In the shale gouge, 98% of grains less than or equal to 5.01 µm diameter (Appendix A ; Table 2.1). However, the shale gouge contains coarse grained quartz grains which make up a small proportion of the gouge by number percent, but comprise a significant proportion by volume percent. 50% of the shale by volume is less than or equal to 17.4 µm in diameter, while 95% of the shale by volume is less than or equal to 120.2 µm in diameter. In comparison, 50% of the carbonate is less than or equal to 5.0 µm in diameter, and 95% is less than or equal to 17.4 µm in diameter (Appendix A). The fine grain size of the carbonate was chosen to simulate the grain size found in micrites, which commonly form the hanging wall in foreland thrust systems. A 1.27 cm grid was drawn on the polyolefin jacket to record strain. Prior to every experiment, each sample was weighed, photographed and detailed dimensions were recorded. To avoid ambiguities associated with unknown moisture content, assembled samples were saturated with distilled water for at least 20 hours prior to experiments [Dieterich and Conrad, 1984; Logan and Rauenzahn, 1987]. Experiments were performed on each endmember composition as well as 75%, 50% and 25% mixtures of shale and calcite gouge by volume. 4 separate types of experiments were conducted: 1) gouge material was saturated, and Badshot Dolomite comprised both forcing blocks, 2) gouge material was saturated, and porous Berea Sandstone (ϕ ~ 17%) 15 Figure 2.1: Sample Prep Figure 2.1: Room temperature, triaxial frictional sliding experiments were conducted on 25.4 mm diameter by 50 mm length cores containing a 1 mm gouge layer spread along a polished 35° angle saw cut. Samples were jacketed in polyolefin heatshrink tubing. A 1.27 cm grid was drawn on the polyolefin jacket to record strain, and to confirm that the majority of strain in the samples is accommodated along by precut surface by deformation within the gouge zone. a. Undeformed starting sample, b. Run product, after ~ 6mm axial displacement. This particular experiment was conducted using two non-porous dolomite forcing blocks. 16 Table 2.1: Starting Material Shale Carbonate Composition 31% quartz 39% illite 18% clinochlore 6% albite 5% microcline 80% calcite 20% dolomite Grain size 98% ≤ 5.01µm 98% ≤ 4.37µm Velocity steps 20-3.8-20-100-20 17 comprised the upper forcing block while impermeable Badshot Dolomite comprised the lower forcing block, 3) oven-dried gouge material was used with both forcing block combinations 4) “strain percent” experiments, performed with one sandstone and one dolomite forcing block, in which experiments were terminated at different points throughout the velocity stepping sequence. 2.2.2. Large Sample Rig (LSR) The triaxial rock press used in these experiments is a modified version of the machine used at Texas A & M [Handin et al., 1972; Shimamoto, 1977], and has been described in detail by Austin [2003; 2005]. The sample assembly consists of upper and lower steel pistons, plus several spacers (Figure 2.2). A coat of MoS2 is applied to the ends of the sample to prevent frictional effects at the end of the sample. The sample is wrapped in 2 polyolefin heat shrinkwrap jackets to separate the internal pore space from the applied confining pressure. The entire sample assembly is placed inside a hardened steel pressure vessel. Confining pressures up to 3 kilobars can be achieved using argon gas as the confining medium. The sample is loaded axially using a gear and screw driven press. Force and displacement are measured using an external load cell and external displacement transducer (DCDT). Output data is recorded using National Instruments Labview™ 8.0 Software. The coefficient of friction (µ) was extracted from this data using a program written for MathWorks’ MATLAB v. 7.4., and corrected for the change in contact area during sliding (Appendix C ; Appendix D) 18 Figure 2.2: LSR Figure 2.2: The Large Sample Rig (LSR) is a gear-driven, triaxial rock press that uses argon gas as a confining medium. The LSR can accommodate cylindrical samples up to 50 mm in diameter. The sample assembly consists of upper and lower hardened steel pistons plus several spaces, enclosed by a polyolefin jacket which separates the internal pore pressure of the sample from the applied confining pressure. A through-going pore fluid system allows pore fluid pressure to be added to the sample. A compensator vessel allows the confining pressure to remain constant during displacement. 19 Experiments were conducted at 70 MPa confining pressure. Displacement rates varied between 100 to 102 µm s-1. Velocity stepping procedure follows the fast velocity stepping sequence outlined by Logan and Rauenzahn [1987] (Table 2.1). In each experiment, samples were held at 70 MPa confining pressure for 30 minutes prior to starting the velocity stepping sequence in order to expedite the evolution to steady-state sliding [Chester et al., 1985; Logan and Rauenzahn, 1987; Shimamoto, 1985]. Each sliding velocity in the sequence was sustained until steady-state sliding was achieved. Sliding velocity was then changed as rapidly as possible to avoid changes in frictional strength due to hold time [Dieterich, 1978; Ruina, 1983]. Microstructure analysis was performed to elucidate the effects of strain, composition and forcing block porosity on the mechanical strength and behavior of gouges, as well as on the sequence and types of microstructures developed. Microstructure analysis was conducted using both standard optical microscopy and a Philips XL30 scanning electron microscope (SEM) at the University of British Columbia. Both backscatter electron imaging and element mapping were used to determine the distribution of mineral phases. A specific effort was made to correlate microstructure development with recorded mechanical data. 2.3. Results 2.3.1. Mechanical Data 2.3.1.1. Effects of Compositional Variation Figure 2.3 shows mechanical data for all compositions deformed with 2 non-porous forcing blocks as well as those deformed with 1 porous and one non-porous forcing block. Under the observed experimental conditions, all gouge compositions deformed using one 20 porous forcing block display steady-state or near steady-state sliding at maximum shear strains. There are a few stress drops that may correspond to failure in the forcing block. Strain hardening follows yielding, and evolves to steady-state or near steady-state sliding. True steady-state sliding is only observed in carbonate-rich gouges (100% carbonate and 75% carbonate). 100% carbonate gouge is the strongest while the composite gouge compositions are weakest. 100% carbonate gouge displays pronounced strain hardening at the lowest, 3.8 µm/s, strain rate. At maximum shear strains, the 50% shale gouge is the weakest material. A change in sliding velocity results in a change in the value of the coefficient of friction, showing that all gouge materials are velocity dependant. Replicate experiments are shown in Appendix E. Although there is some variation in the coefficient of friction between replicate experiments, the 50% and 75% composites gouges are as weak or weaker than the 100% shale gouge at maximum shear strain (Figure E-1). True stick-slip behavior is not observed in any gouge materials despite previous work that suggests carbonate gouge should display stick-slip sliding under similar conditions [Logan et al., 1992; Shimamoto, 1977; Shimamoto and Logan, 1981]. 100% carbonate gouge in the present study does, however, exhibit velocity weakening behavior under some conditions, suggesting that this gouge material may promote seismic faulting rather than aseismic creep [e.g. Scholz, 1998 and references therein]. All other compositions display velocity strengthening behavior at the conditions tested. Velocity strengthening behavior, in which an increase in displacement rate corresponds with an increase in the coefficient of friction, indicates that faults containing these gouge materials at similar conditions will slip aseismically [e.g. Scholz, 1998 and references therein]. 21 2.3.1.2. Effects of Forcing Block Porosity There is a marked contrast in observed coefficient of friction between experiments performed with 1 porous forcing block and those performed with 2 non-porous forcing blocks. Experiments performed with two non-porous forcing blocks and those performed with one porous forcing and one non-porous forcing block show the same general trends related to gouge composition; 100% carbonate gouge is the strongest while the composite gouge compositions are the weakest. However, samples deformed using a porous forcing block typically display a coefficient of friction .10-.20 higher than samples of the same composition deformed using 2 non-porous forcing blocks (Figure 2.3, Table 2.2). The exception, 100% carbonate gouge samples deformed with two non-porous forcing blocks, is .20 stronger than 100% carbonate experiments performed with one porous forcing block. The relative strength of samples with a porous forcing block compared to those with 2 non-porous forcing blocks can be explained by localized changes in pore fluid pressure. Although no additional pore fluid pressure was added, experiments were conducted on water saturated samples. During deformation, porosity in the sandstone forcing block allows fluid to escape from the gouge zone into the forcing block, transiently lowering the effective pore pressure in the gouge zone and leading to an increase in sample strength. In contrast, non- porous forcing blocks trap pore fluid in the gouge zone leading to transiently higher effective pore pressures and decreased sample strength. 100% carbonate samples with two non-porous forcing blocks attain true steady-state sliding at maximum shear strains. However, experiments performed with two non-porous forcing blocks on all other compositions do not attain steady-state or near steady-state 22 Figure 2.3: Mechanical Data 0 1 2 3 4 5 60 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Displacement (mm) Co ef fic ie nt o f F ric tio n 100% Shale, 17% FB Φ 75% Shale 17% FB Φ 50% Shale 17% FB Φ 25% Shale 17% FB Φ 100% Carb.17% FB Φ 100% Shale no FB Φ 75% Shale no FB Φ 50% Shale no FB Φ 25% Shale no FB Φ 100% Carb. no FB Φ velocity stepping sequence (µ/s) =20 - 3.8 - 20 - 100 - 20 Figure 2.3: Velocity-stepping experimental data. Displacement vs. coefficient of friction (µ) for all compositions showing samples deformed with 2 non-porous Badshot Dolomite forcing blocks (dashed line) and samples deformed with 1 porous Berea Sandstone forcing block (~17% ϕ) and 1 non-porous dolomite forcing block (solid line). All samples were saturated and deformed at 70 MPa confining pressure and with displacement rates between 100 to 102 µ s-1. All samples display velocity dependency. When possible, steady-state or near steady- state sliding was attained before sliding velocity was changed. Samples with two non-porous forcing blocks continued to strain harden throughout. Composite gouges are weakest and the carbonate gouge is the strongest. 100% shale gouge is intermediate in strength. There is a notable difference between the coefficient of friction observed in samples with 2 non-porous forcing blocks compared to the coefficient of friction observed in corresponding samples with 1 porous forcing block. 23 Table 2.2: Experimental Data Sample Composition Forcing Blocks Pore Fluid Conditions Displacement Rate (µ s-1) Peak Differential Stress (MPa) Peak Coef. of Friction Yield Strength Coef. Of Friction at Maximum Shear Strain Steady State or Strain Hardening? 100% Shale sandstone/ dolomite saturated 20-3.8-20-100- 20 159 0.68 0.46 0.68 strain hardening 75% Shale sandstone/ dolomite saturated 20-3.8-20-100- 20 127 0.61 0.42 0.6 strain hardening 50% Shale sandstone/ dolomite saturated 20-3.8-20-100- 20 141 0.65 0.49 0.62 strain hardening 25% Shale sandstone/ dolomite saturated 20-3.8-20-100- 20 171 0.69 0.53 0.69 strain hardening 100% Carbonate sandstone/ dolomite saturated 20-3.8-20-100- 20 185 0.72 0.43 0.71 steady state 100% Shale dolomite/ dolomite saturated 20-3.8-20-100- 20 103 0.56 0.27 0.55 strain hardening 75% Shale dolomite/ dolomite saturated 20-3.8-20-100- 20 82 0.5 0.25 0.5 strain hardening 50% Shale dolomite/ dolomite saturated 20-3.8-20-100- 20 74 0.47 0.26 0.47 strain hardening 25% Shale dolomite/ dolomite saturated 20-3.8-20-100- 20 92 0.48 0.25 0.47 strain hardening 100% Carbonate dolomite/ dolomite saturated 20-3.8-20-100- 20 237 0.82 0.47 0.81 steady state 50% Shale sandstone/ dolomite dry 20-3.8-20-100- 20 178 0.72 0.52 0.71 strain hardening 50% Shale dolomite/ dolomite dry 20-3.8-20-100- 20 167 0.73 0.46 0.73 strain hardening 100% Carbonate sandstone/ dolomite dry 20-3.8-20-100- 20 211 0.74 0.49 0.72 steady state 100% Carbonate dolomite/ dolomite dry 20-3.8-20-100- 20 234 0.81 0.56 0.8 steady state 24 sliding; they continue to strain harden throughout. This is in contrast to experiments performed with one porous forcing block in which steady-state or near steady-state sliding is achieved. At low displacement rates, experiments performed with two non-porous forcing blocks show more pronounced strain hardening with increased carbonate content. Presumably, additional sliding distance is required to complete the evolution to steady state, however the amount of displacement in these experiments is limited by the amount of strain the polyolefin jacket can accommodate before it ruptures. 2.3.1.3. Effects of Pore Fluid Dry experiments were aimed at separating the effects of pore fluid pressure from the effects of forcing block porosity. Dry experiments using both forcing block configurations were performed on the weak, 50% shale gouge mixture and the strong, 100% carbonate gouge mixture. Figure 2.4 shows that all dry experiments record coefficients of friction that are higher than their saturated counterparts. The dry 50% shale experiment with 1 porous forcing block records a coefficient of friction that is .09 higher than the saturated experiment, while the dry 50% shale experiment with 2 non-porous forcing blocks records a coefficient of friction that is .26 higher than the corresponding saturated experiment (Figure 2.4a). These observations support that localized changes in pore fluid pressure cause the observed weakness of the non-porous samples relative to the porous samples. Both the dry and saturated experiments performed with 1 non-porous forcing block attain steady-state sliding while both the dry and saturated experiments performed with 2 non-porous forcing blocks strain harden throughout the experiment. This observation suggests that the presence of a porous forcing block promotes the evolution to steady state sliding. This idea will be investigated in more detail in the microstructure section of this paper. 25 Figure 2.4 (following page): Velocity stepping experiments aimed at separating the effects of pore fluid pressure from the effects of forcing block porosity. Oven dried samples are shown by solid lines, saturated samples are shown by dashed lines. a. Dry 50% shale experiments record higher coefficients of friction than their saturated counterparts. The dry 50% shale experiment with 1 porous forcing block records a coefficient of friction that is .08 higher than the saturated experiment, while the dry 50% shale experiment with 2 non-porous forcing blocks records a coefficient of friction that is .35 higher than the corresponding saturated experiment. b. The 100% carbonate, dry experiment with two non-porous forcing blocks also displays a coefficient of friction approximately .35 higher than the saturated experiment. The dry, 100% carbonate gouge deformed with 1 porous forcing block has a coefficient of friction that is .08 lower than the dry sample with two non porous forcing blocks. Note the abrupt ~.03 decrease in the coefficient of friction after 2.25 mm of displacement visible in the dry 100% carbonate sample with one porous forcing block. c. When plotted together, the coefficient of friction for the 100% carbonate samples deformed with a porous forcing block, as well as the dry, 50% shale sample deformed with a porous forcing block plateaus at a steady-state value of ~.7, regardless of composition or the presence of pore fluid, suggesting that in experiments with a porous forcing block forcing block strength provides the upper limit on sample strength. 26 Figure 2.4: Dry Graphs 27 100% carbonate with two non-porous forcing blocks displays the highest coefficient of friction. At maximum shear strains, the dry experiment with two non-porous forcing blocks records approximately the same coefficient of friction as the saturated experiment with two non-porous forcing blocks. The dry, 100% carbonate gouge deformed with 1 porous forcing block has a coefficient of friction that is .08 lower than the dry sample with two non porous forcing blocks. In addition, the dry sample with one porous forcing block records an abrupt ~.03 decrease in the coefficient of friction after 2.25 mm of displacement (Figure 2.4b). The steady-state coefficient of friction following the stress drop is approximately the same as the steady-state coefficient of friction recorded by the saturated sample with 1 porous forcing block, suggesting that in the 100% carbonate gouge experiments forcing block porosity, rather than pore fluid pressure controls sample strength. The steady-state coefficient of friction recorded by the 50% shale sample with 1 porous forcing block is ~.71, which also corresponds to the steady-state coefficient of friction observed in both the dry and saturated carbonate experiments with 1 non-porous forcing block (Figure 2.4c), providing further evidence that forcing block fracture strength provides a limit on peak sample strength. 2.3.2. Microstructure Analysis 2.3.2.1. Compositional Variations and Microstructure Development Experimental products were examined using both standard optical microscopy and a Philips XL30 scanning electron microscope at the University of British Columbia to determine how variations in gouge composition affect microstructure development, and the resulting mechanical response. In the following section microstructure observations on 28 samples with one sandstone and one dolomite forcing block will be presented. The following microstructure descriptions will use the Riedel shear zone terminology outlined and used by Skempton [1966], Bartlett et al. [1981], Logan et al. [1992], and Rutter [1986] (Figure 2.5). SEM element maps of a 100% shale gouge sample shows that mineral phases are homogenously distributed throughout the gouge zone. Figure 2.6a shows quartz clasts define a shape (P-) fabric within the gouge zone. Scanning electron microscopy shows evidence for brittle grain size reduction along P-shears, parallel to the inclined quartz grains defining the shape fabric. Large quartz grains in the middle of the gouge zone show evidence of some brittle fracturing. Phyllosilicates across the middle of the gouge zone change color at a uniform angle when using the λ-plate on the optical microscope, suggesting that phyllosilicates across the gouge zone are preferentially aligned. These aligned phyllosilicates suggest that deformation is distributed throughout the gouge zone [Logan, 2007; Rutter et al., 1986]. At the gouge zone margins, phyllosilicates are preferentially aligned with either other, but not with phyllosilicates in the middle of the gouge zone, providing evidence of strain localization at the gouge zone margin (Figure 2.6a). Backscatter electron imaging shows compaction and grain comminution (Figure 2.6b) at the gouge zone margin. The relative strength of the 100% shale is likely due to the diffuse character of deformation within the sample, and the failure of strain to localize along discrete shear bands away from the gouge zone margin. Deformation within 75% shale gouge is more heterogeneous than that observed in the 100% shale gouge. Quartz defines a weak shape fabric with several P-shears parallel to this fabric. Optical petrography shows discontinuous bands of aligned phyllosilicates that form 29 Figure 2.5: Riedel Shear Zone Terminology Figure 2.5: Schematic diagram showing types of strain localization in riedel shear zones 30 Figure 2.6: Microstructure I 1 mmdol sst 1 mm 100% Shale b c a d 100% Shale 75% Shale 75% Shale Shear Zone Margin dol sst PP 300 μm 100 μm cracking in qtz cracking in qtz Figure 2.6: Microstructures of run products. a. 100% shale gouge, saturated with a sandstone (sst) hanging wall (top) and dolomite footwall (bottom). Sample shows approximately 12% strain that is primarily distributed throughout the gouge zone. Quartz clasts define a shape fabric that leans towards the direction of shearing (cross-polarized optical microscope with gypsum plate). b. SEM BSE image showing some strain localization near the shear zone margin. c. 75% shale sample, saturated, displays discontinuous margin parallel Y- shears (cross-polarized optical microscope with gypsum plate). d. SEM element of a 75% shale gouge sample, saturated, shows compositional banding (green = Si; blue = Ca). 31 anastomosing shear zone bands at a low angle to the shear zone margin (Y-shears) (Figure 2.6c). SEM element mapping shows the development of compositional banding interpreted to represent mechanical separation of phyllosilicate and carbonate minerals Figure 2.6d. SEM backscatter images show evidence of grain size reduction in thin, near margin parallel, phyllosilicate-rich bands suggesting strain localization in these bands. The resulting fabric is a composite P-Y fabric, defined by aligned phyllosilicates. Brittle comminution of large quartz grains can be seen in places. Both the compositional banding and mineral preferred orientations allow for strain localization within the gouge zone, resulting in the low coefficient of friction observed in 75% shale gouge samples. A zone of strain localization is observed at the gouge zone margin. In 50% shale gouge samples, margin parallel Y-shears containing preferentially aligned phyllosilicates are visible using both the SEM and optical microscope (Figure 2.7a, b). These Y- surfaces show evidence of grain size reduction in bands up to several hundred µm thick. Within these zones are bands of even more intense grain size reduction, suggesting further strain localization. SEM element mapping indicates that the most fine grained bands contain phyllosilicates and quartz, and that these bands contain very little porosity. Much of the gouge from this sample did not survive the thin sectioning process, however at least 2 of these Y- shears appear to extend the entire length of the gouge zone. One of these is at the gouge zone margin, the other is in the middle of the gouge zone. 32 Figure 2.7 (following page): Microstructures of run products. a. 50% shale saturated gouge shows through going margin parallel Y-shears (cross-polarized optical microscope with gypsum plate). b. 75% saturated shale gouge. SEM element map shows phyllosilicates (red) define a shape (P-) fabric as well as fill the margin parallel Y shears (green = Si; blue = Ca; red = Al) . c. 25% saturated shale gouge. R1 shears form at an angle to the shear zone and dip towards the direction of shearing (cross-polarized optical microscope with gypsum plate). d. In 25% shale gouge, saturated, phyllosilicates define the shape fabric (SEM element map, red = Al). e. R1 shears also form in 100% carbonate samples, saturated (cross-polarized optical microscope with gypsum plate), f. However, most strain is distributed across the shear zone into many, very thin, margin parallel shear bands (SEM BSE image). 33 Figure 2.7: Microstructure II sst sst 50% Shaleb d 25% Shale f 100% Carbonate Shear Zone Margin 115 μma 50% Shale dol sst e 100% Carbonate 1 mm sst dol 1 mm R1 R1 c 25% Shale sst dol Shear Zone Margin 40 μm 40 μm 100 μm P P R1 R1 Y Y P Y Shear Zone Margin 34 Outside of the Y-shears, the average grain size is larger, and porosity is notably higher. The decrease in porosity within the Y-shears is likely due to brittle grain size reduction and phyllosilicate alignment within these bands. Asymmetric lenses of phyllosilicates define a shape fabric. Figure 2.7b shows where this shape fabric transitions into margin parallel Y- shears. Also in this zone are poorly developed R1-shears that crosscut the shape fabric. The ability of strain to primarily localize into phyllosilicate-rich, margin parallel shear bands likely accounts for relatively low coefficient of friction observed in this gouge composition. In 25% shale gouge samples, strain within the gouge zone is accommodated primarily by R1 shears dipping approximately 25° from the gouge zone margin (Figure 2.7c). Phyllosilicate – rich lenses define a shape fabric that forms at a high angle to the shear zone boundary, and leans towards the direction of shearing (Figure 2.7d). Phyllosilicates also preferentially fill R1 shear bands. Scanning election microscopy shows grain size reduction in R1 shears and at both gouge zone margins. It is significant to note that phyllosilicates are not present in sufficient quantities to form a through going fabric. 100% carbonate gouge displays well-developed R1 shears (Figure 2.7e). Backscatter electron imaging shows grain size reduction along these R1 shears, and in addition shows many more low angle R1 shears that are too small to see with an optical microscope (Figure 2.7f). Grain size reduction is also visible at the dolomite margin, while R1 shears from the gouge zone extend into the sandstone forcing block. Within the sandstone forcing block there is evidence for brittle fracturing of quartz grains and porosity reduction. Additional 35 descriptions of strain localization within the sandstone forcing block will be made in section 2.3.1.2. 2.3.2.2. Strain and Microstructure Development Experimental products record only the microstructures present during the final stages of deformation. Microstructures that formed earlier in the deformation history are not preserved [Logan, 1992]. Mechanical data (Figure 2.3) shows 50% shale gouge to be the weakest material at the highest shear strains. In order to correlate recorded mechanical data with microstructures, experiments were conducted that attained 25% and 80% of the maximum displacement. Strain percent experiments were terminated at the end of the first 20 µm/s velocity step, the end of the 100 µm/s velocity step and at the end of the entire velocity stepping sequence in order to determine whether microstructure development corresponds with the observed changes in sample strength and sliding behavior. Experiments were performed on saturated 50% shale gouge with one sandstone forcing block. This composition was chosen because it is the weakest material at high shear strains and showed the most diverse microstructure. In 50% percent shale experiments, strain hardening occurs throughout the first 3 velocity steps (Figure 2.8). In the middle of the fourth (100 µm/s) velocity step there is a decrease in the rate of velocity-strengthening and near steady-state sliding is achieved. A pronounced decrease in coefficient of friction occurs at start of the final (20 µm/s) velocity step. Experimental data shows good reproducibility of experiments (Figure 2.8). A reference sample of 50% gouge material that only records the pre-compaction process shows a moderately well-developed shape fabric evidenced by leaning quartz clasts, in an 36 otherwise homogenous mixture of phyllosilicate and carbonate gouge material (Figure 2.9a). Following the first, 20 µm/s velocity step, gouge material shows initial strain localization at the margin and weakly developed R1 shears (Figure 2.9b). With increasing strain, well-developed R1 shears form (Figure 2.9c). Finally, well- developed, through going, margin parallel Y-shears develop (Figure 2.9d), and some compositional banding is visible. Decreased frictional strength and the evolution to steady- state sliding correspond to strain localization along the Y-shears. This sequence of microstructure development is well documented in both monomineralic and polymineralic gouges [Logan and Rauenzahn, 1987; Logan et al., 1992; Rutter et al., 1986]. My findings are consistent with these studies in which a P-fabric (also known as a shape fabric), followed by Riedel shear development and finally by the development of through going Y-shears, or boundary shears [Gu and Wong, 1994; Logan and Rauenzahn, 1987; Logan et al., 1992]. Based on the sequence observed above, gouge zone maturity and the degree of strain localization within the gouge zone can be determined. 37 Figure 2.8: Strain Percent Graph Figure 2.8: Coefficient of friction vs. displacement graph of 50% shale experiments. A series of experiments was conducted to analyze the effects of strain on microstructure development. Experiments were terminated at the end of the first 20 µm/s velocity step (b), the end of the 100 µm/s velocity step (c) and at the end of the entire velocity stepping sequence (d) in order to determine whether microstructure development corresponds with the observed changes in sample strength. Experiments were performed with one porous forcing block and one non- porous forcing block. In experiment d, there is a noticeable stress drop preceding the final velocity step followed by unsteady sliding. Note good experiment reproducibility. 38 Figure 2.9: Strain Percent Photos .25 mm strain localization at margin .25 mm R1 shear development .25 mm Margin parallel shears b 300 μm P a c d P P R1 Figure 2.9: Run products derived from experiments a, b, c and d shown in Figure 2.8. All experiments were performed on saturated 50% shale gouge using one porous forcing block and one non-porous forcing block. a. SEM element map showing the effects of confining pressure (Pc). Mineral phases remain distributed throughout the gouge zone however a weak shape fabric forms (green = Si; blue = Ca; red = Al). b. With increasing displacement, strain begins to localize at the shear zone margin (cross-polarized optical microscope with gypsum plate), c. Followed by the development of R1 shears (cross-polarized optical microscope), d. and finally the formation of through going margin parallel Y-shears (cross-polarized optical microscope with gypsum plate). 39 2.3.2.3. Effects of Forcing Block Porosity Mechanical data shows that samples deformed using a porous forcing block typically have a coefficient of friction .10-.20 higher than their non-porous counterparts, likely due to transient variations in pore fluid pressure within the gouge zone. Forcing block porosity allows pore fluid to escape into the forcing block, increasing sample strength. Microstructure observations, however, suggest that forcing block porosity may also have several weakening or sample strength limiting effects that are overprinted by the magnitude of the pore fluid pressure effect. Microstructure observations indicate that increased forcing block porosity leads to strain localization within the gouge zone resulting in more rapid microstructure development and higher shear strains within the gouge zone. In addition, significant deformation that occurs within the porous forcing block also serves to limit sample strength. Porosity in the sandstone creates a cuspate-lobate gouge zone margin (Figure 2.10a). In contrast, the gouge zone margin with the dolomite forcing block is planar (Figure 2.10b). Optical petrography shows gouge material filling sandstone porosity at the gouge zone margin, which effectively decreases the gouge zone thickness. Porosity in the sandstone means increased sliding block surface roughness, leading to increased traction at the gouge zone margin and promoting strain localization. As strain is localized and the effective gouge zone thickness is reduced, the shear strain in these bands increases dramatically resulting in more well-developed microstructure at comparable displacements. For example, margin parallel Y-shears develop in the 50% shale gouge deformed with a porous forcing block (Figure 2.10c), while shear zone margin oblique Riedel shears develop in 50% shale gouge 40 Figure 2.10: Porosity I R1 100 μmd 50% Shale, non-Φ 1 mm dol sst gouge 100% Carbonate, Φa 1 mm gouge dol dol 100% Carbonate, non-Φb c 50% Shale, Φ P Y Y P P 100 μm Figure 2.10: Photomicrographs showing the effects of forcing block porosity on deformation within the gouge zone. a. Gouge fills porosity in the sandstone creating a cuspate-lobate margin and effectively decreasing the shear zone thickness. b. In contrast, the dolomite shear zone margin is planar. As a result, more well developed microstructures form at lower displacements and the evolution to steady-state sliding occurs over lower displacements in samples with one Berea Sandstone forcing block. c. Well-developed, margin parallel Y-shears form in 50% shale gouge deformed with 1 porous forcing block, while d. Riedel shears from comparable displacements in 50% shale gouge deformed using 2 non-porous forcing blocks. 41 Figure 2.11: Porosity II Figure 2.11: Photomicrographs showing the effects of forcing block porosity on deformation within forcing block. a. R1 shears developed in the gouge zone extend into the sandstone forcing block. b. In some cases, gouge material is injected into the sandstone forcing block. This is in contrast to the dolomite gouge margin, in which no significant forcing block deformation is observed. 42 deformed with 2 non-porous forcing blocks (Figure 2.10d) at comparable displacements. More rapid strain localization allows samples with a porous forcing block to attain steady- state or near steady state sliding within the observed sliding distance, while samples with 2 non-porous forcing blocks continue to strain harden throughout (Figure 2.3). Microstructure observations on the 25% shale and 100% carbonate samples show significant deformation within the sandstone forcing block. R1 shears extend from the gouge zone into the sandstone forcing block (Figure 2.11a). In places, gouge material is injected into the fractures in the sandstone forcing block (Figure 2.11b). Mechanical data shows that the 25% shale and 100% carbonate gouge materials are the strongest. Data from dry experiments shows that the observed frictional strength of the 100% carbonate gouge peaks at the at the observed fracture strength of the Berea Sandstone. In addition, the dry, 100% carbonate sample records an abrupt .03 drop in the coefficient of friction. The observed stress drop likely records failure within the sandstone, visible in thin section as well-developed R1 shears in both the gouge zone and in the Berea Sandstone forcing block. Microstructure observations, combined with the mechanical data that shows that samples with a Berea Sandstone forcing block display a coefficient of friction that plateaus at approximately .7 regardless of gouge composition, suggest that the frictional strength of the relatively strong, carbonate-rich gouge materials exceeds the strength of a porous forcing block. In this situation, deformation extends into the forcing block, widening the zone of deformation. 43 2.4. Discussion 2.4.1. Comparison to Previous Experimental Data Coefficient of friction values observed in this study are similar to the coefficients of friction observed in previous studies on similar materials. The measured µ values of .71-.81 for the 100% carbonate gouge are comparable to published experimental data that record friction coefficients between .7 - .85 [Kawamoto and Shimamoto, 1998; C Morrow et al., 2000] (Appendix F). Previous studies record both stable sliding [Brace, 1972; Logan et al., 1992] and stick-slip behavior of carbonate gouge [Kawamoto and Shimamoto, 1998; Logan et al., 1992] at comparable confining pressures and room temperature humidity; my saturated experiments displayed stable sliding. Previous experiments [Logan et al., 1992; Logan et al., 1979; Moore et al., 1988; Shimamoto and Logan, 1981] suggest that strain localization, typically in the form of through going Y-shears, is required for stick-slip behavior to occur. In coarse grained gouges, grain size reduction and compaction are required before strain localization in discrete bands can occur [Logan and Rauenzahn, 1987; Logan et al., 1981] suggesting that the fine grain size (<5 µm) in the carbonate gouge material should promote strain localization and stick-slip sliding. However, it is likely that due to the limited amount of displacement that is possible using a triaxial apparatus I was unable to reach sufficient shear strain for stick-slip behavior to occur. Furthermore, porosity reduction and strain localization in the porous forcing block occurs at the expense of porosity reduction and strain localization in the gouge zone and may help to account for the absence of stick-slip sliding in samples with a porous forcing block. The 100% shale gouge has a coefficient of friction of .52-.67 depending on the forcing blocks used, and displays stable sliding at all conditions tested. These values are comparable 44 to the coefficient of friction (µ = .43-68) recorded when natural illite bearing shale is deformed at a wide variety of normal stresses (σn = 5 – 150 MPa) [Saffer and Marone, 2003]. The 50% and 75% shale experiments performed in this study have coefficients of friction of ~.40 or ~.6 (sst/dol and dol/dol forcing blocks respectively). The lower µ observed in the experiments with the dolomite forcing blocks reflects locally higher pore fluid pressures (i.e. lower effective pressures). These values are consistent with the coefficient of friction observed in natural illite bearing shale [Saffer and Marone, 2003], and similar to or slightly higher than the values observed in pure illite (µ = .41 -. 48) deformed at effective confining pressures (Pe) of 200-300 MPa [Morrow et al., 1992]. In the present study, gouges containing phyllosilicates strain hardened under all conditions tested. Published experimental data on bimineralic gouge mixtures containing clay minerals demonstrate that the addition of small amounts of velocity strengthening clay minerals, results in stable sliding and strain hardening of the composites (e.g. quartz/montmorillonite [Logan and Rauenzahn, 1987], calcite/halite [Kawamoto and Shimamoto, 1998], kaolinite/quartz [Rutter et al., 1986]) . In addition, these experiments show that adding a phyllosilicate to gouge material fundamentally alters the dominant deformation mechanism. For example, the primary mechanism of deformation in quartz gouge is brittle grain size reduction [Logan and Rauenzahn, 1987; Rutter et al., 1986] as a result of grain to grain contact. The presence of phyllosilicates can result in the formation of a pervasive foliation, which in turn, decreases the amount of contact between quartz grains, limiting the amount of brittle grain size reduction that can occur in quartz [Logan and Rauenzahn, 1987]. In homogenously mixed bimineralic fault gouges, it is reported that 45 ~75% of phyllosilicate gouge is needed for phyllosilicates to form a through-going foliation [Logan and Rauenzahn, 1987]. This is in contrast to the present experiments in which a through-going phyllosilicate foliation develops in the 50% shale (~29% phyllosilicate) and 75% shale (~43% phyllosilicate) gouge mixtures. 2.4.1.1. Microscale Deformation Mechanisms and Foliation Development Strong end member gouge compositions display more distributed deformation than composite gouges, with little evidence of strain localization away from the gouge zone margin. In the 100% carbonate gouge, grain rotation and grain boundary sliding rather than brittle grain size reduction are the dominant deformation processes, likely because the fine initial grain size inhibits further grain size reduction. The extreme weakness of the composite fault gouges is due to the ability of strain to localize along phyllosilicate-rich shear bands, thereby decreasing the effective thickness of the zone within which displacement occurs. That multiple mineral phases are needed to promote strain localization is well documented for several gouge compositions [Barnhoorn et al., 2005; Handy, 1990; Holyoke III and Tullis, 2006]. The 100% shale gouge deforms primarily through cataclastic flow, with grain boundary sliding and brittle grain size reduction as endmember cataclastic flow processes [Blenkinsop, 1991; Evans, 1988; Passchier and Trouw, 2005; Sibson, 1977]. In the more coarse-grained components of the 100% shale fault gouge, fracturing and comminution, visible in relatively large quartz grains, are important deformation mechanisms that limit the rate of cataclastic flow [Passchier and Trouw, 2005]. In the present experiments, even without the addition of externally-imposed pore fluid pressures, flow is homogeneously distributed, and hence, foliation development, defined by aligned phyllosilicates oblique to the shear zone boundary, 46 is also homogeneous, at the scale of the gouge zone. As a result, even at maximum shear strains, deformation is unable to effectively localize and remains diffuse. Hence, shear strain is not localized along discreet layers, which makes the gouge apparently ‘stronger’, than a gouge with strain localization (i.e. composite gouges). The 50% and 75% shale gouge materials may be exceptionally weak because flow within the fine grained carbonate gouge facilitates the formation of compositional banding and the rotation of phyllosilicate grains into shear zone parallel Y shear surfaces. At high shear strains, most of the displacement is accommodated by frictional sliding along these margin parallel bands of phyllosilicate foliae. Strain localization along the comparatively weak phyllosilicate phases reduces the strength of the 50% and 75% samples relative to the 100% shale sample. These experiments show the initial phases of compositional stratification and strain localization. It is expected that with increased displacement a pervasive shear zone –parallel foliation would form. In natural shear zones, a pervasive foliation can develop with small amounts of phyllosilicates (6-15% talc [Collettini et al., 2009c]; 10-15% chlorite [Jefferies et al., 2006]), which results in significant fault zone weakening owing to strain localization along phyllosilicate-rich Y- surfaces [Collettini et al., 2009b; Collettini et al., 2009c]. 2.4.2. Fault Core Width Experiments in the present study suggest that the contrast in mechanical strength between gouge material and host rock provides an important control on fault zone core thickness, composition and geometry. In these experiments, the strength contrast between the strong, carbonate gouge material and the porous sandstone forcing block is much less than the contrast in strength between the carbonate gouge material and the non-porous dolomite 47 forcing block. As a result, in carbonate-rich gouge samples deformed with a porous forcing block, deformation extends from the gouge zone into the porous forcing block. Forcing block material is then incorporated into the gouge zone, changing the composition of the gouge zone and creating a wider zone of deformation. Dry, 100% carbonate experiments performed with a porous forcing block record a µ ~.1 lower than 100% carbonate experiments performed with a non-porous forcing block (Figure 2.3). This is because the fracture strength of the sandstone, rather than the frictional strength of the carbonate, limits the strength of the sample. In contrast, in experiments with strong, non-porous forcing blocks all deformation is restricted to the gouge layer, which exhibits, sharp, straight contacts with the forcing blocks These experiments support field studies that demonstrate the importance of competence contrast between the host rock and fault zone core in the resultant geometry and thickness of the fault zone core, and in the amount of mixing that occurs in the fault zone core [Faulkner et al., 2003; Faulkner et al., 2010; Wibberley et al., 2008], which in turn affects fault zone core rheology. Fault zones that form in low porosity, strong, crystalline rocks contain a single or few gouge zones that are relatively planar (~50cm wide in the Punchbowl Fault, California [Chester et al., 1993]). Fault zones formed in high-porosity sedimentary rock and foliated metamorphic rock, for which the contrast in strength between the fault zone core and the host rock is much smaller, more deformation occurs in the areas adjacent to the fault zone core and blocks of host rock are incorporated into the fault core zone (e.g. the Carboneras Fault, Spain[Faulkner et al., 2003]). This leads to a wider, in some cases anastamosing, fault core zone whose composition evolves throughout time as more host rock material is added to the gouge zone [Faulkner et al., 2003; Faulkner et al., 2010]. 48 2.4.3. Implications for Natural Thrust Faults Experiments from this study indicate that shale-carbonate composite gouge is weaker than either the shale or carbonate end member gouges. In many fold and thrust belts, carbonates comprise the hanging wall of fault zones and shale forms the footwall [Burkhard, 1990; Erickson, 1994; Kennedy and Logan, 1998; Wojtal and Mitra, 1986]. Generally, a cataclasite is developed in both the carbonate and shale, with a combined thickness of less than 1 meter. Typically, there is mechanical mixing between the hanging wall carbonate cataclasite and the footwall shale cataclasite, such that shale cataclasites contain limestone and limestone cataclasite clasts and vice versa [Erickson, 1994; Kennedy and Logan, 1998; Wojtal and Mitra, 1986]. The contact between the two cataclasites is commonly sharp, whereas the contacts between the cataclasites and their respective country rock are gradational. These experiments indicate that the composite cataclasite is the weakest material and is most likely to accommodate significant amounts of strain. However, experiments were performed at room temperature, with no added external pore fluid pressures. Field observations demonstrate that elevated pore fluid pressures certainly are present in most upper level fault zones, particularly in thrust faults [Neuzil, 1995; Sibson, 2003; Wibberley et al., 2008]. In the present study, in the absence of pore fluid pressure and elevated temperature, carbonate gouge is much stronger than shale gouge and composite gouge which would infer that deformation would be localized at the hanging wall-footwall contact, and/or within the shales. However, the development of cataclasites within limestone thrust sheets attests to deformation within the carbonate [Erickson, 1994; Kennedy and Logan, 1997; Kennedy and Logan, 1998]. Thus, in order to have the carbonate thrust sheet develop a cataclasite and 49 accommodate significant shear strain (γ), pore fluid pressure, and likely, elevated temperature must be present. Field studies show that under a pore fluid pressure and at elevated, but low T (<200°C) [Erickson, 1994; Kennedy and Logan, 1997; Kennedy and Logan, 1998; Wells et al., 2010; Wojtal and Mitra, 1986], solution transfer processes and grain boundary sliding act in parallel to significantly reduce the strength of the overlying limestone cataclasite [Durney, 1972; Elliott, 1976; Gratier et al., 1999; Rutter, 1983]. The experiments in the present study suggest that even at small grain sizes (~5 µm), pore fluid pressure and temperature are required to generate strain softening of the carbonate. Is strain accommodated homogeneously through the shale cataclasite and limestone cataclasite during displacement of thrust sheets, or is displacement localized in one versus the other as a function of physical conditions (i.e. pore fluid pressure)? It is likely that fluctuations in pore fluid pressure conditions through time will promote switches in strain localization between the carbonate and shale cataclasites. In this scenario, elevated pore fluid pressure promotes strain localization in the hanging wall cataclasite by processes of grain boundary sliding and solution transfer: hence, the fault zone is relatively weak. However, low (near hydrostatic) pore fluid pressures strengthen the carbonate cataclasite relative to the underlying shale cataclasite and shear strain is localized in the footwall shale cataclasite by processes of frictional sliding: hence the fault zone is relatively strong. This switch in strain localization, and fault zone strength, is likely cyclical over the timescales of thrust fault displacement. 50 2.5. Conclusions 1) 100% carbonate gouge is the strongest while the composite gouge compositions are weakest. At the highest shear strain, the 50% shale gouge is the weakest material. All gouge compositions tested show velocity dependency at the conditions tested. Calcite exhibit velocity weakening behavior under some conditions, but all other compositions exhibit velocity strengthening behavior. 2) In thrust faults, carbonate-rich shale cataclasite is the weakest material. Under experimental conditions is most likely to accommodate significant amounts of strain. However, field evidence suggests that significant displacement is accommodated by both the shale and carbonate cataclasites, and that pressure solution and temperature play in important role in deformation within the carbonate cataclasite. Fluctuations in the pressure, temperature and pore fluid pressure conditions through time may be needed to promote strain localization in one type of cataclasite over the other. 3) There is a marked contrast in the observed coefficient of friction between experiments with 1 porous forcing block and those with 2 non-porous forcing blocks. Porosity in the forcing block allows fluid to escape. This lowers the effective pore pressure in the gouge zone, which leads to increased sample strength - provided the gouge is weaker than the forcing blocks. 4) Porosity in the sandstone means increased sliding block surface roughness and increased traction at the gouge zone margin, resulting in more rapid microstructure 51 development and increased strain localization in the gouge zone. As a result, there is a decrease in effective gouge zone thickness leading to higher shear strains in samples with high forcing block porosity. 5) When the frictional strength of the gouge material exceeds the fracture strength of a porous forcing block, deformation can extend into the forcing block, creating a wider of the zone of deformation. This suggests that forcing block porosity may control the width of the zone of deformation. 52 3. Part II: Temperature and Pore Fluid Pressure Controls2 3.1. Introduction In Part I, the results of room temperature, velocity stepping frictional sliding experiments on shale and carbonate fault gouges were presented. These experiments were designed to look at the strength, stability, velocity dependence and microstructural evolution of these gouge materials. These baseline experiments provide insight into the behavior of carbonate- shale gouges at room temperature, saturated conditions and changing displacement rates. Most fault zones at seismogenic depths are under elevated temperature and pore fluid pressure conditions. Therefore, Part II of this thesis presents the results of frictional sliding experiments designed to address the effects of elevated temperature and pore fluid pressure on shale, carbonate and shale/carbonate composite gouges. Room temperature, saturated experiments were also performed to provide a baseline with which to compare the results of experiments performed at elevated temperature and elevated pore fluid pressure. All experiments presented in Part II were conducted at the University of Liverpool (U of L). The experiments at the U of L were performed under different conditions than the experiments at UBC. For example, to simplify the loading path and reduce the number of independent variables, experiments conducted at the U of L were performed at a constant axial strain rate of 4.8 µm s-1, as opposed to the variable strain rate experiments performed at UBC. The results of elevated temperature and pore fluid pressure experiments performed at the U of L will be compared to room temperature, saturated experiments also performed at the U of L as an independent study to eliminate any uncertainty due to differences between 2
A
version
of
this
chapter
will
be
submitted
for
publication.
Haywood,
J.,
Kennedy,
L.
A.,
Faulkner,
D.
R.
Temperature‐fluid
pressure
controls
on
the
mechanical
evolution
of
shale‐carbonate
fault
gouge.

 53 experimental conditions and apparatuses. The experimental parameters and conditions used for both UBC and U of L experiments are outlined in Table 3.1. In the Discussion chapter of this thesis (section 5), the results of experiments performed at UBC will be compared to the results of experiments performed at the U of L. Field studies have demonstrated that most limestone-on-shale thrust faults develop two cataclasites: a shale cataclasite and a limestone cataclasite. Both cataclasites are foliated, and provided evidence of strain accommodation by both rock types [Burkhard, 1990; Erickson, 1994; Kennedy and Logan, 1998; Wojtal and Mitra, 1986]. The mechanisms by which these thrust faults accommodate strain depends on complex interactions between, for example, mineral composition, pore fluid pressure, permeability and temperature. The strength of thrust faults, and shallow, crustal, fault zones in general, is reflected by the coefficient of friction. In carbonates, µ ~.7 - .85 [Shimamoto, 1977; Shimamoto and Logan, 1981], whereas it ranges from ~.27 - .68 in natural chlorite and/or illite bearing shales [Ikari et al., 2009; Saffer and Marone, 2003]. The addition of shale (with abundant phyllosilicates) is therefore expected to decrease strength and promote stable sliding of faults [Kawamoto and Shimamoto, 1998; Logan and Rauenzahn, 1987]. In shale and shale bearing gouges, compaction, homogenous granular flow, brittle grain size reduction, phyllosilicate rotation and frictional sliding along phyllosilicate foliae are the dominant deformation mechanisms [Groshong, 1988; Passchier and Trouw, 2005]. On a grain scale these mechanisms are brittle, meaning that there is little crystal plastic deformation [Groshong, 1988]. 54 Table 3.1: Experimental Conditions Variable UBC Liverpool experiment type velocity stepping constant strain rate strain rate (µm s-1) 2.8 - 100 4.5 core diameter (mm) 25.4 20 height:diameter ratio 2:1 2.5:1 angle of saw cut (corrected for during processing) 35° 30° shale grain size (by volume %) 50% between 5.7 µm and 52.5 µm min = 83 µm max = 239.9 µm 50% between 8.7 µm and 120.2 µm min = 83 µm max = 478.6 µm type of jacket polyolefin copper end of sample lubricant MoS2 PTFE (teflon) avg. total displacement (mm) 4.5 - 5 2.8 - 3.5 55 In carbonates at low temperature, compaction, homogenous granular flow, brittle grain fracturing, cataclasis, and grain rolling can occur[Groshong, 1988; Passchier and Trouw, 2005]. Twinning can also occur in calcite at room T [Schmid et al., 1987; Schmid et al., 1980] resulting in the formation of thin (1-10 µm thick) twins with planar margins [Groshong, 1988; Schmid et al., 1980]. However, the maximum amount of shear strain that can be accommodated by twinning is .35, and as a result the efficacy of twinning as a strain accommodation mechanism is limited [Twiss and Moores, 1992]. At elevated temperature (> ~100°C) dislocation glide can be activated in calcite, and can work in tandem with twinning to accommodate displacement [Schmid et al., 1980], resulting in permanent deformation of the crystal. The rate of dislocation glide is limited by the ability of dislocations to climb over obstacles. Dislocation climb, activated at slightly higher T (> ~ 250°C), allows deformation to continue in a different glide plane resulting in larger intracrystal strain accommodation [Poirier, 1985]. As such, dislocation creep is a thermally activated process [Poirier, 1985]. Because dislocation creep is more temperature dependant then twinning, dislocation creep and dynamic recrystallization becomes increasingly important with increasing T [Schmid et al., 1987; Schmid et al., 1980]. Dislocation glide can also occur in phyllosilicates and calcite at room T (i.e. biotite, muscovite, illite etc. [Beeler et al., 2007; Mares and Kronenberg, 1993]), and the rate of dislocation glide is expected to scale linearly with temperature [Beeler et al., 2007; Nakatani, 2001; Rice et al., 2001]. Recovery of the crystal lattice – a process that reduces internal strain energy – can occur by the reorganization of dislocations into cell-like structures in calcite at T > ~100°C [Molli et al., 2011], which can facilitate further glide, and hence, further deformation in the crystal lattice. 56 In the presence of fluids, at elevated temperature, solution transfer is an important deformation mechanism in calcite ([Knipe, 1981; Robin, 1978; Wheeler, 1987]). The presence of phyllosilicates enhances pressure solution in calcite [Hippertt, 1994; Meike, 1990], and may actually be required to required to activate solution transfer [Bos and Spiers, 2000; Bos et al., 2000]. Solution transfer is a linear viscous process and as such can occur at lower stresses than crystal plasticity or cataclasis. As a result, increases in temperature and pore fluid pressure conditions can serve to weaken the thrust sheet. Here, results are presented from a suite of experiments designed to test the effects of elevated temperature and pore fluid pressure on the mechanical strength and stability of shale, carbonate and shale-carbonate composite fault gouges. Experiments were performed at a temperature of 150°C, a confining pressure of 85 MPa and a pore fluid pressure of 15 MPa to simulate conditions in thrust faults at the upper limit of the seismogenic zone. A suite of experiments performed at saturated, room temperature conditions, provides a baseline with which to compare my high T and Pf results. Previous experiments on carbonate and phyllosilicate gouges indicate that changing temperature and pressure conditions have a profound impact on the strength and stability of gouge materials. Carbonate gouges are unstable when deformed under a wide variety of conditions [Kawamoto and Shimamoto, 1998; Logan et al., 1992; Shimamoto, 1977; Shimamoto and Logan, 1981]. Kawamoto and Shimamoto [1998] showed that calcite can display violent stick-slip behavior at temperatures up to 700°C: in contrast, stable sliding has been observed at low confining pressures (< 50 - 100 MPa [Logan et al., 1981; Logan et al., 1992; Shimamoto and Logan, 1981]). Pure dolomite gouge also displays violent stick-slip at 57 confining pressures above 60 – 90 MPa, and stable sliding below these confining pressures [Shimamoto and Logan, 1981]. Phyllosilicates typically display stable sliding and strain hardening behavior at low confining pressure and temperature (montmorillonite [Ikari et al., 2011; Logan and Rauenzahn, 1987; Morrow et al., 1992; Tembe et al., 2010], illite [Morrow et al., 1992; Saffer and Marone, 2003; Tembe et al., 2010], muscovite [Mariani et al., 2006; Scruggs and Tullis, 1998; van Diggelen et al., 2009], chlorite [Shimamoto, 1977], kaolinite [Bos and Spiers, 2000]), however they can display stick-slip behavior at high confining pressures (illite, talc, chlorite [Byerlee and Summers, 1973]), slow sliding rates (illite, montmorillonite [Moore et al., 1986]) or moderate temperatures (muscovite, 400°C & 500°C [Mariani et al., 2006]) and strain weakening behavior at high temperatures (muscovite [van Diggelen et al., 2009]). It is clear from these experiments that the strength and stability of both carbonate and phyllosilicate bearing fault gouges are strongly affected by changes in temperature, confining pressure and sliding velocity. It is unclear how changes in these conditions affect carbonate and phyllosilicate bearing composites. Although the experiments in the present study are designed to address the role of composition, temperature and pore fluid pressure on shallow (~3-6 km depth) thrust faults, many other types of shallow continental fault zones (e.g. San Andreas Fault, USA; Carboneras Fault, Spain, etc.) are rich in phyllosilicate-rich material and carbonate. Consequently, the results of this study should be applicable not only to shallow foreland thrust faults, but also to other shallow continental faults containing carbonate and phyllosilicates. In addition, these experiments will provide more realistic coefficient of friction values that can be used in numerical fold and thrust belt simulations in which the 58 geometry and deformation style of fold and thrust sheets depends on the basal coefficient of friction. 3.2. Methodology 3.2.1. Sample Description and Preparation Gouge material was created from quartz-bearing phyllosilicate-rich shale (XRD analysis using the Rietveld Method [Rietveld, 1967; 1969] indicates: 32% quartz, 40% illite, 17% clinochlore, 10% feldspar (Appendix B)) combined in various volumetric proportions with reagent grade calcite powder (80% calcite, 20% dolomite) from Ward’s Scientific. To create the shale gouge, a natural quartz-bearing phyllosilicate-rich shale was powdered using a ring mill to pass through a 147 µm sieve. Any carbonate present in the original rock was dissolved using 1M acetic acid. XRD analysis of the shale gouge before and after carbonate dissolution ensured complete removal of any pre-existing calcite and dolomite, and confirmed that acetic acid dissolution has no effect on the phyllosilicates. Laser grain size analysis using a Mastersizer 2000 at the University of British Columbia shows that 98% of the carbonate grains are less than or equal to 4.4 µm diameter. In the shale gouge, 98% of grains are less than or equal to 5.0 µm diameter. However, the shale gouge contains coarse grained quartz grains which make up a small proportion of the gouge by number percent, but comprise a significant proportion by volume percent. 50% of the shale by volume is less than or equal to 34.7 µm in diameter, while 95% of the shale by volume is less than or equal to 239.9 µm in diameter. In comparison, 50% of the carbonate is less than or equal to 5.0 µm in diameter, and 95% is less than or equal to 17.4 µm in diameter (Appendix A). The fine grain size of the carbonate was chosen to simulate the grain size found in micrites, which commonly form the hanging wall in foreland thrust systems. 59 3.2.2. Experimental Procedure For each experiment, a 1 mm thick water saturated gouge layer was spread along a polished 30° angle saw cut in a 20 mm diameter by 50 mm length right-cylinder core (Figure 3.1a and b). Gouge was mixed with distilled water and spread along the saw cut using a specially machined vise that held the core at a 30° angle (Figure 3.1a). Calipers and a razor blade were used to ensure the gouge layer was exactly 1 mm thick. Experiments were performed on each end member composition as well as 75%, 50% and 25% mixtures of shale and carbonate. Experiments were conducted at 70 MPa effective confining pressure and a displacement rate of 4.5 µm/s. Baseline, room temperature experiments on water saturated gouge were conducted at 70 MPa confining pressure. 150°C experiments were conducted with a pore fluid pressure of 15 MPa and confining pressure of 85MPa, to give an effective confining pressure of 70 MPa. Porous Berea Sandstone (ϕ ~ 17%) comprised the upper forcing block while impermeable Badshot Dolomite marble comprised the lower forcing block. An additional suite of experiments was conducted at 150°C with a pore fluid pressure of 15 MPa using impermeable Badshot Dolomite for both the lower and upper forcing block material. A 1 mm diameter hole was drilled in the middle of the upper Badshot Dolomite forcing block to ensure good communication between the pore fluid system and the gouge zone. Pore fluid pressure was increased one hour prior to the start of the experiment to allow the pore fluid pressure in the gouge zone to completely equilibrate [Smith and Faulkner, 2010]. 60 Figure 3.1: Sample Assembly b ca Figure 3.1: Frictional sliding experiments were conducted on 20 mm by 50 mm cores containing a polished, 30° angle saw cut. a. 1 mm of distilled water saturated gouge material was spread on the saw cut using a vise and a razor blade to ensure gouge thickness was the same for each experiment. b. Each sample was then wrapped in a copper jacket to separate the silicon oil confining medium from the pore fluid pressure system. Thin wafers of PTFE were used to reduce friction effects at the ends of the samples. c. Experimental run product. Blue and white squares are 1 cm x 1 cm. 61 Triaxial frictional sliding experiments were conducted on shale, carbonate and carbonate shale composite gouges using the High Pressure High Temperature Triaxial Deformation Apparatus with Fluid Flow in the Rock Deformation Laboratory, Department of Earth and Ocean Sciences, University of Liverpool. This apparatus is a motor and gear driven triaxial rock deformation press (Figure 3.2). It is capable of performing triaxial experiments up to 250°C and 250 MPa confining pressure, and maximum displacement rates of 20 µm s-1. Force is measured using an external force gauge LVDT, and displacement is measured by motor revolutions per second. For a detailed description see Mitchell [2007] and Mitchell and Faulkner [2008]. The axial loading, confining pressure system and pore fluid system are servo-controlled. Silicon oil is used as a confining medium, and distilled water is used as the pore fluid. Temperature is provided by an external Chromalox 1000W/240V ceramic knuckle band heater (CKH). Thermocouples located on the furnace itself and inside the pressure vessel at the top of the sample ensure accurate measurement of the temperature of the sample throughout deformation. Each sample was wrapped in an inner jacket of copper tubing, then surrounded by a PVC jacket that seals the pore fluid system and separates the pore fluid system from silicon oil confining medium. Porous steel spacers placed between the pistons and the sample allow fluid flow but limit the amount of gouge that can extrude into the pore fluid system. Thin wafers of polytetrafluoroethylene (PTFE) were used to reduce friction between the sample and the spacers. 62 Figure 3.2 (following page): The High Pressure High Temperature Triaxial Deformation Apparatus with Fluid Flow at the University of Liverpool is a motor and gear driven triaxial rock deformation apparatus. It can accommodate cylindrical samples up to 20 mm in diameter and 50 mm in length. A PVC jacket seals the pore fluid system and separates from pore fluid system from the silicon oil confining medium. Porous steel spacers at the top and bottom of the sample ensure that pore fluid can be added to the sample, and limit the amount of gouge that can extrude into the pore fluid pressure system. Both the pore fluid pressure system and the confining pressure system are servo-controlled (diagram courtesy of Tom Mitchell). 63 Figure 3.2: Rig Diagram (from [Mitchell, 2007]) 64 Samples were oven dried for 24 hours at 80°C, and the pore fluid system was vacuumed prior to the experiment to ensure no moisture was present. For experiments conducted at elevated temperature and pore fluid pressure, the confining pressure was first increased to 70 MPa. The pore fluid pressure was then increased and the system monitored to ensure a good seal before confining pressure was increased to 85 MPa. This sequence ensured that the 70 MPa was the maximum effective confining pressure ever felt by the sample. Finally the temperature was increased to an internal temperature of 150°C, while the servo-controlled pore fluid and confining pressure pumps worked to offset increases in pressure due to heating induced volume changes in these fluids. Samples were displaced 2.8- 3.67 mm (Figure 3.1c). This axial displacement translates to 3.23-4.24 mm of displacement along saw cut surface. The shear strain across the 1 mm thick gouge zone is 3.23-4.24 – assuming no thickening or thinning during sliding – and the shear strain rate along the surface is 5.5 x10-3 s-1. All mechanical data was recorded using a program written in National Instruments’ LabVIEW™ 7.2 and the coefficient of friction (µ) was extracted from this data using a program written for MathWorks’ MATLAB v. 7.4 (Appendix D). The coefficient of friction was calculated using the ratio of shear stress to normal stress resolved on the saw cut, and corrected for the change in contact area during sliding (Appendix D). 3.2.3. Petrographic Observations Microstructural analysis was performed on experimental run products to elucidate the effects of temperature, pore fluid pressure, composition and forcing block porosity on the deformation microstructure of gouge. Polished thin sections were cut perpendicular to the saw cut and parallel to the direction of maximum displacement (i.e. the kinematic plane). 65 From these observations, the active deformation mechanisms that accommodated displacement are inferred. Microstructure analysis was conducted using both standard optical microscopy and a Philips XL30 scanning electron microscope (SEM) at the University of British Columbia. Both backscatter electron imaging and element mapping were used to determine the distribution of mineral phases. 3.3. Results The mechanical data for all composite mixtures are presented in Table 3.2, and in Figure 3.3 and Figure 3.4. In effect, the composite mixtures of shale and carbonate represent a ternary mixture: the shale is composed mostly of phyllosilicate and quartz. Thus strain can be partitioned into the carbonates, the phyllosilicates or the quartz grains, which are typically larger than the other minerals. 3.3.1. Frictional Strength and Sliding Behavior of Gouge Mixtures All experiments begin with gouge compaction, represented by a sharp increase in the coefficient of friction with little displacement, during which brittle grain size reduction and increased packing intensity of the grains reduce porosity [Marone et al., 1990]. Yielding occurs between .5 and 1 mm displacement, after which significant displacement is recorded without a significant increase the coefficient of friction. In all experimental suites, shale gouge displays the lowest yield strength while the carbonate gouge displays the highest yield strength. In the composite gouges, yield strength of the gouge increases with increasing carbonate content (Figure 3.3a, b and c). The ultimate strength – taken at 2.8 mm displacement, does not linearly increase with carbonate content. Rather there is a moderate increase in strength when carbonate content is increased to up to 75%, and an abrupt increase in strength for the 100% carbonate endmember (Figure 3.5), 66 Table 3.2: Experimental Data Sample Composition Forcing Blocks T ( °C) Pore Fluid Conditions Pc (Mpa) Displacement Rate (µ s-1) Max Displ. (mm) Peak σ1 - σ3 (MPa) Peak Coef. of Friction Yield Strength Coef. Of Friction at Maximum Shear Strain Strain Hardening Rate (from .75 -2.0 mm) 100% Shale sst/dol room saturated 70 4.8 3.53 140 0.66 0.5 0.66 0.0511 75% Shale sst/dol room saturated 70 4.8 3.67 155 0.71 0.55 0.71 0.0418 50% Shale sst/dol room saturated 70 4.8 3.67 161 0.71 0.59 0.71 0.0307 25% Shale sst/dol room saturated 70 4.8 3.53 176 0.75 0.65 0.75 0.0133 100% Carbonate sst/dol room saturated 70 4.8 3.03 244 0.95 0.84 0.95 0.0915 100% Shale sst/dol 150°C 15 MPa 85 4.8 3.67 142 0.66 0.5 0.66 0.0436 75% Shale sst/dol 150°C 15 MPa 85 4.8 3.22 156 0.69 0.52 0.69 0.0525 50% Shale sst/dol 150°C 15 MPa 85 4.8 3.75 174 0.75 0.6 0.75 0.0447 25% Shale sst/dol 150°C 15 MPa 85 4.8 2.8 190 0.78 0.67 0.78 0.0478* 100% Carbonate sst/dol 150°C 15 MPa 85 4.8 3.26 274 0.83 0.8 0.71 Not calculated 100% Shale dol/dol 150°C 15 MPa 85 4.8 3.48 147 0.67 0.43 0.67 0.087 75% Shale dol/dol 150°C 15 MPa 85 4.8 3.62 173 0.73 0.52 0.73 0.0655 50% Shale dol/dol 150°C 15 MPa 85 4.8 3.46 192 0.78 0.52 0.78 0.0758 25% Shale dol/dol 150°C 15 MPa 85 4.8 3.31 193 0.78 0.61 0.78 0.0441 100% Carbonate dol/dol 150°C 15 MPa 85 4.8 3.5 221 0.79 0.73 0.78 Not calculated 100% Carbonate dol/dol 150°C dry 70 4.8 3.95 234 0.81 0.79 0.65 -0.0319 *
calculated
for
1.25
mm
following
yielding
 67 Figure 3.3 (following page): Experimental data showing displacement vs. coefficient of friction (µ). All experiments were performed at displacement rates of 5 µm s-1. a. Data from room temperature saturated experiments performed with one sandstone and one dolomite forcing block. At these conditions, the 100% shale gouge is the weakest while the 100% carbonate is the strongest. The composites are intermediate in strength. All gouges show stable sliding and slight strain hardening at maximum displacements. At 150°C with 15 MPa pore fluid pressure, yield strength increases with increasing carbonate content b. Data from experiments performed with one dolomite and one sandstone forcing block at 150°C with 15 MPa pore fluid pressure. All gouges containing phyllosilicates display stable sliding, and strain harden throughout the duration of the experiment. Following the peak stress, the 100% carbonate displays pronounced strain weakening followed by the evolution to stick-slip sliding at regularly spaced intervals. At maximum shear strain, the 100% carbonate gouges is weaker than both the 25% and 50% shale gouges. c. Data from experiments performed with 2 dolomite forcing blocks at 150°C with 15 MPa pore fluid pressure display similar behavior to experiments with one sandstone forcing block and one dolomite forcing block deformed under the same conditions. Yield strength increases with increasing carbonate content, and the 100% carbonate displays strain weakening and the evolution to stick-slip sliding following the peak strength. However, all gouges deformed with two dolomite forcing blocks display more rapid strain hardening, and at maximum shear strain display a higher coefficient of friction than their counterparts deformed with 1 sandstone and one dolomite forcing block. In addition, carbonate displays less pronounced strain weakening and less regular stick-slip sliding following the peak strength. 68 Figure 3.3: Mechanical Data 0 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Co ef fic ie nt o f F ric tio n 0 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 100% Shale, room T 75% Shale, room T 50% Shale, room T 25% Shale, room T 100% Carb., room T Co ef fic ie nt o f F ric tio n a b c 100% Shale, 150°C 75% Shale, 150°C 50% Shale, 150°C 25% Shale, 150°C 100% Carb., 150°C 100% Shale, 150°C 75% Shale, 150°C 50% Shale, 150°C 25% Shale, 150°C 100% Carb., 150°C Co ef fic ie nt o f F ric tio n Displacement (mm) 69 Figure 3.4: Mechanical Data II 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Displacement (mm) 0 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Displacement (mm) 100% Carbonate, 150C,dry 100% Carbonate, 150C dol, 15 MPa Pf 100% Carbonate, 150C, 15 MPa Pf 100% Carbonate, room T, sat 0 1 2 3 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 P P R1 R1 Y Y Co ef fic ie nt o f F ric tio n 100% Shale, room T, sat 100% Shale, 150C, 15 MPa Pf 100% Shale, 150C dol, 15 MPa Pf 75% Shale, room T, sat 75% Shale, 150C, 15 MPa Pf 75% Shale, 150C dol, 15 MPa Pf a b e c d 25% Shale, room T, sat 25% Shale, 150C, 15 MPa Pf 25% Shale, 150C dol, 15 MPa Pf 50% Shale, room T, sat 50% Shale, 150C, 15 MPa Pf 50% Shale, 150C dol, 15 MPa Pf Co ef fic ie nt o f F ric tio n Co ef fic ie nt o f F ric tio n 0 0 Figure 3.4: Displacement vs. coefficient of friction (µ) graphs displaying the effects of temperature, pore fluid pressure and forcing block variations on the mechanical strength of each gouge composition. 70 indicating that the addition of shale significantly decreases the strength of the carbonate gouge. Reasons for this increase will be discussed in sections 3.3.3 and 3.4.1. All gouges that contain shale exhibit stable sliding (Figure 3.3). Only the 100% carbonate gouge exhibited stick slip. In the 100% carbonate gouge at increased temperature (T) and pore fluid pressure (Pf), significant strain weakening and the evolution to regular stick-slip sliding follow the yield strength. The strength after 3.2 mm displacement in the 100% carbonate gouge at elevated T and Pf is less than the strength of the 25% and 50% shale composite gouges. Steady state sliding only occurred in 100% carbonate gouges, and in 3 composite gouges (100% shale, 75% and 50% shale) deformed at elevated temperature and pore fluid pressure with 2 dolomite forcing blocks. All other gouges continued to strain harden at maximum shear strain. 3.3.2. Effects of Pf and T on Mechanical Strength Increased temperature and increased pore fluid pressure do not have a significant effect on the strength of the 100% shale gouge. The stress-strain curve for the 100% shale gouge at 150°C with 15 MPa pore fluid pressure performed with one sandstone and one dolomite forcing block parallels the stress-strain curve for the same gouge composition at room T, saturated conditions (Figure 3.4a). At maximum shear strain all the 100% shale samples display the same strength (µ ~ .66). With increasing carbonate content, the effects of increases in temperature and pore fluid pressure become more pronounced. Increased temperature and pore fluid pressure increases the ultimate strength of 50% and 25% shale gouge compositions (50% and 75% carbonate) by µ ~ .04 relative to room T gouges of the same compositions (Figure 3.5). Increased temperature and pore fluid pressure conditions most profoundly affect the 100% 71 Figure 3.5: Coefficient of Friction vs. Carbonate Content Room TRoom T 150 C, Pf 150 C Pf, dol Percent Carbonate Co ef fic ie nt o f F ric tio n 1009080706050403020100 .65 .7 .75 .8 .85 .9 .95 1 .6 Room T 150°C, Pf 150°C, dol Figure 3.5: Coefficient of friction vs. carbonate content. The coefficient of friction – taken at 2.8 mm displacement does not increase linearly with carbonate content. At room T (blue), the strength shale gouge gradually increases as carbonate content is increased up to 75%, followed by an abrupt increase in strength for the 100% carbonate endmember. At elevated T and Pf, the absolute strength of the composite gouges is higher than the room T saturated strength (red and green). Elevated T and Pf significantly weaken the carbonate gouge at maximum shear strain. 72 carbonate gouge. While the 100% carbonate gouge displays stable sliding at room T, saturated conditions, both the 100% carbonate samples deformed at elevated temperature and elevated pore fluid pressure display pronounced strain weakening and stick-slip sliding following the peak strength. In the sample deformed with one sandstone and one dolomite forcing block, these stick-slip events are regularly spaced and all have a stress drop of about 10 MPa (.02 decrease in the coefficient of friction). Forcing block composition affects the pore fluid conditions inside the gouge zone. An increase in pore fluid pressure results in a decrease in the effective confining pressure (Pc) according to the relationship: Pe = Pc – Pf, in which Pc is the applied confining pressure (Figure 3.6). The porous sandstone forcing block allows pore fluid to escape into the forcing block, and may lead to transient changes in local pore fluid pressure during sliding. Thus, the Pf in the gouge zone is transiently lower than the Pf in the hanging wall, as well as that measured by the external pore fluid pressure transducer. The effective confining pressure in the sandstone forcing block is lower than the effective confining pressure in the gouge zone, increasing the strength of the gouge zone relative to the sandstone forcing block. When 2 non-porous dolomite forcing blocks are used, pore fluid is trapped inside the gouge zone, which may lead to elevated pore fluid pressure (i.e. decreased effective confining pressure) in the gouge zone and has no effect on the effective confining pressure in the impermeable forcing block. In this case, the strength of the gouge zone relative to the impermeable forcing block will be reduced. Experiments in the present study do not show evidence of weakening due to transient pore fluid pressures in the gouge zone. Rather, experiments performed at elevated T and Pf are typically stronger than experiments performed at room T, saturated conditions. 73 Figure 3.6: Effects of Pf on Mohr-Coulomb Failure τ σn Pf φ σ1σ1e σ2e = Pe σ2 = Pc Pe = Pc - Pf Mo hr- Cou lom b F ailu re E nve lop e Figure 3.6: Mohr-Coulomb failure diagram showing the importance of pore fluid pressure on rock strength. An increase in pore fluid pressure results in a decrease in the effective confining pressure (Pc) according to the relationship: Pe = Pc – Pf. The addition of a pore fluid pressure shifts the circle towards the left, and closer to failure. All experiments in this study were performed at an effective confining pressure of 70 MPa, so differences in effective confining pressure during experiments are strain induced. Φ = angle of internal friction. 74 All shale bearing gouges deformed at elevated temperature with two non-porous forcing blocks show an increase in the strain-hardening rate relative to the room T gouges. Strain hardening rate is defined as Δµ/(Δγ) in which shear strain, γ, is defined by the axial displacement resolved on the saw cut divided by the thickness of the gouge zone (in this case 1 mm)[Tembe et al., 2010]. Figure 3.7 shows strain-hardening rate as a function of gouge composition. At room T, there is a decrease in the strain-hardening rate with increasing carbonate content in all phyllosilicate-bearing gouges, with a sharp increase in strain hardening rate for the 100% carbonate gouge. The observed decrease in strain hardening rate is the function of strain weakening following the yield strength (Figure 3.3a). For samples deformed at elevated T and Pf conditions, there is no systematic variation in strain hardening rate with composition, however shale-bearing samples deformed with 2 dolomite forcing blocks typically display the highest strain hardening rates. Phyllosilicate-rich experiments (shale ≥ 50%) performed at elevated temperature and elevated pore fluid pressure display significant strain hardening prior to the evolution to steady-state sliding (Figure 3.4). Potentially, there is more quartz cataclasis at the dolomite margin than at the sandstone margin and this may cause strain hardening [Biegel et al., 1989]. More likely, however, the increased strength and rate of strain hardening in these samples is due to dilatancy hardening. Dilatancy hardening occurs when the gouge is deformed above a certain critical strain rate. The sample dilates, creating porosity, but the permeability is insufficient to allow fluid to flow into the new pore space creating a transiently lower fluid pressure in the gouge zone; the net effect is an increase in sample strength. Dilatancy hardening is well documented in low permeability rocks and soils at elevated 75 Figure 3.7: Strain Hardening Rate Percent Shale 0102030405060708090100 .01 .02 .07 .06 .05 .04 .03 .08 St ra in H ar de ni ng R at e (∆ µ/ ∆γ fo r d is pl = .7 5 - 2 m m ) Room T 150°C, Pf 150°C, dol .09 .10 Figure 3.7: At room T, the strain hardening rate decreases with increased carbonate content to a minimum at 75% carbonate, with an abrupt increase at 100% carbonate (blue). The decrease in strain hardening rate is a function of the strain weakening that occurs after yielding. There is no systematic variation in strain hardening rate with carbonate content at elevated T and Pf. However, samples deformed at elevated T and Pf with 2 dolomite forcing blocks are typically stronger than samples deformed with 2 non-porous forcing blocks. 76 fluid pressures [Brace and Martin, 1968; Rice, 1975; Samuelson, 2010; Verruijt and van Baars, 2007]. Marone et al. [1990] indicate that for fine grained quartz gouge at 190 MPa confining pressure .5 µm s-1 is the critical sliding speed above which gouge is not permeable enough to conduct drained experiments (i.e. experiments in which there is no change in pore fluid pressure due to loading). The 4.5 µm s-1 sliding velocities in this study are likely high enough to promote undrained conditions and dilatancy hardening in both the carbonate and shale gouge deformed with 2 non-porous forcing blocks. Although dilatancy hardening is most likely to occur in low permeability gouges with a high pore fluid pressure and 2 non-porous forcing blocks, it is possible that dilatancy hardening also occurs in the carbonate-rich samples at elevated temperature and pore fluid pressure with one porous forcing block and one non-porous forcing block. Mechanical data for experiments conducted with one porous and one non-porous forcing block show an increase in strength relative to room temperature values for high carbonate content samples (50% and 75% carbonate, Figure 3.5). The strength of both 75% carbonate samples at elevated temperatures and pore fluid pressures are the same, regardless of forcing block porosity. Since dilatancy hardening occurs if the permeability of the gouge is insufficient to allow fluid to fill porosity as it is created, the increase in strength of the carbonate-rich composite samples (50% & 75% carbonate) relative to the saturated samples of the same composition suggests that the carbonate gouge has a lower permeability than the shale gouge. 77 The 100% carbonate sample with two dolomite forcing blocks displays a lower peak strength, but a higher residual strength (i.e. post strain weakening strength) than the sample deformed with one sandstone forcing block (Figure 3.4e). In addition, the magnitude and frequency of the stress drops during stick-slip sliding is more irregular than those observed using a porous forcing block. These stress drops have an average magnitude of ~ 5 MPa, compared with the ~10 MPa magnitude of the stress drops observed in the sample with one porous forcing block, suggesting that pore fluid pressure conditions affect the character and magnitude of stick-slip sliding in these samples. A 100% carbonate gouge, vacuum dried and deformed at 150°C, displays strain weakening after peak strength, but does not display stick- slip sliding. The 100% carbonate weakens throughout the duration of the experiment, and at maximum shear strain displays the lowest coefficient of friction recorded for any gouge composition in this study. 3.3.3. Description of Gouge Textures The mechanical data presented in the previous section show significant strength variations with temperature, pore fluid pressure and forcing block changes, but these factors do not have a major influence on the microstructure developed in each sample. Compositional variation provides the most important control on microstructure development on the gouges in this study. In the following section, microstructures are documented for the various gouge compositions. Fabric geometry is described in context of the Riedel shear zone terminology outlined and used by Skempton [1966], Bartlett et al. [1981], Logan et al. [1992], and Rutter [1986] (Figure 2.5). 78 All samples show evidence of grain size reduction of all mineral phases at the forcing block gouge margin, and samples containing shale show phyllosilicate alignment at the gouge zone margin. These textures indicate strain localization at the shear zone boundary. Samples with high shale content display more distributed deformation than samples with higher carbonate content. In the 100% shale gouge, two different styles of strain accommodation are observed (Figure 3.8a and b): 1) quartz shows evidence of significant brittle grain size reduction. Quartz clasts typically have sigmoidal ‘tails’ of comminuted quartz grains, providing food kinematic indicators (Figure 3.8d). In addition, phyllosilicates are rotated into the P orientation (Figure 3.8a and b). Bands of grain size reduced quartz and carbonate, occur parallel to inclined phyllosilicate and quartz clasts and contribute to the P fabric (also known as the shape fabric). The above structures are discontinuous or can extend across the gouge zone.; 2) strain is homogenously distributed across the entire thickness of the gouge zone and phyllosilicates display a preferred orientation, defining a shape (P) fabric. Quartz grains in these zones are larger than those observed in sections of the gouge with more localized deformation (Figure 3.8b). Poorly developed R1 shears are visible in both domains, however they do not cross the entire width of the gouge zone (Figure 3.8a). In the 100% shale gouge, the coarse quartz grain size and the platy geometry of the phyllosilicates inhibit strain localization, and promote homogenous flow [Logan, 2007]. Phyllosilicates across the gouge zone are dominantly oriented parallel to the P direction, providing evidence of homogenous deformation. 79 Figure 3.8 (following page): Microstructure of 100% shale gouge run products. a. Two styles of strain localization visible in the 100% shale gouge. An R1 shear (yellow) cuts across part of the gouge zone (cross polarized optical microscope). b. White lines separate zones of more intense deformation with a well-developed shape fabric from quartz-rich zones where the shape fabric is less well defined (cross polarized optical microscope with gypsum plate). c. SEM element map shows bands of comminuted quartz (dotted green line), phyllosilicates and carbonate form compositional banding in the 75% shale gouge. White dotted lines highlight bands of more intense brittle grain size reduction in quartz (blue = Ca). d. Carbonate matrix grains wrap around quartz clasts with asymmetric tails of comminuted quartz. Phyllosilicates form discontinuous margin parallel Y shears (dotted black) (SEM element map: red = Al, blue = Ca, green = Si). e. Lenses of phyllosilicates define the shape fabric, a discontinuous margin parallel Y- fabric and R1 shear bands (SEM element map). f. R1 shears are prominent in the 100% carbonate gouge (cross polarized optical microscope). 80 Figure 3.8: Microstructure 30 μm dol qtz 75% Shale c d50% Shale dol P Y sst 90 μm e25% Shale 100 μm R1 P 200 μm a100% Shale dol sst b100% Shale sst dol P Pqtz f100% Carbonate dol sst Boundary Shear R1 R1 R1 P P Y 200 μm 200 μm qtz 81 With increasing carbonate content, the relative importance of the shape (P) fabric decreases and R1 shears become more prominent (Figure 3.8e and f). Brittle cataclasis of isolated quartz clasts and subsequent alignment of fine comminuted quartz grains define layers (Figure 3.8c). The quartz layers are segregated from the phyllosilicates and carbonate forming a discontinuous and undulatory compositional layering (Figure 3.8d). Asymmetrical phyllosilicate and quartz-rich lenses define the shape fabric in composite gouges (Figure 3.8d and e). Fragments of quartz form tails on larger quartz grains, and provide good kinematic indicators. Comminuted quartz grains and phyllosilicates can occur in layers to form margin-parallel bands (i.e. Y slip surfaces). Both isolated quartz grains and quartz grains in contact with other quartz grains show evidence of cataclasis. Typically, at quartz grain-grain contacts, Hertzian-style fracturing (tensile fractures formed at high stress contacts) promotes fragmentation of grains. Comminuted carbonate appears to wrap around the quartz grains that define the shape fabric and at elevated temperature may show evidence of ductility (Figure 3.8e). R1 shears offset the shape fabric and margin parallel quartz bands suggesting that they were active in the late stages of deformation (Figure 3.8e). The larger R1 shears dip moderately in the middle and flatten towards the margin. These R1 shears are defined either by aligned phyllosilicate grains, or as bands of more intense grain size reduction in quartz, carbonate or phyllosilicates. R1 shears can extend through the entire gouge zone thickness, or only part of it. Although they are present in all composite gouge compositions, they are most visible in the 25% shale gouge. The textures observed in composite gouges closely resemble those observed in experiments conducted by Tembe et al. [2010] on quartz, illite and montmorillonite mixtures. 82 These textures most closely resemble those in Regime 2 and Regime 3, defined by Tembe et al. [2010], in which they see grain size reduction and phyllosilicate entrapment in R1 and boundary shears, while a shape (P) fabric and Y shears link R1 shears. In Regime 2, the coarse grain size of the quartz inhibits strain localization and the development of full thickness R1 shears because the R1 shears propagation stops when the R1 shear intersects a large quartz grain (Figure 3.9). Increased phyllosilicate content and decreased quartz content and allows full thickness R1 shears to develop. In the present study, strain localization increases with the addition of carbonate. Grain size, however, decreases with increasing carbonate content as carbonate replaces the coarse grained quartz in the mixture. The fine-grained carbonate does not block the formation of R1 shears, and as such the fine grain size of the carbonate promotes strain localization in the composite gouges. This observation is in agreement with previous studies on fine grained monomineralic quartz gouge [Gu and Wong, 1994; Logan and Rauenzahn, 1987] that illustrate that while larger grain sizes promote compaction and brittle grain size reduction, smaller grain sizes promote strain localization and strain weakening or steady-state sliding. In the 100% carbonate gouge, visible deformation is primarily restricted to R1 shear bands (Figure 3.8f) and brittle grain size reduction at the shear zone margin (Figure 3.10). R1 shears can extend across the entire thickness of the gouge zone, and are visible in the SEM as zones of brittle grain size reduction (Figure 3.10). In the absence of phyllosilicate and quartz phases, a shape (P) fabric does not form. 83 Figure 3.9 (following page): Schematic of microstructure changes with increasing clay content (from [Tembe et al., 2010]). In Regime 1, zones of large quartz grains and aggregates concentrate near the gouge zone margin, and are separated from zones of more intense grain size reduction by margin parallel Y-shears. In Regime 2 quartz grains concentrate in the middle of the of the gouge zone. R1 shears and a shape fabric are well defined. In Regime 3 clay minerals are ‘pervasively aligned’, and there is little visible evidence of quartz grain size reduction. 84 Figure 3.9: Regimes Increasing clay content Regime 1 Regime 2 Regime 3 quartz grains R R Y Y P P R quartz grain boundary shear X [Tembe et al., 2010] 85 Figure 3.10: Carbonate grain size reduction at the margin grain size reduction in R1 dol dol cal cal cal R1 100% Carbonate 30 µm Figure 3.10: SEM BSE image of 100% carbonate gouge showing grain size reduction in R1 shears and near the gouge zone margin. Darker grains are dolomite and lighter grains are calcite. 86 3.3.1. Orientation of Riedel Shear Localization R1 shears are most well developed in carbonate bearing gouges, however, they form in all gouge compositions in this study. Gu and Wong [1994] suggest that for a cohesionless material, the angle (α) between the R1 shear and the gouge margin is related to the angle of internal friction (µi) by the following equation: 
 Equation 3.1
 While the coefficient of friction (µ) records the macroscopic strength of the entire sample (gouge zone plus forcing blocks), the coefficient of internal friction (µi) records the strength of the gouge itself. Figure 3.11 plots coefficient of friction versus the riedel shear angle. Lines of constant internal friction are plotted for reference. The coefficient of friction at the maximum displacement was used to plot room temperature gouges from this study, as the microstructure preserved in the gouge zone records the final stages of deformation. Figure 3.11 shows µi is greater than .7 for all gouge compositions at room temperature. The coefficient of friction recorded in the 100% carbonate gouge (.94) is too high to put on this graph, however the corresponding µi value would also be high. Although all compositions display relatively high µi, the values calculated from composite gouges are slightly below those calculated for endmember gouges. The 50% and 100% shale mixtures have the lowest calculated µi. 87 Figure 3.11: Riedel Shear Localization Angle (after Tembe et al. 2010) 40º 0º 10º 20º 30º Ri ed el A ng le ( α) Coe!cient of Friction (μ) 0.2 0.8 0.7 0.6 0.5 0.4 0.3 µi = 0.4 0.6 0.8 1.0 1.2 100 SH 75 50 25 this study, Room T Regime 2 mont/ill and mont/ill/qtz from [Tembe et al. 2010] illite/qtz [Tembe et al. 2010] Regime 1 mont/ill and mont/ill/qtz from [Tembe et al. 2010] qtz [Gu and Wong,1994] Figure 3.11: Riedel shear angle vs. coefficient of friction (after [Tembe et al., 2010 and Gu and Wong, 2004]). Lines show the internal coefficient of friction calculated using Equation 3.1. Results from this study are plotted in blue squares. Numbers next to the square indicate gouge composition in percent shale. Results from previous studies on quartz and saturated binary and ternary mixtures of illite/montmorillonite and quartz are plotted in black and white. Results from the present study show that the composite gouges have the lowest internal friction angle while the endmember gouges have the highest internal friction angle. The calculated internal friction angles of all phyllosilicate bearing gouges in this study are close to the values reported for pure quartz and quartz-rich mixtures. The coefficient of friction recorded in the 100% carbonate gouge is too high to plot on this graph, however the corresponding μi would also be high. 88 3.4. Discussion 3.4.1. Effect of Increasing Carbonate Content on Frictional Strength Previous studies of carbonate gouges record both stable sliding [Logan et al., 1992] and stick-slip behavior of carbonate gouge (calcite: [JM Logan et al., 1992; Shimamoto, 1977; Shimamoto and Logan, 1981]; dolomite [Shimamoto, 1977; Shimamoto and Logan, 1981] at comparable confining pressures and room temperature humidity (i.e. nominally dry); my saturated experiments displayed stable sliding. Because the experiments in the present study were performed at saturated conditions, it is possible that increased adsorbed water lubricated the particle surfaces [Morrow et al., 2000] and inhibited the onset of stick-slip sliding. Alternately, the shear strain attained in the experiments was insufficient to complete the evolution to stick-slip sliding. Stick-slip sliding will be discussed in more detail in section 3.4.2. A recent experimental study conducted on a wide variety of gouge compositions Ikari et al. [2011], concluded that: 1) most phyllosilicate gouges display stable sliding or conditional stable sliding, and that 2) the absolute frictional strength of gouge materials correlates with the potential for the gouge to display unstable sliding. That is, gouges with a low coefficient of friction (e.g. sheet silicates like talc, smectites etc. where µ is typically less than .5) always display stable sliding, while gouges with higher friction coefficients display both stable and unstable sliding behavior (e.g. quartz, kaolinite etc where µ is often greater than or equal to .5). Their observations provide some indication of gouge stability based on the absolute 89 frictional strength of a gouge. Although useful for predicting the stability of monomineralic gouges, these observations cannot be easily transferred to studies of composite gouges. At room temperature and saturated pore fluid pressure conditions, all gouge materials exhibit stable sliding (Figure 3.3a). Previous experiments on phyllosilicate gouges show stable sliding, or conditionally stable sliding at similar experimental conditions (montmorillonite [Ikari et al., 2011; JM Logan and Rauenzahn, 1987; Morrow et al., 1992; Tembe et al., 2010], illite [Morrow et al., 1992; Saffer and Marone, 2003; Tembe et al., 2010], muscovite [Mariani et al., 2006; Scruggs and Tullis, 1998; van Diggelen et al., 2009], chlorite [Shimamoto, 1977], kaolinite [Bos and Spiers, 2000]). In the present study, all gouges containing shale (i.e. chlorite and illite) display stable sliding at all conditions. The results of the present study are in agreement with the results of Kawamoto and Shimamoto [1998] and Logan and Rauenzahn [1987] in which they observe that very small amounts (< 5%) of a stable mineral (i.e. montmorillonite, halite) can be enough to stabilize and weaken a gouge. Most previous experiments on gouge mixtures analyze bimineralic gouges in which endmembers display significant contrasts in mechanical strength and/or stability. Previous experiments on bimineralic gouge mixtures typically show an increase in yield strength with increasing proportions of the strong endmember (quartz/montmorillonite [Logan and Rauenzahn, 1987], halite/calcite [Kawamoto and Shimamoto, 1998]). Previous studies have shown small amounts of a weak mineral forming a continuous layer can reduce the frictional strength of an entire sample to the frictional strength of the weak endmember (e.g. 5% talc [Shimamoto, 1977], 5% halite [Kawamoto and Shimamoto, 1998]; 6-15% talc [Collettini et 90 al., 2009c]; 10-15% chlorite [Jefferies et al., 2006]). In my study, the 100% shale endmember gouge displays the lowest yield strength while the 100% carbonate gouge displays the highest. The composite gouges are intermediate in strength, and their yield strength increases with increasing carbonate content. In Kawamoto and Shimamoto’s [1998] experiments on halite/calcite composite gouges, the yield strengths of the composites increased with increasing calcite. However, at high shear strains (displacement ~ 20mm) the coefficient of friction for all composites merged, and the strength of the composites was reduced to near the strength of the weaker halite endmember. At the high shear strains, through going layers of halite were developed, which resulted in a dramatic reduction in the residual strength of these samples. In the present study, it was not possible to reach high shear strains due to the experimental setup. If high strains were possible, all composites may reach the strength of the weakest component: illite - if additional strain causes through going bands of phyllosilicates to form. Because the shale gouge is a composite gouge itself, it is likely that with additional shear strain, the strength of the 100% shale gouge would be reduced as well. In the present study, the composite mixtures of shale and carbonate represent a ternary mixture: the shale is composed mostly of phyllosilicates and quartz (57% phyllosilicate, 32% quartz). Thus strain can be partitioned into the carbonates, the phyllosilicates or the quartz grains, which are typically larger than the other minerals. In these experiments, grain size is an important control on sample strength. The presence of large quartz grains inhibits the propagation of full thickness R1 shears [Tembe et al., 2010]. Cataclasis of quartz and subsequent generation of fine grains promotes homogenous flow and strain hardening in shale-rich gouge mixtures [Biegel et al., 1989; Gu and Wong, 1994]. 91 The behavior of ternary mixtures is more complex. Tembe et al. [2010] illustrated that the addition of illite to quartz causes a linear decrease in frictional strength while the addition of montmorillonite to quartz causes a sigmoidal (i.e. little change with small amounts of montmorillonite, but a sharp change with 30% montmorillonite) decrease in frictional strength (Figure 3.12). An average of these two curves does not fit the coefficient of friction observed in a ternary mixture of quartz, illite and montmorillonite. Rather, the montmorillonite more strongly controls the behavior of the ternary mixtures, since it is the weakest mineral phase and can organize into margin parallel slip surfaces that accommodate the majority of the deformation. Figure 3.12 compares µ values determined for room T gouge from the present study to results from experiments on kaolinite/quartz and illite/quartz ([Crawford et al., 2008; Tembe et al., 2010]) composite gouges. This graph shows a linear relationship between phyllosilicate content and coefficient of friction for both illite/quartz and kaolinite/quartz composites. In contrast, my study shows that the addition of small amounts of phyllosilicate (~14% - .58% of the 25% shale) reduces the frictional strength of the carbonate gouge by 25%. The addition of up to 57% phyllosilicates (i.e. the 75% shale) only reduces the strength of the carbonate gouge by an additional 5%. A ‘framework’ model is invoked to explain the distribution of mineral phases in quartz and phyllosilicate bearing gouges [Crawford et al., 2002; Crawford et al., 2008; Cumberland and Crawford, 1987; Handy, 1990; Logan and Rauenzahn, 1987; Lupini et al., 1981; Tembe et al., 2010]. Typically, strong mineral phases (e.g. quartz) in high concentrations form a stress supporting framework, while phyllosilicates fill the pore space between larger clasts (this is known as Regime 1, Figure 3.9). Above a certain threshold, the phyllosilicates form 92 Figure 3.12: Comparison with Other Experimental Data Illite/Quartz, Tembe et al., 2010 7.98 mm displ. Kaolinite/Quartz, Crawford et al., 2008 3.0-3.4 mm displ. this study, friction at 2.8 mm displ. % Phyllosilicate 0 1.0 20 40 60 80 100 0.8 0.6 0.4 0.2 0.0 Co e! ci en t o f F ri ct io n after Tembe et al, 2010 Montmorillonite/Quartz, Tembe et al., 2010 2.3 mm displ. Montmorillonite/Quartz, Takahashi et al., 2007, 2.2 mm slip-parallel displ. Figure 3.12: Percent phyllosilicate vs. coefficient of friction. Blue squares show the results from the present study. Results from previous studies on quartz and phyllosilicate composites are shown in black and results from previous studies on quartz and montmorillonite composites are shown in pink. Previous studies on quartz/illite and quartz/kaolinite composites show a linear decrease in strength with increasing phyllosilicate content, while previous studies on quartz/montmorillonite composites show a sigmoidal decrease in strength with increasing phyllosilicate content. In my study, the strength of the carbonate gouge decreases by 25 % with the addition of 15% phyllosilicates (i.e. the 25% shale) while the addition of up to 57% phyllosilicates (i.e. the 75% shale) only reduces the strength of the carbonate gouge by an additional 5%. 93 a continuous matrix that reduces contact points between quartz grains and weakens the sample. Tembe et al. [2010] suggest that at this point, the presence of still significant amounts coarse grained quartz inhibits the development full thickness of R1 shears (Regime 2). However once phyllosilicate content reaches another critical threshold, the quartz grains are in such a low concentration that they do not significantly affect R1 shear propagation (Regime 3). The framework model described above assumes that the strong mineral phase is also coarse grained. In this model, fine-grained clay particles fill porosity between larger quartz clasts (Figure 3.13). The experiments in the present study provide an example of a ternary mixture in which the strongest mineral phase is also fine grained. With > 50% carbonate, the fine grained carbonate gouge forms a stress supporting matrix as described in the framework models (Figure 3.13). However due to their fine grain size, there is very little porosity between carbonate grains. Rather, the carbonate fills spaces between quartz and phyllosilicate-rich lenses. Within these lenses, phyllosilicates typically fill spaces between quartz grains. As deformation continues, tails on these lenses elongate and connect to each other forming margin parallel bands containing aligned phyllosilicates. Because the strong carbonate is already fine-grained, little energy goes into comminution of the carbonate, nor does the grain size of the carbonate clasts inhibit phyllosilicate rotation and the development of margin parallel shears. As a result, the fine-grained carbonate gouge facilitates the formation of phyllosilicate-rich bands, and increased strain localization relative to 100% shale gouge that contains only coarse grained quartz and phyllosilicates. The comparatively 94 Figure 3.13 (following page): Schematic diagram of the framework model (green = quartz; red = phyllosilicates; blue = carbonate). a. At high quartz content, the coarse grained quartz forms a stress supporting framework. b. Phyllosilicates fill in the void space between the quartz grains. Hertzian-style tensile fracturing occurs at grain-to- grain contacts in quartz. c. With increasing phyllosilicate content, the phyllosilicates separate the quartz grains and reduce the number of strong grain-to-grain contacts. Little quartz grain size reduction occurs, and strain localizes in the phyllosilicate matrix. In the present study the strong carbonate and the weak phyllosilicates are the same size, and the quartz is coarse grained. d. In carbonate rich gouges, the carbonate forms a stress supporting matrix. Phyllosilicates may fill void spaces between the carbonate and the coarse grained quartz, or be interspersed in the carbonate matrix. e. With increased phyllosilicate content, both phyllosilicates and carbonate form the matrix separating the quartz clasts. The fine grained carbonate facilitates phyllosilicate rotation and strain localization. f. In this study as phyllosilicate content increases, the amount of coarse grained quartz also increases. The coarse grained quartz prevents phyllosilicate rotation and inhibits strain localization within the shale-rich gouges in the present study. 95 Figure 3.13: Framework Model Framework Model - Binary clay volume = Φ between quartz grains framework supported matrix supported carbonate and phyllosilicate matrix supported carbonate matrix supported phyllosilicate matrix supported This study - Modified for ternary mixture a b c d e f 96 weak phyllosilicate-rich bands accommodate deformation, resulting in a significant decrease in carbonate gouge strength with the addition of only 14% phyllosilicates and < 4 mm displacement. R1 shear angle measurements suggest that the composite gouges have a lower angle of internal friction than endmember gouges which supports the idea that the fine grained carbonate facilitates strain localization in the phyllosilicates in the form of R1 and Y shears. Figure 3.11 shows the relationship between the measured coefficient of friction, the internal friction angle, and the angle of riedel shear localization. Data from mixed binary and ternary mixtures of quartz, illite and montmorillonite bearing gouges from Tembe et al. [2010] are plotted on Figure 3.11. Gouges from Tembe et al. [2010] that exhibit microstructure characteristic of Regime 1, in which clay particles fill the pore space between adjacent quartz grains, are shown by open squares. Gouges from Tembe et al. [2010] that exhibit microstructure characteristic of Regime 2, in which quartz grains are no longer touching, but R1 shear development is impeded by large quartz grains, are shown by solid squares. Regime 1 samples display a higher internal friction angle than Regime 2 samples. The gouges in the present study all plot close to the Regime 1 samples from the Tembe et al. [2010] study, despite the presence, in some cases, of significant amounts (i.e. up to 57%) of phyllosilicates, likely reflecting the relatively high coefficient of friction of the carbonate (Figure 3.11). Tembe et al. [2010] observe a decrease in the internal friction angle (µi) with increasing phyllosilicate content. In contrast, the composite gouges in the present study display the lowest internal friction angle, while the endmember gouges display the highest (Figure 3.11). This relationship indicates that combining the shale and carbonate gouges decreases the internal friction angle relative to the endmember gouges, and may help explain 97 how the addition of the comparatively strong carbonate helps to facilitate the formation of margin parallel Y shears in the composite gouges. The result is a significant decrease in strength with the addition of small amounts of phyllosilicates (~14% in the 25% shale gouge). 3.4.2. Effects of T and Pf on Carbonate Gouge Strength and Stability At elevated temperature and pore fluid pressure, the carbonate gouge displays pronounced strain weakening followed by the evolution to stick-slip sliding. The presence of stick-slip in carbonate gouge at elevated temperature and pore fluid pressure suggests strain localization in Y-shears [JM Logan et al., 1992; JM Logan et al., 1979; Moore et al., 1988], or margin parallel boundary shears [Gu and Wong, 1994]. Logan et al. [1992] performed experiments on black and white colored carbonate gouge to analyze the sequence of microstructure development and the relative proportion of strain accommodated by discrete zones of intense grain size reduction and shearing in specific orientations (i.e. shear bands). In these experiments, R1 shear development accompanied strain softening following the peak stress. With continued displacement the R1 shears transitioned to Y-shears, which accommodated most of the subsequent deformation. Stick-slip sliding only began once Y- shears were well developed. Logan et al. [1992] suggests that when Y-shears form, the micromechanical processes occurring on the Y-shear surface is analogous to the processes that occur when 2 intact rocks slide against each other. Therefore, the presence of stick-slip sliding suggests strain localization in margin parallel shear bands. Logan and Rauenzahn [1987] observe stick-slip sliding in room T, saturated quartz gouge. Stick-slip sliding in quartz transitions to stable sliding between 400°C and 600°C at 300 MPa confining pressure [Brace, 1972; Stesky et al., 1974]. In contrast, Olsson [1974] 98 observed stable sliding in a nominally dry, bioclastic limestone at room T, with a transition to stick-slip behavior at confining pressures of 10 – 60 MPa, and 200°C to 300°C. While Kawamoto and Shimamoto [1998] observe stick-slip behavior in calcite gouge at temperatures up to 700°C. Similarly, Drennon and Handy [1972] observe stable sliding in nominally dry, bioclastic limestone at room T and very low confining pressure (Pc = 73 kPa – 19 MPa), and stick-slip sliding at T > 100 °C. Drennon and Handy [1972] suggest that elevated temperature drives off adsorbed water films, resulting in the transition to stick-slip sliding. This proposed mechanism is not the cause of the stick-slip observed in the experiments in the present study, since elevated pore fluid pressures were present in all carbonate gouges displaying stick-slip behavior. In my experiments, stick-slip only occurred in carbonate gouge with both elevated pore fluid pressure and elevated T. My experiments performed on dry carbonate gouge at elevated temperature display pronounced strain weakening, but no stick-slip behavior. The peak coefficient of friction observed in the dry carbonate gouge at elevated temperature is essentially the same as that observed in the saturated gouge at elevated temperature. At maximum shear strains, however, the dry sample is ~18% weaker than the sample with elevated pore fluid pressure at comparable displacements. Morrow et al. [2000], observed the exact opposite: an 18-20% decrease in the frictional strength of a saturated calcite gouge relative to a dry calcite gouge. Their sample is saturated, however, while the sample in the present study has an elevated pore fluid pressure. As with the phyllosilicate bearing gouges in this study, the increased strength of the carbonate sample with elevated pore fluid pressure relative to the dry sample is likely due to dilatancy hardening. Transient pore fluid pressure effects may also promote stick-slip behavior in carbonate 99 gouge under elevated pore fluid pressure conditions. Experiments on quartz gouge show that during stable sliding increased displacement and increased displacement rates result in an increase in porosity [Marone et al., 1990]. During stick-slip sliding however, there is a decrease in porosity during the slip event, and an increase in porosity during the stick event [Lubert and de Ryck, 2001]. Porosity in the gouge zones increase during the stick portion of the stick slip event, which leads to dilatancy hardening. After a certain amount of displacement, the increase in permeability allows a rapid increase in pore fluid pressure, which reduces the effective confining pressure and leads to failure. With continuing displacement, dilatancy hardening would again increase the sample strength until permeability is sufficient to allow pore fluid pressure to increase, leading to failure. This process is similar to hydraulic jacking of a glacier in which increases in pore fluid pressure cause periodic sliding because of the subsequent reduction in effective stress. Segall and Rice [1995] suggest that the magnitude of the stress drop in samples displaying dilatancy hardening correlates with hydraulic diffusivity. In the experiments in the present study, samples with a sandstone forcing block have the highest hydraulic diffusivity since hydraulic diffusivity increases with increasing porosity. The magnitude of the stress drops in samples with 1 porous forcing block is larger (~ 10 MPa) and more regular than the stress drops recorded in samples with two dolomite forcing blocks (~5 MPa avg. stress drop magnitude). Based on the model by Segall and Rice [1995], the highest magnitude stress drops should be observed in completely drained experiments. However, room T, saturated experiments performed on 100% carbonate gouge do not exhibit stick-slip sliding, indicating that increased temperature is required to promote stick-slip sliding in the carbonate gouge. All phyllosilicate-bearing gouges exhibit stable sliding at elevated temperature and pore 100 fluid pressure indicating that: 1) phyllosilicates suppress stick-slip sliding, and 2) transient pore fluid pressure effects alone cannot be responsible for stick-slip sliding. Based on the observations above, it is likely that temperature is key to promoting strain weakening and stick-slip behavior in the carbonate gouge and that increased pore fluid pressure can promote stick-slip sliding, however, that material must already be prone to stick-slip sliding. 3.4.3. Implications for Natural Fault Zones 3.4.3.1. Strain Partitioning in Thrust Faults In many fold and thrust belts, the hanging wall carbonate consists of micritic limestone [Burkhard, 1990], and in most cases, the carbonate matrix in the cataclasites is very fine grained (< 10 µm in the Hunter Valley Thrust, Appalachians [Kennedy and Logan, 1998; Wojtal and Mitra, 1986], and < 50 µm in the matrix dolomite of the Lewis Thrust, Rocky Mountains[Erickson, 1994]). In the Hunter Valley Thrust, the limestone thrust sheet is heterogeneous, and contains both dolomite and quartz in addition to calcite [Kennedy and Logan, 1998]. In addition, the limestone cataclasites in the McConnell Thrust, Canadian Rockies are enriched in dolomite relative to the limestone hanging wall [Kennedy and Logan, 1997]. These observations indicate that both the fine grain size and composition of the carbonate gouge used in the present study are representative of the materials found in natural thrust faults. Furthermore, field studies of foreland thrust sheets deformed at temperatures between 100°C-175°C ([Erickson, 1994; Kennedy and Logan, 1998; Wojtal and Mitra, 1986]) show that illite and chlorite, with minor amount of smectite comprise the dominant shale mineralogy [Erickson, 1994; Kennedy and Logan, 1998; Vrolijk and van der Pluijm, 1999]. 101 Despite sometimes large amounts of smectite in underlying footwall shales, Vrolijk and van der Pluijm [1999] show that the illite/smectite ratio increases from the 30:70 ratio in underlying footwall shales to 80:20 approaching the fault zone (the McConnell Thrust and the Lewis Thrust, both in the Canadian Rockies). These observations indicate that the mineralogy in the shale gouge in this study is a reasonably appropriate proxy for the materials actually found in these fault zones. Field studies show that deformation in the hanging wall begins as localized fractures, and as deformation progresses, cataclasis along these fractures widens the zone of deformation [Erickson, 1994]. In contrast, deformation in the footwall begins as more diffuse deformation resulting in the development of a foliation and broad folds, followed by strain localization in shear bands [Erickson, 1994]. These observations support the observations of the present study that suggest that discrete slip surfaces are likely to develop first in the carbonate-rich material, while deformation in the shale material is likely to initially exhibit more homogeneous flow. The texture Erickson [1994] observed in the shale cataclasite of the Lewis Thrust, in which anastamosing bands of aligned clay minerals surrounding domains of randomly oriented minerals is reminiscent of the heterogeneous deformation style observed in the 100% shale gouge samples in the present study. Within the mixed carbonate-shale cataclasite of the Lewis Thrust, alternating bands of carbonate and aligned phyllosilicates form compositional layering [Erickson, 1994] similar to that formed in the composite gouge samples in the present study. Textures observed in both natural thrust faults and the experiments in the present study suggest that the grain size of the various mineral phases in the gouge is a dominant 102 control on the deformation style. In the Hunter Valley Thrust cataclasites, Kennedy and Logan [1998] observed that quartz grains are more coarse grained than either the shale or carbonate minerals, a grain size distribution reminiscent of the gouge materials in the present study. In both cases, the coarse grained quartz deformed primarily by microfracturing and cataclasis, and these microfractures are filled by matrix material. Wojtal and Mitra [1986] observe that areas adjacent to rigid grains that are normally reserved for pressure fringes are filled by the fined grained calcite matrix. They interpret this texture to represent low T grain boundary sliding and particulate flow in the matrix. This texture is similar to that observed in the composite gouges in the present study where the fine grained carbonate matrix appears to wrap around fractured and elongated quartz clasts. Field studies of these fault zones suggest that increased displacement causes changes in strain partitioning and the dominant deformation mechanisms operating within these faults zones [Erickson, 1994; Kennedy and Logan, 1998; Wojtal and Mitra, 1986]. Erickson[1994] suggests that at the Lewis Thrust, strain hardening in both the shale and carbonate cataclasites results in shear zone abandonment, and the development of new shear zones. In contrast, Wojtal and Mitra [1986] and Kennedy and Logan [1998] suggest that in the Hunter Valley Thrust initial grain size reduction by cataclasis is followed by the transition to intracrystalline dislocation processes and solution transfer in fine grains. This causes weakening and more localized deformation without additional grain size reduction. Experiments in the present study indicate that at room T (i.e. shallow depths), and low shear strain, deformation is likely to localize in the weak shale and carbonate/shale composite material. At 150°C, shale and shale-rich cataclasites are likely to initially accommodate most of the deformation. However, with continuing displacement, the 103 carbonate rich-hanging wall is likely to undergo significant strain softening caused by more rapid strain localization, and deformation is likely to be transferred to the carbonate hanging wall. These experiments suggest that increased displacement, especially in the presence of elevated pore fluid pressure, could lead to seismic faulting. Subsequent coseismic frictional heating can cause the carbonate to decompose into CO2 and nanoparticles of lime (CaO) or periclase (MgO) which have extremely low frictional strengths (µ~.1) [De Paola et al., 2011; Han et al., 2007]. This process, called thermal decomposition, dramatically reduces the strength of the carbonate hanging wall during seismic faulting. The presence of stylolites, calcite veins, and evidence of both calcite and quartz precipitation provides evidence of fluid flow in fault zones [Erickson, 1994; Kennedy and Logan, 1998; Vrolijk and van der Pluijm, 1999; Wojtal and Mitra, 1986]. Solution transfer processes are known to weaken quartz and carbonate during deformation and hence, if solution transfer is the dominant deformation mechanism that accommodates displacement, the strength carbonate cataclasite would be greatly reduced [Durney, 1972; Elliott, 1976; Gratier et al., 1999; Rutter, 1983]. These processes are important in natural fault zones, but were not observed on the timescales of the experiments in the present study. 3.4.3.2. Implications for Thrust Belt Models In numerical simulations of thrust sheets, the coefficient of friction plays an important role in determining the basal wedge geometry and deformation style. The rheology of the basal detachment controls: 1) the location and propagation of the detachment zone, 2) the development of secondary faults, 3) the amount of penetrative deformation in the hanging wall and 4) ramp spacing [Couzens et al., 1997; Kukowski et al., 2002; Nemcok et al., 2005; Riley et al., 1995; Smart et al., 1999; Strayer and Hudleston, 1997]. 104 In two dimensional thrust wedge models of the mechanics of thrust belt deformation, the basal coefficient of friction has a significant impact on the taper and thickness of the wedge [Davis et al., 1983]. The critical taper describes the minimum thickness a thrust belt can be to advance without internal deformation. Deformation occurs within the wedge (e.g. by basin inversion, backthrusting, development of duplexes) in order for the wedge to maintain the critical taper [Nemcok et al., 2005]. An increase in the basal coefficient of friction results in an increase in the critical taper (Figure 3.14), meaning that wedges with a higher basal coefficient of friction will be thicker than wedges with a lower coefficient of friction [Davis et al., 1983]. The experiments in this study suggest that friction coefficients in the range of .6-.7 would be appropriate values if a similar shale forms the basal detachment in these models. These values are less than the µ = .85 values that many thrust belt models use in accordance with Byerlee’s Law [Byerlee, 1978; Davis et al., 1983]. The use of the friction coefficients measured in this study would result in a smaller critical taper, and decreased thrust wedge thickness. In a different thrust sheet simulation, Smart [1999] examines the effect of basal coefficient of friction on the relative importance of internal deformation in the thrust sheet versus displacement along the basal detachment. In Smart’s study, a higher basal coefficient of friction results in little displacement along the basal detachment and instead results in backthrusting. A higher basal coefficient of friction results in more backthrusting while a lower basal coefficient results in more forethrusting. Smart [1999] determines that a coefficient of friction of less than .35 is required to produce significant offset along the basal 105 Figure 3.14: Critical Taper (after [Davis et al., 1983]) = decrease in wedge thickness after Davis et al. 1983 !xed wedge shape supercritical subcritical increase in basal friction decrease in basal friction basal dip β su rf ac e sl op e 0° 2° 4° 6° 8° 10° 12° 14°0° 2° 4° 6° 8° initial critical taper Figure 3.14: Basal dip vs. surface slope diagram [Davis et al., 1983] showing the effect of basal coefficient of friction on modeled thrust wedge thickness. This graph shows that a decrease in the basal coefficient of friction results in a decrease in critical taper, and a thinner thrust wedge. Using the coefficient of friction of .6 - .7 determined for shale in this study to represent the strength of the basal detachment would result in a decrease in the modeled thrust wedge thickness. 106 detachment. Using the coefficient of friction values determined for any material in the present study would result in backthrusting without significant displacement along the basal detachment. The models described above highlight the effect that the basal coefficient of friction has on the geometry of fold and thrust belts in thrust fault models. Typically, a higher basal coefficient of friction means increased thrust sheet deformation, while a lower basal coefficient of friction results in increased displacement along the basal thrust fault. However, the values used in these models vary wildly, and assume that the coefficient of friction stays constant through time. The experiments in the present study provide more realistic values for these models, and show that at initial displacements, the coefficient of friction in the shales will be relatively low. The basal coefficient of friction is expected to evolve with time due to mixing between the hanging wall carbonate and footwall shales. The models predict that this change would result in decreased displacement along the basal detachment and increased internal deformation in the hanging wall, possibly in the form of backthrusting. 3.5. Conclusions 1) At room T, saturated conditions gouge strength increases with increasing carbonate content. All gouges display stable sliding. 2) At 150°C and 15 MPa pore fluid pressure, carbonate undergoes significant strain weakening following the peak strength, and displays regular stick-slip sliding at maximum shears strains 107 3) Experimental results indicate that in thrust faults at deeper levels (i.e. at elevated temperature and pore fluid pressure), deformation is likely to localize in the carbonate-rich hanging wall and carbonate cataclasite at higher shear strain. 4) At lower shear strain and more shallow crustal levels, deformation is more likely to be partitioned into the shale and carbonate-shale composite cataclasites 5) This study provides realistic friction coefficient values for thrust fault materials, and has an effect on the geometry and mechanical behavior of thrust sheets used in these models. 108 4. Discussion 4.1. Comparison of the Two Experimental Suites and Conclusions Results from room temperature, saturated experiments conducted at the University of British Columbia (Part I of thesis) are compared to the results from room T, saturated experiments and elevated temperature and pore fluid pressure experiments conducted at the University of Liverpool. Possible reasons for differences or similarities between the two suites of experiments are reported. Experimental conditions and procedures for the experiments are listed in Table 3.1. 4.1.1. Comparison Between Relative Strengths of Shale and Composite Gouges At room temperature, 100% carbonate gouge is the strongest material in experiments performed at both labs. The shale-calcite composite gouges are the weakest materials at the UBC laboratory. In contrast, the 100% shale is the weakest at the Liverpool lab and, the composites are intermediate in strength (Figure 4.1; Figure 4.2). The shale gouge is the same composition for both experimental suites, however, the Liverpool shale gouge is coarser grained than that used at UBC (Appendix A). The larger grain size will increase the strength of shale bearing gouges since more energy goes in grain size reduction and cataclasis. In addition, more displacement is required for strain localization in coarse grained gouges [Gu and Wong, 1994; Logan and Rauenzahn, 1987]. Since the coarse grained quartz in the shale gouges appears to inhibit strain localization in the phyllosilicates, strain localizes more effectively in the experiments performed at UBC, and 109 Figure 4.1: Carbonate Content vs. Coefficient of Friction Room TRoom T 150 C, Pf 150 C Pf, dol Percent Carbonate Co ef fic ie nt o f F ric tio n 1009080706050403020100 .65 .7 .75 .8 .85 .9 .95 1 .6 Room T 150°C, Pf 150°C, dol Room T, UBC Figure 4.1: Percent carbonate vs. coefficient of friction. Room T, saturated experiments from UBC typically display a lower coefficient of friction than experiments performed at U of L. In experiments performed at UBC (yellow) carbonate is the strongest, the composites are the weakest and the shale is intermediate in strength. In contrast, in room T, saturated experiments performed at U of L carbonate is the strongest but the shale is the weakest and the composites are intermediate in strength. 110 Figure 4.2: Comparison to Previous Data Illite/Quartz, Tembe et al., 2010 7.98 mm displ. Kaolinite/Quartz, Crawford et al., 2008 3.0-3.4 mm displ. Part 2, this study, friction at 2.8 mm displ. % Phyllosilicate 0 1.0 20 40 60 80 100 0.8 0.6 0.4 0.2 0.0 Co e! ci en t o f F ri ct io n after Tembe et al, 2010 Montmorillonite/Quartz, Tembe et al., 2010 2.3 mm displ. Montmorillonite/Quartz, Takahashi et al., 2007, 2.2 mm slip-parallel displ. Part 1, this study, friction at 2.8 mm displ. Figure 4.2: Percent phyllosilicate vs. coefficient of friction. Orange squares show the results of room temperature, saturated experiments from Part I, and blue squares show the results of room temperature, saturated experiments from the Part II. All experiments shown from this study were performed with one Berea Sandstone forcing block. Results from previous studies on quartz and phyllosilicate composites are shown in black and results from previous studies on quartz and montmorillonite composites are shown in pink. The copper jacket used for experiments in Part II likely accounts for the increased strength of Part II samples relative to the experiments performed in Part I that used a weaker polyolefin jacket. The experiments from Part I show an initially linear decrease in strength with increasing phyllosilicate content, followed by a subsequent increase in strength at 57% phyllosilicate. The experiments from Part II show a rapid initial decrease in strength with the addition of small amounts of phyllosilicates, while the addition of more phyllosilicates leads to little subsequent decrease in strength. The more coarse grained quartz used in experiments at the U of L may account for the differences in behavior between these two experimental suites. 111 gouge is subsequently weaker. The effects of grain size on the strength of the composite gouges are discussed in more detail in section 4.1.4. 4.1.2. Effects of the Copper Jacket on Frictional Strength In general, the coefficient of friction values from room temperature, saturated experiments conducted at the University of Liverpool (Part II; Table 3.2) are high relative to published data, and to the same gouge material deformed at UBC (Part I; Table 2.2). Experiments conducted at the University of Liverpool specifically to address the effect of the jacket strength on the sample strength indicate that at 70 MPa confining pressure, the jacket strength has an effect of less than .1 on the total coefficient of the sample [Behnsen, 2011; Smith and Faulkner, 2010]. Although this effect is minimal, it does explain the slightly high friction coefficients observed in these experiments compared to published experiments and my experiments from Part I of the thesis. 4.1.3. Effects of Pore Fluid Pressure Experiments at UBC were performed saturated, but without an elevated pore fluid pressure. During the initial stage of loading, compaction and brittle grain size reduction reduce porosity in the gouge zone. In samples with a sandstone (porous) forcing block, fluid from the gouge zone is expelled into the sandstone forcing block. This reduces the effective confining pressure in the forcing block (since Pf increases), but increases the effective confining pressure in the gouge zone (since Pf decreases); The gouge becomes stronger while the forcing block becomes weaker. 112 In samples deformed with 2 non-porous forcing blocks, fluid is trapped in the gouge zone. Compaction at the start of the experiment increases pore fluid pressure in the gouge zone; Gouge zone volume is reduced, but the pore fluid is trapped. As a result, the Pf increases in the gouge zone increases, thereby decreasing the effective stress in the gouge zone, which lowers the yield strength of gouges deformed with 2 non-porous forcing blocks. After yielding, compaction in the gouge is replaced by dilation as particles rearrange to accommodate displacement along the saw cut. In samples with 2 non porous forcing blocks, permeability is not sufficient to allow Pf to equilibrate during sliding, and dilatancy hardening commences and continues through the maximum displacement (~.5 – 5 mm). As a result, experiments performed with two non-porous forcing blocks continue to strain harden throughout the experiment. This is in contrast to experiments performed with one non-porous forcing block, where dilatancy hardening does not occur and the gouge attains steady state sliding. All room temperature, saturated experiments at Liverpool were conducted with one porous forcing block, and results are in agreement with the UBC data; the samples did not undergo dilatancy hardening. While the experiments performed with two non-porous forcing blocks at UBC were weaker than experiments performed with one non-porous forcing block, experiments performed with two non-porous forcing blocks at the University of Liverpool are stronger than those deformed with one non-porous forcing block. A 1 mm diameter hole was drilled into the upper dolomite forcing block at the Liverpool lab. This hole promotes communication between the pore fluid system and the gouge zone, such that Pf can escape from the gouge zone during experiments. Consequently, the transient increase in Pf and subsequent weakening in the gouge zone that was observed at UBC is not observed in the 113 samples deformed with 2 non-porous forcing blocks at the University of Liverpool. Instead, dilatancy hardening occurs, which increases the strength of samples with two non-porous forcing blocks relative to samples with one porous forcing block. 4.1.4. Microstructure Despite the differences in the mechanical strength between the room T, saturated experiments performed at UBC and those performed at the University of Liverpool, similar microstructures developed in both studies. However, the 75% and 50% shale experiments from UBC typically display slightly more evolved microstructure than experiments performed on the same compositions at the University of Liverpool. With increased carbonate content, inclined quartz clasts and aligned phyllosilicates define a shape (P) fabric in samples from both studies. In the 75% shale samples from UBC, a significant portion of the phyllosilicates align parallel to the shear zone boundary to produce a composite P-Y fabric, while the P- fabric dominates the microstructure in the 75% shale samples from Liverpool. In both studies, mineral phases separate to form compositional banding in 75% and 50% shale samples. In addition, Y shears form in 50% shale gouges deformed at room T, saturated conditions in both studies. In the Liverpool experiments comminuted quartz forms asymmetric tails on quartz grains, creating margin parallel, quartz-rich bands. In contrast, asymmetric tails on quartz grains are rare in samples from UBC. The relative abundance of these features in the U of L samples may be a result of the coarser initial grain size and more intense cataclasis in the Liverpool composite gouges. 114 Microstructure in the 25% shale gouge deformed at both UBC and the University of Liverpool is very similar. A well developed shape (P) fabric and R1 shears characterize the microstructure developed in this composition in both studies. The microstructure formed in these experiments most closely resemble each other, and the strength of the 25% shale gouges at both UBC and the University of Liverpool is also the most similar. In both studies, the grain size contrast between the coarse grained quartz in the shale and the fine grained carbonate has a significant effect on microstructure development and deformation mechanism. The coarse grained quartz promotes brittle grain size reduction and makes strain localization in quartz-rich gouges difficult [Gu and Wong, 1994; Logan and Rauenzahn, 1987]. The replacement of coarse grained quartz with fine grained carbonate facilitates the development of R1 shear bands and the development of aligned phyllosilicates. R1 shear bands become increasingly prominent as carbonate content increases. In the 100% carbonate gouge from both studies, R1 shear bands are the dominant strain accommodation feature visible in gouges at room T, saturated conditions. Increasing carbonate content in composite gouges, leads to a merging of the coefficient of friction observed in both experimental suites. This relationship suggests that the shale endmember is likely responsible for the differences in microstructure and mechanical behavior observed between the two studies. The porous sandstone forcing block in the UBC experiments underwent extensive deformation for 100% and 75% carbonate gouge, while there is no deformation in the sandstone forcing block in similar Liverpool experiments. The UBC experiments show riedel shears that extend from the gouge zone into the forcing block, and gouge is injected into 115 these fractures. In addition, there is significant quartz grain crushing and porosity reduction in parts of the forcing block adjacent to the gouge zone. Experiments performed at UBC accommodated ~5-6 mm of axial displacement, while experiments performs at the University of Liverpool accommodate ~3-4 mm of axial displacement. The relationship between fault zone thickness and displacement is well established [Childs et al., 2009; Evans, 1990; Hull, 1988; Knott, 1994; Scholz, 1987; Shipton et al., 2006]. Broadly, thickness typically scales linearly with displacement, and increased displacement results in a wider fault zone. Since experiments at UBC accommodated ~ 1.5 times the displacement that the samples from the University of Liverpool accommodated, it is possible that the damage recorded in the forcing block in carbonate-rich samples from UBC is due to this extra displacement. Alternatively, the faster strain rate used in experiments at UBC may have promoted fracturing in the forcing block. Velocity–stepping experiments were performed at UBC but constant displacement rate was used at Liverpool - hence the UBC experiments had a complicated loading path and a wide range of displacement rates: nonetheless, the microstructures between the labs are similar. Likewise, Tembe et al. [2010] observe that the microstructures developed in their quartz/montmorillonite composites deformed at a constant displacement rate closely resemble those formed in the velocity stepping experiments conducted quartz/montmorillonite composites [Logan and Rauenzahn, 1987]. Combined, these observations suggest that the loading path does not significantly alter the fabric development within these gouges. 116 4.1.5. Ideas for Future Work The previous sections highlight the potential effects that laboratory variables can have on the outcome of an experiment. More work needs to be done to more effectively apply the results of this study to natural systems. In this section, I will make several suggests for future experiments that could isolate the effects of the aforementioned variables, or test some of the ideas put forth in this thesis. Future experiments would seek to address the relative importance of pore fluid pressure versus temperature on the behavior of the carbonate gouge. This suite would include experiments on 100% carbonate gouge at the following conditions: 1) oven-dried, room T, 2) elevated Pf, room T, 3) saturated, 150°C. From this suite of experiments, the effects of pore fluid pressure could be isolated from the effects of elevated temperature. In addition, these experiments could address the how mechanical behavior in a fault zone might change as the result of a sudden decrease in pore fluid pressure. In this thesis, significant attention was devoted to the relative strengths of each gouge type. This was to determine which gouge composition is the weakest under experimental conditions, with the aim of assessing which gouge composition is most likely to accommodate displacement in a fault zone. Each experiment was conducted on a single gouge composition. Future experiments in which the gouges are layered within a single sample could provide insight into which gouge composition actually accommodates deformation. For example, the 100% shale gouge from UBC could be layered with the 50% shale gouge from UBC to test the hypothesis proposed in Part I that strain localized most effectively in the 50% shale gouge. 117 The experiments in this thesis were conducted on a natural, polymineralic shale and a reagent grade calcite powder that is also bimineralic. Although there is merit in using natural materials, the presence of many mineral phases adds a significant amount of uncertainty. Mineral phases cannot be varied independently, and the effects of specific mineral phases cannot be isolated. In future experiments I would reduce the number of mineral phases. Specifically, I would investigate ternary mixtures of quartz, calcite and illite to determine how each endmember affects the strength of the gouge mixture. In addition, it would be useful to conduct a suite of experiments on calcite and illite gouges in which the grain size of calcite is varied. In rock friction studies of composite gouges, the strong endmember is typically coarse grained (i.e. > 10 µm) [Crawford et al., 2008; Kawamoto and Shimamoto, 1998; Logan and Rauenzahn, 1987; Tembe et al., 2010]. The purpose of this experiment would be to examine the effects of grain size on frictional strength and strain localization in bimineralic mixtures. These additional experiments would provide more insight into strain accommodation and the mechanical behavior of carbonate on shale thrust faults, as well as other shallow, crustal carbonate and phyllosilicate bearing fault zones. 118 5. Conclusion Combined, the results of both studies echo the results of Part II: At room T (i.e. shallow depths), and low shear strain, deformation is likely to localize in the weak shale and carbonate/shale composite material. At 150°C, shale and shale-rich cataclasites are likely to initially accommodate most of the deformation. However, with continuing displacement, the carbonate rich-hanging wall is likely to undergo significant strain softening caused by more rapid strain localization, and deformation is likely to be transferred to the carbonate hanging wall. The differences between the microstructure developed in the samples deformed at UBC and those deformed at the University of Liverpool are subtle, despite significant differences in mechanical strength. As a result, care must be taken when trying to assess the mechanical strength of gouges from microstructure observations. In addition, margin parallel shear bands and R1 shears develop in gouges that do not display stick-slip sliding (i.e. composite gouges, room T carbonate), as well as gouges that do (i.e. carbonate at elevated T and Pf). Although the presence of these structures may be required for stick-slip sliding to occur (e.g. [Gu and Wong, 1994; Logan and Rauenzahn, 1987]), their presence alone does not signify that stick slip behavior has occurred. Since the absolute coefficient of friction values from the present study vary wildly, it is difficult to suggest a specific value that should be used in thrust belt models. However, it is clear from these experiments that for all shale bearing gouges, Byerlee’s standard coefficient of friction value of .85 is too high and should not be used to approximate the basal coefficient of friction in thrust sheet models. 119 Works Cited Austin, N. (2003), An Experimental Investigation of Textural Controls on the Brittle Deformation of Dolomite, M.Sc. thesis, 95 pp, University of British Columbia, Vancouver. Austin, N., L. Kennedy, J. Logan, and R. Rodway (2005), Textural controls on the brittle deformation of dolomite: the transition from brittle faulting to cataclastic flow, Geological Society London Special Publications, 243(1), 51, Doi:doi:10.1144/GSL.SP.2005.243.01.06. Barnhoorn, A., M. Bystricky, K. Kunze, and L. Burlini (2005), Strain localisation in bimineralic rocks: Experimental deformation of synthetic calcite-anhydrite aggregates, Earth and Planetary Science Letters 240, 748–763 Doi:doi:10.1016/j.epsl.2005.09.014. Bartlett, W. L., M. Friedman, and J. M. Logan (1981), Experimental folding and faulting of rocks under confining pressure Part IX. Wrench faults in limestone layers, Tectonophysics, 79(3-4), 255-277, Doi:10.1016/0040-1951(81)90116-5. Beeler, N., T. Tullis, and A. Kronenberg (2007), The instantaneous rate dependence in low temperature laboratory rock friction and rock deformation experiments, J Geophys Res, 112, Doi:10.1029/2005JB003772. Behnsen, J. (Spring 2011), Ph.D. Candidate, University of Liverpool, Personal Communication. Biegel, R. L., C. G. Sammis, and J. H. Dieterich (1989), The frictional properties of a simulated gouge having a fractal particle distribution, Journal of Structural Geology, 11(7), 827-846, Doi:10.1016/0191-8141(89)90101-6. Blenkinsop, T. G. (1991), Cataclasis and processes of particle size reduction, Pure and Applied Geophysics, 136(1), 59-86, Doi:10.1007/bf00878888. 120 Bos, B., and C. Spiers (2000), Effect of phyllosilicates on fluid-assisted healing of gouge- bearing faults, Earth and Planetary Science Letters, 184(1), 199-210 Doi:10.1016/S0012- 821X(00)00304-6. Bos, B., and C. Spiers (2002), Frictional-viscous flow of phyllosilicate-bearing fault rock: Microphysical model and implications for crustal strength profiles, J Geophys Res, 107, 2028, Doi:10.1029/2001JB000301. Bos, B., C. Peach, and C. Spiers (2000), Frictional-viscous flow of simulated fault gouge caused by the combined effects of phyllosilicates and pressure solution, Tectonophysics, 327(3-4), 173-194, Doi:doi:10.1016/S0040-1951(00)00168-2. Bowden, F., and D. Tabor (1950), The friction and lubrication of solids, Clarendon Press, Oxford. Brace, W. (1972), Laboratory studies of stick-slip and their application to earthquakes, Tectonophysics, 14(3-4 ), 189-200, Doi:10.1016/0040-1951(72)90068-6. Brace, W., and J. Byerlee (1966), Stick-slip as a mechanism for earthquakes, Science, 153, 990-992, Doi:10.1126/science.153.3739.990. Bridgman, P. W. (1936), Shearing phenomena at high pressure of possible importance for geology, Journal of Geology 44, 653-669, Doi:10.1086/624468. Burkhard, M. (1990), Ductile deformation mechanisms in micritic limestones naturally deformed at low temperatures (150–350ºC), in Deformation Mechanisms, Rheology and Tectonics, edited by R. Knipe and E. Rutter, pp. 241–257, Geol. Soc. Spec. Publ. . Byerlee, J. (1978), Friction of rocks, Pure and applied geophysics, 116(4), 615-626, Doi:10.1007/BF00876528. 121 Byerlee, J. D., and R. Summers (1973), The effect of fault gouge on the stability of sliding on sawcuts in granite (Abstract), EOS Trans. AGU., 54, 1210, Doi:10.1007/BF00876528. Carpenter, B. M., C. Marone, and D. M. Saffer (2009), Frictional behavior of materials in the 3D SAFOD volume, Geophysical Research Letter, 36(5), L05302, Doi:10.1029/2008GL036660. Chester, F., M. Friedman, and J. Logan (1985), Foliated cataclasites, Tectonophysics, 111(1- 2), 139-146 Doi:10.1016/0040-1951(85)90071-X. Chester, F., J. Evans, and R. Biegel (1993), Internal structure and weakening mechanisms of the San Andreas fault, J Geophys Res, 98( B1), 771-786, Doi:10.1029/92JB01866. Childs, C., T. Manzocchi, J. Walsh, and C. Bonson (2009), A geometric model of fault zone and fault rock thickness variations, Journal of Structural Geology, 29(2), 223-354, Doi:10.1016/j.jsg.2006.06.016 Chinnery, M. A. (1964), The Strength of the Earth's Crust under Horizontal Shear Stress, Journal Geophysical Research, 69(10), 2085-2089, Doi:10.1029/JZ069i010p02085. Collettini, C., N. D. Paola, and D. R. Faulkner (2009a), Insights on the geometry and mechanics of the Umbria–Marche earthquakes (Central Italy) from the integration of field and laboratory data, Tectonophysics, 476(1-2), 99-109, Doi:10.1016/j.tecto.2008.08.013. Collettini, C., A. Niemeijer, C. Viti, and C. Marone (2009b), Fault zone fabric and fault weakness, Nature, 462, 907-910, Doi:10.1038/nature08585. Collettini, C., C. Viti, S. A. F. Smith, and R. E. Holdsworth (2009c), Development of interconnected talc networks and weakening of continental low-angle normal faults, in Geology, edited, pp. 567-570. 122 Crawford, B., R. Myers, A. Woronow, D. Faulkner, and E. Rutter (2002), Porosity- permeability relationships in clay-bearing fault gouge, SPE/ISRM Rock Mechanics Conference, Doi:10.2523/78214-MS. Crawford, B. R., D. R. Faulkner, and E. H. Rutter (2008), Strength, porosity, and permeability development during hydrostatic and shear loading of synthetic quartz-clay fault gouge, J Geophys Res, 113(B3), B03207, Doi:10.1029/2006jb004634. Cumberland, D. J., and R. J. Crawford (1987), The packing of particles, Medium: X; Size: Pages: 150 pp. Davis, D., J. Suppe, and F. Dahlen (1983), Mechanics of fold-and-thrust belts and accretionary wedges, J Geophys Res, 88 ( B2), 1153-1172, Doi:10.1029/JB088iB02p01153. De Paola, N., T. Hirose, T. Mitchell, G. Di Toro, C. Viti, and T. Shimamoto (2011), Fault lubrication and earthquake propagation in thermally unstable rocks, Geology, 39(1), 35-38, Doi:10.1130/G31398.1. Delle Piane, C., C. Wilson, and L. Burlini (2009), Dilatant plasticity in high-strain experiments on calcite-muscovite aggregates, Journal of Structural Geology, 31(10), 1084- 1099, Doi:10.1016/j.jsg.2009.03.005. Dieterich, J. (1978), Time-dependent friction and the mechanics of stick-slip, Pure and applied geophysics, 116( 4-5), 790-806, Doi:10.1007/BF00876539. Dieterich, J. (1979), Modeling of rock friction 1. Experimental results and constitutive equations, J Geophys Res, 84(B5), 2161–2168, Doi:10.1007/BF00876539. . 123 Dieterich, J., and G. Conrad (1984), Effect of humidity on time-and velocity-dependent friction in rocks, J Geophys Res, 89 (B6), 4196-4202, Doi:10.1029/JB089iB06p04196. Drennon, C. B., and R. L. Handy (1972), Stick-slip of lightly loaded limestone, International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 9(5), 603-608, Doi:10.1016/0148-9062(72)90011-3. Durney, D. W. (1972), Solution-transfer, an Important Geological Deformation Mechanism, Nature, 235(5337), 315-317, Doi:10.1038/235315a0. Elliott, D. (1976), The energy balance and deformation mechanisms of thrust sheets, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 283, 1312 1289-1312 Doi:10.1098/rsta.1976.0086. Engelder, J. T. (1974), Cataclasis and the generation of fault gouge, Bulletin of the Geological Society of America, 85(10), 1515, Doi:10.1130/0016- 7606(1974)85<1515:CATGOF>2.0.CO;2. Erickson, S. (1994), Deformation of shale and dolomite in the Lewis thrust fault zone, northwest Montana, USA, Canadian Journal of Earth Sciences, 31, 1440-1448, Evans, J. (1990), Thickness-displacement relationships for fault zones, Journal of Structural Geology, 12(8), 1061-1065, Doi:10.1016/0191-8141(90)90101-4. Evans, J. P. (1988), Deformation mechanisms in granitic rocks at shallow crustal levels, Journal of Structural Geology, 10(5), 437-443, Doi:10.1016/0191-8141(88)90031-4. Faulkner, D., A. Lewis, and E. Rutter (2003), On the internal structure and mechanics of large strike-slip fault zones: field observations of the Carboneras fault in southeastern Spain, Tectonophysics, 367(3-4), 235-251, Doi:10.1016/S0040-1951(03)00134-3. 124 Faulkner, D. R., C. A. L. Jackson, R. J. .J. Dunn, R. W. Schlische, Z. K. Shipton, C. A. J. Wibberley, and M. O. Withjack (2010), A review of recent developments concerning the structure, mechanics and fluid flow properties of fault zones, Journal of Structural Geology, 32(11), 1557-1575, Doi:10.1016/j.jsg.2010.06.009. Gratier, J., F. Renard, and P. Labaume (1999), How pressure solution creep and fracturing processes interact in the upper crust to make it behave in both a brittle and viscous manner, Journal of Structural Geology, 21(8-9), 1189-1197, Doi:10.1016/S0191-8141(99)00035-8. Groshong, R. (1988), Low-temperature deformation mechanisms and their interpretation, Bulletin of the Geological Society of America, 100, 1329-1360, Doi:10.1130/0016- 7606(1988)100<1329:LTDMAT>2.3.CO;2. Gu, Y., and T. Wong (1994), Development of shear localization in simulated quartz gouge: effect of cumulative slip and gouge particle size, Pure and applied geophysics, 143 (1-3), 387-423, Doi:10.1007/BF00874336. Han, R., T. Shimamoto, T. Hirose, J. Ree, and J. Ando (2007), Ultralow friction of carbonate faults caused by thermal decomposition, Science, 316(5826), 878, Doi:10.1126/science.1139763. Handin, J., M. Friedman, J. M. Logan, L. J. Pattison, and H. S. Swolfs (1972), Experimental folding of rocks under confining pressure; Buckling of single-layer rock beams Flow and Fracture of Rocks, Geophysical Monograph, 16, 1-28, Doi:10.1130/0016- 7606(1976)87<1035:EFORUC>2.0.CO;2. Handy, M. (1990), The solid-state flow of polymineralic rocks, Journal Geophysical Research, 95(B6), 8647-8661, Doi:10.1029/JB095iB06p08647. 125 Hippertt, J. (1994), Grain Boundary Microstructures in Micaceous Quartzite: Significance for Fluid Movement and Deformation Processes in Low Metamorphic Grade Shear Zones, The Journal of Geology, 102(3), 331-348, Doi:10.1086/629675. Holyoke III, C., and J. Tullis (2006), Mechanisms of weak phase interconnection and the effects of phase strength contrast on fabric development, Journal of Structural Geology, 28, 621-640, Doi:10.1016/j.jsg.2006.01.008. Hull, J. (1988), Thickness-displacement relationships for deformation zones, Journal of Structural Geology, 43, 257-291, Doi:10.1016/0191-8141(88)90020-X. Ikari, M., D. Saffer, and C. Marone (2009), Frictional and hydrologic properties of clay-rich fault gouge, in J. Geophys. Res., edited. Ikari, M., C. Marone, and D. Saffer (2011), On the relation between fault strength and frictional stability, Geology, 39(1), 83-86, Doi:10.1130/G31416.1. Jefferies, S., R. Holdsworth, T. Shimamoto, and H. Takagi (2006), Origin and mechanical significance of foliated cataclastic rocks in the cores of crustal-scale faults: Examples from the Median Tectonic Line, Japan, J Geophys Res, Doi:10.1029/2005JB004205. Kawamoto, E., and T. Shimamoto (1998), The strength profile for bimineralic shear zones: an insight from high-temperature shearing experiments on calcite-halite mixtures, Tectonophysics, 295(1-2), 1-14, Doi:10.1016/S0040-1951(98)00112-7. Kennedy, L., and J. Logan (1997), The role of veining and dissolution in the evolution of fine-grained mylonites: the McConnell thrust, Alberta, Journal of Structural Geology, 19(6), 785-797, Doi:10.1016/S0191-8141(97)00005-9. 126 Kennedy, L., and J. Logan (1998), Microstructures of cataclasites in a limestone-on-shale thrust fault: implications for low-temperature recrystallization of calcite, Tectonophysics, 295(1-2), 167, Doi:10.1016/S0040-1951(98)00119-X. Knipe, R. (1981), The interaction of deformation and metamorphism in slates, Tectonophysics, 78(1-4), Doi:10.1016/0040-1951(81)90016-0. Knott, S. (1994), Fault zone thickness versus displacement in the Permo-Triassic sandstones of NW England, Journal of the Geological Society, 151, 17, Doi:10.1144/gsjgs.151.1.0017. Lockner, D., H. Naka, H. Tanaka, and H. Ito (2000), Permeability and strength of core samples from the Nojima fault of the 1995 Kobe earthquake, in Proceedings of the International Workshop on the Nojima Fault Core and Borehole Analysis, Open-file Report edited by U. S. G. Survey, pp. 00-129. Logan, J. (2007), The progression from damage to localization of displacement observed in laboratory testing of porous rocks, Geological Society London Special Publications, 289, Doi:doi:10.1144/SP289.5. Logan, J., and K. Rauenzahn (1987), Frictional dependence of gouge mixtures of quartz and montmorillonite on velocity, composition and fabric, Tectonophysics, 144(1-3), 87-108, Doi:10.1016/0040-1951(87)90010-2. Logan, J., N. Higgs, and M. Friedman (1981), Laboratory studies on natural gouge from the US Geological Survey Dry Lake Valley No. 1 well, San Andreas fault zone, in Mechanical behavior of crustal rocks: Am. Geophys. Union Geophys. Mon. , edited by N. Carter and e. al., pp. 121-134 Logan, J., C. Dengo, N. Higgs, and Z. Wang (1992), Fabrics of experimental fault zones: Their development and relationship to mechanical behavior, in Fault Mechanics and 127 Transport Properties of Rocks, International Geophysics Series edited by B. Evans and T. F. Wong, pp. 33-67, Academic Press Ltd London. Logan, J., M. Friedman, N. Higgs, C. Dengo, and T. Shimamoto (1979), Experimental studies of simulated gouge and their application to studies of natural fault zones, in Conference VIII: Analysis of Actual Fault Zones in Bedrock, edited, Natl. Earthquake Hazards Reduction Program, Menlo Park, CA. Lubert, M., and A. de Ryck (2001), Slip events and dilatancy in a sheared fine noncohesive powder, Physical Review E, 63(2), 021502, Doi:10.1103/PhysRevE.63.021502. Lupini, J., A. Skinner, and P. Vaughan (1981), The drained residual strength of cohesive soils, Geotechnique, 31, 181–213, Doi:10.1680/geot.1981.31.2.181. Mair, K., and C. Marone (1999), Friction of simulated fault gouge for a wide range of velocities and normal stresses, Journal Geophysical Research, 104(28), 914, Doi:doi:10.1029/1999JB900279. Mares, V., and A. Kronenberg (1993), Experimental deformation of muscovite, Journal of Structural Geology, 15(9-10), 1061, Doi:10.1016/0191-8141(93)90156-5. Mariani, E., K. Brodie, and E. Rutter (2006), Experimental deformation of muscovite shear zones at high temperatures under hydrothermal conditions and the strength of phyllosilicate- bearing faults in nature, Journal of Structural Geology, 28(1569), Doi:10.1016/j.jsg.2006.06.009. Marone, C. (1998), Laboratory-derived friction laws and their application to seismic faulting, Annual Review of Earth and Planetary Sciences, 26(1), 643-696, Doi:10.1146/annurev.earth.26.1.643. 128 Marone, C., C. Raleigh, and C. Scholz (1990), Frictional behavior and constitutive modeling of simulated fault gouge, in J Geophys Res, edited, pp. 7007–7025. Meike, A. (1990), A micromechanical perspective on the role of dislocations in selective dissolution, Geochimica et Cosmochimica Acta, 54(12), 3347-3352, Doi:10.1016/0016- 7037(90)90289-w. Mitchell, T. (2007), The fluid flow properties of fault damage zones, Ph.D. thesis, 1-214 pp, University of Liverpool, Liverpool. Mitchell, T. M., and D. Faulkner (2008), Experimental measurements of permeability evolution during triaxial compression of initially intact crystalline rocks and implications for fluid flow in fault zones, J Geophys Res, 113(B11), Doi:10.1029/2008JB005588. Mizoguchi, K., M. Takahashi, W. Tanikawa, K. Masuda, S. Song, and W. Soh (2008), Frictional strength of fault gouge in Taiwan Chelungpu fault obtained from TCDP Hole B, Tectonophysics, 460(1-4), 198-205, Doi:10.1016/j.tecto.2008.08.009. Molli, G., J. C. White, L. Kennedy, and V. Taini (2011), Low-temperature deformation of limestone, Isola Palmaria, northern Apennine, Italy - The role of primary textures, precursory veins and intracrystalline deformation in localization, Journal of Structural Geology, 33(3), 255-270, Doi:10.1016/j.jsg.2010.11.015. Moore, D., R. Summers, and J. Byerlee (1986), The effects of sliding velocity on the frictional and physical properties of heated fault gouge, Pure and applied geophysics, Doi:10.1007/BF00875718. Moore, D., R. Summers, and J. Byerlee (1988), Relationship between textures and sliding motion of experimentally deformed fault gouge: Application to fault zone behavior, paper 129 presented at The 29th U.S. Symposium on Rock Mechanics (USRMS), Minneapolis, MN, June 13 - 15. Morrow, B. Radney, and J. Byerlee (1992), Frictional strength and the effective pressure law of montmorillonite and illite clays, in Fault Mechanics and Transport Properties of Rocks, edited by B. Evans and T. F. Wong, pp. 69-88, Academic Press Ltd, London. Morrow, C., D. Moore, and D. Lockner (2000), The effect of mineral bond strength and adsorbed water on fault gouge frictional strength, Geophysical Research Letters, 27(6), 815– 818, Doi:10.1029/1999GL008401. Morrow, C., J. Solum, S. Tembe, D. Lockner, and T. Wong (2007), Using drill cutting separates to estimate the strength of narrow shear zones at SAFOD, Geophys. Res. Lett., 34, L11301, Doi:10.1029/2007GL029665. Nakatani, M. (2001), Conceptual and physical clarification of rate and state friction: Frictional sliding as a thermally activated rheology, Journal Geophysical Research, 106(B7), 13347-13380, Doi:10.1029/2000jb900453. Nemcok, M., Schamel, and R. Gayer (2005), Thrustbelts: Structural Architecture, Thermal Regimes and Petroleum Systems, 541 pp., Cambridge University Press, New York. Neuzil, C. (1995), Abnormal pressures as hydrodynamic phenomena, American Journal of Science, 295(6), 742, Doi:10.2475/ajs.295.6.742. Niemeijer, A., and C. Spiers (2005), Influence of phyllosilicates on fault strength in the brittle-ductile transition: insights from rock analogue experiments, in High-Strain Zones: Structure and Physical Properties, edited by D. Bruhn and L. Burlini, pp. 303-327, Geological Society, London. 130 Ohtani, T., K. Fujimoto, H. Ito, H. Tanaka, N. Tomida, and T. Higuchi (2000), Fault rocks and past to recent fluid characteristics from the borehole survey of the Nojima fault ruptured in the 1995 Kobe earthquake, southwest Japan, J Geophys Res, 105(B7), 16,161-116,171, Doi:10.1029/2000JB900086 Olsson, W. A. (1974), Effects of temperature, pressure and displacement rate on the frictional characteristics of a limestone, International Journal of Rock Mechanics and Mining Science & Geomechanics Abstracts, 11(7), 267-278, Doi:10.1016/0148-9062(74)90228-9. Passchier, C., and R. Trouw (2005), Microtectonics, 366 pp., Springer, Berlin Heidelberg. Paterson, M. S. (1978), Experimental Rock Deformation- The Brittle Field, 254 pp., Springer-Verlag Berlin. Paterson, M. S., and T. Wong (2005), Experimental Rock Deformation – The Brittle Field, 2nd ed., 348 pp., Springer-Verlag, Berlin, Heidelberg, New York. Peyaud, J., M. Pagel, J. Cabrera, and H. Pitsch (2006), Mineralogical, chemical and isotopic perturbations induced in shale by fluid circulation in a fault at the Tournemire experimental site (Aveyron, France), Journal of Geochemical Exploration, 90(1-2), 9-23, Doi:10.1016/j.gexplo.2005.09.001 Poirier, J. (1985), Creep of Crystals: High temperature deformation processed in metals, ceramics and minerals, Cambridge University Press, New York. Reid, H. F. (1911), The Elastic-Rebound Theory of Earthquakes, Bulletin of the Department of Geology, University of California Publications, Bulletin of the Department of Geology, University of California Publications, 6(19), 413-444, 131 Rice, J., N. Lapusta, and K. Ranjith (2001), Rate and state dependent friction and the stability of sliding between elastically deformable solids, Journal of the Mechanics and Physics of Solids, 49(9), 1865-1898 Doi:10.1016/S0022-5096(01)00042-4. Rietveld, H. M. (1967), Line profiles of neutron powder-diffraction peaks for structure refinement, Acta Cryst, 22, 151-152, Doi:10.1107/S0365110X67000234. Rietveld, H. M. (1969), A Profile Refinement Method for Nuclear and Magnetic Structures, J. Appl. Cryst., 2, 65-71, Doi:10.1107/S0021889869006558. Robin, P.-Y. F. (1978), Pressure solution at grain-to-grain contacts, Geochimica et Cosmochimica Acta, 42(9), 1383-1389, Doi:10.1016/0016-7037(78)90043-1. Ruina, A. (1983), Slip Instability and State Variable Friction Laws J Geophys Res, 88(B12), 10359-10370, Doi:10.1029/JB088iB12p10359. Rutter, E. (1983), Pressure solution in nature, theory and experiment, Journal of Geological Society, 140(5), 725, Doi:10.1144/gsjgs.140.5.0725. Rutter, E. (1986), On the nomenclature of mode of failure transitions in rocks, Tectonophysics, 122(3-4), 381, Doi:10.1016/0040-1951(86)90153-8. Rutter, E., R. Maddock, S. Hall, and S. White (1986), Comparative microstructures of natural and experimentally produced clay-bearing fault gouges, Pure and applied geophysics, 124(1), 3-30, Doi:10.1007/BF00875717. Saffer, D., and C. Marone (2003), Comparison of smectite-and illite-rich gouge frictional properties: application to the updip limit of the seismogenic zone along subduction megathrusts, in Earth and Planetary Science Letters, edited, pp. 219-235. 132 Schmid, S., R. Panozzo, and S. Bauer (1987), Simple shear experiments on calcite rocks: rheology and microfabric, Journal of Structural Geology, 9(5-6), 747, Doi:10.1016/0191- 8141(87)90157-X. Schmid, S. M., M. S. Paterson, and J. N. Boland (1980), High temperature flow and dynamic recrystallization in carrara marble, Tectonophysics, 65(3-4), 245-280, Doi:10.1016/0040- 1951(80)90077-3. Scholz, C. (1987), Wear and gouge formation in brittle faulting, Geology, 15(6), Doi:10.1130/0091-7613(1987)15<493:WAGFIB>2.0.CO;2. Scholz, C. (1998), Earthquakes and friction laws, Nature, 391(6662), 37-42, Scruggs, V., and T. Tullis (1998), Correlation between velocity dependence of friction and strain localization in large displacement experiments on feldspar, muscovite and biotite gouge, Tectonophysics, 295(1-2), 15, Doi:10.1016/S0040-1951(98)00113-9. Segall, P., and J. Rice (1995), Dilatancy, compaction, and slip instability of a fluid-infiltrated fault, J Geophys Res, 100(B11), 22155-22171, Doi:10.1029/95JB02403. Shimamoto, T. (1977), Effects of Fault Gouge on the Frictional Properties of Rocks: An experimental study, Ph.D. thesis, 1-215 pp, Texas A&M, College Station. Shimamoto, T. (1985), Confining pressure reduction experiments: A new method for measuring frictional strength over a wide range of normal stress, International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, Doi:10.1016/0148- 9062(85)92950-X. 133 Shimamoto, T., and J. Logan (1981), Effects of simulated fault gouge on the sliding behavior of Tennessee sandstone: nonclay gouges, J Geophys Res, 86(B4), 2902, Doi:10.1029/JB086iB04p02902. Shipton, Z., A. Soden, J. Kirkpatrick, and A. Bright (2006), How thick is a fault? Fault displacement-thickness scaling revisited, AGU. Sibson, R. (2003), Brittle-failure controls on maximum sustainable overpressure in different tectonic regimes, AAPG Bulletin, 87(6), Doi:10.1306/01290300181. Sibson, R. H. (1977), Fault rocks and fault mechanisms, Journal of the Geological Society, 133(3), 191-213, Doi:10.1144/gsjgs.133.3.0191. Skempton, A. (1966), Some observations of tectonic shear zones, paper presented at Proc. 1st Int. Conf. Rock Mech. Smart, K. J., R. D. Krieg, and W. M. Dunne (1999), Deformation behavior during blind thrust translation as a function of fault strength, Journal of Structural Geology, 21(7), 855- 874, Doi:10.1016/s0191-8141(99)00050-4. Smith, S., and D. Faulkner (2010), Laboratory measurements of the frictional properties of the Zuccale low-angle normal fault, Elba Island, Italy, J Geophys Res, 115(B2), B02407, Doi:10.1029/2008JB006274. Solum, J., S. Hickman, and D. Lockner (2006), Mineralogical characterization of protolith and fault rocks from the SAFOD main hole, Geophysical Reseach Letters, 33(L21314), Doi:10.1029/2006GL027285. 134 Stesky, R. M., W. F. Brace, D. K. Riley, and P. Y. F. Robin (1974), Friction in faulted rock at high temperature and pressure, Tectonophysics, 23(1-2), 177-203, Doi:10.1016/0040- 1951(74)90119-x. Tembe, S., D. Lockner, and T. Wong (2010), Effect of clay content and mineralogy on frictional sliding behavior of simulated gouges: Binary and ternary mixtures of quartz, illite, and montmorillonite, Journal Geophysical Research, 115(B3), B03416, Doi:10.1029/2009JB006383. Tembe, S., D. Lockner, J. Solum, C. Morrow, T. Wong, and D. Moore (2006), Frictional strength of cuttings and core from SAFOD drillhole phases 1 and 2, Geophysical Reseach Letters, 33(23), Doi:10.1029/2006GL027626. Tsutsumi, A., S. Nishino, K. Mizoguchi, and T. Hirose (2004), Principal fault zone width and permeability of the active Neodani fault, Nobi fault system, Southwest Japan, Tectonophysics, 379(1-4), Doi:10.1016/j.tecto.2003.10.007. Tullis, T. E., et al. (2007), Rheology of Fault Rocks and Their Surroundings in 95th Dahlem Workshop, edited by M. R. Handy, pp. 183-203, The MIT Press, Berlin, Germany. Twiss, R., and E. Moores (1992), Structural Geology, 532 pp., W. H. Freeman and Company, New York. van Diggelen, E., H. de Bresser, and C. Peach (2009), High shear strain behaviour of synthetic muscovite fault gouges under hydrothermal conditions, Journal of Structural Geology, 32(11), 1685, Doi:10.1016/j.jsg.2009.08.020. Vrolijk, P., and B. van der Pluijm (1999), Clay gouge, Journal of Structural Geology, 21(8- 9), 1039-1048, Doi:10.1016/S0191-8141(99)00103-0. 135 Wells, R., J. Newman, and S. Wojtal (2010), Microstructure and rheology of limestone-shale fault rocks, paper presented at 2010 Fall Meeting, AGU, San Francisco, Calif., Dec 13-17 Wheeler, J. (1987), The significance of grain-scale stresses in the kinetics of metamorphism, Contributions to Mineralogy and Petrology, 97(3), 397-404, Doi:10.1007/bf00372002. Wibberley, C., G. Yielding, and G. Di Toro (2008), Recent advances in the understanding of fault zone internal structure: a review, Geological Society London Special Publications, 299(1), 5, Doi:10.1144/SP299.2. Wojtal, S., and G. Mitra (1986), Strain hardening and strain softening in fault zones from foreland thrusts, Geological Society of America Bulletin, 97(6), 674, Doi:10.1130/0016- 7606(1986)97<674:SHASSI>2.0.CO;2. 136 Appendices 137 Appendix A Grain Size Distribution Laser grain size analysis was conducted using a Mastersizer 2000 Laser Grain Size Analyzer at the University of British Columbia. The grain size distribution for each endmember gouge is an average of ten runs and is presented as both volume percent and number percent. The grain size distribution for composite gouges is calculated from endmember gouge grain size distributions. 138 Figure A-1: Volume Percent 0 1 2 3 4 5 6 Vo lu m e Pe rc en t Liverpool Shale UBC Shale Carbonate 75% Shale, UBC 75% Shale, Liverpool 50% Shale, UBC 50% Shale, Liverpool 25% Shale, UBC 25% Shale, Liverpool 0 1 2 3 4 5 4.5 3.5 2.5 1.5 0.5 103 Grain Size (μm) 10210110-1 100 a b Figure A-1: Grain size vs. volume percent. a. Volume percent for each endmember gouge. b. Calculated volume percent for composite gouges. Dotted lines show UBC composites. Solid lines show U of L composites. 139 Figure A-2: Cumulative Volume Percent Cu m ul at iv e Vo lu m e Pe rc en t Liverpool Shale UBC Shale Carbonate Grain Size(µm) 10 -3 10 0 10 1 10 2 10 3 0 10 20 30 40 50 60 70 80 90 100 Figure A-2: Cumulative grain size distributions for endmember gouge compositions. 140 Table A-1: Grain Size Distribution Percent UBC Shale Carbonate Liverpool Shale 5% ≤ 1.9 µm 1.3 µm 1.9 µm 25% ≤ 5.7 µm 2.5 µm 8.7 µm 50% ≤ 17.4 µm 5.0 µm 34.7 µm 75% ≤ 52.48 µm 8.7 µm 120.2 µm 95 ≤ 120.2 µm 17.4 µm 239.9 µm 50% between 5.7 - 52.5 µm 2.5 - 8.7 µm 8.7 - 120.2 µm min 0.83 µm .72 µm .83 µm V o lu m e % max 239.9 µm 8.7 µm 478.6 µm N u m b e r % 99% ≤ 6.61 µm 5.0 µm 6.61 µm 141 Appendix B XRD Analysis X-Ray Diffractometry was conducted using the Rietveld Method. The following section includes output XRD spectra - including mineral percentages – for each endmember composition. Composite gouge compositions were calculated from measurements on endmembers. 142 Figure B-1: Shale - UBC Illite: 39.23 % Quartz: 31.64 % Clinochlore II: 18.29 % Microcline: 5.31 % Albite: 5.70 % 143 Figure B-2: Carbonate Calcite: 70.19% Dolomite: 18.85% Fluorite: 10% Amorphous: .96% 144 Figure B-3: Shale - Liverpool 145 Table B-1: Endmember Composition % of mineral in gouge Mineral UBC 100% Shale 100% Carbonate U of L 100% Shale quartz 32 0 32 illite 40 0 40 clinochlore 18 0 17 feldspar 11 0 10 calcite 0 80 0 dolomite 0 20 0 % Phyll 58 0 57 % Qtz 32 0 32 % Carbonate 0 100 0 Normalized % Phyll 64% 0% 64% % Qtz 36% 0% 36% % Carbonate 0% 100% 0% Table B-1: Endmember mineral percentages determined from Rietveld analysis. (Phyll = clinochlore + illite; Carbonate = calcite + dol) 146 Table B-2: UBC Composite Composition % of mineral in gouge Mineral 100% Shale 75% Shale 50% Shale 25% Shale 100% Carbonate quartz 32 24 16 8 0 illite 40 30 20 10 0 clinochlore 18 13.5 9 4.5 0 feldspar 11 8.25 5.5 2.75 0 calcite 0 20 40 60 80 dolomite 0 5 10 15 20 % Phyll 58 43.5 29 14.5 0 % Qtz 32 24 16 8 0 % Carbonate 0 25 50 75 100 Normalized % Phyll 64% 47% 31% 15% 0% % Qtz 36% 26% 17% 8% 0% % Carbonate 0% 27% 53% 77% 100% Table B-2: UBC composite gouge compositions calculated from endmember gouge compositions (Phyll = clinochlore + illite; Carbonate = calcite + dolomite) 147 Table B-3: Liverpool Composite Compositions % of mineral in gouge Mineral 100% Shale 75% Shale 50% Shale 25% Shale 100% Carbonate quartz 32 24 16 8 0 illite 40 30 20 10 0 clinochlore 17 12.75 8.5 4.25 0 feldspar 10 7.5 5 2.5 0 calcite 0 20 40 60 80 dolomite 0 5 10 15 20 % Phyll 57 42.75 28.5 14.25 0 % Qtz 32 24 16 8 0 % Carbonate 0 25 50 75 100 Normalized % Phyll 64% 47% 30% 15% 0% % Qtz 36% 26% 17% 8% 0% % Carbonate 0% 27% 53% 77% 100% Table B-3: UBC composite gouge compositions calculated from endmember gouge compositions (Phyll = clinochlore + illite; Carbonate = calcite + dolomite) 148 Figure B-4: Ternary Diagram Figure B-4: Ternary diagram showing normalized gouge compositions and the corresponding coefficient of friction values for room T, saturated experiments performed at the University of Liverpool 149 Appendix C Data Acquisition For all deformation experiments performed at UBC, load (lbs) and displacement (lbs) are measured using an external load cell and external displacement transducer (DCDT). Output data is recorded using National Instruments Labview™ 8.0 Software. For all deformation experiments performed at U of L, load (kgs) is measured using an external force gauge LVDT, and displacement is measured by motor revolutions per second. All experimental data files are compiled on a CD at the back of the thesis. Included in the following section are examples of output data from UBC and U of L. 150 Table C-1: Raw UBC Output File 151 Table C-2: Raw Liverpool Output Data File 152 Appendix D Data Processing The following section provides the calculations and MATLAB code used to process raw experimental data files. 153 Calculations This section outlines the calculations used to convert the data output by the rig to the coefficient of friction and displacement values used for analysis. 154 Coefficient of Friction: Flbs = applied force (i.e. axial load in lbs ft s-2) Fkg = applied force (i.e. axial load in kg m s-2) r = radius of core (in m) A = surface area of the top of the core (in m2) σdiff (kg) = differential stress (in kg m-1 s-2) σdiff = differential stress (in MPa) Pc = confining pressure (in MPa) ! Fkg = Flbs * .45359237 ! r m = r mm * .001 A = \"r2( ) ! \" diff kg m # s2 $ % & ' ( ) = Fkg A \" diff MPa( ) =\" diff kg m # s2 $ % & ' ( ) * .00000980665 ! \" 3 = Pc \" 1 =\" diff +\" 3 155 Correction for the change in surface area due to displacement along saw cut : [He et al., 2006]* = angle between core axis and saw cut r = radius of core σ1app = applied axial stress not corrected for the change in surface area during sliding Convert σ1 and σ3 to normal stress (σn) and shear stress (σs) (2 Dimensional Mohr Circle): θ = angle between axial load (σ1) and the pole to the saw cut surface σ1app = applied axial stress corrected for change in surface area Coefficient of Friction (µ) *
see
page
158
to
calculate
the
change
in
surface
area
during
sliding
 ! \" 1corr = \" 1app 1# 2 $ arcsin %L tan& 2r ' ( ) * + , + %L tan& 2r 1# %L tan& 2r ' ( ) * + , 2' ( ) ) * + , , ! \" ! \" S = \" 1 corr #\" 3 2 sin 2$( ) ! \" n = \" 1 corr +\" 2 2 + \" 1 corr #\" 2 2 cos(2$) ! µ = \" s \" n 156 Displacement: dmm = din * 25.4 din = axial displacement (in inches) dmm = axial displacement (in mm) kkg/cm = rig stiffness (in kg/cm)* kkg/mm = rig stiffness (in kg/mm) dR = deformation in the rig (in mm) dcorr = axial displacement corrected for deformation in the rig (in mm) * see page 173 for details of LSR stiffness calibration 
 ! kkg /mm = kkg / cm 10 dR = Fkg kkg /mm dcorr = dmm \" dR 157 Shear Strain (γ): d = displ (mm) Φ = angle of saw cut to axial load TT = precompaction gouge zone thickness (mm) ! \" = d cos(#) $ % & ' ( ) TT 158 Correction for Change in Surface Area During Sliding The following pages show the derivation of the equation for the correction for the change in surface area during sliding found (courtesy of Changrong He). In this equation the axial displacement is resolved on the saw cut, and area of contact during sliding is determined by the intersection of two offset ellipses. 159 160 Matlab Code The following pages provide the MATLAB code used to process the raw experimental data files. Commented sections of the code (preceded by the percent symbol %) provide explanations for the data processing steps. Only the code used to process data from the U of L is shown in this section. Additional code is on the supplemental CD. 161 Liverpool
Data
Processing
Program
 %% Liverpool Rig Data Processing %{ Code reads in Experimental datasets from prepared output RAW files to: Convert data to SI units Extract data that corresponds to experiment Calculate the stress from load Calculate the displacement from motor revolutions Calculate coefficient of friction Code must be accompanied by the following function files: datapoint_select.m frictional_slide_Liv.m, File names follow the following format: 2 initials, 2 digit sample number, 4 character composition, description of expt i.e.: JH55_SH100_roomT %} clear all S = cd; %import .txt file display('Select Experiment') [file_name,dir_name] = uigetfile('*.*','Enter data file'); full_name = fullfile(dir_name, file_name); uiimport([full_name]); file_short = file_name; file = full_name; display( 'Hit Return after File Selection - twice if necessary') pause exp_data = data; datetime = textdata; l_mm = 50; 162 d_mm = 20; pe_ideal= 70; %Pc from experiment is imported below. % Select type of experiment performed (fracture or frictional sliding) p = input('Which type of experiment did you perform? \\n Press: 1 for fracture experiment '); if isempty(p) p = 2; end if p == 2 phi = input('What is the angle between the fracture plane \\n and the axial load (in degrees)\\n'); if isempty(phi) phi = 30; end end display('Please wait while Matlab plots your data. This make take some time.'); strFileSpec = [dir_name 'fig' file_short ]; tf = isdir(strFileSpec); if tf == 0; mkdir(strFileSpec); end cd([dir_name 'fig' file_short]); %% Assign Variable names to observed data no_obs = length(exp_data); Expt_secs = exp_data(:,1); cpress_mpa = exp_data(:,13); loadf_kN = exp_data(:,15); loadf_N = loadf_kN * 1000; loadf_kg = loadf_N/9.80665; % convert to loadf_kg to make compatible with frictional slide function poreTP_mpa = exp_data(:,11); poreBT_mpa = exp_data(:,12); 163 Pe = round(mean(cpress_mpa))- round(mean(poreTP_mpa)) temp_C = exp_data(:,16); % Preliminary Conversions to useful units encoder_counts = exp_data(:,10); raw_displ_mm = encoder_counts/35322400; core_area_mm= pi*(0.5*d_mm)*(0.5*d_mm); % Stiffness (rig room T) Pc_ideal = round(mean(cpress_mpa)); distortion = 3e-05*(Pc_ideal)^2 - 0.0128*(Pc_ideal) + 8.4785; % this is room T %temperature value from new sample assembly calibrations by sergio llana funez -units are microns/kN %convert from um/kN to cm/kgf distortion_cm_kgf = 9.8067e-07 * distortion; %stiffness in kgf/cm (need for constistency with LSR processing program) stiffness = 1/distortion_cm_kgf; %% Processing Frictional Sliding Data if p == 2 figure(01) plot(raw_displ_mm, loadf_kg), xlabel('Displacement (mm)'), ylabel ('Load (kg)') name = [ file_short '_frac_raw.eps']; print( '-depsc2', name); title('Choose the start and end of experiment'); %choose the start and end of the experiment graph = figure(01); [target, position, data_index] = datapoint_select(graph, 2); a = data_index(1); b = data_index(2); 164 clear target; clear position; clear data_index; % calculates coefficient of friction using frictional_slide_Liv function [temp_out_slide, temp_out2_slide] = frictional_slide_Liv (raw_displ_mm, ... file_short, stiffness, Expt_secs, loadf_kg, Pe, d_mm, S, phi, core_area_mm, a, b); end 165 Datapoint
Select
Function
 function[target, position, data_index] = datapoint_select(graph, j ) % graph = figure(1); % j = number of datapoints you would like to select % target = handle of graphics object; position = x, y (,z) % coordinates; % data_index = index of datapoint % function allows user to select a specified number of data % points (j) from a figure (graph) % outputs position, target and data indices for each % selection for i = 1:j %specify number of points to select datacursormode on k = waitforbuttonpress; for k = 0; dcm_obj = datacursormode(figure(1)); info_struc = getCursorInfo(dcm_obj); target(i) = info_struc.Target; position(i,:) = info_struc.Position; data_index(i) = info_struc.DataIndex; end datacursormode off delete(findall(figure(1),'Draggable', 'on')) end 166 Liverpool
Frictional
Slide
Function
 %% Selecting Start and End of Experiment function [temp_out, temp_out2, strFileSpec] = frictional_slide_Liv (raw_displ_mm, ... file_short, stiffness, Expt_secs, loadf_kg, pe, d_mm, S, phi,core_area_mm, a, b) %% %convert units kg_force_mm2= loadf_kg/core_area_mm; stress_mpa = 9.80665*kg_force_mm2; % Correction for Rig Stiffness % Extract and re-zero experimental data stiffness_mm = stiffness/10; standard = 1/stiffness_mm; %Shimamoto 1977 phd thesis displ_mm = raw_displ_mm - (loadf_kg*standard);%shimamoto 1977 %thesis displ_mm2 = displ_mm(a:b); displ_mm = displ_mm2- displ_mm2(1); stress_mpa2 = stress_mpa(a:b); stress_mpa = stress_mpa2- stress_mpa2(1); Expt_secs2 = Expt_secs(a:b); Expt_secs = Expt_secs2 - Expt_secs2(1); figure(1) plot (displ_mm, stress_mpa,'k*') axis([0 max(displ_mm + .5) 0 max(stress_mpa + 15)]) title(file_short); name = [ file_short 'slide_disp_dstrs.eps']; saveas (1, name, 'eps'); close (1) %% Computes Sigma 1 (Max Principle Stress) sigma1 = stress_mpa + pe; 167 sigma3 = pe; %Accounts for change in sigma 1 due to change in sliding surface area %with displacement % uses phi input in Liverpool_Processing.m %equation from He et al 2006 x = (displ_mm.*(tand(phi)))/(2*(.5 * d_mm)); sigma1_C = sigma1./(1-(2/pi*(asin(x) + (x).*sqrt(1- ((x).^2))))); %% Computes Shear and Normal Stress and Coefficient of Friction % = angle between normal to plane and principle stress theta = 90 - phi; %Calculate shear and normal stresses using 2D Mohr equations sigma_s = ((((sigma1_C-sigma3)/2)* sind(2*theta)))-20; sigma_n = ((sigma1_C + sigma3)/2) + ((sigma1_C - sigma3)/2)*(cosd(2*theta)); coef_f = sigma_s./sigma_n; %% Calculate Shear Strain TT = .1; % TT = thickness of shear zone in mm shear_strain = displ_mm/TT; figure (5) title(file_short); plot (displ_mm, coef_f), xlabel('Displacement (mm)'), ylabel ('Coefficient of Friction') %axis([0 max(displ_mm + .5) 0 max(coef_f + .1)]) name = [ file_short 'slide_disp_mew.eps']; saveas (5, name, 'eps'); %% Save Data Files no_obs = length (Expt_secs); sigma3 = linspace(pe,pe, no_obs)'; temp_out = [Expt_secs, displ_mm, stress_mpa, sigma1_C, 168 sigma3, coef_f, sigma_n, sigma_s, sigma1, shear_strain]; temp_out2 = {'Time (s)'; 'Displacement (mm)'; 'Stress (MPa)'; texlabel('sigma_1'); texlabel('sigma_3'); 'Coefficient of Friction'; texlabel('Normal Stress, sigma_n (MPa)'); texlabel('Shear Stress, sigma_s (MPa)'); texlabel('sigma_1 _a_p_p_a_r_e_n_t'); 'Shear Strain'}; % saves with all graphs in same folder as raw data file strFileSpec = [ file_short 'slide_outnames']; save( strFileSpec, 'temp_out2' ) strFileSpec = [ file_short 'slide_out']; save( strFileSpec, 'temp_out', '-ascii' ) % Saves to processed data folder in programs file strFileSpec = [ S '/processed_data/' ]; tf = isdir(strFileSpec); if tf == 0; mkdir(strFileSpec); end strFileSpec = [ S '/processed_data/' file_short '_out']; save( strFileSpec, 'temp_out', '-ascii' ) 169 Data
Plotting
Program
 %% Data plotting program %{ - requires expt_names.txt data file to create legend - program imports multiple, user selected -ascii matrices (filename_out files) - plots filename_out files created in data processing program - gives an option to save with user input filename - axes default to shear strain and coefficient of friction %} %% clear all S = cd; %imports expt names file to be used in legend % expt_names2.txt file is set up as such: % experiment number = row number % cell contains description of experiment % i.e. cell55: '100% Shale, Room T, Liverpool' file = [S '/expt_names2.txt']; expt_names = importdata(file, ';', 85); A = []; B = []; N = []; K = 1; x = input('What would you like your x axis to be? Press: \\n 1 for time \\n 2 for Displacement \\n 3 for Differential Stress \\n 4 for Sigma 1 \\n 5 for Sigma 3 \\n 6 for Coefficient of Friction \\n 7 for Normal Stress \\n 8 for Shear Stress \\n 9 for Shear Strain \\n' ); y = input('What would you like your y axis to be? Press: \\n 1 for time \\n 2 for Displacement \\n 3 for Differential Stress \\n 4 for Sigma 1 \\n 5 for Sigma 3 \\n 6 for Coefficient of Friction \\n 7 for Normal Stress \\n 8 for Shear Stress \\n 9 for Shear Strain \\n' ); labels = {'Time (s)'; 'Displacement (mm)'; 'Stress (MPa)'; texlabel('sigma_1'); texlabel('sigma_3'); 'Coefficient of Friction'; texlabel('Normal Stress, sigma_n (MPa)'); 170 texlabel('Shear Stress, sigma_s (MPa)'); texlabel('sigma_1 _a_p_p_a_r_e_n_t'); 'Shear Strain'}; % Program defaults to Displ_mm vs Coef. of Friction if isempty(x) varx = 2; else varx = x; end if isempty(y) vary = 6; else vary = y; end display('Please Select Your First File') while K == 1 display('Select Experiment') [file_name,dir_name] = uigetfile('*.*','Enter data file'); full_name = fullfile(dir_name, file_name); data = importdata( [full_name]); cd ([dir_name]) display( 'Hit Return after File Selection - twice if necessary') pause D = data; N = ([N 1]); n = sum(N); a = linspace(NaN,NaN, 10000)'; b = linspace(NaN,NaN, 10000)'; A = ([A a]); B = ([B b]); no_obs = length (D(:,1)); A(1:no_obs, n) = D(:,varx); B(1:no_obs, n) = D(:,vary); 171 % finds expt number in filename % so experiment can be read from expt names %file and used for legend num = file_name(3:4); num2 = str2num(num); expt_num(n) = num2; Q = input('Hit enter to read another file or hit a number \\n'); if isempty(Q) K = 1; else K = 0; end end %% x_var = labels(varx); y_var = labels(vary); TITLE = cellstr([x_var 'vs.' y_var]); % finds axes x = max (A); y = max (B); X = max (x); Y = max (y); % uses index number in file name to reference experiment names file for i = 1:n names(i) = [expt_names(expt_num(i))]; end %% figure1 = figure('XVisual',... '0x24 (TrueColor, depth 24, RGB mask 0xff0000 0xff00 0x00ff)'); axes1 = axes('Parent',figure(1),'FontSize',16) plot(A,B, 'LineWidth', 2.5, 'Parent', axes1) axis([0 (X + 1) 0 (Y + .1)]) 172 % Create xlabel xlabel(x_var,'FontWeight','bold','FontSize',16); % Create ylabel ylabel(y_var,'FontWeight','bold','FontSize',16); % Create legend legend(names, 4, 'FontSize', 16) answer = input('Please type your filename \\n', 's'); if isempty(answer) saved = 'no' else strFileSpec = [ S '/processed_data/graphs' ]; tf = isdir(strFileSpec); % Check to see if file already exists if tf == 0; % if it does not, creates a new file for output data mkdir(strFileSpec); end print( '-depsc2', [strFileSpec '/' answer]); % use to export data to create color eps file openable in illustrator % using user input % 'answer' as filename saveas(1, [strFileSpec '/' answer], 'fig'); % save as .fig for future use in matlab end cd (S); 173 LSR Stiffness Calibration Procedure Because displacement is measured externally, the displacement transducer records deformation within the sample as well as elastic deformation within the loading column. In order to determine the amount of deformation within the sample, the amount of elastic deformation within the loading column needs to be corrected for during data processing. Stiffness (k) is a measurement of the amount of resistance to elastic deformation and is a product of both the applied force and the confining pressure. And is calculated using the relationship between the applied force and the resulting deformation (Equation D-1). k = Δforce/Δdisplacement Equation D-1 The stiffness of the entire system (loading column plus sample) can be determined by dividing the applied axial load (in kg-force) by the amount of deformation (in cm) measured by the displacement transducer. The rig outputs force in lbs-force and displacement in inches so these units need to be converted. As of August 19, 2010 the Center for Experimental Studies of the Lithosphere (CESL) at UBC has 1 steel rod and 1 aluminum rod to be used for calibrations (see attached calibration reports). The stiffness of the calibration rods can be determined using the reported Young’s Modulus (E). The steel calibration bar has been tested for an applied load of 100,000 lbs (pound-force) while the aluminum calibrations bar has been tested up to 10,000 lbs of applied load. The Young’s modulus is the ratio of applied stress to the resulting strain and is measured in units of pressure. The Young’s Modulus, unlike the stiffness is an intrinsic 174 material property that does not depend on the size of the sample. The stiffness of the calibration rods can be determined using the following equation: krod = EA/L Equation D-2 where E is the Young’s modulus, A is the area of the end of the calibration rod, and L is the length of the calibration rod. The stiffness of the LSR loading column can be determined by deforming right cylinders with known physical properties at a variety of confining pressures. The difference between the stiffness calculated for the rig plus the sample and the stiffness expected based on the physical properties of the right cylinders is the stiffness of the loading column.* ksystem = krig + krod Equation D-3 Table 1 shows the minimum number of experiments necessary to calibrate the LSR.** Both the steel and aluminum calibration rods are used because they have different material properties. Combined, they allow stiffness to be calculated more accurately at both the high and low ends of the range of load and confining pressures than if either material was used alone. *
Details
about
the
derivation
of
this
relationship
can
be
found
in
Shimamoto,
1977
**
After
Austin,
2003
 175 Table D-1: Calibration Experiment Conditions Material Confining Pressure (MPa) Maximum Load (in lb- force) Aluminum .1 50,000 Steel .1 100,000 Steel 25 100,000 Steel 50 100,000 Steel 75 100,000 Steel 100 100,000 1) Prior to calibration, the length and diameter of each calibration rod should be recorded. 2) Calibration rod should be placed into sample assembly and wrapped in polyolefin heat shrink wrap. Calibration rods may be longer than normal samples; Only use spacers if necessary. 3) Apply specified confining pressure and note the increase load in measured axial force due to the confining pressure. 3) Increase the load at a rate of 40μm/s (see control box calibrations to determine % dial should be on) until the load is 100,000 lbs higher than the load recorded due to confining pressure. Unload and reload the sample 15 times. 176 4) BEFORE CONDUCTING ANOTHER TEST AT A DIFFERENT CONFINING PRESSURE THE CALIBRATION ROD SHOULD BE REMOVED FROM THE RIG, LEFT FOR 30 MINUTES AND REJACKETED to release any stored elastic energy. 177 Calibration
Calculations:

System
Stiffness:

 Kg‐force
=
lb‐force
*
.45359237
(output
rig
data)
Displacement(cm)
=
displacement
(inches)
*
2.54
krig
+
sample
=
Δkg‐force/Δdisplacement

(slope
of
force‐displacement
curve
)

(k
is
in
kg/cm)

 
Calibration
Rod
Stiffness:


 E
(MPa)
=
0.00689475729
*
E(psi)

 Kg_force
=
E(MPa)
*
πr2
(m2)

 Krod
=
kg_force/length
of
rod
(cm)


(k
is
in
kg/cm)

Loading
Column
Stiffness:

 krig = ksystem - krod Once the loading column stiffness is determined at a variety of confining pressures, a stiffness vs. confining pressure plot can be constructed to more accurately calculate stiffness at a variety of confining pressure.

 
 178 179 ________________________________________________________________________________________________ VACS LTD. 15 REGAN ROAD, UNIT 6, BRAMPTON, ONTARIO L7A 1E3 TEL: (905) 840-7651 FAX: (905) 840-9293 TORONTO (Head Office) VANCOUVER EDMONTON QUEBEC www.calibrations.ca CALIBRATION CERTIFICATE No. 51880 Page 2 of 2 Calibration Date: June 29, 2009 None Requested Due Date: Client: UBC Dept. of Earth & Ocean Sciences Device Under Test (DUT): BAR 2 UBC, Model Aluminum, Serial no. NA, Identification no. #2 Sensor Under Test: 12\" X 1.850\" UBC, Model CYLINDRICAL, Serial no.NA, Identification no. #2 Measurement Display: Load vs. Deflection, Lb inch graduations, resolution ____________________________________________________________________________________________________ CALIBRATION PROCEDURE 001-ASTM-E4-08 Standard Practices for Force Verification of Testing Machines MEASUREMENT RESULTS: Tolerance 1.0% of Reading For measurement results associated with the conformance to a tolerance, the uncertainty in the measurement system did not exceed 25% (4:1 test uncertainty ratio) of the acceptable tolerance for each characteristic calibrated, unless otherwise noted. Calibration Standards Used: WL06 Asset , VACS LEVER-CELL, s/n WL06, calibrated 03/26/09 by VACS, Certificate no. 51207, 2.9 lbf UNC REF. LBF DIRECTION: COMPRESSION STD. REF. STD. DUT. # 1 DUT. # 2 Avg Def lection Deflection Young's NO. READINGS Thousandths Thousandths Thousandths per 10k steps Modulas WL06 0 0.000 0.000 0.000 WL06 1000 1.015 0.945 0.980 0.980 4.56E+006 WL06 2000 1.525 1.445 1.485 0.505 6.01E+006 WL06 3000 2.040 1.955 1.998 0.513 6.70E+006 WL06 4000 2.465 2.405 2.435 0.438 7.33E+006 WL06 5000 2.875 2.810 2.843 0.408 7.85E+006 WL06 6000 3.290 3.220 3.255 0.412 8.23E+006 WL06 7000 3.690 3.625 3.658 0.403 8.54E+006 WL06 8000 4.095 4.030 4.063 0.405 8.79E+006 WL06 9000 4.500 4.430 4.465 0.403 9.00E+006 WL06 10000 4.895 4.830 4.863 0.398 9.18E+006 DUT dimensions: Length 12.003\" Diameter 1.850\" 180 181 ________________________________________________________________________________________________ VACS LTD. 15 REGAN ROAD, UNIT 6, BRAMPTON, ONTARIO L7A 1E3 TEL: (905) 840-7651 FAX: (905) 840-9293 TORONTO (Head Office) VANCOUVER EDMONTON QUEBEC www.calibrations.ca CALIBRATION CERTIFICATE No. 51473 Page 2 of 2 Calibration Date: May 27, 2009 None Requested Due Date: Client: UBC Dept. of Earth & Ocean Sciences Device Under Test (DUT): BAR 1 UBC, Model Steel, Serial no. NA, Identification no. #1 Sensor Under Test: 11\" X 1.875\" UBC, Model CYLINDRICAL, Serial no.NA, Identification no. #1 Measurement Display: Load vs. Deflection, Lb inch graduations, resolution ____________________________________________________________________________________________________ CALIBRATION PROCEDURE 001-ASTM-E4-08 Standard Practices for Force Verification of Testing Machines MEASUREMENT RESULTS: Tolerance 1.0% of Reading For measurement results associated with the conformance to a tolerance, the uncertainty in the measurement system did not exceed 25% (4:1 test uncertainty ratio) of the acceptable tolerance for each characteristic calibrated, unless otherwise noted. Calibration Standards Used: WL07 Asset , VACS Calibrator, s/n WL07, calibrated 09/11/08 by VACS, Certificate no. 50035, 17 lbf UNC REF. LBF DIRECTION: COMPRESSION STD. REF. STD. DUT. # 1 DUT. # 2 Avg Def lection Def lection Young's NO. READINGS Thousandths Thousandths Thousandths per 10k steps Modulas WL07 0 0.000 0.000 0.000 WL07 10000 1.925 1.935 1.930 1.930 2.07E+007 WL07 20000 3.145 3.150 3.148 1.218 2.53E+007 WL07 30000 4.395 4.380 4.388 1.240 2.73E+007 WL07 40000 5.645 5.625 5.635 1.248 2.83E+007 WL07 50000 6.870 6.860 6.865 1.230 2.90E+007 WL07 60000 8.085 8.100 8.093 1.228 2.96E+007 WL07 70000 9.320 9.305 9.313 1.220 3.00E+007 WL07 80000 10.510 10.530 10.520 1.208 3.03E+007 WL07 90000 11.750 11.745 11.748 1.228 3.06E+007 WL07 100000 12.970 12.980 12.975 1.228 3.07E+007 DUT dimensions: Length 11.002\" Diameter 1.875\" 182 Appendix E Replicate Experiments The following section provides a graph of replicate room temperature, saturated experiments performed with one porous forcing block at UBC (i.e. Part I). Replicate experiments were performed on the same gouge compositions and under the same conditions as those reported in Part I of the thesis. The biggest variation in the strength between duplicate experiments is the yield strength. Yield strength variation ranges between Δμ ~ .025 in the 50% shale gouge, and Δμ ~ .1 in the 100% carbonate gouge. Notably, one of the 100% shale gouge experiments displays the lowest observed yield strength. The observed range in coefficient of friction between duplicate experiments at maximum shear strain is less than μ = .05. Although one of the 100% shale samples displays the lowest yield strength, this sample strain hardens throughout the duration of the experiment. At maximum shear strain, its strength exceeds the strength of the 75% shale gouge, and is approximately equal to the strength of the 50% shale gouge. The duplicate experiments show that there is some variation in the observed coefficient of friction, however many of the same trends described in Part I can still be observed. Some of the variation in the observed coefficient of friction may be the result of unquantifiable, transient pore fluid pressure differences. Performing these experiments dry may reduce the range in coefficient of friction between experiments on the same gouge compositions. 183 Figure E-1: Replicate Experiments 0 1 2 3 4 5 6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Displacement (mm) C o e ff ic ie n t o f F ri c ti o n 100% Shale, 17% FB ! 75% Shale, 17% FB ! 50% Shale, 17% FB ! 25% Shale, 17% FB ! 100% Carbonate, 17% FB ! Figure E-1: Velocity-stepping experimental data. Displacement vs. coefficient of friction (μ) for all compositions deformed at UBC with one porous forcing block and one non-porous forcing block showing experimental reproducibility. All sample were saturated and deformed at 70 MPa confining pressure and with displacement rates between 100 and 102 μm s-1 (as in Part I). When possible, steady-state or near steady- state sliding was attained before sliding velocity was changed. 184 Appendix F Compilation of Previous Experimental Data 185 Table F-1: Compilation of Previous Work Composition Peak Strength (µ) Confining Pressure (Pc, MPa) Pore Pressure (Pf, MPa) Temp (°C) Displacement Rate (µ s-1) Grain size (µm) Apparatus Author etc Phyllosilicates 100% montmorillonite .08 - .14 25 - 70 MPa saturated room T 200 - .002 µm/s 70% < .01mm; 30% .01 - .08mm triaxial [Logan and Rauenzahn , 1987] montmorillonite ~ .29* 300 saturated room T .866 µm/s 65% < 74µm triaxial [Morrow et al., 1992] montmorillonite ~ .29 * 150 saturated room T triaxial [Morrow et al., 1992] 50% montmorillonite/ illite mix ~ .29* - .33* 150 saturated room T triaxial [Morrow et al., 1992] illite ~ .41* 300 saturated room T triaxial [Morrow et al., 1992] illite ~ .48 300 100 room T .01 - .1 µm/s triaxial [Morrow et al., 1992] muscovite µss ~ .43 20-100 MPa effective normal stress 100 20 1 µm/s 13 µm median ; 90% betwee n 3 and 50 µm in size rotary shear, normal stress stepping [ Van Diggelen et al., 2009] 186 Composition Peak Strength (µ) Confining Pressure (Pc, MPa) Pore Pressure (Pf, MPa) Temp (°C) Displacement Rate (µ s-1) Grain size (µm) Apparatus Author etc muscovite µss ~ .38 100 MPa 100 20 .03 - 3.7 µm/s 13 µm median; 90% between 3 and 50 rotary shear, velocity stepping [Van Diggelen et al., 2009] muscovite ~ .38 5 MPa constant normal stress room T .001–13 µm/s 13 µm median rotary shear, no velocity dependance [Niermeijer and Spiers, ????] chlorite ~ .38* 100 MPa triaxial [Shimamoto, 1977] Carbonates calcite ~ .8 max 35MPa normal stress max 700°C 0.1–0.25 mm biaxial [Kawamoto and Shimamoto, 1998] calcite ~ .72 (wet) - .85 (dry) room T triaxial [Morrow et al., 2000] calcite ~ .63* - .80* 75 MPa room T (3-8) x10-4 triaxial [Shimamoto, 1977] 187 Composition Peak Strength (µ) Confining Pressure (Pc, MPa) Pore Pressure (Pf, MPa) Temp (°C) Displacement Rate (µ s-1) Grain size (µm) Apparatus Author etc Quartz 100% quartz .49 - .62 25 - 70 MPa saturated room T 200 - .002 µm/s 70% < .01mm; 30% .01 - .08mm triaxial [Logan and Rauenzahn, 1987] Natural Mixtures natural illite shale 0.42-0.68 5 to 150 MPa normal stress none room T 0.1 to 200 µm/s 2 to 500 µm double direct shear [Saffer and Marone, 2003] powdered illite schist (illite (59%), quartz (23%), kaolinite/dickite (9%), and plagioclase (4%)) 0.27 to 0.32 37 MPa 5 MPa room T 0.5-300 < 106 µm true triaxial [Ikari et al., 2009] powdered chlorite schist ( chlorite (46%), plagioclase (35%), quartz (12%), and illite (6%)) 0.27 to 0.32 37 MPa 5 MPa room T 0.5-300 < 106 µm true triaxial [Ikari et al., 2009] 188 Composition Peak Strength (µ) Confining Pressure (Pc, MPa) Pore Pressure (Pf, MPa) Temp (°C) Displacement Rate (µ s-1) Grain size (µm) Apparatus Author etc Quartz 100% quartz .49 - .62 25 - 70 MPa saturated room T 200 - .002 µm/s 70% < .01m m; triaxial [Logan and Rauenzahn , 1987] Natural Mixtures natural illite shale 0.42-0.68 5 to 150 MPa normal stress none room T 0.1 to 200 µm/s 2 to 500 µm double direct shear [Saffer and Marone, 2003] powdered illite schist (illite (59%), quartz (23%), kaolinite/dickite (9%), and plagioclase (4%)) 0.27 to 0.32 37 MPa 5 MPa room T 0.5-300 < 106 µm true triaxial [Ikari et al., 2009] powdered chlorite schist ( chlorite (46%), plagioclase (35%), quartz (12%), and illite (6%)) 0.27 to 0.32 37 MPa 5 MPa room T 0.5-300 < 106 µm true triaxial [Ikari et al., 2009]"@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2011-05"@en ; edm:isShownAt "10.14288/1.0053178"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Geological Sciences"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@en ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Carbonate and shale fault gouge : experimental insights into the role of gouge composition, temperature and pore fluid pressure on the mechanical strength of faults"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/33956"@en .