PLATE IMPACT EXPERIMENTS TO INVESTIGATE DYNAMIC SLIP, DEFORMATION AND FAILURE OF MATERIALS

by FUPING YUAN

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Dissertation Advisor: Prof. Vikas Prakash

Department of Mechanical and Aerospace Engineering CASE WESTERN RESERVE UNIVERSITY

January 2008

CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of Fuping Yuan ______________________________________________________ Ph.D. candidate for the ________________________________degree *.

Vikas Prakash (signed)_______________________________________________ (chair of the committee) Christopher J. Hernandez ________________________________________________ John J. Lewandowski ________________________________________________ Dwight T. Davy ________________________________________________

________________________________________________

________________________________________________

10/12/2007 (date) _______________________

*We also certify that written approval has been obtained for any proprietary material contained therein.

Dedicate to my parents, Chaofu and Yunlan, and my brother fubing.

iii

TABLE OF CONTENTS

Title Page Signature Sheet Dedication Page Table of Contents List of Tables List of Figures Acknowledgements Abstract

i ii iii iv xi xiv xxix xxx

Chapter 1: INTRODUCTION 1.1 Co-seismic Slip Weakening in Earth Faults 1.2 High Speed Friction at metal-on-metal Interfaces 1.3 Glass Fiber Reinforced Composites (GRP) 1.4 Bulk metallic glasses (BMG) 1.4.1 Zr41.25Ti13.75Ni10Cu12.5Be22.5 (LM-1 or Vitreloy 1) 1.4.2 The Dynamic Response of BMGs 1.4.3 Pressure Effects on Flow and Fracture Behavior of BMGs 1.5 Dissertation Outline References

1 3 11 14 18 22 23 27 32 35

iv

Chapter 2: SLIP WEAKENING IN ROCKS AND ANALOG MATERIALS AT CO-SEISMIC SLIP RATES 2.1 Introduction 2.2 Pressure Shear Plate Impact Friction Experiments 2.2.1 Experimental Configuration 2.2.2 Tribo-pair Materials and Design of the Pressure-shear Experiments 2.2.3 Wave analysis of Pressure-shear Friction Experiments on Soda-lime Glass: Calculation of Interfacial Tractions, Slip Velocity, and Temperature 2.3 Modified Torsional Kolsky-bar Friction Experiment 2.3.1 Experimental Configuration 2.3.2 Wave analysis: Calculation of Interfacial Tractions, Slip Velocity, Slip Distance 2.3.3 Calculation of Interfacial Temperature Rise 2.4 Experimental Results 2.4.1 Plate-impact Pressure-shear Friction Experiment 2.4.2 Experimental Results Obtained by Using the Modified
v

46

46 52 52 56

60

67 67 70

75 81 81 101

Torsional Kolsky-bar Friction Experiments 2.5 Summary References 125 131

Chapter 3: HIGH-SPEED FRICTION AT METAL-ON-METAL INTERFACES 136 3.1 Introduction 3.2 Experimental Methods 3.2.1 Experimental Configuration 3.2.2 Tribo-pair Materials and Specimen Preparation 3.2.3 Design of the Experiments: Wave Propagation in the Flyer and Target Plates 3.2.4 Waveanalysis of Pressure-shear Friction Experiments Calculation of Interfacial Tractions, Slip Velocity, Slip Distance and Temperature 3.3 Finite-Element Based Methodology to Simulate the Impact of an Elastic-plastic Flyer with an Elastic Target Plate 3.3.1 Finite Element Implementation 3.3.1.1 Contact Algorithm 164 166 151 146 136 139 139 142 144

vi

3.3.1.2 Friction 3.3.1.3 Constitutive Model and Stress-update Algorithm 3.3.2 Finite Element Discretization 3.3.3 Remeshing Algorithm 3.4 Experimental and Computational Results and Discussion 3.5 Summary References

170 176 178 181 185 204 207

Chapter 4: SPALL STRENGTH OF GLASS FIBER REINFORCED POLYMER COMPOSITES 4.1 Introduction 4.2 Material 4.3 Experimental Procedure 4.3.1 Experimental Configuration and Setup 4.3.2 Target Assembly 4.4 Wave Propagation in the Flyer and the Target plates 4.5 Determination of Spall Strength and the Impact Stress 4.6 Experimental Results
vii

212

212 215 218 218 220 221 225 231

4.7 Discussion and Summary References

243 246

Chapter 5: DYNAMIC RESPONSE OF A ZR-BASED BULK METALLIC GLASS UNDER PLANAR SHOCK LOADING 5.1 Introduction 5.2 Experimental Procedure 5.2.1 Experimental Configuration of Spall Experiments 5.2.2 Experimental Configuration of High-strain-rate Plate-impact Pressure-shear Experiments 5.2.3 Wave propagation in the Flyer and the Target Plates: The t-X (time versus distance) Diagram and the S-V (stress versus particle velocity) Diagram for Spall Experiments 5.2.4 Wave Analysis for Propagation of both Longitudinal and Shear Waves in the Flyer and the Target Plates for High-strain-rate Plate-impact Pressure-shear Experiments

249

249 260 261 262

265

268

viii

5.2.5 Calculation of Hydrostatic Pressure in the BMG Specimen for High-strain-rate Plate-impact Pressure-shear Experiments 5.3 Experimental Results and Discussions for spall experiments 5.3.1 Structure of Shock Waves in BMG 5.3.2 Calculation of Spall Strength in Shock Compressed Zr-Based BMG 5.3.3 Effect of of Shock Compression on the Spall Strength of the Zr-Based BMG

272

274 276 280

283

5.3.4 Effect of the Combined Shock Compression and Shear Loading 286 on the Spall Strength of the Zr-Based BMG 5.3.5 Stress-strain Curves for Zr-based BMG under Shock Compression 5.3.6 Scanning Electron Microscopy of the Spall/Fracture Surfaces 5.4 Experimental Results and Discussions for High-strain-rate Plate-impact Pressure-shear Experiments 5.5 Summary References
ix

289

294

298

312 315

Chapter 6: CONCLUSIONS 6.1 Co-seismic slip weakening in earth faults 6.2 High speed friction at metal-on-metal interfaces 6.3 Spall strength of Glass fiber reinforced composites 6.4 Dynamic reponse of a Zr-based bulk metallic glasses References

320 321 323 326 329 333

APPENDIX A: CALCULATION OF SHEAR STRAIN IN THE GRP TARGET UNDER COMBINED PRESSURE-SHEAR LOADING

336

APPENDIX B: FREE SURFACE PARTICLE VELOCITY PROFILES FOR ALL PLATE-IMPACT EXPERIMENTS

341

BIBLIOGRAPHY

358

LIST OF TABLES Table 1.1: Physical properties of the Zr-based BMG (Vit-1) used in the present investigation Table 2.1: Summary of thermal properties of soda-lime glass, Quartz and fine grained Arkansas rock samples Table 2.2: Summary of the plate impact pressure shear friction experiments on soda-lime glass and fine grained Arkansas rock samples Table 2.3: Summary of modified torsional Kolsky-bar friction experiments conducted on quartz Table 2.4: Summary of modified torsional Kolsky-bar friction experiments conducted on soda-lime glass Table 3.1: Physical properties of the flyer and target plates used in the pressure-shear plate impact friction experiments in the present study Table 3.2: Physical properties and model parameters for 7075-T6 Al alloy and the CH steel tribo-pair materials used in the finite element constitutive model Table 3.3: Explicit time integration algorithm 175 164 143 111 102 82 66 23

xi

Table 3.4: Summary of the plate-impact pressure shear friction experiments conducted in the present study Table 3.5: Summary of the plate-impact pressure shear friction experiments conducted in a previous study by Liou et al. (2004) and Okada et al. (2001) Table 4.1: Summary of all the normal plate impact and the pressure-shear plate impact experiments conducted to obtain the spall strength of S2 glass GRP Table 4.2: Summary of all the normal plate impact and the pressure-shear plate impact experiments conducted to obtain the spall strength of E-glass GRP Table 4.3: Summary of normal stress, shear strain and spall strength for S2 glass GRP Table 4.4: Summary of normal stress, shear strain and spall strength for E-glass GRP Table 5.1: Summary of the normal spall experiments on the Zr-based bulk metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5

186

187

233

234

240

241

275

xii

Table 5.2: Summary of the pressure-shear spall experiments on the Zr-based bulk metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5 Table 5.3: Physical properties of the Zr-based BMG (Vit-1) used in the present investigation. The data is taken from Lu (2002) Table 5.4: Summary of high-strain-rate plate-impact pressure-shear experiments on the Zr-based bulk metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5

276

276

298

xiii

LIST OF FIGURES Figure 1.1: Flash heating at asperity contacts Figure 1.2: Cross-section of an example of composite integral armor system Figure 1.3: Plot of strength vs. elastic limit for various materials, including metallic glasses Figure 2.1: Schematic of the plate-impact pressure-shear friction experiment Figure 2.2: Schematic of the combined normal and transverse displacement interferometry Figure 2.3: Plate-impact configurations for (a) soda-lime glass; (b) Arkansas Novaculite rock Figure 2.4: Wave propagation in flyer and target plates: time-distance diagram Figure 2.5: Schematic of the modified torsional Kolsky-bar friction experiment Figure 2.6: Alignment fixture to ensure parallelism of the tribo-pair surface Figure 2.7: Wave propagation diagram for the modified torsional Kolsky bar Figure 2.8: Loci of all the torque and angular velocity states that can be attained at the tribo-pair interface Figure 2.9: Transverse particle velocity history of the rear surface of the CH tool steel target plate for Shot # 1
xiv

8 16 19

54 56

59

59 69 70 74 75

83

Figure 2.10: History of slip velocity and coefficient of kinetic friction for Shot # 1 Figure 2.11: Estimated temperature rise at tribo-pair interface for Shot # 1 Figure 2.12: Normal and transverse particle velocity history at the rear surface of the target plate for Shot # 2 Figure 2.13: History of slip velocity and coefficient of kinetic friction for Shot # 2 Figure 2.14: Estimated temperature rise at tribo-pair interface for Shot # 2 Figure 2.15(a): AFM scan of soda-lime glass film before the pressure-shear friction experiment for Shot #2 Figure 2.15(b): Post test AFM scan of soda-lime glass film showing regions that have melted away during frictional slip for Shot # 2 Figure 2.16: Transverse particle velocity history at the rear surface of the target plate for Shot # 3 Figure 2.17: History of slip velocity and coefficient of kinetic friction for Shot # 3 Figure 2.18: Estimated temperature rise at tribo-pair interface for Shot # 3

85

86 88

90

91 92

92

95

95

96

xv

Figure 2.19: Normal and transverse particle velocity history at the rear surface of the target plate for Shot FY008 Figure 2.20: History of slip velocity and coefficient of kinetic friction for plate-impact pressure-shear experiment on novaculite rock samples (Shot FY008) Figure 2.21: Estimated temperature rise at tribo-pair interface for the plate-impact pressure-shear friction experiment on novaculite rock samples (Shot FY008) Figure 2.22: History of slip velocity and coefficient of kinetic friction for plate-impact pressure-shear experiment on novaculite rock samples (Shot FY009) Figure 2.23: Estimated temperature rise at tribo-pair interface for the plate-impact pressure-shear friction experiment on novaculite rock samples (Shot FY009) Figure 2.24: Experimentally measured torque at the strain gage location A for Q01 Figure 2.25: History of interfacial normal pressure, shear stress and slip velocity for Q01
xvi

98

98

99

100

100

104

104

Figure 2.26: Interfacial normal pressure, slip velocity and coefficient of kinetic friction vs. interfacial slip distance for Q01 Figure 2.27: Estimated temperature rise at tribo-pair interface for Q01 Figure 2.28: High-speed camera images for Q01 Figure 2.29: Experimentally measured torque at the strain gage location A for Q02 Figure 2.30: History of interfacial normal pressure, shear stress and slip velocity for Q02 Figure 2.31: Interfacial normal pressure, slip velocity and coefficient of kinetic friction vs. interfacial slip distance for Q02 Figure 2.32: Estimated temperature rise at tribo-pair interface for Q02 Figure 2.33: High-speed camera images for Q02 Figure 2.34: Experimentally measured torque at the strain gage location A for G01 Figure 2.35: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G01 Figure 2.36: Estimated temperature rise at tribo-pair interface for G01

105

105 106 108

108

109

109 110 112

113

113

xvii

Figure 2.37: Experimentally measured torque at the strain gage location A for G02 Figure 2.38: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G02 Figure 2.39: Estimated temperature rise at tribo-pair interface for G02 Figure 2.40: Experimentally measured torque at the strain gage location A for G03 Figure 2.41: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G03 Figure 2.42: Estimated temperature rise at tribo-pair interface for G03 Figure 2.43: Experimentally measured torque at the strain gage location A for G04 Figure 2.44: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G04 Figure 2.45: Estimated temperature rise at tribo-pair interface for G04 Figure 2.46: Experimentally measured torque at the strain gage location A for G05

115

115

116 116

117

117 119

120

120 121

xviii

Figure 2.47: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G05 Figure 2.48: Estimated temperature rise at tribo-pair interface for G05 Figure 2.49: Experimentally measured torque at the strain gage location A for G06 Figure 2.50: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G06 Figure 2.51: Estimated temperature rise at tribo-pair interface for G06 Figure 2.52: Highest coefficient of kinetic friction versus interfacial slip velocity obtained from experiments conducted on the soda lime glass/soda lime glass tribo-pair Figure 2.53: SEM micrograph for sliding surface of soda-lime glass thin film tested in G01 Figure 2.54: SEM micrograph for sliding surface of soda-lime glass disk specimen tested in G01 Figure 2.55: Flash heating at asperity contacts Figure 3.1: Schematic of the plate-impact pressure-shear friction experiment

121

122 122

123

123 124

125

125

129 140

xix

Figure 3.2: Schematic of the combined normal and transverse velocity interferometer Figure 3.3: Time-distance diagram detailing the wave propagation in the tribo-pair plates Figure 3.4: Shear stress versus transverse particle-velocity diagram for an elastic flyer plate impacting an elastic target plate Figure 3.5: Flash heating at asperity contacts Figure 3.6: (a) Flow-stress versus the plastic-strain-rate dependence used in the viscoplastic model for the flyer and the target plates; (b) Flow-stress versus the temperature dependence used in the visco-plastic model for the flyer and the target plates; (c) Flow-stress versus the equivalent-plastic-strain dependence used in the visco-plastic model for the flyer and the target plates Figure 3.7: (a) Predictor configuration of contact pair; (b) Corrected configuration after normal acceleration corrections are introduced which eliminate the unwanted penetration Figure 3.8: The finite element mesh used to simulate the plate impact pressure-shear friction experiments
xx

142

145

148

149 163

168

180

Figure 3.9: Re-mesh and re-map algorithm for mesh with periodic boundary condition Figure 3.10: Schematic of the update of the current configuration Figure 3.11: Normal and transverse particle velocity history at the rear surface of the target plate for Shot Fuping05008 Figure 3.12: History of the applied normal stress, friction stress and the slip velocity for Shot Fuping05008 Figure 3.13: Predictions based on finite element simulations for the evolution of effective stress for Shot Fuping05008 Figure 3.14: Temperature profile in the aluminum alloy flyer plate in the vicinity of the sliding interface for Shot Fuping05008 Figure 3.15: History of the applied normal stress, friction stress and the slip velocity. (a) Shot Fuping05001; (b) Shot Fuping05004; (c) Shot Fuping05007 Figure 3.16: Predictions based on finite element simulations for the evolution of effective stress. (a) Shot Fuping05001; (b) Shot Fuping05004; (c) Shot Fuping05007

182

185 190

191

191

192

194

195

xxi

Figure 3.17: Temperature profile in the aluminum alloy flyer plate in the vicinity of the sliding interface. (a) Shot Fuping05001; (b) Shot Fuping05004; (c) Shot Fuping05007 Figure 3.18: Kinetics of growth of molten 7075-T6 Al layer at the tribo-pair interface as a function of impact velocity Figure 3.19: Deformed mesh in 7075-T6 Al in the vicinity of the frictional interface prior to the melt transition for Shot MO9906. Direction of slip is from left to right Figure 3.20: Deformed mesh in 7075-T6 Al in the vicinity of the frictional interface after the melt transition for Shot MO9906. Direction of slip is from left to right Figure 3.21: Micrograph of the slip surface of the Al alloy flyer plate. (a) Shot MO9903; (b) Shot MO9906; (C) Shot Fuping05008; (d) Shot Fuping05007 Figure 3.22: Coefficient of kinetic friction versus interfacial slip-velocity obtained from experiments conducted on the 7075-T6 Al alloy/CH tool-steel tribo-pair Figure 4.1: SEM micrograph of the S2 glass fiber woven roving layer
xxii

196

197

200

201

202

203

218

Figure 4.2: SEM micrograph of the 5-harness satin weave E-glass fiber woven roving layer Figure 4.3: Schematic of the plate impact experimental configuration used in the present study to investigate the spall strength in the GRP under normal shock compression and combined shock compression and shear loading Figure 4.4: Photograph showing a typical GRP specimen mounted on the aluminum target plate Figure 4.5: Time-distance diagram showing the wave propagation and the stress states in the flyer and the target plates. The spall plane occurs approximately in the middle of the target plate Figure 4.6: Stress-velocity diagram showing the loci of all the stress and particle velocity states that can be achieved in a typical plate-impact spall experiment Figure 4.7: Time-distance diagram paired with the measured free surface particle velocity profile for Experiment FY06001 to illustrate the pull-back phenomenon in the free surface particle velocity profile for a typical plate-impact spall experiment
xxiii

218

220

221

224

224

226

Figure 4.8: Free surface particle velocity profile for Experiment FY06001 showing the calculation of the spall strength
Figure 4.9: Shock velocity vs. Particle velocity for E-glass GRP

227

230

Figure 4.10: Spall strength vs. Impact stress obtained from the normal plate-impact experiments Figure 4.11: Free surface particle velocity profiles for Experiments FY06003 and FY06010. The effect of the superimposed shear strain on the spall strength of the E-glass GRP is emphasized Figure 4.12: Spall strength as a function of the shear strain in the S2 glass GRP for selected experiments each having a normal component of the impact stress of about 200 Mpa Figure 4.13: Spall strength illustrated in relationship with normal stress and shear strain for the S2 glass GRP Figure 4.14: Spall strength illustrated in relationship with normal stress and shear strain for the E-glass GRP Figure 5.1: Schematic of the normal and combined pressure-shear plate-impact spall experiments

236

238

239

242

243

262

xxiv

Figure 5.2: Schematic of the high-strain-rate plate-impact pressure-shear experiments Figure 5.3: Schematic of the combined normal and transverse velocity interferometer Figure 5.4: Wave Propagation in the flyer and the target plates (t-X diagram) for a typical plate-impact spall experiment Figure 5.5: Stress versus particle velocity (S-V) diagram for a typical plate-impact spall experiment. The dashed-line shows the hypothetical case for the no-spall condition Figure 5.6: Wave Propagation in the flyer and the target plates (t-X diagram) for a typical high-strain-rate plate-impact pressure-shear experiment Figure 5.6: Free-surface particle velocity versus time profiles for the four normal plate-impact spall experiments conducted in the present study Figure 5.7: Longitudinal stress versus time profiles for Shots FY06012 and FY 06013. The HEL for the Zr-based BMG is estimated to

263

265

266

267

270

278

279

be ~ 6.15 Gpa Figure 5.8: Time-distance diagram paired with the measured free-surface
xxv

282

particle velocity profile for experiment FY06011 to illustrate the pull-back spall signal Figure 5.9: Free surface particle velocity data for experiment Shot FY06011 Figure 5.10: Spall strength as a function of impact Stress for three different Zr-based BMGs Figure 5.11: Free-surface particle velocity versus time profiles for the one normal spall experiments and five pressure-shear spall experiments with normal stress about 5 Gpa Figure 5.12: Spall strength as a function of shear strain for the six experiments conducted in the present study. The normal stress is maintained at ~ 5 GPa in the six experiments Figure 5.13: Time distance diagram for elastic-plastic impact showing the plastic wave fan Figure 5.14: Normal (Longitudinal) stress vs. normal (longitudinal) strain curves for the four normal plate impact experiments Figure 5.15: Strain and strain-rate histories in the Zr-based BMG for Shot FY06012 293 293 291 288 288 282 286

xxvi

Figure 5.16: Normal stress and strain rate vs. normal strain in the Zr-based BMG for Shot FY06012 Figure 5.17: Scanning electron microscope pictures of spall surfaces for two normal plate-impact spall experiments Figure 5.18: Scanning electron microscope pictures of spall surfaces for four pressure-shear plate-impact spall experiments Figure 5.19: Longitudinal and transverse free surface velocities for Shot Exp. 1 Figure 5.20: History of normal stress, the shear stress and the shear strain rate in the specimen for Shot Exp. 1 Figure 5.21: Dynamic shear stress versus shear strain for Shot Exp. 1 Figure 5.22: Longitudinal and transverse free surface velocities for Shot Exp. Figure 5.23: History of normal stress, the shear stress and the shear strain rate in the specimen for Shot Exp. 2 Figure 5.24: Dynamic shear stress versus shear strain for Shot Exp.2 Figure 5.25: Longitudinal and transverse free surface velocities for Shot Exp. 3 Figure 5.26: History of normal stress, the shear stress and the shear strain rate in the specimen for Shot Exp. 3 Figure 5.27: Dynamic shear stress versus shear strain for Shot Exp.3
xxvii

294

296

297

300 300

301 303 303

304 306 306

307

Figure 5.28: Shear yielding stress vs. normal pressure for data obtained from Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002) Figure 5.29: Shear yielding stress vs. normal pressure for data obtained from Lewandowskis work (Lewandowski, J. J. And Lowhaphandu, P., 2002) and present study Figure A.1: Oblique impact configuration

309

310

336

xxviii

ACKNOWLEDGEMENTS

First, I would like to thank my advisor, Prof. Vikas Prakash, for his continuous guidance and encouragement through out my stay in Case Western Reserve University. His endless support has put me through numerous difficulties. I would also like to thank

Prof. John J. Lewandowski for his support and the useful discussions for my research work. I would also like to thank all my committee members: Dr. Dwight T. Davy, and

Dr. Christopher J. Hernandez for spending time in their tight schedule to attend my dissertation defense.

To all my colleagues and friends who helped me academically and emotionally through my time here: Liren Tsai, Xin Tang, George Sunny, Mostafa Shazly, David Nathenson, Pankaj Kaul, Naoto Utsumi, thank you all, you are all so special to me.

Finally, I would like to thank my parents and my brother for their love and support through out my life. I love you all.

xxix

Plate Impact Experiments to Investigate Dynamic Slip, Deformation and Failure of Materials

Abstract By FUPING YUAN

Determination of co-seismic slip resistance in earth faults is critical for understanding the magnitude of shear-stress reduction and hence the near-fault acceleration that can occur during earthquakes. In the present study plate-impact pressure-shear friction

experiments and modified torsional Kolsky-bar friction experiments were employed to investigate the frictional resistance in rocks and analog materials (quartz, soda lime glass), at relevant normal pressures and co-seismic slip rates. The results of this study have

relevance not just to mechanisms of dynamic fault weakening, but also to the constitutive description of the behavior that can be used to critically examine the rate and state dependent dynamic friction models of earthquake rupture.

In an attempt to better understand the nature of dynamic friction for metal tribo-pairs,
xxx

plate-impact pressure-shear friction experiments were employed to investigate dynamic slip resistance and time-resolved growth of molten metal films during dry metal-on-metal (tool-steel on 7075-T6 Al alloy) slip under extreme interfacial conditions. A Lagrangian finite-element code was developed to understand the evolution of the thermo-mechanical fields and their relationship to the observed slip response.

In order to better design and develop GRP-based light-weight integral armor, a series of plate-impact experiments were designed to study spall strength in two glass-fiber reinforced polymer composites (GRP) S2 glass woven roving in Cycom 4102 polyester resin matrix and a balanced 5-harness satin weave E-glass in a Ciba epoxy (LY564) matrix. The spall strengths of the two GRP composites were observed to The E-glass GRP was

decrease with increasing compression stress and shear strain.

found to have a much higher level of spall strength when compared to the S2-glass GRP.

The dynamic response of bulk metallic glasses is of considerable interest for potential applications, such as kinetic energy penetrator. A series of plate-impact

experiments were conducted on a Zr-based BMG (Zr41.25Ti13.75Ni10Cu12.5Be22.5) to: (a) better understand the structure of shock waves in the BMG, (b) estimate residual spall
xxxi

strength of the BMG as a function of normal stress and shear strain, (c) obtain the Hugoniot elastic limit (HEL) of the material, and (d) investigate normal pressure effect on the shear yield stress of the BMG.

xxxii

Chapter 1

INTRODUCTION

The dynamic behavior of materials is an area of study at the confluence of many scientific disciplines. The processes that occur when bodies are subjected to rapidly changing loads can differ significantly from those that occur under static or quasi-static situations. Inertia and inner kinetics of materials become an important factor in the dynamic events. There are many applications of dynamic response of materials, such as

high speed machining, crashworthiness of vehicles, earthquakes, design of blast and impact resistant structures, integral armor used in the military programs, and the dynamic damage to structures used in the aerospace industry. These applications require a

thorough knowledge of the various modes of deformation, slip and failure in engineering material and their interfaces.

Dynamic slip and friction refer to the large and important class of truly dynamic problems which include such effects as dynamic sliding, stick-slip motion, slip weakening and strengthening, lubrication, chattering, wear and plowing, and melting of slip interfaces. In view of the scientific and technological importance of understanding on dynamic friction and given its current state of understanding, new experiments and

friciton models are required which can simulate the local conditions of slip interfaces.

Dynamic deformation and failure of materials include a large amount of physical processes such as large plasticity, wave propagation, crack initiation and propagating, dynamic fracture such as spalling, and phase transformation. Recently, the need for

military, navy and aerospace operations has placed renewed demands on new and improved materials such as fiber-reinforced polymer composite (GRP) and bulk metallic glasses (BMG). In order to better apply these new materials in their operations, new

experiments need be designed and employed to investigate dynamic deformation and failure of these new materials.

This thesis describes experimental and computational work that is focused on obtaining a better understanding of such dynamic effects in materials and interfaces that are of considerable scientific and technical importance. The first part of the thesis focuses on better understanding co-seismic slip weakening and/or strengthening mechanisms in both geological and analog materials that is of importance in understanding fault and earthquake dynamics. In the second part, plate-impact

pressure-shear friction experiments are used to investigate dynamic slip resistance and time-resolved growth of molten metal films that are created during dry metal-on-metal slip under extreme interfacial conditions. The results of these studies are important in

developing a better understanding of the slip and frictional behavior at tool-workpiece

2

interfaces in high-speed machining applications, design of high tolerance interfaces operating under extreme interactions, penetration and perforation of armor, design of rail systems for high speed sleds on rails, to name a few. The third part of the thesis involves the study of dynamic failure, in particular dynamic spall in potential armor materials, such as, light weight glass fiber reinforced polymer composites. In this

regards spall strength in two different architectures of the GRP composites (S2 glass woven roving in Cycom 4102 polyester resin matrix and a 5-harness satin weave E-glass in a Ciba epoxy matrix) are investigated. While in the fourth part, results are presented

on the shock response of a zirconium-based bulk metallic glass (BMG), Zr41.25Ti13.75Ni10Cu12.5Be22.5, subjected to shock compression and combined shock compression and shear loading. Bulk metallic glasses are amorphous metals with extraordinary high hardness and strength that can potentially be used in the design of light weight armor for the navy and the army.

In the following a brief introduction is provided to each of these topics. Besides providing a literature survey of the previous work done in these areas it also provides a summary of the outstanding issues and motivation for their study.

1.1 Co-seismic Slip Weakening in Earth Faults Determination of co-seismic slip resistance in earth faults is critical for understanding the magnitude of shear stress reduction and hence the near-fault
3

acceleration that can occur during earthquakes. Friction stresses, at constant normal pressure and sliding speeds of less than 1 mm/s, have been well studied for a wide range of materials. The characterization of these results in terms of rate and state variable

friction laws has allowed a better understanding of a wide variety of aspects of the mechanics of earthquakes (Scholz et al., 1972, Dieterich, 1979, 1981, Rice and Ruina, 1983, Ruina, 1983, Dieterich and Kilgore, 1994, Tullis, 1994, Dieterich and Kilgore, 1996, Blanpied et al., 1998, Marone, 1998, Scholz, 1998, Di Toro et al., 2004), including, for example, what to expect in terms of premonitory slip and Omoris Law for aftershock decay (Scholz, 1998).

Experimental data suggest that frictional resistance in geo-materials at slip speeds 1 mm/s and slip distance < 1mm is quite high ( ~ 0.6 to 0.85) (Byerlee, 1978, Dieterich, 1978, Dieterich and Kilgore, 1996). However, seismic inversions provide evidence that frictional resistance of major faults at co-seismic slip speeds (~1-2 m/s) may be quite low (Heaton, 1990, Rice, 2006). Moreover, very little data exist for the simultaneously high slip rates and large slip displacements characteristic of co-seismic slip, and the data that do exist suggest that the frictional behavior at these slip speeds is dramatically different and the dynamic slip weakening occurs (Sibson, 1973, Tsutsumi and Shimamoto, 1997, Goldsby and Tullis, 2002, Di Toro et al., 2004, Mizoguchi et al., 2006, O'Hara et al., 2006).

Recent field observations suggest that slip in individual events of earthquake are extremely localized, and occur primarily within a thin shear zone, 0.2-5 mm thick (Chester et al., 2003; Chester et al., 2005; Chester and Goldsby, 2003; Wibberley and Shimamoto, 2003). Given that earthquake slips are often accommodated within thin

zones, but that evidence of melting is not pervasive, especially at the shallow depths of activity represented by surface exposures, it is reasonable to suspect that strong weakening mechanisms must exist during rapid, large slip. Based on observation and theory, Ranjith and Rice (2001), Rudnicki and Rice (2006), Segall and Rice (2006), Rempel and Rice (2006) and Rice (2006) have recently summarized some important thermal weakening mechanisms which are assumed to act in combination during fault events: (1) Flash heating and consequent weakening at highly stressed asperity contacts during rapid slip; (2) Thermal pressurization of pore fluid within the fault core; (3) Macroscopic melting; (4) Silica gel formation; (5) Reductions of normal stress.

Flashing heating at asperity contacts The highest temperatures, which occur close to the areas of true contact at which the energy is dissipated, are of short duration and are sometimes called the flash temperature. Flash heating and consequent weakening at highly stressed asperity contacts during rapid slip which reduces the friction coefficient, a phenomenon studied for many years as the key of understanding the slip rate dependence of dry friction in metals at high slip rates (Bowden and Thomas, 1954, Archard, 1958, Barber, 1976, Kuhlmann-Wilsdorf, 1985,
5

Ashby et al., 1991, Irfan and Prakash, 1994), has also been considered recently in seismology as a mechanism that could be active in controlling fault friction during seismic slip before macroscopic melting (Andrews, 2002, Hirose and Shimamoto, 2005, Wibberley and Shimamoto, 2005).

Recent measurements of contact area and contact indentation strength, C , in transparent materials (such as quartz) by light scattering (Dieterich and Kilgore, 1994, 1996), confirmed earlier suggestions by Boitnott et al.(1992) that the shear strength, C , of the asperity junctions is very high (estimated to be of order 10% of the shear modulus G) in typical rock systems, and thus when forced to shear, they generate intense but highly localized heating during their life time. The local shear strength, C , of the

asperity contact presumably degrades continuously with increasing flash temperature, Ta . An elementary model considering flash heating at asperity contacts had been proposed recently by Rice (2006). The model considers contacts of uniform size L , and hence life-time L /V , where L is the slip needed to renew the asperity contact population, and V is slip rate; and assumes that their shear strength remains at the room temperature value, C , until temperature has reached a critical value, Tcrit , above which the weakened shear strength W is assumed to has a negligibly small compared to its room temperature value C (see Figure 1.1). The temperature rise of the asperities is

estimated from a simple one-dimensional transient heat conduction equation, with heating power CV per unit area at the sliding contact surface. Tullis and Goldsby
6

(2003) observed that a better fit of Rices model to data showing strong rate-weakening of friction was to assume a small but non-negligible W for Ta > Tcrit . The model

thus defines a critical slip rate Vcrit such that there is no weakening if V < Vcrit , but strong weakening if V > Vcrit . That is, the friction coefficient , which has the value

0 at low slip rates and W at high slip rates, is given in the following simple model

= 0 if V < Vcrit , = (0 W )

Vcrit V

+ W if V > Vcrit ; .

(1.1)

Vcrit

= L

c (Tcrit Tbulk ) C

(1.2)

where is thermal diffusivity, c is heat capacity per unit volume.

Using Equations (1.1) and (1.2), a Vcrit = 0.12 m/s can been estimated to be onset of severe thermal weakening in glass and other geo-materials, when it is recognized that L = 5 m , C = 3.0 GPa and temperature range from ambient up to ~ 900 0C.

Figure 1.1: Flash heating at asperity contacts

Thermal pressurization of pore fluid within the fault core Thermal pressurization of pore fluid within the fault core is a weakening mechanism by frictional heating which reduces the effective normal stress and hence reduces the shear resistance associated with any given friction coefficient (Sibson, 1973, Lachenbruch, 1980, Mase and Smith, 1985, Lee, 1987, Mase and L., 1987, Andrews, 2002, Wibberley, 2002, Noda and Shimamoto, 2005, Sulem et al., 2005).

This mechanism assumes that fluids (water, typically) are present within the thin shear zone of fault, and that the shear strength during seismic slip can still be represented by the classical effective stress law = f ( p) , where is normal stress and p is pore pressure (Cocco and Rice, 2002). Frictional heating causes the fluid (if it was unconstrained rather than caged by the densely packed solid particles) to expand
8

in volume.

Thus a pressure increase must be induced in the pore fluid. Since

typically remains constant during slip, that means that strength is reduced, ultimately towards zero, as shear heating continues to raise temperature so that p approaches .

Macroscopic melting Under the right conditions, the first two mechanisms are understood to become important immediately after seismic slip initiates (Segall and Rice, 2006), but as large slip develops they may not always remain the most significant weakening mechanisms. Macroscopic melting (i.e., a coherent melt layer has formed along the whole sliding surface) may occur too for sufficiently large combinations of slip speeds and effective normal stress. In this regime, thermal power generated during the solid-on-solid slip overwhelms the ability of thermal conduction to carry the frictional heat away from the slip interface, and the slip resistance is expected to be primarily controlled by the shear-strain rate, the thickness and viscosity of molten layers (Bowden and Persson, 1960, Okada et al., 2002, Liou et al., 2004, Hirose and Shimamoto, 2005). Molten layers have a low viscosity and may lubricate faults reducing dynamic friction. However, it has been observed that interface melting is not a simple weakening mechanism for rock and analog materials. Hirose et al. (2005) conducted a series of experiments on Indian gabbro at slip rates of 0.85-1.49 m/s and normal stresses of 1.2-2.4 MPa by using a rotary-shear apparatus. The experiments on gabbro revealed two stages of slip weakening separated by a marked strengthening regime. By examination of microstructures of simulated fault
9

zone under scanning electron microscopy (SEM) at different total slip displacements, they proposed that the initial slip weakening is due to the thermal weakening induced by flash heating at the asperity contacts and early stages of melting; this phase is followed by slip strengthening caused by the coalescence of melt patches into a thin molten layer, while the growth of molten layer during friction melting is the primary cause for the observed second slip weakening.

Silica gel formation Another weakening mechanism, i.e. by the formation of silica gel, has been identified by Di Toro et al. (2004). They studied the slip resistance of Arkansas Novaculite rock at a slip rate of 0.03 m/s and a normal stress of 5 MPa using a servo-controlled compression-torsion apparatus. They assume that granulation within

the shear zone produces fine silica particles which absorb water to their surfaces and form a thin gel layer. Then they attributed the initial slip weakening mechanism to the

formation of silica gel, and the time-dependent recovery of shear strength to the thixotropic behavior of the silica gel. Moreover, within a shearing amount and slip rate

range for which the mechanism was plausibly established for pure quartzite rocks (Arkansas Novaculite), weakening for other rock systems seems to be ordered by their silica content (Roig Silva et al., 2004): quartzite (novaculite) > granite > gabbro.

10

Reductions of normal stress In addition to the principal problem that we do not have adequate data or constitutive descriptions for large and rapid sliding on a fault at constant normal stress, we do not have much knowledge of the effect on the resistance of sudden changes in normal stress on frictional slip resistance, and this is an important problem, since during seismic slip there can be abrupt changes in normal stress as well as shear stress. Data exist at low slip speed showing that the simple Coulomb representation for the effect of normal stress is inadequate and that there are memory effects following a change in normal stress (Linker and Dieterich, 1992, Richardson and Marone, 1999, Bureau et al., 2000), but few data exist at high slip velocity (Prakash and Clifton, 1993, Prakash, 1998, Irfan and Prakash, 2000, Ullah et al., 2007). Theoretical studies of slip at dissimilar material interfaces (Weertman, 1980, Adams, 1995) have shown that spatially inhomogeneous slip causes an alteration of normal stress, so this, as well as effects from non-planar faults and interactions between nearby faults, can cause changes in normal stress. Reductions of

normal stress can reduce frictional resistance and thus lead to possible destabilization of slip (Andrews and Ben-Zion, 1997, Simes and Martins, 1998, Cochard and Rice, 2000, Ben-Zion, 2001, Ranjith and Rice, 2001).

1.2 High Speed Friction at Metal-on-metal Interfaces High-speed sliding at metal interfaces plays an important role in the design of several tribo-elements including bearings, gears, high performance brake liners, high speed

11

machining process, rocket driven vehicles on rails, and frictional model for penetration. Most tribologists agree that the nature of dynamic friction for most metal tribo-pairs is a complex process and is affected by a long list of factors including the interface constitution, roughness of the contacting surfaces, slip velocity, slip distance, normal pressure, and inertia and thermal effects.

When two contacting solids slide, work is done against friction. Almost all the frictional work appears as heat, generated at or very close to the surface at which the two solids meet. Frictional heating at relatively low slip speeds is a transient and localized phenomenon and does not have a great influence on the resultant friction mechanism. In marked contrast to this, at relatively high sliding speeds the metal surfaces are subjected to intense frictional heating which profoundly changes the state of the tribo-pair interface (Bowden and Ridler, 1936; Bowden and Tabor, 1950; Bowden and Thomas, 1954; Archard, 1958; Bowden and Persson, 1960; Barber, 1970; Kennedy, 1984; Kuhlmann-Wilsdorf, 1985; Carslaw and Jaeger, 1986; Irfan and Prakash, 1994). Under these conditions shearing of the material at the frictional interface plays a fundamental part during the dynamic slip process and results in a sharp local temperature rise. Since

the sheared layer is very thin and the sliding speeds are very high, extraordinary high rates of strain are obtained.

The dynamics of metal-on-metal sliding at high slip speeds and/or high normal pressures have been investigated by many investigators. In some of the early works, Shugarts (1953) developed an experimental apparatus in which a rod was pressed against

12

the rim of a rapidly revolving disk.

Krafft (1955) investigated the phenomena of high

speed friction in a series of ballistic tests where friction occurred between the surface of a rotating bullet and the steel target it penetrated. Williams and Griffen (1964),

Montgomery (1976a), Kadhim and Earles (1966), Earles and Powell (1966), Lim et al. (1989) and Sternlicht and Apkarian (1960) provided evidence of melt in high speed sliding situations using a high-speed pin-on-disk test device. Montgomery (1976b) investigated the phenomenon of high speed friction by studying the acceleration of projectiles in gun barrels and estimated friction coefficient values as low as 0.02 between gilding metal (90% Cu and 10% Zn) and steel.

Bowden and Freitag (1958) and Bowden and Persson (1960) spun a steel ball to a very high rotational speed and then grabbed it with other frictional samples or dropped it on another to achieve slip speeds as high as 800 m/s. In these experiments, velocity weakening of the frictional force as a function of increasing slip velocities was observed. However, the normal pressures were quite low (less than 0.015 MPa). Higher interfacial

pressures (approximately 50 to 100 MPa) but at relatively lower slip speeds (approximately 5 m/s) were obtained in experiments by Ogawa (1997), who modified a split Hopkinson pressure bar to study impact friction by axially impinging the input tube on a rotating output tube. By replacing the transmitter bar with a rigid disk,

Rajagopalan and Prakash (1999) modified the conventional torsional Kolsky bar to investigate high speed friction at normal pressures of approximately 50 to 100 MPa and slip velocities up to 10 m/s.

13

Most recently, plate impact pressure-shear friction experiments were conducted by Irfan and Prakash (2000) to understand time resolved frictional characteristics during high-speed sliding of metal on metal (Carpenter Hampden tool-steel/Ti-6Al-4V tribo-pair). In their experiments interfacial slip speeds of 1 to 60 m/s at normal pressures ranging from 1 to 3 GPa were obtained. The combination of the high slip speeds and

normal pressures led to interfacial temperatures of approximately 800oC and higher in approximately 3-4 s of frictional slip. However, because of the relative high melt

point temperature of the tribo-pair materials employed in the study, near surface melting during the high speed slip was not observed. By employing tribo-pairs comprising hard tool-steel against relatively low melt-point metals such as 7075-T6 aluminum alloy, Okada (2001; 2002) and Liou et al. (2004) investigated time-resolved growth of molten metal films during dry metal-on-metal slip under extreme interfacial conditions (normal stress ranging from 1 to 3 GPa and interfacial slip speeds in the range of 1 to 100 m/s).

1.3 Glass Fiber Reinforced Composites (GRP) World War II ballistic threats and multifunctional survivability requirements coupled with a U.S. strategy of forward-presence of ground-based forces have encouraged the evolution of ground fighting vehicles to their present 70+ ton status (Fink, 2000). However, global political dynamics and a return of public sentiment for a stronger U.S. role in global peace-keeping have made the ability to fast deploy ground forces around the world essential for future U.S. Army strategy, tactics, and weaponry development. Therefore, several programs have been proposed by U.S. military to decrease the weight

14

of combat vehicles.

In particular, in the late 1980s, polymer-matrix composites (PMCs)

were considered to be the prime candidates to replace the aluminum hull in the Bradley Infantry Fighting Vehicle, and also for the Composite Armored Vehicle (CAV) program in the mid-1990s. In 2003, a joint U.S. military Future Combat Systems (FCS) program was initiated in an effort to develop combat vehicles with less weight and better survivability.

Under the U.S. Armys Composite Armor Vehicles (CAV) program and the Future Combat Systems (FCS) program, various composite material systems have been investigated to understand and optimize the performance of potential Composite Integral armor (CIA) systems (Betheney, 1998; Vaidya et al., 1999). These composites generally comprise of several layers that contribute to specific functions, e.g. ballistic resistance, shock resistance, structure stability, etc. An example of CIA is illustrated in Figure 1.2. The cross-section view of this particular CIA demonstrates the arrangement of the constituents. In this architecture, the ceramic tile absorbs most of the kinetic energy of the incident projectile, while the rubber layer improves multi-hit ballistic performance (by shock-wave attenuation) and it also enables damage isolation to single ceramic tile. The S2-Glass/Epoxy layers in this structure not only provides structural stability and ballistic resistance, but also prevents bending failure of the ceramic tiles during low velocity impact. The phenolic liner at the bottom of the composite provides spall

reduction, and fire, smoke, and toxicity protection (Gama, 2001).

15

Figure 1.2: Cross-section of an example of composite integral armor system.

In the 20th century, modern composites were used in the 1930s with glass fiber reinforced resins. In general, long fibers in various forms are inherently much stiffer Ordinary plate glass fractures at

and stronger than the same material in bulk form.

stresses of only a few tens of MPa, yet glass fibers, in commercially available forms, have strengths of 3 to 5GPa and about 7GPa in laboratory prepared forms. The paradox

of a fiber having different properties from bulk is due to the more perfect structure of a fiber. The crystals are aligned in the fiber along the fiber axis. Moreover, there are

fewer internal defects in fibers than in bulk material (Jones, 1999). Because of its light weight and high strength, glass fiber has long been adopted intensively in composite manufacturing in several applications. In order to take advantage of its directional

properties, glass fibers are also woven into layers to enhance its overall performance as the S2-glass fibers in S2 Glass Fiber Reinforced Polymer Composites.

16

Although GRPs were introduced in the 1930s, the dynamic response of these systems was not the focus until the 1970s when drop-weight testing machines were to estimate their impact strength. Lifshitz (1976) investigated the tensile strength and failure modes of unidirectional and angle-ply E-glass fiber-reinforced epoxy matrix composites at strain rates between 0.1 ~ 200 s-1. The failure stresses under impact loading conditions were found to be considerably higher when compared to those obtained under quasi-static loading conditions. In recent years the dynamic response of glass-fiber reinforced composites has been investigated utilizing the Split Hopkinson Pressure Bars (SHPBs) under relatively simple states of stress, e.g., uniaxial compression, uniaxial tension, and pure shear (Elhabak, 1991; Agbossou et al., 1995; Tay et al., 1995; Barre et al., 1996; Sierakowski, 1997; Gama et al., 2001; Song et al., 2002; Vural and Ravichandran, 2004). In these studies the failure and ultimate strength of the GRP composites were found to increase with increasing strain rates.

Most GRP material systems have excellent strength along the fiberglass direction. However, the cohesion between the fiberglass reinforcement and the resin matrix is not very strong, thereby making them susceptible to spall during a typical impact process. Spallation is the failure of material due to the action of tensile stresses developed in the interior of a sample through the interaction (overlap) of two release waves (Gray, 2000), or more specifically the process of internal failure or rupture of continuum media through a mechanism of decohesion due to stresses in excess of the tensile strength of the material
17

(Grady and Kipp, 1993). In the past, plate impact experiments and/or direct contact explosives methodologies have been employed to investigate the spall strength in materials. The main advantage of these techniques is that in these experiments Consequently, during the time

nominally plane waves of uniaxial strain are utilized.

duration of interest, the applied loading is homogeneous in the central part of the specimen. The spall strength determined in this manner is thus the pure tensile stress

necessary to pull the constituents of the composite apart. Additionally, the location of the spall plane in the specimen (where the tensile stresses are operative), can be precisely controlled by proper selection of the experimental configuration. In the past, using plate

impact experiments with Al 6061-T6 flyer plates, Dandekar et al. (1998) studied the spall strength in S2 glass woven roving in Cycom 4102 polyester resin matrix subjected to shock compression and combined shock compression and shear loading. Zaretsky et al. (2004) also utilized Al 6061-T6 flyer plates to conduct plate impact experiments on a woven glass-fiber reinforced composite in a 7781 epoxy resin matrix. The measured spall strengths were observed to vary from 60 MPa (Dandekar et al., 1998) to about 190 MPa (Zaretsky et al., 2004).

1.4 Bulk Metallic Glasses (BMG) Amorphous metals, also referred to as metallic glasses, differ from ordinary metals in that their constituent atoms are not arranged on a crystalline lattice. Due to their

randomly-ordered atomic structures, metallic glasses are known to exhibit unusual

18

mechanical properties, such as near theoretical strength, large elastic strains (Figure 1.3), high hardness, excellent wear and corrosion resistance, and increased fracture toughness when compared to other brittle, high compressive strength materials (Bruck et al., 1996, Lowhaphandu and Lewandowski, 1998). In crystalline materials, macroscopic plastic

deformation is achieved by the movement and interaction of numerous dislocations that have definite slip systems or by other mechanisms such as twinning. Moreover, dislocations also interact with inclusions and grain boundaries in polycrystalline materials, which give rise to work hardening behavior. However, metallic glasses have long-range disordered atomic configuration and dislocation mechanisms can not be used to explain the observed inelastic deformation in such systems (Leamy et al., 1972).

Figure 1.3: Plot of strength vs. elastic limit for various materials, including metallic glasses.

19

Unlike amorphous polymers, metallic glass is extremely difficult to be formed in most alloy systems. There does not exist a universal rule for predicting the glass The discovery of the first metallic glass at

forming ability of any given metallic system.

Caltech by Duwez and coworkers was somewhat of an unexpected result during their study of solid solubility in a binary noble metallic system (Klement et al., 1960). Since this first report of metallic glass formation by rapid solidification of liquid metallic alloys, there has been extensive effort to develop metallic alloys which exhibit high resistance to crystallization in the undercooled liquid state.

The maximum size of a metallic glass specimen that can be achieved without crystallization is mainly dependent on the critical cooling rate of the glass. No metallic

glass has ever been made in laboratory using a pure metallic element like Ni whose critical cooling rate is estimated to be as high as 1010 K/s (Chen, 1976). As mentioned earlier, the first metallic glass (Klement et al., 1960), Au75Si25 binary system, was formed at Caltech by rapid gun quenching technique that provides a quenching rate of 107 K/s. This 10 m thick Au-Si metallic glass thus obtained was unstable even at room temperature. Cohen and Turnbulls theory (Cohen and Turnbull, 1961; Turnbull, 1969) They pointed

provided a roadmap for the development of other new metallic glasses.

out that strong metallic glass formers favor deep-eutectic compositions and high reduced glass transition temperature, Trg=Tg/Tm, where Tm is the melting temperature of the glass.

20

Early searches for metallic glass formers were focused primarily on binary late transition metal-metalloid type, binary transition metal type deep eutectic systems until Chen and Turnbull (Chen and Turnbull, 1969; Chen, 1974, 1976) discovered thick metallic glass (1 to 3 mm) in Pd-M-Si and Pd-T-P systems (where M=Rh, Au, Ag and Cu and T=Ni, Co, Fe) using water quenching technique which had a cooling rate of around 102 to 103 K/s. Drehman et al. (1982) demonstrated the making of an amorphous Pd40Ni40P20 rod of 5.3 mm in diameter at a cooling rate of 1.4 K/s, using surface etching and thermal cycling technique. Kui et al. (1984), by using a flux treatment (molten

surface flux of dehydrated boron oxide) and thermal cycling to suppress heterogeneous peculations, increased the maximum thickness of Pd40Ni40P20 to 10 mm at an approximate cooling rate of 4 K/s. This appears to be the largest metallic glass ever made before 1990s. However, this tedious fabrication technique is much more suitable

for laboratory synthesis, rather than for industrial productions. In addition, the thermal stability of Pd40Ni40P20 was not improved at all and further search for better formers was necessary.

The development of bulk metallic glasses, which are strong glass formers that can be easily formed in centimeter scale, started in the late 1980s and early 1990s in two research groups at Tohoku University and Caltech. Inoue (2001) discovered

multi-component liquid alloys (Mg-TM-Ln, Ln-Al-TM and Zr-Al-TM, TM= transition metal) with very deep eutectics that were capable of freezing to a glassy state of several
21

centimeters in thickness by utilizing conventional cooling methods with critical cooling rates ranged from 101 to 102 K/s. At around the same time, at Caltech, Johnsons group developed Zr-Ti-based and other sizable amorphous metals. These Zirconium-based

bulk metallic glasses have been of particular interest to the engineering community because of their low critical cooling rate (1 K/s) requirements, allowing thick samples (up to 10 mm) to be processed (Peker and Johnson, 1993).

1.4.1 Zr41.25Ti13.75Ni10Cu12.5Be22.5 (LM-1 or Vitreloy 1)

Peker and Johnson at Caltech discovered a family of bulk metallic glasses based on the Zr-Ti-Cu-Ni-Be system (Vitreloy family). One extensively studied example of this

glass family is Vitreloy 1, Zr41.25Ti13.75Ni10Cu12.5Be22.5 (Peker and Johnson, 1993). Direct experimentally study has shown that this glass has a critical cooling rate of 1 K/s and it can be cast in rods with a diameter of up to 5 cm (Johnson, 1996). This alloy is one of the strongest glass forming alloys to date without a noble metallic element.

The Zr-based metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5, was alloyed by induction melting on a water cooled silver boat under a Ti-gettered argon atmosphere. Ingots of typical size of 5-6 g were prepared. The initial samples were generally found to freeze without any crystallization during preparation resulting in a glassy ingot. The ingot

samples were further processed by casting into copper molds under an inert gas

22

atmosphere and by melting of several ingots together in sealed silica tubes followed by water quenching. Uniform plates, strips, and bars, as well as rods were made by these methods in a variety of shapes and sizes.

Vickers hardness measurements on the Vitreloy 1 samples show a typical value of 585 kg/mm2 (Peker and Johnson, 1993). The details of the physical and mechanical

properties of Vitreloy 1 are listed in Table 1.1 (Peker and Johnson, 1993; Lu, 2002). The elastic wave speeds of the materials were obtained using ultrasonic measurements.

Density

Longitudinal wave speed

Shear wave speed

Bulk wave speed 4335 m/s

Elastic Modulus

Shear Modulus

Bulk Modulus

Poissons ratio

6000 Kg/m3

5185 m/s

2464 m/s

98.6 GPa 36.4 GPa 113 GPa

0.354

Table 1.1: Physical properties of the Zr-based BMG (Vit-1) used in the present investigation.

1.4.2 The Dynamic Response of BMGs The dynamic response of BMGs is of considerable interest for gaining insight into the high strain-rate response of this class of materials and for potential applications, such as kinetic energy penetrators (Johnson, 1999). However, in the past, the mechanical properties and structural performance of BMGs under dynamic loading conditions have

23

been studied by a limited number of investigators.

For a Zr-based BMG, i.e.

Zr41.25Ti13.75Ni10Cu12.5Be22.5, Lu (2002) reported the strength under dynamic and quasi-static loading conditions to be quite similar, with failure being characterized by predominantly inhomogeneous inelastic flow. In addition, temperature increases of up to 500o K were detected during the dynamic deformation and failure, suggesting that perhaps local heating of the material contributed to material softening. Moreover, in both Zr57Ti5Cu20Ni8Al10 and Zr41.25Ti13.75Ni10Cu12.5Be22.5 BMGs, strain-rate softening behavior has been reported (Hufnagel et al., 2002). Such behavior was also observed More recently, Sunny

when zirconium was replaced by hafnium (Subhash et al., 2003).

et al. (2005a, 2005b, 2006a, 2006b) have investigated the high strain-rate behavior of Zr41.25Ti13.75Ni10Cu12.5Be22.5 by utilizing a Split-Hopkinson Pressure Bar. In their study,

in-situ video was performed to examine the deformation modes and the models of failure under uniaxial dynamic compression. The fully amorphous material was observed to

exhibit catastrophic failure with the formation of a dominant shear band and the corresponding slip event, while the annealed BMG was observed to fail by extensive fragmentation after the formation of an initial crack.

Even though the deformation and damage mechanisms of BMGs under moderate dynamic loading conditions have been investigated, only a few studies have addressed the shock response in BMGs. Bach et al. (1991) studied parameters for the shock wave

consolidation of a metallic glass powder, such as the effects of the shock wave energy
24

and shock wave duration.

Conner et al. (2000) investigated the high strain rate behavior

of fiber-reinforced Zr41.2Ti13.8Cu12.5Ni10Be22.5 composites, and demonstrated their excellent potential as armor penetrators. Zhuang et al.(2002) conducted plate impact experiments to investigate the shock response of a Zr-based bulk metallic glass, Zr41.2Ti13.8Cu12.5Ni10Be22.5 (Vitreloy-1), and its particulate composite,

Zr56.3Ti13.8Ni5.6Cu6.9Be12.5 (in-situ dendritic -phase reinforced Vitreloy (Hays et al., 2000)), which are to be respectively referred to as Vit-1 and -Vit hereafter. A

surprisingly low amplitude elastic precursor and bulk wave were observed to precede the rate-dependent large deformation shock wave. Moreover, a concave downward

curvature in the shock Hugoniot for the BMG was observed, suggesting that a phase-change-like transition occurred during the shock compression event. The spalling

in Vit-1 was induced by shear localization, while in -Vit it was due to debonding of the -phase boundary from the matrix. The spall strengths, at a stain rate of 2106 s-1, were

determined to be 2.35 GPa and 2.11 GPa for Vit-1 and -Vit, respectively. Turneaure et

al. (2004) conducted plane shock wave experiments up to 13 GPa on a Zr-based bulk
amorphous alloy, Zr56.7Cu15.3Ni12.5Nb5.0Al10.0Y0.5, with a quasi-static strength of 2.6 GPa. From the measured particle velocity histories, the Hugoniot elastic limit (HEL) was determined to be ~ 7.1 GPa, corresponding to an elastic strain of approximately 4%. For the experiments in which the peak stress exceeded the HEL, a clear two-wave structure consisting of an elastic precursor followed by a plastic wave was observed. Moreover, the experimental results suggested that the shear strength of the Zr-based
25

BMG is reduced as it is shocked above the elastic limit.

Yang et al. (2005) employed a

two stage light gas-gun to investigate the effects of planar shock compression on void formation and cracking in a Zr-based BMG, Zr41Ti14Ni10Cu12.5Be22.5 under hypervelocity impact conditions. Cracking was proposed to be a result of void linkage in a direction that did not coincide with the maximum shear stress plane. Changing the state of stress within the BMG by employing a spherical nosed projectile was proposed to lead to the formation of a different nanovoid distribution. The authors proposed that the nucleation of the nanovoids was possibly initiated by release of excess free-volume under shock wave compression. More recently, Yang et al. (2006) investigated the damage features in Zr41Ti14Ni10Cu12.5Be22.5, subjected to hyper-velocity impact using a two-stage light gas-gun. Using scanning electron microscopy they showed that both radial and

symmetric cracks were formed on the shocked surface of the target plate when impacted by an aluminum flyer with an impact velocity of 2.7 Km/s. Shear bands/cracks parallel

to each other, on the cross section close to the shocked surface of the target, were also observed. The damage features under the projectiles nose were also examined. It was shown that besides the formation of adiabatic shear bands/cracks, craters and lamination cracks were also formed, and the depth of the craters increased with increasing projectile velocity. Mashimo et al. (2006) extended the investigation of BMG under planar shock compression to pressures up to 50 GPa. Using a powder gun and an inclined-mirror

photographic technique they investigated the yield behavior and the phase change in a Zr-based BMG, Zr55Al10Ni5Cu30. They measured an HEL of 6.2 GPa.
26

Moreover, a

kink was observed in the shock velocity versus particle velocity relationship (Hugoniot) at about 400 m/s in particle velocity, suggesting the occurrence of a shock-induced phase transformation, although no determination of the resulting phase(s) were provided.

1.4.3 Pressure Effects on Flow and Fracture Behavior of BMGs The study of pressure effects on flow and fracture behavior of BMGs is important because of a number of reasons. One key reason is that deformation in amorphous

metals is a dilatant process (Lewandowski, J. J. and Greer, A. L., 2006); while the atoms in the alloy move relative to each other, free volume is created and destroyed. Under

the influence of hydrostatic compressive stress, one might expect that the free volume generation to be inhibited, thereby increasing the plastic strain-to-failure of the alloy. Another key reason comes from potential applications, such as underwater vehicles (because of both the corrosion resistance and the hydrostatic stresses present) and composites. Separate papers by Connor et al. (1977) and Vormelker et al. (1998) show In such cases, deformation of the

promise for composites incorporating a metallic glass.

tungsten provides confinement on the metallic glass and leads to high hydrostatic stresses in the metallic glass.

The earliest known paper to characterize the effects of pressure on a metallic glass was by Davis and Kavesh (1975), who studied the effects of hydrostatic pressure on Pd77.5Cu6Si16.5 metallic glass. Quasi-static compression tests were performed on 2 mm
27

diameter rods with the L/D = 2.0, with a tapered shape similar to that of a dog-bone specimen. length. The gage section of the specimen was 0.84 mm in diameter and 1.1 mm in

These test were performed by a liquid-based high-pressure mechanical testing It is a miniature, synchronous motor-driven device which operates

apparatus.

completely enclosed in a high pressure vessel. In these tests, an increase in pressure from 100 MPa to 650 MPa led to a small increase (from 1.45 GPa to 1.55 GPa) in the compressive strength of the material. While no attempt was made to characterize the

fracture angle in compression and no macroscopic fracture surfaces were presented. The slight increase in the compressive stress with hydrostatic pressure suggests that there is some dependence on the normal stress at the fracture plane, and this dependence is best modeled by a Mohr-Coulomb criterion of the following form, n = 0 n . Where

n and n refer to the shear stress and normal stress along the fracture plane, 0 is
the shear stress under pure shear, and is the value of the normal stress dependence. Since the value of is small (0.044), Davis and Kavesh did not see any influence of normal stress and the pressure dependence on fracture strength and/or mechanism. By

using the same miniature testing apparatus as compression tests, 0.3 mm diameter fibers/foils were loaded in tension at 6 x 10-5/s at both ambient pressure and 700 MPa superimposed pressure. In both tests, the tensile stress-strain behavior was observed to very similar to brittle materials, such as ceramics, with mostly linear elastic behavior and very limited nonlinearity just prior to specimen failure. Moreover, the addition of 700

MPa of hydrostatic pressure led to a slight increase in the tensile stress (from 1.3 GPa to
28

1.45 GPa, only slightly below the fracture stress in compression).

Fracture occurred

near the loading grips, and a fracture angle of approximately 50 was exhibited along a curved fracture surface

Another early attempt to infer the effects of pressure was performed by Donovan (1988) on samples of Pd40Ni40Pd20 metallic glass. In her work, 1-2 mm diameter

samples of Pd40Ni40P20 metallic glass were tested under uniaxial compression, modified plane-strain compression, pure shear and tension. The results suggest that yielding in

Pd40Ni40Pd20 also follows a Mohr-Coulomb criterion by the form n = 0 n , rather than the von Mises criterion which is appropriate to polycrystalline metals. The sign of the normal stress acting across the slip-plane has a significant effect on the yield strength so that the yield strength in compression is substantially higher than in tension, but the hydrostatic pressure has only a small effect on yielding. After considering all

experimental data, a value of of 0.113 was determined, along with a flow stress in shear of 0.795 GPa.

By using a gas-based high pressure apparatus, Lewandowski and Lowhaphandu et al. (1999, 2000a, 2000b, 2002) investigated the pressure effects on fracture behavior of Zr41.25Ti13.75Cu12.5Ni10Be22.5 and Zr63Cu18Ni10Al9 under both quasi-static compression and tension testing. The compression testing was conducted under a large range of

hydrostatic pressures (from 100 MPa to upwards of 1 GPa), while the tension testing was
29

conducted under both ambient pressure and superimposed hydrostatic pressures up to 700 MPa. In the compression testing, the addition of large hydrostatic pressures increased

the strain-to-failure of the metallic glasses. The possible reason is that free volume generation may be inhibited under the influence of hydrostatic pressures, thereby increase the plastic strain-to-failure of the metallic glasses. The compression specimens While

exhibited fracture angles of approximately 41 regardless of confining pressure.

the addition of pressure changed fracture angles significantly in the tension testing. Tension specimens tested in ambient conditions exhibited a fracture angle largely oriented at 90, but with two small shear lips oriented at approximately 50. However,

all of the tension specimens tested with hydrostatic pressure except for one exhibited fracture angles between 56 and 59, while one that failed near the grips exhibited a fracture angle close to 50. Even with a net compressive hydrostatic stress, a 15 difference was exhibited between the fracture angles of the tension specimens (55) and the compression samples (40), again providing evidence for yielding behavior following the Mohr- Coulomb criterion. After considering both the tension and compression data,

a value of of 0.04 was determined, along with an expected flow stress in shear of 0.97 GPa, which, while lower than that measured during a torsion test, corresponds well with the 2.0 GPa compressive yield stress that was measured in the same paper.

In order to investigate the effect of confinement pressures on the quasi-static flow behavior of Zr41.25Ti13.75Cu12.5Ni10Be22.5, Lu et al. (2002) utilized maraging steel
30

confining sleeves around disk shaped specimens.

By varying (increasing) the ratio of

the outer radius of the confining sleeve to the inner radius, they were able to achieve confining pressures up to to about 2 GPa on the specimens. However, in their

configuration, unlike the experiments with superimposed liquid or gas pressures, a constant pressure cannot always be achieved on the specimen since the confinement depends on the lateral expansion of the metallic glass (specimen) and the confining sleeve during compression. In the experiments, it was reported that the flow stress of

metallic glass was increased substantially when the pressure was increased -- from 1.9 GPa under ambient pressure and room temperature conditions to 2.5 GPa under 2.0 GPa confining pressure.

More recently, experiments have been performed by Zhang and Eckert (2003a, 2003b) to investigate the normal-stress and pressure dependency of flow stress in Zr59Cu20Al10Ni8Ti3 metallic glass. In addition to the different fracture morphologies in compression and tension, they claimed that the observed larger deviation (from 45) in tension (54) when compared to that in compression (40) require the tensile normal stresses to have a more severe effect on the flow stress than the compressive stresses. Furthermore, to account for this tension/compression asymmetry they postulated a different Mohr-Coulomb relationship in tension than in compression (2003a), i.e.

n = 0 c n ,c and n = 0 t n ,t for compression and tension, respectively,

where c and t are the constants representing the normal stress dependency in
31

compression and tension, respectively, with c < t .

However, when looking at the behavior of a different metallic glass (Lowhaphandu

et al., 1999; Lowhaphandu et al., 2000a; Lowhaphandu et al., 2000b; Lewandowski, J. J.,
and Lowhaphandu, P., 2002; Wesseling, et al., 2003), it appears that the values of these two constants are nearly identical.

1.5 Dissertation Outline In chapter 2, a series of plate-impact pressure-shear friction experiments and the modified torsional Kolsky bar friction experiments are conducted to investigate frictional resistance in Arkansas Novaculite rock and analog materials such as quartz and soda-lime glass. The objective of these experiments is to investigate the frictional resistance of

relevant geo- and analog materials at co-sesimic slip speeds and higher, so as to better understand the role of high slip speeds in leading to slip weakening at the slip interface by the mechanism of asperity flash heating and associated thermal softening with slip. The results of this work have relevance not just to mechanisms of dynamic fault weakening, but also to the constitutive description of the behavior that can be used to critically examine the rate and state dependent dynamic friction models of earthquake rupture.

In chapter 3, we capitalize on the aforementioned recent developments (Prakash and

32

Clifton, 1993; Prakash, 1995; Irfan and Prakash, 2000; Okada et al., 2001) on the plate impact pressure-shear friction experiments to investigate dynamic slip resistance and time-resolved growth of molten metal films during dry metal-on-metal slip under extreme interfacial conditions. By employing a tribo-pair comprising of hard tool-steel against relatively low melt-point metal (7075-T6 Al alloy), interfacial friction stress of up to 300 MPa and slip speeds of approximately 250 m/s have been achieved. These relatively extreme interfacial conditions are conducive to the development of molten metal films at the tribo-pair interface. A Lagrangian finite-element code is used to understand the evolution of the thermo-mechanical fields and their relationship to the observed slip response. The code accounts for dynamic effects, heat conduction, contact with friction, and full thermo-mechanical coupling. At temperatures below the melting point the

material is described as an isotropic thermally softening elastic-viscoplastic solid. For material elements with temperatures in excess of the melt temperature a purely Newtonian fluid constitutive model is employed. The results of this hybrid

experimental-computational approach provide new insights into the thermo-plastic interactions during the high-speed metal-on-metal slip.

In chapter 4, normal impact and the combined pressure-shear plate impact experiments are conducted to investigate the spall strength in two different architectures of the GRP composites S2 glass woven roving in Cycom 4102 polyester resin matrix and a 5-harness satin weave E-glass in a Ciba epoxy (LY564) matrix.
33

The spall strength

of the GRP was obtained as a function of the normal component of impact stress and the applied shear strain by subjecting the GRP specimens to shock compression and combined shock compression and shear loading, respectively.

In chapter 5, Results are presented on the shock response of a zirconium-based bulk metallic glass (BMG), Zr41.25Ti13.75Ni10Cu12.5Be22.5, subjected to shock compression and combined shock compression and shear loading. The first series of experiments are

conducted to study the spall strength of the Zr-based BMG under shock compression and combined shock compression and shear loading. The particle velocity profiles,

measured at the back surface of the target plate by using the VALYNTM VISAR, are analyzed to (1) better understand the structure of shock waves in BMG subjected to planar shock compression, (2) obtain the Hugoniot elastic limit (HEL) of the material, (3) estimate residual spall strength of the BMG under different levels of normal stress and shear strain. The second series of experiment involve the skew impact of a thick WC

flyer plate with a stationary target. The target comprises a front and a back WC plate with a thin BMG specimen sandwiched in between. This leads to pressure-shear shock

loading of the sandwiched specimen. From the measured free surface particle velocities, the flow stress in shear under ultrahigh strain rate (~105 /s) , normal pressure (up to 9 GPa) and hydrostatic pressure (up to 6 GPa) is obtained for the Zr-based BMG.

The results, discussion are summarized in Chapter 6.

34

REFERENCES

Adams, G. G., 1995. Self excited oscillations of two elastic half-spaces sliding with a constant coefficient of friction. Journal of applied mechanics 62, 867-872. Agbossou, A., Cohen, I., and Muller, D., 1995. Effects of Interphase and Impact Strain Rates on Tensile Off-Axis Behavior of Unidirectional Glass-Fiber Composite Experimental Results. Engineering Fracture Mechanics 52 (5), 923. Andrews, D. J., and Ben-Zion, Y., 1997. Wrinkle-like slip pulse on a fault between different materials. Journal of Geophysical Research 102, 553-571. Andrews, D. J., 2002. A fault constitutive relation accounting for thermal pressurization of pore fluid. Journal of Geophysical Research 107, 2363, doi: 2310.1029 / 2002JB001942, ESE 001915-001941001915-001948. Archard, J. F., 1958. The temperature of rubbing surfaces. Wear 2, 438-455. Ashby, M. F., Abulawi, J., and Kong, H. S., 1991. Temperature maps for frictional heating in dry sliding. Tribology transactions 34, 577-587. Bach, J., Krueger, B., and Fultz, B., 1991. Shock wave consolidation of a Ni-Cr-Si-B metallic-glass powder. Materials Letters 11, 383-388. Barber, J. R., 1970. The conduction of heat from sliding solids. International Journal of Heat and Mass Transfer 13, 857-869. Barber, J. R., 1976. Some thermoelastic contact problems involving frictional heating. The Quarterly Journal of Mechanics and Applied Mathematics 29, 1-13. Barre, S., Chotard, T., and Benzeggagh, M. L., 1996. Comparative study of strain rate effects on mechanical properties of glass fibre-reinforced thermoset matrix composites. Composites Part a-Applied Science and Manufacturing 27 (12), 1169-1181. Ben-Zion, Y., 2001. Dynamic ruptures in recent models of earthquake faults. Journal of Mechanics and Physics of Solids 49, 2209-2244. Betheney W., DeLuca E., Prifti J. and Chou S. C., 1998. Ballistic impact damage of S2-glass reinforced plastic structural armor. Composites Science and Technology 58, 1453-1461. Blanpied, M. L., Tullis, T. E., and Weeks, J. D., 1998. Effects of slip, slip rate, and shear heating on the friction of granite. Journal of Geophysical Research-Solid Earth 103, 489-511. Boitnott, G. N., Biegel, R. L., Scholz, C. H., N., Y., and W., W., 1992. Micromechanics of rock friction 2: Quantitative modeling of initial friction with contact theory. Journal of
35

Geophysics Research 97, 8965-8978. Bowden, F. P., and Ridler, K. E. W., 1936. Physical properties of surfaces, III. Proceedings of Royal Society of London A 154, 640-656. Bowden, F. P., and Tabor, D., 1950. The Friction and Lubrication of Solids. Oxford University Press, London. Bowden, F. P., and Thomas, P. H., 1954. The surface temperature of sliding solids. Proceedings of Royal Society of London A 223, 29-36. Bowden, F. P., and Freitag, E. H., 1958. The friction of solids at very high speeds, I. Metal on metal, II. Metal on diamond. Proceedings of Royal Society of London A248, 350-367. Bowden, F. P., and Persson, P. A., 1960. Deformation heating and melting of solids in high speed friction. Proceedings of Royal Society of London A260, 433-458. Bruck, H. A., Rosakis, A. J., and Johnson, W. L., 1996. The dynamic compressive behavior of beryllium bearing bulk metallic glasses. Journal of Materials Research 11, 503-511. Bureau, L., Baumberger, T., and Caroli, C., 2000. Shear response of a frictional interface to a normal load modulation. Physics Review Letters E 62, 6810-6820. Byerlee, J. D., 1978. Friction of rocks. Pure Applied Geophysics 116, 615-626. Carslaw, H. S., and Jaeger, J. C., 1986. Conduction of heat in solids. Oxford University Press, London. Chen, H. S. and Turnbull, D., 1969. Formation, stability and structure of palladium-silicon based alloy glasses. Acta Metall. 17, 1021-1031. Chen, H. S., Metallic glass. 1976. Mater. Sci. Eng., 25, 59-69. Chen, H. S., 1974. Thermodynamic considerations on the formation and stability of metallic glasses. Acta Metall. 22, 1505-1511. Chester, J. S., Kronenberg, A. K., Chester, F. M., and Guillemette, R. N., 2003. Characterization of natural slip surfaces relevant to earthquake mechanics. EOS Trans. AGU, 84(46), Fall Mtg. Suppl., Abstract S42C-0185. Chester, J. S., Chester, F. M., Kronenberg, A. K., and Guillemette, R. N., 2005. Extreme localization of slip and implications for dynamic weakening of faults, Manuscript in preparation. Chester, J. S., and Goldsby, D. L., 2003. Microscale characterization of natural and experimental slip surfaces relevant to earthquake mechanics, SCEC Annual Report.

36

Cocco, M., and Rice, J. R., 2002. Pore pressure and poroelasticity effects in Coulomb stress analysis of earthquake interactions. Journal of Geophysical Research, 107(B2), cn:2030, doi:10.1029/2000JB000138, pp. ESE 2-1 to 2- 17. Cochard, A., and Rice, J. R., 2000. Fault rupture between dissimilar materials: Ill-posedness, regularization, and slip-pulse response. Journal of geophysical research 105, 891-907. Cohen, M. H. and Turnbull, D., 1961. Composition requirements for glass formation in metallic and ionic system. Nature 189, 131-132. Connor, R. D., Dandliker, R. B. and Johnson, W. L., 1998. Mechanical properties of tungsten and steel reinforced Zr41.25Ti13.75Cu12.5Ni10Be22.5 bulk metallic glass composites. Acta Mat. 46 (17), 6089-6102. Conner, R. D., Dandliker, R. B., Scruggs, V., and Johnson, W. L., 2000. Dynamic deformation behavior of tungsten-fiber/metallic-glass matrix composites. International Journal of Impact Engineering 24, 435-444. Dandekar, D. P., and Beaulieu, P. A., 1995. Compressive and tensile strengths of glass reinforced polyester under shock wave propagation. In: Rajapakse, Y. D. S., and Vinson, J. R., (Eds.), High strain-rate effects on polymer, metal, and ceramic matrix composites and other advanced materials. ASME, New York, NY, pp. 63-70. Dandekar, D. P., Boteler, J. M., and Beaulieu, P. A., 1998. Elastic constants and delamination strength of a glass-fiber-reinforced polymer composite. Composites Science and Technology 58 (9), 1397-1403. Davis, L. A. and Kavesh, S., 1975. Deformation and fracture of an amorphous metallic alloy at high pressure. J. Mat. Sci.10 (3), 453-459. DeLuca, E., Prifti, J., Betheney, W., and Chou, S. C., 1998. Ballistic impact damage of S 2-glass-reinforced plastic structural armor. Composites Science and Technology 58 (9), 1453-1461. Di Toro, G., Goldsby, D. L., and Tullis, T. E., 2004. Friction falls towards zero in quartz rock as slip velocity approaches seismic rates. Nature 427, 436-439. Dieterich, J. H., 1978. Time dependent friction and mechanics of stick-slip. Pure and Applied Geophysics 116, 668-675. Dieterich, J. H., 1979. Modeling of rock friction: I, experimental results and constitutive equations. Journal of Geophysical Research 84, 2161-2168. Dieterich, J. H., 1981. Constitutive properties of faults with simulated gouge. In: Carter, N. L., Friedman, M., Logan, J. M., and Stearns, D. W., (Eds.), vol. 24. American Geophysical Union, Washington, D.C., pp. 103-120.
37

Dieterich, J. H., and Kilgore, B. D., 1994. Direct observation of frictional contacts; new insights for state dependent properties. Pure Applied Geophysics 143, 283-302. Dieterich, J. H., and Kilgore, B., 1996. Implications of fault constitutive properties for earthquake prediction. Proceedings of the National Academy of Sciences of the United States of America 93, 3787-3794. Dieterich, J. H., and Kilgore, B. D., 1996. Imaging surface contacts: Power law contact distributions and contact stresses in quartz, calcite, glass and acrylic plastic. Tectonophysics 256, 219-239. Donovan, P.E., 1988. A yield criterion for Pd40Ni40P20 metallic glass. Acta Met. 37, 445-456. Drehman, A. L., Greer, A. L. and Turnbull, D., 1982. Bulk formation of a metallic glass: Pd40Ni40P20. Appl. Phys. Lett. 41 (8), 716-717. Earles, S. W. E., and Powell, D. G., 1966. Variations in friction and wear between unlubricated steel surfaces. 181, Part 30, 171-179. Elhabak, A. M. A., 1991. Mechanical-Behavior of Woven Glass Fiber-Reinforced Composites under Impact Compression Load. Composites 22 (2), 129-134. Fink, B. K., 2000. Performance Metrics for Composite Integral Armor. Journal of Thermoplastic Composite Materials 13 (5), 417-431. Gama, B. A., Bogetti, T. A., Fink, B. K., Yu, C. J., Claar, T. D., Eifert, H. H., and Gillespie, J. W., 2001. Aluminum foam integral armor: A new dimension in armor design. Composite Structures 52 (3-4), 381-395. Gama, B. A., Gillespie, J. W., Mahfuz, H., Raines, R. P., Haque, A., Jeelani, S., Bogetti, T. A., and Fink, B. K., 2001. High Strain-Rate Behavior of Plain-Weave S-2 Glass/Vinyl Ester Composites. Journal of Composite Materials 35 (13), 1201-1228. Goldsby, D., and Tullis, T. E., 2002. Low frictional strength of quartz rocks at sub-seismic slip rates. Geophysical research letters 29, 1844. Grady, D. E., and Kipp, M. E., 1993. High-Pressure Shock Compression of Solids. Springer-Verlag, Berlin, Germany. Grady, D. E., and Kipp, M. E., 1993. Dynamic fracture and fragmentation. In: Asay, J. R., and Shahinpoor, M., (Eds.), High-Pressure Shock Compression of Solids. Springer-Verlag, New York, pp. 265-322. Gray, G. T., 2000. Shock wave testing of ductile materials, Mechanical Testing and Evaluation Handbook, vol. 8. American Society for Metals, Materials Park, pp. 530-559. Hays, C. C., Kim, C. P., and Johnson, W. L., 2000. Enhanced plasticity of bulk metallic
38

glasses containing ductile phase dendrite dispersions. Metastable, Mechanically Alloyed and Nanocrystalline Materials-- Parts 1 and 2, vol. 343-3, pp. 191-196. Heaton, T. H., 1990. Evidence for and implications of self-healing pulses of slip in earthquake rupture. Physics of the Earth and Planetary Interiors 64, 1-20. Hirose, T., and Shimamoto, T., 2005. Growth of molten zone as a mechanism of slip weakening of simulated faults in gabbro during frictional melting. Journal of Geophysical Research-Solid Earth 110. Hirose, T., and Shimamoto, T., 2005. Slip-weakening distance of faults during frictional melting as inferred from experimental and natural pseudotachylytes. Bulletin of the Seismological Society of America 95, 1666-1673. Hsiao, H. M., and Daniel, I. M., 1996. Nonlinear elastic behavior of unidirectionai composites with fiber waviness under compressive loading. Journal of Engineering Materials and Technology-Transactions of the Asme 118 (4), 561-570. Hsiao, H. M., and Daniel, I. M., 1996. Effect of fiber waviness on stiffness and strength reduction of unidirectional composites under compressive loading. Composites Science and Technology 56 (5), 581-593. Hufnagel, T. C., Jiao, T., Li, Y., Xing, L. Q., and Ramesh, K. T., 2002. Deformation and failure of Zr57Ti5Cu20Ni8Al10 bulk metallic glass under quasi-static and dynamic compression. Journal of Materials Research 17, 1441-1445. Inoue, A., and Hashimoto, K., 2001. Amorphous and Nanocrystalline Materials: Preparation, Properties and Applications. Springer-Verlag, Berlin, Germany. Irfan, M. A., and Prakash, V., 1994. Contact temperatures during sliding in pressure shear impact. Proceedings Society of Experimental Mechanics Conference, Baltimore, MD, pp. 173-182. Irfan, M. A., and Prakash, V., 2000. Time resolved friction during dry sliding of metal on metal. International journal of solids and structures 37, 2859-2882. Johnson, W. L., 1999. Bulk Metallic Glasses. In: Johnson, W. L., Inoue, A., and Liu, C. T., (Eds.), MRS Symposium Proceedings, vol. 554. Materials Research Society, pp. 311-339. Jones, R. M., 1999. Mechanics of Composite Materials. Taylor & Francis, Philadelphia. Kadhim, M. J., and Earles, S. W. E., 1966. Unlubricated sliding at high speeds between copper and steel surfaces. Proceedings institute of mechanical engineers 181, Part 30, 157-162. Kennedy, F. E., 1984. Thermal and thermomechanical effects in dry sliding. Wear 100, 453-476.

39

Klement, W. J., Willens, R. H. and Duwez, P., 1960. Non-crystalline structure in solidified gold-silicon alloys. Nature, 187, 869-870. Krafft, J. M., 1955. Surface friction in ballistic penetration. Journal of applied physics 26(10), 1248-1253. Kuhlmann-Wilsdorf, D., 1985. Flash temperatures due to friction and joule heat at asperity contact. Wear 105, 187-198. Kui, H. W., Greer, A. L. and Turnbull, D., 1984. Formation of bulk metallic-glass by fluxing. Appl. Phys. Lett. 45 (6), 615-616. Lachenbruch, A. H., 1980. Frictional heating, fluid pressure, and the resistance to fault motion. . Journal of Geophysical Research 85, 60976122. Leamy, H. J., Chen, H. S. and Wang, T. T., 1972. Plastic flow and fracture of metallic glass. Metallurgical Transaction, 3, 699-708. Lee, T. C., and Delaney, P. T., , 1987. Frictional heating and pore pressure rise due to a fault slip Geophys J. Roy. Astronom. Society 88, 569-591. Lewandowski, J. J., and Lowhaphandu, P., 1998. Effects of hydrostatic pressure on mechanical behavior and deformation processing of materials. International material reviews 43, 145-187. Lewandowski, J. J., and Lowhaphandu, P., 2002. Effects of hydrostatic pressure on the flow and fracture of a bulk amorphous metal. Philosophical magazine A82, 3427-3441. Lewandowski, J. J. and Greer, A. L., 2006. Temperature rise at shear bands in metallic glasses. Nat. Mat. 5, 15-18. Lifshitz, J. M., 1976. Impact strength of angle ply fiber reinforced materials. Journal of Composite Materials 10, 92-101. Lim, S. C., Ashby, M. F., and Brunton, J. H., 1989. The effects of sliding conditions on the dry friction of metals. Acta Materialia 37, 767-772. Linker, M. F., and Dieterich, J. H., 1992. Effects of variable normal stress on rock friction: observations and constitutive equations. Journal of Geophysical Research 97, 4923-4940. Liou, N. S., Okada, M., and Prakash, V., 2004. Formation of Molten Metal Films During Metal-on-Metal Slip Under Extreme Interfacial Conditions. Journal of the Mechanics and Physics of Solids 52, 2025-2056. Lowhaphandu, P., and Lewandowski, J. J., 1998. Fracture toughness and notched toughness of bulk amorphous alloy: Zr-Ti-Ni-Cu-Be. Scripta Materialia 38, 1811-1817. Lowhaphandu, P., Montgomery, S. L. and Lewandowski, J. J., 1999. Effects of superimposed hydrostatic pressure on flow and fracture of a Zr-Ti-Ni-Cu-Be bulk
40

amorphous alloy. Scr. Mat. 41 (1), 19-24. Lowhaphandu, P. et al., 2000a. Deformation and fracture toughness of a bulk amorphous Zr-Ti-Ni-Cu-Be alloy). Intermetallics 8, 487-492. Lowhaphandu, P., 2000b. Mechanical Behavior of a Zirconium-based Bulk Metallic Glass. Ph. D. Thesis, Case Western Reserve University, Cleveland, OH . Lu, J., 2002. Mechanical Behavior of a Bulk Metallic Glass and its Composites Over a Wide Range of Strain Rates and Temperatures, (Ph.D. Dissertation), California Institute of Technology, Pasadena, CA. Mahfuz, H., Zhu, Y., Haque, A., Abutalib, A., Vaidya, V., Jeelani, S., Gama, B., Gillespie, J., and Fink, B., 1999. Investigation of high-velocity impact on integral armor using finite element method. International journal of impact engineering 24 (2), 203-217. Marone, C., 1998. Laboratory-derived friction laws and their application to seismic faulting. Annual Review Earth and Planetary Science 26, 643-696. Mase, C. W., and Smith, L., 1985. Porefluid pressures and frictional heating on a fault surface. Pure Appl. Geophys. 122, 583-607. Mase, C. W., and L., S., 1987. Effects of frictional heating on the thermal, hydrologic, and mechanical response of a fault. J. Geophys. Res. 92, 62496272. Mashimo, T., Togo, H., Zhang, Y., Uemura, Y., and Kawamura, Y., 2006. Shock-compression behavior of Zr-based metallic glass In: Khan, A. S., and Kazmi, R., (Eds.), 12th International Symposium on Plasticity and its Applications. Anisotropy, Texture, Dislocations and Multiscale Modeling in Finite Plasticity and Viscoplasticity and Metal Forming. Neat, Inc., MD, USA, Halifax, Nova Scotia, Canada, pp. 157-159. Mizoguchi, K., Hirose, T., Shimamoto, T., and Fukuyama, E., 2006. Moisture-related weakening and strengthening of a fault activated at seismic slip rates. Geophysical Research Letters 33. Montgomery, R. S., 1976a. Friction and wear at high sliding speeds. Wear 36, 275-298. Montgomery, R. S., 1976b. Surface melting of rotating bands. Wear 38, 235-243. Noda, H., and Shimamoto, T., 2005. Thermal pressurization and slip-weakening distance of a fault: An example of the Hanaore fault, southwest Japan. Bulletin of the Seismological Society of America 95, 1224-1233. Ogawa, K., 1997. Impact friction test method by applying stress wave. Experimental Mechanics 37(4), 398-402. O'Hara, K., Mizoguchi, K., Shimamoto, T., and Hower, J. C., 2006. Experimental frictional heating of coal gouge at seismic slip rates: Evidence for devolatilization and
41

thermal pressurization of gouge fluids. Tectonophysics 424, 109-118. Okada, M., Liou, N.-S., Prakash, V., and Miyoshi, K., 2001. Tribology of high speed metal-on-metal sliding at near-melt and fully-melt interfacial temperatures. Wear 249, 672-686. Okada, M., Liou, N. S., and Prakash, V., 2002. Dynamic shearing resistance of molten metal films at high pressures. Experimental Mechanics. Peker, A., and Johnson, W. L., 1993. A highly processable metallic glass Zr41.2Ti13.8Cu12.5Ni10.0Be22.5. Applied Physics Letters 63, 2342-2344. Prakash, V., and Clifton, R. J., 1993. Time Resolved Dynamic Friction Measurements in Pressure-Shear. In: Ramesh, K. T., (Ed). Experimental Techniques in the Dynamics of Deformable Bodies, vol. AMD Vol. 165. ASME, New York, pp. 33-47. Prakash, V., 1998. Friction response of sliding interfaces subjected to time varying normal pressures. Journal of Tribology 120, 97-102. Rajagopalan, S., and Prakash, V., 1999. A Modified Torsional Kolsky Bar for Investigating Dynamic Friction. Experimental Mechanics 39(4), 295-303. Ranjith, K., and Rice, J. R., 2001. Slip dynamics at an interface between dissimilar materials. Journal of the mechanics and physics of solids 49, 341-361. Rempel, A. W., and Rice, J. R., 2006. Thermal pressurization and onset of melting in fault zones. Journal of Geophysical Research-Solid Earth 111. Rice, J. R., and Ruina, A., 1983. Stability of steady frictional sliding. Journal of Applied Mechanics 50, 343-349. Rice, J. R., 2006. Heating and weakening of faults during earthquake slip. Journal of Geophysical Research-Solid Earth 111. Richardson, E., and Marone, C., 1999. Effects of normal force vibrations on friction healing. Journal of Geophysical Research 104, 28859-28878. Roig Silva, C., Goldsby, D. L., Toro, G. Di, and Tullis, T. E., 2004. The role of silica content in dynamic fault weakening due to gel lubrication (abstract), 2004 Southern California Earthquake Center Annual Meeting, Proceedings and Abstracts Volume XIV, p. 150. Rudnicki, J. W., and Rice, J. R., 2006. Effective normal stress alteration due to pore pressure changes induced by dynamic slip propagation on a plane between dissimilar materials. Journal of Geophysical Research-Solid Earth 111. Ruina, A., 1983. Slip stability and state variable friction laws. Journal of Geophysical Research 88, 10359-10370.
42

Scholz, C. H., Molnar, P., and Johnson, T., 1972. Detailed studies of frictional sliding of granite and implications for the earthquake mechanism. Journal of geophysical research 77, 6392-6406. Scholz, C. H., 1998. Earthquakes and friction laws. Nature 243, 37-42. Segall, P., and Rice, J. R., 2006. Does shear heating of pore fluid contribute to earthquake nucleation? Journal of Geophysical Research-Solid Earth 111. Shugarts, 1953. Measuring friction at high speeds. Journal of Franklin Institute 256, 187-189. Sibson, R. H., 1973. Interaction between temperature and pore-fluid pressure during earthquake faulting A mechanism for partial or total stress relief. Nature 243, 66-68. Sierakowski, R. L. C. S. K., 1997. Dynamic Loading and Characterization of Fiber-reinforced Composites. John Wiley & Sons, New York, NY, USA. Simes, F. M. F., and Martins, J. A. C., 1998. Instability and ill-posedness in some friction problems. International journal of Engineering Science 36, 1265-1293. Song, B., Chen, W., and Weerasooriya, T., 2002. Impact Response and Failure Behavior of a Glass/Epoxy Structural Composite Material. In: Wang, C. M., Liu, G. R., and Ang, K. K., (Eds.), Proceedings of the 2nd International Conference on Structural Stability and Dynamics. World Scientific Publishing Co., Singapore, Singapore, pp. 949-954. Spaepen F., 1977. A microscopic mechanism for steady state inhomogeneous flow in metallic glasses. Acta Metall. 25, 407-415. Sternlicht, B., and Apkarian, H., 1960. Investigation of "Melt Lubrication". ASLE transactions 2, 248-256. Subhash, G., Zhang, H., and Li, H., 2003. Thermodynamic and Mechanical Behavior of Hafnium/Zirconium Based Bulk Metallic Glasses, Proceedings of the International Conference of Mechanical Behavior of Materials (ICM-9). Kenes International, Geneva, Switzerland. Sunny, G. P., Lewandowski, J. J., and Prakash, V., 2006a. Dynamic compression of amorphous and annealed bulk metallic glass, Paper # 349, Proceedings of the 2006 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, St. Louis, MO, USA. Sunny, G. P., Prakash, V., and Lewandowski, J. J., 2005a. Effects of annealing on dynamic behavior of a bulk metallic glass, Paper # IMECE2005-83016, Proceedings of the 2005 International Mechanical Engineering Conference and Exposition, ASME. American Society of Mechanical Engineers, New York, NY, Orlando, FL. Sunny, G. P., Prakash, V., and Lewandowski, J. J., 2006b. Results from a novel insert
43

design for high starin-rate compression of a bulk metallic glass, Paper # IMECE2006-15414, Proceedings of the 2006 International Mechanical Engineering Conference and Exposition, ASME. American Society of Mechanical Engineers, New York, NY, Chicago, IL. Sunny, G. P., Yuan, F., Lewandowski, J. J., and Prakash, V., 2005b. Dynamic Stress-Strain response of a Zr-based bulk metallic glass, Paper # 324, Proceedings of the 2005 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Portland, Oregon USA. Tay, T. E., Ang, H. G., and Shim, V. P. W., 1995. An empirical strain rate-dependent constitutive relationship for glass-fibre reinforced epoxy and pure epoxy. Composite Structures 33 (4), 201-210. Tsai. L., 2006. Shock wave structure and spall strength of layered heterogeneous glass/polyer composites, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, Ohio. Tsutsumi, A., and Shimamoto, T., 1997. High-velocity frictional properties of gabbro. Geophysical Research Letters 24, 699-702. Tullis, T. E., 1994. Predicting earthquakes and the mechanics of fault slip. Geotimes 39, 19-21. Tullis, T. E., and Goldsby, D. L., 2003. Flash melting of crustal rocks at almost seismic slip rates. Eos Trans. AGU 84(46), Fall Mtg. Suppl., Abstract S51B-05. Turnbull, D., 1969. Contemp. Phys. 10, 473. Turneaure, S. J., Winey, J. M., and Gupta, Y. M., 2004. Compressive shock wave response of a Zr-based bulk amorphous alloy. Applied Physics Letters 84, 1692-1694. Ullah, H., Irfan, M. A., and Prakash, V., 2007. State and rate dependent friction laws for modeling high-speed frictional slip at metal-on-metal interfaces Journal of Tribology 129, 17-22. Vaidya U. K., Hosur M. V., Kumar P., Mahfuz H., Haque A. and Jeelani S., 1999. Impact damage resistance of innovative functional sandwich composites. In: ASME 1999 Mechanics and Materials Confberence Blacksburg, VA. Vural, M., and Ravichandran, G., 2004. Failure mode transition and energy dissipation in naturally occurring composites. Composites Part B-Engineering 35 (6-8), 639-646. Weertman, J., 1980. Unstable slippage across a fault that separates elastic media of different elastic constants. Journal of Geophysical Research 85, 1455-1461. Wesseling, P., Lowhaphandu, P. and Lewandowski, J. J., 2003. Effects of superimposed pressure on flow and fracture of two bulk amorphous metals. Mat. Res. Soc. Symp. Proc.
44

754, 275-279. Wibberley, C. A. J., 2002. Hydraulic diffusivity of fault gouge zones and implications for thermal pressurization during seismic slip. . Earth Planets Space 54, 1153-1171. Wibberley, C. A. J., and Shimamoto, T., 2003. Internal structure and permeability of major strikeslip fault zones: the Median Tectonic Line in Mid Prefecture, Southwest Japan, J. Struct. Geol. 25, 5978. Wibberley, C. A. J., and Shimamoto, T., 2005. Earthquake slip weakening and asperities explained by thermal pressurization. Nature 436, 689-692. Williams, K., and Griffen, E., 1964. Friction between unlubricated steel surfaces at sliding speeds upto 750 feet per second. Proceedings institute of mechanical engineers 178, Part 3N, 24-36. Yang, C., Liu, R. P., Zhang, B. Q., Wang, Q., Zhan, Z. J., Sun, L. L., Zhang, J., and Gong, Z. Z., 2005. Void formation and cracking of Zr41Ti14Cu12.5-Ni10Be22.5 bulk metallic glass under planar shock compression. Journal of Materials Science 40, 3917-3920. Yang, C., Wang, W. K., Liu, R. P., Zhang, X. Y., and Li, X., 2006. Damage features of Zr41Ti14Cu12.5Ni10Be22.5 bulk metallic glass impacted by hypervelocity projectiles. Journal of Spacecraft and Rockets 43, 565-567. Zaretsky, E., deBotton, G., and Perl, M., 2004. The response of a glass fibers reinforced epoxy composite to an impact loading. International Journal of Solids and Structures 41 (2), 569-584. Zhang, Z. F., Eckert, J. and Schultz, L., 2003a. Difference in compressive and tensile fracture mechanisms of Zr59Cu20Al10Ni8Ti3 bulk metallic glass. Acta Mat. 51, 1167-1179. Zhang, Z. F., et al., 2003b. Fracture mechanisms in bulk metallic glassy materials. Phys. Rev. Lett. 91 (4-045505), 1-4 . Zhuang, S. M., Lu, J., and Ravichandran, G., 2002. Shock wave response of a zirconium-based bulk metallic glass and its composite. Applied Physics Letters 80, 4522-4524. Zhuk, A. Z., Kanel, G. I., and Lash, A. A., 1994. Glass Epoxy Composite Behavior under Shock Loading. Journal De Physique Iv 4 (C8), 403-407.

45

Chapter 2

SLIP WEAKENDING IN ROCKS AND ANALOG MATERIALS AT CO-SEISMIC SLIP RATES

2.1 Introduction Determination of co-seismic slip resistance in earth faults is critical for understanding the magnitude of shear-stress reduction and hence the near-fault acceleration that can occur during earthquakes. Friction stresses, at constant normal pressure and sliding speeds of less than 1 mm/s, have been well studied for a wide range of materials. The characterization of these results in terms of rate and state variable

friction laws has allowed a better understanding of a wide variety of aspects of the mechanics of earthquakes (Scholz et al., 1972, Dieterich, 1979, 1981, Rice and Ruina, 1983, Ruina, 1983, Dieterich and Kilgore, 1994, Tullis, 1994, Dieterich and Kilgore, 1996, Blanpied et al., 1998, Marone, 1998, Scholz, 1998, Di Toro et al., 2004), including, for example, what to expect in terms of premonitory slip and Omoris Law for aftershock decay (Scholz, 1998).

Experimental data suggest that frictional resistance in geo-materials at slip speeds 1 mm/s and slip distance < 1mm is quite high ( ~ 0.6 to 0.85) (Byerlee, 1978, Dieterich, 1978, Dieterich and Kilgore, 1996). However, seismic inversions provide evidence that
46

frictional resistance of major faults at co-seismic slip speeds (~1-2 m/s) may be quite low (Heaton, 1990, Rice, 2006). Moreover, very little data exist for the simultaneously high slip rates and large slip displacements characteristic of co-seismic slip, and the data that do exist suggest that the frictional behavior at these slip speeds is dramatically different and the dynamic slip weakening occurs (Sibson, 1973, Tsutsumi and Shimamoto, 1997, Goldsby and Tullis, 2002, Di Toro et al., 2004, Mizoguchi et al., 2006, O'Hara et al., 2006).

Earthquakes occur because fault strength falls down with increasing slip or slip rate, so the understanding of dynamic fault weakening during nucleation and the propagation of a seismic rupture is a major task for researchers involved with fault and earthquake physics. Rudnicki and Rice (2006), Segall and Rice (2006), Rempel and Rice (2006)

and Rice (2006) have recently summarized two primary thermal weakening mechanisms which are assumed to act in combination during fault events: (1) Flash heating and consequent weakening at highly stressed asperity contacts during rapid slip which reduces the friction coefficient - a phenomenon studied for many years as the key of understanding the slip rate dependence of dry friction in metals at high slip rates (Bowden and Thomas, 1954, Archard, 1958, Barber, 1976, Kuhlmann-Wilsdorf, 1985, Ashby et al., 1991, Irfan and Prakash, 1994), and which has also been considered recently in seismology as a mechanism that could be active in controlling fault friction during seismic slip before macroscopic melting (see also Andrews, 2002, Hirose and
47

Shimamoto, 2005, Wibberley and Shimamoto, 2005); and (2) Thermal pressurization of pore fluid within the fault core by frictional heating which assumes the presence of water within shallow crustal fault zones such that the effective normal stress n ( n = n p , where n is the compressive normal stress on the fault, and p is the pore fluid pressure) controls frictional strength, and which reduces the effective normal stress and hence reduces the shear resistance associated with any given friction coefficient (Sibson, 1973, Lachenbruch, 1980, Mase and Smith, 1985, Lee, 1987, Mase and L., 1987, Andrews, 2002, Wibberley, 2002, Noda and Shimamoto, 2005, Sulem et al., 2005). Under the right conditions, these two mechanisms are understood to become important immediately after seismic slip initiates (Segall and Rice, 2006).

For sufficiently large combinations of slip speeds and effective normal stress, thermal power generated during solid-on-solid slip overwhelms the ability of thermal conduction to carry the frictional heat away from the slip interface, and macroscopic melting may occur. In this regime, the slip resistance is expected to be primarily

controlled by the shear-strain rate, the thickness and viscosity of molten layers (Bowden and Persson, 1960, Okada et al., 2002, Liou et al., 2004, Hirose and Shimamoto, 2005). Molten layers have a low viscosity and may lubricate faults reducing dynamic friction. However, it has been observed that interface melting is not a simple weakening mechanism for rock and analog materials. Hirose et al. (2005) conducted a series of experiments on Indian gabbro at slip rates of 0.85-1.49 m/s and normal stresses of 1.2-2.4
48

MPa by using a rotary-shear apparatus.

The experiments on gabbro revealed two stages By examination of

of slip weakening separated by a marked strengthening regime.

microstructures of simulated fault zone under scanning electron microscopy (SEM) at different total slip displacements, they proposed that the initial slip weakening is due to the thermal weakening induced by flash heating at the asperity contacts and early stages of melting; this phase is followed by slip strengthening caused by the coalescence of melt patches into a thin molten layer; while the second slip weakening is attributed to the growth of molten layer during friction melting. Another weakening mechanism, i.e. by the formation of silica gel, has been identified by Di Toro et al. (2004). They studied the slip resistance of Arkansas Novaculite rock at a slip rate of 0.03 m/s and a normal stress of 5 MPa using a servo-controlled compression-torsion apparatus. They attributed the initial slip weakening mechanism to the formation of silica gel, and the time-dependent recovery of shear strength to the thixotropic behavior of the silica gel. Although different physical processes, such as, flash heating at asperity contacts, formation of silica gel, and frictional melting have been proposed that could lower shear resistance during fast co-seismic slip, these mechanisms and/or their applicability to earthquakes are still poorly understood.

In addition to the principal problem that we do not have adequate data or constitutive descriptions for large and rapid sliding on a fault at constant normal stress, we do not have much knowledge of the effect on the resistance of sudden changes in normal stress
49

on frictional slip resistance, and this is an important problem, since during seismic slip there can be abrupt changes in normal stress as well as shear stress. Data exist at low slip speed showing that the simple Coulomb representation for the effect of normal stress is inadequate and that there are memory effects following a change in normal stress (Linker and Dieterich, 1992, Richardson and Marone, 1999, Bureau et al., 2000), but few data exist at high slip velocity (Prakash and Clifton, 1993, Prakash, 1998, Irfan and Prakash, 2000, Ullah et al., 2007). Theoretical studies of slip at dissimilar material interfaces (Weertman, 1980, Adams, 1995) have shown that spatially inhomogeneous slip causes an alteration of normal stress, so this, as well as effects from non-planar faults and interactions between nearby faults, can cause changes in normal stress. Reductions of

normal stress can reduce frictional resistance and thus lead to possible destabilization of slip (Andrews and Ben-Zion, 1997, Simes and Martins, 1998, Cochard and Rice, 2000, Ben-Zion, 2001, Ranjith and Rice, 2001).

Nevertheless, the current state of knowledge is so insufficient that all of the processes responsible for frictional resistance during co-seismic slip are not yet known, nor any reliable constitutive description of the behavior that can be used in dynamic models of earthquake rupture. In the present study plate-impact pressure-shear friction

experiments and modified torsional Kolsky-bar friction experiments are conducted to investigate frictional resistance in Arkansas Novaculite rock and analog materials such as soda-lime glass. The objective of these experiments is to investigate the frictional
50

resistance in relevant geo- and analog materials at co-sesimic slip speeds and higher, so as to better understand the role of high slip speeds in leading to slip weakening at the slip interface by the mechanism of asperity flash heating and associated thermal softening with slip. Besides providing fundamental friction data in the normal stress, slip speed and slip distance range not achievable by any other method, the plate impact pressure-shear friction configuration and modified torsional Kolsky-bar configuration represent considerable improvements over conventional dynamic friction experiments. A primary concern with use of conventional dynamic friction experiments has been the influence of stiffness of the dynamic load train in the measurement of the frictional characteristics (Bell and Burdekin, 1969, Antoniou et al., 1976, Martins et al., 1990,
Armstrong-Helouvry et al., 1994).

The use of the pressure-shear waves to load the

frictional interface results in high loading stiffness and any non-uniformity in geometry or alignment of the specimen leads, in the worst case, to a slight spreading of the step-function load in time. Another attractive feature of these configurations is that it

allows time-resolved measurements of interfacial normal and shear tractions and slip-speeds to be made without the use of transducers at the frictional interface (Prakash, 1995; Rajagopalan and Prakash, 1999). In addition, the plate impact pressure-shear

friction experimental configuration can be easily modified to introduce step alterations in the interfacial normal stress. The results of this study have relevance not just to

mechanisms of dynamic fault weakening, but also to the constitutive description of the behavior that can be used to critically examine the rate and state dependent dynamic
51

friction models for earthquake rupture.

2.2 Pressure Shear Plate Impact Friction Experiments In the present study a series of plate impact pressure shear friction experiments were conducted on soda-lime glass (analog material) and fine grained Arkansas Novaculite rock samples to investigate frictional resistance in geo- and analog materials. The

choice of soda-lime glass as an analog material was dictated by a number of previous studies (Weeks et al., 1991, Dieterich and Kilgore, 1994, Dieterich and Kilgore, 1996), which have shown that the frictional behavior of glass is almost identical to that of rock. Plate-impact pressure-shear friction experiments, at both constant normal-stress and with normal-stress alterations at the slip interface, were designed and conducted. Details of the plate impact pressure shear friction experimental configuration and the specimen preparation are provided next.

2.2.1 Experimental Configuration The plate-impact pressure-shear friction experiments were conducted using the 82.5 mm bore single stage gas gun facility at Case Western Reserve University. schematic of the experimental configuration is shown in Figure 2.1. The

A fiberglass

projectile carrying the flyer plate is accelerated down the gun barrel by means of compressed air. The rear end of the projectile has sealing O-ring and a Teflon key that

slides in a key-way inside the gun barrel to prevent any rotation of the projectile. In

52

order to reduce the possibility of an air cushion between the flyer and target plates, impact takes place in a target chamber that has been evacuated to 50 m of Hg prior to impact. To ensure the generation of plane-waves with wave-front sufficiently parallel to

the impact face, the flyer and the target plates are carefully aligned to be parallel to within 210-5 radians by using an optical alignment scheme developed by Kumar and Clifton (1977). The actual tilt between the two plates is measured by recording the times at

which four, isolated, voltage-biased pins, that are flush with the surface of the target plate, are shorted to ground. Impact takes place at an angle relative to the direction of

approach. This results in pressure-shear loading at the flyer-target (tribo-pair) interface. By controlling the skew angle , the impact velocity, a variety of friction states (with normal stress varying from 200 MPa to 1 GPa and slip speeds from no-slip to 100 m/s) can be obtained. During the experiment both normal and transverse particle velocity

histories of the rear surface of the target plate are measured by laser interferometer. These measurements are used to infer the normal and shear tractions, slip-velocity and temperature at the tribo-pair interface. Also, all measurements of the particle velocity

are made before the arrival of the release waves from the lateral boundary of the specimen. In view of this, and during the time interval of interest, the flyer and target

plates can be considered to be essentially infinite in their spatial dimensions and the tribo-pair to be modeled as a semi-infinite half-plane sliding on another. This

simplification in the tribo-pair geometry allows one-dimensional wave theory to be used in the interpretation of the experimental results.
53

Other details regarding the

experimental configuration and execution of the experiments can be found elsewhere (Prakash, 1995).

Figure 2.1: Schematic of the plate-impact pressure-shear friction experiment.

A LASER interferometric technique is used to measure the combined normal and transverse particle velocities at the rear-surface of the target plate. Figure 2.2 shows the schematic of the combined normal and transverse displacement interferometry. A

COHERENT VERDI solid-state laser with a wavelength of 532 nm is used to provide the monochromatic light source. The normal displacement is measured by using Normal

Displacement Interferometry (NDI) while the transverse velocity is measured by using Transverse Displacement Interferometry (TDI) (Kim et al., 1977). The NDI monitors the

time history of the displacement of the free surface of the target plate by combining a reference beam with a reflected beam from the target such that one peak to peak variation in
54

light intensity corresponds to a normal displacement of /2 (0.25725 m), where is the wavelength of the laser light. The TDI monitors two nth order diffracted beams from a

grating deposited on the rear surface of the target plate and measures the phase difference between them. The measured phase difference is related to the history of the transverse The sensitivity of the measured

displacement at the free surface of the target plate.

transverse displacement, i.e. the transverse displacement represented by one peak-to-peak variation in the measured light intensity, is given by = pitch of the grating.
= 0.4165 m .

d , where d=1/D and D is the 2n which gives

For the present experiment, D=1200, and n=(1-(-1)) = 2,

The interferometric signals are detected by NEWPORT silicon pin When necessary the

detector (Model 818-BB-21) having a rise time of less than 1 ns.

output of the photo detectors are amplified by suitable bandwidth digital amplifiers before they are fed to the oscilloscopes. Hewlett Packard 54542A oscilloscope (2 GS/s, 500 MHz) was used for this purpose. A specially developed Matlab program is then used to process

the interferometric traces and calculate the history of the normal and transverse particle velocities at the rear surface of the target plate.

55

Figure 2.2: Schematic of the combined normal and transverse displacement interferometry.

2.2.2 Tribo-pair Materials and Design of the Pressure-shear Experiments In order to investigate the dynamic frictional response of glass versus glass slip, thin films of soda-lime glass (5 m in thickness) were deposited on the impacting faces of the flyer and target plates comprising of Ti-6Al-4V and CH tool-steel, respectively (Figure 2.3a). The thin glass films were deposited on the tribo-surfaces by Thin-Films Research, Inc., Westford, MA, by employing a vapor deposition procedure. The impact velocity in

the experiments was controlled such that the flyer and target plates remain elastic during impact and, thus, the measured friction stress represents the sliding resistance of essentially glass-on-glass slip. Also, given the relatively high wave speed in glass, the

less than 5 micron thickness of the glass films preclude any significant wave dispersion effects. Moreover, the thickness of the flyer and the target plates are designed such that
56

the time for longitudinal wave propagation through the thickness of the flyer plate is greater than the corresponding round-trip time of the longitudinal waves in the target plate and the unloading waves generated at the lateral boundary. Under these conditions,

when the longitudinal wave reflected from the rear surface (free surface) of the target plate arrives at the target/flyer interface, the normal stress at the frictional interface is changed instantaneously. Since the longitudinal impedance of the flyer plate (Ti-6Al-4V) is less than the longitudinal impedance of the target plate (CH tool-steel), it results in a step drop in the applied normal stress.

The wave propagation in the target and flyer is illustrated schematically in the time-distance diagram shown in Fig. 2.4. The abscissa represents the spatial position of

the wave front at any particular time and the ordinate represents the temporal location of the wave front. At impact, both longitudinal and shear waves are generated at the

tribo-pair interface and travel through the thickness of the flyer and the target plates. The longitudinal wave fronts are represented by solid lines and the shear wave fronts are represented by dashed lines. The slope of the solid line represents the inverse of the

longitudinal wave speed and the slope of the dashed line represents the inverse of the shear wave speed in the material. In State 1 the tribo-pair interface is under a

compressive pressure 1, and depending on the friction stress a shear stress 1 is transmitted across the interface. Thus, State 1 allows investigation of dynamic sliding Upon

characteristics of the frictional interface under constant normal pressure.

57

reflection of the compressive wave at the rear (free) surface of the target plate an unloading wave is generated, which propagates back towards the tribo-pair interface. When this unloading wave arrives at the frictional interface it reduces the applied normal pressure from 1 to 2. The corresponding friction stress is denoted by 2. The new

frictional state hence produced is called State 2 and allows investigation of dynamic sliding characteristics of frictional interfaces subjected to step changes in normal pressure.

The experiments on Arkansas Novaculite rock were conducted at constant normal stress using the symmetric plate-impact pressure-shear configuration (Figure 2.3b). In this case an Arkansas Novaculite rock plate is impacted against another Novaculite rock plate. Sliding initiates at the slip interface following the pressure-shear impact loading

at a constant normal stress.

58

(a)

(b)

Figure 2.3: Plate-impact configurations for (a) soda-lime glass; (b) Arkansas Novaculite rock.

Figure 2.4: Wave propagation in flyer and target plates: time-distance diagram.

59

2.2.3 Wave Analysis of Pressure-shear Friction Experiments on Soda-lime Glass: Calculation of Interfacial Tractions, Slip velocity, and Temperature Using the method of characteristics for 1-D hyperbolic wave equations the normal and transverse components of interfacial traction, slip velocity, and the slip displacement can be related to the normal and transverse components of particle velocity of the free surface of the target plate, ufs(t) and vfs(t), respectively, and the shear and longitudinal impedances of the flyer and the target materials.

STATE 1: Before impact the flyer and the target plates are unstressed. The target is held

stationary while the flyer is accelerated and impacts the target at a measured velocity V. From the knowledge of the angle of inclination of the projectile, the initial normal and transverse particle velocities uo and vo, respectively, of the flyer plate are given by

uo = V cos and

vo = V sin , respectively.

When the flyer impacts the target, both the normal and transverse components of velocity are imposed on the impact face of the target. A longitudinal compression stress

wave with wave speed c1 and a shear stress wave with a wave speed c2 propagates into the flyer and the target plates. From one dimensional analysis of the governing

hyperbolic partial differential equations the stress and the particle velocity relations which hold along the characteristics are given by

60

 ( c1 ) u = constant ;

along

dx = c1 , dt

(2.1)

( c2 ) v = constant ;

along

dx = c2 . dt

(2.2)

In Eqs. (2.1) and (2.2), and are the normal and shear stresses, u and v are the normal and transverse components of the particle velocity, is the mass density, and (c1) and (c2) are the longitudinal and shear impedance, respectively.

Using initial conditions it can be shown that all states at the impact face of an elastic flyer plate must satisfy the following characteristic relations:

+ ( c1 )f u = ( c1 )f uo ,

(2.3)

+ ( c2 )f v = ( c2 )f vo .

(2.4)

Moreover, all states on the impact face of an elastic target plate must satisfy the following characteristic equations:

+ ( c1 )t u = 0 ,

(2.5)

61

 + ( c2 )t v = 0 .

(2.6)

In Eqs. (2.5) to (2.6), the subscripts f and t refer to the flyer and the target, respectively.

Using Eqs. (2.5) to (2.6), the components of traction at the interface between the flyer and the target can be expressed as

(t ) =

(c2 )t
2

v fs (t ) ,

(2.7)

(t ) =

(c1 )t
2

u fs (t ) ,

(2.8)

where v fs (t ) and u fs (t ) are the transverse and the normal particle velocities at the rear surface of the target plate.

Based on the elementary definition of friction between two dry contact surfaces under no slip condition, the coefficient of static friction, s, satisfies the inequality

(2.9)

62

For a fully elastic impact with no-slip at the interface, Eq. (2.9) provides a lower bound for the static coefficient of friction, s. When slipping occurs at the flyer-target interface,

the measured free surface velocity of the target plate can be used along with Eqs. (2.7) and (2.8) to obtain the coefficient of kinetic friction

k (t ) =

(2.10)

From the knowledge of the impact velocity V, the skew angle , the shear impedances of the flyer and the target plates, and the measured free-surface transverse velocity vfs(t), the slip velocity can be expressed as

( c ) + ( c2 ) f Vslip = V sin 2 t Vfs (t) . 2 ( c2 )f

(2.11)

The accumulated slip distance can be obtained from Eq. (2.11) by integrating the slip velocity in time

slip =

V
0

slip

(t ) dt .

(2.12)

Moreover, under conditions of no slip the free surface particle velocities of the rear surface of the target plate are related to the flyer velocity by
63

u fs =

2 ( c1 )f ( c ) + ( c ) 1 t 1 f

V cos ,

(2.13)

v fs =

2 ( c2 )f ( c ) + ( c ) 2 t 2 f

V sin .

(2.14)

STATE 2: When the compressive longitudinal wave reflects from the free surface of the target plate it reduces the compressive normal stress at the interface from a stress level 1 to 2

( c ) ( c ) L f L t . 2 = ( c ) + ( c ) 1 L f L t

(2.15)

For the Ti6Al4V/CH tool-steel tribo-pair employed in the present study the ratio 2 / 1 ~ 0.25.

The corresponding expressions for the friction stress and slip velocity are obtained by solving the characteristic relations for State 2, and can be expressed as

2 (t ) =

(c2 )t
2

vb (t ) ,

(2.16)

64

and

slip 2

(c2 ) + (c2 ) t f = V sin 2 (c2 )f

v (t ) . b

(2.17)

Bulk Temperature at the Tribo-Pair Interface: The bulk temperature rise at the tribo-pair interface is estimated by solving the following one dimensional transient heat conduction equation

2T 1 T , = x 2 t

(2.18)

With the initial condition

T(x,0) = 0 ,

(2.19)

and the boundary conditions

T x

(x = 0, t ) = q(t ) ,

(2.20)

T (x = , t ) = 0 .

(2.21)

65

In Eqs. (2.18) to (2.21) T is the temperature rise, k is the thermal conductivity, is the thermal diffusivity and x represents the distance perpendicular to the interface. Table

2.1 provides a summary of thermal properties of soda-lime glass, quartz and Arkansas Novaculite rock. Using Eqs. (2.18) to (2.21) the temperature rise distribution as a

function of time and position can be obtained as

1 T (x , t ) = k

q ( )
0

x 2 exp( ) d . 4 (t ) (t )

(2.22)

In order to calculate the temperature distribution in the tribo-pair, an estimate for the heat source q (t ) is required. Using the experimentally measured friction stress , and the

slip velocity Vslip, the heat generated at the interface can be estimated to be

q flyer (t ) = 0.5 (t )Vslip (t ) .

(2.23)

Materials Soda-lime glass Quartz Novaculite rock

Thermal Conductivity k, ( W / m.K ) 1.38 5.46 5.46

Thermal Diffusivity , m 2 / s 0.89110-6 1.8410-6 1.8410-6

Table 2.1: Summary of thermal properties of soda-lime glass, quartz and fine grained Arkansas rock samples.

66

2.3 Modified Torsional Kolsky-bar Friction Experiment In the present study, in addition to the plate-impact pressure-shear experiments, a series of modified torsional Kolsky-bar friction experiments were conducted on rock-analog material (soda-lime glass) samples. In these experiments interfacial normal

stress ranging from 30 to 80 MPa and slip velocities from 0 to 10 m/s were obtained. The results of the modified torsional Kolsky-bar friction experiments are expected to bridge the gap between the low slip-speed friction experiments (<1 mm/s) and the high-speed plate-impact friction experiments (~20 m/s). Details of the modified

torsional Kolsky-bar friction experimental configuration and the specimen preparation are provided next.

2.3.1 Experimental Configuration The schematic of the modified torsional Kolsky-bar apparatus is shown in Figure 2.5. In this setup, a thin-walled tubular quartz specimen or a thin-walled tubular aluminum specimen with a thin film of soda-lime glass deposited on one of its faces ( the soda lime glass film thickness is ~ 1 m, and is deposited by using a vapor deposition procedure at the Electronics Design Center, CWRU) is mounted at the end of the solid incident bar, while the transmitter bar of the conventional torsional Kolsky-bar is replaced by a quartz or glass disk connected to a rigid support. Besides providing a rigid boundary condition, the disk also represents the other half of the tribo-pair.

67

To conduct the dynamic friction experiments, the specimen on the incident bar is slides axially in the alignment fixture and comes into contact with the disk on the rigid support by a static axial compressive force of predetermined magnitude applied by a hydraulic actuator at the pulley end of the modified torsional Kolsky-bar apparatus. An important consideration in the implementation of the experiment is that while the interfacial sliding is in progress, the sliding face of the tubular specimen should remain parallel and in contact with the other face representing the tribo-pair, at all times. This is achieved by using an alignment fixture, schematically illustrated in Figure 2.6, which ensures that the tubular specimen is aligned perpendicular to the other surface of the tribo-pair, i.e. the surfaces in contact which are lapped flat prior to the experiment are parallel to each other, at all times. The alignment fixture has a Teflon bearing which The glass disk is

allows free rotation in either direction as well as normal motion.

mounted to the supporting rigid block by a two part epoxy (Hysol 143), and thus can be modeled as a rigid support.

Next, a torsional loading pulse is generated by a sudden release of stored torque. This requires a torque pulley system at the end of the incident bar and a frictional clamp positioned a short distance from the pulley end. The torque is generated by employing a hydraulic actuator to twist the pulley attached to the end of the incident bar. The frictional clamp allows the desired torque to be held without slipping and releases the torque rapidly enough when the pre-notched pin breaks to release to a sharp fronted
68

stress pulse which travels towards the tribo-pair specimens.

During the experiment the

input and the reflected torsional pulses from the tribo-pair interface are measured by strain gages attached to the surface of the incident bar. To perform in situ recording of the deformation and failure of specimens during the slip processes, an Imacon 200 ultra-high-speed digital camera (DRS Technologies) was employed. Other details

regarding the design, execution and data analysis of the experiments can be found in (Rajagopalan and Prakash, 1999).

Figure 2.5: Schematic of the modified torsional Kolsky-bar friction experiment.

69

Figure 2.6: Alignment fixture to ensure parallelism of the tribo-pair surfaces.

2.3.2 Wave analysis: Calculation of Interfacial Tractions, Slip Velocity, Slip Distance The wave propagation diagram in the modified Torsional Kolsky bar apparatus is illustrated in Fig. 2.7. Position of the wave front versus time is detailed. The duration

of the loading pulse is the time required for the pulse to travel twice the distance between the pulley and the frictional clamp. the frictional clutch is released. The time t = 0 corresponds to the time at which

At this instant the state in the solid bar to the left of the The solid

frictional clamp is equal to the applied torque T0 and zero angular velocity.

bar to the right of the clamp has zero torque and zero angular velocity. When the clamp is released half of the input torque propagates as a torsional pulse to the left and the other half propagates to the right in the solid bar towards the tribo-pair interface. Upon the

release of the clamp the state is given by 1. The reflected wave from the pulley end

70

travels towards the clamp unloading the bar to state 3 which corresponds to a completely unloaded state. The strain gage at station A sees state 1 until the wave reflected from the The wave

pulley end of the solid bar returns to the gage station and results in state 3.

reflected from the tribo-pair interface returns to gage station A and results in state 5. This returning wave carries information of the frictional state at the tribo-pair interface. By measuring the torsional strain on the incident bar at gage location A, the critical interfacial frictional parameters at the tribo-pair interface such as the frictional stress, the interfacial slip speed and the accumulated slip displacement can be interpreted using the framework of one-dimensional plane-wave analysis.

For the case in which there is full sticking at the tribo-pair interface the state of frictional interface is represented by point A in Fig. 2.8. For the other extreme case, i.e.,

when the tribo-pair interface can support zero frictional stress, the angular velocity at the frictional interface is represented by the point B. For any other case in between, i.e.,

when the torque supported by the tribo-pair interface is given by Tint erface , the corresponding angular slip speed at the frictional interface is given by

interface =

(Tinterface 2T1 ) . ( CJ ) tube

(2.24)

where T1 is the known input torque at the tribo-pair interfaces.

71

In order to obtain the frictional torque at the tribo-pair interface Tint erface , a backward drawn characteristic joining the state ( Tint erface ,int erface ) and ( T5 , 5 ), and a forward drawn characteristic joining the state ( T3 , 3 ) and ( T5 , 5 ) are used. Moreover, using the result that state 3 is a zero state the torque at the tribo-pair interface can be expressed in terms of the measured torques T1 and T5 , i.e.,

Tinterface = T1 + T5 ,

(2.25)

and

interface =

(Tinterface 2T1 ) . ( CJ ) bar

(2.26)

Once the interfacial torque and the interfacial angular slip speed are obtained, the average interfacial friction stress and the average interfacial slip speeds at the tribo-pair interface can be obtained using

ro

interface (t ) =

r (r , t )dr
ri ro

, where (r , t ) = r

rdr
ri

Tinterface (t ) , J specimen

(2.27)

72

r0

Vslip (t ) =

r
2
ri

interface
r0

(t )dr . (2.28)

rdr
ri

In Eqs. (2.27) and (2.28) ri and ro are the inner and outer radii of the thin walled tubular specimen, respectively.

The normal stress at the interface can be obtained from the measured axial strain in the incident bar, i.e,

interface = E bar

Abar Aspecimen

(2.29)

where E is the elastic modulus of the incident bar, and bar is the measured axial strain in the bar.

The accumulated slip distance can be evaluated by integrating the slip velocity history, i.e.

73

 slip (t ) = Vslip (t )dt .

0

(2.30)

Then, by using the interfacial shear stress (Eq. 2.27), and normal stress (Eq. 2.29), the coefficient of kinetic friction k can be obtained. i.e.,

k (t ) =

interface (t ) . interface

(2.31)

Figure 2.7: Wave propagation diagram for the modified torsional Kolsky bar.

74

Figure 2.8: Loci of all the torque and angular velocity states that can be attained at the tribo-pair interface.

2.3.3 Calculation of Interfacial Temperature Rise 2.3.3 (a) Temperature Distribution in the Tubular Specimen:

By neglecting radiation and convection effects, and since the thermal conductivity of soda-lime glass and/or quartz (1.38 and 5.46 W/m.k, respectively) are much larger than that of air (0.024 W/m.k), the heat conduction in the tubular specimen is essentially reduced to one-dimensional heat flow in a semi-infinite solid, and is governed by the 1-D heat equation

75

1 T 2T , = x 2 t

(2.32)

with the initial condition

T(x,0) = 0 ,

(2.33)

and the boundary conditions

T x

(x = 0, t ) = q(t ) ,

(2.34)

T (x = , t ) = 0 .

(2.35)

In Eqs. (2.32) to (2.35) T is the temperature rise, k is the thermal conductivity, is the thermal diffusivity, q (t ) is heat source and x represents the perpendicular distance
away from the interface.

By solving Eqs. (2.32) to (2.35) in the semi-infinite solid, the temperature rise distribution, as a function of time and position, can be expressed as

76

t x2 1 T ( x,t ) = q( ) exp d . k0 (t ) 4 ( t )

(2.36)

In view of Eq. (2.36), the interfacial temperature (i.e. at x=0) can be written as

1 T ( 0 ,t ) = q ( ) d . k0 (t )
t

(2.37)

2.3.3 (b) Temperature Distribution in the Disk Specimen:

In order to obtain the temperature distribution in the disk, a semi-infinite solid with a heat source over a circular area ri r ro , and x =0 , is considered. Note that ri and ro are the inner and outer diameter of the tubular specimen, respectively. By solving the heat conduction problem, the temperature rise distribution as a function of time and position can be expressed as

T ( r , x, t ) =

1 2k

q ( )
0

d (t )

ro 3 2 ri

( r 2 + r 2 + x 2 ) 4 ( t )

I0 (

rr )r dr . 2 (t )

(2.38)

In view of Eq. (2.38), the temperature distribution at the slip interface (i. e. x=0) can be expressed

T (r , 0, t ) =

1 2k

q( )
0

d (t )

ro 3 2 ri

( r 2 + r 2 ) 4 ( t )

I0 (

rr )r dr , 2 (t )

(2.39)

77

where I 0 ( z ) is the modified Bessel function of order zero.

In the present study, the total time duration is less than 1000 s , r , r are of the order 20~30 mm, and ~10-6 m2/s, so that
rr ~105. For large number z 2 (t )

I0 ( z)

ez . 2 z

(2.40)

Applying Equation (2.40) to Equation (2.39)

1 d o 4 (t ) T (r , 0, t ) = q( ) e r dr (t ) 2 k r ri 0
t
i d 4 ( t ) = q ( ) 1 d (t ) r r e 2 k 0 r o

( r r )2

r r

(2.41)

Because the thickness of tubular specimen is small compared to the radius of tubular specimen, we can approximate
1

1.

In view of this, Eq. (2.41) becomes

T (r , 0, t )

i d q ( ) e 4 ( t ) d . 2 k (t ) r ro 0

r r

(2.42)

78

a2

By expressing the integral

a2

a1

x2 ) a3

dx

in terms of error function, i.e.

a1

x ) a3

dx =

a3 a a (erf ( 1 ) +erf ( 2 )) , Equation (2.42) can be written as 2 a3 a3

t T (r , 0, t ) = q( ) 2k 0

ro r r ri ) + erf ( ) . erf ( 1 4 (t ) 4 (t ) (t ) 2

(2.43)

When know

ro r r ri 3, 3 i.e. when ri + 0.1 mm < r < ro 0.1 mm , we 4 (t ) 4 (t )

erf (

ro r ) 4 (t )

1, erf (

r ri ) 4 (t )

1 .

(2.44)

In the present study, ri

22 mm, ro

25 mm , so for most of region in the circular

area ri r ro , x =0 , Equation (2.44) holds and Equation (2.43) can be written as

T (0, t ) =

1 q( ) (t ) d . k0
t

(2.45)

2.3.3 (c) Temperature Rise at the slip Interface:

In order to calculate the temperature distribution in the tribo-pair materials, an estimate for the heat source q (t ) is required. Using the experimentally measured

79

friction stress and slip velocity V slip , the frictional power can be written as

qtubular (t ) = (t )V slip (t ) ,

(2.46)

qdisk (t ) = (1 ) (t )V slip (t ) .

(2.47)

where, is the factor that governs the partitioning of heat in the tribo-pair materials.

The factor is estimated by equating the temperatures at the tribo-pair interface to yield

k tubular disk k tubular disk + kdisk tubular

(2.48)

In the present study, for glass on glass slip, the tribo-pair are the same material, so

= 0.5 . Then, the temperature rise at the interface can determined by

t 1 slip T (0, t ) = 0.5 ( )V ( ) (t ) d . k 0

(2.49)

Thermal properties of soda-lime glass and quartz are provided in Table 2.1.

80

2.4 Experimental Results 2.4.1 Plate-impact Pressure-shear Friction Experiment

Table 2.2 provides a summary of the plate-impact pressure-shear friction experiments discussed in the present paper. Three experiments, Shot #1, Shot #2 and Shot #3 were conducted on soda lime glass, while the experiments FY008 and FY009 were conducted on fine grained Arkansas Novaculite rock. In all experiments, except for Shot #1, a skew angle of 35o was utilized to promote high slip speeds at the slip interface. For Shot #1 a skew angle of 30o was employed. The surface roughness of the impacting surfaces of the flyer and target plates was kept relatively smooth. The RMS surface roughness of the soda lime glass plate was 0.12 m while the RMS surface roughness of the fine grained Arkansas novaculite rock was 0.10 m. The projectile velocities for Shot #1, Shot #2, Shot #3, FY008 and FY009 were 66.3 m/s, 55.4 m/s, 54.3 m/s 75.4 m/s and 64.1 m/s, respectively. At these impact speeds both the flyer and the target plates are For the glass on glass experiments the

expected to remain elastic during impact.

impedance mismatch between the flyer (Ti6Al4V) and the target (tool-steel) plates result in a drop of normal stress at the slip interface. The normal stress in the new state (State 2) is approximately 0.25 times the normal stress in State 1. The experiments on the fine grained Arkansas rock were conducted at a constant normal stress of during the entire duration of the slip process.

81

SHOT # Shot #1 Shot #2 Shot #3

Skew Angle , deg 30 35 35

Impact Velocity V, m/s 66.3 55.4 54.3

Flyer: Ti-6Al-4V (mm) 13.1 13.1 13.1

Target: CH Tool Steel (mm) 6.1 6.1 6.1

Roughness of Soda-Lime Glass Layer, m 0.12 0.12 0.12

SHOT # FY008 FY009

Skew Angle , deg 35 35

Impact Velocity V, m/s 75.4 64.1

Normal Stress , MPa 475 404

Flyer: Fine grained Arkansas Novaculite Rock (mm) 12.3 10.9

Target: Fine grained Arkansas Novaculite Rock (mm) 10.4 9.6

Roughness of Impact surfaces, m 0.10 0.10

Table 2.2: Summary of the plate impact pressure shear friction experiments on soda-lime glass and fine grained Arkansas rock samples. The experimental results for Shot # 1 are shown in Figures 2.9 to 2.11. The

experiment was conducted at an impact velocity of 66.3 m/s and a skew angle of 30o. Figure 2.9 shows the measured normal and transverse components of the particle velocity history at the free surface of the CH tool-steel target plate. The abscissa represents the time after impact. The normal and transverse components of the particle velocity

histories at the rear surface of the target plate were obtained by using the combined normal and transverse displacement interferometer (Kim et al., 1977). Also shown in the figure are the normal stress and the elastic prediction of the transverse component of the particle velocity for the case of no-slip at the interface. The normal stress was obtained from the normal component of the free surface particle velocity by using the
82

1-D elastic wave theory presented in Section 2.2.3. In State 1 the normal stress is 985 MPa. The arrival of shear wave at the free surface of the target plate occurs at

approximately 1.8 s. At the arrival of the transverse wave the transverse component of the free surface particle velocity jumps to 26 m/s, which also represents the no-slip level at the slip interface. The small dip in particle velocity at approximately 2.9 s is a consequence of a small tilt between the impacting plates. The transverse velocity stays at the no-slip level until the arrival of the unloading longitudinal wave at which point the normal stress drops from 985 MPa to 247 MPa; in response to this drop in normal stress the transverse particle velocity falls to a level of about 12 m/s. This transverse velocity level is lower than that predicted for a no-slip condition, indicating the initiation of slip at the glass-glass interface.

Figure 2.9: Transverse particle velocity history of the rear surface of the CH tool steel target plate for Shot # 1.

83

Figure 2.10 shows the history of the normal stress, coefficient of kinetic friction, and interfacial slip velocity as a function of time for Shot # 1. In order to calculate the coefficient of kinetic friction, the friction stress was obtained from the measured transverse components of the free surface particle velocity history by using the 1-D elastic wave theory presented in Section 2.2.3. Two distinct friction states are obtained during the experiment: State 1 which begins just after impact and ends at ~ 2.1 s, and State 2 which ends at ~ 2.8 s. During State 1, a no-slip condition exists at the frictional interface. In State 2 the sudden alteration in normal stress leads to the initiation of slip and the slip velocity increases from no-slip to as high as 20 m/s. It is interesting to note the range of coefficient of friction attained in the experiment, which varies from 0.5 to as high as 1.3. These considerably high values of coefficient of friction are in sharp contrast to metals where the coefficient of dynamic friction has been observed to be in between 0.1 to 0.25 under similar interfacial conditions (Irfan and Prakash, 2000).

84

Figure 2.10: History of slip velocity and coefficient of kinetic friction for Shot # 1. Figure 2.11 shows the estimate for the temperature rise during the dynamic slip event. By assuming slip to occur between two semi-infinite glass blocks, the average temperature rise at the tribo-pair interface during State 2 was estimated by solving the one-dimensional transient heat conduction equation (Equation (2.22)). At the initiation of slip in State 2 considerable frictional heat is generated and the interfacial temperature increases monotonically with time to about 900oC. The diffusive nature of the solution is clearly evident from the sharp decay in the bulk temperatures with increasing distance perpendicular to the frictional interface. It is interesting to note that even though these estimated interfacial temperatures will not lead to large scale melting of the glass, they are still high enough to result in considerable softening of the glass films. It must be mentioned that the aforementioned temperature rise distributions are an estimate of the actual global temperatures existing during the frictional sliding process. A number of

85

assumptions have been made in arriving at these estimates: (a) the thermal properties of soda-lime glass are assumed to be constant at their room temperature values at all temperatures, and (b) all the interfacial frictional work is assumed to be converted to heat. The assumption that all the frictional work is converted to heat is expected to lead to an overestimation of the actual tribo-pair temperature rise distribution. Moreover, all inelastic processes during the slip process are expected to occur within a thin layer at the frictional interface, the thickness of which is of the order surface roughness of the soda lime glass films. In view of these arguments, the computed temperature profiles are expected to represent the actual temperature distributions with acceptable level of accuracy.

Figure 2.11: Estimated temperature rise at tribo-pair interface for Shot # 1. The experimental results for Shot # 2 are shown in Figures 2.12 to 2.14. The experiment was conducted at an impact velocity of 55.4 m/s and a skew angle of 35
86

degrees. It is to be noted that in Shot #1 a skew angle of 30o was employed, and the use of a higher skew angle in Shot # 2 is expected to promote higher levels of slip in this case. Figure 2.12 shows the normal and transverse components of the particle velocity history at the free surface of the CH tool-steel target plate. At the arrival of the longitudinal wave at the free (rear) surface of the target plate, the normal component of the particle velocity jumps to ~ 28 m/s at about 1050 ns. The second step-rise in normal component of the particle velocity occurs at approximately 3000 ns, which is coincident with the arrival of the reflected (release) wave from the rear surface of the steel target plate at the tribo-pair interface. The arrival of shear wave at the free surface of the target plate occurs at approximately 1800 ns. With the arrival of the shear wave at the free surface of the target plate the transverse component of the particle velocity jumps to the no-slip predicted level of 25 m/s. After a momentary stay at this no-slip level the transverse velocity falls precipitously to a level of about 8 m/s before increasing again to a no slip level of 25m/s. The transverse velocity stays at this no-slip level until the arrival of the unloading wave at the slip interface at which the transverse velocity again falls to a level of about 12 m/s.

87

Figure 2.12: Normal and transverse particle velocity history at the rear surface of the target plate for Shot # 2.

Figure 2.13 shows the history of the normal stress, coefficient of kinetic friction, and interfacial slip velocity as a function of time for Shot # 2. Two distinct friction states are obtained in the experiments: State 1 which begins just after impact and ends at ~ 2.1 s, and State 2 which ends at ~ 3.2 s. During State 1 the normal stress at the friction interface is constant at 778 MPa while the normal stress in State 2 is 195 MPa. During State 1, after a momentary initial stick, the interface shows abrupt slip weakening during which the coefficient of friction falls from 0.4 to ~ 0.2 in about 200 ns. The slip velocity during this time increases from no-slip to about 20 m/s. The slip weakening phase is followed by a distinct slip strengthening phase, during which the friction coefficient increases gradually to 0.4 and eventually leading to seizure of the interface.

88

The slip velocity decreases monotonically from 20 m/s to the no-slip level during this slip strengthening phase. It is interesting to note, during State 1, the coefficient of kinetic friction stays in the range of 0.2 and 0.4 (slip speeds < 20 m/s), which is much lower than those observed under quasi-static slip conditions ( ~ 0.6 to 0.85) at slip speeds of 1mm/s (Byerlee, 1978, Dieterich, 1978). The beginning of State 2 is marked by the arrival of the unloading longitudinal wave from the free surface of the target plate. The arrival of the unloading wave results in a step drop in normal stress from 778 MPa to 195 MPa at the slip interface. This sudden alteration in normal stress leads to the However, the coefficient of

re-initiation of interfacial slip at the friction interface.

kinetic friction during this slip is the range 0.6 to 0.8, which is considerably higher than the friction coefficients observed in State 1. Also, in contrast to State 1, the slip velocity in State 2 stays in the range of 16 to 21 m/s. These considerably high levels of slip speeds combined with the high coefficient of friction leads is expected to lead to a substantial increase in heat generation during the slip process in State 2.

89

Figure 2.13: History of slip velocity and coefficient of kinetic friction for Shot # 2. Figure 2.14 shows an estimate for the temperature rise during the dynamic slip event. During state 1, the maximum temperature rise attained at the frictional interface (x=0) occurs during the slip strengthening phase (prior to the seizure of the interface) and is estimated to be ~ 1000oC. After the re-initiation of slip (State 2) considerable frictional heat is generated and the interfacial temperatures at the slip interface increase monotonically to about 1400oC till the end of the window time of the experiment.

90

Figure 2.14: Estimated temperature rise at tribo-pair interface for Shot # 2. Figures 2.15(a) and 2.15(b) represent AFM scans of the glass surface before and after the experiment (Shot #2). A maximum bulk temperature rise of 1500oC was estimated during the experiment. Even though these interfacial temperatures are not expected to lead to any large scale melting of the glass, they are expected to result in considerable softening at the slip interface. Indeed, patches of sheared away molten glass can be clearly seen on the sliding surface of the glass layer in Figure 2.15(b). These

observations are consistent with the attainment of the high levels of the coefficient of friction attained during the later part of slip during which the interfacial temperatures were near the peak of the estimated temperatures.

91

Figure 2.15(a): AFM scan of soda-lime glass film before the pressure-shear friction experiment for Shot #2.

Figure 2.15(b): Post test AFM scan of soda-lime glass film showing regions that have melted away during frictional slip for Shot # 2.

92

The experimental results for Shot # 3 are shown in Figures 2.16 to 2.18. The experiment was conducted at an impact velocity of 54.3 m/s and a skew angle of 35 degrees. Figure 2.16 shows the transverse component of the particle velocity history at the free surface of the CH tool-steel target plate. Also shown in the figure are the normal stress and the elastic prediction of the transverse component of the particle velocity for the case of no-slip at the interface. The normal stress was obtained from the normal component of the free surface particle velocity by using the 1-D elastic wave theory presented in Section 2.2.3. In State 1 the normal stress is 762 MPa, while in State 2 normal stress is 191 MPa. The arrival of shear wave at the free surface of the target plate occurs at approximately 1900 ns. With the arrival of the shear wave at the free surface of the target plate the transverse component of the particle velocity jumps to the no-slip predicted level of about 23 m/s. After a momentary stay at this no-slip level the transverse velocity falls precipitously to a level of about 10 m/s before increasing again to a no slip level of 23 m/s. The transverse velocity stays at this no-slip level until the arrival of the unloading wave at the slip interface at which the transverse velocity again falls to a level of about 3 m/s

Figure 2.17 shows the history of the normal stress, coefficient of kinetic friction, and interfacial slip velocity as a function of time for Shot # 3. Two distinct friction states are obtained in the experiments: State 1 which begins just after impact and ends at ~ 1.9 s, and State 2 which ends at ~ 2.8 s. During State 1 the normal stress at the friction interface is constant at 762 MPa while the normal stress in State 2 is 191 MPa. During

93

State 1, after a momentary initial stick, the interface shows abrupt slip weakening during which the coefficient of friction falls from 0.4 to ~ 0.2 in about 200 ns. The slip velocity during this time increases from no-slip to about 20 m/s. The slip weakening phase is followed by a distinct slip strengthening phase, during which the friction coefficient increases gradually to 0.4 and eventually leading to seizure of the interface. The slip velocity decreases monotonically from 20 m/s to the no-slip level during this slip strengthening phase. It is interesting to note, during State 1, the coefficient of kinetic friction stays in the range of 0.2 and 0.4 (slip speeds < 20 m/s), which is much lower than those observed under quasi-static slip conditions ( ~ 0.6 to 0.85) at slip speeds of 1mm/s (Byerlee, 1978, Dieterich, 1978). The beginning of State 2 is marked by the arrival of the unloading longitudinal wave from the free surface of the target plate. The arrival of the unloading wave results in a step drop in normal stress from 762 MPa to 191 MPa at the slip interface. This sudden alteration in normal stress leads to the

re-initiation of interfacial slip at the friction interface.

94

Figure 2.16: Transverse particle velocity history at the rear surface of the target plate for Shot # 3.

Figure 2.17: History of slip velocity and coefficient of kinetic friction for Shot # 3.

95

Figure 2.18 shows an estimate for the temperature rise during the dynamic slip event. During state 1, the maximum temperature rise attained at the frictional interface (x=0) occurs during the slip strengthening phase (prior to the seizure of the interface) and is estimated to be ~ 850oC.

Figure 2.18: Estimated temperature rise at tribo-pair interface for Shot # 3.

Figures 2.19 and 2.20 show the experimental results for Shot FY008 obtained by using the plate impact friction experiment on fine grain Arkansas Novaculite rock. To conduct the experiments the multi-beam VALYN VISAR (Barker and Hollenbach, 1972) was employed. The use of VISAR allows measurement of combined transverse and normal components of particle velocity without the use of a diffraction grating on the rear surface of the target plate. The normal and transverse components of the particle

96

velocity particle are shown in Figure 2.19. Due to the heterogeneous nature of the rock sample some dispersion and spreading of the shear wave occurs. This dispersion

manifests itself as a rise time during the first 150 to 200 ns as the stress wave propagates in the Novaculite rock specimen. The history of the slip velocity and the coefficient of kinetic friction are shown in Figure 2.20. The slip velocity is observed to range from 5 m/s to 20 m/s during the dynamic slip event. As was the case in the glass-on-glass experiments, the friction stress in the case of novaculite rocks also shows an initial slip weakening followed by slip strengthening. Again, as in the case of the glass on glass experiments, the Novaculite rock samples show much lower levels of the coefficient of kinetic friction ( ~ 0.22-0.4) when compared to those obtained by Byerlee (1978) and Dieterich (1978) for rocks at slower slip rates ( 1mm/s). However, unlike the case for the glass-on-glass slip experiments, the final seizure condition was not attained during the available window time. The absence of the final seizure condition for the Novaculite rock is most likely because of the interfacial temperatures attained during slip at the rock interface, which are much lower than the melting/softening point of the novaculite rock. This can be seen more clearly from Figure 2.21, which shows the estimated temperature rise as a function of depth from the frictional interface in the rock sample. The

maximum temperature rise at the slip interface is predicted to be only 375 oC. Even though this temperature rise cannot lead to melting at the slip interface, the elevated temperatures do give rise to slip strengthening during the later part of the slip event.

97

Figure 2.19: Normal and transverse particle velocity history at the rear surface of the target plate for Shot FY008.

Figure 2.20: History of slip velocity and coefficient of kinetic friction for plate-impact pressure-shear experiment on novaculite rock samples (Shot FY008).

98

Figure 21: Estimated temperature rise at tribo-pair interface for the plate-impact pressure-shear friction experiment on novaculite rock samples (Shot FY008).

Figure 2.22 and 2.23 show the experimental results for Shot FY009 on fine grain Arkansas Novaculite rock. The slip velocity is observed to range from 5 m/s to 15 m/s during the dynamic slip event. As was the case in Shot FY008, the friction stress in the novaculite rocks also shows an initial slip weakening followed by slip strengthening. However, the slip weakening and the slip strengthening process are slower than that of Shot FY008. This is because of the lower normal pressure, lower slip velocity and lower shear stress in this experiment.

99

Figure 2.22: History of slip velocity and coefficient of kinetic friction for plate-impact pressure-shear experiment on novaculite rock samples (Shot FY009).

Figure 2.23: Estimated temperature rise at tribo-pair interface for the plate-impact pressure-shear friction experiment on novaculite rock samples (Shot FY009).

100

The first-weakening, with in the range of 0.4-0.2, is understood to be most likely due to thermal weakening induced by flash heating and incipient melting at asperity junctions while the second-strengthening, with in the range of 0.1 to 0.4, is understood to be a result of material softening and coalescence and solidification of melt patches. For the case of the fine grain Arkansas Novaculite rocks a similar range of slip conditions is observed.

2.4.2 Experimental Results obtained by Using the Modified Torsional Kolsky-bar Friction Experiments

Using Modified Torsional Kolsky-bar, the first series of experiments were conducted to investigate frictional resistance of quartz on quartz at co-seismic slip rates. Table 2.3 summarizes the two experiments conducted on quartz in the present study. In these two experiments, the tribo-pair comprised of a thin-walled tubular quartz specimen and a quartz disk, the RMS surface roughness of the quartz tube was 220 nm while the RMS surface roughness of the quartz disk was 3 nm. The outer and inner diameters for the tubular quartz specimen were 25 mm and 22 mm respectively. Experiment Q01 was conducted at input torque of 38.5 N.m and interfacial normal stress of 66.8 MPa, while experiment Q02 was conducted at input torque of 66.3 N.m and interfacial normal stress of 72.2 MPa. In this series of experiments, the aim was to keep the normal stress nearly the same but vary the input torque dramatically. By doing this, a higher slip speed was

101

expected in Q02, which can led to high flash temperatures, and thus lower the coefficient of friction due to thermal softening.

SHOT # Q01 Q02

Input Torque (N.m) 38.5 66.3

Normal Stress , MPa 66.8 72.2

Roughness of Quartz Tube (nm) 220 220

Roughness of Quartz Disk (nm) 3 3

Table 2.3: Summary of modified torsional Kolsky-bar friction experiments conducted on quartz.

The experimental results for Q01 are shown in Figures 2.24 to 2.28. Figure 2.24 shows the experimentally measured torque at the strain gage location A for Q01. The

input torsional pulse has a magnitude of 38.5 Nm and a pulse duration of 1000 s . Figure 2.25 shows history of interfacial normal pressure, shear stress and slip velocity. Figure 2.26 show the coefficient of kinetic friction, interfacial normal pressure and slip velocity as a function of slip distance. In order to calculate the interfacial normal pressure, shear stress, slip velocity, slip distance and coefficient of kinetic friction, the 1-D elastic wave theory, presented in Section 2.3.2 was used. Figure 2.27 shows the average temperature rise at the tribo-pair interface estimated by using the one-dimensional transient heat conduction equation (2.49). Figure 2.28 shows

sequential images of tribo-pair specimens from the high-speed camera with an inter-frame time of 50 s . In Figure 2.25, it is interesting to note that the interfacial

102

normal pressure maintains steady-state level of approximately 67 MPa until 250 s , then starts to decrease gradually. The drop of interfacial normal pressure is understood to be due to the initiation and propagation of micro-cracking in the tubular quartz specimen, which can be observed from the high-speed camera images, in Figure 2.28. The initiation of micro-cracking can be observed from the third frame of high-speed camera images, which corresponds to starting point of interfacial normal pressure drop (250 s ) in Figure 2.25. The interfacial shear stress rises from zero to a steady-state level of approximately 15 Mpa, and then decreases gradually to zero with the drop in normal stress. Due to the micro-cracking and fracture of tubular quartz specimen, the useful window time is only about 120 s . The coefficient of kinetic friction has an average value of 0.23 during the steady state, and then decreases to zero gradually. The average interfacial slip velocity is 2 m/s during the steady state. The corresponding accumulated slip distance is 1 mm. The average temperature rise at the tribo-pair

interfaces shown in Figure 2.27 is 140 0C, and is not high enough for macroscopic melting at the slip interface, so the much lower friction coefficients ( ~ 0.2) when compared to those obtained by Byerlee (1978) and Dieterich (1978) at quasi-static slip rates ( 1 mm/s, ~ 0.6 to 0.85) and the observed associated slip weakening are understood to be due to thermal weakening induced by the high flash temperatures at the asperity contacts.

103

60

Q01 Tribo-pair: Quartz/Quartz

40

20

Torque (N.m)

-20

-40

-60

Time (s)

1000

2000

Figure 2.24: Experimentally measured torque at the strain gage location A for Q01.

100 90 80 70

Q01 Tribo-pair: Quartz/Quartz Corresponding to First Frame Normal Pressure Initiation of Microcracking

10 9 8

60 50 40 30 20 10 0 0

6 5 4

Slip Velocity

3 2

Shear Stress
100 200

1 300 400 500 0 600

Time (s)

Figure 2.25: History of interfacial normal pressure, shear stress and slip velocity for Q01.

104

Slip Velocity (m/s)

Stress (MPa)

1 0.9

Q01 Tribo-pair: Quartz/Quartz

10 9 8

100 90 80 70 60 50 40 30 20 10

Coefficient of Kinetic Friction, k

0.8

0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5

Normal Pressure

Slip Velocity (m/s)

0.7

7 6 5 4 3 2 1

Slip Velocity

Coefficient of Kinetic Friction

0 0 1.5

Slip Distance (mm)

Figure 2.26: Interfacial normal pressure, slip velocity and coefficient of kinetic friction vs. interfacial slip distance for Q01.

300

Temperature Rise at the Interface (C)

Q01 Tribo-pair: Quartz/Quartz

250

200

150

100

50

100

200

300

400

500

600

Time (s)

Figure 2.27: Estimated temperature rise at tribo-pair interface for Q01.

105

Normal Pressure (MPa)

Figure 2.28: High-speed camera images for Q01.

The experimental results for Q02 are shown in Figures 2.29 to 2.33. Figure 2.29 shows the experimentally measured torque at the strain gage location A for Q02. The

input torsional pulse has a magnitude of 66.3 Nm and a pulse duration of 1000 s . Figure 2.30 shows history of interfacial normal pressure, shear stress and slip velocity. Figure 2.31 show the corresponding coefficient of kinetic friction, interfacial normal pressure and the slip velocity as a function of slip distance. In order to calculate the interfacial normal pressure, shear stress, slip velocity, slip distance and coefficient of kinetic friction, the 1-D elastic wave theory, presented in Section 2.3.2 was used. Figure 2.32 shows the average temperature rise at the tribo-pair interface estimated by using the one-dimensional transient heat conduction equation (2.49). Figure 2.33 shows

sequential images of tribo-pair specimens from the high-speed camera with an inter-frame time of 50 s . In Figure 2.30, it is interesting to note that the interfacial

106

normal pressure maintains steady-state level of approximately 72 MPa until 150 s , and then starts to decrease gradually. The drop of interfacial normal pressure can again be attributed to the initiation and propagation of micro-cracking on the tubular quartz specimen, which can be observed from the high-speed camera images in Figure 2.33. The initiation of micro-cracking occurs from the fifth frame of high-speed camera images, which corresponds to starting point of interfacial normal pressure drop (150 s ) in Figure 2.30. The interfacial shear stress rises from zero to a steady-state level of

approximately 14 Mpa, and then decrease gradually to zero because of the drop in normal stress. Due to micro-cracking and fracture of tubular quartz specimen, the window time for the steady state is only about 100 s for this experiment. The coefficient of kinetic friction has an average value of 0.20 (Figure 2.31) during the steady state, then decreases to zero gradually. The average interfacial slip velocity is ~ 4 m/s during the steady state. The corresponding accumulated slip distance is approximately 0.8 mm. The coefficient of kinetic friction during the steady state (~0.2) in this experiment is a little lower than that of experiment Q01 (~0.23). This can be attributed to the higher slip velocity attained in this experiment. The higher slip velocity is expected because of the higher input torque used in this experiment.

107

100 80 60 40

Q02 Tribo-pair: Quartz/Quartz

Torque (N.m)

20 0 -20 -40 -60 -80

-100

Time (s)

1000

2000

Figure 2.29: Experimentally measured torque at the strain gage location A for Q02.

100 90 80 70

Q02 Tribo-pair: Quartz/Quartz Corresponding to Second Frame Normal Pressure

10 9 8

60 50 40 30

Initiation of Microcracking

5 4 3

Slip Velocity
20 10 0 2 1

Shear Stress
0 100

Time (s)

200

0 300

Figure 2.30: History of interfacial normal pressure, shear stress and slip velocity for Q02.

108

Slip Velocity (m/s)

Stress (MPa)

1 0.9

Q02 Tribo-pair: Quartz/Quartz

10 9 8

100 90 80 70 60 50 40 30 20 10

Coefficient of Kinetic Friction, k

0.8

0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1

Slip Velocity (m/s)

0.7

Normal Pressure

7 6 5 4 3 2 1

Slip Velocity

Coefficient of Kinetic Friction

0 0 1.5

Slip Distance (mm)

Figure 2.31: Interfacial normal pressure, slip velocity and coefficient of kinetic friction vs. interfacial slip distance for Q02.

300

Temperature Rise at the Interface (C)

Q02 Tribo-pair: Quartz/Quartz

250

200

150

100

50

100

200

300

Time (s)

Figure 2.32: Estimated temperature rise at tribo-pair interface for Q02.

109

Normal Pressure (MPa)

Figure 2.33: High-speed camera images for Q02.

In experiments Q01-Q02, due to micro-cracking and fracture of the tubular quartz specimen, the accumulated slip distance was only 0.8~1 mm. In order to study the slip mechanism at larger accumulated slip distance, the thin-walled tubular quartz specimen is replaced by the thin-walled aluminum tube with a thin film of soda-lime glass (thickness ~ 1 m) deposited on one of its faces. Consequently the quartz disk specimen is replaced by a soda-lime glass disk. Table 2.4 summarizes the 6 experiments conducted on soda-lime glass on glass tribo-pair. The outer and inner diameters for the aluminum specimen were 25.4 mm and 21.6 mm respectively. In experiments G01 to G03 the interfacial normal stress was about 70 MPa, but the input torque was varied from 44.8 N.m to 81.5 N.m. In the experiments G04 to G06 the normal stress was kept ~ 40 Mpa, and the input torque was varied from 43.2 N.m to 68.8 N.m. For all six

experiments conducted on soda-lime glass, the RMS surface roughness of the soda lime

110

glass coating was kept as 18 nm while the RMS surface roughness of the soda lime glass disk was 4 nm.

Input SHOT # Torque (N.m) G01 G02 G03 G04 G05 G06 81.5 53.1 44.8 68.8 60.1 43.2

Normal Stress , MPa 76.1 69.2 71.9 44.5 38.0 37.8

Roughness of Soda-lime Glass Thin Film (nm) 18 18 18 18 18 18

Roughness of Soda-lime Glass Disk (nm) 4 4 4 4 4 4

Table 2.4: Summary of modified torsional Kolsky-bar friction experiments conducted on soda-lime glass.

The experimental results for G01 are shown in Figures 2.34 to 2.36. Figure 2.34 shows the experimentally measured torque at the strain gage location A for experiment G01. The input torsional pulse has a magnitude of 81.5 N.m and a pulse duration of 1000 s . Figure 2.35 shows the coefficient of kinetic friction and the interfacial slip velocity as a function of the slip distance. In order to calculate the interfacial slip velocity and coefficient of kinetic friction, the 1-D elastic wave theory, presented in Section 2.3.2, was used. Figure 2.36 shows the average temperature rise at the

tribo-pair interface estimated by using the one-dimensional transient heat conduction

111

equation.

The interfacial normal pressure is nearly constant throughout the entire The interfacial slip velocity rises from zero to

duration of the experiments.

approximately 5 m/s and then oscillates in the vicinity of this level. The accumulated slip distance is around 3.8 mm. The coefficient of kinetic friction during the early part of slip has an average value of 0.19; however, with the accumulation of slip, decreases from 0.19 to 0.12 indicating slip weakening of the slip interface, finally increases again to 0.14 indicating slip strengthening followed the slip weakening. The average

temperature rise at the tribo-pair interfaces shown in Figure 2.36 is 500 0C, and is not high enough for macroscopic melting, so the much lower friction coefficients ( ~ 0.2) can be attributed to thermal weakening induced by the highe flash temperature at asperity contacts.

120 100 80 60

G01 Tribo-pair: Soda-lime glass/Soda-lime glass

Torque (N.m)

40 20 0 -20 -40 -60 -80

-100

Time (s)

1000

2000

Figure 2.34: Experimentally measured torque at the strain gage location A for G01.

112

0.5

G01 Tribo-pair: Soda-lime glass/Soda-lime glass

10 9 8

Coefficient of Kinetic Friction, k

0.4

0.3

Interfacial Slip Velocity

6 5

0.2

Normal Stress: 76.1 MPa

4 3

0.1

Coefficient of Kinetic Friction, k

0 0 1 2 3 4

1 0

Slip Distance (mm)

Figure 2.35: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G01.

1000

Temperature Rise at the Interface (C)

900 800 700 600 500 400 300 200 100 0 0

G01 Tribo-pair: Soda-lime glass/Soda-lime glass

200

400

600

800

1000

Time (s)

Figure 2.36: Estimated temperature rise at tribo-pair interface for G01.

113

Slip Velocity (m/s)

The experimental results for G02-G03 are shown in Figures 2.37 to 2.42. These two experiments were designed tohave a similar normal stress but a lower slip velocity when compared to G01. For experiment G02, the input torsional pulse has a magnitude of 53.1 N.m and a pulse duration of 1000 s , the interfacial slip velocity rises from zero to approximately 2.5 m/s and then oscillates in the vicinity of this level. The

accumulated slip distance is 2.2 mm for this experiment. The coefficient of kinetic friction during the early part of slip has an average value of 0.29, and then decreases to 0.21 with the accumulation of slip indicating slip weakening at the slip interface, finally increases again to 0.27 indicating slip strengthening followed the slip weakening. For experiment G03, the input torsional pulse has a magnitude of 44.8 N.m and a pulse duration of 1000 s . The interfacial slip velocity reaches ~ 2.0 m/s and then oscillates in the vicinity of this level. The accumulated slip distance is around 1.7 mm. The coefficient of kinetic friction during the early part of slip has an average value of 0.28, then decreases to 0.22 with the accumulation of slip indicating slip weakening at the slip interface, finally increases again to 0.27 indicating slip strengthening followed the slip weakening. In both experiments, the average temperature rises at the tribo-pair

interfaces are also much lower than the melting point of soda lime glass.

114

80

60

G02 Tribo-pair: Soda-lime glass/Soda-lime glass

40

Torque (N.m)

20

-20

-40

-60

Time (s)

1000

2000

Figure 2.37: Experimentally measured torque at the strain gage location A for G02.

0.5

G02 Tribo-pair: Soda-lime glass/Soda-lime glass

10 9 8

Coefficient of Kinetic Friction, k

0.4

Coefficient of Kinetic Friction, k

0.3

6 5

0.2

Normal Stress: 69.2 MPa

4 3

0.1

Interfacial Slip Velocity

1 0 0 1 2 3 0

Slip Distance (mm)

Figure 2.38: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G02.

115

Slip Velocity (m/s)

1000

Temperature Rise at the Interface (C)

900 800 700 600 500 400 300 200 100 0 0

G02 Tribo-pair: Soda-lime glass/Soda-lime glass

200

400

600

800

1000

Time (s)

Figure 2.39: Estimated temperature rise at tribo-pair interface for G02.

60 50 40 30

G03 Tribo-pair: Soda-lime glass/Soda-lime glass

Torque (N.m)

20 10 0 -10 -20 -30 -40 0 1000 2000

Time (s)

Figure 2.40: Experimentally measured torque at the strain gage location A for G03.

116

0.5

G03 Tribo-pair: Soda-lime glass/Soda-lime glass

10 9 8

Coefficient of Kinetic Friction, k

0.4

0.3

6 5

0.2

Normal Stress: 71.9 MPa

0.1

3 2

Interfacial Slip Velocity

0 0 0.5 1 1.5 2

1 0

Slip Distance (mm)

Figure 2.41: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G03.

1000

Temperature Rise at the Interface (C)

900 800 700 600 500 400 300 200 100 0 0

G03 Tribo-pair: Soda-lime glass/Soda-lime glass

200

400

600

800

1000

Time (s)

Figure 2.42: Estimated temperature rise at tribo-pair interface for G03.

117

Slip Velocity (m/s)

Coefficient of Kinetic Friction, k

The experimental results for G04-G06 are shown in Figures 2.43 to 2.51. These three experiments were designed to run at a normal stress of ~ 40 Mpa, which is lower than that obtained in G01-G03 (~ 70 MPa). For experiment G04, the input torsional pulse was N.m with a pulse duration of 1000 s , the interfacial slip velocity rises from zero to approximately 4.0 m/s and then oscillates in the vicinity of this level. The accumulated slip distance is around 3.3 mm for this experiment. The coefficient of kinetic friction during the early part of slip has an average value of 0.27, and then decreases to 0.21 with the accumulation of slip indicating slip weakening at the slip interface, finally increases again to 0.25 indicating slip strengthening followed the slip weakening. For experiment G05, the input torsional pulse has a magnitude of 60.1 N.m and a pulse duration of 1000 s ; the interfacial slip velocity rises from zero to approximately 3.3 m/s and then oscillates in the vicinity of this level. The

corresponding accumulated slip distance is 2.5 mm for this experiment. The coefficient of kinetic friction during the early part of slip has an average value of 0.30, and then decreases 0.15 with the accumulation of slip indicating slip weakening at the slip interface, finally increases again to 0.24 indicating slip strengthening followed the slip weakening. For experiment G06, the input torsional pulse has a magnitude of 43.2 N.m and a pulse duration of 1000 s ; the interfacial slip velocity rises from zero to approximately 2.3 m/s and then oscillates at this level. The accumulated slip distance is ~ 1.9 mm for this experiment. The coefficient of kinetic friction during the early part of slip has an average value of 0.31, and then decreases to 0.15 with the accumulation of
118

slip indicating slip weakening at the slip interface, finally increases again to 0.26 indicating slip strengthening followed the slip weakening. The average temperature rise at the tribo-pair interface (400, 300, 200 0C, respectively) in all the three experiments is much lower than the melting point of soda lime glass in these three experiments.

100 80 60 40

G04 Tribo-pair: Soda-lime glass/Soda-lime glass

Torque (N.m)

20 0 -20 -40 -60 -80 0 1000 2000

Time (s)

Figure 2.43: Experimentally measured torque at the strain gage location A for G04.

119

0.5

G04 Tribo-pair: Soda-lime glass/Soda-lime glass

10 9 8

Coefficient of Kinetic Friction, k

0.4

0.3

Coefficient of Kinetic Friction, k

6 5

0.2

Interfacial Slip Velocity

0.1

3 2

Normal Stress: 44.5 MPa

1 0 0 1 2 3 4 0

Slip Distance (mm)

Figure 2.44: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G04.

1000

Temperature Rise at the Interface (C)

900 800 700 600 500 400 300 200 100 0 0

G04 Tribo-pair: Soda-lime glass/Soda-lime glass

200

400

600

800

1000

Time (s)

Figure 2.45: Estimated temperature rise at tribo-pair interface for G04.

120

Slip Velocity (m/s)

100 80 60 40

G05 Tribo-pair: Soda-lime glass/Soda-lime glass

Torque (N.m)

20 0 -20 -40 -60 -80 0 1000 2000

Time (s)

Figure 2.46: Experimentally measured torque at the strain gage location A for G05.

0.5

G05 Tribo-pair: Soda-lime glass/Soda-lime glass

10 9 8

Coefficient of Kinetic Friction, k

0.4

0.3

Coefficient of Kinetic Friction, k

6 5

0.2

4 3

0.1

Interfacial Slip Velocity Normal Stress: 38.0 MPa

2 1 3 0

Slip Distance (mm)

Figure 2.47: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G05.

121

Slip Velocity (m/s)

1000

Temperature Rise at the Interface (C)

900 800 700 600 500 400 300 200 100 0 0

G05 Tribo-pair: Soda-lime glass/Soda-lime glass

250

500

750

1000

Time (s)

Figure 2.48: Estimated temperature rise at tribo-pair interface for G05.

80

60

G06 Tribo-pair: Soda-lime glass/Soda-lime glass

40

Torque (N.m)

20

-20

-40

-60

Time (s)

1000

2000

Figure 2.49: Experimentally measured torque at the strain gage location A for G06.

122

0.5

G06 Tribo-pair: Soda-lime glass/Soda-lime glass

10 9 8

Coefficient of Kinetic Friction, k

0.4

0.3

Coefficient of Kinetic Friction, k

6 5

0.2

Interfacial Slip Velocity

4 3

0.1

Normal Stress: 37.8 MPa

1 0 0 0.5 1 1.5 2 0

Slip Distance (mm)

Figure 2.50: Slip velocity and coefficient of kinetic friction vs. interfacial slip distance for G06.

1000

Temperature Rise at the Interface, C

900 800 700 600 500 400 300 200 100 0 0

G06 Tribo-pair: Soda-lime glass/Soda-lime glass

200

400

600

800

1000

Time (s)

Figure 2.51: Estimated temperature rise at tribo-pair interface for G06.

123

Slip Velocity (m/s)

Figure 2.52 summarizes the highest coefficient of kinetic friction versus interfacial slip velocity obtained from all 6 experiments conducted on the soda lime glass/soda lime glass tribo-pair. It is interesting to note that the highest coefficient of kinetic friction decreases with the increasing slip velocity. Moreover, the interfacial normal pressure has negligible effect on the highest coefficient of kinetic friction.

0.4

Highest Coefficient of Kinetic Friction, k

0.3

0.2

0.1

Experimental data with normal stress of about 70 MPa Experimental data with normal stress of about 40 MPa

Interfacial Slip Velocity (m/s)

Figure 2.52: Highest coefficient of kinetic friction versus interfacial slip velocity obtained from experiments conducted on the soda lime glass/soda lime glass tribo-pair.

Figures 2.53 and 2.54 show SEM micrographs for sliding surfaces of the soda-lime glass specimens tested in G01, slip direction is from the bottom to the top. Sliding wear tracks and brittle cleavage fracture due to thermal softening can be clearly observed on

124

the slip surfaces.

Figure 2.53: SEM micrograph for sliding surface of soda-lime glass thin film tested in G01.

Figure 2.54: SEM micrograph for sliding surface of soda-lime glass disk specimen tested in G01.

2.5 Summary

In the present study plate-impact pressure-shear friction experiments and modified

125

torsional Kolsky-bar friction experiments are employed to investigate the frictional resistance in rocks and analog materials (quartz and soda lime glass), at relevant normal pressures and co-seismic slip rates. The results of the plate-impact pressure-shear

friction experiments indicate that a wide range of frictional slip conditions exist at the frictional interface. These conditions range from initial no-slip followed by slip

weakening, strengthening, and seizure all during a single slip event for the case of glass vs. glass experiments. Moreover, the range of the coefficient of kinetic friction during slip is much lower than those obtained at quasi-static slip rates ( 1 mm/s).

In the modified torsional Kolsky-bar friction experiments on quartz, the interfacial normal pressure maintains steady-state level of approximately 70 MPa for a while, then starts to decrease gradually. The drop of interfacial normal pressure could be attributed to the initiation and propagation of micro-cracking on the tubular quartz specimen, which can be observed from the high-speed camera images. The coefficient of kinetic friction has an average value of approximately 0.20-0.23 during the steady friction state, then decreases to zero gradually. Due to the micro-cracking and fracture of tubular quartz specimen, the window time for the steady-state friction is only 100-120 s for these experiments, and the accumulated slip distance is only 0.8~1 mm. In order to solve this problem and study the slip mechanism at larger slip distance, the thin-walled tubular quartz specimen is replaced by the thin-walled tubular aluminum specimen with a thin film of soda-lime glass (thickness ~ 1 m) deposited on one of its faces in the second
126

series of experiments. The quartz disk specimen also is replaced by the soda-lime glass disk specimen correspondingly. In the modified torsional Kolsky-bar friction

experiments on soda-lime glass, the coefficient of kinetic friction (0.2~0.3) is much lower than those obtained by Byerlee (1978) and Dieterich (1978) at quasi-static slip rates ( 1 mm/s, ~ 0.6 to 0.85). It is also interesting to note that the highest coefficient of kinetic friction decreases with the increasing slip velocity. In every single experiment, the coefficient of kinetic friction goes down to a lower level and slip weakening is observed when the larger slip develops during the tribo-pair interfaces. Also, following the slip weakening, there is a slip strengthening. This slip mechanism can also be seen in the pressure-shear experiments.

Recent measurements of contact area and contact indentation strength, C , in transparent materials (such as quartz) by light scattering (Dieterich and Kilgore, 1994, 1996), confirmed earlier suggestions by Boitnott et al.(1992) that the shear strength, C , of the asperity junctions is very high (estimated to be of order 10% of the shear modulus
G) in typical rock systems, and thus when forced to shear, they generate intense but

highly localized heating during their life time. The local shear strength, C , of the asperity contact presumably degrades continuously with increasing flash temperature,
Ta .

An elementary model considering flash heating at asperity contacts had been

127

proposed recently by Rice (2006). The model considers contacts of uniform size L , and hence life-time L /V , where L is the slip needed to renew the asperity contact population, and V is slip rate; and assumes that their shear strength remains at the room temperature value, C , until temperature has reached a critical value, Tcrit , above which the weakened shear strength W is assumed to has a negligibly small compared to its room temperature value C (see Figure 2.55). The temperature rise of the asperities is estimated from a simple one-dimensional transient heat conduction equation, with heating power CV per unit area at the sliding contact surface. Tullis and Goldsby (2003) observed that a better fit of Rices model to data showing strong rate-weakening of friction was to assume a small but non-negligible W for Ta > Tcrit . The model thus defines a critical slip rate Vcrit such that there is no weakening if V < Vcrit , but strong weakening if V > Vcrit . That is, the friction coefficient , which has the value

0 at low slip rates and W at high slip rates, is given in the following simple model

= 0 if V < Vcrit , = (0 W )

Vcrit V

+ W if V > Vcrit ;

(2.50)

Vcrit =

c (Tcrit Tbulk )

(2.51)

where is thermal diffusivity, c is heat capacity per unit volume.

128

Figure 2.55: Flash heating at asperity contacts

Using Equations (2.50) and (2.51), a Vcrit = 0.12 m/s can been estimated to be onset of severe thermal weakening in glass and other geo-materials, when it is recognized that L = 5 m , C = 3.0 GPa and temperature range from ambient up to ~ 900 0C. Recent laboratory studies (Goldsby and Tullis, 2002, Di Toro et al., 2004, Hirose and Shimamoto, 2005) and the results presented in the present study on rapid slip on geological and analog materials are indeed consistent with such an expected reduction in friction stress. The results suggest that the friction coefficient is in the range of 0.2 to 0.4 at typical seismic slip rates of > 1 m/s in geo-materials.

Based on the aforementioned model, we propose that the first slip-weakening in soda-lime glass and dense novaculite rock is most likely due to the thermal weakening induced by flash heating and incipient melting at asperity contacts, while the

129

strengthening is a result of material softening and coalescence and solidification of melt patches at the slip surface. The much lower friction coefficients in soda-lime glass and analog materials (i.e. ~ 0.2 - 0.4) when compared with ~ 0.6 to 0.85 obtained by Byerlee (1978) and Dieterich (1978) at slow slip rates ( 1mm/s) is caused by flash heating. Moreover, in the modified torsional Kolsky-bar friction experiments, the

reduction of the coefficient of kinetic friction with increasing slip velocities and slip weakening when the larger slip develops during the tribo-pair interfaces are also due to thermal weakening at asperity contacts. These relatively low values of for rocks and analog materials obtained in the current experiments may explain fault weakening during seismic slip. Also, the large decrease in at higher slip rates is consistent with rupture propagating as a self-healing slip pulse (Beeler and Tullis, 1996). Moreover, the experimental results for rocks and analog materials at such high slip rates (0~20 m/s) will help to establish the constitutive description of the behavior that can be used in dynamic models of earthquake rupture.

130

References

Adams, G. G., 1995. Self excited oscillations of two elastic half-spaces sliding with a constant coefficient of friction. Journal of applied mechanics 62, 867-872. Andrews, D. J., and Ben-Zion, Y., 1997. Wrinkle-like slip pulse on a fault between different materials. Journal of Geophysical Research 102, 553-571. Andrews, D. J., 2002. A fault constitutive relation accounting for thermal pressurization of pore fluid. Journal of Geophysical Research 107, 2363, doi: 2310.1029 / 2002JB001942, ESE 001915-001941001915-001948. Antoniou, S., Cameron, A., and Gentle, C., 1976. The friction-speed relation from stick-slip data. Wear 36, 235-254. Archard, J. F., 1958. The temperature of rubbing surfaces. Wear 2, 438-455. Armstrong-Helouvry, B., Dupont, P., and Canudas de Wit, C., 1994. A survey of models, analysis tools and compensation methods for the control of machines with friction. Automatica 30, 1083-1153. Ashby, M. F., Abulawi, J., and Kong, H. S., 1991. Temperature maps for frictional heating in dry sliding. Tribology transactions 34, 577-587. Barber, J. R., 1976. Some thermoelastic contact problems involving frictional heating. The Quarterly Journal of Mechanics and Applied Mathematics 29, 1-13. Barker, L. M., and Hollenbach, R. E., 1972. Laser interferometer for measuring high velocities of any reflecting surface. Journal Applied Physics 43, 4669-4675. Beeler, N. M., and Tullis, T. E., 1996. Self-healing slip pulses in dynamic rupture models due to velocity-dependent strength. Bulletin of the Seismological Society of America 86, 1130-1148. Bell, R., and Burdekin, M., 1969. A study of stick-slip motion of machine tool feed drives. Proceedings Institute of Mechanical Engineers 184: Part I, 543-560. Ben-Zion, Y., 2001. Dynamic ruptures in recent models of earthquake faults. Journal of Mechanics and Physics of Solids 49, 2209-2244. Blanpied, M. L., Tullis, T. E., and Weeks, J. D., 1998. Effects of slip, slip rate, and shear heating on the friction of granite. Journal of Geophysical Research-Solid Earth 103, 489-511. Boitnott, G. N., Biegel, R. L., Scholz, C. H., N., Y., and W., W., 1992. Micromechanics of rock friction 2: Quantitative modeling of initial friction with contact theory. Journal of Geophysics Research 97, 8965-8978.

131

Bowden, F. P., and Thomas, P. H., 1954. The surface temperature of sliding solids. Proceedings of Royal Society of London A 223, 29-36. Bowden, F. P., and Persson, P. A., 1960. Deformation heating and melting of solids in high speed friction. Proceedings of Royal Society of London A260, 433-458. Bureau, L., Baumberger, T., and Caroli, C., 2000. Shear response of a frictional interface to a normal load modulation. Physics Review Letters E 62, 6810-6820. Byerlee, J. D., 1978. Friction of rocks. Pure and Applied Geophysics 116, 615-626. Cochard, A., and Rice, J. R., 2000. Fault rupture between dissimilar materials: Ill-posedness, regularization, and slip-pulse response. Journal of geophysical research 105, 891-907. Di Toro, G., Goldsby, D. L., and Tullis, T. E., 2004. Friction falls towards zero in quartz rock as slip velocity approaches seismic rates. Nature 427, 436-439. Dieterich, J. H., 1978. Time dependent friction and mechanics of stick-slip. Pure and Applied Geophysics 116, 668-675. Dieterich, J. H., 1979. Modeling of rock friction: I, experimental results and constitutive equations. Journal of Geophysical Research 84, 2161-2168. Dieterich, J. H., 1981. Constitutive properties of faults with simulated gouge. In: Carter, N. L., Friedman, M., Logan, J. M., and Stearns, D. W., (Eds.), vol. 24. American Geophysical Union, Washington, D.C., pp. 103-120. Dieterich, J. H., and Kilgore, B. D., 1994. Direct observations of frictional contacts--new insights for state-dependent properties. Pure and Applied Geophysics 143, 283-302. Dieterich, J. H., and Kilgore, B. D., 1996. Implications of fault constitutive properties for earthquake prediction. Proceedings of the National Academy of Sciences of the United States of America 93, 3787-3794. Dieterich, J. H., and Kilgore, B. D., 1996. Imaging surface contacts: Power law contact distributions and contact stresses in quartz, calcite, glass and acrylic plastic. Tectonophysics 256, 219-239. Goldsby, D. L., and Tullis, T. E., 2002. Low frictional strength of quartz rocks at sub-seismic slip rates. Geophysical research letters 29, 1844-1855. Heaton, T. H., 1990. Evidence for and implications of self-healing pulses of slip in earthquake rupture. Physics of the Earth and Planetary Interiors 64, 1-20. Hirose, T., and Shimamoto, T., 2005. Growth of molten zone as a mechanism of slip weakening of simulated faults in gabbro during frictional melting. Journal of Geophysical Research-Solid Earth 110.

132

Hirose, T., and Shimamoto, T., 2005. Slip-weakening distance of faults during frictional melting as inferred from experimental and natural pseudotachylytes. Bulletin of the Seismological Society of America 95, 1666-1673. Irfan, M. A., and Prakash, V., Contact temperatures during sliding in pressure shear impact. Proceedings Society of Experimental Mechanics Conference, Baltimore, MD, 1994, pp. 173-182. Irfan, M. A., and Prakash, V., 2000. Time resolved friction during dry sliding of metal on metal. International journal of solids and structures 37, 2859-2882. Kim, K. S., Clifton, R. J., and Kumar, P., 1977. A combined normal and transverse displacement interferometer with an application to impact of Y-cut Quartz. Journal of Applied Physics 48, 4132-4139. Kuhlmann-Wilsdorf, D., 1985. Flash temperatures due to friction and joule heat at asperity contact. Wear 105, 187-198. Kumar, P., and Clifton, R. J., 1977. Optical alignment of impact faces for plate impact experiments. Journal of Applied Physics 48, 1366-1367. Lachenbruch, A. H., 1980. Frictional heating, fluid pressure, and the resistance to fault motion. . Journal of Geophysical Research 85, 60976122. Lee, T. C., and Delaney, P. T., , 1987. Frictional heating and pore pressure rise due to a fault slip Geophys J. Roy. Astronom. Society 88, 569-591. Linker, M. F., and Dieterich, J. H., 1992. Effects of variable normal stress on rock friction: observations and constitutive equations. Journal of Geophysical Research 97, 4923-4940. Liou, N. S., Okada, M., and Prakash, V., 2004. Formation of Molten Metal Films During Metal-on-Metal Slip Under Extreme Interfacial Conditions. Journal of the Mechanics and Physics of Solids 52, 2025-2056. Marone, C., 1998. Laboratory-derived friction laws and their application to seismic faulting. Annual Review Earth and Planetary Science 26, 643-696. Martins, J. A. C., Oden, J. T., and Simes, F. M., 1990. A study of static and kinetic friction. International Journal of Engineering Science 28, 29-92. Mase, C. W., and Smith, L., 1985. Porefluid pressures and frictional heating on a fault surface. Pure Appl. Geophys. 122, 583-607. Mase, C. W., and L., S., 1987. Effects of frictional heating on the thermal, hydrologic, and mechanical response of a fault. J. Geophys. Res. 92, 62496272. Mizoguchi, K., Hirose, T., Shimamoto, T., and Fukuyama, E., 2006. Moisture-related weakening and strengthening of a fault activated at seismic slip rates. Geophysical

133

Research Letters 33. Noda, H., and Shimamoto, T., 2005. Thermal pressurization and slip-weakening distance of a fault: An example of the Hanaore fault, southwest Japan. Bulletin of the Seismological Society of America 95, 1224-1233. O'Hara, K., Mizoguchi, K., Shimamoto, T., and Hower, J. C., 2006. Experimental frictional heating of coal gouge at seismic slip rates: Evidence for devolatilization and thermal pressurization of gouge fluids. Tectonophysics 424, 109-118. Okada, M., Liou, N. S., and Prakash, V., 2002. Dynamic shearing resistance of molten metal films at high pressures. Experimental Mechanics. Prakash, V., and Clifton, R. J., 1993. Time Resolved Dynamic Friction Measurements in Pressure-Shear. In: Ramesh, K. T., (Ed). Experimental Techniques in the Dynamics of Deformable Bodies, vol. AMD Vol. 165. ASME, New York, pp. 33-47. Prakash, V., 1995. A pressure-shear plate impact experiment for investigating transient friction. Experimental Mechanics 35, 329-336. Prakash, V., 1998. Friction response of sliding interfaces subjected to time varying normal pressures. Journal of Tribology 120, 97-102. Rajagopalan, S. and Prakash, V., 1999. A modified Kolsky bar for investigating dynamic friction. Experimental Mechanics 39(4), 295-303. Ranjith, K., and Rice, J. R., 2001. Slip dynamics at an interface between dissimilar materials. Journal of the mechanics and physics of solids 49, 341-361. Rempel, A. W., and Rice, J. R., 2006. Thermal pressurization and onset of melting in fault zones. Journal of Geophysical Research-Solid Earth 111. Rice, J. R., and Ruina, A., 1983. Stability of steady frictional sliding. Journal of Applied Mechanics 50, 343-349. Rice, J. R., 2006. Heating and weakening of faults during earthquake slip. Journal of Geophysical Research-Solid Earth 111. Richardson, E., and Marone, C., 1999. Effects of normal force vibrations on friction healing. Journal of Geophysical Research 104, 28859-28878. Rudnicki, J. W., and Rice, J. R., 2006. Effective normal stress alteration due to pore pressure changes induced by dynamic slip propagation on a plane between dissimilar materials. Journal of Geophysical Research-Solid Earth 111. Ruina, A., 1983. Slip stability and state variable friction laws. Journal of Geophysical Research 88, 10359-10370. Scholz, C. H., Molnar, P., and Johnson, T., 1972. Detailed studies of frictional sliding of
134

granite and implications for the earthquake mechanism. Journal of geophysical research 77, 6392-6406. Scholz, C. H., 1998. Earthquakes and friction laws. Nature 243, 37-42. Segall, P., and Rice, J. R., 2006. Does shear heating of pore fluid contribute to earthquake nucleation? Journal of Geophysical Research-Solid Earth 111. Sibson, R. H., 1973. Interaction between temperature and pore-fluid pressure during earthquake faulting A mechanism for partial or total stress relief. Nature 243, 66-68. Simes, F. M. F., and Martins, J. A. C., 1998. Instability and ill-posedness in some friction problems. International journal of Engineering Science 36, 1265-1293. Sulem, J., Vardoulakis, I., Ouffroukh, H., and Perdikatsis, V., 2005. Thermo-poro-mechanical properties of the Aigion fault clayey gouge - Application to the analysis of shear heating and fluid pressurization. Soils and Foundations 45, 97-108. Tsutsumi, A., and Shimamoto, T., 1997. High-velocity frictional properties of gabbro. Geophysical Research Letters 24, 699-702. Tullis, T. E., 1994. Predicting earthquakes and the mechanics of fault slip. Geotimes 39, 19-21. Tullis, T. E., and Goldsby, D. L., 2003. Flash melting of crustal rocks at almost seismic slip rates. Eos Trans. AGU 84(46), Fall Mtg. Suppl., Abstract S51B-05. Ullah, H., Irfan, M. A., and Prakash, V., 2007. State and rate dependent friction laws for modeling high-speed frictional slip at metl-on-metal interfaces Journal of Tribology 129, 17-22. Weeks, J. D., Beeler, N. M., and Tullis, T. E., Frictional behavior; glass is like a rock. 1991, pp. 457-458. Weertman, J., 1980. Unstable slippage across a fault that separates elastic media of different elastic constants. Journal of Geophysical Research 85, 1455-1461. Wibberley, C. A. J., 2002. Hydraulic diffusivity of fault gouge zones and implications for thermal pressurization during seismic slip. . Earth Planets Space 54, 1153-1171. Wibberley, C. A. J., and Shimamoto, T., 2005. Earthquake slip weakening and asperities explained by thermal pressurization. Nature 436, 689-692.

135

Chapter 3

HIGH-SPEED FRICTION AT METAL-ON-METAL INTERFACES

3.1 Introduction When two contacting solids slide, work is done against friction. Almost all the

frictional work appears as heat, generated at or very close to the sliding interface. Frictional heating at relatively low slip speeds is a transient and localized phenomenon and does not have a great influence on the resultant friction mechanism. In marked

contrast to this, at relatively high sliding speeds the metal surfaces are subjected to intense frictional heating which profoundly changes the state of the tribo-pair interface (Bowden and Ridler, 1936; Bowden and Tabor, 1950; Bowden and Thomas, 1954; Archard, 1958; Bowden and Persson, 1960; Barber, 1970; Kennedy, 1984; Kuhlmann-Wilsdorf, 1985; Carslaw and Jaeger, 1986; Irfan and Prakash, 1994). Under these conditions shearing of material at the frictional interface plays a fundamental role in controlling the dynamic slip process, resulting in a sharp local rise in temperature. Since the sheared layers are very thin and the sliding speeds are very high, extraordinary high rates of strain are obtained.

In some of the early works to investigate the phenomenon of high speed frictional slip, Shugarts (1953) developed an experimental apparatus in which a rod was pressed against the rim of a rapidly revolving disk. Krafft (1955) investigated high speed

friction in a series of ballistic tests where friction occurred between the surface of a
136

rotating bullet and the steel target it penetrated.

Williams and Griffen (1964),

Montgomery (1976a), Kadhim and Earles (1966), Earles and Powell (1966), Lim et al. (1989) and Sternlicht and Apkarian (1960) provided evidence of melt in high speed sliding situations using a high-speed pin-on-disk test device. Montgomery (1976b)

investigated the acceleration of projectiles in gun barrels and estimated friction coefficient for several tribo-pair materials. He estimated coefficient of friction to be as low as 0.02 for a gilding metal (90% Cu and 10% Zn) and steel tribo-pair.

In some of the most celebrated studies of high speed friction, Bowden and Freitag (1958) and Bowden and Persson (1960) spun a steel ball to a very high rotational speed and then grabbed it with other frictional samples or dropped it on another to achieve slip speeds as high as 800 m/s. They observed velocity weakening of the frictional force as a function of increasing slip velocity. However, in their experiments the normal Higher interfacial pressures

pressures were quite low (less than 0.015 MPa).

(approximately 50 to 100 MPa) but at relatively lower slip speeds (approximately 5 m/s) were obtained in experiments by Ogawa (1997), who modified a split Hopkinson pressure bar to study impact friction by axially impinging the input tube on a rotating output tube. By replacing the transmitter bar with a rigid disk, Rajagopalan and

Prakash (1999) modified the conventional torsional Kolsky bar to investigate high speed friction at normal pressures of approximately 50 to 100 MPa and slip velocities up to 10 m/s.

Most recently, plate impact pressure-shear friction experiments were conducted by

137

Irfan and Prakash (2000) to understand time resolved frictional characteristics during high-speed sliding of metal on metal (Carpenter Hampden tool-steel/Ti-6Al-4V tribo-pair). In their experiments interfacial slip speeds of 1 to 60 m/s at normal pressures ranging from 1 to 3 GPa were obtained. The combination of the high slip speeds and high normal pressures led to interfacial temperatures of ~ 800 oC in approximately 3-4 s of frictional slip. However, because of the relative high melt-point

temperatures of the tribo-pair materials employed in the study, near surface melting was not observed. In a later study, by employing tribo-pairs comprising hard tool-steel

against relatively low melt-point metals such as 7075-T6 aluminum alloy, Okada (2001; 2002) and Liou et al. (2004) investigated time-resolved growth of molten metal films during dry metal-on-metal slip under extreme interfacial conditions (normal stress ranging from 1 to 3 GPa and interfacial slip speeds in the range of 1 to 100 m/s).

In the present study we capitalize on the aforementioned recent developments in the plate impact pressure-shear friction experiments (Prakash and Clifton, 1993; Prakash, 1995; Irfan and Prakash, 2000; Okada et al., 2001) and investigate dynamic slip resistance and time-resolved growth of molten metal films during dry metal-on-metal slip under extreme interfacial conditions. By employing tribo-pairs comprising hard

tool-steel against relatively low melt-point metals such as 7075-T6 aluminum alloy, interfacial friction stress of up to 300 MPa and slip speeds of approximately 250 m/s have been achieved. These relatively extreme interfacial conditions are conducive to In order to interpret

the development of molten metal films at the tribo-pair interface.

the experimental results a Lagrangian finite element based computational procedure is

138

developed.

Scanning electron microscopy is carried out to understand the Besides being useful in the study of

microstructures formed during the sliding process.

high speed friction, the results of the plate impact friction experiments also provide fundamental information on the dynamic shearing resistance of molten films at normal pressures of approximately 14.5 GPa and shear strain rates of ~ 107 s-1.

3.2 Experimental Methods 3.2.1 Experimental Configuration The plate-impact pressure-shear friction experiments were conducted using the 82.5 mm bore single-stage gas gun facility at Case Western Reserve University. schematic of the experimental configuration is shown in Figure 3.1. The

A fiberglass

projectile carrying the flyer plate is accelerated down the gun barrel by means of compressed air. The rear end of the projectile has sealing O-ring and a Teflon key that slides in a key-way inside the gun barrel to prevent any rotation of the projectile. In

order to reduce the possibility of an air cushion between the flyer and target plates, impact takes place in a target chamber that is evacuated to a pressure of ~ 50 m of Hg prior to impact. To ensure the generation of plane waves with wave-front sufficiently

parallel to the impact face, the flyer and the target plates are carefully aligned to be parallel to within 210-5 radians by using an optical alignment scheme developed by Kumar et al. (1977). The actual tilt between the two plates is measured by recording the times at which four, isolated, voltage-biased pins, that are flush with the surface of the target plate, are shorted to ground. direction of approach. Impact takes place at an angle relative to the

This results in pressure-shear loading at the flyer-target

139

(tribo-pair) interface. By controlling the skew angle, , the impact velocity, a variety of friction states with normal stress varying from 1 GPa to 4.5 GPa and slip speeds from 1 m/s to 250 m/s can be obtained. Other details regarding the design, execution and data analysis of the experiments can be found elsewhere (Prakash, 1995).

Figure 3.1: Schematic of the plate-impact pressure-shear friction experiment.

A new laser interferometer, based on the multi-beam VALYNTM VISAR (Barker, 1999), is used in the present study to measure the combined normal and transverse particle velocities at the rear surface of the target plate. These measurements are used to infer the normal and shear components of the interfacial tractions, the slip-velocity and temperature at the tribo-pair interface. interferometer is shown in Figure 3.2. The schematic of the light path of the laser Three VALYNTM VISAR fiber-optic probes are

employed to obtain the normal and transverse components of the free surface of the

140

target plate by collecting the 0 and 1 order diffraction beams from a holographic grating deposited on the rear surface of the target plate. In the preset experiments a holographic diffraction grating with a pitch of 1200 lines/mm is utilized.

Following Chhabildas et al. (1979), the longitudinal, and the transverse components of the free-surface particle velocity, i.e. U ( t ) and V ( t ) , can be expressed in terms of the measured particle velocities along the beams, i.e. V ( t ) , and the diffraction angle,

VISAR , as

U (t ) =

V + (t ) + V (t )

(1 + cos VISAR )

(3.1)

V (t ) =

V + (t ) V (t ) . sin VISAR

(3.2)

141

Figure 3.2: Schematic of the combined normal and transverse velocity interferometer.

3.2.2 Tribo-pair Materials and Specimen Preparation

In the present study the tribo-pair comprises of a 7075-T6 Al alloy (low melting point metal), and a high hardness Carpenter Hampden (CH) tool-steel (AISI grade D3) plate. The 7075-T6 Al alloy is used in the as-received condition. The heat treatment schedule for CH steel includes annealing at 1100 oC for 40 minutes, followed by oil quench to room temperature. Finally, the steel is tempered at 200 0C for 1 hour. This heat-treatment procedure results in steel with a hardness of ~ 62 on the Rockwell scale. The physical properties of the two materials are provided in Table 3.1.

142

Flyer Material Aluminum Alloy 7075-T6 Longitudinal Wave Speed CL, mm/s Transverse Wave Speed CS, mm/s Acoustic Impedance CL, GPa/mm/s Shear Impedance CS, GPa/mm/s Mass Density , kg/m3 Youngs Modulus E, GPa Shear Modulus , GPa Tensile Strength , MPa Thermal Conductivity k, W/mK Thermal Diffusivity , m2/s Specific Heat Cp, J/kgK Melt Temperature, C
0

Target Material CH Steel 5.98 3.264 45.50 24.90 7612 207 81.1 2758 42.3 12.210-6 422 1400

6.23 3.100 17.44 8.68 2800 71 26.9 572 130 48.410-6 960 635

Table 3.1: Physical properties of the flyer and target plates used in the pressure-shear plate impact friction experiments in the present study. In all the experiments conducted in the present study the CH tool-steel is used as the target plate while the 7075-T6 Al alloy is used as the flyer. Prior to conducting the experiment both sides of the flyer and the target plates are ground flat by using a surface grinder to within 250 microns across the diameter. They are then lapped to within 2-4 Newtons rings (1-2 microns) over the diameter. Lapping is carried out on a Lapmaster machine using a slurry of 15 m diamond powder in mineral oil. To obtain the desired surface roughness, both flyer and the target plates are polished on Texmeth cloth using 1 m diamond paste. This procedure allows the root mean square surface roughness (Rq) of CH tool-steel plates to be maintained between 0.044 and 0.088 m and that of the

143

7075-T6 Al plates to be between 0.035 and 0.09 m without loosing flatness. Before conducting the experiments the impacting surfaces of the flyer and target plates are cleaned by employing a three-step process. In the first step the surfaces are cleaned by Tri-chloro-ethylene. Next, the surfaces are cleaned with acetone. These two steps remove much of the grease and oil from the surface. However, they leave behind a thin residual film, which is removed by cleaning the surface with methanol. In order to use the combined normal and transverse displacement interferometer to measure the normal and transverse particle velocity history at the rear surface of the target plate, a holographic diffraction grating with a pitch of 1200 lines/mm is utilized on the rear surface of the target plate.

3.2.3 Design of the Experiments: Wave Propagation in the Flyer and Target Plates

The wave propagation in the flyer and target plates is illustrated by using a time-distance diagram, shown in Figure 3.3. The abscissa represents the spatial position of the wave front while the ordinate represents the temporal location of the wave front. When the flyer plate impacts the target plate (point A on the time-distance diagram), both longitudinal and shear waves are generated in the flyer and the target plates. The solid lines represent the longitudinal wave fronts while the dashed lines represent the shear wave fronts. The slopes of the solid and the dashed lines represent the inverse of the longitudinal wave speed CL, and shear wave speed, CS. Since, in the experiments

conducted in the present study the CH tool-steel plate remains essentially elastic the characteristics representing the wave-fronts in the target plate are shown as straight lines. However, depending on the impact velocity, the state of stress in the 7075-T6 Al plate
144

could be elastic-plastic. In the case when the 7075-T6 Al plate is in the plastic range, a curved line is used to represent its wave characteristics.

In the present experiments tool-steel disks with a diameter of 76 mm and a thickness of ~ 9.5 mm were used as the target plates. These dimensions were chosen to maximize the available window time for interfacial slip. The dimensions of the 7075-T6 Al alloy flyer plates were 76.2 mm in diameter and 20 mm in thickness. The 20 mm thickness was chosen so that the release wave from the back of the flyer plate does not arrive at the slip interface during the duration the window time of the experiment.

Figure 3.3: Time-distance diagram detailing the wave propagation in the tribo-pair plates.

145

3.2.4 Waveanalysis of Pressure-shear Friction Experiments: Calculation of Interfacial Tractions, Slip Velocity, Slip Distance and Temperature Impact of Elastic Flyer with Elastic Target Plate

The normal and transverse components of the velocity of the flyer plate, i.e. u0 and v0 , respectively, can be expressed in terms of the impact velocity V, and the skew angle

, as

u0 = V cos and v0 = V sin .

(3.3)

Using the method of characteristics for 1-D hyperbolic wave equations (Achenbach, 1973), the normal and transverse components of interfacial traction and slip velocity can be related to the measured normal and transverse components of particle velocity of the free surface of the target plate, u fs (t ) and v fs (t ) , respectively, as

(t ) =

( C L )t u fs (t ) , 2

(3.4)

(t ) =

( Cs )t v fs (t ) . 2

(3.5)

Based on the elementary definition of coefficient of kinetic friction between two dry contact surfaces under no-slip condition, K can be written as

146

K (t ) =

(t ) (t )

(3.6)

Figure 3.4 shows the loci of all-possible shear stress and transverse particle velocity states that can be attained for the elastic flyer-target condition. For a no-slip condition, the state of the interface is represented by the letter A. If the interface slips, the Also,

transmitted shear stress at the flyer-target interfaces is reduced from A to *.

because of interfacial slip there is a discontinuity in particle velocity across the frictional interface. Let the particle velocity on the flyer and target sides of the frictional interface be denoted by VB and VC, respectively. The slip velocity V slip can be expressed in terms of the shear impedance of the flyer and target plates and the measured transverse particle velocity at the free surface of the target plate as

( Cs ) t + ( C s ) f V slip (t ) = VB VC = V sin 2( Cs ) f

v fs (t )

(3.7)

147

Figure 3.4: Shear stress versus transverse particle-velocity diagram for an elastic flyer plate impacting an elastic target plate.
Bulk Temperature at the Tribo-pair Interface

Flash temperature models used to explain asperity level flash heating invariably use the empirical and numerical formulation provided by Archard (1958) --- the analysis assumes circular contacts slide on a flat surface with the contact area determined by the applied load and yield pressure of the material. Recently, an elementary model (Rice, 2006) was proposed to explain flash heating and subsequent weakening at micro-asperity contacts. In his model, Rice considers contacts of uniform size L, hence a lifetime of ~

L/V, and assumes that the asperity shear strength remains at the lower-temperature value

C , until the asperity contact temperature reaches a critical value Tw, above which shear
strength of the material at the interface is taken to be zero (Figure 3.5). Using this

148

model, for most metal tribo-pairs, a critical slip velocity of ~ 1-3 m/s is estimated after which the effects of flash heating become important.

Figure 3.5: Flash heating at asperity contacts On the other hand, the bulk temperature rise, in the vicinity of the tribo-pair interface, can be estimated by solving the one-dimensional transient heat-conduction equation (Carslaw and Jaeger, 1986; Irfan and Prakash, 1994; Irfan and Prakash, 2000) to yield

T (x , t ) =

1 k

q ( )

(t )

exp

x 2 d 4 (t )

(3.8)

In Eq. (3.8), T is the temperature rise; t is time; k is the thermal conductivity; is the thermal diffusivity; and q (t ) is the heat source and x is the distance perpendicular to the interface.

149

In order to calculate the temperature distribution in the tribo-pair materials, an estimate for the heat source q (t ) is required. Using the experimentally measured friction stress , and slip velocity Vslip, the frictional power can be written as

q 7075-T6Al (t ) = (t )V slip (t )

(3.9)

q CH tool-steel (t ) = (1 ) (t )V slip (t )

(3.10)

In these equations, is the factor that governs the partitioning of heat in the tribo-pair materials. The factor is estimated by equating the temperatures at the in the contacting materials at the tribo-pair interface to yield

k1 2 k1 2 + k2 1

(3.11)

In the Eq. (3.11), k1 and 1 are the thermal conductivity and the thermal diffusivity of the 7075-T6 Al alloy, respectively; and k2 and 2 are the thermal conductivity and the thermal diffusivity of the CH steel, respectively. For the tribo-pair employed in the present investigation the factor is approximately 0.607. Thus, 60.7% of the heat

generated at the tribo-pair interface due to the frictional sliding goes into the 7075-T6 Al alloy while 39.3% is partitioned into the CH tool-steel plate.

150

3.3 Finite-Element Based Methodology to Simulate the Impact of an Elastic-plastic Flyer with an Elastic Target Plate

In the present experiments, the 7075-T6 Al alloy flyer plate is expected to undergo elastic-plastic deformation. As long as the impact speeds are restricted such that the CH tool steel plate remains elastic during impact, the normal and transverse components of the interfacial tractions and the coefficient of kinetic friction can be obtained by employing Equations (3.3)-(3.7). However, in order to obtain the interfacial slip speed a full elastic-plastic finite-element simulation of the experiment is required. Besides providing an estimate for the interfacial slip speeds, the finite element simulations also provide information on local elastic-plastic interactions at the sliding interface including distribution of equivalent stress, and interfacial temperatures.

To accomplish these objectives, a transient finite-element code is developed that accounts for finite geometry changes, material inertia, material rate sensitivity, heat conduction and contact-impact initial and boundary conditions.

In accordance with the elastic-plastic multiplicative decomposition of the deformation gradient as presented by Lee et al. (1969), when a work-hardening ductile material is stressed to the finite deformation regime, both elastic and plastic deformations comprise the total deformation. When the load is removed, the material is assumed to

return to zero stress state and the elastic strains are recovered. The remaining strain at the zero stress state is the plastic strain.

151

Next, consider a stress-free reference configuration of the body. The decomposition defines three configurations of the body: an undeformed configuration Bo , an intermediate configuration Bt , and a deformed configuration Bt . In what follows we shall use letters with a superimposed
( )

for quantities pertaining to Bt .

The

super-scripts e and p are used to denote the elastic and plastic parts of a quantity.

Guided by the kinematics of elastic-plastic deformation in crystalline metals (Hill, 1966; Teodosiu, 1970; Hill and Rice, 1972; Hill and Rice, 1973; Teodosiu and Sidoroff, ; Asaro and Rice, 1977; Asaro, 1983; Moran et al., 1990), it is assumed that a multiplicative decomposition of the deformation gradient tensor, F, of the form F = Fe Fp exists, in which Fp is the deformation solely due to plastic flow and represents the aggregate effect of dislocation motion and leaves the lattice structure undistorted and un-rotated. F e is the remaining contribution to F associated with lattice distortion and includes any rigid body rotation of the lattice and the material. Also, it is assumed that it is possible to write a set of equations of evolution that enables Fp to be uniquely determined from the history of the deformation.

In the present formulation the right Cauchy Green deformation tensor C , is defined in the reference configuration Bo , i.e. C = FT F . In addition to the right Cauchy Green tensor based on the total deformation gradient F , we adopt an elastic right Cauchy-Green deformation tensor given by C = FeT Fe as the fundamental deformation tensor on the
152

intermediate configuration Bt . In the spatial configuration the velocity gradient tensor

L = FF1 , can be used to define the rate of deformation and spin tensors, i.e. D = sym L

and W = skew L , respectively.

Next, consider a solid with volume Vo in the reference configuration and a deformation process characterized by the mapping x(X, t ) during which a typical material point initially at X is mapped to a location x = X + u after deformation, where u represents the displacement vector field.

For the particular body under consideration a displacement based finite-element formulation can be obtained from the weak form of the momentum balance equation, which is also commonly referred to as the dynamic principle of virtual work, to yield

Vo

P : o dV

Vo

o ( bo a ) dV

St

t dS = 0 .

(3.12)

In Eq. (3.12), P is the First Piola-Kirchhoff stress tensor; is the virtual velocity field; o represents the material density per unit volume in the reference configuration;
bo and a are the body force and the acceleration vectors on the volume Vo ; and t is the

traction on the boundary St .

153

Consider the quantities S and as the stress measures of the intermediate configuration, and E as the elastic strain measure on the intermediate configuration, where S = Fe1 FeT , = CS , and E =
1 (C I) . 2

During the course of frictional

sliding, substantial heat is generated due to frictional slip and elastic-plastic interactions at the sliding interface. The rate of heat supply due to plastic work can be estimated as W p , where W p = : Lp is the plastic power per unit deformed volume and is the Taylor-Quinney coefficient (Kobayashi, 1989). The rate at which heat is generated at the frictional contact can be expressed as fslip = t Vslip , where Vslip is the jump in velocity across the contact.

With body forces and external heat energy source being neglected, the balance of energy can be expressed as integrals over the reference configuration

c dV = : L dV J N F
0 p

q dS

f
St

slip

dS c .

(3.13)

In general, the constitutive relation for the heat flux is formulated on the intermediate configuration

q = k ,

(3.14)

154

where represents the gradient of the temperature field in the intermediate configuration, q represents the heat flux vector, and k the thermal conductivity tensor on the intermediate configuration.

Formulation of the constitutive relation on the intermediate configuration is necessary to insure that the principle of material frame indifference is not violated. Next, rewriting (3.14) in terms of the gradient of the temperature field in the current configuration yields

q = F e q = F e kF eT .

(3.15)

Since the plastic deformation of metal, does not change the lattice structure, the conductivity tensor can be taken to be independent of Fp such that on the intermediate configuration the thermal conductivity tensor can be approximated as k kI . Thus, restricting (3.13) to cases where thermal conductivity can be assumed to be isotropic, yields the energy balance equation in the reference configuration as

c dV = : L dV + Jk (F
p
o p

F e F eT FT

) dS f
St

slip

dS ,

(3.16)

where is the gradient operator in the reference configuration.

155

The stress rate versus strain rate relation on the intermediate configuration can be expressed as

S = L : E L:I ( 0 ) ,

(3.17)

where L is the elastic modulus, and E is the elastic strain measure on the intermediate configuration.

In general, the choice of stress and strain measures in the formulation of constitutive laws is not unique. Several choices are available in the formulation of hyperelasticity constitutive law. The only requirement is that the stress and strains should be conjugate. We adopt the following constitutive law proposed by Weber and Anand (1990). This law is widely used by many investigators like, to name a few, Cuitino and Ortiz (1992); Marusich and Ortiz (1995), and has the following form

S = L ( 1 log Ce ) , 2

(3.18)

where, L is the fourth order isotropic elasticity tensor

L = (k 2 ) I I + 2 , 3

(3.19)

156

and, and are the elastic shear and bulk moduli, respectively; I and are the second-order and fourth-order identity tensors, respectively. Like for the case of the Neo-Hookean material model, and are constants which may be determined from experimental data at infinitesimal strains, and all moderate strain non-linearities are incorporated in the Hencky or the logarithmic strain measure. It has also been shown (Steinberg, 1996) that for isotropic elastic solids the hyperelastic stress-strain relation, (3.19) is an excellent generalization of the classical Hooke's law for infinitesimal isotropic elasticity to moderately large elastic strains involving principal stretches in the range 0.7 to 1.3.

Recall that we have assumed the existence of a set of evolution equations which enables Fp , and therefore the intermediate configuration Bt to be unambiguously determined from the deformation history. Specifically, we require a plastic flow rule and an evolutionary equation for the internal variables. In the present study the plastic flow is given by a flow rule of the type

Lp = Fp Fp1 = R (S, Q) ,

(3.20)

where, is an effective plastic strain rate, R is the plastic flow direction, and Q is some set of internal variables defined over Bt . The evolution of these latter quantities is assumed to be governed by a hardening law of the form
157

Q = H (S, Q) ,

(3.21)

where, H is the hardening modulus.

For a simple phenomenological plasticity model, representative of the early stages of deformation of polycrystalline metals prior to the development of substantial texture, Taylor averaging of polycrystalline plasticity gives W p = 0 , and it suffices to specify the flow rule as

Dp = sym R (S, Q) .

(3.22)

For J2 flow theory model with isotropic-hardening, the internal variables consist of the accumulated effective plastic strain and a scalar parameter that is used as a measure of isotropic hardening. The plastic part of the rate of deformation tensor can be expressed as

Dp =

3 S , 2

(3.23)

where S is the deviatoric part of the stress tensor S , defined on the intermediate configuration, and is the equivalent stress, i.e. = 3 2 S : S .

158

The material is characterized by an isotropically hardening, and thermally softening elastic-viscoplastic von-Mises solid. A combined power law and exponential plastic

strain-rate relation (Clifton, 1990), that gives rise to enhanced strain-rate hardening at ultra-high strain rates is employed. Under high speed frictional sliding, it is anticipated

that the reduction in material strength with increasing temperature will be the key mechanism that plays the dominant role in controlling the interfacial resistance. Under such conditions, a thermal softening Mises flow rule is generally adequate. For the

present simulations the particular form for the effective plastic strain rate, p , that is chosen is based on the shear-stress shear-strain relationship over the strain rate range from
103 s 1 to 107 s 1 .

The effective plastic strain rate, , is taken to be of the form

1 2 1 + 2

(3.24)

where,

a g( , ) m , , 2 = m exp 1 = 0 g( , )

(3.25)

and
159

g( ,T ) = 0 (1 + / 0 )N 1 ( / 0 )k 1 .

(3.26)

Here, = dt is the equivalent plastic strain; m is a reference strain rate; m and a

0

are the rate sensitivity parameters, respectively; 0 is a reference stress; 0 is a reference strain; N is the strain hardening exponent; o is a reference temperature; and and k are the thermal softening parameters. The function g( , ) represents the stress-strain relation at a quasi-static strain-rate of 0 and at temperature . Equation (3.24), provides a smooth transition between the measured response = 1(, , ) at strain rates less than 103 s-1, and the limiting behavior = 2 (, , ) at strain rates greater than, say, 105 s-1.

It must be noted that an explicit yield criterion is not employed in this model and the viscoplastic strain-rate is non-zero at all levels of stress. As shown by Peirce et al. (1984), however, for large values of the rate sensitivity exponent m the plastic strain rate is exceedingly small for as long as is smaller than g( , ) . The solid then exhibits yield like behavior and g( , ) plays the role of a yield stress. We furthermore note that taking
m in Eq. (3.25) gives the rate independent limit, wherein the effective stress is

constrained to remain less or equal to the yield stress g( , ) at all times.

At temperatures above the melt temperatures the material is modeled as a purely Newtonian fluid for which the dependence of the stress on the stretching tensor is linear. In this case the stress tensor simplifies to
160

(1 J + k (T To ))
J

E I + D , (1 )

(3.27)

where, is the stiffness parameter; J is the Jacobian of the deformation is equal to the determinant of the total deformation gradient tensor; k is the thermal conductivity; T is the temperature; E is the elastic modulus; is the Poissons ratio; and is the dynamic shear viscosity of the fluid. The most common naturally-occurring fluids are well

represented by the Newtonian constitutive equation. But there is abundant evidence of fluids behaving in a manner which cannot be described by the theory of Newtonian viscous flow. For such a class of fluids, the dynamic viscosity of fluids is not a constant and may depend upon temperature, pressure and the instantaneous stretch rates. In the present investigation, the computations are restricted to the purely Newtonian viscosity model given by Eq. (3.27).

The material strain-rate hardening response and temperature dependence of flow stress for the flyer and the target plate materials are given in Fig. 3.6. The material parameters for CH tool-steel and 7076-T6 Al alloy used in the model are given in Table 3.2.

161

2.0E+04 1.8E+04 1.5E+04 1.3E+04 1.0E+04 7.5E+03 5.0E+03 2.5E+03 0.0E+00 0 10

CH steel 7075-T6 aluminum Temperature 20oC Equivalent Plastic Strain 0.01

Flow Stress (MPa)

10

10

10

10

10

10

10

10

Equivalent Plastic Strain Rate

(a)
3.0E+03

2.5E+03

CH Steel 7075-T6 aluminum

Equivalent Plastic Strain 0.01 Equivalent Plastic Strain Rate 0.01 s
-1

Flow Stress (MPa)

2.0E+03

1.5E+03

1.0E+03

5.0E+02

0.0E+00

500

1000

1500

Temperature ( C)

(b)

162

8000 7500 7000

Flow Stress (MPa)

6500 6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 0

CH steel 7075-T6 aluminum 6 -1 Equivalent Plastic Strain Rate 10 s

20oC 400oC 800 C

o

20oC

200oC
o

400 C

0.1

0.2

0.3

0.4

0.5

Equivalent Plastic Strain

(c) Figure 3.6: (a) Flow-stress versus the plastic-strain-rate dependence used in the viscoplastic model for the flyer and the target plates; (b) Flow-stress versus the temperature dependence used in the visco-plastic model for the flyer and the target plates; (c) Flow-stress versus the equivalent-plastic-strain dependence used in the visco-plastic model for the flyer and the target plates.

163

FLYER Material

TARGET Material

Aluminum Alloy 7075-T6 CH Tool Steel Young's Modulus E , GPa Mass Density , kg/m3 Tensile Strength y , MPa 71 2800 207 7612

572 960 130 2.2 10-5 293 0.33 100.0 5.2 1.0 10-4 5.0 106 0.01 3.6 0.2 0.1

2758 422 42.3 1.0 10-5 293 0.3 100.0 10.0 1.0 10-4 5.0 108 0.014 4.5 0.12 0.1

Specific Heat Cp , J/kg.K Thermal Conductivuty k , W/mK Coefficient of Thermal Expansion , 1/K Reference Temperature To , K Poisson Ratio Rate Sensitivity Parameter m Rate Sensitivity Parameter a Reference strain rate 0 , s 1 Reference strain rate m , s 1 Reference Strain 0 Thermal Softening Parameter Thermal Softening Parameter Strain Hardening Exponent N

Table 3.2: Physical properties and model parameters for 7075-T6 Al alloy and the CH steel tribo-pair materials used in the finite element constitutive model.

3.3.1 Finite Element Implementation

As discussed by Budiansky (1969), the principle of virtual work (Eq. (3.12)) can be used as the variational principle for a solid continuum undergoing arbitrary large

164

displacements and deformations.

Moreover, the variational equation governing the

thermo-mechanical energy balance can be obtained from the balance of energy, Eq. (3.16), as

c dV = : L dV + Jk ((F
p
0 p

p 1

F p T o ) N dS

Jk (F p1F p T ) (o ) dV .

(3.28)

Once we have established the equations of momentum balance at the element level, we assemble these equations to obtain global equations of momentum balance as

Ma = Fext Fint ,

(3.29)

where M is a lumped mass matrix; Fext s the external force vector; and Fint is the internal force vector arising from the current state of stress. A lumped mass matrix is used instead of the consistent mass matrix, since this has been found preferable for explicit time integration procedures from the point of view of computational efficiency and accuracy (Krieg and Key, 1973). The acceleration can be found by inverting the mass matrix and multiplying it by the residual load vector. Since M is diagonal, its inversion is trivial, and the matrix equation is the set of independent equations for each degree of freedom.

165

In a similar way, the element level equations of energy balance can be assembled to obtain the following global equations

C = K + HB + H .

(3.30)

In Eq. (3.30), C is the lumped heat conductance matrix; and H and HB are the heat source vectors associated with plastic work and friction work, respectively. Equations (3.29) and (3.30) are coupled through the heat source vectors H and HB .

3.3.1.1 Contact Algorithm

In order to account for dynamic friction during contact-impact situations, we make use the predictor-corrector contact algorithm developed by Taylor and Flanagan (1987) for the PRONTO2D explicit dynamics. In this approach, the bodies in contact can be

deformable or rigid. The contacting surfaces are designated as master and slave surfaces and the contact algorithm continuously modifies the nodal accelerations along the contact surface such that the kinematic constraints are satisfied.

The method starts with the calculation of the predictor nodal positions, velocities and
pred pred pred accelerations xn +1 , vn +1 , and an +1 , respectively, assuming that no contact has occurred.

A predictor configuration where penetration has occurred is shown in Figure 7(a). The

166

subscript m and s refer to nodal quantities on the master and the slave surfaces, respectively. The unit tangent vector to the master surface is orthogonal to the outward normal and, therefore, is expressed as s = n . x
n y

A partitioned kinematic approach is used to enforce compliance between two contact surfaces. This means that each surface acts as a master for a fraction of each time step and as a slave for the remainder. This fraction is refereed to as , and the roles of the surfaces are reversed for the remaining fraction (1 ) .

We define the normal contact forces as a fraction of the forces that would be imposed by the slave nodes if the master surface was rigid. The normal contact force for a slave node is expressed as

fp,j =

M s,j

(t )

(n (x

pred s,j

pred xm,k ) n ,

(3.31)

pred pred where, M s,j is the mass of the jth slave node, and xs,j and xm,k are the predicted

coordinates of the jth slave and kth master node respectively.

167

Figure 3.7: (a) Predictor configuration of contact pair. (b) Corrected configuration after normal acceleration corrections are introduced which eliminate the unwanted penetration.

Next we want to obtain the response of the master surface to these contact forces, such that the response of the jth slave node is constrained by the master nodes defining the contact segment, i.e. master nodes k and k+1. In order to achieve this we distribute the mass of the slave nodes and the normal contact force to the two master nodes defining the contact segment. For master node k

M m ,k +

(1- ) M (a
j
s,j

corr
m,k

n n =

(1- ) f
j j

p,j

(3.32)

and for master node k+1

168

 M m ,k +1 +

jM s,j acorr 1 n n = m ,k +

f
j

j p,j

(3.33)

where, the summation is over all slave nodes in contact, and the interpolation variable j is given by (Figure 7b)

j =

pred s (x pred xm,k ) p,j pred pred s (xm,k +1 xm,k )

(3.34)

Equations (3.32) to (3.34) represent a set of uncoupled equations, one for each master node.

After assembling and solving equations (3.32) and (3.34) to obtain the corrections applied to the master nodes, the corrections to the slave accelerations are interpolated using

(acorr n) = (1 j )(acorr n) + j (acorr+1 n) , m ,k s,j m,k

(3.35)

Note that this slave response restricts the motion induced by the contact force given in Eq. (3.31).

169

In the absence of friction, the corrected nodal accelerations of the master and slave nodes, respectively, are given by

pred corr am,k =am,k + (am,k n ) n ,

(3.36)

and

pred as, j =as . j + (acorr n ) n fp,j / M s,j . s, j

(3.37)

3.3.1.2 Friction

Friction resists the relative tangential motion of the contacting slave nodes. The tangential component of the relative predicted velocity of the jth slave node with respect to the master surface is given in terms of the unit tangent vector of the contact segment on the master surface and its (jth slave node) position relative the contact segment

pred pred pred vspred s = s vs, j (1 j ) vm,k j vm,k +1 s . ,j

( (

))

(3.38)

As with the penetration force (3.31), we define the tangential contact force as a fraction of the force which must be applied to the slave node to cancel its relative tangential velocity. This force is given by

170

fs,j =

M s,j t

vspred s , ,j

(3.39)

where, the minus sign reflects that this force would be applied in the direction of s , but opposite to the direction of motion.

The magnitude of the tangential force exerted by the master surface on a slave node cannot exceed the maximum friction force. This constraint is expressed as

fric fs,j =

fs,j fs,j

norm min fs,j , fs,j ,

(3.40)

norm is the magnitude of the normal contact where is the friction coefficient and fs,j

force at the jth slave node as given below

norm corr fs,j = M s,j (as, j n ) n fp,j / M s,j .

(3.41)

Applying this force to the slave node and balancing forces to the master nodes, then dividing by appropriate nodal mass yields the following expressions for the corrections to the tangential accelerations to these nodes

171

s,j (acorr s) s = M s, j

f fric
s,j

(3.42)

fric (acorr s) s = (1 j ) fs,j /Mm,k , m ,k

(3.43)

fric (acorr+1 s) s = j fs,j /Mm,k+1 . m ,k

(3.44)

Finally, by adding the above tangential accelerations to equation (3.36) and (3.37), the corrected total acceleration of the contact nodes is expressed in general form for master and slave nodes by

pred corr corr am,k =am,k + (am ,k n ) n + (am ,k s) s ,

(3.45)

pred as, j =as . j + (acorr n ) n s, j

fp,j M s,j

+ (acorr s) s . s, j

(3.46)

An explicit time integration scheme based on the Newmark -method, with = 0, and
= 0.5 is used to integrate the equations of motion to obtain the nodal velocities and

displacements and the energy equation to obtain the nodal temperatures (Belytschko, 1983; Hughes, 1987). Explicit integration is particularly attractive in impact problems, since the resolution of the various waves in the solution necessitates the use of small time steps

172

well under the stability limit (Hughes, 1983). Table 3.3 shows the main steps of the time integration algorithm employed in the computational code.

For the present study the constitutive equations always set the critical time step for stability. It therefore suffices to integrate the energy equation by forward Euler algorithm, with the result

n+1 = n + t n ,

(3.47)

where,

n = C1 [Kn n +HBn n + Hn ] .

(3.48)

A staggered procedure (Park and Felippa, 1983) is adopted for the purpose of coupling the mechanical and thermal equations. Mechanical and thermal computations are staggered assuming constant temperature during the mechanical step and constant heat generation during the thermal step. Following Lemonds and Needleman (1986), a mechanical step is taken first based on the current temperature distribution. The

corresponding heat source vectors H and HB are computed. The computed heat sources are used in the thermal step where temperature is calculated by the forward Euler scheme (3.47) and (3.48). The resulting temperature is incorporated into the thermal-softening
173

model described in Eq. (3.26)and then used to calculate displacement in the mechanical steps. The displacement and temperature are updated, and the process is repeated.

174

Explicit Time Integration Algorithm

1. Prescribe initial conditions, n = 0

u0 = u0 v0 = v0 a 0 = M1 (f0ext f0int ) .

2. Compute displacement predictor for contact un +1 .

un +1 = un + tvn + 1 2 t 2 a n

3. Modify accelerations due to surface contact based on un +1 . an = a n + a n

4. Update displacements:
un +1 = un + tvn + 1 2 t 2 a n

5. Update 2nd Piola-Kirchoff stress tensor S at each element using material constitutive law.

6. Compute internal force vector.

7. Solve for accelerations: Table 3.3: Explicit time integration algorithm.

175

3.3.1.3 Constitutive Model and Stress-update Algorithm

As discussed in earlier, the elastic response is formulated in terms of a hyper-elastic potential on the intermediate configuration. The stress update algorithm, as detailed below, is especially tailored to take advantage of this fact. The algorithm is explicit in both the plastic flow direction and the scalar increment of plastic flow .

1. Geometric update

n+1 = n + u

(3.49)

Fn+1 = n+1 / X .

(3.50)

In the above equations, the deformation mapping is denoted by and u represents the incremental displacements.

2. Given Fn+1, Fn, Fnp , Sn , Qn and t , compute

= t Sn , Qn .

(3.51)

e 3. Update Fnp+1, Fn+1, Cn+1, Sn+1, Qn+1

176

(Fnp+1 Fnp )
t

Fnp1 = p Rn (assuming W p = 0 )

(3.52)

p p Fn+1 = (I + R n ) Fn

(3.53)

e p Fn+1 = Fn+1Fn+1

-1

(3.54)

e e Cn+1 = Fn+1Fn+1

(3.55)

Sn+1 = S (Cn+1 )

(3.56)

Qn+1 = Qn + H n .

(3.57)

A limitation of this algorithm is that the plastic flow direction remains fixed during an increment and the incremental of effective plastic strain is calculated at the beginning of the increment. For an implicit finite element method, if the time increment t is too large, at points within the body experiencing highly non-proportional loading, this algorithm may not be defined for accurate plastic strain increments. Although accuracy considerations normally rule out the time step for implicit finite element method, for explicit finite element method, the stability considerations rule out the time step which is much smaller than the time step needed for accuracy considerations. So the limitation of

177

this algorithm will not be an issue of explicit finite element program. This is also a reason for explicit finite element methods are used extensively to solve highly nonlinear problems.

3.3.2 Finite Element Discretization

The finite element discretization is based on triangular elements arranged in crossed

triangle quadrilaterals, such that the displacements and temperature vary linearly over the

triangular elements. Nagtegaal et al. (1981) have shown that an element of this type can accommodate isochoric deformations. This is of significance since plastic strain is

volume preserving, so that total deformation at large strain is nearly isochoric. Figure 3.8 shows the finite element mesh used in conducting the numerical simulations of the experiments. Plane strain conditions are assumed to prevail and the flyer and the target plate materials are modeled by using a single column of finite elements. In the vicinity of the frictional interface the elements have dimensions of 1m 1.5m . Such small

element size is necessary to resolve the sharp thermal and mechanical gradients in the vicinity of the sliding interface. The entire finite element mesh consists of 5302 quadrilateral elements with 31820 degrees of freedom.

With the origin of the coordinate system placed at the frictional, the boundary conditions can be written as

178

v1 (x1 = 0, x 2 , t ) = v1 (x1 = h, x 2 , t )

(3.58)

v2 (x1 = 0, x 2 , t ) = v2 (x1 = h, x 2 , t )

(3.59)

f1 (x1 = 0, x 2 , t ) = f1 (x1 = h, x 2 , t )

(3.60)

f2 (x1 = 0, x 2 , t ) = f2 (x1 = h, x 2 , t )

(3.61)

T (x1 = 0, x 2 , t ) = T (x1 = h, x 2 , t )

(3.62)

In equations (3.58)-(3.62) H steel x 2 H 7075T 6Al .

The initial conditions are

v1 (x 1, 0 x 2 H 7075T 6Al , t = 0) = V cos

(3.63)

v2 (x 1, 0 x 2 H 7075T 6Al , t = 0) = V sin

(3.64)

v1 (x1 , H steel x 2 0, t = 0) = 0

(3.65)

v2 (x1 , H steel x 2 0, t = 0) = 0

(3.66)
179

T (x1 , x 2 , t = 0) = To

(3.67)

Besides these boundary and initial conditions, contact boundary conditions are applied on all flyer and target nodes occupying (x1 = 0, 0 x 2 h ) , the contact surfaces. In Equation (3.67), To is the room temperature.

X2 X2

H aluminum
Flyer

Flyer
X1

H steel
(h,0)

Target

(0,0)

X1 FEM Mesh

Impact Surface Target

Element width 2 m

Figure 3.8: The finite element mesh used to simulate the plate impact pressure-shear friction experiments.

180

3.3.3 Remeshing Algorithm

One difficulty that arises in applying Lagrangian finite element based formulations to problems involving large thermo-mechanical plastic flow at high strain rates, such as those expected to occur during the formation of molten metal films at frictional interfaces under high speed impact-contact conditions is that the mesh may become severely distorted. Element distortion can severely reduce the stable computational time-steps used in the explicit time integration scheme and also reduce the accuracy of computations considerably.

A commonly employed procedure to overcome this problem is to re-mesh the computational domain with an undistorted mesh and continue with the calculation. The prevailing adaptive meshing methods can be roughly classified as r, h, p and h-p types. The r-method seeks to maintain the number of finite elements and element nodes fixed while adapting the nodal positions. The h-method keeps the order of the elements

unchanged while seeking to improve the solution by adaptive mesh refinement and coarsening. In the p-approach, the number of elements is held fixed and their order A

increased or decreased adaptively in accordance with accuracy requirements. combination or h and p paradigms is used in h-p methods.

In the present computations a r-adaptation strategy is employed along with a re-mapping algorithm which is customized to take advantage of the periodicity of the
181

problem. This computational strategy allows calculations to be extended to relatively large deformation regimes in the vicinity of the frictional interface, especially during the transition from near-melt to the fully-melt temperature regime. Central to this

computational strategy is the ability to replace the skewed mesh by an un-skewed new mesh satisfying the same periodic boundary conditions (shown in Fig 3.9). In this way the mapping of the constitutive state variables from the old to the new mesh is element to element. This avoids the numerical dispersion due to the interpolation of state variables, in particular the large spurious plastic strain rates for rate-sensitive constitutive laws (Camacho and Ortiz, 1997).

Figure 3.9: Re-mesh and re-map algorithm for mesh with periodic boundary condition.

182

The setting of configurations used in the computations is shown in Figure 3.10. Regular geometric updates of the old current configuration are used to remove the pathology associated with the distortion of the computational mesh by using the re-mapping algorithm. In this manner, the old current configuration with the severely distorted mesh is periodically re-mapped to a new distortion-free reference configuration. The displacements in this new reference configuration are assigned to be zero every time the skewed mesh is replaced by a distortion free mesh.

Next, defining

x = X+u,

(3.68)

u = 0,

(3.69)

x = x+u,

(3.70)

where, X and u are the reference particle positions and displacements in the old
configuration and x , and u are the reference particle positions and displacements in the

new reference configurations. After the update, the new current particle position x is

183

 defined by the new reference vector x and displacement u . Recall that the deformation

gradient relates the current configuration to the reference configuration by

F=

(X + u ) x u = =I+ . X X X

(3.71)

When the reference configuration is updated, the deformation gradient can be rewritten as

F=

x u u u = I + + x I + X . X X

(3.72)

If we save the old deformation gradient Fold at each update as

Fold = I +

u , X

(3.73)

u F = Fold I + . x

(3.74)

In (3.74)

u can be discretized as x

NEN u i = N a ,k uia , xk a =1

(3.75)

184

 where NEN is the number of nodes per element and N a ,k is the derivative of the shape

function with respect to the new reference configuration.

Figure 3.10: Schematic of the update of the current configuration.

3.4 Experimental and Computational Results and Discussion

A series of plate-impact pressure-shear friction experiments was conducted to investigate the phenomenon of high speed slip between a 7075-T6 Al alloy and CH tool-steel tribo-pair. Table 3.4 summarizes the experiments conducted in the present study. In all experiments a skew angle of 35o was utilized to promote high slip speeds at the tribo-pair interface. Moreover, the surface roughness of the flyer and target plates

(tribo-pair) was kept relatively smooth to promote high slip speeds (Ra ~0.07 to 0.12 m

185

for 7075-T6 Al alloy plates and 0.01 to 0.02 m for the CH tool-steel plates). The stress state at the tribo-pair interface was changed by varying the projectile velocity from 115 m/s to 462 m/s. In the experiments conducted in the present study the sliding surface of the 7075-T6 Al alloy (flyer plate) was driven into the fully melt regime. The results of these experiments were compared with those obtained for the same tribo-pair, but at lower impact velocities, in a previous study by Liou et al. (2004) and Okada et al. (2001). A summary of these later experiments is presented in Table 3.5.

Skew SHOT # Angle

, deg

Impact Velocity V, m/s 273 340 413 462

Normal Stress
, GPa

Roughness of 7075-T6 Al alloy (Flyer)

m

Roughness of Carpenter Hampden tool-steel (Target)

m

Fuping05008 Fuping05001 Fuping05004 Fuping05007

35 35 35 35

2.72 3.44 4.08 4.39

0.035 0.05 0.09 0.048

0.044 0.046 0.063 0.088

Table 3.4: Summary of the plate-impact pressure shear friction experiments conducted in the present study.

186

Skew SHOT # Angle

, deg

Impact Velocity V, m/s 115 137 171 193 217

Normal Stress
, GPa

Roughness of 7075-T6 Al alloy (Flyer)

m

Roughness of Carpenter Hampden tool-steel (Target)

m

MO9907 MO9903 MO9904 MO9905 MO9906

35 35 35 35 35

1.214 1.436 1.762 1.987 2.242

0.07 0.12 0.11 0.09 0.07

0.02 0.02 0.01 0.02 0.02

Table 3.5: Summary of the plate-impact pressure shear friction experiments conducted in a previous study by Liou et al. (2004) and Okada et al. (2001). The experimental results for Shot Fuping05008 are shown in Figures 3.11 to 3.14. The experiment was conducted at an impact velocity of 317 m/s, which is the lowest of the impact velocities used in the four experiments conducted in the present study. Figure 3.11 shows the normal and transverse components of the particle velocity history at the free surface of the CH steel target plate. The abscissa represents the time after impact. At about 1600ns, in response to the arrival of the longitudinal wave at the free (rear) surface of the target plate the normal component of the particle velocity jumps to ~ 120.0 m/s.. The second step-rise in normal component of the particle velocity occurs at approximately 4760 ns, which is coincident with the arrival of the reflected (release) wave at the rear surface of the steel target plate from the tribo-pair interface. The arrival of shear wave at the free surface of the target plate occurs at approximately 2920 ns. With the arrival of the shear wave at the free surface of the target plate the

transverse component of the particle velocity jumps to 17 m/s and then decreases

187

monotonically to ~ 7.8 m/s in about 2 s.

Figure 3.12 shows the history of the normal stress, friction (shear) stress and interfacial slip velocity as a function of time for Shot Fuping05008. The normal and transverse components of the interfacial stress were obtained from the measured normal and transverse components of the free surface particle velocity profiles in the tool steel target plate (as shown in Fig. 3.11), using 1-D elastic wave theory. In order to include the effects of plasticity in the flyer and the target plates in the interpretation of the measured particle velocity profiles and also in the understanding the transition of the aluminum alloy from solid to the molten state, finite-element simulations of experiment Shot Fuping05008 were conducted with appropriate material models for the Al alloy flyer plate and the CH tool steel target plate, in both the solid and the fully melt regime. Based on the measured shear stress a slip-speed of approximately 130 m/s is predicted by assuming both target and the flyer plates remain elastic for the duration of the experiment. However, due to the extreme heat generated at the sliding interface, the interfacial temperature approaches the melt-point of the aluminum alloy flyer plate. When this condition is reached the dynamic shearing resistance of the molten aluminum in the vicinity of the frictional interface falls below the frictional resistance required to sustain slip and leads to seizure of the slip interface. This occurs at approximately 600 ns, and is shown in Figure 3.12 as a precipitous fall in the interfacial slip velocity from approximately 127 m/s to zero.

Figure 3.13 shows the results of finite element simulations detailing the transition of

188

the aluminum alloy from the solid to the molten phase.

The effective stress is For

calculated by assuming a perfectly viscous Newtonian model for molten aluminum.

a dynamic viscosity of 10 Pa-s the computed shearing resistance in the early part of the transition from solid to molten aluminum is an excess of the measured shearing resistance. However, the computed and the measured particle velocity profiles are in much better agreement when a molten Al layer is formed at the slip interface. It is interesting to note the relatively high shearing resistance (approximately 100 MPa) of the molten Al layer under the given interfacial conditions.

Figure 3.14 shows the history of the evolution of interfacial temperature at various depths in the 7075-T6 Al flyer plate for Shot Fuping05008. The dashed curves show the temperature profile in the aluminum alloy obtained by using Equations (3.8) to (3.11). In this latter approach, in order to estimate the frictional power at the slip interface, the friction stress and slip velocity histories are estimated using the measured normal and transverse components of the particle velocity and the elastic 1-D hyperbolic wave theory. The solid lines correspond to the temperature profiles obtained by employing elastic-plastic finite element simulations of the experiment. During the early part of slip, the temperature profiles based on both the elastic and elastic-plastic (FEM) analysis are in good agreement. This fact is consistent with the limited plasticity during the early part of the slip process. However, during the later part of the slip process, because of the presence of localized plastic flow in the vicinity of the sliding interfaces, the temperature profiles based on the FE simulations deviate markedly from those solely based on the elastic analysis. As shown in Fig. 3.14, the interfacial temperature at x=0

189

reaches the melting point of aluminum first. As the shearing of the molten Al proceeds, the temperatures at the interior points within the flyer plate also asymptotically approach the melt temperature of the Al alloy. The oscillations in the computed temperature profiles are understood to be numerical artifacts, and occur every time a material finite element transforms from the solid state to the fully melt state.

80

Free Surface Transverse Velocity (m/s)

70 60 50 40 30 20 10 0

250

200

Transverse Velocity Normal Velocity

150

100

50

Time after Impact (s)

Figure 3.11: Normal and transverse particle velocity history at the rear surface of the target plate for Shot Fuping05008.

190

Free Surface Normal Velocity (m/s)

Shot Fuping05008 Flyer: Al 7075-T6 (R q=0.035m) Target: CH Steel (R q=0.044m) Impact Velocity: 273 m/s Skew Angle: 35

300

0.8 0.7

200

0.6 0.5 0.4 0.3 0.2 0.1 0

Experimental Interfacial Stress Computed Slip Velocity

Normal Pressure =2.72 GPa

150

100

50

0 0 0.5 1 1.5 2 2.5 3

Time after Impact (s)

Figure 3.12: History of the applied normal stress, friction stress and the slip velocity for Shot Fuping05008.

200

1.5

Stress (GPa)

Interfacial Stress Effective Stress (=10 Pa.s) Interfacial Slip Velocity

150

100

0.5 50

Slip
0 0

Stick
1 2 3 0

Time after Impact (s)

Figure 3.13: Predictions based on finite element simulations for the evolution of effective stress for Shot Fuping05008.

191

Interfacial Slip Velocity (m/s)

Shot Fuping05008 Flyer: Al 7075-T6 (R q=0.035m) Target: CH Steel (R q=0.044m) Impact Velocity: 273 m/s Skew Angle: 35

250

Interfacial Slip Velocity (m/s)

Interfacial Stress (GPa)

Shot Fuping05008 Flyer: Al 7075-T6 (R q=0.035m) Target: CH Steel (R q=0.044m) Impact Velocity: 273 m/s Skew Angle: 35

250

Temperature Rise in 7075-T6 Al Alloy (C)

1600 1400 1200 1000 800 600 400 200

Shot Fuping05008 Flyer: Al 7075-T6 (R q=0.035m) Target: CH Steel (R q=0.044m) Impact Velocity: 273 m/s Skew Angle: 35 Solid line: Numerical Dashed line: Theoretical X=4m X=0m

X=8m
0 0 0.5 1 1.5

X=12m
2 2.5 3

Time after Impact (s)

Figure 3.14: Temperature profile in the aluminum alloy flyer plate in the vicinity of the sliding interface for Shot Fuping05008. The history of the applied normal stress, the friction stress and the slip velocity for Shots Fuping05001, Fuping05004 and Fuping05007 are shown in Figures 3.15. The three experiments were conducted at impact velocities higher than those employed for Shot Fuping05008 -- Shot Fuping05001 was conducted at an impact velocity of 340 m/s, while Shots Fuping05004 and Fuping05007 were conducted at 413 m/s and 462 m/s, respectively. In all the three cases a transition from metal-on-metal slip to seizure occurs at the slip interface. As depicted in Figure 3.16, the computed shearing

resistance of the molten Al film is in a good agreement with the experimental results when assuming a perfectly viscous Newtonian model for the molten Al with a dynamic viscosity of 10 Pa-s, 2 Pa-s and 1 Pa-s for Shots Fuping05001, Fuping05004 and Fuping05007, respectively. These values of dynamic viscosity correspond to a dynamic shearing resistance of 116 MPa, 53 MPa and 39 MPa for Shots Fuping05001,

192

Fuping05004 and Fuping05007, respectively.

Figure 3.17 illustrates the evolution of the Lagrangian temperatures profiles at various depths in the Al alloy flyer plate. As observed in the case of the experiment Shot Fuping05008, the slip interface x=0 reach the melting point of aluminum first, and as the shearing of molten Al proceeds, the temperatures at the various depths within the flyer plate also approach the melt temperatures.

193

0.8 0.7

0.6 0.5 0.4

200

Experimental Interfacial Stress Computed Slip Velocity

Normal Pressure =3.44 GPa

150

0.3 0.2

100

50 0.1 0 0 0 1 2 3 4

Time after Impact (s)

(a)
0.8 0.7

0.6 0.5 0.4 0.3

250 200 150 100

Experimental Interfacial Stress Computed Slip Velocity

Normal Pressure =4.08 GPa

0.2 0.1 0 50 0 0 0.5 1 1.5 2 2.5 3

Time after Impact (s)

(b)
0.8 0.7

0.6 0.5 0.4 0.3 0.2 0.1 0

300 250 200 150 100 50 0

Experimental Interfacial Stress Computed Slip Velocity

Normal Pressure =4.39 GPa

0.5

1.5

2.5

Time after Impact (s)

(c)

Figure 3.15: History of the applied normal stress, friction stress and the slip velocity. (a) Shot Fuping05001; (b) Shot Fuping05004; (c) Shot Fuping05007.

194

Interfacial Slip Velocity (m/s)

Interfacial Stress (GPa)

Shot Fuping05007 Flyer: Al 7075-T6 (R q=0.048m) Target: CH Steel (R q=0.088m) Impact Velocity: 462 m/s Skew Angle: 35

400 350

Interfacial Slip Velocity (m/s)

Interfacial Stress (GPa)

Shot Fuping05004 Flyer: Al 7075-T6 (R q=0.09m) Target: CH Steel (R q=0.063m) Impact Velocity: 413 m/s Skew Angle: 35

350 300

Interfacial Slip Velocity (m/s)

Interfacial Stress (GPa)

Shot Fuping05001 Flyer: Al 7075-T6 (R q=0.05m) Target: CH Steel (R q=0.046m) Impact Velocity: 340 m/s Skew Angle: 35

300

250

200

1.5

Interfacial Stress Effective Stress (=10 Pa.s) Interfacial Slip Velocity

150

100

0.5 50

Slip
0 0

Stick
1 2 3 4 0

Time after Impact (s)

(a)
2

1.5

250

Interfacial Stress Effective Stress (=2 Pa.s) Interfacial Slip Velocity

200

150

0.5

100

50

Slip
0 0 1

Stick
2 3 0

Time after Impact (s)

(b)
2

1.5

300 250 200 150

Interfacial Stress Effective Stress (=1 Pa.s) Interfacial Slip Velocity

0.5

100 50 0

Slip
0 1

Stick
2 3

Time after Impact (s)

(c)

Figure 3.16: Predictions based on finite element simulations for the evolution of effective stress. (a) Shot Fuping05001; (b) Shot Fuping05004; (c) Shot Fuping05007.

195

Interfacial Slip Velocity (m/s)

Shot Fuping05007 Flyer: Al 7075-T6 (R q=0.048m) Target: CH Steel (R q=0.088m) Impact Velocity: 462 m/s Skew Angle: 35

400 350

Stress (GPa)

Interfacial Slip Velocity (m/s)

Shot Fuping05004 Flyer: Al 7075-T6 (R q=0.09m) Target: CH Steel (R q=0.063m) Impact Velocity: 413 m/s Skew Angle: 35

350

300

Stress (GPa)

Interfacial Slip Velocity (m/s)

Shot Fuping05001 Flyer: Al 7075-T6 (R q=0.05m) Target: CH Steel (R q=0.046m) Impact Velocity: 340 m/s Skew Angle: 35

250

Stress (GPa)

Temperature Rise in 7075-T6 Al Alloy (C)

2000 1800 1600 1400 1200 1000 800 600 400 200 0 0

Shot Fuping05001 Flyer: Al 7075-T6 (R q=0.05m) Target: CH Steel (R q=0.046m) Impact Velocity: 340 m/s Skew Angle: 35 X=4m X=0m Solid line: Numerical Dashed line: Theoretical

X=8m X=12m
1 2 3 4

Time after Impact (s)

(a)
Temperature Rise in 7075-T6 Al Alloy (C)
1600 1400 1200 1000

Shot Fuping05004 Flyer: Al 7075-T6 (R q=0.09m) Target: CH Steel (R q=0.063m) Impact Velocity: 413 m/s Skew Angle: 35 X=4m X=0m Solid line: Numerical Dashed line: Theoretical

800 600 400 200 0

X=8m
0 1

X=12m
2 3

Time after Impact (s)

(b)
Temperature Rise in 7075-T6 Al Alloy (C)
1600 1400 1200 1000

Shot Fuping05007 Flyer: Al 7075-T6 (R q=0.048m) Target: CH Steel (R q=0.088m) Impact Velocity: 462 m/s Skew Angle: 35 X=4m X=0m Solid line: Numerical Dashed line: Theoretical

800 600 400 200 0

X=8m
0 0.5 1 1.5 2

X=12m
2.5 3

Time after Impact (s)

(c)

Figure 3.17: Temperature profile in the aluminum alloy flyer plate in the vicinity of the sliding interface. (a) Shot Fuping05001; (b) Shot Fuping05004; (c) Shot Fuping05007.

196

Figure 3.18 shows the growth of the molten aluminum layer as a function of impact velocity. Simulations are presented for experiments Shot MO9904, MO9905, MO9906, Fuping05008 and Fuping05001 each with a dynamic viscosity of 10 Pa-s, Shot Fuping05004 with dynamic viscosity of 2 Pa-s, and for Shot Fuping05007 with a dynamic viscosity of 1 Pa-s. These values of dynamic viscosity were chosen since they provide the best fit to the measured levels of the dynamic shearing resistance of the fully melt Al alloy films.

40 35

Skew Angle 35
SHOT MO9904 Imp. Vel. 170.9 m/s (Viscosity=10 Pa.s) SHOT MO9905 Imp. Vel. 192.5 m/s (Viscosity=10 Pa.s) SHOT MO9906 Imp. Vel. 217.1 m/s (Viscosity=10 Pa.s) SHOT FY05008 Imp. Vel. 273.0 m/s (Viscosity=10 Pa.s) SHOT FY05001 Imp. Vel. 340.0 m/s (Viscosity=10 Pa.s) SHOT FY05004 Imp. Vel. 413.0 m/s (Viscosity=2 Pa.s) SHOT FY05007 Imp. Vel. 463.0 m/s (Viscosity=1 Pa.s)

Molten Layer Depth (m)

30 25 20 15 10 5 0

Time after Impact (s)

Figure 3.18: Kinetics of growth of molten 7075-T6 Al layer at the tribo-pair interface as a function of impact velocity. As observed from the simulations, the thickness of the molten metal film generated

197

during the slip process at a particular impact velocity increases parabolically with time. It is interesting to note that the growth rate is highest at early times and then decreases as the thickness of the molten metal layer increases. This behavior can be attributed to the higher background temperatures away from the slip interface at the latter times. Moreover, at all times, the thickness of the molten film is observed to increase with an increase impact velocity up to 340 m/s. In the case of the higher impact velocity experiments, i.e. with impact velocities in excess of 340 m/s, the thickness of the molten Al film is observed to decrease with the impact velocity. This decrease in film

thickness is understood to be due to the more localized shearing behavior at the slip interface at the higher impact velocities, i.e. higher slip speeds. The more localized shearing deformation is also reflected in the decrease of dynamic viscosity from ~ 10 Pa-s to ~ 1 Pa-s at the highest impact velocities used in the present experiments.

Figures 3.19 and 3.20 show snapshots of the deformed finite element mesh as a function of time showing deformation in 7075-T6 Al flyer plate in the vicinity of the frictional interface for Shot MO9906. The tribo-pair interface is located at the bottom of the deformed mesh. During the early part of the slip process, even though the interfacial temperature increases substantially, very little plastic deformation is observed in the flyer (Figure 3.19). However, as the thermal stresses grow, appreciable levels of deviatoric stresses are generated in the vicinity of the slip interface. These large deviatoric stresses along with the relatively high interfacial temperatures generate conditions conducive to the localization of shearing deformation at the frictional interface. Figure 3.20 shows

shearing deformation in Al after the transition from dry metal-on-metal slip to formation of

198

fully molten layer. It is to be noted that the re-meshing algorithm employed in the present computations allow arbitrarily large shearing deformations of molten metal layer to be simulated with no or little loss in computational accuracy.

199

Figure 3.19: Deformed mesh in 7075-T6 Al in the vicinity of the frictional interface prior to the melt transition for Shot MO9906. Direction of slip is from left to right.

200

Figure 3.20: Deformed mesh in 7075-T6 Al in the vicinity of the frictional interface after the melt transition for Shot MO9906. Direction of slip is from left to right.

201

(a)

(b)

(c)

(d)

Figure 3.21: Micrograph of the slip surface of the Al alloy flyer plate. (a) Shot MO9903; (b) Shot MO9906; (C) Shot Fuping05008; (d) Shot Fuping05007. In order to evaluate the post-test microstructures of the tribo-pair surfaces scanning electron microscopy was used. One of primary interests was to understand the extent of the heat and deformation affected zone and the changes in microstructure in the lower melt point metal, i.e. 7075-T6 Al alloy, during the high speed slip event. Figure 3.21 shows the SEM micrographs of the slip surfaces of Al alloy for Shot MO9903, MO9906, Fuping05008 and Fuping05007. Sliding wear tracks extending from left to right (slip

202

direction) can be clearly observed on the slip surfaces for all experiments. Unlike lower impact velocity experiment (MO9903, Figure 3.21(a)), molten aluminum can be clearly seen smeared on the sliding surface for Shot MO9906, Fuping05008 and Fuping05007 (Figure 3.21(b), 3.21(c) and 3.21(d)). Moreover, as the impact velocities increase the molten area increases, which can explain why the dynamic shear resistance of molten aluminum goes down and the dynamic shear viscosity of molten layers decrease when the impact velocities increase.

Highest Coefficient of Kinetic Friction, k

0.5

0.4

0.3

Quasi-Static (Bhusham and Gupta, 1991) MO9907 (Impact Velocity V=115 m/s) MO9903 (Impact Velocity V=137 m/s) MO9904 (Impact Velocity V=170.9 m/s) MO9905 (Impact Velocity V=192.5 m/s) MO9906 (Impact Velocity V=217.1 m/s) FY05008 (Impact Velocity V=273 m/s) FY05001 (Impact Velocity V=340 m/s) FY05004 (Impact Velocity V=413 m/s) FY05007 (Impact Velocity V=462 m/s)

0.2

0.1

0 0 10

Interfacial Slip Velocity, m/s

101

102

103

Figure 3.22: Coefficient of kinetic friction versus interfacial slip-velocity obtained from experiments conducted on the 7075-T6 Al alloy/CH tool-steel tribo-pair. Figure 3.22 summarizes the highest coefficient of kinetic friction as a function of the logarithmic of the slip-velocity obtained. It is interesting to note that the linear

203

dependence of the coefficient of kinetic friction on the logarithmic of the slip velocity at slip speeds as high as 250 m/s. Previous work by Dieterich (1978) and Ruina (1983) had showed a linear dependence of the coefficient of kinetic friction on logarithmic of the slip velocity. However, these friction experiments were conducted at quasi-static slip speeds of ~ 102 mm / s .
More recently, He and Robbins (2001), based on

molecular dynamic simulations of dynamic slip, have reported a linear dependence of the coefficient of kinetic friction and the logarithmic of the quasi-slip speeds. In view of the present results, it is evident that the friction-stress versus slip-speed dependence can be modeled as essentially logarithmic over several decades in change of slip velocity.

3.5 Summary

In the present study plate-impact pressure-shear friction experiments were conducted to investigate the dynamic slip resistance and time-resolved growth of molten metal films during dry metal-on-metal slip under extreme interfacial conditions. By employing a tribo-pair comprising a hard tool-steel against relatively low melt-point metal, ( 7075-T6 Al alloy), interfacial friction stress of up to 300 MPa and slip speeds of approximately 250 m/s have been achieved. These relatively extreme interfacial conditions are

conducive to the development of molten metal films at the tribo-pair interface.

A Lagrangian finite-element code is developed to understand the evolution of the thermo-mechanical fields and their relationship to the observed slip response. The code accounts for finite elastic-plastic deformations, affects of material inertia, heat

204

conduction, contact with friction, stain hardening, strain rate hardening, thermal softening, and full thermo-mechanical coupling. At temperatures below the melting point the material is described as an isotropic thermally softening elastic-viscoplastic solid. For material elements with temperatures in excess of the melt temperature a purely Newtonian fluid constitutive model is employed.

During the early part of frictional slip, the measured coefficient of kinetic friction is observed to decrease with increasing slip velocity, and is relatively low when compared to the reported values of 0.4 obtained for the coefficient of kinetic friction between mild steel and aluminum under quasi-static slip conditions (Bhushan and Gupta, 1991). A plausible reason can be attributed to the combination of relatively high values of normal pressure and high slip-speeds which generate substantial heat at the frictional interface. This heat produces a flash temperature and a thermal softening at the asperity-asperity junctions, thus reducing their shear strength, which in turn leads to lower values of friction stress (Archard, 1958; Kuhlmann-Wilsdorf, 1985; Kuhlmann-Wilsdorf, 1987; Ashby et al., 1991; Rice, 2006).

During the later part of the experiments interfacial slip occurs at near-melt to the fully-melt temperature regime of the 7075-T6 Al alloy. Under these fully melt

conditions the interfacial resistance approaches the shear strength of the molten aluminum alloy at normal stress in the range of 1-4.5 GPa and shear strain rates of ~ 107
205

s-1.

The results of the study indicate that under these extreme conditions molten

aluminum films maintain a shearing resistance as high as 50-100 MPa. Moreover, the simulations of the experiments reveal that the dynamic viscosity of the molten aluminum alloy decreases with the increasing impact velocity so that the purely Newtonian model can capture the dynamic shearing resistance of the molten aluminum alloy with sufficient accuracy. Scanning electron microscopy of the slip surfaces confirm the existence of molten Al to be smeared on the interface. Moreover, the molten area increases with the increasing impact velocity, which will give a plausible reason why the dynamic shear resistance of molten aluminum goes down when the impact velocities increase.

206

References

Achenbach, J. D., 1973. Wave Propagation in Elastic Solids. North-Holland, Amsterdam. Archard, J. F., 1958. The temperature of rubbing surfaces. Wear 2, 438-455. Asaro, R. J., and Rice, J. R., 1977. Strain localization in ductile single crystals. Journal of the mechanics and physics of solids 25, 309-338. Asaro, R. J., 1983. Micromechanics of Crystals and Polycrystals. In: Hutchinson, J. W., and Theodore, Y. W., (Eds.), vol. 23. Academic Press, New York, pp. 1-115. Ashby, M. F., Abulawi, J., and Kong, H. S., 1991. Temperature maps for frictional heating in dry sliding. Tribology transactions 34 (4), 577-587. Barber, J. R., 1970. The conduction of heat from sliding solids. International Journal of Heat and Mass Transfer 13, 857-869. Barker, L. M., The development of the VISAR, and its use in shock compression science. Shock compression of condensed matter-1999, Snowbird, UT, 1999. Belytschko, T., 1983. An overview of semidiscretization and time integration procedures. In: Belytschko, T., and Hughes, T. J. R., (Eds.). North-Holland, Amsterdam, pp. 1-65. Bhushan, B., and Gupta, B. K., 1991. Handbook of tribology: material coatings and surface treatments. McGraw Hill Inc., NY. Bowden, F. P., and Ridler, K. E. W., 1936. Physical properties of surfaces, III. Proceedings of Royal Society of London A 154, 640-656. Bowden, F. P., and Tabor, D., 1950. The Friction and Lubrication of Solids. Oxford University Press, London. Bowden, F. P., and Thomas, P. H., 1954. The surface temperature of sliding solids. Proceedings of Royal Society of London A 223, 29-36. Bowden, F. P., and Freitag, E. H., 1958. The friction of solids at very high speeds, I. Metal on metal, II. Metal on diamond. Proceedings of Royal Society of London A248, 350-367. Bowden, F. P., and Persson, P. A., 1960. Deformation heating and melting of solids in high speed friction. Proceedings of Royal Society of London A260, 433-458. Budiansky, B., Remarks on theories of solid and structural mechanics. 1969, pp. 77-83. Camacho, G. T., and Ortiz, M., 1997. Adaptive Lagrangian modelling of ballistic penetration of metallic targets. Computer Methods in Applied Mechanics and Engineering 142, 269-301.

207

Carslaw, H. S., and Jaeger, J. C., 1986. Conduction of heat in solids. Oxford University Press, London. Chhabildas, L. C., Sutherland, H. J., and Asay, J. R., 1979. A velocity interferometer technique to determine shear-wave particle velocity in shock-loaded solids. Journal of applied physics 50 (8), 5196-5201. Clifton, R. J., 1990. High strain rate behavior of metals. Applied mechanics reviews 43 (5, Part II), S9-S22. Cuitino, A. M., and Ortiz, M., 1992. A material-independent method for extending stress update algorithms from small-strain plasticity to finite plasticity with multiplicative kinematics. Engineering computations 9, 437-451. Dieterich, J. H., 1978. Time dependent friction and mechanics of stick-slip. Pure and Applied Geophysics 116, 668-675. Earles, S. W. E., and Powell, D. G., 1966. Variations in friction and wear between unlubricated steel surfaces. 181, Part 30, 171-179. He, G., and Robbins, M. O., 2001. Simulations of the kinetic friction due to adsorbed surface layers. Tribology Letters 10, 7-14. Hill, R., 1966. Generalized constitutive relations for incremental deformation of metal crystal by multislip. Journal of the mechanics and physics of solids 14, 95-102. Hill, R., and Rice, J. R., 1972. Constitutive analysis of elastic-plastic crystals at arbitrary strain. Journal of the mechanics and physics of solids 20, 401-443. Hill, R., and Rice, J. R., 1973. Elastic potentials and the structure of inelastic constitutive laws. SIAM, Journal of applied mathematics 25, 448-461. Hughes, T. J. R., Analysis of transient algorithms with particular reference to stability behavior. 1983, pp. 67-155. Hughes, T. J. R., 1987. The Finite Element Method. Prentice Hall, Englewood Cliffs, NJ. Irfan, M. A., and Prakash, V., Contact temperatures during sliding in pressure shear impact. Proceedings Society of Experimental Mechanics Conference, Baltimore, MD, 1994, pp. 173-182. Irfan, M. A., and Prakash, V., 2000. Time resolved friction during dry sliding of metal on metal. International journal of solids and structures 37, 2859-2882. Kadhim, M. J., and Earles, S. W. E., 1966. Unlubricated sliding at high speeds between copper and steel surfaces. Proceedings institute of mechanical engineers 181, Part 30, 157-162. Kennedy, F. E., 1984. Thermal and thermomechanical effects in dry sliding. Wear 100,
208

453-476. Kobayashi, S. O. S. I. A. T., 1989. Metal forming and Finite Element Method. Oxford University Press, London. Krafft, J. M., 1955. Surface friction in ballistic penetration. Journal of applied physics 26 (10), 1248-1253. Krieg, R. D., and Key, S. W., 1973. Transient shell response by numerical time integration. International journal of numerical methods in engineering 7, 273-286. Kuhlmann-Wilsdorf, D., 1985. Flash temperatures due to friction and joule heat at asperity contact. Wear 105, 187-198. Kuhlmann-Wilsdorf, D., 1987. Temperatures at interfacial contact spots:Dependence on velocity and on role reversal of two materials in sliding contact. Journal of tribology 109, 321-329. Kumar, P., and Clifton, R. J., 1977. Optical alignment of impact faces for plate impact experiments. Journal of Applied Physics 48, 1366-1367. Lee, E. H., 1969. Elastic-plastic deformation at finite strains. Journal of Applied Mechanics 36, 1-6. Lemonds, J., and Needleman, 1986. A finite element analysis of shear localization in rate and temperature dependent solids. Mechanics of Materials 5, 339-361. Lim, S. C., Ashby, M. F., and Brunton, J. H., 1989. The effects of sliding conditions on the dry friction of metals. Acta Materialia 37, 767-772. Liou, N. S., Okada, M., and Prakash, V., 2004. Formation of molten metal films during metal-on-metal slip under extreme interfacial conditions. Journal of the Mechanics and Physics of Solids 52 (9), 2025-2056. Marusich, T. D., and Ortiz, M., 1995. Modeling and simulation of high speed machining. International Journal for Numerical Methods in Engineering 38, 3675-3694. Montgomery, R. S., 1976a. Surface melting of rotating bands. Wear 38, 235-243. Montgomery, R. S., 1976b. Friction and wear at high sliding speeds. Wear 36, 275-298. Moran, B., Ortiz, M., and Shih, C. F., 1990. Formulation of implicit finite element methods for multiplicative finite deformation plasticity. International Journal for Numerical Methods in Engineering 29, 483-514. Nagtegaal, J. D., and DeJong, J. E., 1981. Some computational aspects of elastic-plastic large strain analysis. International Journal for Numerical Methods in Engineering 17, 15-41.

209

Ogawa, K., 1997. Impact friction test method by applying stress wave. Experimental Mechanics 37 (4), 398-402. Okada, M., Liou, N.-S., Prakash, V., and Miyoshi, K., 2001. Tribology of high speed metal-on-metal sliding at near-melt and fully-melt interfacial temperatures. Wear 249, 672-686. Okada, M., Liou, N. S., and Prakash, V., 2002. Dynamic shearing resistance of molten metal films at high pressures. Experimental Mechanics. Park, K. C., and Felippa, C. A., 1983. Chapter 3: Partitioned analysis of coupled systems. In: Belytschko, T., and Hughes, T. J. R., (Eds.), Computational Methods for Transient Analysis. Nort-Holland, Amsterdam, pp. 157-219. Peirce, D., Shih, C. F., and Needleman, A., 1984. A tangent modulus method for rate dependent solids. Computers and Structures 18 (5), 875-887. Prakash, V., and Clifton, R. J., 1993. Time Resolved Dynamic Friction Measurements in Pressure-Shear. In: Ramesh, K. T., (Ed). Experimental Techniques in the Dynamics of Deformable Bodies, vol. AMD Vol. 165. ASME, New York, pp. 33-47. Prakash, V., 1995. A pressure-shear plate impact experiment for investigating transient friction. Experimental Mechanics 35 (4), 329-336. Rajagopalan, S., and Prakash, V., 1999. A Modified Torsional Kolsky Bar for Investigating Dynamic Friction. Experimental Mechanics 39 (4), 295-303. Rice, J. R., 2006. Heating and weakening of faults during earthquake slip. Journal of Geophysical Research-Solid Earth 111 (B5). Ruina, A., 1983. Slip stability and state variable friction laws. Journal of Geophysical Research 88 (B12), 10359-10370. Shugarts, 1953. Measuring friction at high speeds. Journal of Franklin Institute 256, 187-189. Steinberg, D., 1996. Equation of state and strength properties of selected materials. Lawrence Livermore National Laboratory. Sternlicht, B., and Apkarian, H., 1960. Investigation of "melt lubrication". ASLE transactions 2, 248-256. Taylor, L. M., and Flanagan, D. P., 1987. Pronto 2D: A two dimensional transient solid dynamics program. Sandia National Laboratories, Albuquerque, New Mexico. Teodosiu, C., A dynamic theory of dislocations and its applications to the theory of the elastic-plastic continuum. 1970, pp. 837-876. Teodosiu, C., and Sidoroff, F., 1976. A finite theory of the elastoviscoplasticity of single
210

crystals. International Journal of Engineering Science 14, 713-723. Weber, G., and Anand, L., 1990. Finite deformation constitutive equations and a time integration procedure for isotropic, hyperealstic-viscoplastic solids. Computer Methods in Applied Mechanics and Engineering 79, 173-202. Williams, K., and Griffen, E., 1964. Friction between unlubricated steel surfaces at sliding speeds upto 750 feet per second. Proceedings institute of mechanical engineers 178, Part 3N, 24-36.

211

Chapter 4

SPALL STRENGTH OF GLASS FIBER REINFORCED POLYMER COMPOSITES 4.1 Introduction The utilization of layered heterogeneous material systems in the development of integral armor provides a potential for a major improvement in the ballistic performance in a variety of lightweight armor applications. Some of the notable recent examples demonstrating the success of synthetic heterogeneous material systems include composite materials with organic matrices reinforced by glass fibers to achieve lightweight and enhanced ballistic resistance. Under the U.S. Armys Composite Armor Vehicles (CAV)

and the Future Combat Systems (FCS) programs, various light-weight and highly damage-tolerant composite material systems have been investigated to understand and optimize the performance of potential Composite Integral Armor (CIA) systems (DeLuca et al., 1998; Mahfuz et al., 1999; Fink, 2000). Due to their light-weight, high stiffness,

and good ballistic resistance, various GRP composites have been chosen in composite integral armor as the main structural support behind the ceramic plates (Gama et al., 2001a,b).

Although GRPs were introduced in the 1930s, the dynamic response of these material systems was not the focus until the 1970s when drop-weight testing machines were
212

utilized to estimate their impact strength.

Lifshitz (1976) investigated the tensile

strength and failure modes of unidirectional and angle-ply E-glass fiber-reinforced epoxy matrix composites at strain rates between 0.1 ~ 200 s-1. The failure stresses under impact loading conditions were found to be considerably higher when compared to those obtained under quasi-static loading conditions. In recent years the dynamic response of glass-fiber reinforced composites has been investigated utilizing the Split Hopkinson Pressure Bars (SHPBs) under relatively simple states of stress, e.g., uniaxial compression, uniaxial tension, and pure shear (Elhabak, 1991; Agbossou et al., 1995; Tay et al., 1995; Barre et al., 1996; Sierakowski, 1997; Gama et al., 2001a,b; Song et al., 2002; Vural and Ravichandran, 2004). In these studies the failure and ultimate strength of the GRP composites were found to increase with increasing strain rates.

Most GRP material systems have excellent strength along the fiberglass direction. However, the cohesion between the fiberglass reinforcement and the resin matrix is not very strong, thereby making them susceptible to spall during a typical impact process. Spallation is the failure of material due to the action of tensile stresses developed in the interior of a sample through the interaction (overlap) of two release waves (Gray, 2000), or more specifically the process of internal failure or rupture of continuum media through a mechanism of decohesion due to stresses in excess of the tensile strength of the material (Grady and Kipp, 1993). In the past, plate impact experiments and/or direct contact explosives methodologies have been employed to investigate the spall strength in
213

materials.

The main advantage of these techniques is that in these experiments Consequently, during the time

nominally plane waves of uniaxial strain are utilized.

duration of interest, the applied loading is homogeneous in the central part of the specimen. The spall strength determined in this manner is thus the pure tensile stress

necessary to pull the constituents of the composite apart. Additionally, the location of the spall plane in the specimen (where the tensile stresses are operative), can be precisely controlled by proper selection of the experimental configuration. In the past, using plate

impact experiments with Al 6061-T6 flyer plates, Dandekar et al. (1998) studied the spall strength in S2 glass woven roving in Cycom 4102 polyester resin matrix subjected to shock compression and combined shock compression and shear loading. Zaretsky et al. (2004) also utilized Al 6061-T6 flyer plates to conduct plate impact experiments on a woven glass-fiber reinforced composite in a 7781 epoxy resin matrix. The measured spall strengths were observed to vary from 60 MPa (Dandekar et al., 1998) to about 190 MPa (Zaretsky et al., 2004).

In the present investigation normal impact and the combined pressure-shear plate impact experiments are conducted to investigate the spall strength in two different architectures of the GRP composites S2 glass woven roving in Cycom 4102 polyester resin matrix and a 5-harness satin weave E-glass in a Ciba epoxy (LY564) matrix. The

GRP specimens were shock loaded by utilizing the 82.5 mm bore single stage gas-gun at the Case Western Reserve University. The thicknesses of the flyer and target plates
214

were carefully designed so as to produce a state of tension near the center of the GRP target plates. Normal plate impact and combined pressure and shear plate impact

experiments with skew angles ranging from 12o to 20o, were performed to study the effects of normal compression and combined compression and shear on the spall strength of the GRP composites. The results of these experiments were used to develop a failure

map for the two GRP composites. 4.2 Material In the present investigation two different types of GRP specimens were investigated: (a) S2 glass woven roving in Cycom 4102 polyester resin matrix, and (b) a balanced 5-harness satin weave E glass in a Ciba epoxy (LY564) matrix. The S2 glass GRP

composites were fabricated at the Composites Development Branch, US Army Research Laboratory, Watertown, MA, USA, while the E-glass GRP composite was fabricated by the DRA Land Systems, Great Britain. The S2 fiberglass fibers (in which S stands for

higher-strength glass fiber), are known to be stronger and stiffer than the E-glass fiber reinforcement-- they have a 40% higher tensile strength, 10 to 20% higher compressive strength, and much greater abrasion resistance when compared to the E-glass fibers (Wallenberger et al., 2001).

The S2 glass GRP laminates used in the present study were made from S2 glass woven roving in CYCOM 4102 polyester resin matrix with a resin content of 322% by

215

weight.

The individual laminate plies were 0.68 mm in thickness.

Composites of the

desired thickness were manufactured by stacking an appropriate number of plies in a 90o sequence. The desired number of laminates was stacked between two steel plates The stacked layers were then vacuum bagged and subjected to the

with release film.

following heat cycle:

(1) Initially heated to 339 4 K for 45 minutes. (2) Temperature raised to 353 2 K for 2 hours. (3) Temperature raised to 398 4 K and held for 2 hours. (4) Cooled to 312 12 K at the rate of 7 K per minute.

The curing cycle was initiated with a gradual temperature increase under vacuum conditions so that the volatile gases including the water vapor can be driven off. Next, the curing temperature was gradually increased to its maximum and held constant for a couple of hours to develop a high degree of cross-linking, followed by application of pressure to consolidate the laminate (Jones, 1999). The final density of S2 glass GRP

was 1.959 0.043 kg/m3, while the longitudinal wave speed in the composite obtained from phase velocities of ultrasonic waves was 3.2 0.1 km/s in the thickness direction (Dandekar et al., 1998a,b).

216

The E-glass laminates comprised of a balanced 5-harness satin weave E-glass with Ciba epoxy (LY564) as the matrix. The resin content was 50% by volume. individual laminate plies were 1.37 mm in thickness. The

The composite was manufactured

by using the resin transfer molding process, in which an appropriate number of plies were stacked in 90o sequence to achieve the desired thickness. temperature was used to produce a reasonably tough matrix. A low cure-time and The final density of the

E-glass GRP was 1.885 kg/m3, while the longitudinal wave speed in the composite was 3.34 km/s in the thickness direction.

Figures 4.1 and 4.2 show SEM micrographs of the S2 glass and the E-glass fiber woven roving for the two composites, respectively. The E-glass GRP has a much Each fiberglass

smaller fiberglass bundle size when compared to the S2 glass GRP.

bundle is approximately 5 mm in width for the S2 glass GRP, while it was approximately 1.25 mm for the E-glass GRP.

217

Figure 4.1: SEM micrograph of the S2 glass fiber woven roving layer.

Figure 4.2: SEM micrograph of the 5-harness satin weave E-glass fiber woven roving layer.

4.3 Experimental Procedure 4.3.1 Experimental Configuration and Setup In the present study a series of plate-impact experiments were conducted to study the spall strength in GRP using the 82.5 mm bore single-stage gas-gun facility at the Case Western Reserve University. Figure 4.3 shows the schematic of the experimental

configuration used for the normal plate impact and the combined pressure-shear plate

218

impact experiments.

For the case of the normal plate impact experiments the skew

angle of the flyer plate is zero. A fiberglass projectile carrying the flyer plate is accelerated down the gun barrel by means of compressed nitrogen. The maximum projectile velocity attainable with a typical projectile weighing 1.0 Kg is 600 m/s. The

rear end of the projectile has sealing O-ring and a Teflon key that slides in a key-way inside the gun barrel to prevent any rotation of the projectile. In order to conduct the plate impact experiments a metallic flyer-plate (Al 7075-T6) is impacted with the GRP target plate at both normal and oblique incidence. In order to reduce the possibility of an air cushion between the flyer and target plates, impact takes place in a target chamber that has been evacuated to 50 m of Hg prior to impact. A laser-based optical system,

utilizing a UNIPHASE Helium-Neon 5mW laser (Model 1125p) and a high frequency photo-diode, is used to measure the velocity of the projectile. To ensure the generation

of plane-waves with wave-front sufficiently parallel to the impact face, the flyer and the target plates are carefully aligned to be parallel to within 210-5 radians by using an optical alignment scheme developed by Kim et al. (1977). The actual tilt between the two plates is measured by recording the times at which four, isolated, voltage-biased pins, that are flush with the surface of the target plate, are shorted to ground. A velocity

interferometer is used to measure the history of the normal particle velocity at the rear surface of the target plate. The multi-beam VALYN VISAR is used as the

interferometer system (VISAR stands for Velocity Interferometer for Any Reflector, and was first developed by Barker and Hollenback (1972)).
219

A COHERENT VERDI 5W

solid-state diode-pumped frequency doubled Nd:YVO4 CW laser with wavelength of 532 nm is used to provide a coherent monochromatic light source. Other details regarding

the design, execution and data analysis of the experiments can be found elsewhere (Prakash, 1995).

Figure 4.3: Schematic of the plate impact experimental configuration used in the present study to investigate the spall strength in the GRP under normal shock compression and combined shock compression and shear loading.

4.3.2 Target Assembly In all experiments, the Al alloy flyer plates were 3 inches in diameter, while the GRP target plates were 2.5 by 2.5 inches in square. A typical target holder with the GRP specimen is shown in Figure 4.4. The target holder is made of 6061-Al alloy. Besides

being useful in holding and aligning the target, the target holder also provides the ground for the trigger and the tilt measurement systems. One ground pin and four trigger pins are mounted near the periphery of the GRP specimen. The GRP specimen and the

220

ground and the trigger pins are all glued in place by epoxy and lapped flush with the impact surface, shown face-down in Figure 4.4. In all the experiments conducted in the present study a thin (60~125 nm) aluminum coating is applied to the rear surface of the GRP specimen so as to facilitate laser based diagnostics using the multi-beam VALYN VISAR.

Figure 4.4: Photograph showing a typical GRP specimen mounted on the aluminum target plate. 4.4 Wave Propagation in the Flyer and the Target Plates A schematic of the time versus distance diagram (t-X diagram), which illustrates the propagation of compression waves and tensile waves through the target and flyer plates

221

during the plate impact spall experiments, is shown in Figure 5. The abscissa represents the distance in the flyer and the target plates from the impact surface while the ordinate represents the time after impact. The arrows indicate the direction of wave propagation.

Upon impact of the flyer and the target plates, two compressive waves are generated. These waves propagate from the impact surface into the flyer and the target plates with wave speeds that are characteristic of the flyer and target plate materials. Since the flyer has a smaller thickness than the target and the Al alloy flyer has a higher longitudinal wave speed (6.23 km/s) than that of the GRP targets, the compressive wave in the flyer reflects as a release wave from its free surface, part of which is transmitted into the GRP target plate. Similarly, the compressive wave in the target reflects from its back surface as a rarefaction wave and interacts with the release wave from the flyer to generate a state of tensile stress at a predetermined plane in the target (represented as State 7 in the target). If the amplitude of the tensile wave is large enough, the GRP target will

undergo spall failure. Since the spall failure is associated with the creation of a free surface, the tensile stress wave is reflected back from this surface as a compression wave, as shown in Figure 4.5.

The stress vs. particle velocity (S-V) diagram, shown in Figure 4.6, details the locus of the stress and particle-velocity states that can be attained during a typical plate-impact experiment. The abscissa represents the particle velocity in the target and the flyer For the case

plates, while the ordinate represents the stress in the target and flyer plates.
222

in which the spall strength is larger than the tensile strength, the stress and particle velocity in the GRP moves along the dashed lines from State (5) to the no-spall state denoted by State (7). However, if the tensile stress is greater than the spall strength of the GRP (spall indicated by the short dashed lines), the GRP will spall and the tensile stress in State (7) will unload to the stress free state denoted by State (7). The

compressive end of spall wave from State (7) arrives at the free surface of the GRP and brings the free surface particle velocity to State (10), which is the same as that in State (6) and also in State (7). The free surface particle velocity in States 6, 7 and 10,

is referred to as Vmax, and the corresponding free surface particle velocity in State (8) is referred to as Vmin.

223

Figure 4.5: Time-distance diagram showing the wave propagation and the stress states in the flyer and the target plates. The spall plane occurs approximately in the middle of the target plate.

Figure 4.6: Stress-velocity diagram showing the loci of all the stress and particle velocity states that can be achieved in a typical plate-impact spall experiment.

224

4.5 Determination of Spall Strength and the Impact Stress Figure 4.7 shows the measured free-surface particle velocity and the t-X diagram for a typical plate impact spall experiment, FY06001, on the E-glass GRP. The abscissa

represents the time after impact while the ordinate represents the free surface particle velocity measured at the rear surface of the GRP target plate. At time T1, when the

compression wave arrives at the free surface of the GRP plate, the free surface particle velocity rises to the level Vmax, which is consistent with the Hugoniot stress and particle velocity state corresponding to the impact velocity used in the experiment. At time T2,

the release waves from the back of the target and the flyer plates intersect at the middle of the GRP plate; the corresponding unloading tensile wave and the end of spall compressive wave propagate and arrive at the free surface of the GRP plate at times T3 and T4, respectively. At time T3, the free surface particle velocity in the GRP plate starts to decrease and reaches a level Vmin, at the time T4, before recovering to its Hugoniot state level of Vmax. This initial decrease followed by a recovery in the free

surface particle velocity, is also referred to as the pull-back characteristic of the spall signal, and is useful in the calculation of the materials spall strength, as detailed in the following.

225

Figure 4.7: Time-distance diagram paired with the measured free surface particle velocity profile for Experiment FY06001 to illustrate the pull-back phenomenon in the free surface particle velocity profile for a typical plate-impact spall experiment. The method applied for calculating the spall strength from the measured free surface particle velocity history is illustrated in Figure 4.8. The free surface particle velocity The

data for experiment FY06001 (shown in Figure 4.7) is used as an example.

abscissa represents the time after impact and the ordinate represents the free surface particle velocity measured by the VISAR. Due to the oscillatory nature of the measured

free surface particle velocity profiles in GRP, Vmax was taken to be the average free surface particle velocity during the shocked Hugoniot state. This level is also consistent with the prediction of the particle velocity in the Hugoniot state as obtained by using the

226

EOS for the flyer and the target materials.

After the spall event, the free surface particle In most spall experiments, Vo is

velocity drops to Vmin, followed by a pull back to Vo.

expected to be equal to Vmax; however in experiments where Vo is observed to be smaller than Vmax, the occurrence of a partial spall is indicated. Vno spall corresponds to State (7)

in Figure 4.6, when the tensile stress is not high enough to create spall.

120 110 100

V max=103.3 m/s V 0=91.5 m/s Spall

Free Surface Velocity (m/s)

90 80 70 60 50 40 30 20 10 0 2 4

V min=66.6 m/s

V no spall=45.7 m/s Shot FY06001 Impact Velocity = 71 m/s Flyer: 7075-T6 Al (12.5 mm) Target: E-glass GRP (10.34 mm)
6 8 10 12

Time after Impact (s)

Figure 4.8: Free surface particle velocity profile for Experiment FY06001 showing the calculation of the spall strength. The spall strength of the GRP can be estimated by

spall = Z GRP (Vmax Vmin ) / 2 ~119.5 MPa .

(4.1)

227

In Equation (4.1), ZGRP is the acoustic impedance of the GRP in the zero stress condition and is calculated from the initial density and longitudinal wave speed of the GRP. The

S2 glass GRP has an acoustic impedance of 6.288 MPa/(m/s), and the E-glass GRP has an acoustic impedance of 6.296 MPa/(m/s).

The Hugoniot is the locus of all the shock states in a material and essentially describes the shock response of a material. In the present work, in order to estimate the Hugoniot stress state (impact stress) at the flyer and the target interface the Equation of States (EOS) for the flyer and the target materials are utilized. For most materials, the EOS can be approximated as a linear relationship between the shock velocity and the particle velocity (U s vs. u p ) given by

U s = C 0 + Su p

(4.2)

where, S is experimental determined parameter and C 0 is the sound velocity in the material at zero pressure (Meyers, 1994).

The EOS for the E-glass GRP is estimated from the shock velocity vs. particle velocity data obtained from the present experiments, as shown in Figure 4.9. The

abscissa represents the normal component of the particle-velocity within the shock compressed GRP while the ordinate represents the shock velocity.
228

The shock velocity is

estimated from the thickness of the GRP target plates and the shock arrival times at the free surface of the GRP plate. The particle-velocity, u p , is estimated from the

measured free surface particle velocity profiles (Vmax ) in the GRP target plates in the shocked state,

u p = 1/ 2Vmax .

(4.3)

The linear fit of the U s vs. u p data (shown in Figure 4.9) provides the Equation of State for the E-glass GRP

U s = 3.3 + 0.90u p .

(4.4)

229

Shock Velocity (km/s)

2
Experimental data for E-glass GRP Linear fit: U s = 3.3 + 0.90 up

0.1

0.2

0.3

0.4

Particle Velocity (km/s)

Figure 4.9: Shock velocity vs. Particle velocity for E-glass GRP.

The Equation of State for the S2 glass GRP is taken from Tsai and Prakash (Tsai and Prakash, 2005)

U s = 3.2 + 0.96u p .

(4.5)

The HEL of Al alloy flyer plate is 640 MPa while the Equation of State is given by Lundergran (Lundergran, 1963).

U s = 5.37 + 1.34u p .

(4.6)

230

From the Rankine-Hugoniot conservation relationships, the Hugoniot stress, H , under plate impact, can be determined by the following relations

H = 0GRPU sGRP u p = 0GRP (C 0GRP + S GRP u p )u p ,

(4.7)

H = 0AlU s Al (u p uI ) .

(4.8)

In Equations (4.7) and (4.8), 0GRP and 0Al are initial densities of GRP and aluminum alloy, respectively; C 0GRP and S GRP are constants in the Equation of State of the GRP; and uI is the impact velocity. In equation (4.8), when the Hugoniot stress level is below the Hugoniot Elastic Limit (HEL) of the Al alloy flyer plate, U s AL is taken to be the elastic longitudinal impedance of the Al alloy. However, when the Hugoniot stress level is above the HEL of the Al alloy, U s AL represents the shock velocity and is determined from the Equation of State of the Al alloy.

4.6 Experimental Results In the present study results of a series of plate impact experiments designed to study spall strength in glass fiber reinforced polymer composites are presented. Two GRP

architectures are investigated S2 glass woven roving in Cycom 4102 polyester resin matrix and a balanced 5-harness satin weave E-glass in a Ciba epoxy (LY564) matrix. The spall strengths in these two composites were obtained as a function of the normal

231

component of impact stress and the applied shear strain by subjecting the GRP specimens to shock compression and combined shock compression and shear loading. were used to develop a failure surface for the two GRP composites. The results

Table 4.1 provides a summary of all the experiments conducted on the S2 glass GRP in the present study. It shows the Experiment No, the flyer and the target plate materials,

the thickness of the flyer and target plates, the impact velocity, and the skew angle of impact. In this series of experiments the impact velocity was varied from 8.5 m/s to In the case of the combined pressure and shear plate-impact experiments,

138.8 m/s.

skew angles of 12o, 15o, and 20o were utilized. Table 4.2 shows the corresponding experiments on the E-glass GRP. varied from 71 m/s to 448.8 m/s. In this series of experiments the impact velocity was Moreover, as for the case of the S2 glass GRP, skew

angles of 12o, 15o, and 20o were utilized.

232

Experiment No.

Flyer Thickness: Al 7075-T6 (mm)

Target Thickness: S2 glass GRP (mm) 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95 12.95

Impact velocity (m/s) 8.5 38.1 39.1 43.9 108.1 133.2 138.8 48.4 59.9 68.1 75.7 42.3 43.4 82.8 104.7 31.9 47.3 68.9

Skew angle (0)

LT38 LT39 LT37 LT36 LT40 LT53 LT52 LT60 LT57 LT61 LT56 LT43 LT58 LT55 LT42 LT59 LT45 LT44

13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59 13.59

0 0 0 0 0 0 0 12 12 12 12 15 15 15 15 20 20 20

Table 4.1: Summary of all the normal plate impact and the pressure-shear plate impact experiments conducted to obtain the spall strength of S2 glass GRP.

233

Experiment No.

Flyer Thickness: Al 7075-T6 (mm) 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5 12.5

Target Thickness: E-glass GRP (mm) 10.34 10.34 10.34 10.34 10.34 10.34 10.34 10.34 10.34 10.34

Impact velocity (m/s)

Skew angle (0)

FY06001 FY06002 FY06003 FY06004 FY06005 FY06007 FY06006 FY06008 FY06009 FY06010

71 141 199.8 300.1 448.8 113.6 213.3 128.1 177.2 180.2

0 0 0 0 0 12 12 15 15 20

Table 4.2: Summary of all the normal plate impact and the pressure-shear plate impact experiments conducted to obtain the spall strength of E-glass GRP.

Figure 4.10 shows the spall strength data collected from all the normal plate-impact experiments on the E-glass and the S2-glass GRP composites conducted in the present work. The abscissa represents the impact stress while the ordinate shows the estimated

spall strength obtained from the experiments using Equation (4.1). Amongst the seven normal plate-impact experiments conducted on the S2 glass GRP composite, in experiments LT38 and LT39 ( impact stresses lower than 180 MPa) the resultant tensile stress was not sufficient to cause spallation in the specimens. In experiments LT36, LT37 and LT40, (i.e. with impact stresses in the range from 180 to 500 MPa), a finite

234

spall strength was measured. In experiments LT52 and LT53 (with impact stresses greater than 600 MPa), no pull-back signal in the free surface particle velocity profile was observed, indicating that during shock compression the GRP was damaged to such an extent that it could not support any tensile stress (i.e. delamination of the composite occurred with a negligible spall strength).

In all the five normal plate-impact spall experiments that were conducted on the E-glass GRP composite (impact stresses ranging from 330.7 MPa to 2213.8 MPa), a finite spall strength was measured. Also, these levels of spall strength are significantly

higher when compared to the spall strengths measured in S2 glass GRP composites. However, like in the case of the S2 glass GRP, the spall strengths in the E-glass GRP composite was observed to decrease with increasing levels of applied shock compression.

235

140 120

E-glass GRP S2 glass GRP

Spall Strength (MPa)

100 80 60 40 20 0

500

1000

1500

2000

2500

Impact Stress (MPa)

Figure 4.10: Spall strength vs. Impact stress obtained from the normal plate-impact experiments.

In order to illustrate the effect of combined shock compression and shear loading on the spall strength, results of one normal impact and one oblique impact experiment on the E-glass GRP are presented in Figure 4.11. The figure shows the free surface particle

velocity profiles for a normal plate impact experiment (FY06003) and a 20o pressure-shear plate impact experiment (FY06010). The normal component of the

impact stress in the two experiments, FY06003 and FY06010, were 978.0 MPa and 871.4 MPa, respectively. The magnitude of the shear-strain, 13, in the sample for experiment

FY06010 was 1.465%. The shear strain was calculated by using the analysis presented in the Appendix A (Dandekar et al., 1998).

236

13 =

0 4 4 1 2 C 11 sin + C 33 cos + C 13 + C 44 sin 2 2

33 sin cos

(4.9)

In Equation (4.9), 33 is the impact stress along the gun barrel direction and is calculated from the impact velocity and the impedance of the flyer and the target materials; and 0 are the densities of the GRP after and before impact, respectively, and

can be determined by shock velocity and particle velocity; Cij are the elastic 0

constants of GRP and are taken from Dandekar et al. (1998); and is the skew angle of the pressure-shear plate impact experiments.

The spall strengths estimated in the two experiments with and without the presence of shear strain, i.e. experiments FY06003 and FY06010, were 105.1 MPa and 40.4 MPa, respectively. From these results it is quite evident that the presence of shear-strain decreases the spall strength of the E-glass GRP dramatically. For example, in

experiment FY06006 on the E-glass GRP, the spall strength is reduced to essentially zero when the specimen is impacted at a normal stress of 1052.9 MPa and a shear strain of 1.056%.

237

350

300

Spall

Free Surface Velocity (m/s)

250

200
Shot FY06003 Impact Velocity = 199.8 m/s Normal Stress = 978.0 MPa Normal Impact Shear Strain = 0 Spall Strength: 105.1 MPa Shot FY06010 Impact Velocity = 180.2 m/s Normal Stress = 871.4 MPa Skew Angle 20 Shear Strain = 1.465% Spall Strength: 40.4 MPa

150

100

50

Time after Impact (s)

10

Figure 4.11: Free surface particle velocity profiles for Experiments FY06003 and FY06010. The effect of the superimposed shear strain on the spall strength of the E-glass GRP is emphasized.

To illustrate the effects of the shear stress on the spall strength of the S2 glass GRP, results of four pressure-shear plate impact spall experiments (conducted at a normal impact stress of approximately 200 MPa), are shown in Figure 4.12. The abscissa represents the shear-strain while the ordinate represents the spall strength. The normal

components of the impact stresses in these experiments were 187.9MPa, 204.4MPa, 192.9MPa, and 217.5MPa, respectively. As seen from the figure, the spall strength in

these experiments drops very rapidly, i.e. from 39.4 MPa to essentially zero, as the shear strain is increased from 0.229% to 0.353%. These results indicate that for the E-glass

238

GRP much higher levels of normal stress and shear strains are required to reduce its spall strength to essentially zero when compared to the S2 glass GRP.

50 45 40

Shot LT60 Normal Stress 217.5 MPa Shear Strain 0.229% Spall Strength 39.6 MPa Shot LT43 Normal Stress 187.9 MPa Shear Strain 0.245% Spall Strength 33.8 MPa Shot LT58 Normal Stress 192.9 MPa Shear Strain 0.252% Spall Strength 18.3 MPa Shot LT45 Normal Stress 204.4 MPa Shear Strain 0.353% Spall Strength 0 MPa
0.25 0.3 0.35 0.4

Spall Strength (MPa)

35 30 25 20 15 10 5 0 0.2

Shear Strain (%)

Figure 4.12: Spall strength as a function of the shear strain in the S2 glass GRP for selected experiments each having a normal component of the impact stress of about 200 MPa.

Table 4.3 provides a summary of normal stress, shear strain and the measured spall strength from all the experiments conducted in the present study on S2 glass GRP. In

these experiments, the normal stress was varied from 39.0 MPa to 637.9 MPa, while the shear strain was varied from 0% to 0.615%. Table 4.4 shows the corresponding data for the E-glass GRP. The normal stress was varied from 330.7 MPa to 2213.8 MPa, and the

239

shear strain varied from 0.549% to 1.465%.

Exp No.

Normal stress (MPa)

Shear Strain (%)

Spall strength (MPa)

LT38 LT39 LT37 LT36 LT40 LT53 LT52 LT60 LT59 LT43 LT58 LT57 LT61 LT45 LT56 LT55 LT44 LT42

39.0 175.1 179.7 201.6 496.6 612.0 637.9 217.5 137.9 187.9 192.9 269.5 306.3 204.4 340.4 367.6 297.7 464.6

0 0 0 0 0 0 0 0.229 0.237 0.245 0.252 0.283 0.323 0.353 0.359 0.484 0.516 0.615

No spall No spall 46.1 35.8 45.7 0 0 39.6 22.7 33.8 18.3 53.7 0 0 0 0 0 0

Table 4.3: Summary of normal stress, shear strain and spall strength for S2 glass GRP.

240

Exp No.

Normal stress (MPa)

Shear Strain (%)

Spall strength (MPa)

FY06001 FY06002 FY06003 FY06004 FY06005 FY06007 FY06006 FY06008 FY06009 FY06010

330.7 668.4 978.0 1467.8 2213.8 534.8 1052. 605.3 855.3 871.4

0 0 0 0 0 0.549 1.056 0.771 1.094 1.465

119.5 108.1 105.1 78.7 69.7 86.1 0 85.1 73.9 40.4

Table 4.4: Summary of normal stress, shear strain and spall strength for E-glass GRP.

Figures 4.13 and 4.14 show the spall strengths as a function of the applied shear-strain and the normal stress obtained from all the experiments conducted on S2 glass and the E-glass GRP composites. The abscissa represents the normal stress during The

impact while the ordinate represents the shear-strain obtained in each experiment.

Z-axis represents the spall strength. The failure surface shows that the spall strength decreases with increasing shear-strain and with increasing normal stress for the two GRP composites. As noted earlier, the E-glass GRP shows much larger levels for the spall strength when compared to the S2 glass GRP. The maximum spall strength measured

241

for the E-glass GRP was 119.5 MPa, while the maximum measured spall strength for the S2 glass GRP was 53.7 MPa.

Figure 4.13: Spall strength illustrated in relationship with normal stress and shear strain for the S2 glass GRP.

242

Figure 4.14: Spall strength illustrated in relationship with normal stress and shear strain for the E-glass GRP.

4.7 Discussion and Summary A series of normal plate-impact and pressure-shear plate impact experiments were conducted to study the spall strength in two different glass fiber reinforced polymer composites. Based on the experimental results the seven normal plate-impact

experiments on the S2-glass GRP were placed in three different categories. Experiments in the first category were conducted at an impact stress between 0 and 175MPa. In these experiments the resultant tensile stress was too low to cause

spallation within the specimens and the free surface particle velocity profiles were observed to unload completely to their no-spall predicted levels. Experiments in the second category were conducted at impact stresses in the range of 175 MPa and 600

243

MPa; the resulting tensile stresses within the specimen were high enough to result in spall. In these experiments a clear pull-back signal was observed in the measured free surface particle velocity profiles. In the third category of the experiments, the incident These relatively high

compression stress pulse amplitude was larger than 600 MPa.

levels of shock compression resulted in enough damage in the GRP specimens such that no resistance to spall (i.e. zero spall strength) was registered in the experiments. The

corresponding free surface particle velocity profiles for these experiments show no signs of pull-back or unloading of the free surface particle velocity, and it remains at a level corresponding to the predicted Hugonoit state, Vmax. On the other hand, experiments

conducted on the E-glass GRP composites (at impact stresses ranging from 330.7 MPa to 2213.8 MPa) showed a finite spall strength. However, like in the case of the S2-glass GRP, the spall strength of the E-glass GRP composite was observed to decrease with increasing levels of shock compression.

Under the combined compression and shear loading (pressure-shear plate impact experiments), the spall strengths in the two GRP composites were found to decrease with increasing levels of applied normal and the shear stress. A zero spall strength condition

was found for the E-glass GRP when the specimen was impacted at a normal stress of 975 MPa and a shear strain of 1.056%, which is much higher than that obtained for the case of the S2 glass GRP composite. Based on these results, the spall strength for the

244

two GRP composites is illustrated as a failure surface in the shear strain and the normal stress space.

The measured spall strengths are much lower than those observed in monolithic metals, ceramics, polymer etc. In such homogeneous materials, the conventional spall

process is thought to proceed from the coalescence/growth of inherent defects, such as impurities, micro-cracks, pre-existing pores, etc. However, damage in GRP materials is complicated by the presence of additional heterogeneities due to the composite materials microstructure, and failure under impact loading is understood to the proceed by

various mechanisms -- the incident energy is dissipated through the spread of failure laterally as well as through the thickness. Moreover, due to their inherent multi-material

heterogeneous composition of the GRPs, several distinctive modes of damage are observed which includes extensive delamination and fiber shearing, tensile fiber failure, large fiber deflection, fiber micro-fracture and local fiber buckling. In particular, local fiber waviness is understood lead to inter-laminar shear failure in such materials (Hsiao and Daniel, 1996; Hsiao and Daniel, 1996). Moreover, strong wave-reflection-effects, between components with different shock impedance, lead to significant shock wave dispersion resulting in an overall loss of spall strength (Zhuk et al., 1994; Dandekar and Beaulieu, 1995; Zaretsky et al., 2004).

245

References

Agbossou, A., Cohen, I., and Muller, D., 1995. Effects of Interphase and Impact Strain Rates on Tensile Off-Axis Behavior of Unidirectional Glass-Fiber Composite Experimental Results. Engineering Fracture Mechanics 52 (5), 923-934. Barker, L. M., and Hollenbach, R. E., 1972. Laser interferometer for measuring high velocities of any reflecting surface. Journal Applied Physics 43 (11), 4669-4675. Barre, S., Chotard, T., and Benzeggagh, M. L., 1996. Comparative study of strain rate effects on mechanical properties of glass fibre-reinforced thermoset matrix composites. Composites Part a-Applied Science and Manufacturing 27 (12), 1169-1181. Dandekar, D. P., and Beaulieu, P. A., 1995. Compressive and tensile strengths of glass reinforced polyester under shock wave propagation. In: Rajapakse, Y. D. S., and Vinson, J. R., (Eds.), High strain-rate effects on polymer, metal, and ceramic matrix composites and other advanced materials. ASME, New York, NY, pp. 63-70. Dandekar, D. P., Boteler, J. M., and Beaulieu, P. A., 1998. Elastic constants and delamination strength of a glass-fiber-reinforced polymer composite. Composites Science and Technology 58 (9), 1397-1403. DeLuca, E., Prifti, J., Betheney, W., and Chou, S. C., 1998. Ballistic impact damage of S 2-glass-reinforced plastic structural armor. Composites Science and Technology 58 (9), 1453-1461. Elhabak, A. M. A., 1991. Mechanical-Behavior of Woven Glass Fiber-Reinforced Composites under Impact Compression Load. Composites 22 (2), 129-134. Fink, B. K., 2000. Performance Metrics for Composite Integral Armor. Journal of Thermoplastic Composite Materials 13 (5), 417-431. Gama, B. A., Bogetti, T. A., Fink, B. K., Yu, C. J., Claar, T. D., Eifert, H. H., and Gillespie, J. W., 2001a. Aluminum foam integral armor: A new dimension in armor design. Composite Structures 52 (3-4), 381-395. Gama, B. A., Gillespie, J. W., Mahfuz, H., Raines, R. P., Haque, A., Jeelani, S., Bogetti, T. A., and Fink, B. K., 2001b. High Strain-Rate Behavior of Plain-Weave S-2 Glass/Vinyl Ester Composites. Journal of Composite Materials 35 (13), 1201-1228. Grady, D. E., and Kipp, M. E., 1993. Dynamic fracture and fragmentation. In: Asay, J. R., and Shahinpoor, M., (Eds.), High-Pressure Shock Compression of Solids. Springer-Verlag, New York, pp. 265-322. Gray, G. T., 2000. Shock wave testing of ductile materials, Mechanical Testing and Evaluation Handbook, vol. 8. American Society for Metals, Materials Park, pp. 530-559.
246

Hsiao, H. M., and Daniel, I. M., 1996. Nonlinear elastic behavior of unidirectionai composites with fiber waviness under compressive loading. Journal of Engineering Materials and Technology-Transactions of the Asme 118 (4), 561-570. Hsiao, H. M., and Daniel, I. M., 1996. Effect of fiber waviness on stiffness and strength reduction of unidirectional composites under compressive loading. Composites Science and Technology 56 (5), 581-593. Jones, R. M., 1999. Mechanics of Composite Materials. Taylor & Francis, Philadelphia. Kim, K. S., Clifton, R. J., and Kumar, P., 1977. A combined normal and transverse displacement interferometer with an application to impact of Y-cut Quartz. Journal of Applied Physics 48, 4132-4139. Lifshitz, J. M., 1976. Impact strength of angle ply fiber reinforced materials. Journal of Composite Materials 10, 92-101. Lundergan, C. D., 1963. Equation of State of 6061-T6 aluminum at low pressures. Journal of Applied Physics 34(7), 2046-2052. Mahfuz, H., Zhu, Y., Haque, A., Abutalib, A., Vaidya, V., Jeelani, S., Gama, B., Gillespie, J., and Fink, B., 1999. Investigation of high-velocity impact on integral armor using finite element method. International journal of impact engineering 24 (2), 203-217. Meyers, M. A., 1994. Dynamic Response of Materials. John Wiley & Sons, Inc, New York. Prakash, V., 1995. A pressure-shear plate impact experiment for investigating transient friction. Experimental Mechanics 35 (4), 329-336. Sierakowski, R. L. C. S. K., 1997. Dynamic Loading and Characterization of Fiber-reinforced Composites. John Wiley & Sons, New York, NY, USA. Song, B., Chen, W., and Weerasooriya, T., 2002. Impact Response and Failure Behavior of a Glass/Epoxy Structural Composite Material. In: Wang, C. M., Liu, G. R., and Ang, K. K., (Eds.), Proceedings of the 2nd International Conference on Structural Stability and Dynamics. World Scientific Publishing Co., Singapore, Singapore, pp. 949-954. Tay, T. E., Ang, H. G., and Shim, V. P. W., 1995. An empirical strain rate-dependent constitutive relationship for glass-fibre reinforced epoxy and pure epoxy. Composite Structures 33 (4), 201-210. Tsai, L., Prakash, V., 2005. Shock response of S2 glass fiber reinforced polymer composites. In: Khan, A.S., Khoei, A.R. (Eds.), 11th International Symposium on Plasticity and its Current applications: Dislocations, Plasticity, Damage and Metal Forming: Material Rsponse and Multiscale Modeling. Neat Press, MD, USA, Kauai, Hawaii, pp. 364366.

247

Tsai. L., 2006. Shock wave structure and spall strength of layered heterogeneous glass/polyer composites, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, Ohio. Vural, M., and Ravichandran, G., 2004. Failure mode transition and energy dissipation in naturally occurring composites. Composites Part B-Engineering 35 (6-8), 639-646. Wallenberger, F. T., Watson, J. C., Li, H., 2001. Glass fibers. In: Miracle, D.B., Donaldson, S.L. (Eds.), . In: ASM HandbookComposites, Vol. 21. ASM, Metals Park, OH, pp. 2734. Zaretsky, E., deBotton, G., and Perl, M., 2004. The response of a glass fibers reinforced epoxy composite to an impact loading. International Journal of Solids and Structures 41 (2), 569-584. Zhuk, A. Z., Kanel, G. I., and Lash, A. A., 1994. Glass Epoxy Composite Behavior under Shock Loading. Journal De Physique Iv 4 (C8), 403-407.

248

Chapter 5

DYNAMIC RESPONSE OF A ZR-BASED BULK METALLIC GLASS UNDER PLANAR SHOCK LOADING

5.1 Introduction Amorphous metals, also referred to as metallic glasses, differ from ordinary metals in that their constituent atoms are not arranged on a crystalline lattice. Due to their

randomly-ordered atomic structures, metallic glasses are known to exhibit unusual mechanical properties, such as near theoretical strength, large elastic strains, high hardness, excellent wear and corrosion resistance, and increased fracture toughness when compared to other brittle, high compressive strength materials (Bruck et al., 1996, Lowhaphandu and Lewandowski, 1998). Nevertheless, until recently, amorphous

metals have largely been manufactured in the form of thin ribbons, usually less than 1 mm in thickness, because fast cooling rates (~ 106 K/sec) are required for retaining the meta-stable amorphous phase. The first reported bulk amorphous metals were Pd-based

alloys designed in the early 1980s. The real breakthrough came during the period from 1988 to 1990, when Inoue (2001) discovered multi-component liquid alloys with very deep eutectics that were capable of freezing to a glassy state of several centimeters in thickness by utilizing conventional cooling methods. At around the same time, at

Caltech, Johnsons group developed Zr-Ti-based and other sizable amorphous metals.
249

These Zirconium-based bulk metallic glasses have been of particular interest to the engineering community because of their low critical cooling rate (1 K/s) requirements, allowing thick samples (up to 10 mm) to be processed (Peker and Johnson, 1993).

The dynamic response of BMGs is of considerable interest for gaining insight into the high strain-rate response of this class of materials and for potential applications, such as kinetic energy penetrators (Johnson, 1999). However, in the past, the mechanical properties and structural performance of BMGs under dynamic loading conditions have been studied by a limited number of investigators. For a Zr-based BMG, i.e.

Zr41.25Ti13.75Ni10Cu12.5Be22.5, Lu (2002) reported the strength under dynamic and quasi-static loading conditions to be quite similar, with failure being characterized by predominantly inhomogeneous inelastic flow. In addition, temperature increases of up to 500o K were detected during the dynamic deformation and failure, suggesting that perhaps local heating of the material contributed to material softening. Moreover, in both Zr57Ti5Cu20Ni8Al10 and Zr41.25Ti13.75Ni10Cu12.5Be22.5 BMGs, strain-rate softening behavior has been reported (Hufnagel et al., 2002). Such behavior was also observed More recently, Sunny

when zirconium was replaced by hafnium (Subhash et al., 2003).

et al. (2005a, 2005b, 2006a, 2006b) have investigated the high strain-rate behavior of Zr41.25Ti13.75Ni10Cu12.5Be22.5 by utilizing a Split-Hopkinson Pressure Bar. In their study,

in-situ high speed video was used to examine the deformation modes and the models of failure under uniaxial dynamic compression.
250

The fully amorphous material was

observed to exhibit catastrophic failure with the formation of a dominant shear band and the corresponding slip event, while the annealed BMG was observed to fail by extensive fragmentation after the formation of an initial crack.

Even though the deformation and damage mechanisms of BMGs under moderate dynamic loading conditions have been investigated, only a few studies have addressed the shock response in BMGs. Bach et al. (1991) studied parameters for the shock wave consolidation of a metallic glass powder, such as the effects of the shock wave energy and shock wave duration. Conner et al. (2000) investigated the high strain rate behavior

of fiber-reinforced Zr41.2Ti13.8Cu12.5Ni10Be22.5 composites, and demonstrated their excellent potential as armor penetrators. Zhuang et al.(2002) conducted plate impact

experiments to investigate the shock response of a Zr-based bulk metallic glass, Zr41.2Ti13.8Cu12.5Ni10Be22.5 (Vitreloy-1), and its particulate composite,

Zr56.3Ti13.8Ni5.6Cu6.9Be12.5 (in-situ dendritic -phase reinforced Vitreloy (Hays et al., 2000)), which are to be respectively referred to as Vit-1 and -Vit hereafter. A

surprisingly low amplitude elastic precursor and bulk wave were observed to precede the rate-dependent large deformation shock-wave. Moreover, a concave downward

curvature in the shock Hugoniot for the BMG was observed, suggesting that a phase-change-like transition occurred during the shock compression event. The spalling

in Vit-1 was induced by shear localization, while in -Vit it was due to debonding of the -phase boundary from the matrix. The spall strengths, at a stain rate of 2106 s-1, were
251

determined to be 2.35 GPa and 2.11 GPa for Vit-1 and -Vit, respectively. Turneaure et al. (2004) conducted plane shock wave experiments up to 13 GPa on a Zr-based bulk amorphous alloy, Zr56.7Cu15.3Ni12.5Nb5.0Al10.0Y0.5, with a quasi-static strength of 2.6 GPa. From the measured particle velocity histories, the Hugoniot elastic limit (HEL) was determined to be ~ 7.1 GPa, with a corresponding elastic strain of approximately 4%. For the experiments in which the peak stress exceeded the HEL, a clear two-wave structure consisting of an elastic precursor followed by a plastic wave was observed. Moreover, the experimental results suggested that the shear strength of the Zr-based BMG is reduced as it is shocked above the elastic limit. Yang et al. (2005) employed a

two stage light gas-gun to investigate the effects of planar shock compression on void formation and cracking in a Zr-based BMG, Zr41Ti14Ni10Cu12.5Be22.5 under hypervelocity impact conditions. Cracking was proposed to be a result of void linkage in a direction that did not coincide with the maximum shear stress plane. Changing the state of stress within the BMG by employing a spherical nosed projectile was proposed to lead to the formation of a different nanovoid distribution. The authors proposed that the nucleation of the nanovoids was possibly initiated by release of excess free-volume under shock wave compression. More recently, Yang et al. (2006) investigated the damage features in Zr41Ti14Ni10Cu12.5Be22.5, subjected to hyper-velocity impact using a two-stage light gas-gun. Using scanning electron microscopy they showed that both radial and

symmetric cracks were formed on the shocked surface of the target plate when impacted by an aluminum flyer with an impact velocity of 2.7 Km/s.
252

Shear bands/cracks parallel

to each other, on the cross section close to the shocked surface of the target, were also observed. The damage features under the projectiles nose were also examined. It was shown that besides the formation of adiabatic shear bands/cracks, craters and lamination cracks were also formed, and the depth of the craters increased with increasing projectile velocity. Mashimo et al. (2006) extended the investigation of BMG under planar shock compression to pressures up to 50 GPa. Using a powder gun and an inclined-mirror

photographic technique they investigated the yield behavior and the phase change in a Zr-based BMG, Zr55Al10Ni5Cu30. They measured an HEL of 6.2 GPa. Moreover, a

kink was observed in the shock velocity versus particle velocity relationship (Hugoniot) at about 400 m/s in particle velocity, suggesting the occurrence of a shock-induced phase transformation, although no determination of the resulting phase(s) were provided.

The study of pressure effects on flow and fracture behavior of BMGs is important because of a number of reasons. One key reason is that deformation in amorphous

metals is proposed to be accompanied by a highly dilatant process (Spaepen, 1977, Argon, 1979); while the atoms in the alloy move relative to each other, free volume is created and destroyed. Under the influence of hydrostatic compressive stress, one might expect free volume generation to be inhibited, thereby increasing the resistance to plastic flow and also the plastic strain-to-failure of the alloy. Another key reason comes from

potential applications of BMG, such as in underwater vehicles (because of both the corrosion resistance) and structural composites.
253

In both cases hydrostatic pressure

effects can play an important role in controlling the overall structural design and performance. Separate papers by Connor et al. (1977), Vormelker et al. (1998) and

Lewandowski et al. (2006) show promise for composites incorporating a metallic glass. In such cases, deformation of the tungsten is expected to provide confinement in the metallic glass that can potentially lead to the development of high hydrostatic stresses, and thus strength enhancements in the metallic glass.

The earliest known paper to characterize the effects of pressure on a metallic glass was by Davis and Kavesh (1975), who studied the effects of hydrostatic pressure on Pd77.5Cu6Si16.5 metallic glass. Quasi-static compression tests were performed on 2 mm

diameter rods with the L/D = 2.0, with a tapered shape similar to that of a dog-bone specimen. length. The gage section of the specimen was 0.84 mm in diameter and 1.1 mm in

These tests were performed by a liquid-based high-pressure mechanical testing

apparatus. In these tests, an increase in pressure from 100 MPa to 650 MPa led to a small increase (from 1.45 GPa to 1.55 GPa) in the compressive strength of the material. The slight increase in the compressive stress with hydrostatic pressure suggests that there is some dependence on the normal stress at the fracture plane, and this dependence is best modeled by a Mohr-Coulomb criterion of the following form, n = 0 n , where

n and n refer to the shear stress and normal stress along the fracture plane, 0 is
the shear stress under pure shear, and is the value of the normal stress dependence.

254

Based on their experimental results, Davis and Kavesh determined the value of to be 0.044. By using the same miniature testing apparatus as used in the

aforementioned compression tests, 0.3 mm diameter fibers/foils were loaded in tension at quasi-static strain rates of 6 x 10-5/s at both ambient pressure and 700 MPa superimposed pressure. In both tests, the tensile stress-strain behavior was observed to very similar to

brittle materials, such as ceramics, with mostly linear elastic behavior and very limited nonlinearity just prior to specimen failure. Moreover, the addition of 700 MPa of

hydrostatic pressure led to a slight increase in the tensile strength (from 1.3 GPa to 1.45 GPa, which is only slightly below the fracture stress in compression). Fracture occurred

near the loading grips, and a fracture angle of approximately 50 was exhibited along a curved fracture surface

Another early attempt to infer the effects of pressure on the flow and fracture behavior of BMGs was performed by Donovan (1988) on samples of Pd40Ni40Pd20 metallic glass. In her work, 1-2 mm diameter samples of Pd40Ni40P20 metallic glass

were tested under uniaxial compression, modified plane-strain compression, pure-shear and tension. The results suggest that yielding in Pd40Ni40Pd20 also followed a

Mohr-Coulomb criterion of the form n = 0 n , rather than the von Mises criterion, which is more appropriate for polycrystalline metals. The sign of the normal stress

acting across the slip-plane has a significant effect on the yield strength so that the yield strength in compression is substantially higher than in tension, but the hydrostatic
255

pressure has only a small effect on yielding. After considering all their experimental data, a value of of 0.113 was determined, along with a flow stress in shear of 0.795 GPa.

By using a gas-based high pressure apparatus, Lewandowski et al. (1999, 2000a, 2000b, 2002) investigated the pressure effects on fracture behavior of

Zr41.25Ti13.75Cu12.5Ni10Be22.5 and Zr63Cu18Ni10Al9 under both quasi-static compression and tension. The compression testing was conducted under hydrostatic pressures in the range, from 100 MPa to upwards of 1 GPa, while tension testing was conducted under both ambient pressure and superimposed hydrostatic pressures up to 700 MPa. In the

compression testing, it was observed that the addition of hydrostatic pressure increased the strain-to-failure of the BMG specimen. Inhibition of the generation of free volume

under the influence of hydrostatic pressure, was attributed to the increase the plastic strain-to-failure. However, the compression specimens exhibited fracture angles of approximately 41 regardless of confining pressure; this was markedly different from the observations of the fracture angles in tension, which showed significant deviations from the 45o angle, as predicted by a typical von Mises type of material. It is to be noted that tension specimens, tested in ambient pressure conditions, exhibited a fracture angle largely oriented at 90, but with circumferential shear lips oriented at approximately 50. However, all tension specimens tested under hydrostatic pressure, except for one, exhibited fracture angles between 56 and 59, while one that failed near the grips
256

exhibited a fracture angle close to 50, providing evidence for yielding behavior following the Mohr-Coulomb criterion. After considering both the tension and

compression data, a value of of 0.04 was determined, along with a flow-stress in shear of 0.97 GPa, which, while lower than that measured during a torsion test on a solid sample, corresponds well with the measured 2.0 GPa compressive yield stress.

In another study, in order to investigate the effect of confinement pressures on the quasi-static flow behavior of Zr41.25Ti13.75Cu12.5Ni10Be22.5, Lu et al. (2002) utilized maraging steel confining sleeves around disk shaped specimens. By varying (increasing) the ratio of the outer radius of the confining sleeve to the inner radius, they were able to achieve confining pressures up to to about 2 GPa on the BMG specimens. However, in

their configuration, unlike the experiments with superimposed liquid or gas pressures, a constant pressure cannot always be maintained on the specimen, since during compression, the confinement depends on the lateral expansion of the metallic glass specimen and the confining sleeve. The experimental results indicated that the flow

stress of metallic glass increased substantially when the pressure was increased-- from 1.9 GPa under ambient pressure and room temperature conditions to 2.5 GPa under 2.0 GPa confining pressure.

More recently, experiments have been performed by Zhang and Eckert (2003a, 2003b) to investigate the normal-stress and pressure dependency of flow stress in
257

Zr59Cu20Al10Ni8Ti3 metallic glass. In addition to the different fracture morphologies in compression and tension, they claimed that the observed larger deviation in tension (54) when compared to that in compression (40) requires the tensile normal stresses to have a more severe effect on the flow stress than the compressive stresses. In order to account for this tension/compression asymmetry, they postulated (2003a) a different Mohr-Coulomb relationship in tension than in compression, i.e. n = 0 c n ,c and n = 0 t n ,t for compression and tension, respectively, where c and t are the constants representing the normal stress dependency in compression and tension, respectively, with c < t . It is to be noted that this result is in contradiction with the

observations of ,Lewandowski et al., who based on experimental observations on several different metallic glasses (Lowhaphandu et al., 1999; Lowhaphandu et al., 2000a; Lowhaphandu et al., 2000b; Lewandowski, and Lowhaphandu, 2002; Wesseling, et al., 2003), postulated that the value of the two constants were nearly identical.

In the present study, an 82.5 mm bore single stage gas-gun is utilized to obtain shock response in a zirconium-based BMG, Zr41.25Ti13.75Ni10Cu12.5Be22.5, subjected to planar shock compression and combined shock compression and shear loading. The first series of experiments are conducted to study the spall strength of the Zr-based BMG under both normal shock compression and combined shock compression and shear loading. A Ti6Al4V flyer plate was used to shock the BMG target plate to approximately 7 GPa under normal shock compression. The flyer thickness in each experiment was carefully
258

designed to produce a state of tension near the center of the BMG target plate.

For the

combined pressure and shear experiments skew angles ranging from 60 to 240 were utilized. During the experiments, the free-surface particle velocity history at the back of These

the bulk metallic glass target plate was recorded using a VALYNTM VISAR.

measurements were analyzed to gain : (a) a better understanding of the structure of shock waves in BMG subjected to planar shock compression; (b) estimate residual spall strength of the BMG as a function of normal stress and shear strain; and (c) obtain the Hugoniot elastic limit (HEL) of the material.

A second series of experiments was conducted to study the sensitivity of flow stress of the BMG to high normal compressive stresses, hydrostatic pressures and ultra-high strain rates. Although the pressure effects on the flow and fracture behavior of Zr41.25Ti13.75Cu12.5Ni10Be22.5 have been investigated by Lewandowski and Lowhaphandu et al. (1999, 2000a, 2000b, 2002) under quasi-static compression and tension testing, the information of the flow stress in shear at ultra-high strain rates, high normal stresses and high hydrostatic pressures are not clear yet. This is in part due to the availability of

mechanical testing equipment that can achieve such extreme states of stress on samples. In the present study, plate-impact pressure-shear experiments are utilized in order to investigate the behavior of BMG specimens under strain rates ~ 105 s-1, normal pressure up to 9 GPa and hydrostatic pressure of up to 6 GPa; such combined states of stress cannot be achieved by any other experimental configuration. The experiments involve
259

the skew impact of a thick WC flyer plate with a stationary target.

The target comprises

a front and a back WC plate with a thin BMG specimen sandwiched in between. Upon impact, pressure-shear loading is generated in the sandwiched BMG specimen. From

the measured normal and transverse components of the free surface particle velocity histories, the dependence of flow stress in shear at ultra-high strain rates, high normal stress and high hydrostatic pressures can be obtained for the Zr-based BMG specimens.

5.2 Experimental Procedure The material used in the present study was LM-1, supplied by Liquidmetal, Inc., Lake Forest, CA. In the first series of experiments, the material is in the form of rectangular plates with dimensions 90mm63mm5mm. Reproducibility of the samples was established through metallurgical characterization by Liquidmetal technologies and in our laboratory. Moreover, these plates were determined to be fully amorphous by

differential scanning calorimetry and X-ray diffraction in previous work (Lowhaphandu and Lewandowski, 1998). The as-received plate was then electrical discharge machined

(EDM) into smaller circular disks with a diameter of 25.4 mm and thickness of approximately 4.5 mm. The disks were then carefully lapped and polished to a surface

finish of 5 m and used as target plates in the first series of plate impact experiments. In the second series of experiments, the material is in the form of thin square plates with dimensions 38mm38mm0.7mm. The as-received plate was then carefully lapped and polished to a surface finish of 5 m and thickness of approximately 0.6 mm. Then the
260

thin BMG sample was sandwiched by a front and a rear WC plate into target assembly.

5.2.1 Experimental Configuration of Spall Experiments The plate-impact experiments were conducted using the 82.5 mm bore single-stage gas-gun facility in the Department of Mechanical and Aerospace Engineering at Case Western Reserve University. Figure 5.1 shows the schematic of the experimental

configuration used for normal plate impact and the combined pressure-shear plate impact spall experiments. For the case of the normal plate impact experiments the skew angle

of the flyer plate is zero, while the skew angle varies from 60 to 240 for the combined pressure-shear plate impact experiments. A fiberglass projectile carrying a Ti-6Al-4V flyer plate is accelerated down the gun barrel by means of compressed air. The rear end

of the projectile has a sealing O-ring and a Teflon key that slides in a key-way inside the gun barrel to prevent any rotation of the projectile. In order to reduce the possibility of

an air cushion between the flyer and target plates, impact takes place in a target chamber that has been evacuated to 50 m of Hg prior to impact. To ensure the generation of

plane waves with wave-front sufficiently parallel to the impact face, the flyer and the target plates are carefully aligned to be parallel to within 210-5 radians by using an optical alignment scheme developed by Kim et al. (1977). The actual tilt between the two plates is measured by recording the times at which four, isolated, voltage-biased pins, that are flush with the surface of the target plate, are shorted to ground. A laser-based

optical system, utilizing a UNIPHASE Helium-Neon 5mW laser (Model 1125p) and a
261

high frequency photo-diode, was used to measure the velocity of the projectile. A velocity interferometer was used to measure the history of the normal particle velocity at the rear surface of the target plate. The multi-beam VALYN VISAR was used as the interferometer system (Barker and Hollenbach, 1972). A COHERENT VERDI 5W

solid-state diode-pumped frequency doubled Nd:YVO4 CW laser with wavelength of 532 nm was used to provide a coherent monochromatic light source. Other details regarding

the design, execution and data analysis of the experiments can be found elsewhere (Prakash, 1995).

Figure 5.1: Schematic of the normal and combined pressure-shear plate-impact spall experiments.

5.2.2

Experimental

Configuration

of

High-strain-rate

Plate-impact

Pressure-shear Experiments The high-strain-rate plate-impact pressure-shear experiment involves the impact of a thick WC flyer plate mounted on a projectile with a stationary target.
262

The target

comprises a front and a back WC plate with a thin BMG specimen sandwiched in between. The corresponding experimental configuration is shown in Figure 5.2. The impacting

Impact takes place at an angle relative to the direction of approach.

flyer and target plates, are ground and lapped flat, and are aligned parallel to each other prior to impact. Upon impact, both longitudinal and shear waves are generated within

the flyer and target plates. This leads to pressure-shear loading of the sandwiched specimen. In all the tests the projectile velocity and the skew angle are controlled such

that the flyer, the front and rear target plates remain elastic during impact.

Figure 5.2: Schematic of the high-strain-rate plate-impact pressure-shear experiments. In the present study, a new laser interferometer, based on the multi-beam VALYNTM VISAR (Barker and Hollenbach, 1972), is used to measure the combined normal and transverse particle velocities at the free surface of the target plate.
263

The schematic of the

light path of the laser interferometer is shown in Figure 5.3.

Three VALYNTM VISAR

fiber-optic channels are simultaneously employed to obtain the normal and transverse components of the free surface of the target plate by collecting the 0 and 1 order diffraction beams from a holographic grating deposited on the rear surface of the target plate. In the preset experiments a holographic diffraction grating with a pitch of 1200 lines/mm is utilized. Following Chhabildas et al. (1979), the longitudinal, and the transverse components of the free-surface particle velocity, i.e. U ( t ) and V ( t ) , can be expressed in terms of the measured particle velocities along the beams, i.e. V ( t ) , and the diffraction angle,

VISAR , as

U (t ) =

V + (t ) + V (t )

(1 + cos VISAR )

(5.1)

V + (t ) V (t ) V (t ) = . sin VISAR

(5.2)

264

Figure 5.3: Schematic of the combined normal and transverse velocity interferometer.

5.2.3 Wave propagation in the Flyer and the Target Plates: The t-X (time versus distance) Diagram and the S-V (stress versus particle velocity) Diagram for Spall Experiments

A schematic of the t-X (time vs. distance) diagram, which illustrates the propagation of compression and tensile waves through the target and flyer plates during a typical plate-impact spall experiment, is shown in Figure 5.4. The abscissa represents the

distance in the target and the flyer plates from the impact surface, while the ordinate represents the time after impact. The arrows indicate the direction of wave propagation.

Upon impact, compressive stress-waves are generated in both the target and the flyer plates. The interaction of the corresponding release waves from the free surface of flyer and the target plates brings the shocked material into a state of tensile stress at a

265

pre-determined location within the BMG specimen.

If the tensile stress exceeds a

critical strength value for the material (spall strength), a tensile damage process is initiated and eventually spall (material separation) occurs if both the amplitude and the pulse duration of the tensile wave are sufficient. The occurrence of spall can be inferred from the free-surface particle velocity versus time profiles measured using the VALYNTM VISAR.

Figure 5.4: Wave Propagation in the flyer and the target plates (t-X diagram) for a typical plate-impact spall experiment.

266

Figure 5.5: Stress versus particle velocity (S-V) diagram for a typical plate-impact spall experiment. The dashed-line shows the hypothetical case for the no-spall condition.

The S-V (stress vs. particle velocity) diagram details the locus of all the stress and particle-velocity states that can be attained during a typical plate-impact experiment. It illustrates the stress and particle velocity in the various shock states, as depicted in Figure 5.5. The abscissa represents the particle velocity in the target and the flyer plates while

the ordinate represents the stress in the target and flyer. For the case in which the spall strength is greater than the applied tensile stress, the stress and particle velocity states in the BMG moves along the dashed lines from State 5 to the no-spall-state, denoted by State 7. However, if the tensile stress is greater than the spall strength of the BMG (spall indicated by the short dashed lines), the BMG will spall and the tensile stress in State (7) will unload to the stress-free state denoted by State (7). The compressive end of spall

wave from State (7) arrives at the free surface of the BMG, and brings the free surface velocity to State (10), which is the same as that in State (6) and also in State (7). The
267

free surface particle velocity in States 6, 7, and 10 are referred to as Vmax, and the corresponding free surface particle velocity in State 8 is denoted by Vmin. The spall strength, spall, can then be calculated from the measured free surface particle velocities Vmax and Vmin (Grady and Kipp, 1993), by using

spall = Z BMG (Vmax Vmin ) / 2 .

(5.3)

In Eq. (5.3) ZBMG represents the acoustic impedance of the BMG in the zero stress condition.

5.2.4 Wave Analysis for Propagation of both Longitudinal and Shear Waves in the Flyer and the Target Plates for High-strain-rate Plate-impact Pressure-shear Experiments

The time-distance diagram showing the propagation of both longitudinal and transverse waves in the flyer and the target plated is shown in Figure 5.6. Since the flyer, the front and rear target plates remain elastic during impact, the normal stress and the particle velocity in the BMG specimen (State 4) can be obtained by using the forward characteristic equations

4 (t ) ( C L ) u 4 = 5 ( C L ) u fs (t ) .

(5.4)

268

u4 =

1 u fs (t ) . 2

(5.5)

Using Equations (5.4) and (5.5), and noting that the rear surface of the target plate is traction free, i.e. 5 = 0 , the normal stress within the specimen can be expressed in terms of the measured normal component of the free surface particle velocity u fs (t ) and the longitudinal impedance of the WC target plates C L , as

4 =

1 ( C L ) u fs (t ) . 2

(5.6)

The shear stress and the transverse particle velocity in State 2 can be obtained by using the forward and backward characteristic equations

2 ( C S ) v2 = ( C S ) v1 .

(5.7)

2 + ( C S ) v2 = 0 .

(5.8)

Such that

2 =

1 ( Cs ) v1 and 2

1 v2 = V0 sin . 2

(5.9)

269

In Equation(5.9), and Cs are the density and shear wave speed of the WC target plates respectively. V0 and are the impact velocity and the skew angle of the impact.

Figure 5.6: Wave Propagation in the flyer and the target plates (t-X diagram) for a typical high-strain-rate plate-impact pressure-shear experiment.

In order to obtain the transverse particle velocity and shear stress in State 4, we note that the shear stress across the specimen is continuous while the transverse component of the particle velocity is discontinuous. Denoting the transverse velocity at the

+ front-plate/specimen interface by v4 and that at the specimen/rear-plate interface by v 4 ,

270

we can write the characteristic equations governing the state of stress and particle velocity in State 4 as

+ 4 ( C S ) v 4 = ( C S ) v fs (t ) .

(5.10)

+ + 4 + ( C S ) v 4 = 3 + ( C S ) v3 = 0 .

(5.11)

Which yields

4 =

1 1 ( Cs ) v fs and v4+ = v fs (t ) . 2 2

(5.12)

In Equation (5.12), v fs (t ) is the measured transverse component of the free surface particle velocity.

The stress and particle velocity states to the left of the specimen can be obtained from the following characteristic equations

4 ( C s ) v 4 = 2 ( C s ) v2 = ( C s ) v1 .

(5.13)

Combining Equations (5.12) and (5.13), it gives

271

 v4 (t ) =

1 1 1 ( 4 + C S v ) = v fs + v1 = V0 sin v fs . C S 2 2

(5.14)

In view of the kinematics for pressure-shear loading, the deformation occurring in the specimen during the pressure-shear loading is essentially simple-shear. The nominal shear rate in the specimen can be expressed as

(t ) =

+ v4 v4 (t )

h(t )

V0 sin v fs (t ) h(t )

(5.15)

Where h(t ) is the thickness of the thin BMG specimen.

Using Equation (5.15), the shear strain (t ) in the specimen can be obtained by integrating the shear strain-rate over the total time duration, i.e.

(t ) = dt .
0

t .

(5.16)

5.2.5 Calculation of Hydrostatic Pressure in the BMG Specimen for High-strain-rate Plate-impact Pressure-shear Experiments

In Equation (5.6), the normal stress within the specimen can be expressed in terms of the measured normal component of the free surface particle velocity u fs (t ) . Since the normal stress is also one of the principal stresses within the specimen (along the impact
272

direction), it can be rewritten as 11 . The other two principal stresses can be written as

22 and 33 respectively. It is also noted that 22 = 33 due to symmetry of plate

impact experiments.

Since plate-impact pressure-shear experiments generate fully plane strain conditions before the arrival of the release waves from the lateral boundary of the specimen,the lateral principal strain 22 should be equal to zero during the window time of experiments. Using Hookes law, the principal stress 22 within the thin BMG

specimen can be estimated by using the plane strain condition

22 =

22 v( 11 + 33 )
E

= (1 v) 22 v 11 = 0 .

(5.17)

Combining equations (5.6) and (5.17), we can express the principal stress 22 as

22 =

(1 )

11 =

1 ( CL ) u fs (t ) . 2 (1 )

(5.18)

So, the hydrostatic pressure within the BMG specimen can be expressed as

p=

11 + 22 + 33
3

(1 + ) ( CL ) u fs (t ) . 6(1 )

(5.19)

273

5.3 Experimental Results and Discussions for spall experiments

In the present study, the first series of experiments comprise four normal spall experiments and six pressure-shear spall experiments conducted on a Zr-based BMG, Zr41.25Ti13.75Ni10Cu12.5Be22.5. These experiments were designed to: (a) better understand the structure of shock waves in BMG subjected to planar shock compression, (b) estimate spall strength of the BMG following normal shock compression and combined shock-compression and shear loading, and (c) obtain the Hugoniot elastic limit (HEL) of the material. The four normal plate impact spall experiments are designated as

FY06011, FY06012, FY06013 and FY06014. Table 5.1 lists the key parameters for the four experiments -- it provides the Shot #, thickness of the BMG target plate, thickness of the Ti-6Al-4V flyer plate, impact velocity, and the shock-induced stress. The

shock-induced stresses were estimated from the measured impact velocity from the knowledge of the elastic longitudinal impedance of the BMG and the shock Hugoniot for the Ti-6Al-4V flyer plate. The dimensions of the flyer and target plates were chosen so as to avoid the arrival of the release waves from the lateral boundary at the monitoring point of the target plate during the time duration of interest. The impact velocities were chosen so as to span the elastic to the elastic-plastic range of the Zr-based BMG during impact. The BMG is expected to remain elastic at the lowest impact velocities of 328.9 m/s and 383.8 m/s, but show significant inelasticity at the higher impact velocities, i.e. 463.7 m/s and 541.8 m/s. Also, these impact velocities are expected to generate high enough tensile stresses so as to create spall (tensile) failure within the BMG at the four
274

impact velocities. Studying the residual spall strength in BMG at different levels of impact stresses is expected to provide a better understanding of the accumulation of damage within the BMG during planar shock compression. The six combined pressure and shear spall experiments are designated as FY06015, FY06016, FY06017, FY06018, FY06019 and FY06020. Table 5.2 lists the key parameters for the six experiments -- it

provides the Shot #, thickness of the BMG target plate, thickness of the Ti-6Al-4V flyer plate, impact velocity, skew angle, the shock-induced normal stress and shear strain. In order to study the effect of shear strain on the spall strength of the BMG, the normal stress for all six experiments is kept approximately constant - 5 GPa. The skew angle varies from 6o to 24o, and the corresponding shear strain varies from 0.90% to 3.18%. Table 5.3 lists the mechanical properties, such as the elastic wave speeds, of the Zr-based BMG material, Zr41.25Ti13.75Ni10Cu12.5Be22.5, used in the plate impact experiments conducted in the present study. Flyer Ti-6Al-4V (mm) 3.1 3.1 3.1 3.1 Target BMG (mm) 4.3 4.6 4.1 4.2 Impact Velocity (m/s) 328.9 383.8 463.7 541.8 Impact Stress (GPa) 4.39 5.13 6.06 7.06

Exp No. FY06014 FY06011 FY06012 FY06013

Table 5.1: Summary of the normal spall experiments on the Zr-based bulk metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5.

275

Exp No. FY06020 FY06015 FY06019 FY06018 FY06016 FY06017

Flyer Ti-6Al-4V (mm) 3.1 3.1 3.1 3.1 3.1 3.1

Target BMG (mm) 4.4 4.2 4.5 4.5 3.9 3.9

Impact Velocity (m/s) 411.8 399.7 383.9 389.1 404.8 391.1

Skew Angle (0) 6 12 15 18 20 24

Impact Stress (GPa) 5.44 5.06 4.87 4.81 4.95 4.83

Shear Strain (%) 0.90 1.66 2.03 2.40 2.76 3.18

Table 5.2: Summary of the pressure-shear spall experiments on the Zr-based bulk metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5.

Density

Longitudinal wave speed 5185 m/s

6000 Kg/m3

Shear wave speed 2464 m/s

Bulk wave speed 4335 m/s

Elastic Modulus 98.6 GPa

Shear Modulus 36.4 GPa

Bulk Modulus 113 GPa

Poissons ratio 0.354

Table 5.3: Physical properties of the Zr-based BMG (Vit-1) used in the present investigation. The data is taken from Lu (2002).

5.3.1 Structure of Shock Waves in BMG

Figure 5.7 shows the measured free-surface particle-velocity profiles for four normal impact spall experiments -- FY06011, FY06012, FY06013, and FY06014 -- conducted at impact velocities of 383.8 m/s, 463.7 m/s, 541.8 m/s, and 328.9 m/s, respectively. The corresponding shock-induced compressive stresses for the four experiments were 5.13, 6.06, 7.06 and 4.39 GPa. In the figure the abscissa represents the time after impact, while the ordinate provides the measured particle velocity at the free surface of the BMG target plate. The predicted arrival times of the elastic precursors for Shots FY06011,
276

FY06012, FY06013, and FY06014, based on the elastic longitudinal wave speed in the BMG and the thickness of the BMG target plates used in the four experiments, were 887 ns, 790.7 ns, 810 ns, and 830 ns, respectively, and are in close agreement with the arrival times shown in Figure 5.7. For the experiments conducted at the low impact velocities, i.e. experiments FY06011 and FY06014, upon impact the elastic precursor jumps to a level of 340 m/s and 289 m/s, respectively, and then remains nearly constant up to the arrival of the spall wave at the free surface of the target plate. At these stress levels the structure of the shock profiles are consistent with the nearly elastic response of the BMG target plates. It also indicates that for the Zr-based BMG, the stress level at the HEL (which represents the critical threshold for the onset of inelasticity in the material) is higher than 5.13 GPa. At the higher impact velocities, the measured free surface

particle velocity versus time profiles show a clear two-wave structure indicative of an elastic-plastic response --- a step-like elastic precursor to the HEL level, a peak at the elastic front followed by a steeply rising plastic wave, and then a gradual transition to a final plateau. It is interesting to note that the peak at the elastic precursor front becomes much sharper as the impact velocity is increased from 463.7 m/s (impact stress of 6.06 GPa) to 541.8 m/s (impact stress of 7.06 GPa). Also, behind the elastic precursor, the stress rise associated with the plastic wave to the final plateau is not steady. The stress rise is comparatively gradual at an impact velocity of 463.7 m/s, but becomes much steeper as the impact velocity is increased to 541.8 m/s. Such behavior in the particle velocity profiles is typical of highly time-resolved shock-wave measurements on
277

elastic-plastic solids with a limited strain hardening capability (Mashimo, 1998). The corresponding longitudinal stress profiles for Shots FY06012 and FY06013 are shown in Figure 5.7. From the stress profiles the stress at the Hugoniot Elastic Limit (HEL) for Zr41.25Ti13.75Ni10Cu12.5Be22.5 can be estimated to be HEL ~ 6.15 MPa . The HEL level and the clear two-wave structure observed in the present experiments are in sharp contrast to the results reported previously (Zhuang et al., 2002) for a similar Zr-based BMG composition, where a very low HEL was reported. However , the present HEL levels are a bit lower than 7.1 GPa measured by Turneaure et al. (2004) for Zr56.7Cu15.3Ni12.5Nb5.0Al10.0Y0.5, a material with a quasi-static strength of ~ 2.0 GPa.

600 550

Free Surface Velocity (m/s)

500 450 400 350 300 250 200 150 100 50 0 0 0.5

Shot FY06011 (impact velocity 383.8 m/s) Shot FY06012 (impact velocity 463.7 m/s) Shot FY06013 (impact velocity 541.8 m/s) Shot FY06014 (impact velocity 328.9 m/s)

1.5

2.5

Time after Impact (s)

Figure 5.6: Free-surface particle velocity versus time profiles for the four normal plate-impact spall experiments conducted in the present study.

278

10 9

Longitudinal Stress (GPa)

8 7 6 5 4 3 2 1 0 0 0.5

Shot FY 06012 Shot FY 06013

HEL= 6.15 GPa

1.5

Time after Impact (s)

Figure 5.7: Longitudinal stress versus time profiles for Shots FY06012 and FY 06013. The HEL for the Zr-based BMG is estimated to be ~ 6.15 GPa.

Using the von Mises yield criterion, the HEL (under uniaxial strain shock loading) can be related to the yield stress Y , under uniaxial stress loading (Fowles, 1961) as:

HEL = Y

1 . 1 2

(5.20)

Using an ambient value for the Poisonss ratio, , of 0.354 (Lu, 2002), a yield strength of 2.78 GPa is obtained. For comparison, dynamic yield strength data under uniaxial stress conditions on the same composition of Zr-based BMG yields Y ~ 1.6 1.8 GPa (Sunny

et al., 2005b), with a quasi-static strength in the range of 2.0 GPa (Lewandowski and
279

Lowhaphandu, 1998, 2002).

Such discrepancies in yield stress obtained under different loading conditions have been previously observed in other amorphous glasses. For example in pyrex glass, a yield strength of 1.5 GPa was obtained from uniaxial stress bar impact experiments (Espinosa et al., 1997). In contrast, plate impact experiments (uniaxial strain) on Pyrex provide an HEL of 7.5 GPa (Cagnoux and Longy, 1988), which gives a yield stress of 5.1 GPa from Eq. (5.20).

5.3.2

Calculation of Spall Strength in Shock Compressed Zr-Based BMG

Figure 5.8 shows the measured free-surface particle velocity profile and the corresponding t-X diagram for experiment FY06011. At time T1, , the free surface particle velocity increases to a level of Vmax at the arrival of the compression wave at the free surface of the BMG target plate, which is consistent with the Hugoniot State corresponding to the impact velocity used in the experiment. At time T2, the release waves from the back of the target plate and the flyer plate intersect at a pre-determined plane within the BMG target plate; the corresponding unloading tensile wave and the end of spall compressive wave propagate and arrive at the free surface of the BMG plate at times T3 and T4, respectively. At time T3, the particle-velocity at the free surface of the BMG plate starts to decrease, and at time T4, reaches a level of Vmin. Eventually the particle velocity recovers to its Hugoniot State level of Vmax.
280

This

decrease in the particle velocity is followed by a recovery, which is also referred to as the pull-back characteristic of the spall signal; the magnitude of the pull-back signal is used in the calculation of the materials spall strength, as detailed in the next section.

The calculation of the spall strength from the measured free surface particle velocity history requires the estimation of the magnitude of the pull-back signal from the measured particle velocity profile. The method applied for calculating the spall strength is illustrated in Figure 5.9. The free surface velocity data for experiment FY06011 is used as an example. The particle velocity increases to Vmax at the arrival of the first compression wave at the free surface of the target plate. After the spall event, the particle velocity drops to Vmin followed by a pull back to Vo. In general, in most spall experiments, V0 should equal Vmax; however in some experiments Vo has been observed to be smaller than Vmax, indicating the occurrence of partial spall or plasticity (Shazly, 2005). For the case in which there is no spall, the free surface particle velocity remains at its steady state level of Vno_spall. Thus, using the pull-back signal, the spall strength of the BMG can be estimated by using Equation (5.3).

281

Figure 5.8: Time-distance diagram paired with the measured free-surface particle velocity profile for experiment FY06011 to illustrate the pull-back spall signal.

500

Free Surface Particle Velocity (m/s)

450 400 350 300 250 200 150 100 50 0 0

Shot FY06011 Flyer: Ti-6Al-4V (3.1 mm) Target: BMG (4.6mm) Impact Velocity: 383.8 m/s
V max=325.9 m/s V 0=299.8 m/s

Spall signal V min=146.4 m/s

V no spall= 0 m/s
0.5 1 1.5 2 2.5 3

Time after Impact (s)

Figure 5.9: Free surface particle velocity data for experiment Shot FY06011.
282

5.3.3 Effect of of Shock Compression on the Spall Strength of the Zr-Based BMG

Figure 5.10 shows the spall strength data collected from the four normal plate impact spall experiments conducted in the present study on Zr41.25Ti13.75Ni10Cu12.5Be22.5 as well as the two normal plate impact experiments conducted by Zhuang et al. (2002) on Zr41.2Ti13.8Ni10Cu12.5Be22.5 (Vit-1) and its composite Zr56.3Ti13.8Ni5.6Cu6.9Be12.5 (-Vit). It is interesting to note that the spall strength of the Zr-based BMG decreases with increasing levels of shock compression. The measured spall strengths at shock

compression levels of 7.06 GPa, 6.06 GPa, 5.13 GPa and 4.39 GPa are 2.33 GPa, 2.35 GPa, 2.72 GPa, and 3.5 GPa, respectively. These levels of spall strength are in close

agreement to those measured by Zhuang et al. (2002) on Vit-1 and -Vit in the stress range of 5-7 GPa, as illustrated in Figure 5.10. Since, Zhuang et al. did not conduct any spall experiments in the shock-induced stress range < 5 GPa, they did not observe the decrease in spall strength reported presently with increasing levels of shock compression. The decrease in the spall strength with increasing levels of shock compression provides an indication of accumulation of damage within the BMG as the initial shock-induced compression wave propagates through the BMG.

Moreover, it must be mentioned that similar variations in spall strength in nominally brittle materials, e.g. sintered and hot pressed silicon carbide (SiC), subjected to planar
283

shock compression have been reported in the past (Bartkowski and Dandekar, 1996, Dandekar, 2004). The spall threshold of both the SiC materials increased when impact stress was increased from 1.6 GPa to 3.7 GPa; above 3.7 GPa, the spall threshold decreased with an increase in impact stress to 12 GPa; both materials appear to peak in spall threshold at an impact stress of about 3.7 GPa. For example, at an impact stress of 1.6 GPa, the Cercom SiC exhibits a spall threshold of 0.90 GPa. At the stress of 3.75 GPa, the spall threshold increases to 1.30 GPa but at an impact stress of 12.2 GPa, the spall threshold is only 0.90 GPa. This unusual trend in the spall strength of silicon carbides is understood to be due to the competing roles of (i) localized plasticity, (ii) generation and propagation of cracks taking into consideration their relative dominance below and above a given magnitude of stress. The initial increase in spall strength of silicon carbide, with an increase in shock-induced stress, is likely due to the dominance of localized plastic deformation over crack-dominated brittle deformation, while the corresponding decline in spall strength with an increase in shock-induced stress is attributed to the dominance of crack-induced brittle deformation over plastic deformation. Similar mechanical behavior was also observed in the work of Nathenson et al. (2005) on ASA-800 Si3N4, where the spall strength was observed to decrease with an increase in the shock-induced normal impact stress. The spall strength at stress levels below the HEL were found to decrease by almost 37% from its maximum measured value of 0.9 GPa as the impact velocity was increased from 65 m/s to 599 m/s. Moreover, as also observed for the BMG in the present investigation, the rate of drop in spall strength with the
284

increase in shock-induced stress was high in the beginning and then leveled off as the stress levels approach the HEL and lead to an accumulation of plastic flow. Besides ceramics, the above behavior has also been observed in S2-glass fiber reinforced polymer composites (GRP) (Tsai and Prakash, 2005). Under planar shock compression the spall strength of GRP was only measured to be between 30 ~ 60 MPa, and was observed to decrease with increasing levels of shock-induced impact stress. In particular, for

shock-induced stresses greater than 600 MPa the spall strength was essentially negligible, while no-spall was observed at stress levels below ~100 MPa.

It is noted that the Zr-based BMG has a very high spall strength when compared to other materials (ASA-800 Si3N4 only has a spall strength of 0.5~1.0 GPa under impact stress of 1.12 ~13.3 GPa (Nathenson et al., 2005); Gamma-Met PX only has a constant spall strength of 1.8 GPa under impact stress of 2.7 ~ 4.7 GPa (Shazly, 2005)). The main reason for the high spall strength of BMGs is understood to be due to their randomly-ordered atomic structures and the absence of defects in their microstructure.

285

4 3.5

Spall Strength (GPa)

3 2.5 2 1.5 1 0.5 0

Zr41.25Ti13.75Ni10Cu12.5Be22.5 (present study) Zr41.2Ti13.8Ni10Cu12.5Be22.5 ( Zhuang et al. 2002) Zr56.3Ti13.8Ni5.6Cu6.9Be12.5 ( Zhuang et al. 2002)
0 1 2 3 4 5 6 7 8

Impact Stress (GPa)

Figure 5.10: Spall strength as a function of impact Stress for three different Zr-based BMGs.
5.3.4 Effect of the Combined Shock Compression and Shear Loading on the Spall Strength of the Zr-Based BMG

Figure 5.11 shows the measured free-surface particle-velocity profiles for one normal impact spall experiment (FY06011) and six pressure-shear spall experiments (FY06015, FY06016, FY06017, FY06018, FY06019 and FY06020) conducted at impact velocities of 383.8 m/s, 399.7 m/s, 404.8 m/s, 391.1 m/s, 389.1 m/s, 383.9 m/s and 411.8 m/s respectively. The corresponding shock-induced stresses for these seven experiments were 5.13, 5.06, 4.95, 4.83, 4.81, 4.87 and 5.44 GPa. The skew angles for these seven experiments are 00, 120, 200, 240, 180, 150 and 60 respectively. Since these experiments are conducted at similar normal stress and different shear strain, the results can be used to
286

study the effect of shear strain on the spall strength of the Zr-based BMG.

For

experiments FY06018 and FY0619,, the free surface particle velocity profiles unload completely to their no-spall levels (zero velocity) at around 2.2 s after impact, indicating that the applied tensile state of stress is not enough to create spall in the BMG specimens for the two experiments, and the spall strengths under the two experimental conditions are greater than the applied tensile stress of 4.75 GPa. Figure 5.12 shows the spall strength versus. shear strain data collected from the seven experiments. With increasing levels of shear strain, the spall strength of the BMG is found to decrease initially, and then increase dramatically, and decrease again as the shear strain is increased from 2.40% to 3.18%. This unusual trend in the shear strain on the spall strength of the Zr-based BMG can be explained by the competing roles of localized plasticity and the generation and defects taking into consideration their relative dominance below and above a given magnitude of shear strains --- Dandekar explained a similar trend for the normal shock compression dependence of the spall strength on a sintered and hot pressed silicon carbide (SiC) (Bartkowski and Dandekar, 1996, Dandekar, 2004). The sudden increase in spall strength of silicon carbide, with an increase in shear strain (skew angle of impact), is likely due to the initiation of shear induced plastic deformation over crack-dominated brittle deformation during the planar shock compression loading, while the corresponding decline in spall strength with a further increase in shear strain is attributed to the subsequent growth of crack-dominated brittle deformation under normal shock compression and the subsequent shear loading.
287

500 450

Free Surface Velocity (m/s)

400 350 300 250 200 150 100 50 0 0

FY06011, skew angle 0, Imp. Vel. 383.8 m/s FY06015, skew angle 12, Imp. Vel. 399.7 m/s FY06016, skew angle 20, Imp. Vel. 404.8 m/s FY06017, skew angle 24, Imp. Vel. 391.1 m/s FY06018, skew angle 18, Imp. Vel. 389.1 m/s FY06019, skew angle 15, Imp. Vel. 383.9 m/s FY06020, skew angle 6, Imp. Vel. 411.8 m/s

V no spall =0 m/s
1 2 3

Time after Impact (s)

Figure 5.11: Free-surface particle velocity versus time profiles for the one normal spall experiments and five pressure-shear spall experiments with normal stress about 5 GPa.

5 4.5 4

Spall Strength (GPa)

3.5 3 2.5 2 1.5 1 0.5 0 0 1 2

20 0 No Spall 6 12 24

Shear Strain (% )

Figure 5.12: Spall strength as a function of shear strain for the six experiments conducted in the present study. The normal stress is maintained at ~ 5 GPa in the six experiments.
288

The effects of shear loading on the spall strength of other engineering materials have been investigated before. In the work of Nathenson (2006) on ASA-800 Si3N4, the presence of shear strain following normal shock compression resulted in a severe degradation of the spall strength. In particular, the spall strength was reduced to

essentially zero at a normal stress of 6.93 GPa and a shear strain of approximately 0.4%. In contrast, Si3N4 has a spall strength of ~ 600 MPa at a normal stress of 13.3 GPa, but at zero shear-strain. Besides Si3N4, similar behavior has also been observed in S2-glass and E-glass fiber reinforced polymer composites, as discussed in detail in Chapter 4.

5.3.5 Stress-strain Curves for Zr-based BMG under Shock Compression

To obtain the stress versus strain response of the Zr-based BMG under uniaxial strain compression, we consider the elastic-plastic wave profiles generated upon impact, as shown in Figure 5.13. In this figure, the x-t (distance versus time) diagram for a target with a thickness h, is shown. The elastic precursor arrives at the back of the target plate with a wave speed of CL , followed by plastic wave with speed a , the speed of which depends on the plastic response of the material. Based on elementary wave mechanics, the speed of plastic wave propagation, a , can be expressed as

a =

d / d

(5.21)

289

where d / d is the slope of the stress versus strain curve at a particular level of stress, and 0 is the material density in the reference configuration.

Furthermore, from geometrical considerations (refer to Figure 9), a can be expressed as

a =

h . t ( ) + h / CL

(22)

Moreover, based on the method of the characteristics, the following stress versus particle velocity relationship is expected to hold within the BMG target plate

0 a v p = const

(5.23)

In Eq. (5.23), v p is the particle velocity in the shocked region in the target plate, and can be related to the measured free-surface particle-velocity v fs , in the target plate by

1 v p = v fs . 2

(5.24)

290

Figure 5.13: Time distance diagram for elastic-plastic impact showing the plastic wave fan.

Using Eqs. (5.23) and (5.24), and substituting for v p yields

1 d (t ) = m 0 a dv fs . 2

(5.25)

Rewriting Eq. (5.21), in view of Eq. (5.25) yields,

d =

1 d . 2 0 a

(5.26)

Figure 5.14 shows the longitudinal stress-strain curves for the four normal impact experiments conducted in the present study. From the figure, the HEL can be estimated

291

to be ~ 6.15 GPa. A larger region of non-linearity beyond the HEL can be observed for the higher impact velocity experiments, i.e. Shots FY06012 and FY06013. The

dynamic elastic modulus of the material is estimated to be ~ 150 GPa, which is significantly (~ 50%) higher than that reported under quasi-static deformation (Table 5.3). Figure 5.15 and 5.16 show the strain-rate and strain histories, the normal stress and strain rate vs. normal strain for Shot FY06012 obtained by using the corresponding measured free surface particle velocity profile and Eq. (5.26), respectively. As shown in the figures, the total strain-rate in the sample (shown by the red curve) is as high as 5106 s-1; however, the duration for this high strain rate is short and coincides with the rise time of the shock front, after which the strain-rate becomes negligibly small. accumulation of strain within the specimen is ~ 4%. The total

292

Normal Stress (GPa)

Shot FY06011 Shot FY06012 Shot FY06013 Shot FY06014

Normal Strain (%)

Figure 5.14: Normal (Longitudinal) stress vs. normal (longitudinal) strain curves for the four normal plate impact experiments.

Strain Rate Longitudinal Strain

4

Strain Rate (10 /s)

4 3 3 2 2

Shot FY06012
1

Impact Velocity: 463.7 m/s

0 0 0.5 1 1.5 2

Time after Impact (s)

Figure 5.15: Strain and strain-rate histories in the Zr-based BMG for Shot FY06012.

293

Normal Strain (%)

Normal Stress (GPa)

Normal stress

6 5

Strain rate

3 2

1 0 1 2 3 4 5 0

Normal Strain (%)

Figure 5.16: Normal stress and strain rate vs. normal strain in the Zr-based BMG for Shot FY06012.
5.3.6 Scanning Electron Microscopy of the Spall/Fracture Surfaces

In order to examine the micro-structural effects of normal stress and shear strain on the material, fragments with a face that is a surface of the spall plane were examined under the scanning electron microscope (SEM).

To observe the effect of normal stress on the spall strength, two normal plate-impact spall experiments FY06012 and FY06014 were chosen. The experiment FY06012 was conducted with a normal stress below HEL, while the experiment FY0614 was conducted with a normal stress above HEL. Figure 5.17 shows the SEM pictures for the spall surface of these two experiments. From the fracture surface, the extensive veining is
294

Strain Rate (10 /s)

Shot FY06012 Flyer: Ti-6Al-4V (3.1 mm) Target: BMG (4.1mm) Impact Velocity: 463.7 m/s

10 9 8

present, providing evidence of the failure of the specimen in shear locally. The density of veining is very low when normal stress is below HEL, indicating very low plasticity and reflecting almost macroscopic elastic behavior for the structure of shock wave. The density of veining is much higher when normal stress is above HEL (more number of veining is seen with even higher magnification), indicating onset of plasticity and reflecting a two-wave structure of shock wave. So, the decrease in the spall strength with increasing levels of normal stress provides an indication of accumulation of damage within the BMG as the initial shock-induced compression wave propagates through the BMG. Moreover, as also observed for the BMG in the present investigation, the rate of drop in spall strength with the increase in shock-induced normal stress was high in the beginning and then leveled off as the stress levels approach the HEL and lead to an accumulation of plastic flow as shown in the SEM pictures (Figure 17).

295

Shot FY06012 (Normal stress below HEL)

Shot FY06014 (Normal stress above HEL)

Figure 5.17: Scanning electron microscope pictures of spall surfaces for two normal plate-impact spall experiments. In order to observe the effect of shear strain on the spall strength, four pressure-shear plate-impact spall experiments FY06015, FY06018, FY06016 and FY06017 were chosen. These four experiments were conducted under similar normal stresses, about 5 GPa. The experiments FY06015, FY06018, FY06016 and FY06017 were conducted with a skew angle of 120180200and 240 respectively. The corresponding shear strain is 1.66%, 2.40%, 2.76% and 3.18% respectively. Figure 5.18 shows the SEM pictures of the spall surfaces from these four experiments. From these micrographs it can be seen that in the case of the combined pressure and shear plate impact spall experiments, even when the normal stress is below HEL, a much higher density of veining (when compared to that observed in experiment FY06012 (normal spall, below HEL)) is obtianed. Besides higher density of veining compared to normal spall experiments, experiments FY06016 and FY06018 also show some larger sized vein structures (higher roughness)

296

in the SEM micrographs of the fracture/spall plane when compared to the other pressure-shear spall experiments, which might provide a pausible reason for the sudden increase of spall strength in these two experiments.

Shot FY06015 (skew angle 120)

Shot FY06018 (skew angle 180)

Shot FY06016 (skew angle 200)

Shot FY06017 (skew angle 240)

Figure 5.18: Scanning electron microscope pictures of spall surfaces for four pressure-shear plate-impact spall experiments.

297

5.4 Experimental Results and Discussions for High-strain-rate Plate-impact Pressure-shear Experiments

The motivation for this series of experiments is to obtain information on the shearing strength of the Zr-based BMG as a function of shearing rates, normal stress and superimposed pressure. In view of this a second series of experiments was conducted that involves the skew impact of a thick WC flyer plate with a stationary target. The target comprises a front and a back WC plate with a thin BMG specimen sandwiched in between. Upon impact, a pressure-shear loading is generated in the sandwiched BMG specimen. From the measured normal and transverse components of the free surface particle velocity histories, the dependence of flow stress on strain rate, normal stress and the normal pressure can be obtained for the Zr-based BMG specimens. Table 5.4

summarizes the key parameters for the three experiments conducted in this series of experiments -- it provides the Shot #, thickness of the thin BMG sample, thickness of the WC flyer plate, thickness of the front and rear target plates, impact velocity and the skew angle.

Exp No. Exp. 1 Exp. 2 Exp. 3

BMG sample (mm) 0.58 0.60 0.44

Flyer WC (mm) 6.59 6.59 6.55

Front WC Target Plates (mm) 3.65 3.60 3.61

Rear WC Target Plates (mm) 4.83 4.83 4.81

Impact Velocity (m/s) 189.9 178.4 143.4

Skew Angle (0) 22 15 22

Table 5.4: Summary of high-strain-rate plate-impact pressure-shear experiments on the Zr-based bulk metallic glass, Zr41.25Ti13.75Ni10Cu12.5Be22.5.
298

Figures 5.19 to 5.21 summarize the results for Shot Exp. 1. Figure 5.19 shows the history of the longitudinal and transverse components of the free surface particle velocities measured using the VISAR probes. Figure 5.20 shows the corresponding history of the normal stress, shear stress and the shear strain rate within the specimen, obtained by using the theory and equations in the section 5.2.4. The abscissa shows the time after the arrival of the longitudinal wave at the specimen plane, while the ordinate represents the magnitude of the stresses and shear strain rate in the specimen. Upon arrival of the longitudinal wave

at the specimen plane, the normal stress in the specimen builds up to a level of approximately 8.85 GPa in about 0.85 seconds, the corresponding hydrostatic pressure is 6.18 GPa. The details of the pressure build-up are shown as a sequence of steps in the normal stress profile before it reaches a plateau. Since the speed of transverse wave in the front target plate is smaller than the longitudinal wave speed, the shear wave arrives at the specimen after approximately 0.9 s of the arrival of the longitudinal pulse. On the arrival of the shear wave at the specimen plane, the thin sandwiched BMG specimen undergoes a combined pressure and shear deformation. The shearing rate during this pressure-shear deformation is ~

0.75 105 s 1 . Figure 21 shows the shear stress vs. shear strain response of the specimen. It
is important to note that by the time the shear-wave arrives at the specimen plane a state of essentially uniform normal stress has been attained in the specimen. The incident shear wave reverberates a few times within the specimen after which the specimen is sheared at a near constant shearing rate. The flow stress is noted to be about 1.02 GPa in shear.
299

200 180

160 140 120 100 80 60 40 20 0 0

40

Normal velocity

35 30 25 20

Transverse velocity

15 10 5

Time after impact (s)

Figure 5.19: Longitudinal and transverse free surface velocities for Shot Exp. 1.

10 9

1.5

Normal stress

6 5 4 3 2 1 0 0 0.5 1

Shear strain rate Shot Exp. 1 Impact Velocity: 189.9 m/s Skew Angle: 22 Specimen: Zr-based BMG (LM-1) Shear stress

0.5

0 1.5

Time after arrival of longitudinal wave (s)

Figure 5.20: History of normal stress, the shear stress and the shear strain rate in the specimen for Shot Exp. 1.

300

Shear strain rate (X 10 s )

Stress (GPa)

-1

Free surface transverse velocity (m/s)

Free surface normal velocity (m/s)

Shot Exp. 1 Impact Velocity: 189.9 m/s Skew Angle: 22 Specimen: Zr-based BMG (LM-1)

50 45

1.5

Shot Exp. 1 Impact Velocity: 189.9 m/s Skew Angle: 22 Specimen: Zr-based BMG (LM-1)

Shear stress (GPa)

0.5

Shear Strain (%)

Figure 5.21: Dynamic shear stress versus shear strain for Shot Exp. 1.

The results of the high-strain-rate pressure-shear plate-impact experiment for Experiment 2 are shown in Figures 5.22 to 5.24. Experiment 2 was conducted with a smaller skew angle, i.e. 150, when compared to 220 for Experiment 1. As a result, the shearing rate in the specimen is much smaller, (~ 0.32 105 s 1 ) when compared to that obtained in Experiment 1. Figure 5.22 shows the history of longitudinal and transverse free surface velocities measured using the VISAR probes. Figure 5.23 shows the history of normal stress, shear stress and the shear strain rate within the specimen by using the theory and equations in the section 5.2.4. The abscissa shows the time after the arrival of the longitudinal wave at the specimen plane, while the ordinate represents the magnitude of the stresses and the shear strain rate in the specimen. Upon arrival of the normal stress pulse, the normal stress in the
301

specimen builds up to a level of approximately 8.36 GPa in about 0.81 seconds, the corresponding hydrostatic pressure calculated by Section 5.2.5 is 5.84 GPa. The details of the normal stress build-up are shown as a sequence of steps in the normal stress profile before it reaches a plateau. Since the speed of transverse wave in the front target plate is smaller than the longitudinal wave speed, the shear wave arrives at the specimen after approximately 0.9 s of the arrival of the longitudinal pulse. At the arrival of the shear wave the specimen undergoes combined pressure-shear deformation. Figure 24 shows the shear stress vs. shear strain response of the sandwiched BMG specimen. It is important to

note that by the time the shear-wave arrives at the specimen plane a state of essentially uniform normal stress has been attained in the specimen. The shear wave reverberates a few times within the specimen after which the specimen is sheared at a constant shearing rate. The flow shear stress is noted to be about 0.96 GPa for Experiment 2.

302

200 180

160 140 120 100 80 60 40 20 0 0

40 35 30 25 20

Normal velocity

Transverse velocity

15 10 5

Time after impact (s)

Figure 5.22: Longitudinal and transverse free surface velocities for Shot Exp. 2.

10 9

1.5

6 5 4 3 2 1 0 0 0.5

Shear strain rate Shot Exp. 2 Impact Velocity: 178.4 m/s Skew Angle: 15 Specimen: Zr-based BMG (LM-1) Shear stress
1 0 1.5 0.5

Time after arrival of longitudinal wave (s)

Figure 5.23: History of normal stress, the shear stress and the shear strain rate in the specimen for Shot Exp. 2.

303

Shear strain rate (X 10 s )

Stress (GPa)

-1

Normal stress

Free surface transverse velocity (m/s)

Free surface normal velocity (m/s)

Shot Exp. 2 Impact Velocity: 178.4 m/s Skew Angle: 15 Specimen: Zr-based BMG (LM-1)

50 45

1.5

Shot Exp. 2 Impact Velocity: 178.4 m/s Skew Angle: 15 Specimen: Zr-based BMG (LM-1)

Shear stress (GPa)

0.5

Shear Strain (%)

Figure 5.24: Dynamic shear stress versus shear strain for Shot Exp. 2.

The results of the high-strain-rate pressure-shear plate-impact experiment for Experiment 3 are shown in Figures 5.25 to 5.27. Experiment 3 was conducted at a skew angle of 220 and at a lower impact velocity, 143.4 m/s, when compared to Experiments 1 and 2. As a result, the normal stress is smaller in this experiment, which is about 6.92 GPa. Figure 5.25 shows the history of longitudinal and transverse free surface particle velocitiy profiles measured using the VISAR probes. Figure 5.26 shows the history of normal stress, shear stress and the shear strain rate within the specimen by using the theory and equations in the section 5.2.4. The abscissa shows the time after the arrival of the longitudinal wave at the specimen plane, while the ordinate represents the magnitude of the stresses and shear strain rate in the specimen. Upon the arrival of the pressure pulse at the specimen plane, normal
304

stress in the specimen builds up to a level of approximately 6.92 GPa in about 0.79 seconds, the corresponding hydrostatic pressure calculated by Section 5.2.5 is 4.83 GPa. Again, the details of the pressure build-up are shown as a sequence of steps in the normal stress profile before it reaches a plateau. The shear wave arrives at the specimen after approximately 0.83 s of the arrival of the longitudinal pulse. At the arrival of the shear wave the specimen undergoes a combined pressure-shear deformation. Figure 27 shows the shear stress vs. shear strain response of the specimen. It is important to note that by the time the shear-wave arrives at the specimen plane a state of essentially uniform normal stress is attained in the specimen. The shear wave reverberates a few times within the specimen after which the specimen is sheared at a near constant shearing rate of ~ 0.65 105 s 1 . The average flow stress in the specimen is noted to be about 0.94 GPa.

305

200 180

160 140 120 100 80 60 40 20 0 0

40 35 30 25 20

Normal velocity

Transverse 15 velocity
10 5 1 2 3 0

Time after impact (s)

Figure 5.25: Longitudinal and transverse free surface velocities for Shot Exp. 3.

10 9 8 7

6 5 4 3 2 1 0 0

Normal stress

Shear strain rate

0.5

Shear stress
0.5 1 0 1.5

Time after arrival of longitudinal wave (s)

Figure 5.26: History of normal stress, the shear stress and the shear strain rate in the specimen for Shot Exp. 3.
306

Shear strain rate (X 10 5 s-1)

Shot Exp. 3 Impact Velocity: 143.4 m/s Skew Angle: 22 Specimen: Zr-based BMG (LM-1)

1.5

Stress (GPa)

Free surface transverse velocity (m/s)

Free surface normal velocity (m/s)

Shot Exp. 3 Impact Velocity: 143.4 m/s Skew Angle: 22 Specimen: Zr-based BMG (LM-1)

50 45

1.5

Shot Exp. 3 Impact Velocity: 143.4 m/s Skew Angle: 22 Specimen: Zr-based BMG (LM-1)

Shear stress (GPa)

0.5

Shear Strain (%)

Figure 5.27: Dynamic shear stress versus shear strain for Shot Exp. 3.

With regard to amorphous metals, observations of shear band angles deviant from 450 (Donowan, 1989) have prompted proposals of a pressure-modified Mohr-Coulomb yield criterion of the form c = k0 n

(1 + 2 + 3 ) . 3

Where c is the shear

stress on the slip plane at yielding, n is the stress component in the direction normal to the slip plane, k0 is the yield stress in shear, and are constants. In

Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002), the flow and fracture behaviors of the same bulk amorphous alloy, Zr41.25Ti13.75Ni10Cu12.5Be22.5, have been determined in tension and compression at room temperature with levels of superimposed hydrostatic pressure ranging from 0.1 to 1500 MPa. Also, torsion

specimens were tested at atmospheric pressure (i.e. 0.1 MPa) in order to approach
307

conditions of pure shear. The results of these experiments clearly indicated that for the Zr41.25Ti13.75Ni10Cu12.5Be22.5 BMG, the parameter , that characterizes the pressure sensitivity term, is negligible over the pressure range used in the experiments, suggesting a Mohr-Coulomb criterion of the form: c = k0 n . Figure 5.28 shows the all shear yield stress vs. normal pressure data from Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002). The calculated values for k0 =950 MPa is close to the value of flow stress obtained from a typical torsion test (i.e. 1081 MPa) and probably represents an upper bound owing to the stress and strain gradient present in the solid specimen. On

the other hand, the constant is determined to be 0.38, as shown in Figure 5.28. The normal stress dependence on the flow stress was also demonstrated by the macroscopic orientation of the fracture plane. The tensile fracture surfaces were oriented at 50-590, while compression fracture surfaces were oriented at 400.

308

1200

1000

800

c (MPa)

600

c = 0 - n-m 0 = 950 MPa = 0.038, =0

400

LM-1 (Lewandowski et al., 2002) Linear fit of Lewandowski's data

200

0 -1500

-1000

-500

500

1000

1500

n (MPa)

Figure 5.28: Shear yielding stress vs. normal pressure for data obtained from Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002).

Figure 5.29 shows the yield stress in shear vs. the normal pressure data on Zr41.25Ti13.75Ni10Cu12.5Be22.5 obtained from the pressure shear plate impact experiments in the present study along with the data from (Lewandowski, and Lowhaphandu, 2002). The superimposed hydrostatic pressure m for Lewandowskis work varies from 0.1 to 1500 MPa, while the hydrostatic pressure in the plate-impact pressure-shear experiments conducted in the present study are in the range 4.8 ~6.2 GPa. The results indicate that the response of the Zr-based BMG Zr41.25Ti13.75Ni10Cu12.5Be22.5 is nearly independent of both the hydrostatic pressure and the normal stress (i.e. = = 0 in compression) over a wide range of hydrostatic pressures.( 0.1 6.2 GPa) and normal stress (0.78 8.8

309

GPa) used in the experiments. The results also suggest that the Zr-based BMG used in the present study may have a different sensitivity to the normal stress in compression and tension, i.e. compression tension , although they are both very small.

1200

1000

c = 0 - n-m = 0, =0

800

c (MPa)

c = 0 - n-m 0 = 950 MPa = 0.038, =0

600

400

200

LM-1 (Lewandowski et al., 2002) Linear fit of Lewandowski's data LM-1 (Present study) Linear fit of all compression data
-8000 -6000 -4000 -2000 0

0 -10000

n (MPa)

Figure 5.29 Shear yielding stress vs. normal pressure for data obtained from Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002) and present study.

It is not yet clear why the magnitude of is small in the currently tested amorphous metal. The structural units which control the flow in BMG are probably at the level of atomic dimensions. The highest value for the flow stress of the BMG in shear under quasi-static loading conditions has been reported to be 1081MPa; the ratio of the flow stress in shear to the shear modulus (G =33 GPa) gives G / =30, which is close to

310

the model calculations of Maeda and Takeuchi (1978) of G / =20, performed on a static two-dimensional array of several hundred atoms arranged randomly. In that

model, yield occurred along the line of maximum shear via the collapse of holes in the structure. The present results appear consistent with a concept of flow that requires a certain number of atoms to move cooperatively as the structural units, similar to that in granular solids, although on a much finer size scale. Early dislocation models of

plasticity in amorphous metals (Li and Wu, 1976) predicted plasticity to occur on planes of maximum resolved shear stress, no normal stress dependence would be predicted under such a model. On the other hand flow mechanisms based on free volume and dilatation predict (Spaepen,1977; Argon, 1979; and Steif, 1983) a significant pressure dependence to the flow behavior, which are in contrast to the results obtained in the present study up to pressures of 7.6 GPa. However, it is important to emphasize that the material fails close to its theoretical strength.

Unlike the negligible hydrostatic pressure sensitivity and small normal stress dependency of flow stress of the Zr-based BMG studied in the present investigation, many polymeric systems exhibit very high hydrostatic pressure sensitivity in addition to normal stress dependence on their flow and fracture behavior (Bowden and Jukes, 1972; Parry and Wronski, 1985 and Mehta, 2003). The main reason for such a behavior in the polymeric systems can be attributed to their highly compressible nature, while amorphous metals are essentially incompressible.
311

5.5 Summary

In the present study the first series of experiments are conducted to study the spall strength of the Zr-based BMG under shock compression and combined shock compression and shear loading. The bulk metallic glass samples were shock loaded by utilizing Ti-6Al-4V flyer plates to impact stresses around 7 GPa. Normal plate impact and combined pressure and shear plate impact experiments with skew angles ranging from 60 to 240, were employed to study the effects of normal compression and combined compression and shear on the spall strength of the Zr-based BMG.. Under normal shock compression, the measured particle velocity versus time profiles show a step-like elastic precursor to the HEL level, a peak at the elastic front, followed by a steeply rising plastic wave and a gradual transition to a final plateau at the higher impact velocities. The peak at the elastic precursor front becomes much sharper as the impact velocity is increased from 463.7 m/s (impact stress of 6.06 GPa) to 541.8 m/s (impact stress of 7.06 GPa). Also, behind the elastic precursor, the rise associated with the plastic wave to the final plateau is not steady. The stress rise is relatively gradual at an impact speed of 463.7 m/s and becomes much steeper as the impact velocity is increased to 541.8 m/s. The spall strength of BMG was found to decrease with increasing normal stress. The

decrease in spall strength can be attributed to damage accumulation as the impact velocity is increased. The spall strength at normal stress 4.4 GPa was 3.5 GPa, while the spall strengths at normal stresses of 5.1, 6.0 and 7.0 GPa were 2.72, 2.35 and 2.33
312

GPa, respectively. The HEL of Zr41.25Ti13.75Ni10Cu12.5Be22.5 was found out to be 6.15 GPa. In the pressure-shear plate-impact spall experiments, with increasing levels of shear strain, the spall strength of the BMG was found to decrease initially, and then increase dramatically, and decrease again as the shear strain is increased from 2.40% to 3.18%. This unusual trend in the shear strain on the spall strength of the Zr-based BMG can be explained by the competing roles of (i) localized plasticity, (ii) generation and propagation of cracks taking into consideration their relative dominance below and above a given magnitude of shear strain.

The second series of experiment involve the skew impact of a thick WC flyer plate with a stationary target. The target comprises a front and a back WC plate with a thin BMG specimen sandwiched in between. This leads to pressure-shear shock loading of the sandwiched BMG specimen. From the measured free surface particle velocities, the flow stress under ultrahigh strain rate (~105 /s) and normal pressure (~9 GPa) is obtained for the Zr-based BMG. The shear yielding stresses were found out to be around 1 GPa at a a normal pressure ~ 9 GPa and a hydrostatic pressure ~6.0 GPa. Combining the

present results with the results in Lewandowski and Lowhaphandu,(2002), suggests that hydrostatic pressure has a negligible influence on the flow stress of the BMG in shear. Moreover, the applied normal stress has a negligible influence ( = 0 ) on the shear yield stress in compression.. This result along with the results of Lewandowski and

313

Lowhaphandu (2002), which showed that normal pressure has a small influence ( = 0.038 ) as we move from compression to the tension state, suggests that the Zr-based BMG tested in the present study has a different normal stress sensitivity in compression and tension, although they are both very small.

314

REFERENCES

Argon, A. S., 1979. Plastic deformation in metallic glasses. Acta Metalli. 27 (n1), 47-58. Bach, J., Krueger, B., and Fultz, B., 1991. Shock wave consolidation of a Ni-Cr-Si-B metallic-glass powder. Materials Letters 11, 383-388. Barker, L. M., and Hollenbach, R. E., 1972. Laser interferometer for measuring high velocities of any reflecting surface. Journal Applied Physics 43, 4669-4675. Bartkowski, P. T., and Dandekar, D. P., 1996. Spall strengths of sintered and hot pressed silicon carbide In: Schmidt, S. C., and Tao, W. C., (Eds.), Shock Compression of Condensed Matter-1995. American Institute of Physics, New York, pp. 535-539. Bruck, H. A., Rosakis, A. J., and Johnson, W. L., 1996. The dynamic compressive behavior of beryllium bearing bulk metallic glasses. Journal of Materials Research 11, 503-511. Bowden, P. B., Jukes, J. A., 1972. Plastic flow of isotropic polymers. Journal of Materials Science 7 (n1), 52-63. Cagnoux, J., and Longy, F., 1988. Spallation and shock-wave behavior of some ceramics. Journal de Physique Colloque 49, 3-10. Chhabildas, L. C., Sutherland, H. J., and Asay, J. R., 1979. A velocity interferometer technique to determine shear-wave particle velocity in shock-loaded solids. Journal of applied physics 50 (8), 5196-5201. Connor, R. D., Dandliker, R. B. and Johnson, W. L., 1998. Mechanical properties of tungsten and steel reinforced Zr41.25Ti13.75Cu12.5Ni10Be22.5 bulk metallic glass composites. Acta Mat. 46 (17), 6089-6102. Conner, R. D., Dandliker, R. B., Scruggs, V., and Johnson, W. L., 2000. Dynamic deformation behavior of tungsten-fiber/metallic-glass matrix composites. International Journal of Impact Engineering 24, 435-444. Dandekar, D. P., 2004. Spall strength of silicon carbide under normal and simultaneous compression-shear shock wave loading. International Journal of Applied Ceramic Technology 1, 261-268. Davis, L. A. and Kavesh, S., 1975. Deformation and fracture of an amorphous metallic alloy at high pressure. J. Mat. Sci.10 (3), 453-459. Donovan, P.E., 1988. A yield criterion for Pd40Ni40P20 metallic glass. Acta Met. 37, 445-456. Espinosa, H. D., Xu, Y., and Brar, N. S., 1997. Micromechanics of failure waves in glass:
315

I, experiments. Journal of American Ceramic Society 80, 2061-2073. Fowles, G. R., 1961. Shock wave compression of hardened and annealed 2024 Aluminum. Journal of Applied Physics 32, 1475-1487. Grady, D. E., and Kipp, M. E., 1993. High-Pressure Shock Compression of Solids. Springer-Verlag, Berlin, Germany. Hays, C. C., Kim, C. P., and Johnson, W. L., 2000. Enhanced plasticity of bulk metallic glasses containing ductile phase dendrite dispersions. Metastable, Mechanically Alloyed and Nanocrystalline Materials-- Parts 1 and 2, vol. 343-3, pp. 191-196. Hufnagel, T. C., Jiao, T., Li, Y., Xing, L. Q., and Ramesh, K. T., 2002. Deformation and failure of Zr57Ti5Cu20Ni8Al10 bulk metallic glass under quasi-static and dynamic compression. Journal of Materials Research 17, 1441-1445. Inoue, A., and Hashimoto, K., 2001. Amorphous and Nanocrystalline Materials: Preparation, Properties and Applications. Springer-Verlag, Berlin, Germany. Johnson, W. L., 1999. Bulk Metallic Glasses. In: Johnson, W. L., Inoue, A., and Liu, C. T., (Eds.), MRS Symposium Proceedings, vol. 554. Materials Research Society, pp. 311-339. Kim, K. S., Clifton, R. J., and Kumar, P., 1977. A combined normal and transverse displacement interferometer with an application to impact of Y-cut Quartz. Journal of Applied Physics 48, 4132-4139. Lewandowski, J. J., and Lowhaphandu, P., 1998. Effects of hydrostatic pressure on mechanical behavior and deformation processing of materials. International material reviews 43, 145-187. Lewandowski, J. J., and Lowhaphandu, P., 2002. Effects of hydrostatic pressure on the flow and fracture of a bulk amorphous metal. Philosophical magazine A82, 3427-3441. Lewandowski, J. J. and Greer, A. L., 2006. Temperature rise at shear bands in metallic glasses. Nat. Mat. 5, 15-18. Lewandowski, J. J., Shazly, M. and Nouri, A. S., 2006. Intrinsic and extrinsic toughening of metallic glasses. Scripta Materialia 54 (3), 337-341. Li, J. C. and Wu, J. B. C., 1976. Pressure and normal stress effects in shear yielding. Journal of Materials Science 11 (n3), 445-457. Lowhaphandu, P., and Lewandowski, J. J., 1998. Fracture toughness and notched toughness of bulk amorphous alloy: Zr-Ti-Ni-Cu-Be. Scripta Materialia 38, 1811-1817. Lowhaphandu, P., Montgomery, S. L. and Lewandowski, J. J., 1999. Effects of superimposed hydrostatic pressure on flow and fracture of a Zr-Ti-Ni-Cu-Be bulk amorphous alloy. Scr. Mat. 41 (1), 19-24.
316

Lowhaphandu, P. et al., 2000a. Deformation and fracture toughness of a bulk amorphous Zr-Ti-Ni-Cu-Be alloy). Intermetallics 8, 487-492. Lowhaphandu, P., 2000b. Mechanical Behavior of a Zirconium-based Bulk Metallic Glass. Ph. D. Thesis, Case Western Reserve University, Cleveland, OH . Lu, J., 2002. Mechanical Behavior of a Bulk Metallic Glass and its Composites Over a Wide Range of Strain Rates and Temperatures, (Ph.D. Dissertation), California Institute of Technology, Pasadena, CA. Maeda, K., Nakagawa, K., and Takeuchi, S., 1978. Activation analysis of plastic deformation of CdTe below room temperature. Phys Status Solidi (a) 48 (n2), 587-591. Mashimo, T., 1998. Effect of shock compression on ceramic materials In: Davison, L., and Shahinpoor, H., (Eds.), High-Pressure Shock Compression of Solids III. Springer-Verlag NewYork, Inc., New York, NY, pp. 101-146. Mashimo, T., Togo, H., Zhang, Y., Uemura, Y., and Kawamura, Y., 2006. Shock-compression behavior of Zr-based metallic glass In: Khan, A. S., and Kazmi, R., (Eds.), 12th International Symposium on Plasticity and its Applications. Anisotropy, Texture, Dislocations and Multiscale Modeling in Finite Plasticity and Viscoplasticity and Metal Forming. Neat, Inc., MD, USA, Halifax, Nova Scotia, Canada, pp. 157-159. Metha, N., 2003. Pressure and Strain Rate Dependency of Flow Stress of Glassy Polymers under Dynamic Loading, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, OH. Nathenson, D. I., Prakash, V., and Dandekar, D. P., 2005. Dynamic response of silicon nitride under combined pressure and shear impact, Paper # 315 (s22). Proceedings of the 2005 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Bethel, Connecticut, USA, Portland, Oregon USA. Nathenson, D. 2006. Experimental Investigation of High Velocity Impacts on Brittle Materials, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, OH. Parry, T. V. and Wronski, A. S., 1985. Effect of hydrostatic pressure on the tensile properties of pultruded CFRP. Journal of Materials Science 20 (n6), 2141-2147. Peker, A., and Johnson, W. L., 1993. A highly processable metallic glass Zr41.2Ti13.8Cu12.5Ni10.0Be22.5. Applied Physics Letters 63, 2342-2344. Prakash, V., 1995. A pressure-shear plate impact experiment for investigating transient friction. Experimental Mechanics 35, 329-336. Shazly, M., 2005. Dynamic Deformation and Failure of Gamma-met PX at Room and Elevated Temperatures, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, OH.
317

Spaepen F., 1977. A microscopic mechanism for steady state inhomogeneous flow in metallic glasses. Acta Metall. 25, 407-415. Steif, P. S., 1983. Ductile versus brittle behavior of amorphous metals. Journal of the Mechanics and Physics of Solids 31 (n5), 359-388. Subhash, G., Zhang, H., and Li, H., 2003. Thermodynamic and Mechanical Behavior of Hafnium/Zirconium Based Bulk Metallic Glasses, Proceedings of the International Conference of Mechanical Behavior of Materials (ICM-9). Kenes International, Geneva, Switzerland. Sunny, G. P., Lewandowski, J. J., and Prakash, V., 2006a. Dynamic compression of amorphous and annealed bulk metallic glass, Paper # 349, Proceedings of the 2006 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, St. Louis, MO, USA. Sunny, G. P., Prakash, V., and Lewandowski, J. J., 2005a. Effects of annealing on dynamic behavior of a bulk metallic glass, Paper # IMECE2005-83016, Proceedings of the 2005 International Mechanical Engineering Conference and Exposition, ASME. American Society of Mechanical Engineers, New York, NY, Orlando, FL. Sunny, G. P., Prakash, V., and Lewandowski, J. J., 2006b. Results from a novel insert design for high starin-rate compression of a bulk metallic glass, Paper # IMECE2006-15414, Proceedings of the 2006 International Mechanical Engineering Conference and Exposition, ASME. American Society of Mechanical Engineers, New York, NY, Chicago, IL. Sunny, G. P., Yuan, F., Lewandowski, J. J., and Prakash, V., 2005b. Dynamic Stress-Strain response of a Zr-based bulk metallic glass, Paper # 324, Proceedings of the 2005 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Portland, Oregon USA. Tsai, L., and Prakash, V., 2005. Dynamic response and spall strength of S2-glass fiber reinforced polymer composites, Paper # 322 (s57). Proceedings of the 2005 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Bethel, Connecticut, USA, Portland, Oregon USA. Turneaure, S. J., Winey, J. M., and Gupta, Y. M., 2004. Compressive shock wave response of a Zr-based bulk amorphous alloy. Applied Physics Letters 84, 1692-1694. Wesseling, P., Lowhaphandu, P. and Lewandowski, J. J., 2003. Effects of superimposed pressure on flow and fracture of two bulk amorphous metals. Mat. Res. Soc. Symp. Proc. 754, 275-279. Yang, C., Liu, R. P., Zhang, B. Q., Wang, Q., Zhan, Z. J., Sun, L. L., Zhang, J., and Gong, Z. Z., 2005. Void formation and cracking of Zr41Ti14Cu12.5-Ni10Be22.5 bulk metallic glass
318

under planar shock compression. Journal of Materials Science 40, 3917-3920. Yang, C., Wang, W. K., Liu, R. P., Zhang, X. Y., and Li, X., 2006. Damage features of Zr41Ti14Cu12.5Ni10Be22.5 bulk metallic glass impacted by hypervelocity projectiles. Journal of Spacecraft and Rockets 43, 565-567. Zhang, Z. F., Eckert, J. and Schultz, L., 2003a. Difference in compressive and tensile fracture mechanisms of Zr59Cu20Al10Ni8Ti3 bulk metallic glass. Acta Mat. 51, 1167-1179. Zhang, Z. F., et al., 2003b. Fracture mechanisms in bulk metallic glassy materials. Phys. Rev. Lett. 91 (4-045505), 1-4 . Zhuang, S. M., Lu, J., and Ravichandran, G., 2002. Shock wave response of a zirconium-based bulk metallic glass and its composite. Applied Physics Letters 80, 4522-4524.

319

Chapter 6

CONCLUSIONS

This thesis describes experimental and computational work that is focused on obtaining a better understanding of dynamic behavior in materials and interfaces that are of considerable scientific and technical importance. The first part of the thesis focuses

on better understanding co-seismic slip weakening and/or strengthening mechanisms in geological and analog materials that is of importance in understanding fault and earthquake dynamics. In the second part, plate-impact pressure-shear friction

experiments are used to investigate dynamic slip resistance and time-resolved growth of molten metal films that are created during dry metal-on-metal slip under extreme interfacial conditions. The third part of the thesis involves the study of dynamic failure, in particular dynamic spall in potential light weight armor materials, such as, glass fiber reinforced polymer composites. In this regards spall strength in two different architectures of the GRP composites (S2 glass woven roving in Cycom 4102 polyester resin matrix and a 5-harness satin weave E-glass in a Ciba epoxy matrix) are investigated. In the fourth part of the thesis, results are presented on the shock response of a zirconium-based bulk metallic glass (BMG), Zr41.25Ti13.75Ni10Cu12.5Be22.5, subjected to shock compression and combined shock compression and shear loading. Bulk metallic

glasses are amorphous metals with extraordinary high hardness and strength that can

320

potentially be used in the design of light weight armor and anti-armor materials for the navy and the army.

6.1 Co-seismic slip weakening in earth faults Earthquake occurs when fault strength falls down with increasing slip (slip weakening) or slip rate (slip velocity weakening), so the understanding of dynamic fault weakening during the nucleation and the propagation of a seismic rupture is a major task for researchers involved with fault and earthquake physics.

In the present study plate-impact pressure-shear friction experiments and modified torsional Kolsky-bar friction experiments are employed to investigate the frictional resistance in rocks and analog materials (quartz and soda lime glass), at relevant normal pressures and co-seismic slip rates. The results of the plate-impact pressure-shear

friction experiments indicate that a wide range of frictional slip conditions exist at the frictional interface. These conditions range from initial no-slip followed by slip

weakening, strengthening, and seizure all during a single slip event for the case of glass vs. glass experiments. For the case of dense Novaculite rock vs. Novaculite rock, the Moreover, the range of the coefficient of kinetic friction

similar trend is observed.

(0.1~0.4) during slip is much lower than those obtained at the nucleation phase, i.e., at quasi-static slip rates (0.6~0.85) (Byerlee, 1978; Dieterich, 1978).

321

In the modified torsional Kolsky-bar friction experiments on quartz, the window time for the steady-state friction is limited to only 100-120 s for these experiments, and the accumulated slip distance is only 0.8~1 mm due to the micro-cracking and fracture of tubular quartz specimen. As a result, the normal stress at the slip interface maintains

steady-state level of approximately 70 MPa for a while, and then starts to decrease. The coefficient of kinetic friction has an average value of approximately 0.20-0.23 during the steady friction state, then decreases to zero gradually. In order to solve the problem of microcracking and propagation of the tubular quartz specimen, and study the slip mechanism at larger slip distances, the thin-walled tubular quartz specimen is replaced by the thin-walled tubular aluminum specimen with a film of soda-lime glass (thickness ~ 1 m) deposited on one of its ends. Moreover, the quartz disk specimen is replaced by the soda-lime glass disk. In the modified torsional Kolsky-bar friction experiments on

soda-lime glass, the measured coefficient of kinetic friction (0.2~0.3) is also much lower than those obtained by Byerlee (1978) and Dieterich (1978) at quasi-static slip rates ( 1 mm/s, ~ 0.6 to 0.85). It is also interesting to note that the highest coefficient of kinetic

friction decreases with the increasing slip velocity. Also, in every experiment, the slip weakening is followed by strengthening at the tribo-pair interfaces; this behavior is very similar to what was seen in the plate-impact pressure-shear friction experiments on Novaculite rock and soda-lime glass.

An elementary model considering flash heating at asperity contacts had been

322

proposed recently by Rice (2006).

Using this model, a critical velocity,

Vcrit = 0.12 m/s , can been estimated to be onset of severe thermal weakening in glass
and other geo-materials. Recent laboratory studies (Goldsby and Tullis, 2002, Di Toro et al., 2004, Hirose and Shimamoto, 2005) and the results presented in the present study on rapid slip on geological and analog materials are indeed consistent with such an expected reduction in friction stress. The results suggest that the friction coefficient is

in the range of 0.2 to 0.4 at typical seismic slip rates of > 1 m/s in geo-materials.

Based on Rices model, we propose that the first slip-weakening in soda-lime glass and dense novaculite rock is most likely due to the thermal weakening induced by flash heating and incipient melting at asperity contacts, while the strengthening is a result of material softening and coalescence and solidification of melt patches at the slip surface. The much lower friction coefficients in soda-lime glass and analog materials (i.e. ~ 0.2 - 0.4) when compared with ~ 0.6 to 0.85 obtained by Byerlee (1978) and Dieterich (1978) at slow slip rates ( 1mm/s) is caused by flash heating at the asperity junctions. These relatively low values of for rocks and analog materials obtained in the current experiments may explain the large stress drop during fault rupture.

6.2 High speed friction at metal-on-metal interfaces High-speed sliding at metal interfaces plays an important role in the design of several tribo-elements, inculding bearings, gears, high performance brake liners, high speed
323

driven rocket on rail, geological fault modeling and frcitional model for penetration. In the present study we capitalize on the recent developments in the plate impact pressure-shear friction experiments (Prakash and Clifton, 1993; Prakash, 1995; Irfan and Prakash, 2000; Okada et al., 2001) and investigate dynamic slip resistance and time-resolved growth of molten metal films during dry metal-on-metal slip under extreme interfacial conditions. By employing tribo-pairs comprising hard tool-steel against

relatively low melt-point metals such as 7075-T6 aluminum alloy, interfacial friction stress of up to 300 MPa and slip speeds of approximately 250 m/s have been achieved. These relatively extreme interfacial conditions are conducive to the development of molten metal films at the tribo-pair interface. In order to interpret the experimental results a Lagrangian finite element based computational procedure is developed. The code accounts for finite elastic-plastic deformations, affects of material inertia, heat conduction, contact with friction, stain hardening, strain rate hardening, thermal softening, and full thermo-mechanical coupling. At temperatures below the melting point the

material is described as an isotropic thermally softening elastic-viscoplastic solid. For material elements with temperatures in excess of the melt temperature a purely Newtonian fluid constitutive model is employed.

During the early part of frictional slip, the measured coefficient of kinetic friction is observed to decrease with increasing slip velocity, and is relatively low when compared to the reported values of 0.4 obtained for the coefficient of kinetic friction between mild
324

steel and aluminum under quasi-static slip conditions (Bhushan and Gupta, 1991). A plausible reason can be attributed to the combination of relatively high values of normal pressure and high slip-speeds which generate substantial heat at the frictional interface. This heat produces a flash temperature and a thermal softening at the asperity-asperity junctions, thus reducing their shear strength, which in turn leads to lower values of friction stress (Archard, 1958; Kuhlmann-Wilsdorf, 1985; Kuhlmann-Wilsdorf, 1987; Ashby et al., 1991; Rice, 2006).

During the later part of the experiments interfacial slip occurs at near-melt to the fully-melt temperature regime of the 7075-T6 Al alloy. Under these fully melt

conditions the interfacial resistance approaches the shear strength of the molten aluminum alloy at normal stress in the range of 1-4.5 GPa and shear strain rates of ~ 107 s-1. The results of the study indicate that under these extreme conditions molten

aluminum films maintain a shearing resistance as high as 50-100 MPa. Moreover, the FEM simulations of the experiments reveal that the dynamic viscosity of the molten aluminum alloy decreases with the increasing impact velocity so that the purely Newtonian model can capture the dynamic shearing resistance of the molten aluminum alloy with sufficient accuracy. Scanning electron microscopy of the slip surfaces

confirm the existence of molten Al to be smeared on the interface.

325

6.3 Spall strength of Glass fiber reinforced composites The utilization of layered heterogeneous material systems in the development of integral armor provides a potential for a major improvement in the ballistic performance in a variety of lightweight armor applications. Due to their light-weight, high stiffness, and good ballistic resistance, various GRP composites have been chosen in composite integral armor as the main structural support behind the ceramic plates (Gama et al., 2001a,b).

In the present study, a series of normal plate-impact and pressure-shear plate impact experiments were conducted to study the spall strength in two different glass fiber reinforced polymer composites S2 glass woven roving in Cycom 4102 polyester resin matrix and a 5-harness satin weave E-glass in a Ciba epoxy (LY564) matrix. Based on

the experimental results the seven normal plate-impact experiments on the S2-glass GRP were placed in three different categories. Experiments in the first category were In these experiments the

conducted at an impact stress between 0 and 175MPa.

resultant tensile stress was too low to cause spallation within the specimens and the free surface particle velocity profiles were observed to unload completely to their no-spall predicted levels. Experiments in the second category were conducted at impact stresses

in the range of 175 MPa and 600 MPa; the resulting tensile stresses within the specimen were high enough to result in spall. In these experiments a clear pull-back signal was

326

observed in the measured free surface particle velocity profiles.

In the third category of

the experiments, the incident compression stress pulse amplitude was larger than 600 MPa. These relatively high levels of shock compression resulted in enough damage in

the GRP specimens such that no resistance to spall (i.e. zero spall strength) was registered in the experiments. The corresponding free surface particle velocity profiles for these

experiments show no signs of pull-back or unloading of the free surface particle velocity, and it remains at a level corresponding to the predicted Hugonoit state, Vmax. On the

other hand, the normal plate-impact experiments conducted on the E-glass GRP composites (at impact stresses ranging from 330.7 MPa to 2213.8 MPa) showed a finite spall strength. However, like in the case of the S2-glass GRP, the spall strength of the

E-glass GRP composite was observed to decrease with increasing levels of shock compression.

Under the combined compression and shear loading (pressure-shear plate impact experiments), the spall strengths in the two GRP composites were found to decrease with increasing levels of applied normal and the shear stress. A zero spall strength condition

was found for the E-glass GRP when the specimen was impacted at a normal stress of 975 MPa and a shear strain of 1.056%, which is much higher than that obtained for the case of the S2 glass GRP composite.

327

Based on these results from both normal plate impact and pressure-shear plate impact experiments, the spall strength for the two GRP composites is illustrated as a failure surface in the shear strain and the normal stress space. The failure surface shows that

the spall strength decreases with increasing shear-strain and with increasing normal stress for the two GRP composites. Moreover, the E-glass GRP shows much larger levels for the spall strength when compared to the S2 glass GRP. The maximum spall strength

measured for the E-glass GRP was 119.5 MPa, while the maximum measured spall strength for the S2 glass GRP was 53.7 MPa.

The measured spall strengths of GRPs are much lower than those observed in monolithic metals, ceramics, polymer etc. In such homogeneous materials, the

conventional spall process is thought to proceed from the coalescence/growth of inherent defects, such as impurities, micro-cracks, pre-existing pores, etc. However, damage in

GRP materials is complicated by the presence of additional heterogeneities due to the composite materials microstructure, and failure under impact loading is understood to the proceed by various mechanisms -- the incident energy is dissipated through the spread of failure laterally as well as through the thickness. Moreover, due to their inherent

multi-material heterogeneous composition of the GRPs, several distinctive modes of damage are observed which includes extensive delamination and fiber shearing, tensile fiber failure, large fiber deflection, fiber micro-fracture and local fiber buckling. In

328

particular, local fiber waviness is understood lead to inter-laminar shear failure in such materials (Hsiao and Daniel, 1996a; Hsiao and Daniel, 1996b). Moreover, strong

wave-reflection-effects, between components with different shock impedance, lead to significant shock wave dispersion resulting in an overall loss of spall strength (Zhuk et al., 1994; Dandekar and Beaulieu, 1995; Zaretsky et al., 2004).

6.4 Dynamic reponse of a Zr-based bulk metallic glasses Amorphous metals, also referred to as metallic glasses, differ from ordinary metals in that their constituent atoms are not arranged on a crystalline lattice. Due to their

randomly-ordered atomic structures, metallic glasses are known to exhibit unusual mechanical properties, such as near theoretical strength, large elastic strains, high hardness, excellent wear and corrosion resistance, and increased fracture toughness when compared to other brittle, high compressive strength materials (Bruck et al., 1996, Lowhaphandu and Lewandowski, 1998). The dynamic response of bulk metallic

glasses is of considerable interest for gaining insight into the high strain-rate response of this class of materials and for potential applications, such as kinetic energy penetrators (Johnson, 1999).

In the present study the first series of experiments are conducted to study the spall strength of the Zr-based BMG under shock compression and combined shock compression and shear loading. In the normal impact spall experiments at the low impact velocities, the structure of the shock profile shows the nearly elastic response of
329

the target material prior to the arrival of the spall wave at the free surface of the target plate. While in the normal impact spall experiments at the higher impact velocities, the

measured particle velocity versus time profiles show a step-like elastic precursor to the HEL level, a peak at the elastic front, followed by a steeply rising plastic wave and a gradual transition to a final plateau at the higher impact velocities. The peak at the elastic precursor front becomes much sharper as the impact velocity is increased from 463.7 m/s (impact stress of 6.06 GPa) to 541.8 m/s (impact stress of 7.06 GPa). Also, behind the elastic precursor, the rise associated with the plastic wave to the final plateau is not steady. The stress rise is relatively gradual at an impact speed of 463.7 m/s and The spall

becomes much steeper as the impact velocity is increased to 541.8 m/s.

strength of BMG was found to decrease with increasing normal stress. The decrease in spall strength can be attributed to damage accumulation as the impact velocity is increased. The spall strength at normal stress 4.4 GPa was 3.5 GPa, while the spall strengths at normal stresses of 5.1, 6.0 and 7.0 GPa were 2.72, 2.35 and 2.33 GPa, respectively. The HEL of Zr41.25Ti13.75Ni10Cu12.5Be22.5 was found out to be 6.15 GPa. In the pressure-shear plate-impact spall experiments, with increasing levels of shear strain, the spall strength of the BMG was found to decrease initially, and then increase dramatically, and decrease again as the shear strain is increased from 2.40% to 3.18%. This unusual trend in the shear strain on the spall strength of the Zr-based BMG can be explained by the competing roles of (i) localized plasticity, (ii) generation and propagation of cracks taking into consideration their relative dominance below and above
330

a given magnitude of shear strain, as Dandekar explained the similar trend in the normal stress dependence of the spall strength on a sintered and hot pressed silicon carbide (SiC) (Bartkowski and Dandekar, 1996, Dandekar, 2004). The sudden increase in spall

strength of silicon carbide, with an increase in shear strain, is likely due to the dominance of localized plastic deformation over crack-dominated brittle deformation, while the corresponding decline in spall strength with an increase in shear strain is attributed to the dominance of crack-induced brittle deformation over plastic deformation.

The motivation for the second series of experiments is to obtain information on the shearing strength of the Zr-based BMG as a function of shearing rates, normal stress and superimposed pressure. In view of this a series of experiments was conducted that involves the skew impact of a thick WC flyer plate with a stationary target. The target

comprises a front and a back WC plate with a thin BMG specimen sandwiched in between. Upon impact, a pressure-shear loading is generated in the sandwiched BMG specimen. From the measured free surface particle velocities, the shear flow stress

under ultrahigh strain rate (~105 /s), normal pressure (up to 9 GPa) and hydrostatic pressure (up to 6 GPa) is obtained for the Zr-based BMG. With regard to amorphous

metals, observations of shear band angles deviant from 450 (Donowan, 1989) have prompted proposals of a pressure-modified Mohr-Coulomb yield criterion of the form

c = k0 n

(1 + 2 + 3 ) . 3

Where c is the shear stress on the slip plane at

yielding, n is the stress component in the direction normal to the slip plane, k0 is the
331

yield stress in shear, and are constants. In Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002), the flow and fracture behaviors of the same bulk amorphous alloy, Zr41.25Ti13.75Ni10Cu12.5Be22.5, have been determined in tension and compression at room temperature with levels of superimposed hydrostatic pressure ranging from 0.1 to 1500 MPa. Also, torsion specimens were tested at atmospheric

pressure (i.e. 0.1 MPa) in order to approach conditions of pure shear. The results of these experiments clearly indicated that for the Zr41.25Ti13.75Ni10Cu12.5Be22.5 BMG, the parameter , that characterizes the pressure sensitivity term, is negligible over the pressure range used in the experiments, suggesting a Mohr-Coulomb criterion of the form:

c = k0 n .

Combined the present results with Lewandowskis work (Lewandowski, J. J. and Lowhaphandu, P., 2002), it seems that the hydrostatic pressure has a negligible influence on the shear yielding stress over a wild range (0~6 GPa). The normal pressure has a

negligible influence ( = 0 in compression) on the shear yield stress during compression (-8.8 GPa ~ -0.78 GPa), while the normal pressure has a small influence ( = 0.038 ) across from compression to tension state. It may suggest that this material

has a different normal stress sensitivity in compression and tension although they are both very small.

332

References

Archard, J. F., 1958. The temperature of rubbing surfaces. Wear 2, 438-455. Ashby, M. F., Abulawi, J., and Kong, H. S., 1991. Temperature maps for frictional heating in dry sliding. Tribology transactions 34 (4), 577-587. Bartkowski, P. T., and Dandekar, D. P., 1996. Spall strengths of sintered and hot pressed silicon carbide In: Schmidt, S. C., and Tao, W. C., (Eds.), Shock Compression of Condensed Matter-1995. American Institute of Physics, New York, pp. 535-539. Bhushan, B., and Gupta, B. K., 1991. Handbook of tribology: material coatings and surface treatments. McGraw Hill Inc., NY. Bruck, H. A., Rosakis, A. J., and Johnson, W. L., 1996. The dynamic compressive behavior of beryllium bearing bulk metallic glasses. Journal of Materials Research 11, 503-511. Byerlee, J. D., 1978. Friction of rocks. Pure Applied Geophysics 116, 615-626. Dandekar, D. P., and Beaulieu, P. A., 1995. Compressive and tensile strengths of glass reinforced polyester under shock wave propagation. In: Rajapakse, Y. D. S., and Vinson, J. R., (Eds.), High strain-rate effects on polymer, metal, and ceramic matrix composites and other advanced materials. ASME, New York, NY, pp. 63-70. Dandekar, D. P., 2004. Spall strength of silicon carbide under normal and simultaneous compression-shear shock wave loading. International Journal of Applied Ceramic Technology 1, 261-268. Di Toro, G., Goldsby, D. L., and Tullis, T. E., 2004. Friction falls towards zero in quartz rock as slip velocity approaches seismic rates. Nature 427, 436-439. Dieterich, J. H., 1978. Time dependent friction and mechanics of stick-slip. Pure and Applied Geophysics 116, 668-675. Donovan, P.E., 1988. A yield criterion for Pd40Ni40P20 metallic glass. Acta Met. 37, 445-456. Gama, B. A., Bogetti, T. A., Fink, B. K., Yu, C. J., Claar, T. D., Eifert, H. H., and Gillespie, J. W., 2001a. Aluminum foam integral armor: A new dimension in armor design. Composite Structures 52 (3-4), 381-395. Gama, B. A., Gillespie, J. W., Mahfuz, H., Raines, R. P., Haque, A., Jeelani, S., Bogetti, T. A., and Fink, B. K., 2001b. High Strain-Rate Behavior of Plain-Weave S-2 Glass/Vinyl Ester Composites. Journal of Composite Materials 35 (13), 1201-1228. Goldsby, D., and Tullis, T. E., 2002. Low frictional strength of quartz rocks at
333

sub-seismic slip rates. Geophysical research letters 29, 1844-1852. Hirose, T., and Shimamoto, T., 2005. Growth of molten zone as a mechanism of slip weakening of simulated faults in gabbro during frictional melting. Journal of Geophysical Research-Solid Earth 110. Hsiao, H. M., and Daniel, I. M., 1996a. Nonlinear elastic behavior of unidirectionai composites with fiber waviness under compressive loading. Journal of Engineering Materials and Technology-Transactions of the Asme 118 (4), 561-570. Hsiao, H. M., and Daniel, I. M., 1996b. Effect of fiber waviness on stiffness and strength reduction of unidirectional composites under compressive loading. Composites Science and Technology 56 (5), 581-593. Irfan, M. A., and Prakash, V., 2000. Time resolved friction during dry sliding of metal on metal. International journal of solids and structures 37, 2859-2882. Johnson, W. L., 1999. Bulk Metallic Glasses. In: Johnson, W. L., Inoue, A., and Liu, C. T., (Eds.), MRS Symposium Proceedings, vol. 554. Materials Research Society, pp. 311-339. Kuhlmann-Wilsdorf, D., 1985. Flash temperatures due to friction and joule heat at asperity contact. Wear 105, 187-198. Kuhlmann-Wilsdorf, D., 1987. Temperatures at interfacial contact spots:Dependence on velocity and on role reversal of two materials in sliding contact. Journal of tribology 109, 321-329. Lewandowski, J. J., and Lowhaphandu, P., 2002. Effects of hydrostatic pressure on the flow and fracture of a bulk amorphous metal. Philosophical magazine A82, 3427-3441. Lowhaphandu, P., and Lewandowski, J. J., 1998. Fracture toughness and notched toughness of bulk amorphous alloy: Zr-Ti-Ni-Cu-Be. Scripta Materialia 38, 1811-1817. Okada, M., Liou, N.-S., Prakash, V., and Miyoshi, K., 2001. Tribology of high speed metal-on-metal sliding at near-melt and fully-melt interfacial temperatures. Wear 249, 672-686. Prakash, V., and Clifton, R. J., 1993. Time Resolved Dynamic Friction Measurements in Pressure-Shear. In: Ramesh, K. T., (Ed). Experimental Techniques in the Dynamics of Deformable Bodies, vol. AMD Vol. 165. ASME, New York, pp. 33-47. Prakash, V., 1995. A pressure-shear plate impact experiment for investigating transient friction. Experimental Mechanics 35 (4), 329-336. Rice, J. R., 2006. Heating and weakening of faults during earthquake slip. Journal of Geophysical Research-Solid Earth 111. Zaretsky, E., deBotton, G., and Perl, M., 2004. The response of a glass fibers reinforced
334

epoxy composite to an impact loading. International Journal of Solids and Structures 41 (2), 569-584. Zhuk, A. Z., Kanel, G. I., and Lash, A. A., 1994. Glass Epoxy Composite Behavior under Shock Loading. Journal De Physique Iv 4 (C8), 403-407.

335

APPENDIX A

CALCULATION OF SHEAR STRAIN IN THE GRP TARGET UNDER COMBINED PRESSURE-SHEAR LOADING

The GRP sample is oriented such that the z-direction (001) of the sample is orthogonal to the principal axis of the two fiber plies lying along the x (100) and y (010) directions, as indicated in Figure A.1 below. The GRP thickness is along the z-direction. For pressure-shear experiments, the axis of the gun barrel axis is oriented at a skew angle of relative to the z-direction. We introduce a new primed coordinate frame that is

rotated by an angle about the y-axis. In this primed coordinate frame, the axis of the gun is oriented along the z-direction, as shown in Figure A.1. The transformation matrix is given by

Figure A.1: Oblique impact configuration.

336

cos a= 0 sin

0 sin 1 0 . 0 cos

(A.1)

In the primed frame, the resulting uniaxial strain tensor is given by

0 0 0 = 0 0 0 . 0 0 33

(A.2)

Alternatively, the components of the strain tensor in the unprimed coordinate frame can be expressed as

sin 2 33 ij = a3i a3 j33 = 0 33 sin cos

0 33 sin cos 0 0 . 0 33 cos 2

(A.3)

Also, in the unprimed frame, the 2nd Piola-Kirchoft stress tensor, tij , can be calculated by making use of the components of the fourth order stiffness constant tensor of GRP, C ijkl , and the Lagrangian strain kl ,

tij = Cijkl kl .
337

(A.4)

In the primed coordinate frame, the 2nd Piola-Kirchoft stress tensor, tij , and Cauchy stress tensor kl , can be expressed as

tij = aik a jl tkl ,

(A.5)

And

kl =

1 Fkj Fli tij . J

(A.6)

In Equation (A.6), Fkj are the (k, j) components of the deformation gradient tensor, which can be related to the strain tensor . J is the determinant of deformation gradient tensor, and assumes the simple form ( 0 / ) . Note that J is invariant under coordinate transformation.

Next, using Equations (A.3)-(A.6), the components of the Cauchy stress tensor in the primed coordinate frame can be written as

338

 11 =

{( C11 + C33 4C44 ) sin 2 2 + 4C13}33 4 0 22 = {C12 sin 2 + C13 cos 2 }33 0 1 33 = 0 C11 sin 4 + C33 cos 4 + C13 + C44 sin 2 2 33 2

(A.7)

13 = C11 sin 2 C33 cos 2 + ( C13 + C44 ) cos 2 sin 2 33

Moreover, combining Equations (A.3) and (A.7), the shear strain 13 in the GRP specimen can be expressed as

13 =

0 1 2 4 4 C11 sin + C33 cos + C13 + C44 sin 2 2

33

sin cos ,

(A.8)

where,

0 = (U s u p )/U s .

(A.9)

In Equation (A.9), 33 is the impact stress along the gun barrel direction, and is calculated from the impact velocity and the impedance of the flyer and the target materials; and 0 are the densities of the GRP after and before impact respectively;
is the skew angle of the pressure-shear experiments; and Cij are the elastic constants

339

of the GRP composite, and are C 11 = 31.55 GPa, C 33 = 20.12 GPa, C 13 = 9.75 GPa and

C 44 =4.63 GPa (Dandekar et al., 1998).

340

APPENDIX B

FREE SURFACE PARTICLE VELOCITY PROFILES FOR ALL PLATE-IMPACT EXPERIMENTS

40

Free Surface Transverse Velocity (m/s)

35 30 25 20 15 10 5 0

0.5

1.5

Time after Impact (s)

2.5

3.5

4.5

Experiment Shot1 for glass-on-glass slip

341

50

60 55 50 45

Free Surface Transverse Velocity (m/s)

40 35 30 25 20 15 10 5 0 0 1 2

40

Normal

35 30 25 20 15 10

Transverse
3 4 5

5 0

Time after Impact (s)

Experiment Shot2 for glass-on-glass slip

40

Free Surface Transverse Velocity (m/s)

35 30 25 20 15 10 5 0

0.5

1.5

Time after Impact (s)

2.5

3.5

4.5

Experiment Shot3 for glass-on-glass slip

342

Free Surface Normal Velocity (m/s)

45

100 90 80 70 60 50 40 30 20 10 0 0 1 2 3 4 5 6

100 90 80

Normal

70 60 50 40

Transverse

30 20 10 0

Time after Impact (s)

Experiment ShotFY008 for Novaculite-on-Novaculite slip

50

Free Surface Transverse Velocity (m/s)

40

30

20

Transverse
10

Time after Impact (s)

Experiment ShotFY009 for Novaculite-on-Novaculite slip

343

Free Surface Transverse Velocity (m/s)

Free Surface Normal Velocity (m/s)

80

300

Free Surface Transverse Velocity (m/s)

250

60 50 200

Normal
40 30 20 150

100

Transverse
10 0

50

Time after Impact (s)

Experiment ShotFY05001 for steel-on-aluminum slip

80

300

Free Surface Transverse Velocity (m/s)

250

60 50 40 30 20

Normal

200

150

100

Transverse
10 0

50

Time after Impact (s)

Experiment ShotFY05004 for steel-on-aluminum slip

344

Free Surface Normal Velocity (m/s)

70

Free Surface Normal Velocity (m/s)

70

80

300

Free Surface Transverse Velocity (m/s)

250

60 50 40 30 20 10 0

Normal
200

150

100

Transverse

50

Time after Impact (s)

Experiment ShotFY05007 for steel-on-aluminum slip

80

300

Free Surface Transverse Velocity (m/s)

250

60 50 40 30 20 10 0 200

Normal

150

100

Transverse
50

Time after Impact (s)

Experiment ShotFY05008 for steel-on-aluminum slip

345

Free Surface Normal Velocity (m/s)

70

Free Surface Normal Velocity (m/s)

70

120 110 100

Free Surface Velocity (m/s)

90 80 70 60 50 40 30 20 10 0 0 2 4 6 8 10

Time after Impact (s)

Experiment ShotFY06001 for GRP

250

Free Surface Velocity (m/s)

200

150

100

50

Time after Impact (s)

10

Experiment ShotFY06002 for GRP

346

350

300

Free Surface Velocity (m/s)

250

200

150

100

50

Time after Impact (s)

10

Experiment ShotFY06003 for GRP

450 400

Free Surface Velocity (m/s)

350 300 250 200 150 100 50 0 0 2 4 6 8 10

Time after Impact (s)

Experiment ShotFY06004 for GRP

347

700

600

Free Surface Velocity (m/s)

500

400

300

200

100

Time after Impact (s)

10

Experiment ShotFY06005 for GRP

350

300

Free Surface Velocity (m/s)

250

200

150

100

50

Time after Impact (s)

10

Experiment ShotFY06006 for GRP

348

200

Free Surface Velocity (m/s)

150

100

50

Time after Impact (s)

10

Experiment ShotFY06007 for GRP

200 175

Free Surface Velocity (m/s)

150 125 100 75 50 25 0

Time after Imapct (s)

10

Experiment ShotFY06008 for GRP

349

300

250

Free Surface Velocity (m/s)

200

150

100

50

Time after Impact (s)

10

Experiment ShotFY06009 for GRP

300

250

Free Surface Velocity (m/s)

200

150

100

50

Time after Impact (s)

10

Experiment ShotFY06010 for GRP

350

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

0.5

Time after Impact (s)

1.5

2.5

Experiment ShotFY06011 for BMG

500 450

Free Surface Velocity (m/s)

400 350 300 250 200 150 100 50 0 0 0.5 1 1.5 2 2.5 3

Time after Impact (s)

Experiment ShotFY06012 for BMG

351

500

Free Surface Velocity (m/s)

400

300

200

100

0.5

Time after Impact (s)

1.5

2.5

Experiment ShotFY06013 for BMG

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

0.5

Time after Impact (s)

1.5

2.5

Experiment ShotFY06014 for BMG

352

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

Time after Impact (s)

Experiment ShotFY06015 for BMG

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

Time after Impact (s)

Experiment ShotFY06016 for BMG

353

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

Time after Impact (s)

Experiment ShotFY06017 for BMG

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

Time after Impact (s)

Experiment ShotFY06018 for BMG

354

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

Time after Impact (s)

Experiment ShotFY06019 for BMG

400 350

Free Surface Velocity (m/s)

300 250 200 150 100 50 0

Time after Impact (s)

Experiment ShotFY06020 for BMG

355

200 180

50

160 140 120 100 80 60 40 20 0 0 1 2 3

40

Normal velocity

35 30 25 20

Transverse velocity

15 10 5 0

Time after impact (s)

Experiment ShotExp.1 for BMG

200 180

50

160 140 120 100 80 60 40 20 0 0 1 2 3

40

Normal velocity

35 30 25 20

Transverse velocity

15 10 5 0

Time after impact (s)

Experiment ShotExp.2 for BMG

356

Free surface transverse velocity (m/s)

45

Free surface normal velocity (m/s)

Free surface transverse velocity (m/s)

45

Free surface normal velocity (m/s)

200 180

50

160 140 120 100 80 60 40 20 0 0 1 2 3

40

Normal velocity

35 30 25 20

Transverse velocity

15 10 5 0

Time after impact (s)

Experiment ShotExp.3 for BMG

357

Free surface transverse velocity (m/s)

45

Free surface normal velocity (m/s)

BIBLIOGRAPHY

Achenbach, J. D., 1973. Wave Propagation in Elastic Solids. North-Holland, Amsterdam. Adams, G. G., 1995. Self excited oscillations of two elastic half-spaces sliding with a constant coefficient of friction. Journal of applied mechanics 62, 867-872. Agbossou, A., Cohen, I., and Muller, D., 1995. Effects of Interphase and Impact Strain Rates on Tensile Off-Axis Behavior of Unidirectional Glass-Fiber Composite Experimental Results. Engineering Fracture Mechanics 52 (5), 923. Andrews, D. J., and Ben-Zion, Y., 1997. Wrinkle-like slip pulse on a fault between different materials. Journal of Geophysical Research 102, 553-571. Andrews, D. J., 2002. A fault constitutive relation accounting for thermal pressurization of pore fluid. Journal of Geophysical Research 107, 2363, doi: 2310.1029 / 2002JB001942, ESE 001915-001941001915-001948. Antoniou, S., Cameron, A., and Gentle, C., 1976. The friction-speed relation from stick-slip data. Wear 36, 235-254. Archard, J. F., 1958. The temperature of rubbing surfaces. Wear 2, 438-455. Armstrong-Helouvry, B., Dupont, P., and Canudas de Wit, C., 1994. A survey of models, analysis tools and compensation methods for the control of machines with friction. Automatica 30, 1083-1153. Argon, A. S., 1979. Plastic deformation in metallic glasses. Acta Metalli. 27 (n1), 47-58. Asaro, R. J., and Rice, J. R., 1977. Strain localization in ductile single crystals. Journal of the mechanics and physics of solids 25, 309-338. Ashby, M. F., Abulawi, J., and Kong, H. S., 1991. Temperature maps for frictional heating in dry sliding. Tribology transactions 34, 577-587. Bach, J., Krueger, B., and Fultz, B., 1991. Shock wave consolidation of a Ni-Cr-Si-B metallic-glass powder. Materials Letters 11, 383-388. Barber, J. R., 1970. The conduction of heat from sliding solids. International Journal of Heat and Mass Transfer 13, 857-869. Barber, J. R., 1976. Some thermoelastic contact problems involving frictional heating. The Quarterly Journal of Mechanics and Applied Mathematics 29, 1-13. Barker, L. M., and Hollenbach, R. E., 1972. Laser interferometer for measuring high velocities of any reflecting surface. Journal Applied Physics 43, 4669-4675.
358

Barker, L. M., The development of the VISAR, and its use in shock compression science. Shock compression of condensed matter-1999, Snowbird, UT, 1999. Barre, S., Chotard, T., and Benzeggagh, M. L., 1996. Comparative study of strain rate effects on mechanical properties of glass fibre-reinforced thermoset matrix composites. Composites Part a-Applied Science and Manufacturing 27 (12), 1169-1181. Bartkowski, P. T., and Dandekar, D. P., 1996. Spall strengths of sintered and hot pressed silicon carbide In: Schmidt, S. C., and Tao, W. C., (Eds.), Shock Compression of Condensed Matter-1995. American Institute of Physics, New York, pp. 535-539. Beeler, N. M., and Tullis, T. E., 1996. Self-healing slip pulses in dynamic rupture models due to velocity-dependent strength. Bulletin of the Seismological Society of America 86, 1130-1148. Bell, R., and Burdekin, M., 1969. A study of stick-slip motion of machine tool feed drives. Proceedings Institute of Mechanical Engineers 184: Part I, 543-560. Belytschko, T., 1983. An overview of semidiscretization and time integration procedures. In: Belytschko, T., and Hughes, T. J. R., (Eds.). North-Holland, Amsterdam, pp. 1-65. Ben-Zion, Y., 2001. Dynamic ruptures in recent models of earthquake faults. Journal of Mechanics and Physics of Solids 49, 2209-2244. Betheney W., DeLuca E., Prifti J. and Chou S. C., 1998. Ballistic impact damage of S2-glass reinforced plastic structural armor. Composites Science and Technology 58, 1453-1461. Bhushan, B., and Gupta, B. K., 1991. Handbook of tribology: material coatings and surface treatments. McGraw Hill Inc., NY. Blanpied, M. L., Tullis, T. E., and Weeks, J. D., 1998. Effects of slip, slip rate, and shear heating on the friction of granite. Journal of Geophysical Research-Solid Earth 103, 489-511. Boitnott, G. N., Biegel, R. L., Scholz, C. H., N., Y., and W., W., 1992. Micromechanics of rock friction 2: Quantitative modeling of initial friction with contact theory. Journal of Geophysics Research 97, 8965-8978. Bowden, F. P., and Ridler, K. E. W., 1936. Physical properties of surfaces, III. Proceedings of Royal Society of London A 154, 640-656. Bowden, F. P., and Tabor, D., 1950. The Friction and Lubrication of Solids. Oxford University Press, London. Bowden, F. P., and Thomas, P. H., 1954. The surface temperature of sliding solids.
359

Proceedings of Royal Society of London A 223, 29-36. Bowden, F. P., and Freitag, E. H., 1958. The friction of solids at very high speeds, I. Metal on metal, II. Metal on diamond. Proceedings of Royal Society of London A248, 350-367. Bowden, F. P., and Persson, P. A., 1960. Deformation heating and melting of solids in high speed friction. Proceedings of Royal Society of London A260, 433-458. Bowden, P. B., Jukes, J. A., 1972. Plastic flow of isotropic polymers. Journal of Materials Science 7 (n1), 52-63. Bruck, H. A., Rosakis, A. J., and Johnson, W. L., 1996. The dynamic compressive behavior of beryllium bearing bulk metallic glasses. Journal of Materials Research 11, 503-511. Budiansky, B., Remarks on theories of solid and structural mechanics. 1969, pp. 77-83. Bureau, L., Baumberger, T., and Caroli, C., 2000. Shear response of a frictional interface to a normal load modulation. Physics Review Letters E 62, 6810-6820. Byerlee, J. D., 1978. Friction of rocks. Pure Applied Geophysics 116, 615-626. Cagnoux, J., and Longy, F., 1988. Spallation and shock-wave behavior of some ceramics. Journal de Physique Colloque 49, 3-10. Camacho, G. T., and Ortiz, M., 1997. Adaptive Lagrangian modelling of ballistic penetration of metallic targets. Computer Methods in Applied Mechanics and Engineering 142, 269-301. Carslaw, H. S., and Jaeger, J. C., 1986. Conduction of heat in solids. Oxford University Press, London. Chen, H. S. and Turnbull, D., 1969. Formation, stability and structure of palladium-silicon based alloy glasses. Acta Metall. 17, 1021-1031. Chen, H. S., Metallic glass. 1976. Mater. Sci. Eng., 25, 59-69. Chen, H. S., 1974. Thermodynamic considerations on the formation and stability of metallic glasses. Acta Metall. 22, 1505-1511. Chester, J. S., Kronenberg, A. K., Chester, F. M., and Guillemette, R. N., 2003. Characterization of natural slip surfaces relevant to earthquake mechanics. EOS Trans. AGU, 84(46), Fall Mtg. Suppl., Abstract S42C-0185. Chester, J. S., Chester, F. M., Kronenberg, A. K., and Guillemette, R. N., 2005. Extreme localization of slip and implications for dynamic weakening of faults, Manuscript in
360

preparation. Chester, J. S., and Goldsby, D. L., 2003. Microscale characterization of natural and experimental slip surfaces relevant to earthquake mechanics, SCEC Annual Report. Chhabildas, L. C., Sutherland, H. J., and Asay, J. R., 1979. A velocity interferometer technique to determine shear-wave particle velocity in shock-loaded solids. Journal of applied physics 50 (8), 5196-5201. Clifton, R. J., 1990. High strain rate behavior of metals. Applied mechanics reviews 43 (5, Part II), S9-S22. Cocco, M., and Rice, J. R., 2002. Pore pressure and poroelasticity effects in Coulomb stress analysis of earthquake interactions. Journal of Geophysical Research, 107(B2), cn:2030, doi:10.1029/2000JB000138, pp. ESE 2-1 to 2- 17. Cochard, A., and Rice, J. R., 2000. Fault rupture between dissimilar materials: Ill-posedness, regularization, and slip-pulse response. Journal of geophysical research 105, 891-907. Cohen, M. H. and Turnbull, D., 1961. Composition requirements for glass formation in metallic and ionic system. Nature 189, 131-132. Connor, R. D., Dandliker, R. B. and Johnson, W. L., 1998. Mechanical properties of tungsten and steel reinforced Zr41.25Ti13.75Cu12.5Ni10Be22.5 bulk metallic glass composites. Acta Mat. 46 (17), 6089-6102. Conner, R. D., Dandliker, R. B., Scruggs, V., and Johnson, W. L., 2000. Dynamic deformation behavior of tungsten-fiber/metallic-glass matrix composites. International Journal of Impact Engineering 24, 435-444. Cuitino, A. M., and Ortiz, M., 1992. A material-independent method for extending stress update algorithms from small-strain plasticity to finite plasticity with multiplicative kinematics. Engineering computations 9, 437-451. Dandekar, D. P., and Beaulieu, P. A., 1995. Compressive and tensile strengths of glass reinforced polyester under shock wave propagation. In: Rajapakse, Y. D. S., and Vinson, J. R., (Eds.), High strain-rate effects on polymer, metal, and ceramic matrix composites and other advanced materials. ASME, New York, NY, pp. 63-70. Dandekar, D. P., Boteler, J. M., and Beaulieu, P. A., 1998. Elastic constants and delamination strength of a glass-fiber-reinforced polymer composite. Composites Science and Technology 58 (9), 1397-1403. Dandekar, D. P., 2004. Spall strength of silicon carbide under normal and simultaneous
361

compression-shear shock wave loading. International Journal of Applied Ceramic Technology 1, 261-268. Davis, L. A. and Kavesh, S., 1975. Deformation and fracture of an amorphous metallic alloy at high pressure. J. Mat. Sci.10 (3), 453-459. DeLuca, E., Prifti, J., Betheney, W., and Chou, S. C., 1998. Ballistic impact damage of S 2-glass-reinforced plastic structural armor. Composites Science and Technology 58 (9), 1453-1461. Di Toro, G., Goldsby, D. L., and Tullis, T. E., 2004. Friction falls towards zero in quartz rock as slip velocity approaches seismic rates. Nature 427, 436-439. Dieterich, J. H., 1978. Time dependent friction and mechanics of stick-slip. Pure and Applied Geophysics 116, 668-675. Dieterich, J. H., 1979. Modeling of rock friction: I, experimental results and constitutive equations. Journal of Geophysical Research 84, 2161-2168. Dieterich, J. H., 1981. Constitutive properties of faults with simulated gouge. In: Carter, N. L., Friedman, M., Logan, J. M., and Stearns, D. W., (Eds.), vol. 24. American Geophysical Union, Washington, D.C., pp. 103-120. Dieterich, J. H., and Kilgore, B. D., 1994. Direct observation of frictional contacts; new insights for state dependent properties. Pure Applied Geophysics 143, 283-302. Dieterich, J. H., and Kilgore, B. D., 1996. Implications of fault constitutive properties for earthquake prediction. Proceedings of the National Academy of Sciences of the United States of America 93, 3787-3794. Dieterich, J. H., and Kilgore, B. D., 1996. Imaging surface contacts: Power law contact distributions and contact stresses in quartz, calcite, glass and acrylic plastic. Tectonophysics 256, 219-239. Donovan, P.E., 1988. A yield criterion for Pd40Ni40P20 metallic glass. Acta Met. 37, 445-456. Drehman, A. L., Greer, A. L. and Turnbull, D., 1982. Bulk formation of a metallic glass: Pd40Ni40P20. Appl. Phys. Lett. 41 (8), 716-717. Earles, S. W. E., and Powell, D. G., 1966. Variations in friction and wear between unlubricated steel surfaces. 181, Part 30, 171-179. Elhabak, A. M. A., 1991. Mechanical-Behavior of Woven Glass Fiber-Reinforced Composites under Impact Compression Load. Composites 22 (2), 129-134.

362

Espinosa, H. D., Xu, Y., and Brar, N. S., 1997. Micromechanics of failure waves in glass: I, experiments. Journal of American Ceramic Society 80, 2061-2073. Fink, B. K., 2000. Performance Metrics for Composite Integral Armor. Journal of Thermoplastic Composite Materials 13 (5), 417-431. Fowles, G. R., 1961. Shock wave compression of hardened and annealed 2024 Aluminum. Journal of Applied Physics 32, 1475-1487. Gama, B. A., Bogetti, T. A., Fink, B. K., Yu, C. J., Claar, T. D., Eifert, H. H., and Gillespie, J. W., 2001a. Aluminum foam integral armor: A new dimension in armor design. Composite Structures 52 (3-4), 381-395. Gama, B. A., Gillespie, J. W., Mahfuz, H., Raines, R. P., Haque, A., Jeelani, S., Bogetti, T. A., and Fink, B. K., 2001b. High Strain-Rate Behavior of Plain-Weave S-2 Glass/Vinyl Ester Composites. Journal of Composite Materials 35 (13), 1201-1228. Goldsby, D. L., and Tullis, T. E., 2002. Low frictional strength of quartz rocks at sub-seismic slip rates. Geophysical research letters 29, 1844-1855. Grady, D. E., and Kipp, M. E., 1993. High-Pressure Shock Compression of Solids. Springer-Verlag, Berlin, Germany. Grady, D. E., and Kipp, M. E., 1993. Dynamic fracture and fragmentation. In: Asay, J. R., and Shahinpoor, M., (Eds.), High-Pressure Shock Compression of Solids. Springer-Verlag, New York, pp. 265-322. Gray, G. T., 2000. Shock wave testing of ductile materials, Mechanical Testing and Evaluation Handbook, vol. 8. American Society for Metals, Materials Park, pp. 530-559. Hays, C. C., Kim, C. P., and Johnson, W. L., 2000. Enhanced plasticity of bulk metallic glasses containing ductile phase dendrite dispersions. Metastable, Mechanically Alloyed and Nanocrystalline Materials-- Parts 1 and 2, vol. 343-3, pp. 191-196. He, G., and Robbins, M. O., 2001. Simulations of the kinetic friction due to adsorbed surface layers. Tribology Letters 10, 7-14. Heaton, T. H., 1990. Evidence for and implications of self-healing pulses of slip in earthquake rupture. Physics of the Earth and Planetary Interiors 64, 1-20. Hill, R., 1966. Generalized constitutive relations for incremental deformation of metal crystal by multislip. Journal of the mechanics and physics of solids 14, 95-102. Hill, R., and Rice, J. R., 1972. Constitutive analysis of elastic-plastic crystals at arbitrary strain. Journal of the mechanics and physics of solids 20, 401-443.

363

Hill, R., and Rice, J. R., 1973. Elastic potentials and the structure of inelastic constitutive laws. SIAM, Journal of applied mathematics 25, 448-461. Hirose, T., and Shimamoto, T., 2005. Growth of molten zone as a mechanism of slip weakening of simulated faults in gabbro during frictional melting. Journal of Geophysical Research-Solid Earth 110. Hirose, T., and Shimamoto, T., 2005. Slip-weakening distance of faults during frictional melting as inferred from experimental and natural pseudotachylytes. Bulletin of the Seismological Society of America 95, 1666-1673. Hsiao, H. M., and Daniel, I. M., 1996. Effect of fiber waviness on stiffness and strength reduction of unidirectional composites under compressive loading. Composites Science and Technology 56 (5), 581-593. Hufnagel, T. C., Jiao, T., Li, Y., Xing, L. Q., and Ramesh, K. T., 2002. Deformation and failure of Zr57Ti5Cu20Ni8Al10 bulk metallic glass under quasi-static and dynamic compression. Journal of Materials Research 17, 1441-1445. Hughes, T. J. R., Analysis of transient algorithms with particular reference to stability behavior. 1983, pp. 67-155. Hughes, T. J. R., 1987. The Finite Element Method. Prentice Hall, Englewood Cliffs, NJ. Hsiao, H. M., and Daniel, I. M., 1996. Nonlinear elastic behavior of unidirectionai composites with fiber waviness under compressive loading. Journal of Engineering Materials and Technology-Transactions of the Asme 118 (4), 561-570. Hsiao, H. M., and Daniel, I. M., 1996. Effect of fiber waviness on stiffness and strength reduction of unidirectional composites under compressive loading. Composites Science and Technology 56 (5), 581-593. Hufnagel, T. C., Jiao, T., Li, Y., Xing, L. Q., and Ramesh, K. T., 2002. Deformation and failure of Zr57Ti5Cu20Ni8Al10 bulk metallic glass under quasi-static and dynamic compression. Journal of Materials Research 17, 1441-1445. Inoue, A., and Hashimoto, K., 2001. Amorphous and Nanocrystalline Materials: Preparation, Properties and Applications. Springer-Verlag, Berlin, Germany. Irfan, M. A., and Prakash, V., 1994. Contact temperatures during sliding in pressure shear impact. Proceedings Society of Experimental Mechanics Conference, Baltimore, MD, pp. 173-182. Irfan, M. A., and Prakash, V., 2000. Time resolved friction during dry sliding of metal on metal. International journal of solids and structures 37, 2859-2882.
364

Johnson, W. L., 1999. Bulk Metallic Glasses. In: Johnson, W. L., Inoue, A., and Liu, C. T., (Eds.), MRS Symposium Proceedings, vol. 554. Materials Research Society, pp. 311-339. Jones, R. M., 1999. Mechanics of Composite Materials. Taylor & Francis, Philadelphia. Kadhim, M. J., and Earles, S. W. E., 1966. Unlubricated sliding at high speeds between copper and steel surfaces. Proceedings institute of mechanical engineers 181, Part 30, 157-162. Kennedy, F. E., 1984. Thermal and thermomechanical effects in dry sliding. Wear 100, 453-476. Kim, K. S., Clifton, R. J., and Kumar, P., 1977. A combined normal and transverse displacement interferometer with an application to impact of Y-cut Quartz. Journal of Applied Physics 48, 4132-4139. Klement, W. J., Willens, R. H. and Duwez, P., 1960. Non-crystalline structure in solidified gold-silicon alloys. Nature, 187, 869-870. Kobayashi, S. O. S. I. A. T., 1989. Metal forming and Finite Element Method. Oxford University Press, London. Krafft, J. M., 1955. Surface friction in ballistic penetration. Journal of applied physics 26(10), 1248-1253. Krieg, R. D., and Key, S. W., 1973. Transient shell response by numerical time integration. International journal of numerical methods in engineering 7, 273-286. Kuhlmann-Wilsdorf, D., 1985. Flash temperatures due to friction and joule heat at asperity contact. Wear 105, 187-198. Kuhlmann-Wilsdorf, D., 1987. Temperatures at interfacial contact spots:Dependence on velocity and on role reversal of two materials in sliding contact. Journal of tribology 109, 321-329. Kui, H. W., Greer, A. L. and Turnbull, D., 1984. Formation of bulk metallic-glass by fluxing. Appl. Phys. Lett. 45 (6), 615-616. Kumar, P., and Clifton, R. J., 1977. Optical alignment of impact faces for plate impact experiments. Journal of Applied Physics 48, 1366-1367. Lachenbruch, A. H., 1980. Frictional heating, fluid pressure, and the resistance to fault motion. . Journal of Geophysical Research 85, 60976122. Leamy, H. J., Chen, H. S. and Wang, T. T., 1972. Plastic flow and fracture of metallic glass. Metallurgical Transaction, 3, 699-708.
365

Lee, E. H., 1969. Elastic-plastic deformation at finite strains. Journal of Applied Mechanics 36, 1-6. Lee, T. C., and Delaney, P. T., , 1987. Frictional heating and pore pressure rise due to a fault slip Geophys J. Roy. Astronom. Society 88, 569-591. Lemonds, J., and Needleman, 1986. A finite element analysis of shear localization in rate and temperature dependent solids. Mechanics of Materials 5, 339-361. Lewandowski, J. J., and Lowhaphandu, P., 1998. Effects of hydrostatic pressure on mechanical behavior and deformation processing of materials. International material reviews 43, 145-187. Lewandowski, J. J., and Lowhaphandu, P., 2002. Effects of hydrostatic pressure on the flow and fracture of a bulk amorphous metal. Philosophical magazine A82, 3427-3441. Lewandowski, J. J. and Greer, A. L., 2006. Temperature rise at shear bands in metallic glasses. Nat. Mat. 5, 15-18. Li, J. C. and Wu, J. B. C., 1976. Pressure and normal stress effects in shear yielding. Journal of Materials Science 11 (n3), 445-457. Lifshitz, J. M., 1976. Impact strength of angle ply fiber reinforced materials. Journal of Composite Materials 10, 92-101. Lim, S. C., Ashby, M. F., and Brunton, J. H., 1989. The effects of sliding conditions on the dry friction of metals. Acta Materialia 37, 767-772. Linker, M. F., and Dieterich, J. H., 1992. Effects of variable normal stress on rock friction: observations and constitutive equations. Journal of Geophysical Research 97, 4923-4940. Liou, N. S., Okada, M., and Prakash, V., 2004. Formation of Molten Metal Films During Metal-on-Metal Slip Under Extreme Interfacial Conditions. Journal of the Mechanics and Physics of Solids 52, 2025-2056. Lowhaphandu, P., and Lewandowski, J. J., 1998. Fracture toughness and notched toughness of bulk amorphous alloy: Zr-Ti-Ni-Cu-Be. Scripta Materialia 38, 1811-1817. Lowhaphandu, P., Montgomery, S. L. and Lewandowski, J. J., 1999. Effects of superimposed hydrostatic pressure on flow and fracture of a Zr-Ti-Ni-Cu-Be bulk amorphous alloy. Scr. Mat. 41 (1), 19-24. Lowhaphandu, P. et al., 2000a. Deformation and fracture toughness of a bulk amorphous Zr-Ti-Ni-Cu-Be alloy). Intermetallics 8, 487-492. Lowhaphandu, P., 2000b. Mechanical Behavior of a Zirconium-based Bulk Metallic Glass. Ph. D. Thesis, Case Western Reserve University, Cleveland, OH .
366

Lu, J., 2002. Mechanical Behavior of a Bulk Metallic Glass and its Composites Over a Wide Range of Strain Rates and Temperatures, (Ph.D. Dissertation), California Institute of Technology, Pasadena, CA. Lundergan, C. D., 1963. Equation of State of 6061-T6 aluminum at low pressures. Journal of Applied Physics 34(7), 2046-2052. Maeda, K., Nakagawa, K., and Takeuchi, S., 1978. Activation analysis of plastic deformation of CdTe below room temperature. Phys Status Solidi (a) 48 (n2), 587-591. Mahfuz, H., Zhu, Y., Haque, A., Abutalib, A., Vaidya, V., Jeelani, S., Gama, B., Gillespie, J., and Fink, B., 1999. Investigation of high-velocity impact on integral armor using finite element method. International journal of impact engineering 24 (2), 203-217. Marone, C., 1998. Laboratory-derived friction laws and their application to seismic faulting. Annual Review Earth and Planetary Science 26, 643-696. Martins, J. A. C., Oden, J. T., and Simes, F. M., 1990. A study of static and kinetic friction. International Journal of Engineering Science 28, 29-92. Marusich, T. D., and Ortiz, M., 1995. Modeling and simulation of high speed machining. International Journal for Numerical Methods in Engineering 38, 3675-3694. Mase, C. W., and Smith, L., 1985. Porefluid pressures and frictional heating on a fault surface. Pure Appl. Geophys. 122, 583-607. Mase, C. W., and L., S., 1987. Effects of frictional heating on the thermal, hydrologic, and mechanical response of a fault. J. Geophys. Res. 92, 62496272. Mashimo, T., 1998. Effect of shock compression on ceramic materials In: Davison, L., and Shahinpoor, H., (Eds.), High-Pressure Shock Compression of Solids III. Springer-Verlag NewYork, Inc., New York, NY, pp. 101-146. Mashimo, T., Togo, H., Zhang, Y., Uemura, Y., and Kawamura, Y., 2006. Shock-compression behavior of Zr-based metallic glass In: Khan, A. S., and Kazmi, R., (Eds.), 12th International Symposium on Plasticity and its Applications. Anisotropy, Texture, Dislocations and Multiscale Modeling in Finite Plasticity and Viscoplasticity and Metal Forming. Neat, Inc., MD, USA, Halifax, Nova Scotia, Canada, pp. 157-159. Metha, N., 2003. Pressure and Strain Rate Dependency of Flow Stress of Glassy Polymers under Dynamic Loading, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, OH. Meyers, M. A., 1994. Dynamic Response of Materials. John Wiley & Sons, Inc, New York.
367

Mizoguchi, K., Hirose, T., Shimamoto, T., and Fukuyama, E., 2006. Moisture-related weakening and strengthening of a fault activated at seismic slip rates. Geophysical Research Letters 33. Montgomery, R. S., 1976a. Friction and wear at high sliding speeds. Wear 36, 275-298. Montgomery, R. S., 1976b. Surface melting of rotating bands. Wear 38, 235-243. Moran, B., Ortiz, M., and Shih, C. F., 1990. Formulation of implicit finite element methods for multiplicative finite deformation plasticity. International Journal for Numerical Methods in Engineering 29, 483-514. Nagtegaal, J. D., and DeJong, J. E., 1981. Some computational aspects of elastic-plastic large strain analysis. International Journal for Numerical Methods in Engineering 17, 15-41. Nathenson, D. I., Prakash, V., and Dandekar, D. P., 2005. Dynamic response of silicon nitride under combined pressure and shear impact, Paper # 315 (s22). Proceedings of the 2005 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Bethel, Connecticut, USA, Portland, Oregon USA. Nathenson, D. 2006. Experimental Investigation of High Velocity Impacts on Brittle Materials, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, OH. Noda, H., and Shimamoto, T., 2005. Thermal pressurization and slip-weakening distance of a fault: An example of the Hanaore fault, southwest Japan. Bulletin of the Seismological Society of America 95, 1224-1233. Ogawa, K., 1997. Impact friction test method by applying stress wave. Experimental Mechanics 37(4), 398-402. O'Hara, K., Mizoguchi, K., Shimamoto, T., and Hower, J. C., 2006. Experimental frictional heating of coal gouge at seismic slip rates: Evidence for devolatilization and thermal pressurization of gouge fluids. Tectonophysics 424, 109-118. Okada, M., Liou, N.-S., Prakash, V., and Miyoshi, K., 2001. Tribology of high speed metal-on-metal sliding at near-melt and fully-melt interfacial temperatures. Wear 249, 672-686. Okada, M., Liou, N. S., and Prakash, V., 2002. Dynamic shearing resistance of molten metal films at high pressures. Experimental Mechanics. Park, K. C., and Felippa, C. A., 1983. Chapter 3: Partitioned analysis of coupled systems. In: Belytschko, T., and Hughes, T. J. R., (Eds.), Computational Methods for Transient Analysis. Nort-Holland, Amsterdam, pp. 157-219.
368

Parry, T. V. and Wronski, A. S., 1985. Effect of hydrostatic pressure on the tensile properties of pultruded CFRP. Journal of Materials Science 20 (n6), 2141-2147. Peirce, D., Shih, C. F., and Needleman, A., 1984. A tangent modulus method for rate dependent solids. Computers and Structures 18 (5), 875-887. Peker, A., and Johnson, W. L., 1993. A highly processable metallic glass Zr41.2Ti13.8Cu12.5Ni10.0Be22.5. Applied Physics Letters 63, 2342-2344. Prakash, V., and Clifton, R. J., 1993. Time Resolved Dynamic Friction Measurements in Pressure-Shear. In: Ramesh, K. T., (Ed). Experimental Techniques in the Dynamics of Deformable Bodies, vol. AMD Vol. 165. ASME, New York, pp. 33-47. Prakash, V., 1995. A pressure-shear plate impact experiment for investigating transient friction. Experimental Mechanics 35, 329-336. Prakash, V., 1998. Friction response of sliding interfaces subjected to time varying normal pressures. Journal of Tribology 120, 97-102. Rajagopalan, S., and Prakash, V., 1999. A Modified Torsional Kolsky Bar for Investigating Dynamic Friction. Experimental Mechanics 39(4), 295-303. Ranjith, K., and Rice, J. R., 2001. Slip dynamics at an interface between dissimilar materials. Journal of the mechanics and physics of solids 49, 341-361. Rempel, A. W., and Rice, J. R., 2006. Thermal pressurization and onset of melting in fault zones. Journal of Geophysical Research-Solid Earth 111. Rice, J. R., and Ruina, A., 1983. Stability of steady frictional sliding. Journal of Applied Mechanics 50, 343-349. Rice, J. R., 2006. Heating and weakening of faults during earthquake slip. Journal of Geophysical Research-Solid Earth 111. Richardson, E., and Marone, C., 1999. Effects of normal force vibrations on friction healing. Journal of Geophysical Research 104, 28859-28878. Roig Silva, C., Goldsby, D. L., Toro, G. Di, and Tullis, T. E., 2004. The role of silica content in dynamic fault weakening due to gel lubrication (abstract), 2004 Southern California Earthquake Center Annual Meeting, Proceedings and Abstracts Volume XIV, p. 150. Rudnicki, J. W., and Rice, J. R., 2006. Effective normal stress alteration due to pore pressure changes induced by dynamic slip propagation on a plane between dissimilar materials. Journal of Geophysical Research-Solid Earth 111.

369

Ruina, A., 1983. Slip stability and state variable friction laws. Journal of Geophysical Research 88, 10359-10370. Scholz, C. H., Molnar, P., and Johnson, T., 1972. Detailed studies of frictional sliding of granite and implications for the earthquake mechanism. Journal of geophysical research 77, 6392-6406. Scholz, C. H., 1998. Earthquakes and friction laws. Nature 243, 37-42. Segall, P., and Rice, J. R., 2006. Does shear heating of pore fluid contribute to earthquake nucleation? Journal of Geophysical Research-Solid Earth 111. Shazly, M., 2005. Dynamic Deformation and Failure of Gamma-met PX at Room and Elevated Temperatures, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, OH. Shugarts, 1953. Measuring friction at high speeds. Journal of Franklin Institute 256, 187-189. Sibson, R. H., 1973. Interaction between temperature and pore-fluid pressure during earthquake faulting A mechanism for partial or total stress relief. Nature 243, 66-68. Sierakowski, R. L. C. S. K., 1997. Dynamic Loading and Characterization of Fiber-reinforced Composites. John Wiley & Sons, New York, NY, USA. Simes, F. M. F., and Martins, J. A. C., 1998. Instability and ill-posedness in some friction problems. International journal of Engineering Science 36, 1265-1293. Song, B., Chen, W., and Weerasooriya, T., 2002. Impact Response and Failure Behavior of a Glass/Epoxy Structural Composite Material. In: Wang, C. M., Liu, G. R., and Ang, K. K., (Eds.), Proceedings of the 2nd International Conference on Structural Stability and Dynamics. World Scientific Publishing Co., Singapore, Singapore, pp. 949-954. Spaepen F., 1977. A microscopic mechanism for steady state inhomogeneous flow in metallic glasses. Acta Metall. 25, 407-415. Steif, P. S., 1983. Ductile versus brittle behavior of amorphous metals. Journal of the Mechanics and Physics of Solids 31 (n5), 359-388. Steinberg, D., 1996. Equation of state and strength properties of selected materials. Lawrence Livermore National Laboratory. Sternlicht, B., and Apkarian, H., 1960. Investigation of "Melt Lubrication". ASLE transactions 2, 248-256. Subhash, G., Zhang, H., and Li, H., 2003. Thermodynamic and Mechanical Behavior of
370

Hafnium/Zirconium Based Bulk Metallic Glasses, Proceedings of the International Conference of Mechanical Behavior of Materials (ICM-9). Kenes International, Geneva, Switzerland. Sulem, J., Vardoulakis, I., Ouffroukh, H., and Perdikatsis, V., 2005. Thermo-poro-mechanical properties of the Aigion fault clayey gouge - Application to the analysis of shear heating and fluid pressurization. Soils and Foundations 45, 97-108. Sunny, G. P., Lewandowski, J. J., and Prakash, V., 2006a. Dynamic compression of amorphous and annealed bulk metallic glass, Paper # 349, Proceedings of the 2006 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, St. Louis, MO, USA. Sunny, G. P., Prakash, V., and Lewandowski, J. J., 2005a. Effects of annealing on dynamic behavior of a bulk metallic glass, Paper # IMECE2005-83016, Proceedings of the 2005 International Mechanical Engineering Conference and Exposition, ASME. American Society of Mechanical Engineers, New York, NY, Orlando, FL. Sunny, G. P., Prakash, V., and Lewandowski, J. J., 2006b. Results from a novel insert design for high starin-rate compression of a bulk metallic glass, Paper # IMECE2006-15414, Proceedings of the 2006 International Mechanical Engineering Conference and Exposition, ASME. American Society of Mechanical Engineers, New York, NY, Chicago, IL. Sunny, G. P., Yuan, F., Lewandowski, J. J., and Prakash, V., 2005b. Dynamic Stress-Strain response of a Zr-based bulk metallic glass, Paper # 324, Proceedings of the 2005 SEM Annual Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Portland, Oregon USA. Tay, T. E., Ang, H. G., and Shim, V. P. W., 1995. An empirical strain rate-dependent constitutive relationship for glass-fibre reinforced epoxy and pure epoxy. Composite Structures 33 (4), 201-210. Taylor, L. M., and Flanagan, D. P., 1987. Pronto 2D: A two dimensional transient solid dynamics program. Sandia National Laboratories, Albuquerque, New Mexico. Teodosiu, C., A dynamic theory of dislocations and its applications to the theory of the elastic-plastic continuum. 1970, pp. 837-876. Teodosiu, C., and Sidoroff, F., 1976. A finite theory of the elastoviscoplasticity of single crystals. International Journal of Engineering Science 14, 713-723. Tsai, L., and Prakash, V., 2005. Dynamic response and spall strength of S2-glass fiber reinforced polymer composites, Paper # 322 (s57). Proceedings of the 2005 SEM Annual
371

Conference and Exposition on Experimental and Applied Mechanics. Society of Experimental Mechanics, Bethel, Connecticut, USA, Portland, Oregon USA. Tsai, L., Prakash, V., 2005. Shock response of S2 glass fiber reinforced polymer composites. In: Khan, A.S., Khoei, A.R. (Eds.), 11th International Symposium on Plasticity and its Current applications: Dislocations, Plasticity, Damage and Metal Forming: Material Rsponse and Multiscale Modeling. Neat Press, MD, USA, Kauai, Hawaii, pp. 364366. Tsai. L., 2006. Shock wave structure and spall strength of layered heterogeneous glass/polyer composites, (Ph.D. Dissertation), Case Western Reserve University, Cleveland, Ohio. Tsutsumi, A., and Shimamoto, T., 1997. High-velocity frictional properties of gabbro. Geophysical Research Letters 24, 699-702. Tullis, T. E., 1994. Predicting earthquakes and the mechanics of fault slip. Geotimes 39, 19-21. Tullis, T. E., and Goldsby, D. L., 2003. Flash melting of crustal rocks at almost seismic slip rates. Eos Trans. AGU 84(46), Fall Mtg. Suppl., Abstract S51B-05. Turnbull, D., 1969. Contemp. Phys. 10, 473. Turneaure, S. J., Winey, J. M., and Gupta, Y. M., 2004. Compressive shock wave response of a Zr-based bulk amorphous alloy. Applied Physics Letters 84, 1692-1694. Ullah, H., Irfan, M. A., and Prakash, V., 2007. State and rate dependent friction laws for modeling high-speed frictional slip at metal-on-metal interfaces Journal of Tribology 129, 17-22. Vaidya U. K., Hosur M. V., Kumar P., Mahfuz H., Haque A. and Jeelani S., 1999. Impact damage resistance of innovative functional sandwich composites. In: ASME 1999 Mechanics and Materials Confberence Blacksburg, VA. Vural, M., and Ravichandran, G., 2004. Failure mode transition and energy dissipation in naturally occurring composites. Composites Part B-Engineering 35 (6-8), 639-646. Wallenberger, F. T., Watson, J. C., Li, H., 2001. Glass fibers. In: Miracle, D.B., Donaldson, S.L. (Eds.), . In: ASM HandbookComposites, Vol. 21. ASM, Metals Park, OH, pp. 2734. Weber, G., and Anand, L., 1990. Finite deformation constitutive equations and a time integration procedure for isotropic, hyperealstic-viscoplastic solids. Computer Methods in Applied Mechanics and Engineering 79, 173-202.
372

Weeks, J. D., Beeler, N. M., and Tullis, T. E., Frictional behavior; glass is like a rock. 1991, pp. 457-458. Weertman, J., 1980. Unstable slippage across a fault that separates elastic media of different elastic constants. Journal of Geophysical Research 85, 1455-1461. Wesseling, P., Lowhaphandu, P. and Lewandowski, J. J., 2003. Effects of superimposed pressure on flow and fracture of two bulk amorphous metals. Mat. Res. Soc. Symp. Proc. 754, 275-279. Wibberley, C. A. J., 2002. Hydraulic diffusivity of fault gouge zones and implications for thermal pressurization during seismic slip. . Earth Planets Space 54, 1153-1171. Wibberley, C. A. J., and Shimamoto, T., 2003. Internal structure and permeability of major strikeslip fault zones: the Median Tectonic Line in Mid Prefecture, Southwest Japan, J. Struct. Geol. 25, 5978. Wibberley, C. A. J., and Shimamoto, T., 2005. Earthquake slip weakening and asperities explained by thermal pressurization. Nature 436, 689-692. Williams, K., and Griffen, E., 1964. Friction between unlubricated steel surfaces at sliding speeds upto 750 feet per second. Proceedings institute of mechanical engineers 178, Part 3N, 24-36. Yang, C., Liu, R. P., Zhang, B. Q., Wang, Q., Zhan, Z. J., Sun, L. L., Zhang, J., and Gong, Z. Z., 2005. Void formation and cracking of Zr41Ti14Cu12.5-Ni10Be22.5 bulk metallic glass under planar shock compression. Journal of Materials Science 40, 3917-3920. Yang, C., Wang, W. K., Liu, R. P., Zhang, X. Y., and Li, X., 2006. Damage features of Zr41Ti14Cu12.5Ni10Be22.5 bulk metallic glass impacted by hypervelocity projectiles. Journal of Spacecraft and Rockets 43, 565-567. Zaretsky, E., deBotton, G., and Perl, M., 2004. The response of a glass fibers reinforced epoxy composite to an impact loading. International Journal of Solids and Structures 41 (2), 569-584. Zhang, Z. F., Eckert, J. and Schultz, L., 2003a. Difference in compressive and tensile fracture mechanisms of Zr59Cu20Al10Ni8Ti3 bulk metallic glass. Acta Mat. 51, 1167-1179. Zhang, Z. F., et al., 2003b. Fracture mechanisms in bulk metallic glassy materials. Phys. Rev. Lett. 91 (4-045505), 1-4 . Zhuang, S. M., Lu, J., and Ravichandran, G., 2002. Shock wave response of a zirconium-based bulk metallic glass and its composite. Applied Physics Letters 80, 4522-4524.

373

Zhuk, A. Z., Kanel, G. I., and Lash, A. A., 1994. Glass Epoxy Composite Behavior under Shock Loading. Journal De Physique Iv 4 (C8), 403-407.

374