<%BANNER%>

Mixed-Mode Fracture in Brittle Materials

Permanent Link: http://ufdc.ufl.edu/UFE0043736/00001

Material Information

Title: Mixed-Mode Fracture in Brittle Materials
Physical Description: 1 online resource (140 p.)
Language: english
Creator: Gopalakrishnan, Karthik
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: brittle-fracture -- fractography -- fracture-mechanics -- glass-ceramic -- mixed-mode -- r-curve -- t-stress
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Engineering applications of ceramics often involve mixed-mode loading conditions, where both tensile and shear stresses act simultaneously. Thus, the study of fracture of ceramics under the complex mixed-mode (I/II) loading conditions is of considerable practical interest. Though mixed-mode fracture has been investigated for a wide range of ceramic specimens, there are several aspects of mixed-mode fracture that is still unknown. The present study aims to evaluate the mixed-mode fracture parameters in ceramic specimens, perform fractographic studies on the mixed-mode fracture surfaces and derive an appropriate mixed-mode fracture criterion. Mixed-mode fracture was investigated for three materials, namely soda lime silica glass (amorphous, isotropic material), silicon nitride (polycrystalline material) and mica glass ceramic (R-curve material). Crack geometry effects on mixed-mode fracture in soda lime silica glass disks in diametral compression were first studied. The results indicate that the chevron notch type cracks are more sensitive to mode II loading than surface cracks. Mixed-mode fracture studies on silicon nitride disks showed that silicon nitride had greater fracture resistance in comparison to soda lime silica glass even under mixed-mode loading conditions. Mixed-mode fracture in mica glass ceramic, an R-curve material, from indent cracks in flexure showed a rising crack growth resistance behavior. These results indicate that microstructure and crack geometry influences mixed-mode fracture in ceramics. The existing mixed-mode fracture theories based on the singular stress terms could not adequately explain the mixed-mode fracture of the disks in diametral compression. The conventional minimum strain energy density was modified to incorporate the non singular or T-stress terms. The results from the modified theory could explain the experimental results. The values of the T-stress terms were found to be geometry dependent. The mixed-mode fracture surfaces were characterized by an absence of the mist region and the presence of distinct hackle markings, termed lances. Quantitative fractography studies showed that fracture mechanics and fractography principles remain the same in both pure mode I and mixed-mode loading conditions. The major practical implication of this work is that fracture stress can be evaluated from the fractographic measurement of the branching radius without any prior knowledge of the loading conditions.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Karthik Gopalakrishnan.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Mecholsky, John J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043736:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043736/00001

Material Information

Title: Mixed-Mode Fracture in Brittle Materials
Physical Description: 1 online resource (140 p.)
Language: english
Creator: Gopalakrishnan, Karthik
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: brittle-fracture -- fractography -- fracture-mechanics -- glass-ceramic -- mixed-mode -- r-curve -- t-stress
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Engineering applications of ceramics often involve mixed-mode loading conditions, where both tensile and shear stresses act simultaneously. Thus, the study of fracture of ceramics under the complex mixed-mode (I/II) loading conditions is of considerable practical interest. Though mixed-mode fracture has been investigated for a wide range of ceramic specimens, there are several aspects of mixed-mode fracture that is still unknown. The present study aims to evaluate the mixed-mode fracture parameters in ceramic specimens, perform fractographic studies on the mixed-mode fracture surfaces and derive an appropriate mixed-mode fracture criterion. Mixed-mode fracture was investigated for three materials, namely soda lime silica glass (amorphous, isotropic material), silicon nitride (polycrystalline material) and mica glass ceramic (R-curve material). Crack geometry effects on mixed-mode fracture in soda lime silica glass disks in diametral compression were first studied. The results indicate that the chevron notch type cracks are more sensitive to mode II loading than surface cracks. Mixed-mode fracture studies on silicon nitride disks showed that silicon nitride had greater fracture resistance in comparison to soda lime silica glass even under mixed-mode loading conditions. Mixed-mode fracture in mica glass ceramic, an R-curve material, from indent cracks in flexure showed a rising crack growth resistance behavior. These results indicate that microstructure and crack geometry influences mixed-mode fracture in ceramics. The existing mixed-mode fracture theories based on the singular stress terms could not adequately explain the mixed-mode fracture of the disks in diametral compression. The conventional minimum strain energy density was modified to incorporate the non singular or T-stress terms. The results from the modified theory could explain the experimental results. The values of the T-stress terms were found to be geometry dependent. The mixed-mode fracture surfaces were characterized by an absence of the mist region and the presence of distinct hackle markings, termed lances. Quantitative fractography studies showed that fracture mechanics and fractography principles remain the same in both pure mode I and mixed-mode loading conditions. The major practical implication of this work is that fracture stress can be evaluated from the fractographic measurement of the branching radius without any prior knowledge of the loading conditions.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Karthik Gopalakrishnan.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Mecholsky, John J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043736:00001


This item has the following downloads:


Full Text

PAGE 1

1 MIXED MODE FRACTURE IN BRITTLE MATERIALS By KARTHIK GOPALAKRISHNAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Karthik Gopalakrishnan

PAGE 3

3 To my parents and my loving wife for believing in me

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I sincerely thank my advisor, Dr. John J. Mecholsky Jr for giving me the opportunity to work in his lab and be a part of his research team. His valuable advice and guidance has helped me to learn how to learn and become a successful researcher I really appreciate his under standing and patience which made my journey to obtain my PhD degree a pleasant and memorable one. I thank my supervisory committee members Dr Hassan El Shall, Dr. Wolfgang Sigmund, Dr. Gerhard Fuchs and Dr. Bhavani Shankar. Their suggestions have helped me a lot in writing this dissertatio n. I wou ld also like to thank Dr.Nagaraj Arakere for the discussions that we had regarding the mixed mode fracture in silicon nitride. I thank Robert Ben Lee and Allyson Barrett of the Dental Biomaterials department for t he i r friendliness and help in fabri cating my test specimens. My special thanks to Dr. Kenneth Anusavice, for allowing me to use instruments in the Dental Biomaterials department I would like to thank the MAIC staff especially Dr. Eric Lambers for giving me ample training to use the Atomic Force Microscope I wish to extend my thanks to all the administrative and secretarial staff of Materials Science and E ngine ering department for making my life easier here. I wish to thank my lab mate s especially Matthew Strasberg, for the ir friendship and all the discussions that we had in the lab. I would also wish to thank my fellow graduate students for their help, friendship and fun filled moments I had with them. Last but not the least, I wish to thank my parents Gopalakrishnan K.S. and Shylaja Gopalakrishnan for their love, sacrifices, support and encouragement in every

PAGE 5

5 step of my life. I also thank my wife Parvathy Varma for her dedication, patience and support which motivates me in pursuing my goals.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBR EVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 2 BACKGROUND ................................ ................................ ................................ ...... 22 2.1 Brittle Fracture ................................ ................................ ................................ .. 22 2.2 R Curve Behavior in Ceramics ................................ ................................ ......... 22 2.3 Fractography ................................ ................................ ................................ ..... 24 2.4 Mixed Mode Fracture ................................ ................................ ........................ 27 2.5 Mixed Mode Fracture Theories ................................ ................................ ......... 28 2.6 Statistical Analysis ................................ ................................ ............................ 31 3 CRACK GEOMETRY EFFECTS ON MIXED MODE FRACTURE BEHAVIOR OF SODA LIME SILICA GLASS ................................ ................................ ............. 33 3.1 Experimental Details ................................ ................................ ......................... 35 3.2 Stress Analysis for Disks in Diametral Compr ession ................................ ........ 37 3.3 Mixed Mode Fracture Test Results ................................ ................................ ... 38 3.4 Crack Turning Angles ................................ ................................ ....................... 39 3.5 Expressions for Mixed Mode Stress Intensity Factors for Surface Cracks ........ 40 3.6 Expressions for Mixed Mode Stress Intensity Factors for Chevron Notch Cracks ................................ ................................ ................................ ................. 42 3.7 Mixed Mode Stress Intensity Factor Envelopes ................................ ................ 43 3.8 Summary ................................ ................................ ................................ .......... 45 4 MIXED MODE FRA CTURE IN SILICON NITRIDE ................................ ................. 51 4.1 Experimental Details ................................ ................................ ......................... 52 4.2 Stress Corrosion Effects ................................ ................................ ................... 53 4.3 Biaxial Flexure Test Results ................................ ................................ ............. 54 4.4 Mixed Mode Brazilian Disk Test Results ................................ ........................... 55 4.5 Summary ................................ ................................ ................................ .......... 58

PAGE 7

7 5 MIXED MODE FRACTURE IN AN R CURVE MATERIAL ................................ ..... 64 5.1 Experimental Details ................................ ................................ ......................... 65 5.2 Expressions for Stress Intensity Factors ................................ ........................... 67 5.3 Three Point Flexure Results ................................ ................................ ............. 68 5.4 Evaluation of the Mixed Mode Fract ure Theories ................................ ............. 69 5.5 Discussion ................................ ................................ ................................ ........ 70 5.6 Summary ................................ ................................ ................................ .......... 72 6 FRACTOGRAPHY O F MIXED MODE FRACTURE IN SODA LIME SILICA GLASS ................................ ................................ ................................ .................... 77 6.1 Experimental Details ................................ ................................ ......................... 77 6.2 Fracture Surface Features ................................ ................................ ................ 78 6.3 Stress Intensity at Microbranching ................................ ................................ .... 81 6.4 Implications ................................ ................................ ................................ ....... 82 6.5 Summary ................................ ................................ ................................ .......... 82 7 FRACTOGRAPHY OF MIXED MODE FRACTURE IN AN R CURVE MATERIAL ................................ ................................ ................................ .............. 87 7.1 Experimental Details ................................ ................................ ......................... 88 7.2 Fracture Surface Features ................................ ................................ ................ 88 7.3 Crack to Mirror Size Ratio ................................ ................................ ................. 89 7.4 Stress Intensity at Microbranching ................................ ................................ .... 90 7.5 Effective Mixed Mode Geometry Factors ................................ .......................... 91 7.6 Fractal Dimensional Increment, D* ................................ ................................ ... 95 7.7 Summary ................................ ................................ ................................ .......... 97 8 MODIFIED STRAIN ENERGY DENSITY CRITERION ................................ ......... 101 8.1 Modified Minimum Strain Energy Density Theory ................................ ........... 103 8.2 Application of Modified MSED Criterion to Brazilian Disks .............................. 106 8.3 Results and Discussion ................................ ................................ ................... 107 8.4 Summary ................................ ................................ ................................ ........ 112 9 CONCLUSIONS ................................ ................................ ................................ ... 118 APPENDIX A GEOMETRIC FACTORS AND T* VALUES FOR CENTRALLY CRACKED BRAZILIAN DISK SP ECIMEN ................................ ................................ .............. 122 B MIXED MODE STRESS INTENSITY FACTORS FOR SODA LIME SILICA GLASS DISKS IN DIAMETRAL COMPRESSION ................................ ................ 124 C MIXED MODE STRESS INTENSITY FACTORS FOR MICA GLASS CERAMIC BARS IN FLEXURE ................................ ................................ .............................. 126

PAGE 8

8 D PUBLICATIONS AND PRESENTATIONS FROM THIS THESIS ......................... 131 LIST OF REFERENCES ................................ ................................ ............................. 132 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 140

PAGE 9

9 LIST OF TABLES Table page 4 1 Results from the biaxial flexure test using piston on three ball arrangement. ..... 59 4 2 Mixed mode SIFs from the diametral compression test ................................ ...... 59 5 1 Strengths for as indented and annealed MGC flexure specimens for different crack sizes along with the standard deviations ................................ ................... 73 6 1 Fracture surface and stress measurements ................................ ....................... 83 7 1 Stress intensity at branching, K B2 MGC along with the standard deviation ................................ .............................. 98 7 2 Geometric factors, Y c for the various glass and MGC ................................ ................................ ................................ ... 98 7 3 Effective mixed mode fracture toughness, K c evaluated for differen t ceramic specimens. ................................ ................................ ................................ ......... 98 8 1 Comparison of predictions of the normalized mixed mode stress intensity factors from the modified MTS theory to the experimental values .................... 113 A 1 First five values of T i as function of crack length ratio. ................................ ...... 122 A 2 First five values of S i as function of crack length ratio. ................................ ...... 122 A 3 First five values of Ai. ................................ ................................ ....................... 123 A 4 First five values of B i ................................ ................................ ........................ 123 A 5 Values of T* for different crack orientations and crack length ratios. ................ 123 B 1 Mixed mode stress intensity factors for soda lime silica glass disks with chevron notches. ................................ ................................ .............................. 124 B 2 Mixed mode stress intensity factors for soda lime silica glass disks with as indented surface cracks. ................................ ................................ ................... 125 B 3 Mixed mode stress intensity factors for soda lime silica glass disks with annealed surface cracks. ................................ ................................ .................. 125 C 1 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =90 ................................ ................................ .............................. 126 C 2 Mixed mode stres s intensity factors for mica glass ceramic bars with crack orientation, =80 ................................ ................................ .............................. 127

PAGE 10

10 C 3 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =70 ................................ ................................ .............................. 128 C 4 Mixed mode stress intensity factors for mica glass ceramic bars with cra ck orientation, =55 ................................ ................................ .............................. 129 C 5 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =45 ................................ ................................ .............................. 130

PAGE 11

11 LIST OF FIGURES Figure page 2 1 Variation of fracture toughness with crack size ................................ ................... 31 2 2 Schematic of fracture surface of a brittle material ................................ ............... 32 2 3 Potential modes of loading. ................................ ................................ ................ 32 3 1 Geometry of the chevron notches introduced in the disk specimens. ................. 46 3 2 Optical micrographs of the chevron notch region showing the direction of crack growth. ................................ ................................ ................................ ...... 46 3 3 Schematic of diametral compression specimen ................................ .................. 47 3 4 Grid used for aligning the precracks in soda lime silica glass disks. ................... 47 3 5 Apparatus for the diametral compression test ................................ .................... 48 3 6 Stresses acting in a Brazilian disk ................................ ................................ ...... 48 3 7 Fractured glass disk spe cimen with the chevron notch ................................ ...... 49 3 8 Fracture paths for surface cracks under mixed mode loading.. .......................... 49 3 9 Crack turning angles for glass disks with chevron notch and surface cracks. .... 50 3 10 Comparison of normalized stress intensity factor envelopes for glass disks with chevron notch and surface cracks with the mixed mode fracture theories .. 50 4 1 Silicon nitride ball material in application showing the elliptical contacts ............ 59 4 2 Schematic of piston on three ball configuration ................................ .................. 60 4 3 Fracture surface of silicon nitride showing equiaxed grains. .............................. 60 4 4 ................................ ........... 61 4 5 Crack turning angles for sil icon nitride disks as a function of crack orientation.. ................................ ................................ ................................ ......... 61 4 6 SEM micrograph of silicon nitride fracture surface showing the linkage of multiple indents to form a single semi elliptical surface crack. ........................... 62 4 7 Variation of mixed mode SIFs of silicon nitrid e with crack orientation. ............... 62

PAGE 12

12 4 8 Comparison of normalized mixed mode fracture toughness envelopes for indented silicon nitride di sks to the mixed mode fracture theories. ..................... 63 4 9 Comparison of normalized mixed mode fracture toughness envelopes between glass and silicon nitride Brazilian disk specimens with as indented surface cracks.. ................................ ................................ ................................ .. 63 5 1 SEM micrograph showing microstructure of MGC. ................................ ............. 73 5 2 Schematic of three point flexure specimen with the inclined crack at the center of the tensile surface. ................................ ................................ .............. 74 5 3 SEM micrographs of two different fracture surfaces in the MGC showing the semielliptical cracks. ................................ ................................ ........................... 74 5 4 Fracture toughness of MGC in pure mode I as a function of crack size ............. 75 5 5 Crack turning angle for all crack sizes as function of the crack orientation.. ....... 75 5 6 Mode I a nd Mode II SIFs for mixed mode fracture from different crack sizes in MGC ................................ ................................ ................................ ............... 76 5 7 Effective fracture toughness of MGC, estimated f rom the mixed mode fracture theories for all crack sizes at different mode mixities.. .......................... 76 6 1 Fracture surface of soda lime silica glass. ................................ .......................... 84 6 2 Hackle markings seen on fracture surfaces of soda lime silica glass disks. ....... 84 6 3 Twist hackle marking at crack origin indicated by the arrow. .............................. 85 6 4 Stress intensity at microbranching, K B1 for soda lime silica glass as a function ................................ ................................ ....................... 85 6 5 AFM images at different areas of the mixed mode fracture surface.. ................. 86 7 1 Fracture surfaces of mica glass ceramic.. ................................ .......................... 99 7 2 Schematic of features on fracture surfaces. ................................ ....................... 99 7 3 Crack to mirror size ratio for different mode mixities i n MGC as a function of crack size.. ................................ ................................ ................................ ........ 100 7 4 Fractal dimensional increment, D* for different mode mixities in MGC as a function of crack size. ................................ ................................ ....................... 100 8 1 Crack tip coordinates and two modes of loading acting on the crack. .............. 113

PAGE 13

13 8 2 Comparison of predictions of modified MSED theory for the crack turning angles in the mixed mode fracture of soda lime silica glass disk specimens to experimental r esults presented in Chapter 3. ................................ ................... 113 8 3 Comparison of predictions of modified MSED theory for mixed mode fracture in soda lime sil ica glass disk specimens to experimental results presented in Chapter 3. ................................ ................................ ................................ ......... 114 8 4 Comparison of predictions of modified MSED theory for the crack turning angles in the mixed mode fracture of soda lime silica glass disk specimens to experimental results of Shetty et al. and modified MTS theory by Smith et al. 114 8 5 Comparison of predictions of modified MSED theory for mixed mode fracture in soda lime silica glass disk specimens to experimental results of Shetty et al. ................................ ................................ ................................ ...................... 115 8 6 Comparison of predictions of modified MSED theory for mixed mode fracture in soda lime float glass disk specimens to experimental results of Awaji et al. 115 8 7 Comparison of predictions of modified MSED theory for the crack turning angles in the mixed mode fracture of alumina and CeO 2 TZP disk specimens to experimental results of Singh et al. ................................ ............................... 116 8 8 Comparison of predictions of modified MSED theory for mixe d mode fracture in alumina disk specimens to experimental results of Singh et al. .................... 116 8 9 Comparison of predictions of modified MS ED theory for mixed mode fracture in CeO 2 TZP disk specimens to experimental results of Singh et al. ................ 117

PAGE 14

14 LIST OF ABBREVIATION S BD Brazilian Disk CSERR Coplanar Strain Energy Release Rate MGC Mica Glass Ceramic MSED Minimum Strain Energy Density MTS Maximum Tangential Stress NCSERR Non Coplanar Strain Energy Release Rate SIF Stress Intensity Factor

PAGE 15

15 Abstract of Dissertation Presented to the Graduate S chool of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MIXED MODE FRACTURE IN BRITTLE MATERIALS By Karthik Gopalakrishnan December 2011 Chair: John J. Mecholsky, Jr. Major: Materials Science and Engineering Engineering applications of ceramics often involve mixed mode loading conditions, where both tensile and shear stresses act simultaneously. Thus, the study of fracture of ceramics under the complex mixed mode (I/II) load ing conditions is of considerable practical interest. Though mixed mode fracture has been investigated for a wide range of ceramic specimens, there are several aspects of mixed mode fracture that is still unknown. Th e present study aims to evaluate the mixed mode fracture parameters in ceramic specimens perform fractographic studies on the mixed mode fracture surfaces and derive an appropriate mixed mode fracture criterion. Mixed mode fracture was investigated for three materials, namely soda lime silic a glass (amorphous, isotropic material) silicon nitride (polycrystalline material) and mica glass ceramic (R curve material) C rack geometry effects on mixed mode fracture in soda lime silica glass disks in diametral compression were first studied The re sults indicate that the chevron notch type cracks are more sensitive to mode II loading than sur face cracks. M ixed mode fracture studies o n silicon nitride disks showed that silicon nitride had greater fracture resistance in comparison to soda lime silica glass even under mixed mode loading conditions Mixed mode fracture in mica glass ceramic, an R curve material, from indent

PAGE 16

16 cracks in flexure showed a rising crack growth resistance behavior. These results indicate that microstructure and crack geometry i nfluences mixed mode fracture in ceramics. The existing mixed mode fracture theories based on the singular stress terms could not adequately explain the mixed mode fracture of the disks in diametral compression. The conventional minimum strain energy dens ity was modified to incorporate the non singular or T stress terms. The results from the modified theory could explain the experimental results. The values of the T stress terms were found to be geometry dependent. The mixed mode fracture surfaces we re characterized by an absence of the mist region and the presence of distinct hackle markings, termed lances. Quantitative fractography studies showed that f racture mechanics and fractography principles remain the same in both pure mode I and m ixed mode loading conditions The major practical implication of this work is that fracture stress can be evaluated from the fractographic measurement of the branching radius without any prior knowledge of the loading conditions.

PAGE 17

17 CHAPTER 1 INTRODUCTION Engineering applications involve three potential modes of loading: tensile loading (mode I) in plane shear loading (mode II) and out of plane shear loading (mode III) Structural components of ceramics can fail from cracks or flaws inclined at an y arbitrary angle wit h respect to the overall stresses applied. For many cases, the cracks simultaneously experience two major modes of loading: mode I and mode II. Thus, the study of fracture of ceramics under the complex mixed mode (I/II) loading conditions is of considerabl e practical interest. Though mixed mode fracture in ceramics has been a topic of discussion for quite some time there are several aspects of mi xed mode fracture that has not been investigated. Mixed mode fracture studies to date are restricted to ceramics which have no crack growth resistance effects with crack extension. But there are several ceramics in application s which have crack growth resistance with increasing crack length (R curve behavior). To knowledge there is no literature on the mixed mode fracture behavior in such R curve ceramics. Fractography is a tool used in failure an alysis to identify the fracture origins and fracture stress levels. Quantitative f ractography studies, however, have been perfor med only on sur faces failed in pure tension ( mode I ) One of the major objective s of the present work wa s to evaluate the mixed mode fracture parameters of different ceramic specimens (both R curve and non R curve) and perform fractographic studies on the mixed mode fracture surfaces Under mixed mode loading conditions, the crack is seen to deviate from its initial crack plane and propagate in a direction perpendicular to the maximum tensile stress. The angle by which the

PAGE 18

18 crack deviates from its initial p lane is defined as the crack turning angle and is a significant param eter investigated in the present work The second major object ive wa s to derive a mixed mode fracture criterion that would describe fracture under mixed mode loading conditions and give a n effective value for the mixed mode fracture toughness. The existing mixed mode fracture theories are all based on the singular stress terms (which include the mixed mode stress intensity factors ) and neglects the non singular stress terms (often termed a s T stress) in the series expansion of the stresses ahead of the crack. Williams et al [1], Ueda et al [2] and Smith et al [3] have shown that t he nonsingular stress terms at the tip of the crack, often referred to as the T stresses, have an influence on mixed mode fracture of brittle materials along with the singular stresses, which are characterized by the SIFs, K I and K II The objective is to modify the existing theories to incorporate t he non singular stress terms as well T he re are five maj or aspe cts of this dissertation, the first one being to study the effects of crack geometry on mixed mode stress intensity factors and crack turning angles in soda lime silica glass Second major aim was t o determine the mixed mode stress intensity factors in cer amics with non R curve and R curve behavior. The non R curve ceramic chosen was silicon nitride with equiaxed grains and mica glass ceramic was the representative R curve ceramic. Another major objective of this work was t o perform quantitative fractograp hy studies on mixed mode fracture surfaces. The final aim was t o demonstrate the significance of T stress in mixed mode fracture of brittle materials and develop a three parameter mixed mode fracture theor y which include s the effect of T stress terms

PAGE 19

19 C hapter 2 of the thesis is l iterature review explaining the concepts of brittle fracture, R curve behavior in ceramics, quantitative fractography, mixed mode fracture and mixed mode fracture theories. Chapter 3 deals with the c rack geometry effects on the mixed mode fracture behavior of soda lime silica glass. Soda lime silica glass was primarily chosen to eliminate microstructural effects. The mixed mode stress intensity factors were evaluated for soda lime silica glass disks with s urface cracks and ch evron notch cracks. Chevron notch cracks were found to be more sensitive to mixed mode loading conditions. The extent of deviation of the crack from its initial plane, defined as the crack turning angle showed different trends for each type of crack. M ixed mode stress intensity factors and the crack turning angles for both the type of precracks deviate from the predic tions of the two parameter mixed mode fracture theories Chapter 4 is the discussion of m ixed mode fracture in silicon nitride S ilicon nitr ide disk s with multiple Vickers indents that link up to create a long semi elliptical crack at the center was loaded in diametral compression, to determine the effect of mixe d mode loading on the stress intensity factors. Stress intensity factors were comp uted analytically. Comparisons showed that the stress intensity factors were underestimated by the mi xed mode fracture theories. Microstructural influence and significance of the T stress terms in mixed mode fracture is discussed. Chapter 5 is discussion o f m ixed mode fracture in an R curve material. A m ica glass ceramic was chosen as a representative R curve material. Vickers indentation cracks of varying sizes were introduced on the tensile surface of glass ceramic bars fractured in three point flexure. M ixed mode fracture from indent cracks was governed

PAGE 20

20 by a rising crack growth resistance behavior. Minimum strain energy density theory based on the singular stress terms was effective in describing mixed mode fracture in the glass ceramic and in previously studied non R curve ceramics in flexure. This agreement implies that the effect of T stress terms in mixed mode fracture from relatively small surface cracks in flexure is negligible. Chapter 6 deals with the f ractography of mixed mode fracture in soda li me silica glass The fracture features on the mixed mode fracture surfaces of soda lime silica glass are identified and differentiated from those observed on mode I fracture surfaces. T he mixed mode fracture surfaces are characterized by an absence of the mist region and the presence of distinct twist hackle markings (also called lances ) Quantit ative fractography performed on the mixed mode fracture surface s showed that the stress intensity at microbranching is a constant irrespective of the mode mixity introduced Chapter 7 deals with the q uantitative f ractography of mix ed mode fracture in an R curve material. Mica glass ceramic was the representative R curve material chosen. The parameters investigated were the crack to mirror size ratios, stress intens ity at branching, mixed mode geometric factors and fractal dimensional increments. The crack to mirror size ratios were found to be a function of the crack length and mode mixity. The stress intensity at branching for the mirror hackle transition was a con stant in mixed mode fracture and was less than the corresponding stress intensity at branching in mode I loading. An empirical relationship is derived for the effective geometric factors in mixed mode fracture of ceramics from surface cracks in flexure and the relationship is shown to be effective for calculating the mixed mode fracture toughness of previously studied ceramics fractured in flexure. Fractal dimensional

PAGE 21

21 increments were found to be a function of the crack size but were determined to be a const ant for a specific crack size irrespective of the mode mixity. Chapter 8 is the development of m odified strain energy density theory. A mixed mode fracture criterion based on th e strain energy density concept was derived which takes into account both the s ingular as well as the non singular stress terms in the series expansions for the stress fields ahead of the crack. A good agreement is seen between the experimental values of the crack turning angles and the mixed mode stress intensity factors and the pre dictions from the modified strain energy density theory for ceramic disks in diametral compression. Chapter 9 is an outline of t he major conclusions of this research Experimental details are included indiv idually in each of the chapters 3 7 Suggestions for future work, wherever needed, are included at the end of each chapter. Every chapter concludes with a summary of the results.

PAGE 22

22 CHAPTER 2 BACKGROUND 2.1 Brittle Fracture Fracture occurs in brittle materials at a critical crack size when the stress intensity in mode I reach a critical value, denoted by K IC Mode I or tensile mode refers to the crack opening mode where tensile stresses are applied normal to the crack faces. The critical stress intensity factor (SIF) in mode I is often referred to as the fracture toughness and is a measure of the resistance to crack propagation. The fracture toughness, K IC can be related to the critical crack size, c and the fracture stress le vel, f through Eq uation ( 2 1 ). (2 1 ) where Y is a geometric factor which depends on the loading conditions and crack geometry. Fracture toughness is a material property independent of the crack length, geometry or the loading system, and is usually taken as a constant. However, in certain materials due to the crack microstructure interactions, the fracture toughness values change with the crack length. Such kind of behavior is known as R curve behavior. 2.2 R Curve Behavior in Ceramics Crack microstructure interactions can lead to increasing values of fracture toughness with crack extension, thus improving the fracture resistance in ceramics Figure 2 1 A shows the trend of fracture toughness for non R curve (also known as flat R curve) materials like glass as a function of crack size and Fig ure 2 1 B shows the same for an R curve material The fracture toughness remains a constant f or flat R curve materials irrespective of the crack size whereas the fracture toughness is a

PAGE 23

23 function of crack size for the R curve material. Several interactions between the crack and the microstructure can arrest crack propagation. Some of the mechanisms by which the propagating crack can be shielded include crack bridging [4 5], crack deflection [6 ], m icrocracking [7, 8 ] a nd transformation toughening [9 ] (e.g., in zirconia) Some of these mechanisms are described below. Y ttria stabilized tetragonal zirco nia (Y TZP) has metastable tetragonal grains in its microstructure. When the tetragonal grains encounter a stress field associated with the propagating crack, the tetragonal phase transforms to a monoclinic phase. This martensitic phase transformation is a n energy consuming process, which decreases the driving force for crack propagation. The monoclinic phase expands by 4%which exerts compre ssive stresses on the crack tip, which leads to crack closure. Several studies have shown that t he major toughening mechanism which is responsible for the R curve ef fect is crack bridging [10, 11 ]. Pre vious studies have shown that crack bridging and the R curve effect is observed in brittle materials which have elongated g rains with high aspect ratio [12, 13 ]. Examples of R curve materials with crack bridging are silicon nitride (TSN 03H) [12] and canasite glass ceramic [14] These materials have randomly oriented rod like grains in its microstructure which are interwoven with one another. The propagating crack has to overcome the mechanical interlock force between the grains, thus resisting the grains to be pulled apart. The elongated grains thus act as bridges across the crack faces and exert a closure stress on the crack, thus inhibiting crack propagation and catastrophic failure. There is a critical size beyond which crack bridging does not increase the resistance to crack propagation. Crack bridging effect occurs only for relatively small sized cracks, usually

PAGE 24

24 in the micron range For example, the maximum cr ack length below which crack bridging is effective f or silicon nitride (TSN 03H) [12 d for canasite glass ceramic [14 cracks for evaluating R curve behavior in materials where crack bridging i s the major toughening mechanism There is an upper limit to the crack length, beyond which the R curve reaches a plateau and there is no further resistance to crack propagation. The fracture toughness measured from large crack fracture mechanics specimen s usually fall in the plateau region [15] Thus, to measure R curve behavior, usually small indent cracks are preferred. The most effective way to measure R curve in a material is to measure strength and the critical crack size for the material and compute the fracture toughness corresponding to the particular crack size [12,16] A plot of the fracture toughness with the crack size would be an indication of the rising R curve behavior (as seen in Fig ure 2 1B ) Materials with R curve behavior i.e., increased crack growth resistance with increased crack size, are being used in many engineering applications due to their improved damage tolerance and impro ved thermal shock resistance [17 ]. T oughened zirc onia used as dental ceramics [18 ] is a good example f or an R curve material in application 2.3 Fractography Fractography is an important tool used to identify failure origins and identify features in the fracture process. Fractography studies have shown that the fracture surfaces of materials that fail in a brittle manner from surface cracks are characterized by a sequence of three distinct fracture features surrounding the critical crack which initiated failure. The critical crack is surrounded by a relatively flat and smooth region

PAGE 25

25 called the mirror, which is followed by a slightly tortuous region called the mist, which is then bounded by a highly tortuous region consisting of large radial ridges called the hackle. Macroscopic crack b ranching occurs after the hackle region. A schematic of the features seen on the fracture surfaces of brittle materials are shown in Fig ure 2 2 Depending on the material discipline, i.e., metals, polymers, etc., the name for these regions change, but the characteristics o f the phenomenon are the same. These regions provide useful information about the stress state of the brittle material and are associated with specific stress intensity and crack velocity [19, 20, 21] The presence of stress corrosion can also be determined from the size of fracture features. The distance of the various boundaries from the crack origin is empirically related to the applied stress at fracture [22] : ( 2 2 ) f is the stress at failure, r j corres ponds to the distance from the crack origin to mirror mist boundary (j = 1), mist hackle boundary (j = 2) and crack branching boundaries (j = 3), respectively, M j are a series of constants corresponding to r 1 r 2 and r 3 respectively The mirror mist boundary is usually considered the region at which the microbranching of the crack is first observed under visible light. In reality the crack starts deviating from a plane at the beginning of crack propagation. In the mirror region, the crack deviates l ess than about a quarter of the wavelength of light and is not observed using optical microscopy. However, if other means are used, such as an atomic force microscope (AFM) or a scanning electron microscope (SEM), then deviations are observed [23].

PAGE 26

26 Several criteria have been developed to explain the crack branching phenomena in brittle materials. A stress intensity criterion [22] was proposed by Kirchner and Conway to distinguish the regions of microcrack branching in isotropic materials, i.e., the regions between mirror mist, mist hackle, etc. According to this criterion, microcrack branching occurs at a critical value of the stress intensity factor. The stress intensity criterion is represented by Eq uation (2 3 ): (2 3 ) where K Bj corresponds to the stress intensity factor at r 1 r 2 and r 3 respectively. Y j are geometric constants related to loading and crack geometry and may be different for the three regions [24] Kirchner [25] proposed a st rain intensity criterion to describe the formation of crack branching in brittle materials where he stated that the crack branching s tarts when the strain intensity reaches a critical value The criterion is given by Eq uation ( 2 4 ) (2 4 ) where K j corresponds to the strain intensity factor at r 1 r 2 and r 3 respectively E is the elastic modulus and f is the strain at fracture A third criterion based on the fracture energy was developed to incorporate all the events that led to energy dissipation d uring the fracture process [26] According to the fracture energy criterion, branching boundaries are formed at critical values of the fracture energy. The fracture energy criterion can be stated as follows: (2 5)

PAGE 27

27 (2 6) where j is the value of the fracture energy corresponding to r 1 r 2 and r 3 respectively Tsai et al [27] showed that criteria based on the fracture energy [26] stress intensity [22] and strain intensity [25] are all valid to estimate the position of the fracture mirror boundary and hence all the three criteria can be used to describe microcrack branching in isotropic materials. However, for anisotropic materials, only the fracture energy criterion, was found to best estimate t he position of the fracture mirror formation and hence only the criterion based on fracture energy can be used to describe crack branching in anisotropic materials [27] 2.4 Mixed Mode Fracture Figure 2 3 shows the three potential modes of loading that a crack in a material can experience. Mode I or the tensile mode is the crack opening mode where tensile stresses are applied in the direction normal to the crack faces. Mode II or the in plane shear mode refers to a shear stress applied in the plane of the crack but normal to the leading edge of the crack. Mode III or the tearing mode is the out of plane shear mode where shear stress is applied parallel to the leading crack edge. M ost ceramic structural applications involve a combination of tensile (mode I) and shear (mode II) loading conditions. The knowledge of mixed mode fracture behavior of ceramics is thus, very significant. Mixed mode (I/II) fracture has been investigated for a wide range of ceramic specimens using different test geometries. The test g eometries are selected depending on the shape or geometry of the test specimen. Some of the test specimens used are flexure bars [28 31] angled internal cracked plates [32] compact tension shear

PAGE 28

28 specimens [33] semicircular bend specimens [34 ] and circul ar discs in diametral compression (often referred to as the Brazilian disc test) [35 37 ] T he mixed mode fracture studies presented in this dissertation majorly employs flexure bars and circular discs in diametral compression. Mixed mode loading situations discussed in th is thesis refer to combinations of tensile (mode I) and shear (mode II) loading. Such mixed mode loading conditions involve two SIFs, each corresponding to a particular loading mode. The individual SIFs can be combined to yield an effective value for the mixed mode fracture toughness using mixed mode fracture theories. 2.5 Mixed Mode Fracture Theories Mixed mode fracture theories have been developed primarily to describe crack propagation in brittle materials under mixed mode loadi ng conditions and evaluate an effective mixed mode f racture toughness parameter These fracture criteria are all based on two parameters mode I SIF K I and mode II SIFs, K II The most popular amongst these are the Non Coplanar Strain E nergy Release Rate criterion [38 ] (NCSERR), Maximum Tangential Stress criterion [39 ] (MTS), Coplanar Strain Energy Release Rate criterion [28 30, 40 ] (CSERR) and Minimum Str ain Energy Density criterion [41 ] (MSED). The fracture criteria are based on the assumption that crac k growth starts at a critical value of a parameter related to the strain energy release rate (SERR) or state of stress near the crack tip and propagate along a plane at which the chosen parameter is at its maximum or minimum. The NCSERR criterion assumes that crack growth starts at a critical value of the SERR in a direction along which the SERR is maximum. An analysis based on the maximum SERR was firs t performed by Hussain et al [38 ]. They analyzed the non

PAGE 29

29 coplanar crack extension of inclined cracks bas ed on Griffith energy balances. The condition for fracture under the mixed mode loading in this case is given by: (2 7 ) ( 2 8 ) (2 9 ) (2 10 ) The combination of K I and K II results in effective mixed mode fracture toughness and is given by K C as shown in Equation (2 7 ). Erdogan and Sih [39 ] hypothesized that fracture commences at the tip of the crack in a radial direction and crack growth occurs in the plane which is normal to the direction of maximum tensile stress. This hypothesis implies that crack growth starts at a critical value of the tangential stress along the direction of maximum tangential stress. The tangential stress in the ne ighborhood of the crack tip, in cylindrical coordinates, is given by: (2 11 ) The MTS criterion can be expressed as: at the crack turning angle, (2 12 )

PAGE 30

30 The angle at which the tangential stress reaches a maximum value can be evaluated in terms of K I and K II which is used to derive the general expression for mixed mode fracture, given below. (2 13 ) Equation (2 13 ) represents the conventional MTS criterion. The CSERR criterion states that the crack growth starts along the initial crack plane when the SERR reaches a critical value, leading to the relation below [28 30, 40 ]. (2 14 ) Sih [41 ] proposed that crack growth starts at a critical value of strain energy density, S crit and propagates in the direction of MSED. The S crit value was postulated as a material constant and hence it could be used as a measure of the fracture toughne ss of the material under mixed mode conditions. Based on this concept an effective fracture toughness K C can be derived as (2 15 ) (2 16 ) (2 17 ) (2 18 ) where is the Poisson s ratio of the material.

PAGE 31

31 2.6 Statistical Analysis All statistical analyses were conducted using Minitab and Design Expert Two sample t test was used to determine if there was a ny significant statistical difference or similarity bet ween two mean values significant difference between a group of mean values using ANOVA. A significance for all the statistical comparisons A p 0.05) indicates that there is a statistically significant difference between the values compared A p values compared are statistically the same. Figure 2 1. Variation of fractu re toughness with crack size. A) Fla t R curve. B) R ising R curve ceramics.

PAGE 32

32 Figure 2 2. Schematic of fracture surface of a brittle material Figure 2 3. Potential modes of loading. A) Mode I (tensile mode). B) Mode II (in plane shear). C) Mode III (out of pla ne shear )

PAGE 33

33 CHAPTER 3 C RACK GEOMETRY EFFECT S ON MIXED MODE FRACTURE BEHAVI OR OF SODA LIME SILICA GLASS Mixed mode fracture studies are generally performed on ceramic specimens by introducing precracks. Two approaches are mainly used for generating precracks in ceramic specimen s to de termine its fracture toughness: (1) Controlled surface cracks are introduced through indentations, which generate smaller prec racks in the specimens, and (2) Larger p recracks are introduced through a sharp edged notch or chevron notch for which there is stable crack growth during the initial precracking. There have been quite a numbe r of studies conducted on mixed mode fracture o f ceramics by using either of these approaches. Awaji et al [42] employed the diametral compression test to examine the fracture toughness of graphi te, plaster and marble in mixed mode loading using machined central cracks. Petrovic et al [40] ha ve examined mixed mode fracture toughness in hot pressed s ilicon nitride using surface cracks produced by Kno op indentation in a four point bending test. Khand elwal et al [43] ha ve found the high temperature mixed mode fracture toughness of hot isostically pressed PY6 silicon nitride using four point bend bars which contained indentation cracks produced usin g the Vickers hardness indenter Glaesemann et al [31] Freiman et al [28] and Ikeda et al [44] have examined the mixed m ode fracture toughness of s oda lime silica glass using incline d Knoop indentation cracks. Shetty et al [37] and Awaji et al [45] det ermined the mixed mode fracture toughness of soda lime silica glas s using chevron notched cracks. This literature review shows that mixed mode fracture behavior of different types of precracks has not been studied on the same specimen geometry and material

PAGE 34

34 We propose to study the difference, if any, in the fracture toughness and crack turning angle for mixed mode fracture of the same specimen using two different crack geometries. To the author this is the first time that mixed mode fracture be havior is studied for chevron notch and surface cracks on test specimens of the same material Soda lime silica glass was used primarily to eliminate all microstructural effects, if any. The diametral compression test, often known as the Brazilian Disk (BD) test was employed to study the mixed mode fracture of the diff erent type of precracks in soda lime silica glass disks. The BD test was first proposed by Carniero and Barcellos [46] in 1953 and is commonly used to measure the tensile strength and frac ture toughness of brittle materials such as r ocks, concrete and ceramics. The test uses a circular disk loaded diametrally in compression The diametral loading develops tensile stress es at the centre (perpendicular to th e loading direction) of the disk T he major ad vantage of the BD test is that the test can be performed under a range of mode mixities, in principle, ranging from pure mode I to pure mode II and thus, the fracture toughness in tension, shear and in any combination of tension and shear c an be evaluated. Simple specimen geometry, minimal fixture requirements and its potential for use at higher temperatures are ot her major merits of this test [47 ]. Surface cracks (through indentation) and chevron notch cracks were introduced at the center of th e glass disks tested in diametral compression. The mixed mode stress intensity factors were evaluated for disk specimens with each type of precrack and t h e chevron notch cracks were found to be more sensitive to mixed mode loading. The angle by which the c rack deviates from its initial plane was def ined as the crack turning angle, which was measured and compared for both types of cracks.

PAGE 35

35 3.1 Experimental Details Soda lime silica glass disks of diameter 32 mm and thickness 3 mm was used as the test material for the study. The soda lime glass disks were supplied by Valley Design Corp. and had a standard transparent finish of 80/50 60/40 scratch/dig. Glass disks used for this study were all from the same batch. A set of 75 g lass disks were divided into three groups one third of the set with ch evron notc hes, the next one third with surface cracks containing local residual stress due to the indentation process, and the rest with surface cracks which are relieved from residual stress The geometry of the chevron notched disk specimens used is illustrated in Fig ure 3 1 Chevron notches were machined at the center of the disk using diamond wheels of radius 25 mm. The glass specimens were first loaded in compression i n pure mode I (i.e. with the notches parallel to the loading direction). This initial loading initiated stable crack growth from the tip of the notch. The crack was allowed to grow till the crack length equals a 1 after which the disk was unloaded. Figure 3 2 shows the chevron notch region and the direction of stable crack growth. For crack length s 2c > a 1 crack s grow unstable and rapid fracture occur s Shetty et al [ 47] has determined that this transition from stable to unstable fracture at 2c =a 1 agrees with the prediction from a compliance analysis performed for the chevron notch cracks using an equivalent straight through crack assumption. The precracking was thus primarily perfo rmed to obtain a crack length, 2c close to a 1 so that the straight throug h crack assumption can be used. The initial precracking also relieves the residual stress associated with the machining of the notches, to a certain extent. Singh et al [35] has observed that mode II fracture toughness is signi ficantly underestimated when the mixed mode fracture tests are performed on as notched test specimens. Thus,

PAGE 36

36 precracking of the chevron notched specimens is essential for mixed mode fracture test s. The value of the crack length, 2c was 10 mm after precrack ing and thus the crack length ratio, defined as c/R equals 0.3. The prec racked disk specimens were tested in compression at a loading rate of 250 N/min, by orienting the notch plane at an angle (90 ) with respect to the loading direction, thus creating mi xed mode loading conditions. The schematic of the diametral compression specimen is shown in Fig ure 3 3. The grid with 5 increments that was used for aligning the cracks is shown in Figure 3 4. is the inclination of the notch plane with respect to the h orizontal, i.e. the direction of tensile stress and was varied from 90 to 60 at intervals of 5 A minimum of three disk specimens were tested for each crack orientation. Thus, =90 would correspond to pure mode I loading and <90 would represent mixed mode loading conditions. Surface cracks were introduced at the center of glass disks using multiple Vickers indentations. The indentations were created by a Vickers diamond 2kgf load using a Zwick Hardness Testing Machine The multi ple adjacent Vickers indents were introduced in a coordinated manner such that the individual cracks link up with neighboring cracks to form a single long semi elliptical crack. This single semi elliptical crack increases the probability of failure at the BD center, rather than at the diametral load contacts, during loading. Half of the glass disks indented was annealed at 550 C for 30 minutes to relieve the residual stress associated with the indentation. The penny crack was aligned at the disk center at various angles, ( ) with respect to the horizontal using a grid (Figure 3 4) Schematic of the crack orientation is shown in Figure 3 3. A s indented and annealed specimen s

PAGE 37

37 were then tested in load control at a loading rate of 30 0 N/min in diametral compression Three glass disks were tested for each crack orientation corresponding to a particular condition (as indented or annealed) Samples that failed from the contact points were rejected (~ 10) The semi minor axis or ( in our case ) critical crack depth, a and major axis or the semi elliptical crack width 2b were measured for the samples that failed from the indents at the center The indented cracks can be approximated as semi circular surface cracks, and the equiv alent semi circular crack length or the critical crack size c is given by Equation (3 1) [48, 49 ] (3 1) The apparatus used for testing the glass disks is shown in Fig ure 3 5 The fixtures were designed with slots in them in order to distribute the high stresses at the contacts. This prevents the crushing and inelastic deformation of the specimen at the contacts. 3.2 Stre ss Analysis f or Disks in Diametral Compression A biaxia l stress state is generated in the specimen under diametral compression, as shown in Fig ure 3 6 The biaxial stress field solutions in a thin isotropic disc without a flaw are given by Mitchell [5 0]: ( 3 2 ) ( 3 3 ) (3 4 ) ( 3 5 )

PAGE 38

38 ( 3 6 ) P is the load applied, D is the diameter of the disk specimen and B is t he thickness of the disk specimen. At the cent er of the disk : ( 3 7 ) ( 3 8 ) ( 3 9 ) x (0,0) given by Equation (3 7 ) is the maximum tens ile stress magnitude in the disk For the glass disks used, the value of stress x (calculated from Eq uation (3 2)) remains the same un til the crack tip reaches a distance of 10 mm from the center of the disk The precracked glass disks with the chevron notch have a half crack length, c of 5 mm and thus, the magnitude of the failure stress for the disks with the chevron notch can be estimated from Eq uation ( 3 7 ). 3.3 Mixed Mode Fracture Test Results Diametral c ompression tests were conducted on the glass disk specimens for varying crack angles to vary mode mixity at the crack tip, and the fracture load s were recorded. Figure 3 7 shows a representative glass disk specimen with chevron notch fractured in mixed mode. It can be seen that for crack orientations with mode mixity, the crack deviates from its initial crac k plane and propagates in a direction perpendicular to the maximum normal stress. The same phenomenon is observed for the as indented and annealed samples too. However, the re is a difference in the mechanism by which the crack deviates from its initial pla ne.

PAGE 39

39 Figure 3 8 A shows the fractured disk surface for the as indented glass disks where the crack plane gradually reorients in the direction perpendicular to the maximum normal stress. Thus, for the as indented cracks, though mixed mode loading conditions were introduced initially, the crack at instability fails in pure mode I. This is due to the influence of the residual stress due to indentation. The initial stable crack propagation which leads to the curving of the crack paths is most likely due to stre ss corrosion effects [29]. However for the annealed indent cracks (stress free cracks) the crack deviates abruptly from the initial crack plane (Fig ure 3 8 B ) and there is no stable coplanar crack growth. This indicates that the annealed cracks were subjected to true mixed mode loading conditions. Marshall [29] also found similar results in his study of mixed mode fracture in hot pressed silicon nitride bars. 3.4 Crack Turning Angles For crack orientation s <90 the crack deviate s from its initial plane and propagate s along the direction of loading, which is perpendicular to the maximum normal stress as seen in Fig ure 3 7 The angle by which the crack deviates from its initial plane is defined as the crack turning angle. The crack turning angles are plotted as a function of the crack orientation for the indent cracks and the chevron notch cracks in Fig ure 3 9 There is no significant difference between t he crack turning angles, for as indented and annealed indent surface cracks at a p (p>0.05) For the se surface cracks, to the following relation : ( 3 10 )

PAGE 40

40 The crack turning angles for glass disks with larger chevron notch crack s in diametral compression, are greater than, and follow an entirely different trend from that seen with the crack turning angles for indent cracks (Fig ure 3 9 ) The difference in the crack turning ang les indicate that the three dimensional surface cracks behave differently from the two dimensional chevron notch cracks in mixed mode loading. T he crack turning angles from the predictions of any mixed mode fracture theory deviate greatly from the measured angles for both the indent as well as the chevron notch type cracks The deviation of the experimental from the predictions of MSED theory is much greater for the surface cracks in comparison to the chevron notch cracks. The reason for deviation of the crack turning angles from the conventional mixed mode fracture theories is the influence of non singular stress (T stress) terms in mixed mode loading, the effects of which are neglected in the c onventional mixed mode fracture theories. Difference in the crack turning angles between the chevron notch and surface cracks is due to the different contribution s of the T stress terms for the different crack geometries. T he details about inclusion of the T stress terms will be covered in detail in Chapter 8. 3.5 Expressions for Mixed Mode Stress Intensity Factors for Surface Cracks The analytical expressions for the stress intensity factors (SIF) for annealed surface cracks under mixed mode loadi ng condi tions are given by [36, 51 ]: (3 11 ) (3 12 ) c is the c ritical crack size given by Eq uation ( 3 1 ) Y 1 and Y 2 are geometric constants M 1 and M 2 are the free surface correction factors. For as indented cracks, the mode I

PAGE 41

41 SIF should include the residual stress effects due to indentation and hence an additional factor of 4/3 is included [52] The expression for K 1 fo r as indented cracks is given by Equation (3 13). (3 13) For a crack inclined at an angle with respect to the direction of maximum tensile stress, the tensile stress t and shear stress, are given by [36] : (3 14) (3 15 ) From Eq uation (3 14 ), it can be seen that at =60 the tensile stress is zero and a small surface crack is only subjected to shear stresses, creating pure mode II conditions. At =90 the crack is subjected to pure mode I loading since the shear stress reduc es to zero, as seen from Eq uation ( 3 15 ) For 90 > >60 cracks are subjected to both tensile and shear stresses leading to mixed mode loading conditions at these crack orientations The crack orientations used for this study was therefore in the range from 90 to 60 For <60 the crack experiences compressive stresses. For a semi circular surface crack, the geometric constant, Y 1 is given by [ 5 3 ]: (3 16 ) Kassir and Sih developed the expression for K II f or penny shaped cra cks with the geometric constant Y 2 given by [54 ]: (3 17 )

PAGE 42

42 The value of Y 2 comes out to be 1.28 for the soda lime silica glass which has a Poisson s ratio, = 0.23 [55 ]. Studies conducted by Randall [49] and other researchers [56 ] agree that the correction factor is within 10% 12% for surface cracks. Based on this fact, the value of M 1 is taken to be 1.1 for th e current study. Smith et al [57 ] have shown that the values of M 2 are approximately equivalent to M 1 for equivalent crack sizes and shapes. The ratio M 2 /M 1 is thus taken to be unity. Thus, once the applied stress and the crack dimensions are measured, the SIFs can be calculated from Eq uation (3 11) and Eq uation (3 12 ). 3.6 Expressions for Mixed Mode Stress In tensity Factors for Chevron Notch Cracks The SIFs for the diametral compression specimens with the chevron notch are given by the following expressions [35, 37] : (3 18 ) (3 19 ) f is the failure stress of the disk specimens calculated from the expression for x (0,0) ( Eq uation ( 3 7 )) and c is the half crack length which is 5 mm. The values for the geometric factors, Y 1 and Y 2 have been calculated as a function of the crack length ratio (c/R) and t he crack orientation, by various researchers using different techniques (Atkinson et al [58] Awaji et al [42] Ayatollahi et al [59] ). The solutions from the various techniques agree with each other and hence the geometric factors determined from any technique can be used to evaluate the SIFs. The Y 1 and Y 2 chosen for this study were taken from the paper of Atkinson et al [58] which gives the values for the geometric factors for c/R < 0.6 (Appendix A) The c/ R for the disk specimens was 0.3.

PAGE 43

43 3.7 Mixed Mode Stress Intensity Factor Envelopes The mode I and mode II SIFs were calculated for the glass disks with the surface cracks as well as the chevron notch type cracks the values of which are included in Appendix B The mode I and mode II SIFs evaluated f or the disks with a specific precrack were normalized with the critical mode I fracture toughness, K IC of the glass disk with the same type of precrack. The normalized mixed mode SIFs are graphed in Fig ure 3 10 The critical mode II SIF, K IIC can be determ ined from the x intercept of the normalized fracture toughness envelope. The ratio for the critical mode II SIF, K IIC to the critical mode I SIF, K IC for the glass with chevron notches was determined to be 1.2 which is in close agreement with the values obtained by Awaji et al [42] and Shetty et al [37] for soda lime silica glass disks tested with chevron notches. As can be seen from Fig ure 3 10 as indented and annealed surface cracks shift towards greater values of K II in compari son to the chevron n otch cracks. K IIC /K IC for the su rface cracks was found to be 1.4 5 This indicates that mode II conditions had more of an effect on the chevron notch (or larger) cracks than on the sur face cracks. The reason for the difference in sensitivity to mode II l oading for the different crack geometries is because the contributions of the T stress terms are different for different crack geometries The effect of T stress is greater for disk specimens with relatively small surface cracks, in comparison to disks wit h chevron notch cracks, which leads to the difference in the mixed mode SIF envelopes for the different crack geometries. Details will be covered later in Chapter 8. Normalized mixed mode SIF envelopes for as indented and annealed surface cracks are the same and overlap with each other This trend is different from that seen for silicon nitride flexure bars by Marshall [29] Silicon nitride flexure bars, with annealed

PAGE 44

44 indent cracks, however was seen to give a greater value for K II when compared with silic on nitride flexure bars with as indented cracks. From acoustic scattering results, Marshall showed that annealed indent cracks have crack surfaces in contact at the asperities. Shear is resisted at the contact points due to which the shear loading is not e ffective in producing stress intensification around the crack tip, which reduces the actual level of K II for the silicon nitride [30] Eq uation (3 12 ) calculate s the values of K II assuming the surfaces are traction free. Thus, K II in the case of silicon ni tride (or for any polycrystalline ceramic) would be overestimated if K II is calculated using Eq uation (3 12 ). For soda lime silica glass, which is free from any microstructural effects, there are no asperities on the crack surfaces large enough to come in contact with each other. Thus, Eq uation (3 12 ) does not provide any overestimates for the values of K II for the annealed indent cracks and hence there is no significant difference from the K II values obtained for as indented cracks. As seen from Fig ure 3 10 ensity (MSED) theory and the Coplanar Strain Energy Release Rate theory (CSERR) have the closest predictions to the experimental results in comparison to the Non Coplanar Strain Energy Release Rate (NCSERR) theory and the Maximum Tangential Stress (MTS) theory. However, n one of the 2 parameter mixed mode fracture theories (described i n Sec tion 2.5 ) can adequately explain t he results from the disk tests for both the types of precra c ks Shetty et al [ 37] had suggested that t he influence of the non singular stress terms was the possible reason for the overestimation of mixed mode SIFs in soda lime silica glass di sks with chevron notch cracks Recently, there have been discussions on the incorporation of non singular stress (or T stress) terms at the crack tip, and thus

PAGE 45

45 modifying the existing two parameter (K I and K II ) fracture criteria into three parameters (K I K II and T) fracture criteria. Ayatollahi et al [ 59, 60 ] has found that the values of the T s tress terms for ceramic disk specimens with a through crack ( under mixed mode loading conditions ) is significant and cannot be neglected. Further discussions about T stress es are i n cluded in C hapters 4, 5 and 8 Chapter s 4 and 5 discuss the significance of test geometry in estimating the T stress values A three parameter mixed mode fracture theory based on the strain energy density have been developed in Chapter 8 and it is found that t he three parameter theory give s better predictions about crack turning angles and the mixed mode fracture parameters, K I and K II for the glass disks with chevron notches 3.8 Summary Chevron notches and surface cracks were i ntroduced at the center of soda lime silica glass disks, which were tested in diametral compression. The crack turning angles and the mixed mode stress intensity factors were compared for both type s of precracks. Residual stress played an important role in crack propagation for disks with surface cracks Surface cra cks show ed less sens itivity to mode II loading than chevron notch cracks. Mixed mode s tress i n tensity factor envelopes for as indented disks and disks with annealed indents wer e identical. Crack geometry was seen to have an influence on the mixed mode frac ture in soda lime silica glass. The direction of extension of surface cracks is much less than the corresponding extensions for chevron notch cracks. The crack turning angles and the stress intensity factors for both types of precracks deviate d from the predictions of the mixed mode fracture theories.

PAGE 46

46 Figure 3 1. Geometry of the chevron notches introduced in the disk specimens Figure 3 2. Optical micrographs of the chevron notch region showing the direction of crack growth.

PAGE 47

47 Figure 3 3. Schematic of diametral compression specimen Figure 3 4. Grid used for aligning the precracks in soda lime silica glass disk s

PAGE 48

48 Figure 3 5 Apparatus for the diametral compression test Figure 3 6 Stresses acting in a Brazilian disk

PAGE 49

49 Figure 3 7 Fracture d glass disk specimen with the chevron notch Figure 3 8 Fracture paths for surface cracks under mixed mode loading. A) As indented glass disk specimen B) Annealed glass disk specimen

PAGE 50

50 Figure 3 9 Crack turning angles for glass disks with chevron notch and surface cracks (Curves are the best polynomial fits) Figure 3 10 Comparison of n ormalized stress in tensity factor envelopes for glass disks with chevron notch and surface cracks wi th t he mixed mode fracture theories (Curves are the best polynomial fits)

PAGE 51

51 CHAPTER 4 MIXED MODE FRACTURE IN SILICON NITRIDE S ilicon nitride is used in hybrid ball bearing applications because it has several desirable properties including high hardness, a lo w coefficient of thermal expansion, corrosion resistance and a third of the d ensity of common bearing metals [ 61 ]. S ilicon nitride balls used in high performance hybrid ball bearings are susceptible to failure from fatigue spalls emanating from pre existing partial cone surface cracks generated during the manufacturing process The balls are subjected to Hertzian rolling contact loading, leading to compressive and shear traction loading at the contact surface along with tensile stresses that exis t at the periphery of the elliptical contact. Figure 4 1 shows a silicon nitride ball in appli cation with elliptical contacts The surface cracks are therefore subject to complex mixed mod e rolling contact fatigue loading potentially leading to crack propa gation, surface fatigu e spalls, and bearing failure [62, 63 ]. There is considerable interest in the hybrid ball bearing industry towards evalua ting a critical flaw size or the largest allowable manufacturing flaw in the ball that does not grow under servi ce conditions. The critical flaw size is used for defining limits for non destructive evaluation methods for silicon nitride ball quality control and for developing a fracture mechanics based life prediction methodology for hybrid bearings. An important co mponent in evaluating the critical flaw size is the determination of critical mixed mode stress intensity factor (SIF) for silicon nitride ball material [64, 65] This material property in combination with fracture mechanics principles is used to evaluate/ compute the critical flaw size for a ball subject to specific contact fatigue loading conditions.

PAGE 52

52 The aim of the present study is to determine the mixed mode fracture toughness of NBD300 silicon nitride disks. The results generated are of immediate enginee ring relevance to the hybrid ball bearing industry since NBD300 is among the most widely used aerospace grade silicon nitride material s in gas turbine engine bearings [61] Biaxial flexure studies were first conducted on NBD 300 silicon nitride discs for d eterminin g the mode I fracture toughness The mixed mode SIFs were evaluated using the Brazilian disc or the diametral compression test Since silicon nitride is very difficult to cut or grind, introducing surface cracks via indentation wa s the preferred approach to produce precracks in the specimen [61] Indentation cracks are preferred also because they represent the size of the actual strength controlling surface flaws in ceramics. The study also aims to make a comparison between the predictions of the critical SIFs from four widely used mixed mode fracture theories. SIFs computed using analytical expressions are used for evaluating four mixed mode fracture parameters (from each of the four mixed mode fracture theories) and compared with mixed mode exper imental fracture data. Comparisons are also made between experimentally measured and theoretically predicted crack turning angles. These comparisons provide insight towards identifying effective mixed mode fracture theories for brittle materials an d silico n nitride in particular. 4.1 Experimental Details The material used for the study was NBD 300 silicon nitride disks manufactured by Saint Gobain and supplied by Timken Corp All materials used for this study came from the same batch. Biaxial flexure studie s were performed on NBD 300 silicon nitride disk s of diameter 25 mm and thickness 2 mm to evaluate the fracture toughness or cr itical SIF of the disk in mode I. A Vickers indentation crack was introduced at the ce nter of

PAGE 53

53 the silicon nitride disk using a 1 96 N indent load. The disk s were loaded in biaxial flexure using a piston on three ball configuration with a loading rate of 0.5 mm/minute. A loading piston with a diameter of 2 mm and an effective ring diameter of 15 mm was used for the experiment. A sche matic for the loading configuration is shown in Fig ure 4 2. Silicon nitride disks were tested in d iametral compression to evaluate the mixed mode SIFs. The schematic of the test geometry is shown in Fig ure 3 3. Collinear m ultiple indents were introduc ed on the silicon nitride discs using a diamond Vickers indenter at a 196 N load on the Zwick Hardness Testing Machine The cracks from individual indents link up to form a larger crack which increase s the probability of failure at the BD center, rather t han at the diametral load contacts, during loading. Using a printed grid the indented cracks were align ed at the disc center at various angles, ) with respect to the horizontal axis. The specimen was then tested in load co ntrol at a loading rate of 4 5 0 N/min in diametral compression Figure 4 3 shows t he fracture surface of silicon nitride under the SEM. The microstructure contains mostly equiaxed grains, which indicates that the silicon nitride tested is a flat R curve ma terial. 4.2 Stress Corrosion Effects In order to determine any stress corrosion effects in the material, the indented silicon nitride disc was kept in water and the crack length was monitored after 5 minutes, 10 minutes, 1 hour, 24 hours and 48 hours in o rder to see if there is any effect of slow crack growth. Previous work has shown that there is enough residual stress present at the indent to propagate the crack if stress corrosion due t o the environment is a factor [66 ]. The length of the crack introduc ed to test the stress corrosion susceptibility

PAGE 54

54 was ~ 0.4 mm. There was no change in the crack length observed even after 48 hours. This indicates that for large cracks, there is very little or no effect of water environment on crack propagation 4.3 Biaxia l Flexure Test Results The critical mode I SIF for the NBD 300 silicon nitride discs is first determined using the biaxial flexure test. For biaxial flexure, the fracture stress can be calculated from the fracture load using the following equation [67 ]: (4 1 ) where P is the failure load, is B is the specimen thickness, a is the support ring radius, b is the loading piston radius, and R is the specimen radius. T he strength indentation technique was used to calculate the mode I fracture toughness, K IC (SI) [68 ]. (4 2 ) where E is the elastic modulus of the specimen, H is the fracture stress and Q is the indent load. The value of is averaged to be 0.88 for many materials [68 ]. This estimate adds no more than 10% to the error in the K IC evaluation for a material whose elastic/plastic parameters are totally unknown. The major advantage of this technique is that crack size is eliminat ed as a test variable, instead the indent load is used and can be monitored more easily. Table 4 1 gives the value for the fracture load, fracture stress (calculated from Equation ( 4 1 ) ) and the fracture toughness calculated using the strength indentation

PAGE 55

55 technique The mean K IC (SI) value is 4.83 + 0.31 MPa m 1/2 and compares favorably with the fracture toughness value of 4.850.36 MPa m 1/2 obtained for silicon nitride from the studi ] using a Vickers indented half inch diameter silicon nitride ball in diametral compression in pure mode I loading. 4.4 Mixed Mode Brazilian Disk Test Results Compress ive loading tests were conducted on BD specimens for varying crack at the crack tip, and the fracture load recorded. Figure 4 4 shows representative fractured specimens. The fracture of the sp e ure 4 4 A, occurred along the initial plane of the crack since the loading was close to pure Mode I. However for other crack angles with mixed mode loading, the crack turned from its initial plane and propagated in a direction normal to the maximum pri n cipal stress, as seen in Fig ure 4 4 B The angle by which the crack deviates from its initial plane is de fined as the crack turning angle, The crack turning angles for the various crack orientations, are graphed in Fig ure 4 5 The trend for the crack turning angles is similar to that given by Eq uation (3 10) which is: (4 3) Fig ure 4 4 C ), the discs shattered, very likely because of crack closure. Note that these angles correspond to 35 and 40 respectively, from the loading axis. 3D finite element analysis performed on silicon nitride disks by George et al [61] confirms c rack closure at these angles. Figure 4 6 shows a representative micrograph of the fracture surface with the semicircular surface crack clearly visible. The linking of the multiple Vickers indent cracks can be seen to form the l arger semi

PAGE 56

56 elliptical surface flaw with depth a (semi minor axis) and half length b (semi major axis). The effective crack length can b e estimated from the expression [48, 49] : (4 4 ) The normal and shear stresses actin g on an inclined crack is given by the expressions given in Eq uations (3 14 ) and (3 15 ) respectively. The mixed mode SIFs, K I and K II were calculated using Eq uation s (3 13 ) (3 12 ) respectively. Eq uation (3 13) was used instead of Eq uation (3 11) in order to include the residual stress due to indentation. The value for Y 1 was taken as 1.24. Y 2 was calculated to be 1.29 from Eq uation (3 17) (using for silicon nitride as 0.25 [29] ). The mixed mode SIF values K I and K II for the discs from t he BD test calculated using Eq uation s (3 13 ) and (3 12) are shown in Table 4 2 The values for the critical SIF obtained from the biaxial flexure test using strength indentation, K IC (Equation (4 2) ) were similar to that obtaine d from the BD test using E quation (3 13 ) at =90 With I values decreased and K II values increased, as expected (Fig ure 4 7 ) Figure 4 8 shows the normalized mode I SIF, K I /K IC plotted with the normalized mode II SIF, K II /K IC On extrapolating the normalized fracture toughness envelopes, the value of the apparent K IIC can be determined and in this case approximates K IIC as 1.7 K IC Petrovic [30] had measured K IIC as 1.1 K IC for hot pressed silicon nitride in bend tests using surface cracks from Knoop inde nts. Shetty et al [36] determined K IIC to be equal to 2 K IC for Pyroceram glass ceramic with Knoop indent cracks which was tested in diametral compression. These comparisons show that K IIC is not a consistent material property and depends on the test geom etry and the

PAGE 57

57 st ress state in the test specimen. This also indicates that there is no unique relationship between K IIC and K IC As seen from Fig ure 4 8 t he mixed mode fracture theories underestimate the value of the mixed mode SIFs for the current diametral compression studies on silicon nitride disks Shetty et al [36 ] has shown that ceramic disks with indentation precracks tested in diametral compression, give greater values for K I and K II than those obtained from the four point flexure tests on the same materials. Thus, the normalized fracture toughness envelopes shift for the different test geometries due to the different stress states in the specimens. The comparison of the SIFs from the different test geometries highlights the influence of tes t geometry on mixed mode fracture studies. C entrally cracked square plates give much lesser values for mixed mode SIFs in comparison to the centrally cracked disks [70 ] which again signifies the influence of test geometry As described in Sec tion 3.7 t h e T stress terms for the semi elliptical surface cracks in the disks should be high, due to which the two parameter mixed mode fracture theories underestimate the values of the mixed mode SIFs. knowledge, there are no calculations of the T stress terms for semi elliptical surface cracks in mixed mode loading conditions. T stress calculation for indent cracks is challenging because of the residual stress associated with the indentation process and the three dimensional geometry of th e indents. Further details about incorporation of the T stress terms in the m ixed mode fracture theories are described in Chapter 8. T he stress intensity factor envelopes for the as indented silicon nitride disks from this study are compared with as indented soda lime silica glass disks (from C hapter 3) which were both tested in diametral compress ion (Fig ure 4 9 ) Silicon nitride disk

PAGE 58

58 specimens shift towards greater values of K II when compared to the K II values of soda lime silica glass. Marshall [29] and Petrovic [30] had suggested that polycrystalline materials like silicon nitride resist shear deformation generated from the mode II loading due to the asperities on crack faces. Thus, the effective shear stresses acting on the cracks are lesser th an the applied stress for polycrystalline materials. It has been seen that friction effects vary with mode mixity in polycrystalline materials [36] The lower sensitivity of silicon nitride towards mode II loading could be due to the friction effects arisi ng from the resistance of shear stresses by the crack asperities. Thus, material microstructure plays a vital role in mixed mode loading of ceramics. 4.5 Summary Diametral compression tests were performed on NBD 300 silicon nitride BD specimens in order to evaluate effective fracture toughness under mixed mode loading conditions. SIFs of NBD300 discs were calculated for various mode mixities by using the measured crack dimensions under the load that induced failure and the analytical expressions F our widel y used mixed mode fracture criteria were compared The fracture toughness of the NBD300 silicon nitride discs under pure mode I loading conditions was evaluated to be 4.70.3 MPa m 1/2 The mixed mode fracture theories underestimate d the mixed m ode fractur e toughness for the silicon nitride disks Silicon nitride disks with surface cracks show ed less sensitivity to mode II loading in comparison to soda lime silica glass disks with surface cracks. This indicates the influence of microstructure in mixed mode fracture studies. The critical mode II stress intensity factor is not a consistent material property and is dependent on the test geometry.

PAGE 59

59 Table 4 1. Results from the biaxial flexure test using pi sto n on three ball arrangement. Sample No. Indent load, Q (N) Fracture (MPa) K IC (SI) Strength Indentation (MPa m 1/2 ) 1 2 3 196 196 186 176 151 174 5.1 4.5 4.9 Table 4 2. Mixed mode SIFs from the diametral compression test (deg) Measured crack size, c (mm) Normal (MPa) Shear stress, (MPa) K I (MPa m 1/2 ) K I I (MPa m 1/2 ) 90 90 80 80 80 70 70 70 65 65 65 0.629 0.770 0.760 0.698 0.687 0.303 0.269 0.219 0.471 0.772 0.237 113 97 100 98 104 98 110 120 4 5 36 63 0 0 78 77 81 238 267 291 243 194 340 4.7 4.4 4.5 4.3 4.5 2.8 3.0 2.9 1.6 1.5 1.6 0 0 2.8 2.6 2.7 5.3 5.6 5.6 6.8 6.5 6.8 Figure 4 1. Silicon nitride ball material in application showing the elliptical contacts

PAGE 60

60 Figure 4 2. Schematic of piston on three ball configuration (Court es y: Nicholas Mecholsky) Figure 4 3. Fracture surface of silicon nitride showing equiaxed grains.

PAGE 61

61 Figure 4 4 Figure 4 5 Crack turning angles for silicon nitride disk s as a function of crack orientation Error bars represent standard deviation.

PAGE 62

62 Figure 4 6 SEM micrograph of silicon nitride fracture surface showing the linkage of multiple indents to form a single semi elliptical surface crack Figure 4 7 Variation of mixed mode SIFs of silicon nitride with crack orientation

PAGE 63

63 Figure 4 8 Comparison of n ormalized mixed mode fracture toughness envelopes for indented silicon nitride disks to the mixed mode fracture theories (Curves are best polynomial fits) Figure 4 9 Compar ison of normalized mixed mode fract ure toughness envelopes between glass and silicon nitride Brazilian disk specimens with as indented surface cracks (The curve is best polynomial fit) .. 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 K I /K IC K II /K IC Glass disk with as indented surface cracks Silicon nitride disk with as-indented surface cracks

PAGE 64

64 CHAPTER 5 MIXED MODE FRACTURE IN AN R CURVE MATERIAL Previous studies on mixed mode fracture were focused on materials which have a constant value for fracture toughness and have no crack growth resistance (or R curve) behavior. To the author s knowledge there is no literature on mixed mode fracture studies on materials whi ch show rising crack growth resistance (R curve) behavior with crack extension as a function of crack size This research aims to study the behavior of R curve materials in mixed mode (I/II) loading including varying the crack size A mica glass ceramic (M GC) was chosen as the representative R curve material. The MGC used has wide applications because of its high machinability, high electrical resistivity, zero porosity and capability to withstand high temperatures [ 71 7 3 ]. The areas of application include ultra high vacuum environments such as with electric al insulators, in the aerospace industry, with nuclear related materials with welding nozzles and in industrial high heat electrical cutting operations such as electrode supports. There was an attempt b y Singh et al. [35 ] to determine the effect of mixed mode loading in alumina and CeO 2 TZP ceramic disk specimens which show R curve behavior. However, because of the large chevron notch type cracks introduced in the specimens, no crack growth resistance ef fect could be observed. This emphasizes the significance of using relatively small precracks for assessing R curve behavior in cerami cs (also described in Sec tion 2.2 ).Hence, relatively small indentation precracks were used to measure the R curve behavior in the MGC. The four popularly used fracture criteria (described in Sec tion 2. 5 ) were evaluated for mixed mode fracture in MGC and the results were compared to flexure tests performed on non R curve materials using surface cracks The incorporation of the T

PAGE 65

65 stress terms in the mixed mode fracture theories and the effect of the T stress terms in mixed mode fract ure from controlled surface crack s in flexure are briefly discussed. The objective of the present study was to determine the effects of mode I and c ombined mode (I/II) loading on a commercially available MGC 1 using controlled surface cracks in flexure. The glass ceramic studied contains interlocking mica flakes of high aspect ratio randomly oriented in a borosilicate glass matrix [12, 13, 30]. The mic rostructure of this glass ceramic result in an R curve fracture behavior, i.e., a rising resistance curve with increasing crack size. 5 .1 Experimental Details The MGC used has a composition of 46% SiO 2 17% MgO, 16% Al 2 O 3 10% K 2 O, 7% B 2 O 3 and 4% F [7 3 ]. T he microstructure of MGC shows the elongated grains randomly distributed in the matrix (Fig ure 5 1) The elongated mica crystals are interlocked and dispersed in the glass matrix, which accounts for the easy machinability of MGC bars [ 71 7 2 ]. MGC bars of dimensions 40 mm X 4mm X 3 mm were used for the study. Vickers indent ation s were introduced at the center of the tensile surface of the specimens using 0.05, 0.1, 0.2, 0.3, 0.5, 1 and 2 kg loads in order to create varying crack sizes. Multiple indents were also introduced to create larger cracks. Indents were introduced on the t ensile surface of the MGC specimens at using a grid. Fig ure 5 2 shows the bottom view of the tensile surface of the MGC specimen with an inclined crack represents the cra ck loaded in pure mode I. Mode mixity is introduced at the crack tip The as indented bars were tested in three point flexure 1 Macor, Corning Glass Works, Corning, NY

PAGE 66

66 The flexure tests were conducted at room temperature on a span of 25 mm at a loading rate of 150 N /min. In order to determine the influence of residual stress due to indentation, some of the indented MGC bars (80 bars) wi th different sizes of precracks were a [74, 75 fracture surf aces were examined using an optical microscope to determine the failure origins. The flexural strength, i.e., the stress at calculated from the following equation : ( 5 1) where P is the failure load, L is th e support span, w is the specimen width and d is the specimen thickness. Only samples failed from indentation cracks were examined. Samples that failed from edges or from other inherent flaws in the material were rejected. The semi minor axis in this case the critical crack depth, a and major axis or the semi elliptical crack width 2b were measured for the samples that failed from the indents. The indented cracks can be approximated as semi circular surface cracks, and the equivalent semi circular crack length or the critical crack size, c [c = (a b) 1/2 ; Equation (3 1) ]. Figure 5 3 shows representative micrographs of two fracture surfaces of MGC showing semielliptical surface cracks of the sizes expected in service. Semi elliptical surface cracks are r epresentative of the type of cracks that will be observed in most applications. The strengths of the annealed samples were compared to those of the as indented samples with similar crack sizes.

PAGE 67

67 5 .2 Expressions for Stress Intensity Factors The strengths fo r annealed and as indented samples of MGC are shown in Table 5 1. For similar crack sizes, the strengths for annealed and indented samples were statistically the same (p>0.05). At first this agree ment between the strengths for annealed and as indented cra cks may seem surprising. However, microstructural effects are known to be able to relieve the residual stress from indentation. A major toughening mechanism seen in glass ceramics is microcracking near the crack tip; this microcracking relieves the stres s near the cr ack tip and shields the crack [7 6 ]. The local residual stress from indentation gets relieved due to the microcracking, and hence, as indented samples have similar strengths to the annealed samples. A similar phenomenon was seen to occur in Li 2 O 2SiO 2 glass ceramics [77 ] and canasite glass ceramics [1 4]. The analytical expressions for the stress intensity factors (SIF) for surface cracks under mixed mode loading conditions are given by [3 0, 36 ]: (5 2 ) (5 3) he applied stress given by Eq uation (5 1 ), c is the critical crack size given by Eq uation (3 1 ), Y 1 and Y 2 are the geometric constant and M 1 and M 2 are the free surface correction factors. Similar equations apply for the SIFs for both as indented and annealed samples in the case of glass ceramics as opposed to non R curve materials, wherein the effect of residual stress needs to be accounted for in the as indented samples. The expressions for Y 1 and Y 2 are gi ven by Eq uation s (3 16 ) and (3 17 ) respectively. Y 1 is calculated to be 1.24 and t he value of Y 2 comes out to be 1.32 for the

PAGE 68

68 MGC which has a Poisson s ratio, = 0.29 [7 3 ]. The details for the surface correcti on factors can be found in Sec tion 3. 5 The value of M 1 and M 2 was taken to be 1.1. 5.3 Three Point Flexure Results calculated for the various critical crack sizes is designated as K IC Fracture toughness, K IC for the MGC is shown to be a fu nc tion of the crack size (Fig ure 5 4 ). K IC values increase from 0.7 MPa m 1/2 m 1/2 increase in resistance to crack growth as crack length increases. The value of K IC remains constant at 1.65 MPa m 1/2 as the toughness value of 1.58 + 0.09 MPa m 1/2 obtained for chevron notched flexure specimens of the MGC reported by Reddy et al [78 ](p > 0.05). Crack bridging is mainly responsible fo r the R cur ve behavior in this material [10, 11 ]. With increasing crack sizes, the bridging force exerts a closure stress leading to an increase in crack growth resistance which results in an increased K IC The MGC used has crack bridging effect till the crack length reaches corresponds to pure Mode I. However for other crack orientations, the crack turned from its initial plane and propagated in a direction normal to the maximum principal tensile stress. The angle by which the crack deviates from its initial flaw plane is defined as the crack turning angle. The crack turning angles, were found to be nearly a constant for all the crack sizes tested. However, increases ure 5 5 ) and varies according to the same relation as expressed in Eq uation (3 10 ) i.e

PAGE 69

69 K I and K II (calculated from Eq uation s (5 2) (5 3 )) for the various crack orientations show increasing stress intensity with crack extension un til c ure 5 6 ). This trend of the SIFs indicates that there is resistance to crack growth with increasing crack length in mixed mod e loading conditions. For a particular crack size, mode I SIF, K I II The values of the critical crack sizes, f ailure stress and the mixed mode stress intensity factors for the different crack orientations in the MGC are included in Appendix C. 5 4 Evaluation of the M ixed Mode Fracture T heories The mixed mode fracture theories for non R curve materials have been evaluated based on the fact that mixed mode fracture toughness, K C is a material property not related to the conditions of loading and should be equal to mode I fracture toughness, K IC irrespecti ]. For the MGC, which is an R curve materia l, this criterion should be modified to accommodate the fact that fracture toughness is a function of crack size. The criterion for evaluating the mixed mode fracture theories in R curve materials is that the K C om the theories) should overlap with the K C rather, the K C values (calculated from the theories) for a particular crack size should The effective mixed mode fracture toughness, K C was evaluated using the various theories and graphed as a function of crack size for the various crack orientation (Fig ure 5 7 ). The values of the effective fracture toughness, K C calculated using the NCSERR the ory (Eq uations (2 7) (2 10 )) corresponding to each crack size, at different levels of mode mixity, is statistically different (p<0.05). The curves for the effective fracture

PAGE 70

70 toughness (from the NCSERR theory) as a function of crack size are thus, different for the various crack orientations which con tradicts the fact that the effective fracture toughness value should be a constant for a particular crack size (Fig ure 5 7 A ). The MTS theory (Eq uation (2 13 )) gives negative values for the effective K C the various crack sizes, and hence cann ot be used to describe the mixed mode fracture in this MGC (Fig ure 5 7 B ). The predictions of K C for the various cr ack sizes from the CSERR (Eq uation (2 14)) and MSED (Eq uation s (2 15) (2 18 )) theories are close. The average values of K C for all crack sizes corresponding standard deviations for the two theories are: CSERR (Fig ure 5 7 C ): 1.53 + 0.11 MPa m 1/2 and MSED (Fig ure 5 7 D ): 1.59 + 0.04 MPa m 1/2 .There is no statistical difference between these average K C values (p> 0.05). However, CSERR theory cannot be accepted since this theory assumes the crack to propagate in the same plane as that of initial crack, which is contrary to that observed. The mean fracture toughness from the MSED theory is statistically the same as the fracture toughness value for the MGC from previously published results (1.58 + 0.09 MPa m 1/2 [78 ]) and the standard deviation is less than the CSERR theory. The MSED criterion, thus, is the most applicable fracture criteria in determining the effective fracture toughness in mixed mode loading for MGC bars tested using relatively small cracks and fractured in flexure. This result also implies that a crack in MGC subjected to mixed mode loading conditions starts to grow at a critical value of the strain en ergy density at the crack tip. 5.5 Discussion Previous mixed mode fracture studies on non R curve ceramic flexu re specimens (soda lime glass [28 ] hot pressed silicon nitride [30 ] ) using surface cracks have shown that the crack turning angles follow the trend si milar to Eq uation (3 10 ). A similar

PAGE 71

71 relationship between the crack turning angle and the crack orientation (Eq uation (3 10 )) is also seen for ceramic disk specimens with indent cracks in diametra l compression ( soda lime silica glass (Sec tions 3.4 & Fig ure 3 9 ), Silicon nitride NBD 300 (Sec tion 4.4 & Fig ure 4 4 ) Pyroceram 9606 [36] and hot pressed alumina [36 ]). Thus, irrespective of the test geometry, indent cracks under mixed mode loading condit ions follow the same equation for the crack turning angles as a function of the initial crack orientation A review of literature on the comparison of mixed mode fracture theories in non R curve materials tested in flexure using indentation precracks show that MSED theory provides the best description for mixed mode failure in ceramic flexure specimens. The mixed mode fracture theories are again compared based on the criterion that the fracture toughness should remain a constant at all crack orientations. A study by Freiman et al 2 [28 ] on mixed mode fracture in soda lime glass in four point flexure using Knoop indentation flaws showed that the MSED criterion provided the best fit for the K C values. Petrovic [3 0 ] performed similar test on hot pressed silicon nitride and again came to the same conclusion. The normalized SIFs also agreed well with the predictions from the MSED theory. Marshall [2 9 ] compared the relative failure strengths in mixed mode fracture estimated from the fracture theories to the experim ental values for as indented hot pressed silicon nitride fractured in flexure, and found that the experimental values agree very well to the theoretical predictions from the MSED theory. Glaesemann et al [31 ] also came to a similar conclusion when the exp erimental relative failure strengths for hot pressed silicon nitride (from ref [3 0 ]) and for soda lime glass tested in liquid nitrogen were compared with those from the MSED theory. 2 The angles are mislabeled i n the manuscript by Freiman et al. The values of K c were recalculated.

PAGE 72

72 I and K II ) MSED theory is able to pr ovide constant values for the fracture toughness of the many materials tested in flexure, irrespective of the loading conditions, implies that the effect of the T stress terms for flexure specimens with relatively small surface cracks is negligible, at lea st for the cases considered in this paper. However, for disk specimens with indent cracks, the two parameter mixed mode fracture theories largely underestimate the values of the mixed mode SIFs, possibly because of the greater influence of the T stress in the disk specimens. This indicates that the value of the T stress is dependent on the test geometry and the stress state of the specimen. Future work should aim to determine the T stress for different type of cracks (semi elliptical cracks, straight throug h cracks etc ) in various test geometries under different loading conditions. 5.6 Summary Mica glass ceramic bars with varying indent precrack sizes were tested in three point flexure both in pure mode I and under mixed mode loading states. SIF in pure mo de I for the mica glass ceramic was found to be a function of crack size, which indicated crack growth resistance behavior with increasing crack extension. The mixed mode SIFs were evaluated for the mica glass ceramic and R curve behavior was exhibited in mixed mode loading, similar to that seen in pure mode I. Four popularly used mixed mode fracture theories based on the singular stress terms were investigated : Non Coplanar Strain Energy Release Rate theory, Maximum Tangential Stress theory, Coplanar Str ain Energy Release Rate theory and Minimum Strain Energy Density theory. T he Minimum Strain Energy Density theory provided a very convenient and effective way to describe mixed mode fracture from surface cracks in the mica glass ceramic. Comparison with si milar test results reported for non R curve

PAGE 73

73 materials showed that the Minimum Strain Energy Density theory best described mixed mode fracture in the non R curve materials also Irrespective of the test geometry, crack turning angles for relatively small in dent cracks subjected to mixed mode loading conditions follow: Table 5 1. Strengths for as indented and annealed MGC flexure specimens for different crack sizes along with the standard deviations Strength (MPa) As Indented Annealed 35 115 + 15 117 + 1 0 54 106 + 1 2 105 + 8 97 105 95 + 5 94 + 8 98 + 9 93 + 1 3 122 87 + 6 87 + 5 155 80 + 4 84 + 1 0 170 74 + 6 73 + 7 210 65 + 8 65 + 11 Figure 5 1. SEM micrograph showing microstructure of MGC (taken from fracture surface of MGC)

PAGE 74

7 4 Fig ure 5 2 Schematic of three point flexure specimen with the inclined crack at the center of the tensile surface (bottom view) Figure 5 3 SEM micrographs of two differe nt fracture surfaces in the MGC showing the semielliptical cracks (The arrows border the critical crack)

PAGE 75

75 Fig ure 5 4 Fracture toughness of MGC in pure mode I as a function of crack size (The curve is the best polynomial fit) Fig ure 5 5 Crack turning angle for all crack sizes as function of the crack orientation. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 50 100 150 200 250 K IC (MPa m 1/2 ) Crack size,c (microns)

PAGE 76

76 Fig ure 5 6 Mode I and Mode II SIFs for mixed mode fracture from different crack sizes in MGC. Pure mode I corresponds to best fits to aid the eye ) Fig ure 5 7 Effective fracture toughness o f MGC, estimated from the mixed mode fracture theories for all crack sizes at different mode mixities. A) NCSERR theory. B) MTS theory. C) CSERR theory. D) MSED theory.

PAGE 77

77 CHAPTER 6 FRACTOGRAPHY OF MIXE D MODE FRACTURE IN SOD A LIME SILICA GLASS To date, all studies on fractography were focused on surfaces failed from surface cracks in pure tension or on the quantitative fractographic study of fracture surfaces failed in combined tension and shear (mixed mode) loading. This research aims to study the features of fracture surfaces failed in m ixed mode (I/II) loading and perform fracture surface analyses. The purpose of this paper is to study fractography of room temperature mixed mode fracture from surface cracks on soda lime silica glass, which can be considered as an ideal brittle material. Soda lime silica glass was chosen primarily to eliminate microstructural effects. Surface cracks are introduced on glass disks using indentation with a Vickers diamond indenter and then the disks were tested in mixed mode loading conditions. The fracture s urfaces and the associated measurements in mixed mode loading are compared to those in pure M ode I. The practical implications of these findings are also discussed. 6.1 Experimental Details Multiple indents were introduced at the center of soda lime silica glass disks of diameter 32 mm and thickness 2.5 mm, with a 20 N indent load using a Vickers diamond indenter on a Zwick Hardness Testing Machine. The adjacent Vickers indents were introduced in a coordinated manner such that the individual cracks link up with neighboring cracks to form a single long semi elliptical crack. This single semi elliptical crack increases the probability of failure at the disk center, rather than at the diametral load contacts, during loading. The glass disks were annealed at 550 order to eliminate the residual stresses due to indentation. The specimens were then

PAGE 78

78 tested in load control at a loading rate of 300 N/min in diametral compression by respect to t he horizontal, as shown in Fig ure 3 3 mode I loading. In the crack plane, the tensile stress can be resolved into a normal and a shear stress component leading to mixed mode loadin g conditions at the crack tips. The load to failure is recorded, from which the failure stress, f i.e., strength of th e disk, is calculated using Eq uation (6 1 ). (6 1 ) where P is the load applie d, D is the diameter of th e disk specimen an d B is the thickness of the disk specimen. The fracture surfaces were examined and the distance from the center of the crack to the mirror boundary, defined as the mirror radius, r 1 was measured. Atomic force microscopy (AFM) was perfor med using Nanoscope Dimension 3100 machine at various regions of the mixed mode fracture surfaces in contact mode using a scan size 6.2 Fracture Surface Features The results of measurements of the characteristic markings on the fracture surfaces for flexure bars failed in mode I and mixed mode (I/II) loading are presented in Table 6 1 The fracture surfaces of the glass disk specimens failed in pure mode I have thr ee distinct regions: mirror, mist and hackle (Fig ure s 6 1 A ). However, there was no mist region observed on the fracture surfaces of the glass specimens failed in mixed mode (Fig ure 6 1 B ). The twist hackle markings seen on fracture surfaces of glass failed in mode II are very distinct from those seen on fracture surfaces failed in pure mode I

PAGE 79

79 (Fig ure 6 2 ). Similar twist hackle markings were also seen at the start of crack propagation (Fig ure 6 3 ), when a portion of the crack begins propagating out of the crack plane and eventually return s to the general crack plane. The distinct twist hackle markings, similar to those seen in the present studies have been observed by other researchers. Smekal [79 ] observed such hackle lines in amorphous silica and termed the twist of the features to the medieval lances. Fracture lance features are also formed during the lateral cracking of glass surfaces indented with a hard spherical indenter [80 ] Sommer [81 ] observed lan ces on fracture surfaces of glass rods failed from combined tension and torsion, i.e. mixed mode (I/III) loading conditions and explained that the fracture lances are primarily formed due to the piecewise adjustments of the crack plane to changes in the di rection of the maximum tensile stress, as a result of the superposition of the mode III component. Pons et al [82 ] performed simulations using a continuum phase field method and showed that under mixed mode (I/III) conditions, the crack front undergoes he lical deformations which develop nonlinearly to form the segmented array of daughter cracks (or lances). AFM performed at several areas in the mirror region of the mixed mode fracture surfaces show that lances start to generate in the mirror region itself at distances close to the mirror twist hackle boundary (Fig ure 6 4 ). Comparison of the images from different areas scanned show that the features are very similar, the only difference being the scale. The length s of the lance features increase as we near regions close to the mirror twist hackle boundary.

PAGE 80

80 Beauchamp [83 ] came to a similar conclusion about the appearance of fracture surface features in the mirror region for Mode I loading of lightly tempered soda lime glass fracture surface. He found that f eatures at different regions on the fracture surface are similar but different in size at different length scales. The features found in the mirror region (1/3 rd of the way to th had l ong rounded ridges which differ in appeara nce from the lance like features seen in the mirror region of the mixed mode fracture surfaces. This difference arises due to the different loading conditions. For a semi function of the position along the crack front [61] Only modes I and II are dominant at the surface of the crack, whereas along the crack front, modes I, II and III are active. At the maximum crack depth, only modes I and III are dominant and there is no effect of th e mode II stress intensity factor [84 ] We think that the distinct hackle markings (or lances) seen on the glass disks with inclined semi elliptical cracks, is due to the influence of the mode III component along the crack front. Pons et al [82 ] showed t hat the initial instability wavelength is directly related to the mode I and mode III stress intensity factors and the size of the process zone in front of the crack tip. There is a balance between the destablizing effect of the far field stresses and the stabilizing effect of the forces on the process zone. They estimated that the process zone size was in the range of 10 100 nm. This corresponds to a wavelength of 20 microns for the fastest growing instability, which agrees reasonably well with that obse rved by Sommer (40 microns) [81 ]

PAGE 81

81 6.3 Stress Intensity at M icrobranching Due to the absence of the mist region in the mixed mode fracture surfaces of glass, the mirror radii were measured from the crack origin to the mirror twist hackle boundary. The stres s intensity at microbranching, K B1 would correspond to the mirror mist region for pure mode I and mirror twist hackle region in the case of mixed mode loading. Increasing mode mixity in the soda lime silica glass disks resulted in an increase in fracture stress and a decrease in the mirror size (Table 6 1). K B1 calculated for both cases of pure mode I and mixed mode using Eq uation (2 3 ),was determined to be a constant (Fig ure 6 4 ), the value of which is equal to 2.3 + 0.1 MPa m 1/2 This value is statistically the same as the previously reported value of 2.4 MPa m 1/2 for K B1 of soda lime silica glass disks failed in pure mode I [85 ] (p>0.05). The stress intensity at micro crack branching was determined to be a constant for soda lime silica glass. The value of the stress intensity at branching in mixed mode loading (corresponding to mirror hackle transition) is statistically the same as that in pure mode I (corresponding to mirror mist transition). The stress intensity criterion thus, remains valid to explain the crack branching phenomenon in mixed mode loading conditions. The mirror radius (distance from crack origin to the mirror hackle boundary) in mixed mode loading is related to the far field stress through the same relation as see n for pure m ode I loading (Eq uation (2 2) or Eq uation (2 3 )). This implies that fractographic principles remain the same in both pure mode I and mixed mode loading. The fracture stress in mixed mode conditions can thus be estimated from the knowledge of the stress int ensity at branching and the branching radius. Thus, the result

PAGE 82

82 of this work supports the evidence that forensic analysis can be used without a priori knowledge of the loading conditions for materials failed in a brittle manner. 6.4 Implications The absence of the mist region and the presence of the characteristic twist hackle markings (or lances) on the fracture surfaces can be used to identify surfaces which fail in combined mode loading. This can be used as a powerful characterization tool to identify chi pping failures [86 ] and sample misorientations while testing. Identification of sample misorientations would help to take into account the orientation effects and make the analysis more accurate. Tsai et al [27] performed mode I three point flexure tests on single crystal silicon bars for two orientations: (1) the {100} tensile surface and the {110} fracture plane and, (2) the {110} tensile surface and {110} fracture plane; and observed the lances and only two boundaries (twist hackle and macroscopic crac k branching) on the fracture surfaces. These features are seen on fracture surfaces of other single crystals such as sapphire and spinel in various orientations [87 ]. The absence of the mist region and the presence of the lances indicate that the single cr ystals were subjected to mixed mode loading in these orientations. Mode mixity is naturally introduced in single crystals because of the crystal anisotropy. Fractography thus provides an important tool to characterize the loading conditions. 6.5 Summary Soda lime silica glass disks with indent precracks were tested under mixed mode loading conditions. T he fracture surfaces ob tained from mixed mode failure we re characterized by an absence of the mist region and the existence of distinct twist hackle markin gs, often referred to as lances. The lances are generated in the

PAGE 83

83 mirror region itself when subjected to mixed mode loading and the fracture features at different regions differ by scale. The stress intensity at microbranching was deter mined to be 2.3 + 0.1 MPa m 1/2 and remains constant at all crack orientations for soda lime silica glass. Fractography and fracture mechanics principles used in pure mode I loading studies apply for mixed mode loading. Table 6 1. Fracture surface and stress measurements Crack Orientation, Crack size, c (m) 10 4 Mirror radius, r (m) 10 3 Stress to failure, Stress intensity at microbranching, K B1 (MPa m 1/2 ) 90 2.8 2.81 35 2.30 90 2.4 2.45 38 2.33 90 2.1 2.32 41 2.45 90 2.3 2.09 39 2.21 80 3.2 1.37 49 2.25 80 3.0 1.35 52 2.37 80 2.0 0.98 55 2.13 80 3.1 1.32 51 2.30 80 4.1 1.48 48 2.29 70 5.5 0.94 63 2.40 70 6.3 0.95 66 2.52 70 5.9 0.94 60 2.28 55 6.2 0.59 75 2.26 55 5.7 0.55 78 2.26 55 6.7 0.62 77 2.38 55 4.9 0.53 81 2.31 55 3.0 0.47 80 2.14

PAGE 84

84 Figure 6 1. Fracture surf ace of soda lime silica glass. A) Mode I failure. B) Mixed mode (I/II) failure Figure 6 2. Hackle markings seen on fracture surfaces of soda lime silica glass disks. A) Mode I failure. B) Mixed mode (I/II) failure. (Arrow indicates hackle markings )

PAGE 85

85 Figure 6 3 Twist h ackle marking at crack origin indicated by the arrow. Fig ure 6 4 Stress intensity at microbranching, K B1 for so da lime silica glass as a

PAGE 86

86 Figure 6 5 AFM images at different areas of the mixed mode fracture surface. A)Mirror region. B) Mirror region. C) Just near the twist hackle region D) Twist hackle region. The white dots are debris and should not be mistaken as fracture surface features The areas are marked as A D in the image I to give an idea of the position of differe nt areas scanned and should not be considered exact locations. The arrows in A D indicate direction of crack propagation.

PAGE 87

87 CHAPTER 7 FRACTOGRAPHY OF MIXE D MODE FRACTURE IN AN R CURVE MATERIAL knowledge, there are no fractography studies on m ixed mode failure in R curve materials using a range of crack sizes This chapter describes the quantitative fractograph ic analysis performed on fracture surfaces of R curve materials failed in mixed mode (I/II) loading. A m ica glass ceramic (MGC) was the R curve material chosen for the study As explained before, Tsai et al [27] has show n that criteria based on the fracture energy [26] stress intensity [22] or strain intensity [25] are all valid to describe microcrack branching in isotropic materials. Mica glass ceramic can be considered an isotropic material for relatively large cracks, such as those considered in this study, and hence any of the three criteria could describe the microcrack branching process. The stre ss intensity criterion (represented by Eq uation ( 2 3 )) is thus used in the present study for the sake of convenience The mist region is quite obvious in glasses and the mirror radius refers to the distance from the crack origin to the mirror mist boundary However, the mist region is more difficult to discern in polycrystalline ceramics so the distance from the crack origin to the mirror hackle boundary is usually mea sured. This is generally termed the outer mirror radius. The crack to mirror size ratio is a very significant and useful parameter used in quantitative fractography to identify slow crack growth and residual stress knowledge, the effect of R curve behavior on the crack to mirror size ratio and the stress intensi ty at microcrack branching, in mixed mode fracture is unknown.

PAGE 88

88 There are several objectives for the work presented in this chapter. The effect of mode mixity and R curve behavior on the crack to mirror size ratio and the stress intensity at microbr anching is determined. T he effective mixed mode geometric factors for the crack orientations tested for the MGC are evaluated and compared to the values of the geometric factors obtained for glass under similar test and loading conditions. An empirical relationsh ip is derived between the effective mixed mode geometric factors and crack orientation. The relationship is validated by calculating the mixed mode fracture toughness values of previously studied ceramics under similar test and loading conditions. Finally, the fractal dimensional increments for the MGC specimens were determined from the crack to mirror size ratios and the effective mixed mode geometric factors. 7.1 Experimental Details The fracture surfaces of MGC bars fractured in three point flexure (as described in Sec tion 5 .1 ) were used for this study. Th e outer mirror radius (r 2 ), semi elliptical crack length (2b) and the semi elliptical crack depth (a) were measured from the fracture surfaces of MGC either using optical microscop y or scanning electro n microscop y for specimens that failed from the indentation site. The indented cracks can be approximated as semi circular surface cracks and an equivalent semi circular crack length can be calculated which will be referred to as the critical crack size (c). The critical crack size, c, is given by the square root of a b (as shown in Eq uation ( 3 10 ) ) 7.2 Fracture Surface Features Typical fracture surfaces of the glass ceramic in pure mode I and mixed mode ( I/II) loading are shown in Fig ure 7 1 As can be seen from the micrographs, the mist region was hard to discern in the MGC. Hence, the outer mirror radius, r 2 was measured as the

PAGE 89

89 distance from the crack origin to the mirror hackle boundary. Looking very closely at the mix ed mode fracture surf aces (Fig ure 7 1B ), we can see that, for the mixed mode loading conditions, the hackle markings are different from those in mode I loading. The features appear as twist hackle markings and have been called lances. Such lance like features are also observed on the mixed mode fracture surfaces in soda lime silica glass (as described in Sec tion 6.2 ) and are thought to be generated due to the influence of mode I and mode III stress intensities acting along the crack front [81 82 ] The schematic of the typical f eatures on a mixed mode fracture surface is compared to the features on mode I fracture surface in Fig ure 7 2 7.3 Crack to Mirror Size Ratio The crack to mirror size ratio (c/r 2 ) as a function of the crack length, c for the different crack orientations i n the MGC are presented in Fig ure 7 3 The R curve behavior of MGC (as shown in Fig ure 5 3 ) is mimicked in the c/r 2 versus c plot for the different crack orientations. Mecholsky et al [12] and Chen [88 ] ha ve shown that c/r is directly proportional to the fracture toughness. Thus, for an R curve material where the fracture toughness is a function of the crack length, c/r 2 values would be expected to increase with increasing crack length as observed This is contrary to the constant values of c/r obtained at a particular crack orientation for non or flat R curve materials. The variation of c/r 2 with crack length at a particular crack orientation for the MGC is most likely due to the crack bridging m echanism which leads to the increasing resistance to crack growth with increasing crack extension in the MGC [10,11] In pure mode I loading, at c constant value, the value of c/r 2 is found to be approximately value of c/r 2 for c 2

PAGE 90

90 [89 ] had graphed the r 2 /c and r 1 /c for mixed mode failure from soda l ime silica glass (flat R curve) and had c o me to the conclusion that the ratio systematically decreases with increasing mode mixity. 7.4 Stress Intensity at Microb ranching Since the mist region is hard to recognize on the fracture surfaces of MGC, the stre ss intensities for the mirror hackle transi tion was calculated using Equation (2 3 ). The stress intensity is denoted by K B2 r 2 is the distance from the crack origin to the mirror hackle boundary. Mecholsky et al. has shown that the values Y j can be assum ed to be 1.24 when r j is measured along the tensile surface [85 ] K B2 was thus calculated and tabulated for the different crack orientations (Table 7 1). The K B2 value for the different crack sizes was found to be a constant both in pure mode I and mixed m ode loading conditions. K B2 mode loading conditions) is statistically the same (p>0.05), however these values are statistically different (p<0.05) from the value of K B2 The st udies on mixed mode fracture in soda lime silica glass have shown that mist region is absent in the mixed mode fracture surfaces. It has been found through quantitative fractography on the soda lime silica glass surfaces (C hapter 6) that the stress intensity at branching for the mirror mist transition in pure mode I loading is equal to the stress intensity at branching for the mirror hackle transition in mixed mode loading. For the case of MGC, since the K B2 value in mixed mode loadi ng is less than the corresponding value in pure mode I, it is probable that the stress intensity at branching for the mirror hackle transition in mixed mode fracture of MGC actually corresponds to the stress intensity at branching for the mirror mist tran sition in pure

PAGE 91

91 mode I fracture. This suggests that mist region is absent in the mixed mode fracture surfaces of MGC. 7.5 Effective Mixed Mode Geometry Factors The geometric factor(s) in all fracture mechanics expressions is(are) extremely important for di stinguishing specific crack and loading geometry. After the initial suggestion of Irwin [49] for very specific conditions, many investigators have spent many hours using finite element analysis [90 ] closed form solutions [53, 54, 91 ] and photoeleastic me thods [92 ] to determine the values for the geometric constants. The values for the geometric factors can be different for the three regions based on the ellipticity of the propagating crack and the loading conditions [24] Newman and Raju [90 ] have evalua ted the geometric factors along the boundaries of semi elliptical surface cracks in a plate or beam subjected to tension or bending loads. However, to the knowledge, geometric factors have not been evaluated as a function of the inclination of sur face cracks with respect to the tensile stress direction. From our studies here, we observed that it may be possible to use the quantitative fractographic observations to obtain the effective geometric factors for mixed mode loading. The effective geomet ric factor, Y c for each crack orientation (or mode mixity) in the MGC can be evaluated from the following standard equation. (7 1 ) where K c is the mixed f i s the failure stress (from E q uation (5 1 )) and c is the critical crack size (from Eq uation (3 10 )). As described in Sec tion 5.4 t he mixed mode fracture toughness in the MGC is best evaluated

PAGE 92

92 Minimum Strain Energy Density theory The values of K c for the MGC were taken to be 1.59 0.04 MPam 1/2 The strengths of the annealed and the as indented samples of MGC hav e been shown in Sec tion 5.3 to be similar, which indicates that the residual stress is relieved in the as indented MGC samples. The value of the geomet ric factor s evaluated from Eq uation (7 1 ), thus, will not have any influence of local residual stress. These values are presented in Table 7 2. The Y c values are also calculated from the data for as indented soda lime silica glass bars tested in flexure in ref [28 ] using Eq uation (7 1 ). The Y c values calculated for the soda lime silica glass would include the effects of residual stresses. Y c values evaluated for the glass and MGC are shown in Table 7 2. The Y c values calculated for both the MGC and the glass are a fu nction of the crack orientation angles and decrease with decreasing crack orientation (Table 7 2). It is interesting to note that the effective geometric factor for the glass bars is greater than that for the MGC bars by a factor of 4/3. This difference is the effect of the residual stress due to indentation because the glass specimens were tested as indented. Marshall et al [52] had also reached a similar conclusion that the geometric factors between specimens with or without residual stress differ by a f actor of 4/3. We can fit the data obtained for the geometric factors from Table 7 2 to a polynomial equation. The following third order polynomials were derived to approximately fit the Y c values evaluated: // Soda lime silica glass (7 2 ) (with residual stress) //MGC (no residual stress) (7 3 )

PAGE 93

93 If th e geometric factors (from Eq uation s (7 2) and (7 3 )) are valid, then the mixed mode fracture toughness calculated using the geometric factors corresponding to all fracture toughness. To validate the above calculations, K c values of different ceramic specimens from the literature tested under similar conditions were calculated using the geometric factors from Eq uation (7 2) and Eq uation (7 3 ). Specimens tested under similar conditions were chosen because test and loading conditions are known to influence geometric factors [24] Mixed mode fracture toughness values were evaluated for three previously studied ceramic specimens fractured in f lexure: silicon nitride (ref [30 ] Petrovic ), Pyroceram glass ceramic (ref [36] Singh et al ) and alumina (ref [36 ] Singh et al f was not explicitly stated in the paper (silicon nitride Petrovic) f was back calculated from the values of the mixed mode stress intensity factors. The specimens considered were all tested a fter annealing, which removed the tensile residual stress due to indentation. Hence Eq uation (7 3 ) was used to evaluate the geometric factors for the different crack orientations considered. Table 7 3 gives the values of K c evaluated for the ceramic spec imens in ref [30 ] (Silicon nitride Petrovic) and ref [36 ] (Pyroceram and alumina Singh et al ) for the various crack orientations. The average value of K c estimate d for silicon nitride in ref [30 ] (Petrovic) c evaluated from Eq uation (6) was 5.4 + 0.5 MPa m 1/2 which is statistically the same as the published value of K IC = 5.48 MPa m 1/2 (ref [30 ]). The average values of K c estimated for the glass ceramic and alumina in ref [36 ] (Singh

PAGE 94

94 et al ) + 0.2 MPa m 1/2 and 4.3 + 0.4 MPa m 1/2 respectively. The values of K c calculated (using the values of Y c evaluated from Eq uation (7 3 )) for a particular ceramic specimen, are statistically the same for all crack orientations. This indicates that the empirical relatio nship between the geometric factors and the crack o rientation, represented by Eq uation (7 3 ) are valid, at least for the cases considered in this paper. The relationships in Eq uation (7 2) and Eq uation (7 3 ) can thus, be used as a first approximation for the effective mixed mode geometric factors f or surface cracks subjected to flexure. Mode I or mode II stress intensity factors, however, cannot be evaluated from these expressions. A similar approach can be employed to derive expressions of the effective g eometric factors for surface cracks in other test and loading conditions. The mode I and mode II geometric factors have been calculated by finite element analysis techniques for through cracks, in various test geometries, as a function of the inclination angle of the cracks and the relative crack lengths ratios [58, 59, 93, 94 ] There have been previous works on the evaluation of stress intensity factors at various points along the crack front for an inclined surface crack subjected to mixed mode loading [ 84, 95, 96 ] evaluation of mode I and mode II geometric factors of surface cracks as a function of the inclination angle of the cracks with respect to the tensile stress direction. Future wo rk should aim to model surface cracks in mixed mode loading conditions and evaluate the mode I and mode II geometric factors for the different crack orientations, or different levels of mode mixities.

PAGE 95

95 7.6 Fractal Dimensional Increment, D* Mandelbrot [97 ] u sed fractal geometry to quantitatively describe an irregular shape or surface. Fractal geometry is a non Euclidean geometry which is self similar and scale invariant Self similar objects refer to those which have several features similar to the entire obj ect. Scale invariant objects are those where the features at one magnification look the same at any other magnification level. Fracture in brittle materials can be described as a self similar and scale invariant process and hence can be described quantitat ively by fractal dimensions [98 ] Fracture surfaces have a fractal dimension of 2.D* where D* is known as the fractal dimensional increment. Fractal dimension of 2.1 would denote a relatively flat surface and higher D* components would correspond to higher surface irregularities. D* is thus a quantitative measure of the extent of irregularity of the surface from its Euclidean geometry. Mecholsky et al [98, 99 ] has related t he fracture toughness, K C to D* through the following equation: (7 4 ) The stress intensity at branching, K B is experimentally related to elastic modulus, E by the following expression [100, 101, 102 ] : (7 5 ) From Eq uation s (7 4 ) and (7 5 ), D* can be evaluated from the following expression. (7 6 )

PAGE 96

96 Y b is assumed to be 1.24 as described in Sec tion 7.4 Mecholsky et al has shown that the proportionality constant between a o and b o is approximately 1 for a variety of ceramic materials [103 ] Thus Eq uation (7 6 ) can be modified as: (7 7 ) Y c is evaluated from Eq uation (7 3 ) for the corresponding crack orientation. D* can thus be calculated from the c/r ratio and the values of the effective mixed mode geometric factors. To date, there have been no studies to calculate D* values for the different initial crack sizes in an R curve material. Eq uation (7 7 ) will provide an estimate for the D* values for R curve materials. For the MGC in this study, the outer mirror radius, r 2 wa s measured. Hence the values of c/r 2 are used for the calculation of D* values for the MGC. As shown in Fig ure 7 4 D* is a function of the crack size for the MGC in pure mode I till the crack size reaches 150 m, when D* reaches a constant value of 0.22 This was expected because D* is directly proportional to the fracture toughness which, in turn is proportional to the crack size for an R curve material The R curve behavior is thus mimicked in the D* versus crack size graph This is in contradi ction to the constant values of D* obtained for flat R curve materials like glass. However, D* values were statistically a constant at a specific crack size for all crack orientations (p>0.05) D* varied from 0.05 to 0.22 for the range of crack sizes consi dered. Whether this value has any physical significance is an interesting area for future research. Further studies should aim to actually measure the D* values for the MGC fracture surfaces at various levels of mode mixities using any standard technique l ike the slit island technique [104 ] or the fracture profile technique [105 ] The comparison of the measured D* values with

PAGE 97

97 the D* values calculated from Eq uation (7 7 ) will provide new insights to the application of fractals in failure analysis The results generated would be very interesting and could probably help in the better understanding of how fracture occurs at the atomic level. 7.7 Summary Fracture surface measurements were performed on mica glass ceramic bars fractured in three point flexur e The crack to mirror size ratio was found to be a function of the crack length and mode mixity. For a particular crack length, the crack to mirror size ratio increase s with increasing mode mixity. For a given crack angle, the crack size to mirror size ra tio increases with crack length in this R curve material. The stress intensity at branching for the mirror hackle transition in pure mode I was determined to be 4.1 + 0.1 MPa m 1/2 whereas for mixed mode loading conditions, the stress intensity at mirror tw ist hackle boundary was determined to be 3.78 + 0.05 MPa m 1/2 The geometric factors used for calculation of mixed mode fracture toughness of the mica glass ceramic decrease with increasing level of mode mixity. Empirical relationships were derived for th e effective mixed mode geometric factors used for mixed mode fracture toughness evaluation of ceramic specimens fractured from surface cracks in flexure Fractal dimensional increments were determined to be a function of the crack size for this R curve mat erial. The value of the fractal dimensional increment remains a constant at a specific crack size, irrespective of the crack orientation.

PAGE 98

98 Table 7 1 Stress intensity at branching, K B2 MGC along with the standard deviation Crack orientation, Stress intensity at branching, K B2 (MPa m 1/2 ) 90 4.1 0 + 0.1 0 3.78 + 0.01 3.78 + 0.06 3.79 + 0.05 3.77 + 0.06 80 70 55 45 Table 7 2 Geometric factors, Y c glass and MGC Soda lime silica glass ( Freiman et al ) With residual stress Mica glass ceramic (MGC) Without residual stress Crack orientation, Mixed mode geometric factor, Y c Crack orientation, Mixed mode geometric factor, Y c 90 60 45 30 20 1.65 1.55 1.21 0.79 0.51 90 80 70 55 45 1.24 1.22 1.21 1.07 0.91 Table 7 3 Effective mixed mode fracture toughness, K c evaluated for different ceramic specimen s Silicon Nitride (ref [30 ]) (Petrovic) Glass ceramic (ref [36 ]) (Shetty et al ) Alumina (ref [36 ]) (Shetty et al ) Crack orientation, (degrees) Effective mixed mode fracture toughness, K c (MPa m 1/2 ) Crack orientation, (degrees) Effective mixed mode fracture toughness, K c (MPa m 1/2 ) Crack orientation, (degrees) Effective mixed mode fracture toughness, K c (MPa m 1/2 ) 90 67.5 45 5.3 + 0.4 5.6 + 0.7 5.3 + 0.6 90 65 55 45 40 35 1.6 1.8 1.8 1.6 2.1 2.0 90 67.5 45 40 3.7 4.2 4.5 4.7

PAGE 99

99 Figure 7 1 Fracture surfaces of mica glass ceramic. A) Pure mode I loading. B) Mixed mode (I/II) loading. The inner arrows indicate the critical crack size and the outer arrows indicate the mirror size. Figure 7 2 Schematic o f features on fracture surfaces. A) Pure Mode I loading. B) Mixed mode (I/II) loading

PAGE 100

100 Figure 7 3 Crack to mirror size ratio for different mode mixities in MGC as a function of crack size. (The curves are best polynomial fits) Figure 7 4 Fractal dimensional increment, D* for different mode mixities in MGC as a function of crack size. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0 50 100 150 200 250 Crack to mirror size ratio c/r 2 Crack size(microns) 90 degrees Pure Mode I 80 degrees 70 degrees 55 degrees 45 degrees 0 0.05 0.1 0.15 0.2 0.25 0 50 100 150 200 250 Frractal dimensional increment, D* Crack size(microns) 90 degrees-Pure mode I 80 degrees 70 degrees 55 degrees 45 degrees

PAGE 101

101 CHAPTER 8 MODIFIED STRAIN ENER GY DENSITY CRITERION The mixed mode fracture problem in fracture mechanics is of considerable scientific interest f rom the viewpoint of identifying and establishing a true fracture criterion for brittle materials. Mixed mode fracture theories have been developed primarily to evaluate an effective fracture toughness parameter and describe crack propagation in mixed mode loading conditions. These fracture criteria are all based on two parameters the mode I Stress Intensity Factor (SIF), K I and the mode II SIF K II The most popular amongst these are the Non Coplanar Strain Energy Release Rate criterion (NCSERR), Maxim um Tangential Stress criterion (MTS), Coplanar Strain Energy Release Rate criterion (CSERR) and Minimum Strain Energy Density criterion (MSED) which are discussed in detail in Sec tion 2.5 These fracture theories agree very well with each other for the te nsile or pure mode I condition, but deviate from each other for the mixed mode loading conditions. One of the first experiments o n mixed mode fracture was done on PMMA plate s by Erdogan and Sih in 1963 [39 ]. The MTS theory was found to adequately explain the crack extension direction as well as t he critical loads. Shetty et al [36] used a maximum hoop stress theory to study the mixed mode fracture of a glass ceramic and an alumina ceramic from inclined precracks formed from Knoop indentation using diametr al compression an d four point bend tests The experimental values of stress intensities were close to that calculated from the maximum hoop stress theory [1, 39] But it was observed that the Knoop flaws extended in the direction of the maximum normal tens ion and not in the direction of the maximum hoop stress. Fr e i man et al [28] conducted mixed mode fracture studies on soda lime silica glass bars and calculated the critical

PAGE 102

102 stress intensity factors (SIFs) by combining K I and K II using the NCSERR MTS C SE RR and MSED criteria and found that the MSED theory yielded the best agreement to the K C values obtained using quantitative fractogr a phy Petrovic et al [30 ]. showed that the NCSERR criterion best described the mixed mode fracture o f hot pressed silicon nitride However, the studies of Marshall, which were based on four point flexure tests on hot pressed silicon nitride bars, suggested that MSED and CPSERR crite ria are the most appropriate [29 ]. This review of the mixed mode fracture studies shows that the re is as yet no consen sus on the most effective mixed mode fracture criterion that applies to a range of materials, but that the criteria are application specific. The conventional mixed mode fracture theories consider only the singular terms in the serie s expansions for the stress fields around the crack tip and neglects the non singular stress terms. Experiments by William s et al [1] and Ueda et al [2] have shown that the non singular stress terms, also referred to as the T stress, plays a significant role in mixed mode fracture in brittle materials. However, studies on mixed mode fracture in mica glass ceramic s (Chapter 5) showed that the effect of the non singular stress terms is neg ligible for surface cracks in flexure and the conventional two parameter MSED theory could adequately explain the experimental values of the stress intensity factors. At the same time, s tudies presented in C hapters 3 &4 showed that none of the theories pre dicted the experimentally observed values for the mixed mode SIFs for soda lime silica glass and silicon nitride disks in diametral compression. As explained in Sec tion 3.7 Shetty et al [37] attributed the underestimation of the mixed mode SIFs by the th eories, to the influence of the non singular stress terms.

PAGE 103

103 Smith et al [3 ] modified the MTS criterion to include the effects of both the singular stress terms and the T stress terms on the tangential stress around the crack tip. Recently, Ayatollahi et a l [10 6 ] showed that the use of this modified MTS criterion can explain the greater values of fracture toughness obtained in mixed mode loading of centrally cracked ceramic disk specimens of soda lime glass, sialon, mullite, silicon carbide and porc elain i n diametral compression. However, the major drawback of the MTS criterion is that the fracture process is not described using an energy consideration which has been shown to be the major criterion to determine fracture since fracture is an energy related process [26]. Figures 3 10 and 4 8 show that the predictions from the MSED and CSERR theories are closer to the experimental results in comparison to the MTS or the NCSERR theories. CSERR theory is not logical since it assumes the crack to propagate in the same direction as that of the initial crack plane which is contrary to what is observed. MSED theory seems to be a logical theory since it relates the crack propagation to strain energy density which is an energy related parameter. The MSED theory is m odified here to incorporate the effects of both the singular and non singular stress terms. The predictions from the modified MSED theory are compared to the experimentally obtained values of the stress intensity facto r s and crack turning angles 8.1 Modi fied Minimum Strain Energy Density Theory The derivation for the mo dified MSED theory starts by making use of more accurate expressions for the stress fields ahead of the crack tip. Eftis et al [ 107 ] has given the following expressions for the stre ss fields around the crack tip in polar coordinates :

PAGE 104

104 (8 1 ) (8 2 ) (8 3 ) where r and are the crack tip coordinates shown in Figure 8 1. T he non singular stress (also known as T stress ) terms are present only in the expression for xx T stress represents a constant stress parallel to the crack expressed only in the xx component. There are terms of the order r 1/2 in the expansions of the stress, which are negligible near the crack tip. The total elastic strain energy, dW stored in an elemental volume, dV subjected to 3D stress is given by: (8 4 ) Strain energy density is given by the elastic strain energy per unit volume Under plane strain conditions: (8 5) (8 6) Substituting Eq uation (8 5) and Eq uation (8 6) in Eq uation (8 4), the strain energy density, can be rewritten as given in Eq uation (8 7). (8 7) Substituting the expressions for the stresses, Eq uation (8 7) can be simplified as: (8 8)

PAGE 105

105 where (8 9) (8 10) (8 11) (8 12) (8 13) G is the shear modulus. The strain energy density, can be written in terms of the strain energy density factor, S as: (8 14) (8 15) According to the modified MSED criterion, c rack turning angles or the direction of crack propagation is given when S is minimum. Thus, the crack turning angles can be determined by differentiating Eq uation (8 15) with respect to and (8 16) can be calculated for solved using Eq uation (8 16 ). Using these values of a 11 a 12 a 22 a 33 and a 44 in Eq uation (8 15) would give S critical The fracture toughness, K C is given by [41] :

PAGE 106

106 (8 17 ) Substituting for S critical will give rise to the following equation for K C (8 18 ) The effective mixed mode fracture toughness is now given by K C For pure mode I, K C would be denoted by K IC The conventional M SED criterion has only the first three terms of Eq uation (8 18 ). Equation (8 18 ) represents the modified MSED criterion. r c is the critical distance from the crack tip where the strain energy density reaches its minimum. r c is a material property that is assumed to be independent of the specimen and test geometry. The normalized fracture toughness is given by the following expressions : (8 19 ) (8 20 ) 8.2 Application of Modified MSE D Criterion to Brazilian Disks The mixed mode SIFs for the disks with through cracks are given in Sec tion 3.6 : (8 21 ) (8 22 ) Y 1 and Y 2 are given as a function of the crack size, c, to radius of disk, R, ratio, c/R and the crack inclination angle by several researchers [42, 58, 59] The values for the geometric factors were taken from the work by Atkinson et al [58] where the crack was

PAGE 107

107 represented by a continuo us series of edge dislocations. The failure stress, f is given by: (8 23 ) P is the load applied, R is the radius of the disk specimen and B is the thickness of the disk specimen. The T stress for the BD specimens is given by [59, 60]: (8 2 4 ) T* is the norma lized form of T stress For the calculations used here, the values for T* were taken from Ayatollahi et al [59] where FEA was used to calculate the values of T* The values of T* are given as a function of c/R and in Appendix A The expressions for the normalized fracture toughness envelopes can now be simplified as: (8 25 ) (8 26 ) 8 .3 Results and Discussion The predict ions from the modified MSED criterion are compared to experimental results in this section. Y 1 Y 2 and T values have been computed for BD specimens with through cracks as a function of the crack inclination and c/R ratios For a particular crack ratio, Y 1 Y 2 and T values have been computed for a wide range of crack

PAGE 108

108 orientations, (Appendix A ) At first, the crack turning angle or the angle at which the strain energy density is min imum is determined from Eq uation (8 16 ) for a particular crack orientation Using the value of the crack turning angle determined, a 11 a 12 a 22 a 33 and a 44 can be determined. The normalized fracture toughness values for the particular crack orientation, can then be determined. A similar procedure for a wide range of would generate the normalized fracture toughness envelope for a particular c/R ratio. This would be the predicted loci of fracture toughness from the modified MSED criterion. The calculations can be performed only if the critical distance, r c is known. T he fra cture process zone is the area around the crack tip that affects crack propagation. The size of the fracture process zone can be approximated as the critical distance, r c [ 106, 108 ]. This distance is entirely dependent on the material under consideration. For metals, the size of the process zone can be in the range of millimeters meters. However, the process zone radius for brittle materials or ceramics is found to be much less sometimes in the range of Pons et al [82 ] have estimated the process zone size for soda lime silica glass to be in the range of 10 100 nm. There is no expression, as such, for calculating the radius of the fracture process zone. A model was proposed by Schmidt [ 109 ] for calculating the radius of the process zone in rocks given by: (8 27 ) where t is the tensile strength of the specimen. The model by Schmidt is a plane stress model and assumes a spherical process zone. Ayatollahi et al [59] use d Eq uation (8 27 ) to determine the process zone radius in ceramic specimens. For soda lime silica glass, the value was calculated to be 0.5 mm which is much greater than the estimated process zone size by Pons et al (10 100 nm) The validity of Eq uation (8 27 ) is thus

PAGE 109

109 very qu estionable. This was the other major drawback with the calculations of the modified MTS criterion by Ayatollahi et al. The predictions from the modified MSED theory are compared to the experimental results obtained by using soda lime silica glass disks wit h chevron notches (described in Chapter 3). The disks with the notches were first precracked in pure mode I, till the crack grows up to the base of the notch. Thus, the disks have through cracks when loading in mixed mode loading conditions. Hence, the values of Y 1 Y 2 and T by Atkinson et al and Ayatollahi et al for disk specimens with through cracks can be used. The process zone size, r c used for the calculations for the mixed mode SIFs for soda lime silica glass was 100 nm which gave a better fit to the experimental data than the mixed mode SIFs evaluated using a process zone size of 10nm The crack turning angles and the fracture toughness envelopes from the modified MSED theory is compared to th e experimental values in Fig ure 8 2 and Fig ure 8 3 respectively. Figure s 8 4 and 8 5 show the comparison of the crack turning angles and fracture toughness envelopes from Shetty et al [37] study on BD specimens of soda lime sil ica glass with a c/R ratio of 0.2 There was no statistical difference betwee n the experimental fracture toughness envelopes (Figure 8 3 and Figure 8 5) and the predictions from the modified MSED theory (p>0.05). Figure 8 4 shows that the modified MSED theory deviates from the experimental crack turning angles for increasing levels of mode mixity (i.e. decreasing ), but gives excellent predictions for most values of ( ) A probable reason for the deviation at lower values of is that crack closure occurs as the loading approaches pure mode II conditions ( =60 ), which could influence the crack turning angle. Comparisons to the crack turning angles evaluated from the modified MTS

PAGE 110

110 criterion by Smith et al. [3] is included in Figure 8 4 using the same r c values used for computing the modified MSED criterion. The p redictions from the modified MTS criterion for the crack turning angles are statistically different from the experimentally observed values (p<0.05) The normalized fracture toughness values calculated using the predicted values of the crack turning angles from the modified MTS criterion had large deviations from the experimental results (Table 8 1). Ayatollahi et al. observed closer agreement of the modified MTS criterion to the experimental results possibly due to the greater r c values used for their calc ulations (which were estimated from Equation (8 27)). The experimental res ults of Awaji et al. [45] study on mixed mode fracture in soda lime float glass with c/R ratio of 0.4 are compared to the predictions from the modified MSED in Figure 8 6. The exper im ental results of Awaji et al. are statistically the same as the predictions from modified MSED theory for c/R=0.4 (p>0.05). Figure 8 7 shows the comparisons of the crack turning angles determined from Singh et al [35] study on alumina (c/R=0.5) and CeO 2 TZP (c/R=0.4) disk specimens. Since the critical radius, r c is unknown, different values of r c were used for the calculations. There was only a change of 3 on an average in the values of the crack turning angles when r c was varie d which was not statistically significant (p>0.05) However, the fracture toughness envelopes (Fig ure 8 8 & Fig ure 8 9 ) shifted to greater values of K II at greater values of r c This indicates greater fracture resistance as the radius of the process zone size increases. T he data at r c = 200 m fitted the experimental values for alumina and the values are statistically the same Because of the scatter in the experimental data for CeO 2 TZP specimens, it is hard to point out an exact value of r c which fitted the

PAGE 111

111 experimental d ata. Both values of r c = 60 m and 1 m provided good fits to the experimental data. There was no statistical difference between the predictions at both these values of r c (p>0.05). The results shown in Fig ure s 8 2 to 8 7 indicate that the predictions from the modified MSED theory agrees very well with the experimental results, indicating that the modified MSED criterion is valid to describe the mixed mode fracture behavior in brittle materials (at least for the cases considered in this paper). These results show that the non singular stress terms also play a significant role in the mixed mode fracture in brittle materials testes in diametral compression. The value of the T stresses estimated for the cracked disk specimens is highly negative, the magni tude of which is significant [70 ]. Neglecting the contribution of the T stress values is the reason why the mode II fracture toughness is always underestimated for the disk specimens using the conventional theories. However, the studies on mica glass ceramic (Chapter 5) and other ceramics in flexure showed that the effect of the T stress terms in fracture from surface cracks in flexure is negligible. Other studies have shown that the magnitude of T stress values for angled center cracked plates of sod a lime glass are not as great as in the case of centrally cracked disks of soda lime glass [70 ] T stress has also been estimated for other test specimens and the values have been found to be different for different test geometries [110 ] Thus, magnitude a nd sign of the T stress is dependent on the test geometry and stress state of the specimen. For indent cracks, c/R ap proaches zero and hence Eq uation (8 24 ) can be simplified as: (8 28 )

PAGE 112

112 T* values are found to be highly neg ative and more significant as c/R decreases in a BD specimen. Thus, it can be expected for the indent cracks to have a much greater T stress value, which will explain the greater mixed mode fracture toughness found for disk specimens with indent cracks. Fu ture studies should aim to estimate the T stress for the BD specimen with indent cracks as a function of the crack inclination and mode mixity Calculations of the T stress will help further understand the crack geometry effects in mixed mode fracture. 8.4 Summary The Minimum Strain Energy Density (MSED) theory was modified to incorporate both the singular and the non singular stress (T stress) terms in the series expansions for the stress fields ahead of the crack tip. The predictions from the modified MSE D theory are consistent with the experimental results reported for various Br azilian disk ceramic specimens. This indicates that strain energy density influences crack initiation under mixed mode loading conditions. The T stress is a significant contributi on in mixed mode fracture of Brazilian disc specimens. This contribution explains t he greater fracture resistance observed in the Brazilian disc specimens under mixed mode loading conditions in contrast to three and four point flexure s pecimen s. T stress d epends on the test geometry and the stress state of the specimen. Fracture resistance was found to increase as the size of the process zone increases

PAGE 113

113 Table 8 1. Comparison of p redictions of the normalized mixed mode stress intensity facto rs from the modified MTS theory to the experimental values Predictions from modified MTS theory (Smith et al.) Experimental values (Shetty et al.) K I /K IC K II /K IC K I /K IC K II /K IC 1 0 1 0 0.92 0.47 0.93 0.34 0.57 1.39 0.56 0.86 0.27 1.29 0.23 1.08 0.02 1.62 0.03 1.22 Figure 8 1. Crack tip coordinates and two modes of loading acting on the crack Figure 8 2 Comparison of predictions of modified MSED theory for the crack turning angles in the mixed mode frac ture of soda lime silica glass disk specimens to experimental results presented in Chapter 3.

PAGE 114

114 Figure 8 3 Comparison of predictions of modified MSED theory for mixed mode fracture in soda lime silica glass disk spec imens to experimental results presented in Chapter 3 Figure 8 4 Comparison of predictions of modified MSED theory for the crack turning angles in the mixed mode frac ture of soda lime silica glass disk specimens to experimental results of Shetty et al and to the modified MTS theory by Smith et al. 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 K 1 /K IC K II /K IC Soda lime silica glass disks (Data from Chapter 3) Conventional MSED Theory Modified MSED Theory 0 10 20 30 40 50 60 70 80 90 60 70 80 90 Crack turning angle Crack orientation, Soda lime silica glass BD specimens (Shetty et al.) Conventional MSED Theory Modified MSED Theory Modified MTS theory (Smith et al.)

PAGE 115

115 Figure 8 5 Compa rison of predictions of modified MSED theory for mixed mode fract ure in soda lime silica glass disk specimens to experimental results of Shetty et al Figure 8 6 Comparison of predictions of modified MSED theory for mixed mode frac ture in soda lim e float glass disk specimens to experimental results of Awaji et al 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 K I /K IC K II /K IC Soda lime silica glass BD specimens (Shetty et al.) Conventional MSED Theory Modified MSED Theory 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 K 1 /K IC K II /K IC Soda lime silica glass disks (Awaji et al.) Conventional MSED Theory Modified MSED Theory

PAGE 116

116 Figure 8 7 Comparison of predictions of modified MSED theory for the crack turning angles in the mixed mode fracture of alumina and CeO 2 TZP disk specimens to experimental results of Singh et al Figure 8 8 Comparison of predictions of modified MSED theory for m ixed mode fracture in alumina disk specimens to experimental results of Singh et al 0 10 20 30 40 50 60 70 80 90 60 70 80 90 Crack turning angle Crack orientation, Alumina BD specimens (Singh et al.) Cerium Oxide BD specimens (Singh et al.) Conventional MSED Theory Modified MSED Theory 0 0.2 0.4 0.6 0.8 1 1.2 0 0.4 0.8 1.2 1.6 2 K 1 /K IC K II /K IC Alumina BD specimens (Singh et al.) Conventional MSED Theory Modified MSED Theory(r=200 microns) Modified MSED Theory (r=1 micron) Modified MSED Theory (r= 0.1 micron)

PAGE 117

117 Figure 8 9 Comparison of predictions of modified MSED theory for mixed mode fracture in CeO 2 TZP disk specimens to experimental results of Singh et al 0 0.2 0.4 0.6 0.8 1 1.2 0 0.4 0.8 1.2 1.6 2 K 1 /K IC K II /K IC Ceria BD specimens (Singh et al.) Conventional MSED Theory Modified MSED Theory(r=60 microns) Modified MSED Theory (r=1 micron) Modified MSED Theory (r= 0.1 micron)

PAGE 118

118 CHAPTER 9 CONCLUSIONS Three major aspects of mixed mode fracture in brittle materials were analyzed: Evaluation of mixed mode stress inte nsity factors Fractography of mixed mode fracture Establishing a mixed mode fracture criterion which applies to a wide range of brittle materials. Mixed mode stress intensity factors were evaluated for soda lime silica glass, representing amorphous, isotro pic materials, silicon nitride representing fine grain polycr y stalline materials, and mica glass ceramic representing rising resistance curve (R curve) materials Brazilian disk specimens and three point bend specimens were the test geometries used to p erform the mixed mode fracture studies. Mixed mode loading conditions were introduced by inclining precracks at different angles with respect to the loading direction. Precracks can be either two dimensional through cracks or three dimensional surface crac ks created through indentations. Stable crack growth occurs for an as indented surface crack before unstable fracture occurs, as opposed to an annealed surface crack. The fracture paths for any type of precrack, under mixed mode loading conditions are non coplanar with respect to the initial crack plane. The crack extension directions for the through cracks and surface cracks were found to be different. Irrespective of the test geometry t he crack turning angles remained the same for a particular crack orie ntation for the indent surface cracks. The crack turning angles were, however, found to be much greater for through cracks. The critical mode II stress intensity factor, unlike the critical mode I stress intensity factor or the fracture toughness is not a consistent material property and is dependent on the stress states in the specimen or the test geometry. Surface cracks showed

PAGE 119

119 greater fracture resistance in comparison to chevron notch cracks under mixed mode loading conditions in soda lime silica glass d isks in compression. The normalized mode II fracture toughness in polycry s talline ceramics such as silicon nitride was found to be greater in comparison to soda lime silica glass as is seen in the case for pure mode I loading The reason for the greater fracture resistance for the polycrystalline ceramic in mode II loading is attributed to the microstructural asperities resisting the shear deformation. R curve ceramics such as mica glass ceramics show crack growth resistance behavior with crack extension in both pure mode I and mixed mode loading conditions. These findings suggest that crack geometry and material microstructure influence mixed mode fracture in ceramics. The mixed mode fracture surfaces of soda lime silica glass failed from controlled surfa ce cracks are characterized by an absence of the mist region and presence of distinct twist hackle markings, often called lances For a semi elliptical surface crack, modes I, II and III are active along the crack front. Lances are thought to be generate d due to the influence of the mode III component acting along the crack front. Atomic force microscopy on the mixed mode fracture surfaces revealed that the features at different regions are similar and differ only by scale. However, these features are dif ferent from those seen on fracture surfaces failed in pure mode I Fracture surface measurements for mixed mode loading are compared to the measurements for mode I loading. The stress intensity at microbranching was determined to be a constant at all mode mixities. The practical implications of these observations are that forensic analyses can be used without a priori knowledge of the loading conditions. The crack to mirror size ratios for R curve materials is found to be a

PAGE 120

120 function of crack length as oppos ed to a non R curve material such as glass. For a particular crack length, the crack to mirror size ratio increases with increasing mode mixity. Four conventional fracture theories were evaluated: Non Coplanar Strain Energy Release Rate (NCSERR) theory, Ma ximum Tangential Stress (MTS) theory, Coplanar Strain Energy Release Rate (CSERR) theory and Minimum Strain Energy Density (MSED) theory. The conventional theories are based on the assumption that the singular stress terms characterized by the mixed mode s tress intensity factors, K I and K II dominate near to the crack tip. The predictions from the conventional MSED theory for the mixed mode stress intensity factors agrees very well with the experimental results for flexure specimens with surface cracks. This two parameter (MSED) theory is thus effective in describing mixed mode fracture from surface cracks in flexure specimens. However, none of the conventional theories could explain the experimental mixed mode stress intensity factor values for disk specimen s in compression. The series expansion for the stress fields around the crack tip has non singular stress terms, often called T stress which is assumed to be negligible in the conventional theories. The MSED theory was modified to account for both the singular and the T stress terms. The modified MSED theory is shown to be sufficient to describe the mixed mode fracture toughness envelop es for many ceramic Brazilian disk specimens with through cracks. This indicates that T stress is significantly greater in these disk specimens than for surface cracks in flexural specimens The experimental value of the angle by which the crack deviates f rom its initial plane in Brazilian disk specimens with through cracks could be predicted by the modified MSED theory. This

PAGE 121

121 indicates that crack propagation initiates in the direction of minimum strain energy density. The T stress has a significant i nfluence on mixed mode fracture in ceramic disk specimens. The T stress becomes more significant in disk specimens as the crack length decreases which could possibly explain the greater resistance to mode II loading for relatively small surface cracks. Ho wever, the influence of the T stress is negligible for flexure specimens. This shows the dependence of test and crack geometry on the values of the T stress term T stress calculations for surface cracks for various levels of mode mixities is an important and interesting area for future research.

PAGE 122

122 APPENDIX A GEOMETRIC FACTORS AN D T* VALUES FOR CENT RALLY CRACKED BRAZIL IAN DISK SPECIMEN The mixed mode geometric factors Y 1 and Y 2 for Brazilian Disk (BD) specimens with through crack at the center hav e been evaluated by Atkinson et al. [58]. (A 1) (A 2) where is the orientation of the crack with respect to the horizontal direction (i.e. direction of the ten sile stress ). The schematic of th e crack or ientation is shown in Figure 3 3 T i and S i are functions of c/R, which is the ratio of the half crack length, c to the radius of the disk specimen, R whereas A i and B i are functions of the crack orientation, The first five values of T i S i A i and B i are usually considered for evaluating the geometric factors, the values of which are given in Tables A 1 to A 4. The normalized form of T stress, T were calculated using FEA techniques by Ayatollahi et al [59] the values of which are given as f unctions of c/R and in Table A 5. Table A 1. First five values of T i as function of crack length ratio. c/R T 1 T 2 T 3 T 4 T 5 0.2 1. 0 60049 0.514907 0.382430 0.383392 0.318086 0.3 1.135551 0.533477 0.391640 0.393835 0.325033 0.4 1.243134 0.559734 0.404603 0.408597 0.334831 0.5 1.387239 0.594892 0.421949 0.428353 0.347941 Table A 2. First five values of S i as function of crack length ratio. c/R S 1 S 2 S 3 S 4 S 5 0.2 1.039864 0.509959 0.379956 0.380584 0.316245 0.3 1.089702 0.522272 0.386086 0.387518 0.320834 0.4 1.160796 0.539824 0.394822 0.397403 0.327411 0.5 1.257488 0.563966 0.406869 0.410966 0.336447

PAGE 123

123 Table A 3. First five values of Ai. i A i 1 1 4cos 2 2 8cos 2 4sin 2 3 4cos 2 36sin 2 4 4 16cos 2 1+24sin 2 80sin 4 6 5 20cos 2 40sin 2 4 448sin 6 8 Table A 4. First five values of B i i B i 1 1 2 5+8sin 2 3 3+8(1 2sin 2 3sin 2 4 3+16(1 2sin 2 12(1 2sin 2 2 32(1 2sin 2 3 5 5 16(1 2sin 2 60(1 2sin 2 2 +32(1 2sin 2 3 +80(1 2sin 2 4 Table A 5. Values of T* for different crack orientations and crack length ratios. c/R=0.2 c/R=0.3 c/R=0.4 c/R=0.5 9 0 3.301 3.19 3.038 2.886 88 3.298 3.172 3.011 2.851 86 3.28 3.14 2.96 2.783 84 3.201 3.074 2.868 2.662 82 3.153 2.985 2.741 2.498 80 3.102 2.878 2.596 2.314 78 3.026 2.749 2.424 2.099 76 2.954 2.596 2.229 1.863 74 2.821 2.431 2.021 1.601 72 2.613 2.249 1.802 1.348 70 2.5 2.055 1.581 1.103 68 2.354 1.852 1.359 0.867 66 2.152 1.639 1.131 0.643 64 1.958 1.422 0.924 62 1.75 1.201 60 1.5

PAGE 124

124 APPENDIX B MIXED MODE STRESS INTENSIT Y FACTORS FOR SODA L IME SILICA GLASS DISKS IN DIAMETRAL C OMPRESSION The failure stress and mixed mode stress intensity factors for soda lime silica g lass disks with chevron notches are given in Table B 1. c/R for the disks with the chevron notches was 0.3. Table B 2 and Table B 3 gives the crack size, failure stresses and mixed mode stress inte nsity factors for as indented soda lime silica glass disks and annealed glass disks respectively. Table B 1. Mixed mode stress intensity factors for soda lime silica glass disks with chevron notches. Crack Orientation, Stress to failure, MPa) K I (MPa m 1/2 ) K II (MPa m 1/2 ) 90 5 .0 0.71 0 90 4.9 0. 7 0 0 90 5.1 0.72 0 85 4.4 0.6 0 0.46 85 4.5 0.6 0 0.48 80 4.1 0.49 0.6 0 80 3.7 0.44 0.65 80 3.6 0.43 0.67 80 3.7 0.43 0.65 75 3.8 0.36 0.7 0 75 3.9 0.37 0.69 75 3.7 0.35 0.71 70 3.8 0.23 0.79 70 2.7 0.16 0.81 70 3.7 0.22 0.78 6 5 2.7 0.03 0.82 65 4 .0 0.01 0.84 65 4.2 0.0 1 0.85

PAGE 125

125 Table B 2. Mixed mode stress intensity factors for soda lime silica glass disks with as indented surface cracks Crack Orientation, Normal Stress, Shear stress, (MPa) Crack size, c ( m) K I (MPa m 1/2 ) K II (MPa m 1/2 ) 90 20 0 500 0.73 0.00 85 16 16 550 0.64 0.61 85 17 11 603 0.68 0.45 85 18 18 434 0.61 0.61 85 17 17 517 0.64 0.64 80 15 20 450 0.51 0.72 80 15 20 475 0.55 0.72 80 17 28 278 0.47 0.76 7 5 9 22 583 0.35 0.86 7 5 8 23 532 0.29 0.88 70 4 27 465 0.15 0.95 70 7 26 474 0.24 0.93 70 4 25 546 0.17 0.98 70 6 27 456 0.20 0.94 65 2 30 415 0.06 1.01 65 1 26 598 0.04 1.06 65 2 26 560 0.07 1.02 Table B 3. Mixed mode stress intensity factors for soda lime silica glass disks with annealed surface cracks Crack Orientation, Normal Stress, Shear stress, (MPa) Crack size, c ( m) K I (MPa m 1/2 ) K II (MPa m 1/2 ) 90 28 0 435 0.72 0.00 90 23 12 564 0.69 0.35 85 24 23 459 0.63 0.62 85 29 32 263 0.59 0.63 80 18 27 478 0.48 0.74 80 14 27 550 0.40 0.80 80 14 26 588 0.43 0.79 75 15 37 362 0.36 0.87 75 15 34 407 0.37 0.85 75 12 31 543 0.35 0.89 70 5 32 638 0.16 1.01 70 9 47 286 0.18 0.98 70 7 37 475 0.20 0.99 65 3 36 524 0.09 1.03 65 2 34 598 0.07 1.04 65 2 38 508 0.04 1.06

PAGE 126

126 APPENDIX C MIXED MODE STRESS INTENSIT Y FACTORS FOR MICA G LASS CERAMIC BARS IN FLEXURE Mica glass ceramic bars with oriented indentation cracks were tested in flexure. The critical crack sizes, failure stress and the mixed mode stress intensity factors for the different crack orientations are shown in Tables C 1 to C 5. Schematic of an incli ned crack on the tensile surface of a flexure bar is shown in Figure 5 2. Table C 1. Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =90 Crack size c (microns ) Stress to failure, K I (MPa m 1/2 ) K II (MPa m 1/2 ) 12 170 0.73 0 25 134 0.83 0 49 121 1.05 0 53 127 1.15 0 65 125 1.25 0 71 117 1.22 0 77 119 1.3 0 0 87 110 1.27 0 99 114 1.41 0 107 111 1.42 0 118 115 1.55 0 125 110 1.53 0 139 109 1.6 0 0 143 109 1.62 0 159 106 1.65 0 177 100 1.65 0 185 98 1.65 0 194 95 1.64 0 202 93 1.64 0 240 86 1.65 0 235 87 1.65 0

PAGE 127

127 Table C 2 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =80 Crack size c (microns ) Stress to failure, K I (MPa m 1/2 ) K II (MPa m 1/2 ) 25 135 0.81 0.17 35 128 0.91 0.19 43 131 1.03 0.21 44 130 1.04 0.22 45 130 1.05 0.22 46 131 1.07 0.22 58 130 1.19 0.25 58 130 1.19 0.25 60 129 1.2 0 0.25 65 129 1.25 0.3 0 68 117 1.16 0.24 75 121 1.26 0.26 76 121 1.27 0.26 76 121 1.27 0.26 79 120 1.28 0.27 99 115 1.37 0.28 105 114 1.4 0 0.29 105 107 1.32 0.27 124 111 1.48 0.31 150 108 1.59 0.32 157 105 1.58 0.32 159 105 1.59 0.32 178 100 1.6 0 0.33 198 94 1.59 0.32 216 91 1.6 0 0.32

PAGE 128

128 Table C 3 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =70 Crack size c (microns ) Stress to failure, K I (MPa m 1/2 ) K II (MPa m 1/2 ) 29 136 0.8 0.34 37 134 0.89 0.38 47 132 0.99 0.42 44 134 0.97 0.41 48 131 0.99 0.42 35 125 0.81 0.35 49 124 0.95 0.41 68 131 1.18 0.5 0 60 121 1.03 0.44 64 132 1.16 0.5 0 64 121 1.06 0.45 75 120 1.14 0.49 79 114 1.11 0.47 82 120 1.19 0.51 84 120 1.2 0 0.51 106 114 1.29 0.55 107 116 1.31 0.56 100 115 1.26 0.54 136 107 1.36 0.58 139 106 1.37 0.59 145 105 1.38 0.59 150 103 1.38 0.59 178 95 1.38 0.59 198 90 1.38 0.59 225 83 1.37 0.59

PAGE 129

129 Table C 4 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =55 Crack size c (microns ) Stress to failure, K I (MPa m 1/2 ) K II (MPa m 1/2 ) 23 139 0.66 0.46 29 128 0.68 0.47 33 125 0.71 0.49 37 123 0.74 0.51 48 116 0.79 0.55 51 112 0.79 0.55 57 109 0.81 0.56 65 107 0.85 0.59 66 105 0.84 0.58 69 104 0.85 0.59 69 104 0.85 0.59 76 102 0.88 0.61 78 100 0.87 0.6 0 88 103 0.95 0.66 94 103 0.99 0.68 98 105 1.03 0.71 102 102 1.02 0.71 103 101 1.01 0.7 0 125 100 1.1 0 0.76 139 97 1.13 0.78 148 97 1.17 0.81 150 98 1.18 0.82 168 92 1.18 0.82 186 88 1.19 0.82 200 85 1.19 0.82 229 79 1.18 0.82

PAGE 130

130 Table C 5 Mixed mode stress intensity factors for mica glass ceramic bars with crack orientation, =45 Crack size c (microns ) Stress to failure, K I (MPa m 1/2 ) K II (MPa m 1/2 ) 25 148 0.53 0.54 33 143 0.59 0.6 0 41 126 0.58 0.59 49 124 0.62 0.63 59 129 0.71 0.72 68 124 0.73 0.74 73 122 0.75 0.76 82 122 0.79 0.8 0 95 112 0.78 0.79 98 118 0.84 0.85 103 115 0.84 0.85 105 120 0.88 0.89 113 113 0.86 0.87 119 114 0.89 0.9 0 125 110 0.88 0.89 129 108 0.88 0.89 136 109 0.91 0.92 157 101 0.91 0.92 168 104 0.97 0.98 183 100 0.97 0.98 199 95 0.96 0.97 214 93 0.97 0.98

PAGE 131

131 APPENDIX D PUBLICATIONS AND PRE SENTATIONS FROM THIS THESIS Gopalakrishnan K, Mecholsky Jr JJ. Mixed mode f racture in a n R curve material. Engng Fract Mech Gopa lakrishnan K, Mecholsky Jr JJ. Fractography of mixed mode fracture in soda lime silica glass. J Eur Ceram Soc George L, Nagaraj A, Mecholsky Jr JJ, Gopalakrishnan K. Numerical and Experimental investigation of mixed mode fracture parameter on silicon nitride u sing th e Brazilian disc t est. Fat igue Fract Engng Mate r Struct. Gopalakrishnan K, Mecholsky Jr JJ. Quantitative fractography of mixed mode f racture in an R c urve material. J Am Ceram Soc Mixed mode fracture in silicon nitride disks using the Brazilian disk t est Materials Science and Technology Conference (MS&T) Pittsburgh, PA (2009) Mixed mode (I/II) fracture behavior of surface and chevron notch cracks in diametral c ompression 35th International Conference & Exposition on Advanced Ceramics & Composites (ICACC) Daytona Beach, FL (2011) Fractography of mixed m ode fracture in soda lime silica g lass Fractography of Glasses and Ceramics VI Conference, J acksonville, FL (2011) Mixed mode (I/II) fracture in an R curve material 36th International Conference & Exposition on Advanced Ceramics & Composites (ICACC), Daytona Beach, FL (2012).

PAGE 132

132 LIST OF REFERENCES 1. Williams JG, Ewing PD. Fracture under complex stress the angled crack problem. Int J Fract 1972;8:416 41. 2. Ueda Y, Ikeda K, Yao T, Aoki M. Characteristics of brittle fracture under general combined modes including those under bi axial tensile loads. Eng ng Fract Mech 1983;18:11 31 58 3. Smith DJ, Ayatollahi MR, Pavier MJ. The role of T stress in brittle fracture for linear elastic materials under m ixed mode loading. Fatigue Fract Eng ng Mater Struct 2001;24:137 50. 4. Faber KT, Evans AG. Crack deflection processes I. Theory. Acta Metall 1983; 31: 565 76. 5. Faber KT, Evans AG. Crack deflection processes II. Experiment. Acta Metall 1983 ; 31 : 577 84. 6. Nandy MO, Nose T, Enoki M, Kishi T. R curve and cyclic fatigue behavior in alumina particles reinforced sil icon carbide. Mat er Trans 1996;37(4): 769 75. 7. Thompson JY, Mecholsky Jr JJ, Anusavic e KJ, Balasubramanian KJ. The effect of microcracking on the fracture toughness and fractal dimension of lithia based glass cer amics. J Am Ceram Soc 1995;78: 3045 9. 8. Han LX, Warren R, Suresh S. An experimental study of toughening and degradation due to microcracking in a ceramic composite. Acta Metall Mater 1992; 40 (2): 259 74. 9. Dauskardt RH, Yu W, Ritchie RO. Fatigue crack propagation in transformation toughened zirconia ceramic. J A m Ceram Soc 1987 ; 70 (10):C 248 C 252. 10. Swanson PL. Crack interface traction: A fracture resistance mechanism in brittle polycrystals.In: Varner J Frec hette VD, editors. Advances in c era mics, Vol 22, Fractography of glasses and c eramics, A merican Ceram ic Society, Columbus, OH; 1988. p. 135 55. 11. Fairbanks CJ, Lawn BR, Cook RF, Mai YW. Microstructure and the strength of ceramics. In: Bradt RC Evans AG, Hasselman DP H, Lange FF, editors. Fracture mechanics of c er amics, Vol 8, Plen um Publishing Corpora tion, NY; 1986. p. 23 37. 12. Mecholsky Jr JJ, Hill TJ, Chen Z. Application of quantitative fractography to the characterization of R curve behavior. In: Salem JA, Quinn GD, Jenkins MG, editors. Fracture resistance testing of monolithic and composite b rit t le materials, ASTM STP 1409, American So ciety for Testing and Mat erials, West Conshohocken, PA; 2002. p. 152 68

PAGE 133

133 13. Hill TJ, Mecholsky Jr JJ, Anusavice KJ. Fractal analysis of toughening behavior in 3BaO.5SiO2 glass ceramics. J Am Ceram Soc 2000;83(3): 545 52. 14. Chan JY. Mechanical degradation of canasite glass ceramic un der dynamic and Gainesville (GNV): University of Florida; 1996. 15. Quinn GD. Quantitative analysis. In: Fractography of ceramics and glasses, NIST S pecial Publication 960 16; 2007. p. 7 52. 16. Marshall DB. Strength characteristics of transformation toughened zirconia. Ibid 1986;69(3):173 80 17. Pyzik AJ, Carroll DF. Technology of self reinforced silicon ni tride. Annu Rev Mater Sci 1994; 24:189 214 18. Fischer H, Rentzch W, Marx R. R curve behavior of dental cera mic materials. J Dent Res 2002; 81(8):547 51. 19. Kirchner HP, Kirchner JW. Fracture mechanics of fracture m irrors. J Am Ceram Soc 1979;62(3 4) :198 202 20. Rabinovitch A. On the Mot t der ivation of crack velocity. Phil Mag Lett 1994; 70(4): 231 3 21. Hollan d D, Marder M. Cracks and Atoms. Adv. Mater 1999;11( 10 ) 22. Kirchner HP, Conway Jr JC. Comparison of the stress intensity and Johnson and Holloway criteria for crack branching in rect angular bars. J Am Ceram Soc 1987;70(8): 565 9. 23. Kulawansa DM, Jensen LC, Langford SC, Dickinson JT. Scanning tunneling microscope observations of the mirror region of silicate glass fracture su rfaces. J Mater Res 1994;9(2): 341 70 24. Mecholsky Jr JJ Analysis of fracture origins in g lass. In : Bradt RC, Tressler RE, e d itors. Fractography of glass, Plenum Press, NY ; 1994. 25. Kirchner HP, Conway Jr JC. Fractography of glasses and c e ramics. In: Varner J, Frechette V, editors. Advances in c eramics Vo l 22 The American Ceramic S ociety, Westerville, OH; 1988. p. 187 213 26. Irwin GR. Relation of stresses near a crack to the crack extension f orce. Ninth International Congress of Applied Mechanics, Brussels ; 1957 27. Tsai YL, Mecholsky Jr JJ. Fracture mechanics description of fracture mirror formation in single crystals. Int J Fr act 1992;57: 167 82.

PAGE 134

134 28. Frieman SW, Gonzalez AC, Mecholsky Jr JJ. Mixed mode fracture in soda l ime glass. J Am Ceram Soc 1979; 62(3 4):206 8. 29. Marshall DB. Me chanisms of failure from surface flaws in mixed mode loading. J Am Ceram Soc 1984; 67(2):110 6. 30. Petrovic JJ. Mixed mode fracture of hot pressed Si 3 N 4 J Am Ceram Soc 1985; 68(6): 348 55. 31. Glaesemann GS, Ritter Jr JE, Jakus K. Mixed mode fracture in soda lime glass using indentation flaws. J Am Ceram Soc 1987; 70(9): 630 6. 32. Williams JG, Ewing PD. Fracture under complex stress the angled crack problem. Int J Fract 1972; 8:441 6. 33. Richard HA, Benitz K. A loading devic e for the creation of mixed mode in fractu re mechanics. Int J Fract 1983; 22:R55 8 34. Ayatollahi MR, Aliha MRM, Hassani MM. Mixed mode brittle fracture in PMMA A n experimental study using SCB spe cimens. J Mater Sci Eng A 2006; 417(1 2):348 56. 35. Singh D, Shetty DK. Fracture toughness of polycrystalline ceramics in combined mode I and mode II loading. J Am Ceram S oc 1989; 72(1): 78 84. 36. Shetty DK, Rosenfield AR Duckworth WH. Mixed mode fracture of ceramics in diametral co mpression. J Am Ceram Soc 19 86; 69(6):437 43. 37. Shetty DK, Rosenfield AR, Duckworth WH. Mixed mode fracture in biaxial stress state: Application of the Brazilian disk test. Eng ng Fract Mech 1987; 26(6):1825 40. 38. Hussain MA, Pu S L Underwood J. Fracture Analysis ASTM STP 560;19 74: 2 28. 39. Erdogan F, Sih GC. On the crack extension in plates under plane loading and transverse shea r. J Basic Eng 1963;85D: 519 27. 40. Petrovic JJ, Mendiratta MG. Mixed mode fracture from controlled surface flaws in hot pressed silico n nitride. J A m Ceram Soc 1976; 59(3 4):163 7. 41. Sih GC. Strain energy density factor applied to mixed mode crack problems. Int J Fract 1974; 10( 3):305 31. 42. Awaji H, Sato S. Combined mode fracture toughness measurement by the disk test. J Eng ng Mater and Tech 197 8;100:175 82 43. Khandelwal P, Majumdar BS, Rosenfield AR. Mixed mode high temperature toughness of silicon nitride. J Mater Sci 1995;30:395 8

PAGE 135

135 44. Ikeda K, Igaki H. Mixed mode fracture criterion in soda lime glass. The Japan Soc Mech Eng ng 1990;33(1): 84 8 45. Awaji H, Kato T, Honda S, Nishikawa T. Criterion for combined mode I II brittle fracture. J Ce ram Soc of Japan 1999;107(10): 918 24 46. Carniero FLLB, Barcellos A. Union of testing and research laboratories for materials and s tructures 1953 13 47. Shetty DK, Rose nfield AR, Duckworth WH. Fracture toughness of ceramics measured by a chevron notch diametral compression test. J Am Ceram Soc 1985;68(12) :C325 C327 48. Mecholsky Jr JJ, Frieman SW, Rice RW. Effect of grinding on flaw geometry and fracture of glass. J Am Ceram Soc 1977; 60(3 4):114 7. 49. Randall PN. In: Plane Strain Crack Toughness Testing of High Strength Metallic Materials. ASTM STP 410; 1967. p. 88 126 50. Mitchell NB. The indirect tension for c oncrete. Mater Res and Stds AS TM 1961:780 8 51. Sih GC. Handbook of stress intensity factors. L ehigh University, Bethlehem, PA; 1973. p. 3.1 52. Marshall DB, Lawn BR, Mecholsky Jr JJ. Effect of residual contact stresses on mirror/flaw size r elation. J Am Ceram Soc 1980;63(5 6) :358 360 53. Smith FW, Emery AF, Kobayashi AS. Stress intensity factors for semicircula r cracks II. J Appl Mech 1967; 34(4):953 60. 54. Kassir MK, Sih GC. Three dimensional stress distribution around an elliptical crack under arbitrary loadings. J A ppl Mech Trans ASME Ser.E 1966; 33(3):601 11. 55. Duffrene L, Gy R. Viscoelastic constants of a soda lime silica glass. J Non Crystalline Solids 1997; 211(1 2):30 8 56. Scott PM, Thorpe TW. A critical review of crack tip stress intensity factors for se mi ell iptic cracks. Fatigue Fract Eng ng Mater Struct 1981; 4(4):291 309. 57. Smith FW, Sorensen DR. Mixed mode stress intensity factors for semi ellip tical surface cracks. NASA Report No. NASA CR 134684; 1974. 58. Atkinson C, Smelser RE, Sanchez J. Combined mod e fracture via the cracked Brazili an disk test. Int J Fract 1982;18(4): 279 91. 59. Ayatollahi MR, Aliha MRM. Wide range data for crack tip parameters in two disc type specimens under mixed mod e loading. Comp Mater Sci 2007;38(4): 660 70

PAGE 136

136 60. Fett T. Stress intensity factors and T stress for internally cracked circular disks under various boundary conditions. Eng ng Fract Mech 2001;68: 1119 36 61. Levesque G, Arakere NK, Mecholsky Jr JJ, Gopalakrishnan K. Numerical and experimental investigation of mixed mode fracture parameter on silicon nitride using the brazilian disc test. Fatigue Fract Eng ng Mater Struct 2010 ; 33(8): 490 503. 62. Hadfield M, Stolarski T, Cundill RT, Horton S. Failure modes of ceramic elements with ring crack defects. Tribol Int 1993; 26:157 64 63. Levesque G, Arakere N. An investigation of partial cone cracks in silicon nitr ide balls. Int. J Solids Struct 2008; 45(25 26):6301 15 64. Levesque G, Arakere N. Critical flaw size in silicon nitride ball bearings. Tribol Trans 2010 ;5 3:511 9. 65. Leves que G, Arakere N. Empirical curve fits for surface flaws subj ect to rolling contact fatigue. Tribol Trans 2010 ;53:621 9. 66 Gupta PK, Jubb NJ. Post indentation slow growth of radial cracks in g lasses. J Am Ceram Soc 198 1;64(8): C 112 67 Wachtman JB, Capps W, Mandel J. Biaxial flexure tests of ceramic substrates. J Mater 1972;7(2):188 94 68 Chantikul P, Anstis GR, Law n BR, Marshall DB. A critical evaluation of indentation t echniques f or m ea suring fracture toughness: II, strength m e thod. J Am Ceram Soc 1981;64(9):539 43 69 Piotrowski AE, MJ. A novel test method to measure the fracture toughness of ceramic balls used in bearings Fatig ue Fract Eng ng Mater Struct 2006; 29 : 558 72. 70 Ayatollahi MR, Aliha MRM. Mixed mode fracture in soda lime glass analysed using the gene ralized MTS criterion. Int J Solids Struct 2009 ; 46:311 21. 71 Baik DS, No KS, Chun JS. Mechanical properties of mica glass ceramics. J Am Ceram Soc 1995;78(5): 1217 22. 72 Cai H, Kalceff MAS, Lawn BR. Deformation and fracture of mica containing glass ceramics in hertzian contacts. J Mater Res 199 4; 9(3):762 70. 73 Corn ing, Inc., Brochure for macor. 74 Accuratus Standard Ceramic Products, Macor machinable glass ceramic brochure.

PAGE 137

137 75 Liu Y, Sheng X, Dan X, Xiang Q. Preparation of mica/apatite glass ceramics biom aterials. Mater Sc Eng: C 2006; 2 6(8): 1390 4. 76 Mecholsk y Jr JJ. Fracture mechanics analysis of glass ceramics. In: Simmons JH, Ulhmann DR, Beall GH, editor s. Advances in Ceramics, Vol 4 Am erican Ceram ic Soc iety, Columbus, OH; 1982. p 261 76. 77 Thompson JY, Mecholsky Jr JJ, Anusavice KJ, Ba lasubramanian KJ The effect of microcracking on the fracture toughness and fractal dimension of lithia based glass ceramics. J Am Ceram Soc 1995;78: 3045 9. 78 Reddy KPR, Fontana E, Helfinstine JD. Fracture toughness measurement of glass and ceramic materials using chev ron notched specimens. J Am Ceram Soc 1988; 71( 6): C 310 C 313. 79 Smekal A. Zum bruchvorgang bei spro¨dem stoffverhalten unter ein und mehrachsigen beanspru chungen. Oesterr Ing Arch 1953;7: 49 70 80 Momber AW. Fracture lances in glass loaded wi th spherical indenters. J Mater Sci Lett 2003;22: 1477 81 81 ng Fract Mech 1969;1:539 46 82 Pons JA, Karma A. Helical crack front instability in mi xed mode fracture. Nature 2010; 464: 85 9 83 Be auchamp EK. Mechanisms of hackle formation and crack branching. In: Varner J, Frechette VD, Quinn GD, editors. Fractography of Glasses and Cera mics III, Ceramic Transactions, Vol 64, Am erican Ceram ic Soc iety, Westerville, Ohio; 1996. p. 409 45 84 Noda N A, Kagita M Variations of stress intensity factors of a semi elliptical surface crack subjected to mode I, II, III loading Int J Press Vessel and Pip 2004; 81(7 ): 635 44 85 Mecholsky Jr JJ, Rice RW, Freiman SW. Prediction of fracture energy and f law size in glasses from mi rror size. J Am Ceram Soc 1974;57(10): 440 3 86 Quinn JB, Mohan R. Geometry of edge chips formed at different angles. Ceram Engng Sci Proc 2005;26(2): 85 92 87 Quinn GD. Single crystals. In: Fractography of ceramics and glasses, NIST Special Publication 960 16; 200 7. p. 8 10 8 18 88 Chen Z. Wear mechanisms of silicon nitride bearing materials u nder contact fatigue stresses. (PhD Thesis) Gainesville (GNV): University of Florida; 1995.

PAGE 138

138 89 Rice RW. Cer amic fracture features, observations, mechanisms and u ses. In:.Mecholsky Jr JJ, Powell Jr SR, ed itors. Fractography of c eramic and metal failures, ASTM STP 827, American Society for Testing and Materials, Philadelphia, PA ; 1984. 90 Newman Jr JC, Raju IS. An empirical stress intensity factor equation for the surface c rack. Eng ng Fract Mech 1981;15(1 2):185 92 91 Smith FW, Kobayashi AS, Emery F. Stress intensity factors for penny shaped c racks Part 1 Infi nite solid. J Appl Mech 1967; 34(4):947 52. 92 Ma rl off RH, Leven MM, Ringler TN, Johnson RL. Photoelastic determination of stress intensity factors. Exp Mech 1971;11(12):529 39 93 Hosseini SR, Choupani N, Gharabagh ARM. Geometry calibration factors of m odified Arca n fracture test for welded joint. I nt J Aerospace Mech Engng 2008; 2(4):241 6 94 Suresh S, Shih CF, Morrone A, O 'Dowd NP. Mixed mode fracture toughness of ceramic materials. J Am Ceram Soc 1990;73(5): 1257 67 95 He MY, Hutchinson JW. Surface crack subject to mixed mode loading. Eng ng Fract Mech 2000; 65 : 1 14. 96 Noda NA, Miyoshi S Variation of stress intensity factor and crack opening displacement of semi elliptical surface crack. Int J Fract 1996 ; 75( 1 ): 19 48 97 Mandelbrot BB. The fractal geometry of nature. W.H. Freeman, San Fransisco, CA; 1982 98 Mecholsky Jr JJ, Mackin TJ, Passoja DE. Self similar crack propagation in br ittle materials. Advances in Ceramics, Vol 22, Fractography of glasses and ceramics, The American Ceramic Society, Westerville, O H; 1988. p. 127 34 99 Mecholsky Jr JJ, Passoja DE, Feinherg Ringel KS. Quantitative analysis of brittle fracture surfaces using fractal g eometry. J Am Ceram Soc 1989; 72(1) : 60 5 100 Kirchner HP. Brittleness dependence of crack branching in c eram ics. J Am Ceram Soc 1986; 69(4): 339 42 101 Freiman SW, Mecholsky Jr JJ, Beche r PF. Fractogra phy: A quantitative measure of the fracture process. In: Frechette VD, Varner J, editors. CeramicTra nsactions, Vol 17, Fractography of glasses and ceramics American Cerami c Society, Westerville, OH; 1991. p. 55 78 102. Mecholsky Jr JJ, Rice RW, Freiman SW. Fracture surface analysis of ceramics. J Mater Sci 1976;11:

PAGE 139

139 103 Mecholsky Jr JJ, Freiman SW. Relationship between fractal geometry and fra ctography. J Am Ceram Soc 1991; 74(12):3136 8 104 Hill TJ, Dellabona A, Mecholsky Jr JJ. Establishing a protocol for measurements of fractal dimensions in britt le materials. J Mater Sci 2001;36: 2651 7 10 5 Mandelbrot BB, Passoja DE, Paullay AJ. The Fractal Character of Fracture Surfaces of Metals. Nature 1984; 308(19):721. 106 Ayatollahi MR, Aliha MRM. Fracture analysis of some ceramics under mixed mode l oading. J Am Ceram Soc 2011;94(2): 561 9. 107 Eft is J, Subramonian N, Liebowitz H. Biaxial load effects on the crack border elastic strain energy and strain energy release rate. Engng Fract Mech 1977;9:753 64. 108 Ayatollah i MR, Aliha MRM. On the use of b razilian disc specimen for calculating mixed mode I II fracture toughness of rock materials. Engng Fract Mech 2008;75:4631 41. 109 Schmidt RA. A microcrack model and its significance to hydraulic fracturing and fracture toughness testing. In: Proceedings of 21 st US symposium on rock mechanics; 1980. p. 581 90. 110 Ayatollahi MR, Smith DJ, Pavier MJ. Characterizing mixed mode brittle fracture using near crack tip stress fields. In: Forni di Sopra; 2011. p. 9 16

PAGE 140

140 BIOGRAPHICAL SKETCH Karthik Gopalakrishnan was born in the city of Cochin, also known as Queen of the Arabian Sea, in 1984. After graduating from high school, Karthik passed the Indian Institute of Technology Joint Entrance Examination (IIT JEE) where only the top 0.1% of the students who appear for the exam clears. He got admission in th e Indian Institue of Technology (IIT), Roorkee where he earned his Bachelor of Technology degree in Metallurgical and Materials Engineering. Karthik got admitted to the Department of Materials Science and Engineering at University of Florida in Fall 2007 Within four and a half years, he earned a PhD degree i n materials science and e ngineering under the guidance of Dr. John J Mecholsky, Jr.