Atlas of

Fatigue Curves
Edited by

Howard E. Boyer
Senior Technical Editor
American Society for Metals

ASM lnternatlonal"
Materials Park, Ohio 44073-0002
The Materials www.asminternational.org
Information Society
 Preface
This Atlas was developed to serve engineers who are looking for fatigue
data on a particular metal or alloy. In the past, the first step to locating this
data was an expensive and time-consuming search through the technical liter-
ature. Now, many ofthe important and frequently referenced curves are pre-
sented together in this one volume. They are arranged by standard alloy des-
ignationsand are accompanied by a textual explanation offatigue testing and
interpretation of test results. In each case, the individual curve is thoroughly
referenced to the original source.
Having these important curves compiled in a single book will also facili-
tate the computerization of these data. Plans are currently under way also to
make the data presented in this book available in ASCII files for analysis by
computer programs.
The Atlas of Fatigue Curves is obviously not complete, in that many
more curves could be included. Persons wishing to contribute curves to this
compilation for inclusion in future revisions should contact the Editors,
Technical Books, American Society for Metals, Metals Park, Ohio 44073.
 Contents
Fatigue Testing 1
Introduction I
Fatigue Crack Initiation 4
Fatigue Crack Propagation 12
SECTION 1: S-N Curves That Typify Effects of Major Variables 27
I-I. S-NCurves Typical for Steel 27
1-2. S-NCurves Typical for Medium-Strength Steels 28
1-3. S-NDiagrams Comparing Endurance Limit for Seven Alloys 30
1-4. Steel: Effect of Microstructure 31
1-5. Steel: Influence of Derating Factors on Fatigue Characteristics 32
1-6. Steel: Correction Factors for Various Surface Conditions 33
1-7. Fatigue Behavior: Ferrous vs Nonferrous Metals 34
1-8. Comparison of Fatigue Characteristics: Mild Steel vs Aluminum Alloy 35
1-9. Carbon Steel: Effect of Lead as an Additive 36
1-10. Corrosion Fatigue: General Effect on Behavior 37
I-II. Effect of Corrosion on Fatigue Characteristics of Several Steels 38
1-12. Steel: Effect of Hydrogen on Fatigue Crack Propagation 39
1-13. Relationship of Stress Amplitude and Cycles to Failure 40
1-14. Strain-Life and Stress-Life Curves 41
1-15. Fatigue Plot for Steel: Ultrasonic Attenuation vs Number of Cycles 42
SECTION 2: Low-Carbon Steels: Flat-Rolled, Weldments and Tubes 43
2-1. Typical S-N Curve for Low-Carbon Steel Under Axial Tension 43
2-2. AISI 1006: Effects of Biaxial Stretching and Cold Rolling 44
2-3. AISI 1006: Weldment; FCAW, TIG Dressed 45
2-4. AISI 1006: Weldment; Shear Joints 46
2-5. AISI 1006: Weldment; Lap-Shear Joints 47
2-6. AISI 1015: Effect of Cold Working 48
2-7. A533 Steel Plate: Fatigue Crack Growth Rate 49
2-8. A514F Steel Plate: Fatigue Crack Growth Rates 50
2-9. A514F and A633C: Variation in Fatigue Crack Growth Rate With Orientation 51
2-10. A514F: Scatterbands of Fatigue Crack Growth Rate 52
2-11. A633C Steel Plate: Scatterbands of Fatigue Crack Growth Rates 53
2-12. Low-Carbon Steel Weldment: Effects of Various Weld Defects 54
2-13. Low-Carbon Steel Weldment: Effect of Weld Reinforcement and Lack of
Inclusions 55
2-14. Low-Carbon Steel Weldment: Effect of Weld Reinforcement and Lack of
Penetration 56
2-15. Low-Carbon Steel Weldment: Computed Fatigue Strength; Weldment Contained
Lack of Fusion 57
2-16. Low-Carbon Steel Weldment: Effect of Reinforcement and Undercutting 58
2-17. Low-Carbon Steel: Transverse Butt Welds; Effect of Reinforcement 59
2-18. A36/E60S-3 Steel Plate: Butt Welds 60
2-19. A514F/EllO Steel: Bead on Plate Weldment 61
2-20. A36 and A514 Steel Plates: Butt Welded 62
2-21. A36 Plate Steel: Butt Welded 63
2-22. Low-Carbon Steel Tubes: Effect of Welding Technique 64
2-23. Low-Carbon Steel: Effect of Applied Anodic Currents in 3% NaCI 65
2-24. Low-Carbon Steel: Effect of pH in NaCI and NaOH 66
2-25. Low-Carbon Steel: Effect of Carburization and Decarburization 67

v
VI Contents

2-26. A514B Steel: Effect of Various Gaseous Environments on Fatigue Crack

Propagation 68
2-27. Cast 1522 and 1541 Steels: Effect of Various Surface Conditions 69
2-28. Cast A216 (Grade WCC) Steel: Fatigue Crack Growth Rate 70
SECTION 3: Medium-Carbon Steels, Wrought and Cast 71
3-1. AISI 1030 (Cast) Compared With AISI 1020 (Wrought) 71
3-2. AISI 1035: Effect of Gas and Salt Bath Nitriding 72
3-3. AISI 1040: Cast vs Wrought 73
3-4. AISI 1045: Relationship of Hardness and Strain-Life Behavior 74
3-5. AISI 1141: Effect of Gas Nitriding 75
3-6. Medium-Carbon Steels: Interrelationship of Hardness, Strain Life and Fatigue
Life 76
3-7. Medium-Carbon Steel: Effect of Fillet Radii 77
3-8. Medium-Carbon Steel: Effect of Keyway Design 78
3-9. Medium-Carbon Steel: Effect of Residual Stresses 79
3-10. Medium-Carbon Cast Steel: Effect of Changes in Residual Stress 80
3-11. Medium-Carbon Cast Steel: S-NProjection (Effect of Applied Stress) 81
3-12. Medium-Carbon Cast Steel: Effect of Applied Stress (Shot Blasting) 82
SECTION 4: Alloy Steels: Low- to High-Carbon, Inclusive 83
4-1. Medium-Carbon Alloy Steels, Five Grades: Effect of Martensite Content 83
4-2. Medium-Carbon Alloy Steels, Six Grades: Hardness vs Endurance Limit 84
4-3. Medium-Carbon Alloy Steels: Effect of Specimen Orientation 85
4-4. 4027 Steel: Carburized vs Uncarburized 86
4-5. 4120 Steel: Effect of Surface Treatment in Hydrogen Environment 87
4-6. 4120 Steel: Effect of Surface Treatment in Hydrogen Environment 88
4-7. 4120 Steel: Effect of Various Surface Treatments on Fatigue Characteristics in Air
vs Hydrogen 89
4-8. 4130 Steel: Fatigue Crack Growth Rate vs Temperature in Hydrogen 90
4-9. 4135 and 4140 Steels: Cast vs Wrought 91
4-10. 4135 and 4140 Steels: Cast vs Wrought 92
4-11. 4140,4053 and 4063 Steels: Effect of Carbon Content and Hardness 93
4-12. 4140 Steel: Effect of Direction on Fatigue Crack Propagation 94
4-13. 4140 Steel: Effect of Cathodic Polarization 95
4-14. Cast 4330 Steel: Effects of Various Surface Conditions 96
4-15. 4340 Steel: Scatter of Fatigue Limit Data 97
4-16. 4340 Steel: Strength vs Fatigue Life 98
4-17. 4340 Steel: Total Strain vs Fatigue Life 99
4-18. 4340 Steel: Stress Amplitude vs Number of Reversals 100
4-19. 4340 Steel: Effect of Periodic Overstrain 101
4-20. 4340 Steel: Estimation of Constant Life 102
4-21. 4340 Steel: Effect of Strength Level on Constant-Life Behavior 103
4-22. 4340 Steel: Notched vs Unnotched Specimens 104
4-23. 4340 Steel: Effect of Decarburization 105
4-24. 4340H Steel: Effect of Inclusion Size 106
4-25. 4340 Steel: Influence of Inclusion Size 107
4-26. 4340 Steel: Effect of Hydrogenation; Static Fatigue 108
4-27. 4340 Steel: Effect of Hydrogen 109
4-28. 4340 Steel: Effect of Nitriding 110
4-29. 4340 Steel: Effect of Nitriding and Shot Peening III
4-30. 4340 Steel: Effect of Induction Hardening and Nitriding 112
4-31. 4340 Steel: Effect of Surface Coatings 113
4-32. 4340 Steel: Effect of Temperature on Constant-Lifetime Behavior 114
4-33. 4520H Steel: Effect of Type of Quench 115
4-34. 4520H Steel: Effect of Shot Peening 116
4-35. 4620 Steel: Effect of Nitriding 117
4-36. 4620 Steel: P/M-Forged 118
4-37. 4620 Steel: P/M-Forged at Different Levels 119
 Contents VII

4-38. 4625 Steel: P/M vs Ingot Forms 120

4-39. 4640 Steel: P/M-Forged 121
4-40. High-Carbon Steel (Eutectoid Carbon): Pearlite vs Spheroidite 122
4-41. 52100 EF Steel: Surface Fatigue; Effect of Finish and Additives 123
4-42. 52100 EF Steel: Surface Fatigue; Effect of Surface Finish and Speed 124
4-43. 52100 EF Steel: Surface Fatigue; Effect of Lubricant Additives 125
4-44. 52100 EF Steel: Surface Fatigue; Effect of Lubricant Viscosity, Slip Ratio and
Speed 126
4-45. 52100 EF Steel: Rolling Ball Fatigue; Effect of Oil Additives 127
4-46. 52100 Steel: Carburized vs Uncarburized 128
4-47. 8620H Steel: Carburized; Results From Case and Core 129
4-48. 8620H Steel: Effect of Variation in Carburizing Treatments 130
4-49. 8620 Steel: Effect of Nitriding 131
4-50. 8622 Steel: Effect of Grinding 132
4-51. Cast 8630 Steel: Goodman Diagram for Bending Fatigue 133
4-52. Cast 8630 Steel: Effect of Shrinkage 134
4-53. Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue 135
4-54. Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue 136
4-55. Cast 8630 Steel: Effect of Shrinkage on Plate Bending 137
4-56. Cast 8630 vs Wrought 8640 138
4-57. 8630 and 8640 Steels: Effect of Notches on Cast and Wrought Specimens 139
4-58. Nitralloy 135 Steel: Effect of Nitriding 140
4-59. AMS 6475: Effects of Welding 141
4-60. Medium-Carbon, ICr-Mo-V Steel Forging: Effect of Cycling Frequency 142
4-61. EM 12 Steel: Effect of Temperature on Low-Cycle Fatigue 143
4-62. Cast 0.5Cr-Mo-V Steel: Effects of Dwell Time in Elevated-Temperature
Testing 144
4-63. Cast 0.5Cr-Mo-V Steel: Effect of Environment at 550°C (1022 OF) 145
4-64. Cast C-0.5Mo Steel: Effect of Temperature and Dwell Period on Cyclic Endurance
at Various Strain Amplitudes 146
SECTION 5: HSLA Steels 147
5-1. HI-FORM 50 Steel vs 1006 147
5-2. HI-FORM 50 Steel vs 1006: Stress Response 148
5-3. HI-FORM 50 Steel Compared With 1006, DPI and DP2 149
5-4. HSLA vs Mild Steel: Torsional Fatigue 150
5-5. Proprietary HSLA Steel vs ASTM A440 151
5-6. Comparison of HSLA Steel Grades BE, JF and KF for Plastic Strain Amplitude
vs Reversals to Failure 152
5-7. Comparison of HSLA Steel Grades BE, JF and KF for Total Strain Amplitude
vs Reversals to Failure 153
5-8. Comparison of a Dual-Phase HSLA Steel Grade With HI-FORM 50: Total Strain
Amplitude vs Reversals to Failure 154
5-9. AISI 50 XF Steel: Effects of Cold Deformation 155
5-10. AISI 80 DF Steel: Effects of Cold Deformation 156
5-11. Comparison of Three HSLA Steel Grades, Cb, Cb-V and Cb-V-Si: Strain Life
From Constant Amplitude 157
5-12. Comparison of Stress Responses: DPI vs DP2 Dual-Phase HSLA Steels 158
5-13. Dual-Phase HSLA Steel Grade: Stress Response for As-Received vs
Water-Quenched 159
5-14. Dual-Phase HSLA Steel Grade: Stress Response for As-Received vs Gas-Jet-
Cooled 160
5-15. S-N Comparison of Dual-Phase HSLA Steel Grades DPI and DP2 With
1006 161
5-16. Comparison of Dual-Phase HSLA Steel DP2 With HI-FORM 50 162
5-17. Comparison of Cyclic Strain Response Curves for Cb, Cb-V and Cb-V-Si Grades
of HSLA Steel 163
5-18. Fatigue Crack Propagation Rate: Effect of Temperature for Two HSLA Steel
Grades 164
viii Contents

5-19. Effect of R-Ratio and Test Temperature on Crack Propagation of HSLA Steel
Grade I 165
5-20. Effect of Test Temperature on Fatigue Crack Propagation Behavior for Two
HSLA Steel Grades 166
5-21. Stress-Cycle Curves for Weldments of Different HSLA Steel Grades 167
5-22. Weldments (FCA W): SAE 980 X Steel vs 1006 168
5-23. Weldments (TIG): DOMEX 640 XP Steel Welded Joints vs Parent Metal 169
5-24. Weldments (FCAW Dressed by TIG): Fatigue Life Estimates Compared With
Experimental Data for SAE 980 X Steel 170
5-25. SAE 980 X Steel Weldment (FCAW): Smooth Specimen vs TIG-Dressed vs
As-Welded 171
5-26. SAE 980 X Steel Weldment (FCAW): Lap-Shear Joints 172
5-27. Microalloyed HSLA Steels: Properties of Fusion Welds 173
5-28. Microalloyed HSLA Steels: Properties of Spot Welds 174

SECTION 6: High-Strength Alloy Steels 176

6-1. HY-130 Steel: Effect of Notch Radii 176
6-2. 300 M Steel: Effect of Notch Severity on Constant-Lifetime Behavior 177
6-3. TRIP Steels Compared With Other High-Strength Grades 178
6-4. Corrosion Fatigue: Special High-Strength Sucker-Rod Material 179
6-5. Corrosion Fatigue Cracking of Sucker-Rod Material 180
6-6. Hydrogenated Steel: Effect of Baking Time on Hydrogen Concentration 181
6-7. Hydrogenated Steel: Effect of Notch Sharpness 182

SECTION 7: Heat-Resisting Steels 183

7-1. 0.5%Mo Steel: Effect of Hold Time in Air and Vacuum at Different
Temperatures 183
7-2. DIN 14 Steel (1.5 Cr, 0.90 Mo, 0.25 V): Effect of Liquid Nitriding 184
7-3. 2.25Cr-1.0Mo Steel: Influence of Cyclic Strain Range on Endurance Limit in
Various Environments 185
7-4. 2.25Cr-1.0Mo Steel: Effect of Elevated Temperature 186
7-5. 2.25Cr-I.OMo Steel: Effect of Elevated Temperature and Strain Rate 187
7-6. 2.25Cr-1.0Mo Steel: Effect of Temperature on Fatigue Crack Growth Rate 188
7-7. 2.25Cr-1.0Mo Steel: Effect of Cyclic Frequency on Fatigue Crack Growth
Rate 189
7-8. 2.25Cr-1.0Mo Steel: Fatigue Crack Growth Rates in Air and Hydrogen 190
7-9. 2.25Cr-1.0Mo Steel: Effect of Holding Time 191
7-10. Cast 2.25Cr-1.0Mo Steel, Centrifugally Cast: Fatigue Properties at 540°C
(1000 OF) 192
7-11. HII Steel: Crack Growth Rate in Water and in Water Vapor 193
7-12. 9.0Cr-1.0Mo Steel: Creep-Fatigue Characteristics 194
7-13. 9.0Cr-1.0Mo Modified Steel: Stress Amplitudes Developed in Cycling 195
7-14. 9.0Cr-1.0Mo Modified Steel: Effect of Deformation 196

SECTION 8: Stainless Steels 197

8-1. Type 301 Stainless Steel: Scatter Band for Fatigue Crack Growth Rates 197
8-2. Type 301 Stainless Steel: Effects of Temperature and Environment on Fatigue
Crack Growth Rate 198
8-3. Type 304 Stainless Steel: Effect of Temperature on Frequency-Modified
Strains 199
8-4. Type 304 Stainless Steel: Fatigue Crack Growth Rate-Annealed and Cold
Worked 200
8-5. Type 304 Stainless Steel: Effect of Humidity on Fatigue Crack Growth Rate 201
8-6. Type 304 Stainless Steel: Effect of Aging on Fatigue Crack Growth Rate 202
8-7. Type 304 Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rate 203
8-8. Type 304 Stainless Steel: Damage Relation at 650°C (1200 OF) 204
 Contents ix

8-9. Type 304 Stainless Steel: Fatigue Crack Growth Rate at Room and Subzero
Temperatures 205
8-10. Types 304 and 304L Stainless Steel: Effect of Cryogenic Temperatures on Fatigue
Crack Growth Rate 206
8-11. Type 304 Stainless Steel: Fatigue Crack Growth Rate in Air With Variation in
Waveforms 207
8-12. Type 304 Stainless Steel: Effect of Hold Time on Cycles to Failure 208
8-13. Type 304 Stainless Steel: Effect of Hold Time and Continuous Cycling on Fatigue
Crack Growth Rates 209
8-14. Type 304 Stainless Steel: Effect of Cyclic Frequency on Fatigue Crack Growth
Rate 210
8-15. Type 304 Stainless Steel: Effect of Frequency on Fatigue Crack Growth
Behavior 211
8-16. Type 304 Stainless Steel Welded With Type 308: Fatigue Crack Growth Rates 212
8-17. Types 304 and 310 Stainless Steel: Effect of Direction on S-N 213
8-18. Types 304, 316, 321, and 348 Stainless Steel: Effects of Temperature on Fatigue
Crack Growth Rates 214
8-19. Type 309S Stainless Steel: Effect of Grain Size on Fatigue Crack Growth
Rate 215
8-20. Type 310S Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rate 216
8-21. Type 316 Stainless Steel: Growth Rate of Fatigue Cracks in Weldments 217
8-22. Type 316 Stainless Steel: Fatigue Crack Growth Rates-Aged vs Unaged 218
8-23. Type 316 Stainless Steel: Fatigue Crack Growth Rates-Effect of Aging 219
8-24. Type 316 Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rate 220
8-25. Type 316 Stainless Steel: Effect of Cyclic Frequency on Fatigue Crack Growth
Rate 221
8-26. Type 316 Stainless Steel: Fatigue Crack Growth Rate in the Annealed
Condition 222
8-27. Type 316 Stainless Steel: Effect of Environment (Sodium, Helium, and Air) on
Cycles to Failure 223
8-28. Types 316 and 321 Stainless Steel: Effects of Gaseous Environments on Fatigue
Crack Growth Rates 224
8-29. Type 32I Stainless Steel: Effect of Hold Time on Fatigue Crack Growth Rates 225
8-30. Type 403 Stainless Steel: Effect of Environment on Fatigue Crack Growth
Rate 226
8-3I. Type 403 Modified Stainless Steel: Scatter of Fatigue Crack Growth Rates 227
8-32. Type 422 Stainless Steel: Fatigue Crack Growth Rates in Precracked
Specimens 228
8-33. Type 422 Stainless Steel: Fatigue Strength-Longitudinal vs Transverse 229
8-34. Type 422 Stainless Steel: Effect of Temperature on Fatigue Strength 230
8-35. Type 422 Stainless Steel: Effects of Delta Ferrite on Fatigue Strength 231
8-36. 17-4 PH Stainless Steel: Fatigue Crack Growth Rates in Airvs Salt Solution 232
8-37. 15-5 PH Stainless Steel: Fatigue Crack Growth Rates in Air vs Salt Solution 233
8-38. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates at Room
Temperature 234
8-39. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates in Air and Sump Tank
Water 235
8-40. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates at Subzero
Temperatures 236
8-41. PH 13-8 Mo Stainless Steel: Constant-Life Fatigue Diagram 237
8-42. Types 600 and 329 Stainless Steel: S-NCurves for Two Processing Methods 238
8-43. Grade 21-6-9 Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rates 239
8-44. Kromarc 58 Stainless Steel: Effect of Cryogenic Temperatures on Weldments 240
8-45. Pyromet 538 Stainless Steel: Effects of Welding Methods on Fatigue Crack
Growth Rates 241
8-46. Duplex Stainless Steel KCR 171: Corrosion Fatigue 242
x Contents

SECTION 9: Maraging Steels 243

9-1. Grades 200, 250, and 300 Maraging Steel: S-N Curves for Smooth and 'Notched
Specimens 243
9-2. Grade 300 Maraging Steel: Fatigue Life in Terms of Total Strain 244

SECTION 10: Cast Irons 245

10-1. Fatigue of Cast Irons as a Function of Structure-Sensitive Parameters 245
10-2. Gray Iron: Fatigue Life, and Fatigue Limit as a Function of Temperature 246
10-3. Gray Iron: S-N Curves for Unalloyed vs Alloyed 247
10-4. Gray Iron: Effect of Environment 248
10-5. Class 30 Gray Iron: Modified Goodman Diagram for Class 30 249
10-6. Class 30 Gray Iron: Fatigue Crack Growth Rates for Class 30 250
10-7. Gray Irons: Torsional Fatigue for Various Tensile Strength Values 251
10-8. Gray Irons: Torsional Fatigue Data for Five Different Compositions 252
10-9. Gray Irons: Thermal Fatigue-Effect of Aluminum Additions 253
10-10. Gray Irons: Thermal Fatigue-Effect of Chromium and Molybdenum
Additions 254
10-11. Gray Irons: Thermal Fatigue-Room Temperature and 540°C (1000 OF) 255
10-12. Gray Irons: Thermal Fatigue Properties-Comparisons With Ductile Cast Iron
and Carbon Steel 256
10-13. Cast Irons: Thermal Fatigue Properties for Six Grades 257
10-14. Ductile Iron: Effect of Microstructure on Endurance Ratio-Tensile Strength
Relationship 258
10-15. Ductile Iron: Effect of Microstructure on Endurance Ratio-Tensile Strength
Relationship 259
10-16. Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades, Using V-Notched
Specimens 260
10-17. Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades, Using Unnotched
Specimens 261
10-18. Ductile Iron: Fatigue Diagrams for Bending Stresses and Tension-Compression
Stresses 262
10-19. Ductile Iron: Effect of Surface Conditions-As-Cast vs Polished Surface 263
10-20. Ductile Iron: Fatigue Limit in Rotary Bending as Related to Hardness 264
10-21. Ductile Iron: Effect of Rolling on Fatigue Characteristics 265
10-22. Ductile Iron: Effect of Notches on a 65,800-psi-Tensile-Strength Grade 266
10-23. Ductile Iron: Fatigue Crack Growth Rate Compared With That of
Steel 267
10-24. Malleable Iron: S- N Curve Comparisons of Four Grades 268
10-25. Pearlitic Malleable Iron: Effect of Surface Conditions on S-N Curves 269
10-26. Pearlitic Malleable Iron: Effect of Nitriding 270
10-27. Ferritic Malleable Iron: Effect of Notch Radius and Depth 271

SECTION 11: Heat-Resisting Alloys 272

II-I. A286: Effect of Environment 272
11-2. A286: Effect of Frequency on Life at 593°C (1095 OF) 273
11-3. A286: Fatigue Crack Growth Rates at Room and Elevated Temperatures 274
11-4. Astroloy: S-N Curves for Powder vs Conventional Forgings 275
11-5. Astroloy: Powder vs Conventional Forgings Tested at 705°C (1300 OF) 276
11-6. FSX-430: Effect of Grain Size on Cycles to Cracking 277
11-7. FSX-430: Effect of Grain Size on Fatigue Crack Propagation Rate 278
11-8. HS-31: Effect of Testing Temperature 279
11-9. IN 738 LC Casting Alloy: Standard vs HIP'd Material 280
11-10. IN 738 LC: Effect of Grain Size on Cycles to Failure 281
II-II. IN 738 LC: Effect of Grain Size on Cycles to Cracking 282
11-12. IN 738 LC: Effect of Grain Size on Fatigue Crack Propagation Rate 283
11-13. IN 738 LC: Fatigue Crack Growth Rate at 850°C (1560 OF) 284
11-14. Inconel 550: Axial Tensile Fatigue Properties in Air and Vacuum at 1090 K 285
 Contents xi

11-15. Inconel625: Effect of Temperature on Cycles to Failure 286

I 1-16. Inconel 706: Effect of Temperature on Fatigue Crack Growth Rate 287
11-17. Inconel "7I3C": Effect of Elevated Temperatures on Fatigue Characteristics 288
II-18. Inconel "7I3C" and As-Cast HS-31: Comparison of Two Alloys for Number of
Cycles in Thermal Fatigue to Initiate Cracks 289
11-19. Inconel 718: Effect of Frequency on Fatigue Crack Propagation Rate 290
1I-20. Inconel 718: Relationship of Fatigue Crack Propagation Rate With Stress
Intensity 29I
I 1-21. Inconel 718: Relationship of Fatigue Crack Growth Rate With Load/Time Wave-
forms 292
1I-22. Inconel 718: Fatigue Crack Growth Rate in Air vs Helium 293
11-23. Inconel 718: Effect of Environment on Fatigue Crack Growth Rate 294
11-24. Inconel 718: Fatigue Crack Growth Rate in Air Plus 5% Sulfur Dioxide 295
I1-25. lnconel 7I8: Fatigue Crack Growth Rate in Air at Room Temperature 296
11-26. Inconel 718: Fatigue Crack Growth Rate in Air at 316°C (600 OF) 297
11-27. Inconel 718: Fatigue Crack Growth Rate in Air at 427°C (800 OF) 298
11-28. Inconel 718: Fatigue Crack Growth Rate in Air at 538°C (1000 OF) 299
II-29. Inconel 718: Fatigue Crack Growth Rate in Air at 649°C (1200 OF) 300
11-30. Inconel 718: Fatigue Crack Growth Rates at Cryogenic Temperatures 301
1I-31. Inconel 718 and X-750: Fatigue Crack Growth Rates at Cryogenic
Temperatures 302
I 1-32. Inconel X-750: Effect of Temperature on Fatigue Crack Growth Rates 303
I 1-33. Jethete M I52: Interrelationship of Tempering Treatment, Alloy Class, and Testing
Temperature With Fatigue Characteristics 304
11-34. Lapelloy: Interrelationship of Hardness and Strength With Fatigue
Characteristics 305
11-35. MAR-M200: Effect of Atmosphere on Cycles to Failure 306
11-36. MAR-M509: Correlation of Initial Crack Propagation and Dendrite Arm
Spacing 307
11-37. MAR-M509: Correlation Between Number of Cycles Required to Initiate a Crack
and Dendrite Arm Spacing 308
1I-38. MERL 76, P/M: Axial Low-Cycle Fatigue Life of As-HIP'd Alloy at 540°C
(1000 OF) 309
11-39. Nickel-Base Alloys: Effect of Solidification Conditions on Cycles to Onset
of Cracking 310
11-40. Rene 95 (As-HIP): Cyclic Crack Growth Behavior Under Continuous and Hold-
Time Conditions 3I I
11-41. Rene 95: Effect of Temperature on Fatigue Crack Growth Rate 312
I 1-42. S-8 I6: Effect of Notches on Cycles to Failure at 900°C (1650 ° F) 313
11-43. Udimet 700: Fatigue Crack Growth Rates at 850°C (1560 OF) 314
11-44. U-700 and MAR-M200: Comparison of Fatigue Properties 315
1I-45. Waspaloy: Stress-Response Curves 316
I 1-46. X-40: Effect of Grain Size and Temperature on Fatigue Characteristics 317
11-47. Cast Heat-Resisting Alloys: Ranking for Resistance to Thermal Fatigue 318
SECTION 12: Aluminum Alloys 319
12-1. Corrosion-Fatigue Properties of Aluminum Alloys Compared With Those of
Other Alloys 319
12-2. Comparisons of Aluminum Alloys With Magnesium and Steel: Tensile Strength
vs Endurance Limit 320
12-3. Aluminum Alloys (General): Yield Strength vs Fatigue Strength 321
12-4. Comparison of Aluminum Alloy Grades for Crack Propagation Rate 322
12-5. Alloy 1100: Relationship of Fatigue Cycles and Hardness for HO and H 14
Tempers 323
12-6. Alloy 1100: Interrelationship of Fatigue Cycles, Acoustic Harmonic Generation
and Hardness 324
12-7. Alloy 2014-T6: Notched vs Unnotched Specimens; Effect on Cycles to Failure 325
12-8. Alloy 2024-T3: Effect of Air vs Vacuum Environments on Cycles to Failure 326
12-9. Alloy 2024-T4 Alclad Sheet: Effect of Bending on Cycles to Failure 327
xii Contents

12-10. Alloy 2024-T4: High-Cycle vs Low-Cycle Fatigue 328

12-11. Alloy 2024-T4: Relationship of Stress and Fatigue Cycles 329
12-12. Alloy 2024-T4: Dependence of the Average Rocking Curve Halfwidth 7J on Dis-
tance From the Surface 330
12-13. Alloys 2024 and X2024: Effect of Alloy Purity on Cycles to Failure 331
12-14. Alloys 2024 and 2124: Relationship of Particle Size and Fatigue
Characteristics 332
12-15. Alloys 2024-T4 and 2124-T4: Comparison of Resistance to Fatigue Crack
Initiation 333
12-16. Alloys 2024-TJ and 7075-T6: Summary of Fatigue Crack Growth Rates 334
12-17. Alloys 2024-T4 and 7075-T6: Effect of Product Form and Notches 335
12-18. Alloys 2024-T351 and 7075-T73XXX: Comparison of P / M Extrusions and
Rod 336
12-19. Alloy 2048-T851: Longitudinal vs Transverse for Axial Fatigue 337
12-20. Alloy 2048-T851: Notched vs Unnotched Specimens at Room and Elevated
Temperatures 338
12-21. Alloy 2048-T851: Fatigue Crack Propagation Rates in LT and TL
Orientations 339
12-22. Alloy 2048-T85I: Modified Goodman Diagram for Axial Fatigue 340
12-23. Alloy 2219-T851: Dependence of Relaxation Behavior on the Cyclic Hardening
Parameter 341
12-24. Alloy 2219-T851: Effect of Strain Amplitude on the Relaxation of Residual
Surface Stress With Fatigue 342
12-25. Alloy 2219-T851: Relationship of Fatigue Cycles to Different Depth Distributions
of Surface Stress 343
12-26. Alloy 2219-T851: Probability of Fatigue Failure 344
12-27. Alloys 3003-0, 5154-H34 and 6061-T6: Effect of Alloy on Fatigue Characteristics
of Weldments 345
12-28. Alloy 5083-0 Plate: Effect of Orientation on Fatigue Crack Growth Rates 346
12-29. Alloy 5083-0 Plate: Effect of Temperature and Humidity on Fatigue Crack Growth
Rates 347
12-30. Alloys 5086-H34, 5086-H36, 6061-T6, 7075-T73 and 2024-T3: Comparative
Resistance to Axial-Stress Fatigue 348
12-31. Alloys 5083-0/5183: Fatigue Life Predictions and Experimental Data Results for
Double V-Butt Welds 349
12-32. Alloys 5083-0/5183: Predicted Effect of Stress Relief and Stress Ratio on Fatigue
Life of Butt Welds 350
12-33. 7XXX Alloys: Cyclic Strain vs Crack Initiation Life 351
12-34. Alloy 7050: Influence of Alloy Composition and Dispersoid Effect on Mean
Calculated Fatigue Life 352
12-35. Alloy 7050: Effect of Grain Shape on Cycles to Failure 353
12-36. Alloy 7075 (TMP, T6 and T651): Effect of Thermomechanical Processing on Cycles
to Failure 354
12-37. Alloys 7075 and 7475: Effect of Inclusion Density on Cycles to Failure 355
12-38. Alloy 7075: Effect of TMT on Cycles to Failure 356
12-39. Alloys 7075 and 7050: Relative Ranking for Constant Amplitude and Periodic
Overload 357
12-40. Alloy 7075: Effect of Environment and Mode of Loading 358
12-41. Alloy 7075-T6: Effects of Corrosion and Pre-Corrosion 359
12-42. Alloy 7075-T73: Effect of a 3.5% NaCl Environment on Cycles to Failure 360
12-43. Alloy 7075: Effect of Cathodic Polarization on Fatigue Behavior 361
12-44. Alloy 7075-T6: Effect of Surface Treatments and Notch Designs on Number of
Cycles to Failure 362
12-45. Alloy 7075-T6: Effect of R-Ratio on Fatigue Crack Propagation 364
12-46. Alloy 7075: Effect of Predeformation on Fatigue Crack Propagation Rates 365
12-47. Alloys 7075 and 2024-T3: Comparative Fatigue Crack Growth Rates for Two
Alloys in Varying Humidity 366
12-48. Alloy 7075-T65I: Fatigue Life as Related to Harmonic Generation 367
12-49. Alloys 7075-T6 and 7475-T73: Effect of Laser-Shock Treatment on Fatigue
Properties 368
12-50. Alloy 7075-T6: Effect of Laser-Shock Treatment on Hi-Lok Joints 369
 Contents xiii

12-51. Alloy 7075 (High Purity): Effect of Iron and Silicon on Cycles to Failure 370
12-52. Alloy X-7075: Effect of Grain Size on Cycles to Failure 371
12-53. Alloy X-7075: Effect of Grain Size on Stress-Life Behavior 372
12-54. Alloy X-7075: Effect of Environment; Air vs Vacuum 373
14-55. Alloy X-7075: Effect of Environment on Two Different Grain Sizes 374
12-56. Alloy X-7075: Effect of Grain-Boundary Ledges on Cycles to Failure 375
12-57. Alloys X-7075 and 7075: Effects of Chromium Inclusions on Fatigue Crack
Propagation 376
12-58. Alloy 7475-T6: S-N Diagram for a Superplastic Fine-Grain Alloy 377
12-59. Alloy 7475: Effect of Alignment of Grain Boundaries on Cycles to Failure 378
12-60. Alloy 7475-T6: Superplastic vs Nonsuperplastic, as Related to Fatigue Crack
Growth 379
12-61. Alloys X-7075 and 7075: Effect of Chromium-Containing Inclusions on Cycles to
Failure 380
12-62. Aluminum Forging Alloys: Stress Amplitude vs Reversals to Failure 381
12-63. AI-5Mg-0.5Ag: Effect of Condition on Fatigue Characteristics 382
12-64. AI-Zn-Mg and AI-Zn-Mg-Zr: Effect of Grain Size on Strain-Life Behavior 383
12-65. AI-Zn-Mg: Strain-Life Curves of a Large-Grained Alloy 384
12-66. Aluminum With a Copper Overlay: Stress Amplitude vs Cycles to Failure 385
12-67. P/M Alloys 7090 and 7091 vs Extruded 2024 386
12-68. P / M Alloys 7090 and 709I vs 1/ M 7050 and 7075 Products 387
12-69. P/M Aluminum Alloys: Typical Fatigue Behavior 388
12-70. P / M Aluminum Alloys: Comparison With Specimens Made by Ingot
Metallurgy 389
12-71. P/M Aluminum Alloys: Comparison With Forged 7175 for Cycles to
Failure 390
12-72. Various Aluminum Alloys: Comparison of Grades for Corrosion-Fatigue Crack
Growth Rates; Air vs Salt Water 391
12-73. Various Aluminum Alloys: Comparison of Grades for Corrosion-Fatigue Crack
Growth Rates in Salt Water 392
12-74. Various Aluminum Alloys: Wrought vs Cast, and Influence of Casting Method on
Fatigue Life 393
12-75. Aluminum Casting Alloy AL-195: Interrelationship of Fatigue Properties With
Degree of Porosity 394
12-76. Aluminum Casting Alloy LM25-T6: Squeeze Formed vs Chill Cast; Effect on
Reversals to Failure 395

SECTION 13: Copper Alloys 396

13-1. Copper: Effect of Air and Water Vapor on Cycles to Failure 396
13-2. Copper: Applied Plastic-Strain Amplitude vs Fatigue Life 397
13-3. Copper Alloy CI 1000 (ETP Wire): Effect of Temperature on Fatigue
Strength 398
13-4. Copper Alloy C26000 (Cartridge Brass): Influence of Grain Size and Cold Work on
Cycles to Failure 399
13-5. Copper Alloy C83600 (Leaded Red Brass): S-N Curves; Scatter Band 400
13-6. Copper Alloy C86500 (Manganese Bronze): S-N Curves; Scatter Band 401
13-7. Copper Alloys C87500 and C87800 (Silicon Brasses): S-N Curves; Scatter
Band 402
13-8. Copper Alloy C92200 (Navy "M" Bronze): S-N Curves; Scatter Band 403
13-9. Copper Alloy C93700 (High-Leaded Tin Bronze): S-NCurves; Scatter Band 404
13-10. Copper Alloy No. 192: Effect of Salt Spray on Tubes 405
13-1 I. Copper Alloy 955: Goodman-Type Diagram 406

SECTION 14: Magnesium Alloys 407

14-1. Magnesium Casting Alloy QE22A-T6: Effects of Notches and Testing
Temperature 407
14-2. Magnesium Casting Alloy QH2 IA-T6: S- N Curves; Effects of Notches and Testing
Temperature 408
14-3. Mg-AI-Zn Casting Alloys: Effects of Surface Conditions on Fatigue
Properties 409
xiv Contents

SECTION 15: Molybdenum 410

15-1. Molybdenum: Fatigue Limit Ratio vs Temperature 410
SECTION 16: Tin Alloys 411
16-1. Tin-Lead Soldering Alloy: S-N Data for Soldered Joints 411
16-2. Babbitt: Variation of Bearing Life With Babbitt Thickness 412
16-3. SAEI2 Bearing Alloy: Effect of Temperature on Fatigue Life 413
SECTION 17: Titanium and Titanium Alloys 414
17-1. Unalloyed Titanium, Grade 3: S-N Curves for Annealed vs Cold Rolled 414
17-2. Unalloyed Titanium, Grade 4: S-N Curves for Three Testing
Temperatures 415
17-3. Ti-24V and Ti-32V: Stress Amplitude vs Cycles to Failure 416
17-4. Ti-5AI-2.5Sn: Effects of Notches and Types of Surface Finish 417
17-5. Ti-5AI-2.5Sn and Ti-6AI-4V: Fatigue Crack Growth Rates 418
17-6. Ti-6AI-6V-2Sn: Effects of Machining and Grinding 419
17-7. Ti-6AI-6V-2Sn (HIP): S-N Curves for Titanium Alloy Powder Consolidated
by HIP 420
17-8. Ti-6AI-6V-2Sn (HIP): S-N Curves for Annealed Plate vs HIP 421
17-9. Ti-6AI-2Sn-4Zr-2Mo: Bar Chart Presentation on Effects of Machining and
Grinding 422
17-10. Ti-6AI-2Sn-4Zr-2Mo: Constant-Life Fatigue Diagram 423
17-11. Ti-6AI-2Sn-4Zr-6Mo: Low-Cycle Axial Fatigue Curves 424
17-12. Ti-8Mo-2Fe-3AI: S-NCurves; 'Solution Treated and Aged Condition 425
17-13. Ti-IOV-2Fe-3AI: S-N Curves; Notched vs Unnotched Specimens in Axial
Fatigue 426
17-14. Ti-IOV-2Fe-3AI and Ti-6AI-4V: Comparison of Fatigue Crack Growth
Rates 427
17-15. Ti-IOV-2Fe-3AI: S-N Curve; Notched Bar Fatigue Life for a Series of Forgings
Compared With Ti-6AI-4V Plate 428
17-16. Ti-I3V-IICr-3AI: Constant-Life Fatigue Diagrams 429
17-17. Ti-6AI-4V: Effect of Condition and Notches on Fatigue Characteristics 430
17-18. Ti-6AI-4V: Effect of Direction on Endurance 431
17-19. Ti-6AI-4V: Effect of Isothermally Rolled vs Extruded Material on Cycles to
Failure 432
17-20. Ti-6AI-4V: Comparison of Wrought vs Isostatically Pressed Material for Cycles
to Failure 433
17-21. Ti-6AI-4V: Effect of Fretting and Temperature on Cycles to Failure 434
17-22. Ti-6AI-4V (Beta Rolled): Effect of Finishing Operations on Cycles to
Failure 435
17-23. Ti-6AI-4V: Effect of Yield Strength on Stress-Life Behavior 436
17-24. Ti-6AI-4V: Effect of Stress Relief on Cycles to Failure 437
17-25. Ti-6AI-4V: Interrelationship of Machining Practice and Cutting Fluids on Cycles to
Failure 438
17-26. Ti-6AI-4V: Relative Effects of Machining and Grinding Operations on Endurance
Limit 439
17-27. Ti-6AI-4V: Effects of Various Metal Removal Operations on Endurance
Limit 440
17-28. Ti-6AI-4V: Effect of Texture on Fatigue Strength 441
17-29. Ti-6AI-4V: Effect of Complex Texture on Cycles to Failure 422
17-30. Ti-6AI-4V: Effect of Texture and Environment on Cycles to Failure 443
17-31. Ti-6AI-4V: Fatigue Crack Growth Rates 444
17-32. Ti-6AI-4V: Fatigue Crack Growth Rates for ISR Tee, and Extrusions 445
17-33. Ti-6AI-4V: Fatigue Crack Growth Rates 446
17-34. Ti-6AI-4V: Effect of Final Cooling on Fatigue Crack Growth Rates 447
17-35. Ti-6AI-4V: Effect of Dwell Time on Fatigue Crack Growth Rates 448
17-36. Ti-6AI-4V: Fatigue Crack Growth Data 449
17-37. Ti-6AI-4V P / M: Comparison of HIP'd Material With Alpha-Beta Forgings for
Cycles to Failure 450
 Contents xv

17-38. Ti-6AI-4V PI M: Comparisons of HIP'd Material With Annealed Plate for Cycles
to Failure 45 I
17-39. Ti-6AI-4V P/M: Effect of Powder Mesh Size on Fatigue Properties 452
17-40. Ti-6AI-4V P/M: Comparison of Blended Elemental, Prealloyed and Wrought
Material for Effect on Cycles to Failure 453
17-41. Ti-6AI-4V: P/M Compacts vs 11M Specimens: Cycles to Failure 454
17-42. Ti-6AI-4V: Comparison of Specimens Processed by Various Fabrication Processes
for Cycles to Failure 455
17-43. Ti-6AI-4V: Comparison of Fatigue Crack Growth Rate, PI M vs II M 456
17-44. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for Cycles to Failure 457
17-45. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for Cycles to Failure 458
17-46. Ti-6AI-4V EB Weldments: Base Metal Compared With Flawless Weldments 459
17-47. Ti-6AI-4V EB Weldments: Effects of Porosity on Cycles to Failure 460
17-48. Ti-6AI-4V Gas Metal-Arc Weldments: Effects of Porosity on Cycles to
Failure 461
17-49. Ti-6AI-4V: Unwelded vs Electron Beam Welded Material for Cycles to
Failure 462
17-50. Ti-6AI-4V: S-N Diagram for Laser-Welded Sheet 463
17-51. Ti-6AI-4V (Cast): S-N Diagram for Notched Specimens 464
SECTION 18: Zirconium 465
18-1. Zirconium 702: Effects of Notches and Testing Temperature on Cycles to
Failure 465
SECTION 19: Steel Castings 466
(For other data on steel castings see Sections 3,4 and 5, on carbon and
alloy steels.)
19-1. Steel Castings (General): Effect of Design and Welding Practice on Fatigue
Characteristics 466
19-2. Steel Castings (General): Effects of Discontinuities on Fatigue
Characteristics 467
SECTION 20: Closed-Die Forgings 468
(See also under specific grades of alloys.)
20-1. Closed-Die Steel Forgings: Effect of Surface Condition on Fatigue Limit 468
SECTION 21: Powder Metallurgy Parts 469
(See also under specific alloys.)
21-1. P/M: Relation of Density to Fatigue Limit and Fatigue Ratio 469
21-2. PI M: Relation of Fatigue Limit to Tensile Strength for Sintered Steels 470
21-3. PI M (Nickel Steels): As-Sintered vs Quenched and Tempered for Cycles to
Failure 471
2 I-4. PI M (Nickel Steels): Relation Between Fatigue Limit and Tensile Strength for
Sintered Steels 472
21-5. P/M (Nickel Steels): Effect of Notches on Cycles to Failure for the As-Sintered
Condition 473
21-6. PI M (Nickel Steels): Effect of Notches on Cycles to Failure for the Quenched and
Tempered Condition 474
21-7. P/M (Low-Carbon, 1-5%Cu): Effects of Notches and Nitriding on Cycles to
Failure 475
2 I-8. PI M (Sintered Iron, Low-Carbon, No Copper): Effect of Density and Nitriding on
Cycles to Failure 476
21-9. P/M: Effect of Nitriding on Ductile Iron and Sintered Iron (3%Cu) for Cycles to
Failure 477
SECTION 22: Composites 478
22-1. Brass/ Mild Steel Composite: Comparison of Brass-Clad Mild Steel With Brass and
Mild Steel for Cycles to Failure 478
22-2. Stainless Steell Mild Steel Composite: Comparison of Stainless-Clad Mild Steel
With Stainless Steel and Mild Steel for Cycles to Failure 479
xvi Contents

SECTION 23: Effects of Surface Treatments 480

23-1. Carbon and Alloy Steels (Seven Grades): Effects of Nitrocarburizing on Fatigue
Strength 480
23-2. Carbon and Alloy Steels (Seven Grades): Effects of Tufftriding on Fatigue
Characteristics 481
23-3. Carbon and Alloy Steels (Six Grades): Effects of Nitriding on Fatigue
Strength 482
23-4. Carbon-Manganese Steel: Effects of Nickel Coating on Fatigue Strength 483
SECTION 24: Test Results for Component Parts 484
24-1. Coil Springs, Music Wire (Six Sizes): Data Presented by Means of a Goodman
Diagram 484
24-2. Coil Springs: S-N Data for Oil-Tempered and Music Wire Grades 485
24-3. Coil Springs: Effects of Shot Peening on Cycles to Failure 486
24-4. Coil Springs, 8650 and 8660 Steels: Relation of Design Stresses and Probability of
Failure 487 .
24-5. Coil Springs, HSLA Steels: Effects of Corrosion on Cycles to Failure 488
24-6. Leaf Springs, 5160 Steel: Maximum Applied Stress vs Cycles to Failure 489
24-7. Front Suspension Torsion Bar Springs, 5160H Steel: Distribution of Fatigue
Results for Simulated Service Testing 490
24-8. Gears, Carburized Low-Carbon Steel: Relation of Life Factor to Required
Life 491
24-9. Gears, Carburized Low-Carbon Steel: Bending Stress vs Cycles to Failure 492
24-10. Gears, Carburized Low-Carbon Steel: Effect of Shot Peening on Cycles to
Failure 493
24-11. Gears, Carburized Low-Carbon Steel: Probability-Stress-Life Design Curves 494
24-12. Gears, 8620H Carburized: Bending or Contact Stress vs Cycles to Fracture or
Pitting 495
24-13. Gears, 8620H Carburized: A Weibull Analysis of Bending Fatigue Data 496
24-14. Gears, 8620H Carburized: T-N Curves for Six-Pinion, Four-Square Tests 497
24-15. Hypoid Gears, 8620H Carburized: Minimum Confidence Level; Stress vs Cycles to
Rupture 498
24-16. Hypoid, Zero I and Spiral Bevel Gears, 8620H Carburized: S-NScatter Band and
Minimum Confidence Level 499
24-17. Spiral Bevel and Zero I Bevel Gears, 8620H Carburized: S-N Scatter Band and
Minimum Confidence Level 500
24-18. Gears, 8620H Case Hardened: Relation of Life Factor to Cycles to Rupture 501
24-19. Bevel Gears, Low-Carbon Steel Case Hardened: Relation of Life Factor to Cycles
to Rupture for Various Confidence Levels 502
24-20. Gears, AMS 6265: S-N Data for Cut vs Forged 503
24-21. Spur Gears, 8620H: S-N Data for Cut vs Forged 504
24-22. Gears and Pinions: PIM 4600V vs 4615; Weibull Distributions 505
24-23. Gears and Pinions: PIM Grades 4600V and 2000 vs 4615; Percent Failure vs
Time 506
24-24. Gear Steel AMS 6265: Parent Metal vs Electron Beam Welded 507
24-25. Gears, 42 CrMo4 (German Specification): S-N Curves for Various Profiles 508
24-26. Gears, 42 CrMo4 (German Specification): Endurance Test Results in the Weibull
Distribution Diagram 509
24-27. Bolts, 1040 and 4037 Steels: Maximum Bending Stress vs Number of Stress
Cycles 510
24-28. Bolts: S-N Data for Roll Threading Before and After Heat Treatment 511
24-29. Power Shafts, AMS 6382 and AMS 6260: Electron Beam Welded vs Silver Brazed
Joints 512
24-30. Axle Shafts, 1046, 1541 and 50B54 Steels: S-N Data for Induction Hardening vs
Through Hardening 513
24-31. Steel Rollers, 8620H Carburized: Effects of Carburizing Temperature and
Quenching Practice on Surface Fatigue 514
 Contents xvii

24-32. Steel Rollers, 8620H Carburized: Effects of Carburizing Temperature and

Quenching Practice on Surface Fatigue 515
24-33. Linkage Arm, Cast Low-Carbon Steel: Starting Crack Size vs Cycles to
Failure 516
24-34. Notched Links, Hot Rolled Low-Carbon Steel: S-N Data for Component Test
Model 517
24-35. Fuselage Brace, Ti-6AI-6V-2Sn: Fatigue Endurance of HIP-Consolidated
Powder 518

Fatigue Testing
Introduction
Fatigue is the progressive, localized, perma-
nent structural change that occurs in materials
subjected to fluctuating stresses and strains that
may result in cracks or fracture after a sufficient
number of fluctuations. Fatigue fractures are
caused by the simultaneous action of cyclic
stress, tensile stress and plastic strain. If anyone
of these three is not present, fatigue cracking will
not initiate and propagate. The cyclic stress
starts the crack; the tensile stress produces crack
growth (propagation). Although compressive
stress will not cause fatigue, compression load
may do so.
The process of fatigue consists of three stages:
• Initial fatigue damage leading to crack nu-
cleation and crack initiation
• Progressive cyclic growth of a crack (crack
propagation) until the remaining un cracked
cross section of a part becomes too weak to
sustain the loads imposed
• Final, sudden fracture of the remaining
cross section
Fatigue cracking normally results from cyclic
stresses that are well below the static yield
strength of the material. (In low-cycle fatigue,
however, or if the material has an appreciable
work-hardening rate, the stresses also may be
above the static yield strength.)
Fatigue cracks initiate and propagate in re-
gions where the strain is most severe. Because
most engineering materials contain defects and
thus regions of stress concentration that intensify
strain, most fatigue cracks initiate and grow
from structural defects. Under the action of cy-
clic loading, a plastic zone (or region of deforma-
tion) develops at the defect tip. This zone of high
deformation becomes an initiation site for a fa-
tigue crack. The crack propagates under the ap-
plied stress through the material until complete
fracture results. On the microscopic scale, the
most important feature of the fatigue process is
nucleation of one or more cracks under the influ-
1
ence of reversed stresses that exceed the flow
stress, followed by development of cracks at per-
sistent slip bands or at grain boundaries.
Prediction of Fatigue Life
The fatigue life of any specimen or structure is
the number of stress (strain) cycles required to
cause failure. This number is a function of many
variables, including stress level, stress state, cy-
clic wave form, fatigue environment, and the
metallurgical condition of the material. Small
changes in the specimen or test conditions can
significantly affect fatigue behavior, making ana-
lytical prediction of fatigue life difficult. There-
fore, the designer may rely on experience with
similar components in service rather than on
laboratory evaluation of mechanical test speci-
mens. Laboratory tests, however, are essential in
understanding fatigue behavior, and current
studies with fracture mechanics test specimens
are beginning to provide satisfactory design
criteria.
Laboratory fatigue tests can be classified as
crack initiation or crack propagation. In crack
initiation testing, specimens or parts are sub-
jected to the number of stress cycles required for
a fatigue crack to initiate and to subsequently
grow large enough to produce failure.
In crack propagation testing, fracture mechan-
ics methods are used to determine the crack
growth rates of preexisting cracks under cyclic
loading. Fatigue crack propagation may be
caused by cyclic stresses in a benign environ-
ment, or by the combined effects of cyclic stresses
and an aggressive environment (corrosion fa-
tigue).
Fatigue Crack Initiation
Most laboratory fatigue testing is done either
with axial loading, or in bending, thus producing
only tensile and compressive stresses. The stress
usually is cycled either between a maximum and
a minimum tensile stress, or between a maximum
tensile stress and a maximum compressive stress.
2 Fatigue Testing

The latter is considered a negative tensile stress, 1100

I
I
is given an algebraic minus sign, and therefore is 1000 2340 steel 150 =
known as the minimum stress. o~ '"' 48 HRC
The stress ratio is the algebraic ratio of two <ll 900 ~ (unnotchedl'r
a.. 125 'iii
specified stress values in a stress cycle. Two :2: 800 1::-- - - - .:.!.

commonly used stress ratios are the ratio, A, of <I)" 700

Fatigue limit Sf
1 -- <I)"
the alternating stress amplitude to the mean 2340 steel-= 100
stress (A = Sal S m) and the ratio, R, of the min- u
Q)
600 ~ 48 HRC _ u
Q)

~
imum stress to the maximum stress (R = Sminl .; '1... (notched) .~
500 _. 75
Smax)'
0.
E
400
f-
Fatigue limit, Sf -cr.... 0.
E
If the stresses are fully reversed, the stress ratio <ll
l
<ll

R becomes -1; if the stresses are partially re-

versed, R becomes a negative number less than 1.
If the stress is cycled between a maximum stress
and no load, the stress ratio R becomes zero. If
the stress is cycled between two tensile stresses,
the stress ratio R becomes a positive number less
than 1. A stress ratio R of 1indicates no variation
in stress, making the test a sustained-load creep
test rather than a fatigue test.
Applied stresses are described by three pa-
rameters. The mean stress, S m' is the algebraic
average of the maximum and minimum stresses
in one cycle, S m = (S max + Smin) / 2. In the com-
pletely reversed cycle test, the mean stress is zero.
The range of stress, S" is the algebraic difference
between the maximum and minimum stresses in
one cycle, S, = Smax - Smin' The stress amplitude,
S a' is one half the range of stress, Sa = S,/ 2 =
(Smax - Smin)/2.
During a fatigue test, the stress cycle usually is
maintained constant so that the applied stress
conditions can be written Sm± SO' where S mis the
static or mean stress, and Sa is the alternating
stress, which is equal to half the stress range.
Nomenclature to describe test parameters in-
volved in cyclic stress testing are shown in Fig. 1.
Smax I---~:------;...----r---,----
Time
Ol-\-.........,~-----Ic-I-'--~r--f--+----
Smin L..- --""- ......._ - ' -_ _
Fig. 1 Nomenclature to describe test
parameters involved in cyclic stress
testing
S-N Curves. The results offatigue crack initia-
tion tests usually are plotted as maximum stress,
minimum stress, or stress amplitude to number
of cycles, N, to failure using a logarithmic scale
for the number of cycles. Stress is plotted on
either a linear or a logarithmic scale. The result-
<Il
<Il Aluminum alloy 50 <Il
<Il
~ 300 """Q; ~ ,}075-T6 ~,
U5 U5
200 25
j-a~-
Stress ratio (R) = - 1
Number of cycles to fracture, N
Fig. 2 Typical S-N curves for constant
amplitude and sinusoidal loading
ing plot of the data is an S-N curve. Three typical
S-N curves are shown in Fig. 2.
The number of cycles of stress that a metal can
endure before failure increases with decreasing
stress. For some engineering materials such as
steel (see Fig. 2) and titanium, the S-N curve be-
comes horizontal at a certain limiting stress.
Below this limiting stress, known as the fatigue
limit or endurance limit, the material can endure
an infinite number of cycles without failure.
Fatigue Limit and Fatigue Strength. The hor-
izontal portion of an S-N curve represents the
maximum stress that the metal can withstand for
an infinitely large number of cycles with 50%
probability of failure and is known as the fatigue
(endurance) limit, Sf' Most nonferrous metals do
not exhibit a fatigue limit. Instead, their S-N
curves continue to drop at a slow rate at high
numbers of cycles, as shown by the curve for
aluminum alloy 7075-T6 in Fig. 2.
For these types of metals, fatigue strength
rather than fatigue limit is reported, which is the
stress to which the metal can be subjected for a
specified number of cycles. Because there is no
standard number of cycles, each table of fatigue
strengths must specify the number of cycles for
which the strengths are reported. The fatigue
strength of nonferrous metals at 100million (108)
or 500 million (5 X 108) cycles is erroneously
called the fatigue limit.
Low-Cycle Fatigue. For the low-cycle fatigue
region (N< 104 cycles) tests are conducted with
controlled cycles of elastic plus plastic strain,
 Introduction 3

rather than with controlled load or stress cycles.

Under controlled strain testing, fatigue life be-
havior is represented by a log-log plot of the total
strain range, dE, versus the number of cycles to
failure (Fig. 3).
The total strain range is separated into elastic
and plastic components. For many metals and
alloys, the elastic strain range, dE eo is equal to the
stress range divided by the modulus of elasticity.
The plastic strain range, dE p' is the difference be-
tween the total strain range and the elastic strain
range. 10- 3 L..-_ _--I. --L. -l.. ..J

Stress-Concentration Factor. Stress is concen- 10- 1 10 103 105

trated in a metal by structural discontinuities, Cycles to failure
such as notches, holes, or scratches, which act as
stress raisers. The stress-concentration factor, Fig.3 Typical plot of strain range versus
cycles-to-failure for low-cycle fatigue
K" is the ratio of the area test stress in the region
of the notch (or other stress concentrators) to the cling of notched specimens that have been pre-
corresponding nominal stress. For determina- cracked in fatigue. Crack length is measured as a
tion of K" the greatest stress in the region of the function of elapsed cycles, and these data are
notch is calculated from the theory of elasticity, subjected to numerical analysis to establish the
or equivalent values are derived experimentally. rate of crack growth, da l d N,
The fatigue notch factor, Kf> is the ratio of the Crack growth rates are expressed as a function
fatigue strength of a smooth (unnotched) speci- of the crack tip stress-intensity factor range, dK.
men to the fatigue strength of a notched speci- The stress-intensity factor is calculated from ex-
men at the same number of cycles. pressions based on linear elastic stress analysis
Fatigue notch sensitivity, q, for a material is and is a function of crack size, load range, and
determined by comparing the fatigue notch fac- cracked specimen geometry. Fatigue crack growth
tor, K J, and the stress-concentration factor, K" data are typically presented in a log-log plot of
for a specimen of a given size containing a stress da/dNversus s« (Fig. 4).
concentrator of a given shape and size. A com-
mon definition of fatigue notch sensitivity is q = :J.K. ksivTn.
(KJ - l)f(K, - 1), in which q may vary between o 5. 10 15 20 25 30 35
1.-----r-"T"T"--r--r-~--,.,..._,_-,-,
zero (where K J = 1) and 1 (where KJ = K,). This
value may be stated as percentage.
10- 1 I---+_--+----;:--et-'~-__l
Fatigue Crack Propagation
In large structural components, the existence 10- 3
Q)
of a crack does not necessarily imply imminent 10- 2 U
failure of the part. Significant structural life may >-
~
remain in the cyclic growth of the crack to a size c::
10- 4 <:-
at which a critical failure occurs. The objective of
10- 3 "tl
fatigue crack propagation testing is to determine
the rates at which subcritical cracks grow under ~
cyclic loadings prior to reaching a size critical for 10 -5
fracture. 10-4 I-~=---+_--+----+---__l
The growth or extension of a fatigue crack
under cyclic loading is principally controlled by o~
maximum load and stress ratio. However, as in 10- 6
10- 5 L..-_ _.L-_ _..I.--- _ _-L._ _- '
crack initiation, there are a number of additional o 10 20 30 40
factors that may exert a strong influence, includ-
ing environment, frequency, temperature, and J.K. MPa\ m
grain direction. Fatigue crack propagation test- Fig. 4 Fatigue crack propagation rate
ing usually involves constant-load-amplitude cy- data in 7075-T6 aluminum alloy (R <0)

LIVE GRAPH
Click here to view
4 Fatigue Testing
Fatigue Crack Initiation
Crack initiation tests are procedures in which
a specimen or part is subjected to cyclic loading
to failure. A large portion of the total number of
cycles in these tests is spent initiating the crack.
Although crack initiation tests conducted on
small specimens do not precisely establish the fa-
tigue life of a large part, such tests do provide
data on the intrinsic fatigue crack initiation be-
havior of a metal or alloy. As a result, such data
can be utilized to develop criteria to prevent fa-
tigue failures in engineering design. Examples of
the use of small-specimen fatigue test data can be
found in the basis of the fatigue design codes for
boilers and pressure vessels, complex welded, riv-
eted, or bolted structures, and automotive and
aerospace components.
Fatigue Cracking·
Fatigue cracks normally result from cyclic
stresses that are below the yield strength of the
metal. In low-cycle fatigue, however, the cyclic
stress may be above the static yield strength, es-
pecially in a material with an appreciable work-
hardening rate. Generally, a fatigue crack is in-
itiated at a highly stressed region of a component
subjected to cyclic loading of sufficient magni-
tude. The crack then propagates in progressive
cyclic growth through the cross section of the
part until the maximum load cannot be carried,
and complete fracture results.
Crack Nucleation. A variety of crystallo-
graphic features have been observed to nucleate
fatigue cracks. In pure metals, tubular holes that
develop in persistent slip bands, slip band extru-
sion-intrusion pairs at free surfaces, and twin
boundaries are common nucleation sites. Grain
boundaries in polycrystalline metals, even in the
absence of inherent grain boundary weakness,
are crack nucleation sites. At high strain rates,
this appears to be the preferred site. Nucleation
at grain boundaries appears to be a geometrical
effect, whereas nucleation at twin boundaries is
associated with active slip on crystallographic
planes immediately adjacent and parallel to the
twin boundary.
The foregoing processes also occur in alloys
and heterogeneous materials. However, alloying
and commercial production practices introduce
segregation, inclusions, second-phase particles,
and other features that disturb the structure. All
of these phenomena have a significant influence
on the crack nucleation process. In general, al-
loying that (1) enhances cross slip, (2) enhances
twinning, or (3) increases the rate of work hard-
ening will stimulate crack nucleation. On the
other hand, alloying usually raises the flow stress
of a metal, thus offsetting its potentially detri-
mental effect on fatigue crack nucleation.
Crack Initiation. Fatigue cracks initiate at
points of maximum local stress and minimum
local strength. The local stress pattern is deter-
mined by the shape of the part and by the type
and magnitude of the loading. In addition to the
geometric features of a part, features such as sur-
face and metallurgical imperfections can act to
concentrate stress locally. Surface imperfections
such as scratches, dents, burrs, cuts, and other
manufacturing flaws are the most obvious sites
at which fatigue cracks initiate. Except for in-
stances where internal defects or special surface-
hardening treatments are involved, fatigue cracks
initiate at the surface.
Relation to Environment. Corrosion fatigue
describes the degradation of the fatigue strength
of a metal by the initiation and growth of cracks
under the combined action of cyclic loading and
a corrosive environment. Because it is a synergis-
tic effect of fatigue and corrosion, corrosion fa-
tigue can produce a far greater degradation in
strength than either effect acting alone or by su-
perposition of the singular effects. An unlimited
number of gaseous and liquid mediums may af-
fect fatigue crack initiation in a given material.
Fretting corrosion, which occurs from relative
motion between joints, may also accelerate fa-
tigue crack initiation.
Fatigue Testing Regimes
The magnitude of the nominal stress on a cy-
clically loaded component frequently is mea-
sured by the amount of overstress-that is, the
amount by which the nominal stress exceeds the
fatigue limit or the long-life fatigue strength of
the material used in the component. The number
of load cycles that a component under low over-
stress can endure is high; thus, the term high-
cycle fatigue is often applied.
As the magnitude of the nominal stress in-
creases, initiation of multiple cracks is more
likely. Also, spacing between fatigue striations,
which indicate the progressive growth of the
crack front, is increased, and the region of final
fast fracture is increased in size.
 Fatigue Crack Initiation
Low-cycle fatigue is the regime characterized
by high overstress. The arbitrary, but commonly
accepted, dividing line between high-cycle and
low-cycle fatigue is considered to be about 104 to
105 cycles. In practice, this distinction is made by
determining whether the dominant component
of the strain imposed during cyclic loading is
elastic (high cycle) or plastic (low cycle), which in
turn depends on the properties of the metal as
well as the magnitude of the nominal stress.
Presentation of Fatigue Data. High-cycle fa-
tigue data are presented graphically as stress (S)
versus cycles-to-failure (N) in S-N diagrams or
S- N curves. These are described in the Introduc-
tion to this Section along with the symbols and
nomenclature commonly applied in fatigue test-
ing. Because the stress in high-cycle fatigue tests
is usually within the elastic range, the calculation
of stress amplitude, stress range, or maximum
stress on the S-axis is made using simple equa-
tions from mechanics of materials; i.e., stress cal-
culated using the specimen dimensions and the
controlled load or deflection applied axially, in
flexure, or in torsion.
Figure 5 illustrates a stress-strain loop under
controlled constant-strain cycling in a low-cycle
fatigue test. During initial loading, the stress-
strain curve is O-A-B. Upon unloading, yielding
begins in compression at a lower stress C due to
Fig. 5 Stress-strain loop for constant-
strain cycling
5
the Bauschinger effect. In reloading in tension, a
hysteresis loop develops. The dimensions of this
loop are described by its width df (the total
strain range) and its height da (the stress range).
The total strain range df consists of an elastic
strain component df e = dalE and a plastic strain
component df p •
The width of the hysteresis loop depends on
the level of cyclic strain. When the level of cyclic
strain is small, the hysteresis loop becomes very
narrow. For tests conducted under constant df,
the stress range da usually changes with an in-
creasing number of cycles.
The common method of presenting low-cycle
fatigue data is to plot either the plastic strain
range, df p' or the total strain range, df, versus N.
When plotted using log-log coordinates, a straight
line can befit to the dfp-Nplot. The slope of this
line in the region where plastic strain dominates
has shown little variation for the large number of
metals and alloys tested in low-cycle fatigue, the
average value being Y2. This power-law relation-
ship between dfpand Nis known as the Coffin-
Manson relationship. Figure 6 is an example of
the typical presentation of low-cycle fatigue test
results.
Classification of Fatigue
Testing Machines
Fatigue test specimens are primarily described
by the mode of loading:
• Direct (axial) stress
• Plane bending
• Rotating beam
• Alternating torsion
• Combined stress
Testing machines, however, may be universal-
type machines that are capable of conducting all
of the above modes ofloading, depending on the
fixturing used.
Fatigue Testing Machine Components
Whether simple or complex, all fatigue testing
machines consist of the same basic components:
a load train, controllers, and monitors. The load
train consists of the load frame, gripping devices,
test specimen, and drive (loading) system. Typi-
cal load train components in an electrohydraulic
axial fatigue machine are shown in Fig. 7.
The load frame is the structure of the machine
that reacts to the forces applied to the specimen
by the drive system.
6 Fatigue Testing

Ia-...
IU
r--..... ......
Q. r-, <,
'"
"1 n
<Ii ........ --,
01
C
10- 1
~ ~
~
C

s ~
1Il

.~
"- ..........
1Il
<II 10- 2 ~
a:::
~ ~ O.n
00
~ r-,
0
.......
10
"
Cycles to failure, N

Fig. 6 Low-cycle fatigue curve (t1€p versus N) for type 347 stainless steel

Control Systems. The controls and controllers

Fig. 7 Schematic of the load train in an
electrohydraulic axial fatigue machine
The drive system is the most significant feature
of a fatigue testing system and usually is electri-
cally powered. The simplest systems use electric
motors to act on test specimens via cams, levers,
or rotating grips. In electrohydraulic machines,
the motors drive hydraulic pumps to provide
service pressure for control of the motion and
force of a hydraulic piston actuator. Electro-
magnetic excitation can be used to excite a mass
or inertia system to load a specimen.
manually or automatically initiate power and
test, adjust, and maintain the controlled test pa-
rameter(s). Controllers also terminate the test at
a predefined status (failure,' load drop, exten-
sion, or deflection limit). The control of time-
varying deflection or displacement can be ob-
tained in mechanical systems by cam-operated
deflection levels, a rotating eccentric mass, or
hydraulically through a piston limited by stops.
Control in most simple machines and drive
systems is obtained via the open-loop mode. In
such systems, the magnitude of force and dis-
placement initially set by the control system re-
mains constant throughout the test.
Sensors are required to measure the load,
strain, displacement, deflection, and cycle count.
Some devices provide an output signal to the
controller, or to a readout device in the case of
uncontrolled parameters. Common sensors are
load cells (resistance strain gage bridges cali-
brated to load) inserted in the load train. Pres-
sure transducers are used in hydraulic or pneu-
matic actuator devices.
Loading fixtures to alter the mode of loading
provide versatility. Fixtures can be designed to
convert the axial force provided by a hydraulic
actuator to perform four-point bending or tor-
sion testing. Similarly, fixtures attached to an
oscillating platen of a rotating-eccentric-mass-
type machine can facilitate axial, bending, and
torsion fatigue testing of specimens.
 Fatigue Crack Initiation
{a} {bl {e}
(a) Standard grip body for wedge-type grips. (b) V-grips for rounds for use in standard grip body. (c) Flat grips for
specimens for use in standard grip body. (d) Universal open-front holders. (e) Adapters for special samples (screws,
bolts, studs, etc.) for use with universal open-front holders. (f) Holders for threaded samples. (g) Snubber-type wire
grips for flexible wire or cable.
Fig. 8 Grip designs used for axial fatigue testing
Grips. Proper gripping is not simply the at-
tachment of the test specimen in the load train.
Grip failure sometimes occurs prior to specimen
failure. Frequently, satisfactory gripping evolves
after specimen design development. Care must
be taken in grip design and specimen installation
in the grips to prevent misalignment. The grips
shown in Fig. 8 are typical of those used for axial
fatigue tests.
Axial (Direct-Stress)
Fatigue Testing Machines
The direct-stress fatigue testing machine sub-
jects a test specimen to a uniform stress or strain
through its cross section. For the same cross sec-
tion, an axial fatigue testing machine must be
able to apply a greater force than a static bending
machine to achieve the same stress.
Electromechanical systems have been devel-
oped for axial fatigue studies. Generally, these
are open-loop systems, but often have partial
closed-loop features to continuously correct
mean load.
In crank and lever machines, a cyclic load is
applied to one end of the test specimen through a
deflection-calibrated lever that is driven by a
variable-throw crank. The load is transmitted to
the specimen through a flexure system, which
provides straight-line motion to the specimen.
The other end of the specimen is connected to a
hydraulic piston that is part of an electrohydraul-
ically controlled load-maintaining system that
senses specimen yielding. This system automati-
cally and steplessly restores the preset load
through the hydraulic piston.
Servohydraulic closed-loop systems offer op-
timum control, monitoring, and versatility in fa-
tigue testing systems. These can be obtained as
7
{dl lei Ifl (g)
component systems and can be upgraded as re-
quired. A hydraulic actuator typically is used to
apply the load in axial fatigue testing.
Electromagnetic or magnetostrictive excita-
tion is used for axial fatigue testing machine
drive systems, particularly when low-load ampli-
tudes and high-cycle fatigue lives are desired in
short test durations. The high cyclicfrequency of
operation of these types of machines enables test-
ing to long fatigue lives (> 108 cycles) within
weeks.
Bending Fatigue Machines
The most common types of fatigue machines
are small bending fatigue machines, In general,
these simple, inexpensive systems allow labora-
tories to conduct extensive test programs with a
low equipment investment.
Cantilever beam machines, in which the test
specimen has a tapered width, thickness, or di-
ameter, result in a portion ofthe test area having
uniform stress with smaller load requirements
than required for uniform bending or axial fa-
tigue of the same section size.
Rotating Beam Machines. Typical rotating
beam machine types are shown in Fig. 9. The
R. R. Moore-type machines (Fig. 9a) can operate
up to 10000 rpm. In all bending-type tests, only
the material near the surface is subjected to the
maximum stress; therefore, in a small-diameter
specimen, only a very small volume of material is
under test.
Torsional Fatigue Testing Machines
Torsional fatigue tests can be performed on
axial-type machines using the proper fixtures if
the maximum twist required is small. Specially
8 Fatigue Testing

~Load
(a) (b)

(a) Four-point loading R.R. Moore testing machine. (b) Single-end rotating cantilever testing machine.
Fig. 9 Schematic of rotating beam fatigue testing machines

Program

- - - - - Torque feedback - - ---I

I
I
Angular
L----
I
I
I
display feedback ...-----i----r--...., I
I

Hydraulic
power Specimen

.,.,C:::}:;;:;;;:;;;:;:l:::;~~,.
supply
Displacement
Rota ry actu ato r
Hydraulic transducer
service
manifold

Fig. 10 Schematic of a servohydraulic torsional fatigue testing machine

designed torsional fatigue testing machines con-
sist of electromechanical machines, in which lin-
ear motion is changed to rotational motion by
the use of cranks, and servo hydraulic machines,
in which rotary actuators are incorporated in a
closed-loop testing system (Fig. 10).
Special-Purpose Fatigue
Testing Machines
To perform fatigue testing of components that
are prone to fatigue failure (gears, bearings, wire,
etc.), special devices have been used, sometimes
as modifications to an existing fatigue machine.
Wire testers are a modification of rotating beam
machines, in which a length of the test wire is
used as the beam and is deflected (buckled) a
known amount and rotated.
Rolling contact fatigue testers usually are
constant-load machines in which a Hertzian con-
tact stress between two rotating bearings is ap-
plied until occurrence of fatigue failure by pitting
or spalling is indicated by a vibration or noise
level in the system. Rolling contact fatigue of ball
and roller bearings under controlled lubrication
conditions is a specialized field of fatigue testing.
Multiaxial Fatigue Testing Machines
Many special fatigue testing machines have
been designed to apply two or more modes of
loading, in or out of phase, to specimens to de-
 Fatigue Crack Initiation 9

termine the properties of metals under biaxial or D

( ~=====-t-$
triaxial stresses.

Fatigue Test Specimens

A typical fatigue test specimen has three areas:
4.8 mm (3116 in.) ~R
D, selected on basis of ultimate strength
the test section and the two grip ends. The grip of material R, 12.7 mm (0.50 in.)
ends are designed to transfer load from the test
machine grips to the test section and may be (a)
identical, particularly for axial fatigue tests. The
transition from the grip ends to the test area is
designed with large, smoothly blended radii to
30 mm (13/16 in.)"] k 50 ~m_1
eliminate any stress concentrations in the tran-
sition.
The design and type of specimen used depend
-$-O~2m.) :cD
on the fatigue testing machine used and the ob-
jective of the fatigue study. The test section in the Tapered D, 12.7 mm (0.50 in.)
specimen is reduced in cross section to prevent Ib)
failure in the grip ends and should be propor-
tioned to use the upper ranges of the load capac-
ity ofthe fatigue machine; i.e., avoiding very low I'" 90 mm (3:6 i ~ 19 mm (% in.)
load amplitudes where sensitivity and response l
of the system are decreased. Several types of fa-
tigue test specimens are illustrated in Fig. 11. ~~-®-
R~ ~ '\ I
12 mm (0.48 in.)
Effect of Stress Concentration
D, 5 to 10 mm (0.20 to 0.40 in.) selected on
Fatigue strength is reduced significantly by the basis of ultimate strength of material
introduction of a stress raiser such as a notch or R, 90 to 250 mm (3.5 to 10 in.)
hole. Because actual machine elements invari- [c]
ably contain stress raisers such as fillets, key-
ways, screw threads, press fits, and holes, fatigue
cracks in structural parts usually initiate at such .25 mm (1.0 in.) D 38 mm (1V2 in.)

~---~ ~3-
geometrical irregularities.
An optimum way of minimizing fatigue failure
is the reduction of avoidable stress raisers
through careful design and the prevention of ac-
cidental stress raisers by careful machining and
fabrication. Stress concentration can also arise
-~ ~ . 5"43'

38 mm (1'12 in.)
from surface roughness and metallurgical stress
(d)
raisers such as porosity, inclusions, local over-
heating in grinding, and decarburization.
The effect of stress raisers on fatigue is gener-
ally studied by testing specimens containing a
notch, usually a V-notch or a U-notch. The pres-
ence of a notch in a specimen under uniaxial load
introduces three effects: (1) there is an increase or
concentration of stress at the root of the notch,
D, selected on basis of ultimate strength of material
(2) a stress gradient is set up from the root of the R, 75 to 250 mm (3 to 10 in.)
notch toward the center of the specimen, and (3)
a triaxial state of stress is produced at the notch Ie)
root. (a)Torsional specimen. (bl Rotating cantilever beam spec-
The ratio of the maximum stress in the region imen. (c) Rotating beam specimen. (d) Plate specimen for
of the notch (or other stress concentration) to the cantilever reverse bending. Ie) Axial loading specimen.
corresponding nominal stress is the stress-con- Fig. 11 Typical fatigue test specimens
10 Fatigue Testing
cent ration factor, K,(see the Introduction to this
Section). In some situations, values of K,can be
calculated using the theory of elasticity, or can be
measured using photoelastic plastic models.
The effect of notches on fatigue strength is de-
termined by comparing the S-N curves of notched
and unnotched specimens. The data for notched
specimens usually are plotted in terms of nomi-
nal stress based on the net cross section of the
specimen. The effectiveness of the notch in de-
creasing the fatigue limit is expressed by the
fatigue-notch factor, K p This factor is the ratio
of the fatigue limit of unnotched specimens to the
fatigue limit of notched specimens.
For materials that do not exhibit a fatigue
limit, the fatigue-notch factor is based on the fa-
tigue strength at a specified number of cycles.
Values of KJhave been found to vary with (1)
severity of the notch, (2) type of notch, (3) mate-
rial, (4) type of loading, and (5) stress level.
Effect of Test Specimen Size
It is not possible to predict directly the fatigue
performance oflarge machine members from the
results oflaboratory tests on small specimens. In
most cases, a size effect exists; i.e., the fatigue
strength of large members is lower than that of
small specimens. Precise determination of this
phenomenon is difficult. It is extremely difficult
to prepare geometrically similar specimens of in-
creasing diameter that have the same metallurgi-
cal structure and residual stress distribution
throughout the cross section. The problems in fa-
tigue testing of large specimens are considerable,
and few fatigue machines can accommodate
specimens with a wide range of cross sections.
Changing the size of a fatigue specimen usu-
ally results in variations oftwo factors. First, in-
creasing the diameter increases the volume or
surface area of the specimen. The change in
amount of surface is significant, because fatigue
failures usually initiate at the surface. Secondly,
for plain or notched specimens loaded in bending
or torsion, an increase in diameter usually de-
creases the stress gradient across the diameter
and increases the volume of material that is
highly stressed.
Experimental data on the size effect in fatigue
typically show that the fatigue limit decreases
with increasing specimen diameter. Horger's
data for steel shafts tested in reversed bending
(Table 1) show that the fatigue limit can be ap-
preciably reduced in large section sizes.
Table 1 Effect of specimen size on the fatigue
limit of normalized plain carbon steel in
reversed bending
Specimen diameter Fatigue limit
mm in. MPa ksi
7.6 0.30 248 36
38 l.50 200 29
152 6.00 144 21
Surface Effects and Fatigue
Generally, fatigue properties are very sensitive
to surface conditions. Except in special cases
where internal defects or case hardening is in-
volved, all fatigue cracks initiate at the surface.
Factors that affect the surface of a fatigue speci-
men can be divided into three categories: (1) sur-
face roughness or stress raisers at the surface, (2)
changes in the properties of the surface metal,
and (3) changes in the residual stress condition of
the surface. Additionally, the surface may be
subjected to oxidation and corrosion.
Surface Roughness. In general, fatigue life in-
creases as the magnitude of surface roughness
decreases. Decreasing surface roughness mini-
mizes local stress raisers. Therefore, special at-
tention must be given to the surface preparation
of fatigue test specimens. Typically, a metallo-
graphic finish, free of machining grooves and
grinding scratches, is necessary. Figure 12 illus-
trates the effects that various surface conditions
have on the fatigue properties of steel.
Effect of Mean Stress
A series of fatigue tests can be conducted at
various mean stresses, and the results can be
plotted as a series of S-N curves. A description of
applied stresses and S-N curves can be found in
the Introduction to this Section. For design pur-
poses, it is more useful to know how the mean
stress affects the permissible alternating stress
amplitude for a given life (number of cycles).
This usually is accomplished by plotting the al-
lowable stress amplitude for a specific number of
cycles as a function of the associated mean stress.
At zero mean stress, the allowable stress am-
plitude is the effective fatigue limit for a specified
number of cycles. As the mean stress increases,
the permissible amplitudes steadily decrease. At
a mean stress equal to the ultimate tensile
strength of the material, the permissible ampli-
tude is zero.
The two straight lines and the curve shown in
 Fatigue Crack Initiation 11

1000 ,...---,....----,----r---,-------r-----r------,-----,...-----,

800
900 r~~;;::~~;;;;t;;~==~===~~=~===t===~==~
ro 700 100 'iii
a.. 90 -""
~ 600 ~--_t_---~2.......,_1_--~~=_---="""" .......= _ - _ + - - _ 1 _ - - - - _ t _ - _ _ l vi
~ W ~
~ 500 70 U;
~ ~

~ 60 ~
~ 400 Q)
~ >
> 50 ~
~ >
Qi
~ 300 t-----t-----t---+----+----+--"""""~--~""""""----+--___I 40 Qj
~ ~
~ E
E
o
0
u
u 30
200 I----t----+---+---+-----+---+---+--~.,___f-----~

2 5 10 20 50 100 200 500

Life, 1000 cycles

Fig. 12 Effect of surface conditions on the fatigue properties of steel (302 to 321 HB)

Fig. 13 represent the three most widely used em-

pirical relationships for describing the effect of
mean stress on fatigue strength. The straight line
joining the alternating fatigue strength to the
tensile strength is the modified Goodman law.
Goodman's original law included the assump-
tion that the fatigue limit was equal to one third
of the tensile strength; this has since been gener-
alized to the relation shown in Fig. 13, using the
fatigue strength as determined experimentally.
Stress Amplitude. Because stress amplitude
varies widely under actual loading conditions, it
is necessary to predict fatigue life under various
stress amplitudes. The most widely used method
of estimating fatigue under complex loading is
provided by the linear damage law. This is a hy-
pothesis first suggested by Palmgren and re-
stated by Miner, and is sometimes known as
Miner's rule.
The assumption is made that the application
of n.cycles at a stress amplitude S;, for which the
average number of cycles to failure is N;, causes
an amount of fatigue damage that is measured by
the cumulative cycles ratio n;/N;, and that fail-
ure will occur when "'i.(n;/ N;) = 1.
This method is not applicable in all cases, and
numerous alternative theories of cumulative lin-
ear damage have been suggested. Some consid-
erations of redistribution of stresses have been
clarified, but there is as yet no satisfactory ap-
proach for all situations.
Fatigue strength, S
CI)" / Gerber's parabola
~
vi / Modified Goodman line
~
OJ Tensile
C
strength, Su
Mean stress, Sm
As shown by the modified Goodman line. Gerber's parab-
ola. and Soderberg line. See text for discussion.
Fig. 13 Effect of mean stress on the al-
ternating stress amplitude
The effect of varying the stress amplitude (lin-
ear damage) can be evaluated experimentally by
means of a test in which a given number of stress
cycles are applied to a test piece at one stress am-
plitude. The test is then continued to fracture at a
different amplitude. Alternatively, the stress can
be changed from one stress amplitude to another
at regular intervals; such tests are known as
block, or interval, tests. These tests do not simu-
late service conditions, but may serve a useful
purpose for assessing the linear damage law and
indicating its limitations.
12 Fatigue Testing

Corrosion Fatigue Stress-intensity factor range UK). ksi\. fil.

10 20 50 100

Corrosion fatigue is the combined action of

, q
repeated or fluctuating stress and a corrosive en- I
I
vironment to produce progressive cracking. Us-
ually, environmental effects are deleterious to fa-
tigue life, producing cracks in fewer cycles than
a
~ ~
would be required in a more inert environment. ] 6~
i .9' r;~
Once fatigue cracks have formed, the corrosive
aspect also may accelerate the rate of crack I
~~o
Region 1: I
growth. slow crack I I
I
In corrosion fatigue, the magnitude of cyclic
stress and the number of times it is applied are . I- growth
da
dN
~ C(~Kln

dO
I~ 1
not the only critical loading parameters. Time- i Region 3:
rapid
I
r- -
dependent environmental effects also are of
prime importance. When failure occurs by cor- t---- sx; : W '0
unstable
crack
growlh
I
! r--
rosion fatigue, stress-cycle frequency, stress- 5
I
Ie? I
I
wave shape, and stress ratio all affect the crack-
ing processes.
I if
I
'I Region 2: power-taw behavior
1

I,c
Fatigue Crack 6

1;1
Propagation I,j>
I

Fatigue failure of structural and equipment I

components due to cyclic loading has long been a 8 10 20 30 40 50 60 80 100

major design problem and the subject ofnumer-
ous investigations. Although considerable fa-
tigue data are available; the majority has been
concerned with the nominal stress required to
cause failure in a given number of cycles-
namely, S-N curves. Usually, such data are ob-
tained by testing smooth or notched specimens.
With this type of testing, however, it is difficult to
distinguish between fatigue crack initiation life
and fatigue crack propagation life.
Preexisting flaws or crack-like defects within a
material reduce or may eliminate the crack initia-
tion portion of the fatigue life of the component.
Fracture mechanics methodology enhances the
understanding of the initiation and propagation
of fatigue cracks and assists in solving the prob-
lem of designing to prevent fatigue failures.
Fatigue Crack
Propagation Test Methods
The general nature of fatigue crack propaga-
tion using fracture mechanics techniques is
summarized in Fig. 14. A logarithmic plot of the
crack growth per cycle, da/ dN, versus the stress-
intensity factor range, I:!..K, corresponding to the
load cycle applied to a specimen is illustrated.
The da/ dN versus I:!..K plot was constructed of
Suess-tntensnv teeter range UK), MPa \ m
Yield strength of 470 MPa (70 ksi). Test conditions: R =
0.10; ambient room air, 24°C (75 OF).
Fig. 14 Fatigue crack growth behavior
of ASTM A533 B1 steel
data on five specimens of ASTM A533 HI steel
tested at 24 0 C (75 0 F). A plot of similar shape is
anticipated with most structural alloys; the abso-
lute values of da/dNand I:!..K, however, are de-
pendent on the material.
Results of fatigue crack growth rate tests for
nearly all metallic structural materials have
shown that the da/ dN versus I:!..K curves have
three distinct regions. The behavior in Region I
(Fig. 14) exhibits a fatigue crack growth thresh-
old, I:!..K"" which corresponds to the stress-
intensity factor range below which cracks do not
propagate.
At intermediate values of I:!..K (Region II in
Fig. 14), a straight line usually is obtained on a
log-log plot of I:!..K versus da/ dN. This is de-
scribed by the power-law relationship:
da
dN = C(I:!..K)"
where C and n are constants for a given material
and stress ratio.
 Fatigue Crack Propagation 13

Fatigue crack growth rate data for some steels Stress-intensity factor range (.:lK), ksiV'Ti1."
show that the primary parameter affecting
10 102
growth rate in Region II is the stress-intensity
factor range and that the mechanical and metal-
lurgical properties of these steels have negligible 10- 2
o 12 Ni steel
effects on the fatigue crack growth rate in a
o 10 Ni steel
room-temperature air environment. Data for
four martensitic steels fall within a single band, Ql • HY-80 steel Ql
"0
as shown in Fig: 15. The upper bound of scatter > • HY-130 steel 10- 4 "0
>
~ ~
can be obtained from: E c::
E 10- 3
da ......: <:
dN = 0.66 X 1O-8(~K)2.25 <: ~
~

where a is given in inches, and ~K is given in
ksiyTr;.
For some steels, the stress ratio and mean
stress have negligible effects on the rate of crack
growth in Region II. Also, the frequency of cyclic
loading and the waveform (sinusoidal, triangu-
lar, square, trapezoidal) do not affect the rate of
crack propagation per cycle of load for some
steels in benign environments.
At high ~Kvalues (Region III in Fig. 14), un-
stable behavior occurs, resulting in a rapid in-
crease in the crack growth rate just prior to com-
plete failure of the specimens. There are two
possible causes of this behavior. First, the in-
creasing crack length during constant load test-
ing causes the peak stress intensity to reach the
fracture toughness, K'n of the material, and the
unstable behavior is related to the early stages of
brittle fracture. Second, the growing crack re-
duces the uncracked area of the specimen suffi-
ciently for the peak load to cause fully plastic
limit load behavior. The first possibility is opera-
tive for high-strength, low-toughness metals, in
which specimen sizes normally used for fatigue
crack growth rate testing behave in a linear elas-
tic manner at K levels equal to K/c. The second
possibility, plastic limit load behavior, is com-
mon for ductile metals, particularly if K/cis high.
When plastic limit load behavior causes un-
stable crack growth, ~K values have no mean-
ing, because the limitations of linear elastic frac-
ture mechanics have been exceeded. Here, the
use of the J-integral concept, crack-opening dis-
placement, or some other elastic-plastic fracture
mechanics approach is more appropriate than
~K for correlating the data.
Standardized testing procedures for measur-
ing fatigue crack growth rates are described in
ASTM Standard E 647. This method applies to
medium to high crack growth rates-that is,
above 10-8 tu] cycle (3.9 X 10-7 ui.] cycle). Proce-
dures for growth rates below 10-8 in] cycle are
~ ~
10- 5 i
2:! ~
~
s:
s: 10- 4
~ ~
e e
Ol
Ol
-"
-"
0
0
co
~ 10- 6 L.
U
U
10- 5 t--t-l~tf--+--+--+----I
2 x 10- 6 '---_..I..-_..L-_..I..-_....L_--L--'
2 5 10 20 50 100 200
Stress-intensity factor range (.:lKl. MPaVm
Fig. 15 Summary of fatigue crack
growth data for martensitic steels
under consideration by ASTM. For applications
involving fatigue lives of up to about 106 load
cycles, the procedures recommended in ASTM E
647 can be used. Fatigue lives greater than about
106 cycles correspond to growth rates below 10-8
in] cycle, and these require special testing proce-
dures, which are related to the threshold of fa-
tigue crack growth illustrated in Fig. 14.
ASTM E 647 describes the use of center-
cracked specimens and compact specimens (Fig.
16 and 17). The specimen thickness-to-width ra-
tio, B/W, is smaller than the 0.5 value for K/c
tests; the maximum B/W values for center-
cracked and compact specimens are 0.125 and
0.25, respectively. With the thinner specimens,
crack length measurements on the sides of the
specimens can be used as representations of
through-thickness crack growth behavior.
For tension-tension fatigue loading, the K;
loading fixtures frequently can be used. For this
type of loading, both the maximum and min-
imum loads are tensile, and the load ratio, R =
P min/Pmax' is in the range 0 < R < 1. A ratio of
14
Two holes W/3 diam
~----+-----J_l
!I ! !I
I I
2a n i.sthe machined notch; a is the crack length; B is the
specimen thickness.
Fig. 16 Standard center-cracked ten-
sion specimen for fatigue crack propaga-
tion testing when the width (WI of the
specimen ';;;;75 mm (3 in.)
R= 0.1 is commonly used for developing data for
comparative purposes.
Testing often is performed in laboratory air at
room temperature; however, any gaseous or liq-
uid environment and temperature of interest
may be used to determine the effect of tempera-
ture, corrosion, or other chemical reaction on
cyclic loading.
Data Analysis. For constant-amplitude load-
ing, a set of crack-length versus elapsed-cycle
data (a versus N) is generated, with the specimen
loading, Pmax and P min' generally held constant.
Figure 18illustrates a typical a versus N plot. The
minimum crack-length interval, 6.a, between
data points (see Fig. 18)should be 0.25 mm (0.01
in.) or ten times the crack-length measurement
precision, which is defined as the standard devia-
tion on the mean value of crack length deter-
mined for a set of replicate measurements. This
prevents the measurement of erroneous growth
rates from a group of data points that are spaced
too closely relative to the precision of data mea-
surement and relative to the scatter of data.
Crack measurement intervals are recom-
mended in ASTM E 647 according to specimen
type. For compact-type specimens:
6.a ~ 0.04 Wfor 0.25 ~!!.- ~ 0.40
W the frequently observed increase in da/ dN asso-
Fatigue Testing
6.a ~ 0.02 W for 0.40 ~ !!.- ~ 0.60
W
-r 6.a~
a
0.01 Wfor -;;::: 0.60
W
For center-cracked tension specimens:
W
6.a ~ 0.03 Wfor 2a < 0.60
W
6.a ~ 0.02 W for ~ > 0.60
W
Fatigue crack growth rate data can be calcu-
lated by several methods. The most commonly
used methods, however, are the secant and in-
cremental polynomial methods. The secant meth-
od consists of the slope of the straight line con-
necting two adjacent data points. This method,
although simpler, results in more scatter in mea-
sured crack growth rate.
The incremental polynomial method fits a
secon~-order ,Polynomial expression (parabola)
to typically five to seven adjacent data points,
and the slope of this expression is the growth
rate. The incremental polynomial method elimi-
nates some of the scatter in growth rate that is
inherent in fatigue testing.
Numerous relationships have been generated
to correlate crack growth rate and stress-intensity
data. The most widely accepted relationship is
that proposed by Paris. This is a linear relation-
ship when plotted on log-log coordinates and
generally yields a reasonable fit to the data in
Region II (see Fig. 14) of the crack growth
regime.
Other relationships based on the Paris equa-
tion, such as the commonly used Forman equa-
tl?n, are used to represent the variation of da/ dN
with other key variables, including load ratio, R
and the critical K value, K" at which rapid frac~
ture of the specimen occurs (Region III in Fig.
14). The Forman equation is:
da = C(6.K)"
dN (1 - R)(K,. - 6.K)
where Cand n are material constants of the same
types as those in the Paris equation, but of differ-
ent values. An advantage of the Forman equa-
tion is that it describes the type of accelerated
da/dNbehavior that is often observed at high
values of 6.K, which is not described by the Paris
equation.
Additionally, the Forman equation describes
 Fatigue Crack Propagation 15

Two holes
0.25Wdiam

r
t
0.6W

0.275W

t 1
0.275W
t
t r
0.6W

~~_1_
1 + - - - a ----;~
Allowable thickness: W/20 s B s W/4
1( - + - - - - - - w-------;~I
)
Minimum dimensions: W = 25 mm (1.0 in.)
an = 0.20W
~------1.25W-------:;~I

Fig. 17 Standard compact-type specimen for fatigue crack propagation testing (see Fig.
16 for explanation of symbols)

ciated with an increase in R from 0 toward I. 55 2.2

When it is necessary to describe the effect of K (
approaching Kc> or the effect of R on da / dN, the 50 2.0
Forman equation can be used to represent the
da/dNbehavior. When only ~Kin Region II is E
E 45
~ 1.8 C
involved, the less complex Paris equation may be
used.
.r= 40 I 1.6
s:
<iI

OJ i~' 0,
Cyclic Crack Growth Rate ",,0
1.4 Q)c
Testing in the Threshold Regime ~ 35
.><
(J 'rP0a- J .><
(J

Cyclic crack growth rate testing in the low- ~ 30 1.2 ...

ro
u 00
U
growth regime (Region I in Fig. 14) complicates Oo c
25 n
acquisition of valid and consistent data, because v 1.0
the crack growth behavior becomes more sensi-
tive to the material, environment, and testing 20
procedures under this regime. Within this re- 125 150 175 200 225 250 275 300
3
gime, the fatigue mechanisms of the material that Cycles. 10 N
slow the crack growth rates are more significant. Fig. 18 Crack growth versus constant-
The precise definition of the cyclic crack amplitude stress cycles for a Fe-10Ni-
growth rate threshold, ~K'h' varies significantly. 8Co-1 Mo high-strength steel
16 Fatigue Testing
The most accurate definition would be the stress-
intensity value below which fatigue crack growth
will not occur. It is extremely expensive to obtain
a true definition of IJ.K,h, and in some materials a
true threshold may be nonexistent. Generally,
designers are more interested in the near-thresh-
old regime, such as the IJ.Kthat corres~onds to a
fatigue crack growth rate of 10-8 to 10- 0 tn] cycle
(3.9 X 10-7 to 10-9 in.j cycle). Because the dura-
tion of the tests increases greatly for each addi-
tional decade of near-threshold data (10- 8 to 10-9
to 10-10, etc., m/cycle), the precise design re-
quirements should be determined in advance of
the test.
Behavior of Short Cracks
Recently, it has been well documented that
short cracks may behave differently from large
cracks when plotted in the standard form of cy-
clic crack growth rate versus stress intensity.
A short crack is difficult to define. It may be
small compared to the microstructure of the
material to be studied (I to 50 /.Lm) when the con-
cepts of continuum mechanics are of interest. It
can also be small compared to the plastic zone
size (10 to 1000 /.Lm). In this situation, linear elas-
tic fracture mechanics might be replaced with
elastic-plastic fracture mechanics. The crack
may also be physically small (500 to 1000 /.Lm)
when crack closure, crack tip shape, environ-
ment, and growth mechanisms are of concern.
Figure 19 schematically illustrates the possible
behavior of short cracks.
Selection of Test Specimens
Selection of a fatigue crack growth test speci-
men is usually based on the availability of the
material and the types of test systems and crack-
monitoring devices to be used. The two most
widely used types of specimens are the center-
cracked tension specimen and the compact-type
specimen (see Fig. 16 and 17). However, any
specimen configuration with a known stress-
intensity factor solution can be used in fatigue
crack growth testing, assuming that the appro-
priate equipment is available for controlling the
test and measuring the crack dimensions. Stress-
intensity factor solutions for center-cracked ten-
sion and compact-type specimens are given in
Table 2.
Consideration of the range of application of
the stress-intensity solution of a specimen con-
figuration is very important. Many stress-inten-
Short cracks
that join long
q/;/"
crack behavior
// / \
/ Short cracks that behave
\', / as long cracks
\ ./
\---
\ Short cracks that become
\ non propagating cracks
Stress-intensity factor range
Fig. 19 Typical short crack behavior
sity expressions are valid only over a range of the
ratio of crack length to specimen width (a/W).
For example, the expression given in Table 2 for
the compact-type specimen is valid for a/ W >
0.2; the expression for the center-cracked tension
specimen is valid for 2a/W< 0.95. The use of
stress-intensity expressions outside their appli-
cable crack-length region can produce signifi-
cant errors in data.
The size of the specimen must also be appro-
priate. To follow the rules of linear elastic frac-
ture mechanics, the specimen must be predomi-
nantly elastic. However, unlike the requirements
for plane-strain fracture toughness testing, the
stresses at the crack tip do not have to be main-
tained in a plane-strain state. The stress state is
considered to be a controlled test variable. The
material characteristics, specimen size, crack
length, and applied load will dictate whether the
specimen is predominantly elastic. Because the
loading mode of different specimens varies sig-
nificantly, each specimen geometry must be con-
sidered separately.
Notch Preparation. The method by which a
notch is machined depends on the specimen
 Fatigue Crack Propagation 17

Table 2 Stress-intensity factor solutions for standardized (ASTM E 647)

fatigue crack growth specimen geometries
Center-cracked tension specimens (Fig. 16)
sr ~
sec
tra
I:!.K= Ii V 2W 2
where a = - . va1'"
2a; expression 2a 0.95
id 10r-<
W W
Compact-type specimens (Fig. 17)
_
I:!.K -
I:!.P(2
ru;
+ a)
3/2 (0.886 + 4.64a - 13.32a
2
+ 14.Tl« 3 - 4
5.6a )
By W(l-a)

. va liId f or-»
where a = -a ; expression a 0.2
W W

material and the desired notch root radius (p).
Sawcutting is the easiest method, but is generally
acceptable only for aluminum alloys. For a
notch root radius of p ~ 0.25 mm (0.010 in.) in
aluminum alloys, milling or broaching is re-
quired. A similar notch root radius in low- and
medium-strength steels can be produced by
grinding. For high-strength steel alloys, nickel-
base superalloys, and titanium alloys, electrical
discharge machining may be necessary to pro-
duce a notch root radius of p ~ 0.25 mm (0.010
in.).
Precracking of a specimen prior to testing is
conducted at stress intensities sufficient to cause
a crack to initiate from the starter notch and
propagate to a length that will eliminate the ef-
fect of the notch. To decrease the amount of time
needed for precracking to occur, common prac-
tice is to initiate the pre cracking at a load above
that which will be used during testing and to sub-
sequently reduce the load.
Load generally is reduced uniformly to avoid
transient effects. Crack growth can be arrested
above the threshold stress-intensity value due to
formation of the increased plastic zone ahead of
the tip of the advancing crack. Therefore, the
step size of the load during precracking should
be minimized. Reduction in the maximum load
should not be greater than 20% of the previous
load condition. As the crack approaches the final
desired size, this percentage may be decreased.
Gripping of the specimen must be done in a
manner that does not violate the stress-intensity
solution requirements. For example, in a single-
edge notched specimen, it is possible to produce
a grip that permits rotation in the loading of the
specimen, or it is possible to produce a rigid grip.
Each of these requires a different stress-intensity
solution. In grips that are permitted to rotate,
such as the compact-type specimen grip, the pin
and the hole clearances must be designed to min-
imize friction. It is also advisable to consider lat-
eral movement above and below the grips.
Gripping arrangements for compact-type and
center-cracked tension specimens are described
in ASTM E 647. For a center-cracked tension
specimen less than 75 mm (3 in.) in width, a single
pin grip is generally suitable. Wider specimens
generally require additional pins, friction grip-
ping, or some other method to provide sufficient
strength in the specimen and grip to prohibit
failure at undesirable locations, such as in the
grips. Grips designed for compact-type speci-
mens are illustrated in Fig. 20.
Crack-Length
Measurement Techniques
Several different techniques have been devel-
oped to monitor the initiation, growth, and in-
stability of cracks, including optical (visual and
photographic), electrical (eddy current and re-
sistance), compliance, ultrasonic, and acoustic
emission monitoring techniques.
Optical Crack Measurement Techniques
Monitoring of fatigue crack length as a func-
tion of cycles is most commonly conducted visu-
ally by observing the crack at the specimen sur-
faces with a traveling low-power microscope at a
magnification of 20 to 50X. Crack-length mea-
surements are made at intervals such that a
nearly even distribution of da] dN versus ~K is
achieved. The minimum amount of extension be-
18 Fatigue Testing
Fig. 20 Grips designed for fatigue crack
propagation testing of compact-type
specimens (courtesy of MTS Systems
Corp.)
tween readings is commonly about 0.25 mm
(0.10 in.).
The optical technique is straightforward and,
if the specimen is carefully polished and does not
oxidize during the test, produces accurate re-
sults. However, the process is time consuming,
subjective, and can be automated only with
complicated and expensive video-digitizing
equipment. In addition, many fatigue crack
growth rate tests are conducted in simulated-
service environments that obscure direct obser-
vation of the crack.
Compliance Method of
Crack Extension Measurement
The compliance of an elastically strained spec-
imen containing a crack of length a measured
from the load line to the crack tip is usually ex-
pressed as the quotient of the displacement, 0,
and the tensile load, P, with the displacement
measured along, or parallel to, the load line. Fig-
(a)
~--al~
p
(b) p
(a) C(aD) = 0D/P, (b) C(a1) = 01 /P
Fig. 21 Schematic of the relationship
between compliance and crack length
ure 21 illustrates that the more deeply a specimen
is cracked, the greater the amount of 0 measured
for a specific value of tensile load. Compliance
can also be defined for shear and torsional loads
applied to cracked specimens, and crack exten-
sion under these loading modes can be similarly
determined.
Specimen load is simultaneously measured by
an electronic load cell and conditioner / amplifier
system, and the output is directed to the same
data-acquisition system. A generalized schematic
of the circuits involved is shown in Fig. 22.
 Fatigue Crack Propagation
x-v recorder
Load
t a, a 1 a2 a Gage
/P'~
condition
I
± 10 V dc
± 10 V dc
Fig. 22 Components of a compliance measurement system
The required sensitivity of the systems de-
pends on specimen geometry and size; in general,
noise-free, amplified output on the order of I V
dc per I mm (0.04 in.) of deflection is satisfac-
tory. Similarly, for the load range applied to the
specimen, an approximately I V de change in
signal from the load cell is required for accurate
calculation of the compliance.
Electric Potential
Crack Monitoring Technique
The electrical potential, or potential drop,
technique has gained increasingly wide accep-
tance in fracture research as one of the most ac-
curate and efficient methods for monitoring the
initiation and propagation of cracks. This meth-
od relies on the fact that there will be a distur-
bance in the electrical potential field about any
discontinuity in a current-carrying body, the
magnitude of the disturbance depending directly
on the size and shape of the discontinuity.
For the application of crack growth monitor-
ing, the electric potential method entails passing
a constant current (maintained constant by ex-
ternal means) through a cracked test specimen
and measuring the change in electrical potential
across the crack as it propagates. With increasing
crack length, the uncracked cross-sectional area
of the test piece decreases, its electrical resistance
increases, and thus the potential difference be-
tween two points spanning the crack rises. By
monitoring this potential increase, Va' and com-
paring it with some reference potential, Vo , the
19
Specimen
Displacement
gage -+----... )
....
'Lr
J=1
Load
cell
Load cell
condition
crack length to width ratio, a/ W, can be deter-
mined through the use of the relevant calibration
curve for the particular test piece geometry
concerned,
Crack Growth Studies. By far the most useful
application of the electrical potential method has
been in measurements of crack length during
crack propagation, where it has been utilized to
monitor almost all mechanisms of subcritical
crack growth and most notably to follow fatigue
crack growth. Typical crack propagation rates
derived from direct current potential measure-
ments are shown in Fig. 23 for tests on a 2.25Cr-
IMo steel in air, gaseous hydrogen, and hydro-
gen sulfide environments.
Electromechanical
Fatigue Testing Systems
The primary function of electromechanical fa-
tigue testers is to apply millions of cycles to a test
piece at oscillating loads up to 220kN (50000 lbf)
to investigate fatigue life, or the number of cycles
to failure under controlled cyclic loading condi-
tions. Variables associated with fatigue-life tests
are frequency of loading and unloading, ampli-
tude of loading (maximum loads and minimum
loads), and control capabilities. The fundamen-
tal data output requirement is the number of cy-
cles to failure, as defined by the application,
A variety of electromechanical fatigue testers
have been developed for different applications.
Forced-displacement, forced-vibration, rota-
20 Fatigue Testing

Table 3 Comparison of electromechanical fatigue systems

Parameter Forced displacement Forced vibration

Tension . Yes Yes

Compression . Yes Yes
Reverse stress . Yes Yes
Bending . Yes Yes
Frequency range . Fixed Fixed, 1800 rpm
Load range . Typically < 450 N Up to 220 kN
« 1001bf) (500001bf)
Type:
Control . Open-loop Open-loop
Mode . Displacement Load

Maximum deflection . 25.4 mm (l.00 in.

Advantages Simple, straight- Versatile, efficient,
forward durable
Disadvantages No load control, very Fixed frequency, Iii
limited applications ited control (open
(soft samples) loop)

Alternating stress intensity (:>K), ksi vln,

678910 20 30 40 50 60 10 80
E I I I
Environment
I I I 'I II ,
R
r I

Frequency. Hz
I

~#
~I

rPJ pO
I I

.
.. ..
- 10" ,
•
0
Moist air
Dry hydrogen
0.05
0.05
50
50
/
I"'"
nO
od
~ .
• Air 0.1 5 '" !II ~ U
>
0 Dry hydrogen 0.1 5 # • fill' 10-. u
Hydrogen sulfide 0.1 ~ .~
•'"
5 00 '"

l> Molstair 0.75

i-
t4,f- ~
Dry hydrogen 0.75
- 10'
.
1)

~ , I oOd9

- 10- 1
..
~

~
~.'l>~ ~ s:
~
ern
~ V~
'U

rP One lattice -..

I~rt -- 10-.
..
~
spacing ~

I per cycle
u

U
:: lJ
- •
t 0
t
0

6 7
Threshold :>K",

8 9 10 20 30 40 50 60 10 80 90
10

Alternating stress intensity (:>K), MPa\m

Data derived from direct current potential measurements in martensitic 2.25Cr-1 Mo steel (SA542-C12) at R =0.05
to 0.75 in air, hydrogen. and hydrogen sulfide at ambient temperature.
Fig. 23 Fatigue crack propagation data over a wide spectrum of growth rates

tional bending, resonance, and servomechanical
systems are discussed in this article and are com-
pared in Table 3. Other specialized electrome-
chanical systems are available to perform spe-
cific tasks.
Forced-Displacement Systems
Forced-displacement motor-driven systems
are the simplest type of electromechanical fa-
tigue testers. They effectively reproduce service
environments that impart fixed, reciprocating
displacements to a component or test piece. An
electric motor-driven flywheel is used to carry a
loading arm at a variable distance from the cen-
ter of rotation, much in the same manner as a
connecting rod in an automotive engine. This ro-
tational displacement is transformed into a
guided, vertical displacement and is used to fa-
tigue the specimen.
Although load can be monitored in such sys-
 Fatigue Crack Propagation 21

Rotational bending Resonance Servomechanical

No Yes Yes
No Yes Yes
Yes Yes Yes
Yes Yes Yes
0-10000 rpm 40-300 Hz 0-1 Hz
Up to 180 kN Up to 90 kN
(400001bf) (200001bf)

Open-loop Closed-loop Closed-loop

Rotation/ bending Load Load, displacement,
strain
l.0 mm (0.040 in.) 100 mm (4 in.)
fficient, durable, Fully closed-loop, Fully closed-loop,
simple extremely efficient high precision
.otational bending Operating frequency Low frequency only
only, limited appli- directly proportional
cations to sample stiffness

tems, the fixed displacement precludes the ability
to control load, which is a function of specimen
characteristics. Therefore, the load generally
drops as failure progresses. These systems typi-
cally are custom-built, inexpensive fatigue ma-
chines, used primarily for bend tests on soft sam-
ples in which load control, high frequencies, and
large loads are not required.
Forced-Vibration Systems
Forced-vibration motor-driven systems were
the first production fatigue testers in commercial
use. The centrifugal forces of an imbalanced ro-
tor is used to impart a cyclic load to the test piece.
In operation, an electric motor is used to ro-
tate an eccentric mass via flexible couplings. The
rotating mass is mounted in a frame that is
guided by flexure plates to restrict movement to
vertical motion only. The centrifugal force pro-
duced by the rotating eccentric mass (m) is
transmitted through the vertically guided frame
to the test piece. The horizontal component of
the centrifugal force is absorbed by the restrain-
ing flexure plates.
Because the centrifugal force usually is totally
absorbed by the mounting frame (of mass M),
the inertial reaction is separated from the cen-
trifugal force in such a way as to transmit only
the centrifugal forces to the specimen. This tech-
nique involves the use of frame-support compen-
sator springs; the natural frequency of the spring
(K)/mass (M) system is tuned to the revolutions
per minute of the motor. Thus, neither the spec-
imen nor the rotating eccentric mass (m) "sees"
an inertial reaction from the frame, because the
inertial effects of the frame are totally compen-
sated for by the frame support springs (not the
specimen).
This technique has two requirements: the ro-
tating frequency (w) must be kept constant and
the mass of mounting frame (M) must be kept
constant. Consequently, the loading frequency
of the device is fixed at 1800 rpm, and masses
must be added or removed from the frame to
compensate for fixturing to keep M constant.
The magnitude of the dynamic load is deter-
mined by placing the rotating mass at a known
distance from the axis ofrotation (r). Because w,
m, M, and K are known, the force on the speci-
men, F, is calibrated directly as a function of r as
follows:
F = Mw 2r (centrifugal) - Ma, (inertial)
+ Kz (spring compensated)
where a, is the acceleration of the frame in the z
direction, and Kz is the spring-compensated dis-
placement in the z direction. Because Ma, is
tuned to equal Kz, F = Mw 2r.
Thus, the forced-vibration rotating eccentric
mass system is an open-loop, load-controlled
system with the ability to accommodate up to 25
mm (1.0 in.) of total sample deflection at loads
up to 220 kN (50000 lbf) using special fixtures.
The mean or static load, onto which the dynamic
load is superimposed, is achieved by preloading
the inertia compensator spring, K.
Through special fixturing, forced-vibration
devices are capable of testing in tension, com-
22 Fatigue Testing
pression, bending, torsion, or reverse st~ess.
Although servo-controlled, mean-load-mainte-
nance systems are available, the open-loop na-
ture of the system prevents direct load measure-
ment or control, which is characteristic of
closed-loop systems. The load applied to the
specimen is assumed to be a function of r, and a
graduated scale is provided to permit reasonably
accurate setup.
Rotational Bending Systems
Rotational bending systems effectively apply
reversed loading to the outer surface of rods or
shafts. The basic operating principle ofthe rotat-
ing beam consists of the use of a motor to rotate a
shaft of known dimensions around its longitudi-
nal axis. By applying a known static force at the
end of the shaft, a bending moment can be ap-
plied to the test section, the outer surface of
which oscillates between tension and compres-
sion during each rotation.
The cantilevered specimen, however, is sub-
jected to a nonuniform bending moment, which
is large at the supported end of the specimen ~nd
zero at the free end. To produce a more meanmg-
ful uniform bending moment throughout the
tes~ piece, a specially designed tapered specimen
should be used or bending moments should be
applied to each end of the specimen. Figure 24
illustrates the rotating-beam operating mecha-
nism and the resulting stress distribution in the
specimen.
Gage
area
Drive motor
I Tension
Hotation 0~:f: '----.. )
A
Bending
mom,",
Compression
Fig.24 Schematic ofthe rotating-beam
operating mechanism and the resulting
stress distribution in the specimen
Resonance Systems
A high-speed fatigue testing system was devel-
oped by Amsler that operated at 40 to 300 Hz,
achieved high loads (up to 90 kN, or 20000 lbf),
and consumed minimal energy. It is based on a
resonant spring/mass system, in which the spec-
imen is used, like a spring, as an integral part of
the oscillating mechanism.
The fatigue load, in the form of a sine wave, is
achieved by preloading the sample in the frame
via a complex optomechanical procedure and
dynamically loading the sample at the natural
oscillating frequency of the spring/ mass system.
The preload is maintained automatically during
the test. The dynamic load is achieved by pulsing
an electromagnet at the natural frequency of the
spring/ mass system. During resonance, the elec-
tromagnet restores any hysteresis energy lost
during the previous cycle, thereby maintaining a
constant, controllable dynamic load. Capable of
tension, compression, bending, torsional, and
reverse-stress fatigue tests, the Amsler resonant
fatigue testers were instrumental in obtaining the
vast amount of fatigue data currently available.
The resonant system is based on a similar prin-
ciple, but incorporates solid-state technology to
achieve fully closed-loop control of mean and
dynamic loads. This system uses dual opposing
masses (unlike the single oscillating mass/ seismic
base of earlier systems), linked by the specimen
to achieve vibration-free resonance. A strain
gage load cell, in series with the specimen, senses
the load and automatically triggers the electro-
magnet to achieve self-tuning capability.
The mean load is achieved by physically mov-
ing the upper mass up or down to achieve tension
or compression, respectively; the dynamic load is
achieved by varying the width of the pulse to the
magnet beneath the lower mass. The dynamic
load, like the mean load, is electronically ma~n­
tained at a preset command level through solid-
state closed-loop circuitry. The remainder of the
controls and mechanisms associated with the res-
onant fatigue system maintain a preset air gap
between the magnet and the oscillating lower
mass, maintain preset loading conditions (S~1Ut­
ting down at preset load levels or frequencies),
and power the electromagnet.
The high efficiency of resonant systems makes
them well suited to high-cycle fatigue tests, in
which closed-loop load control, high loads (up to
180 kN, or 40000 lbf), low power consumption
(around 750 W maximum for closed-loop sys-
 Fatigue Crack Propagation
Electronic
demand
signal
(a) Typical components. (b) Transfer functions. See text for details and explanation of symbols.
Fig. 25 Simplified block diagram for a negative-feedback closed-loop testing machine
tems), and high throughput are required. These
systems tolerate minimal hysteresis and produce
optimum testing results when used with stiff me-
tallic samples.
Closed-Loop Servomechanical Systems
The most recent development in electrome-
chanical fatigue testers is based on an electric ac-
tuator/load frame assembly. The system closely
resembles its servohydraulic counterpart in that
it consists of an actuator, a load frame, a load
cell, a power supply, and a solid-state closed-
loop electronic control console. Closed-loop sys-
tems compare live feedback signals to an input
command signal to maintain accurate control of
preset conditions. The closed-loop servomechan-
ical system is, by virtue of its design, primarily
intended for low-cycle and creep-fatigue studies.
Servohydraulic Fatigue
Testing Systems
Servohydraulic testing machines are particu-
larly well suited for providing the control capa-
bilities required for fatigue testing. Extreme de-
mands for sensitivity, resolution, stability, and
reliability are imposed by fatigue evaluations.
Displacements may have to be controlled (often
for many days) to within a few microns, and
forces can range from 100 kN to just a few new-
tons. This wide range of performance can be ob-
tained with servomechanisms in general and, in
particular, with the modular concept of servohy-
draulic systems.
23
1-----,---- c
(bl
Usually, the problem of selecting the appro-
priate system is simply a matter of optimizing the
various components to form a system best suited
to the given testing application. In this section,
the principles underlying closed-loop servo sys-
tems are discussed briefly. In addition, the inter-
action between system components is illustrated,
and a brief description of their operating princi-
ples and characteristics is provided.
With any type of control system, the objective
is to obtain an output that relates as closely as
possible to the programmed input. In a fatigue
testing system, it may be desired to vary the force
on a specimen in a sinusoidal manner, at a fre-
quency of 1 Hz over a force range of 0 to 100 kN
(0 to 22000 lbf). The only practical means to ac-
complish this with precision is through the use of
a negative-feedback closed-loop system. An over-
view of the basic principles of operation of
negative-feedback systems is provided in Fig. 25.
The blocks shown in Fig. 25(a) represent a group
oftypical components of a testing machine. The
transfer functions of each of these blocks can be
combined to produce the more simplified dia-
gram shown in Fig. 25(b).
Placement of the switch, S I' has been added to
the diagram to permit analysis of the system
when it is open (no feedback, or an open-loop
condition) and when it is closed (providing feed-
back to the system). The equation governing this
simplified open-loop system is:
C= KoD
where Crepresents the controlled output, K; rep-
24 Fatigue Testing

resents the open-loop transfer function, and D • Auxiliary functions such as recorder signal
represents the electronic demand signal. There- conditioning, calibration, and system start-
fore, the output is simply proportional to the sys- up and shutdown
tem demand if K is a constant. Unfortunately,
The servo-valve controls the volume and di-
K is seldom a constant, because it can be influ-
rection of flow of hydraulic fluid between the hy-
en"ced by several common system variations. The
draulic power supply and the hydrauI~c ram.
electronic components may drift slightly, or the~r
Within the control loop, it is the intermediary be-
gain may vary. The behavior of the hydraulic
tween the low-power servo-controller and the
components may change with tempera~ure, con-
hydraulic ram, which can supply large force~ a~d
tamination, or wear, and the mechamcal com-
displacements to the specimen. Characte~ISt1CS
ponents may vary because of thermal effects or
of the device are such that the output flow IS ap-
friction.
proximately proportional to the input current
Servohydraulic System Components when the output pressure is constant. Also, the
output pressure is approximately proportlOn~1
Many commercially manufactured units are
to the square of the input current when the flow IS
available for each component in a typical servo-
constant.
hydraulic testing system. .
Hydraulic rams, or actuators or cylinders,
The programmer supplies the command signal
furnish the forces and displacements required by
to the system, which is generally an analog of the
the testing system. These rams usually are double
desired behavior of the controlled parameter.
ended to provide the greatest lateral rigidity and
For example, assume the same test conditions as
to produce the balanced flow and f~rce charac-
previously discussed (control the force on the
teristics desirable for push-pull testmg. The ef-
specimen in a sinusoidal manner at a frequency
fective area of the piston is therefore equal to the
of I Hz and a force range of 0 to 100 kN). In this
cross-sectional area of the piston minus the
instance, the programmer might be set to pro-
cross-sectional area of the piston rod. Under
duce an electronic signal with a sinusoidal wave-
static conditions (very little flow), the maximum
form that has a frequency of I Hz and a voltage
force capability of the ram will approach the hy-
output of 0 to 10 V. The analog is: 1 V represents
draulic supply pressures multiplied by the effec-
1000 N. The system can then be adjusted to pro-
tive area.
duce the correct output. Any change in the pro-
The force available during dynamic operation
grammer signal will result in a corresponding
depends on the pressure drop and flow character-
change in the controlled parameter.
istics of the servo-valve. Reference should be
The servo-controller makes most of the ad-
made to the load/flow/pressure characteristics
justments necessary to optimize system perfor-
supplied by the servo-valve manufacturer.
mance. For example, it compares the command
Load Cells. The strain gage load cell is the
signal with a signal produced by the controlled
most widely used force-measuring and feedback
parameter (stress or strain, for example) and re-
device in closed-loop fatigue machines. An ex-
lays a correction signal, if needed, to the control
ternal applied force causes the elastic deforma-
device in the system (usually a flow-control
tion of an internal member to which a strain gage
servo-valve). A servo-controller incorporates
bridge has been attached. An electronic sign~l
numerous other compensatory features, such as:
that is proportional to the resistance change m
• Means to adjust the gain or proportional the bridge and to the applied force can thus .b.e
band of the system produced. Some load cells are designed specifi-
• Controls to modify the feedback or correc- cally for fatigue evaluations. Variable features
tion signals for improved stability include sensitivity, natural resonant frequency,
• Controls to adjust the mean level and am- temperature stability, fatigue rating, linearity,
plitude of the command signal(s) hysteresis, deflection constant, load capacity,
• Controls to enhance and adjust servo-valve overload rating, resistance to extraneous load-
response . ing, and compatibility with t?e testing machine
• Means to monitor the system error signal (a and fixtures. Most commercially available cells
measure of how well the command and are very competitive with respect to these
feedback signals agree) features.
• Capability to select various command and Load Frames. In a fatigue machine, the reac-
feedback signals tion forces to the specimen and to the housing of
 Fatigue Crack Propagation
the ram are supplied by the load frame. Many
styles ofload frames are available, but for fatigue
purposes the frames should be customized. The
requirements of good high-frequency response
demand that there be high axial stiffness in the
load frame. When a deflection occurs in the load
frame, additional flow is required from the
servo-valve. Therefore, this deflection should be
minimal in comparison with the deflection im-
parted to the specimen.
In addition, because fatigue specimens must
be subjected to fully reversed loading (i.e., com-
pressive as well as tensile forces), lateral rigidity
25
must be increased to resist bending. This is gen-
erally considered necessary in the design of fa-
tigue machines. The extra rigidity can be ob-
tained by increasing the diameter of the support
columns or by utilizing three- or four-column
configurations.
Exceptional alignment is required of load
frames used in fatigue evaluations to minimize
undesirable bending forces. In addition, some
means is usually provided to refine the alignment
with manual adjustments when necessary. A
strain-gaged specimen can be used to make this
evaluation.
 l-1. S-N Curves Typical for Steel 27

Schematic S-.V curves for a material at various

stress ratios. l!TS and I’S indicate ultimate tensile
strength and yield strength, respectiveI), in uniax-
ial tensile testing.

The results of fatigue tests are usually plotted as maximum stress or stress amplitude to
number of cycles. .V. to fracture using a logarithmic scale for the number of cycles. Stress is
plotted on either a linear or a logarithmic scale. The resulting tune of data points is called an
S-,Vcur\e. A family of S-,Z’curves for a material tested at various stress ratios is shown schemati-
cally in the above curves. Stress ratio is the algebraic ratio oftwo specified stress Lalues in a stress
cycle. Twocommonl~ used stress ratiosare the ratio. A.ofthealternatingstressamplitudeto the
mean stress (A = Sa,‘Sm) and the ratio. R. of the minimum stress to the maximum stress (R=
S,,,/ S,,,). If the stresses are fully reversed. the stress ratio R becomes -I: ii the stresses are
partially re\,ersed. R becomes a negatke number less than I. If the stress is cycled between a
maximum stressand no load. the stress ratio R becomes zero. lithestress isqcled bewssn two
tensile stresses. the stress ratio R becomes a positive number less than I. .A stress ratio R oi I
indicates no variation in stress. and the test \\ould becomea sustained-load creep test rather than
a iatigue test. For carbon and lo\{-allo! steels. S-.Vcur\es typically halea fairI> straight slanting
portion at low cycles changing into a straight. horizontal line at higher cycles. with a sharp
transition between the two.
An S-.I’cur!e usually represents the median life for a given stress-the life that half the speci-
mens attain. Scatter of fatigue li\es can cover a \ery ibide range.

Source Metals Handbook.%h Edann. Volume I. Proprrr~rrnnd Srlei~~on. lronsand Sreelr. .Amcrwan So&t! ior hlctals. hlcr&
Park OH. 1978. p 667
28 1-2. S-N Curves Typical for Medium-Strength Steels

100

Fracture region
(all specimens fractured)
80
s:
rn Fatigue - Iraclure band
lii rz:
~ CIl
70
ViS:
~
0;-
i! 60 Finile-Iife region
'" CIl (no specimens fractured)
:: (; 50 II ..
0_
.,,g
g>:; 40
c~
~-'
Q; 30
0..
Fatigue limit Infinile-lile region
20

10
O~ _ _-L-_ _-L-_ _- L_ _- ' - - L - _ - L_ _- L_ _- L _

10° 10' 103 10' 10' 10 6 107

(1) (10) (1,000) (10.000) (100,000) (1.000.000) (10,000.000)

Number of cycles to Iracture

S·N curves that typify fatigue test results for testing of

medium-strength steels.

As an explanation, if the single-load fracture strength of the specimens is considered

to be 100 percent, for purposes of illustration this is the starting place, for the speci-
mens can sustain no higher load without fracture. If ten specimens are fractured, the
results are placed as points at the top of the left axis at one load application.
Intuitively it is known that if the maximum load (or stress) is lowered to 90 percent of
the tensile strength, it will require more than one load application to fracture the speci-
mens. The ten points shown in the diagram at 90 percent represent the possible life to
fracture of each of the ten specimens. Because the scale is logarithmic, the points ap-
pear to be relatively close, but in fact the scatter in life from longest to shortest is on the
order of more than 2 to I. At this high stress, plastic deformation of the test specimen is
likely to be great, such as in bending a paper clip or wire coat hanger to make it
fracture. Actual parts are not intentionally designed to operate in this regime, and
normal fatigue fractures have no obvious plastic deformation.
If the load is dropped to 80 percent of the single-load fracture strength and ten more
specimens are tested, they will run longer with a fatigue life scatter of perhaps 3 to I,
which is not unusual, even for theoretically identical specimens (which, of course, they
are not). When the load is dropped to 70 percent, the lives get longer and the scatter in
fatigue life increases to perhaps·5 to I. Again dropping the load, now to 60 percent of
the single-load fracture strength, the fatigue lives again increase, as does the scatter
from longest to shortest life. Invariably, in actual fatigue testing, there is at least one
specimen that inexplicably fractures far earlier than any of the others in the same
group. One such specimen is shown at the 60 percent level fracturing at about 150
cycles, while the other supposedly "identical" specimens or parts had lives of from
about 1,000 to 10,000 cycles. The cause of such an early "anomaly" is often sought in
vain, although it is possible that some metallurgical reason, such as a large inclusion on
the surface, might be found. Frequently, this lone early fracture specimen is simply
ignored.
Dropping now to 50 percent of the single-load fracture strength, the fatigue lives
increase dramatically, as the S-N curve starts to flatten out. This flattening out is
characteristic of ferrous metals oflow and moderate hardness; many nonferrous met-
als and some very high-hardness ferrous metals tend to continue their downward path
at very large numbers of cycles. Now, the problem is when to stop the tests. The test
 1-2. S-N Curves Typical for Medium-Strength Steels (continued) 29

machine will be needed for another test specimen after a very long test time, depending
upon the rate ofloading, or cycles per minute. If ten million is selected as the end point,
the test must be stopped at that figure even if a specimen is unbroken, and the point
shown with an arrow pointing to higher values, for it did not actually fracture. Fre-
quently, five million, or even one million, cycles is selected as the end point, depending
upon the metal, purpose, and urgency of the tests. For example, five hundred million
cycles is sometimes used in the aluminum industry.
The region below the lowest portion of the S-N curve is called the infinite-life region,
because specimens that are tested at stresses below the curve should run indefinitely;
that is, they should have infinite life. The leveling of the S-N curve is the fatigue limit,
characteristic of ferrous metals but not of most nonferrous metals. However, the re-
gion to the left of the sloping part ofthe S-N curve is called the finite-life region, for at
the higher stress levels the test specimens or parts will eventually fracture in fatigue.
This is typical of certain structural parts in aircraft which have their histories carefully
recorded so that they may be inspected and/ or replaced as their fatigue lives are used
up in service. Also, growing fatigue cracks must not be permitted to exceed the critical
flaw size characteristic of the metal and the stress state.

Source: Donald J. Wulpi, Understanding How Components Fail, American Society for Metals, Metals Park OH, 1985,P 135-137
30 1-3. S-N Diagrams Comparing Endurance Limit for
Seven Alloys

100,000
1.1
........... ao'tLerMcarOon.sttel, ~~
90,000
\ Oil quenchedan 1 drawn'

80,000 ~
K"'!> ~,

10,000 ~~e>/
'wI. 0,7quenchedolnd -
l' n
'~6QOOO fir
J II, I,
!
~.
0. dJ.S.Jpe,.l.
O'i''7ve", cen/-carbon 'steel r; c-J.-
.t50.000
<I)
....
("o$e<tand drawn

£ r-,
~ 40,000 <J'~

'x RC'-1,
<Q-/'r.t>e:
" 30POO
~ ~ '! ts« as rolled
:l,fio..L....!. I .
<:0,"",;." ~
C I· v.".,·c ~ •
20,000 0..
~...Q'
r- ~~n
~aJfn I i?o;C:P~
.1
l'fo~"ea-lea
III
10,000
.I II. J.I III . I I,. I II. I I
a
Not: rr (nd(cO/fis flf!f,mfn ~/c.;"lot rUf(furr I
104 10 5 106 10' 108
Number of c~c1e5 for rupture.Ioq scale

Typical S-N diagrams for determining endurance

limit of metals under reversed flexural stress.

To determine the endurance limit of a metal, it is necessary to prepare a number of

similar specimens that are metals tested, and for most nonferrous metals, the S-N
diagrams become horizontal, as nearly as can be determined, for values of N ranging
from 1,000,000 to 50,000,000 cycles, thus indicating a well-defined endurance limit.
The S- N diagrams for duralumin and monel metal do not indicate well-defined endur-
ance limits. The first specimen is tested at a relatively high stress so that failure will
occur at a small number of applications of stress. Succeeding specimens are then
tested, each one at a lower stress. The number of repetitions required to produce failure
increases as the stress decreases. Specimens stressed below the endurance limit will not
rupture. The results of fatigue tests are commonly plotted on diagrams in which values
of stress are plotted as ordinates and values of number of cycles of stress for fracture
are plotted as abscissas. Such diagrams are called S-N diagrams (S for stress, N for
number of cycles). In general, the S- N diagrams are drawn using semilogarithmic plot-
ting as shown in the above diagram, which presents the results for various typical
materials.

Source: Fatigue and Creep Tests of Metals, P 220

 1-4. Steel: Effect of Microstructure 31

I. 0 r-- - - - - - - - - - - - - - - - - ----,
" 0.9
"i
:--:.. 0.8
'<::
"i 0.7
o
'e 0.6
ec 0.5
e 0.4
:>
-g
w
0.3
0.2
0.1
o
Effect of steel microstructure on endurance ratio.

One of the more extensive investigations on influence of microstructure was con-

ducted by Cazaud. The results of some of his work are summarized in the above bar
chart. His data confirm that 0.5 is a conservative number; he found ratios varying from
0.55 to 0.62 for highly tempered martensites. These data were also for steels in the
0.40% carbon range. When untempered martensite is included, the total ratio range is
from 0.26 to 0.62. Untempered 0.40% carbon martensite is about 55 HRC. Above 40
HRC, factors other than microstructure become more significant, especially nonme-
tallic content and residual stress.
Many believe that tempered martensite gives optimum fatigue properties. However,
much of the early work was with medium-carbon steels with intermediate hardnesses.
Only limited data are available for other structures, including low-carbon martensites.
Borik and Chapman determined the endurance limit of bainite and martensite in the
range 36 to 61 HRC. They used 5ll00, a 1.00% carbon steel. They concluded that
above 40 HRC, bainite had better fatigue properties at the same hardness than did
martensite, whereas below 40 HRC the reverse was true. They explained the results in
terms of carbide morphology and distribution. Below 40 HRC, the carbides in the
martensite are spheroidal. Above 40 HRC, the carbide associated with the bainite was
very fine and well-distributed, but below 40 HRC the carbides had a "pearlitic mode,"
which was less favorable in resisting fatigue.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure.
American Society for Metals, Metals Park OH, 1979,P 77
32 l-5. Steel: Influence of Derating Factors on Fatigue Characteristics

Derating factors for influence of surface condition

on fatigue.

The graph above gibes C, factors for various surface conditions. It should be ob-
vious that these factors are approximate. since it is impossible to represent such vari-
able conditions by a single cur\e. C,. the size factor. is significant. Earl) work b>
Horger firmly established that large-diameter samples of the same metallurg) were not
asgood in bending fati@e as weresmall samples. In the presence ofa stress gradient. as
in bending. a larger volume of metal is subject to high stress in a large part than in a
small-diameter part. Since a large volume is subject to maximum stress. there is a
higher probability of a critical-size nonmetallic inclusion to be in that volume. The
fatigue properties established by testing large specimens are thought to represent the
lower bound for a large number of small samples. Sinceaxial tests. b> their nature. test
fairly large volumes at maximum stress. they also gibe lower-bound results. C,, is usu-
all) taken at I.0 for diameters less than O.J’inches and 0.9 for diameters between 0.1
and 3.0 inches. It must be borne in mind that this is a \er) rough estimateand that the
cur\es shown in the above graph are thought to be touard the conser\ati\e side of
scatter bands. The 0.5 relation for S,and S,, is onl! reasonably accurate in the low and
intermediate hardness ranges because of limitations related to microstructure.
nonmetallic-inclusion content. and carbon content 31 higher hardnesses.

Source: 0. H. Brcen and E. hf. b’ene. “Fatigue in hlashines and S~ruc~urcs~Ground Vshlslcs. ‘. in Far~gue and hlcroaucture.
American So&t) for Metals. hlerals Park OH. 1979. p 72
 l-6. Steel: Correction Factors for Various Surface Conditions 33

Value for loading in

Factor Bending Torsion Tension

K, . 1.0 0.58 0.9(a)

Kd, where:
d c 0.4 in. 1.0 1.0 1.0
0.4 in. < d
5 2 in. . . . 0.9
0.9 1.0
K. . . . . . . . . . From chart above
(aI A lower value 10.06 to 0.85) may be used to
account for known or suspected undetermined
bending because of load eccentricity.

Correction factors for surface roughness (kJ,

type of loading (14). and poti diameter (KJ, for
fatigue life of steel ports.

Comparative effects of various surface conditions on fatigue limit of steels at

various levels of tensile strength.

Source. Metals Handbook.9rh Edlrlon. Volume I. Properr~erand Selcc[ia,n. Ironsand Srscls. American SO~ISI) ior hlclalr. hleralr
Parh OH. 19% p 6’1
34 l-7. Fatigue Behavior: Ferrous vs Nonferrous Metals

S (stress)--N (cycles to failure) curves. 4. ferrous

metals; B. nonferrous metals. S, is the endurance
limit.

Traditionally. the behavior of a material under conditions of fatigue has been

studied by obtaining the S-.‘l’cur\es (see above), where S is the stress and .E- is the
number of cycles to failure. For steels. in general. one obsenes a fatigue limit or endur-
ance limit (curve A above) which represents a stress level below which the material does
not fail and can be cycled infinitely. Such an endurance limit does not exist for nonfer-
rous metals (curve B above). The relation between Sand ,V. it must be pointed out. is
not a single-ralue function but serves to indicate a statistical tendency.
Up until the 196Os, almost all fatigue failures. and consequently all the research in
the field, was confined to moving mechanical components (e.g.. axles. gears, etc.).
Starting in the late 1950s. entire structures or very large structural elements (e.g.. pres-
sure Lessels. rockets, airplane fuselages, etc.) have been studied and tested for fatigue.
This can beattributed to the use of materials such as high-strength alloys. together with
the advances in the fabrication technolom. resulting in monolithic structures meant to
undergo high cyclic stresses in service. It is this class of materials which has shown
catastrophic failures in fatigue. and it is for this kind of material that fracture mechan-
ics is being applied. with considerable success. to fatigue problems.

Source: Marc Andrt Meyersand Krishan KumarChauls. hlechanlcal hle~allurg): Prmaplesand .Appllcaionr. Prcm~ce-Hall. Inc..
Engleuood Chfk NJ. 198-l. p 689
l-8. Comparison of Fatigue Characteristics: Mild Steel vs 35
Aluminum Alloy

Typical bending (R = -1) fatigue curves for fer-

rous and nonferrous metals.

Here it is noted the lack of the “knee” for the aluminum allo)
compared with steel: that is. the point on the cun’e where the
number of cycles to failure becomes a straight line-essential11
infinity.
36 1-9. Carbon Steel: Effect of Lead as an Additive

120~-...,...--r----r--r---,---,---.-----,---.--...,.-----,

IOo/-----1I--+----1--+-----+~~
V>
n.
o BOI-------t---t----r-----:
o
o

Fatigue limit of leaded and nonleaded alloy steels as a

function of ultimate tensile strength.

Lead is often added to steels to improve machinability, al-

though usually at the cost of a minor (usually) loss in mechanical
properties,. The interrelationship of lead additions with tensile
strength and fatigue limit is summarized in the above graph.

Source: George M. Sinclair, "Some Metallurgical Aspects of Fatigue."in Fatigue-An Interdisciplinary Approach. John J. Burke,
Norman L. Reed and Volker Weiss, Eds., Syracuse University Press, Syracuse NY, 1964. p 68
 1-10. Corrosion Fatigue: General Effect on Behavior 37

t Fatigue without corrosion

Fatigue limit

Logarithm of number of cycles needed for fracture _

Effect of alternating stresses with and without corrosion.

If a specimen is subjected to alternating stress (ten- below the fatigue limit, the life, in absence of corro-
sion and compression in turn) over a range insufficient sion, should be indefinitely long as shown above.
to cause immediate fracture, gliding may occur within
In the presence of a corrosive environment the situa-
some of the grains, but when the disloca tions reach a
tion will be different. Disorganized atoms along a
grain-boundary they are halted, retracing their move-
gliding-plane may require less activation energy to
ment along the gliding-plane when the stress is re-
pass into a liquid than more perfectly arrayed atoms
versed. If the material were ideal, it might be hoped
elsewhere; certainly, while the atoms are in motion
that the dislocations would merely move to and fro
along a gliding-plane, preferential attack may reason-
along the plane, and that no damage would result. In
ably be expected even below the fatigue limit. This
practice a large number of cycles can be withstood
means that there is no "safe stress range" within which
without apparent damage, but in material as we know
the life should be infinite. It is, however, convenient to
it, slight irregularities will prevent smooth gliding in-
determine an endurance limit-namely, the stress
definitely, and roughening along the original gliding-
range below which the material will endure some spec-
plane will make movement difficult, so that gliding
ified number of cycles (the number must be stated).
will then start on another parallel plane. In the end,
bands of material will have become disorganized, and It should be noted that, although stress-corrosion
ultimately one of two things must happen: (1) if the cracking is often intergranular, corrosion-fatigue
stress range is low, gliding will cease altogether, the cracks are usually transgranular, following gliding-
only changes still produced by the alternating stress planes inclined at such an angle as to provide high re-
being elastic, (2) if it exceeds a certain level (the fatigue solved shear stress. There are exceptions to both rules.
limit) the gliding will become so irregular, as to cause Whitwham, studying corrosion-fatigue cracks on steel,
separation between the moving surfaces, first locally, found that, although mainly transgranular, they fol-
producing gaps, which later will join up into cracks. lowed grain-boundaries for short distances, where
Thus above the fatigue limit (after a time which is such boundaries chanced to run in a convenient
shorter at high stress ranges), there will be failure; direction.

Source: Ulick R. Evans, An Introduction to Metallic Corrosion, 3d Edition, Edward Arnold (Publishers) Ltd and American Society
for Metals, Metals Park OH, 1982, P 160
38 1-11. Effect of Corrosion on Fatigue Characteristics of Several Steels

l00r------------,
'Copper' steel
l' 09')(, C Steel

~
(annealed) 40
Tensile: 103500 Ib.lsq.in.
80 (hardened and
tempered)
:c 20 12·9')(, Cr,0·11')(, C
u :c
u Tensile: 89600 Ib./sq.in
60 .5 0·14')(, C,O'9B')(, Cu 45
.5

~
ii.5 g- O
Tensile:61500 Ib./sQ.in.
g-
Ii;
c. 104 105 106 107 lOB Ii; 66
g-

.
c.

-
40
~
,e 100 ,e 35
"c.
,e
'0
20
0
i:1<:
ill 80
'0
i:1<:
ill
.,
w :> :>
"0 0 0 25
e
ill
:> 0 :E :E 104 105 106 107 lOB
0 103 104 105 106 107 lOB "'"<: 60 "'e"
:E e
\4
i:
l!!
"'"e ~ ~
50
e

"R"."~
~

,,~
40 40
.1£
~
'E Chrome- E
"
(/)
vanadium "
(/)

'E 20 20 steel chromium

"
(/)
(hardened and tempered I
(hardened and temperedl
O'BB')(, Cr,0·14% Vo,0'46')(, C
steel
27')(, Cr.0'2')(, C
Tensile: 65700 lb.Zsq.in,
Tensile: 150600 Ib.lsq,in. Tensile: 81000 Ib.lsQ,ln.
0 0 30
103 104 105 106 107 lOB 104 105 106 107 lOB 104 105 106 107 lOB

N = Cyclesto fracture 1I0g scale)

---Denotes testsmadein air

6------ " "stream of fresh water
II saline riverwater

Typical curves showing the number of cycles needed to produce fracture at different
stress ranges in absence and presence of corrosion.

Two main procedures are available for corrosion-fatigue

tests:
One-stage tests. Here the corrosion fatigue is continued until
breakage. The logarithm of the number of cycles needed to pro-
duce breakage is generally plotted against the stress range, as in
the above curves selected by Gough from McAdam's experi-
mental data.
Two-stage tests. Here the corrosion fatigue is interrupted af-
ter a definite number of cycles, and the residual strength is esti-
mated by measuring either (a) the endurance limit in the absence
of corrosive influences (i.e., the stress which can be withstood
for some definite number of cycles, (b) the number of cycles
needed to produce fracture in the absence of corrosive influ-
ences at some definite stress, (c) the tensile strength, or (d) the
shock resistance (Izod number).

Source: Ulick R. Evans, An Introduction to Metallic Corrosion, 3d Edition. Edward Arnold (Publishers) Ltd and American Society
for Metals, Metals Park OH, 1982, P 165
 1-12. Steel: Effect of Hydrogen on Fatigue Crack Propagation 39

I
PR PAGAIION LIfE

-,---_.
I
I
i

1+--.,--,---,--..c--,...-,-,--.-.+---.--.---.--.,..-,-,....,.......,r+--...,.----,.---..
1 6 7 0 9 0

S-N type of fatigue curve.

In the majority of all cases, the external load changing with

time, whereby low frequencies «10- 2 Hz) have the highest prac-
tical importance. Under these circumstances a structural com-
ponent can be subject to fatigue which is conventionally de-
scribed by an S-N curve relating the cycle life, N, to applied
stress, S, as in the above chart. In non-aggressive environments
an endurance limit can be defined below which no fatigue failure
occurs. A disadvantage of this approach is that S-N curves do
not differentiate between crack initiation and crack propaga-
tion. The number of the cycles corresponding to the endurance
limit presents initiation life primarily, whereas the number of
cycles for crack initiation at a high value of applied stress is
negligible. Consequently S-Ntype data do not necessarily pro-
vide information regarding safe-life predictions in structural
components. Particularly, if the structure contains surface ir-
regularities different from those of the test specimens, these are
likely to reduce or even eliminate the crack initiation portion of
the fatigue life.

Source: M. Kesten and K.-F. Windgassen, "Design of Equipment to Resist Hydrogen Fatigue Service," in Current Solutions to
Hydrogen Problems in Steels, C. G. Interrante and G. M. Pressouyre, Eds., American Society for Metals, Metals Park OH, 1982,
P 390
40 l-1 3. Relationship of Stress Amplitude and Cycles to Failure

(a) Finite life ASR diagram. showing R = -I equivalent stress for R = 0.6 loading. (b)
S-.Y diagram, showing life prediction for R = 0.6 loading using R = -I equivalent
stress.

The ASR diagrams normally use theendurance-limit fatigue-

strength value. but substitution of the fatigue strength at specific
finite li\es can also be used (see chart a abole). Life estima-
tions from the diagram can be done using such information as is
shottn in charts a and b abo\,e. Here the hno\\n stress range at
some Rvalue is conierted to an equivalent completely reversed
(R= - I) stress. and thisequkalent stress is applied to the mate-
rial’s R = -I S-.Vcur\e for the life estimate.
Designers have the ability to calculate the component’s
stresses using classical formulas or the computer-based finite-
element-analysis (FEA) techniques. Both of these methods ex-
amine the elements for the maximum stresses that are normall>
in the areas of a discontinuity. or stress concentration.

Source. D. H Brscn and E hl Hens. “Faugue m hlachlner and StrucrtmpGround \ ehlcles.” in Fatigue and hlwrosrruc~ure.
Amencan So&l! ior hkrals. Metals Park OH. IYTY. p 6’
 1-14. Strain- Life and Stress- Life Curves 41

2Nf
Reversals to failure (lag scale)

Strain-life and stress-life curves,

Fatigue damage is caused by cyclic plastic strain, and consequently, the fatigue life
should be related to the plastic-strain amplitude. Coffin and Manson independently
proposed a relationship between the plastic-strain amplitude and the cycles to failure
of the form:
AEp/ 2 = EJ (2NJ),
where Ej is the fatigue-ductility coefficient, 2NJ is the number of reversals to failure, and
c is the fatigue-ductility exponent. Their equation is very similar to the Basquin equa-
tion relating the elastic-strain or true-stress amplitude to the number of load reversals
to failure:
I::.E,E/2 = aa = aJ (2NJ)b
where I::.E,/2 is the elastic-strain amplitude, E is the modulus of elasticity, his the
fatigue-strength exponent, and a;
is the fatigue-strength coefficient. A schematic rep-
resentation of these relationships and their superposition is shown in the above dia-
gram. The summation curve is analogous to the stress-life, Wohler diagram, if the
strain amplitudes are replaced by their respective stress amplitudes. The intersection of
the Basquin and Coffin-Manson plots is normally defined as the transition between
high- and low-cycle fatigue. Consequently, the regime of low-cycle fatigue depends on
the properties (for example, the ductility) of a particular material.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, p 211
42 1-15. Fatigue Plot for Steel: Ultrasonic Attenuation vs
Number of Cycles

0.264

0.249
E
~ 0.232
CD
:5!
t:I
z 0.216 ADDITIONAL PULSE

l
0
;::: 0.200 OBSERVED
«
:::>
Z 0.4 dBATIENUATION
0.184
~
«
0.168
CHANGE OBSERVED
l
0.152

5
10.8xl0

Typical plot of ultrasonic attenuation versus num-

ber of fatigue cycles for steel.

Joshi and Green determined the attenuation coefficient IX for longitudinal bulk
waves in aluminum and steel at 10 and 5 MHz, respectively. The measurements have
been performed in a pulse-echo mode, with the acoustic pulse reflected at the back
surface of the material. The above chart shows their results obtained on cold rolled
steel bars. The attenuation started to increase at about 6 X 105 fatigue cycles (65% of
fatigue life). At roughly 7.5 X 105 cycles (85%), an additional pulse was observed,
arriving earlier than the one reflected from the back surface.
Results are interpreted in terms of a series of microcracks being formed, probably at
the surface. As soon as the microcracks are sufficiently deep, they will change the bulk
attenuation. As soon as a macrocrack has been formed (by coalescence of micro-
cracks), it will reflect part ofthe pulse. After that, the attenuation is primarily deter-
mined by the transmission coefficient of this single crack. Thus, the attenuation curve
(versus fatigue cycles) becomes discontinuous, as may be noticed in the above chart.

Source: O. Buck and G. A. Alers, "New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstruc-
ture, American Society for Metals. Metals Park OB, 1979. p 135
 2-1. Typical S-N Curve for Low-Carbon Steel Under Axial Tension 43

28
TYPICAL FATIGUE CURVE

26
I FOR M.S. UNDER
REPEATED AXIAL "TENSION
~
c (f MIN. = 0)
~
1.
~ 24
I
rc-
)(
«
~ 22
I
I

't-
'-'
III
i
t
III
uJ
0:: 20
\
\
J-
III
0
W
18

--
.J
0.
0.
<t: I~
~
~
16 --fATIG-U-E-LlMi.:r-(REPEATEO- TEtlSION)

14
o
I
2 4
I G
I 8 10 \2.
NUMBER OF STRESS APPLlC",.\ONS- MILLIONS

The term "fatigue" refers to the failure of metals from repetitions of stress rather than from a
single application, as occurs for example in a simple tensile test or with a brittle failure. The value
of the stress necessary to cause failure of a material from fatigue is lower than its nominal tensile
strength. For example, a sample of mild steel may have a maximum stress of 27 t.p.s.i. when
subjected to a single application ofload as in an ordinary tensile test. If, however, a stress of say
25 t.p.s.i. is applied repeatedly to the same material, failure will not take place until this has been
done a certain number of times, while at a lower stress still, the number ofload cycles required to
cause failure will be even greater. If testing is continued in this manner, a stress value will ulti-
mately be found at which fracture will not occur, no matter how many stress repetitions are
applied. This value is known as the fatigue limit of the material. If the results from such a series of
tests are plotted, a graph such as the one above will be obtained, the curve tending to run parallel
to the abscissa after approximately IOmillion cycles (for steel), the corresponding value ofthe
stress being known as the fatigue limit. Under conditions of repeated tension the value of the
fatigue limit for the above mild steel which has a tensile strength of approximately 27 t.p.s.i.
would be of the order of 16 t.p.s.i. If the same steel was tested under conditions of reversed
bending stresses a value of the order of± 12 t.p.s.i. may be found.
It must also be pointed out that where corrosive conditions operate in addition to fluctuating
stresses, failure from "corrosion-fatigue"may occur and, in these circumstances, the concept of a
fatigue limit does not apply, since if the stress applications are continued for a sufficient number
of times, ultimate failure will occur. Further, most nonferrous metals and alloys do not possess
a fatigue limit.

Source: F. R. Hutchings, "Fatigue Failure of Components of Lifting Machinery," in Failure Analysis: The British Engine Technical
Reports, F. R. Hutchings and Paul Unterweiser, Eds., American Society for Metals, Metals Park OH, 1981, P 344
44 2-2. AISI 1006: Effects of Biaxial Stretching and Cold Rolling

..
C'l
b
)(
cP
Eoff

.. 0.2
~ 00.4
..r a 0.6
s 6
"j
c
:::>
f-
:::i
Do
::l!: 4
oct
z
«
a:
lii
..J
oct 2
f-
0
f- a

10 3 104 loS 10 6
REVERSALSTO FAILURE, 2N,

..
C'l
b
)(
8
Eeff

~ .. 0.2
..r
s 00.4
w· a 0.6
c 6
:::>
f-
:::i
Do
::l!:
oct
z 4
«
a:
f-
Vl
..J
oct
f-
0 2 Runouts
f-

r
103 104 loS 10 6
REVERSALSTO FAILURE, 2N,

Strain-life plots for two modes of deformation for 1006 steel.

Plots in the top chart are for biaxial stretching; those in the bottom chart are for cold
rolling. Included is the data band for the undeformed material. The effect of balanced
biaxial stretching on fatigue life was as follows: at large strain amplitudes (/:;.EI/2~
~ 2.5 X 10-3 ) , the fatigue life remained approximately the same or decreased slightly
when compared to that ofthe undeformed material; in contrast, at small strain ampli-
tudes the fatigue life increased as a result of the prior deformation.
After cold rolling, the fatigue life was approximately the same as in the undeformed
material at large strain amplitudes (short lives) but it was longer at small strain ampli-
tudes (long lives). Thus, unlike BBS, CR appeared to cause no reduction in fatigue life
at short lives. Another difference between the two deformation modes was that the
scatter ofthe data was larger after BBS than after CR. Thus, BBS was somewhat more
detrimental to the fatigue life than CR.

Source: John M. Holt and Philippe L. Charpentier. "Effect of Cold Formingon the Strain-Controlled Fatigue Properties ofHSLA
Steel Sheets," in H'Sl.A'Stccls-c-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 217
 2-3. AISI 1006: Weldment; FCAW, TIG Dressed 45

SAE-1006
R' 0.1. t' 0.13" 13.3mml
--~-- Smooth Specimen
--<>-- TIG-Dressed
_.--fr-.- As-Welded

_ 00 0 - - - - -
._....g 0 0 0 roo 6_ --rrtr-lS" i:J_ _
tr 6
~D;!.~_ 0 0 0 --
o 0 .-._ oro ~
'-.0 0-.
o DO ~._._
'-''''lJ..._
'--0
..............

10 5 10 6
NT' CYCLES TO FAILURE

Fatigue strengths of FCA W/TlG- dressed joints compared to those without TlG
dressing for AISI I006steel (unwelded). The improvement in fatigue provided by TlG
dressing tbe welds is obvious.

Source: Kon-Mei Ewing, Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens. "Weld Fatigue ofTlG-Dressed
SAE-98QX HSLA Steel,"in HSLA Steels-Technology & Applications. American Society for Metals, Metals Park OH. 1984,p 557
46 2-4. AISI 1006: Weldment; Shear Joints

CJ)
~.or
• • .
(a,.-32KSI
<J • \;
\0 1
SAE 1006 LAP-SHEAR WELDS ur= 32 KSI
Kfmax = 2.77 I R = 0.1
• EXPERIMENT
- PREDICTION

10 5 106
NT I CYCLES
Total fatigue life predictions and experimental results for FCA W, AISII0061ap-shear
joints. Note that the results and predictions compare closely.

Source: Kon-Mei Ewing. Pei-Chung Wang, Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue ofTIG-Dressed
SAE-980X HSLA Steel," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984,p 562
 2-5. AISI 1006: Weldment; Lap-Shear Joints 47

~
a.
~
(f)
• • . €~"-'2KSI vi
<I • ;;; 10
2
<l
10 1
SAE 1006 LAP-SHEAR WELDS CTr =32 KSI
Kfmax = 2.77. R = 0.1

-
• EXPERIMENT
PREDICTION

101

NT, CYCLES
Total fatigue life predictions and experimental results for FCAW, AISI 1006 lap-shear
joints. Here, the prediction and actual results are very close.

Source: Kon-Mei Ewing. Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue ofTIG-Dressed
SAE-980X HSLA Steel, "in HSLA Steels-Technology &Applicalions, American Society for Metals, Metals Park OH, 1984,p 562
48 2-6. AISI 1015: Effect of Cold Working

5.0 . 0.20

E
~
E 4.0
.x:
<, 0.15
M

E <II
E 0-

'"
M
I
0
3.0 I-
ui :E
I- -O.I0:J
<l:
a: uJ
a: 2.0 ::J
<l:
lLJ
'"
I-
:: 0.05
<l:
u,

1.0

oL--!------~---_:::_---_::::__'O
o 20 40 60
DEGREE OF COLD-WORKING, %
Comparison of effects of cold working on wear rate and fatigue limit of fully
annealed 0.15%C mild steel. Wear was determined in sliding between the end
surfaces of cylinders at a speed of 0.56 mjs under the loads .:82 N, ():124 N
and 0:147 N in machine oil. Fatigue limit (.) was determined by reversed
bending fatigue tests of notched plate specimens 25 mm wide and 4 mm thick
having a central hole 1.5 mm in diameter.

Attempts have been made to determine effects of cold-working on the

resistance to wear and fatigue of a O.15%C mild steel. Fully annealed material
was then cold-worked to different degrees and the specimens were machined
from it. Wear experiments were conducted in a rotating cylinder machine as
described above with a machine oil as the lubricant. Care had been taken to
avoid the effects of work hardening during machining by electrolytically pol-
ishing the sliding surface. Reversed bending fatigue tests were carried out by
using notched test pieces of the same material. The wear rate and the fatigue
limit are compared with the degree of cold-working in the above chart, which
shows a definite correlation.

Source: Yoshitsugu Kimura, "The Role of Fatigue in Sliding Wear," in Fundamentals of Friction and Wear of Materials, David A.
Rigney, Ed., American Society for Metals, Metals Park OH. 1981, P 215
 2-7. A533 Steel Plate: Fatigue Crack Growth Rate 49

Stross-intensity factor rango, .0. K, ksi • in. 1/ 2

50

10- 2

10- 4

a
<9':!
10- 3
01
~ 00 I
~ Region 1: 1 10- 6 U'"
E slow crack I Region 3: ~
E growth I rapid .5
Z I unstable
~
I crack Z
...
co
",'
10- 4 growth ~
...
co

1! I ti
I l!
ie I
10- 6
~
at
I e
~
eu I ...u'"
...e 10- 6 I l!u
...l!!
I
~
'"
:iE 10-7 ~
Region 2: power-law behavior '"
:iE

10- 6

Stross-intensity factor rango, .0. K, MPa . m 1/2

Fatigue crack growth behavior of AS33 steel. The material was ASTM
AS33 B-1 steel, with a yield strength of 470 MPa (70 ksl), Test condi-
tions: R= 0.10; ambient room air; 24°C (75 OF).

The general nature of fatigue crack growth and its description using fracture mechanics can be
briefly summarized by the example data shown in the above chart. This figure, based on the
work of Paris et al, shows a logarithmic plot of the crack growth per cycle, daj dN, versus the
stress-intensity factor range, t::.K, corresponding to the load cycle applied to a sample. The
da] dN versus t::.K plot shown is from five specimens of ASTM A533 B-1 steel tested at 24°C
(75 OF). A plot of similar shape is expected with most structural alloys; the absolute values of
daj d N and t::.K are dependent on the material. Results of fatigue crack growth rate tests for
nearly all metallic structural materials have shown that the da I dN versus t::.K curves have the
following characteristics: (a) a region at low values of dald N and t::.K in which fatigue cracks
grow extremely slowly or not at all below a lower limit of t::.K called the threshold of t::.K, t::.K,,,;
(b) an intermediate region of power-law behavior described by the Paris equation:
~=C(t::.KJ'
dN

Source: J. H. Underwood and W. W. Gerberich, "Concepts of Fracture Mechanics." in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH. 1982. P 18
50 2-8. A514F Steel Plate: Fatigue Crack Growth Rates

MPavm 100 20 100

20
10'

A514F
CON Quality '(J '
I A514F
CaT Quality
~ 10"
<

~
~
E

I
E

/;
z!. lS----1
u- -::'O··

.,
15-----·<
TL-- ---- .. - --- -_ :
/,. ~I=-:-:=:':::- i

100 10 '00
l'.K. ksl\ ii1.

Plots of fatigue crack growth rate versus range of stress

intensity factor (best fit lines) for A514F plates.

The increased isotropy in the CaT over the CON steels is evi-
dent with the through thickness (ST, SL) orientation having the
fastest growth rate in the CON steel and showing the greatest
improvement by CaT.

Source: Alexander D. Wilson. 'The Effect of Inclusions on the Properties of Constructional Steels." in Wear and Fracture Preven-
tion. American Society for Metals. Metals Park OH, 1981. p 196
 2-9. A514F and A633C: Variation in Fatigue Crack Growth Rate 51
With Orientation

5.10-'

~Kof 50 kst/iil (55MPa.frii)

CON~
A633e A633C
CaT 0
CON~ ~§~~
A514F
caTD ICaT I A514F

1110"' 3.10-' 4.10·' 8x10" 12110" 16.10" 20110-'

daldN, inches/cycle
Comparison of fatigue crack growth rate variation with
orientation for A633C and AS14F plates at two tJJ( levels.

These data show that the CaT improvement in FCP growth rate
takes place only at higher L1Klevels. Additionally, this figure indicates
that there is a more substantial enhancement in FCP behavior for
ASI4F. Also there generally appears to be more anisotropy in the
ASl4F steels of both quality levels. It has previously been shown that
higher strength level steels tend to be more adversely affected by inclu-
sions associated in groups, such as present in CON steels.

Source: Alexander D. Wilson, "The Effect ofInclusions on the Properties of Constructional Steels, "in Wear and Fracture Preven-
tion, American Society for Metals, Metals Park OH, 1981, P 197
52 2-10. A514F: Scatterbands of Fatigue Crack Growth Rate

MPavm
20 100
10. 0 :------ .....,
I
I
I I
, I I

,:
I
,
I
10"
, I
I
I

,,,
I
I I
I

I
, I

I ~
I

I
I
I ~
I

A514F }
2'/.ln(57mm)Gage E
E
CjCON
ClCaT 10"

10·' L-_ _-----''--_-'------'_'-'-'--'---'-L-_ _-----'

10 100
6 K , ksi\/fil.

Summary scatterbands of fatigue crack growth

rate versus range of stress intensity factor encom-
passing all data points in 6-orientation testing
comparing CON and CaT quality A514F plates.

Source: Alexander D. Wilson, "The Effect oflnclusions on the Properties of Constructiona ISteels,"in Wear and Fracture Preven-
tion, American Society for Metals, Metals Park OH, 1981,p 197
 2-11. A633C Steel Plate: Scatterbands of Fatigue Crack Growth 53
Rates

20
MPavm 100
10-'

10-'

GI

I
'fi 10"'
,5
A633C
4In(102mm)Gage
GI

I
C=JCON E
z
~
DCaT 10-'

10-' L-_ _---'-_---''---'------'-----'----'--'--'---'--_ _------"-'

10 100
,c.. K • kslv'ln:

Summary scatterbands of fatigue crack growth

rate versus range of stress intensity factor encom-
passing all data points in 6-orientation resting
comparing CON and CaT quality A633C plates.

In this presentation the generally faster FCP growth rates for

the CON steels at higher 6.K levels are displayed, as well as the
improved isotropy of the CaT steels.

Source: Alexander D. Wilson,"The Effect of Inclusions of the Properties of Constructional Steels," in Wear and Fracture Preven-
tion, American Society for Metals, Metals Park OR, 1981, P 196
54 2-12. Low-Carbon Steel Weldment: Effects of Various Weld Defects

100

80 I I
poli:he~ Plai~ pl~te
-
60
r- -,~ 0
- -- ....c
I I

I,t:
I I
+ 1 standard deviat ion

40
... 1"-1-
...
0

--- uv
-
'."C
-

ri
_C ~-
r----- . --
... • r-9 ·0
data bank

o VV if.. ...ll ... Data bank

-in
IA ;--I-- r-- mean
.0/.
OJ"
e» ., " f..' curve
~
~ 20
on
~ Low-carbon steel ., ~
'flo
ch reinforcement off
9-
LC?cation of failure:
10
o Plate or weld edge
• Porosity in weld
... Slag near surface
-
V
V

11 111
8
b. Slag at midthickness V
f-
V Lack of penetration 11
I~
6 f-
1<:;1-
I I
2 4 6 B 10' 2 4 6 8 • 2 4 6 8 ,
10' 10 10
Fatigue life, cycles

S-N curves showing effect of various weld defects on fatigue life of a

low-carbon steel weldment, presented as a comparison with fatigue life
of the plate.

Source: Metals Handbook, 9th Edition, Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 848
 2-13. Low-Carbon Steel Weldment: Effect of Weld Reinforcement 55
and.Lack of Inclusions

Reinforcement intact
-e Reinforcement removed
;j!.
"0'
0;
;:
"0
C
::J
g
(;
-S
'"
~ 40 _---"" __
_-~......=c-I_---~I""'-_~=__iI_---___l -----
1;; ----
CI>

.,
::J
.'"
'"
u, 201_----1_----1--

Ol.-- .l.- ..l- ...J...... ...l- -J

o 0.2 0.4 0.6 0.8 1.0

Slag inclusion length, in.

Fatigue strength of a weldment containing slag inclusions as a

percentage of the mean fatigue strength of a sound low-carbon
steel weld.

Source; Metals Handbook, 9th Edition, Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 850
56 2-14. Low-Carbon Steel Weldment: Effect of Weld Reinforcement
and Lack of Penetration

100

\\ r-....
\\
-,~'" <,
R . fl.
80
- ~I
I
I
em orcement mtact
I
Reinforcement remived

60
"
1'-,
...........
.....
.....
<, ~ t--.

--.- r--r-.r---
.................. ~ r--
----<.
.i
A 100 oob cycles
~

-. -
40

\/ -:
/
-.-
2 000 000 cycles
20

o
o 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
Lack of penetration half depth, in.

Fatigue strength ofa weldment containing lack of penetration as

a percentage ofthe mean fatigue strength ofa sound low-carbon
steel weld.

Source: Metals Handbook, 9th Edition. Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 849
 2-15. Low-Carbon Steel Weldment: Computed Fatigue Strength; 57
Weldment Contained Lack of Fusion

100
\ ~ 1\
80 \'\ III
'\
r-,
....
Reinforcement intact
I I

"
' I
;fi.
,,'
Qj
<, <, - . - Reinforcement rejOved

s <, <,
-g 60
~
'0
"""-
" - -::::--- --~
....... ..........""'" ~ ~-- ~
b..

r-._ r--
J
1 100 000 cycles

--
-S ~.- ::-- ....t.
'"
~ 40 •
Ii:
Ql }
".,'"5. 2 000 000 cycles
LL 20

o
o 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
Lack of fusion half depth, in.

Computed fatigue strength of a weldment containing lack offu-

sion as a percentage ofthe mean fatigue strength oCa sound low-
carbon steel weld.

Source: Metals Handbook, 9th Edition, Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
QH, 1983, P 849
58 2-16. Low-Carbon Steel Weldment: Effect of Reinforcement and
Undercutting

100
~ I
~~ ~ . fl. I
""""- ern orcement Intact
80
~ \~
. - Reinforcement removed
*'
'0'

~ <,
'0
5 60
\ [\ ~ ....... -
;00 000 cycles
g
...o
I" '-,......r- r': t::'- r--- I

fia>
~ 40 ""'-
r-; :::::-
10-. ~
r::: t:-- r--.r-- .r:;'/
to-.
t:
t::-I-- i'---
'

-.
""'--
~ t:- t--
Cl>
:::l
a>
';;
co
u..
20
r-- 1-. --. r • t--

"k 00 000 cycles

a
a 0.Q2 0.04 0.06 0.08 0.10 0.12 0.14 0.16
Undercut depth, in.

Effect of depth of undercut in terms of percentage of fatigue

strength of a sound low-carbon steel weld.

Source: Metals Handbook, 9th Edition. Volume 6. Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH. 1983. P 848
 2-17. Low-Carbon Steel: Transverse Butt Welds; Effect of 59
Reinforcement

• • 'I r , '1-0--- , - - - , - - r ,-,.-.-,--""--"~J.------,---r---"I---y---r-T"""T""

........ "-

600 • ~"""""O<UNWELDED
.0 0 0 •
500 ~<, O~ ~REINFORCEMENT OFF
• • .~.,-.-3
o __• ~ Q.... ~_

• 1.5 •• 0 - -~- 't:..

Q..
~ 2.3 - .
*• .• REINFORCEMENT ON
••
3'-............ <,

.
x
~--.
ro
E
(/)
300
h =3.8mm ..~ •• j
.........
t.:
----t--~
<,
...
200
r h .'6~-------~ __ C
Tr==r?J ~. • ~
100
~ • -.r:~

CYCLES

Influence of weld reinforcement on fatigue strength (R=O) of transverse butt welds of

quenched and tempered carbonsteels, From these data it is evident that removal ofthe
reinforcement (weld dressing) improves fatigue strength and fatigue life.

Source: Drew V. Nelson, "Fatigue Considerations in Welded Structure," in Proceedings of the SAE Fatigue Conference P-109,
Society of Automotive Engineers, Inc., Warrendale PA, 1982, P 206
60 2-18. A36/E60S-3 Steel Plate: Butt Welds

IOOr:---,----,.--..-r-.-rn"----r--r--r-OT,,""TT--..--r-r-,T"T,,,
80 600

6Or---====:=--_
40
400
300
CT, =0
200
20

CT,' +35 ksl 100~

10 7O:e
~ 8
vi
5O<l
6 /!l361 E60S-3 Double-V Bull Welds

4 ~ma.'2.5I, R=0, I = 5/8 in. 30

OCT, = + 35 ksl 20
S S
2

10
7
10
N1 ' Cycles

Fatigue crack initiation life predictions and exper-

imental results for ¥a-in. (16-mm) A36/E60S-3
butt welds.

Source: F. V. Lawrence, "The Predicted Influence of Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981,P 114
 2-19. A514F/E1·10 Steel: Bead on Plate Weldment 61

~14F/EII(j Begd On plgte

Kfmal • 31~1 DJ • 0.01In.• R -0. t -112 ln.
- - . "".120kl'
---0 "" ·-120kli
200
200

. 100
10 ~
l.
vi
<I
-:---_-
• --- 00 0 - 0 - - - _
"'----
so <I

30
vi

20

20

10

Total fatigue life predictions and experimental re-

sults for A514F /EllO weldments with tensile and
compressive residual stresses.

Source: F. V. Lawrence, "The Predicted Influence of Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 113
62 2-20. A36 and A514 Steel Plates: Butt Welded

514F / E 110 Bull Wold

100 ~
K,...3.15, 01- D.OIIn. (0254mm)
10 70 ::f.
~ 8 .-90·, , -60·, I -1/2 In.. (I2.7mml
50 '"
~
6 --.,.s,
- - - ' , -0

s~s
~ .. 30

20

K,..... r

10

NT' Cy(lu

Predicted effect of stress relief and stress ratio on

A514/EllO butt weld fatigue life.

1~'r--~--r-r""""~"'--~-~~""""~"'--~-'-"""'~~"'6oo

60 400
300
40
R'O

20
A36/E60S-3 Bull Weld
K,,"_ ·2.35, Ot' 0.01 In (O.254mmJ
10 .-90·,' -60·, 1-I/Zin,,(I27mm)
8
--.,-s),
6
- - - w,-O

~-.
30

20
s~s
K r_ . '
10
10< 10'
NT' Cycle,

Predicted effect of stress relief and stress relief and

stress ratio on A36/E60S-3 butt weld fatigue life.

The results for the high-strength, quenched-and-tempered steels (upper chart), indicate that such
materials can sustain high residual stresses which do not relax. The total fatigue life of such materials is
strongly influenced by both residual stress (a,) and stress ratio (R). Stress relief or mechanically in-
duced compressive residuals should be highly effective. An intermediate case is mild steel as shown in
the lower chart. Mild steels can have appreciable residual stresses; but, since the transition fatigue life
(N,,) is often very long (= 500,000 cycles), there are large amounts of plasticity at the notch root even at
long lives (106 cycles); this notch-root plasticity tends to relax rapidly the notch-root residual and mean
stresses with the result that N[is little affected forlives less than 106 cycles. The observed dependence of
N» on stress ratio does, however, result in a predicted variation oftotal fatigue life with stress ratio R.

Source: F. V. Lawrence, "The Predicted Influence of Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 112
 2-21. A36 Plate Steel: Butt Welded 63

100
80 600
60
Zero Mean Slress
40
200

...;;
20

10
--\---====::::::::d Mean Slress Effects With No Rela.atian
100
70 2:
e
~~
VI 8
<I
6

4 30
A36 Butt Weld (HAl)
KImox =3 I R-O 20
2 a r =+35 ksi

10
IIO~ 10'
N • Cycles
I

Mean stress relaxation behavior influence on fatigue

crack initiation life (A36 HAZ material, K f = 3, R = 0,
a r = +35 ksi (242 MPa) ).

Materials such as high-strength steels exhibit very little notch-

root plasticity; consequently, a os may be larger than a r- The re-
sults obtained using the model agree with the experimentally
observed behavior. The above chart shows the qualitative be-
havior of N[ predictions.

Source: F. V. Lawrence, "The Predicted Influence of Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, pIli
64 2-22. Low-Carbon Steel Tubes: Effect of Welding Technique

.,.--'

" '-
- - - - f----- _._ •..• _ -
1·0
----- '--'-
.-,
""
,
1-. - ' ,,""' ~,~'-. "
"
.
I\. ....
"
,
, .,~:,
f--------.--

0·8
~ " ','~ --
i'..'.. " ,~' ~~
~/ ~/ -
..,~ ~'/~ .,~

O·B '-'0 ~ ~
.-:~~ /.--~--
~ -----_.-
"-
I~'"
~
%-. ~/. ~ A_
V'~ '~r::-, '/8 ~.ij'/W
~., ~c ,-
0·4 ..,~
b-.,
~.
'~
~D
0·2

o 10
I I I I I I I I I I I I I I I I I I I I I I

Fatigue strength of welded tubes: A - unwelded or welded without filler metal;

B - helical welding (700 angle); C - longitudinal or helical welding (550 or 600
angle); D - helical welding (500 angle).

Source: R. V. Salkin, "Low Cycle Fatigue of Welded Structural Steels: A Material Manufacture and Design Approach," in Proceed-
ings of the Conference of Fatigue of Welded Structures, Vol 2, The Welding Institute, Abington Cambridge, 1971, P 193
 2-23. Low Carbon Steel: Effect of Applied 65
Anodic Currents in 3% NaCI

.. 120
E
~
~
100
>--
t--
~ 80
w
o
~ 60
w
IX:
IX:
~
040

20

OL-_ _- '_ _-'--------'-_'--.l-L....L--L..J'--_ _- '_ _-'--------'-_'--L-JL..l...-'LIlk<

10' 10' 10'
CYCLES TO FAILURE
Effect of applied anodic currents on the fatigue lives of low-carbon steel in
deaerated 3% NaCI solution. The corrosion rate ofthe steel in this solution is
virtually zero in the absence of applied currents. Note the independence of fa-
tigue life at currents greater than-: 40 /-LA/cm 2, the absence of an applied stress
effectand the reappearance of a fatigue limit at currents less than ~ 0.2 /-LA/cm-,

The effects of salt concentration and temperature on the fatigue behavior of steels
have been studied. Experiments performed on mild steel specimens in distilled water
and in various concentrations of potassium chloride have shown that solutions ranging
from 2 molal to 1/40 molal have virtually identical effects on corrosion-fatigue lives,
but that at concentrations below 1/40 molal, the effect approaches that of distilled
water, although corrosion rates increase in an almost linear manner with solution ion
concentration. A similar result has been reported for deaerated 3% NaCI solution in
which corrosion rates were controlled by applied anodic currents (see above chart).
These observations indicate that a critical corrosion rate is a necessity to initiate
corrosion-fatigue failures. Additionally, increasing over-all corrosion rates over a long
range of rates has little effect on corrosion-fatigue resistance.

Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys." in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 344
66 2-24. Low-Carbon Steel: Effect of pH in NaCI and NaOH

60 I I

55
/ I NORMAL SOLUTION

50

45

.,~ 40
Q
"'.H =12.1

~ 35
UJ
II:
t-
'" 30

25
pH :10

20
.5 5 8 10

The effect of pH on the fatigue behavior of low-carbon steel in

NaCI+NaOH.

The effect of stress frequency on corrosion fatigue has been studied by a number of investiga-
tors but is still not completely understood. For example, an early review of corrosion fatigue
noted that it is difficult to compare the corrosion-fatigue properties of metals exposed to like
environments because data reported are usually taken at different frequencies. In general, a
given time was found to produce more damage at a higher frequency, but a given number of
cycles was found to produce greater damage at low frequencies. For low-alloy steels in fresh
water, a frequency of 1450 cycles/min produced failure in 106 cycles or II Y2 hours, but at a
frequency of 5 cycles/min, failure occurred in 0.11 X 106 cycles, or 400 hours.
To date, the effect of pH of aqueous solutions on corrosion-fatigue behavior has not received
extensive study. A study of the effect of 0.1 N HCl on the fatigue life of steels showed greater
damage in this medium than in neutral potassium chloride solutions. Tests conducted in alkaline
media, at a pH above 12.1, showed that a fatigue limit is regained, this limit improving at still
higher pH values (above chart). These investigators suggested that corrosion fatigue is a result of
differential aeration cells, which produce pits in the metal surface, and that a high pH provides
diffusion barriers (ferrous hydroxide) to oxygen on the surface. Higher fatigue limits at high pH
are explained in terms of a "better and more perfect film barrier."

Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys." in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 346
 2-25. Low-Carbon Steel: Effect of Carburization 67
and Decarburization

CJ)A
CJ)
• ~•••• NON-DECARBED MATERIAL

~ B ...............~
t)
~
INTRINSIC'
, . , ...DECARBED
FOR DECARB • -COMPOSITE
MATERIAL

LIFE
Influence on fatigue SoN curve of soft surface
caused by decarburization.

Parts that were made from low-carbon steel, but have high-carbon sur-
faces resulting from carburizing, have special microstructural factors that
must be considered. From the carburizing process an intergranular oxide
network may develop. This oxide may be an alloy oxide which causes alloy
depletion in grain-boundary areas. As a rule, this condition is thought to
detract from fatigue properties. The two exceptions may be in combination
rolling and sliding contact fatigue, where the oxide network may enhance
low-cycle bending fatigue-somewhat the same as does decarburization. The
effect on high-cycle bending fatigue is deleterious, as is decarburization.
These concepts are shown schematically in the above chart.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979,P 80
68 2-26. A514B Steel: Effect of Various Gaseous Environments
on Fatigue Crack Propagation

r------;;;:;:;;;::::::j""". 18
16
0.4
14
2
AK = 29.7 MN/m 3/
0.3

0.2

0.1 4

The fatigue-crack propagation of ASTM A514B

steel in various gaseous environments.

The origin of the element (such as sulfur) Onthe surface could result from
its presence in the gas phase (such as for hydrogen sulfide). It could also
originate as an enriched sulfur layer associated with a propagating crack, as
would be the case for sulfur segregated to a grain boundary. Oxygen alone on
the surface tends to drive the hydrogen-dissociation reaction rates in the op-
posite direction from the sulfur. The above bar chart shows how a mixture of
environments can influence the fatigue-crack growth of an alloy when all the
loading factors are kept constant.
The main influence ofthe environment is to supply the active atoms to the
vicinity of the crack tip. Subsequent interaction with the crack allows the
degradation mechanism to take place. The next step in the environmental
interaction is the transport of the active species to the location in the vicinity
of the crack tip where the degradation mechanism takes place.

Source: H. L. Marcus. "Environmental Effects 11:Fatigue-Crack Growth in Metals and Alloys," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979. P 371
 2-27. Cast 1522 and 1541 Steels: Effect of Various 69
Surface Conditions

NONE 0.006 0.012 NONE 0.006 0.012

SHOT PEENING INTENSITY - C, ALMEN

The effect of shot peening, carburization, and decarburi-

zation on the endurance ratio of normalized and tem-
pered cast steel with cast surfaces. Plate bending fatigue
.specimens were used to secure these data.

Decarburization of the surface lowers fatigue resistance. This

effect, along with the beneficial effects of carburization and shot
peening on the endurance ratio of cast low alloy plate specimen
in bending, is shown in the above diagram. The nominally 1.2%
Mn steels with 0.22% C and 0.4 I% C, respectively, were normal-
ized and tempered to 78 and 95 ksi (538 and 65~ MPa) ultimate
tensile strength, respectively. The depth of decarburization
(0.05% C at the surface) was 0.06 in. (1.5 mm); that of carburiza-
tion (1.15% C at the surface) was 0.08 in. (2 mm).

Source: Steel Castings Handbook. 5th Edition. Peter F. Weiser. Ed.. Steel Founders' Society of America. Rocky River OH. 1980.
P 15-29
70 2-28. Cast A216 (Grade WCC) Steel: Fatigue Crack Growth Rate

20 40 60 80 100 200
3r---.---.--__,_~__,_,...,.,_r_--"

lTyS : 48 ksi (331 MPo) 6

2 TEST TEMP: 7soF ( 24°C)
TEST FREQUENCY: 600 cpm - 4
WOL TYPE SPECI MENS
UPPER SCATTER BAND w
-.I
( SLOPE n : 3 ) """ U
2 r
u
-,
E
I
10- 6
Z
- 8 "0
<,
Z 0
"0 6
<, "0
o 2 I
"0 4 w
I I-
W <l:
~ 10- 5
0::
0::
8 2 I
I I-
I- 6 ~
~ 0
o 10-7{.?
0::
0:: 4
{.?
8 ~
~ ~. u
'<l
0::
2
o. 6 <l:
0::
U
U

0
. 4

10- 6 '--_ _-'-_-'----'_L.-I-L.........--'-_ _---'

10 20 40 60 100 200
STRESS INTENSITY FACTOR RANGE,
~K, ksi ~
Fatigue crack growth rate as a function of ilK for
A216 (grade weC) cast steel.

The equation do I dN= CoI1K" is sometimes referred to as the

Paris law and predicts a linear plot of log dol dN versus log 11K
with slope n. This is observed for a wide variety of materials and
is illustrated in the above diagram for an ASTM-A2l6, Grade
WCC cast steel. Some materials show a significant influence of
the mean load or Klevel on fatigue crack growth rates. The ratio
ofKmin to K max is used to express the mean load conditions.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
p4--16
 3-1. AISI 1030 (Cast) Compared With AISI 1020 (Wrought) 71

(\J
<,
I.O....--r--..,.....--r---r--....... --r--...,
I-
III
<I
'f "''- /WROUGHT SAE 1020
I
0.1 ....
ILl
C ,
C 1 ", /CAST SAE 1030
~
I-
:J 0.01 PLASTIC
/O,W
O\' '
Q. 0,'
~ '00"::0."
<t 1)'0'
~O
Z
<i
0:
0.001 "
t;
0.0001 ~_L--..l....::-~-..l...:-....l..:~....L.:=--~
7
100 10 102 106 10
REVERSALS TO FAILURE - 2N f

Low-cycle strain-control fatigue behavior of car-

bon steel.

A number of techniques are available for computing the low-

cycle fatigue life, although a straightforward approach is simply
to compute the fatigue life from the expected cyclic plastic strain
amplitude in service. Errors in computing or estimating 6.Ep
produce a smaller change in the computed cyclic life than similar
errors in the elastic strain range. Note that there is a large differ-
ence in slopes "c" and "b" in the above diagram. Plastic strain
ranges may be computed using sophisticated finite element
techniques, estimated from simple approximations such as
Neuber's rule or experimentally measured in component or
model tests.

Source: Steel Castings Handbook, 51hEdition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,P
4-13
72 3-2. AISI 1035: Effect of Gas and Salt Bath Nitriding

50

A1MQSPH£RE
NITRIDED
40

SALT BATH PRQCT_'S

0
§
QUENCIU:D AND TEl1PERID
lC AT l050 F (565 C)

'0
~

HUKBER OF CYCLES

Torsional fatigue strength of AISI 1035 steel-stress vs number of cycles for

completely reversing torsional fatigue, featuring the effects of gaseous atmo-
sphere and salt bath nitriding on fatigue strength.

Source: J. A. Riopelle. "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals. Metals
Park OH, 1977, P 286
 3-3. AISI 1040: Cast vs Wrought 73

TENSILE YIELD
STRENGTH STRENGTH ELONG. HARDNESS
ksi MPo ksi MPo % BHN
CAST 94 (6481 56 (386) 25 187
WROUGHT 90(62~ 56 (3B6) 27 170

350 ~
en 50 ~
.><: ,
en
J, 45 WROUGHT } NO 300 f3
f3 0:
g: 40 CAST NOTCH
~
en 250 ~
~ 35 ::::>
::::> ~
~
5(30 200 X
<t
<t NOTCHED
~
~ 25
150
5 6 7
10 10 10
CYCLES TO FAILURE

Fatigue characteristics (S-N curves) for cast and wrought

1040 steel in the normalized and tempered condition,
both notched and unnotched. R. R. Moore rotating beam
tests, K, = 2.2.

Cast steel suffers less degradation offatigue properties due to

notches than equivalent wrought steel. When the ideallabora-
tory test conditions are replaced with more realistic service con-
ditions, the cast steel shows much less notch sensitivity to varia-
tions in the values of the test parameters than wrought steel.
Under the ideal laboratory test conditions and test preparation
(uniform section size, polished and honed surfaces, etc.), the en-
durance limit of wrought steel is higher. The same fatigue char-
acteristics as those of cast steel, however, are obtained when a
notch is introduced, or when standard lathe-turned surfaces are
employed in the rotating beam bending fatigue test. These ef-
fects are illustrated above.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 15-10
74 3-4. AISI 1045: Relationship of Hardness and Strain-Life Behavior

0.1
.
C\l
.....
<l
SAE 1045

or
"0

.~
Q.
E 0.01
C3:
>-;~- _ _ _ 1l911HB

.
c:
1:1
+0
(/)
'.:::--=----_410
..... __ - - - 3 3 0
- - - - 280
2211
0.001
I 10 102 103 104 lOll 6
10 107
Reversals la Failure, 2N
f

Strain-life behavior of medium-carbon steel as a

function of hardness.

Strain-life curves at various hardnesses are presented in the

diagram above to demonstrate the range of properties attain-
able by tempering. Such information, used in conjunction with
life-prediction models, provides guidelines for optimizing mate-
rial processing for specific situations.

Source: R. W. Landgraf. "Control of Fatigue Resistance Through Microstructure-Ferrous Alloys," in Fatigue and Microstruc-
ture, American Society for Metals, Metals Park OH. 1979.P 458
 3-5. AISI 1141: Effect of Gas Nitriding 75

ATHOSI'HERE
NIUIDED

ATHOSI'HERE NITRIDED niDI

GROUND to IlDlOVE COHPOUND
QU!NCBED AND TDO'UlD I.\YEII
AT 1050 r (565 C)
40

NlHIEI OF C'l'CLlS

S-N curves for 1141 steel-gaseous-atmosphere nitrided vs not nitrided

(quenched and tempered only)-showing stress vs number of cycles for
completely reversing torsional fatigue.

Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 287
76 3-6. Medium-Carbon Steels: Interrelationship of Hardness,
Strain Life and Fatigue Life

1.0~

\. \.
1\ \.
\\ \
0.1
1\\'\\
600
~500~00
.
en
" " ,,30~ 200
~
c; -, "'-.." .'\. \ Hardness, HB

<; ~ t\
.~

t:
~ 0.01
S
'0
n;
.
s:

0
c: ~ ~ 400
600
.

0.001
<,
<, 300-
-200-
-

0.0001
1 10 1M 1~ 1~ 1~
Stress reversals to failure

Predicted plots of strain versus fatigue life for typical medium-

carbon steels at the hardness levels indicated above.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH. 1978, P 673
 3-7. Medium-Carbon Steel: Effect of Fillet Radii 77
,---

60 <-,
t-- --~~p~~lf:o
~ ~
c
~/lf:
1ft'! d
~ ~O'47"
on r--,
o
~ ~

gj 50 .
o
x
l-
I
oil
oil
III
0:
~
li;
~(l'O'I7 DIcJ.:2
~40
z
~
I~
I"""
- d. r 'Y-. I
Q ...
~S- D/d: 2
{)r
~

~
r- u
o
Z d.·2·/3"~
o d (i·Q·If;
O/cl.ys 0+
----

30 I I I II
300,000 1.000,000 10.000,000 50,000.000
I'lUMBlR OF C'{CLES ,0 FAILURE (LOG. SCALE)

~A1 UNMODIFIED
~&j RE-EI1TRAN, FILLET

EXTERNAL 5TRE55- REl.IEVER

~4 5EPARATE COLLAR

The fillet radius at a change in diameter should be made as

great as possible. This cannot always be done; e.g., if the inner
race of a rolling bearing must abut against a shoulder formed by
the change in diameter. In such cases the stress-raising effect can
be moderated very considerably by adopting one of the expe-
dients illustrated above.

Source: G. A. Cottell, "Some Common Stress Raisers in Engineering Parts," in Failure Analysis: The British Engine Technical
Reports, F. R. Hutchings and Paul Unterweiser, Eds., American Society for Metals, Metals Park OH, 1981, P 108
78 3-8. Medium-Carbon Steel: Effect of Keyway Design

5LEO-RUNNl:.R KEYWAY PROFILED Kf..YWAY

70,000

PROFILED ...EV.......'(

MEDIUM-CARBON S,.EEL
(NORMALISE-D)
I I I ~ I

o
200.000 1.000.000 10,000.000
NUMBER. DF C'(CLES,.O FI'.\LURE (LOG.StALE)

Keyways are severe stress raisers from which fatigue cracks

are very liable to develop. Where bending stresses are predomi-
nant the cracks usually run transversely in the region of the key-
way end, but where torsional stresses predominate they origi-
nate at the root at one side and may cause a portion of the shaft
to peel off or they may lie diagonally across the bottom. Effects
of various keyway designs on fatigue life are shown above.

Source: G. A. Cottell, "Some Common Stress Raisers in Engineering Parts," in Failure Analysis: The British Engine Technical
Reports. F. R. Hutchings and Paul Unterweiser. Eds., American Society for Metals, Metals Park OH. 1981, p 109
 3-9. Medium-Carbon Steel: Effect of Residual Stresses 79

+600

+400

+200
NEAR-SURFACE
RESIDUAL
O~-----------~-------
STRESS
MPa (= +3000 po IN/IN
-200
TESTS STOPPED AT •
-400
10 7 CYCLES MAX. ••

-600L---------....,...L,.-----------,J.
0.1 1.0 10

UFE CYCLES Ie 10 6

Fatigue life relationship to near-surface residual stress.

Fully reversed fatigue tests on smooth bar specimens in me-

dium carbon steels fully hardened show, as expected, that fa-
tigue life increases directly with surface and near-surface resid-
ual compressive stress (see above chart). Residual stress mea-
surements are usually made in the direction of the applied stress.
The achievement of high residual compressive stress in a part
requires a careful balance of the factors which affect this prop-
erty and often involves a number of trade-offs which vary with
the application.

Source: J. Alan Burnett. "Prediction of Stresses Generated During the Heat Treating of Case Carburized Parts," in Residual Stress
for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 44
80 3-10. Medium-Carbon Cast Steel: Effect of Changes in
Residual Stress

25

211
NDR"""L.IZE ~N~

.
.>:
15 :a---~
I
.4'~ <,
~
e;
-, -, ~HOT BLASi'
i- III
lJl lil....
-' '4 <,
:§ A ' "a...
~
Ul
5 NORH"LIZE - TEMP~ ~
ILl

'" ~ti
II

-5
4
111

CYCLES

Residual stress at completion of testing,

III

-411 .1.--+- +--_ _--< --+ --+- -+- +--'

Ill"
CYCLES

Change in residual stress with cycles at constant applied

stress,

The upper chart shows residual stresses existing on the completion of individual tests. The
similarity to S-N curves is apparent with the exception that the curve for normalized bars
(R 1 =+22 ksi) is inverted. Since the initial residual stress was known, there was a question on the
manner in which the residual stress changed during the progress oftesting. To explore this point,
two shot blasted bars were tested with applied stress levels of 40 and 55 ksi. The test on each
specimen was interrupted periodically to measure the residual stress at that time. The results are
shown in the lower chart. It is apparent that the change in residual stress is proportional to the
number of cycles when the latter is represented on a logarithmic basis. The lower chart also
points to the fact that the rate of change in residual stress is dependent on the level of applied
stress. Since the initial and final residual stress values were known for all bars, the slope for each
line could be determined.

Source: P. J. Neff. "A Quantitative Evaluation of Surface Residual Stress and Its Relation to Fatigue Performance," in Residual
Stress for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, /981, pp 127-128
 4-1. Medium-Carbon Alloy Steels, Five Grades: 83
Effect of Martensite Content

100
100

650

.
c,
::;;;;
600
~ <, 01340
.4042
90 ·Vi
~
~.

~
]
_4340 :~
I--.I'--- 05140

"'"
.~ 550
6.
6.
6. 80840

'"I'--. ---
u.

6.
n
-
500
I--.
70
All specimens 36 HRC
450

60
400
100 80 60 40 20
Martensite. %

For specimens having comparable strength levels, resistance

to fatigue depends somewhat on the microstructure. A tempered
martensite structure provides the highest fatigue limit. How-
ever, if the structure as-quenched is not fully martensitic, the
fatigue limit will be lower (see graph above). Pearlitic structures,
particularly those with coarse pearlite, have poor resistance to
fatigue. S- N curves for pearlitic and spheroidized structures in a
eutectoid steel are provided in chart 4-40 (p 122).

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 676
84 4-2. Medium-Carbon Alloy Steels, Six Grades:
Hardness vs Endurance Limit

160 H-ll
1----1I---t---I----1t-"'1 Aus t ernpered
150
140
Vi
a.
130
0
0 120
0
110
E 100
--'
Q)
90
u
c
~ • - SAE 4063
:J
"0
c
6- SAE 5150
w .a._ SAE 4052
0 - SAE 4140
50 G - SAE 4340
Q - SAE 2340

20 30 40 50 60
Rockwell "c" Hordness

Relation of hardness and fatigue strength for sev-

eral steels.

The above chart and other data can be used to show the importance of
limiting the system to low and intermediate hardnesses as well as to point out
the importance of residual stress in fatigue. These data from Garwood, Zur-
burg and Erickson show a very tight linear relation up to about 40 HRC.
Above that hardness, the relation deviates from linearity, seemingly depend-
ing on carbon content. Carbon, however, is in an intermediate role here,
because it affects temperability. Because response to tempering is dependent
on carbon and alloy levels, it was necessary for samples of different grades to
be tempered at different temperatures to achieve the same hardness; conse-
quently, a variety of residual-stress conditions resulted. The tempering
temperatures were necessarily sufficiently high to obtain 40 HRC; the resid-
ual stresses were reduced to a very low level, making all samples similar in
that usually the tensile strength for small sections decreases with increasing
section size and I or decreasing hardenability to compressive values. The se-
quence of transformation from surface to center, together with the tempera-
ture gradients, governs the outcome.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 73
 4-3. Medium-Carbon Alloy Steels: Effect of Specimen Orientation 85

Avg
tensile
No. of strength Hardness,
BOO
Steel tests(a) MPa ksl HRC

100
Longitudinal Tests
s:
::;: .>! 4027 11 1179 171 37 to 39
600
.- BO 4063 12 1682244 47 to 48
.
~ ~
. 4032 11 1627236 46 to 48
5, 5,
.~ 400 60 "g
u, u, Transverse Tests
40 4027 10 1130 164 34 to 39.5
200 4063 9 1682 244 47 to 48.5
4140 X4340 4027 4063 4032
Hardness, HRC 30 32 44 46 4B
4032 10 1254 182 47.5 to 48.5
(a) Number of fatigue specimens. For 4140 steel, 50
longitudinal and 50 transverse specimens were
tested; for 4340 steel, 10 longitudinal and 10 trans-
verse.

It must always be considered that in rolled steels fatigue be-

havior is affected significantly by specimen orientation. Shown
above is the effect of orientation relative to fiber axis resulting
from hot working on the fatigue limit of low-alloy steels.
Through hardened and tempered specimens, 6.3 mm (0.250 in.)
in diameter, were taken from production billets. Specimens for
each grade were from the same heat of steel, but the tensile and
fatigue specimens were heat trated separately, thus accounting
for one discrepancy in hardness readings between the chart and
the tabulation above. Fatigue limit is for 100 million cycles.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 677
86 4-4. 4027 Steel: Carburized vs Uncarburized

Bending Fatigue
0.02 • 4027
Carburlzed
b. O.OOS"Ca..
(\J
..... 0.01
D O.OIS" Ca.e

<l'"
o 0.03S"Ca..

,,;
"0
~ 0.005
ji
E
~
c
.~
.. 0.002
Ul

0.001 L...-_--'-_----'--;.----'-....-_'-;;-_........_--'-;;_--'
1 10 102 103 104 lOS
Reversals 10 Follure,2N f

Bending-fatigue results for uncarburized and car-

burized 4027 steel.

Bending-fatigue results supporting the validity of the effect of

carburizing are presented in the above curves. An uncarburized
baseline curve is compared with curves for three case thick-
nesses. As predicted, all carburized specimens show inferior
low-cycle resistance. At longer lives, the thinnest case offers
some improvement but tends toward the baseline as a result of
subsurface failure initiation. The thickest case, which shows the
greatest life improvement and has been found to exhibit surface
failure initiation, seems close to optimum.

Source: R. W. Landgraf, "Control of Fatigue Resistance Through Microstructure-Ferrous Alloys," in Fatigue and Microstruc-
ture, American Society for Metals, Metals Park OR, 1979, P 463
 4-5. 4120 Steel: Effect of Surface Treatment in 87
Hydrogen Environment

200°6
1600 25 Cr '"'A 4 - annealed
.4~.= 1.2%. H1 (10MPo)
600 £ =aOO3 sec:'
Cl bose metal - mech. pot.
m bose metot - chem. pol.
prior to surface treatmMt

surtoce treatments

Effect of surface treatment on fatigue life in hy-

drogen environment for a O.23C-O.98Cr-O.22Mo
steel.

In the above bar chart, the effect of surface treatment on fa-

tigue life is summarized. The base metal was mechanically pol-
ished before surface treatment. For comparison, pot galvaniz-
ing and Ni-plating have been performed after mechanical as well
as after chemical polishing of the base metal. The results after
chemical polishing are given above in the form of dashed
columns.
The galvanizing such as Ni- and ZN-plating is by no means an
appropriate method to increase the fatigue life in hydrogen in
spite of the reduced surface roughness and protecting effect.
This is because the galvanizing produces a relatively high tensile
residual stress and the deposits possess generally poor ductility.

Source: Kyong-Tschong Rie and Werner Kohler, "Improvement of the Resistance of Metals to Cyclic Plastic Loading in High
Pressure Hydrogen Environment," in Current Solutions to Hydrogen Problems in Steels, C. G. Interrante and G. M. Presouyre,
Eds., American Society for Metals, Metals Park OH. 1982, P 380
88 4-6. 4120 Steel: Effect of Surface Treatment in
Hydrogen Environment

500'.-----~---~-----,-------,
,
hammered
I
/ alvanized
I

e- 300
III
I
~
III
I Ia/hed
200 I - -1 -
Ni- la/ed '25 erMa, -annealed
<a=!o.6%
t =0.003 s-'
H2 (10 MPaJ

number af cycles N

Effect of surface treatment on cyclic now curve in

hydrogen environment.

The above graph shows the cyclic strain hardening and soften-
ing curves for different surface treatments. It can be seen that the
fatigue behavior in hydrogen environment can be improved by
some surface treatments.

Source: Kyong-Tschong Rie and Werner Kohler, "Improvement of the Resistance of Metals to Cyclic Plastic Loading in High
Pressure Hydrogen Environment," in Current Solutions to Hydrogen Problems in Steels, C. G. Interrante and G. M. Pressouyre,
Eds., American Society for Metals, Metals Park OH, 1982, P 379
 4-7. 4120 Steel: Effect of Various Surface Treatments on Fatigue 89
Characteristics in Air vs Hydrogen

.. ,

.
2.0
..e
"'l
''';'. .....
..
0.
,:~
1.0
~ 25 erMa' -annealed -+-
i: =0.003 sec-'
S -H1,IOMPo)
g --air
<II . "," chem. polished .........-+-.J-
0.'
0,_ pol galvanized
~
s 0•• Ni-ptatea
0,+ mech. polished
0,2
60 100 200 500 1000 2000 5000 MXJ
cntxa! nurrtJer of cycles Na

Fatigue life curve for various surface treatments of steel in

hydrogen environment and in air. Steel contained 0.23 C,
0.98 Cr and 0.22 Mo (4120).

Source: Kyong-Tschong Rie and Werner Kohler, "Improvement of the Resistance of Metals to Cyclic Plastic Loading in High
Pressure Hydrogen Environment," in Current Solutions to Hydrogen Problems in Steels, C. G. Interrante and G. M. Pressouyre,
Eds., American Society for Metals, Metals Park OH, 1982, p 380
90 4-8. 4130 Steel: Fatigue Crack Growth Rate vs Temperature in
Hydrogen

10- 3
I 0 O"ty = 1330 MN m- 2 I O"tu = 1660 MN m- 2

o O"ty = 1190 MN m- 2 I O"tu =1310 MN m-

2

K =40 MN m- 3 / 2

.,
U
<Il
V>
E

I I
3.4 3.8 4.2 4.6 5.0
liT, oK-I

Crack growth rate versus temperature in hydrogen

gas, for 4130 steel with yield strengths of 1330 and
o2
1190 MN m •

The striking characteristic of hydrogen which sets it apart from other

causes of embrittlement is its large diffusivity. Although the diffusivity of
hydrogen does vary significantly among metals and alloys, it is nevertheless
always several orders of magnitude larger than the diffusivities of other spe-
cies. Consequently, hydrogen transport is a prominent feature of discussions
of hydrogen-induced crack growth kinetics, and ofthe unique strain rate and
temperature dependence of hydrogen embrittlement.
Nelson and Williams reported the first complete investigation ofthe kinet-
ics of crack growth in high strength steel exposed to hydrogen gas (see graph
above).

Source: Herbert H. Johnson, "Keynote Lecture: Overview on Hydrogen Degradation Phenomena," in Hydrogen Embrittlement
and Stress Corrosion Cracking, R. Gibala and R. F. Hehemann, Eds., American Society for Metals, Metals Park OH, 1984, P 17
 4-9. 4135 and 4140 Steels: Cast vs Wrought 91

85

80
UNNOTCHED NOTCHED
4135 CAST STEEL 6. A
••
75
4140 (LONGITUDIAL) }NO.32 0 100

4140 (TRANSVERSE) 0
70
D~-&
65 °o~............... o

60
~"
"<, 80--0~
06..,
0-- 80
<Il
.... 6,6.. 0
.>t:
o ..... o,
I
55
""8<,6 . . - 6 .__ ~
(f)
(f) O-B~
UJ (f)
a: 50
••• (f)

\~\A
I- UJ
(f) a:
I-
(f)
45 60

\\.\.
.
40

35
,,~& .........
~~ ·-A-A_
30 ~.-- 40

25
'.~-.-
104 105 106 107
CYCLES TO FAILURE

SoN curves of a normalized and tempered AISI 4140

wrought steel in the longitudinal and transverse direction
and cast 4135 steel normalized and tempered. Tensile
strength for wrought steel: longitudinal, 110.0 ksi (758
MPa); transverse, 110.7 ksi (763 MPa); cast steel: 112.7
ksi (770 MPa).

In general, if the longitudinal and transverse ductility, impact,

or fatigue property values of rolled steel are averaged, they will
be about the same as properties of cast steel. One example of this
is shown in the S-N curves presented above. For these, a 4140
rolled steel was tested in fatigue in the longitudinal and trans-
verse position and compared with a similar Cr-Mo cast steel.

Source: Steel Castings Handbook, 5th Edition. Peter F. Weiser, Ed.. Steel Founders' Society of America, Rocky River OH, 1980.
P 3-16
92 4-10. 4135 and 4140 Steels: Cast vs Wrought

STRENGTH
TENSILE YIELD ELONG
ksi (MPo) ksi (MPol
% BHN
CAST 4135 113 (779) B7 (560) 43 223
WROUGHT 4140-L 110 (758) 80 (552) 61 217
-T III (7651 81 (558) 30 217

80 WROUGHT _
75 CAST
500
~ 70 ........... ' ......
b
I

<l
65
UNNOTCHED'
fh . . . . . . . . . .
LONGITUDINAL
------~
o
W 60 -,
400 11.
o <, <, ,~RANSVERSE ~
~ 55 ,1__ -. I
::i
11. 50
~
« 45
\
\
\
--- lr
en
en
w
en \ 300 I-
en en
""
I ',
w 40
lr
I- NOTCHED
en 35
LONGITUDINAL ''-...... =*
30 AND TRANSVERSE - ----
NO FAILURE 200

106
CYCLES TO FAI LURE

Fatigue characteristics (S-N curves) for cast and

wrought 4100 series steels, quenched and tem-
pered to the same hardness, both notched and
unnotched.

The number of cycles to failure ofa structure subjected to the

above stress history can be expressed in terms of the SoN curve
shown above. The fatigue life increases as the cyclic stress ampli-
tude decreases. For ferrous alloys a true endurance or fatigue
limit is reached below which fatigue failure is not observed.
The data presented in the above S-N curves illustrate several
important points. First, a fatigue limit is evident. That is, below
a certain cyclic stress amplitude, fatigue failure will not occur for
any arbitrarily large number of cycles. Secondly, while the fa-
tigue properties of cast steel are lower than those obtained with
the wrought steel, it has less anisotropy. And, finally, the pres-
ence of a notch equalizes the fatigue properties of cast and
wrought steels.
The above data also illustrate that the fatigue limit of notched
test specimens is substantially below that of unnotched samples
when the fatigue limit is computed on the basis of nominal
stress.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 4-8
 4-11. 4140, 4053 and 4063 Steels: Effect of Carbon Content and 93
Hardness

1000,-------------,--------,--------.--------;
140
o to 2 micro-in.finish
900.1--------/-------t--------t 130

120
~ 8001--------/-------t--,
Q. ·in
::;; sc
~.
110 .~
·E ~
~
:J
7001-------j-----;
100
.
:J

.~ .~
u,
U.
90
6001-------:

80

70

60
30 40 50 60 70
Hardness, HAC

Effect of hardness and carbon level on fatigue

limit of alloy steels.

As shown above, when steels are hardened to 45 HRC or

higher an increase in carbon content can increase fatigue limit.
Although other alloying elements may be required in order to
attain desired hardenability, they have little effect on fatigue
behavior.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 676
94 4-12. 4140 Steel: Effect of Direction on Fatigue Crack Propagation

10' 10-'

L T

.....
Q)
.....
Q)
~ 10-' ~ 10:'
U U
>-
U
>-
U
..... .....
E E
....
E
....E
C121(j' cI2 Ill"' -
'tJ'tJ 'tJ'tJ

10· 1-1-_ _' - - _ - - - ' - _ - ' - - ' - - ' - - _ - - - - ' - ' 10· L.L_ _-'--_---J'---'--'--'-_--,-lJ
10 20 304050 100 5 100

AK (MPaV'ffi) AK(MPaV'ffi)
a b

Fatigue crack propagation in an AISI 4140 steel: (a) longitudinal di-

rection (parallel to rolling direction); (b) transverse direction (perpen-
dicular to rolling direction).

The Paris power law, which describes the crack propagation rate in stage II
for a series of metals, is very useful because of its extreme simplicity. For
example, it has been observed experimentally that data points in the form of
log (do / dN) versus log ~K for a given material (constant metallurgical struc-
ture) from three different samples-edge crack in a compact tension sample,
through-thickness central crack in a plate, and plate containing a partially
through-thickness crack -all fall on the same line. Also, there is experimental
evidence that shows that the stress level by itself does not influence the fatigue
crack growth rate for stress levels below the general yielding. Thus, it can be
considered that the parameter ~K describes uniquely the crack growth rates
for many engineering applications. However, the structure of material can
influence fatigue crack growth rates drastically; the value of m can change a
lot. The above charts illustrate the directionality in the fatigue crack propa-
gation rate in an AISI 4140 steel. The exponent m has a much higher value in
the transverse direction than in the longitudinal (rolling) direction, due to the
presence of elongated inclusions.

Source: Marc Andre Meyers and Krishan Kumar Chawla, Mechanical Metallurgy: Principles and Applications, Prentice-Hall, Inc.,
Englewood Cliffs NJ, 1984, p 714
 4-13. 4140 Steel: Effect of Cathodic Polarization 95

Carras.....
LLi-O.4 Potential . - . _ " - - -
I .~
(fj •
(/) -0.6
•
~
o
>-0.8
.r
<{

~ -1.0 •
w
I-
•
o
0... -1.2

10 5 106
CYCLES TO FAILURE

The effect of cathodic polarization on the fatigue behavior of

4140 steel (heat treated to HRC 52) in 3% NaCI solution at a
stress level below the fatigue limit in air is shown above. The use
of cathodic protection to prevent corrosion fatigue of steels de-
pends sensitively on the hardness of the steel. For example, ca-
thodic protection of a 4140 steel was shown to be feasible for
hardness values of Rockwell C 40. At higher hardness values, an
improvement in fatigue resistance is observed for moderate ca-
thodic potentials, but complete protection is not possible. At po-
tentials large enough to inhibit corrosion fatigue for softer
steels, a decrease in fatigue resistance is observed, presumably
due to hydrogen embrittlement (note above chart).

Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OR, 1979, P 360
96 4-14. Cast 4330 Steel: Effects of Various Surface Conditions

SURFACE ROUGHNESS- RMS - rnrn- 10- 2

.5 1.5 2 2.5 3

FUL L Y MACHINED
SILICA PRIME INV.

o /
f===:J~
,
ZIRCON INV.
SILICA ALUMINA INV
.... 0.2 C 6 C 16 WASH
<l: ZIRCON GR.S.
0:: CRYOLITE HIGH

L,\
~
w
COPE a
""""';;:::::====~~:.:;:=~:::..CHROM ITE

DRAG MACHINED
WASH
GR. S.
PROPR. WASH
_ _ _ _==--========,JS~IL~ICAGR.

200 400 600 800 1000

SURFACE ROUGHNESS - RMS- MICROINCHES·

Relationship between surface roughness and endurance ratio (endur-

ance limit divided by tensile strength) of quenched and tempered cast
4330 steel rUTS = 165-185 ksi (1138-1276 MPa)]. Fully reversed plate
bending tests. .

Plate bending tests for quenched and tempered low alloy cast
4330 steel indicate that investment cast surfaces, or conven-
tional castings produced with special mold washes, performed
equal to, or better than, fully machined and polished plate spec-
imens. The data also suggest a tapering off of the surface effects
on the endurance ratio at 600 or more RMS surface roughness
as indicated in the above diagram.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America. Rocky River OH, 1980,
P 15-29
 4-15. 4340 Steel: Scatter of Fatigue Limit Data 97

Tensile strength, ksi

12 160 200 240 280
800 iir=-
0- - - - ---'-T-'------ - - ----,rr-- - - - rr-r -- - - .,---,
110

7001-------1---------11--------1--------:::J 100
50% survival
e:;; .""'"
90
Ii 600 f-------t--------::;;;Io--"""'-----t------=90% ~
~ 80 c
1;:
'"
c 99% '" .~
.~ 500 1----.....~l=____=,....",~~~k:;;o>'-=::::::=:1f===-----l 70
E
~ ~
~ ~

400 f - - - - - - - - ' - - - t - - - - - + - - - - - - t - - - - - - - i 60
Approximately 1000 specimens. 1 heat
50
300 '-- L- -----''----- ~~ _:_:_'

800 1100 2000

Interrelationships of alternating stress, tensile strength and ex-

pected percent survival for heat treated 4340 steel.

These data show survival after I a million cycles of AISI-SAE

4340 steel with tensile strengths of 995, 1320, and 1840 MPa
(144, 191, and 267 ksi). Rotating-beam fatigue specimens tested
at 10 000 to II 000 rpm. Coefficients of variation range from
0.17 to 0.20. From these data it is evident that scatter increases
as strength level is increased.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 678
98 4-16. 4340 Steel: Strength vs Fatigue Life

'.

0', = 0, = 174 ksi

/ ' F..tigue strength eOr'ie;ent
0. - 0',{2N,)b - 17412N,ro.o9
/
I
Fatigue strength
exponent = slope = b = -0.09
/
-

Reversals to failure. 2N f

Typical data for strength versus fatigue life for annealed 4340 steel.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 672
 4-17. 4340 Steel: Total Strain vs Fatigue Life 99
100, - - - -....- - - - - - , , - - - -....------,-----r-------,------,

10-11----+---"O:-"'.--I-----r- 2
~ = l1€p +
2
~
E
= 0 5Bl2N r O.57 + 0 0062(2N r O.09
. , . ,

10-2 1 - - - - - + - - - - 1 - - - - - - - 1 " - " ".......:----1----+-----+------1

10-4l::- --'-:- ---lL::- -'-:: -'--:-- -'-::-- -..L:--~"___J

100

Typical data for total strain versus fatigue life for an-
nealed 4340 steel.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 672
100 4-18. 4340 Steel: Stress Amplitude vs Number of Reversals

4
10

f
( V = Uf =1200 MPa
C
Q.
3
~ 10
C
b
~
0 2
10
~IN
1
10

Stress amplitude (!:J.aj2) versus number of reversals (2 N J) for AISI 4340 steel.

It is convenient to consider separately the elastic and the plastic components of strain. The elastic compo-
nent can be readily described by means of a relation between the true stress amplitude and the number of
reversals (i.e., twice the number of cycles):

~E
2
a (a'J)
_e_=_o_= _ _ (2N)b
E E f

where ~Ee12 is the elastic strain amplitude, a o the true stress amplitude, aj-the fatigue strength coefficient
(equal to stress intercept at 2NJ = 1), NJthe number of cycles to failure, and b the fatigue strength exponent.
This relation is an empirical representation of the S- N curve above the fatigue limit. The above chart shows
an application of this relation to SAE 4340 steel. It was observed that fatigue life increased with decreasing
b. Morrow, based on energy considerations, showed that the fatigue strength exponent is given by:
b=---
n'
1 + 5n'
where n' is the cyclic hardening coefficient.
Thus, the fatigue life under elastic cyclic conditions (whether stress- or strain-controlled) increases with a
reduction in n'. Of course, the higher the material coefficient aj, the better it is for fatigue. There is evidence
that ajis approximately equal to aJ' the monotonic fracture strength.
The plastic strain component is better described by the Manson-Coffin relation:
~E
-p-=
2
Ej( 2NJ r
where ~Ep12 is the plastic strain amplitude, Ej is the ductility coefficient in fatigue and is equal to strain
intercept at 2NJ = 1, 2NJ is the number of reversals to failure, and c is the ductility exponent in fatigue.

Source: Marc Andre Meyers and Krishan Kumar Chawla, Mechanical Metallurgy: Principles and Applications. Prentice-Hall, Inc.,
Englewood Cliffs NJ, 1984. p 697
 4-19. 4340 Steel: Effect of Periodic Overstrain 101

2000

250

1500
e
::;;
200
.>l
1200 1J,
"c
'" c
~ ~
~ o No overstrain or single over- 150 ~
e 1000 strain at beginning of test
0
0 0 e
Vi Vi
• Periodicoverstrain
800

100

600
2 3 7
10 10 104 105 106 10
Number of cycles to failure

Overstrain superimposed on constant strain may have a sig-

nificant effect on fatigue life. Shown above is the effect of peri-
odic large strain cycles on fatigue life of AISI-SAE 4340 steel
hardened and tempered to a yield strength of 1100 MPa (160
ksi).

Source: Metals Handbook. 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH. 1978, P 681
102 4-20. 4340 Steel: Estimation of Constant Life

Minimum SIren, ksi

600 --400 ~200 200 .00 600 800 1000 1200 1'00
Minimum Sites'. MPa

Potter has described a method for approximating a constant-lifetime fatigue dia-

gram for unnotched specimens. Using this method, a series of points corresponding to
different lifetimes are calculated and plotted along the diagonal line on the left side
(R = -1). Each of these points is connected by a straight line to the point of the other
diagonal (R = 1.0) that corresponds to the ultimate tensile strength. A comparison
between the estimated constant-lifetime diagram and the experimentally determined
diagram is given in the above illustration. Here is presented a comparison between a
calculated constant-life fatigue diagram (solid lines) and experimentally determined
data (dashed lines). The calculated lines correspond well with the experimental lines.
Generally, the predicted lines represent lower stresses than the actual data. Estimating
fatigue parameters from the Brinell hardness number provides more conservative es-
timates. These results are only approximations, and the methods may not apply for
every material.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 681
 4-21. 4340 Steel: Effect of Strength Level on Constant- Life Behavior 103

Minimum stress. ksi

1.0

200

.
c,
:;;;
1000 150
"'"
~
E
800 ~
E
"E E
'x
'x "1""""') 100
~ 600
O't;..; 600 ~
:9~
"""~~'1;1.()
400
.. 10 6 cycles lifetime 50

200

o'----_---'-__---'--__--'--__-'---__ _ __''"____ ~ ____<_ ____L_ ____'___ _ ___'____ _ _~ L __ _ _ '_ _____' 0

-1200 -1000 "'1100 -tiOO -400 -200 200 400 600 800 1000 1200 1400 1600
Minimumstress, MPa

Constant-lifetime fatigue diagram for AISI-SAE 4340 alloy steel (bar), hardened and
tempered to tensile strength levels of 860 MPa (125 ksi), 1035 MPa (150 ksi), 1380 MPa
(200 ksi) and 1790 MPa (260 ksi), All lines represent fatigue lifetimes of one million
cycles.

It may be noted that lives of the specimens at the three higher

strength levels are about the same; the scatter in data is at least as
great as any real differences in fatigue life among specimens.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Jronsand Steels, American Society for Metals, Metals
Park OH, 1978, P 669
104 4-22. 4340 Steel: Notched vs Unnotched Specimens

Minimum stress,ksl
o

160

~
~ 800
~
1;; ~
,
E 100 E
E ~
"~ 600 "w
~ :!E

400
60

200

O'--_ _....L- ' -_ _---'- ...L-_ _- ' " ' " -_ _--'---_ _- - ' ' - -_ _....L- -'--_ _- ' - - ' -_ _-'0
-1000 -een -il00 -400 -200 200 400 600 800 1000 1200 1400
Minimum stress.MPa

Constant-lifetime fatigue diagram for AISI-SAE 4340alloy steel (bar), hardened and
tempered to a tensile strength of 1035 MPa (150 ksi), Solid lines represent data ob-
tained from unnotched specimens; dashed lines represent data from specimens having
notches with K, = 3.3.

Source: Metals Handbook, 9th Edition. Volume I, Properties and Selection: Irons and Steels, American Society [or Metals, Metals
Park OH. 1978, p 667
 4-23. 4340 Steel: Effect of Decarburization 105

1800
250

1500
0 200

If. 1200
:;:
Ie.
... 0

150""
0
0 ::i
~ •• o Not decarburized e'
0
- :;;
; 900 • Decarburized 0>
c
.~ • 0
0
0
C
.~

~ •• 100 ~
~ 600 0 r.
<t
~

300
.....
~
>-
50

Number of cycles to failure

Decarburization is the removal of carbon from the surface of

a steel part; as indicated in the above S-N curve, it significantly
reduces the fatigue limits of steel. Decarburization offrom 0.08
to 0.76mm (0.003 to 0.030 in.) on AISI-SAE 4340 notched spec-
imens that were heat treated to a strength level of 1860 MPa (270
ksi) reduces the fatigue limit almost as much as a notch with
K,=3.
When subjected to the same heat treatment as the core of the
part, the decarburized surface layer is weaker and therefore less
resistant to fatigue than the core. Hardening a part with a decar-
burized surface can also introduce residual tensile stresses,
which reduce the fatigue limit of the material. Results of re-
search studies have indicated that fatigue properties lost through
decarburization can be at least partially regained by recarburi-
zation (carbon restoration in the surfaces).

Source: Metals Handbook, 9th Edition. Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 674
106 4-24. 4340H Steel: Effect of Inclusion Size

1100

--.r--.. o Small inclusions

• Large inclusions

<,
1000
140
s::;;
~ 900
1;:
en
c
.~

s
E 800
;;:

700
<,
~
~ ~.
~
- 1

100

Number of cycles to failure

Fatigue life of two lots of AISI-SAE 4340H steel; one lot (lower curve) contained
abnormally large inclusions; the other lot (upper curve) contained small inclusions.

Points on the lower curve represent the cycles to failure for a

few specimens from one bar selected from a lot consisting of
several bars of 4340H steel. Large spherical inclusions, about
0.13 mm (0.005 in.) in diameter, were observed in the fracture
surfaces of these specimens; the inclusions were identified as co-
rundum or silicate particles. No spherical inclusions larger than
0.02 mm (0.00075 in.) were detected in the other specimens.
Large nonmetallic inclusions can often be detected by non-
destructive inspection; steels can be selected on the basis of such
inspection. Vacuum melting, which reduces the number and size
of nonmetallic inclusions, increases the fatigue limit of 4340.

Source: Metals Handbook, 9th Edition, Volume 1, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 673
 4-25. 4340 Steel: Influence of Inclusion Size 107

3
Xl0
200

190

180
STRESS
RANGE. 170 • ••
Ib/on 2
160 tal

150

140

130 •
i20
103

" " ,;: :I

• SURFACE INCLUSIONS
a SUBSURFACE INCLUSIONS

DIAMETER,
on 0001-

I--~~ CYCLES

SoN curve and dependence oflife on inclusion size

for AISI 4340 steel.

Typical initiated crack sizes are l-lO Mm. As this is an order-of-magnitude greater
than dislocation substructure sizes, such an initiated crack will behave as in a contin-
uum. For materials with lower stacking fault energy cross-slip and PSB formation is
difficult, thus inhibiting initiation. For such materials crack initiation can occupy a
significant fraction of life. Other microstructural sites for initiation are discontinuities
such as grain and twin boundaries, the latter being particularly operative in hcp metals.
Usually, however, at ambient temperatures it is the dislocation substructures which
dominate initiation.
For stronger, more complex alloys planar slip behavior dominates, making local-
ized slip bands the initiation sites cfthe random notch-peak topography generated by
shearing a pack of cards. The interaction of slip bands with second-phase particles
(inclusions, precipitates) can produce a local stress concentration which cracks the
interface, producing a surface crack. The above SoN curves show the results of this
process for a low alloy steel. Variations in fatigue life relate to variations in inclusion
size. As well as debonding, oxide or carbide particles can crack under concentrated
localized stresses.

Source: B. Tomkins, "Fatigue: Mechanisms," in Creep and Fatigue in High Temperature Alloys, J. Bressers, Ed., Applied Science
Publishers Ltd, London, England, 1981, p 115
108 4-26. 4340 Steel: Effect of Hydrogenation; Static Fatigue

--
Normal notch strength = 300 ksi
Ai>

.~, + Uncharged

-.
+ +-
275 I'\.. '" +

~
+-
'~ \
250 Bake 24 hr -
'---
225
"\\ 0\ \
1 <, Bake 18 hr
"\
'00
.x:
ul
(/) 200
\ -
~
(;) e\ \ 1 Bake 12 hr
"0 175
.~
a.
C-
« 150
\ \.
125
\ \ t-, Bake 7 hr

~ --
"..
:\ Bake 3 hr
100

75 • Bake 0.5 hr -.-..

50
0.01 0.1 1 10 100 1000
Fracture lime, hours

Static fatigue curves for quenched and tempered 4340 notched

specimens charged with hydrogen and baked at 150°C (300 OF)
for the times shown.

There are many embrittling effects of hydrogen on steels: the ultimate strength of a steel may be
reduced; ductility as measured by total elongation to fracture or reduction of area may be decreased;
and crack growth may be significantly accelerated. The hydrogen responsible for these effects may be
present in the environment external to the steel or may be present internally as a reslt of steelmaking or
processing operations such as pickling or electroplating. Hydrogen may promote a transition from a
ductile to brittle fracture mode or it may reduce ductility without a change in fracture mode.
The graph above shows the effects of baking at 150°C (300 OF) on the static fatigue (sustained load-
ing) of the hydrogen-charged specimens. Increasing baking time effectively lowers hydrogen content
even in the plated specimens, and sufficient baking eventually restores the strength of charged speci-
mens to that of uncharged specimens. The horizontal portions of the curves in the graph above are
designated as static fatigue or endurance limits, i.e., the stress level below which failure would not occur
no matter what the duration of stress application. As hydrogen content is decreased by baking, the
static fatigue limit increases.
The specimens used to obtain the above data were notched and therefore the static fatigue limits hold
for that particular notch geometry. In general, the sharper the notch, the lower the static fatigue limits,
an indication that a critical combination of hydrogen concentration and triaxial stress state is required
for crack initiation.

Source: George Krauss, Principles of Heat Treatment of Steel, American Society for Metals, Metals Park OH, 1980, p 223
 4-27. 4340 Steel: Effect of Hydrogen 109

400
4340
250,000 PSI
350

300

250
~
~

•
H
co ......
P-t .....
......
0
0
200
0
r-I

Cf) 150
Cf)

~
co 100
RATE UNEMB. EMBRIT.
1000 RPM
50 250 RPM 0
<>
PLATED IN LAB
COMMERCIALLY PLATED ••
0
200 RPM
.33 RPM
'V PLATED IN LAB
PLATED IN LAB •.
10 100 1000 10,000
CYCLES TO FAILURE
Schematic representation ofthe effect of cycling rate on theS-N curve of
hydrogen-containing 4340 steel, heat treated to a strength level of
250,000 psi.

Source: George Sachs, "Test Methods for Evaluating Hydrogen Ernbrittlement," in Materials Evaluation in Relation to Compo-
nent Behavior (Proceedings of the Third Sagamore Ordnance Materials Research Conference). Syracuse University, Syracuse NY,
1956, P 508
110 4-28. 4340 Steel: Effect of Nitriding

----- -- --- ATMOSPHERE

HITfUDED

---------
ATMOSPHERE NITRJDED TJfEN
GROUND TO RDlOVE COMPOUND
----
lAYER

5'>

QUENCHED AND TD1PERED

AT 1050 f (565 C)

HUHBER OF CYCLES

10 ~ 10 5

S-N curves for 4340 steel, gaseous atmosphere nitrided versus

not nitrided (quenched and tempered only), showing stress ver-
sus number of cycles for completely reversing torsional fatigue.

Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 287
 4-29. 4340 Steel: Effect of Nitriding and Shot Peening 111

1200 r-----,------,-----,,---,---,,------,------,-----------,
160
Nitridedcrankshafts

120
800

100

:Ii
~ 6001----+----'

Normally
heat treated
500 I----+-crankshafts --+--+--+----+----+-----,---------'5,....---1

350 L -_ _--'- - ' - _ - - - ' _ - - ' - - - - ' ' - -_ _---... ----'_.L.>... ----'

105

Cvcles to failure

Comparison between fatigue limits of crankshafts (S-Nbands) and fa-

tigue limits for separate test bars, which are indicated by plotted points
at right.

Mechanical working of the surface of a steel part effectively increases the

resistance to fatigue. Shot peening and skin rolling are two methods for de-
veloping compressive residual stresses at the surface of the part. The im-
provement in fatigue life of a crankshaft that results from shot peening is
illustrated in the above curves. Shot peening is useful in recovering the fa-
tigue resistance lost through decarburization of the surface; decarburized
specimens were shot peened, raising the fatigue limit from 275 MPa (40 ksi)
after decarburizing to 655 MPa (95 ksi) after shot peening.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 674
112 4-30. 4340 Steel: Effect of Induction Hardening and Nitriding

160
\ 'r-. As ni ide
._ 150 \
'"
a.. \ \
:5 140 \.
C!.

~-130 \ "I' ilrid ~ DU nducti n har en d

e
u;
120 "1\"'r--. Indu tio a dened
110
10 5 10 6 10 7
Cycles to Failure

As demonstrated in the above S-N curves, fatigue tests of

AISI 4340 steel in various surface hardened conditions show
that combined treatments produce endurance limits between
those developed by separate treatments.

Source: Sander A. Levy, Kenneth E. Barnes and Joseph F. Libsch, "Combining Nitriding With Induction Heating Pays a Bonus," in
Source Book on Nitriding, American Society for Metals, Metals Park OH, 1977, P 241
 4-31. 4340 Steel: Effect of Surface Coatings 113

190r---------------------------,

180

170
0---

6
_

----
O~---------<>-

~
~

O .. ~
<>-
Cr + SFl

Cd+
~Chromate

~160 ~1f8are
j
ISO
o
140
R • 0.8
2000 cpm
~TUngsten
Carbide +
RT SFl
Cr + we - Shot Peened
lJOL-----,---,-.w..u.uJ.,-----,---,-.w..u.u.l.;-----'---'-.L..LlllJJ.,-----'---'-.L..Lu.u-,;---'----'-L.LJL..LUJ

SoN curves (axial tension) of bare and coated 4340

steel in air environment.

Air 3.5% NaC1

Stress Stress
Change, Change,
Test Condi tion ksi MN/m 2 % ks; MN/m 2 %

Rotating Bending Bare 105 724 20 13B -81

R = -1 Cd + Chromate 105 724 0 80 552 -24
Cr + Ory Film' 95 655 -9.5 so 621 -14.3
WC + Dry Film' 90 621 -14.3 so 621 -14.3
WC + Cori cone + so 621 -14.3 90 621 -14.3
Dry Film'
Axial Tension Bare 160 1103 110 758 -31. 2
R = 0.8 Cd + Chromate 165 1138 +3.1 165 1138 0
Cr + SFL' 175 1207 +9.4 90 621 -43.8t
-4B. 6~
WC + SFL' 140 965 -12.5 60 414 -62.5t
-57~
WC + Cori cone + 140 965 -12.5 60 4.4 -62.5 t
SFL' -57f

'Shot peened
tCompared to bare alloy air value
~Compared to coated alloy air value

The above SoN curves, in conjunction with the table, contain the data obtained in air for 4340 steel, bare
and coated. Fatigue data at 107 cycles showed that the cadmium and chromium electroplates, particularly the
chromium, improved the fatigue strength. They were similar in both rotating bending and axial tension
fatigue tests. But in NaCl solution, significantly greater reductions in axial fatigue strength of the coated
alloys were observed due to environmental effects, which remains to be elucidated. Since the Cr and WC hard
(brittle) coatings have a relatively low intrinsic fatigue strength in comparison with the steel, they will become
discontinuous at a relatively low stress level owing to the development of fatigue cracks. (The Cr normally
contains internal cracks.) These cracks will permit access ofthe corrosive NaCI solution to the steel base at the
root of the fatigue crack. In the case of the axial tension test (high steady tensile load), it may be easier for the
environment to reach the crack tip.

Source; M. Levy and C. E. Swindlehurst. Jr., "Corrosion Fatigue Behavior of Coated 4340 Steel for Blade Retention Bolts of the
AH-I Helicopter.Yin Risk and Failure Analysis for Improved Performance and Reliability.John J, Burke and Volker Weiss. Eds .•
Plenum Press. New York NY. 1980. P 275
114 4-32. 4340 Steel: Effect of Temperature on
Constant-Lifetime Behavior

Minimumstress, ksi

150

'"
c,
:2
";;;
800 '"
e
1;; ~
100
E E
:J
"
E 600 E
"x "1'/.
~~.,
"00 "x
~ Q)lj...; 107 cycles lifetime ~

400
".9s
~~.. J'J' 6'a
..
'1"..0 1:> 50

200

0"--_ _-'----_ _--'--_ _---'--_ _---'_ _--'''"---_ _--'---_ _---'--_ _----' '"-_ _--'---_ _---'--_ _----'0
-1000 -800 --1l00 -400 -200 200 400 600 800 1000 1200 1400
Minimum stress, MPa

Constant-lifetime fatigue diagram for AISI-SAE 4340 alloy steel (bars) hardened and. tempered to a
tensile strength of 1035 MPa (150 ksi) and tested at the indicated temperatures. Solid lines represent data
obtained from unnotched specimens; dashed lines represent data from specimens having notches withK1
= 3.3. All lines represent lifetimes of ten million cycles.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 669
 4-33. 4520H Steel: Effect of Type of Quench 115

z 40
....I
30
p:\
....:l
0 20 ~4",~... DIRECf (OIL) QUENCH
~.;~~.'::......
0
0
...... (COLD OIL) /
~.... /
10 ~1ARQUENCH /-.:::••••••••••••
~
....:l
8 (400 0F OIL)' -........:.::-~ •••••••••••••
-----.....:.....
l=l

~
6
5 ---
4
~ 3

2
10 10 3 10 4
CYCLES - N
Effect of quench type on fatigue of carburized dif-
ferential cross.

Since 4520H steel is relatively low in hardenability for the part

involved, the depth-hardening characteristics of the two groups
were significantly different. The marquenched group had shal-
lower case depths, which resulted in fracture origins below the
surface at the case-core junctures. However, when the compari-
son was made with higher-hardenability steels, with sufficient
gradient strengths and thus all fracture origins at the surface, the
difference was very slight, though still in favor of direct quench-
ing. This is consistent with what is known concerning the differ-
ences in residual stress, which in this case would have been the
only other contributing factor. In other instances-such as for
gears, where distortion could be a factor-the results might turn
out differently for the marquenching.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 92
116 4-34. 4520H Steel: Effect of Shot Peening

40
~, 30
CQ
,...J

0
0
20
~':".':'.": .
0
......, ~.;....... SHOT PEE~D
10
---~
:::.:.: / .
§
,...J 8
lJNPEE\'ED~ __••••••••••••
----...:....
c:
u..;
if;
6 --
l:>: 5
F 4
~ 3

2
10 10 2
CYCLES - r\

Effect of shot peening on fatigue of carburized dif-

ferential cross.

Shot peening is known to increase fatigue strength; hence,

tests were run to determine the amount of increase to be ex-
pected. The above chart shows some of the results. Shot peening
was found to provide significant fatigue-strength improve-
ments. Peening surfaces that had suffered grinding damage was
found to be very beneficial, although not recommended because
of the high risk of having grinding cracks. Peening parts that
had marginal strength gradients improved the strength at the
surface but moved the failure origin to a subsurface location.
The net gain was small. It was also determined that to gain sig-
nificant improvement the hardness of the shot used was very
important. Since carburized surfaces are very hard, the shot
must also be hard to be effective.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 93
 4-35. 4620 Steel: Effect of Nitriding 117

-- ---- -- -- -- ATMOSPHERE NITRIDED THEN

GROUND TO RI1'IOV£ COMPOUND

--- r.~Y~p

Al110SPHERE
NITRIDED

QUENCHED AND TEMPERED

AT 1050 F (565 CJ

HUHBER OF CYCLES

10~ 1') 5

S-N curves for 4620steel, nitrided versus not nitrided (quenched

and tempered only), showing stress versus number of cycles for
completely reversing torsional fatigue.

Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 286
118 4-36. 4620 Steel: P1M-Forged

ksi

100

Height strain
x 80
oE

r.-------'-----Time
60

LI-! , - - I_ _- ' - -_ _' - - - _ - - - - - ' -

4.0 5.0 6.0 7.0

_

Log cycles

Axial fatigue of P 1M-forged 4620 steel as a func-

tion of height strain during forging.

In general, sensitive properties improve as the level of upset-

ting is increased during the forging process. The diagram above
shows the effect for fatigue resistance, although the cyclic stress
state also influences fatigue behavior. An interesting feature of
P / M-forged parts is the fact that deformation does not signifi-
cantly affect through-thickness properties as it does detrimen-
tally for wrought material. For re-pressed parts, through-
thickness toughness is slightly lower than longitudinal tough-
ness. Upsetting increases longitudinal toughness while toughness
in the through-thickness direction remains at a relatively con-
stant level.

Source: B. Lynn Ferguson, "Part II: Fully Dense Parts and Their Applications." in Powder Metallurgy: Applications, Advantages
and Limitations, American Society for Metals, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 100
 4-37. 4620 Steel: P/M-Forged at Different Levels 119

700 ,.------,-----....---------,r--------,-------,
100

600
co
a..
:2: Height strain, HIH o 80
x
.'+-----+ 71%
E (JrnaK 56 & 65%
.£ 500
CI)
CI) 42%
~ 1'------+ 30%
cil
60
400

o~ 0 7.0 8.0
Log cycles

SoN curves for P 1M-forged 4620steel at various levels offorging

deformation. As shown, fatigue limit increases as deformation
(level of strain) increases.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, p416
120 4-38. 4625 Steel: P1M vs Ingot Forms

100

50
1/
V
10
Ingot material
L,o 192 h /
\

o
1/ P/M material
L,o 5

163
h I

100 200 500 1000 2000 5000

Life, h

Fatigue life characteristics of P 1M roller bearing

cups, as shown by a typical Weibull plot. Shown is
a 10% life (L IO ) of 563 hr for P/M material
compared with L IO life of 192 hr for ingot
material.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals. Metals Park OH, 1984, P 620
 4-39. 4640 Steel: P1M-Forged 121

100 ,-------.----.r------...-------.-----.------.
0.365-1
r- '" 'Scatter band of SAE 4340
',~,Wrought steel, tested in the
':3. 80 dir '/
'0
"
"
longitudinal direction
-,
,,
8 3-1/4
f ~

'0, '------------

~ u~
,,
,,
c'i)
9-7/8 R. ,
60
'~
'------_Q:
21
32
Specimen
configuration
40 '--------'---------''------'-------'---'
103 10' 105 10· 10'
Cycles of stress

R. R. Moore fatigue curve for P/M-forged 4640

steel hardened and tempered to 33 HRC and a
yield strength of 138,000 psi.

Water-atomized 4600 steel powder was blended with graphite

and compacted in the split punch tooling. Fatigue data for P/M
forged 4640 are shown above, and these data fall within the scat-
ter band for 4340 steel. The most impressive statistic is that the
P / M-forged parts passed the Army ambient and low-temperature
firing endurance tests.

Source: B. Lynn Ferguson, "Part II: Fully Dense Parts and Their Applications,"in Powder Metallurgy: Applications, Advantages
and Limitations, American Society for Metals, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, p 103
122 4-40. High-Carbon Steel (Eutectoid Carbon): Pearlite vs Spheroidite

Property Spheroidite Pearlite

350. - - - - - - - - , - - - - - - . - - - - - - - , 5 0 Tensile strength,
MPa (ksi) 641(93) 676(98)
45 '"
~.
Yield strength, MPa
i------j ~ (ksi) 490(71)(b) 248(36)(c)
40 ~
Elongation in 2 in.,
'f; % .............. 28.9 17.8
c:
Reduction in area, % 57.7 25.8
35 ~
« Hardness, HB ..... 92 89
(a) Composition 0.78 C, 0.27 Mn, 0.22 Si, 0.016 S
6 7 and 0.011 P. (b) Lower yield point. (c) 0.1% offset
10 10 yield strength.
Number of cycles to failure

Both pearlitic and spheroidized structures have notably lower

fatigue strength than martensitic structures (see 4-1, on p 83). As
is shown above, the fatigue properties of spheroidized structures
are superior to those of pearlitic structures for eutectoid steels.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 677
 4-41. 52100 EF Steel: Surface Fatigue; Effect of Finish and Additives 123

1 MLNG
2 MU\ G
600
3 MLCG
4 SLD G 4.14
4 2 8 6 12 10 5 MLNH
6 MU\ H
7 MLCH 3.79
8 SLD H
MLN P
... 500
9
10 MU\ P 3.45
'"
.:< 11
12
MLCP
SLD P
.;
..'" 450
....
~ 3.10

til N
E

..
~ 400 2. 76 ~
'"
E

.~x 350 2.41

~

300 2.07

Mean Predicted Cycles to Failure

Effect of surface finish and additive on mean predicted surface

fatigue life. 52100 EF steel, high slip, high speed.

The mean predicted fatigue life is highest with a polished sur-

face and least with a ground finish (9 versus I, etc.). Polished
surface has about 6 times and honed surface about 3 times the
fatigue life of ground finish. No interaction effect between addi-
tives and surface finish is revealed.

Source: S. Bhallacharyya, F. C. Bock, M. A. H. Howes and N. M. Parikh, "Chemical Effects of Lubrication in Contact Fatigue-
Part II; The Statistical Analysis, Summary, and Conclusions," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 277
124 4-42. 52100 EF Steel: Surface Fatigue; Effect of Surface
Finish and Speed

II
600 4.14

3.79

... 500 3.45

Ul

""
.
::: 450
:;
/---e---i
1 5E GLl
3.10

2 5E GL2
'" 3 5E PLl
N
e
..
rl 400 4
5
5E
5V
PL2
GU
2. 76 ~
to

'"e 6
7
5V
5V
GL2
PLl
~ 350 8
9
5V
8V
PL2
GLl
x
:2 10 8V GL2
11 8V PLl
12 8V PL2
13 5E GH2
300 14 5V GH2 9 5 2 2.07
15 8V GH2
16 5E PH2
17 5V PH2
18 8V PH2

10 7
Mean Predicted Cycles to Failure

Effect of steel, surface finish and speed on mean predicted fatigue

life. Low-viscosity mineral oil, no additive.

Interaction effects of steel with speed and of surface finish

with slip and speed on fatigue life are shown in the above graph.
The direct steel effects are nonsignificant. The effect of surface
finish is shown in the difference between the two line groups 13,
14, 15 (ground) and 16, 17, 18 (polished). The difference in the
line groups 4, 8, 12 (low slip) and 16, 17, 18 (high slip) again
brings out the very large detrimental effect of high slip on life.
Higher speed decreases life with the maximum effect observable
on 8620 CV steel (compare lines II and 12), on polished speci-
mens at low slip ratio.

Source: S. Bhattacharyya, F. C. Bock, M. A. H. Howes and N. M. Parikh, "Chemical Effects of Lubrication in Contact Fatigue-
Part II: The Statistical Analysis, Summary, and Conclusions," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed .. American Society for Metals, Metals Park OH, 1980, P 277
 4-43. 52100 EF Steel: Surface Fatigue; Effect of Lubricant Additives 125

HLN 1.2
HL\ 1.2
600 HLC L2 4.14
~1.0 1.2
HI.N 112
HL\ 112 J.79
HLC 112
~LO 112
.... J.45
'"
.:<
90% Con fidence
Band
'"
'" J.IO
"
~
~

N
E

.
::l
400 2. 76 ~
"
x
E

"
.~
.x J50
:E
2.41

JOO 2.07

250 L...;--'----'--'L..L__'_'..u..L..;-_ _'____'L....
L..L..L.J..u.~--'----'L....IL..L--'-'..u.....--'----'L....IL..L--'-'L..1..L! I. 72

J 6
IO 10 10 7
:ican l'r-ed i c t ed Cyc l e s to Fa i lure

SoN diagram for 52100 EF steelshowing the effect oflubricant additives

on surface fatigue. The specimens had a ground finish, and a low-
viscosity oil was used. Additives were used for I, 2, 3 and 4; the favorable
effect of the additives is obvious.

Source: S. Bhattacharyya, F. C. Bock, M. A. H. Howes and N. M. Parikh, "Chemical Effects of Lubrication in Contact Fatigue-
Part II: The Statistical Analysis, Summary, and Conclusions," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 275
126 4-44. 52100 EF Steel: Surface Fatigue; Effect of Lubricant Viscositv,
Slip Ratio and Speed

1 MLN Ll
2 MLN L2
3 MHN Ll
600 4 MHN L2 4.14
5 SLD Ll
6 SLD L2
550 7 MLN Hi j.79
8 MLN H2
9 MHN Hi
-;: 500 10 MHN H2
) .45
.Yo 11 SLD Hi
. '
:: 450
12 SLD H2

f---<T---i 3.10
t
.
l/} N

-!:
.
:: 400 2.76 G
'"
E
~
~ 350 2.41
:>!

300 2.07

250 '-c--'-------'---L-...L.J...J....IJ.Lr--.1_L-LJ...L.I...l..L.L...,-_...L--..L----'-...L-L.LJL..U.,,--........J_-'----.LJ~L.LJ.J
10 5 10 10
Mean Predicted Cycles to failure

SoN diagram for 52100 EF steel.

The effects of lubricant viscosity, slip ratio, and speed on fatigue life are
shown in the diagram. The 12lines in the figure are separated in two distinct
groups, low slip (lines I to 6) and high slip (lines 7 to 12). In each group the
effects of viscosity and speed may be noted. Viscosity X speed interaction
produces complex effects on mean predicted lives which under low slip con-
ditions are not statistically significant in their differences. Only under high
slip condition, lines 9 versus 10indicate a small statistically significant lower-
ing in mean fatigue life in high-viscosity oil under higher speed. A comparison
of lines 11 and 12 shows that the lesser life in synthetic oil with additive is a
statistically borderline case, though the trend is similar to that with mineral
oil under the present operating conditions. The regression analysis shows
that in the present tests both speed and viscosity have nonsignificant direct
effect on life, and a few small interaction effects with steel, surface finish,
viscosity, and slip were observed.

Source: S. Bhattacharyya, F. C. Bock, M. A. H. Howes and N. M. Parikh, "Chemical Effects of Lubrication in Contact Fatigue-
Part II: The Statistical Analysis, Summary, and Conclusions," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, p 276
 4-45. 52100 EF Steel: Rolling Ball Fatigue; Effect of Oil Additives 127

SOO , - - - - - - - - - - - - - - - : : - - - - - - - - - - - - - - , 3 4 5

(OMe bl & 68'"

400 2.76
BASI 011 OA'"

300 2.07
~

vi
~
w
N
~ E
<,

~
z
D<
\ e
W Co
::c
200
"'4 1.38
"
:>
.}Is-,f....

~o

"
)(
ot
o ()"''''
""'.. F

" -,

----- 'HASiIllOAU.

'I(DI(1I0 BY MUllIPll UQRfUION ANAlYSIS

10 100

l,. LIFE. 10 6 CYClES

Comparison of stress/life data for the mineral oil with and without the
ZnDDTP additive in surface fatigue; 85 percent confidence bands for the LSD
life estimates are shown and compared with the stress/life relation predicted
from regression analysis.

The synthetic and mineral oil no-additive conditions had about the same
life. However, the life at all stress levels tested was significantly reduced for
the mineral oil with additive below that without additive, by almost a factor
of three at the L so level, further indicating a detrimental effect of the
ZnDDTP additive on life.
Both the synthetic and the mineral oil tests had lives almost two orders of
magnitude below the standard Lundberg-Palmgren calculated life. A life re-
duction factor is used with the Lundberg-Palmgren theory when applied to
rolling bearings having high contact angles and thus high slip; but rarely does
the slip at bearing contacts approach that level used in these tests, so it is not
surprising that the life reductions observed are much greater than the life
reduction factors normally used for bearings.
The stress/Iife plot shown above is particularly revealing. There is no
doubt that the stress / life slope for the additive oil is significantly steeper than
for the base stock, which seems to approach the Lundberg-Palmgren theory
in stress/fife slope except for the highest stresses where it is even shallower.

Source: w. E. Littmann, B. W. Kelley, W. J. Anderson, R. S. Fein, E. E. Klaus, L. B. Sibley and W. O. Winer, "Chemical Effects or
Lubrication in Contact Fatigue-Part III: Load-Life Exponent, Life Scatter, and Overall Analysis," in Source Book on Gear
Design, Technology and Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 285
128 4-46. 52100 Steel: Carburized vs Uncarburized

99.
95.
90.
80.
70.
60.
50.
40.
0
ILl
...J
30.
ex 20.
IL.

...
Z
ILl 10.
0
a:::
ILl
a.. 5.0
4.0
3.0
2.0

I. 0 L..-_-'------'-----'----'''--'--'l..J...J...L--_-'------'-----'----'''--'--'LJ...LJ
I 10 100
MILLIONS OF STRESS CYCLES

Rolling Contact Fatigue Tests

Bar specimens, 0.973 em (0.383 in.) in diameter, about 8 em
long, were machined from spheroidize-annealed 52100 steel.
Two pieces were copper plated to prevent carburizing, then,
along with two unplated pieces, were austenitized at 815°C
(1500 OF) for two h in a carburizing atmosphere, oil-quenched
and tempered for I.5 h at 175°C (350 OF). After finish grinding
to 0.953 mm (0.375 in.), pieces were fatigue tested using a Poly-
met Model RCF-I testing machine with a computed maximum
hertzian contact stress of 503 MPa (729 ksi).
A Weibull plot, shown above, of the 16 tests on each type of
specimen shows that pieces with a carburized surface had a fa-
tigue life about 50% longer at all failure rate levels than pieces
which were subjected to the same thermal cycle, but not carbu-
rized. The nonparametic Walsh test for statistical significance
indicated at a 99.5% level of confidence, the two batches of fa-
tigue test data came from different populations.

Source: C. A. Stickels and A. M. Janotik, "Controlling Residual Stresses in 52100 Bearing Steel by Heat Treatment," in Residual
Stress for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, 1981, p 34
 4-47. 8620H Steel: Carburized; Results From Case and Core 129

CASE----
CORE----
CARBURIZED----

10' 10· 10'

REVERSALS TO FAILURE,

Summary plots of total strain amplitude: reversals-to-

failure data for simulated case, simulated core, and car-
burized materials.

Carburized material is seen to have low-cycle fatigue resist-

ance intermediate between the simulated case and core material,
a common intersection with simulated core material at interme-
diate lives; and in the long-life regime, carburized material spec-
imens are more fatigue resistant than either simulated core or
case material specimens.
Plotting the strain-life curves for both case and core simulated
materials on a common set of axes , as shown in the above chart,
reveals an interesting feature. It has been observed that curves of
these materials intersected at a life of approximately 2.NJ = 105
reversals. This is in agreement with the results of this investiga-
tion. Intersection of the life curves for simulated case and core
materials accounts for a shift of failure location in carburized
components.

Source: J. M. Waraniak and D. F. Socie, "Cyclic Deformation and Fatigue Behavior of Carburized Steel," in Wear and Fraclure
Prevention, American Society for Metals, Metals Park OH, 1981, P 249
130 4-48. 8620H Steel: Effect of Variation in Carburizing Treatments

300

Single --.........
reheat .........

•
•
-------~rr"~
<lID
- /""": .0
o
Direct quench ---./ 0
o o

100 L....---L--L....L..JL...U.UJ..._...L......l-JL...L.I..LU.l...----l.---l.....L...L.L.I........ _-'--'-.L..L.J..L.LJ.J---J

103 105 107

Cycles to failure

The above S- N curves show results of a study of the effect of

martensite morphology, including the effects of micro cracking
on fatigue resistance of a carburized 8620 steel. These speci-
mens, which were directly quenched from the carburizing
temperature, had the coarsest structure and the highest density
of microcracks, some of which were directly exposed on the
specimen surfaces bychemical polishing. The single-reheat spec-
imens had a finer austenite grain structure and therefore finer
martensite plates and a lower density of microcracks. Since the
retained austenite content and hardness profiles of the direct
and single-reheat specimens were identical, the improved fatigue
resistance of the single-reheat specimens is attributed to the
smaller size of the microcracks and their lower density in the
finer structure. The best fatigue resistance was shown by the
double-reheat specimens.

Source: George Krauss, Principles of Heat Treatment of Steel, American Society for Metals, Metals Park OH. 1980, P 264
 4-49. 8620 Steel: Effect of Nitriding 131

50
§ .... _-- .... _--
~
~

;
--
-- .... -- .THO~l'\IrRr NITRIDw TNr.N
GROUND TO RJ>1OVr COMPO''''"
lAnR

AnlOSI'IlERE
:;; NITRIDED

iO
QlJ[HCHED AHD TD4PER£D
AT 1010 F (\6\ C)

tHJKBER OF CYCUS

7
10

S-N curves for 8620 steel; nitrided versus not nitrided (quenched
and tempered only), showing stress versus number of cycles for
completely reversing torsional fatigue.

Source. J. A. Riopelle, "Short Cycle Atmosphere Nitriding,"in Source Book on Nitriding, American Society for Metals, 1977,p 286
132 4-50. 8622 Steel: Effect of Grinding

EFFECT OF GRINDING BURN

C/) 40
~ 30
...... 20

~ ... . ....
-.-.-.-.
"':'.:.:-. - • _ 8622 MATERIAL
_ 10 ............... . ..... ............
o 8
o
o
..... 6 ...................
4 ....
........
GRINDING WITHam BURl, - • - . - '"
2 SEVERE DAHAGE BY GRINDING •••••••• ••••••
(REFER FIG. (~4))

10 10 2 10 3 10 4 10 5 106
CYCLES TO FAILURE

Influence of grinding quality on fatigue properties of car-

burized differential cross.

C/) 40

z~ 30
~~
...... 20 "2'::.".::- • 8822 MATERIAL
I
....-::-.:.......
§
P- 10
.....::: ........
...... ........
........ ::: .
~
o 8
0
0 6 '"
.....
~ 4

~
t::l
2
GRIND AFfER H. T.
GRIND BEFORE H.T.
- . _.-

Z
......
ffi
~
10 10 2 10 3 10 4 lOS 10 6
CYCLES TO FAILURE

Influence of grinding sequence on fatigue of carburized

differential cross.

As shown in the above charts, grinding has an important influence on fatigue.

Elimination of grinding damage resulted in drastic improvement in fatigue perfor-
mance (upper chart). However, it was also determined that a high-quality ground
part gave better fatigue performance than when the carburized surface was left
unground (lower chart).

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, pp 91-92
 4-51. Cast 8630 Steel: Goodman Diagram for Bending Fatigue 133

LEGEND
-. - CAViTIES
- 0 - CAST STEEL - SOUND
- X- R.R. MOORE FATIGUE SPECIMEN I UNNOTCHED I
1.0 - a- R. R. MOORE FATIGUE SPECIMEN (NOTCHED 0.001!! in. R I
CONTAINED WITHIN THE BAND ARE THE
FOLLOWING DISCONTINUiTIES
WELD -INCOMPLETE PENETRATION
0.9 WELD - UNDERCUT
WELD - SLAG
WELD - MACHINE - SOUND
OoB AS WELDED - SOUND
SLAG INCLUSION
HOT TEARS

0.7

0.6

0.5

l: l:
l- I-
e> ~ 0.4
Z
III III
0:: 0:: ;.:.:.:=.::.::.....:.:.:=--=-""-=~ RANGE
I-
Ul :n 0.3

III III
.J
=>
e> ~O.2
~
lL
III
I-
0.1

0
o.s
MEAN STRESS
- 0.1 TENSILE STRENGTH

- 0.2

- 0.3

- 0.4

Goodman diagram for bending fatigue for normalized and tempered cast 8630 steel.

Data here show that severe discontinuities lower the fatigue strength of
cast steel. However, it will be observed from the Goodman diagram above
that the results of the notched [0.0015 in. (0.0381 mm) radius] R. R. Moore
fatigue specimen fall below those of the other bending fatigue values. Good-
man diagrams for torsion fatigue and for a quenched and tempered heat
treatment show similar conditions with the notched fatigue values below the
surface discontinuity values. In many cases, therefore, design, based upon
notched R. R. Moore fatigue data, introduces a safety factor.
It must be remembered that the discontinuities were very severe and ex-
ceeded all ASTM classes of nondestructive inspection standards. The allow-
able discontinuities described in the ASTM standards are therefore expected
to exert a somewhat less damaging effect on fatigue behavior.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed.. Steel Founders' Society of America, Rocky River OH, 1980,
P 15-32
134 4-52. Cast 8630 Steel: Effect of Shrinkage

0.6 ,--------,----,--------,r------,
UTS END.
!.!l (~) ~
138 (951) 0.17
0.5 138 (951) 0.13
(/)
135 (931) 0.26
(/) 137 (945) 0.28
W
g:(/)1-::c 0.4
o
WZ
::::>W
00::
i= Ii; 0.3
ItW I:::=0-
0-.J 6. CLASS 2 SHRIN K
z-
_(/) • t. (SUB SURFACE)
~ ~ 0.2 "" ~ flLASS 2 SHRINK
Zl- CLASS 6 SHRIN~~A~~ 6. (EXTENg~ SURF.)
0::
w ( EXTENDS TO ~URFACE )".
I-
-.J
<3: 0.1
~.

0'--------'-----...1.--_ _---'
--.-~

-
NO FAILURE

----'
104 10 5 106 10 7
CYCLES TO FAILURE

Effect of shrinkage on plate bending fatigue of quenched

and tempered cast 8630 Ni-Cr-Mo steel.

As shown in the chart here, plate bending tests (completely

reversed tension and compression with cast-to-size specimens)
oflow-alloy cast 8630 steel indicate only minor effects of Class 2
internal shrinkage.

Source: Steel Castings Handbook, 5th Edition. Peter F. Weiser. Ed., Steel Founders' Society of America. Rocky River OH, 1980,
P 15-30
 4-53. Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue 135

0.5 ...------,-------r-----r--------,
I I I

(f) TENSILE STRENGTH

(f)
w 0.4 - .84 - 91 ksi (579 - 627 MPa) -
OCr
t;;t;
oc Z
w 0.3 -
<1
w OC ° rSOUND
rt;; • 0 """"-- x I
(f) <. X~ 0
. . . . . . . . . ~- 9t,--:;, O'::CLASS 2
w
~....J
Z(f) 0.2 I-
~~SHRINK
......
CLASS 6 SHRINK-./ -1-
~~ -..
zl- NO FAILURE
OC
W
I-
....J
0.1 r- ....
0-0

x-x
NO. 15
NO. 13
NO. 13
SOUND
CLASS 6 SHRINK
CLASS 2 SHRINK
-
<1

CYCLES TO FAI LURE

Effect of shrinkage on torsion fatigue properties of an-

nealed cast 8630 steel.

Source: Steel Castings Handbook. 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 15-31
136 4-54. Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue

0.5 ,....------r----~----...,__---___.
I I I
T ENSILE STRENGTH
en
en 0.4 I- 133-138 ksi (917-951 MPo)
W
lr
I-I
en I- ....
o . . XC
lr Z o ,~?...
<t W 0.3 ~ o Xo ..... _ 0
W lr
II- o 0 o~--o_
en en
~--1
W «<. o~ 0
X_X_

-
-
Z in 0.2 I- 00 - - - 0 _
j::Z
<tW
ZI- 0-0 NO. 15 SOUND NO FAILURE
lr
W 0--<>
~ 0.1 I-
x-x
NO. 12

NO. 12
CLASS 6

CLASS 2
SHRINK

SHRINK
-
<t

I I I

10 5 10 6 10 7
CYCLES TO FAILURE

Effect of shrinkage on torsion fatigue properties of water

quenched and tempered cast 8630 steel.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
p 15-31
 4-55. Cast 8630 Steel: Effect of Shrinkage on Plate Bending 137

0.6 l:i.-,-------,-----,-------.
en
en \

~
AXSOUND
W
a::
I-J:
en t; 0.5
WZ
:::>W
~a:: CLASS 2 6,
~tn SHRINK ----- o_"'\..
Lt W
0.4
( SUB SURFACE 1 ° 6 "'-
~...J
0'0 6-
Z -
_en ''00, '--6_
0 0_
I-Z
<l:W
Z I- 0.3 UTS END.
a:: STEEL ~ I MPo 1 RATIO
W
I- 6 II 83 (5721 .35
...J 14 B4 (5791 .32
<l: 0.2 L..-
° ..J....... ----' ......... .....J

104

CYCLES TO FAILURE

Effect of shrinkage on plate bending fatigue of normal-

ized and tempered cast 8630 Ni-Cr-Mo steel.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 15-31
138 4-56. Cast 8630 vs Wrought 8640

600,--------.----------,-----------,
Notched Unnotched
Wrought 8640 0 • 80
Cast 8630 '" '"
5001--------,,------=""""'ct----------t---------;

.0;
0:
:;; 60 "
::i 400 1-----------+-----2-"""=-+-~---------l ~
~ ~
IJ)

300 I-----------+"""-~""'_=------t------------j
40
Normalized and
tempered to 220 HB

0.1 1 10
Millions of cycles to failure

600
Ouenched and
tempered to 80
286 HB
500

~ .0;
c,
sc
:;;
60 ~.
400
e U5
U5

300
40

200L----------'----------'-- -----J
0.01 0.1 10
Millions of cycles to failure

The fatigue limit for smooth-machined specimens is generally

about one half the tensile strength, but is reduced considerably
by notches or a rough cast surface. The S- N curves in the graphs
above compare wrought 8640 and cast 8630 steel in two different
conditions of heat treatment. In both of these comparisons, the
wrought 8640 is superior, but the two steels are practically iden-
tical in the notched fatigue test. This is significant because most
articles fabricated from either wrought or cast steel contain
more than one notch and more than one type of notch.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 397
 4-57. 8630 and 8640 Steels: Effect of Notches on 139
Cast and Wrought Specimens

Tensile strength. ksi

80 100 120 140 160
700 r---.-,-----..------.-.---.----.----.------y-----,
100
,
• Cast steel
600 I---I-------Ir----t-----t-----..~!i'±i~_+--___l

80

5001---1--

"'----+---t----j 60 ]
"
.~
u.
300
40

20
100 L -_ _-'---_ _-'----_ _--'---_ _--'---_ _-'---_ _-'---_ _---l

500 600 700 800 900 1000 1100 1200

Tensile strength, MPa

The effect of notches on fatigue limit is apparent when com-

paring similar wrought and cast steels with regard to fatigue
limit at selected static tensile strength levels; note curves above.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 397
140 4-58. Nitralloy 135 Steel: Effect of Nitriding

60 .....
..........
..... ......
.... .....
...... ...... ATHOSPHEAE NITAIDED THEN
CADUND TO RDlDVI:
CDHPOUND lAYER

50
AntDSPHERE
NITRIDED
---

40
QUENCIlED AND mtPERED
AT 1010 r (16' C)

MlHIER OF CYCLES

7
10

S-N curves for aluminum-bearing nitriding steel (Nitralloy 135),

gaseous atmosphere nitrided versus not nitrided (quenched and tem-
pered only), showing stress versus number of cycles for completely re-
versing torsional fatigue.

Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals. 1977.p 287
 4-59. AMS 6475: Effects of Welding 141

AMS b 4-75 (CE.VM)f ~TI r;.UE

120 " R:.R Me OR 1:.- CSMOO Ir-H i)Pec,C

_110
\ "", ----- ;OR~

'\.'~ "
- - - - ~~ \~EL.\:)ED ( E.B')
WEWDE't>+ST~~~!l ~eL..
V)
e,
; i i
':I. 100
, <, "-- --+ -- ---- -
If)
(/) 90
" '., ....... I

lJJ
~ " i"o .. _
~ 80 ~--

IO~ 106 107 10 8

CYCLES TO FA'LUR~

Fatigue strengths for case-hardened materials as well as

through-hardened may be satisfactorily defined using the R. R.
Moore rotating specimen test. The smooth unnotched Moore
specimen is ideally suited for studying many of the effects of
manufacturing and processing variables upon fatigue endur-
ance. An example of the use of this testing technique in the eval-
uation of electron beam welding and postwelding aging effects
upon the endurance limit of basic AMS 6475 material is shown
in the above S-N diagram.

Source: Charles W. Bowen, "Review of Gear Testing Methods," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 346
142 4-60. Medium-Carbon, 1Cr-Mo-V Forging: Effect of
Cycling Frequency

600

500 'E,'! 0.75 %

400

300

w 500 .-..---r-.-n'TTTT"--'--'-rTT'1~---.-.--rTTTTrr-....,.-;".,."""""
a::
::>
--l400
~
g 300
(f)
w 200
--l
U
t> 100 L-..L......L--L.L..I.lllL----'--l-L..L1.JL.U.L...-..L.......L.L.L.U.Lll....---'---'L.J...J.J..WJ

~:] ~': ;,:,~:~ ::~ ::::' : : : :J

10- 2 10- 1
FREQUENCY - CYCLES PER MINUTE
1.0 10.0 100.0

Influence of cycling frequency on the fatigue

properties of forged lCr-Mo-V steel at 1049 OF
(566°C); no dwell period.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH. 1980,
P 15-55
 4-61. EM12 Steel: Effect of Temperature on Low-Cycle Fatigue 143

1= 1lef"rmal:ion ""'90-
I- li'T (%)
I-

I-
I-

I- -~ 1550°

O/ll---I---+-+-H--+---+-+-H--+---+----I---If-+--+---+-I-H--t---+---I-l--l
I-----I-++t--+--+--HH----I--+--t--t-+---t----t-+--HI--+---t--t-t--l

I I I r I

10 10~

Low-cycle fatigue ofEM12 at 20 and 550°C (68 and 1020 OF).

As holds true for other ferritic steels, the effect of hold time in
compression is slightly detrimental to fatigue life.

Source: Philippe Berge, Jean-Roger Donati, Felix Pellicani and Michel Weisz, "Properties of EM 12." in Ferritic Steels for High-
Temperature Applications, Ashok K. Khare, Ed., American Society for Metals, 1983, p 114
144 4-62. Cast O.5Cr-Mo-V Steel: Effects of Dwell Time in
Elevated- Temperature Testing

~
+,' 3.0 , - - - - - r - , . - , - - , - ,r-tr-'r-r-r-t-t-t-r-r-r-r-r-r-r-r-r-......-r-r-r-rr-r-r-r-r-r-r-r-r--r
Z REVERSE BENDING
« 6 MATERIAL A
~ o
"J
MATERIAL B
MATERIAL C
-l 1.0
~ PUSH - PULL
g ° MATERIAL A

LL
o ............CONTI NUOUS CYCLE
W
t:l
Z
<l
0::
LL
-l
<l
I 0.1 L-_----'-_..L-....l-..I......JL..LLJ...JL-_-'-_-'---...l......l.....J.-L-L..LJ--:-_-'

10 2 10 3
CYCLES TO FAILURE Nf

___ I
'f--
-6-----6_
..... ~ 0.5 % STRAIN
z
W
0::
::J -6----_6 _
-l
~ 100 f0-
--- A-=
~ 1.5 % STRAIN
g
(f)
W
.J
U
>-
U
I I
1.0 10
DWELL PERIOD - h

Effect of dwell periods on fatigue characteristics of

low-alloy cast steel.

As the upper diagram shows, when a D.5Cr-Mo-V steel was

tested at 1022 ° F (550 0C), a 20% drop in fatigue life in reverse
bending resulted when a D.5-hour dwell was added to each cycle.
The lower diagram shows that extended dwell periods, up to 10
hours, have relatively little additional effect beyond that in-
duced by the D.5-hour dwell.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
pp 15-56 - 15-57
 4-63. Cast 0.5Cr-Mo-V Steel: Effect of Environment 145
at 550°C (1022 OF)

o"e
..,' " I
REVERSEO BEND TESTS
z 0. IN AIR
~
+ IN STEAM
~ x IN VACUUM
PUSH - PULL TESTS
-l 1 0
~ ° IN AIR
~
u,
o
W
o
Z
«
a::
o I '-::2--'-----'-...L..JL....L..L.L.L-'--;:---'-----'-...L..J'--'--'c...L.L'--;---'---'--'--'--'-'..........
10

CYCLES TO FAILURE - N

Fatigue endurance behavior of cast 0.5Cr-Mo-V steel at

1022 of (550°C) in air, steam, and vacuum (no dwell
period).

Source: Steel Castings Handbook. 5th Edition. Peter F. Weiser. Ed .• Steel Founders' Society of America, Rocky River OR 1980,
p 15-55
146 4-64. Cast C-O.5Mo Steel: Effect of Temperature and Dwell Period
on Cyclic Endurance at Various Strain Amplitudes

TEMPERATURE - 0 c
100 200 300
10 4 c:---,-----r---,-----.--....--...----,

103
If)
0.6 %
W
...J
U 0.7 %
~
U
I 1.0 %
W
U
Z 1.5 %
<t
0::
=> 0.5 %
0
z
w
10 2

1.0 %

CONTINUOUS CYCLE
30 MIN. DWELL TIME

10 '--_ _. L -_ _- ' -_ _...L-_ _--'-_ _- ' -_ _- ' - - - '

200 400 600 800 1000 1200
TEMPERATURE - OF

Influence of temperature and dwell period on the cyclic

endurance of C-O.5Mosteel at various strain amplitudes.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980
pIS-55
 5-1. HI-FORM 50 Steel vs 1006 147

o HI-FORM 50 I STRAINED
~ 0.010
t.1006 lAND AGED
....::::;
::l

...
::E
<I:
Z
<I:
....
~

VI
U
.....
u
>-
u 0.001 '-- ---'- ....L.-:- _
3 10 4
105
10
REVERSALS TO FAILURE, 2Nf

Strain-life data for AISI 1006 and HI-FORM 50 (a columbium-bearing

HSLA steel) in the strained-and-aged condition.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals, Metals Park OR, 1981, P 214
148 5-2. HI-FORM 50 Steel vs 1006: Stress Response

0500L

~ ~1.FORM50
~400 o-o-~
~~ ~
"-"-~C)oO-OC>-_~0.002
0.0065

gJOO
~
1006

STRAINED AND AGED

OL.-----.L..;-----...I..::------'--;:------L-;------'
2 3
1 10 10 10~
CYCLE NUMBER, N

Stress response of strained-and-aged AISI 1006 and HI-

FORM 50 steels.

The imposed constant total strain amplitudes are indicated on

the graph. The degree of softening of these two steels is less
compared with that of dual-phase steels, which simply reflects the
significantly lesser degree of strain hardening of the 1006 and
HI-FORM 50 compared with the dual-phase steels.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals, Metals Park OH, 1981, P 209
 5-3. HI-FORM 50 Steel Compared With 1006, DP1 and DP2 149

500
_000-0 DP 1
3=
cyo-~
.......o 400 DP 2

U 0
~HI-FORM50
a~ 300
>-
Uv; /0
~~ 200
~:;;
II>

! '-1006

..:
~ 100 STRAINED-AND-AG ED
«
II>

OL...-_...L-_--'-_----'L-_...J..-_--'-_ _
o .002 .004 .006 .008 .010
CYCLIC STRAIN AMPLITUDE

Comparison offour steels: AISIl006, HI-FORM

50 (a columbium-bearing HSLA steel), a lean
phosphorus-bearing dual-phase HSLA steel
(DPl), and a carbon-manganese dual-phase HSLA
steel (DP2).

Here it can be seen that all three high-strength steels offer

substantial increase in load carrying capacity at the same gauge
when compared to the plain low-carbon steel. This confirms the
potential for gauge, and consequently weight, reduction that
can be realized from the use of higher-strength steels.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals, Metals Park OH, 1981, p 212
150 5-4. HSLA vs Mild Steel: Torsional Fatigue

300
\,
\\.
~ -. ~
A

I
-
IWIK A
SPC(1.21)
-'-0-- IAI'I'C4011.111
I
spc(o,81)
spc(O.81)
..·06-- I APFC45It.Ol ISPcIO.81
--0-- APFC5O(1.01 SPC(O,811

0: 1'-.~"" " . ~~ ~
SPC : MLD STEEL
APFC: IIGH STllEHOTH STEEL
I : THICKNESS(mm)

-.......:;.r......:.~ ~.
~ ~.'- • _
D

--~
...
50

10' 10'
NUMBER OF CYCLES

SoN curves showing torsional fatigue of automobile frame steels.

To determine whether the foregoing basic test results apply to

the frame models, experiments were conducted. The above
chart presents the torsional fatigue behavior of the frame mod-
els fabricated with the mild steel (0.8 mm) and each ofthe three
high-strength steels. In the high-stress, low-cycle range, fatigue
strength differs with the class of high-strength steel but virtually
no differences of that nature are seen in the low-stress, high-
cycle range. The three high-strength steel combinations showed
virtually the same torsional fatigue strength values as those of
the mild steel (1.2 mm) combination, indicating the possibility
of gauge reduction.

Source: M. Takahashi. "Criteria of High Strength Steels for Applying to Automobile Frame Components," in HSLA Steels-
Technology & Applications, American Society for Metals, Melals Park OH, 1984, P 498
 5-5. Proprietary HSLA Steel vs ASTM A440 151

0.03 .-----,----,----T""""--T""""----,

0.01 t---=~~f___t--_t--_+----l

.,
"tl
.e
Q.
E
'"
e
en'i! Proprietary HSLA
690 MPa (100 ksi] min UTS
0.001 t---t------jt---t---t--~

0.0004 L..-_~L..-_~L..- _ _L -_ _.L-_ _____I

102 103 104 105

Cycles to failure

Total strain versus fatigue life for ASTM A440

having a yield strength of about 345 MPa (50 ksi)
and for a proprietary quenched and tempered
HSLA steel having a yield strength of about 750
MPa (110 ksi),

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 672
152 5-6. Comparison of HSLA Steel Grades BE, JF and KF for Plastic
Strain Amplitude vs Reversals to Failure

2.0
••
1.0

t
'ill-

~N
ui 0.1

~
::J
/.Aj-pl= 1.871(2Nf)-o·8396

e, R 2= 0.968
::iE
-e
z·
:(
a:
I-
en 0.Q1 •
0 •
i=
en • BE(Cb-)
:5e, • JF(Cb-V)
• KF(Cb-V-SI)

0.001
10 2 10 3 10 4 10 5 10 6 10 7
REVERSALS TO FAILURE. 2Nf

Plastic strain amplitude vs reversals to failure for Cb (BE), Cb- V

(JF) and Cb-V-Si (KF) steels.

For plastic strain-life relationship the statistical analysis indi-

cates that there are no significant differences between the three
steels (F-ratio is not significant). This is further illustrated in the
above chart, where all the plastic strain data are plotted as a
function of reversals to failure. It is clear that a single straight
line can adequately describe all the data. Such a regression line is
drawn as the solid line in this chart.

Source: Shrikant P. Bhat, "Influence of Composition Within a Grade on the Fatigue Properties of HSLA Steels," in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 588
 5-7. Comparison of HSLA Steel Grades BE, JF and KF for Total Strain 153
Amplitude vs Reversals to Failure '

.,.
w
o 2
::::I-
I-
::::i
e,
:::l:
<
z
<
a:
I-
en
...J
-e
I- 8
12 0.1'------'----'----'----'-----'
10 2 10 3 10 4 10 5 10 6
REVERSALS TO FAILURE. 2N f

Total strain amplitude vs reversals to failure for Cb (BE),

Cb-V (JF) and Cb-V-Si (KF) steels.

Strain-life behavior: The strain-life curves for the three steels

are compared in this graph. It is clear that when plotted as total
strain versus reversals to failure, the three steels behave similarly
and the differences between them are minor.

Source: Shrikant P. Bhat, "Influence of Composition Within a Grade on the Fatigue Properties of HSLA Steel," in HSLA Steels-
Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 587
154 5-8. Comparison of a Dual-Phase HSLA Steel Grade With
HI-FORM 50: Total Strain Amplitude vs Reversals to Failure

~ 1.0 0 AS-RECEIVED
w- DUAL PHASE 1 16 STRAINED AND AGED
o
...='::::; HI-FORM SO - - -
".
:e
«
z
«
...'"
'"
...u
u
>- 0 . 1 ' - : : - - - - - - - - - - - ' - - : - - .1--:- __
u 103 10
4
10
5

REVERSALS TO FAILURE, 2Nf

Total strain amplitude versus life data for DPI (a lean-phosphorus

HSLA steel) in the as-received and strained-and-aged conditions. Data
for HI-FORM 50 (a columbium-bearing HSLA steel) are included for
comparison.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals. Metals Park OH. 1981, P 213
 5-9. AISI 50 XF Steel: Effects of Cold Deformation 155

8
'? IJo tJ>
...
0
I( A
Eeff
0.2
~ D0.4
.r
s 6 00.6
w·
0
~ 00
I-
::::i
e,

~ 4
z
cta:
I-
CIl
...J
<l: 2
l-
e
I-

103 104 105 106

REVERSALS TO FAILURE, 2Nf

8
'? Eeff
...
0
I(
A 0.20
~ 00.30
.r 0.57
s 6 D

w·
0
~
!:
...J
e,
~ 4
<l:
z
ct
a:
lii
...J 2
<l:
I- 0
o
I-
103 104 105 106
REVERSALS TO FAILURE, 2Nf

Total strain amplitude versus reversals to failure for AISI 50 XF

HSLA steel. Upper chart: after balanced biaxial stretching;
lower chart: after cold rolling.

Although the effects of prior deformation by BBS or CR on the

strain-life behavior of 50 XF were generally similar to those in 1006,
some specific differences were apparent; for example, the effect of
prior deformation was stronger for 50 XF than for 1006 in that both
the decrease in life at large strain amplitudes and the increase in life at
small strain amplitudes were greater in 50 XF than in 1006.

Source: John M. Holt and Philippe L. Charpentier, "Effect of Cold Forming on the Strain-Controlled Fatigue Properties of HSLA
Steel Sheets," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 218
156 5-10. AISI 80 OF Steel: Effects of Cold Deformation

Effective Str8in 8nd Mode

8 'lI of Deformation
b • 4
po
)( 60.0
'90.06
1 Uniaxi81
T8nsion
~ a 0.08
-r
s 6 00.16 881. Biaxial
Stretching
w·
c
::l Solid Symbols-D8ta for
I-
:::i Specimens Tr8nsverse to
a- Hot Rolling Direction
~ 4
z
«
II:
In Runoutl
..J
~ 2
b
I-

REVERSALS TO FAILURE, 2Nf

Strain-life curves after deformation for AISI 80

DF HSLA steel.

In this steel, the fatigue life appeared to remain unchanged or

to increase very slightly as a result of deformation, at least for
the effective strain levels investigated (see graph). Also, the fa-
tigue life appeared to be unaffected by the mode of deformation
and the specimen orientation.

Source: John M. Holtand Philippe L. Charpentier, "Effect of Cold Forming on the Strain-Controlled Fatigue Properties ofHSLA
Steel Sheets," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, p 218
 5-11. Comparison of Three HSLA Steel Grades, Cb, Cb-Vand 157
Cb-V-Si: Strain Life From Constant Amplitude

2,..,-----------------, 2,-----------------,

~
t t
w
c
~
:::>
I-
w
c
0.1 :::> 0.1
::::i I-
0.
::l!i ::::i
4( 0. • TOTAL
::l!i .PLASTIC
~ 4(
AELASTIC
4(
a: z
I- :;:
en a:
I-
en
0.01 0.01

•
10 3 10 4 10 5 10 6 10 7
REVERSALS TO FAILURE, 2 N f -

Strain-life curves for the Cb steel. 0.001 '-----''---------'-----'-----'-----'

10 2 10 3 10 4 10 5 10 6
REVERSALS TO FAILURE, 2 N f -

-Strain-life curves for the Cb-V steel.

~ 0.1
l!f
E
..J

~ Strain-life curves: Strain-life data from constant-

4( amplitude tests for the three steels are plotted in the
Z three charts here respectively in the form of total
:;:
a: 0.01 .1OTAL strain amplitude versus the number of complete
lii • PLASTIC reversals to failure.
AELASTIC

0.001'----'----'----'----L....:..---'
10 2 10 3 10 4 105 10 6
RI;VERSALS TO FAILURE, 2Nf

Strain-life curves for the Cb- V-Si steel.

Source: Shrikant P. Bhat, "Influence of Composition Within a Grade on the Fatigue Properties of HSLA Steels," in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 583
158 5-12. Comparison of Stress Responses:
DP1 vs DP2 Dual-Phase HSLA Steels

600
------ =........... --
_11 1
_ '==... - ....
:-..
"'- ...:=:.-~
: 5 0 ~ -

.............
:E
:::
. ..- 0.0065 --;::J" 1
... ~_ ..... -.:-----0.003
~ .. 009--o_-.>-""""_ _ ~-0.0065 0' 2
~

...~boo 0--- C : c I °0.0025

~
v
~200

STRAINED AND AGED

0'-::- '-;- '-::- '-::- ---'''--:- --1

10° 10'

Comparison of stress response of strained-and-aged DPl (a lean

phosphorus HSLA steel) with that ofDP2 (a carbon-manganese HSLA
steel) for the total strain amplitudes indicated.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention.
American Society for Metals, Metals Park OR, 1981, p 209
 5-13. Dual-Phase HSLA Steel Grade: Stress Response for 159
As-Received vs Water-Quenched

..; 0.005
~ 400'r-o-OCo--o-----..,.--<-->--e..-"d""--ov-...o-O-
6 u 0.002
'"
~
o
~300
v
::::;
v
>-
v200

Ol.- ..L..,- ....L..:-- ........-::- --'-:- --'

I 10 2 10 3
CYCLE NUMBER, N

Stress response of a water-quenched dual-phase steel in the as-received

condition for total strain amplitudes of 0.002 and 0.005.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention.
American Society for Metals, Metals Park OH. 1981, P 208
160 5-14. Dual-Phase HSLA Steel Grade: Stress Response for
As-Received vs Gas-Jet-Cooled

.500
Go
:e
E.. oo ~-----o.--o_o(a
o ..__ 0.006

=
'"
~ 3 0 0 r - - o - - _....- ....- ...._ -....._-"''-oA.-o''-_..o-_ _....._0.0025
~ I~
v
~hoo
.
>-
v

..>
e
Cl00

C
OL- ----''-;- ----'-:;- --l.-:;- --l.-:;- ---'
I
CYCLE NUMBER, N

Stress response of the gas-jet-cooled dual-phase steel in the as-received

condition for total strain amplitudes of 0.0025 and 0.006.

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals, Metals Park OH. 1981. P 208
162 5-16. Comparison of Dual-Phase HSLA Steel DP2 With HI-FORM 50

0.010 0 AS-RECEIVED
DUAL PHASE 2 1'" STRAINED AND AGED
HI-FORM SO - - - -

0.001L..:---------"---=,.-- ...L..,,--- ..J

10 3 10 4 10 5
REVERSALS TO FAILURE, 2Nf

Strain-life curves for DP2 (a carbon-manganese HSLA steel) in two

conditions compared with HI-FORM 50 (a columbium-bearing HSLA
steel).

Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals, Metals Park OH. 1981, p 214
164 5-18. Fatigue Crack Propagation Rate: Effect of Temperature for
Two HSLA Steel Grades

HSLA-1
G.S. =lOfLm
R =0.1
o 300K
t> 233K
!'/;f/:/. HSLA- 2
G.S. = 10JLrn
R=O.1
o 300K
t> 233K ,
01:
I/,~
1//,
I,
f

10 ,
Cl 173K .
/
o 173K o~
Q)
o 123K

/?~ I
'/:
1/ t>A Q)
0123K~I1!I: ',,
(PlI
, I
u o I
>. >.
o 0,: t>. o
<,
"-
E 1/ / E b
I: •
g' i Prof
, 6'
01 '

.,pI
I ,
/ t>,
I ,
,
/,/ Pi
P, T(K) n
T(K) n 300 7.6
300 3.6 233 8.8
233 5.1 173 12,1
173 6.5 123 15.6
123 10.8

14 18 22 26 30 36
6K, MPa -m 1/2

The effect of test temperature on the fatigue crack propagation rates in the Paris law regime for two HSLA
steels in the as-received condition.

The only significant difference between HSLA-I and HSLA-2 is that HSLA-2 contains double the
amount of Nb that HSLA-I contains (see compositions on p 165).
The effect of temperature is seen to decrease the crack propagation rate with decreasing temperature
at low values of 6.K. However, as the stress intensity increases, a crossover occurs wherein higher
growth rates were observed, as shown in the above charts. This crossover is further reflected in the
increase in the Paris law exponent, n, where it ranged from 3.6at room temperature to 10.8at l23K for
HSLA-l. The large increase is a result ofthe change in the fracture mechanism from ductile transgran-
ular fracture to cleavage. This behavior has also been seen in iron binary alloys where n increased from
3.5 at room temperature to 20.9 at l23K.

Source: Khlefa A. Esaklul, William W. Gerberich and James P. Lucas, "Near-Threshold Behavior of HSLA Steels." in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, p 569
 5-19. Effect of R-Ratio and Test Temperature on Crack Propagation 165
of H SLA Steel Grade 1

HSLA-1 HSLA-1
G.S. 10pm G.S. 10pm
T-300K T -123K
• R-O.IO • R-O.IO
10R-0.35 10' 10 R-
0.35
• R-0.70 • R-0.70

z e.
~
c lO
-,
"0

4 II II 7 II 9 10 12 14 It III 20 24 4 II II 7 II 910 1214161920 24

6K, MPa-m" 2 6K, MPa-m l12

The effect of R-ratio on fatigue crack propagation behavior of HSLA-l at test temperature of 300
and 123K in the as-received condition.

Compositions of HSLA-l and HSLA-2

Alloy C Mn Nb SI P S Al Ni Cr Fe

HSLA-1 .. 0.07 0.51 0.014 0.03 <0.005 0.005 0.01 0.01 Rem
HSLA-2 . 0.06 0.35 0.03 0.03 0.01 0.01 0.01 Rem

Source: Khlefa A. Esaklul, William W. Gerberich and James P. Lucas, "Near-Threshold Behavior of HSLA Steels," in HSLA
Steels-Technology & Applications. American Society for Metals, Metals Park OH, 1984, p 571
166 5-20. Effect of Test Temperature on Fatigue Crack Propagation
Behavior for Two HSLA Steel Grades

HSLA-1
G.S. IOlJm HSLA-2
G.S. 10JLm
R-O.I
R' 0.1
e300K o 300K
.233 K ~ 233K
.173 K II)
o 173K
.123 K u>-
u o 123K
.....
E

4 1I 6 7 8 9 10 12 14 16 18 20 24 4 5 6 7 8 9 10 12 14 16 1820 24
.6K, MPa-m"2 6K, MPa - m112

Fatigue crack propagation behavior of two HSLA steels tested at temperatures of 300,233, 173 and
123K in the as-received condition.

The only significant difference between HSLA-I and HSLA-2 is that HSLA-2 contains twice
as much Nb as HSLA-I (for compositions of the steels, see p 165).
Near-threshold crack growth and threshold stress intensities for both steels in the as-received
condition are depicted in the above charts for all test temperatures. Comparison of crack growth
rates and threshold stress intensities at room temperature indicate that HSLA-2 has a higher
resistance to fatigue crack propagation than HSLA-l. The stress intensities amplitude, 11K, for
constant growth rates of IO- s and 10-9 ta] cycle are 2.0-2.5 MPa-m 1/2 higher in HSLA-2 than in
HSLA-l. The threshold stress intensityl1K'h' is also higher for HSLA-2 (8.0 MPa-m 1/ 2) com-
pared to HSLA-I (5.5 MPa-m 1/2). The 2.0-2.5 MPa-m 1/2 difference in threshold and for the two
growth rates clearly demonstrates that there is an inherent difference in the fatigue crack propa-
gation behavior of these two HSLA steels. This difference is also reflected at low temperatures,
where HSLA-2 showed lower crack propagation rates and higher threshold stress intensities
than HSLA-I. Furthermore, by comparison of threshold stress intensities for these two steels in
relation to the effect of decreasing temperature on increasing 11K/I" it was found that the ratios
of I1K,h (D t 11K,,, (300K) are the same for both steels.

Source: Khlefa A. Esaklul, WilliarnW. Gerberich and James P. Lucas, "Near-Threshold Behavior of HSLA Steels," in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 569
 5-21. Stress-Cycle Curves for Weldments of Different 167
HSLA Steel Grades

I~
lHlER OYIW.IC
S"mESS - LMTS OF WB.DED SAt.flI.ES lCW> - cotI:XTJ:)N:S
'II'EI.~E

""'-'~"---dE]r
oOJtrElDl!1 lotAlERrAl
.~:~,:,., =Jl . I~.~ ~ USIW 22->€.....

'..po'- '.
~ ""
OSIE 3eON
~~ ~
-'ed'" , "
5' 52.
(00 ~~~ I
IoU5

JO~

.=-
2C~

on
'<:
....., ~
...... ~ ~--.....--.:.
<,
10:
.....,~ .q
....
~ - - - f=:-~
~-..::. 00

I-- ~

I I i
$~ ,,~ NJIIB[A ~ CYCLES "

Stress-cycles curves of welded samples of different mate-

rials under tension load.

T~
"._.-"---dEl
lHlER OYIW.IC
S"mESS - LMTS OF WB.DED SAt.flI.ES LOAD - 00NXT10N9

, " ~.
...,•.,M"A.E- wac, WATEAW.
:;).('..,100 ,.-1 . I~%
, USIW 22J.E"'B

""
,

i~!
E

I~ .. tiS
,'·0;;',
'-PO -, -,
OStEJOOn-l
5152.3
'00 •
...::""
)00

""" "- 1--.-

."" F--",,~
",
~~
~
---
.";" c-c
"'''''':
0-

0=
.",
.-

-
I I
sc.J
"".,.B£R ~ CYCLES

Stress-cycles curves of welded samples of different mate-

rials under alternating bending load.

Fatigue data were derived from testing welded samples under tension and bend-
ing loads. It was surprising that under both types of load the HSLA steel and the
soft unalloyed steel hardly differed in fatigue strength; thus it can be said that the
use of HSLA steels is not justified if a component has a weld in the highest-stressed
area. An explanation for this is the loss of the thermal-mechanical effect, which is
responsible for increased strength, by the heat influence during the welding opera-
tion; and it is thought that a higher-strength manganese-alloyed steel, such as St
52.3 (according to DIN), the strength of which results from the chemical composi-
tion, would be more favorable in this respect.

Source: Klaus E. Richter, "Cold and Hot-Rolled Microalloyed Steel Sheets in Opel Cars-Experience and Applications,"in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 487
168 5-22. Weldments (FCAW): SAE 980 X Steel vs 1006

R=0.1. toO.13" 13.3mml

--6-- SAE-1006. Smooth
-'-0-'- SAE·1006. FCAW
- - . - - SAE-980X. Smooth
- - - - SAE-980X. FCAW

- ..----.-........ -..
----.1;,.1---...- __
•
-----6- -'"l:>"'&--C'IS"" 6-"tl. _

10 5 106
NT' CYCLES TO FAILURE
Fatigue properties of smooth and FCA W SAE 1006 and SAE 980 X steels.

The fatigue strengths of the smooth HSLA steel were higher than that of the low-carbon steel.
The I06-cyclefatigue limit stress of the smooth SAE 980 X steel was 469 MPa (68 ksi) and that
for the SAE 1006 steel was 283 MPa (41 ksi).
However, after welding, SAE 1006 and SAE 980 X steels exhibited similar fatigue properties
over the 104-1 06-cyclelife range studied. The I06-cycle fatigue limit stresses for FCAW SAE 1006
and SAE 980 X steels were between 114 MPa (16.5 ksi) and 117 MPa (17 ksi).

Source: Kon-Mei W. Ewing. Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue of TIG-
Dressed SAE-980X HSLA Steel." in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH,
1984, P 556
 5-23. Weldments (TIG): DOMEX 640 XP Steel Welded Joints vs 169
Parent Metal

1000
000
"00
100 r---~ __
600 Parent IIletal

sou

400
TIC-treated bull
weld
JOO

!
TlG·treUed fillet
weld
~ '.00
Untreated butt weld

Untreated fillet

100

50 '-- .L- .L- -'- .L..-........

105 106
CYClES TO FAILURE

Fatigue strength for DOMEX 640 XP. Standard-Wohler-

diagram (log-log scale) with pulsating load (R=min
stress/max stress=O). Sheet thickness 5 mm and ultimate
tensile strength 767 MPa. .

For unwelded parent metal the fatigue strength of a steel is

improved with increasing static strength. For welded joints the
fatigue strength in the endurance range 105-2 X 106 is mainly
dependent upon the weld geometry and is therefore roughly the
same irrespective of the static strength of the steels. For making
full use of an increased static strength for a steel subjected to
severe fatigue, special attention must be paid to the configura-
tion of the welds. After welding, grinding or TI G-treatment can
be used to improve the weld geometry. The notch effect at the
weld toe is decreased and the fatigue properties can be im-
proved. Another solution is to place the welds in areas where the
stresses are low.

Source: Tony Nilsson, "Formable Hot-Rolled Steel With Increased Strength," in HSLA Steels->Technology & Applications,
American Society for Metals, Metals Park OH, 1984, P 259
170 5-24. Weldments (FCAW Dressed by TIG): Fatigue Life Estimates
Compared With Experimental Data for SAE 980 X Steel

103
2
10
--.:::::::::::: ..•••. (U r =-87KSI
-,
~

CJ)
~ .e'!: J!.
•
en
<I
10 1
TIG DRESSED SAE 980 LAP-SHEAR WELDS
. ..'=ur = 87 KSI
~ 10 2
:E
en
<I
Kf = 2.52 I R = 0.1
•EXPERIMENT
- - PREDICTION

101
I 3
10 105 106 108
NT, CYCLES
Total fatigue life estimates compared to the experimental data for the FCA WITIG-dressed
SAE 980 X steel.

It should be emphasized that life estimates made on the FCAW{TlG-dressed welds were
based on geometry changes brought about by TlG-dressing. The other beneficial effects such as
removal of slag intrusions and inclusions were not considered. The close agreement between the
calculated and observed long-life fatigue properties suggested that the majority of fatigue im-
provement seen in TIG-dressed joints was attributable to the geometry change. The smaller
flank angle contributed significantly to the increased fatigue strengths of TIG-dressed
weldments.

Source: Kon-Mei W. Ewing, Pei-Chung Wang, Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue of T1G-
Dressed SAE-980X HSLA Steel," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH,
1984, P 563
 5-25. SAE 980 X Steel Weldment (FCAW): Smooth Specimen vs 171
TIG-Dressed vs As-Welded

SAE-980X
R= 0.1, t= 0.13" 13.3mml
--06--- Smooth Specimen
---0--- TIG-Dressed
_.-(}-.- As-Welded

~-------4..II __ ~
-----ZP-n---..I\
~o
o '-'il... 0
~00o0 o
,-.~ 0 0
-0....
o Jgt. '1I'o-."'"b- 0
'[jCD-.,_. 0
-c._
00
o
'- ''''"00'_
,-"",..It.0
o'-'n-
0

10 5 10 6
NT' CYCLES TO FAILURE
Fatigue properties of FCA WITI G-dressed SAE 980 X steel compared to the smooth speci-
men and as-welded data, From these data, a significant improvement in fatigue characteristics
can be obtained by TlG-dressing the welds,

Source: Kon-Mei W. Ewing. Pei-Chung Wang, Frederick V. Lawrence. Jr., and Albert F. Houchens, "Weld Fatigue of TIG-
Dressed SAE-980X HSLA Steel," in HSLA Steels-Technology & Applications, American Society for Metals. Metals Park OH,
1984. P 558
172 5-26. SAE 980 X Steel Weldment (FCAW): Lap-Shear Joints

en
<]
10 1
SAE980 LAP-SHEAR WELDS
Kfmax =3.49 I R = 0.1
• EXPERIMENT
- PREDICTION

105 106
NT, CYCLES
Total fatigue life predictions and experimental results for the FCAW, SAE 980 X lap-shear joints.

Source: Kon-Mei W. Ewing, Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue of TIG-
Dressed SAE-980X HSLA Steel,"in HSLA Steels-Technology and Applications, American Society for Metals, Metals Park OH.
1984, p 562
 5-27. Microalloyed HSLA Steels: Properties of Fusion Welds 173

a: WELDED SN.f'lES I~
lHlER OYNAlolC
STRESS - LMTS COIClIJ1DN8

"-'-dEl
Lp"o -
. v; ...~ :.~"",.TC'SS'O\·T~ST
w(lO~O'"" oi~hO£l ......l[R'.l1.
10 )10.5
",-
I~""
·.• ':._':'0 .~ . I l:SIW n."E."~'"

~~
'0(>0'· -. 051E 38CN

::. ~
" 101" , , ",
"00', >l "3
'00

-
iec

-=
lC~

nn
'l;:
:.-........-.:..: ~
I~:
'~ ......,
'-.I,!~
~ ..... -...;:..-" ~~ 00
q
.............
'--- -- c«
, , I I I I I II
~~ -;.' I lA.IM8ER(JFC'fC~t'S_

Stress-cycles curves of fusion welded samples of different

materials under tension load.

'-~
lHlER OYNAlolC
STRESS - LMTS a: WELDED SN.f'lES COIClIJ1DN8

"-"-dEi
Lp"o -

1.C.... '00 ..
".~' C' lQl.Orh(i· BfK>'Mj·Y«=ST
10 )IQ.S
~~~-"£'S~~n
I~% ~%
IUTERlAl

USIW 22."'£.',

.. n '. o~· -. OSIE380T'4

,co
~~: ,.11' ' 10&1 0'"
", SI S23

'00 '-="
""""
~--'= =---=
'" f"........C'l>~

---
'00

~~ r-''!!::='- .. -
~
~ "=
I
-- --
-
"., , I I III II
I fUo'8(A OF CVCl£S- ~

Stress-cycles curves of fusion welded samples of different

materials under alternating bending load.

Due to the preferred crack location in the welded areas, it

seemed necessary to examine the dynamic strength of fusion
welded joints of HSLA steels in more detail, testing the steel
used for the crossmember at a minimum yield strength of 380
N /mm 2, in comparison to a soft unalloyed, hot rolled steel
sheet. Tensile load and alternating bending load were the se-
lected types for dynamic test. The above charts show the respec-
tive stress-cycle curves.

Source: Klaus E. Richter, "Cold and Hot-Rolled Microalloyed Steel Sheets in Opel Cars-Experience and Applications,"in HSLA
Steels-s-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 487
174 5-28. Microalloyed HSLA Steels: Properties of Spot Welds

LOAD AMPLITUDE
6P, kips
AND NUGGET ROTATION
68 N , degree

5.0
0.2
1.0
0.5

GroupC

Fatigue test results for the 2.06 mm (0.081 inch) thick

sheet with various weld diameters.

5.0
LOAD AMPLITUDE 2.0
6P. kips 1.0 LOAD
AND NUGGET ROTATION 0.5 AMPLITUDE
68 N • degree 6P, kN

5.0
2.0
1.0
0.5

Group K

4 5 6 7
LOG CYCLES TO FAILURE, N,

Fatigue test results for 1.02 mm (0.040 inch) thick sheet of

different stiffnesses.

Results of spot-weld fatigue tests are presented in the four plots (above and
on the facing page) for the stated conditions. Each curve shows the load
amplitude, I1P, and nugget rotation values, 116 N' for each test as a function of
cyclic life. Straight lines were fitted through the data.
 1·~oe""",roOlI" I
175
LOAD AMPLITUDE 5.0 o ~::S';=I~~:"
dP, kips LOAD
AND NUGGET ROTATION 2.0 AMPLITUDE
d(~N' degree 1.0 dP, kN
0.5
Group E

5.0
2.0
1.0
0.5

5.0
d0
0.2 dP
1.0
0.5

Fatigue test results for variations in specimen width and

thickness.

5.0
2.0
LOAD AMPLITUDE
dP, kips 1.0
AND NUGGET ROTATION 0.5
:>SN. degree

~
-- - -
"W-IOI61M11~O"1
' 02 _ " " . '
5.0 o 0-66In"lf02fj1fl1

2.0
1.0 ::~'
0.5
GroupN ~
5.0
2.0 de
dP
1.0
0.5
Group 0

5.0
2.0
1.0
0.5

1,.1 0211Yn1004..., ,
W .. l01611'WT1140n,
0 .. 66 ........,0416 ... ,
dP
de

4 5 6 7
LOG CYCLES TO FAILURE. Nf

Fatigue test results for 1.02 mm (0.040 inch) thick sheet

with single and multiple welds.

Source: James A. Davidson, "Design-Related Methodology to Determine the Fatigue Life and Related Failure Mode of Spot-
Welded Sheet Steels," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, p 542
176 6-1. HY-130 Steel: Effect of Notch Radii

~::;;
b 800
<l 600
w·
g 400
I-
:J
Q.
9.5mm
~ 200
6.4
~
0:100
3.2
tii80 1.6
~ 60 0.80
«z
40
~z
0.40
0.20
20

2 3
2 3 4 5 6 8 10 2 3 4 5 6 8 10 2 3 4 5 6 8 10
CYCLES TOFATIGUE-CRACK-INITIATION, NjXI03

Cycles to fatigue-crack initiation versus nominal stress amplitude, ~a,

for notched specimens with various radii of curvature.

~:::!:
4000
3000
2000

1000
~ 800
" 600 p=0.20mm
~ 500
<l 400
300
200

2 3 4 5 6 8 10 2 3 4 56 8 I 2 3 4 56 8 I~
CYCLES TO FATIGUE-CRACK-INITIATION, NI xI0 3

Same data as in upper graph but plotted versus Ill( /.JP rather than ~a.
Curvature ofthe notch and I:1Kis the stress-intensity amplitude computed for an imagi-
nary crack whose length is the same as the notch depth, a. Barsom and McNicol used this
parameter to compare N j , the cycles to fatigue-crack initiation, in HY-130 steel for
notches of constant depth but various radii of curvature. The results are shown in the
above graphs. In the upper graph, N, is plotted versus l:1a, where N, is defined as the
number of cycles to give a 50-J..!m side notch. There is a wide spread in the curves. As
expected, the sharpest notch, lowest p, gave the most rapid initiation at a given stress. The
lower graph shows I:1K/ vp plotted versus N;. A narrow-spread family of curves results;
these converge as the value of I:1K/vp is decreased to a threshold value I:1K/ vp I'h' the
minimum value to initiate fatigue cracks in notches.

Source: M. E. Fine and R. O. Ritchie, "Fatigue-Crack Initiation and Near-Threshold Crack Growth," in Fatigue and Microstruc-
ture, American Society for Metals, Metals Park OR, 1979, pp 256-257
 6-2. 300 M Steel: Effect of Notch Severity on 177
Constant-Lifetime Behavior

Minimum stress,ksi

300

s:::;: 1500 ~
~.

~E 200 1;;
E
:>
:> E
E .~
.~
::;:
::;: 1000

100
Notch
500 severity
(107 cycle
IifeUmel

_~00'::0:----...,."..::-----,.L..._-_--L._------"L-_-_.L-_-_--L._-----,,------..,..L...------'25~0
-1500 -1000 --500 1000 1500 2000

Constant-lifetime fatigue diagram for 300 M alloy steel, hardened and tempered to a tensile strength of 1930
MPa (280 ksi). Solid lines represent lifetimes obtained from unnotched specimens. Dashed lines represent
lifetime of ten million cycles for specimens having the indicated notch severity.

Source: Metals Handbook. 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals. Metals
Park OH, 1978, P 670
178 6-3. TRIP Steels Compared With Other High-Strength Grades

MN/m2
IlXXl 1200 1400 1600 1800 2lXXl 2200

200 1400
R • 0.1
TRIP
~175 1200
i
~
u
~150
.,. SR 4340
IlXXl"/;

~ i
~125
.
v;
:J
.2'
4l«l 800

.l!loo
600

Fatigue strength at 10' cycles (R = 0.1) vs ultimate tensile strength for

TRIP steels compared with other high-strength steels.

Studies on fatigue-crack propagation (FCP) conducted under controlled stress-

intensity amplitude (~K) conditions indicate that deformation-induced transforma-
tion retards crack growth in lower-strength metastable austenites, particularly at low
~K, and also exerts a beneficial influence in high-strength TRIP steels, although to a
much lesser extent. This growth retardation may be due to crack-closure effects arising
from the transformation volume change, which may be particularly effective in the
fatigue-threshold regime. Smooth-bar fatigue properties appear to be dominated by
transformation hardening, which is desirable under stress-control conditions (reduc-
ing strain amplitude) but generally undesirable under strain-control conditions (in-
creasing stress amplitude). In lower-strength austenites, transformation reduces fa-
tigue life under conditions of controlled plastic strain amplitude; under controlled
total strain amplitude, transformation is detrimental to low-cycle fatigue life, but a
small amount of transformation may be beneficial at high cycles. Similarly, the low-
cycle fatigue properties of high-strength TRIP steels are found to be degraded by trans-
formation under controlled total strain amplitude. Under stress control, the fatigue life
of lower-strength austenites is greatly enhanced by transformation; for a stress ratio
(R= amin/ a m• x ) of 0, fatigue limits in excess of the yield strength are observed. Investi-
gation of the smooth-bar fatigue properties of high-strength TRIP steels at R= 0.1, in
which thermodynamic stability was varied by heat treatment, also revealed transfer-
mation enhancement of fatigue life. Such enhancement allows the achievement of ex-
ceptional fatigue strength at high ultimate strength levels, as illustrated by comparison
with other high-strength steels in the above graph.

Source: G. B. Olson, "Transformation Plasticity and the Stability of Plastic Flow," in Deformation, Processing, and Structure,
George Krauss, Ed.. American Society for Metals, Metals Park OH, 1984, p 419
 6-4. Corrosion Fatigue: Special High-Strength Sucker-Rod Material 179

10.000

80,000

.Q,

UI
50,000
UI
....
a:
l-
UI 40.000
B
....0
::i
lL
lL 30.000 F
ce
AB- IN AIR
20,000 C0 - IN ACID BRINE 0
EF- IN ACID BRINE WITH INHIBITOR
10.000
102 t04 10~ 10' 10' to'
NUMBER OF CYCLES FAILURE

Effect of corrosion and corrosion inhibitors on the SoN curve for high-
strength steel (sucker-rod material).

After the first brittle crack is initiated, No.2 is the slow step in the
process and electrochemical action is the slowest part of this step.
Thus, the effect of corrosion can be illustrated with curves of stress vs
logarithm-of-number-of-reversals-to-failure for sucker-rod steel. Cor-
rosion accelerates cracks propagation, so the fatigue curve drops from
AB to CD, as shown in the graph. Deceleration of the slow stage with a
corrosion inhibitor will raise the S-Nfatigue curve from CD to EF.

Source: Joseph F. Chittum, "Corrosion Fatigue Cracking of Oil Well Sucker Rods," in Corrosion: Source Book, Seymour K.
Coburn, Ed., American Society for Metals, Metals Park, OH, 1984, P 380
180 6-5. Corrosion Fatigue Cracking of Sucker-Rod Material

.8

Ul
~.6
0
~
Ii
0
ct:
~.4
...
0
:J:
I- I
~.2 I
I
UJ
oJ II
I
I
II
0 !THIRD,
FIRST STAGE SECOND STAGE ISTAGEI
0
RELATIVE NUMBER OF CYCLES, PERCENT OF FAILURE

Corrosion fatigue cracking of sucker rods.

This graph shows typical progress of a crack at high stress plotted

against number of cycles, showing stages in the fatigue process. Obser-
vations of sucker-rod crack penetration as a function of reversal ac-
cumulation are possible using a bending apparatus and a magnetic
fluorescent powder technique. Penetration vs reversal curves resemble
the one shown above when the stress is well in excess ofthe endurance
limit.
During bending, no penetration is apparent in the first 40-60% of the
specimen's fatigue life, even though intrusions and extrusions may
form earlier. A crack eventually appears and progresses through the
specimen. When the penetration reaches a certain percentage of the
cross section, the cracking accelerates until catastrophic failure occurs.

Source: Joseph F. Chittum, "Corrosion Fatigue Cracking of Oil Well Sucker Rods," in Corrosion: Source Book, Seymour K.
Coburn, Ed., American Society of Metals, Metals Park OH, 1984, P 378
 6-6. Hydrogenated Steel: Effect of Baking Time on 181
Hydrogen Concentration

Normal Notch Strength = 300,000 psi

o.. ~
300
i---:::::::::::: i'... ~.
+
+ :---:- +
Uncharged
+--

-,
i'-..
275
o~ ~~
\
+-

250 " Bake 24 h r -

---
a. 225
(f)
"\\'\ 0\ \
<, Bake 18 hr

~
0
1
0
0 200 .- ,~
-
(f)
(f)
-\ \ 1 Bake 12 hr
-...
Q)

( /)
175

\ \
-\ ~
"0
150

\
Q)

a.
a. Bake 7 hr
<t 125

100
l
-,.
..1 \
~
Bake 3 hr
<-
--
75 • Bake 0.5 hr
1-
-
50
0.01 0.1 1 10 100 1000
Fracture Time, Hours
Static fatigue curves for various hydrogen concentrations obtained by baking different times
at 150°C (300 OF). Sharp-notch specimens. 230,000 psi strength level.

These are, in essence, static fatigue curves, and the lower critical stress may be considered a
static endurance limit-that is, a stress below which failure will not occur for an indefinite period
of time. This behavior is sensitive to hydrogen concentration as shown above, where it may be
seen that all delayed-failure parameters-notch strength, rupture time, and static fatigue limit
increase with decreasing hydrogen concentration. Also, even after 24 hours at 150°C (300 OF),
there is still a substantial stress range, of the order of 60,000 psi, over which delayed failure will
occur. In an unnotched specimen, full recovery ofthe ductility as measured by the reduction of
area can be attained in less than 20 hours at 150°C (300 OF), yet delayed failure will occur after
24 hours or longer of baking time at 150 °C (300 OF).

Source: Alexander R. Troiano. "The Role of Hydrogen and Other Interstitials in the Mechanical BehaviorofMetals,"in Hydrogen
Damage Source Book, Cedric D. Beachem, Ed., American Society for Metals, Metals Park OH, 1977, P 154
182 6-7. Hydrogenated Steel: Effect of Notch Sharpness

300

275
T~.
250
I~~ 1\
0
0
en
0.. 225 a
'\
~\
-a_
""'"
0
01'\. Notch Radius = 0.25 in.

; a1lus
J. = 2'In.
>--
>--
0 200 Notch
en
en
175
i'\ \\
-
Q)

\ ~\\
L..

( J)
Notch Radiu's = 0.020 in.
"0 150 -\.. Notch Radiu~ = 0.010 in. It-- - - -

\
Q)

0.. >--
0..
« 125
\
100

75 . ~. • Notch Radius = 0.001 in .

.....-
50
0.01 0.1 I 10 100 1000
Fracture Time, Hours

Static fatigue curves for specimens of different notch sharpness. Baked 0.5 hour at 150°C
(300 oF).

The variation of lower critical stress with notch severity is shown in this diagram. It is evident
that the static fatigue limit rises as notch severity (radius) decreases for hydrogen-charged
high-strength steels (using the same baking time).

Source: Alexander R. Troiano, "The Role of Hydrogen and Other Interstitials in the Mechanical Behavior of Metals," in Hydrogen
Damage Source Book, Cedric D. Beachem, Ed., American Society for Metals, Metals Park OH, 1977, P 155
 7-1. O.5%Mo Steel: Effect of Hold Time in Air and Vacuum at 183
Different Temperatures

+ 10 I
- KEY CYCliNG I::NV. TEMP.
CONTINUOUS AIR 275 K
..J VAC . ..
< AIR 775 K
l-
....
....... ' ........
e VAC. II
l-
LL
o loO
~ 10 0
.. =:::::.:.~.........:.~ . ... :--;:
~
30 on HOLD AIR
VAC ...
..........,,:~:-:-: ....
'
w Z

__
l.'
<
Z IX: " .""'"
".. . . ..«:»r.
. . :--;; -
. "=':..,:-:-
<
IX: I-
r
w
VI
VI " .',~""-
--............:..
........ --.
.....::---
._~.

CYCLES TO FAILURE

Effect of hold time in air and vacuum upon the fatigue endurances of a
O.5%Mo steel at 275 and 775 K.

Source: R. H. Cook and R. P. Skelton, "Environment-Dependence of the Mechanical Properties of Metal sal High Temperature,"
in Source Book on Materials for Elevated- Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals
Park OH, 1979, p 83
184 7.2 DIN 14 Steel (1.5 Cr, 0.90 Mo, 0.25 V): Effect of
Liquid Nitriding

1000

'\
- 140
900 :'\.

""-8 - 120
800

700

I'
- 100 ';;;

. .
-'"

~ 600 0- ~
.s A .~
13 - 80 13
>-
IS U
500

400 \ - 60
\
300 ~
\
C - 40

105 106 107

Number of load cycles

Effect of nit riding on fatigue behavior of DIN 14 CrMoV

69 steel (0.14 C, 1.5 Cr, 0.90 Mo, 0.25 V). Curves A and C
are for hardened and tempered (not nitrided) specimens;
Band D are for liquid nitrided specimens. A and B are for
smooth specimens; C and D are for notched specimens
K,=2.

Nitriding introduces residual compressive stresses at the

surface of steel parts; these residual stresses, together with the
increased strength of the nitrided layer, increase the fatigue
resistance of the part. The increase in fatigue strength that re-
sults from nitriding is illustrated in these S-N curves.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 541
 7-3. 2.25Cr-1.0Mo Steel: Influence of Cyclic Strain Range 185
on Endurance Limit in Various Environments

10.0
8.0

6.0 KEY ENVIRONMENT

• Na (30 ppm OF °21
U X AIR
• HELIUM

-.
3.0 • Na \300 ppm OF 02)

2.0 ~XX . ,

z
..:
'~
x~ 'U ...~~~.
a: 1.0
f-
VI
u 0.8
........~
-~
.........=-
.J
u 0.6
>- - .. ~
u 11K
.... •
0.4

0.3

0.2

C.I
10~

CYCLES TO FAILURE

Influence of cyclicstrain range upon fatigue endurance of 2.25Cr-l.OMo steel in

sodium, air, and helium at 865 K. (Cycle used was approximately up 5 s, hold
5 s, down 5 s, hold 5 s.)

Source: R. H. Cook and R. P. Skelton, "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
in Source Book on Materials for Elevated-Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals
Park OH, 1979, P 83
186 7-4. 2.25Cr-1.0Mo Steel: Effect of Elevated Temperature

Testing temperature
~C °FI
o 425
, 800,
?f!. • 540 1000
OJ' 2
"Q
~
l1595 1100
.'=
a.
E
'"
c
'iii
J5
0.2

0.1 2
10 103
Cycles to failure

The results of strain-controlled fatigue tests of 2.25Cr-l.OMo

steel at 425, 540 and 595°C (800, 1000 and 1100 OF) on speci-
mens of annealed 2.25Cr-I.OMo steel are presented in these S-N
curves. Within this range, test temperature had relatively little
effect on number of cycles to failure.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 659
 7-5. 2.25Cr-1.0Mo Steel: Effect of Elevated Temperature 187
and Strain Rate

Temperature,Oc

e
~

104
:e
B
:G
u
>
u

103 350 400 450 500 550

Temperature,OC

I I I
600 700 800 900 1000 1100
Temperature, of

Effect of elevated temperature on strain-controlled fatigue behavior of annealed 2.25Cr-l.OMo steel.

Strain-controlled fatigue tests have also shown (note above) that reducing
carbon content to 0.03% results in a reduction in fatigue strength. Further-
more, because of variations in strain aging effect, specimens from one heat
with a higher carbon content ran longer at 427°C (800 OF) than at 316 °C
(600 OF).

Source: Metals Handbook, 91h Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 659
188 7-6. 2.25Cr-1.0Mo Steel: Effect of Temperature on Fatigue Crack
Growth Rate

6K, stress intensity factor, ksi..Jiil.

10 20 60
10.4
I--+-------+-------+--------+-----i
Testing temperature
10-31--+--oC- of - - - + - - - - - - - + - - - - - - - - - - - f + - - - l

• 205 400
o 370 700
1---+-6-455-850---+--------+--------,.~____J<-----,_F____I1____l

6595 1100

10-6

Q>
U Q>

~ U
E 10-4 ~
E .£
:i
\!
s: ~
s:
~ 6 ~
...e
.><
0
l;,
u .><
eu l;l
10.6 1;
:e' :e'
..
:!'!
"D .
:!'!
"D

10-6

10 20 60
s«, stress intensity factor, MPa vmm
Variations in fatigue crack growth rate with test temperature for specimens of 2.25Cr-l.OMo steel tested in air.

Specimens were subjected to cyclic loading at a constant maximum load.

Stress ratio was 0.05; cyclic frequency was 400 per minute. As shown, the
stress-intensity factor range increased as the crack length was increased.

Source: Metals Handbook. 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978. P 660
 7-7. 2.25Cr-1.0Mo Steel: Effect of Cyclic Frequency 189
on Fatigue Crack Growth Rate

b.K, siren intemity laClor. ksi...rm:

2 5
2

I 0-4

4C"7f'/ 5

5
/,r:/
/ ! vr:~400C~ I

40 CfKO
~
5
/ 2

Frequency, 1
cvcles/mln
o 4
• 40 6

2 5
ilK. stressintensity laflOr, MPaym
(.1
IJJ.K. stress intensity factor, ksi vrn:
2 5
2

I
2

3 1-// 5

. 5
III 2

4C~
~E
E
2
1/ I

II;
t
5

40c"m
400c:pm
2

Frequency. I
cycles/min
o 4
• 40 5

2 5
4K. slreu intensily 'actor. MPa V;;
Ibl

Data shown above indicate that in elevated temperature tests at a given

stress-intensity factor range, crack growth rate increases as cyclic frequency
is decreased. These fracture mechanics data may be applied to the design of
structural components that may contain undetected discontinuities, or that
may develop cracks in service. Stress ratio was 0.05. (a) Tested at 510 °C
(950 OF); (b) tested at 595°C (1100 OF).

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society ForMetals, Metals
Park OH, 1978, p 661
190 7-8. 2.25Cr-1.0Mo Steel: Fatigue Crack Growth Rates in Air
and Hydrogen

10-5
I

10-6 v
~ ,• -

;. ••
i>.
i>.V
Gl D
1i vi>.
~
-. Vi>.
i>.
E 10-7 .Jl ~
z· i>. ~
-
't:l
-. ~
'"
't:l
0'
.....
al
oil'
'"
s: 10-8 .dt -
i0 •
.Ol
.l' o~ ,
°•
oOe

U o , 2% Cr - 1 Mo Steel
~
0°.,.
f: R = 0.05
u
Gl 10-9 l- -
:::I
Ol
.;::; Frequency Environment
'"
u..
8
• • 2 Hz, 50 Hz Air

§ °0 50 H, }
10- 10 f-
• 5 Hz
138 kPa H2 -
• V
i>.
2 Hz
0.5 Hz
•
~I Threshold
10- 11 I ,f I I I J I I J I j I I I
3 4 5 6 78910 20 40 60 80 100

Stress-intensity factor range, L1K, MPa . m 1/2

Fatigue crack growth rates in 2.25Cr-l.OMo steel in air and in hydrogen.

Corrosion fatigue descriptions are further complicated by the fact that the envi-
ronment may produce multiple effects. For example, Suresh et a/demonstrated that
dry hydrogen may produce afrequency-sensitiveenvironmental effect analogous to
SCC at intermediate t:J.K values and a frequency-insensitive environmental effect
near the threshold. This is illustrated in the above graph for 2.2SCr-1Mo steel tested
in air and in 138-kPa hydrogen gas. Because the sustained-load threshold for this
steel is on the order of 90 M Pa . m 112 (82 ksi . in. 112), the K""'j of about 22 MPa .
m 1/2 ( 20 k' . 1/2) gives
SI . Ill.
•
K ,h ({) ...?
~
K th : v.

lt can be seen for t:J.Kvalues greater than K ,h ({) that there is a large increase in
growth rate for the low test frequencies but not for the higher ones. Therefore, this
regime may be considered to be one where superposition might apply. In addition,
however, there is a true threshold, t:J.K,h , which appears to be frequency-insensitive
but which nevertheless decreased by about 30% to SA MPa . m 112 (4.9 ksi in. 112 )
because ofthe hydrogen environment. Such mulitple effects are poorly understood
and are clearly possible in a large number of material/environment systems.

Source: W. W. Gerberich and A. W. Gunderson, "Design, Materials Selection and Failure Analysis," in Application of Fracture
Mechanics for Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds..
American Society for Metals, Metals Park OH, 1982, P 333
 7-9. 2.25Cr-1.0Mo Steel: Effect of Holding Time 191

120days

c
E

Compressive or
tensile hold

Type of strain hold

Strain
range.% Compressive Tensile Both
2.0
loJ 104
C.}
1.0 •
Cycles to failure 0.5 • o

:::2 } .
X indicates zero hold lime

Time-to-failure/cycles-to-failure dia-
grams for annealed 2.25Cr-l.0Mo steel
tested in strain-controlled cyclic load-
ing at (a) 480°C (900 OF) and (b) 540
°C (1000 OF). Hold time indicated on
graph is length oftime that specimens
were held (during each cycle) in the
state of maximum tensile strain (open
symbols) or compressive strain (filled
symbols). Strain amplitude indicated
by shape ofsymbols and figures along
Cycles to failure (b) zero-hold-time line.

In these "time-to-failure/cycles-to-failure" diagrams, the lowest curve (zero hold time) indicates the cor-
responding time period and number of cycles to failure for continuous strain-controlled fatigue tests over the
strain range from 0.4 to 2.0% with no holding period at maximum strain. The other curves, which are
approximately parallel, are for increasing periods of holding time at maximum strain levels in either tension
or compression. The vertical curves are drawn through the number of cycles to failure for each particular
cyclic strain. For all tests at 2% strain, failure occurred in less than I000 cycles regardless of holding time or
whether the stress was tensile or compressive. The effect of reducing the strain increment and increasing the
holding time on number of cycles to failure can be determined from the appropriate curves in the figures.

Source: Metals Handbook, 9th Edition, Volume I. Properties and Selection: Irons and Steels, American Society for Metals. Metals
Park OH. 1978. pp 662-663
192 7-10. Cast 2.25Cr-1.0Mo Steel, Centrifugally Cast: Fatigue
Properties at 540°C (1000 OF)

60,------r----r---r---r--r---.,.------,

\
5.0 • • •

I 4.0
lLJ
o
Z
<l
cr
z
3.0 •••
<r
cr
I-
(f) 2.0
...J
~
g
1.0

OL...._ _-L-_ _---1._---1._.l.-...L-_ _--'-_ _- . J

100 200 400 600 1000 2000 4000
CYCLES TO FAILURE

Fatigue properties of 2.25Cr-1.0Mo centrifugally cast pipe,

A2l7, Grade WC9, at 540°C (1000 OF).

Source: Steel Castings Handbook, 5th Edition. Peter F. Weiser, Ed.• Steel Founders' Society of America. Rocky River OH. 1980,
P 15-55
 7-11. H11 Steel: Crack Growth Rate in Water and in Water Vapor 193

0.5 ,........----,-----,--"'"""'T--"'"""'T--~--....,

0.3 "0 H -II Steel

230 ksi Y. S.
K - 30 ksi IT'"

c
E 0.1
..... o
c Activation Energy
9,000 cal./gm-atom
-Q)

o
a:::
.05

-...
s: .03
~
o •
o
.Jl:
U
o...
u .01 o Water
o Relative Humidity 100% at Test Temperature
• II II II II 80° F,
0.005 Tested at Higher Temperature

Crack growth rate versus temperature for an HII steel in water and
water vapor.

It is of considerable interest that the strain rate and temperature

dependence of hydrogen embrittlement, as determined by ductility
measurements after rising load tests on hydrogen-charged materials,
show a characteristic behavior that resembles closely that seen with
crack growth rate measurements and external hydrogen environments.

Source: Herbert H. Johnson. "Keynote Lecture: Overview on Hydrogen Degradation Phenomena," in Hydrogen Embrittlement
and Stress Corrosion Cracking, R. Gibala and R. F. Hehemann, Eds., American Society for Metals. Metals Park OH. 1984, P 18
194 7-12. 9.0Cr-1.0Mo Steel: Creep-Fatigue Characteristics

9% Cr 1% Mo AT 525°C WITH OR WITHOUT

TENSILE HOLD TIMES TO hOWELL t
(TOTAL STRAIN RANGE; FROM (22))

~ LOW C 9% Cr 2% Mo AT 550°C
UJ
CONTINUOUS CYCLING
l-' (TOTAL STRAIN RANGE; FROM (2311
Z
« 10°
cr
z
~
cr
~
Vl 9% Cr 2% Mo NbV
AT 550°C ~--5 MIN. TENSILE DWELL
(PLASTIC STRAIN
RANGE; FROM (24)) CYCLING

10'
CYCLES TO FAILURE

Illustrating the elevated temperature low-cycle fatigue and creep-fatigue prop-

erties of normalized and tempered 9% Cr Mo variants.

In this chart are presented the elevated-temperature-fatigue and creep-

fatigue data for the 9%Cr-1%Mo steel as a single curve in terms of total strain
range against cycles to failure; also shown for direct comparison are the con-
tinuous cycling fatigue data for the low-C, 9%Cr-2%Mo variant which, al-
though inferior at relatively high strain ranges, suggests superior endurance
may be attained in the high-cycle region. From the limited evidence, it seems
probable that normalized and tempered 9%Cr-I%Mo steel may be used in
reactor-quality sodium at service temperatures with little effect on tensile
properties and stress rupture strengths or ductility and that the short term
low-cycle fatigue endurance will be increased and fatigue crack growth rate
reduced. This behavior is a consequence of the structural stability of the
material with respect to interstitial element transfer in liquid sodium and also
the low oxygen potential of the overall system which may be expected to
preclude oxide penetration and enable partial recohesion of the crack faces
during fatigue.

Source: S. J. Sanderson, "Mechanical Properties and Metallurgy of 9%Cr I%Mo Steel," in Ferritic Steels for High-Temperature
Applications, Ashok K. Khare, Ed., American Society for Metals, Metals Park OH, 1983, P 95
 7-13. 9.0Cr-1.0Mo Modified Steel: Stress Amplitudes Developed 195
in Cycling

Fe 9Cr IMo BV, Nb

--------------------------
200 t - - - - - -__
------ --- ..
----Continuous Cycle
100
e
a..
::E
en o
(/)
w
a::
l-
(/)
-100
- - with 30 sec T H
---
-200
r,.,..,,"'="::=-:-:.....---......:_:'":_;:"::-=-=_,,:_::- - - _...
10 100 1000 10000
CYCLES
This chart shows stress amplitudes (tensile and compressive)
that developed in the course of cycling the modified Fe-9.0Cr-
l.OMo steel through a total strain range of 0.5% at 649°C
(1200 OF). Fatiguing was carried out in vacuum. Dotted curve
indicates continuous cycling; solid curve indicates cycling with a
30-s hold at maximum tensile strain.

Source: S. Kim, J. R. Weertman, S. Spooner, C. J. Glinka, v. Sikka and W. B. Jones, "Microstructural Evaluation ofa Ferritic
Stainless Steel by Small Angle Neutron Scattering," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck
and Stanley M. Wolf, Eds., American Society for Metals, Metals Park OH, 1984, p 175
196 7-14. 9.0Cr-1.0Mo Modified Steel: Effect of Deformation

cb:
dQ
s Fe 9Cr IMo 8 V,Nb
0 Not deformed (N aT)
In 10 8
• 0 Fotl(~ued (649°C, A Et = 0.5 -/0,
10000 cycles, continuously
•
8
•, 6 Foti~ed ( 649°C, A€t
79
= 0.5 %,
cycles with 30 sec tension hold
6
•'i 0
Crepl R~9~~,M~fb28 holn)

4 8
8
2
~
~
0
~
~ 0

! 0
-2 8 0
S 0 0
0
0
~ ~
-4 S

0.02 0.04 0.06 0.08 q <I")

Curves of dI.ldO vs q for specimens of modified Fe-9.0Cr-1.0Mo steel which have
undergone various types of deformation. A magnetic field of ~28 kg was applied to the
specimens during the SANS measurements. A = 0.48 nm,

Source: S. Kim, J. R. Weertman, S. Spooner, C. J. Glinka, V. Sikka and W. B. Jones, "Microstructural Evaluation of a Ferritic
Stainless Steel by Small Angle Neutron Scattering," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck
and Stanley M. Wolf, Eds., American Society for Metals, Metals Park OR, 1984, P 175
 8-1. Type 301 Stainless Steel: Scatter Band for 197
Fatigue Crack Growth Rates

K.If , ksi • in.1/2

20 40 60 80 100

10- 3
/
GI
i3 /
~
...... 10- 2 /
E
E
/
z· /
......
"t:l
/ / GI
i3
'"
"t:l

......
GI' I / 10- 4 >-
u
'<,

'" / /
I .~
..c z·
i0
... 10- 3
// / "t:l
......
/ / '"
"t:l

~'"
u
...u / /
'" / /
GI
:::I / /
'"
.;; / / 10-5
'"
u.. / /
/ /
/ 1/2 hard tvpe 301
/ 24°C (75°F)
10- 4 / /
/ 0.063 < R < 0.807

L-.J'- --'-_ _.1...---J'---'---'---'-...J...........L - - J 10- 6

80
Effective stress-intensity factor,
K.1f = Kmax [1 - Rl o.66 7 , MPa • m 1/2

Scatter band offatigue crack growth rates of Y2·hardtype 301 stainless

steel, tested at 24 0 C (75 OF), 10 Hz, andR ratios of 0.063 to 0.807 based
on effective stress-intensity factor, Kerr'

Fatigue crack growth rate data reported by Walker for Y2-hard type 301 stainless steel sheet are
summarized in the above graph. The data were obtained in air at room temperature over a series ofload
ratios (R) from 0.063 to 0.807 at a frequency of 10 Hz. These data are based on the "effective stress
intensity factor," Kerr, rather than on fj,K, to account for the effect of the range of stress ratios. Kerr is
defined as follows:
Kerr = K max (I - R)m
where m is determined empirically and R is the load ratio (minimum load/maximum load) on cyclic
loading. The crack growth rate law then becomes:
da/dN= C[Kmax(l- R)my
Results of fatigue crack growth rate tests on austenitic stainless steels have shown that the crack growth
rate tends to increase as the R ratio is increased, when compared at given values of fj,K. If tests are made
at several load ratios to determine m, then the effects of other load ratios may be estimated.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
Metals. Metals Park OH, 1982, P 114
198 8-2. Type 301 Stainles Steel: Effects of Temperature and
Environment on Fatigue Crack Growth Rate

tlK, ksi . in. 1/ 2

20

10- 5 Gl
U
>
--.E
c.>

2'
"'"
I"'" --'"
"tl

"tl

10- 6
20°C (68°F)
L-T

Type 301
2 X 10- 7

30 40 50 60
Stress-intensity factor range, tl K, MPa . m 1/2

-----Annealed, tested in argon

- - - - - Annealed, tested in air
- - - - - Warm worked, tested in argon

Fatigue crack growth rates for type 301 stainless steel have been
reviewed by Pineau and Pelloux in the temperature range from - 30 to
+95 °C (-22 to +203 OF).The results, summarized in this graph, were
obtained on compact specimens 7 mm (0.28 in.) thick at a cyclic fre-
quency of 20 Hz with a sinusoidal waveform at a load ratio (R) of 0.0 1.
All specimens were tested in dry argon except one series that was tested
in laboratory air. For the annealed specimens tested in argon, fatigue
crack growth rates at a given t::.K value increased as the temperature
increased over the testing temperature range. Fatigue crack growth
rates in laboratory air at 20°C (68 OF) were higher than for corres-
ponding conditions in argon, indicating that the humidity and/ or ox-
ygen in the air influenced the growth rates.
The warm worked specimens were reduced 65% at 450 to 500°C
(840 to 930 OF), resulting in a substantial increase in strength. Fatigue
crack growth rates for the warm worked specimens (above) indicate
that the fatigue crack propagation properties of the warm worked al-
loy are different from those of the annealed alloy. This effect of warm
working has been observed for other austenitic stainless steels. These
differences are attributed to the extent of the strain-induced transfor-
mation at the crack tip. This transformation effect would be most no-
ticeable in type 301, because it is less stable than the other alloys in the
UNS S3xxxx series.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OR, 1982, P 113
 8-3. Type 304 Stainless Steel: Effect of Temperature on 199
Frequency-Modified Strains

AISI 304 STAINLESS STEEL

o • 430°C
o • 650°C
LIJ
.....J 6 .. 8160C
«
<.,;)
(f)

z
«
a:
.....
(f)

Data of Berling and Slot for AISI 304stainless steel, showing frequency-modified elastic and plastic
strains at three temperatures in air,

In contrast to most other segments of our technology, interest in the fa-

tigue problem in the power-generation industry generally involves elevated
temperature. Laboratory testing on both smooth specimens and specimens
designed for crack growth is performed with temperature and frequency or
strain rate as parameters. The importance offrequency or strain-rate effects
is shown in this chart. These data are for solution-treated AISI 304 stainless
steel subject to triangular wave shapes at equal-loading and reverse-loading
strain rates. Representation of the behavior here utilizes fatigue equations
known as frequency-modified fatigue equations. They describe the elastic
and plastic strains versus fatigue life and include the frequency or strain rate
of the test. For the present purposes they are useful in showing how increas-
ing temperature acts to change the cyclic stress-strain response and the strain-
life fatigue response of this alloy.

Source: L. F. Coffin, "Fatigue in Machines and Structures-Power Generation," in Fatigue and Microstructure, American Society
for Metals, Metals Park OH, 1979. P 13
200 8-4. Type 304 Stainless Steel: Fatigue Crack Growth Rate-
Annealed and Cold Worked

~K, ksi . in. 1/ 2

10 20 30 40 60 80 100
2 X 10-4
Cold worked 25%, /
tested at 427°C (800°F)~ I

Annealed, tested at I
I ,1:" 10-4

427°C (800°F) - /

., I I /
/

I ~//
U 10- 3
>
u
-<;
E

y/~
E
z' .,

I'
"t:l
'<, U
'"
"t:l >
u
....,'~
'<,

.~
Cold worked z'
...s:
s: / 25%, tested
10-5
"t:l
-..
e / at 25°C '"
"t:l
Cl
-"u
IJI,' (77°F)
~
.,u
::> /it'
"
Cl
'':;
'" 10- 4
U.
/
/ Annealed,
tested at
25°C (77°F)

Type 304

........ -'--_--'-_....L.._---'_....L.----lU 10- 6

10 20 30 40 60 80 100
Stress-intensity factor range, ~K, MPa • m 1/2

Fatigue crack growth rates for annealed and cold worked type
304 stainless steel at 25 and 427°C (77 and 800°F), 0.17 Hz, and
an R ratio of O.

In some applications, type 304 stainless steel components are fabricated in

the cold worked condition to improve strength properties. A comparison of
fatigue crack growth rate data by Shahinian, Watson, and Smith, illustrated
in this graph, shows that the high-~K crack growth rates were lower for the
cold worked specimens than for the annealed specimens. Crack growth rates
were higher for the specimens tested at 427°C (800 OF) than for correspond-
ing specimens tested at room temperature.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels." in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell, William w. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, p 120
 8-5. Type 304 Stainless Steel: Effect of Humidity on Fatigue Crack 201
Growth Rate

CiK, ksi . in.1/2

10 20 40 60 80 100

10- 4

10- 5

~Roomair
10- 4
/
/
/ Type 304
25°C mOF)

20 30 40
Stress-intensity factor range, CiK, MPa • m 1/2

Effect of humidity on fatigue crack growth rates for type 304

stainless steel tested at room temperature, 0,17 Hz, and an R
ratio of O.

The effects of humid air environments on the room temperature fatigue crack growth rates of
specimens of annealed type 304stainless steel are shown in the above chart for specimens cycled
at 0.17 Hz with an R ratio of zero (Shahinian, Watson, and Smith). At the lower end of the t:.K
range, fatigue crack growth rates in humid air are substantially greater than crack growth rates
in dry air. However, fatigue crack growth rates of specimens oftype 304stainless steel tested in a
pressurized water reactor environment at 260 to 315 °C (500 to 600 °F) with R ratios of O. 2 and
0.7 were no greater than the fatigue crack growth rates in air at the same temperature with an R
ratio less than 0.1 (Bamford). However, variations in R ratios influenced the fatigue crack
growth rates in the pressurized water reactor environment.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH. 1982. p 122
202 8-6. Type 304 Stainless Steel: Effect of Aging on Fatigue Crack
Growth Rate

l1K, ksi . in.1/2

10-4

10- 3

.,
u
>
--
lJ

.~
10- 5 Z
~
'"
"t:J

10- 4
Unaged Aged Hold time
o • Zero

• 0.1 min

A 1.0 min
10- 6

Type 304
593°C (1100°F)
10-5 L-..JL-.I--'- .l--_ _....L..._--'-_....I....----I_L-.L-J........J

10 20 40 60 80 100
Stress-intensity factor range, l1K, MPa . m 1/2

Effect ofaging at 593 ° C (1100 OF)for 5000 h, and hold times of 0.1 and
1.0 min for each cycle, on fatigue crack growth rates of L-T oriented
specimens of type 304 stainless steel tested in air at 0.17 Hz and an R
ratio of O.

Because austenitic stainless steels are expected to give long service life, an evaluation of
the effect oflong-time aging at service temperatures is important. Results offatigue crack
growth rate tests on specimens that were tested in the unaged and aged conditions (5000
hours at 593 °C, or 1100 OF) are shown in this graph, as reported by Michel and Smith.
After aging for 5000 hours at this temperature, precipitation of M 23C 6 carbides is essen-
tially complete. These results indicate that at 593 °C (1l00 OF) there are no deleterious
effects of aging on the crack growth rates of specimens that are continuously cycled. When
a holding time of 0.1 or 1.0 minute is included in each loading cycle, there tends to be a
slight increase in the fatigue crack growth rate at a given 11Klevel.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 121
 8-7. Type 304 Stainless Steel: Effect of Temperature on Fatigue 203
Crack Growth Rate

.1K, ksi . in,1!z

10 20 40 60 100

II>
c:;
>-
u
649O C l 1 2
y
,
10-4

....... 10-3
E 538"C (1000"-><-/
E
z·
~
"tJi'"
..
~
E
i0
.'" 10- 4
t<,
1/ .:
10- 5
.E
z·
II>
c:;
>-
u
.......

"'"
.......

"".
u
f! ~~~
"
'"

t'I~
u
II>
~'
...'"
:::l
10-6
u,'"
10- 5
316°C 1600°F) Type 304

10 20 40 60 100
Stress-intensity factor range, .1K, MPa . m l/Z

Effect of testing temperature on fatigue crack growth rates for annealed

type 304stainless steel tested in air at 0.066 Hz and anR ratio of 0 to 0.05.

Results offatigue crack growth rate tests on types 304 and 304L stainless
steel at room temperature and at elevated temperatures have been reported by
James and Schwenk, and by others. As shown in this graph, increasing the
exposure temperature from room temperature to 650 °C (1200 ° F) increases
the fatigue crack growth rates at any ~Klevel within the range ofthe tests in
an air environment. These data, reported by James and Schwenk, are for
specimens of both the L-T and T-L orientations, for several different maxi-
mum alternating loads, for load ratios of 0 to 0.05, and for cyclic frequencies
from 0.033 to 6.66 Hz for the room-temperature tests and 0.067 Hz for the
elevated-temperature tests.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park DB, 1982, P 115
204 8-8. Type 304 Stainless Steel: Damage Relation at 650°C (1200 OF)

STRAIN RATE, HOLDPERIOD, MINUTES

IN.lIN.lSEC TENSION COMPRESSION
o 4 x 10'3 o 0
o I 0
o 10 0
(). 30 0
~ 60 0
'i7 180 0
Ll o 30
-tI I I
o 3 3
Q 30 30
o 30 3
~ 4 x 10'3 0.1 0
o 4 x 10'4 0 0
(3 4 X 10'5 0 0

O"rtl€pN, I3 v l3 ( k. I ) = C

C = 1.158 x 10 5
13 = 0.895
k = 0.756

101' - : - - -.........- -.........- - ' -.........L.;:-----''------''-----'---'--'-=----'----'----'---'---'-:-----'----'--.........- I -.....

101 105

Ostergren's damage relation for AISI 304 at 650°C (1200 OF).

The damage function was proposed by Ostergren and is based on the frequency-
modified fatigue approach. A damage function is approximated by the quantity U,!:iE p ,
where u, is the maximum stress in the cycle and !:iE p is the inelastic strain range. The tensile
hysteresis energy is employed to account for the facts that low-cycle fatigue is essentially a
crack-growth process and that crack growth and damage occur only during the tensile
part of the cycle. The use of the tensile-stress quantity, in conjunction with the plastic-
strain range, provides a means of accounting for loop unbalance, since, for the same
inelastic strain, a positive mean stress provides a greater hysteresis energy than does a
compressive mean stress. The method is effective in accounting for hold-time effects, as
indicated in the chart above.

Source: L. F. Coffin. "Fatigue in Machines and Structures-Power Generation, "in Fatigue and Microstructure, American Society
for Metals, Metals Park OH, 1979, P 23
 8-9. Type 304 Stainless Steel: Fatigue Crack Growth Rate at Room 205
and Subzero Temperatures

5
10-4

10-3 1-- +-- - +-- - - - - - +-'''''10,......,

• 24°C (75 of)

o -196°C (-320 OF)
0-269 °c (-452 OF)

10-4 f--+---+-------4I::L-+------=l

5 10 50 100
Stress intensity factor
range, 11K, MPa vm
Fatigue crack growth rate data for type 304 austenitic stainless steel
(annealed) at room temperature and at subzero temperatures. For this
alloy, crack growth rates are nearly the same at room and cryogenic
temperatures.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, p 756
206 8-10. Types 304 and 304L Stainless Steel: Effect of Cryogenic
Temperatures on Fatigue Crack Growth Rate

~K. ksi • in. 1/ 2

5 20 40 60 80

Gl
U
~
......
E 10- 3
E
Z
......
"C
Cll
Gl
"C U
>
~i
E
Type 304L
22°C (72°FI
10- 5 -.
u
.E
~
...i!= z
...0
tJ)
.>t!.
--
"C
Cll
"C

u Type 304L
...u'" -196, -269°C
Gl
~ 10- 4 (-320, -452°FI
tJ)
.;;
'"
IL

10- 6

Stress-intensity factor range, ~K, MPa • m 1/2

Fatigue crack growth rates for annealed types 304 and 304L stainless
steel at room and cryogenic temperatures, 20 to 28 Hz, and an R ratio
of 0.1.

Fatigue crack growth rate data obtained by Tobler and Reed on

specimens of types 304 and 304L stainless steel (annealed) at tempera-
tures in the range from room temperature to liquid helium tempera-
ture (-269°C, or -452 OF)are shown in this graph. The data for type
304 were scattered over the range shown, while for type 304L, the data
at room temperature described one curve and the data at the cryogenic
temperatures described the other curve. These results indicate that
cryogenic fatigue crack growth rates for type 304 do not deviate signif-
icantly from room temperature fatigue crack growth rates over the
/:!"K range studied. Furthermore, if design calculations for type 304L
are based on room temperature fatigue crack growth rates, the calcula-
tions will be conservative for cryogenic exposure.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 123
 8-11. Type 304 Stainless Steel: Fatigue Crack Growth Rate in Air 207
With Variation in Waveforms

~K, ksi . in. 2

'/
10 20 30 40 50

10- 4

CIl
U
~
.......
E 10- 3
E / /
Z· /
'1::J
.......
/ /
co
'1::J
I /
2l' / / .!!
u
e II >-
u
10- 5 .......
s: / / Waveforms .E
i0 ( (
..
."
/ / rYYY\
Z·
~
~ } 0.067 Hz co
..
u
co
u 10- 4
/ / /VVV\
'1::J

CIl
:J
."
/ /
.;;
co / /
u,
/
/ Type 304
/ 538°C (1 0000 F) 10- 6
/ R = 0.05

10 20 30 40 50
Stress-intensity factor range, ~J<, MPa . m 1/2

Scatter band of fatigue crack growth rates for annealed type 304 stain-
less steel at 538 °C (IOOO°F)in air at anR ratio of 0.05 with two different
waveforms at 0.067 Hz.

The data presented in this graph were obtained in tests with a saw-
tooth waveform. Changing from a sawtooth waveform to a waveform
with a short holding period at maximum load did not influence the
overall fatigue crack growth rates according to additional data re-
ported by James and shown above.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society For
Metals, Metals Park OH, 1982,P 117
208 8-12. Type 304 Stainless Steel: Effect of Hold Time on Cycles to
Failure

30
LABORATORY TESTS
z
......
Z
~
180
AISI304 STAINLESS STEEL
650°C

ILl
z 0.01
C>
«
a:::
z 60 60 10
« o DO 01 0 o 0
a::: 30 30 I I
~
en
-J o NO HOLD TIME
~ o TENSILE HOLD TIME IN MINUTES-
g AS INDICATED

0.001 L...-_--L._ _.L..---L----L---l.-_ _.L..-_----L_..L-...l..-.L..-_-..l..._ _--'--~

100 1000 10,000
LIFE - CYCLES TO FAILURE
Effect of hold time on life for AISI 304 stainless steel.

Wave-shape effects are also important in fatigue crack

growth, as has been studied by Barsom. He observed that the
crack growth rates were greater as the loading rate increased and
the unloading rate decreased, given a fixed period of cycling.
Overload effects are also important in retarding crack growth.
Substantial damage can result from these wave shapes, particu-
larly when the hysteresis loop is severely unbalanced, as can oc-
cur in long tensile-strain hold-time tests.

Source: L. F. Coffin, "Fatigue in Machines and Structures-Power Generation," in Fatigue and Microstructure, American Society
for Metals, Metals Park OH, 1979, P 19
 8-13. Type 304 Stainless Steel: Effect of Hold Time and Continuous 209
Cycling on Fatigue Crack Growth Rates

6K, ksi . in. 1/ 2

10 20 40 60 80 100
10- 1

.:E.
E
E
I
I
0.1 min hOld-...f
I
I
J
"
"

,r~ 10- 2

/ 1.0minhold
,
Type 304
593°C (1100°F)
10- 1 '--_ _---'_ _....L-_...L-_---'_---'_...L...I
20 30 40 60 80 100
Stress-intensity factor range, 6K, MPa . m 1/2

Fatigue crack growth rates per unit of time ida]dt) for annealed
type 304 stainless steel for continuous cycling (0.17 Hz), for 0.1
and 1.0-min hold times at maximum load for each cycle at 593 ° C
(1100 OF), and for anR ratio of O.

As shown above, the fatigue crack growth rate is greater for

specimens tested with no holding time (continuous cycling) than
for specimens held at maximum load for 0.1 or 1.0 minute per
cycle. The lowest fatigue crack growth rates occurred for speci-
mens with the longest holding time, based on dal dt, The same
trend was observed for tests at 593°C (liOO OF), as shown here.
Therefore, cyclic loading has a more damaging effect than static
loading on crack growth per unit of time.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 118
210 8-14. Type 304 Stainless Steel: Effect of Cyclic Frequency on
Fatigue Crack Growth Rate

dK, ksi . in.1/ 2

10 20 30 40 50

0.0014 Hz 0.0067 Hz 10- 4

Q)
U
>-
--EE
o
10-3

z'
--'"
"tl

"tl
l!l' Q)

~
u
>-
...::
s: 10- 5
--.5
u

een z'
..10:
u
~
u 10- 4
--'"
"tl

"tl
Q)
;:,
en
'':;
'"
u..

Type 304
538°C (lOOO°F) 10- 6
R = 0.05

10 20 30 40 50
Stress-intensity factor range, d K, MPa . m 1/2

Effect of variation in cyclic frequency on fatigue crack growth

rates for annealed type 304stainless steel at 538 ° C (1000 0 F) for
an R ratio of 0.05 in air with a sawtooth waveform.

For fatigue crack growth rate tests on specimens of annealed

type 304 stainless steel at elevated temperatures, increasing the
cyclic frequency will decrease the crack growth rate over part of
the ~K range, as shown here for tests at 538°C (1000 OF).

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William w. Gerberich and John H. Underwood. Eds., American Society for
Metals, Metals Park OH, 1982, P 116
 8-15. Type 304 Stainless Steel: Effect of Frequency on Fatigue Crack 211
Growth Behavior

SiRESS INTENSITY FACTOR RA:\iGE, ~K, kg/lmm?/2

5x 101 Il

..,
~
.l::!
s:
u
c:
z·
::!2
"'
't:l

....-c.
~

co: 10-5
....
:I:
;:
0
co:
<.:l
:><:
u
e:(
co: ANNEALED TYPE 304 S S.
y
~ TESTED IN AIR AT 53S"C 1l000°f)
:::I
<.:l R· 0.05, Ref. [45]
....
-e
...... + 0.003 cpm
'il 0.4 cpm
00 4cpm
10-6
o 40cpm
li. 400 cpm
o 4000cpm

4
10

STRESS INTENSITY FACTOR RANGE, ~K, Ib/lin)3/2

5x101
STRESS INTENSITY FACTOR RANGE, oM, foIW/(m?/2

Effect of frequency on the fatigue crack growth behavior of type 304 tested in an air
environment at 538°C (1000 OF).

Source: L. F. Coffin, "Fatigue in Machines and Structures-Power Generation," in Fatigue and Microstructure, American Society
for Metals, Metals Park OH, 1979, P 14
212 8-16. Type 304 Stainless Steel Welded With Type 308: Fatigue Crack
Growth Rates

t.K. ksi . in. 1I2

t.K. ksi . in. 1/ 2
20 40 60 80
20 40 60 80
10- 2

Ql
c:;
10- 4 > 10- 4
Ql
c:;
~
--
u
E
E
Shielded
E z·
--. 10-
E metal arc "t:l

Z 10- 3 "t:l Shielded

...
"t:l 3
co j!l' metal arc Ql
-e
Ql c:;
c:; >
...eoi
--.S ...::.. --.S
>
u s: u

0
s: z·
z· '"
i0
~
~
10- 5 --.. e
"t:l

"t:l
~
u
u Submerged 10- 5
--..
"t:l

"t:l

Ql

..~
~ arc
u

~
..'"
.;:;
u..
.;:;
• 10- 4
u.. 10- 4
o SMAW 1
• SMAW 2
24°C (75°F) 593°C (l100°F)
10-6 10-6
20 40 60 80 100 20 40 60 80 100

Stress-intensity factor range, t.K, MPa . m 1/2 Stress-intensity factor range, t.K, MPa . m 1/2

Fatigue crack growth rates for annealed type 304 base metal and type 308 weld metal at 24 and
593°C (75 and 1100 OF), 0.17 Hz, and an R ratio of O.

Type 308 stainless steel is the alloy that is usually used for welding rod for weldments in
type 304 stainless steel when those weldments are to be exposed to room temperature or to
elevated temperatures in service. Because service experience has shown that failures are
more likely to originate in weld metal or in heat affected zones than in the base metal, it is
important to have fracture information on weldments. In general, fatigue studies at ele-
vated temperatures on specimens from type 304 weldments have shown that the fatigue
crack growth rates in the type 308 weld metal and heat affected zones are no greater than
in comparable specimens of the base metal. Fatigue crack growth rate data obtained by
Shahinian for specimens of type 304 welded with type 308 rod by the submerged arc and
shielded metal arc processes are shown above for tests at room temperature and at 593°C
(llOO°F).

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels." in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood. Eds.. American Society for
Metals. Metals Park OH. 1982. P 125
 8-17. Types 304 and 310 Stainless Steel: Effect of Direction on S-N 213

400
<II Type 310, transverse
Q. 50 'iii
Type 310, longitudinal
~" 300 ..><:
~ 40 '" ","

'"... .......'"
~E 200
:::l ~~-~
Type 304, longitudinal
30 '"
E
:::l
,~ 20 ,~
x
~ 100
~ 10 ~"'
0
105 106 108
No. of stress cycles

.SoN curves for two grades of stainless steel,

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 32
214 8-18. Types 304, 316, 321, and 348 Stainless Steel: Effects of
Temperature on Fatigue Crack Growth Rates

AK, ksi . in. 1/ 2

10 20 40 60 100
10-2 ...-r----r----.------r---.---. 4 X 10-4

10- 4
Q)
u
>
--EE
Ll

z· 10- 3
--'a>"
"0

"0
.!!!
...
~
Ll
>
...3:
s: --
Ll

.~
e z·
~
Cl

..
Ll
l'Il
--
"0
l'Il
"0
Ll 10-5
Q)
::::I
Cl
.~
l'Il
IL.

10-4

L -_ _....L. 10-6
....I...._--l._...l.-..J.J

10 20 40 60 80 100
Stress-intensity factor range, AK, MPa • m 1/2

Fatigue crack growth rates for annealed types 304, 316, 321, and 348
stainless steel in air at room temperature and 593°C (1100 OF), L-T
orientation, 0.17 Hz, and an R ratio of O.

As reported by Shahinian, Smith, and Watson, fatigue crack growth rate tests were made on single-
edge-notch cantilever specimens oftypes 321 and 348 stainless steel from L-T orientation at 0.17 Hz
with an R ratio of zero at room temperature and at elevated temperatures to 593°C (1100 OF). As for
types 304 and 3I6, fatigue crack growth rates in air increased with increasing testing temperature. The
curves above show that, at room temperature, the fatigue crack growth rates for types 304, 316,321,
and 348 all fall within a narrow band. For tests at 593°C (1100 OF),however, specimens of type 3I6 had
the least fatigue crack propagation resistance, whereas specimens of type 348 had the highest fatigue
crack propagation resistance, over the 11Krange studied. Results of tests on specimens of types 304 and
321 were nearly the same at 593°C (1100 OF) in air.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 138
 8-19. Type 309S Stainless Steel: Effect of Grain Size on Fatigue 215
Crack Growth Rate

~K, ksi • in.1/2

10 20 40 100
r - - , - - - r - - - , - - - r - - - - - - - - - - , 10- 3

10- 2

10-4

Type 309S

Testing
frequency, Grain size
Hz 45 fJm 480 fJm
10 -0- -0- 10- 6
15
20
25
30
10- 7
10 20 40 100
Stress-intensity factor range, ~K, MPa • m 1/2

Fatigue crack growth rates for annealed type 309S stainless steel
for two grain sizes, at frequencies from 10 to 30 Hz and an R
ratio of 0.05 at room temperature in air.

Types 309S and 3 lOS stainless steel are the low-carbon ver-
sions of types 309 and 310. They have higher chromium and
nickel contents than type 304 and consequently have better cor-
rosion resistance and more stable austenite than type 304. Fa-
tigue crack growth rate data have been reported by Thompson
for tests made at room temperature on compact specimens from
plate of type 309S and the L-T orientation after heat treating to
a grain size of 45 Jlm in one set and 480 Jlm in the second set.
Specimens with the smaller grain size had substantially higher
yield and ultimate tensile strengths than the specimens with the
larger grain size. Fatigue crack growth rates were obtained on
tension-tension loading at frequencies from 10 to 30 Hz and at
an R ratio of 0.05. The results are plotted above. These data
provide further evidence that a wide variation in grain size, and
the associated variation in strength level, does not affect the re-
sults of fatigue crack growth rate tests.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
Metals, Metals Park OH, 1982. P 126
216 8-20. Type 31 OS Stainless Steel: Effect of Temperature on Fatigue
Crack Growth Rate

~K, ksi • in. 1/ 2

20 40
.-------r---r-..---..---.......,."""T""'l~---.______. 10-4

.,
U 10- 3
>
--("l

E
E
z· Base metal
.,
--.,'
"C

"C
co
-196, -269°C
(-320, -452°F)
u
>
......co
10- 5
--.5
("l

...
.c z·
~
een --
"C
co
"C

-"o
...
co
.,
("l
10- 4
:::l
en
'':;
co
u..

Type 3105

20 40 60 100
Stress-intensity factor range, ~K. MPa • m 1/2

Fatigue crack growth rates for annealed type 31OSstainless steel

at 22, -196, and - 269 0 C (72, - 320, and -452 0 F), 10 to 28 Hz,
and anR ratio of 0.1, with corresponding data for SMAW welds
with type 316 filler metal.

Because of its high nickel content, type 3lOS stainless steel is

completely stable at all cryogenic temperatures and with any
amount of cold working. Therefore, it is often considered for
cryogenic applications that require a high degree of austenite
stability on thermal cycling and strain cycling. Fatigue crack
growth rate at various temperatures is illustrated above.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
Metals, Metals Park OH, 1982, P 127
 8-21. Type 316 Stainless Steel: Growth Rate of Fatigue Cracks 217
in Weldments

liK, ksi • in. 1/ 2

20 40 60 100
10-2 4 X 10-4

Ql
13 10-4
>
u
'<,

E
E
z'
~ 10- 3 Ql
13
<0
"C >
u
'<,
fl'
E .~
s: z'
~ "C

..
0
Cl
'<,
<0
"C

~
u
10- 5
E
u
Ql
:::J
Cl
'.J
<0
u.. 10- 4

Type 316
593°C (1100°F)
10- 6
20 40 60 100
Stress-intensity factor range,
liK, MPa . m 1/2

Fatigue crack growth rates in type 316 base metal and

weld metal in the unirradiated and irradiated conditions
at 593 ° C (BOO OF)in air [fluence 1.2 X 1022 n] cm-, >0.1
MeV at 410°C (770 OF)].

Results of fatigue crack growth rate tests on weldments of type 316 stainless steel have
shown that the crack growth rates in the weld metal are generally no higher than in the
base metal and may be somewhat lower at elevated temperatures (Shahinian, Smith, and
Hawthorne). The curve shown above for unirradiated weld metal tested at 593°C (1100 OF)
represents fatigue crack growth rates substantially lower than those for the unirradiated
base metal at any given I:>.K level (Shahinian). The weld was produced by the submerged
arc method using type 316 welding rod. Weldments were stress-relief annealed at 482°C
(900 OF). Specimens were single-edge-notch specimens for cantilever loading and were
tested at 0.17 Hz and at an R ratio of zero. Irradiation slightly reduced the fatigue crack
growth resistance of the weld metal, but its fatigue crack growth resistance was better than
that of the unirradiated base metal.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds .. American Society for
Metals, Metals Park OH, 1982, p 134
218 8-22. Type 316 Stainless Steel: Fatigue Crack Growth Rates-
Aged vs Unaged

AK. ksi . in.1/2

10 20 40 60 80
10 2 ."......---r------,;-------r--,--..,

Type 316, cold worked,

0 0
tested at 593 C (1100 F )

III
U
e
EE
z 10- 2
~
"lJ 10- 1 U
III

>
J!l'
E
..c:
--.S
u

10- 3 Z
i
;
e 10- 2 --..
"lJ

"lJ
u
E
u
III
:J 10-4
..
en
.;;
u..

10- 5
Aged Hold time
10- 4
• Zero
0
• 0.1 min
Ii.
• 1.0 min
v 10.0 min
10- 5 L-_....L- --l. ...L-_--l._......L~ 10- 6
10 20 40 60 80 100
Stress- intensity factor range, AK. MPa . m 1/2

Effect of exposure at 593 °C (1100 OF) for 5000 h, and hold times during
cycling, on fatigue crack growth rate of20% cold worked type 316 stain-
less steel at 593 ° C in air.

Results also have been reported by James for fatigue crack growth rate
tests in 20% cold worked specimens of type 316 stainless steel which were
cycled at frequencies of 0.0055 to 6.66 Hz, at 538°C (lIDO OF) and at an R
ratio of 0.05. Over the 11Krange studied, the fatigue crack growth rates were
highest for the specimens subjected to the lowest cyclic frequency.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Material, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
Metals, Metals Park OH, 1982, p 133
 8-23. Type 316 Stainless Steel: Fatigue Crack Growth Rates- 219
Effect of Aging

.1K, ksi • in. 112

10 20 40 60 80
10-1.......- --r-----.-------,----.--,....,

10-3

10-2
Q)
U
-.~ 593°C (1100°F)
E
E
z· 10- 4
~
'"
"tl
U
Q)

...e
Q)'
u>-
'<,
10-3
..c .~
i z·
een "tl
'<,

~
u '"
"tl

~
u
Q) 10-5
:::l
en
';:;
'"
LL. Type 316
10-4 593°C (1100°F)
Unaged Aged Hold time
0
• Zero
G
A
•... 0.1 min
1.0 min 10-6

10-5
10 20 40 60 80 100
Stress-intensity factor range, .1K, MPa . m 1/2

Effect of exposure in air at 593°C (1100 OF) for 5000 h, and hold times,
on fatigue crack growth rates for annealed type 316 stainless steel at
593°C in air.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, p 130
220 8-24. Type 316 Stainless Steel: Effect of Temperature on Fatigue
Crack Growth Rate

dK, ksi • in. 1/ 2

10 20 40 60 100

10- 4

10- 3

Ql
1)
>-
10- 5 --.5
u

2"

10-4
--
"'C
III
"'C

10-6

Type 316
10- 5 Cold worked

10 20 40 60 100

Stress-intensity factor range, dK, MPa . m 1/2

Fatigue crack growth rates of20% cold worked type 316stainless

steel for various temperatures. Curves are averages for L-T and
T·L specimens at each temperature in air; 3 Hz at 24°C, 0.67 Hz
at elevated temperatures; R = 0.05.

Results of tests on compact specimens of 20% cold worked

type 316 stainless steel at frequencies of 0.67 and 3.0 Hz and at
an R ratio of 0.05 are summarized in the above graph. Similar
results have been reported by Shahinian for tests on cold
worked type 316 at 427°C (800 OF).

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 132
 8-25. Type 316 Stainless Steel: Effect of Cyclic Frequency on 221
Fatigue Crack Growth Rate

dK. ksi • in. 1I2

10 20 30 40 50 60
10- 4
0.0067 Hz----... 1,1
U
II>

>
--
o
E
E
z·
10- 3

I
)7 ~0.67"'
~
'"
"C 0067 "'-y' //J II>

'~
...eai 10- 5 U
>
...;: --.5
u

f
s:

e z·
Cl "C
<,

"'..u'"" 10- 4 '"

"C

U
II>
::>
Cl
.;;
'"
u,

Type 316 10- 6

538°C (lOOO°F)

10
Stress-intensity factor range, dK, MPa • m 1/2

Effect of variation in cyclic frequency on fatigue crack growth rate of

annealed type 316 stainless steel in air at 538°C (1000 OF) and an R ratio
of 0.05.

As may be observed above in tests at frequencies in the range from

0.0067 to 6.67 Hz at 538°C (1000 OF), the trend is for the crack growth
rate to increase as the frequency is decreased, but there is more scatter
than for type 304.
In studying heat-to-heat variations in fatigue crack growth rates for
specimens from three heats of type 316stainless steel, James has shown
that the spread from high to low values of fatigue crack growth rates is
no greater than that represented by a factor of 2.6 over the range of 11K
values studied. One heat was produced by air melting, another by va-
cuum arc remelting, and a third by double vacuum melting.

Source: J. E. Campbell, "Fracture Properties or Wrought Stainless Steels," in Application or Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood. Eds .• American Society for
Metals, Metals Park OH, 1982, P 131
222 8-26. Type 316 Stainless Steel: Fatigue Crack Growth Rate in the
Annealed Condition

6K, ksi . in. 1I2

10 20 30 40 60 80 100
10- 3

370·C (700.F) /1
10- 2
---t!
""1/;/
I ,

482·C {900·

,1 /1
p'"
//'
fi,' I
10- 4

593·C .'
:,/
l'
/ ..
u
>-
(1100·F) / "I ~

y: /: .5
z·
./ ;,'I ~
"
:i I'"I
/ /'1 10- 5
I

10- 4

316
' - -_ _'-_J.----'_L-l......J-J....I-J..J 10- 6
10 20 30 40 60 80 100
Stress-intensity factor range, 6K, MPa • m 112

Effect of testing temperature on fatigue crack growth

rates for annealed type 316 stainless steel tested in air at
0.17 Hz and an R ratio of O.

Most of the fa tigue crack growth rate testing on type 316 stainless steel has been or-
iented toward its use in components for nuclear reactors, but the data also are applicable
to design of equipment for fossil fuel power stations, petrochemical refineries, and chemi-
cal plants. Its improved yield strength compared with that of type 304 stainless steel is an
advantage for these applications. The austenite stability in type 316 is greater than that in
type 304, so it is advantageous to use type 316 rather than type 304 for critical cryogenic
applications. Effects of elevated temperature on crack growth rate are summarized in the
graph above.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 129
 8-27. Type 316 Stainless Steel: Effect of Environment (Sodium, 223
Helium, and Air) on Cycles to Failure

10.0
8.0
DATA FATIGUE
6.0 POINTS TEST MEDIA
o Na (10 ppm OF 02)
4,0 x AIR
a HELIUM
t'
z 2.0
~
ex:
l-
V>
u 1.0
...J
u 0.8
>-
u 0.6
0.4
EXPOSED SPECIMENS
• FATIGUED IN SODIUM
EXPOSURE MEDI~ EXPOSURE CONDITION
• FATIQJEDIN AIR
ue ppm OF 02)
0.2
• FATIGUED IN HELIUM Na 286 hrs AT 92S K

0.1
3 4 6 2 6 8 104
102 8 103

CYCLES TO FAILURE

Effect of environment on fatigue characteristics of type 316 stainless steel at 92SK;

based on cyclic strain and cycles to failure.

Source: R, H. Cook and R. P, Skelton, "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
in Source Book on Materials for Elevated-Temperature Applications, Elihu F, Bradley, Ed" American Society for Metals, Metals
Park OH, 1979, P 84
224 8-28. Types 316 and 321 Stainless Steel: Effects of Gaseous
Environments on Fatigue Crack Growth Rates

aK, ksi . in.1I2

8 10 20 40 60 100
10-3 r----,,-..----.... ---...,.:.-......:.=----..:..;

Type 316
649°C 10-5
(1200°F)
U"> Roomair
u
E
E
z· Type 316
~
."
10-' 2SoC 177°F)
Dry air
fi Wet nitrogen U">
f! Dry nitrogen -!:!
.c .~
~ z·
12tn
~
.>< <0
."
u
f!
u

":tn>
.~ 10-6
u, Types 316 and 321
2SoC (77°F)
Room air
Wet air

10- 5 Types 316 and 321

649°C (1200°F)
Dry nitrogen
Dry argon

8 10 20 40 60 100
Stress.intensity factor range, 6K, MPa • m 1/2

Effect of gas environments on fatigue crack growth rates for types 316 and 321
stainless steel at 25 and 649°C (77 and 1200 OF).

Fatigue crack growth rate data at 25°C (77 OF) show that crack growth
rates increased slightly with increased humidity when oxygen was present but
that high humidity in an inert gas had no significant effect. Fatigue crack
growth rates in room air at room temperature were the same for types 316
and 321 stainless steel. Furthermore, in tests at 649 °C (1200 OF)in dry nitro-
gen, fatigue crack growth rates for types 316 and 321 also were the same. In
air, however, fatigue crack growth rates in type 316 specimens increased by a
factor of about 22 over rates in an inert environment at the same temperature.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, p 135
 8-29. Type 321 Stainless Steel: Effect of Hold Time on Fatigue 225
Crack Growth Rates

.6K, ksi . in. 1/ 2

10 20 40 60 80
10-'

10- 3
Type 321
10- 2 593°C (1100°F)

..
U
>
u
E 10-'
E
z·
.
~ 10- 3 .
u
"
B
>
~
~ .~
~ 10- 5 z·
e
'"
..
~

""eu 10-' "

..
u
::I
'"
.~
u,
10- 6
Unaged Aged Hold time
10- 5 0
• Zero
•... 0.1 min
1.0min

10-
'

10- 6 '--_-'- ..J.... .......L_ _.l----JL-..J

10 20 40 60 80 100
Stress-intensitv factor range,ll.K, MPa . m'l2

Fatigue crack growth rates for annealed type 321 stainless steelunaged
and aged at 593 ° C (1100 OF)for 5000 h and tested in air with continuous
sawtooth waveform (0.17 Hz), with 0.1 and 1.0-min hold time at anR
ratio of 0 at 593°C (1100 OF).

Results oftests by Michel and Smith on specimens of annealed type

321 stainless steel that had been aged at 593°C (1100 OF) for 5000
hours and then tested at 593°C have shown that long-time exposure at
the service temperature does not reduce the fatigue crack propagation
resistance in air. Aged specimens tested with zero holding time had
lower crack growth rates than corresponding specimens that were not
aged (see above graph). Fatigue cycling with holding times of 0.1 and
I,D minute on each cycle increased the crack growth rates slightly, as
shown in the figure,

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics ForSelection of
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H, Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 139
226 8-30. Type 403 Stainless Steel: Effect of Environment on Fatigue
Crack Growth Rate

~K, ksi • in. 1/ 2

10 20 30 40 60 80
10-3 rr----,-----,--,--...........-,------,
I
/
/
I
II /
/ 10-5
/
/ Air
10-4 Type 403
z
In H20 ~
III
"1:J
pH 7, 25°C
pH 10, 25°C
-- - - pH 7, 100°C
pH 10, 100°C
10- 6
In 1M NaCI solution
- - - - pH2tol0, 100°C

10- 5 L..-_ _---L_ _L.--L---L----I---L...J......L.J. --'

10 20 30 40 60 80 100
Stress-intensity factor range, ~K, MPa . m 1/2

Fatigue crack growth rates in type 403 stainless steelin air, water,
and aIM NaCI solution at 10 Hz and an R ratio of 0.5.

Exposure to water at 25°C (77 OF) resulted in intermediate

crack growth rates between those in air and those in water at
100 DC, as shown on a different scale in the above graph.
Tests in the 0.01 M (molar) and 1.0 M sodium chloride solu-
tions were made with the solutions at pH levels of 2, 7, and 10
and with an open circuit. Fatigue crack growth rates in 0.01 M
sodium chloride at pH 10 and 100°C were the same as those in
water at 100 "C. At lower cyclic frequencies, the fatigue crack
growth rates were higher than at 40 Hz at 6.K values above 20
MPa· m 1/2 (18 ksi· in. 1/2). For tests in the 1.0 M sodium chlo-
ride solution at 100°C (212 OF) (see graph), fatigue crack
growth rates were the same as for water at the same temperature.
At 100 °C (212 OF), fatigue crack growth rates in 1.0 M sodium
phosphate solution at pH 10 and at 10 and 40 Hz and in 1.0 M
sodium silicate at pH 10 and at 10 Hz were practically the same
as those in air.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell, William W. Gerberich and John H. Underwood. Eds., American Society for
Metals. Metals Park OH, 1982. P 147
 8-31. Type 403 Modified Stainless Steel: Scatter of Fatigue Crack 227
Growth Rates

'--_-'-_-'-......L.......L.....L....w....LJ..._ _- ' - _ - ' - ----I 10-6

20 40 60 100 200 300
Stress-intensity factor range, ll. K, MPa . m 1/2

- - - - Heat 484 in room temperature air

Heat 634 in room temperature air
Heat 933 in room temperature air
Heat 933 in 271°C (520°F) distilled water at
8.3 MPa (1200 psi)

Upper boundaries of fatigue crack growth rate scatter bands for

three heats oftype 403 modified stainless steel in the heat treated
condition, tested at 10 Hz and an R ratio of 0.083 or 0.067.

The curves representing the upper boundaries of the scatter

bands of the fatigue crack growth rate data indicate that there is
some heat-to-heat variation in fatigue crack growth rate proper-
ties for these heats. Furthermore, exposure at 27 1°C (520 OF)in
distilled water at a pressure of 8.3 MPa (1200 psi) increased the
fatigue crack growth rate.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 145
228 8-32. Type 422 Stainless Steel: Fatigue Crack Growth Rates in
Precracked Specimens

Koff' ksi • in. 1/2

10 20 40 60 80 100 200

71°e (160°F)

57°e (135°F)

10-
4
I a>

c
.:.::::
z
~
~
"tl

Type 422

~_---l ---L_---l_....L.........L. l..- ..J 10-5

20 200

Fatigue crack growth rates in precracked round rotating beam speci-

mens of type 422 stainless steel in 4.5% NaCl solution at room and ele-
vated temperatues, 10Hz, and an R ratio of-I.

Type 422 stainless steel contains nickel, molybdenum, and tungsten,

as well as 12%chromium to improve properties. The effects of sodium
chloride solutions and elevated-temperature exposure on fatigue crack
growth rates were determined by Eisenstadt and Rajan in tests of
notched round rotating beam specimens in which the numbers of test
cycles were marked by minor stress interruptions that produced mark-
ing rings. Calculations for maximum stress-intensity factors were
based on equations for solid round bars subjected to bending loads.
The material for these tests apparently had been heat treated to a yield
strength of approximately 827 MPa (120 ksi). The specimens were one
inch in diameter in the test sections. Each specimen was rotated at 600
cycles per minute (10 Hz) while at constant load with the salt water
solution flowing over the notched section. Tests with several concen-
trations of salt solution indicated that the maximum corrosive effect
was obtained with the 4.5% solution. Results oftests with specimens in
the 4.5%sodium chloride solution at room temperature, 57°C (135 OF),
and 71°C (160 OF) are shown above. Increasing the temperature of the
solution substantially increased the fatigue crack growth rates.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood. Eds., American Society for
Metals, Metals Park OH. 1982, P 150
 8-33. Type 422 Stainless Steel: Fatigue Strength-Longitudinal 229
vs Transverse

CLASS II
(Crucible 422)
turbine-wheel forgings, 12 to 30 inches in diameter,
ksi 1900 F (l 040 C)oil quench + 1200/1400 F(650/760 C)
100..--------------------------.
UNNOTCHED
x

80- o

A
•
60r- o

1-----------... VIBRATING CANTILEVER TESTS

NOTCHED FOR 108 CYCLES
x
401- Kt = 2.1 o x 0% delta ferrite
60· notch .x
o • 5% delta ferrite
0.030·inch
o o 15/16% delta ferrite
root radius A
A 20% delta ferrite

20 .....

I I I I
20 40 60 80 100 ksi
TRANSVERSE FATIGUE STRENGTH
Transverse fatigue strength as related to longitudinal fatigue strength for type
422 stainless steel, including effects of varying amounts of delta ferrite.

Source: J. Z. Briggs and T. D. Parker, "The Super 12%Cr Steels," in Source Book on Materials for Elevated-Temperature Applica-
tions, Elihu F. Bradley, Ed.. American Society for Metals, Metals Park OH, 1979, p 121
230 8-34. Type 422 Stainless Steel: Effect of Temperature on
Fatigue Strength

CLASS II
ksi (Crucible 422) (VacuumMelted)
140
15% delta ferrite
130 f- 1800 F (980 C)oil quench + tempered
to a tensile strength of 131/138 ksi
120 ,....-

(/)
(/)
IJJ
110 ,....-

a:: 100 "-

J-
'.,""'-ee_
....... -.-
700 F(370 C)

...
(/)
90 -
80 - •
room temperature
•
70 - ROTATING CANTILEVER-BEAM TESTS
I I I
601()~ 105 106 101 108
CYCLES

SON curves for vacuum-melted type 422 stainless steel with 15% delta ferrite,
showing effect of temperature on fatigue strength.

Source: J. Z. Briggs and T. D. Parker, "The Super 12% Cr Steels," in Source Book on Materials for Elevated-Temperature
Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH, 1979, P 121
 8-35. Type 422 Stainless Steel: Effects of Delta Ferrite on 231
Fatigue Strength

CLASS II
(Crucible 422)
3/4 -inch-diameter bar stock or 3/4 -inch-thick plate
1900 F (1 040 C)oil quench
ksi % DeltaFerrite TensileStrength, ksi

~ 15~20
110' 155
160

'~~v_
.... 140

en
en
LLI
100

\ ~...
00% Ferrite,Tempered 1150 F -~
--.. . . . . .-v..
5% Ferrite,Tempered 1150 F
(620 C)
V;;
0: 90
"-.. 0 (620 C)
I-
en o 0.:
15/20% Ferrite,Tempered 1200 F(650 C)
a
lr-+

80 0- 0-
'LONGITUDINAL ELECTROMAGNETIC
CANTILEVER-BEAM TESTS
70

lOS 107
CYCLES

S-N curves for type 422 stainless steel, which demonstrate the adverse effects of
delta ferrite on fatigue strength.

Source: J. Z. Briggs and T. D. Parker. "The Super 12%Cr Steels," in Source Book on Materials (or Elevated-Temperature Applica-
tions, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH. 1979, p 121
232 8-36. 17-4 PH Stainless Steel: Fatigue Crack Growth Rates in Air
vs Salt Solution

.o.K, ksi • in.'/2

10 100

10-2
I

'---Hll00

'I
,

,~
R =0.05
1 min hold
Salt soln 10-4
I~
I
I
10- 3

il!
'"
~u
10- 4
"5,
."u..

10-6

10- 5

17·4 PH

'-- -'- -'- ...1..1 10-1

10 100
Stress-intensity lactor range, .o.K, MPa • m1/2

Fatigue crack growth rates in WOL specimens of 17-4 PH stainless steel in the HI050 and
HllOO conditions in room temperature air and in a 3.5% NaCI solution,

Results of fatigue crack growth rate tests on specimens of 17-4PH stainless steel under
comparable conditions are presented here. Those specimens that were tested in the HI 050
condition at a stress ratio of 0.67 with a one-minute holding period at maximum load in
each cycle had the highest fatigue crack growth rates (as for 15-5PH) in the upper levels of
I:!.Kvalues. Specimens in the H II 00 condition tested in a salt solution with a one-minute
holding period, however, had fatigue crack growth rates only slightly higher than those of
comparable specimens tested in air with continuous cycling.

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels." in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals. Metals Park OH. 1982, P 156
 8-37. 15-5 PH Stainless Steel: Fatigue Crack Growth Rates in Air 233
vs Salt Solution

LlK, ksi • in. 1/2

10 100

)
"
Hl~50 ~I //'I
R - 0.67 I:/,
1 min hold I 1//
. RT.ir I/Il
E
'li
~

E
z·
10- 3

1
,,1
l,' l'
---Hl050
R = 0.05
~
..
."
Hll00--.......,'
R = 0.05 ~ I
10Hz
Sine wave
!l' 1 min hold " I RT air
e Salt soln I 'I
.t::
i VI
e
'"
'I1 ,r--HllOO
R = 0.05
""ut! 10-' ~ 1 min hold
..
u

.§.
I'
,
RT air

TO
u,
I 10- 6
H 1100 :----""'/'
R = 0.05

10-5
10 Hz
Sine wave
RTair
I
I
15·5 PH

1- --L. ...1... .J1O-

7
10 100
Stress·intensity factor range, LlK,MPa • m 1/2

Fatigue crack growth rates in WOL specimens of 15-5 PH stainless steel in the HI050 and
HllOO conditions in room temperature air and in a 3.5% NaCI solution.

For specimens in the HI 050 condition, increasing the R ratio from 0.05 to 0.67 and incorpo-
rating a one-minute holding period at maximum load in each cXcle substantially increased the
crack growth rates at LiKvalues over 40 MPa· m 1/2 (36 ksi- in. I 2). For specimens in the H 1100
condition, exposure to a salt solution environment during tests with a one-minute holding pe-
riod at maximum load increased the fatigue crack growth rates over those of specimens tested in
air with one-minute holding time or with continuous cycling (see graph).

Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels." in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood. Eds .• American Society for
Metals, Metals Park OH. 1982, pISS
234 8-38. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates
at Room Temperature

~K. ksi . in.1/2

10 20 40 60 80100 200
10-2 4 X 10-4

Gl
U
>
--u
E
E 10- 4
Z

--
-e
l'O
Gl
U

.....
"tl
Gl'
>
--
U

.~
..
l'O

.s:
;:
10- 3 Z'

e
Cl
~
--
"tl
l'O
"tJ

u
E
u
Gl
:::l
Cl
';;
l'O
u.. 10-5
PH 13-8 Mo
H1100

10- 4
20 40 60 80 100 200

Stress-intensity factor range, ~K. MPa • m 1/2

Fatigue crack growth rates in cantilever beam specimens of PH

13-8 Mo (HII00) stainless steel, at L-T orientation, 0.17 Hz, and
an R ratio of 0, in room temperature air. Data are based on the
stress-intensity-factor range as shown.

Source: J. E. Campbell. "Fracture Properties of Wrought Siainiess Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 159
 8-39. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates in Air 235
and Sump Tank Water

toK, ksi • in. 1/2

6 8 10 20 40 60

10-2
R = 0.3
STW
L-T
10- 4

10- 3

10- 4

10- 6

PH 13-8 Mo
H1000

6 8 10 20
Stress-intensity factor range, toK, MPa • m 1/2

Fatigue crack growth rates in compact specimens of PH 13-8 Mo

stainless steel in the HI 000 condition for room temperature tests
at I Hz, R ratios of 0.08 and 0.3, L-T and T-L orientations, in
low-humidity air (LHA) or sump tank water (STW).

Effects of increasing the load ratio,R, on fatigue crack growth

rates in low humidity air (LHA) in sump tank residue water
(STW) for specimens of PH 13-8 Mo (H 1000)are shown above.
The highest fatigue crack growth rates in this series were ob-
tained on specimens tested at an R ratio of 0.3 in STW. Increas-
ing the load ratio from 0.08 to 0.3 had a marked effect on the
growth rates.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, p 158
236 8-40. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates at
Subzero Temperatures

~K, ksi . in. 1/ 2

4 6 8 10 20 40 60 100
10- 4

10- 5 Q)
u
>
u
'-:
.!:
z
1:1
"-
nl
1:1
10-6

PH 13-8 Mo
H1000
10- 7
4 6 8 10 20 40 60 100
Stress-intensity factor range, ~K. MPa . m1/2
Fatigue crack growth rate scatter band for compact specimens from rolled bar and
extrusions of PH 13-8 Mo stainless steel in the HIOOO condition for room temperature
tests in low-humidity air and in sump tank water at frequencies of 1 and 6 Hz and anR
ratio of 0.08 for L-T and T-L orientations.

Fatigue crack growth rate data for room temperature tests on specimens from
rolled bar and extrusions of PH 13-8 Mo (H 1000)stainless steel make up the scatter
band in the above graph. Specimens of L-T and T-L orientations were tested in
low-humidity air and in sump tank residue water at frequencies of I and 6 Hz and at
an R ratio of 0.08. Under these conditions, variations in frequency and environment
had little effect on fatigue crack growth rates. For tests at -54°C (-65 OF),the rates
of fatigue crack growth were lower than those at room temperature over most of the
~Krange.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 157
 8-41. PH 13-8 Mo Stainless Steel: Constant-Life Fatigue Diagram 237

Minimum stress, ksi

-150 -100 o +50 100 150 200
1600

1400 200

1200

'"
a..
150 Jl
:2 1000
",'

...~'"'" 800
E S)<::l l'l'b E
:l 'li~ :l
E 100 .S
'x 600 S)<::l
\e"""
.,-,; x
'"
:2 (c) 'b~ '"
:2
~fl, Axial fatigue
400 Unnotched specimens
Longitudinal and 50
transverse or-
200
ientations

o '--_-'--_...l-__....J...._----L_--''------''L-_-'--_-L.._--'--_----L_--'_ _-'----_...l-_-'
-1200 -1000 -800 -600 -400 -200 0 +200 400 600 BOO 1000 1200 1400 1600
Minimum stress, MPa

Constant-life fatigue diagram for PH 13-8 Mo stainless steel, condition HI ODD.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, p 32
238 8-42. Types 600 and 329 Stainless Steel: S-N Curves for Two
Processing Methods

1000
800
I
I I I I
600 600 Steel (STAMP)-+--+--I I 100
80
.~ 0.57
'"
o,
::2:
Vi
400
.L
I
I
I I I.
IV

••
. '\1'
60
'iii
-'"
",'

'"
e AISI 329 (electroslag remelted) • 0.51 40 '"
e
Ul Ul
200

20
100
10' 10' 10' 10'
Cycles to failure

Mechonlcal properties
Tensile Yieldstrength Elongationin Impact Fatigue
strength, (0.2% olTset), 50 mm (2 ln.), Reduction energy, strength,
Steel MPa(ksl) MPa (ksi) % in area, % J (R·lb) MPa (ksl)
STAMP 600 760 600 26 54 25 430
(110) (87) (18) (62)
Electroslag-remelted 329 630 500 29 65 35 320
(91) (73) (25) (46)

S-N curves showing test results and mechanical properties of STAMP-processed 600
stainless steel and electroslag-remelted AISI 329 stainless steel. Fatigue ratio
(0 107/Rm) for 600 steel: 0.57. Fatigue ratio for electroslag-remelted 329 steel: 0.51.

Source: Metals Handbook, 9th Edition, Volume 7. Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, p 549
 8-43. Grade 21-6-9 Stainless Steel: Effect of Temperature on 239
Fatigue Crack Growth Rates

AK, ksi • in.1/2

10 20 40 60
10- 4
21-6-9 I
Q)
u 22 to -196°C /
>
u
...... 10- 3 (72 to -320°F) /
E
E -269°C
2: (-452° F)
"t:l
......
/
Q)
III
"t:l U
>
...oj.. 10- 5
......u
III
,E
or.
s..
0
/ 2:
"t:l
......
III

~
Cl

.
u
III
u
10- 4 / "t:l

I
Q)
:l
Cl
'+:i
III
u.

/ 10- 6

10 20 40 60 100
Stress-intensity factor range,
AK, MPa • m 1/2

Fatigue crack growth rates in specimens of annealed 21-6-9 stain-

less steel at 22, -196 and - 269 0 C (72, - 320 and -452 0 F), 20 and
28 Hz, and an R ratio of 0,1.

Similar tests made with specimens of 22-13-5 stainless steel

showed fatigue crack growth rates that were nearly the same as
shown here for 21-6-9.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels,"in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 140
240 8-44. Kromarc 58 Stainless Steel: Effect of Cryogenic
Temperatures on Weldments

dK, ksi • in. 1/ 2

20 40 60 100 200

10- 3

10- 4

I Kromarc 58

I Base metal
Weld metal

......_ -::':::--_--:'-_....L.._----''--_ _.L------' 10- 6

20 40 60 100 200
Stress-intensity factor range, dK, MPa . m 1/2

Fatigue crack growth rates for solution treated Kromarc 58 base

metal in air at room temperature, and base metal and weld metal
at-269°C (-452 OF) in liquid helium, at 10 Hz and anR ratio
of 0.1.

For the fusion zone of a gas tungsten arc weld made with
Kromarc 58 filler metals, the KIJJ) value was 156 MPa. m l / 2
(141 ksi- in. 1/2) at -269°C (-452 OF).Fatigue crack growth rate
data for the base metal at room temperature and at - 269°C and
for the weld metal at -269°C are shown above. The data were
obtained on compact specimens at 10 Hz and at an R ratio of
0.1. Fatigue crack growth rates for tests in liquid helium were
lower than at room temperature at the same t::.Kvalues. There-
fore, if room temperature crack growth rate data are used to
estimate crack growth at cryogenic temperatures, the estimated
values will be conservative.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, P 142
 8-45. Pyromet 538 Stainless Steel: Effects of Welding Methods 241
on Fatigue Crack Growth Rates

AK, ksi . in. 1/ 2

10 20 30 40 60 80 100 200
10- 4

",-SMAW
Gl 24 and -269°C
U
> 10- 3 (75 and -452° F)
......u
E
E
I
Z·
't:l
I
...... Gl
'"
't:l
/"GTAW
u
>
!l u
......
24°C (75°F) .~
'"
.~

..r: I 10- 5 z·
~
0
~ I ~
'"
~'"
u
't:l

E
u
Gl
10- 4
:l
'"
.;:;
'"
u..
Pyromet 538 welds

20 30 40 60 80100 200
Stress-intensity factor range, AK, MPa . m 1/2

Fatigue crack growth rates in weld metal in Pyromet 538 stain-

less steel at room temperature and -269°C (-452 OF) and at
10Hz.

The base metal was solution annealed prior to welding. One

set of welds was made by the gas tungsten arc welding (GT AW)
process with 21-6-9 filler wire, and the other was made by the
shielded metal arc welding (SMA W) process with IN 182 cov-
ered electrodes. Results of these tests are summarized in the
graph above. Specimens with SMA W welds had the same fa-
tigue crack growth rates at room temperature and at -269 "C
(-452 ° F). Specimens welded by the GT AW process had higher
crack growth rates at -269 °C than at room temperature. Exam-
ination of the microstructures near the fracture surfaces for the
specimens tested at - 269 ° C showed tha t there was 6 to 7% delta
ferrite (produced by welding) in the weld metals along with in-
duced martensite. The SMAW weld metal was fully austenitic.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William w. Gerberich and John H. Underwood. Eds., American Society for
Metals, Metals Park OH, 1982, p 141
242 8-46. Duplex Stainless Steel KCR 171: Corrosion Fatigue

KCR 171
Whit. wat.r
400
pH ·4. IS
T • ISO·C

300

CI
Q.
_ 0 .......
~

(\/ 200 0 ...

l;
~
o 6 Hz
t:> 20 Hz
100
o 100 Hz
o 167Hz

Nf eyel ..

Rotating bending S-Ntests were carried out in 50°C (122 OF) white water
at different frequencies (6, 20, 100,and 167Hz) for samples polished with 240
grit emery paper and the results obtained are presented in the above S-N
diagram. The results thus far 0 btained for the two highest frequencies appear
to fall on the same S-N curve, and the indication is that this curve would
present a quite horizontal fatigue limit. In the short life regime (N,«: 106
cycles), the results suggest that decreasing the frequency below 100 Hz dis-
places this portion of the S-N curve to shorter lives without significantly
changing its slope.

Source: M. Ait Bassidi, J. Masounave and J. I. Dickson, "The Corrosion Fatigue Behaviour in White Water of KCR 171," in
Duplex Stainless Steels, R. A. Lula, Ed., American Society for Metals, Metals Park OH, 1983, p 455
 9-1. Grades 200, 250, and 300 Maraging Steel: S-N Curves for 243
Smooth and Notched Specimens

1500r - - - - - - r - - - - -.....----~----_,
CZZl 18Ni(300)
200
Em 18Nj(250)
css 18Ni(200)

1000 _ _-+- ....=::j 150

'"
c..
~
'iii
.:.l

v.
...e'"
If)
500

o L.- --'- ---L L.- ~ 0

104 105 106 107 108
Number of stress cycles

Fatigue properties of maraging steels are comparable to those of

other high-strength steels. Smooth-bar and notched-bar fatigue prop-
erties for I8Ni(200), I8Ni(250), and I8Ni(300) grades are summarized
in the S-N curves shown above, Fatigue crack growth rates in marag-
ing steels obey the da] dN= (t1K)m relationship commonly observed in
steels and are similar to those of conventional steels. Improved fatigue
properties can be obtained by shot peening and by nitriding,

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH. 1978, p 451
244 9·2. Grade 300 Maraging Steel: Fatigue Life in Terms of Total Strain

18% Ni morolling (300)

-1
10

,, • - LOAD CONTROL

\
\

10
-2
--
/(J'f/E
--..
-)-Trrr~ ..c
'\
~\ ~
...
- : TOTAL

ELASTIC A..-
\
PLASTIC~\
h
\
\
-3
10 L-J....l.l..JLlJJ.U-l.....LJ..J.JJJ.U.--I....LJ.J.LWL--I...1.LI..JWJJ..--L..LUJ..LJ.ll..-l-u..uJJ.lJ

1 10

Superposition of elastic and plastic curves gives fatigue life in terms

of total strain. An actual example for this method of determining fa-
tigue life is presented above for a maraging steel.

Source: Marc Andre Meyers and Krishan Kumar Chawla. "Mechanical Metallurgy: Principles and Applications," Prentice-Hall,
Inc.. Englewood Cliffs NJ, 1984, p 700
 10-1. Fatigue of Cast Irons as a Function of 245
Structure-Sensitive Parameters

50

20 iii
a:

10
§ •

...J
I&.:

• "B" BAR
~ ·S~ BAR

..1 I
MEAN FATIGUE LIMIT VERSUS (BHN)
10

Fatigue of cast irons as a function of structure-sensitive parameters: Bhn, elastic modu-

lus (Eo) and damping capacity (D).

Walter has shown that the fatigue properties of irons are highly dependent on
volume of graphite and its morphology and distribution, as well as the matrix struc-
ture. He was able to reduce these factors to some easily measurable parameters, Eo,
D, and Bhn, which gave good correlation with fatigue properties over a rather wide
range of irons (see graph). It is reasonable that these parameters relate to fatigue
performance, since they are measures offatigue-related properties. Eo, the modulus
at very small strains, is controlled mostly by the volume of free graphite and to some
degree by the graphite shape. Since the graphite present detracts from the matrix
load-carrying area, the more graphite, the higher the stress on the remaining
matrix-thus lower fatigue performance. D, the damping capacity, is controlled
mostly by the graphite morphology and to some degree by the graphite volume.
Sharp-edged flakes are greater stress raisers than rounded-edge flakes and spher-
oids; thus the higher the D, the poorer the fatigue performance. Bhn is largely a
measure of the matrix hardness and, to some degree, ofthe graphite volume; thus
the higher the Bhn, the better the fatigue performance. These easily measured prop-
erties are put to good use in industry as specification means and process-control
criteria.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure.
American Society for Metals, Metals Park OH, 1979, p 86
246 10-2. Gray Iron: Fatigue Life, and Fatigue Limit as a
Function of Temperature

Temperature, of
200 400 600 800 1000
400
I
Fatigue life
I
Fatigue limit

350
- - 50
300

- - 40
250

IV
0-
:2
- -

, ".
::;' 200
e
cil -,
.
150

100
<,
.,
_____0

17
'-
~
- ---- -% - " ~ <,~
- 20

"
- 1o
50
~
o
100 200 300 400 500 600
Number of cycles to failure Temperature,oC

o Notched specimens • Un notched specimens • Notched specimens.

stress based on net area

Composition: 2.84 C, 1.52 Si, 1.05 Mn, 0.07 P, 0.12 S, 0.31 Cr, 0.20 Ni, 0.37 Cu. (Ref 5)

Typical fatigue life for as-cast gray iron of the above composition (left). Effect of temperature on
fatigue limit for the same gray iron (right).

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 21
 10-3. Gray Iron: S-N Curves for Unalloyed vs Alloyed 247

34
o Plain Iron 230
• Alloy Iron. 1% Ni. 0.4% Cr.
32

30
""",
'\. 0
0.6% Mo. 1.0% Mn
220
210

~
(f)
a. 200 <U
0 28 o,
0 190 2
0
..-- 26 • 180 en
(f)
en
(f)
<J)
.....
L..
<J) 170 (f)
.....
L.. 24
(f)
Knee 160
22 - /-----------1----0--
0
150
• Endurance or ~
-L
20

18
Fatigue Limit .,.
- ._--_.1. ________________
- 140
130

10' 10 5 10 6 10 7
Cycles To Failure

A fatigue crack starts in an area of high stress concentration after a large number
of loading cycles. It is always a brittle type of fracture even when occurring in
ductile metals. As the crack progresses it increases the stress concentration, and the
rate of propagation under the cyclic loading increases. When the cr~ss section of the
remaining metal becomes insufficient to support the maximum load, complete
failure occurs as it would under an excessive steady stress.
The number of stress applications that will induce a fatigue failure is less at higher
maximum stress values, and conversely a larger number of stress cycles can occur at
a lower maximum stress level before a fatigue crack is initiated. A plot of this
relation for a metal is called an S- N curve and relates the maximum applied stress to
the logarithm of the number of cycles for failure. When the number of cycles with-
out failure exceeds ten million, the endurance life is considered infinite for body-
centered-cubic ferrous metals. The maximum stress that will allow this number of
cycles is established as the endurance limit, or the fatigue strength or fatigue limit.
Two typical S-N curves for a plain and alloy high-strength gray iron are presented
above.

Source: Iron Castings Handbook, Charles F. Walton, Ed.. Iron Castings Society, Inc.. 1981, p 246
248 10-4. Gray Iron: Effect of Environment

21
140
20 -
Fatigue
'ea.n 19 Strength 130
co
Q.
0 18.000 PSI

0
18 1124 MPa) ~
0 120 en
(JJ
17 Q)
en ....
+-'
(JJ
Q) (.f)
....
+-'
16 110
(.f)

15
_ 14.500 DSt
(100 MPal 100
14

10· 10· 10 7 10·

Number of Cycles

The effect of various environments and corrosion inhibitors listed in the

table below on the corrosion fatigue properties of gray iron.

Fatigue strength Fatigue strength

Environment psi MPa reduction factor

Air 17,920 124

Water 14,560 100 1.23
3% sodium
chloride. . . . . . . . . . . . . . . . .. 5,600 39 3.20
1% borax 15,680 108 1.14
3% "Sobenite"* 17,920 124 1.00
3% sodium
carbonate 17,920 124 1.00
3% soluble oil. 17,920 124 1.00
0.25% potassium
chromate 17,920 124 1.00

* "Sobenite" is a mixture oj 10paris sodium benzoate to 1 part sodium

nitrite.

The corrosion fatigue program involved testing in air, a spray of deminer-

alized water, and a spray of three-percent sodium chloride solution; addi-
tional tests were made with a demineralized water spray and various known
corrosion inhibitors. The S-N curves and table above indicate that both the
demineralized water and three-percent sodium chloride sprays reduced the
fatigue strength of a pearlitic gray iron. Of the various alkaline inhibitors and
soluble oils investigated, only borax was not completely effective for the
pearlitic irons. Annealed ferritic gray irons were similarly affected by the
sodium chloride solution.

Source: Iron Castings Handbook. Charles F. Walton. Ed., Iron Castings Society, Inc., 1981, p 255
10-5. Class 30 Gray Iron: Modified Goodman Diagram 249

Mean Stress. MPa

50 100 150 200
30 ~---"""T---r--""---T""""'---"""'"2
200

Area of Finite Life

150
20 I--------+---~;<;;;;;;;;;;;;;;;;+.;;;;;;ijit----~

'(ij
a. eo
o 100 CL
o ~
o
iii
Ul

....~
(f)
Cl
50 c
:il
0
S

-50

-101'- .....L. .l...- ...J

o 10 20 30

Mean Stress. 1000 psi

A modified Goodman diagram relates the endurance limit to an

allowable working stress when it is superimposed on a steady
stress.

In many engineering applications, alternating stress is super-

imposed on a steady stress and requires special consideration. A
method of relating the effect of the combined static and alternat-
ing stresses on the endurance limit has been developed into the
Goodman diagram, of which a modified form is shown here.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981,P 251
250 10-6. Class 30 Gray Iron: Fatigue Crack Growth Rates

Stress Intensity Factor Range. D.K in MPavm

10 20 40 60 100

Q)
o>- Q)
o>-
~ ~
s:
o E
c E
.~ 10-' c
Z z
~ ~
'"
"0
Band for Wrought '"
"0

2iIII Ferrite-Pearlite 10- 4 ~

III
a: Steels a:
s: s:
~ Load Range ~
e
(9
e
• 2000 Ib (910 kg) (9

i3 10-8 o 2200 Ib (1000 kg) .:>L

o
'"
U • 2500
o 2500
Ib (1130 kg)
Ib (1130 kg) U
~

6 3000 Ib (1360 kg)

... 3300 Ib (1500 kg)

10 20 30 40 60 80100
Stress Intensity Factor Range. D.K in ksi VIiicli

Fatigue crack growth rate. The endurance limit approach to

design utilizes fatigue data taken on smooth, defect-free test
specimens. For such specimens, fatigue crack initiation may
take 80 to 90 percent of the total lifetime while crack growth is
only 10 to 20 percent of the lifetime. Such flaws allow fatigue
cracks to initiate in a relatively small number of cycles so that
the lifetime of the component depends principally on the crack
growth rate. If the initial flaw size can be determined from expe-
rience or by utilizing nondestructive inspection and the critical
flaw size calculated using the fracture toughness value K,c' then
crack growth rate data may be used to calculate the number of
cycles required to grow a crack from an initial size to a critical
size where final fracture occurs.
Only limited fatigue crack growth rate data are available on
cast irons. These results are presented in the above chart for a
class 30 gray iron, where dol dN is the crack growth per cycle
and 6.K is the stress intensity range.

Source: Iron Castings Handbook, Charles F. Walton, Ed.. Iron Castings Society, Inc., 1981, p 250
 10-7. Gray Irons: Torsional Fatigue for Various Tensile 251
Strength Values

MPa 1000 psi

60 58.000 psi (400 MPa) /

400

50

300
en
en
~ 40
(j)
"0
Q)
+-'
<U
Q)
li} 200
a::
'0
.~
E
~ 24.000 psi (166 MPa)
Q; 100
~
o
..J 14.000 psi (97 MPa)
"0
c
<U
Q; 1000 psi
~ 0
::J

Mean Stress

-100

Torsional fatigue strength for three levels oftensile strength with various mean
stresses.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, p 253
252 10-8. Gray Irons: Torsional Fatigue Data for Five
Different Compositions

40 r---------------------.,
1.50

en
1.25 Q)
E 30 .I:
E c
U

....OJ
.I:
1.0
C
....OJ
.I:

Q) C
-.J 20 Q)
~ 0.75 -.J
U ~
lO U
L..
lO
U L..

0.5 U
lO
o 10 ....0
lO
I-
0.25 I-

O'--__ '---""'--_-'----_-'-----I._-'-----'-----'- ----l_ _----l --'

300 400 500 600 800 1000 2000

Number of Cycles
Total length of six cracks (the first three cracks in each of two specimens of each iron)
as a function ofthe number of thermal cycles between 1100 and 400 °C (590 and 200
° C). Iron compositions are as follows:

Composition. %
Iron C Si Mn Cr Mo Ni Cu Sn
A 3.43 1.65 0.57
B 3.45 1.74 0.59 0.49 0.60 0.59
C 3.45 1.68 0.63 0.30 0.30 0.97 0.87
0 3.44 1.69 0.58 0.21 0.38 0.30 0.077
E 3.43 1.66 0.58 0.50 0.39

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, pp 288, 289
10-9. Gray Irons: Thermal Fatigue-Effect of Aluminum Additions 253

,..----r----,---,----,-----r----r----.5
Gray Iron
.175
3% AI. 0.6% Si (65% Ferrite)
D

.!: .150
<> 2% Si --+---t"7""'==------j4
o o 0.5% Cu. 0.5% Mo D E
c
A 2% AI. 2% Si (4% Ferrite) E
.125
3
s:
+-'
g> .100 t---+----+---+---.,,>4F---t----+------l
I
Q)
....J _ _--1
~ .075 t---+----+---.",..e..t----+-----.,.,.-=-- 2
o
U
...ro .050 I--------if------""--+---=-'f----+----+---'::=--'""=i

2 4 6 8 10 12
Number of Thermal Cycles (X 1000)
Thermal fatigue resistance of different alloyed gray irons.

This graph shows results of a thermal fatigue test in which notched

disc specimens were alternately heated to 800 OF (425°C) and cooled
to 200 OF (95°C) in two fluid beds, demonstrating that a peariitic gray
iron containing 3.4% carbon and 2% aluminum was highly resistant to
thermal crack propagation.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, p 434
254 10-10. Gray Irons: Thermal Fatigue-Effect of Chromium and
Molybdenum Additions

4
3.7% C - .15

3
I
Bridge Cracked
~.
1;..:-.,;.;;;, ..,. ,~~
.10
. ~-.:

Unalloyed ____~ ~ »->

CrNi _ '?"' . ~ . ..-'
/~
V
.-- !
.-1---
.05
en
Q)
E I ...-:. ..c
o
E CrNiMo0'" c
en
.::.t. CrMo( en
~ 0 o ~
L. <Il
L.
U U
4
'0
-
s:

o
Q.
Q)
3
3.2-3.3% C

I
Bridge Cracked
I
--- .15
'0
..c
a.
Q)
o
.10
CrNiMo / .
2
.
'

.05

o '--__-L._ _- - ' .L- J..._ _----l_ _----J 0

o 100 200 300 400 500 600

Number of Cycles
The depth of cracks resulting from the thermal cycling of gray irons
between 860 of (460°C) and room temperature.

Alloying with molybdenum and chromium provided superior thermal fatigue resis-
tance compared with irons that contained other alloying additions or no alloying at all. In
this case, the improved thermal fatigue resistance is believed to be directly related to the
higher elevated-temperature tensile strength and better stability of the chromium-
molybdenum irons. However, it must be remembered that this improvement is related to
and dependent on the temperature cycle and base iron composition, as shown above. It
has also been indicated that the development of an acicular matrix structure, by adding
relatively large quantities of molybdenum and copper, supplies a less than desirable influ-
ence on thermal fatigue cracking.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Casings Society, Inc., 1981, p 288
10-11. Gray Irons: Thermal Fatigue-Room Temperature 255
and 540°C (1000 OF)

Tensile Strength. MPa

o 100 200 300

1500
<J)
Q)
0
>-
o
'0 1000
Q;
.D
E
:J
Z
500

o L- ' - -_ _---' --'- -'- '---'

o 10 20 30 40 50
Tensile Strength. 1000 psi

Curves showing relation between the number of thermal cycles

for cracking and tensile strength at room temperature and 1000 ° F
(540 o q .

For good resistance to thermal fatigue, gray irons should have

high thermal conductivity, a low modulus of elasticity, high
strength at both room and elevated temperatures, and for
temperatures above 900 OF(500 "C), resistance to oxidation and
structural change. Because some of these properties are in oppo-
sition, a compromise must be made in selecting the most appro-
priate metal for each type of service. As the maximum tempera-
ture to which the gray iron is cycled and number of cycles
increase, the number and size of thermal fatigue cracks become
larger. The above curves illustrate the influence of room- and
elevated-temperature strength on the thermal fatigue resistance
of irons having similar carbon equivalents, thermal conductivi-
ties, matrix structures, and elastic moduli. Those irons with
higher room and elevated-temperature tensile strengths (achieved
by alloying) generally display higher thermal fatigue strength.

Source: Iron Castings Handbook. Charles F. Walton, Ed .• Iron Castings Society, Inc.• 1981, p 285
256 10-12. Gray Irons: Thermal Fatigue Properties-Comparisons
With Ductile Cast Iron and Carbon Steel

Maximum Cycle Temperature. F

1200 1400 1600
1o',---,--------,r--------,---,

z
d>
c
:;;; 10' Steel
o
~
(J
Ductile
B Iron
<f)
<Il
u
>-
(J

'0 Gray
Q> 10'
.0
AI
; 1m",
E
:J
Z

16~0.,...0--------=-.L------~-----~~
700 800 900
Maximum Cycle Temperature. C

The above curves show the variation of the number of

cycles to cracking with the maximum temperature of the
cycles for gray iron, ductile iron, and carbon steel. Com-
positions of the four gray irons are as follows:

Analysis. %
Iron C Si Mn Cr Mo Other
A 3.43 2.37 0.78 0.22 0.32 0.21 Sn
B 3.49 2.37 0.84 0.24 0.22
C 3.48 0.60 0.88 0.23 0.20 2.37 AI
D 3.50 2.38 0.83 0.30 0.77 1.51 Cu

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, pp 286, 287
 10-13. Cast Irons: Thermal Fatigue Properties for Six Grades 257

Ferritic Compacted Graphite

Pearlitic Compacted Graphite

Ferritic Ductile

Pearlitic Ductile

Alloyed Ductile
6 8 10' 2 4 6 8 10'
Number of Cycles

The number of thermal cycles required to produce thermal fa-

tigue cracking in cast irons. Compositions are tabulated below.

Analysis %C %5 %Mn %P %Mg Alloys

Class 35 Gray Iron 2.96 2.90 0.78 0.07 0.12Cr
Ferritic Compacted 3.52 2.61 0.25 0.05 0.015
Graphite
Pearlitic Compacted 3.52 2.25 0.40 0.05 0.015 1.47Cu
Graphite
Ferritic Ductile 3.67 2.55 0.13 0.06 0.030
Pearlitic Ductile 3.60 2.34 0.50 0.05 0.030 0.54 Cu
Alloyed Ferritic 3.48 4.84 0.31 0.07 0.030 1.02Mo
Ductile

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society. Inc., 1981, pp 393, 396
258 10-14. Ductile Iron: Effect of Microstructure on Endurance
Ratio-Tensile Strength Relationship

Tensile Strength. MPa

J.
300 500 700 900 1100 1300
I I I
I
~o
0.5 -o\'ox>
~.~ "\~t x Pearlitic
Ferritic.~.
o
.."
'"
II:
.o~ "x~x:
x
~
Q)
0000
g 0.4 o ,
~
::J
'U
c
\ ·K Tempered Martensite
W <,
0.3 • j( .....
•
• ••
60 100 140 160
Tensile Strength. 1000 psi

In general the fatigue limit for ductile iron increases with ten-
sile strength, but as with other ferrous metals, the increase is less
than proportional. The relation between the tensile strength and
the endurance ratio for the annealed, ferritic irons is different
from that of the irons with a matrix of pearlite or tempered mar-
tensite, as illustrated above.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, p 341
 10-15. Ductile Iron: Effect of Microstructure on Endurance 259
Ratio-Tensile Strength Relationship

Tensile strength, ksi

50 75 100 125 150 175

0.51-----e-..... "ki~---T___+-----__1I_-----+__-----__+--_t

..o,
~

1ie O.4I------_+_---'''--------'''r__+-----.-+--'''o,~ ......I _ - - - - - _ + _ - - - - - _ _ + - - _ _ 1

::>
"0
<:
w

0.31------t------+-----t--------1r---'''''''=----+----j
•
200
Tensile strength, MP.

The influence of tensile strength and structure on the endurance ratio of ductile
iron is indicated in this graph. Endurance ratio is defined as endurance limit divided
by tensile strength. Because the endurance ratio of ductile iron decreases as tensile
strength increases, regardless of structure, there may be little value in specifying a
higher-strength ductile iron for a structure that is prone to fatigue failure. For tem-
pered martensite ductile iron, the improvement in fatigue strength due to an in-
crease in tensile strength is greater than for pearlitic or ferritic structures. This is
indicated in the graph above by the shallower slope for martensite.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels. American Society for Metals, Metals
Park OH, 1978, P 45
260 10-16. Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades,
Using V-Notched Specimens

400'.---------,-------,-------,---------,

3S01-------1-------+------+---------1
F_I6G-40-18 .......odl
~v...otchod

3ool----------+------+-------t--------I

g. 2001--------=zilJ!l
~
ISOI--------+----==
20

loo'f--------f-------+-------1-----------l

106
Fatiguelift. cyeles

400,-------.-----------,--------,r--------,

3S01-----------lI----------l--------l'---------l
Pllrlitlc (80-66-06 •• -e...t)
46° V-notched
3001-----------1r--------l--------l---------l
.r. 40
~ ~
-6 2SOI -- - - - - -+----:
i!' ~e
~o :;

:
il. 2001-----------1r----"
1
ISOI---------1I--------f--------f---------j
20

1001-----------1r-------f-------1--------l

106
Fatiguelife. cycles

Top:S-N curves, including scatter bands, for annealed ductile iron. Bot-
tom: Similar to above except for as-cast pearlitic ductile iron. All test
specimens were V-notched (45°).

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 43
 10-17. Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades, 261
Using Unnotched Specimens

.00r-------,.-------,--------,.-----------,
F.,rillc: (60.4o-18 ann••IId)
unnotc'*l
3501--------t------+------+--------I

3001--------t----::;
e:lE '0 ]
~ 2501--------t-- S
=
~ ~
~
=
J 200
:
1

1501--------t------+------+--------I
20

1001--------+------+------1-------;

106
Faligue life. cYcles

.OOr--------,------.----------,r--------,

3501---------f~

3001--------;------'
t:. '0
:lE ~
ii 2501--------;-------+------......,1--------1
go
a~
J200

1501-------+------+--------1------.,
20
1001-------+------+--------1------.,

5:0'"'·-------'<--------';-------:--',--------'
106 10'
Faltgue lIle,cycles

Unnolched Notched Slreaa

Tenolle Endurance Endurance Endul' coneen-
alrenl!h UmII Endurance Umil ance lrallon
Grade MPa kal MPa kat rallo MPa kat ratio lacier
60·40-18 480 70 205 30 0.43 125 18 0.26 1.67
80-55-06 680 99 275 40 0.40 165 24 0.24 1.67

Top: Similar to upper graph on the opposite page, but here the specimens were unnotched. Bottom: Similar to lower
graph on the opposite page, but here the specimens were unnotched. Data in table pertain to graphs on this and the
opposite page.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 44
262 10-18. Ductile Iron: Fatigue Diagrams for Bending Stresses and
Tension-Compression Stresses

MPa 1000 psi MPa 1000 psi

100
600

600
80
400 5

60
400
JOO

·E 40
::; ~ 200
~ 200 Q)
c: o
c
5
-0
20 5 100
c: -0
W
c
W

o ~--:-f\P'A"':ld'-?-:!:---!c:---f::---,~~I:--- 1001
psi

-20 -100

-200

-40 -200
-JO

MPa -40 L-_-:-'-:-_-=-~_--:-'-,----_-,-:-_-=-'--

MP,
200 400 600 o 100 200 JOO 400 600
Mean Stress Mean Stress

Fatigue diagrams indicating endurance limits for five grades of ductile iron under bending stresses (left) and
tension-compression stresses (right). Minimum properties of the irons are given in the table below.

Min. Tensile Strength Min. Yield Strength

Iron Min. Elongation
No. 1000 psi MPa 1000 psi MPa
Percent
1 55 38 36 25 17
2 61 42 41 28 12
3 72 50 51 35 7
4 87 60 61 42 2
5 102 70 72 50 2

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, pp 344, 345 and 346
 10-19. Ductile Iron: Effect of Surface Conditions- 263
As-Cast vs Polished Surface

Tensile strength. ksi

90 100
350 n60r--_":,,,_-.--~......,.._--, 120 _':"';'::'_...:...r:.-....,....;.,~-, 50
.........._r----r-..........
n8 g

45
'" 3001---+----+---i--7"'q...---f---+-----i
a..
::2:
40 .<:='

200 1----+---+--7"'~_+_---'

25
150 _ __'___ _-'-_ _
~ __L. _ _- - - - ' " ' - _ ___'__ _ __ '

400 500 600 700 800 900 1000 1100

Tensile strength, MPa

Tests made on 10.6-mm (O.417-in.) diameter specimens. Fully reversed stress (R = -1).

Data given in the above graph show that the endurance limit for any given
strength level of ductile iron is significantly affected by surface conditions of
unnotched specimens. The endurance limit is much higher for the polished
specimens than it is for the as-cast specimens, which have relatively rough
surfaces.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 45
264 10-20. Ductile Iron: Fatigue Limit in Rotary Bending
as Related to Hardness

0
0
0
iii
0-
60

... ••
------
•
400
<U
0...
~
...
...
..
360

.....
60 'E
••
'E
::i
Q)
::J
. •~ 300
::i
Q)
::J
0)
';:;

,., •• •
0)
';:; 40 <U

• • U.

• • •
<U

.,....•
u, 0)
0)
260 .s
c "0

----I-
"0
c
Q)

• •
c lD
Q) 30
lD 200 ~
e
<U
• • s0
a:
0
a: 160
20
100 200 300 400
Hardness. Brinell

'iii
0- 60 <U
0...
0 - 400 ~
0
0 ...
...
'E 60 -
/ .~
. .,- 360
'E
::i
Q)

..
::i ::J
0)
Q) ';:;
::J

/
<U
0) u,

.
.~ 300 0)
u,
0)
c
40
r- A. .s
"0
c

~
Q)
"0
c 260 lD
Q)
lD z-
<U
z-
<U 30 I I I I I
0
a:
0 100 200 300 400 600 600 700
a:
Micro-Vickers Hardness Number

Top: Relation between Brinell hardness and fatigue limit in rotary bend-
ing for ductile iron. Bottom: Relation between rotary bending fatigue
limit and matrix hardness for ductile iron.

Source: Iron Castings Handbook, Charles F. Walton. Ed" Iron Castings Society, Inc" 1981, p 347
10-21. Ductile Iron: Effect of Rolling on Fatigue Characteristics 265

Rolling
75 Pressures
(Pounds) 500

70
406
580
65 768

60

'(ii
400
o,
o 55 ,, <0
n,
0
0
,, ~

en
(J)
"
50
~
u:; "",
" 83
45 " "-----------j
,----------- - - 300
Unnotched. Unrolled

40

35

30
Unrolled 200

10' 10' 10' 10'

Number of Cycles of Stress (Log Scale)

Fatigue strength of ductile iron can be increased substantially

by cold working, especially when this method is applied to
stressed radii or notches. More than a 60% increase in the endur-
ance limit was obtained with a rolling pressure that was insuffi-
cient to depress the surface a measurable amount. The im-
provement in fatigue properties obtained by various rolling
pressures on ductile iron is indicated above.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc .• 1981, p 348
266 10-22. Ductile Iron: Effect of-Notches on a 65,800-psi-
Tensile-Strength Grade

I 2.75"R I I
36
t 1 i - 250
<, 0.750" 0}17" 10.4'72"
0
<, ~
t
I
..
- 225
32
0
0"", I
Unnotched ~ ~()OOO - 200
'iij
0. 28 -"'0- eo
n,
o 0
o ~
o '\.

'\.0" W ~
- 175
W
(J)

~
Ul
t
~
"1
0.700" 0.417"
-i - 150

20
V-Notched
I'" ~
0
!
\
0.25 mm Root Had.
.....

o 0 125

1--0-000--
16

- 100
10' 10' 10' 10'
Number of Cycles

The unnotched and notched fatigue properties of an annealed

ductile iron with a tensile strength of 65,800 psi (454 MPa). The
endurance ratio is 0.41 and the notch sensitivity ratio is 1.67.

Source: Iron Castings Handbook. Charles F. Walton, Ed.. Iron Castings Society, Inc.• 1981, p 341
 10-23. Ductile Iron: Fatigue Crack Growth Rate 267
Compared With That of Steel

Stress Intensity Factor Range. t.K in MPa v'rri

20

Q) Q)

<3 <3
>- o
>-
o <,
:? 10-' E
o
c E
c c
Band for Wrought
Z Ferrite-Pearlite
z
"0
<, Steels 10-'
~
eo
'"
"0 "0

<Ii 2
m
a: '"
a:
s: s:

~ ~
o
(910-'
e
<.9
• 3300 Ib (1500 kg)
.:.!
.:.! o
o o 1650 Ib (750 kg)
~ ~
U .. 2700 Ib (1225 kg) U
• 2200 Ib (1000 kg)

10-' '--_...J......_ _...L-_L-...L-..I...-I....L...I..J....L-_ _..I...-_L-.L.....I

10 30 40 60 80100

Stress Intensity Factor Range. t.K in ksi yinch

Fatigue crack growth rate of annealed ferritic ductile iron,

compared with that of ferritic-pearlitic steels.

Source: Iron Castings Handbook. Charles F. Walton. Ed.. Iron Castings Society. Inc.. 1981. p 349
268 10-24. Malleable Iron: S-N Curve Comparisons of Four Grades

60 r--------...,---------,
400

350
50
.iii
0. <tI
0 c...
0
0
:2:
-0
-0 300 ~
<tI --l
0
--l
40

260

30 L- -'- ~

10' 10- 10'

Number of Cycles

The effect of cast surfaces on four grades of malleable iron

was also studied in high-stress, low-cycle fatigue. The results
with a 95% confidence limit are presented in this S-N diagram.
Unmachined and notched surfaces do reduce the fatigue
strength. The reduction factor is as low as 1.2 for the lower-
strength irons to over 2.0 for the higher-strength irons. Inducing
compressive stresses into the surface by rolling, coining, or shot
peening can increase the fatigue life of a component signifi-
cantly. Design with adequate sections that are well blended to
reduce stress concentrations is most effective in reducing the
possibility of a fatigue failure.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981,p 311
10-25. Pearlitic Malleable Iron: Effect of Surface Conditions 269
on S-N Curves

70

'00 60
C- 400 <0
o a..
0 60 ~
0
300 en
III
enIII 40 ~
~ Vi
Vi 30 200

20
10' 10' 10' 10'
Number of Cycles

The influence of as-cast surfaces, smooth ma-

chined surfaces, and machined notches on the fa-
tigue behavior of pearlitic malleable irons. Iron 1
is grade 60003 and Iron 2 is grade 80002.

Surface finish has an important influence on fatigue proper-

ties, as shown above. Samples of malleable grades 60003 and
80002 were tested in fatigue with "as-cast" and machined sur-
faces. Samples of the 60003 grade were also included with a ma-
chined surface containing a sixty-degree notch that was 0.050 in.
(1.25 mm) deep. The resulting data are shown in this diagram.

Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, p 310
270 10-26. Pearlitic Malleable Iron: Effect of Nitriding

1 PIECE
t-----

50.0

9 PIECP.S

ATHOSPIlr.RE
NITRIDED

COHPRESSIOH 11,160 POUND~

TENSION 7.600 POUNDS

1,01.---------------------------------
II) 5
HUMBER OF CYCLES

Effect of gaseous atmosphere nitriding on fatigue characteristics of pearlitic malleable

iron, tested by tension-compression.

Samples were austenitized, oil quenched, and tempered to 241-269 HB prior to

nitriding or testing without nitriding. This chart indicates an increase in fatigue life
of 750,000 to 2,700,000 cycles attained by nitriding.

Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals. Metals
Park OH, 1977, P 287
 10-27. Ferritic Malleable Iron: Effect of Notch Radius and Depth 271

Depth of notch, in.

o 002 004 006 008
250
~ tNotch
rad.ius
o 0.13 mm or 0.005 in.
I I
200 0.25 mm or 0.010 in. 30
~
c, I I
:;;
-5
0>
150
<, • 0.75 mm or 0.030 in. ~
E:
~ ............... - ~

--
20
t;

'"
5,
~
. r-- • '"
5,
.~ 100
u.
~ .~

r--- r-s- 10
u.

50

o o
o 0.5 1.0 1.5 2.0
Depth of notch, mm

Effect of notch radius and notch depth on fatigue strength of ferritic

malleable iron.

Fatigue strength of unnotched ferritic malleable iron is approxi-

mately 50% of the tensile strength, or from 170 to 205 MPa (25 to 30
ksi). The graph above summarizes the effects of notches on fatigue
strength. As a rule, notch radius has little effect on fatigue strength, but
fatigue strength decreases as notch depth increases.

Source: Metals Handbook. 9th Edition. Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 65
272 11-1. A286: Effect of Environment

A286
o AIR 593·C
w • VACUUM 593·C
C)
z VACUUM AVACUUM 20·C
: 0.01
z
c
a:
~
en
~ 0.001
en
c
-J
Q.

Plastic-strain range versus fatigue life for A286 ferrous alloy in air and in vacuum af
593°C (1095 OF). Numbers adjacent to test points indicate frequency in cycles per
minute. Note absence of frequency effects in vacuum.

Coffin has suggested that for a number of materials, virtually aU of the degrada-
tion in fatigue life at elevated temperatures can be attributed to environmental
interactions. He noted that frequency effects in the low-cycle-fatigue law could be
eliminated for a large number of metals and alloys by testing in vacuum (note
above). Additionally, it was noted that tests performed in vacuum showed trans-
granular crack nucleation and propagation versus intergranular nucleation and
propagation in air at elevated temperatures. These results are not unambiguous,
since Koburger has shown a frequency effect in high-cycle fatigue for directionally
solidified eutectic alloys when tested in air and in vacuum, particularly at elevated
temperatures. The primary difference in these results may be related to the lack of
intergranular cracking in eutectic alloys.

Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 343
 11-2. A286: Effect of Frequency on Life at 593°C (1095 OF) 273

/.--------- - ------ --:;....--:;;.--

~ ~---
--
l!>;:;::--:: /' ---
1>-;;:::::::0 /' /' v STANDARD HT-AIR
:0 -: oHT#1 -AIR
".0 ,/ • HT#I -VACUUM
/ 0 HT#2 -AIR
) A 286-593°C I> HT #3 -AIR
/ Kr=3.0 .MI. =60ksi 0 DS HT # 3 -AIR
2 ~DS-STD HT - AIR

Effect of frequency on life of notched fatigue bars of A286 at 593°C (1095 OF)
in air and vacuum. As indicated, decreasing frequency has a degrading effect on
fatigue life ofsamples tested in air, with little or no effect on samples tested in a
vacuum.

Source: L. F. Coffin, "Fatigue in Machines and Structures-Power Generation," in Fatigue and Microstructure, American Society
for Metals, Metals Park OH, 1979, P 13
274 11-3. A286: Fatigue Crack Growth Rates at Room and Elevated
Temperatures

AK, ksi • In.1I2

10 20 30 40 60 SO 100

t 10- 4

A' /
10-3 53S·C (1000.F)~' /
, /
427·C (SOO·F) ,
/ A
/ 01
U
/ / "'-24·C (75·F) 10-6 ~
,/ / .5
2'
10-4 (i /
I" 316·C (600·F)
~
"tl

I, /
ill 10-6

10-6
I
, I

A·2S6

I
10 20 30 40 60 SO 100
Stress-intensity factor range, AK, MPa ' m1/2

Fatigue crack growth rates for specimens of A286 stain-

less steel at room temperature and elevated temperatures
for tests in air at 3 Hz (RT) and 0.67 Hz (elevated temper-
atures), anR ratio ofO.OS, and at L-T, T-L, R-L, and R-C
orientations.

The austenitic precipitation-hardening stainless steel A286

(heat-resistant alloy) is the main representative in this category.
It contains titanium and small amounts of vanadium and alumi-
num, which precipitate as intermetallic compounds such as Ni,
(AI, Ti) and Ni 4Mo(Fe, Cr) Ti on aging. Various mill forms of
the alloy are usually supplied in the annealed condition-
Condition A (980°C, or 1800 OF, for one hour followed by
quenching in oil or water). Precipitation hardening occurs on
aging in the range from 700 to 760°C (1300 to 1400 OF)for 16
hours. Other combinations of heat treatments may be used de-
pending on the application. One variation is to re-solution treat
at 900°C (1650 OF)for two hours, quench in oil or water, and
age at 700°C (1300 OF) for 16 hours. This variation results in
improved room temperature properties but less desirable stress-
rupture properties.

Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, p 161
 11-4. Astroloy: S-N Curves for Powder vs Conventional Forgings 275

100
Conventional Forgings
I
'"I o Powder Forging
2 80
' .
x
~
\I)
\I) 60
w
...
<J<:

'"
>-
<J<: 40
...
0
s<II
s 20

0
104 10 5 10 6
CYCLES

S-N curves for conventional and powder forgings of Astroloy (notched versus
smooth),

Testing was performed using standard methods at 705°C (1300 OF) and a
combination of steady and vibratory stresses for which comparative data
were available. Cycles to first indication (crack) were comparable to con-
ventional material. Crack propagation as judged by the number of additional
cycles from first indication to failure was slower than conventional material,
as shown above.

Source: M. M. Allen, R. L. Athey and J. B. Moore, "Application of Powder Metallurgy to Superalloy Forgings," in Source Book on
Powder Metallurgy, Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979, P 97
276 11-5. Astroloy: Powder vs Conventional Forgings Tested at 705°C
(1300 OF)

Steady Stress = Vibratory Stress = 40,000 psi

x = Crack
o = Failure
Powder
Forging

Conventional
Forging

10 4
CYCLES

Astroloy tested in high-cycle fatigue at 705 °C (1300 OF). Vibratory

stress levels were selected to facilitate a direct comparison between con-
ventional and powder forgings.

Source: M. M. Allen, R. L. Athey and J. B. Moore, "Application of Powder Metallurgy to Superalloy Forgings," in Source Book on
Powder Metallurgy, Samuel Bradbury. Ed., American Society for Metals, Metals Park OH, 1979,p 97
 11-6. FSX-430: Effect of Grain Size on Cycles to Cracking 277

lSI 4)0

0.4 large grains t---O-I- IG

fine grain a
\ , ~-LG

~ I I I 0.01
~
if , HI N HZ
10.1~",1 E
E
• \ \
MIN AVG MAl
'"Z
C '\
\
\
i
..
II:

~
II: 0.01 .,;
.
~
." 0.)
c \
(O.~'I ~
0
c

....& ~cb
II: II:
~ ...
......
(
\ > 0.05
10.7621
c
...
II:

~,
~
!:
~ 0.0"'
11.0161
...
II:
L

"-
0.2

o 40 .0 IZO 160 ZOO 240 zeo no

CYCLES TO CRACK

SoN curves for alloy FSX-430, showing effect of grain size on cycles to cracking,

Source: Eric Bachelet and Gerard Lesouit, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed.. American Society for Metals, Metals Park OH, 1984, P 336
278 11-7. FSX-430: Effect of Grain Size on Fatigue Crack Propagation
Rate
GROWTH RATE IN llNlAl! flAIlGE ,",,,,/CYCLE I
o 0.01 ODZ O.OJ 0.04

~
J--O-1-fG
0.4
~-LG
I I I I
CRACK AVG CflACK 0.01 _
I 2 t-{j----I (O.Z:>41 E
I E
I
I
...
.
~
'" < 0)1
I
I
O.O!
(o.~oeio
a:

!:

,,
z· 0.3
I c
:ca:
....
~ .
'"
C

C
...%
IE
< ,
It' ) 0.03 II:
(O.TUI~
e,
II:
Q,
< )~ 0.04 ~

..
...i
I
I
(1.0151

0.2 large grlolinB

fine grains

o 0.5 1.0 1.5 2.0

GROWlH RAlE IN LINEAR flANGl (IO·"n.lCTCLEI

Fatigue crack propagation rate-effect of grain size on fatigue characteristics of FSX-430.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of SuperalIoys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr.. Ed.. American Society for Metals. Metals Park OH, 1984, P 337
 11-8. HS-31: Effect of Testing Temperature 279

70
! Fatigue strength
60 100 million cycles
.u; ..........
a. 50
( Aged 50 hr at 1350 F
""""-<
o "'- r---.....
o
Q 40
30
20
1100 1200 1300 1400
(["---...:

-- :--

1500
Testing temperature, F

Effect of testing temperature on fatigue strength of HS-31 cast-

ing alloy, after aging at 730°C (1350 OF), for 100 million cycles.

Source: ASM Committee on Heat-Resistant Castings, "Heat-Resistant Alloy Castings,"in Source Book on Materials for Elevated-
Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH, 1979, p 237
280 11-9. IN 738 lC Casting Alloy: Standard vs HIP'd Material

220 y-"STANDARD CONDITlON" MATERIAL

CII
<,
I 200 • '-.<,
E
E <, • MATERIAL
:z; <,
. 180
<,
+1
Po.

Po.
......,
Ul
160

11M>
-"<.

" ...... ......

<,
-
-
Ul
.............. tI$-
~ ..............
E-i
Ul 120

107

CYCLES TO FAILURE
SoN curves for casting alloy IN 738 LC. High-cycle fatigue properties ofnimocast alloy IN 738 LC tested at
850°C (1560 oF).

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Super alloys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, P 340
 11-10. IN 738 LC: Effect of Grain Size on Cycles to Failure 281

240
/ STANDARD CDNDITIDN MATERIAL
N
I
E
E
220
" -,
.
2:
.
a..
200
• ......
<,
.......
• FINE GRAINS
.......
180 ......
+ I
....... •
a.. ......
.......
111
111
160
• ......
' •
. •
-
....... ~
W
0:
t-
•
140
111 •
120
e·
6 7 8
10 10 10
CYCLES TO FAILURE
SoN curves for alloy IN 738 LC. High-cycle fatigue properties of extra-fine-grain and conventional material
tested at 850°C (1560 OF).

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book. Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, p 340
282 11-11. IN 738 LC: Effect of Grain Size on Cycles to Cracking

0.4 'N - 738

.......-o--i - fG
~ -LG 0.0'
o.ZS41 'i
I I I !
M'N. HZ N ••, MAX."2 .!i
~ a:
...
w• 0.02
.;
::l
_ O.J IO.!lOI1 0
! c
c
...
e::
~~)-~......--<O)-----~
, 0.03 a:c
a:
oJ

...
-
10.JUI
:... <,

~
........ 2:
~
f .... 0.04 ...a:
....
c ~ 11.0161 L

fine grains
large grains
0.2

o eo .20 160 200 2.40 :tIO 320

CYCLES TO CRACK

S-N curves for alloy IN 738 Le, showing the effect of grain size on number of cycles to cracking.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, P 335
 11-12. IN 738 LC: Effect of Grain Size on Crack Propagation Rate 283

c.ROWTH IlATE IN L1I1EAR 11&",[ l ..." u n l l

O.GI 0.01 O.OJ 0,04
, , I ,

•. 4

0.01
O.~~I
E
!
s
t; O.GZ
&
.:
(0.!>08)

..•~
.0.) f-
.
! Ci

..•....
~

QOJ
(0.7611
........
.,
... f-
..'"
%

f
•
...
L IN ' ) I
0.0-
1.01'1 .....
i

0.2 f- t-O-i - FG

:E-C»i - LG
I I I
, I CRACI( I /IIIG CU~K I
I
o 0.' 1.0 1.5 1.0
GIIOWTH RAT( IN LINEAR RANG[ llli'in.lCYCLL,

SoN curves for alloy IN 738 Le, showing the effect of grain size on crack propagation rate.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OB, 1984, P 336
284 11-13. IN 738 LC: Fatigue Crack Growth Rate at 850°C (1560 OF)

10- 5
do
dN (m/cycle)
10- 6
100Hz

•
Alloy I IN 738 LC
Temperature I 850°C
.0 fine-grained
10- 10 Abo coarse - grained
R t'<J 0.1
Waveform: Sinusoidal
10- 11
o 10 20 30 40 50 60
-3/2
~K(MNm )
Fatigue crack growth rate at 850°C (1560 OF) in various grain sizes of alloy IN 738 LC.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book,' Matthew J. 000-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984,p 341
 11-14. Inconel 550: Axial Tensile Fatigue Properties in 285
Air and Vacuum at 1090 K
50

40

30
2
(a)

....
C>
20

'iii
0-
10
vi
VI
w 50
a:
t-
VI ·VACUUM
Z °AIR
<Cl:
w 40
::E

30 (b)

20

10
105 107
LIFE. cycles
I I II lid I I "" II

10 100 1000 10000

LIFE. h

Axial tensile fatigue properties oflnconel550 at 1090 K in

air and vacuum. (a) Ratio of cyclicto mean stress= 0.125.
(b) Ratio of cyclic to mean stress = 0.667. Testing fre-
quency = 33 Hz.

In reversed bend tests on lead at 500 cycles/min, Snowden

demonstrated a difference oftwo orders of magnitude in fatigue
life between vacuum, air, and pure oxygen. At all strain levels
vacuum endurances exceeded those in air, which exceeded those
in oxygen. Intermittent stress-free exposure to air had no effect
on the lifetime in vacuum. At high temperature (l090 K) vac-
uum also improved endurance, relative to air, of the Co-base
alloy S-816 and the Ni-base alloy Inconel550, although the ef-
fect was much smaller than that seen in lead. Endurances for the
nickel-base alloy converged at low stresses, indicating a possible
strengthening effect of air, as shown above.

Source: R. H. Cook and R. P. Skelton. "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
in Source Book on Materials for Elevated-Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals
Park OH, 1979, p 81
286 11-15. Inconel 625: Effect of Temperature on Cycles to Failure

600

80

500 427°C Joo of)

,~ e:-- 29°C (85 of)

400 '-
- 538°C (1000 °F)_ 60
<0
a.
:;;
Ii 300
~
200
<,
...
NOICh~~
specimens
--- I-
649°C (1200 of)
760°C ('1400 of)

i
29 Oc (85 of)

(Kt~3.3) ~
871°C (1600 of) _
20
100

Cycles 10 failure

S-N curves for hot rolled solution treated Inconel625 bar

15.9 mm (0.625 in.) in diameter at various temperatures.
Average grain size was 0.10 mm (0.004 ln.),

Source: Metals Handbook, 9th Edition. Volume 3. Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 143
 11-16. Inconel 706: Effect of Temperature on Fatigue Crack 287
Growth Rate

Stress-intensity factor range, aK, ksi • in.1/2

10 20 30 40506080 100 200 300

10- 4

-196°C (-320°F)
10-3
Ql
U Ql
> U
--EE
u
~
--z·,E
z
--
"C
III
"C
10- 5
--
"C
III
"C
l!l ....
..
III

s:
Ql'
III

s:
i i
e
Cl
-269°C (--452°F) e
Cl
~
u 10- 4 ~
u
I!
u I!
U
Ql
::J Ql
Cl ::J
'':; Cl
III '':;
II.. III
II..

10-6

20 30 40 50 60 80 100 200 300

Stress-intensity factor range, aK, MPa . m1/2

Fatigue crack growth rates of Inconel 706 forged billet (vacuum induction
melted/vacuum are remelted) at an R ratio of 0.1 and a frequency of 10 Hz.
Heat treatment: 980°C (1800 OF) I h, AC; double aged 730°C (1350 OF) 8 h,
FC to 620°C (1150 OF), hold 8 h, AC.

Results of FCP tests at room temperature and at temperatures as low as

-269°C (-452 OF) for Inconel 706 are shown above, At equivalent 11K
values, the fatigue crack growth rates for this alloy are slightly lower at sub-
zero temperatures than at room temperature,

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties ofSuperalloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, p 297
288 11-17. Inconel"713C": Effect of Elevated Temperatures on
Fatigue Characteristics

50
"in
c.
g 4 a I-+-+-f-Ht--~H-Yt---+--+
o
:i 3 a 1--+-+++t--~,*h-r--+--P't"+l:>---1--+t-t-l
Q)
"-
iii
Ol 2 a I--+--+--r-f-..---r---t-+-+--i---i--t-++t---+--t-+-+i
~ NOTE: Higher fatigue strength at
o 1500 F than at 1200 F is consistent
E 10 with tensile strength relations in
.2?
<i graph shown above left.
OL-..I...-...L....J.....LJ...---l...~
.............1...--'--...L....J.....LJ...---L.----L~
0.1 I 10 100
Millions of cycles to failure

SoN curves for Inconel "713C." Tests were performed at

two different elevated temperatures as shown.

Source: ASM Committee on Heat-Resistant Castings, "Heat-Resistant Alloy Castings," in Source Book on Materials for Elevated-
Temperature Applications. Elihu F. Bradley, Ed.. American Society for Metals. Metals Park OH. 1979, P 235
 11-18. Inconel "713C" and As-Cast HS-31: Comparison of Two 289
Alloys for Number of Cycles in Thermal Fatigue to Initiate Cracks

I r ,-
Thermal fatigue Thermal fatigue
Material Cycles to first Material Cycles to develop
crack 8I-m,
.
erne k
avg avg1
ri
HS-31
Inconel
1713C"
-1& HS-31
Inconel
1713C"
IIIII
l-rnin cycles, a I 2 3 3 tests a I 2 3
100 to 1700F Thousands of cycles each motertol Thousands of cycles

Thermal fatigue properties of HS-31 compared with those of Inconel "713C." Left: Number of cycles re-
quired to initiate cracks. Right: Number of cycles required to develop VB-in. crack.

Source: ASM Committee on Heat-Resistant Castings. "Heat-Resistant Alloy Castings." in Source Book on Materials for Elevated-
Temperature Applications. Elihu F. Bradley, Ed.. American Society for Metals. Metals Park OH. 1979, P 235
290 11-19. Inconel718: Effect of Frequency on Fatigue Crack
Propagation Rate

I I I I I I

5 X 10-3 f- -

o
2
o 5 X 10- Hz } 0

05 X 10-' H, \ o",~/ / •
o I
5 X 10-4 - 1> & -
o 00 /S
/ili.
o /
o /;".
/&

5 X 10-5 I- o
o
o

a~a / I.•
tL {-- &
20 Hz
2Hz _
a
/
I a 0.5 Hz
~/
&&/
& //

& ......---/

&1
/
5 X 10-6 1..----I._ _--1.._--'_...L...--1..--'---'--'-...L........L.---l
8 10 15 20 30 40 50 60
Stress-intensity factor range, 61<, MPa . m1/2

Variation ofFCP rate (da/dN) with stress-intensity factor range

(LlK) and frequency at 550°C (1025 OF) (sinusoidal load) for
specimens oflnconel 718.

The effect of frequency at 550°C (1025 OF) was studied using

a sine wave; the results are shown above. Below 0.5 Hz, the FCP
rate was more rapid, and the crack surfaces showed an increased
amount of intergranular fracture with decreasing frequency,
with the crack path following the boundaries of the largest
grains. One may be inclined to attribute the increase in FCP to
either creep or environment, but this may not be the case, be-
cause different modes of deformation may have occurred at dif-
ferent strain rates.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 294
 11-20. Inconel718: Relationship of Fatigue Crack Propagation Rate 291
With Stress Intensity

• 25°C (77°F)
• 550°C (1025°F)
20 Hz
5 X 10-4 ••
••
• • ••
..
••
•••
• •
.
5 X 10-5 • •
• ••
.. , • •
••
••
•••••
••• •
5 X 10-6 :-• ••
• •
._~

: Twins No twins
10- 6 ......--L ....L._--'-_.J...-............L.....L.........L.-............

8 10 15 20 30 40 50 60

Stress-intensity factor range, Li K, MPa . m 1/2

Dependence of FCP rate (dol dN) on stress-intensity fac-

tor range (LiK) and temperature at 20 Hz (sinusoidal
wave shape signal) for specimens of Inconel 718.

An important effect is the hydrostatic state of stress in the tip

region. This idea has been considered for Inconel 718. The FCP
response is shown above for 20 Hz, where the effect of tempera-
ture is to increase the FCP rate, especially at LiK levels.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 290
292 11-21. Inconel 718: Relationship of Fatigue Crack Growth Rate
With Load/Time Waveforms

a K. ksi • in. '/2

10 20 30 40 60

• /'v (2)
5 X 10-2 Hz
{ _ rxr (1)

2 Hz D /'v (3) I (2)

I I

Sinusoidal ~ rf· (1)

~
'Jr¥
4l

!
2
5 X 10- Hz ~ - / (3)
5 X 10- 4 -;0 2X 10-6
I.• _ -Q .5

..
r:t
I
I... /
- /
z·
"tl

" tl
I t6
/ rI
,6 ~ Sinusoidal
2 Hz

5 X 10- 6 2 X 10- 6

8 10 15 20 30 40 50 60

aK. MPa • m,/2

Load/time waveforms and FCP rates for specimens of Inconel

718. Top: Various forms of cyclic stress fluctuations used at 550
°C (1025OF)at a frequency of5X 10- 2 Hz. Bottom: FCP rates at
550°C under sinusoidal, triangular and square loads.

To separate out the possible effects of creep or environment from deforma-

tion mode, the authors used triangular and square wave shapes, as shown in
the top graph. The data obtained using the triangular wave at 2 Hz were the
same as the data obtained in other tests using the sine wave at the same
frequency which resulted in the lowest FCP rate. The effect of loading at the
same rate but imposing a lO-second hold time at maximum load was to in-
crease the FCP rate only slightly, as shown in the lower graph.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 295
 11-22. Inconel 718: Fatigue Crack Growth Rate in Air vs Helium 293

~K, ksi • iny2

10 20 3040 6080
5 X 10-1 1'"T'"--"""--T"-"""-T""T""1'""'I""l

• = Air
10- 3
o = He
.,
u
~
--EE 10- 2

z'
--.,'..
"tl

"tl
U
Gl

>
u
..... ......
.5
.I: 10- 4 Z'
i0
..
0>
--.
"tl

"tl

...
.¥
u

.,
u
::;,
10-3

..
0>
.;::;
u..

10-5

10-4 ......._ _L - - J ' - -........I-l'-'-..........

10 20 40 60 80100

~K, MPa . m1/2

Fatigue crack growth rate data for Inconel 718 in

air and in helium. Frequency, 0.1 Hz. Tempera-
ture, 650°C (1200 OF). Here it is evident that crack
growth rate at constant temperature is lower in in-
ert gas.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 287
294 11-23. Inconel 718: Effect of Environment on Fatigue Crack
Growth Rate

~K, ksi • in. 1/ 2

10 20 30 40 60 80
5 X 10-1

10- 3

.,
u
~ 10-2
--E
E
z· .,
~ U
~
III
"'C

....,' --.5
E 10- 4 z·
.t:
i
e
--
"'C
III
"'C
CI

"'u..."
III 10-3
.,
U

:::l
CI
'';::;
III
u..

10- 5
• = He + 0.5% H 2S
0= He + 5% S02
10- 4 L-_----I_~---L___L.....I.....L.J...L.J
10 20 40 60 80100

~K, MPa • m 112

Fatigue crack growth rate data for Inconel 718 in

helium + 0.5% hydrogen sulfide and helium + 5%
sulfur dioxide. Frequency, 0.1 Hz. Temperature,
650°C (1200 OF).

From the data above it becomes obvious that fatigue crack

growth rate increases greatly in aggressive environments com-
pared with exposure to helium alone.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 288
 11-24. Inconel 718: Fatigue Crack Growth Rate in Air Plus 295
5% Sulfur Dioxide

AK. ksi • in. ' / 2

10 20 30 40 60 80
5 X 10- 1

Air + 5% S02 10- 3

II>
Q
> 10- 2
--
u
E
E
z·
--'"
"C

"C
II>
Q
>
~
i .S
...'" 10- 4
~ z
i ~
e
." '"
"C
~
u
eu 10- 3
II>
:s
."
.;;
'"
II..

10-4 '--_ _.J...._-'---'--'-...L..J.....L..L..J

10 20 40 60 80100
AK, MPa • m' /2

Fatigue crack growth rate data for Inconel 718 in

air + 5% sulfur dioxide.

(The effect of air plus 5% S02 was similar to the effect of air alone.) It was observed that in
the helium atmosphere, which was used to establish a baseline, cracking was generally trans-
granular with well-defined striations. In the air, oxygen-bearing and sulfur-bearing environ-
ments, the crack path changed from transgranular to intergranular, indicating that an important
effect of the environment was to degrade the boundary strength by mechanisms that were not
clearly defined. It was suggested that oxygen diffusion along grain boundaries and localized
oxidation may have occurred. Another very important observation was that the effect of a given
environment on FCP could not be predicted on the basis of unstressed exposure tests. The attack
on the surfaces of unstressed specimens in aggressive S02 environments was minimal, but the
S02 environments caused substantial increases in FCP.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 289
296 11-25. Inconel 718: Fatigue Crack Growth Rate in Air
at Room Temperature

Spec. 1290" /
/
(CHT) 'f

III
/
: ,
~
E
Spec. 158 & 803'" (!.
E
10-3 (CHT) / v'
.. ,I:!> 0
/ ,
Z'
~
.. I 0
"~.'" Inconel718
Tested in air at
24°C (75°F)
500 < f < 600 cpm,
R = 0.05

Conventional Modified
heat treatment heat treatment
I:!> Spec. 1290, heat I I:!> Spec. 1283, heat I
0 Spec. 158 } heat II 0 Spec. 253, heat II
V Spec. 803
0 Heat III I
I (b)
20 40 60 80 20 40 60 80
Stress-intensity factor range, ~K, MPa • m 1/2

Fatigue crack growth rate behavior ofInconel718 tested in air at 24 ° C (75OF).CHT=

conventional heat treatment. All testing was done at R = 0.05 and at a frequency of
0.67 Hz.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys,..in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH. 1982, p 276
 11-26. Inconel 718: Fatigue Crack Growth Rate in Air 297
at 316 °C (600 OF)

Spec. 210 (CHT)

Spec. 1294 (CHTl

y "IJ./
-:
.. ,
.I IJ./
/ ,~
Inconel718 .. t:J.
Tested in air at ..I1J.4'
,0
316°C (600°F) / IJ.I 0
f = 40 cpm, R = 0.05 .. IJ. .,
/
./
/

Conventional Modified
heat treatment heat treatment
IJ. Spec. 1294, heat I IJ. Spec. 1282, heat I
o Spec. 210, heat II o Spec. 254, heat II
o Heat III
(a) (b)
10- 6 L..-_ _--I._--'_....L--L.-L-'-...L...J-'- .L..-_....L--''--J.-........--L....J-I

20 40 60 80 20 40 60 80
Stress-intensity factor range, Ll. K, MPa . m 1/2

Fatigue crack growth rate behavior of Inconel 718 tested in air at 316°C (600 OF).
CHT = conventional heat treatment. All testing was done at R = 0.05, and at a fre-
quency of 0.67 Hz.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 277
298 11-27. Inconel 718: Fatigue Crack Growth Rate in Air
at 427 °C (800 OF)

10-1 ...----r--,-__r-..___..___.....-r::I:~--__r--..___.....,..____.____.___r.... T.:I

Inconel718
Tested in air at
427°C (800°Fl
f = 40 cpm, R = 0.05
Conventional Modified
10- 2 heat treatment heat treatment
D. Spec. 1291, heat I D. Spec. 1286, heat I
o Spec. 162, heat II o spec. 255, heat II I
:1
Spec. 162 (CHT1,,/1 0
/'1 0
10-3
/.~
Spec. 1291",: ~ 0
(CHT) lID.
.. I
II
.. I
10-4 II
.. I 0
.II <9
//
10-5 L...- L . - _ - ' - - - ' - _ . L - . L -................... ----'_ _.L---L----'----''---L...............

20 40 60 80 20 40 60 80

Stress-intensity factor range, AK, MPa • m1/2

Fatigue crack growth rate behavior of Inconel 718 tested in air at 427°C (800 OF).
CHT = conventional heat treatment. All testing was done at R = 0.05 and at a fre-
quency of 0.67 Hz.

Source: Stephen D. Antolovich and J. E. Campbell. "Fracture Properties ofSuperalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH. 1982, p 278
 11-28. Inconel 718: Fatigue Crack Growth Rate in Air 299
at 538 °C (1000 OF)

10- 1 1:""""----,--...-~-,.._r_1r_T"T"']r:__--__r--.___r-,.._r_1r_T"""1"":I

Inconel 718
Tested in air at
538°C (1000°F)
f = 40 cpm. R = 0.05
10- 2

10- 3

10- 4
Conventional Modified
heat treatment heat treatment
Spec. 1288, heat I Spec. 1284. heat I
Spec. 165, heat II Spec. 251} heat II
heat III Spec. 250
10- 5 '--_ _----'C--_.1.--L_.l.-.L.....JL..-L....L.J'--_ _----'_ _-'------'_.l.-.L:--JL.....L~
20 40 60 80

Stress-intensity factor range, LiK. MPa . m 1/2

Fatigue crack growth rate behavior of Inconel 718 tested in air at 538°C (1000 OF).
CHT = conventional heat treatment. All testing was done at R = 0.05 and at a fre-
quency of 0.67 Hz.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of'Superalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 279
300 11-29. Inconel 718: Fatigue Crack Growth Rate in Air
at 649°C (1200 OF)

1
10- ~---r---r-'---'-"'T1-r:r-----r----,--,--r""'T"r-r""']

Inconel 718
Tested in air at
649°C (1200°F)
f = 40 cprn, R = 0.05

/
Spec. 1289 V·'
(CHT) •

10- 3 /
./
./ !SJ
/
10- 4 Conventional Modified
heat treatment heat treatment
b. Spec. 1289, heat I Spec. 1281, heat I
o Spec. 156, heat II Spec. 252, heat II
o heat III

20 40
Stress-intensity factor range, l1K, MPa . m 1/2

Fatigue crack growth rate behavior of Inconel 718 tested in air at 649°C (1200 OF).
CHT = conventional heat treatment. All testing was done at R = 0.05 and at a fre-
quency of 0.67 Hz.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties ofSuperalloys."inApplication of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH. 1982, P 280
 11-30. Inconel 718: Fatigue Crack Growth Rates at 301
Cryogenic Temperatures

• 22° C (72° F) }
l> -78° C (-108° F) 2.54 cm
'V -196° C (-320° F) thickness
o -269° C (--452° F)

Q) o 22° C (72° F) 0.51 cm

U 10- 3
e thickness
EE
z
~
<tI
"0

~f
~
.r:
~
e
Cl
.:.!

~ 10- 4
CJ
Q)
:I
Cl
.;:;
<tI
u..

5 10 50 100

Stress-intensity factor range, AI<, MPa • m 1/2

Fatigue crack growth rates of Inconel 718 forged bar at an R

ratio of 0.1 and a frequency of 20 Hz. Heat treatment: 980°C
(1800 OF) '% h, AC; double aged 720°C (1325 OF) 8 h, FC to
620°C (1150 OF), hold 10 h, AC. At the constant frequency the
effect of higher temperature is to increase the FCP rate.

Source: Stephen D. Antolovichand J. E. Campbell, "Fracture PropertiesofSuperalloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 298
302 11-31. Inconel718 and X-750: Fatigue Crack Growth Rates
at Cryogenic Temperatures

Inconel 718 (Ref 8.44)

~ 22 to -269°C (72 to -452°F)

Inconel 718 (Ref 8.49)

• • 22°C (72°F)
Inconel X-750 (Ref 8.48)
~ 27 to -269°C
(80 to -452°F)

5 10 50 100

Stress-intensity factor range, ~K, MPa • m 1/2

Fatigue crack growth rates for Inconel718 and Inconel X-7S0in

the subzero temperature range.

A comparison of FCP values from room temperature to

-269°C (-452 OF)for Incone1718 and Inconel X-750 is shown
in the above chart, along with room temperature FCP data for
Inconel 718 from Shahinian et al. The FCP data for these two
alloys overlap in the t!K range shown. Under some conditions,
the FCP rate for Inconel 706 is slightly less than those for In-
conel 718 and Inconel X-750 at corresponding temperatures and
t!Klevels. However, results of FCP tests depend on both melt-
ing practice and thermomechanical processing.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OB, 1982,P 300
 11-32. Inconel X-750: Effect of Temperature on Fatigue Crack 303
Growth Rates

CIl
• 22°C (72°F)
u A -196°C (-320°F)
!E 10- 3 o -269°C (-452°F)
E
Z
~
"tl
:!l'
E
ie
en
~
CJ
b 10-4
CIl
5,
.~

'"
II..

100

Stress-intensity factor range, llK, MPa . m 1/2

Fatigue crack growth rates of Inconel X-750 at an

R ratio of 0.1 and at frequencies of 20 to 28 Hz.
Heat treatment: solution treated and double aged.
Within this frequency range, the effect of higher
temperature is to increase the FCP rate.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 299
304 11-33. Jethete M 152: Interrelationship of Tempering Treatment,
Alloy Class, and Testing Temperature With Fatigue Characteristics

ksl
100r------------------,
o 0.6

...:::>~ )(--------x 80
'"
~ 0.4
=ksi
100
0-_0______ i!:
...~
0:
Ii;
60

50
x

0_
x tempered at 510 F(300 Cllor one hour
to alensile slren&lh 0labout205 ksl
o temperedal1200 Fl650 ClIorone hour
0'
toa tensilestreneth abouI150ksl
0" o
...
G
~
40

20
LONGITUOINAL SPECIMENS
ROTATING·8EAM TESTS FOR \0' CYCLES

200
°O!----..!:--....,-k".--""*';---=----.od;;--__=!
1200 I
400 600 800 1000 F
I 1 I
50 450 200 300 400 600 C
TEST TEMPERATURE

Left: Interrelationship of prior tempering treatment and testing temperature with limiting fatigue stress, and
with fatigue ratio for Jethete M152. Right: Influence of alloy class and testing temperature on fatigue
strength for the same alloy.

Source: J. Z. Briggs and T. D. Parker, "The Super 12%CrSteels," in Source Book on Materials for Elevated-Temperature Applica-
tions, Elihu F. Bradley, Ed.. American Society for Metals, Metals Park OH, 1979, P 123
 11-34. Lapelloy: Interrelationship of Hardness and Strength With 305
Fatigue Characteristics
ksi
9o,------------..----, CLASS II (Lapelloy) 90,---------------,

2000 F C1 095 Cl sail quench,

marlemper 650 F(345 C) +
lemper 1150/1600 F(620/B70 C)
CANTILEVER ROTATING·BEAM
TESTS
mechanically polished, BO
rms 2.5/4.0 mlcrelnehes
alhersurface Irealmenls,
o Including surface lolline.
rough e:rinding andlough
machining. rms5.5/40
microlnches 70

I 95% conlidencelimits

50 50

4~'='00,..----,-!;"....----:-!-::-----:*"--""IBO ksi

Left: Relation between surface hardness and mean fatigue limit for Lapelloy. Right: Relation between
tensile strength and mean fatigue limit for the same alloy.

Source: J. Z. Briggsand T. D. Parker, "The Super 12%Cr Steels," in Source Book on Materials for Elevaled-Temperature Applica-
tions, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH, 1979, p 123
306 11-35. Mar-M200: Effect of Atmosphere on Cycles to Failure

10
8
• AIR
o DRY AIR
• WET AIR 1
o VACUUM
6 PREOXIDIZED SPECIMEN. VACUUM TESTED

10

10
I 101 8
w
o
z
'"a:
V>
1
V>
w
a:
l-
V> 10

6
4

10
10 5
CYCLES TO FAILURE

S-N curves showing fatigue life at 10 Hz of single-crystal

low-carbon alloy Mar-M200 at 295-1200 K.

Convergence of air and vacuum data was noted for AISI 3I6
steel at 1090 K, and a crossover of the air and vacuum curves
occurred for nickel, where it was suggested that oxide in cracks
could prolong life in air at low stresses. Crossovers have also
been seen in a ferritic stainless steel and a Nil Cr alloy in the
range 875-1025 K, where tests in purified argon gave shorter
endurances than those in air, impure argon, or sulfur dioxide.
Also, in single crystals of the alloy Mar-M200, air endurances
were less than those in vacuum at room temperature whilst the
reverse was true at high temperature (above). A thin oxide film,
formed during testing, suppressed surface crack initiation, but
oxide formed during pre-exposure did not.

Source: R. H. Cook and R. P. Skelton, "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
in Source Book on Materials for Elevated-Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals
Park OH, 1979, P 81
 11-36. Mar-M509: Correlation of Initial Crack Propagation and 307
Dendrite Arm Spacing

5.0

U 3,0 u>.
..
U u
>.
u 4.5 N""-
N""- E
c E

'fg
:il~ 4. 0
-
'1'0

:il~
u
...
:£
or:
2.5 1;
or:
c c
0 0
iii
0\
iii0\
:!. 3.5 :!.
0
~ ...
0
no
......
ij
......
~
u
u u
;;; 2.0
3,0 ~c
;::
-
c
-
2.5 L-.. .L- ...r........ .....

20 40 60
Dendrite Arm Spacing 1~1

Correlation between the initial crack propagation rate and the dendrite arm
spacing for Mar-MS09.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys,' in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, p 338
308 11-37. Mar-M509: Correlation Between Number of Cycles Required
to Initiate a Crack and Dendrite Arm Spacing

300l r-----~---____r---- ......---__r---_,

~
~
u
...
... 2000
;§
c
-a...... - - - PresentStudy

~
u
'0
IlXXl
~
e
:::J
z

OL.-_.L.-_--L. -'- ---' -L.. ---'

o 20 40 60 100
Dendrite Arm Spacing (1111)

Correlation between the number of cycles required to initiate a crack and the
dendritic arm spacing for cast alloy Mar-MS09.

Source: Eric Bachelet and Gerard Lesoult, "Quality or Castings or Superalloys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984. P 337
 11-38. MERL 76, P1M: Axial Low-Cycle Fatigue Life of As-HIP'd 309
Alloy at 540°C (1000 OF)

100

90
00
80
0 0

'in 70 KT = 1.0
~

m'
fA 60

50

40

30
Life, cycles

Axiallow-cycIe fatigue life ofas-HIP'd P/M alloy MERL 76 at 540°C (1000

OF) at notch severities as indicated.

Source: J. H. Moil, V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications," in Powder Metallurgy-
Applications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, p 275
310 11-39. Nickel-Base Alloys: Effect of Solidification Conditions on
Cycles to Onset of Cracking

1000
F"t Solidification SlolI Solidlficotion
(Condition f) (Condillon S)

...
~...
......:>
Q
......... 100
..J
0.
r
...
Q

...'"
..J
~
U
...
s
~ 10
...a:
Q.

......r
...
Q

...a:
CD
r M21 71) 1I~71a IN!U
:> M21 713 IN738 '''939
:z lC lC lC LC
"~_ _..L.----1_......L_..L.._..L-_---1L.-_-L..._"""'---1_......L_-L...

(;;:s;J Creln 10 Ihl anul 01 crackIng

o Talal eycln 10 !raw crack to 2·5 "''''

Bar chart showing effect of solidification conditions on cycles to onset of

cracking and total number of cycles required to grow cracks to 2.5 mm
in several nickel-base casting alloys.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superal1oys: Source Book, Matthew J, Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, P 339
 11-40. Rene 95 (As-HIP): Cyclic Crack Growth Behavior Under 311
Continuous and Hold-Time Conditions

Stress-intensity factor range, L1K, ksi • in. 1/ 2

6 8 10 20 40 60 80 100
10-2

10- 1

Stress-level
dependence
Ql Ql
U 10-3 U
>-
u >-
u
-<; '<,

E .~
E 10-2 Z'
z' 15-minute hold
--
"C
time at maximum --
"C
IV
IV

..
"C
Ql'

f!
tensile stress .....
"C
Ql'
IV

10- 4 ..c
..c
~
~ 2Cl
...0
~
Cl
10- 3 ~
u
u f!
f! u
U Ql
Ql ::3
::3 Cl
Cl '';:;
'';:; IV
IV
10- 5 U.
U.

10- 4

'--.1...--'-_---1._ _........." --_ _- ' -_ _...1-_"'-----'-1 10- 6

8 10 20 40 60 80 100
Stress-intensity factor range, L1K, MPa • m 1/2

Cyclic crack growth behavior for as-HIP Rene 95 under both continu-
ous and hold-time conditions at 650 °C (1200 OF).

The effect of environment need not always lead to more rapid crack growth. It has been
proposed that oxidation products could form in the crack tip region and prevent crack
resharpening during the unloading portion of the cycle. If the stresses are sufficiently low,
the oxidation products in the crack tip region will not be cracked and, in some systems, an
elevation of the threshold might occur. Such effects would be pronounced at high temper-
atures and long hold times and have actually been observed in Rene 95, as shown in the
above chart. Once the stress intensity is high enough to crack the oxides, the rate of crack
growth would be expected to increase due to the severely degraded region in the crack tip
zone.

Source: Stephen D. Antolovich and J. E. Campbell. "Fracture Properties of Superalloys. "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood. Eds.. American
Society for Metals. Melals Park OH. 1982. P 284
312 11-41. Rene 95: Effect of Temperature on Fatigue Crack Growth Rate

Testing temperature, • F

10_2 20 100 1000 2000

U
.. 10-4
~
{ .!
E .5
z·
....
z·
~
10- 3 ...
-.r
:i
...
l! i'
~ ~
e e
11'" 11'"
.."
l!
:::J
10-6
..~
:::J
'"
.~
... '"
...'=
10- 4

o
Testing tempereture, ·C

Effect of temperature on fatigue crack growth rate at constant

bJ( for Rene 95.

That the effect of environment can be large may be inferred from some low-cyclefatigue
studies of Rene 95 in which surface and subsurface cracking was observed at comparable
strain ranges and defect sizes. As expected, the life of the subsurface crack was much
greater than that of the surface crack, leading to the hypothesis of a strong environmental
effect. This possibility is considered in more detail in an analysis of FCP properties of
Rene 95. The FCP rate was plotted as a function of temperature for a given !:J.K range, as
shown in the above chart. It is noteworthy that there is a minimum in the FCP rate at all
!:J.K levels except 22 MPa·M 1/2 (20 ksi-in.b"), where the data are at least suggestive of a
minimum. Because any environmental interaction is thermally activated, the crack
growth rate at a given !:J.K level and frequency may be written as:
da
dN = Aexp - Q(!:J.K)/ RT

where A is a constant and Q(!:J.K) is the apparent activation energy.

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys, "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 282
11-42. 8-816: Effect of Notches on Cycles to Failure at 900°C 313
(1650 OF)

40 ,----,r----,r--....,...-...,..--r---r---r--..---r---,...---. 280
ltJ
a..
~ ~
vi 30 210 vi
~
"- ~
"-
'Iii 'Iii
E
:J
'n. ........... E
:::J
E 0-_-0... K, = 3.4 0 E
.~ 20 A::;oo ----__ r-v 140 'x
--00"-- ltJ
~ Reversed stress fatigue ~

106
Number of cycles

SoN diagram for 8-816 heat-resisting alloy tested at 900°C

(1650 OF), notched (broken curve) versus unnotched (solid
curve).

Source: High-Temperature Fatigue, p 245

314 11-43. Udimet 700: Fatigue Crack Growth Rates at 850°C (1560 OF)

AK, ksi • in. 1/ 2

10 20 40 60 80100

10- 3

10- 2
Q)
u Q)
> u
--EE
t.l >
--.E
t.l

z· z'
"tl
'<, ~
'"
"tl 10- 4 '"
"tl

...~
Q)'
...e
Q)'

s: .t:
~ ~
e01
e
01
10-3
-"
t.l -"t.l
~ ~
t.l t.l
Q)

...
::J
01

'"
u,
.'"
Q)
::J
.01
u,

Stage I 10- 5
0

m = 16
KN Ib
10- 4
c 582 1310
v 711 1600
01067 2400
l> 1244 2800

° 160 3600
10- 6
20 40 60 80 100
AK, MPa • m 1/2

Crack growth rates in terms of stress-intensity factor range for

Udimet 700 at 850 0 C (1560 0 F). Crack growth rates for this alloy
are greatly accelerated by increases in temperature.

Source: Stephen D. Antolovich and J. E. Campbell. "Fracture Properties of Super alloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American
Society for Metals, Metals Park OH, 1982, P 285
 11-44. U-700 and Mar-M200: Comparison of Fatigue Properties 315

I
- - COLUMNAR GRAINE\i
AND SINGlE CRYSTAl
M~R.:-,~OO I I

-- .. 1-..
- - - CONVENTIONAllY CAST
MAR.M200
I I I I I I
w
C> 10-2
-- --. .~
--·-WROUGHT POIYCRYSTAlliNE
I':::--
--_.~~
UDiMEl 700
r--
-.
I
Z
-c
a:
Z
1700 0 F
- '--. I

-c
a: - 1'-.
-. -- ..
.. r---
l-
V)
..... I- I--
~ !--.
-( w- 2 ..,.--
I-
oI- - r-
-. .1---•.
1400 0 F
'.
,
103

CYCLES TO FAILURE

Comparison offatigue properties at 760 0 C (1400 0 F) and 925 0 C

(1700 OF) for a typical wrought nickel-base alIoy (U-700) with
conventionalIy cast, directionally solidified and mono crystal ,
Mar-M200.

Source: Francis L. Versnyder and M. E. Shank, "The Development of Columnar Grain and Single Crystal High Temperature
Materials Through Directional Solidification," in Source Book on Materials for Elevated-Temperature Applications, Elihu F.
Bradley, Ed., American Society for Metals, Metals Park OH, 1979, p 358
316 11-45. Waspaloy: Stress-Response Curves

0
a. 1200 STRAIN CONTROLLED
::iE
Ul 1100
c
:::l
t-
:J
a.
::iE
«
V)
V)
Ul
0:: WASPALOY
t-
V)

10 100 1000
N, CYCLES

Stress-response curves for Waspaloy having non-

shearable precipitates.

During aging of precipitation-hardenable alloys, the coherent

precipitates grow, and accommodation strains build up. At
some point the energy associated with the accommodation
strains exceeds that necessary to create a precipitate-matrix in-
terface, and the precipitates become partly incoherent. This is
accompanied by a change in precipitate-dislocation interaction
from one of shearing to that of dislocation looping or bypassing
the precipitates. Since the reasons for strain localization have
been removed, deformation becomes more homogeneous. Lo-
cal softening is thus prevented, and the cyclic-response curve
shows hardening to saturation, or to failure, as illustrated
above.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 217
 11-46. X-40: Effect of Grain Size and Temperature on Fatigue 317
Characteristics

~ %
2 } at 750 DC
....
Z -'---}
---- at 650 DC
<
~ 0.6
1Il
L!J 0.5
z 0.4
~
z< 0.3
0::.
w Small
I-
...J
-e
0.2

-....
0.1
2 5 2 5
100 1000 10.000
CYCLES TO FAILURE
S -N curves for X-40 showing effects of grain size and temperature on fatigue characteristics of this alloy.

Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Don-
achie, Jr., Ed., American Society for Metals, Metals Park OH. 1984, p 335
318 11-47. Cast Heat-Resisting Alloys: Ranking for Resistance to
Thermal Fatigue

IlJ
a::
:J
='300
it
o
F-200
~150
...J

~IOO
<l
::!:
a::
ILl
~ 50
u, 40
o
ffi 30
m
z~ 20
ILl 15
~
a::
ILl 10
~

10 15 20 30 40 50 100 200 300

CYCLES TO FAILURE OF INDIVIDUAL TESTS

The design of components that are subject to considerable temperature cycling must also
include consideration of thermal fatigue. This is particularly true ifthe temperature changes are
frequent or rapid, and nonuniform within or between casting sections. Fatigue is a condition in
which failure results from alternating load applications in shorter times, or at lower stresses,
than expected from constant-load properties. "Thermal fatigue" denotes the condition when the
stresses are primarily due to hindered expansion or contraction. Good design helps minimize the
external restraint to expansion and contraction. Rapid heating and cooling may, however, im-
pose temperature gradients within the part causing the cooler elements of the component to
restrain the hotter elements. Finite-element computer analysis has shown that, for some indus-
trial applications, these thermally induced stresses may exceed those resulting from the mechan-
icalloads.
An example of results from thermal fatigue data is presented above. This graph offers a rank-
ing of many cast heat-resistant high-alloy grades relative to their resistance to thermal fatigue.
Such rankings are indicative of general alloy properties only because most thermal fatigue tests
are based on an arbitrary set of experimental conditions rather than on their fundamental mate-
rial behavior. Nevertheless, such test results have been useful in considering alloy selection ques-
tions, and in identifying the superior thermal fatigue resistance of nickel predominating grades
and the good performance of some HH type compositions.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 19-7
 12-1. Corrosion-Fatigue Properties of Aluminum Alloys Compared 319
With Those of Other Alloys

500,.------- -.,
typical corrosion fatigue strength 70
(sea water , N • 108 cycles I R· -1 )
ambient temperature

t 400
60
';:;'"""'
-€.
z t
duplex staness steels,
~ titanium alloys
,.....,
50 'iii
(e.g. n -6A1-4V) .><
L-J

0 300

40
0
J:.
0.
c:
e
iii

s 30
t
iii
200
-F nickel alloys .!j
~ .S!'

r.. .·-/ .
(e.g. aJloys 600 and fn»
~
.~
.~
~ 20
~ ferritic _ stainless e
0
100 martensilic_ \ steels 8
copper - nickel alloys
1)
,carbon steels,
I bw alloy
aluminum
m 0

As shown above the corrosion-fatigue strength of bare alumi-

num alloys is superior only to that of magnesium alloys. Careful
surface protection may bring the corrosion-fatigue strength up
into the range of bare stainless steels or copper-nickel alloys.

Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material," in Aluminum Transformation Technology and Appli-
cations (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, R. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OR, 1980, P 617
320 12-2. Comparisons of Aluminum Alloys With Magnesium and Steel:
Tensile Strength vs Endurance limit

250 6.

()~ /
':>.o~'b /
200 "3 / 6.

r0- 6./

.'= ~
ll..
/
E::::: 150
:.= (J)
0)
0)-
u u
c i:)
moo 100
~O
"O~
C )( x Aged aluminium alloys x
UJ~ Non-heat treatable
0
50 aluminium alloys
Magnesium alloys
Steels
•
6.
O~---JL...-_----I. _ _---L_ _-'-_ _....I..-_ _.L......_---l

100 200 300 400 500 600 700

Tensile strength (MPa)

Fatigue ratios (endurance limit/tensile strength)

for aluminum alloys compared with those of mag-
nesium and steel.

It is well known that, in contrast to steels, the increases that

have been achieved in the tensile strength of most nonferrous
alloys have not been accompanied by proportionate improve-
ments in fatigue properties. This feature is illustrated in the
graph above, which shows relationships between fatigue endur-
ance limit (S X 108 cycles) and tensile strength for different al-
loys. It should also be noted that the fatigue ratios are lowest for
age-hardened aluminum alloys and, as a general rule, the more
an alloy is dependent upon precipitation hardening for its total
strength, the lower this ratio becomes.

Source: I. J. Polmear, Light Alloys, Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981, P 39
 12-3. Aluminum Alloys (General): Yield Strength vs Fatigue Strength 321

• ultimate tensile strength

o yield strength
500
o fatigue strength in air , N - 5 X 101
1
• latiguestrength in sea water I N _10
70
• fatigue strength 8
•
in river water I N - 10

R - -1 , .. 60 Hertz I
stnooth specimens I
• 60
4()() ambient temperature

+
50
';:;"'"'
+
6
--z
E
.--.
'iii
""
300
• '--'

0 • corrosion fatigue strength

40
0
01 aluminum alloys
s: s:
e;, e;,
c 30 c
l!! 200 l!!
OJ
• 000
OJ

o 0CF,air , N-5x1Q1 20
-Orr-----l:I
100

000
I I I ~ ~~ ~I
<0
~
I ~
,
~
I
...
!2 <0, <0 <0
I
~

~§ ~
I II
~
Cl)
Ie ~ 'Ie
~ ~~
I
f2, iij R ;::
R
~
An analogous conclusion can be drawn from a review of corrosion fatigue
tests with smooth aluminum alloy specimens as shown in 'the above graph.
Here aluminum alloys are listed in order of increasing yield strength. As the
yield strength goes up, so does the ultimate tensile strength, but the fatigue
strength in air soon reaches a limit which is roughly the same for alloys of
greatly different yield strength. In other words, medium- and high-strength
aluminum alloys all have about the same fatigue strength. The above graph
shows that the same is true for the corrosion-fatigue strength: there is as yet
not a single commercial aluminum alloy available with a high-cycle
corrosion-fatigue strength significantly higher than all the other aluminum
alloys. Thus, corrosion fatigue is still a limiting factor for the application of
aluminum alloys.

Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material," in Aluminum Transformation Technology and Appli-
cations (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 616
322 12-4. Comparison of Aluminum Alloy Grades for Crack Propagation
Rate

1O- 3 r--- - - - - - - - - - - - - - - - - - - ---,

7050 - T 736
7175 - T 736
2219 - T 851
7079 - T 651
2618- T 6

~ 5
Q) 10
~
~
E
'----'
10-6
Z
~
<I
10- 7

8
10
~ typical scatter in experiments

crack orientation L - T
10- 9 specimens CNP, DCB I CT
I

ambient temperature
environment air I

R =0,0-0,1 0,1 -30 Hz

I

1).-.... ---+--------------------4
10 20 40 50 60
° cyclic stress intensity range I .t:.K I [ MN· m- 3f2]

Many commercial aluminum alloys show similar fatigue

crack propagation rates in air, as indicated in the above
comparison.

Source: Markus O. Speidel, "Aluminum asa Corrosion Resistant Material, "in Aluminum Transformation Technology and Appli-
cations (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloniand D. E.
Embury, Eds.. American Society for Metals. Metals Park OH, 1980, p 613
 12-5. Alloy 1100: Relationship of Fatigue Cycles and Hardness for 323
HO and H14 Tempers

50r----.,---.------.,---.-------r--r-----r--r-----r--r--,---r--,---r----,

•
~ 30
LU
-----------
Z
o
cz::
<:
:I:
~ 20
o
z
><::

10 • AI 1100 HO
• AI 1100 H14

Comparison of the Knoop hardness for well-annealed (HO) and cold-

rolled (H14) aluminum as a function of fatigue.

Microcrack initiation is often triggered by a dislocation rearrangement.

For instance, in the case of well-annealed Al 1100 (RO), the material will
harden in the early stages of fatigue (see S-N curves above) as the dislocation
density in the bulk of the material increases, accompanied by pronounced
slip-step formation on the surface. On the other hand, in the case of the
cold-worked material Al 1100 (R 14), the material will soften in the early
stages offatigue (above curves) as the dislocation density, introduced by the
cold work, decreases. Slip-step formation in this situation is much less pro-
nounced than it is during hardening, because the initial dislocation-loop
length is much shorter. In either case, during this initial rearrangement, the
dislocations form a cell structure with individual dislocations of long loop
length shuttling to and fro between the cell walls. This latter part of the fa-
tigue life is called the saturation stage offatigue, during which the dislocation
shuttling leads to local instabilities, or "extrusions-intrusions," and finally to
microcracks, which can be observed after about 25% of the fatigue life has
been expended. The micro crack density is about the same for both materials.

Source: O. Buck and G. A. A1ers,"New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstruc-
ture, American Society for Metals, Metals Park OH, 1979, P 128
324 12-6. Alloy 1100: Interrelationship of Fatigue Cycles, Acoustic
Harmonic Generation and Hardness

8xlf3 3
16xlf 80

...... A
~
4 zlA I
Cla 60
-e
Vl

0 ......
.....
Vl
z
~ 0
40 oe:
-c'" «
::J:
ll.
K.H. 0
0
z~
4 ZO
AI 1100

0 0
0 80 120
160xllY
FATIGUE CYCLES

Normalized second harmonic displacement and

Knoop hardness as a function of fatigue.

The effects of dislocation rearrangements on harmonic gen-

eration within the bulk of the material during fatigue are shown
in the above chart. Using 3D-MHz longitudinal waves, the nor-
malized second harmonic amplitude of an initially compression-
deformed Al 1100 single crystal was monitored and found to
increase as a function of compression-compression fatigue. At
the same time, the surface hardness (Knoop) decreased. Appar-
ently, the dislocation-loop length prior to fatigue was quite
short, since the initial amplitude of harmonic generation was
small. During fatigue softening, the cell structure that developed
(with its individual dislocations within the cells) became quite
large, so that a change of the dislocation-induced harmonic gen-
eration, A 2d , increased. Application of this technique to high-
strength aluminum alloys failed, however, apparently because
of an immediate rep inning of the long loops by interstitials in
this alloy.

Source: O. Buck and G. A. Alers, "New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstruc-
ture, American Society for Metals, Metals Park OH, 1979, P 131
 12-7. Alloy 2014-T6: Notched vs Unnotched Specimens; Effect on 325
Cycles to Failure

50

~ 40
)(

N.
c:
~ 30
VI
....
VI
cr:
:n 20

10

Effect of notch on fatigue of 2014-T6. As is true for most

alloys, notches greatly reduce the fatigue properties of
aluminum alloys.

Source: P. C. Varley, The Technology of Aluminium and Its Alloys, Butterworth & Co. Ltd., London, England, 1970, p 43
326 12-8. Alloy 2024-T3: Effect of Air vs Vacuum Environments on
Cycles to Failure

~ 35
..
40
,
~
<,
.- Ultrahigh Vacuum
0- Air
- 3.5 x 10
-3

Ii
Q)

J; 30
Cl
c
'" .........
..........

<,
~,
<,
r-
•
-3
c
.
"0 ......... '~ - 2.5 x 10 "~
; 25
III r""- ... ...
l/)

~~
E
::J
,§ 20
I -
)(

~ En'durance Limit J
I I I I III"' - 1.5 x 10- 3
15 s 7
10 2 4 6 8 10' 2 4 6 8 10 2 3
Number of Cycles-to-Failure

The effects of air versus vacuum on the fatigue life

of a 2024-T3 aluminum alloy.

For most materials, environment appears to be most effective

early in the crack-growth process, with little or no effect at high
crack-growth rates. Additionally, the majority of S- N curves di-
verge at decreasing stresses, the increase in fatigue life caused by
vacuum becoming greater at lower stresses. In contrast to this
behavior, however, aluminum and aluminum alloys have been
shown to exhibit conflicting results. For example, a 2017-T4
alloy tested in air and at 2 X 10-6 torr and a 2024-T3 alloy tested
in air and at 10-10 torr in rotating bending exhibit convergence of
S- N curves at low stresses, the effect of environment apparently
becoming less important at decreasing stresses, as shown in the
above chart.

Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 337
 12-9. Alloy 2024-T4 Alclad Sheet: Effect of Bending on Cycles to 327
Failure

400

350 50
300
rf. 40 'w
:E ...:
VI'
VI
250 ~.

...e
(/)
30
e
en
200

150
Stress ratio. 0.1 -----2 1
....::.4"-3
20
100
0.01 0.1 10
Millions of cycles to failure

Effects of bending on fatigue characteristics of

aluminum alloy sheet.

For the data here, sheet 1.02 mm (0.040 in.) thick was an-
nealed, solution heat treated and quenched, and then fatigue
tested. The sheet represented by curve I was not bent. All other
sheet was bent 90° in the annealed condition. Flattening (un-
bending) was done in either the annealed condition (curve 2) or
the solution heat treated and quenched condition (curves 3, 4
and 5). Details of bending and flattening were as follows: (I) Not
bent. (2) Bend radius, 3.18 mm (Ys in.); flattened in annealed
condition. (3) Bend radius, 3.18 mm (Ys in.); flattened in
quenched condition after 3 days of storage at-18 to-12°C (0 to
IOOF). (4) Bend radius, 3.18 mm (Ys in.); flattened in quenched
condition after 14 days of storage at -18 to -12°C (0 to 10 OF).
(5) Bend radius, 1.59 mm (1/ 16 in.); flattened in quenched con-
dition after 3 days of storage at - 18 to - I2 ° C (0 to IO° F).

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 35
328 12-10. Alloy 2024-T4: High-Cyclevs Low-Cycle Fatigue

40
~ - LOW CYCLE t R -0 2 C. RAOIAJlON I
III
-_. HIGH CYCLE (R. - 1 Mo RADIATION I
A
'5 I PANGBORN e I ill I
/,
C 30 /,
] /
/

C::P /
% /
/
5 20 /

~ /
/

~ /
/
Q
w 10 /

~a:
/
/
a:
8 0
0 0.5 1.0

FRACTION OF LIFE In/n,1

Dependence of 13 on n/nior low-cycle fatigue and

bulk properties of high-cycle fatigue of AI 2024.

After correcting for the difference in initial /30 values, it can be

seen in the above diagram that the two fatigue processes, al-
though radically different in strain history, exhibit similar be-
havior throughout most of the fatigue life.

Source: Sigmund Weissmann and William E. Mayo, "Determination of Strain Distributions and Failure Prediction by Novel X-ray
Methods," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck and Stanley M. Wolf, Eds., American
Society for Metals, Metals Park OB, 1984, p 195
 12-11. Alloy 2024-T4: Relationship of Stress and Fatigue Cycles 329

32

30

28
u
~ 26
0
.....0
VI
24
-;'"
c:: 22
E
I~ 20
..,.z
sc; 18
.c::
16
e'"
0'

«'" ,,
> 14

12 :

5,000 10,000 15,000 20,000

Number of fatigue cycles, N

Dependence of 13 on number of cyclesN at various

stress levels of AI 2024-T4.

Here is shown that for the maximum stress of241 MPa with R
= 0.1, the 13 value increased during the first several hundred cy-
cles. This was more pronounced for the surface grains (Cr KO'I
radiation).

Source: Sigmund Weissmann and William E. Mayo, "Determination of Strain Distributions and Failure Prediction by Novel X-ray
Methods," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck and Stanley M. Wolf, Eds., American
Society for Metals, Metals Park OB, 1984, P 194
330 12-12. Alloy 2024-T4: Dependence of the Average Rocking Curve
Halfwidth 13 on Distance From the Surface

20..--------------------,

19

18 ~-_:_------,O----..."....----_1
o

ICQ.
o b
14

13 _ _ _ _ _ ~o (inillal ~alf.i~_~_) _

12
IIL..-_ _.......... .L--_ _- - ' - ....l...- _ _---J

o 50 100 150 200 250

I'm

Dependence of the average rocking curve half-

width 11 on depth distance from surface for differ-
ent fractions of corrosion fatigue lives, N F' of Al
2024-T4.

X-ray rocking curve measurements were carried out as a func-

tion of depth distance from the surface, and typical results ofthe
dependence of 13 on depth distance for an alloy cycled with a =
276 MPa, corresponding to the static yield stress, are shown
above. It may be seen that the minimum 13 values at the surface
layers were larger than those in the interior. The 13 values de-
clined up to a depth of about 50 Mm from the surface and subse-
quently retained a plateau value throughout the interior of the
specimen for each fraction of the life.

Source: Sigmund Weissmann and William E. Mayo, "Determination of Strain Distributions and Failure Prediction by Novel X-ray
Methods." in Nondestructive Evaluation: Application to Materials Processing, Otto Buck and Stanley M. Wolf. Eds., American
Society for Metals, Metals Park OH, 1984, P 193
 12-13. Alloys 2024 and X2024: Effect of Alloy Purity on Cycles to 331
Failure

260

tV
a... 0
:2:
CIl
CIl
220
~
CIl

Ol
C
.~
C

~ 180 +
<i:

140 ~ ---'- o ---'----'

105 106
Cycles to failure

Effect of reducing the concentration of submicron

particles in an AI-Cu-Mg alloy. X2024 is a high-
purity version of the commercial alloy 2024.

The disappointing fatigue properties of age-hardened alumi-

num alloys are also attributed to an additional factor, which is
the metastable nature of the metallurgical structure under con-
ditions of cyclic stressing. Localization of strain is particularly
harmful because the precipitate may be removed from certain
slip bands, which causes softening there and leads to a further
concentration of stress, so that the whole process of cracking is
accelerated.
The fatigue behavior of age-hardened aluminum alloys
would therefore be improved if fatigue deformation could be
dispersed more uniformly. Factors which prevent the formation
of coarse slip bands should assist in this regard. Thus it is to be
expected that commercial-purity alloys should perform better
than equivalent high-purity compositions because the presence
of inclusions and intermetallic compounds would tend to dis-
perse slip. This effect has been confirmed for an Al-Cu-Mg al-
loy, and fatigue curves for commercial-purity and high-purity
compositions are shown in the above S-N diagram. Here the
superior fatigue behavior of the former alloy arises because slip
is more uniformly dispersed by submicron dispersoids such as
MnAI 6 •

Source: I. J. Po 1mear, Light Alloys, Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981, P 40
332 12-14. Alloys 2024 and 2124: Relationship of Particle Size and
Fatigue Characteristics

Ii K, ksi • in. 1/ 2

3 5 8 10 15 20
10- 3
10- 2
Q) Q)
u u
> 10-4 >
--
E
E
u

10- 3
--.5
e

z·

t
Z
--'"
'C

'C 10-5 --'"

'C

'C

..... ....
Q)"
'C

1......
Q)"
10- 4 .~
'"
...'"
s: ...
.s: Q)
e,

..
Q)
~
.
UI
~
0 10-6 0
en '"
'0.
o
.. -~
en .:.: 'u
.:.:
'".
u
e
10-5
.'"
o
u
Q)
c,

1
Q)
Q) :l
:l
10- 7 en
en '~
'';:::;
'"
U. 10-6 U.

"---'~_ _"""""_ _"""""---''--_''''''''''_'''''''''---' 10- 8

20
Stress-intensity factor range, LiK, MPa • m 1/2

Comparison of typical particle sizes in aluminum alloys

with crack advance per cycle on fatigue loading.

The above graph represents Staley's work in summarizing the

role of particle size on fatigue crack growth in aluminum alloys,

Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys." in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OK 1982. P 191
 12-15. Alloys 2024-T4 and 2124-T4: Comparison of Resistance to 333
Fatigue Crack Initiation

Mechanically Polished R=-I

700
2.8 2024 T4 oNjj
600
2124 T4 ANjj
~
500
E
<, Nij =Cycles to ~ 15 fLm Crock
0
z 2.6 0
~
400
~ A N

fJl
fJl
~
Z
...
Ql A
~ 300
in 0
fJl
Ql 2.4 AA
fJl
fJl
0 ...
Ql

co
..r:::
in
0 CIl 200
fJl
"0 0
z 2.2
co
0>
0
A Ao ..r:::
0
...J 0 AA "0
z

2.0
102 103 105 106 107

Cycles to fatigue crack initiation for specimens of alumi-

num alloys 2024-T4 and 2124-T4 versus stress at notch
base (computed using Neuber stress-concentration fac-
tor).

The 2124 alloy studied had 1/ 10 the inclusions of the 2024

alloy studied (0.2 vol% compared to 2 vol%) but a larger grain
size (45 /-lm compared to 20 /-lm in the transverse direction nor-
mal to the loading direction). With 2124-T4, slip-band cracks
not associated with inclusions formed at the lowest stress stu-
died. They also formed more easily in 2124-T4 than in 2024-T4
at high stresses, in keeping with the larger grain size. Thus, as
shown in the above chart, at high stresses 2124-T4 is less resis-
tant to fatigue crack initiation than is 2024-T4, but it is more
resistant at low stresses.

Source: M. E. Fine and R. O. Ritchie, "Fatigue-Crack Initiation and Near-Threshold Crack Growth," in Fatigue and Microstruc-
ture, American Society for Metals, Metals Park OH, 1979, P 251
334 12-16. Alloys 2024-T3 and 7075-T6: Summary of Fatigue Crack
Growth Rates

t.K, ksi . in.1/2

10 100
10- 1 r - - - - - - - , - - - - - -....- - - - - - - . ,
7075-T6
9 investigations

10- 4
2024-T3 G>
8 investigations U
>-
u
'<,

10- 5 .~
iCl
10-
4 Note:
Bounds defined
z
"C
......
01
"C
~ by mean curves
u
E
u
of separate 10- 6
~ 10- 5 investigations
Cl
.~
U.
10-7
10- 6

10 100
Stress-Intensity factor range, AK, MPa • m 1/2

Summary of fatigue crack growth rate data for

aluminum alloys 7075-T6 and 2024-T3.

Considerable use has been made of the fracture mechanics

approach in measurement of fatigue crack growth rates in alum-
inum alloys. These data have been generated by methods com-
parable to those of ASTM Method E647 for measuring fatigue
crack growth rates. In general, fatigue crack growth rates are
found to fall within a relatively narrow scatter band, with only
small systematic effects of composition, fabricating practice or
strength level, as illustrated by the data in the above chart.

Source: J. G. Kaufman and J. S. Santner, "Fracture Properties. of Aluminum Alloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerherich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982,P 189
 12-17. Alloys 2024-T4 and 7075-T6: Effect of Product Form and 335
Notches

_ - -----+---j--+_
I
2024-T4 ---+----1
414

'"
e,
::E
345
~---l·~-~~
- --
."
~.,-::'~---+---+---j
276
rn
UI
w
I I,..... 1'I,--'~-r-''"'
~ ..
". - ~.....;-"'~_ _+-_--j
207 I "
a:
l-
rJ)
'"'-I' ,·t'i:::::~"
J -, <,
~
'q".
<, :.:~,.;-.~
".. c ~--•• - ~.
138

69

0
I
414 I---L.--+~-t---+- 7075-16 ~----I

I", -,,
345 I----i---"~~.-:-l.c---+----jf----+----j
\ ~'
'"
c,
::E
rn
rJ)
276
.\ . f~,~~.l,.
~
w 207 1---+---''':-....' \ • ''':".":-- --+ ,,-- - - -

.~":''-1''''.1
• t': ':', .... ~'~r, ::""II'jl
a:
l-
" "

"e'::;"',"-j--""'-f'--..
UI
~. ,', " .,'\ ~,.
138 ~c-'-cc-:--''''.j,-,''-...,.;t-,§:--,:,r-'- ,-
~.," 1'" ---~..!...:'.I,
:• ~gmg ~~~TE
,"00.
69 :• FORGINGS
....::.:,
~ '---->":" -. --}.!
, • EXTRUSIONS • -.#_.s "-,.J.'
o - 010 NOT FAil
----L . _ _ . __ _ _
I.
1~ 1~ 1~ 1~ 1~ 10' 108 109
CYCLES

Fatigue performance of smooth and notched (K, > 17)

rotating-beam specimens from various product forms of
2024-T4 and 7075-T6 aluminum alloys.

Numerous methods have been developed to evaluate response of materials to cyclic deformation.
The earliest method was by use of S-Nplots. Typical examples are depicted above. Basic specimens
include rotating beam, axially stressed and sheet flexure. Notches have been employed to provide stress
concentration, and special specimens have been used to simulate a variety of other conditions. The S- N
response is strongly influenced by a number of conditions, including surface condition, stress ratio, and
environment. The various alloys differ widely in their response to fatigue testing-specifically, in the
number of cycles where a "level out" condition is attained. As shown in the above S- N diagrams, the
SoN response for aluminum alloys tends to level out as the number of applied cycles approaches 500
million.
Based on SoN data of smooth and sharply notched specimens and of similar tests of specimens
designed to simulate joints in structures, the following conclusions have been drawn. From fatigue
results for aluminum alloys obtained with smooth specimens . . . rather wide variations can exist
without causing appreciable differences in fatigue strengths. . . . When severely notched specimens
are used, the effects of composition and temper are even less pronounced and generally are of no
practical significance. . . . As in the case of simple notch fatigue tests, there is a lack of significant
differences in the fatigue strength of the joints of the various alloys.
Despite these laboratory data, users discovered that certain aluminum alloys performed decidedly
better than others in service when fluctuating loads were encountered. For example, airframe manu-
facturers determined that fatigue performance of alloy 7075-T6 was unquestionably inferior to that of
alloy 2024-T3.

Source: T. H. Sanders, Jr.• and J. T. Staley. "Review of Fatigue and Fracture Research on High-Strength Aluminum Alloys." in
Fatigue and Microstructure. American Society for Metals, Metals Park OH, 1979. P 470
336 12-18. Alloys 2024-T351 and 7075-T73XXX: Comparison of P1M
Extrusions and Rod

500 r------,r------,-----,.---~---_,_---__r---__, 70
1.35 mm (0.053 in.) 6.43 mm (0.253 in.)

400 1-_ _ +-__-+-_ _ ---l_+_~=1B

--F
f
Notch tip radius:
-+-_ _--+ t -_ _-1 60

<0 e = 0.33 mm 10.013 in.)

0..
K, = 3
50
::E Stress ratio: R = 0
vi Ambient air
III
~ 300 f----I------1I-------1I-----j----j-----t-----l
III ~
E E
:J
:J
E E
'x 'x
<0
<0 M ::E
::E 200 1 - - - - / - - - + - I f - 1 . * - - - I - - - - - t - - - - - t - - - - - + - - - - - - l

20

100 f -_ _ 2024-T351 -:::~~~~~"'F;;;;;:;:;;:::::j::~-__jl--I

rod and bar band -
~"t----I-~ 7075·T73XXX 10
products band
oL -_ _-.l. ....L ...L. L -_ _---' ....L_ _~ 0
10' 103 10' 105 10' 10' 10' 10'
Cycles

Comparison of room-temperature axial stress notch fatigue strength ofP 1Malloy

extrusions and ingot metallurgy alloy rod, bar and products, 0 ,X7090·T7E71 in
the longitudinal direction; • , X7091-T7E69 in the longitudinal direction; 6.,
X7091-T7E69 in the long transverse direction; - denotes test specimen did not fail,

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,p 468
 12-19. Alloy 2048-T851: Longitudinal vs Transverse for Axial Fatigue 337

500
70
'"
Q..
60 ""'"
:i: 400 ::I"
I;.
::I"
50 ~
~
'" 300
E •
• '"
40 E
:::J
:::J E
E R = 0.1 'x
'x 200 30
'"
:i:
'"
:i:
o 0 I;. Longitudinal
I
• • ... Longtransverse I;. 20
100
103 10 4 105 106 107
Numberof stress cycles

SoN axial fatigue curves for unnotched specimens of

aluminum alloy 2048·T851 plate, showing effects of R
value and direction upon fatigue properties.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals. American Society
for Metals, Metals Park OH, 1979, P 80
338 12-20. Alloy 2048-T851: Notched vs Unnotched Specimens at
Room and Elevated Temperatures

500
10

400 ~
Unnotched
s: R ~ 0.1 60 ]
:::E

~
Ii: 300
r-, 24°C 176of) 50 g
E
E 176°C 13600F)~_ /120oC 1260 of) 40 ~
~ T -=-
.~
:::E
200 30 !
20

101
Number of stress cycles

...:::E.. 400

g ~
Notched (K t '" 3.01
R ~ 0.1 - 50 ]
300
~ s, 24°C (76 of)
- 40 g
E
~ 200 30 §
.~ ~J 120 °c 1260 of) _ 20 E
:::E 100 "i
176°C 1360 OF) --;;;;1 :::E

Number of stress cycles

S-N curves for unnotched (upper graph) versus

notched (lower graph) specimens of aluminum al-
loy 2048-T851 plate.

Source: Metals Handbook. 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals. Metals Park OH, 1979, P 82
 12-21. Alloy 2048-T851 : Fatigue Crack Propagation Rates in LT and 339
TL Orientations

1 1
AK. kSi'in. 2 AK. kSI'in.2
10 10

LT crack oriant!tion
TL crack orianLtion

l~3.4
1 1
{.:
rot"
,
/"
10- 5

10- 6
!
G

.5
~.
..:!
-1i
~

~
E

..:!
10-5 G
"iI
~
.5

~
..:!
0 10-6

10-8
f
M~7.1
M-a.:
61 1 61 1

10 10

Fatigue crack propagation in aluminum alloy 2048-T851 plate, showing

propagation data for both LT and TL (longitudinal and transverse)
crack orientations.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, p 81
340 12-22. Alloy 2048-T851: Modified Goodman Diagram for Axial
Fatigue

Minimum stress,ksi

s:
::;: 60
~
~
1;;
E
~E
"E "
..
'x
::;:
200 E
.~

::;:

Minimum stress, MPa

Modified Goodman diagram for axial fatigue of un notched specimens

of aluminum alloy 2048-T851 plate.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 81
 12-23. Alloy 2219-T851: Dependence of Relaxation Behavior on the 341
Cyclic Hardening Parameter

O~-----r-------,------r-------,r-----...,
o o o
0% RH

•
50% RH

-300 L-- ....J.. ...!- .l..- ---I. ...J

o 10 20 30 40 50

CYCLES (x 10-3)

The dependence ofrelaxation behavior on the cyclic hardening parame-

ter, (J. (J was varied by changing the relative humidity (RH), which af-
fects the near surface ductility in this alloy. Values used were: (J= 6 X 10.5
for 50% RH and (J= 2 X 10-5 for 0% RH. The cyclic stress amplitude was
0.88 a yleld for both samples.

Source: M. R. James and W. L. Morris. "The Relaxation of Machining Stresses in Aluminum Alloys During Fatigue." in Residual
Stress for Designers and Metallurgists. Larry J. Vander Walle. Ed.. American Society for Metals. Metals Park OH. 1981. P 184
342 12-24. Alloy 2219-T851: Effect of Strain Amplitude on the
Relaxation of Residual Surface Stress With Fatigue

Or-------.-------,r------.------,----...,
0.7" YIELD

o
0.64 a YIELD

-300
o~-----'--------l----...J-.-------l---- .....
10 20 30 40 50
CYCLES (x 10- 3)

The effect of strain amplitude on the relaxation of surface residual stress

with fatigue. The symbols are the residual stress value measured by the
x-ray diffraction peak shift technique. The solid curves are the predicted
mean residual stress values during fatigue.

Surface milling produced the shallowest stress gradient and resulted in the
slowest rate of relaxation of the surface stresses. A comparison of measured
to predicted values of residual stress during fatigue is made for four "as ma-
chined" specimens in the above chart. The residual stress values were mea-
sured parallel to the external stress axis. A value of f3 = 0.0004 was used to fit
the data for all specimens. Residual stress measurements were also made in a
direction transverse to the applied stress axis. Within experimental error, the
cyclic relaxation rate was the same as in the longitudinal direction.

Source: M. R. James and W. L. Morris, "The Relaxation of Machining Stresses in Aluminum Alloys During Fatigue," in Residual
Stress for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 182
 12-25. Alloy 2219-T851 : Relationship of Fatigue Cycles to Different 343
Depth Distributions of .Surface Stress

O....------r-----r-----,-----r-----,
o
o

• •

-300 l-- --L -L- ..L....- ----I .....

o 10 20 30 40 50
CYCLES (x 10-3)

The relaxation behavior oftwo samples having different depth distribu-

tions of residual stress. Note the difference in the peak cyclic stress, a.
0= rolling (10%reduction); f3 = 0.012; a= 0.91 ayield'. = sand blasting;
f3 = 0.003; a = 0.71 ayield'

Relaxation of a compressive surface stress requires an expansion of the

material normal to the surface. Of necessity, this involves slip at an acute
angle to the surface. If the slip does not penetrate the surface, the residual
stress cannot relax. Supporting this picture are our observations that the
relaxation rate in Al 2219-T851 is more rapid in dry air. It is known that
humidity increases the rate of cyclic hardening of a thin (less than I Jlm) layer
at the surface. The effect of humidity on relaxation is therefore simply to
make it more difficult for dislocations to penetrate to the surface.

Source: M. R. James and W. L. Morris. "The Relaxation of Machining Stresses in Aluminum Alloys During Fatigue," in Residual
Stress for Designers and Metallurgists. Larry J. Vander Walle. Ed.• American Society for Metals, Metals Park OH. 1981, P 183
344 12-26. Alloy 2219-1851: Probability of Fatigue Failure

LOG (NUMBER OF FATIGUE CYCLES)

Schematic curves of constant probability for failure (ac-

tual failure = 100%).

The solid line in the graph represents failure; the dashed lines
indicate the percentage offatigue life expended. The exact loca-
tion of these lines is highly sensitive to the material and its
microstructure as well as the influences of environment.

Source: O. Buck and G. A. Alers, "New Techniques [or Detection and Monitoring of Fatigue Damage.t' in Fatigue and Microstruc-
ture. American Society for Metals. Metals Park OH, 1979. p 104
 12-27. Alloys 3003-0, 5154-H34 and 6061-T6: Effect of Alloy on 345
Fatigue Characteristics of Weldments

300

40

260

30
If. 200 :t!
::;;

g 160
g
E E
20
"
E
..
'x
::;; 100
E
.."
'x
::;;

10
50

Number of cvcles

The fatigue life of welded joints at high loads varies with the
alloy. As the load is decreased, differences disappear until, at
about one to ten million cycles of axial loading (R = 0), the
fatigue strength of an arc-welded joint is approximately the
same regardless of alloy and is 50 to 70% that of the unwelded
alloy. Typical data are given in the above graph for three alumi-
num alloys. Specimens were from 9.5-mm (Ys-in.) plate; weld
reinforcement removed; axial loading; R = O.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 195
346 12-28. Alloy 5083-0 Plate: Effect of Orientation on Fatigue Crack
Growth Rates

.1K, ksi • in. 112

5

10- 4

10- 3
..
u u
.
~ >
-. -.t.l
E
10-5
.E
E
Z
Z 'C
-.
'C
-.
Cll
Cll
'C
'C 10- 4

10- 6
T-S
10-5 Compact specimen thickness
= 46 mm (1.8 in.)
R = 1/3, f = 13 Hz
Room temperature, dry air

.1K, MPa . m 1/2

Effect of orientation on fatigue crack growth rates

in 180- and 196-mm (7.0- and 7.7-in.) 5083·0
plate.

From the data shown above there is obviously no great effect

of specimen orientation on fatigue crack growth rates.

Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 193
 12-29. Alloy 5083-0 Plate: Effect of Temperature and Humidity on 347
Fatigue Crack Growth Rates

~K, ksi • in. 1/2

10- 2

,•
I
,/ 10- 4
RT, moist air~i
10- 3 ,,
Ql ,,,I Ql
U
~ ,
,
U
~
EE 10- 5 .E
z· z·
~
~ III
"C
"C 10-4

10-6

Compact specimen thickness

10-5 = 46 mm (1.8 in.)
R = 1/3.
f=13&18Hz.
T-L orientation
'-- ...L-_--L. '--~ 10- 7
5 10 50
~K. MPa • m1/2

Effect of temperature and humidity on fatigue

crack growth in 180-mm (7.0-in.) 5083-0 plate.

As shown in the above graph, growth rates for alloy 5083-0

are appreciably higher in moist air than in dry air. Growth rates
in water solutions of sodium chloride are similar to those in
moist air.

Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William w. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 195
348 12-30. Alloys 5086-H34, 5086-H36, 6061-T6, 7075-T73 and
2024-T3: Comparative Resistance to Axial-Stress Fatigue

0.7
R~O

Frequencv = 1.1 kHz

0.6

0.5

U
. 0.4
~
....
~
.
~

.~ 0.3
a:

0.2

0.1

5086- 5086- 6061- 7075- 2024-

H34 H36 T6 T73 T3

Ratio of axial-stress fatigue strength of aluminum alloy

sheet in 3% NaCI solution to that in air.

Fatigue strengths of aluminum alloys are lower in corrosive

environments such as seawater and other salt waters than they
are in air, especially when evaluated by low-stress, long-period
tests. As shown in the above bar chart, such corrosive environ-
ments produce smaller reductions in fatigue strength in alloys of
the more corrosion-resistant types, such as 5xxx and 6xxx al-
loys, than in less resistant alloys, such as those of the 2xxx and
7xxx series. Like stress-corrosion cracking of aluminum alloys,
corrosion fatigue requires the presence of water. In contrast to
stress-corrosion cracking, corrosion fatigue is not appreciably
affected by test direction, because fracture resulting from this
type of attack is predominantly trans granular.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 220
 12-31. Alloys 5083-0/5183: Fatigue Life Predictions and 349
Experimental Data Results for Double V-Butt Welds

IOOr---r-,--...--I'""T'"T'TTT"--,----.-r....T"T"TT,----r-,--...--I'""T'"TT"n
~ ~

60 400

40 300

200

.: 10 CT,=+18kSi
1II 8
<I
6
30
4 KI~
20
5083 -0/5183 Double - V Butt Welds s~s
2
K
' mol =2.60, 0ls0.0Iln., R =0, 1= 3/8 .n.
CT, • + 18 ksi 10
I 4
10 10~

NT' Cycles

Total fatigue life predictions and experimental results for

5083-0/5183 3/8-in. (10-mm) butt welds.

Source: F. V. Lawrence, "The Predicted Influence of Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 114
350 12-32. Alloys 5083-0/5183: Predicted Effect of Stress Relief and
Stress Ratio on Fatigue Life of Butt Welds

1OO~"""""""""""'--''''''''''''''''''''''~-~''''''''''''''''''''''''''''''"T''T"r--''''''---'---'''''''''''''''''''''''''600
80 ~3-o1!l183 Bull Weld
eo Kt..." 3.HI.
°1" 0.0111\ (0.2~41Ml) 400
."90",' "SO", '" 112ir\ 112 7mm) 300
40
- - .,.5, 200
--- .,·0
R"O
j

~-.
30

20

2
S~S
1(,... '
10

Predicted effect of stress relief and stress ratio on

5083-0/5183 butt weld fatigue life.

Because of the high notch-root plasticity during the first few

cycles, before the material cyclically hardens, the aluminum
weld considered here (5083/5183) exhibits little dependence
upon either residual stress or stress ratio, even though the relax-
ation of the stabilized mean stress (uos) is very slow-as indi-
cated in the above chart.

Source: F. V. Lawrence, "The Predicted Influence or Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 113
 12-33. 7XXX Alloys: Cyclic Strain vs Crack Initiation Life 351

100

.fIN UTS TYS ./,

W AllOY kSi (MPa) ks; (MPa) fl. 2"
0
:::l 7075·T6 810(558) 769(530) 95
f- 7050·T6 887(612) 82 7(570) 12.0
::J
Q.. 7075·T7 73.2(505) 658(454) 11.0
::iE 7050-T7 74.1(511) 63.8(440) 14.0
<>:
z
<i:
a:
f-
en
...J
10- 2 /7050.T6
<>:
:-~--
~ ----.
-------
f-
0
f-
--"""""''''''''""""",""
"""'''''''''''''1;;./11
1O-3 ':'":-_ _---'- -'- -':-_ _----J'-_ _--'- -'-_ _............
100 10 7

REVERSALS TO INITIATION, 2N

Cyclic strain versus initiation life for laboratory-fabri-

cated high-strength 7XXX aluminum alloys,

Plots of elastic-strain amplitude versus life have seen rela-

tively little use for commercial alloys, but plots oflog total strain
amplitude versus life have been used more frequently to com-
pare materials. This approach offers the advantage that both
high- and low-strain fatigue may be characterized with one plot.
As illustrated above, fatigue resistance at low total strain ampli-
tude is governed by the elastic-strain amplitude. Fatigue lives
for total strain amplitudes less than about 5 X 10-3 generally
increase with increasing strength. On the other hand, fatigue
lives for total strain amplitudes greater than about 10-2 generally
increase with increasing ductility.

Source: T. H. Sanders, Jr., and J. T. Staley, "Review of Fatigue and Fracture Research on High-Strength Aluminum Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, p 472
352 12-34. Alloy 7050: Influence of Alloy Composition and Dispersoid
Effect on Mean Calculated Fatigue Life

Zn Mg Cu 6%Zn
5.5 2.2 2.3 2.2% Mg
to to to 0.4% Mn
6% 2.4% 2.4%
Low humidity Aged 3 h at 121°C
Aged 3 h at 121°C (250°F) 1.2 (250°F) + 9 h at
+ 9 h at 163°C (325°F) 163°C (325°F)

Alloy 7050 sheet Alloy 7050

1.0 1.0

0.8
High humidity

0.6

0.4

0.2

% Zr 0.1 0.1 %Cu 1.0 2.3 1.0 2.3

%Mn 0.4 0.4

Effect of dispersoid type (based on composition) on fatigue crack prop-

agation life of 7050 alloy sheet.

The influence of alloy composition on dispersoid effect is shown in the

above bar chart. The general trend in this chart is that for more finely
dispersed particles the fatigue crack propagation life is increased. Whereas
dispersoid type appears to have a relatively small effect on mean calculated
life, the smaller precipitates provided by aging produce a much larger effect.
There is some evidence that new processing practices may provide the fine
microstructures needed to enhance fatigue resistance. The potential of inter-
mediate working (commonly referred to as ITMT treatments) remains at-
tractive but has not been proven for notched specimens.

Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, p 192
 12-35. Alloy 7050: Effect of Grain Shape on Cycles to Failure 353

o
Q.
~

.400
UJ
o
::J
I-
...J
Q.

«~ 3
• 7050 AR
(J)
• 7050 HR
(J)
UJ
• •
II:
:n 200

...
105 106
CYCLES TO FAILURE

Stress-life curves for two 7050 alloys having fine,

equiaxed grains (AR) and pancake-shaped grains
(HR).

As indicated in the above graph, grain shape showed no per-

ceptible difference in life over a range of stress amplitudes.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 238
354 12-36. Alloy 7075 (TMP. T6 andT651): Effect of Thermomechanical
Processing on Cycles to Failure

PS TS
• 7075 TMP 587 632
300 ~ od~ ... 7075 TMP 600 627
o 7075-T651 516 573
\ \ 67075-T6 488 567
m
a.. \~
~ \ \
If)
If) '6 0\0
e
tl & \
CJ)
c 200
\ 0 6- 0
.~ 6'h "-
~
E
<i:
\6,
""4_
'
0 0'0Q,
'lIQ .Q..
6 --_~~
100 L..- """'--- """'--- ......... ........

4
10
Cycles to failure

Effect ofthermomechanical processing (TMP).on the un-

notched fatigue properties of the commerical AI-Zn-Mg-
Cu alloy 7075. PS = proof stress (MPa); TS = tensile
strength.

Detailed studies of the processes of fatigue in metals and al-

loys have shown that the initiation of cracks normally occurs at
the surface. It is here that strain becomes localized due to the
presence of pre-existing stress concentrations such as mechani-
cal notches or corrosion pits, coarse (persistent) slip bands in
which minute extrusions and intrusions may form, or at rela-
tively soft zones such as the precipitate-free regions adjacent to
grain boundaries. Density has also been found to improve the
fatigue performance of certain alloys, although this effect arises
in part from an increase in tensile properties caused by such a
treatment (see above diagram). It should be noted, however,
that the promising results mentioned above were obtained for
smooth specimens. The improved fatigue behavior has not been
sustained for severely notched conditions, and it seems that the
resultant stress concentrations override the more subtle micro-
structural effects that have been described.

Source: I. J. Polmear, Light Alloys. Edward Arnold Ltd. London, England, and American Society for Metals, Metals Park OH,
1981.P 41
 12-37. Alloys 7075 and 7475: Effect of Inclusion Density on 355
Cycles to Failure

400
1:1
Q.
~

ILl
0
;:)
I-
300
:J
Q.
~
c:(

I/)
I/) 200
ILl
II::
l-
I/)

100
104 105 106 107 108
CYCLES TO FAILURE

Effect of inclusion density on the stress-life behav-

ior oftwo 7XXX alloys: high-inclusion density, al-
loy 7075; low-inclusion density, alloy 7475,

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 233
356 12-38. Alloy 7075: Effect of TMT on Cycles to Failure

n 400 7075 oY;I~Nm-' ]

'E 7075TMT 600
z
~
u
:gw 300
a:
~
til
Cl
Z
~ 200
z
a:
w
~
«
100-'----,..---.-----,-----.----.---
10' 10' 10' 107 10'
CYCLES TO FAILURE

Influence of TMT on S·N curves (R = -1).

There is evidence in the literature that a uniform dislocation

density introduced by cold working improves the fatigue life
also in connection with FTMT. The above graph shows an ex-
ample taken from the work of Ostermann. Most of these im-
provements are due to an increased yield stress.

Source: G. Lutjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, p 195
 12-39. Alloys 7075 and 7050: Relative Ranking for Constant 357
Amplitude and Periodic Overload

Total life

2.0 X 106
fi
. • .• .• .• .• •
?<
::'::)
c~ns.tant amplitude life
Life Increase due
to retardation
Overload ratio
= 1.8
Applied every
4000 .cycles
,

Overload ratio
= 1.8
Applied every
8000 cycles
,

'0
j 1.0 X 106
E
::::J
Z

0.5 X 106 Constant Overload ratio

amplitude =1.4
(Overload ratio Applied every
= 1.0) 4000 cycles
I

Relative ranking of fatigue life of 7075 and 7050 aluminum alloys under constant
amplitude and periodic single overload conditions.

Crack-growth retardation is caused by tension overloading during fatigue testing.

The variable-amplitude test is believed to be more sensitive to alloy difference, and it
clearly provides more useful information for alloy-development investigations. For
example, as illustrated by the data for alloys 7075 and 7050 in the above graph, quite
different results are obtained in constant-amplitude tests than in tests with single over-
loads every 4000 or 8000 cycles. Thus, information on the variation in load level during
fatigue cycling is required for correct characterization of the fatigue behavior of alum-
inum alloys.

Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982, P 197
358 12-40. Alloy 7075: Effect of Environment and Mode of Loading

o AIR
~ • NaCl

;f
160
'0
-0
--- .0... - - 0 -

.; 120

'"
~
a:
I;;
u 80
::;
u
~

MODE I LDADING

'" ,I
7
10

o AIR
• Nael

;f
.; 120

'"
'"
uJ o
....a: --_~ 0
'"
u 80 • --0
::;
u
>-
u

40 MODE J LOADING
• •

0 ....._ ........-1.-1.........................._ ........- - ' - - '.....................' - - _........- - ' - - ' - ' "....., ...............

10~

Fatigue behavior of 7075 aluminum alloy in air and aerated sodium

chloride solution: (above) under mode 1 loading; (below) under mode 3
loading.

Tests performed on a commercial 7075 alloy in a mode 3 loading condition (torsion) indicated that
the reduction in fatigue resistance associated with cathodic charging was considerably less than it was
under mode I loading (note above charts). Although total immunity to corrosion fatigue was not
observed, the slight reduction in fatigue resistance can be associated with conditions that did produce a
true mode 3 loading condition both on a micro-scale and on a macro-scale.
To summarize the aluminum alloy results, it appears that corrosion reactions liberate hydrogen,
which effectively embrittles the region in the vicinity of a crack tip. The specific details of the embrit-
tlement are not known, but it appears that dislocation transport of the hydrogen is involved. It has been
speculated that hydrogen may collect at the semicoherent precipitate-matrix interface, thus explaining
the reported fracture plane; however, a great deal more research will have to be performed before a
more definitive answer will be available.

Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 356
 12-41. Alloy 7075-T6: Effects of Corrosion and Pre-Corrosion 359

200 -------------------
7075 AI T6 - 28
0.5 M NoCI
276 MN/m (40 ksil mean stress
A RT 24
N-
E 150
<, 20
z
~
'iii

V> - 16
""
V>

~
w
cr 100
l-
V> 12

~
u
:i
u 8
u>- 50 -
A AIR
B CORROSION 4
C PRE CORRODED/AIR FATIGUE
D PRE CORRODED/HEAT TREATED /AIR FATIGUE
L.LLL.J...U _ _L I I l l d _ _ .I_-.l...-
0--- I
0
4 10 5
106 10
7
10

The effects of corrosion and pre-corrosion on the fatigue lives of a 7075-

T6 alloy. Note that re-solutionizing and re-aging the alloy after pre-
corrosion results in a significant increase in fatigue resistance.

Fatigue resistance of high-strength aluminum alloys is severely affected by corrosive

solutions, especially chloride solutions, and this behavior has been attributed either to
preferential dissolution at the tips of the growing cracks or to preferential adsorption
of damaging ionic species. Experiments on a 7075-T6 commercial alloy and on a high-
purity analog of the alloy (AI-5.0Zn-2.5Mg-1.5Cu) indicate that localized hydrogen
embrittlement may be responsible for the poor corrosion fatigue resistance of these
alloys. For example, the above diagram shows the results of fatigue tests performed on
the 7075 alloy under simultaneous exposure to cyclic stresses and a corrosive environ-
ment (curve B) compared to tests performed in laboratory air (curve A). If specimens
are pre-corroded and tested in laboratory air, there is also a significant reduction in
fatigue resistance (curve C). The reduction in life at low Nfis associated with pits which
form at nonmetallic inclusions. If the alloy is re-solutionized and aged, equivalent to a
low-temperature bake, a significant amount offatigue resistance is regained, indicat-
ing at least partial reversibility of the damaging phenomenon and strongly suggesting a
solid-solution effect arising from environmental interaction.

Source: D. J. Duquette, "Fundamentals of Corrosion Fatigue Behavior of Metals and Alloys," in Hydrogen Embrittlement and
Stress Corrosion Cracking, R. Gibala and R. F. Hehernann, Eds., American Society for Metals, Metals Park OH. 1984, P 265
 12-43. Alloy 7075: Effect of Cathodic Polarization on 361
Fatigue Behavior

2U

.-... ,.
-,
~" .
I~

L .. .' ,.
' ~~' ..... 10 ~
. / " " 13v Na2S04
13v NaCI
~

.---- .
N,

Effect of cathodic polarization on the fatigue behavior of

7075 AI alloy in NaCI and Na Z S0 4'

It had been previously observed that halide ions are particu-

larly damaging to the fatigue behavior of Al alloys; however, if
the alloy is cathodically charged during stressing, sulfate ions
prove to be equally damaging, particularly at long NJ , At lower
NJ the slight decrease observed in cr solutions appears to be
associated with damage to the passive film, as shown in the
above S-Ndata, In SO~ solutions, a crack must initiate to break
the protective film to allow access to the bulk alloy, Cathodic
charging of the high-purity analog of the 7075 alloy also shows a
reduction in fatigue resistance, In many cases, fatigue crack
initiation in the equiaxed-grain high-purity alloy is inter-
granular, and at more active cathodic potentials there is a ten-
dency toward a higher percentage of transgranular cracking.

Source: D. J. Duquelle, "Fundamentals of Corrosion Fatigue Behavior of Metals and Alloys," in Hydrogen Embrittlernent and
Stress Corrosion Cracking, R. Gibala and R, F. Hehemann, Eds., American Society for Metals, Metals Park OH, 1984, P 266
362 12-44. Alloy 7075-T6: Effect of Surface Treatments and Notch
Designs on Number of Cycles to Failure

50 o PAlMlIIC ACIO
" ANODIZED AND WATER SEALED

41

~
~ 40
"j
ij
!:1'" }I
I;;;
......'------------~
}O

4
10 I~ 10
4 lOB
NUMBER OF CYCLES

Tension fatigue test of 7075-T6 aluminum alloy sheet, notch factor K T = 1.

·}D
o PAlMIIIC ACID
" ANODIZED AND WATER SEALED

21

~
~20 o
" 0 0

ij
ol-

~15
----------
t;;
10

4
10 I~ 104
NUMBER OF CYClES

Tension fatigue test of 7075-T6 aluminum alloy sheet, notch factor K T = 2.37.
 363

45
• BARE·PQISHffi
o ANODIZm AND NOI WAHR S£AL£D
o PALMIIIC ACtO
40

000

o
o

ANOD Izm AND HOI

WATER S£AUD
20

6
101 10
NUMBER OFCYCLES

Flexure fatigue test of 7075-T6 aluminum alloy sheet.

The three charts show the effects of notch designs and surface treatments on fatigue properties
of aluminum alloy 7075-T6 sheet; the table shows the effects of 17 surface treatments.

No. of cycles No. of cycles

Surface treatment to failure Surface treatment to failure

Polished . 125,000 Palmitic acid ••••••••••••••• 30,000,000

Anodized and water-sealcd ••• 125,000
Propionic acid •••••••••••••• 2,800,000
Vale ric acid •••••••••••••••• 15,000,000
Caproic acid •••••••••••••••• 9,200,000
Octanoic acid ••••••••••••••• 12,300,000
Decanolc acid .. 7,500,000
Lauric acid ••••••••••••••••• 8,600,000
Myristic acid ••••••••••••••• 11,600,000
Stearic acid •••••••••••••••• 8,700,000
Docosanoic acid ••••••••••••• 6,000,000
Sebacic acid •••••••••••••••• 13,700,000
Octyl alcohol ••••••••••••••• 6,000,000
Dodecyl alcohol ••••••••••••• 7,000,000
Dodecylamine •••••••••••••••• 18,500,000
Hexandeiamine ••••••••••••••• 3,000,000

Sheet was anodized: 15% sulfuric acid, 23 °C, 15 amp/sq ft, 40 minutes.
Stress amplitude: 26,000 psi.

Source: Irvin R. Kramer. "Improvement of Metal Fatigue Lifebya ChemicalSurface Treatment, "in Fatigue-An Interdisciplinary
Approach, John J. Burke, Norman L. Reed and Volker Weiss, Eds., Syracuse University Press, Syracuse NY, 1964, pp 250,251
364 12-45. Alloy 7075-T6: Effect of R-Ratio on Fatigue
Crack Propagation

1 - R I I I I 3 R o
10 f-- 0.00
0.00 .20 •
.20x .33e
.33e
.50l>
I; .50l>
.700
c .80 -
.700 a...
2
10- 1 f-- .80- ~ 10 f--
Q)
U
Q) >.
U .....u
>. E
o
..... 10- 2 f--
E
E ~10­
E <I
I
z lie:
"
.....
C 10- 3 f--
~

It
I

" ...
~
~
1-

~
.....
o

.g o

~ 101 _ o
o

I I I
2 5 10 20 30
3 10 30
AK(MPavm) tJ.K (MPa Ifffi )

Fatigue crack propagation in aluminum alloy 7075-T6 showing the effect of R ratio
and the applicability of the Forman, Keraney, and Engle relation. The scatter in the
data is much less in the latter.

The above diagrams show that data scatter is much less when the rate daldN is
computed according to the equation due to Foreman et al. R< O. The proposed Fore-
man equation is:
da C(t::.K)1I1
dN (I -R) K c - s«

Source: Marc Andre Meyers and Krishan Kumar Chawla, "Mechanical Metallurgy: Principles and Applications," Prentice-Hall,
Inc., Englewood Cliffs NJ, 1984, p 716
 12-46. Alloy 7075: Effect of Predeformation on Fatigue 365
Crack Propagation Rates

VACUUM
-2
10

oJ
10
.9!
u
>.
~
E
E
z
"tl
Ci
"tl

,~

7 10 12 15 20 30 40 50
11K [ MNm-JI2 )

Influence of predeformation by cold rolling on fatigue

crack propagation rates for 7075. Ih 100°C, SEN speci-
mens, vacuum, R = O.I,j= 30 Hz.

Cold deformation also increases the fatigue crack propaga-

tion rate as shown in the above graph, which compares an unde-
formed structure with 10% and 20% cold rolled structures.

Source: G. Liitjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds.. American Society for Metals. Metals Park OH, 1980, P 207
366 12-47. Alloys 7075 and 2024-T3: Comparative Fatigue Crack
Growth Rates for Two Alloys in Varying Humidity

10" c:--,---,--,r-r-,.,..-----,---r--____:::>

4 78910 20
a« 1
(MPa m / 2 )

Comparative fatigue crack growth rates for alu-

minum alloys 2024-T3 and 7075-T6 in air ofvary-
ing humidity.

Relationships between rate of growth of fatigue cracks and

stress intensity for the alloys 2024-T3 and 7075-T6 are shown
above. Other 2xxx series alloys show rates of crack propagation
similar to tha t of 2024-T3 over most of the range of test condi-
tions. In general, these alloys have rates of crack growth that are
close to one-third those observed in the 7xxx series alloys. It is
now common to use precracked specimens to assess compara-
tive resistance of alloys to stress-corrosion cracking, since this
type oftest avoids uncertainties associated with crack initiation.

Source: I. J. Polmear, Light Alloys. Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981,p79
 12-48. Alloy 7075-T651: Fatigue Life as Related to 367
Harmonic Generation

FATIGUE LIFE EXPENDED

50 30*> 60*

I..LI
o 0
l:::Ji=? 4
-' - '
<t:a..
::?:::?:
""<t:
3
~~
2
LL.
00
I..LI::?:
=>""
-'<t: 2
<t:::c
>0
>£2
L58
a.. I..LI
Vl

00 1000 2000
FATIGUE CYCLES

Peak value of normalized second harmonic gener-

ation as a function of fatigue life.

Recent experiments on flexural-fatigue specimens (aluminum

alloy 7075-T651) clearly show the potential of harmonic genera-
tion for fatigue monitoring. The above chart shows the peak
value of the harmonic generated as a function of fatigue life. At
60% ofthe fatigue life expended, the harmonic had increased by
about a factor of four.

Source: O. Buck and G. A. Alers, "New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstruc-
ture, American Society for Metals, Metal Park OR, 1979, P 137
368 12-49. Alloys 7075-T6 and 7475-T73: Effect of Laser-Shock
Treatment on Fatigue Properties

.25" OPEN
0_
HOLE

.75" 0.0. LASER SHOCK PROCESSED AREA

2
(26 J/cm / SIDE)

Fatigue Test Specimen Geometry

FATIGUE TEST RESULTS FOR 7075-T6 ALUMINUM

"MAX = 20 KSI NET R = .t

CONTROL SPECIMENS LASER-SHOCKED

SPECIMEN NO. NO. OF CYCLES SPECIMEN NO. NO. OF CYCLES

TO FAILURE TO FAILURE

1 51,500 1 473,800
2 n,loo 2 520,700
3 385,000

AVG. 169,600 AVG. 497,250

SCATTER 7.5 SCATTER 1.1

FATIGUE TEST RESULTS FOR 7475-T73 ALUMINUM

"MAX = 20 KSI NET R= • I

CONTROL SPECIMENS LASER-SHOC KED

SPECIMEN NO. NO. OF CYCLES SPECIMEN NO. NO. OF CYCLES

TO FAILURE TO FAILURE
1 41,500 1 171,800
2 74,300 2 266,200
3 109,300

AVG. 75,033 AVG. 218,950

SCATTER 2.63 SCATTER 1.5

The fatigue test specimens were 0.25 inch thick by 1.5inches wide and approximately 9.5 inches long, as shown
in the above sketch. The specimen blanks were laser-shock processed, and then the 0.25-inch-diameter hole was
bored through the center of the laser-shock-processed area. The diameter of the laser-shock-processed area is
three times the fastener hole diameter. All of the specimens had machined surfaces of less than 125RMS. All of
these open-hole specimens were fatigue tested to failure at a maximum net section stress of20,000 psi, and an R=
0.1 under constant-amplitude load control. Three control specimens for each material were tested to establish
the typical fatigue life for the material. Two LSP specimens were tested for each material to establish the degree
of improvement due to the laser-shock processing.
The fatigue test results for the 7075-T6 material are summarized in the upper tabulation. The LSP specimens
showed three times better fatigue lives on the average and much less scatter than the unprocessed material. The
results for the 7475-T73 material are summarized in the lower tabulation; these show the same typically large
increases in fatigue life and reduced scatter. It should be noted that the 7075-T6 material shows better fatigue
resistance than the 7475-T73 material, whether or not it is laser-shock processed. This is largely due to the
differences in dislocation/precipitate interactions that result from the T6 and T73 heat treatments. The disloca-
tions appear to shear through the precipitate particles in the T6 condition. The precipitate particles are appar-
ently so strong in the T73 condition that the dislocations just loop around the particles.

Source: William F. Bates, Jr., "Laser Shock Processing of Aluminum Alloys," in Source Book on Applications of the Laser in
Metalworking, Dr. Edward A. Metzbower, Ed., American Society for Metals, Metals Park OH, 1981,pp 256-258
 12-50. Alloy 7075-T6: Effect of Laser-Shock Treatment 369
on Hi-Lok Joints

3/16~ HI-LOKS & TAPER-LOKS

(4 PLACES)

0- -0

c:::::=:=========$$:Jr==== .072"

Joint Fatigue Test Specimen Geometry

106 .....-- - - - - - - - - - - - - - - --== 0=----.

IZZI SYD HOLE
~ LASER SHOCKED HOLES

LOAD
CYCLES TO
FAILURE

Iff

"MAX = 25 KSI NEJ "MAX = 17 KSI NET "MAX· 14 KSI NET

Fatigue Test Results for Laser-Shack-Processed

7075-T6 (Clod) Hi-lok Joints

The full load transfer joint shown in the above sketch was made from
7075-T6 clad aluminum alloy and fatigue tested. The purpose of this test was
to evaluate the fatigue life improvement of laser-shack-processed fastener
holes when the holes are loaded by the Hi-Lokfastener in bearing. A second-
ary purpose was to find out if the cheaper Hi-Lok fastener system in a
laser-shack-processed hole would show as good a fatigue life as the much
more expensive Taper-Lok fastener system. The above bar chart shows the
test results for three different stress levels. At each stress level, three speci-
mens with standard holes and three specimens with laser-shack-processed
holes were tested. The specimens tested at the l4-ksi stress level showed se-
vere fretting at the intersection of the hole wall with a badly galled area of the
fretted faying surface. All of the fatigue origins occurred at or near the hole
wall corners on the faying surface.

Source: William F. Bates, Jr., "Laser Shock Processing of Aluminum Alloys," in Source Book on Applications of the Laser in
Metalworking, Dr. Edward A. Metzbower, Ed., American Society for Metals, Metals Park OH, 1981, pp 262-263
370 12-51. Alloy 7075 (High Purity): Effect of Iron and Silicon on
Cycles to Failure

400
n
I
E
z
:::E
u
If) 300
If)
UJ
a:
Iii
o
Z
200
~
Z
a:
UJ
~
<
100-'--,----r-----.,-----.-----,--
10' 10' 10' 10' 10·
CYCLES TO FAILURE

Influence of Fe and Si content on SoN curves (R = -1).

The large Fe- and Si-containing inclusions are detrimental to

the fatigue life of smooth specimens, because these inclusions
serve as easy nucleation sites for cracks. Comparing two alloys,
one containing these inclusions (Commercial Purity 7075) and
the other one not (High Purity 7075), shows the improvement in
fatigue life due to the removal of these inclusions (see the above
SoN curves). The alloy termed High Purity 7075 in this figure
still contains Cr and therefore the small Cr-containing inclu-
sions. This is important because the removal of these small in-
clusions would have the opposite effect on fatigue life.

Source: G. LUtjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds.• American Society for Metals, Metals Park OH, 1980, p 193
 12-52. Alloy X-7075: Effect of Grain Size on Cycles to Failure 371

o
0..
::r 300
UJ
a
...::>
::::;
0..
::r
«
Vl 200
Vl
UJ
a::
Vi

100L~,----------~;----------'-=--------~;- __
o
CYCLES TO FAILURE

Influence of grain size on S-N curves (R = -1,

f= 100 Hz) for X-7075 with PFZ (20 hat 160°C,
or 320 OF).

The resulting improvement in fatigue life due to the grain size

reduction for this crack nucleation mechanism is shown in the
above S-N curve. Again, the tensile yield stress was equal for
both grain sizes. Also for low-cycle fatigue it was found that
red ueing the grain size of 7XXX series alloys results in increased
fatigue life of smooth specimens in the averaged condition.

Source: G. Lutjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys." in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals. Metals Park OH. 1980, p 192
372 12-53. Alloy X-7075: Effect of Grain Size on Stress-Life Behavior

400

"
c,
::E

w 300
0
~
::::i
a,
::E
<I:
lI)
lI)
w 200
0:
l-
ll)

CYCLES TO FAILURE

Aluminum alloy X-7075; 24 h at 100 °C (212 OF).

The above chart shows the grain-size effect in a stress-

controlled test for a high-purity 7075 alloy (X-7075) aged to con-
tain shearable precipitates. Since the flow stress is determined
by the interaction of dislocations with the coherent precipitates,
the yield stress is approximately the same for both alloys. Opti- .
cal examinations of the specimen surfaces show that cracks nu-
cleate much earlier in specimens with the large grain size. Cracks
nucleated at intense slip bands for both grain sizes.

Source: Edgar A. Starke. Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure." in Fatigue and Microstructure.
American Society for Metals. Metals Park OH, 1979. P 225
 12-54. Alloy X-7075: Effect of Environment; Air vs Vacuum 373

A: 24h 100'C VACWM

-I
10 C:48h 180'C AIR

..
n
~ 10
-)

u
E
E _,
u 10
Z
:E
a
'0 -i
10

I06..L..----,--r--..,.....--..--..,.....---.------r----r--..,......

Influence of environment (laboratory air) on fatigue

crack propagation rates for underaged (A) and overaged
(C) condition. X-7075, CT specimens,R = O.I,f= 30 Hz.

A basic correlation between microstructural parameters and

fatigue crack propagation rate can only be determined so clearly
if the tests are performed with the exclusion of any aggressive
environment. To illustrate this point, the above graph shows the
comparison between underaged and overaged microstructure
also for tests performed in laboratory air. The aggressive envi-
ronment has a much more pronounced effect on the underaged
condition, leading even to an opposite ranking of the alloy con-
ditions. In laboratory air the cracks propagate still along slip
bands at low dajdN rates.

Source: G. LOtjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 204
374 12-55. Alloy X-7075: Effect of Environment on Two Different
Grain Sizes

-2
10
GS 200.lJrn

- -J
GS 46.lJrn
~ 10
u
>-
.!2
E -4
E 10
z
~ n
,,/~ 3.5NoCI
.g -5
10

I06.L----.-----.---'-r---..-----r-----.---...----.-
7 10 15 20 30 40 50
t:. K [MNrn-3I2 )

Influence of environment (3.5% NaCI) on fatigue crack

propagation rates for two different grain sizes. X-7075,
24 h 100 DC, CT specimens,R = 0.1,1= 30 Hz.

The same tendency is observed for the grain size dependence

of crack propagation ifthe tests are carried out in a 3.5% NaCI
solution (note above curves). The influence of environment is
larger for the large grain size. For this highly aggressive envi-
ronment the cracks propagate at low dal dN rates along grain
boundaries in a complete brittle fashion.

Source: G. Lutjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 204
 12-56. Alloy X-7075: Effect of Grain-Boundary Ledges on 375
Cycles to Failure

..-
c
Q.
! 300 r--~"-T'"""...-r'1""""'-r---'----r-"'T"""T"T"l"T~---'----"--"-"""""""'"
co.. ••
LLJ \
o 'PO
\
~ 250 I---~ID-----':~_----_--t-----____l
\
:J b
Q. ~o.
....
~
~ 200t------O-~~--""'~-+__---____l "
X-7075

150 t----::-----±-:-~-:-----t-----_l
0---0 If = 0, 8h 160°C
--If = 0.5,4h 160°C

105
CYCLES
Effect of grain-boundary ledges on the stress-life
behavior of an alloy containing nonshearable
precipitates and PFZ.

One method that may be employed to reduce the slip length in

the PFZ is thermomechanical processing. If enough cold de-
formation is employed to introduce steps (or "ledges') into the
grain boundaries, the effective slip length within the PFZ is
drastically reduced (similar to a small grain size) with corre-
sponding improvement in resistance to fatigue-crack nuclea-
tion. The above chart shows the results of a stress-controlled test
for two high-purity 7075 alloys, one cold-worked 50% to pro-
duce grain-boundary steps. The cold work drastically reduced
the incidence of grain-boundary cracking and improved the fa-
tigue life at high stress amplitudes. At low stress amplitudes and
long fatigue lives, crack nucleation occurred at inclusions for
both alloys. This effect is most likely due to stress concentration
at inclusions.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 230
376 12-57. Alloys X-7075 and 7075: Effects of Chromium Inclusions on
Fatigue Crack Propagation

-1 24 h 100·C VACUUM
10

n
~ -)
u 10
~'"
E
E
u -.
10
z
~
a
"0 -5
10

-6
10
7 40 50

Influence of Cr-containing inclusions on fatigue crack

propagation rates by comparing aluminum alloys X-7075
and 7075.24 h 100 °C, CTspecimens, vacuum,R = 0.1,!
= 30Hz.

As shown above, the small inclusions have a much stronger

influence on fatigue crack propagation because they lower the
reversibility of slip and they crack within the plastic zone ahead
of the crack tip. Furthermore, they normally increase fatigue
crack propagation rates also indirectly by their effect on grain
size and shape.

Source: G. Lutjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 207
 12-58. Alloy 7475-T6: S-N Diagram for a Superplastic 377
Fine-Grain Alloy

80 r------------------------------.
o
SPF Conditions: T= 516·C = As pr......d
i=21104S-1 A = SPF - highsup.rpl.stic st"inl
70 o = SPF - low luporpl.stic st"inl
-+ = No,.ilu"

R =+0.1
60
AS PROCESSED S·N CURVE
MAX
STRESS 50 AOC>AO o
(Ksi)

40 o _
-...
0

...
6.

_--------
(]I

3D 0-+0-+

20 I-.....I._..I....I..I-I._..................I-.....I._...."""'"-"u.._"'-..........I."""".........._ ....."""'"-"U
10' 10' 1~ ,~ 10' 10'
CYCLES TO FAILURE

S·N curves for a superplastic aluminum alloy: fine-grain

7475. All testing was done with smooth specimens.

Tests on fine-grain 7475 alloy have shown improved fatigue

life as superplastic strain is increased, as shown in the above SoN
diagram. An even more dramatic improvement is obtained in
damage tolerance.

Source: C. Bampton, F. McQuilkin and G. Stacher, "Superplastic Forming Applications to Bomber Aircraft ." in Superplastic
Forming. Suphal P. Agrawal. Ed.. American Society for Metals. Metals Park OH. 1985. p 77
378 12-59. Alloy 7475: Effect of Alignment of Grain Boundaries on
Cycles to Failure

300
0
,.....
0
Q..
~
-..J 250
I

UJ
0 - II o <9=0 16h 160°C
:J -<9=0.5 6h 160°C
~
200
!
...J
Q.
~
<{

I/l
I/l 0
UJ 150
0:
~
II)
0 o
7475
I 1. I I

CYCLES TO FAILURE

Effect of alignment of grain boundaries-and alignment

plus steps in grain boundaries-on the stress-life behavior
of a 7475 aluminum alloy containing nonshearable pre-
cipitates and PFZ.

If the stress axis is parallel or perpendicular to the long grain

dimension, there will be no shear stress parallel to the grain
boundary, and preferential deformation within the PFZ will be
restricted. Grain-boundary alignment is then as effective in re-
stricting deformation in the PFZ as are steps produced by
thermomechanical treatment: this is shown by the stress-life
curves in the above graph.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 232
 12-60. Alloy 7475-T6: Superplastic vs Nonsuperplastic, as Related 379
to Fatigue Crack Growth

1 X10-] r - - - - - - - - - , . - - r - - - - - - - - ,

7475 T-6
FINE GRAIN

~ 1 X 10-'
(.)

~
.....
(J) 7075 T-6 -
w
::I:
(.)
Z 1 X 10- 5
Z
o
<,
<
o

1 X 10- 6

1 X 10- 1 1.----I._.L.I.....L..J..J.._...l----JL...-..JL.I...l.----I._.....L....J...J..J
1 10 100 1000
K (I-R)M-l (KSI "'lINCH)

This comparison of conventional, coarse-grain, nonsuper-

plastic aluminum alloy 7075 with superplastic alloy 7475 shows
almost an order-of-magnitude reduction in crack growth for the
superplastic material.

Source: C. Bampton, F. McQuilkin and G. Stacher, "Superplastic. Forming Applications to Bomber Aircraft," in Superplastic
Forming, Suphal P. Agrawal, Ed., American Society for Metals, Metal Park OH, 1985, p 77
380 12-61. Alloys X-7075 and 7075: Effect of Chromium-Containing
Inclusions on Cycles to Failure

400

:g 300
~
I-
:::;
Q.
~
<l:
~ 200

~ J~----,--;,------------'--r:---""""'----""'--
10
3
I
10'
I
10
5
CYCLES TO FAILURE
I
la'
•

Influence of Cr-containing inclusions on SoN curves

(R = -1,1= 100 Hz) comparing aluminum alloys x-7075
and 7075, 24 h 100 0 C. (Arrows indicate crack nucleation
visible by LM at u a= ± 200 MNm -2. j

The above S- N curves compare results obtained from testing

commercial 7075 alloy with the alloy X-7075 which does not
contain Cr. These small inclusions, as in the tensile test, inhibit
the formation of intense slip bands, thus retarding crack nuclea-
tion as indicated by arrows on the graph. Due to these small
Cr-containing inclusions, the grain size of the 7075 alloy was
somewhat smaller as compared to that of X-7075, which also
may have contributed to the observed improved fatigue behavior.

Source: G. LUtjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 195
 12-62. Aluminum Forging Alloys: Stress Amplitude vs 381
Reversals to Failure

IJ. 7075-T73 squeeze formed

"t 400 o SF1-T61T73 squeeze formed
E • 6082-T6 squeeze formed
z x 6082-T6 extruded bar
2.:
~ SF1-T6/T73

§300~-T6
S ~
~200 .~~o
~
~

In
x
.
10 3 10 4 10s 10 6
REVERSALS TO FAILURE (2N,)

SON fatigue data for several squeeze-formed

forging-type aluminum alloys compared with ex-
truded AA 6082-T6.

The above chart presents results from push-pull, about mean

zero, fatigue tests which have been carried out on a servohy-
draulically controJled machine. The tests have been carried out
on samples cut from actual components, not from separately
made testpieces. The results from conventionaJly extruded AA
6082 (H30) are included for reference: in this case, the data are in
the longitudinal direction, it not being possible to obtain sam-
ples of sufficient size from the transverse direction.
The results show good fatigue properties for squeeze-formed
material, which in one case compare favorably with conven-
tionally extruded material. This further substantiates the claim
that squeeze formings in general are comparable with forgings
with respect to mechanical performance.

Source: G. Williams and K. M. Fisher, "Squeeze Forming of Aluminium-Alloy Components," in Production to Near Net Shape:
Source Book, C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals, Metals Park OH, 1983, p 367
382 12-63. AI-5Mg-O.5Ag: Effect of Condition on Fatigue Characteristics

180 ~------'--------r------.,----,
0.2% proof Tensile
AI-5%Mg-0.5%Ag stress IMPa) strength IMPa)
0 ST and quenched 85 L60

150 • Aged 1 day 175°C

X Aged 70 days 175°C

200

175
310

270
lU
a..
~

~
(/] 120
(/]
Q)
•
.~
~
(/]

Cl
c
'16 90 oho 0_ 87

~ ~....
c
~
~ X 72
<i:
&0
48
~
10& 7 lOll
10
Number of cycles

Fatigue (SoN) curves for the alloy AI-SMg-O.SAg

in different conditions.

The fact that microstructure can have a greater influence

upon the fatigue properties of aluminum alloys than the level of
tensile properties has been demonstrated for an AI-Mg alloy
containing a small addition of silver. It is well known that binary
AI-Mg alloys such as AI-5Mg, in which the magnesium is pres-
ent in solid solution, display a relatively high level of fatigue
strength. The same applies for an AI-5Mg-0.5Ag alloy in the
as-quenched condition, and the above diagram shows that the
endurance limit after lOS cycles is ±87 MPa, which approxi-
mately equals the 0.2% proof stress. This result is attributed to
the interaction of magnesium atoms with dislocations, which
minimizes formation of coarse slip bands during fatigue. The
silver-containing alloy responds to age hardening at elevated
temperatures due to the formation of a finely dispersed precipi-
tate, and the 0.2% proof stress may be raised to 200 MPa after
aging for one day at 175°C (350 OF).

Source: I. J. Polrnear, Light Alloys. Edward Arnold Ltd. London, England, and American Society for Metals, Metals Park OH.
1981, P 42
 12-64. AI-Zn-Mg and AI-Zn-Mg-Zr: Effect of Grain Size on 383
Strain-Life Behavior

100 r----------------...,
a,
" ", 24h at 150GC
GI
.':::. .. _- " ...........
........ ::---
'
-'.::: ....
...
...... ,
(\J e-- :.:~~

-...
Q.
<.&> ......

<l
o Failure AI-Zn-Mg
• Failure AI-Zn-Mg -Zr
, Crack Initiation AI-Zn-Mg-Zr

10
2Nf

Effect of grain size on the strain-life behavior of an

alloy having nonshearable precipitates plus PFZ.
The AI-Zn-Mg alloy had large grain size; the AI-
Zn-Mg-Zr, small grain size.

The above chart shows Coffin- Manson life plots of two aver-
aged AI-Zn-Mg alloys. The small-grained AI-Zn-Mg-Zr alloy
has a much longer life than does the large-grained AI-Zn-Mg
alloy. The improvement in life is attributed to increasing the
cycles to crack initiation, as indicated in the chart. A conver-
gence is noted for long lives (low plastic-strain amplitudes) for
this strain-controlled test. Since the fine-grained material
hardens more than the other at low strains, the stress to enforce
the applied strain is greater at long lives, and this affects the life
improvement due to the fine grains.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979. P 228
384 12-65. AI-Zn-Mg: Strain-Life Curves of a Large-Grained Alloy

• 4h at 120°C
10.
.96h at 150°C

c..
wiN
<l ..........

I.

AI-Zn-Mg

0.1
10 1

Strain-life curves of large-grained AI-Zn-Mg alloy having

shearable precipitates when underaged (4 h at 120 ° e, or
250 OF) and nonshearable precipitates plus PFZ when
overaged (96 h at 150 "C, or 300 OF).

Since the strain localization occurs in a region free of solute, overag-

ing the matrix precipitates or adding dispersoids does not homogenize
the deformation. This is clearly illustrated by comparing the Coffin
Manson life curves of underaged and overaged specimens of large-
grained AI-Zn-Mg alloy (see above chart). The tensile yield strength
and strain to fracture are approximately the same for both specimens.
The underaged alloy has shearable precipitates, which results in strain
localization, the formation of intense slip bands, and early crack nu-
cleation under cyclic loading. Overaging was one method described for
homogenizing deformation; however, this method is not effective for
large-grained material. Preferential deformation in the PFZ also leads
to strain localization and results, for this particular case, in the same
fatigue life. Dispersoids distributed throughout the matrix would not
inhibit strain localization in the PFZ for the same reason.

Source: Edgar A. Starke, Jr., and Gerd Liltjering, "Cyclic Plastic Deformation and Microstructure,"in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, P 227
 12-66. Aluminum With a Copper Overlay: Stress Amplitude vs 385
Cycles to Failure

70

-g
Q.
D

_-- -----
-b
60
~ ....... c-
0
<,
<,
+1
50 <,
~o 0

o Al <,
D Al (Cu) <,
0.....
40
4 5 6 7
10 10 10 10

Stress amplitude U a versus number of cycles to failureNj

for AI and AI with a Cu layer. Note the pronounced im-
provement in the latter at large N j •

The fraction of fatigue life spent in crack nucleation, N; / NJ ,

increases with decreasing load amplitude (i.e., at high N's). It
would be expected that the treatment suggested above will pro-
duce a great effect in large fatigue life regimes (i.e., under condi-
tions where initiation of fatigue crack is more important than its
propagation). The above graph shows this phenomenon in the
case of pure aluminum and aluminum with a copper surface
layer.

Source: Marc Andre Meyers and Krishan Kumar Chawla. "Mechanical Metallurgy: Principles and Applications," Prentice-Hall,
Inc .. Englewood Cliffs NJ, 1984, P 707
386 12-67. P/M Alloys 7090 and 7091 vs Extruded 2024

60
400
l> 7091·T7E69 (2 lots) 60°
o 7090·T7E71
50 Open symbols-longitudinal
Solid symbols-long transverse
O~;"
'in 40 - - I ndicates did not fail
300 .
o,
::;;;
"'" Notch tip radius = 0.013 in.

~
KT = 3.0 4
/to:D
~-
~
30 200 E
E
OJ OJ
E
..
'x
::;;; 20
E
..
'x
::;;;
Band for 2024·T351 100
10 rod and bar
Stress ratio R " 0
Ambient air
Ol-_ _-'-_ _--'--_ _--'-_ _-'-_ _--'-_ _--'-_ _--'_ _--.J
10' 10' 10' 10· 107 10· 10· 10'0
Cycles

S-N diagram that provides a comparison of notched axial

fatigue strength for P 1M alloy 7090 and 7091 extrusions
vs 11M alloy 2024-T351 rod and bar.

The notched axial fatigue strengths of alloys 7090 and 7091

are 35 to 40% higher than those of alloys 7050, 7075 and 2024 (an
11M alloy often selected for its resistance to fatigue) at one mil-
lion or more cycles.

Source: Robert H. Graham, "Wrought Aluminum PIM Alloys," in Powder Metallurgy-Applications, Advantages and Limita-
tions, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 240
 12-68. P1M Alloys 7090 and 7091 vs 11M 7050 and 7075 Products 387

60 r-----.-----.----.-------.----.-------.---~-------,400
A 7091-T7E69 (2 lots) 60°
o 7090-T7E71
50 Open symbols-longitudinal
Solid symbols-long transverse O~"
Indicates did not fail 300 ..
] 40 a.
Notch tip radius = 0.013 in. 2:
g KT = 3.0
~~
E
:>
30 200 E
:>
E E
~ 20 ..
'x
2:
Band for 7050 and
100
Stress ratio R = 0 7075 products
10 Ambient air'

O'----:----'-,-------'-c-----'-:c-----L,---'--:---'----:----'-,------'
10 2 103 10' 10" 10 7 10· 10 9 10 10
Cycles

S-N diagram that compares notched axial fatigue strength

for P 1M alloy 7090 and 7091 extrusions vs 11M 7050 and
7075 products.

The notched axial fatigue strengths of alloys 7090 and 7091

are 35 to 40% higher than those of alloys 7050,7075 and 2024 (an
11M alloy often selected for its resistance to fatigue) at one mil-
lion or more cycles, as shown above.

Source: Robert H. Graham. "Wrought Aluminum PIM Alloys." in Powder Metallurgy-Applications. Advantages and Limita-
tions. Erhard Klar, Ed.• American Society for Metals, Metals Park OH, 1983. p 240
388 12-69. P1M Aluminum Alloys: Typical Fatigue Behavior

9·71B-in. R ?:
o
o
o
Forged £-=fL-i-----3-"1!I
x 40 Forged 601AB-T6 201 AB-T6 0.300-in. diam
'u;
c.
201AB-T6
202AB_T2~
E 60 1AB-T2'>.
.~ 20

601A~
x
2''"
_ _..::::===_=====a=_=_=_=_=_-1

OL------''------''-----'----'-----'-----'----'
10' 10' 10· 10'
Cycles to failure, N

Typical fatigue behavior of alloys 601AB, 201AB and 202AB.

Fatigue is an important design consideration for P / M parts

subject to dynamic stresses. The above S-N diagram shows typi-
cal fatigue behavior of specimens of alloys 6OIAB, 20lAB and
202AB in the T2 (as-cold-formed after sintering) and/ or T6
tempers.

Source: John D. Generous and Wayne C. Montgomery, "Aluminum PI M-Properties and Applications." in Powder Metallurgy-
Applications. Advantages and Limitations, Erhard Klar, Ed.. American Society for Metals, Metals Park OH, 1983, p 214
 12-70. P1M Aluminum Alloys: Comparison With Specimens Made 389
by Ingot Metallurgy

500
70
251 mm (9'10 in.) 60· sharp' V
: ~ 1.9;mm (0.075 in.) -1.0

E--t--]- E= f-- t
400
(0.300 in.)
tf--+-1=
7.62 mm . _ _ 8.39 mm
(0.3.30 ln.I 12.19 mm (0.480 in.)
60

diarn. diarn. I I
Notch tip radius:
'"
0.. -, e ~ 1.254 mm (0.01 in.)
50
::E
Ii • K, ~ 3 .'"
]

E
~
1;;

::>
E
300
.. 0 40
~
1;;
E
::>
E
'x 0 'x
'"
::E '"
::E
Smooth 0 30
200
a---

-- - 20

100 Solid lines represent

bands for 7075·17352
!three lots), longitudinal direction
.. •
-- 10
Dashed lines represent
bands for 2014-T61 (nine lots smooth.
-- -~

five lots notch I. longitudinal direction

0 0
10' 10' 10' 10' 10' 10'
Cycles

Rotating-beam fatigue strength for die forgings of P 1M alloy X709I-T7E76 and ingot
metallurgy alloys 7075-T7352 and 2014-T6I. For P/M X709I-T7E76: 0 ,smooth,
transverse direction; • ,notched, transverse direction; - denotes test specimen did
not fail in number of cycles indicated.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society ForMetals, Metals Park OH, 1984,p 469
390 12-71. P 1M Aluminum Alloys: Comparison With Forged 7175 for
Cycles to Failure

.500

t." L10."" J,~.L mm 10.,)' in.]

- 70

400
~t~ Notch tip radius:
60

e = 0.33 mm (0.013 in.) 'iii

K, =3 - 50 -"
ul
ul Ul
Ul
e 300 e
1il
1il
E
:>
- 40 E
:>
E E
'x
co
• m:a 'x
co
::E 0 • • 30 ::E
200 -.we • o • • •~
•It •

Band for 7175-T73~ ~

100 _forgings (six lots)
~
20

r r
Stress ratio: R = 0.0
10

o o
10' 103 10' 10' 10' 10' 10' 10'
Cycles

Comparison of axial-stress notch fatigue strength of P 1Malloy X7091-T7E69 die forg-

ings and ingot metallurgy alloy 7175-T736 die forgings. 0 , longitudinal direction,
one lot; • , short transverse direction, two lots; - denotes test specimen did not fail in
number of cycles indicated. Stress ratio: R = 0.1.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,P 469
 12-72. Various Aluminum Alloys: Comparison of Grades for 391
Corrosion-Fatigue Crack Growth Rates; Air vs Salt Water

'(f3_-------------------"'"
crack orientation L - T
specimens: SEN , GNP, DCB , CT
R • 0.0 - 0.1 , 0.1 - 30 Hz
ambient temperature

11 AI - alloys in salt water:

2048-T851 7005 - T63
2219 - T 87 7050 -T736
2618 -T6 7fJ15 - T651
5456-H117 7175 - T736
5456-H321 7475 -T651
6061 -T651
~ 18 AI- alloys in air:
2cro - T651 2618 - T 6 7050 - T 736
2024 - T 3 5456 -H117 7075 - T 6
2024-T~ 5456-H321 7079-T651
2048 - T851 6061 - T651 7106 - T 63
2219 - T851 7'005 - T63 7175 - T 736
2219 - T87 7039 - T6X31 7475 - T 651
20 30 40 50 60

cyclic stress iltensity range, 6K, [MN 'm-¥2]

Comparison of scatterbands of corrosion-fatigue
crack growth rates and fatigue crack growth rates
of many commercial aluminum alloys.

Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material," in Aluminum Transformation Technology and Appli-
cations (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 615
392 12-73. Various Aluminum Alloys: Comparison of Grades for
Corrosion-Fatigue Crack Growth Rates in Salt Water

163~------------------,
7475 - T651 7175 - T 736
7005 -T63 7075 - T 651
2618 - T6 2219 - T 87
5456 - H 321

,----,
~
10- 5
~
~
E
'------'
6061 -T651

~~
-6 2048 - T851
10
7050 - T736
5456 - H 117

10- 7

Q)
~ typical experimental scatter
::J
.21
.... crack orientation L - T
....ro specimens: CNP I DCB I CT
environment: salt water
ambient temperature
-10
R ·0,0-0,1 I 0,1 - 30 Hz
10 0,......--+---+----+----+----+---.....
10 20 30 40 50 60
cyclic stress intensity range, ~K , [MN. m-~]
Corrosion-fatigue crack growth rates in salt water
for aluminum alloys exceed the scatterband.

As shown in the above graph, curves for growth rate are

somewhat higher than the air-test scatterband, but at very low
and very high stress-intensity ranges, no significant difference
between fatigue and corrosion fatigue crack growth rates is
observed.

Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material, "in Aluminum Transformation Technology and Appli-
cations (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OH, 1980, p 614
 12-74. Various Aluminum Alloys: Wrought vs Cast, and Influence of 393
Casting Method on Fatigue Life

500

400

300
250
'"
~ 200
e <;;
~
t:
en
~ 201---"': 150 rf.
~
~ 18t----
II)
~ 161----+~
§ 141----+----,.:-' 100
E 121----'--+---'--~
~ 75
iU 1 0 1 - - - - + - - ' -......- +-..........:
":;'l
1.
8!---'---'-f--'-----if--"""'""-,..-t----'
71-----+--.-:.+------1------+--,,-:~ 50
61-----+---+--...--:'":-...;,......1----'---+..,.....-..:.,......+---~

106 lOS 109

Life N (cycles (log))

Representative S-N curves for various aluminum alloys are

shown in the above graph. Note the absence ofa sharply defined
"knee" and true endurance limit. This is typical of nonferrous
metals. In the absence of an endurance limit, the fatigue strength
at 108 or 5 X 108 cycles is often used. (To give a "feel" for the time
required to accumulate this many cycles, an automobile would
typically travel nearly 400,000 miles before anyone of its cylin-
ders fired 5 X 108 times.) As is true for most metals and alloys,
the wrought versions of aluminum alloys have greater fatigue
strength than the cast (see graph). It will also be noted in the
graph that there is an overlapping of fatigue strength for the
sand and permanent mold casting methods (same alloy).

Source: Robert C. Juvinall, Fundamentals of Machine Component Design, John Wiley & Sons, New York NY, 1983, p 207
394 12-75. Aluminum Casting Alloy AL-195: Interrelationship of Fatigue
Properties With Degree of Porosity

S-N CURVES
CASTING ALLOY AL - 195
26 DEGREE FATIGUE* DEGREE FATIGUE*
... ,,1,- 3, , , 1\ POROSITY STRENGTH POROSITY erRENGTH

24
... ... , ,~ 2 ;t
2 10,500 6
rr;ooo
- - 5 - - 9,000
8,150
... ... 1'>,

,""1\
" 5) 1) 3 9,000 1 1,000
4 8
A~~~
9,000
16, [,>.. 4 ~ 1-1',
IV * AT 10 000 000 CYCLES
22
~ i" \, I
,1"\ ,~ .... REPRESENTATION OF
20 NO FAILURE
" ... I, 'I\.
x 18 ~ " <, ,~
-- f - -
~ ....... ,'\.. r-, .~

f' ~" _~
... .... ~
6
.... -
"'I' ~ "I",

' ... , ...

~ f'
-- -
~I'
41- I - - - . . I-
8./ ... ~
12
"o
"of-.

1'1·
I""' l ...... '" I"'-\l:i:
",I'. ,
;.. "~ r-;
<, ... .. I-
...
"o I' ~ r-, r>-
J'.... ..... ~ ..... I-
10
.......
8
,,"'. .. .. .. r-...... .....-r--
} DEGREE
:-:-- POROSITY
10 Q ,,5
10 10,,6
- 101

Interrelationship offatigue properties with degree of porosity for AL-195

casting alloy.

Source: N. E. Promisel, "Evaluation of Non-ferrous Materials," in Materials Evaluation in Relation to Component Behavior
(Proceedings of the Third Sagamore Ordnance Materials Research Conference), Syracuse University Research Institute, Syracuse
NY, 1956, P 65
 12-76. Aluminum Casting Alloy LM25-T6: Squeeze Formed vs Chill 395
Cast; Effect on Reversals to Failure

N 400
I
E
z
~300
w
o
=>
l-
ii 200
2
«
(/)
~ 100 LM 25-T6 chill cast
a::
In
O'-:--_--L --'- '-::-_ _--'- --'---'
4
10 2
1~ m 1~ 1~
REVERSALS TO FAILURE(2N,)

S-N curves for aluminum casting alloy LM25-T6; chill

cast versus squeeze formed.

Fatigue tests have been carried out with LM25 samples,

which were cut from a bracket component. A servohydraulically
controlled fatigue machine was used to execute push-pull tests
about mean zero. The results are presented in the above chart,
which includes for reference the results of similar tests carried
out on conventionally cast LM25. It can be seen that a signifi-
cant improvement in the fatigue performance has been achieved
by squeeze forming this type of alloy.

Source: G. Williams and K. M. Fisher, "Squeeze Forming of Aluminium-Alloy Components," in Production to Near Net Shape:
Source Book, C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals, Metals Park OH, 1983, p 367
396 13-1. Copper: Effect of Air and Water Vapor on Cycles to Failure

5.6 ~"

s: <, I

u
c ~ -, I
...5.4 - - I---
Vi
Q;
~5.2 --
'~ l' ......
.... iA

~
~

~
'"
c
o <, Q(~I'("
I--
::f 5.0
e
Vi
---
\.\~I _______ T-D
1110
~fiedAir .... - --
~-- 1---
'0 4 .8 \
OJ
\
I' r-...

----
Ol If) 0

~
C ~
<0 .... PlJr i f · iltr)fJ
a: I 'ea"
--I" "~~-
"E - - ---
OJ
(J)
4.6
. II
InAir"_
~
-~
4.4 6 7 7
10 10 5.10
Endurance, Cycles to Fracture (Log. Scale]
- - + Indicates Specimen Unbroken

The effect of air and water vapor on the fatigue life of annealed copper.

The effect of atmospheric oxygen on fatigue life of copper has

been investigated; oxygen and water vapor reduce fatigue life in
copper. Alternate static exposure to air and dynamic exposure
to vacuum do not affect fatigue life, and SoN curves diverge as
applied stresses are reduced (see graph). Based on these experi-
ments, the investigators concluded that:
Fatigue cracks form early, because the majority of life is
concerned with crack propagation (environment has little
or no effect on nucleation and initial growth).
2 Oxygen and water vapor are the primary damaging con-
stituents in air (water vapor alone is effective).
3 Oxygen must be a gas (preoxidation or intermittent expo-
sure is not effective).

Source: D. J. Duquette. "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 336
 13-2. Copper: Applied Plastic-Strain Amplitude vs Fatigue Life 397

Cu
• prestrain 20"1.
• annealed
"prestrain 40/.

_- A-

Applied plastic-strain amplitude versus fatigue-

life curves for copper at long life.

Helgeland was the first to observe and claim a fatigue limit for
copper (actually the plateau stress, although it was not recog-
nized as such at the time). Unfortunately, his results were appar-
ently contradicted by those of Kettunen, who observed failures
at stresses down to 17.7 MPa. This difficulty was resolved by
Laird, who showed that Lukas and Klesnil's long-life Coffin-
Manson plots showed failures to occur only down to the plastic-
strain fatigue limit; at lower strains, no failures were observed in
the testing time available (see above chart). However, Lukas et
al. also carried out stress-cycling tests, in which they monitored
the plastic strain. Specimens that had been stress-cycled yielded
a plot of saturation plastic-strain amplitude versus life, where
failures occurred at strains as low as 10-5 . The difference be-
tween these tests is that in strain cycling, the stress is low in the
initial cycles and increases to saturation, whereas in stress
cycling, full application of the load in the first cycle causes a
large strain in a soft material. This initial large strain creates the
PSB cell structures, which would not otherwise form in a
constant-strain test. Since Kettunen applied the full load to his
specimens, failures were observed at stresses below that of the
plateau. Helgeland, on the other hand, although he was stress
cycling, imposed a low stress at the start of his tests and in-
creased it gradually to the chosen value.

Source: Campbell Laird. "Mechanisms and Theories of Fatigue," in Fatigue and Microstructure, American Society for Metals.
Metals Park OH, 1979, P 195
398 13-3. Copper Alloy C11 000 (ETP Wire): Effect of Temperature on
Fatigue Strength

200 r - - - - - - - - - - - - - - - , - - - - - - - - - - - - - - - ,

26

'"
Cl.
::E
150 22 ]1
s:."
...en
s:." 'C,
e
e
e 18 ~
.
1;;
::J
en
Q)
::J
en
'+>
'+> 100
'"
u. '"
u.
14

10

Stress cycles

Rotating-beam fatigue strength of electrolytic tough

pitch copper, CllOOO wire, 2 mm (0.08 in.) diam, H80
temper when tested at various temperatures.

Source: Metals Handbook, 9th Edition. Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 289 .
 13-4. Copper Alloy C26000 (Cartridge Brass): Influence of Grain 399
Size and Cold Work on Cycles to Failure

60

50 I' <,
"0
'" <, ..... 1'0- ...
0
0
a.
40
........ <,
0 I'--~ .........
r-, ~~~

--
<fl-
f_
...e
<fl

-
.... 30
Q)

-
II!,,,)" ~ ~ a
l J)
+~
....0
~?" t;.1--
I"'--.:!""...
::l
><
Q)
20 1'; I- :: .. r--
r- x_
x
IJ.. -i
-+ +
10

6
10 107
Cycles for Failure ,N
Legend
o Group G (60% Cold Drown)
I:>. Group D (40% Cold Drawn)
o Group A (20% Cold Drawn)
• Anneal I Grain Size 0.0/2 mm.
... Anneal 3 Grain Size 0.026 mm.
x Anneal 4 Grain Size 0.051 mm.
+ Anneal 5 Grain Size 0.131 mm.

Influence of grain size and cold work on fatigue

strength of copper alloy C26000 (cartridge brass).

Changes in grain size and in degree of cold work which result

in increased tensile strength or hardness usually result in im-
proved fatigue strength. The above S-N curves illustrate this and
indicate the influence of grain size and cold work on the fatigue
strength of alpha brass. All specimens were prepared from the
same heat and, therefore, had the same nominal composition.

Source: George M. Sinclair, "Some Metallurgical Aspects of Fatigue.vin Fatigue-An Interdisciplinary Approach, JohnJ. Burke,
Norman L. Reed and Volker Weiss, Eds., Syracuse University Press, Syracuse NY, 1964, P 69
400 13-5. Copper Alloy C83600 (Leaded Red Brass): S-N Curves;
Scatter Band

180 ~------..---------r--------''-------~_-----~
25

160 f---~~~-+-------+------+-------+--------l

140 f------~ 20 ]
rf
:::E .J::.'
t»
~ s
~ 1201--------+------' 1;;
1;; Ql
Ql :::l
:::l

'"
.~
'"
.~

15 u.
u. 100 f - - - - - - - + - - - - - - - + - ------+-------1

80 f-------+--------+-------I-----= -
0- 10

Stress cycles

S-N curves (scatter band) for copper alloy C83600 (leaded red brass).

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 406
 13-6. Copper Alloy C86500 (Manganese Bronze): S-N Curves; 401
Scatter Band

400

<ttIOo. 0
- 50
If 300 ~
:E
~~ - 40 .r."
~c

--
",
.................. c
e
e 200 ~~
30
t:
t: n Cll
Cll
;:)

.~
- 20 .;::;c:
;:)

- 10 '"
100 u.
'"
u.

Number of stress cycles

S-N curves (scatter band) for copper alloy C86500 (man-

ganese bronze). All testing was performed at room
temperature.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, ArnericanSociety
for Metals, Metals Park OH, 1979, p 35
402 13-7. Copper Alloys C87500 and C87800 (Silicon Brasses): S-N
Curves; Scatter Band

300
.
D-
40 ';;;
~
~ 250
.,; Ii
~
1;;
.....~
200 30 Cl
Cl

..
c:
';;
E
.
e
';;
E
...
Ql

<{
150
~
Ql

20

100
104 105 106 107 108 109
Stress cycles

S-N curves (scatter band) for copper alloys C87500 and

C87800 (silicon brasses) tested at room temperature.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 416
 13-8. Copper Alloy C92200 (Navy "M" Bronze): S-N Curves; 403
Scatter Band

200
'"
0- 'iii
::i: 25 .>t

...~..
Ii 150
~ 20
Cl

s
Cl
sc
.....
100 15
'" '"
...'"E 10 Ci:'"
Ci: 50
104 105 106 107 108 109
Number of stress cycles

S-N curves (scatter band) for copper alloy C92200 (Navy

"M" Bronze, or steam bronze) tested at room temper-
ature.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, p 421
404 13-9. Copper Alloy C93700 (High-Leaded Tin Bronze): S-N Curves;
Scatter Band

180
25
.
0-
~
160

't!'
eCl
140 20

.
.g
E
120
l!l
<i: ~_----=l15
100

80
104 105
Number of stress cycles

S-N curves (scatter band) for copper alloy C93700 (high-

leaded tin bronze) tested at room temperature.

Source: Metals Handbook•. 9th Edition. Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH. 1979, P 426
 13-10. Copper Alloy No. 192: Effect of Salt Spray on Tubes 405

liAs Received" Cycles to Failure Cycles to Failure After 180-0ays Exposure to Salt Sorcv
Tube Oia. 3/16-io. 1/4-io. 5/16-io. 3/8-io. Tube Die, 3/16-io. 1/4-io. 5/16-io. 3/8-io.
I
I----+-
I---J---- 0
-o
0
i 0 0 0
-r---- -~
10
0

I
:0 0
10 0 ~ b a
0
0
'Q, 0
- 0

..!
uc-,
U

.....
~

-
0

N
~

Results of fatigue tests on copper alloy tubes before and after salt spray exposure.

The tubes made from the copper alloy failed in the range of
105 to 106 cycles. After exposure for 180 days to salt spray, the
fatigue performance level was not lowered (see plot above).
Brazed steel tubes, prior to salt exposure, failed in the same
test in the range of 105 to 107 cycles. After 30-days exposure to
salt spray, the resistance to fatigue was 105 to 106 cycles. After
90-days exposure, the steel tubes showed no fatigue strength in
this particular test.

Source: Donald K. Miner, "An Effective Solution to the Problem of Hydraulic Brake Line Corrosion," in Source Book on Copper
and Copper Alloys, American Society for Metals. Metals Park OH, 1979, p 356
406 13-11. Copper Alloy 955: Goodman-Type Diagram

1.5 UTS: 152025 ksi

150 1----+----+----+-----t----2F------i

120

UTS: 101.5 ksi

~
0
°ili 90
iii
°ili .....
""
~ 60

30 I-:JIC---I ..,.<........,h~---fI+_-I+_---1t+_-H__-I::l=-_H-+++-_I_If--I

~
o

°i
.3 G _----'
30 '--_ ---'- -'- -L. --'-_ _- - - '

o 30 60 90 120 150 180

Goodman-type diagram (after Creech) for an-

nealed copper alloy 955.

Two variable-speed, flat-plate testing machines of the fixed-

deflection type were used for the test work. These machines have
a speed range of 750-2000 cpm with a maximum deflection of I
in. The yield, ultimate tensile strength and 1.5times the ultimate
are plotted in the graph above. The fatigue limit at zero mean
stress was determined and was found to be 22.0 ksi.

Source: J. M. Cieslewicz. "A Modified Goodman Diagram to Predict the Fatigue Limits of Copper Alloy 955." in Source Book on
Copper and Copper Alloys. American Society for Metals. Metals Park OH, 1979, p 40
 14-1. Magnesium Casting Alloy QE22A-T6: Effects of Notches and 407
Testing Temperature

If.
:2
150

125
-.
\. <, /Unnotched
20

1 ~
+1 100 +1

~
t: 75 ~ U·nolched
/(K t ~ 2)
1
~.
E
::J
E
"",-- V·notched
E
::J
E
..
'x
:2
50
............... /IKt = 31 'x
~
25

Cycles of stress
150

125
<, - 20
If.
<; I
/20 oC (68°FI- 16 ~

-
:2
+1 100 .,
~. 75 <,...... I
i!
200°C 1392°F)_ 12 :;;
E
::J ~25~ °c (482 of) E
::J
E 50 E
'x 'x
~ ~
25

Cycles of stress

S-N curves for magnesium alloy sand castings, QE22A-

T6, showing effects of type of notches (upper graph) and
testing temperature (lower graph). Rotating-beam (Woh-
ler) tests. Machine speed was 2960 Hz.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 589
408 14-2. Magnesium Casting Alloy QH21A-T6: S-N Curves; Effects of
Notches and Testing Temperature

175 25
20 °c (68 IF)
\
150
-, r-, 20

rr.:::;; 125
... ............... '-- Unnolched ]
...
"- 15
g
--
~ 100
<,
1;;
E
::l 75
I'-... U·nolched K t=21-
E
::l
E 10 E
'K .
:::;;
.
'K
:::;;
50

25

106
Cycles of stress

175 25
I
Unnolch.d
150
\ 20
125
rr.:::;;
'I\. "'--- 20°C (68 OF)
15
]
01

g
+~ 100

~
~ 75
-, ~50 °c 1480 OF) 10
E
::l
E
E
I"--. 'K
.
~
:::;;
50

25

Cycles of stress

S-N curves for magnesium alloy castings, QH21A-T6,

showing effects of notches (upper graph) and testing
temperature (lower graph). Rotating-beam (Wohler)
tests; machine speed 2960 Hz.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 590
 14-3. Mg-AI-Zn Casting Alloys: Effects of Surface Conditions on 409
Fatigue Properties

Meanstress. ksi
Meanstress,ksi
20

±20 ±20

.16 .16
~
Ii
e
:t12 ~
-l---I----l ."
~

4-----11-----l
!II
.
!i
~
ell

t---+--+-----j---t---+--....=j±4 f---t---+---+----'=-+--+--, '4

100 125 125

Mean stress,MPa
M&an stress,MPa

Meanstress,ksl Mean stress, ksi

o 10 15 20 O 10 20
.150,------,,--,-----,,---y-,-----...---,---r-. '150 r ---,----i---.---'T...---.:n--r--T--.

±20 ±20
Cast plates
'1251---+---+-- 107 cycles. loaded in bending
.16 "6 Xi
~-
.12 tl2 '"
.~
---f----f----;
.§
!II ~
--+--1---;

I---t---+----f---""'---+-...:....-=J '4 1-----J-=...."""-+-----1F"'-...;...f---+--..., ±4

125
Mean stress,MPa
Mean stress,MPa

Effect of type of surface on fatigue properties of cast Mg-AI-Zn alloys.

Machining improves fatigue properties of castings, as shown in the above

curves. Small radii, notches or fretting corrosion are more likely to reduce
fatigue life than are minor variations in composition or heat treatment.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 532
410 15-1. Molybdenum: Fatigue Limit Ratio vs Temperature

0.9

0.8

~
0.7

! ~r\
0
';:;
a::'"
0.6

0.5
\
0.4
o 100 200 300 400 500 600
Temperature,OC

Ratio of the fatigue limit of molybdenum at various

temperatures to its tensile strength at the same tempera-
ture.

Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH, 1979, P 774
 16-1. Tin-Lead Soldering Alloy: S-N Data for Soldered Joints 411

5000
I
I'~
~
4000

,
...... .......
~~ :--.... ........
A ~ :--.... ~
'iii 3000
..... 'Speed of testing I -
......
--....r- .... ~
I~

a. .... ~/min I I
ui <,
'"
~
in
"'r-..... ~ f4.. ...... . . 'N 2
2000
-- r- --..........., ~~
.........
':).04
..... ...........
0.2
1000
-
a 4 6 8 2 4 6 8 2
2 2 10 3
1 102
Cycles to failure

The fatigue strength of soldered joints is a complex and diffi-

cult subject to examine. Because solder alloys are strain-rate
sensitive and have large elongation capabilities, the perfor-
mance of fatigue tests under constant stress causes progressive
and rapid relaxation of the joint, and conversely, tests under
constant strain do not reflect a practical application situation.
The influence of the rate of stress cycling in terms of rate of
straining on the fatigue life of copper soldered joints with
60%Sn-40%Pb alloy is presented in the above graph.

Source: Metals Handbook, 9th Edition. Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 1095
412 16-2. Babbitt: Variation of Bearing Life With Babbitt Thickness

Babbitt thickness, in.

0.030 0.040
o Bimetal
I
• Trimetal
~
OJ
:::l 200

......
Cl
',J

OJ

..
.~
Q;
100
a:

0.25 0.50 0.75 1.00

Babbitt thickness, mm

Variation of bearing life with babbitt thickness for

lead or tin babbitt bearings. Bearing load was 14
MPa (2000 psi) for all tests.

One of the most useful concepts in bearing-material design

came with the recognition that the effective load capacities and
fatigue strengths of lead and tin alloys were sharply increased
when these alloys were used as thin layers intimately bonded to
strong bearing backs of bronze or steel. Use is made of this prin-
ciple (see graph above), in both two-layer and three-layer con-
structions, in which the surface layer is composed of a lead or tin
alloy, usually no more than 0.13 mm (0.005 in.) thick.

Source: Metals Handbook, 9th Edition, Volume 3. Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society [or Metals, Metals Park OH, 1980, P 806
 16-3. SAE 12 Bearing Alloy: Effect of Temperature on Fatigue Life 413

Bearing temperature, ° F
200 300
300 r---,c-r.----~-----,

~ 200 I------'\-t---+------t
en
.;;
.:?
'"
>
.;;
'" 100 I----~r_--t-----f
Gi
II:

OL-_ _L--_ _L - - _ - - - - l
50 100 150 200
Bearing temperature, °c
Varlanon of bearing life with temperature for
SAE12 bimetal bearings.

The alloy lining was 0.05 to 0.13 mm (0.002 to 0.005 in.) thick,
on steel backing. Bearing load: 14 MPa (2000 psi). As indicated,
operating temperature markedly influences fatigue life.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 813
414 17-1. Unalloyed Titanium, Grade 3: S-N Curves for Annealed vs Cold
Rolled

500
6, r-;
- 70
- 60
400
"'
.~~
a.. .;;;
:2
'" 300
Cold rolled
.-. - 50 .><
::i
~'"
'"
E
J
.§ 200
)(

"'
:2
0
Annealed

- ~.... - 40 ~
J
30 E
'"
E

'x
20 :2"'
100
- 10

Number of stress cycles

SON curve for unalloyed grade 3 titanium. Data were obtained by rotating-beam test-
ing at room temperature.

Source: Metals Handbook, 9th Edition. Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 376
 17-2. Unalloyed Titanium, Grade 4: S-N Curves for Three Testing 415
Ternperatu res

600
- 80

'"
Q.
~

'"
'"
~
E
::>
400
500

300
-----
---
_40°C (-40 ° F)

20°C (6SoF)
-
- 60
- 50
70 .;;;
..><
:i
~
E
::>

.Sx I - 40 .Sx
315°C (600 of)
'"
~
~'"
200 30

- 20
100
104
Number of stress cycles

S-N curves for unalloyed titanium, grade 4, at subzero, room, and elevated tempera-
tures. Data were obtained by rotating-beam testing of unnotched, polished specimens
machined from annealed bar stock.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH. 1980.0378
416 17-3. Ti-24V and Ti-32V: Stress Amplitude vs Cycles to Failure

-CI
a.. STRESS CONTROLLED
::lE
600
ILl
0
::>
I-
:J Ti-24%V, A.Q.
a..
::lE
<l:
400
lI)
lI)
ILl
It:
l-
ll)

200
3
10
CYCLES TO FAILURE

Stress-life curves of two Ti-V alloys that undergo cyclic hardening (Ti- 24%V) and
cyclic softening (Ti-32%V).

Cyclic-response curves indicate that the Ti-24%V alloy undergoes extensive cyclic
hardening, whereas Ti-32%V undergoes cyclic softening, as indicated above. Harden-
ing is caused by incomplete reversibility of twinning.

Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH. 1979, p 236
 17-4. Ti-5AI-2.5Sn: Effects of Notches and Types of Surface Finish 417

800
a-
... 100 .;;;
~l>
::2: 600 ~o Shot peened .>I.

~ 80
e ~
~
rt:- e~
t: 400 60 t:
E Hand finished Ground - E
.§
:I
200
40 E
';(
)(
... 20 ~
::2:
o
8
10
Lifetime, stress cycles

800
a-
... Ultrasonic machined
100 ...
.>I.
::2: Slab milled
l:/"
600
80
......
e
t: 400
Chem milled and annealed
60 ~
Chem mille E
E :I
:I
E 200
40 E
';(
20 ::2:...
';(
...
::2:
0 0
104 105 106 107 108
,:; .. Lifetime, stress cycles

800
...
a-
::2:
100 ]
....... 600 ~are - 80
e notch (K r = 2.4)
t: 400 60 E
~~ 16
E ~
a
:I
,§ l>Al>A A ~ - 40
'x
:I
E
...
)(

::2:
200
V-notch (K r = 3.2) ~l>l> - 20 ::2:...
'I
o
108
Lifetime, stress cycles

SON curves for annealed titanium alloy Ti-5AI-2.5Sn (rotating-beam

tests). Top and center graphs show fatigue strength for different types of
surface finish. Bottom graph shows fatigue strength as affected by type
of notch.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Melals, American Society for Metals, Metals Park OH, 1980, p 382
418 17-5. Ti-SAI-2.SSn and Ti-6AI-4V: Fatigue Crack Growth Rates

Stress intensity factor range, /iK, ksi Viii.

5 10

10-4

Ti-6AI-4V
NI24 to -269 0C - - ---f
c:l 10-31---- (75 to - 452 of) ----=",...j~t;f--+_--__l
~
0C(75 0F)
{24
E ELI -196°C (-32°F)
E -269°C (-452 of)
~'
~
i
~
NI
~ ELI
e
Cl 24 to -269°C
~
u (75 to -452 of)
b 10-41-----+--~~--I~----4-----=1
OJ

.,
:J
.Cl
'"
u.

Stress intensity factor range, s«, MPa .,;m

Fatigue crack growth rates for Ti-5AI-2.5Sn and Ti-6AI-4V.

Data on fatigue crack growth rates for Ti-5AI-2.5Sn and Ti-

6AI-4V alloys are plotted above. These data indicate that the
exposure temperature has no effect on the fatigue crack growth
rates for Ti-5AI-2.5Sn and Ti-6AI-4V(NI). However, over part
of the 11K range, the fatigue crack growth rates for Ti-6Al-
4V(ELI) are higher at cryogenic temperatures than at room
temperature at the same 11Kvalues.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 765
 17-6. Ti-6AI-6V-2Sn: Effects of Machining and Grinding 419

Ti-bAI-bV-lSn (STA, 42 Rcl

Surface
R aug h ne s s , AA

SURFACE Gentle I b5 43
GRIND Cony. I lO 44
Abusive 20 70

SURFACE Gentle )83 43

GRIND + PEEN Ah ... ,v. 150 55

HAND Gentle J67 14

GRIND Abusive
67 14

END MILL· Gentle 73 28

PERIPHERAL CUT Abusive I 45 39

ECM
I", 12 II
Off- Sta nda rd 147 145

ECM +
PEEN
'". 85 48
iOff-S'.n""." ~5 120
I I I I I
o 20 40 60 80

ENDURANCE LIMIT, KSI

Bar chart presentation showing effects of various machining and grinding operations
on fatigue characteristics of titanium alloy Ti-6AI-6V-2Sn.

Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 355
420 17-7. Ti-6AI-6V-2Sn (HIP): S-N Curves for Titanium Alloy Powder
Consolidated by HIP

iii 120
ANNEALED PLATE
"
vi
en
(MINI
w
a:
l-
ll) 10('
:2
:>
:2
X
<{
:2 80

60 -'-- ---, "T" "T'" --,

5
10
NUM8ER OF CYCLES

OJ HIP RUN 1. AS MACHINED (SPEC W "M" SU8SCRIPT WAS EVALUATED 8Y METALLOGRAPHY)

mHIP RUN 2. VAC ANN AT 1300·FI2 HR
rn HIP RUN 2. VAC ANN AT 1300·FI16 HR
o HIP RUN 2. VAC ANN AT 1300·FI24 HR
iii HIP RUN 4. VAC ANN AT 1300"FI24 HR

S-N curves showing endurance limits for titanium alloy powder consolidated by HIP at
900°C (1650 OF).

Note that most data points obtained in this phase fell within the representative data
band for annealed forgings. In the specimen designated with an "M"subscript, low fatigue
endurance was apparently associated with failure initiation at an inclusion, This shows
that a clean powder is required for parts that are fatigue-critical and must operate with the
equivalent of fully forged properties.

Source: R. H. Witt and W. T. Highberger, "Experience With Net-Shape Processes for Titanium Alloys," in Production to Near Net
Shape: Source Book. C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals. Metals Park OH. 1982. P 277
 17-8. Ti-6AI-6V-2Sn (HIP): S-N Curves for Annealed Plate vs HIP 421

ANN PLATf'
120 . / eeo F-16KeI-2 /lR
\ / 22-5'0~-fOl<'Sl-I/-IR
tOO
\
\ <,
80
tgTReeS,
KSI €JO '7------:
RAfIIG[ O~ PROGRAM l?ATA
40
20
~4 ~~ ~~ ~7
CVClP8 TO {"AILURf"
S-N curves showing that HIP Ti-6AI-6V-2Sn is equivalent to annealed plate of
the same composition.

Source: W. Theodore Highberger. "Manufacture of Titanium Components by Hot Isostatic Pressing," in Production to Near Net
Shape: Source Book. C. J. Van Tyne and B. Avitzur, Eds.. American Society for Metals. Metals Park OH, 1983, P 304
422 17-9. Ti-6AI-2Sn-4Zr-2Mo: Bar Chart Presentation on Effects of
Machining and Grinding

Ti-bAl-lSn-4/'.r-lMo (STA, 36 Re!

Surface
Roughness, AA

SU RFACE I-l'G~e.:.:n~tl~e_....,..".... -, 68 39
GRIND Cony
I I 17 41
~IO lZO

END MILL- I-l'G~en~ltI.:i-e "T"" '" 8Z 36

A"'lb~u~.,:.:·v~e
PERIPHERAL CUT ... ...1 47 77

I I I I I
o ZO 40 60 80

ENDURANCE LIMIT, KSI

Bar chart presentation showing the effects of specific machining and grinding opera-
tions on fatigue characteristics of titanium alloy Ti-6AI-2Sn-4Zr-2Mo.

Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 355
 17-10. Ti-6AI-2Sn-4Zr-2Mo: Constant-Life Fatigue Diagram 423

Minimumstress. ksi
-100 -50 0 +50 100 150
1200

150
1000

800
.;;;
.><
If.
:;;
~
t:
10D ~
600 t:
E E
:>
:>
E E
'x 'x
~ 400 ~
50
10 7 cvcles lifetime
200

0 0
-800 -600 --400 -200 +200 400 600 800 1000 1200
Minimumstress, MPa

Constant-life fatigue diagram for duplex annealed Ti-6AI-2Sn-4Zr-2Mo sheet, 1 mm

(0.04 ln.) thick.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 385
424 17-11. Ti-6AI-2Sn-4Zr-6Mo: Low-Cycle Axial Fatigue Curves

1600
220
'"
Q.
:2 1400
--- STA
'"
.><
200 ~.
~.

~ 1200
180 ~
'"
E E
::J ::J
E 160 E
'x 1000
'x
'"
:2 140 '"
:2

800 120
1 103 105
Number of cycles

Low-cycle axial fatigue curves for Ti-6AI-2Sn-4Zr-6Mo. STA

(solution treated and aged) condition: 1 hat 870 ° C (1600 OF),
water quench, age 8 h at 595°C (1100 OF) and air cool. DA
(duplex annealed) condition: 15 min at 870°C, air cool, then 8h
at 540 ° C (1000 ° F) and air cool. All fatigue tests conducted at a
stress ratio of R = 0.1. Open symbols indicate fatigue tests; solid
symbols, tension tests.

Source: Metals Handbook, 9th Edition, Volume 3. Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals. American Society for Metals, Metals Park OH, 1980, P 395
 17-12. Ti-8Mo-2Fe-3AI: S-N Curves; Solution Treated and Aged 425
Condition

600
I 80 -
Tension-tension tests
, -
500 R = 0,25; K r = 3.5
'"
e,
:::E 400
'. 60
'in
~

~
'" '"
'"
'"
~ ~
'" 300 E
•
E
:>
E
'x 200
I'---- .~
,....
- 40 :>
E
'x
'"
:::E'" :::E
- 20
100

Stress cycles

SoN curve for Ti-8Mo-2Fe-3AI titanium alloy in the solution treated

and aged condition. Data are for 1.S-mm (0.060-in.) thick sheet solution
treated 10 min at 790 0 C (14S0 0 F), air cooled, and aged 8 h at 480 0 C
(900 OF).

Source: Metals Handbook. 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 403
426 17-13. Ti-10V-2Fe-3AI: S-N Curves; Notched vs Unnotched
Specimens in Axial Fatigue

1100

1000 ~o
I
Unnotched
~ 140
~
'" 900 r--- 'u;

.---- --
Q. ~
.:,t.
:!:
~ lili 120 ",'

800 '"
...e
'" ~
A ~
'"
E E
:::> 700 100 :::>
.§
)(

'" 600
:!:
<, E
'x
'"
:!:

~ to-.. 80
500

----
Stress cycles
I--

500

400
\~ I
Notched
60
'"
Q. Kr = 3.0 'u;
.:,t.
:!: \0
'"
'" 300
'"
~
E 200
:::>
r-, 40
'"
~
E
:::>
E o RT ~ E
'x ~2050C (400°F)
~ 0 20 'x
'" 100 '"
:!:
:!: o 425 °c (800 of)
I
Stress cycles

Axial fatigue of Ti-l0V-2Fe-3AI bar stock in the STOA (solution

treated and overaged) condition. Specimens were taken from round
bars 7Smm (3 in.) in diameter that had been solution treated 1 h at 760° C
(1400 OF), furnace cooled, overaged 8 h at 565°C (1050 OF) and air
cooled. Tests were conducted at a stress ratio ofR= 0.1 and a frequency
of 20 Hz. Top: results of unnotched bars tested at room temperature.
Bottom: fatigue characteristics of notched specimens tested at elevated
temperature.

Source: Metals Handbook, 9th Edition.Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 399
 17-14. Ti-10V-2Fe-3AI and Ti-6AI-4V: Comparison of Fatigue 427
Crack Growth Rates

do/d~ in/cycle (pm/cycle)

10-4-r-----r---...."7T""-------,
(25)
MA Ti'-6A/-4V

iI
i Ti"-IOV2Fe-3A/
i
10-5
(025)
i
•

70.6
O--.....L.L---
(0025)/i'... J,
...J
20 O'---4..J.O - - - - - - - - '
(II) (22) (33) (44)
LlK, ksi vm. (MPoW;;)

Comparison of fatigue crack growth rates. Data for Ti-

10V-2Fe-3AI,R = 0.05,F= 1 to 30 Hz; for MA Ti-6AI-
4V,R= 0.08,F= 1 to 25 Hz; for RA Ti-6AI-4V,R= 0.08,
F= 6Hz.

Fatigue crack growth rates in air have been found to lie in the
scatter band for mill annealed (MA) Ti-6AI-4V, as shown
above. At high ~Kvalues, Ti-IOV-2Fe-3Al approaches the per-
formance of Ti-6AI-4V in the recrystallized annealed (RA)
condition.

Source: Wayne A. Reinsch and Harry W. Rosenberg, "Three Recent Developments in Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 375
428 17-15. Ti-1 OV-2Fe-3AI: S-N Curve; Notched Bar Fatigue Life for a
Series of Forgings Compared With Ti-6AI-4V Plate

rrmox,IOJpsi(MPo)
70-r-----~----------____..,
(48J)

60
(414)

50
(345)

40 STA Ti·6AI-4V plate

(216)

30
(201)/i.l-O-4----------'-----"':::OO"------.J

Cycles tofaIlure
Comparison of notched fatigue lives for Ti-lOV-2Fe-3AI forg-
ings and Ti-6AI-4V plate. Data for Ti-lOV-2Fe-3AI,R = O.OS,F
=KT= 2.9; for STA Ti-6AI-4V plate,R = O.l,K T= 3.

Fatigue characteristics of Ti-lOV-2Fe-3Al are equal to or su-

perior to those of Ti-6AI-4V. Notched fatigue results are shown
above. Data from a series of die forgings have shown that the
mean value for fracture toughness is 49.1 ksi ~. (54 MPa
Jill), with a standard deviation of 2.3 ksi~. (2.5 MPa Jill).
K/scc in 3.5% NaCl is typically about 90% of the x;

Source: Wayne A. Reinsch and Harry W. Rosenberg, "Three Recent Developments in Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 375
 17-16. Ti-13V-11 Cr-3AI: Constant-Life Fatigue Diagrams 429

Minimum stress, MPa

-100 -80 -60 -40 -20

40 +20
60 80 100 120 140 160 180 200
1400 r-_,_----,~-_,_-_,___,...---.-_r-_r""__,-__,-~""__.--t-er____,"'_,,,._-r____,, 200

180

160

140

120 ~
~~
:;;
E 100 E
"~" 600 ~
80 "K
~ ~
At room
400 temperature 60

40
200
- - Unnotched
- - Edgenotched, Kr = 3.0 20

-600 -400 -200 +200 400 600 800 1000 1200

Minimum stress,ksi

Minimum stress,ksl

1400 -100 -80 -60 -40 -20 +20 40 60 80 100 120 140 160

180

160

140

~ 0;;;
:;; 120 ""
• 800
~~
~E 100 E

~ 600 "E
"K 80 o~
~ :;;

400 60
Al315 -c (600 OF)
40
200
- - Unnotched
- - Edgenotched, Kr=3.0 20

-400 -200 +200 400 600 800 1000 1200

Minimum stress, MPa

Constant-life fatigue diagrams for Ti-13V-llCr-3AI, STA (solution treated and aged) condition, longitudi-
nal orientation. Data arefor axial fatigue of edge-polished sheet specimens of material solution treated and
aged to room-temperature tensile strength of 1203 MPa (174.5 ksi), Corresponding yield strength was 1080
MPa (156.7 ksi); at 315°C (600 OF), the tensile strength was 1078 MPa (156.3 ksi) and the yield strength was
876 MPa (127.0 ksi), Tests were conducted at a speed of 60 Hz.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels. Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, P 401
430 17-17. Ti-6AI-4V: Effect of Condition and Notches on Fatigue
Characteristics

800 .a...- ~ Smooth bar, STA stock

co
e, -r:"---~ 100 ';;;

g
::E 600 a.. Smooth bar, annealedrock ~
--...:
::i ~ 80
~on 400
E
:::l
E
'j( 200
co
::E
- -- Notched bar (K t = 3.5). STA stock
"" Notched bar. annealed stock
60
40
on

106
Number of cycles

S-N curves for titanium alloy Ti-6AI-4V (rotating beam) showing effects
of STA (solution treated and aged) versus annealed conditions, and ef-
fect of notches.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, p 389
 17-18. Ti-6AI-4V: Effect of Direction on Endurance 431

700

650
m
a..
~
CIl
CIl
Q)
;:,
CIl
0)
c: 550
.~

'"c:
L.
2:! 500
~

450

106
Endurance (cycles)

Rotating-cantilever fatigue (S-N) curves for three testing directions in

57 mm thick, forged and annealed Ti-6AI-4V bar.

These curves show that fatigue properties are lowest in the long
transverse direction. This result has been attributed to the fact that
Poisson's ratios are also sensitive to crystal orientation, these ratios
being higher in the longitudinal and short transverse directions be-
cause stressing occurs parallel to the basal planes. Higher ratios imply
greater constraint, which means that the levels of strain will be reduced
and the fatigue strength enhanced in these two directions. The differ-
ences observed in fatigue strengths in the longitudinal and short trans-
verse directions have been attributed to relative changes in grain
shapes that also occur during processing.

Source: I. J. Polmear, Light Alloys. Edward Arnold Ltd, London, England, and American Society for Metals. Metals Park OH,
1981, p 193
432 17-19. Ti-6AI-4V: Effect of Isothermally Rolled vs Extruded Material
on Cycles to Failure

100 , - - - - - - - - - - - - - - - - - - - - - - - - - -

80

iii o!';.
'" 60
.,;
'"
w
....a:
'"
::;
:::l
~ 40
x AXIAL FATIGUE. RoO. I
« ROOM TEMPERATURE
::;
K,' 2.8 FLAT SPECIMEN WITH
0.050·IN. HOLE
20 o ISR TEE
6. EXTRUSION

OL-_ _'--_---JL...--l----JL.......I_ _---L_ _---L_-L.........-..L _ _......._ _......._ ...L.............L _ _-'-_ _...... _'--..L-J

103 105
CYCLES TO FAILURE

S-Nfatigue data for isothermally rolled tees versus extruded material. The notched fatigue behavior of the
ISR tees is as good as or slightly better than that of the extrusion.

Source: W. T. (Ted) Highberger, Govind R. Chanani and Gregory V. Scarich, "Advanced Titanium Metallic Materials and Pro-
cesses for Application to Naval Aircraft Structures," in Production to Near Net Shape: Source Book, C. J. Van Tyne and B. Avitzur,
Eds., American Society for Metals, Metals Park OH, 1983, P 124
 17-20. Ti-6AI-4V: Comparison of Wrought vs Isostatically Pressed 433
Material for Cycles to Failure

1200

I 1111
w'"".""~
150
co 1000 grade Ti-6AI-4V ~

~ ~
o,
~ I ~
""" '"
'" ~~ ~ ~ ~ ~
<I>
<I> 125 ~
~
~ 1il
U; 800
E
~ ~~
~~
E
~~
:J

~ ~"
:J ~
E
"~ 100

~ ~ t:S ~"
"xco
X
co
~ 600
~~ ~ ~ ~ ~i'- ~

~
• Engine mount supports
o Witness blocks I I ~~ :\I'ci ~ :\"
~ ~t'- ~ ~ ~ ~" 75
1lJ1I Radius failures
400
10' 10' 105 10' 10'
Cycles to failure

900
I. I I I I 125
o Engme mount supports

co 700
~ ~~ ~
100

~ ~" ~
n, Wrought "iii

~ ~ ~ -<: ~
~
~
-""
standard
~
e'"
C'\ e'"
<I>
<I> grade
1il 500
Ti-6AI-4V ~ 75 1il

~ ~ ~ ~\ ~ ~
~ C'\~~ ~ ~ r-;~ -:~" ~ ~ ~~
E

~ ~ ~ ~~ ~ ~~
E :J
:J
E E
"xco 'x
co
50
~
~ 300 r-
~ ~ 88 ~ ~~
~ r-,~ ~"
':':";lo,;

25
100
10' 10' 105 10' 10'
Cycles to failure

S-N curves for titanium alloy engine mount supports. Top: Data are for the standard wrought
grade;R = O.I,K,= 1.0, load controlled smooth specimens. Bottom: Data are for isostatically
pressed alloy powder, notched specimens;R = 0.1, K,= 3.

Source: Metals Handbook. 9th Edition, Volume 7. Powder Metallurgy, American Society for Metals. Metals Park OH. 1984.P 654
434 17-21. Ti-6AI-4V: Effect of Fretting and Temperature on Cycles to
Failure

~ • R.T
z • 200°C
~
<Jl
560 o 340°C
Fatigue (all test
<Jl
w
a:
In 420
~ ,om,,,":,,,,
o
z
~280
z
a:
w
~ 140
<l:

°1LO"':'~ ---..L.::~-----~------.L:;------'
1(f 10~ 10~
CYCLES TO FAILURE

Effect of simultaneous fretting and fatigue ofshot-peened Ti-6AI-4V at room tempera-

ture, 200, and 340 °C at a mean tensile stress of 140 MNm·2•

Room temperature fretting was found to have little effect on the fatigue strength
at 107 cycles. Fretting at 200 °C lowered the fatigue strength by approximately IS%;
furthermore, the fretting fatigue life in the overstress region (70 MNm- 2 above the
nonfretted run out stress) was lowered by two orders of magnitude compared with
results in the absence of fretting. At 340 °C, fretting similarly reduced specimen life
7
at overstress levels; however, more importantly, the fretting fatigue strength at 10
cycles was reduced to approximately 40% of that found under room temperature
conditions.
The gross result of fretting normally is fatigue failure brought about by surface
damage in conjunction with normal or transient high stresses in a component. It
should be said at the outset that visual assessment of fretting mildness or severity is
inconclusive by itself, in that the presence of more or less fretting debris on a
microscopic examination is not necessarily relevant to the loss of surface integrity.
It may in fact be misleading and should not be relied upon for assessing the severity
offatigue life degradation. It is the stress state acting in concert with stress raisers
(e.g., pits, tears, cracks) which determines the actual fatigue propensity.

Source: Practical Observations of Fretting Fatigue Cracks. p 180

 17-22. Ti-6AI-4V (Beta Rolled): Effect of Finishing Operations on 435
Cycles to Failure

FATIGUE CHARACTERISTICS OF BETA ROLLED TITANIUM 6AI-4V, az R c

METAL REMOVAL PROCESSES: SURFACE GRINDING. PERIPHERAL END MILLING.
CHEMICAL MILLING
MODE: CANTILEVER BENDING. ZERO MEAN STRESS
TEMPERATURE: 75' F

ENDUR. SURF.

II CONDITION LIMIT FINISH

I<SI -;v;:-
80

~
U;
:<:
en ~I
'"
lol
0: -..; I
I I I I III GENTLE MILL
GENTLE GRIND
66
6l
41
35
f.< 60
'"
\J
Z
""'-- I
~
I III CHEM. MILL 51 zo
E=:
............. .... :0-
I
----
..;
Z
0: 40
lol
f.<
...l
I ABUSIVE MILL az 59
..; "-
r-, .....
zo
- ABUSIVE GRIND 13 65

10 5 10,6
II 107

CYCLES TO FAILURE

S-N curves for beta-rolled titanium alloy Ti-6AI-4V. Curves show the effects of the various
finishing operations on fatigue.

Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982,p 354
436 17-23. Ti-6AI-4V: Effect of Yield Strength on Stress-Life Behavior

700
Ti - 6AI-4V, 24h 500°C
-..
tJ 910 MPo o eooppm 02
a.. 0"0.2
:::E (J"02 990 MPo • 1900ppm 02
650 •
ILl
0
::>
I- 600
...J
Q..
•
:::E
« 550
UI
UI
ILl
a:
I- 500
UI

450 4
10

CYCLES TO FAILURE

Effect of yield strength on the stress-life behavior of two Ti-6AI-

4Valloys.

In order to establish microstructural effects on fatigue behav-

ior, comparisons should be made on materials having the same
yield stress, especially for stress-controlled tests. This is illus-
trated above where it is shown that two titanium samples having
different yield strengths have different stress-life behavior when
tested at 500°C (930 OF).

Source: Edgar A. Starke, Jr., and Gerd LOtjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society For Metals, Metals Park OH, 1979, P 237
 17-24. Ti-6AI-4V: Effect of Stress Relief on Cycles to Failure 437

60

II II
0-AS RECEIVED I
\ e- H T 400"r. 2 HRS.

.~
"'" 40
iii
II:
1'., \.~
o
~ 30 ~ -,
~
I

~
Ul

~
Iii
20
~

- T

0
kr - BlOKE IN GRIPS

CYCLES

Flexural fatigue tests of titanium sheet (123,000 T.S.).

Flexural tests of the sheet specimen were made at 1725 CPM. Results are indicated by the
open points in the above SoN diagram. The endurance limit was not reached at stresses as low as
20,000 psi.lfan estimated limit of 19,000 psi is chosen, the endurance ratio would be only 0.155,
a value considerably lower than for any other known metal or alloy. Most investigators have
obtained normal values around 0.3 in similar tests.
Several of the sheet fatigue specimens developed fatigue cracks away from the milled specimen
edges. The cracks did not appear to be associated with any visible surface imperfection. For these
reasons, it was assumed that the sample was abnormal, rather than the test procedure. Very
careful oil-powder and fluorescent powder tests, supplemented by metallographic examination,
failed to reveal any surface cracks, even when the sheet was flexed to open any incipient hairline
defects.
It was considered possible, though not probable, that residual stresses from cold rolling were
acting in a deleterious manner. If so, a moderate temperature stress relief might help. Brief
experiments soon disclosed that temperatures at least as high as 400 OF(205°C) did not lower
the hardness; in fact, the hardness may have increased very slightly. Knowing this, a set of sheet
fatigue specimens was stress relieved for two hours at 400 OF(205°C). The solid points in the
graph above represent the results obtained with these specimens. The endurance limit was not
altered significantly. A definite shift to the left in the upper portion of the curve was evident,
although the direction of shift was opposite to that, had the heat treatment released undesirable
stresses.

Source: Titanium Symposium, Office of Naval Research, p 97

438 17-25. Ti-6AI-4V: Interrelationship of Machining Practice and
Cutting Fluids on Cycles to Failure

lee.

..
~
--_. _--
E 7EUI . e __ ~-- _ C -- •••••
---==='~- ~_==
II -- --.~--
- ....
-

~~
---- --..-----... --= ----=
e.
.Iee .28e .60e I.ee 2.e8 6.00 le.ee
CYCLES TO FAILURE (IN MILLIONS)

Alternating stress vs cycles to failure in high cycle fatigue of machined titanium

surfaces using neutral, chlorinated, and sulfurized soluble cutting fluids. A =
abusive grinding; B = low stress grinding; C = end milling.

Influence of Chlorinated and Sulfurized Cutting Fluids

On High Cycle Fatigue Properties of Ti-6Al-4V
Machined Surfaces at 75°F

10 7 Cycle Fatigue Strength (ksi)

Sol. Cutting Low Stress Abusive
Fluids End Milling Grinding Grinding

Neutral 75 62.5 12.5

Chlorinated 65 57.5 12.5

Sulfurized 67.5 12.5

Source: V. A. Tipnis and J. D. Christopher, "Machinability Testing for Industry," in Machinability Testing and Utilization of
Machining Data, American Society for Metals, Metals Park OH, 1979, p 26
 17-26. Ti-6AI-4V: Relative Effects of Machining and Grinding 439
Operations on Endurance Limit

Tio 6AI-4V B ETA ROLLED, l2 R c

Surfac e
Roughness. AA

SURFACE Gentle I 62 35
GRIND ~13 65

HAND r;~ntl~ 157 80

GRIND Ab .. etve I 30 80

END MILL- Gentle 164 67

END CUT u ........ 177 84

END MILL- r;~ntl~ I 66 'II

PERIPHERAL CUT Ah ... lu~ I l2 59

CHEMICAL ~""n", ... 51 20

MILLING nu_~· 14 5 165

I I I I I

o 20 40 60 80

ENDURANCE LIMIT, KSI

Bar chart presentation showing relative effects of various machining and grinding op-
erations on fatigue characteristics of titanium alloy Ti-6AI-4V.

Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr .• Ed., American Society for Metals, Metals Park OH, 1982, P 354
440 17-27. Ti-6AI-4V: Effects of Various Metal Removal Operations on
Endurance Limit

Ti- 6Al-4V ANNEALED, 35 R c

Surface
Rouglme9s, AA

SURFACE GRIND Gentle I 67 14

ECM FRONTAL 160 Il

ECM TREPAN I 40 161

I J I I I
o lO 40 60 80

ENDURANCE LIMIT, KSI

Bar chart presentation showing effects of various metal removal operations on the fatigue
characteristics of titanium alloy Ti-6AI-4V.

Source: Norman Zlatin and Michael Field, "Procedures and Precautions in MachiningTitanium Alloys."in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 355
 17-28. Ti-6AI-4V: Effect of Texture on Fatigue Strength 441

EFFECTS OF TEXTURE ON SMOOTH FATIGUE LIFE

130

120

.
..
... 110
I

E
<,
.....
b lOa Load ~
axis II [loio la ........
90 Ti-6AI-4V Re-X Anneal
R =0.1
........................
........ ------
80 L-- .L..- .....L ---J

Nt - cycles

SON curves showing the effect of texture on the fatigue strength of Ti-
6AI-4V. Fatigue strength is greater when the stress axis coincides with
the direction of a high density of basal poles.

Source:J. C. Williams and E. A. Starke.Jr., "The Role of Therrnomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys," in Deformation, Processing, and Structure, George Krauss, Ed., American Society for Metals, Metals Park
OH, 1984, P 334
442 17-29. Ti-6AI-4V: Effect of Complex Texture on Cycles to Failure

Ti-6AI-4V Air

~
E
900 \
\
-[]o-
z \ -o~-
::;:
\
800
.,
-0
\
-,
~ <,
.-
a.
«
E
700
...... ....
-- - - - - - - -
II>
II>

'" 600
<fl
10 4 10 5 106 10 7
Cycles to Failure

S-N curves showing the effect of more complex texture on

fatigue strength of Ti-6AI-4V. These data show that a
mixed texture lowers the high-stress end oftheS-N curve
preferentially.

Source: J. C. Williams and E. A. Starke, Jr., "The Role of Thermomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys, "in Deformation, Processing, and Structure, George Krauss, Ed., American Society for Metals, Metals Park
OH, 1984, p 335
 17-30. Ti-6AI-4V: Effect of Texture and Environment on Cycles to 443
Failure

TI-6AI-4V AIr

-,
-,
<,
............ _-----

lOS
Cycles to Failure
Q.
a
E Ti-6AI-4V 31/2 % Nael Solution
..
<l:

~
900
\
IJ)
\
800 \
\
\
\
700 \
-,
-,
-,
<,
600 ~------

104 105 106 10 7

Cycles to failure
b

SoN curves for O! + ,8- processed Ti·6Al·4V, showing the

effects oftexture and environment on fatigue strength. (a)

Tested in air. (b) Tested in 3'12% NaC\, These data show
that testing in an aqueous 3'12% NaCI solution reduces
fatigue strength when the stress axis is along [0001].

Source:J. C. Williams and E. A. Starke, Jr.. "The Role of Thermomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys," in Deformation. Processing, and Structure, George Krauss, Ed .. American Society for Metals, Metals Park
OH, 1984, p 335
444 17-31. Ti-6AI-4V: Fatigue Crack Growth Rates

Kmin+lJ.KalK 1c .:»:"
1
..,z
.....
..,e
'"
o

10- 10 L-_ _...L<... _

log lJ.K-

Schematic plot showing characteristic shape offa-

tigue crack growth rate (doldN) versus cyclic-
stress-intensity (~) curves.

It can be seen that at higher growth rates there is a linear

portion of the curve. This linear portion was represented as tiK'"
by Paris and Erdogan and is now frequently referred to as the
Paris law regime of fatigue crack growth. Most structural mate-
rials show variations in near-threshold FCP rate and in tiK,,, but
fewer show significant variations in FCP rate in the Paris law
regime. In contrast, Ti alloys show significant variations in FCP
rate over the entire range. At the highest crack growth rates
shown above, the FCP rate curve bends upward. This is con-
trolled by fracture toughness. However, since crack growth
rates are uncontrollably rapid in this latter regime, it is of little
interest and will not be discussed further here. Moreover, since
the majority of the lifetime of a crack component is spent in the
low-FCP-rate regime, factors which control FCP at rates less
than ~ 10-6 tu] cycle are probably most important. These factors
include microstructure and texture.

Source: J. C. Williams and E. A. Starke, Jr., "The Role of Thermomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys," in Deformation, Processing, and Structure, George Krauss, Ed., American Society for Metals, Metals Park
OB, 1984, P 338
 17-32. Ti-6AI-4V: Fatigue Crack Growth Rates for ISR Tee, and 445
Extrusions

2
10- rr========::::::;,-----,
Ti-6AI-4V
R = O.l,ROOM TEMPERATURE 000
TL AND LT ORIENTATION o
• ISR TEE
W
...J
o EXTRUSION
U
>-
~
:r:
o
z
z
~
"tl
w·
I-
«
a:
:r:
~
a
a:
(9
~
o
«
a:
o
w
::J
(9
~
«
lL

o
o 0

-7 ':------:~--........-_:';:-.......-7:........~::_":~
10 10 20 40 60 80 100
STRESS INTENSITY FACTOR RANGE.~K. KSI viN.

Fatigue crack growth results for ISR tee, and extrusions.

Data for both TL and LT orientations are shown above,

along with data for the extrusion. In comparing individual re-
sults, no differences were seen between TL and LT. In the chart,
it can be seen that particularly at lower stress intensities the fa-
tigue crack rate for the (lSR) isothermally rolled tee is faster
than that for the extrusions. This is probably due to the extru-
sions being beta formed while the ISR tees are alpha-beta
formed.

Source: W. T. (Ted) Highberger, Govind R. Chanani and Gregory V. Scarich, "Advanced Titanium Metallic Materials and Pro-
cesses for Application to Naval Aircraft Structures," in Production to Near Net Shape: Source Book, C. J. Van Tyne and B. Avitzur,
Eds., American Society for Metals, Metals Park OH, 1983, P 124
446 17-33. Ti-6AI-4V: Fatigue Crack Growth Rates

~ 0
: : (INCHES/CYCLE) I.IE
0

~
~

_°11 -
,~ o
o

'§
o

10 100
dK (KSI {INCH)

0- HEAT 3 ENVIRONHENT - LOW-HUHIOITY AIR

b - HEAT 2 ORIENTATION - RW
0- HEAT 1 R FACTOR -to. 30
HEAT TREATHENT - RECRYSTALLI ZATI ON ANNEAL

Fatigue crack growth rates for three different heats of Ti-6AI-4V

titanium alloy.

The fatigue crack growth rate in the RW orientation for this

alloy, when recrystallization-annealed, behaved similarly with
decreasing oxygen and aluminum. The crack growth rate is
shown as a function of 6.K tested at an R factor of +0.30.

Source: M. J. Harrigan, M. P. Kaplan and A. W. Sommer, "Effect of Chemistry and Heat Treatment on the Fracture Properties of
Ti-6AI-4V Alloy," in Titanium and Titanium Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals,
Metals Park OH, 1982,P 65
 17-34. Ti-6AI-4V: Effect of Final Cooling on Fatigue Crack Growth 447
Rates

~K, MPa • -.jm

10 100
I
CONSTANT AMPLITUDE FATIGUE:
R = 0.1; 10 TO 20 HZ
MAXIMUM SCATTER OF ACTUAL
DATA POINTS FROM ANY CURVE
IS LESS THAN 40%
WQ WATER-QUENCHED
AC AIR-COOLED
FC FURNACE-COOLED

A'A MILL-ANNEALED (DATA IN- ~

CLUDED FOR COMPARISON)
FC 100
WQ I

10.3 k~
'" '!:/lAC
Y
o
>..
u
':- :/
.s 10- 4 I :'/
:/
Z 1I
~ :/
"U :,:"
10. 5 ,.":'
I
0.1

10 100
~K I ksi- Jin":
Effects of final cooling rate on fatigue crack growth rate in
duplex-annealed Ti-6AI-4V, I-in. plate, 1775 OF (968.3 0q, 1/2
h, air cooled; and 1450 OF (787.4 "C), I h, cooled as noted.

From the data presented above, it can be seen that air cooling, per se, produced
little or no change in the cyclic crack growth compared to the mill-annealed base
material. The slightly decreased crack growth rates above a IJ.K of 20 ksi (22 Jill.
MPa· VITi) are, more probably than not, the result of the higher fracture toughness
of the air-cooled material. However, both water quenching and furnace cooling
resulted in fatigue crack growth rates noticeably different from those measured for
the mill-annealed base material. As shown, furnace cooling had a consistently det-
rimental effect on the crack growth rate while water quenching produced $!eatly
increased crack growth rates above a stress-intensity range of 18 ksi yin. (20
MPa . ~). The accelerated growth rate above 18ksi \!Ill. (20 MPa . may be Jill)
attributed to the proximity of the maximum stress intensity to the critical value.
The critical stress-intensity value for water quenching was an exceptionally low 38
ksi .jll1. (42 MPa . Jill).

Source: R. E. Lewis, J. G. Bjeletich, T. M. Morton and F. A. Crossley, "Effect of Cooling Rate on Fracture Behavior of Mill-
Annealed Ti-6AI-4V," in Titanium and Titanium Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for
Metals, Metals Park OH, 1982, P 90
448 17-35. Ti-6AI-4V: Effect of Dwell Time on Fatigue Crack
Growth Rates

10- 4 ,.-- - - - - - - - - - ...,

~
u
>- 10- 5
u
....
I
Q>
a.
(
.s I
Q>
I
~
.J::. I
'i0 10- 6 I
0,
-"
I
u
m
U - Sinusoidal
loading
--Dwell at
10-7 maximum load
10 100
Stress intensity
factor range (MPa m 1/ 2 )

A phenomenon which may be unique to certain titanium al-

loys is the effect of dwell periods at maximum load on rates of
growth of fatigue cracks. This effect is shown schematically
here, and increases in the rate of crack growth of as much as 50
times may occur compared with results obtained in tests on the
same material subjected only to sinusoidal stress cycles. Dwell
effects are maximized in alloys containing substantial amounts
of the a-phase which have a preferred texture such that stressing
is normal to the basal planes, whereas they appear to be insignif-
icant if stressing occurs parallel to the basal planes of the a-
phase, or if the microstructure is homogeneous and fine grained.
Particular attention has been paid to a / {3 alloys, e.g., Ti-6AI-
4V, in which dwell effects have also been found to decrease with
increasing amounts of the {3-phase in the microstructure. In all
cases, dwell effects disappear when stressing occurs at tempera-
tures above 75°C (165 "Fj.and they are generally considered to
arise from the preferential diffusion of hydrogen, during the
dwell period, to regions of localized hydrostatic tension ahead
of an advancing crack. Such an accumulation of hydrogen
would tend to embrittle this region, and it has even been sug-
gested that brittle plates of TiH 2 may be formed.

Source: I. J. Polmear, Light Alloys, Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981, P 200
 17-36. Ti-6AI-4V: Fatigue Crack Growth Data 449

0
• Annealed 2 hours at 705 C. air-cooled after forging
0
• alP transus 1005 C
" Axial loading: smooth specimens. K = 1.0
t
r ~ I -. ", .' " ..

Ti-6AI-4V

-SA
Left: Fatigue crack growth rates for Ti-6AI-4V rolled
plates in the ,a-annealed (HA) and mill annealed (MA)
conditions. BA = 0.5 h 1038 0 C, air-cool to room temper-
ature. Tests conducted at 5 Hz using compact tension
specimens. Ratio of minimum to maximum load = 0.1.
Above: Branching of fatigue cracks within the Widman-
stiitten packets of the a-laths.
10 20 50 100
Stress-intensity factor range
LlK(MPa m1J2 )

Work on Ti-6Al-4V rolled plate has indicated that the superior fatigue perfor-
mance with the {3-annealed condition is associated with relatively slower rates of
crack propagation (above graph). This effect, in turn, is attributed to the slower
progress of cracks through the Widmanstatten microstructure, particularly at stress
intensities below a critical value at which desirable crack branching occurs within
packets ofthe a-laths.

Source: I. J. Polmear, Light Alloys, Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981, P 179
450 17-37. Ti-6AI-4V P/M: Comparison of HIP'd Material With
Alpha-Beta Forgings for Cycles to Failure

r:

70

~o
o ~OHIP o
------00------

Upper and lower limits of

40 ~-_-:--.. alpha-beta-processed forgings
as listed in AFML TR-73-301

30 LJ...Ll..lll_--.J_L-L.l.....cr::ti±==I==:±=~:::l.-'~
106 10
Cycles to failure

Notched fatigue strength of HIP'd P 1M Ti-6AI-4V compared with fa-

tigue strength ofalpha-beta processed forgings.K, = 3; Hz = 30;R=
0.1.

Source: J. H. Moll, V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications," in Powder Metallurgy-
Applications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 286
 17-38. Ti-6AI-4V P/M: Comparisons of HIP'd Material With 451
Annealed Plate for Cycles to Failure

140

120~
r
100
'GM"S'$, o
K'S/ 80

~ Q I/IPC'tCl.£
o 15S0i:-I-5 1t:.'6/-II-IR

40t •
a
''''50.~-1'51t:'8/- 3/-1J:?
22'50·~-I0/t:!9I-II-IR
ANN P!AT~
I
20 1----,-_ _--'--_ _---"I -'-:-_ _----',
1

10~ /0'- 10-5 /0· /0'

C~ESTO~(L~

SoN curve for HIP'd Ti-6AI-4V and annealed plate. According to the
above data, fatigue results for Ti-6AI-4V are within the required range
for plate properties from MIL-T-9046.

Source: W. Theodore Highberger, "Manufacture of Titanium Components by Hot Isostatic Pressing," in Production to Near Net
Shape: Source Book, C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals, Metals Park OH, 1983, P 304
452 17-39. Ti-6AI-4V P/M: Effect of Powder Mesh Size on
Fatigue Properties

70.3 (100)

56.2 (80)
'Vi
~
m 42.2 (60)
a..
:E
VI'
VI
28.1 (40)
.......
Q.l

en o As received powder (SM 772)

0-80 mesh screened powder (SM 768)
14.1 (20)

0
104 ·10,

70.3 (100 )

'u;
~

~
56.2 (80 )

42.2 (60 )
-:--
- r----
-
li.

-- ~li.
Pli.
Ali.
o~l
:E

....~'
en
28.1 (40 )
O As received powder (SM 772)
li. -80 mesh screened powder (SM 768)
14.1 (20 )

io, 10,

Top: Room-temperature properties. Bottom: Properties at 700 OF (370°C).

High-cycle fatigue (HCF) data were developed on Ti-6AI-4V (Std) by Wil-

liams International in a program to apply near-net-shape HIP technology to
a compressor rotor part for the F-I07 cruise missile engine. In this study, two
size fractions of powder were used: -35 mesh (as-received) and -80 mesh.
There was no difference in HCF test results between the two sizes. Room-
temperature and 700 OF (370°C) S/ N curves are shown above.

Source: J. H. Moll. V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications,"in Powder Metallurgy-
Applications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 286
 17-40. Ti-6AI-4V P/M: Comparison of Blended Elemental, 453
Prealloyed and Wrought Material for Effect on Cycles to Failure

1200

1000
140
co
Q..
'iii
~ .><

iii 800 iii

III III
e
li)
~
100 ti
E E
:J :J
E 600 E
'x 'x
co
co
~ ~

60
400

200
10' 10' 10' 10' 10' 10'
Cycles to failure

SON curves showing comparison of smooth axial fatigue behavior of Ti-6AI-4V blended ele-
mental and prealloyed P 1M compacts with wrought annealed material. Tested at room
temperature, R = O.

The fatigue behavior of titanium PIM compacts is compared to wrought products in

the graph above. The blended elemental material is inferior to prealloyed compacts and
II M materials. This is caused by residual chlorides and consequent porosity; also, chem-
cial heterogeneity may lead to areas of similarly aligned alpha plates. Blended elemental
compacts, however, compete well with many titanium alloy castings in fatigue strength.
Prealloyed powder compacts exhibit fatigue behavior equivalent to that of Ij M materials.
This situation is achieved by careful control of cleanliness (powder handling) and micro-
structure. Cleanliness depends on the environment in which the powder is produced, con-
ditions of subsequent handling, and microstruture developed by compaction. Cleanliness
dictates the amount of contamination contained in the final product; microstructure de-
termines the ability of the compact to accommodate foreign particles and resist crack
initiation.

Source: Metals Handbook. 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,P 753
454 17-41. Ti-6AI-4V: P/M Compacts vs 11M Specimens:
Cycles to Failure

co
Q.. '(ij
~ 1200 .>t:

'"
en
~
1000 140 '"
f/l

~f/l
1ii 800
E 100 E
::J
::J 600
E E
x 400 60 'x
CO
CO
~ 103 104 10 5
106 107
10 8 ~

Number of cycles to failure

Fatigue chart presentation showing a comparison of fa-

tigue behavior of Ti-6AI-4V compacts with ingot metal-
lurgy material.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, P 44
 17-42. Ti-6AI-4V: Comparison of Specimens Processed by Various 455
Fabrication Processes for Cycles to Failure

1200
Axial fatigue smooth 160
Room temperature
R = 0,1
ro Annealed
a.. 'ijj
~ 800 120 ..".

u> vi
III
III
~ ~
1;; 1;;
E 80 E
::J ::J
E E
'x 'x
ro 400 ro
~ ~
40

o"- -'- ...L...- --' ....L.. ---I0

3
10 108
Cycles to failure

SON curves (bands) for titanium alloy Ti-6AI-4V processed by various fabrication processes.
The inconsistentfatigue life ofthe hot isostatically pressed product is usually casued by inclu-
sions in the compact.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,P 439
456 17-43. Ti-6AI-4V: Comparison of Fatigue Crack Growth Rate,
P/M vs 11M

Stress intensity, ksiVTrl.

1 10 102
3
10
11M and P/M Ti-6AI-4V
at room
10-2
temperature
(laboratory air),
R = 0.1, at 5 to
30 Hz
10-4
Q)
Q) u
U 10- ~
3
>
~ E
c E
......ai
III
ai
i...
Recrystallization
.s::. 10- 5 s:
~ anneal ~
...001 ...0
.>0:
10- 4 ~
(J (J

...
III ...
III

o U

10- 6

10- 5

Stress intensity, MPaViTl

Comparison of fatigue crack growth rate of Ti·6AI·4V

P 1M compact with 11M material heat treated to various
conditions. The fatigue crack growth rate of blended ele-
mental and pre alloyed compacts is equivalent to 11M
material with the same microstructure.

Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, p 752
 17-44. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for 457
Cycles to Failure

130

~
TEST CONDITIONS:
CONSTANT AMPLITUDE, R ~.1
<, K ~ 1.0
T
110 <, STRESS RELIEVED, FLUSH WELD BEAD

0---,
0 <, c:
u; <, 0
>l

~
90
l!l'
w
a:
"""- CD
t-
~- ~
IJ)

:E
:::>
:E
x
« 70
:E
LEGEND
---/:::,. BASE METAL

50 --0 SSEB WELDEO

6
10

NUMBER OF CYCLES (LOG SCALE) TO FAILURE

S-N curve for titanium alloy plate-base metal versus SSEB-welded. Results show that the constant-
amplitude fatigue life of SSEB weldments in O.440-in.-thick plate equals that of the base metal.

Source: R. H. Will, J. G. Madora and H. P. Ellison, "Sliding-Seal Electron-Beam Welding of Titanium," in Source Book on
Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American SccietyIor Metals, Metals Park OH, 1981, P 87
458 17-45. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for
Cycles to Failure

120

TEST CONDITIONS:
CONSTANT AMPLITUDE, R ~.1
KT~ 10
1lX' STRESS RELIEVED, FLUSH WELD BEAD

80 o
o
~

LEGEND:

40 --- 0 SSEB WELDED

---- 6 BASE METAL

20

SON curves for titanium alloy plate-base metal versus SSEB-welded. Results show that the constant-
amplitude fatigue life of SSEB weldments in O.940-in.-thick plate equals that of the base metal.

Source: R. H. Witt, J. G. Madora and H. P. Ellison, "Sliding-Seal Electron-Beam Welding of Titanium," in Source Book on
Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, p 87
 17-46. Ti-6AI-4V EB Weldments: Base Metal Compared With 459
Flawless Weldments

140 0 0

o 0 O.OSQ-IN.·THICK BASE·METAL
Ti-6AI-4V STOA CURVES
120

ci o
100
s"
0 o

en o
~
en' o
en SO
o
w
a:
I-
en
«X
::E 60

40 ~
20
103 10 5

CYCLES

SoN curves for EB weldments that were flawless (lower two curves). Upper curve shows scatter band for base
metal (O.080·in.-thick Ti-6AI-4V STOA.).

Source: R. Witt, A. Flescher and O. Paul, "Weldability and Quality of Titanium Alloy Weldments," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982,P 313
460 17-47. Ti-6AI-4V EB Weldments: Effects of Porosity on
Cycles to Failure

140
a O.oaO·IN.·THICK
a a a aASE·METAL
Ti·6AI·4V STOA CURVES
a
120

d 100
II
a:
in
~

~' ao
w
a:
Ii;
X
<l: 60
:E

40

20
103 105
CYCLES

SON curves for ED-welded Ti-6AI-4V titanium alloy showing effects of porosity.

Above are shown experimental data obtained for porosity-containing ED welds which are
superimposed on a set of curves for the base material (0.080-in.-thick Ti-6AI-4V STOA sheet) at
various K, factors. For the points within the boundaries of the band, radiography indicated
scattered porosity (0.003 to 0.005 in. in diameter). For points below the lower boundary of the
band, radiography indicated either linear or heavily scattered porosity.

Source: R. Witt, A. Flescher and O. Paul. "Weldability and Quality of Titanium Alloy Weldments," in Titanium and Titanium
Alloys: Source Book Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 312
 17-48. Ti-6AI-4V Gas Metal-Arc Weldments: Effects of Porosity on 461
Cycles to Failure

140

0.250-IN.-THICK BASE-METAL
D D Ti-6AI-4V STOA CURVES
120

0 D
~ 100
Cii
~
iii
lI)
w BO
a:
l-
ll)

~
::E
60

D D

40 D o
A
o

105
CYCLES

SON curves for porosity-containing gas metal-arc welds. In the above graph the experimental fatigue
data for porosity-containing GMA W weldments are superimposed on SoN graphs for Ti-6AI-4V
STOA material (0.25 in. thick) for various K 1 factors.

Source: R. Witt, A. Flescher and O. Paul. "Weld ability and Quality of Titanium Alloy Weldments," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, p 313
462 17-49. Ti-6AI-4V: Unwelded vs Electron Beam Welded Material for
Cycles to Failure

III
0..
en
III
w
It:
t;
(!)
z
100
80
60
--- . -'" ..
'r.I
~ .. Q=

40
5z
It:
W
o HEAT TREATED TI-6AI-4V
S
C[
20
f· 7500 CPM
• HEAT TREATED
ELECTRON BEAM WELDED
Kr =1.0 a STRESS RELIEVED TI-6AI-4V
I
6
10
CYCLES TO FAILURE

Room temperature rotating-beam fatigue life of unwelded and electron-beam-welded Ti-6AI-4V titanium
alloy in fully heat-treated condition. Decrease in fatigue strength ofthe weldment relative to the parent metal
did not exceed 12%.

Source: S. M. Silverstein, V. Strautman and W. R. Freeman, "Application of Electron Beam Welding to Rotating Gas Turbine
Components," in Source Book on Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals,
Metals Park OH, 1981, P 169
 17-50, Ti-6AI-4V: S-N Diagram for Laser-Welded Sheet 463

120
Legend o 0.230 in, (0.584 cm) sheet
• 0.140 in. (0.356 cm) sheet
..... - Mean curve for plasma-arc welds 700 co.....
.... 100
$1!!~iJfift:
I
0
ci
II
of mean fatigue strength
of all un-welded control ....
)(
a:
..... '@specimens 600 N
'iii
.¥ 80 oco .0 0 ~ • -.....
E
c

.....
Ul
Ul
Gl
-, .....
l3Gl

----
Ul
E 60 <, 4 runouts
Ul

---
:;, E
E
')(
<, 400 :;,
E
IV
<, ')(
::iE 40 IV
Lower bound data for pIa welds J ---- ::iE
300
on flat sheet with filler
20
10 6 10 7 10 8
Cycles to failure
S-N diagram for laser-welded titanium alloy sheet,

The fatigue properties of welds as shown above indicate that under proper
welding conditions, laser welds can be made in Ti-6AI-4V which exhibit base
metal fatigue characteristics. The best laser weld failures initiated at sites in
the base metal, whereas other weld failures originated at undetected small
pores. Where failures initiated in the base metal, it was concluded that no
porosity or weld defects of sufficient size to preferentially initiate fatigue
fractures was present.

Source: E. M. Breinan, C. M. Banas and M. A. Greenfield. "Laser Welding-The Present State-of-the-Art," in Source Book on
Electron Beam and Laser Welding. Melvin M. Schwartz. Ed., American Society for Metals, Metals Park OH, 1981, P 289
464 17-51. Ti-6AI-4V (Cast): S-N Diagram for Notched Specimens

1000
900
o 0 0
800 t- o 120
oooeJO
.. 700 - '';;
Do .>t
0000y 00
:E 600 0 - 90 .n
~.

...'"
~ 500
0
00'17001:>.61:>.
- ...e'"'"
E ~f:>6'17 E
::J
::J 400 +000# 00
- 60 E
E
.
'j(
300
o 00+-tp.~66 '17 'Ii' '17
o 000 01:>.Q)
+0
..
'j(

:E
:E x ooo,+°oOOx O 0
+0 0 o-tCo 0 0
200 00
00 +o~ooo
+ 30
oc:P
100

No. of stress cycles

Notch fatigue strength of as-cast Ti-6AI-4V. Each symbol rep-

resents fatigue data from a different source. Stress ratio,R, typi-
cally was +0.1; stress concentration factor, K/, was mostly 3.0,
but a few tests were run at K/ = 1.0.

Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980,P 411
 18-1. Zirconium 702: Effects of Notches and Testing Temperature on 465
Cycles to Failure

'"0.....
N
e
U
-,
-"
Cl
to
en
en
w
a::
l-
en
0.5

CYCLES
S-N curves for zirconium grade 702, showing effects of notches
and elevated temperature (400°C, or 750 OF) on fatigue
characteristics.

As indicated above, zirconium and its alloys exhibit a fatigue

limit behavior similar to most ferrous alloys.

Source: Donald R. Knittel, "Zirconium." in Corrosion and Corrosion Protection Handbook, Philip A. Schweitzer, Ed., Marcel
Dekker, Inc.• New York NY, 1983. p 198
466 19-1. Steel Castings (General): Effect of Design and Welding
Practice on Fatigue Characteristics

'W:.... S-N CURVE -CYCLES OF

~
X "" STRESS VS COMPUTED
"" .... x STRESS
....
.... ~ x Xx
00 .....x
~~~
::f ....0.1:-:!-::
o ...............
x~x
x

10 6 107
NUMBER OF CYCLES

S-N curve for cast box designs.

u,
o x,oX, S-N CURVE -CYCLES OFSTRESS
~
(5.-
50 'o'x VS COMPUTED STRESS
Q.~ '-.""',
~ I 40 '"'ox
Vl
llJ lbX~',
[3 ~ 30 " X
11::1-
1-0 "'80x.. . XX x
VI Cl:
olE
llJ
20 0",,-- -_
--o-J!l<u_x)(""
-0_
....x....
0_ 0

~ 10 X - 2-BW HAND WELDED, STRESS RELIEVED--

o - 2 - BW MACH WELDED. STRESS RELIEVED-
-I
::::I
3
(3 105 106
NUMBER OF CYCLES

SoN curves of box weldments, comparing hand weldments with

machine weldments. All weldments were stress relieved at 1100 F Q

(593 Qq.

The S-N curves shown above indicate that: (1) the welding
practice is of no great importance; and (2) the cast steel box
design is superior to a weldment design.

Source: Steel Castings Handbook. 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 7-6
 19-2. Steel Castings (General): Effects of Discontinuities on 467
Fatigue Characteristics

1.0

0.9

0.6

0.7

0.6

0.5

:I: :I:
l- I-
e> ~ 0.4
Z
W W
AVERAGE YIELD STRESS RANGE
It: It:
l- I- TENSILE STRENGTH
lI) lI) 0.3

W W
~
.J
e>
~
~ 0.2
<l: W
u, I-

0.1

0
0.5 0.6
MEAN STRESS
- 0.1 TENSILE STRENGTH

- O. 2

- 0.3

-0.4

Goodman diagram for bending fatigue for normalized

and tempered 8630 cast steel. (Machined notch of R. R.
Moore specimen: 60° included angle, 0.0015-in. (0.0381-
mm) root radius.]

Surface condition has a significant effect on fatigue life and fatigue limit. A
highly polished smooth test specimen can exhibit twice the fatigue strength of
a rough machined sample. A good design approach is to use the notched
fatigue limit asa design value. For cast steels a O.OOIS-in. (O.0381-mm)root
radius circumferential notch in a rotating beam fatigue specimen reduces the
fatigue limit by about 0.7 of the unnotched value. This is sufficient to account
for variations in surface finish and minor surface discontinuities. The above
diagram shows that even severe surface discontinuities, not normally permit-
ted by workmanship standards, do not reduce the fatigue limit by much more
than the 0.7 value.
The above emphasis on surface discontinuities is due to the fact that sub-
surface discontinuities which do not have a crack-like sharpness and which
do not significantly reduce the load-bearing area of a component generally
have little effect on fatigue performance.

Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 7-6
468 20-1. Closed-Die Steel Forgings: Effect of Surface Condition on
Fatigue Limit

9oor--....,._~---------,r------------,----------~
Ground and polished 125
8oot-----.::>.,;~~<?' .......:::_II__-___7'''-------_+---------~
7oo~------~......,;~~-..-==-"""'-"""",------,,L----+----------d 100

6oo~--------___I--=~:::""....,--=_=-=""'l-"",""----------l

~ 500 ~--------___I---___;.,L_-=....,.-=.......; ::_t--"'-..=_-:=.""""=__--=l 75 ~

~ 4oot-----------1------------'''''f.,=---'''......, -----'''''-..---=c.j Ii
l'
Ul 50 Ul
3oot-----------1----------+----"'-,--=---~c.j

2oot-----------1----------+---------''''''''',----~c.j
25
loot-----------1----------+------------''-l
0'-------------'------------'-----·-------' 0
103 106
Number of cycles to failure

Tensile strength, ksi

50 100 150 200 250
1000

I II 125

.
D..
::;;
800 MaChine[5h
v ~

100
~
600
]
'!l
.~ 400
u.
/
-:V As-forgedor
decarburized
75

50
]
~
.~
u.

t.> 1,....--- --- >-_L --- ----1---

200
".... ...... -... .;;; 25

o
200 400 600 800 1000 1200 1400 1600 1800 2000
Tensile strength, MPa

Application of small-scale laboratory fatigue testing to the analysis of components or assem-

blies introduces additional variables. One is the effect of surface condition. The curves in the top
curve above demonstrate that the fatigue strength of steel specimens varies markedly, depending
on whether the surface is polished, machined, hot rolled, or as-forged. The steel tested was an
unidentified wrought low-alloy steel heat treated to 269 to 285 HB, equivalent to a tensile
strength of 876 MPa (127 ksi) and a yield strength of 696 MPa (101 ksi). Sample preparation
required that the specimens be machined and polished after heat treatment and that rolling or
forging precede heat treatment. For a fatigue life of one million cycles, the fatigue limit was 393
MPa (57 ksi) for the ground specimens, 317 MPa (46 ksi) for the machined specimens, 207 MPa
(30 ksi) for the as-rolled specimens, and only 152 MPa (22 ksi) for the as-forged specimens.
The curves in the bottom graph apply to steels with tensile strength ranging from 345 to 2070
MPa (50 to 300 ksi) and are approximations from several independent investigations. Sample
preparation for "as-forged or decarburized" specimens at the 965 MPa (140 ksi) tensile-strength
level include 4140-type steels rough machined from bar stock, heated to approximately 900°C
(1650 OF) in a gas-fired muffle for 20 to 30 min, very lightly swaged from an original 7,47-mm
(0.294-in.) diameter to a final diameter of 7.16 mm (0.282 in.), and air cooled. Heat treatment
consisted of austenitizing in a salt bath at approximately 830°C (1525 OF)for 45 min, oil quench-
ing, tempering in air for 1 h at approximately 620°C (1150 OF), and water quenching. Forging
and heat treating produced a surface decarburized to a depth of about 0.06 mm (0.0025 in.).
These specimens exhibited a fatigue strength, at 106 cycles, of about 310 MPa (45 ksi), compared
with 470 MPa (68 ksi) for samples that were not forged but were machined or polished and free
of decarburization. Decarburization lowers the strength levels obtained by heat treatment.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 355
 21-1. P/M: Relation of Density to Fatigue Limit and Fatigue Ratio 469

'" 50
~ (345)
tJl 40
0.. (276)
o
g 30
_ (207)
'E
:.:::i 20
~ (138)
Cl
~ 10
(69)
.50

0 .45
~
tx:
Q)
~
.40
Cl
~
u..
.35

6.4 6.6 6.8 7.0 7.2

Density, gr/cucm

The relationship of fatigue strength to density is shown above.

Fatigue strength is best at high densities. For similar P / M and
wrought parts, the ultimate tensile strength to fatigue strength
ratios are the same. However, fatigue strengths of P/M parts
generally are more stable and uniform than for wrought parts.
Parts containing nickel show improved fatigue resistance com-
pared to iron-carbon steels, and high-density nickel steel parts
can be case hardened to improve wear and fatigue properties.

Source: Kurt H. Miska, "Powder Metal Parts," in Source Book on Powder Metallurgy, Samuel Bradbury, Ed., American Society
for Metals, Metals Park OH, 1979, P 3
470 21-2. P/M: Relation of Fatigue Limit to Tensile Strength for
Sintered Steels

2
N/mm

150

• 0

00
•
100
•o
o
o

200 400 500 2 600

Tensile strength) N / mm

Fatigue limit of different sintered steels as a function of tensile strength.

Triangles are values for materials without phosphorus; open circles cor-
respond to PNC materials, closed circles to PASC materials.

Source: Per Lindskog, "The Effect of Phosphorus Additions on the Tensile, Fatigue, and Impact Strength ofSintered Steels Based
on the Sponge Iron Powder and High-Purity Atomized Iron Powder," in Source Book on Powder Metallurgy, Samuel Bradbury,
Ed., American Society for Metals, Metals Park OH. 1979, P 46
 21-3. P/M (Nickel Steels): As-Sintered vs Quenched and Tempered 471
for Cycles to Failure

r 0,1 Quenched and Tempered

/ _ Tensile Strength 105,000 psi
40

As - Sintered, Tensile
Strength 67,OOOpsi

10

7
106 10
Cycles to Fuilure , N.

SON diagrams representing fatigue behavior of7.0 g/ cm 3 density,

4 Ni-0.48 C steels, and effect of quenching and tempering on ten-
sile and fatigue strength.

One of the characteristics of the fatigue behavior of wrought

steels is that the S- N curve usually shows a distinct fatigue limit.
This is most marked in wrought plain carbon steels and usually
occurs between 105 and 107 cycles. A typical SoN curve for an
as-sintered nickel steel is shown above. As-sintered nickel steels
possess distinct fatigue limits occurring between 106 and 108
cycles.

Source: A. F. Kravic and D. L. Pasquine, "Fatigue Properties of Sintered Nickel Steels," in Source Book on Powder Metallurgy,
Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979, P 28
472 21-4. P/M (Nickel Steels): Relation Between Fatigue Limit and
Tensile Strength for Sintered Steels

60 r---,...-----,---,----...------,---,---....,.--=__--,
LEGEND
• Smooth As-Sinfered
50 0 Smooth Quenched a Tempered
.... 2.2 Kt Notch As-Sinlered
t:J. 2.2 Kt Notch Ouenched-B Tempered

Ui
(L
40
o
o
o
I
~
E 30
--'
.,
:::J
0>

" 20
LL

10 -

o
o 20 40 60 80 100 120 t40 160
Tensile Strength-IOOO PSI

Relation between fatigue limit and tensile strength (fatigue ratio) of sintered nickel
steels.

A plot ofthe fatigue ratio (above) indicates an average smooth value of 0.4 up to
150,000 psi tensile strength. Thus the averagefatigue ratio for sintered nickel steel is
0.4 which is apparently independent of density level, alloy content, and state of heat
treatment and therefore can be used to predict the fatigue behavior of other sintered
nickel steels.

Source: A. F. Kravic and D. L. Pasquine, "Fatigue Properties ofSintered Nickel Steels," in Source Book on Powder Metallurgy,
Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979, P 30
 21-5. P1M (Nickel Steels): Effect of Notches on Cycles to Failure for 473
the As-Sintered Condition

50 r--"-'-"'-''"T'"l"'T'T''r---''-'-'''-''"T'''l''T''Mr---r--''T''""T''"T'''''''''TT''I---'''''

40

Smooth

2.2 Kt Notched

10

106 107
Cycles to Fa.ilure IN.

SoN curves for 0.48% carbon-4.0% nickel alloy steel in the as-sintered condition (7.0
g/ em! density). The two curves demonstrate the effect of a notch on fatigue
characteristics.

Source: A. F. Kravic and D. L. Pasquine, "Fatigue Properties of Sintered Nickel Steels," in Source Book on Powder Metallurgy,
Samuel Bradbury, Ed., American Society for Metals, Metals Park DB, 1979, P 33
474 21-6. P1M (Nickel Steels): Effect of Notches on Cycles to Failure for
the Quenched and Tempered Condition

0--
0--
'in 0--
a.
o 30
o
o 2.2 Kt Notched
If)

~ 20
e
+
If)

10

106 107
Cycles to Failure, N.

S·N curves for 0.48% carbon-4.0% nickel alloy steel in the quenched and tempered
condition (7.0 g/cm! density). The two curves demonstrate the effect of a notch on
fatigue characteristics.

Source: A. F. Kravic and D. L. Pasquine, "Fatigue Properties of Sintered Nickel Steels," in Source Book on Powder Metallurgy,
Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979,P 34
 21-7. P1M (Low-Carbon, 1-5%Cu): Effects of Notches and Nitriding 475
on Cycles to Failure

40
--- Nilrided l""- I'--.
Nitriderl

'"
a.. 30
=
= .........
r-
=- to-.
Not nitrided
'"
'"
e
en 20

Smooth (K=1)
r--

Notched (K=2)
- r---..r-- Not nitrided

10
lQ5
I I III I IIII
10 8 lQ5
Cycles to Failure

S-N curves for sintered powders (low-carbon; 1 to 5% copper, 7.1 g/cm 3 density). As
shown above, notches greatly lower fatigue strength, particularly of those that were
not nitrided.

Source: "Nitriding Improves Fatigue Resistance of P 1M Parts," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 292
476 21-8. P1M (Sintered Iron, Low-Carbon, No Copper): Effect of Density
and Nitriding on Cycles to Failure

30
<;
I I III I I II I III r
Density - 6.4 g per eu em
I II
Density - 7.0 g per eu em
Nitrided

"iii
D.- 20
o
0
0_
r-..... r--...
Not nilrided
<11-
<II
~
""" t--.....
en 10
- Nilrided
Not nitrided
11
10 6 10 8 !OS 10 6
eye les to Fa ilure

SoN curves for P/M parts. As shown above, the greater the density, the higher the
fatigue strength of sintered iron powder (low carbon; no copper; notched; K = 2).
Nitriding in a salt bath is especially beneficial, it will be noted. Bath temperature was
565°C (1050 OF); nitriding time was two hours.

Source: "Nitriding Improves Fatigue Resistance of PI M Parts." in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 292
 21-9. P/M: Effect of Nitriding on Ductile Iron and Sintered Iron 477
(3%Cu) for Cycles to Failure

60
Ductile Iron Sintered Iron
I
I"":iZ ~ I
~'''':': :::
'Vi
::fit k:t Nitrided
.:.:.:.:.:.: ;:::;:::::; :::;::;::
;:;:; ::: :: :.:.:.:.:.:. ~'" .....:.... Nilrided
a.. 40 00

.....':':":
=
=. ~ ~ ::::;: ::: :~ .........•
:;:;:;:;:::::
:.:~:~:~:~:~: ~:~:~ :~:~ ~:

~
=
~ Not nihided

_:
-
::::; :::::'~
::::::;:;::: :.:.:.:.:.:
00

~:::::: ;~~~;~~~~~~ U~ ~ ~ ~ mmmm ~~t~r ~~r m= ~ Not nilrided

m
it ~ : mIt ~~tt :~t~ ~~~~ ::: ~~
~~~~~~~~~~~~~ ~~~~~~~~~ ~~~~;~ .':'
20
"""" <' t t~~ jtm~~ :::::;:;: ~ ~ ~: ~m ~
:::::::::

10 6 10 7 10 8 10 5 106 10 7 108
Cycles to Failure

Left: Effect of salt bath nitriding on fatigue strength of ductile iron. Right: Effect ofsalt
bath nitriding on P 1M parts. Specimens were made from a 3% copper sintered iron
ranging from 6.2 to 7.0 g/cm l in density. All specimens were unnotched and were
heated in a nitriding salt at 565°C (1050 OF) for two hours.

Source: "Nitriding Improves Fatigue Resistance ofP I M Parts," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, p 291
478 22-1. Brass/Mild Steel Composite: Comparison of Brass-Clad Mild
Steel With Brass and Mild Steel for Cycles to Failure

r----------r-------,.---------,15

20
12·5

• Brass clad mild sleel

5 a Mild sleel
o Brass

~0~4;:---------:±<----------,:-:':T------~IOV
5 6
10 10
Cycles 10 failure, log scale

SoN curves for composite of mild steel clad (by the explosion technique) with
brass.

Initially, the composite has greater fatigue strength than either brass or
mild steel alone, but above about 106 cycles the values for the composite drop
to about that of mild steel but still remain substantially higher than for brass
alone.

Source: S. K. Banerjee and B. Crossland, "Mechanical Properties of Explosively-CJadded Plates," in Source Book on Innovative
Welding Processes, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, P 148
 22-2. Stainless Steel/Mild Steel Composite: Comparison of 479
Stainless-Clad Mild Steel With Stainless Steel and Mild Steel for
Cycles to Failure

...-----------r----------,---------,15
• staintess clad mild steel

0____~.
20 ' Slainl e s s
steet
• 0 Mild sleel 12-5
'"
~
2 /5 10 '1:
.e
-e
o~~-'====t
0
= _ OJ'
"l::I
0_
7·5 ~
~
5
'"
:::
~
~

/05 /0 6
Cycles 10 failure, log scale.

SoN curves for composite of mild steel clad (by the explosion technique) with
austenitic stainless steel. Here it is seen that fatigue characteristics of the com-
posite are nearly the same as for stainless steel, and substantially higher than the
fatigue strength of the unclad mild steel.

Source: S. K. Banerjee and B. Crossland, "Mechanical Properties of Explosively-Cladded Plates," in Source Book on Innovative
Welding Processes, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, P 148
480 23-1. Carbon and Alloy Steels (Seven Grades): Effects of
Nitrocarburizing on Fatigue Strength

~ Normalized
c::J Normalized and treated in cyanide-base salt bath (treatment 1).
90 mm (3.5 in.), 570°C (1060 of), water quenched
600,------------------------,

~ 500 1 - - - - - - -
:2
s: 400
a,
1-------
c
~
300
Cl>
.~ 200
'"
u,
100

o
SAE 1000 SAE 1015 SAE 1035 SAE 1045 SAE 1060

600

500
'"
0-
:2
-S' 400
'"
c
300
-
~
Cl>

.g» 200
::l

'"
u,
100

0
SAE 1035 SAE 5134 Nitralloy

Bar charts showing increases in fatigue limit that may be ob-

tained by nitrocarburizing (gas or liquid processes).

The amount of improvement in fatigue strength of nitrocar-

burized materials, as determined with unnotched Wohler test
specimens, depends on the hardness and depth of the diffusion
zone. The potential for improvement in fatigue strength lessens
with increasing carbon and alloy content.

Source: Metals Handbook, 9th Edition, Volume 4, Heat Treating, American Society for Metals, Metals Park OH, 1981, p 269
 23-2. Carbon and Alloy Steels (Seven Grades): Effects of 481
Tufftriding on Fatigue Characteristics

Tufftrided, 90 min, 1050 F, water-quenched Normalized and tufflrided,

90 min, 1050 F, water-quenched
80 "" 1---, ~ I------'r---.....:..;'--T''''''-=..:~..:...;.;.'''''''':....:l.::..:..;,;.:..;;;.;.'__,
~ ~
"is~
en
.".
60 J----,---------t t-----t + I---l~-~----t

'"
c,
g 401-----1
=-
20 1-----11_

5134
SAE

Basic fatigue characteristics are directly related to carbon content, as indicated in the
above bar charts for carbon and alloy steels (unnotched test bars). Tufftriding these steels
shows results which prove that fatigue strength increases inversely with carbon content;
that is, the lower the carbon, the greater percentage increase in fatigue strength by
Tufftriding.

Source: Edward Taylor, "Tufftride: Only Skin Deep"," in Source Book on Nitriding, American Society for Metals, Metals Park
OH, 1977, P 280
482 23-3. Carbon and Alloy Steels (Six Grades): Effects of Nitriding on
Fatigue Strength

60
After Atmoaphere NitridJng 111111/
After Hardening I I
55

....
III
.c
..10:

11111
I
I

NltraUoy
1015 1141 8.620 4620 4340 Il5

Atmosphere nitriding studies showing the interrelationships of steel composition and nitrid-
ing by the gas process, and the effect on fatigue strength from nitriding.

Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 286
 23-4. Carbon-Manganese Steel: Effects of Nickel Coating on 483
Fatigue Strength

Thickness, mils
o 0.4 0.8 1,2 1.6
340

320
\ \ 45

'"
~ 'iii
:2 .:.l

...en
£ 300
c -, J::."
1»
c
~
'"
Q)

~ 280
-, e
1;;
Q)
::s
en
<, 40 'fJ
r-, .....
''::;
'"
u, u.

260
~

240 35
o 10 20 3D 40 50
Thickness, /.1m

Effect of coating thickness on the fatigue strength of a carbon-

manganese steel.

The reduction in fatigue strength produced by electro less

nickel deposits is affected by the thickness of the coating.
Thicker deposits have the greatest effect on fatigue strength,
This is illustrated in the above graph, which shows the reduction
in strength of a carbon-manganese steel (Werkstoff St52) pro-
duced by different thicknesses of a 5% boron-nickel.

Source: Metals Handbook, 9th Edition, Volume 5, Surface Cleaning, Finishing, and Coating, American Society for Metals, Metals
Park OH, 1982, P 232
484 24-1. Coil Springs, Music Wire (Six Sizes): Data Presented by
Means of a Goodman Diagram

40
200 1-----,,(---+----+----+----+--_=
20
o

Minimum stress,MPa

Wire Spring Free

Spring diam on Spring length Total Active Total
No. mm in. mm in. index mm in. turns tums tested
1 0.81 0.032 9.52 0.375 10.7 22.10 0.87 6.0 4.2 16
2 0.81 0.032 6.35 0.250 6.8 26.97 1.062 7.0 5.2 28
3 1.22 0.048 15.88 0.625 12.0 44.45 1.75 7.0 5.2 38
4 2.59 0.102 22.22 0.875 7.6 60.20 2.37 7.0 5.2 43
5 3.07 0.121 22.22 0.875 6.2 57.15 2.25 7.5 5.7 35
6 4.50 0.177 22.22 0.875 4.9 57.15 2.25 7.5 5.7 25

Data are average fatigue Iimita from S-Ncurv.. for 185 unpeened springs of various wire diameters run to 10 million cycles oCetrese.Allstresees were
corrected for curvature using the Wahl correction factor. The springs were automatically coiled, with one tum squared on each end, then baked at 260'C
(500 'F) for 1 h, after which the ends were ground perpendicular to the spring axis. The test load was applied statically to each spring and a check made for
set three times before fatigue testing. The springs were all tested in groups of six on the same fatigue testing machine at ten cycles per second. After
testing. the unbroken springs were again checked for set and recorded. Number 4 springs, tested at 1070 MPa (166 ksi) max strese, had undergone about
2,",%set after 10 million strese cycl.., but the stresees were not recalculated to take this into account. None of the other springs showed appreciable set.
The tensile strengths of the wires were according to ASTM A228.•

By means of the Goodman diagram many fatigue-limit test results can be shown on the same
diagram as indicated above. In this diagram, line OM represents the minimum stress for the
cycle; the plotted points represent fatigue limits for the respective minimum stresses used. The
vertical distances between these points and the minimum stress reference line represent the stress
ranges. Some scatter may be expected, at least partly attributed to normal changes of tensile
strength with wire diameter. Line UT is usually drawn to intersect line OM at the average ulti-
mate shear strength of the various sizes of wire.

Source: Metals Handbook, 9th Edition. Volume I, Properties and Selection: Irons and Steels. American Society for Metals. Metals
Park OH. 1978, P 293
 24-2. Coil Springs: S-N Data for Oil-Tempered and 485
Music Wire Grades

1 5 0 0 , - - - - - - - - - - - - - - , - - - - - - --r-r-- - - - - - - - - ,

200

1250t---------+--------+------------j
175
~
~

::i 1000f-------""....,,-=-::l-::--------+-
f
00
---l150 *
~

-
~
125
750t---------+----.::::!!oo""""'::-l----------l

500L ;--- - - - - - '-;-- - - - - - --'-::-- - - - - - ___==!

104 105 106
Number of cycles to failure
la)

Type of wire

r
Number Avg
of
springs

Oil tempered 6 ~

Music wire 8
~

104 105
Number of cycles to failure
(b)

(a) Springs were made of minimum quality music wire 0.59 mm (0.022 in.) in diameter. Spring
diameter was 5.21 mm (0.205 in.);D/dwas 8.32. Minimum stress was zero. Stresses corrected
by Wahl factor. (b) Life of springs used in a hydraulic transmission. They were made of
oil-tempered wire (ASTM A229) and music wire (A228). Wire diameter was 4.75 mm (0.187
in.), outside diameter of spring was 44.45 mm (1.750 ln.), with 15 active coils in each spring.
The springs were fatigue tested in a fixture at a stress of 605 MPa (88 ksi), corrected by the
Wahl factor.

The upper graph is a typical S- N diagram showing results of compression testing coil
springs, where the minimum stress is zero while maximum stress is shown by points on the
chart (see spring and testing details given in caption). The lower graph shows an alternate
method of presenting fatigue data for steel springs.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 292
486 24-3. Coil Springs: Effects of Shot Peening on Cycles to Failure

100

90

80

~
t<:

e
" 70

~~ r-...
'in

-
<:
s /peened
°
111 60

<,~tressed in bending

50 ~
-
~
""-Not peened
/Peened
I-
40

----- ~Stressed
'\:"Not peened
in torsion

Number of cycles to failure

Shot peening is often used to improve fatigue strength of springs by pre-

stressing the surface in compression as indicated in the chart above. Shot
peening can be applied to wire 1.6 mm (1/16 in.) or more in diameter, and
slightly smaller wire using special techniques. The kind of shot used is impor-
tant; better results are obtained with carefully graded shot having only a few
broken, angular particles. Shot size may be optimum at roughly 20% of the
wire diameter. However, for larger wire, it has been found that excessive
roughening during peening with coarse shot lessens the benefits of peening,
apparently by causing minute fissures. Also, peening too deeply leaves little
material in residual tension in the core; this negates the beneficial effect of
peening, which requires internal tensile stress to balance the surface
compression.
Shot peening is effective in largely overcoming the stress-raising effects of
shallow pits and seams. Proper peening intensity is an important factor, but
more important is the need for both the inside and outside surfaces of the
spring to be thoroughly covered. An Almen test strip necessarily receives the
same exposure as the outside of the spring, but to reach the inside, the shot
must pass between the coils and is thereby much restricted. Thus, for springs
with closely spaced coils, a coverage of 400% on the outside may be required
to achieve 90% coverage on the inside.
Cold wound steel springs normally are stress relieved after peening to re-
store the yield point. A temperature of 230°C (450 OF) is common because
higher temperatures degrade or eliminate the improvement in fatigue
strength.
The extent of improvement in fatigue strength to be gained by shot peen-
ing, according to one prominent manufacturer of cold wound springs, is
shown in the above diagram.

Source: Metals Handbook. 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 297
 24-4. Coil Springs, 8650 and 8660 Steels: Relation of Design 487
Stresses and Probability of Failure

800
9J", p'~b'blllty of f.i1u,~ 110

100

. -
100

<, /Peened sn
600
-.......
<,
o Not peened r-.. l"- BO

500
• Peened ..---- fl--- ~- 10

Not peened
400 I:: 60

50

106
Number of cycles to failure

BO0
50% p,oti.bility 01 f.i1~,e I 10

100 100

600
",0 ..... 90

<, r--
r-, r---- r-- BO

500
r-
r---- r----- ~ 10

400 I- 60

50

106
Number of cycles (0 failure

BOO
J", p,lb.b!lIty 011,"u,l 110

100 100

600
0"" . 90
:ii
~~
.
o'
80

500
. "" f'--.. I'-- t-
0 ......... 10
il

400 r--o r--.. 0 60

60
0
106
Number01 cycles to failure

Design stresses. Springs were made from 15.9 to 27.0-mm (% to 1-1/16-in.) diam 8650 and
8660 hot rolled steel and heat treated to between 429 and 444 HB. Springs were shot peened to
an average arc height of 0.008 in. on the type C almen strip at 90% visual coverage.

The desirability of conservative design in cyclical service is

illustrated in the three charts above, in which the minimum
stress used was low. Such data on springs hot wound from bars
with as-rolled surfaces are limited, and interpretation is there-
fore difficult. The value of peening, however, is made quite ap-
parent. Pertinent test data are given above.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH. 1978, p 304
488 24-5. Coil Springs, HSLA Steels: Effects of Corrosion on
Cycles to Failure

stress: 490 ± 340 t~Pa

~ 540
:<:
- e:tJ- - 78 (71 ± 50 ksi)
Mean
.; 490 - stress: 0&- - 71 52 - •, • \ •
.\.
-o 637 MPa
.,-l
U \
::s (/) p::
(92 ksi) ..>: :r:
~ 440 ocj- - 64 0 0
\
0
~
0. \
,
(/)
e
~ 390
o SUP7
_ SUP7-Nb-V C - 57 ~ 50 \
(/)
I r r
C
'tl \
,
0=, •,
(/) I I H
Q)
111
H
+J 104 10 5 106 :r:
(J) \
48
Cycles to Failure
o SUP7
• SUP7-Nb-V
I I I
2x10 4 5x104 10 5 2x10 5
Cycles to Failure

Fatigue life of coil springs: (left) not corroded and (right) corroded,

Compositions of HSLA Springs Tested

C Si Mn P S Cu Cr AI Nb v
SUP7 0.58 2.09 0.83 0.014 0.008 0.09 0.14 0.025 o o
SUP7-Nb-V " 0.56 1.94 0.79 0.014 0.008 0.09 0.09 0.021 0.15 0.18

Fatigue tests on coil springs at a hardness of 50 HRC were performed to examine the feasibil-
ity of S UP7-Nb-V to the actual suspension coil springs. When the coil springs were free of
corrosion, the result was as shown above (left), in which SUP7-Nb-V has comparable fatigue life
to that of SUP7 in any stress amplitude. When the coil springs were corroded, on the other hand,
the result was rather different. The corroding condition was as follows: an exposure in a chamber
filled with saltwater mist for 1.08 X 104 s (3 h) and a keeping in the atmosphere for 7.56 X 104 s
(21 h). After 10and 20 cycles ofthe corroding, the coil springs were loaded with the surface stress
of 490 ± 340 MPa (71 ± 50 ksi). The fatigue life of the coil springs subjected to 20 cycles of the
corroding are shown above (right). This time, different from the case in the graph at left, there
appears a remarkable difference between SUP7 and SUP7-Nb-V. Measurement ofthe surface
corrosion depth of the two steels showed no difference.

Source: Toshiro Yamamoto, Ryohei Kobayashi, Toshio Ozone and Mamoru Kurimoto, "Precipitation Strengthened Spring Steel
for Automotive Suspensions,"in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984,
P 1022
 24-6. Leaf Springs, 5160 Steel: Maximum Applied Stress vs 489
Cycles to Failure

2070/ 300
U)
U)
PEAK
250 RESIDUAL
ILl
lX:
E-<
U) '.-1
III
1380/200 . ."'.-... . ...... . . -----....-...
'. b----_
~
~ u
- - -
a
STRESS, MPa
-1070
-950
Q::<: ..... " '.
ILl <,
150 "
,.
, • ......d ••c .
,,
>-<<1l -565
...:Il:lo
l:lo;:;:
c,
------- -207
« -.
><
690/100 ., -69
~ , f
" .. • _ . - •• +400
345/5
104 105 106 107

CYCLES
Residual stress and unidirectional bending fatigue data for strain-peened SAE
5160 steel. Applied strain during peening curve a, -0.60%; curve b, -0.30%;
curve c, 0% (conventional peening); curve d, preset only; curve e, +0.30%; and
curve f, +0.60%.

Leaf spring specimens of SAE 5I60 steel quenched and tempered to 48

HRC were shot peened under various conditions of applied strain to intro-
duce a wide range of residual stress; then the S-N curves (see above) were
obtained from the same samples by testing in unidirectional bending. In this
illustration, the endurance limit corresponding to the specimen strain-peened
to produce a residual stress of -565 MPa (-82 ksi) will be used to develop a
stress-free ASR diagram for 5160 spring steel (48 HRC). This stress-free
ASR diagram will be used to predict the endurance limit for the other speci-
mens containing peak stress values of - 1070 (- I55), -950 (- 138), -207
(-30), -69 (- 10), and +400 MPa (+58 ksi). The predicted endurance limit
will be compared with the values determined experimentally.

Source: V. K. Sharma and D. H. Breen, "Some Aspects of Incorporating Residual Stresses in Gear Design." in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 82
490 24-7. Front Suspension Torsion Bar Springs, 5160H Steel:
Distribution of Fatigue Results for Simulated Service Testing

40,-------------------------------.....,
25 lots, 300 parts
'" 30r--------~A_r/hr7'T_V.l+_-------------____l
t:
~
'0
.8 20
E
:l
Z
101------r;..."...,1'A -V/WA-f/.HV/l-V.l1-V'A-t/'J--Y./lH/.,-+-r.,1'A - - - - - - - - _ _ _ _ l

Service life, 1000 cycles

Here are results from simulated service fatigue tests offront suspension torsion bar springs of
5160H steel. Size of hexagonal bar section was 32 mm (1.25 in.), Mean service life, 134,000
cycles; standard deviation, 37,000 cycles; coefficient of variations, 0.28. It must always be
considered that results from actual or simulated service testing are likely to vary considerably
from results of laboratory testing as shown above.

Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 677
 24-8. Gears, Carburized Low-Carbon Steel: Relation of Life Factor to 491
Required Life

5.0
II ~
4.0

3.0

u-' 2.0

-
....
.3
u
'"
u..
Q)

...J

1.0

0.5
10 4 106 10 8
Required Life in Cycles

The life factor depends on the required life in cycles. For a single mesh the number of revolu-
tions and the number of cycles are equal. For a gear which has more than one mating member,
the life must be equal to the required number of revolutions multiplied by the number of mating
gears.

Source: "Bending and Contact Stresses in Hypoid Gear Teeth," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 127
492 24-9. Gears, Carburized Low-Carbon Steel: Bending Stress vs
Cycles to Failure

(/)lI'l
(/)0
LLl ......
0:::
E-><
(/)
,......
~ '~ 6 . 90/ 1. 0
...... 0..
0 .........
Zc;l
LLlo..
a:l::'<:
'-'

3.45/.5

CYCLES
Bending fatigue design curves for carburized gears having different amounts of
circumferential residual stress at the root-fillet surface,

The bending fatigue design curves for case-carburized gears with the cir-
cumferential root-fillet residual stress varying from +138 MPa (20 ksi) to
-690 MPa (100 ksi) are shown above. From these curves the residual stress
factors at various life cycles were calculated as the ratio of the allowable
bending stress for gears with -483 MPa (70 ksi) residual compression to the
allowable stress for gears with + 138 (20), 0, -276 (40), and -690 MPa (100
ksi) residual stresses.

Source: V. K. Sharma and D. H. Breen, "Some Aspects of Incorporating Residual Stresses in Gear Design," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 86
 24-10. Gears, Carburized Low-Carbon Steel: Effect of Shot Peening 493
on Cycles to Failure

L50
L50
L50

3. 45/ • 5L....::-- '-;- ----'----;;- ----'

105

CYCLES

'"
0
.... 1.2
u
«
....
Vl 1.1
Vl
w
'"
....
Vl
SHOT CLEANED
...J
«
1.0 e e e e
:::>

-
Q

~
Vl
w
0.9
'"
o. 8 '--.......4- -L..,,----'----;-----..L-:;--.....L.-:;---I
5 6 7 8
10 10 10 10 10

CYCLES

Top: Allowable (LSO) bending stress design curves for as-carburized,

shot cleaned, and shot peened carburized gears. Bottom: Residual stress
factor computed from the upper chart. K a for carburized shot cleaned
gears equals 1; that is, the allowable SoN curve for carburized shot
cleaned gears is used for design purposes.

Ka Based on Dynamometer Tests

The dynamometer test data obtained on testing sets of gears with different magnitudes of
residual stresses can be used to develop the S- N curves necessary to calculate the Ka factor. The
L50 design curves for as-carburized, shot cleaned, and shot peened gears are shown in the upper
chart. The data for as-carburized and shot cleaned gears were obtained on testing six-pitch test
pinions on a Four Square Dynamometer. The S-N curve for shot peened gears is derived from
the results published by Alman and Black. The lower chart shows the residual stress factors
calculated from the S-N curves in the upper chart. It is assumed that the S-N curve for shot
cleaned gears is used for design purposes; that is, Ka for shot cleaned gears equals one. Accord-
ing to these results the effective bending stress for shot cleaned gears at 108 cycles is approxi-
mately 20% higher than as-carburized gears and approximately 15% lower than shot peened
gears. The value of Ka deviates from unity with increasing cycles, indicating a more significant
effect of residual stress at higher life cycles. At low cycles, the residual stress factor seems to
approach one, which means the residual stress has almost no influence on the fatigue properties
of a material at high loads. This is obviously because ofthe stress relaxation caused by the cyclic
plastic deformation accompanying low-cycle fatigue.

Source: V. K. Sharma and D. H. Breen, "Some Aspects of Incorporating Residual Stresses in Gear Design," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, pp 77, 78
494 24-11. Gears, Carburized Low-Carbon Steel: Probability-Stress-Life
Design Curves

U'l
o
U) ...... 13.80/2.0
U) L90
Jl.lX
c:<: 10.35/1.5
E- .......
U) 'M
Ul L50
t:lp..
z ....... 6.90/1.0
......
00..
z ......
'" L10
Jl.l:'<:
!Xl '-'

3. 45/ . 5 L..- .1....- ...I...- --'

5 8
10 10

CYCLES
Probability-stress-life design curves for shot cleaned carburized steel gears hav-
ing a root-fillet surface finish of 5Jl in.; l.e., KJl = 1.

Bending Stresses in Gears

In designing gears for a new application, a designer usually begins with a
preliminary selection of the tooth widths and other design parameters based
on past practices and empirical approaches recommended by AGMA. The
applied root-fillet bending stress is then calculated to predict the gear life
from the stress-life design curves such as those shown above. The procedure
is reiterated to optimize the design so that the calculated life is just equal to
the required life with an appropriate level of safety. The allowable stress-life
diagram characteristic for each material, heat treatment, and surface treat-
ment is usually obtained on testing acceptable commercial quality gears on a
dynamometer.

Source: V. K. Sharma and D. H. Breen, "Some Aspects of Incorporating Residual Stresses in Gear Design," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 74
 24-12. Gears, 8620H Carburized: Bending or Contact Stress vs 495
Cycles to Fracture or Pitting

500

BENDING 400
(St)
GlO~
OR
CONTACT
(Se)
300
~}s,
STRESS,
200
KSI G90
G5~ ~~
_-}st

10 6 10 7
CYCLES TO FRACTURE OR PITTING
S-N curves showing the wide difference in cycles to failure between bending and con-
tact stress.

Source; D. H. Breen, "Fundamental Aspects of Gear Strength Requirements," in Source Book on Gear Design, Technology and
Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 66
496 24-13. Gears, 8620H Carburized: A Weibull Analysis of
Bending Fatigue Data

5,600
.c
..J
c::
Q>
- 5,200
::I
e-
o
1--
c: 4,800
0
c:
0...
4,400
10 5 10 6
Cycles
Weibull analysis of bending fatigue data from gear tests
indicates that gears made from either the experimental
CH steel or 8620H have equivalent durability.

Metallurgical data gathered on these gears established the

adequacy of the experimental steel (a CRAT steel-computer
harmonizing by application tailoring). Although the experi-
mental steel had a significantly lower case hardenability, it
quenched out to a 100% martensite plus austenitic structure at
the root-fillet surface. Obviously, it had adequate, though not
excessive, case hardenability, thus representing an efficient use
of alloy hardenability in CRAT steels.

Source: G. H. Walter, "Computer Oriented Gear Steel Design Procedure," in Source Book on Gear Design, Technology and
Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 85
 24-14. Gears, 8620H Carburized: T-N Curves for Six-Pinion, 497
Four-Square Tests

8
7
M
0 G90 ---- SPAlLlNG
6
G50 : : : : - - - FATIGUE

~1~~-----
)(
GIO
..0 '" 5
, t
BENDING
c:
4 FRACTURE
w
::::l MIXED .............. '-........
'-........G90
C
CI::
aI-
3
PITTING
1.. . . . . ....
'-.....: G50
GIO

2L-- -L.. ----I -.l

105
CYCLES

T-N curves for carburized six-pitch pinion, four-square gear tests.

The above fatigue data show torque versus cycles to breakage, pitting and
spalling for a six-pitch pinion test. Note that there is a mixed area where
failure can occur from anyone or a mixture of the three modes.

Source: D. H. Breen, "Fundamental Aspects of Gear Strength Requirements," in Source Book on Gear Design, Technology and
Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 66
498 24-15. Hypoid Gears, 8620H Carburized: Minimum Confidence
Level; Stress vs Cycles to Rupture

i ... ..
C1
Ul 105
0:
....Q.
" ......
.. .. . .
....... .·--_ev
M

<,
~ .. .. . ..
<,
ai
...J ..... .. .
or <, ~......
fill...• :
.!!!,.
.......
a 1-..........:
0:
Iii ..... " <,

105 10 5 107
CYCLES FOR RUPTURE - (N)

Fatigue life data for hypoid gears. Sloping line indicates minimum confidence level.

Source: "Gleason Method for Estimating the Fatigue Life of Bevel Gears and Hypoid Gears." in Source Book on Gear Design,
Technology and Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 386
 24-16. Hypoid, Zero I and Spiral Bevel Gears, 8620H Carburized: 499
S-N Scatter Band and Minimum Confidence Level

-------- ----
-..........-..........
..........
---- 1;,0--..
060

<,
0 co

<,
---- -- ............
0 00

--
~<QI o

.a"" . ... . ..........

0..ll........

---
0

<, ..... .; ~"'-"'.

~

....'
:--.. .. ... ----... . . .. --. ..
.. . . • . .
: 0 0

<, ~~~ ~ ·8·~

<, ......
I~ 'l,

"
n n no

PROPOSED DESIGN LIMIT-

(NOT OVER 5 PER CENT FAILURES) .............
1
lI)
lI)
w
a:::
l-
ll)

105 106
CYCLES FOR RUPTURE - (N)

Fatigue data-composite for results obtained by testing various gear designs,

Source: "Gleason Method for Estimating the Fatigue Life of Bevel Gears and Hypoid Gears," in Source Book on Gear Design,
Technology and Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 386
500 24-17. Spiral Bevel and Zerol Bevel Gears, 8620H Carburized:
S-N Scatter Band and Minimum Confidence Level

....,..
~
«
g.
i.
d 1'---.",-
III
10
5 .-
<, .. ..- ..........- .-. ....
... . .'.
a:
w
<, ...
<; ~

. ... . . . . . ...- ..
CL
,..,..~" ~

ai <,
-----...... . .... ,... . . . . . . .
.J

~
a..
I
III
<,
1'--- ...
... ..
:la:
I-
... -.........:
III .........

10 5 10& 107
C.YCLE:S fOR RUPTURE -(N)

Graph offatigue life for spiral bevel and Zerol bevel gears. Sloping line indicates the minimum
confidence level.

Source: "Gleason Method for Estimating the Fatigue Life of Bevel Gears and Hypoid Gears," in Source Book on Gear Design,
Technology and Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 385
 24-18. Gears, 8620H Case Hardened: Relation of Life Factor to 501
Cycles to Rupture

5
4

3
~
~
--~
~
• - 5%
~ 2
I
5 <i
n0 50% >
... 1'000.
.......
.....Gl
Gl
....
Gl
:..:i
...... u
C
Gl
1 95% -0
c;:
0.9 0
C
0.8 U
0.7
0.6

10J 105 106

Cycles for Rupture

Both strength and durability are fatigue phenomena and therefore display a relation-
ship between stress and life. The life factor for strength may be obtained from the above
data.

Source: "Bending and Contact Stresses in Hypoid Gear Teeth," in Source Book on Gear Design, Technology and Performance.
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, p 127
502 24-19. Bevel Gears, Low-Carbon Steel Case Hardened: Relation of
Life Factor to Cycles to Rupture for Various Confidence Levels

5 -

...
~ 2
5%
I
l>
n
0
50%
u, po".

...::;
I>
....... .......
I
I
I

1 95%
, I

~
0.9 I
I
0.8 l
)]
0.6

103 1O~ 106 10 7 108

Cycles for Rupture

The life factor is obtained from the graph above. This depends upon the required life in cycles.
For a single mesh the number of revolutions and the number of cycles are the same. For a gear
which has more than one mating member the life must be equal to the required number of
revolutions multiplied by the number of mating gears. When the required life is less than
6,000,000 cycles on the pinion, the life factors will be different on gear and mating pinion.
In cases where the load varies, the designer may wish to determine the equivalent life at maxi-
mum torque. One suggested method is as follows:

L cp=60L H [k lnpl+ k n n ( T)
2 T~
5.68
+k)npJ
(Tf; ) 5.68
+ .. +kl/npl/
( T)
i; 5.68 ]

where L cp = required equivalent life in pinion cycles at maximum torque.

L H = required total life in hours.
k),k 2,k) k,,=proportion of time at torque loads T" Tb T) . . . T"respectively.
n PI' n n. n P) n PI/= pinion rpm corresponding to torque loads T" T 2, T) . . . TI/
respectively.
T 1, T 2, T) . . . T" = torque loads where T I is maximum torque and T" is minimum
torque which will produce a stress above the endurance limit.
The required equivalent life in gear cycles at maximum torque may be obtained by multiplying
the life in pinion cycles by the gear ratio:
L CG=L cp -
n,
NG

Source: "Bending Stresses in Bevel Gear Teeth," in Source Book on Gear Design. Technology and Performance. Maurice A. H.
Howes, Ed., American Society for Metals, Metals Park OH, 1980, p 149
 24-20. Gears, AMS 6265: S-N Data for Cut vs Forged 503

35

.c
.....J
30
=
=
=-
",,-

,9
""
25
.6.
ll...
.8
"C
,~ 20
CL
CL
<I:
"C
co
0
.....J
15
E 6 CuI gears
:::I
E • Forged gears
';(
co
::;;:
10

5
lQ3 104 10 5 10 6 1Q7
Cycles 10 Failure

Fatigue data shown in this chart proved that teeth on precision forged AMS 6265
helicopter pinions have a higher fatigue limit than cut teeth. Loads shown are applied
actuator loads. Tooth loads are approximately 33% greater.

Source: "How Gearmaking Methods Compare," in Source Book on Powder Metallurgy. Samuel Bradbury, Ed" American Society
for Metals, Metals Park OH, 1979, P 347
504 24-21. Spur Gears, 8620H: S-N Data for Cut vs Forged

120

110

100

"u;
a.. 90
0
0
0_

80
"0
0
0:::
co 70
lJ)
lJ)
e 1% failure
en 60
m
c
'j§
0
z: 50

40
- - - Cui gears
Forged gears
30

10 5
Cycles 10 Failure

S-N curves for spur gears forged and cut from 8620H steel.

As shown above, results of beam fatigue tests indicate that precision flow-forged
gear teeth are about 20% higher in fatigue strength than cut teeth.

Source: "How Gearmaking Methods Compare," in Source Book on Powder Metallurgy, Samuel Bradbury, Ed., American Society
for Metals, Metals Park OH, 1979, P 346
 24-22. Gears and Pinions: P/M 4600Vvs4615; Weibull Distributions 505

LEGEND
SINTERING
~. ALLOY TEMP TIME
99._ 4 4600V 2~F I MIN
5. 4600Y 2100-F 3MIN /
6 4600 V 2350·F 3MIN /
7 4600V 2350·F 6MIN
8 4600V 2100-F 6 MIN /
_5.0 3 4615 (BAR STOCK) /
/
80.0 /
/
/
60.0
/
/
w
400 /
a: 300 /
::>
...J 3 /
¢ --(
u. 20.
/
~ /
z
w /
'00 /
0
a: 8
I
w 60
Q. /
40 /
/
/
20 /
/
/
1.0
.1 .2 .3

TIME (HR.)

LEGEND

.- ~.
,
2
3
!!::ill
2000
2000
4615
SINTERING
TEMP TIME
2350·F "'6MiN
2350·F 3 MIN
(BAR STOCK)
95.0

80.0

600

w 40.0
a:
300
3
¢
u. 20.0
~
Z
W
0 10.0
a: 80
w
Q. 60

40

2.0

1.0 34~6810
.1 .2 .8 1.0

TIME (HR.)

Top: Weibull distribution charts for fatigue testing of actual gears and
pinions made from 4600V alloy with various sintering times and
temperatures as shown, compared with cut pinions (4615 bar stock).
Bottom: Similar to graph at top exceptfor alloy 2000 and 4615bar stock.

Source: P. C. Eloff and L. E. Wilcox, "Fatigue Behavior of Hot Formed Powder Differential Pinions." in Source Book on Powder
Metallurgy, Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1980, p 308
506 24-23. Gears and Pinions: P/M Grades 4600V and 2000 vs 4615;
Percent Failure vs Time

99.9

NO.
LEGEND
ALLOY
/
950 t 4600V /
2
3
2000
4615 8AR STOCK
/
80.0 1-1 95'/. CONFIDENCE /
LIMITS

60.0
/
3 /
40.0 'i
w
0:: 30.0
~
--l
<l: 20.0
lJ...

I-
Z 10.0
w
0 8.0
0::
W 6.0
(L

4.0

2.0

1.0
.1 2 3 4 5 6 8 10

TIME (HR.)

Fatigue data for actual gears tested in specially designed machines. Presented
here are Weibull distributions for the three types of alloys tested.

Since the data from the two powder alloys fell into two groups, it was decided to
fit one Weibull curve to all of the data points from each alloy. This was done to
obtain more data points for each curve. The results are shown in the graph above,
which also graphically indicates the 95%confidence limits on the BID lives. It isplain
that the 4600V pinions have superior fatigue life at the stress level of 92,400 psi, and
the slope ofthe Weibull curve indicates uniform deoxidation of preforms and there-
fore less scatter (steeper Weibull slope) in the fatigue data. In the case ofthe 4600V
alloy, the sintering temperature should have little effect on deoxidation, since the
major alloying constituents, nickel and molybdenum, are readily reducible by CO
at temperatures even below 1150 °C (2100 OF).

Source: P. C. Eloff and L. E. Wilcox, "Fatigue Behavior of Hot Formed Powder Differential Pinions." in Source Book on Powder
Metallurgy, Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1980, P 313
 24-24. Gear Steel AMS 6265: Parent Metal vs Electron Beam Welded 507

2 4 68l 2 4 681 2 4 681 2 4 681

100
I
900
10
10
i~

50'

SON diagram for AMS 6265-parent metal versus electron beam

welded.

The welded specimens failed in the weld zone at 86% joint

efficiency.
In the weld evaluations made, excellent mechanical proper-
ties were found. Other gear materials tested resulted in compar-
a ble weld joint efficiencies. In general, it was demonstrated that
electron beam mechanical properties were comparable or better
than welds made with other fusion welding processes such as gas
tungsten arc and metallic arc welding.

Source: N. F. Bratkovich, W. L. Mcintire and Robert E. Purdy. "Electron Beam Welding-Applications and Design Considera-
tions for Aircraft Turbine Engine Gears.t' in Source Book on Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., Ameri-
can Society for Metals, Metals Park OH, 1981, P 199
508 24-25. Gears, 42 CrMo4 (German Specification): S-N Curves for
Various Profiles

80

z-,

z ~. ;.>.
>-
\0
40
~
.........~
~
" ~. .......
.~
,....
I~
:~~.. :. ;. ........
i
--- -- -
.::--.
O,I..l,fl
(I ~ (; (I
-
30
~:~ ! .,..:.-;:::~ "."!.:' .- . . .- .. --- ..
~~~ ll.i i\:
~~.

I • O.~o2

0, 'lG r\(W/m
20
3.10 4
-- _.. . J ~

1.)C G
-J-L
LW--+

SoN curves for various tooth profiles (50% survival probability in the short life
range).

The Woehler curves shown above are based (in the sloping section) on a
survival probability of 50% at the number of cycles indicated. The horizontal
sections of the curves are based on the highest load that can be carried for a
minimum life of five million cycles.

Source: H. Winter and M. Hirt, "The Measurement of Actual Strains at Gear Teeth, Influence of Fillet Radius on Stresses and
Tooth Strength,"in Source Book on Gear Design, Technology and Performance, Maurice A. H. Howes, Ed .. American Society for
Metals, Metals Park OH, 1980, P 102
 24-26. Gears, 42 CrMo4 (German Specification): Endurance Test 509
Results in the Wei bull Distribution Diagram

95

I I I , ~ ...

f/ II ~
..... I II'
;-- 70
Of J
50
II il
if 71'
30
If
20 f
,I:
I

10 J~!'I A -- 1·-
-- _..
8 , l-
[
. .
6 ----f I - - 1--
I
4 L ~
- ---
.- . -
3.,0'< 4 6 8 1.'0:> 2 3 t. 5 (; 8 1.1
,-W --~

Endurance test results in the Weibull distribution diagram.

The parameters ')I,A are adjusted to the test points. An example is

shown in the graph above.which represents in such a probability grid
the test results for gears of one tooth form. Scale of ordinates is the
failure probability A = 1- W= if (n + I) for test i out of n test results
sorted to the number ofload cycles at which fracture occurred. The test
points are approximated with a straight line. From this curve we are
able to read, respectively, life values L IO , L so, and LJo for 10,50, and 90
percent failure probability, or 90, 50 and 10 percent survival proba-
bility.
A more adequate approximation by the theoretical distribution is
achieved by a three-parameter Weibull distribution. This formulation
produces a minimum endurance Lo, which is reached by all test pieces.
Also, the above chart shows the compensating curve which results
from the formulation of the three-parameter Weibull distribution.

Source: H. Winter and M. Hirt, "The Measurement of Actual Strains at Gear Teeth, Influence of Fillet Radius on Stresses and
Tooth Strength,"in Source Book on Gear Design, Technology and Performance, Maurice A. H. Howes, Ed.. American Society for
Metals, Metals Park OH, 1980, P 102
510 24-27. Bolts, 1040 and 4037 Steels: Maximum Bending Stress vs
Number of Stress Cycles

100
1040 steel
0 • 0
I - 100
• 4037 steel
615

0 • - 95
s: 650 "v;
.
:::iE
"'".
~

~g' 625
~
Cl
0.0 - c
90 :0
:0
~
c
.
~
)(

600
.n . ..
.0
)(

:::iE
~o.
of> .
~~-
~-
- 85
515
o 0 • • 0
>---
1-

Number of stress cycles

The bolts (7'8 by 2 in., 16 threads to the inch) had a hardness of 35 HRC. Tensile properties of
the 1040 steel at three-thread exposure were: yield strength, 1060 MPa (154 ksi); tensile
strength (axial), 1200 MPa (175 ksi); tensile strength (wedge), 1190 MPa (173 ksi). For the
4037 steel: yield strength, 1110 MPa (161 ksi); tensile strength (axial), 1250 MPa (182 ksi);
tensile strength (wedge), 1250 MPa (182 ksi).

In general, if bolts made of two different steels have equivalent hardnesses throughout
identical sections, their fatigue strengths will be similar (see above S- N data), 'as long as
other factors such as mean stress, stress range, and surface condition are the same. If the
results of fatigue tests on standard test specimens were interpreted literally, high-carbon
steels would be selected for bolts. Actually, steels of high carbon content (more than
0.55% carbon) are unsuitable because they are notch sensitive.
The principal design feature of a bolt is the threaded section, which establishes a notch
pattern inherent in the part because of its design. The form of the threads, plus any me-
chanical or metallurgical condition that also creates a surface notch, is much more impor-
tant than steel composition in determining the fatigue resistance of a particular lot of
bolts.

Source: ASM Committee on Carbon and Alloy Steels. "Threaded Steel Fasteners." in Quality Control Source Book. A. K. Hingwe,
Ed .. American Society for Metals, Metals Park OH, 1982. P 206
 24-28. Bolts: S-N Data for Roll Threading Before and After Heat 511
Treatment

450 r - - - , - - - - , - - - - - , - - - ,
Roll threaded before 60
heat treatment
375 1 - - - t - - - - - 1 - - - - t - - - - - j
50

3001---t-----1----t-----j
40

225
30 U5~

150 f-----11-
20

751----t------I---' 10

o L- ! -_ _-'----_ _ ~_~

104
450 , - - - , - - - - , - - - - - . - - - ,

60

375
- 50

300
40
'"
0-
:;;:
::i 225 I - - - t - -
1"
U5

150 f - - - - t - - - - t - - - + - - - - - j 20 ~

Roll threaded after

heat treatment
75 1---+---+---+---::::1 10

Number of stress cycles

S-N curves showing fatigue limits for roll-threaded bolts. Upper

graph represents four different lots of bolts that were roll
threaded, then heat treated to average hardness of 22.7, 26.6,
27.6, and 32.6 HRC. Lower graph represents five different lots
that were heat treated to average hardnesses of 23.3, 27.4, 29.6,
31.7, and 33.0 HRC, then roll threaded. Bolts having higher
hardnesses in each category had higher fatigue strengths.

Other factors being equal, a bolt with threads properly rolled

after heat treatment-that is, free from mechanical imperfec-
tions-has a higher fatigue limit than one with cut threads. This
is true for any strength category. The cold work of rolling in-
creases the strength at the weakest section (the thread root) and
imparts residual compressive stresses, similar to those imparted
by shot peening.

Source: ASM Committee on Carbon and Alloy Steels, "Threaded Steel Fasteners," in Quality Control Source Book, A. K. Hingwe,
Ed., American Society for Metals, Metals Park OH, 1982, P 202
512 24-29. Power Shafts, AMS 6382 and AMS 6260: Electron Beam
Welded vs Silver Brazed Joints

.. ELECTRON BEAM WELDED JOINTS

• SILVER BRAZED JOINTS

TEST TEMPERATURE, 500-600 oF
60

o
o

.
Q
x 50
III
o
Z
:l
o0..
S/N-2

..
S/N-3

..
:I: 40
u
z
~~
+1 SIN-I S/N-2
w
5
--
.. --
ROOM TEMP.

--
30

- r-- l -
0:
o
I- i----
r---- ....
SIN-I

---
>-
0:
o
!;i 20
I-
r--.- :--- --. "-
0:
m
>
• -I---

10

CYCLES TO FAILURE

S-N curves for electron beam welded versus silver brazed power shafts made from AMS 6382 and
6260 alloy steels.

In the welded shafts, failures occurred apart from the weld, while in the brazed units all
failures occurred in the brazed joints.

Source: S. M. Silverstein, V. Strautman and V. R. Freeman, "Application of Electron Beam Welding to Rotating Gas Turbine
Components," in Source Book on Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals,
Metals Park OH, 1981, p 187
 24-30. Axle Shafts, 1046, 1541 and 50854 Steels: S-N Data for 513
Induction Hardening vs Through Hardening

:\, B & C I1'I1lUCf IO~ HARDENED 1541

~on D- 1046
if. E - THROUQI HARDENED 50B54
'""
if.
v: 1 SO
..9."Il!':.? ....
,..... . :~
.... ..........
?i
., ............ --
-, ~
". \
tr. ... "

-. -'-."
..... \', ·'-'-'-B
u:: ] on
>-
c:::
<
~
, .......
0; 50 ......... ::::.:::-:.:::-::.: ~
"
-

APPLIED ~
STRESS GRWIE~T
200 400 600 800
DISTA.'JCE FRO~l SURFACES (.001")

A, B & C - rxructrox HARDE\'ED 1541

D - 1046
..... 140 E - TI-JROllGH HARDE\ED 50B46
IF.
~

n ..... .......~
u: 100
u:
2
l-
i/;
':-.."

.
..
.,
~ 60
"
.~' -, . " :"::..,,:,,'~ c "
f:S
u: .'·-._.-E
----B
20

10 2 ]0 3 10 4 lOS 10 6
NO. OF CYCLES TO FAILURE

Top: Axle-shaft strength gradients in terms of shear yield strength. Bottom: Fa-
tigue performances of axle shafts as a function of strength gradient.

Induction hardened 0.40% carbon steel axle shafts were developed to replace through hardened alloy steel
shafts for both product- and cost-improvement purposes. This was accomplished after a rather comprehensive
bench-test program, which examined variables such as surface hardness, core hardness, gradient strength, dis-
tortion, composition, and surface-condition effects. The more promising approach was then subjected to chas-
sis, proving-ground, and in-service testing. Some interesting reflections can result from examining some of the
fatigue data that were generated. That the through hardened concept was vulnerable can be surmised by consid-
ering the stresses developed in a full-floating splined shaft loaded in torsion. The stresses are a maximum at the
surface and drop linearly to zero at the center. At the spline, the stresses drop more rapidly at the onset due to the
stress concentration caused by the spline. The upper graph shows the stress gradient in the body area when the
shaft is loaded to 110,000-psi shear stress. Also plotted on this graph are the shear yield-strength gradients
(converted from hardness) of the production alloy shaft, along with several experimental induction hardened
shafts. One would expect the through hardened shaft to have a surface-origin failure and to be lower in strength,
since its surface is the lowest hardness. Also, the high strength of the center ofthe shaft is essentially wasted, since
it is lowly stressed. Gradient strengthening by induction hardening provides a means of providing a better
strength match for the stress gradient. The lower graph gives the fatigue curves established for shafts having the
strength gradients shown in the upper graph.

Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals. Metals Park OH. 1979, P 88
514 24-31. Steel Rollers, 8620H Carburized: Effects of Carburizing
Temperature and Quenching Practice on Surface Fatigue

99·9

99·0 SAE 8b2011 - Reheat Quench

95'0
Carbo
of .
Sliding
.. Cor r .
Coef.
Slope
B
0 Grou 0 1750 ~I .89 2.199
90·0
II Group Q 1800 21 .97 I. 463

80·0 • Group R 1900' 21 .98 2.961

·vacuum
70'0 ---- ------
AJ·l
60·0

50·0
Data corrected to
40'0 -

= 400 ksi
Sc___
____
30'0
A= 0.5

20·0
~

.
a:
:>

:;;
.... 10'0
I-
Z
~

U
a:
~
e,
5'0

4'0

3·0

"0

Q 0 R
1'0

0'5
0'4

0'3

0·2

105 8
0" 106 7 10
I 4 5 6 7 8 9 , 2 3 4 5 6 7 • 9 , 2 4 5 6 7,19
c v c i es

S-N data, Weibull probability plot: Effect of three carburizing temperatures on surface fa-
tigue for carburized 8620" steel. All were slow cooled and reheated for quenching. This
technique improved fatigue characteristics compared with direct-quenched rollers.

Source: S. L Rice. "Pilling Resistance of Some High TemperatureCarburized Cases," in Source Book on Gear Design. Technology
and Performance. Maurice A, H, Howes, Ed" American Society for Metals, Metals Park OH, 1980, p 234
 24-32. Steel Rollers, 8620H Carburized: Effects of Carburizing 515
Temperature and Quenching Practice on Surface Fatigue

99 .• 1- =...,
SAE 8620H - Direct Quenched
.'·0
Carbo Sliding Corr.
OF \ Coef.
9"0
1750
'0' 0 1-0=--Gc..r_o_u....
p_N 21 .90 r-
+f. ._ _r----1
o 1800
Group P 21 .97
21 .97
10.0 ~.u!!J"--"S'---'-"=--=----'-".:.-----=:..:..:.:...:..,~J_------+f
1900' ...- r _ - - - - _ j
70.0 ~a"'cull
...umlll-. _f_.=_"~-----_c,.-_r'--:_---__1

6&:~(--------------_:_
. .+-------__,f.;;;;:"M,./_------_I
'O·ol---------------.'f--------r..-f---i~-------__I

4o·01---------------;.t----------tl~-'":f_--------__I
Data Corrected to
30·01--1>:c-~W9__1<!;i_-----.+------__::~+...,,'-----------_f
c
A = 0,5

20·0

. JO·Oj----------r--='---------if---------------_f
~
I-
z
U
0:

.. "0 - - - - - - - - f - : ; . ; o - - - - - - - - - - T - . H........---------------1

"0 t - - - - - r - - - - - - - - + - j ' - - - - - - - - - - - - - - - - - - - _ _ I

S N P
'-0

D"

0"

D"

0·3

0'\ ,105 345678" 2 3 4567." 45.71.

C 't' c i es

S-N data, Weibull probability factor: Effect ofthe three carburizing temperatures on
surface fatigue for carburized 8620H steel. All were direct quenched from the tempera-
tures shown above.

Source: S. L. Rice, "Pitting Resistance of Some High Temperature Carburized Cases," in Source Book on Gear Design, Technology
and Performance, Maurice A. H. Howes, Ed.. American Society for Metals, Metals Park OH, 1980, P 233
516 24-33. Linkage Arm, Cast Low-Carbon Steel: Starting Crack Size vs
Cycles to Failure

III
OJ
s:
u
c .300
E
E
I
7
0
.250 0
W 6
N W
en N
(136MPa) iii
:.:: .200 5
U :.::
<{ af; 0.63 in. 116 m rn ) U
a:: <{
u 4 a::
.150 U
<.9
Z <.9
i= 3
Z
a:: i=
;:! .100 a::
<{
en I-
2 en
.050

1000 10,000 100,000

CYCLES TO FAILURE. N f -

Fatigue life of a linkage arm as a function of starting crack size.

The variation of the fatigue life, NJ, with the starting flaw size a.; is shown in the
diagram above. The fatigue life increases dramatically at very small a i values. The
far curve shows that in the long life regime the final crack size has only a small effect
on Np This is because fatigue crack growth rates are very low at low ~Kvalues and
hence the greatest fraction offatigue life is spent at the smaller crack sizes. Since the
controlling parameter is ~K, low life for small crack sizes is possible at high cyclic
stresses. The second set of curves shows that doubling the cyclic stress range reduces
the fatigue life by about an order of magnitude. Also, if the starting ~K value is
high, the final crack size has a larger effect on the cyclic life. The above diagram
shows the importance of adjusting both the cyclic stress and starting flaw size to
optimize the fatigue life.

Source: Steel Castings Handbook. 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH. 1980,
P 4-17
 24-34. Notched Links, Hot Rolled Low-Carbon Steel: S-N Data for 517
Component Test Model

... - Component Tests

I. 0 r - - - - - - - - - - - . : . . . - - - - - - - - - - - ,

~: ~.., _
...... ~~notched
0.7 ~~~ ~/2Z7!?2Z
0.6 .~ _./Coml
0.5 ---------_--=--~~~nent Tests
<, .l --A
~~
04
. rr .............& -
Local-Strain Model

0.3
~1 ......
S-N / ~@;/?/!?!Z2Z
R=-1 Model I
0.2L..--------1------.........L~------J
7
10· loll lOll ~
Fatigue Life. Nj • cycles

COMPONENT TEST MODEL. The most straightforward life prediction model of a component is
developed from fatigue tests of the component itself. The component is cyclically loaded in a manner
that attempts to simulate service and the model is the plot of the test results. The cyclic load or nominal
net-section stress is plotted versus cycles to crack initiation N; or to failure Np
Example: Component fatigue tests were performed on the notched link ofthe previous examples; the
results are listed in the table and shown graphically above. The fatigue strength for crack initiation at
one million cycles is:
iJ.S
= 0.47
Sli
106 cycles
The three models-S-N, local strain, and component tests-are compared above, and the three
predictions for fatigue strength (iJ.S) at one million cycles are, respectively, 28, 42, and 47% of the
tensile strength Sli' The local strain prediction is closer to the behavior observed in the component test
than the S-N prediction.

Component Test Results, Hot-Rolled Carbon Steel Notched Link

Tensile Yield Cycles to

strength, strength, crack initiation,
Su,MPa S}",MPa N;

417 242 0.818 1.36 X 10'

388 236 0.880 1.69 X 10'
366 239 0.936 2.62 X 10'
366 239 0.690 6.99 X 10'
402 235 0.628 8.24 X 10'
417 242 0.604 1.05 X 105
366 239 0.444 1.64 X 10"
402 235 0.404 2.21 X 10"
388 236 0.418 2.99 X 106

Source: Harold S. Reemsnyder, "Constant Amplitude Fatigue Life Assessment Models," in Proceedings of the SAE Fatigue Con-
ference P-I09, Society of Automotive Engineers. Inc .. Warrendale PA, 1982. P 127
518 24-35. Fuselage Brace, Ti-6AI-6V-2Sn: Fatigue Endurance of
HIP-Consolidated Powder

.;;;
.:.!
tl' 120 Annealed plate

-..;
~ (min)
:;;
E
::>
E 100
'x
~'"
80
m
60
10' 10' 10· 10'
Number of cycles

Fatigue endurance of Ti-6AI-6V-2Sn powder consolidated by HIP at 1650 of

(900°C). Significance of boxed numbers is as follows: 1: HIP run #1, as ma-
chined. 2: HIP run #2, vacuum annealed for 2 h at 1300 of. 3: HIP run #2,
vacuum annealed for 16 h at 1300 ° F. 4: HIP run #2, vacuum annealed for 24 h
at 1300 OF. 5: HIP run #4, vacuum annealed for 24 hat 1300 OF.

A fuselage brace made from HlP'd Ti-6AI-6V-2Sn powder was used to

establish the flight worthiness of a HlP'd P/M airframe component. The
tensile and toughness properties developed compare well with the average
values for forgings. The fatigue endurance limit of H lP powder developed in
this program is given in the SoN diagram above. Here the HIP data points lie
within the band for annealed forgings and plate.

Source: J. H. Moll. V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications," in Powder Metallurgy-
Applications. Advantages and Limitations. Erhard Klar, Ed.. American Society for Metals. Metals Park OH, 1983, P 288