Citation
Computational Design of Nickel Based Superalloys for Industrial Gas Turbine Components

Material Information

Title:
Computational Design of Nickel Based Superalloys for Industrial Gas Turbine Components
Creator:
TAPIA, ALMA STEPHANIE ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Alloys ( jstor )
Carbides ( jstor )
Heat resistant alloys ( jstor )
Heat treatment ( jstor )
Liquidus ( jstor )
Melting ( jstor )
Modeling ( jstor )
Phase transformations ( jstor )
Solidification ( jstor )
Solidus ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Alma Stephanie Tapia. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/24/2006
Resource Identifier:
496180313 ( OCLC )

Downloads

This item has the following downloads:

tapia_a ( .pdf )

tapia_a_Page_051.txt

tapia_a_Page_230.txt

tapia_a_Page_053.txt

tapia_a_Page_073.txt

tapia_a_Page_225.txt

tapia_a_Page_226.txt

tapia_a_Page_011.txt

tapia_a_Page_131.txt

tapia_a_Page_056.txt

tapia_a_Page_171.txt

tapia_a_Page_218.txt

tapia_a_Page_177.txt

tapia_a_Page_193.txt

tapia_a_Page_176.txt

tapia_a_Page_028.txt

tapia_a_Page_058.txt

tapia_a_Page_122.txt

tapia_a_Page_096.txt

tapia_a_Page_006.txt

tapia_a_Page_196.txt

tapia_a_Page_009.txt

tapia_a_Page_187.txt

tapia_a_Page_080.txt

tapia_a_Page_231.txt

tapia_a_Page_126.txt

tapia_a_Page_086.txt

tapia_a_Page_102.txt

tapia_a_Page_130.txt

tapia_a_Page_153.txt

tapia_a_Page_134.txt

tapia_a_Page_189.txt

tapia_a_Page_228.txt

tapia_a_Page_217.txt

tapia_a_Page_033.txt

tapia_a_Page_181.txt

tapia_a_Page_088.txt

tapia_a_Page_175.txt

tapia_a_Page_214.txt

tapia_a_Page_084.txt

tapia_a_Page_071.txt

tapia_a_Page_160.txt

tapia_a_Page_232.txt

tapia_a_Page_166.txt

tapia_a_Page_045.txt

tapia_a_Page_191.txt

tapia_a_Page_104.txt

tapia_a_Page_207.txt

tapia_a_Page_015.txt

tapia_a_Page_180.txt

tapia_a_Page_132.txt

tapia_a_Page_182.txt

tapia_a_Page_017.txt

tapia_a_Page_041.txt

tapia_a_Page_019.txt

tapia_a_Page_047.txt

tapia_a_Page_054.txt

tapia_a_Page_205.txt

tapia_a_Page_185.txt

tapia_a_Page_036.txt

tapia_a_Page_169.txt

tapia_a_Page_149.txt

tapia_a_Page_078.txt

tapia_a_Page_020.txt

tapia_a_Page_037.txt

tapia_a_Page_164.txt

tapia_a_Page_143.txt

tapia_a_Page_061.txt

tapia_a_Page_241.txt

tapia_a_Page_202.txt

tapia_a_Page_087.txt

tapia_a_Page_002.txt

tapia_a_Page_167.txt

tapia_a_Page_114.txt

tapia_a_Page_105.txt

tapia_a_Page_094.txt

tapia_a_Page_113.txt

tapia_a_Page_003.txt

tapia_a_Page_052.txt

tapia_a_Page_012.txt

tapia_a_Page_044.txt

tapia_a_Page_238.txt

tapia_a_Page_022.txt

tapia_a_Page_082.txt

tapia_a_Page_031.txt

tapia_a_Page_074.txt

tapia_a_Page_129.txt

tapia_a_Page_156.txt

tapia_a_Page_125.txt

tapia_a_Page_219.txt

tapia_a_Page_155.txt

tapia_a_Page_008.txt

tapia_a_Page_035.txt

tapia_a_Page_016.txt

tapia_a_Page_183.txt

tapia_a_Page_188.txt

tapia_a_Page_111.txt

tapia_a_Page_215.txt

tapia_a_Page_162.txt

tapia_a_Page_157.txt

tapia_a_Page_237.txt

tapia_a_Page_148.txt

tapia_a_Page_065.txt

tapia_a_Page_222.txt

tapia_a_Page_235.txt

tapia_a_Page_209.txt

tapia_a_Page_121.txt

tapia_a_Page_213.txt

tapia_a_Page_072.txt

tapia_a_Page_010.txt

tapia_a_Page_064.txt

tapia_a_Page_034.txt

tapia_a_Page_119.txt

tapia_a_Page_179.txt

tapia_a_Page_195.txt

tapia_a_Page_029.txt

tapia_a_Page_039.txt

tapia_a_Page_057.txt

tapia_a_Page_067.txt

tapia_a_Page_203.txt

tapia_a_Page_100.txt

tapia_a_Page_146.txt

tapia_a_Page_095.txt

tapia_a_Page_144.txt

tapia_a_Page_108.txt

tapia_a_Page_023.txt

tapia_a_Page_025.txt

tapia_a_Page_128.txt

tapia_a_Page_093.txt

tapia_a_Page_141.txt

tapia_a_Page_135.txt

tapia_a_Page_116.txt

tapia_a_Page_150.txt

tapia_a_Page_133.txt

tapia_a_Page_081.txt

tapia_a_Page_063.txt

tapia_a_Page_221.txt

tapia_a_Page_098.txt

tapia_a_Page_239.txt

tapia_a_Page_030.txt

tapia_a_Page_069.txt

tapia_a_Page_026.txt

tapia_a_Page_120.txt

tapia_a_Page_139.txt

tapia_a_Page_032.txt

tapia_a_Page_211.txt

tapia_a_Page_062.txt

tapia_a_Page_233.txt

tapia_a_Page_216.txt

tapia_a_Page_236.txt

tapia_a_Page_118.txt

tapia_a_Page_165.txt

tapia_a_Page_038.txt

tapia_a_Page_170.txt

tapia_a_Page_076.txt

tapia_a_Page_227.txt

tapia_a_Page_224.txt

tapia_a_Page_158.txt

tapia_a_Page_050.txt

tapia_a_Page_204.txt

tapia_a_Page_101.txt

tapia_a_Page_103.txt

tapia_a_Page_192.txt

tapia_a_Page_077.txt

tapia_a_Page_043.txt

tapia_a_Page_005.txt

tapia_a_Page_024.txt

tapia_a_Page_229.txt

tapia_a_Page_007.txt

tapia_a_Page_066.txt

tapia_a_Page_106.txt

tapia_a_Page_092.txt

tapia_a_Page_085.txt

tapia_a_Page_174.txt

tapia_a_Page_240.txt

tapia_a_Page_070.txt

tapia_a_Page_040.txt

tapia_a_Page_186.txt

tapia_a_Page_242.txt

tapia_a_Page_127.txt

tapia_a_Page_173.txt

tapia_a_Page_091.txt

tapia_a_Page_194.txt

tapia_a_Page_200.txt

tapia_a_pdf.txt

tapia_a_Page_159.txt

tapia_a_Page_107.txt

tapia_a_Page_137.txt

tapia_a_Page_090.txt

tapia_a_Page_014.txt

tapia_a_Page_184.txt

tapia_a_Page_048.txt

tapia_a_Page_042.txt

tapia_a_Page_068.txt

tapia_a_Page_027.txt

tapia_a_Page_172.txt

tapia_a_Page_115.txt

tapia_a_Page_201.txt

tapia_a_Page_099.txt

tapia_a_Page_097.txt

tapia_a_Page_060.txt

tapia_a_Page_152.txt

tapia_a_Page_136.txt

tapia_a_Page_055.txt

tapia_a_Page_001.txt

tapia_a_Page_197.txt

tapia_a_Page_110.txt

tapia_a_Page_089.txt

tapia_a_Page_079.txt

tapia_a_Page_046.txt

tapia_a_Page_049.txt

tapia_a_Page_161.txt

tapia_a_Page_117.txt

tapia_a_Page_220.txt

tapia_a_Page_018.txt

tapia_a_Page_206.txt

tapia_a_Page_147.txt

tapia_a_Page_140.txt

tapia_a_Page_109.txt

tapia_a_Page_145.txt

tapia_a_Page_021.txt

tapia_a_Page_198.txt

tapia_a_Page_190.txt

tapia_a_Page_112.txt

tapia_a_Page_075.txt

tapia_a_Page_154.txt

tapia_a_Page_208.txt

tapia_a_Page_123.txt

tapia_a_Page_151.txt

tapia_a_Page_210.txt

tapia_a_Page_142.txt

tapia_a_Page_234.txt

tapia_a_Page_199.txt

tapia_a_Page_163.txt

tapia_a_Page_013.txt

tapia_a_Page_178.txt

tapia_a_Page_223.txt

tapia_a_Page_124.txt

tapia_a_Page_138.txt

tapia_a_Page_059.txt

tapia_a_Page_083.txt

tapia_a_Page_004.txt

tapia_a_Page_168.txt

tapia_a_Page_212.txt


Full Text











COMPUTATIONAL DESIGN OF NICKEL BASED SUPERALLOYS FOR
INDUSTRIAL GAS TURBINE COMPONENTS














By

ALMA STEPHANIE TAP IA


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2006
































Copyright 2006

by

Alma Stephanie Tapia















ACKNOWLEDGMENTS

The author would like to thank and to acknowledge Dr. Gerhard Fuchs, Dr.

Reza Abbaschian, Dr. Robert DeHoff, and Dr. Hans Jurgen Seifert for their

support and guidance in this project. Special thanks go out to Allister James and

David Hunt of Siemens Westinghouse Power Generation (SWPC) in Orlando,

FL, who dedicated resources and time to make this project possible, and to the

high temperature alloys group. Additional thanks go to Wayne Acree and the

staff of the Major Analytical Instrument Center (MAIC) at the University of Florida.

This material is based on work supported by the Department of Energy.
















TABLE OF CONTENTS

page

ACKNOWLEDGMENTS .......................... ................. ...............iii

LIST OF TABLES....................... .................... ........ix

LIST OF FIGURES ............................................................. xi

ABSTRACT ....................... .......................... xvii

CHAPTER

1 INTRODUCTION.............................. ....... ........1

2 LITERATURE SEARCH......................... ......... ...........5

Microstructure ..................... .................... ........ 5
The y M atrix ............................................... .... 5
The y' Phase............................................ .... 5
The yly' Mismatch...... ....... ................ ......... 7
Carbides ....................... ..................... ......... 7
Phase Instabilities.............. ................ ........ ........... 8
Topologically close packed (TCP) phases .................................. 8
Secondary reaction zones (SRZs) ................................................. 9
Deleterious phase considerations ............................................... 10
Chemical Composition .............................. ............... 11
Strengthening Methods ........................... ..... .................. 11
Solid Solution Strengthening ................... .................. 11
Precipitation Hardening ........._.... ............... ..........12
Alloying Elements ........................ ............. ............ 12
Cobalt........................................... 13
Carbon ...................... ..................... ........ 13
Ruthenium...................... ................ ......... 13
Rhenium........................ .......................... 13
Chromium..................... ........................... 14
Aluminum and Titanium............................... 14
Tantalum and Tungsten ........._. ..........................16
Molybdenum....................................... 16
Hafnium........................ ........................... 17
Casting and Processing .............................. ........... 17









Casting Concerns and Defect Formation................................ 17
Solutionizing/Homogenization Heat Treatments.................................. 18
Predictive Methods....................................... 19
PHACOMP ............... .... ......... ........ ......... 20
New PHACOMP ........... ... ........ ......... .. .......... 22
NASA Rene N6 Model ................ .......... ......... ........ 23
Secondary Reaction Zone (SRZ) Model....................... 24
CALPHAD........................................... 24

3 DESIGN AND EXPERIMENTAL PROCEDURE ............... ............... 28

Alloy Design ................. ............. ...... .............. .................. 28
Alloy Development Model Base Chemistry ................................. 30
"Phase I" Alloy Development Modeled Elemental Variations ............ 31
"Phase II" Alloy Development Computational Alloy Refinement........ 37
Baseline Model A alloy ....................... .............. 37
M odeled elem ental variations........................................... 38
"Phase III" Alloy Development Experimental Validation .................... 41
M materials ............... ............ .......................... ............. 43
Solution Heat Treatment .......... .................. ............................. 44
Differential Thermal Analysis....................................... 46
M icro sco py .........4............... ...................... 4 8
Segregation............................. ................ 49

4 RESULTS ....................................... .................... 54

Phase I Modeled Elemental Variations ............... .............................. 54
Baseline Model Alloy ...... ............... ..... ......... ...... ......... 54
Microstructural stability........................... ...... 54
Phase transformation temperatures....................... ................ 55
Elemental segregation........................... ......... 56
Elemental Variation Effects................ .................... 57
Chrom ium variation effects............................................. 57
Aluminum (and Tantalum) variation effects............... ............... 61
Titanium (and Tantalum, Aluminum) variation effects ..................... 65
Rhenium (and Tantalum, Tungsten) variation effects................... 69
Carbon variation effects ...... ........................ ............... 73
Cobalt variation effects........... ..... ........................... 77
Ruthenium variation effects......... ............... ................... 80
Tungsten (and Molybdenum) variation effects .............................. 83
Gamma prime former (Tantalum, Aluminum, and Titanium)
variation effects ............. ... .... .. ... ... .................. 87
Temperature Range Comparisons for "Phase I" of Alloy Development. 91
Elemental Variation Trend Summary for "Phase I" ........ ........... 93
Phase II Computational Alloy Refinement.................... ........................... 95
Microstructural Stability............................... 95
Phase Transformation Temperatures....... .... ............................... 95









Elemental Segregation .............................. .......... 96
Elemental Variation Effects................ .................... 96
Rhenium variation effects......... ......... .. ... .......... ....... 96
Chromium (Aluminum and Titanium) variation effects................ 101
Gamma prime former (Tantalum, Aluminum, and Titanium)
variation effects ................ ... ....... ......... .......... 105
AI/Ti Ratio variation effects............................. 110
Elemental Variation Trend Summary for "Phase II" ...... ............ 114
Phase III Experimental Validation ........ .............. ......... 114
M icrostructural Stability................ ............. ............... 116
Experimetal results: microstructural characterization .................... 116
Computational results: JMatPro equilibrium phase predictions..... 129
Experimental to computational comparisons............... ............... 131
Elemental variation effects ............. ..................................... 131
Phase Transformation Temperatures .......................................... 133
Experimental results: DTA results ...................................... 133
Computational results: JMatPro phase transformation
temperatures ............. .. ....... ........ ............ ........ .... 140
Experimental to computational comparisons.............................. 141
Temperature range comparisons for "Phase III" of alloy
development ................ .................... ... ........ 144
Elem ental variation effects ................................. ..... ...... ...... 145
Elemental Segregation .......... ............... .. ............ 149
Experimental results: EMPA/WDS microprobe analysis.............. 149
Computational results: JMatPro solidification predictions............ 149
Experimental to computational comparisons.............................. 150
Elem ental variation effects ................................. ..... ...... ...... 152

5 DISCUSSION............................................. 157

Microstructural Stability ............... .................. ........... 159
Elemental Variation Effects.................................. 159
Carbon variation effects .................................. .. ......... ... ....... 159
Cobalt variation effects.................................. ..... ...... ......... 160
Ruthenium variation effects......... .................... 160
Tungsten (and Molybdenum) variation effects ............................ 161
Gamma prime former (Tantalum, Aluminum, and Titanium)
variation effects .............. ... .. ......... .. .......... 161
Aluminum (and Tantalum) variation effects.......................... 163
Titanium (and Tantalum, Aluminum) variation effects ................... 164
AI/Ti ratio variation effects ............. .... .......... .............. 165
Chromium variation effects................................ 166
Rhenium variation effects............................................ 168
Experimental to Computational Material Microstructure Comparisons 170
Phase Formation .......... .......... ................ .... 171
Phase Transformation Temperatures................................... 173
Elemental Variation Effects.................................... 174









Carbon variation effects ...................... .................... 174
Cobalt variation effects....................................... 175
Ruthenium variation effects............... ................ ...... ......... 175
Tungsten (and Molybdenum) variation effects ............................ 176
Gamma prime former (Tantalum, Aluminum, and Titanium)
variation effects ......................... ........... .. .. .......... 177
Aluminum (and Tantalum) variation effects.......................... 179
Titanium (and Tantalum, Aluminum) variation effects ............... 180
AI/Ti Ratio variation effects............ .................... 181
Chromium variation effects................................ 183
Rhenium variation effects.......................... ....... 185
Experimental to Computational Temperature Range Comparisons..... 186
Segregation............................................. 188
Elemental Variation Effects................ ................... 189
Carbon variation effects ......................................... 189
Cobalt variation effects............ ... ..... ........... ............. 190
Ruthenium variation effects............ ........... ......................... 190
Tungsten (and Molybdenum) variation effects ............................ 191
Gamma prime former (Tantalum, Aluminum, and Titanium)
variation effects ......................... ........... .. .. .......... 192
Aluminum (and Tantalum) variation effects.......................... 194
Titanium (and Tantalum, Aluminum) variation effects ............... 195
AI/Ti Ratio variation effects................ .................. 195
Chromium (Aluminum and Titanium) variation effects................ 197
Rhenium (and Tantalum, Tungsten) variation effects................. 199
Experimental to Computational Partitioning Comparisons.................. 202
Compositional Refinement ................................. .............. 204
Compositional Modifications ............... ....................................... 205
Alloy Comparisons .......................... ......................... ........... 206
Future Developm ent ................. ................. ................ .......... 210

6 CONCLUSIONS......................................... 214

Microstructural Stability ............... .................. ........... 214
Phase Transformation Temperatures........ ......... ...... ................ 214
Segregation ............................................ 214
Elemental Variation Effects ............................ ........... 214
Future Development...................................... 215

7 FUTURE WORK ............. .... ........ ................216

Continued Computational Modeling ................................... 216
Microstructural Stability Evaluations..................................... 216
Further Development of Alloy 1 ..................................... 217

LIST OF REFERENCES ........ ................... ..................................... ... ......218









BIOGRAPHICAL SKETCH .................. ........................ 224















LIST OF TABLES


Table page

3-1 Nominal composition in wt% of commercial/experimental Ni-base
superalloys........................................... 30

3-2 Model alloy composition in wt% and at%. .................... ............ 30

3-3 Baseline Model alloy composition and 'Phase I' variant compositions..... 35

3-4 Baseline Model A alloy composition in wt% and at%.............................. 38

3-5 Model A and "Phase II" design alloy's chemical compositions in wt%
and at%.............................................. 40

3-6 "Phase III" compositional variants with respect to the baseline Model A
alloy ........... .. ... ...................... ......... 41

3-7 "Phase III" alloy compositions and variation groups in wt% ..................... 42

3-8 Heat treatment used for the IGT experimental alloys............................. 45

4-1 Baseline Model alloy composition in wt% and at%. ............................... 54

4-2 Predicted partitioning coefficient values (kx,calc) for the 'Phase I'
baseline M odel alloy ............... ...................................... ......... ... 56

4-3 Predicted elemental variation effects on microstructural stability, phase
transformation temperatures and elemental segregation..................... 94

4-4 Baseline Model A alloy composition in wt% and at%.............................. 95

4-5 Predicted elemental variation effects on microstructural stability, phase
transformation temperatures and elemental segregation................... 115

4-6 "Phase III" alloy compositions and variation groups in wt% and at%..... 116

4-7 Predicted phase transformation temperatures (oC) and ranges for
'Phase III' alloys .......... .... ... ......... .............. 141

4-8 Predicted and experimental phase transformation temperatures (oC)
and ranges, for 'Phase III' alloys......... ............. .. ....... ........ .. 141









4-9 'Phase III' phase transformation temperature deviations from
experimental values (calculated experimental) and modeling error
((deviance/experimental)*1 00).................. .................. ........... ..... 142

4-10 Experimental partitioning coefficient values (kx,exp) for as-cast 'Phase
III' alloys ....... ......... ................... ........ 149

4-11 Predicted partitioning coefficient values (kx,calc) for as-cast 'Phase III'
alloys............................................ .. 149

4-12 Partitioning coefficient deviations (calculated experimental) and
modeling error ((deviance/experimental)*100) for as cast 'Phase III'
alloys............................................... 150

6-1 Alloy 1 composition ........... ......... ............ ... ...... 215

7-1 Recommendations in approximate wt%....... .............. ......... 217















LIST OF FIGURES


Figure page

1-1. Mitsubishi 701 Gas Turbine Engine....................................... 1

2-1. Typical FCC L12 y' crystal structure ............... ............... .................. 6

2-2. Transmission electron micrograph showing cuboidal y' particles in a y
matrix for a Ni-9.7AI-1.7Ti-17.1 Cr-6.3Co-2.3W at% alloy ............. 7

2-3. TEM image of a o plate in a SC Ni-based superalloy (SCA) .................... 9

2-4. Particle diameter vs. hardness for Ni-22Cr-2.8Ti-3.1Al wt% alloy ........... 12

2-5. Phase fraction diagram for SAF 2507 Duplex stainless steel.................... 26

3-1. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 2 from 'Phase III' compositions in the heat treated
condition ....... ..................... .. ...................... ........... 47

3-2 Button alloy sectioning and mounting orientation in metallographic analysis 48

3-3. EMPA/WDS compositions in normalized wt% versus line scan
measurement points (pm) for experimental Alloy2................ ............... 51

3-4 Schematic representation of solidification occurring in a eutectic binary
phase diagram. ............. .... ......... ................. 52

4-1. Predicted phase fraction diagram for baseline Model alloy .................... 55

4-2. Predicted phase diagram for the baseline Model alloy............................. 56

4-3 Predicted Cr variation effects on TCP equilibrium phase amounts........... 58

4-4 Predicted Cr variation effects on phase transformation temperatures.......... 59

4-5. Predicted Cr variation effects on elemental segregation ......................... 60

4-6. kcalc comparisons between the baseline Model alloy and Cr variants....... 61

4-7. Predicted Al (and Ta) variation effect on TCP equilibrium phase amount
with respect to Al (wt%) concentration ...... ........................................... 62









4-8. Predicted Al (and Ta) variation effects on phase transformation
temperatures with respect to Al (wt%) concentration.............................. 63

4-9. Predicted Al (and Ta) variation effects on elemental segregation with
respect to Al (wt%) concentration. ................... .................. 64

4-10. kcalc comparisons between the baseline Model alloy and Al (and Ta)
variants. ................... ...................................... ............ ......... 65

4-11. Predicted Ti (and Ta or Al) variation effects on TCP equilibrium phase
amount with respect to Ti (wt%) concentration. ................. ....... ....... 66

4-12. Predicted Ti (and Ta or Al) variation effects on phase transformation
temperatures with respect to Ti (wt%) concentrations .............................. 66

4-13. Predicted Ti (and Ta or Al) variation effects on elemental segregation
with respect to Ti (wt% ) concentration. .............. ..................................... 68

4-14. kcalc comparisons between the baseline Model alloy and Ti variants...... 69

4-15. Predicted Re (with Ta, Al, and W) variation effects on TCP equilibrium
phase amounts with respect to Re concentration (weight percent)............ 70

4-16. Predicted Re (with Ta, Al, and W) variation effects on phase
transformation temperatures with respect to Re (wt%) concentration........ 71

4-17. Predicted Re (with Ta, Al, and W) variation effects on elemental
segregation with respect to Re (wt%) concentration.............................. 72

4-18. kcalc comparisons between baseline Model alloy and Re variants.......... 73

4-19. Predicted C variation effect on TCP equilibrium phase amount ............... 74

4-20 Predicted C variation effects on phase transformation temperatures......... 75

4-21. Predicted C variation effects on elemental segregation. ....................... 76

4-22. kcalc comparisons between the baseline Model alloy and C variants...... 76

4-23. Predicted Co variation effect on TCP equilibrium phase amount .......... 77

4-24. Predicted Co variation effects on phase transformation temperatures.... 78

4-25. Predicted Co variation effects on elemental segregation ...................... 79

4-26. kcalc comparisons between Model alloy and Co variants .................... 80

4-27. Predicted Ru variation effect on TCP equilibrium phase amount .......... 80









4-28. Predicted Ru variation effects on phase transformation temperatures..... 81

4-29 Predicted Ru variation effects on elemental segregation ....................... 82

4-30 kcalc comparisons between Model alloy and Ru variants ..................... 83

4-31. Predicted W (and Mo) variation effect on TCP equilibrium phase amount
with respect to W (wt%) concentration. ............. ................................... 84

4-32. Predicted W (and Mo) variation effects on phase transformation
temperatures with respect to W (wt%) concentration.............................. 84

4-33. Predicted W (and Mo) variation effects on elemental segregation with
respect to W (wt%) concentration. ............... .......... ....... ........ 85

4-34. kcalc comparisons between Model alloy and W (and Mo) variants.......... 86

4-35. Predicted y'-former variation effects on TCP equilibrium phase amounts. 88

4-36. Predicted y'-former variation effects on phase transformation
temperatures................ .................... ........ ......... 89

4-37. Predicted y'-former variation effects on elemental segregation................ 90

4-38. kcalc comparisons between Model alloy and y'-former variants........ 91

4-39. Calculated heat treatment window soliduss y' solvus) vs. melting range
(liquidus solidus) for the baseline Model composition, the compositional
variants in 'Phase I' of alloy development, and selective 1st and 2nd
generation commercial and experimental alloys. ............ .. ............... 92

4-40. Predicted Re variation effects on TCP equilibrium phase amounts with
respect to Re (wt%) concentration ............. ...................... ............ 98

4-41. Predicted Re variation effects on phase transformation temperatures
with respect to Re (wt%) concentration .................................. 100

4-42. Predicted Re variation effects on elemental segregation with respect to
Re (wt%) ........... ........... .............. ......... .......... 102

4-43 Predicted Cr (with Ti and Al) variation effects on TCP equilibrium phase
amounts with respect to Cr (wt%) content......................... ......... 103

4-44 Predicted Cr (with Ti and Al) variation effects on phase transformation
temperatures with respect to Cr (wt%) content. .............. .... ........... 104

4-45. Predicted Cr (and Ti and Al) variation effects on elemental segregation
with respect to Cr (wt%) content. ............... .......... ....... ........ 105









4-46. Predicted y'-former variation effects on TCP equilibrium phase amounts. 106

4-47. Predicted y'-former variation effects on phase transformation
temperatures ....... ...... .................... ......... 108

4-48. Predicted y'-former variation effects on elemental segregation ........... 109

4-49. Predicted elemental variation effects on the amount of TCP equilibrium
phases with respect to AI/Ti ratio ........ ............................... ..... 111

4-50. Predicted elemental variation effects on phase transformation
temperatures with respect to AI/Ti ratio................................... 112

4-51 Predicted elemental variation effects on elemental segregation for AI/Ti
variants with respect to AI/Ti ratio .............. ......... .... ............ 113

4-52. Alloy 1 in the as-cast condition .............. .......... ...... ......... ..... 117

4-53. Microstructural characteristics for Alloy 1 in the heat treated condition.. 118

4-54. Alloy 2 in the as-cast condition......................................... 119

4-55. Alloy 2 in the heat treated condition ............. ........ ..................... 120

4-56. Alloy 3 microstructure in the as-cast condition. .................... 123

4-57. Alloy 3 microstructure in the heat treated condition.............................. 123

4-58 Alloy 4 microstructure in the as-cast condition ............... ... ........... 125

4-59. Material microstructure for Alloy 4 in the heat treated condition........... 126

4-60. Alloy 5 in the as-cast condition......................................... 127

4-61. Alloy 5 material microstructure in the heat treated condition .................. 128

4-62. Predicted AI/Ti ratio (and Ta) variation effects on TCP equilibrium phase
amounts with respect to AI/Ti ratio.. ...... ..... ........................ ............. 131

4-63. Predicted Cr (with Al and Ta) variation effects on TCP equilibrium phase
amounts with respect to Cr content......... ......... ..................... 132

4-64. Predicted Re variation effects on the amount of TCP equilibrium phases 33

4-65. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 1 in the as-cast condition ................... ....... 134

4-66. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 1 in the heat treated condition............................... 134









4-67. DTA temperature difference (AT) vs. specimen temperature curves for
experim ental Alloy 2 in the as-cast condition ........................................... 135

4-68. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 2 in the heat treated condition............................... 135

4-69. DTA temperature difference (AT) vs. specimen temperature curves for
experim ental Alloy 3 in the as-cast condition ........................................... 136

4-70. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 3 in the heat treated condition............................... 136

4-71. DTA temperature difference (AT) vs. specimen temperature curves for
experim ental Alloy 4 in the as-cast condition ........................................... 137

4-72. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 4 in the heat treated condition............................... 137

4-73. DTA temperature difference (AT) vs. specimen temperature curves for
experim ental Alloy 5 in the as-cast condition ........................................... 138

4-74. DTA temperature difference (AT) vs. specimen temperature curves for
experimental Alloy 5 in the heat treated condition............................... 138

4-75. Comparison between experimental and calculated phase transformation
temperatures for 'Phase III' alloys. ....... ........... ... ............. 142

4-76. Comparison between experimental and calculated melting ranges and
heat treatment windows for 'Phase III' alloys............... .............. 143

4-77. Heat treatment window vs. melting range for the baseline Model
composition, the baseline Model A composition, the compositional
variants in 'Phase III', and selective 1st and 2nd generation commercial
and experimental alloys. ............... ............... ............ 144

4-78. Predicted and experimental phase transformation trends for AI/Ti ratio
variants with respect to AI/Ti ratio. .................. ................. 146

4-79. Predicted and experimental phase transformation trends for Cr variants
with respect to Cr (wt%) concentration. .............. ........................... 147

4-80. Predicted and experimental phase transformation trends for Re variants
with respect to Re (wt%) concentration............................. ........ 148

4-81. Comparison between experimental and predicted partitioning coefficient
values (kexp vs kcal) for 'Phase III' alloys............................... 151

4-82. Experimental and predicted partitioning coefficient values for elements .152









4-83. Predicted and experimental segregation trends for AI/Ti variants with
respect to A I/T i (w ith Ta) variations.......................................................... 153

4-84 Predicted and experimental segregation trends for Cr variants with
respect to Cr (wt%) concentration. ..... .... ...................................... 154

4-85. Predicted and experimental segregation trends for Re variants with
respect to Re (wt%) concentration ...... .. .... ...................... .. ............ 156















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

COMPUTATIONAL DESIGN OF NICKEL BASED SUPERALLOYS FOR
INDUSTRIAL GAS TURBINE COMPONENTS

By

Alma Stephanie Tapia

May 2006

Chair: Gerhard E. Fuchs
Major Department: Materials Science and Engineering

Ni-based superalloys play an essential role in the advancement of power

technology. Recent initiatives to increase efficiency and decrease emissions in

power generating industrial gas turbines (IGT's) can only be met by increasing

turbine inlet temperatures and turbine component temperature capabilities. To

reach these target operating temperatures, traditional IGT processing will need to

transition to directional solidification processing, for single crystal turbine blade

production.

The successful use of single-crystal alloys in IGT applications is contingent

upon overcoming processing problems such as defect formation, and maintaining

microstructural stability once in service. Elemental segregation resulting from the

casting process, in particular, is linked to defect formation and the formation of

deleterious phases over the extended lifetime of the IGT component. To avoid

the formation of these deleterious phases, solutionizing heat treatments between









the y solvus and the solidus temperatures are used to reduce or eliminate

segregation from the as-cast materials.

To investigate how typical Ni-based superalloy elemental additions affect

microstructural stability, phase transformation temperatures, and material

segregation behavior, a baseline alloy composition was used as the foundation

from which two iterations of "elemental variation effect" evaluations were

conducted. Elemental trends were assessed using a thermodynamic equilibrium

module in the 3.0 Java-based Materials Properties Program (JMatPro).

Property trends from the "Phase I" and "Phase II" studies were used to

redefine the baseline alloy's chemistry. Compositional modifications resulted in

five experimental alloy compositions that were manufactured and experimentally

tested as a comparison to theoretical results.

The JMatPro thermodynamic equilibrium module was evaluated and

elemental relationships were assessed. Conclusions about elemental effects on

microstructural stability, phase transformation temperatures, and material

segregation were drawn, which may contribute to the development of better

alloys for single crystal IGT use in the future.


xviii














CHAPTER 1
INTRODUCTION

Over the last several decades, Ni-based superalloys have played an

essential role in the advancement of power technology. Characterized by their

high structural, surface, and property stability, Ni-based superalloys are widely

used in the highest temperature components of power generating industrial gas

turbines (IGTs), including turbine discs, turbochargers, blades, and vanes [3].

The need to increase power output (efficiency) and decrease emissions in

power-generating industrial gas turbines (IGT's), can only be met by increasing

turbine inlet temperatures and turbine component temperature capabilities.

Land-based industrial gas turbines (IGTs) operate at inlet temperatures

rapidly approaching 1500 'C for service lifetimes up to, or in excess, of 10,000 h

[45]. IGTs, such as the Mitsubishi 701 seen below (Figure 1-1), are exposed to

high temperatures and corrosive environments for a significant portion of their

lives, making their components susceptible to hot corrosion (or sulfidation)[1,65].












Figure 1-1. Mitsubishi 701 Gas Turbine Engine









Hot corrosion can be described as the accelerated surface attack of

components due to condensed alkali metal salts, such as Na2SO4. When a gas

turbine ingests air from the atmosphere to mix it with injected fuel for burning,

combustion gases remain that may be contaminated with corrosive impurities.

IGTs consume more air than fuel, with air-to-fuel consumption ratios of up to 50

to 1; thus, even a small amount of sodium chloride and sodium sulfate in the

atmosphere can react with residual sulfur in the fuel, leading to severe corrosion

problems [51]. Sodium chloride will react with sulfur to form Na2SO4 as shown

below.

2NaCI + S02 + 02 Na2SO4 + Cl2

Currently, the baseline alloy used for IGT applications is IN738. IN738 is a

polycrystalline material that exhibits good hot corrosion resistance but does not

meet the increasing temperature demands of the power industry. To increase

material temperature capabilities for IGTs, the use of single crystal (SC) turbine

blades will be required. Processing problems, such as castability, that come

from fabricating inherently large IGT components must be factored into the

selection of an appropriate SC IGT alloy. Some of the single crystal alloys

presently considered for IGT applications include CMSX-4 and PWA 1483.

PWA 1483 demonstrates an acceptable level of hot corrosion resistance

and castability for IGT applications but exhibits limited strength in comparison to

other single crystal alloys (i.e., CMSX-4). CMSX-4, developed for aerospace

applications, is a second -generation, single crystal alloy that exhibits high

strength at elevated temperatures but demonstrates poor hot corrosion









resistance and castability. Other alloys such as CMSX-11 B and CMSX-11 C are

first-generation, experimental alloys designed, specifically, for single crystal IGT

use. Both CMSX-11 B and CMSX-11 C demonstrate extremely good blends of

hot corrosion and oxidation resistance but are prone to recrystallization and

freckle formation (along with CMSX-4) during processing [10,34]. SC-16, a first-

generation single crystal alloy developed by Onera, is not commonly used in the

United States and may exhibit poor hot corrosion resistance.

The successful use of single-crystal alloys in IGT applications is contingent

upon overcoming processing problems such as defect formation, and maintaining

microstructural stability once in service. The increased number of elemental

additions in a Ni-based superalloy and the complex interactions of these

additions reveal a need to investigate elemental variation effects on

microstructural stability, phase transformation temperatures, and material

segregation behavior.

The present work uses a design approach aimed toward the development

of a set of alloys for industrial gas turbine application. In the hopes of better

understanding elemental variation effects on the aforementioned material

properties, a baseline alloy composition named the baseline 'Model' alloy (based

on CMSX-4 and PWA 1483) was used as the foundation from which two

iterations of 'elemental variation effect' evaluations were conducted ('Phase I'

and 'Phase II'). The thermodynamic equilibrium module in the 3.0 Java-based

Materials Properties Program (JMatPro) was utilize to evaluated 'Phase I' and

'Phase II' theoretical property trends and determine chemistry modifications to









the baseline 'Model' alloy. Five variant alloy compositions were tailored using

JMatPro modeling techniques and were laboratory tested for validation purposes.

To address the effects of additions previously shown to influence hot corrosion

and material stability, final compositions incorporated characteristic variations of

AI/Ti ratio (with Ta variation), Cr (with Al and Ta variations), and Re content for

comparison [35, 40, 45].

This study evaluates the computational capabilities of the JMatPro

thermodynamic equilibrium module to predict material properties related to defect

formation and microstructural stability. Through computational techniques this

work also contributes to a better understanding of elemental variation effects on

microstructural stability, phase transformation temperatures, and material

segregation behavior to facilitate the development of better alloys for future

single crystal IGT use.














CHAPTER 2
LITERATURE SEARCH

This chapter will provide an overview of the microstructure, chemical

composition, casting, and processing of Ni-based superalloys, as well as a

discussion of the current methods utilizing empirical and computational models to

predict deleterious phase formation and material properties.

Microstructure

A basic Ni-base superalloy consists, mainly, of a two-phase equilibrium

microstructure: the gamma (y) nickel-chromium matrix and the gamma-prime (y')

precipitate. Carbon additions can lead to the formation of carbides and certain

service/heat treatment conditions may result in the formation of deleterious TCP

phases.

The y Matrix

The continuous gamma matrix (y) is a solid solution FCC nickel based

austenitic phase, strengthened by high percentages of Co, Cr, Mo, W, Ti, and Al

[44,51].

The y' Phase

The y' phase is an intermetallic compound that provides strength to the Ni-

base superalloy [44]. The y' phase precipitates coherently out of the y matrix

with an FCC L12 ordered superlattice structure to become the material's major

precipitate (Figure 2-1) [44,51].



















Figure 2-1. Typical FCC L12 y' crystal structure

This L12 structure is of the Cu3Au-type, where Ni atoms occupy the centers

of the cube faces and Al typically resides in the cube corners [13]. Ti, Nb, and Ta

also contribute to y' precipitation and can substitute for up to approximately 50%

of the y'. The binary atomic arrangement has the chemical formula Ni3AI, Ni3Ti,

or Ni3(AI,Ti), which mainly consists of Al, Ti, Nb, or Ta [51]. The y' forming

elements, Ti, Nb, and Ta, also increase the y' anti-phase boundary energy

(yAPB) [7,35,45].

The y' phase is, in large part, responsible for the elevated-temperature

strength in a Ni-based superalloy; since the strength of y' actually increases with

increasing temperature [13]. The total y' former content in most Ni-base

superalloys is usually maintained at about 12-15 at%. The attractive properties

of y/y' superalloys has resulted in a continuous increase in the y' volume fraction.

Recent alloys may contain over 60% and can approach 75% in some Ni-based

superalloys [45]. However, increasing y' volume fractions, must be balanced with

modification of the y-matrix composition since a concentration of the refractory

elements in the y matrix, can lead to deleterious phase formation [34].









The yly' Mismatch

y' morphology is affected by lattice mismatch, strain energy, and interfacial

energy. The lattice mismatch, a result of the differences in y and y' lattice

parameters (ay and ay' respectively) result in an interfacial misfit energy [13].

The unconstrained lattice misfit parameter (d) is defined below.

d = (ay ay') / ay

In most superalloys, the y' precipitate is in tension while the y matrix is

under compression, leading to a small misfit. This small misfit results in a

cuboidal precipitate morphology and helps ensure a low y/y' interfacial energy

[13,33]. The transmission electron micrograph below (Figure 2-2) depicts a

typical cuboidal y' morphology [45,51].






Gamma prime (y')
precipitate

Gamma (y) matrix




Figure 2-2. Transmission electron micrograph showing cuboidal y' particles in a y
matrix for a Ni-9.7AI-1.7Ti-17.1 Cr-6.3Co-2.3W at% alloy

Carbides

Primary carbides, or MC carbides, form as discrete FCC particles during the

solidification of an alloy, and are typically observed throughout the material

[44,45]. MC carbides form due to interactions between carbon and reactive or









refractory metals, such as Ti, Ta, Hf, or Ta apart from other elements like Cr, Mo,

W, and Nb [45]. The formation of carbides in the material matrix consumes

refractory elements that contribute to solid solution strengthening or y' formation,

which could also promote phase instability during service [45].

Carbide morphologies range from cubic to script, but carbides are most

commonly seen as large blocky or spherical particles in superalloys. Primary

FCC close-packed carbides are some of the most stable compounds found in

nature [45].

Secondary carbides, of the M23C6 type, form through the decomposition of

MC type, primary carbides. The degenerations of MCs occur during lower

temperature heat treatments and service in the 760-980'C range in alloys

containing moderate to high amounts of Cr, apart from W and Mo.

Phase Instabilities

Topologically close packed (TCP) phases

Deleterious topologically close packed (TCP) phases can result from

microstructural/chemical instabilities in nickel-based superalloys, during the heat

treatment or service lifetime of a component [54]. TCP phases exist in many

forms, but typically appear in the sigma (o), miu ([t), or Laves form. The phases

have characteristic close-packed atom planes stacked in the sequence ABCABC

which are parallel to the {1 111} planes of the y matrix [45]. An example of a o

TCP phase; identified by Strunz, in an experimental nickel-base superalloy, is

seen below (Figure 2-3) [48].









The chemical formula (Cr,Mo)x(Ni,Co)y has been reported for the o phase,

where x and y vary from 1 to 7. In general, TCP phases are predominately made

up of refractory elements, and the t phase is characteristically dominated by Mo

and Co [45]. Accordingly, TCP phase formation results in the depletion of solid

solution strengthening elements such as W, Mo, Cr, Co and Re from the y matrix.

The depletion of these strengthening elements may produce a marked reduction

in rupture life at high temperatures [8]. The intrinsically brittle nature of the

topologically close-packed (TCP) phases reduces the ductility of an alloy. The

physical hardness and, many times, plate-like morphology of TCP phases also

provide a source for crack initiation and propagation, leading to material failure.














Figure 2-3. TEM image of a o plate in a SC Ni-based superalloy (SCA)

Secondary reaction zones (SRZs)

The occurrence of secondary reaction zones was noted by Walston et al. in

Rene N6 [44,61,62]. Secondary reaction zones (SRZs) are y' regions within a

material that contain y and P phase needles. These regions are referred to as

cellular colonies and can form in dendrite cores and along low angle boundaries,

common in single crystal castings [44].









SRZs are thought to form in areas of local elemental enrichment due to

either coating processes or material segregation (resulting from casting

processes), apart from factors such as strain energy and misfit strains [44,61].

These cellular colonies have been observed in superalloys containing high

concentrations of refractory elements, demonstrating the highest affinity for Re-

bearing alloys [25,62]. The presence of SRZs beneath material coatings can

eventually affect the rupture strength of a material and induce premature failures

when crack initiation occurs at SRZ interfaces [25,62].

Deleterious phase considerations

Overall, deleterious phases tend to nucleate in materials with excessive

additions of refractory elements or in areas that are enriched with high

concentrations of refractory elements [51]. Consequently, careful chemistry

control to balance alloy composition and effective homogenization treatments are

necessary to minimize regions of localized elemental enrichment and,

subsequently, prevent deleterious phase formation.

To restrict microstructural instabilities, limits have been introduced to the

concentrations of solid solution strengtheners [25,34]. Typical Re bearing alloys,

such as CMSX-4, CMSX-10, and Rene N6 exhibit deleterious phase formation of

either the TCP or SRZ type [1,25,62]. To improve 'long-tem' stability, Re bearing

alloys have, in good measure, reduced Cr and Ti concentrations (elements that

can provide the hot corrosion resistance vital to industrial gas turbine

applications) [35,44]. Studies conducted with Rene N6, showed that

microstructure stability could be improved by decreasing the level of Re in a

material and by introducing Mo in order to keep a comparable total amount of









strengthening refractory elements [44,61]. It is evident, therefore, that

modifications to alloy composition as a means to improve material stability can

be carefully balanced with specific IGT material needs.

Chemical Composition

The properties of Ni-base superalloys are strongly dependent upon the

chemical composition of the given material. The high solubility of Ni for a wide

variety of alloying elements is largely due to its partially filled third electron shell.

Typical Ni-based superalloy compositions can contain up to 12 to 13 different

elements, without sacrificing microstructural stability [1,2]. However, the

interaction all of these solutes raises the challenge of balancing alloy

compositions to obtain specific desired material properties.

Strengthening Methods

A key alloying effect is the solid solution strengthening and precipitation

hardening of a material.

Solid Solution Strengthening

Solid solution strengthening arises from solute interactions with dislocations

in the material's matrix. These solutes strengthen the material by introducing

atomic diameter differences, elastic interactions, modulus interactions, electrical

interactions, and short-range/long-range order interactions [44,45]. The lowering

of the stacking fault energy with alloying additions also increases resistance to

cross slip and dislocation motion, thus, increasing material strength. Solid

solution strengtheners include Re, W, Mo, Cr, Co, Ti and Al [45].










Precipitation Hardening

Precipitation hardening of a material results from dislocation interactions

with coherent particles within a material matrix. Material strength increases with

particle size, due to an increased amount of dislocation cutting that occurs in the

larger coherent precipitates in the y matrix. Once a critical particle size is

reached; precipitates become incoherent with the y matrix. At this critical stage,

dislocations begin to bypass precipitates, resulting in decreased material

strength. The relationship between strength and precipitate size was shown in a

study by Mitchell, for a Ni-22Cr-2.8Ti-3.1Al wt% alloy (Figure 2-4)[45].

400
Aging Tempsrlawu, C
650 W1"
a ~55
x 700 x
S750
A 800 x
350 -











10 10 10
Mean Partiia Diametw, A

Figure 2-4. Particle diameter vs. hardness for Ni-22Cr-2.8Ti-3.1AI wt% alloy

Alloying Elements

Common superalloy elemental additions (Co, C, Cr, Mo, W, Al, Ti, Ru, Re,

and Ta) and some of their key characteristic influences on Ni-based superalloys

are mentioned below.









Cobalt

Cobalt concentrations of approximately 2-15 wt% are used in most Ni-

based superalloys [44]. Co additions stabilize the material microstructure,

reduce the y' solvus temperature, and reduce the stacking fault energy (YSFE)

[11,44]. Co is also reported to partition to the dendrite core, and provides a

limited amount of solid solution strengthening [44,45].

Carbon

Minor additions of carbon can be used in Ni-based superalloys. A study of

Rene N4 showed that a 0.05 wt% C addition yielded increased rupture strength

at high temperatures [35]. An increased tolerance for grain boundary

misorientations at low angle boundaries (LABs) was also attributed to C additions

[16,35]. C additions have also been shown to decrease refractory element (W,

Re) partitioning to the dendrite core, improving microstructural stability [50].

Ruthenium

Ruthenium additions of approximately 0 to 9 at% in Ni-based superalloys,

stabilize the material's microstructure and provide the material with solid solution

strengthening [13,28,45]. Ru additions also increase the liquidus temperature

and tends to partition to the dendrite core [45,51].

Rhenium

Rhenium is a strong solid solution strengthener that improves creep

strength, and increases the temperature capabilities of a material [51]. The

amount of Re used in an alloy categorizes it as a first, second, or third generation

alloy, containing 0 wt% Re, 3 wt% Re, or 6 wt% Re, respectively. Rhenium

additions are expected to increase the density and liquidus temperature of a Ni-









base superalloy [51]. The use of this refractory element may result in convective

instabilities during solidification. Re also partitions to the dendrite core, which

can lead to the development of deleterious phases in the dendritic region [44,51].

The high temperature strength supplied by Re additions must be balanced with

the concern for decreasing microstructural stability. Refractory elements,

including Cr and Ti concentrations (additions that increase hot corrosion

resistance), have been reduced in more recent alloys to compensate for Re

additions [51].

Chromium

Chromium concentrations typically range from 10-20 wt% for industrial gas

turbine applications. Cr additions improve hot corrosion and oxidation resistance

due to the formation of a protective Cr203 rich oxide scale [35,44]. The Cr oxide

hinders diffusion and effectively stops environmental reaction with the bulk alloy.

Cr has also been reported to reduce the y' solvus temperature and the anti-

phase boundary energy (yAPB) of the y' phase [9]. Cr tends to partition to the

dendrite core and may promote deleterious phase formation [24].

To improve microstructural stability, 2nd and 3rd generation alloys have

notably reduced Cr concentrations [51]. The low Cr concentrations in the higher

generation aero-alloys could result in hot corrosion concerns for IGT components

that require an extensive amount of hot corrosion resistance [44].

Aluminum and Titanium

* Aluminum

Aluminum concentrations typically used in most Ni-base superalloys range

between approximately 3-6 wt%. Aluminum is a low density addition to Ni-based









superalloys, which acts as a primary y' former, improves material castability, and

partitions to the inderdendritic region [12,45]. Al additions are also considered

essential for oxidation resistance [35]. The Al solute contributes to the

development of an A1203 scale, which increases its oxidation resistance at high

temperatures [45].

* Titanium

Titanium concentrations in most Ni-based superalloys can range from 0 to 5

wt%. Titanium is also a low density addition to Ni-based superalloys, which acts

as a y' former, strengthens the y' phase, and increases the y' anti-phase

boundary energy (YAPB) [7,45]. Ti partitions to the interdendritic region and

generally decreases the oxidation resistance and increases the hot corrosion

resistance of the alloy [24,35].

* Al/Ti Ratio

AI/Ti ratio is used to illustrate the influence of Al and Ti on the oxidation and

corrosion resistance of an alloy. Ross and O'Hara, reported that the AI/Ti ratios

in Rend N4 had a significant impact on oxidation and hot corrosion resistance

[35]. The study on Rend N4 showed that decreasing AI/Ti ratios, increased hot

corrosion resistance, but decreased oxidation resistance [35]. Consequently,

using lower AI/Ti ratios (moderate Al concentrations with increased Ti additions)

can provide IGT components with increased resistance to hot corrosion attack.

Recent Ti reductions in 2nd and 3rd generation alloys could then result in the

degradation of hot corrosion properties that are so important for IGT applications

[51].









Tantalum and Tungsten

* Tantalum

Ta concentrations of approximately 4 to 12 wt% are used in many Ni-based

superalloys [44]. Ta is a y' former and acts as a strong solid solution

strengthener [45]. Ta additions increase the y' anti-phase boundary energy

(YAPB) and tend to partition to the interdendritic region [24,53]. Ta has also been

reported to improve alloy castability [31].

* Tungsten

Tungsten additions of approximately 5-8 wt%, strengthen the y matrix

through solid solution strengthening and synergistic effects with Re strengthening

mechanisms [6]. The use of W in superalloys is reported to increase the incipient

melting point, decrease microstructural stability, and increase hot corrosion

susceptibility [45]. W partitions to the dendrite core and decreases material

castability [6,45].

* Ta/W Ratio

The Ta/W ratio is also used to evaluate an alloy's castability. Increased

Ta/W ratios (from increased Ta or reduced W concentrations), are reported to

decrease the incidence of casting defects, caused by convective instabilities

during processing [34,38].

Molybdenum

Molybdenum additions of 0-3 wt%, are used to increase solid solution

strengthening of the y' matrix [6,15]. Mo decreases microstructural stability and

has been reported to partition to the dendrite core [15,16,21].









Hafnium

Hafnium additions of 0-0.2 wt% are used to increase oxide scale adherence

to the metal substrate. Minimal Hf additions enhance coated oxidation life by

diffusing into a metal's surface oxide [10,45].

Casting and Processing

Casting Concerns and Defect Formation

The first solids to form during solidification are gamma dendrites [54].

Solute and solvent fluxes during dendrite growth cause solvent/solute buildups

that are unable to redistribute completely before solidification is complete. The

supersaturation of the liquid with segregating elements, results in the formation of

secondary solidification constituents, such as MC carbides, in the interdendritic

regions [54]. Elemental build ups in the dendritic solid and interdendritic liquid

result in material segregation.

The degree of segregation in a material is, typically, measured with the use

of elemental partitioning coefficients (k'). The partitioning coefficient for a given

element (x) is expressed as the ratio of interdendritic to dendritic composition; as

seen below.

kx -= Cx,corel Cx,inter

A partitioning coefficient (k') value of one indicates that no partitioning is

present for a given element. More succinctly, an equal amount of the element

was measured in both the dendrite core and the interdendritic region, exhibiting

no preference or "segregation" during solidification. The ratio also allows the

direction of segregation to be determined. Partitioning coefficients (kx') less than









unity, indicate that an element partitions to the interdendritic region. Solutes with

partitioning coefficients greater than unity segregate to the dendrite core.

Elements such as Ni, Ta, and Al, have been previously reported to

segregate toward interdendritic regions. Cr, Co, W, and Re have been

previously reported to segregate to the dendrite core [4,45].

The loss of control over temperature gradients in conventional casting

techniques, can also lead to convex solidification interfaces that result in material

segregation and defect formation [54,64]. Small castings, such as those used in

the aero engine field, can more easily maintain steep temperature gradients as

compared to large IGT components. The difficulty in maintaining these gradients

in large single crystal blades, results in a high propensity towards material

segregation and defect formation. There are several defects associated with the

casting of single crystal structures, such as low angle boundaries, slivers, and

freckles [13,54]. Defect formation and solute partitioning may be controlled

through a combination of alloy design and careful control of the casting process.

Solutionizing/Homogenization Heat Treatments

Elemental segregation, resulting from the casting process, is typically

reduced or eliminated by solutionizing/homogenization heat treatments. Solution

heat treatments are conducted at temperatures high enough to dissolve the y'

phase and homogenize the alloy for the, subsequent, re-precipitation of uniform

y' precipitates in the material.

Solutionizing heat treatments are limited to a temperature range between

the y' solvus and the solidus, called the 'heat treatment window.' Complete

homogenization is dependant on both the temperature and time of the heat









treatment. The ability to homogenize cast structures may be severely restricted

by low or even negative heat treatment windows; which can result in incipient

melting [3]. A study conducted for an experimental Ni-base superalloy (Re3)

demonstrated that solution heat treatments were unable to fully homogenize the

material due to minimized heat treatment times. The Re3 solutionizing heat

treatments were reduced in time, to avoid the risk of incipient melting, due to the

material's narrow heat treatment window [33]. Inadequate solutionizing can

result in areas of residual elemental enrichment which can then lead to

deleterious phase formation.

Predictive Methods

The development of new Ni-based superalloys for the modern gas turbine

have primarily been the result of trial and error processes. The task of identifying

new alloys that provide increasing temperature capabilities yet balance good

microstructural stability is becoming increasingly difficult. Given the high degree

of complexity in nickel based superalloy chemistry, it is of no surprise that a

major area of concern and attention has been microstructural stability [75]. This

was evident in the microstructural stability difficulties faced for 720Li, a high

strength nickel-base superalloy used for turbine disk applications. This turbine

disk alloy, in the powder or conventionally processed cast and wrought form,

suffers rapid precipitation of the TCP o phase above 6500C. Property

degradation concerns encouraged C.J. Small and N. Saunders to investigate

new alloy compositions based on 720 Li and Waspaloy [46].

With the need for a tool that can guide initial alloy-chemistry selection,

semi-empirical and computational models have been developed to predict









deleterious phase formation and/or simulate material properties based solely on

the alloy composition with some success. Existing tools for design applications

will be described in more detail below.

PHACOMP

Early attempts to predict TCP phase formation were based on the PHAse

COMPutation (PHACOMP) method. This conventional calculation tool relates

chemical composition to electron valence theory to predict the formation of

deleterious phases in an alloy.

PHACOMP relies on the importance of electronic interactions (the unpaired

d-electrons or electron vacancies) for each element in an alloy. The average

electron vacancy concentration (Nv) of the solid solution matrix is calculated on

the premise that closely packed phase instabilities (such as o) are electronic

compounds. In the PHACOMP technique, the alloy matrix composition is

calculated by subtracting the normal precipitation phases carbidess, borides, y')

from the total composition before the average electron-hole concentration (Nv) is

computed as follows.[4]

Nv = Zfi (nv)i

where fi is the atomic fraction of an element (i) in the y matrix and (nv)i is its

corresponding electron-hole number. A typical equation to calculate an alloy's

electron-hole weighted average is shown below [45].

Nv = 4.66 (Cr + Mo) + 3.66(Mn) + 2.66(Fe) + 1.71 (Co) + 0.61 (Ni)

The calculated Nv value is compared to some critical value that determines

whether the alloy is prone to sigma-phase precipitation. An average electron









hole number Nv above the threshold would indicate that an alloy is sigma-prone,

while a number below the critical value would deem the material "sigma safe"

and suitably stable for practical applications [71]. The critical value approximates

2.5 for individual alloys but is not necessarily a fixed value common to all metals.

The model's oversimplified nature calculates the matrix composition based

on assumptions of the amount, type, and composition of carbides, borides, and

intermetallic compounds that are expected to precipitate during processing and

service [45]. Dreshfield and Ashbrook studied a wide range of cast and wrought

Inconel alloys using the PHACOMP method. Although electron vacancy number

calculations indicated a tendency to form the sigma phase, no deleterious

phases were observed in any of the examined materials [8]. It is evident that the

accuracy of PHACOMP predictions depends on the validity of the assumptions

made for estimating the composition of the y matrix. Work conducted by Milhalis

on a variety of alloys used experimentally determined matrix compositions in

PHACOMP analysis, resulting in more accurate predictions of material phase

stability [14]. Difficulties with using the PHACOMP method are described by

Murphy et. al. who concluded that PHACOMP calculations are not accurate

unless the compositions of the precipitating phases are available [45,62].

The PHACOMP method lacks the ability to handle the true complexity

associated with topologically close-packed (TCP) phase formation, in addition to

entirely omitting the development of other deleterious phases such as t and

Laves. As a result, PHACOMP techniques do not apply to Re containing alloys,

which can contain o, p, and P type TCP phases. Furthermore, PHACOMP is not









capable of providing any details on stability temperature ranges or phase

boundaries.

New PHACOMP

Several improvements to the PHACOMP technique were suggested to

increase the accuracy of phase instability calculations. A new technique, known

as the new phase computation method (New PHACOMP), was developed by

Morinaga et al., to predict the formation of deleterious phases, such as the o

phase and p phases, in nickel-based superalloys [58]. The New PHACOMP

method takes into account atomic size factors (atomic radius) in electronic

structure calculations (electro-negativity). New PHACOMP uses the molecular

orbital method to obtain two alloying parameters used to predict deleterious

phase formation. One parameter is the d-orbital energy level of alloying

transition metal elements (M) in a base metal (X), known as the Md level. The

other alloying parameter is a measure of the strength of the covalent bonds

between M and X atoms, known as the bond order (Bo) [59].

For Ni-based superalloys, the average d-orbital energy level (Md-value) is

calculated for alloying transition elements in the y matrix. Just as the PHACOMP

method defines a threshold value, the New PHACOMP method defines a critical

Md value, above which instability occurs.

Similar problems to those in the PHACOMP method arise in the New

PHACOMP method. The oversimplified nature of the New PHACOMP

calculations also does not take into account solute interactions.









NASA Rene N6 Model

Work conducted by Frank Ritzert et al. attempted to describe the

occurrence of TCP ( and P) phases in Ni-based superalloys, paying particular

attention to the potential synergistic effects of alloying elements on deleterious

phase formation [60].

In general, refractory metal content in a Ni-base superalloy is thought to

contribute to the formation of TCP phases. On the premise that certain

elements, or combinations of alloying elements, are more potent than others in

forming TCP phases, a regression model was developed on a design-of-

experiments (DOE) methodology for the Ni-based superalloy Rend N6. The

regression model developed for Rene N6 calculated both the linear and pairwise

interactive effects of Al, Co, Cr, Mo, Re, Ta, and W on final TCP phase content

[60]. The resulting relationship to predict TCP phase volume fraction (in terms of

atomic percent) is seen below [60].

(vol% TCP)112 = 16.344782 1.019587(Al) 2.624322(Cr) 3.821997(Mo) +
1.109575(Re) 3.207295(Ta) + 6.462984(W) 2.271803(Co) + 0.052884(AI*Co)
+ 0.214059(AI*Cr) + 0.300698(AI*Mo) + 0.80011(Co*Re) + 0.257108(Cr*Mo) -
5.081598(Re*W) + 1.824441(Ta*W)

The confidence interval around the predicted value for Rend N6 in this study was

approximately 95%.

To simplify the model's application, trace elements in the alloys were

omitted from calculation. Ti was also excluded in the model, since it was

assumed to behave similar to Ta. An attempt to apply the Rend N6 model to

Rend N5 resulted in inflated TCP content predictions, indicating that the 'model'

is not applicable to all 2nd and 3rd generation Ni-base superalloys [60]. It is,









therefore, reasonable to conclude that the usefulness of this relationship is

limited to alloys that lie near Rend N6 design parameters.

Secondary Reaction Zone (SRZ) Model

More extensive studies by Walston et al. on the occurrence of deleterious

phases in Rend N6, also indicated that secondary reaction zone (SRZ) formation

was related to an alloy's chemical composition [61,62]. Re content played a key

role in predicting SRZ formation, due to its extensive segregation during the

casting process. Statistical analysis of SRZ formation, measured by quantitative

metallographic techniques, produced the following empirical expression for the

relationship between alloy chemistry (in atomic percent) and the linear % SRZ

[62].

[SRZ(%)]1/2 = 13.88 (%Re) + 4.10(%W) 7.07(%Cr) 2.94(%Mo) 0.33(%Co) +
12.13

It is interesting to note that the SRZ empirical correlation is based solely on

experimental observations, with no fundamental scientific basis or foundation.

Although the relationship was successfully used to minimize SRZ formation in

Rend N6; it is exclusively applicable to alloys within the alloy's limited

composition ranges [61,62].

CALPHAD

Computer aided thermodynamic phase diagram calculations (CALPHAD)

have been recently used, to predict phase stability in multi-component systems,

including Ni-base superalloys.

Two main CALPHAD models are the substitutional and the multiple

sublattice models. These models predict the properties of higher-order systems









from lower-component systems, assuming that higher order interactions are

small in comparison to those that arise from the binary terms. Both of these

models are broadly represented by the equation below, where AGO is the free

energy of the phase in its pure form, AGideal is the ideal mixing term

corresponding to entropy, and AGxs is the excess free energy of mixing of

components [13,24].

S= G +Giad+G G

Once the thermodynamics of the phases are defined, the phase equilibria

can be calculated by using Gibbs free energy minimizing routines for the multi-

component system, where ni is the number of moles and Gi is the Gibbs energy

of phase i.


G = ni = minimum
i=I

When the minimum Gibbs energy at a given state is achieved, chemical

potential, Un, of each component, n, is the same in all phases and are related to

the Gibbs energy by the equation below.



1=1

Thermodynamic/mathematical CALPHAD models require the use

coefficients that uniquely describe the properties of the various phases in a Ni-

based superalloy. Coefficients for multi-component systems are kept in

databases which are proprietary or based on open literature, and are accessed

by software packages such as Thermo-Calc or JMatPro [37].









Phase equilibria calculations predict temperature and chemistry variation

effects on phase amounts. The CALPHAD method was used to simulate new

turbine disk superalloys, similar to 720 Li and Waspaloy, to increase

microstructural stability and crack propagation resistance. Through the use of

CALHAD calculations, TCP phase formation was minimized as compared to

720Li [46].

The Java-based Materials Properties (JMatPro) software was developed to

facilitate material evaluation by the calculation of phase equilibria in complex

material systems [38]. JMatPro's thermodynamic calculation software uses the

Ni-DATA (ver.6) database for the calculation of phase equilibria in all types of Ni-

based superalloys.


Liq
100-







40-



Cr2N

600 M3C6 900 1200 1500
Temperature (0C)
Figure 2-5. Phase fraction diagram for SAF 2507 Duplex stainless steel

In the thermodynamic calculation module, Gibbs free energy minimization

routines are performed using CALPHAD methods. These minimization systems

routinely calculate multi-component, multi-phase equilibria as a function of

composition or temperature. The thermodynamic model also includes stability









checking for miscibility gaps or potential ordering to find phase boundaries. The

thermodynamic calculations provide a phase fraction diagram for a given alloy

chemistry under equilibrium. An example of one such phase fraction diagram

calculated by N.Saunders and X.Li for SAF 2507 Duplex stainless steel is seen in

Figure 2-5 [40].

Once equilibrium information is obtained through conventional

thermodynamic methods, JMatPro uses physically based models to correlate

equilibrium results to material properties. The more generalized software

package, JMatPro, allows the calculation of a wide range of material properties.

The available material modeling components of this program include: material

properties related to thermodynamic calculations, solidification, thermo-physical

properties, phase transformations, and mechanical properties, using incorporated

theoretical models and proprietary property databases to make quantitative

calculations [46]. This software may prove significant to future material design

but requires validation between thermodynamic calculations and final materials

properties.














CHAPTER 3
DESIGN AND EXPERIMENTAL PROCEDURE

Alloy Design

The rapid increase of inlet temperatures in industrial gas turbines (IGT's)

produces a corresponding need for advancement in heat resistant and high

strength materials [51]. The higher temperature requirements for IGT

applications can only be met through the use of single crystal (SC) superalloys.

Over the last 30 years, the development of single crystal (SC) superalloys has

found wide spread use in aircraft jet engines; however, few alloys have been

tailored specifically for IGT applications [14,45].

IGT operational environments and fuel impurities make sulfidation or hot

corrosion attack an area of major concern. Processing challenges typical of

large SC IGT components include convective instabilities during the casting

process that lead to defect formation. Additionally, elemental segregation that

occurs during solidification requires costly homogenization heat treatments.

Incomplete homogenization resulting from inadequate heat treatments could

even lead to deleterious (TCP) phase formation.

In order to develop new SC IGT alloys, a high strength aero-engine

composition could be modified to facilitate the production of larger components

that are microstructurally stable. Desired microstructural stability, phase

transformation temperatures, and material segregation characteristics could be

obtained through the understanding and subsequent control of alloying elements.









In this study, computational techniques were used to identify compositions

that could be used for IGT applications. The alloy compositions should reflect

the strength of common second generation superalloys such as CMSX-4, and

the hot corrosion resistance and castability of first generation superalloys, such

as PWA 1483. Two iterations of elemental variation effects on microstructural

stability, phase transformation temperatures, and material segregation behavior

properties were investigated using the JMatPro thermodynamic equilibrium

module. Calculated elemental variation trends were used to identify final

compositions for experimental validation.

Compositional adjustments were selected to meet three key property

targets outlined in this study:

* Minimize deleterious topologically close packed (TCP) phase formation

o Deleterious TCP phases can result from microstructural/chemical
instabilities in nickel-based superalloys during the casting, heat
treatment, or service lifetime of a component [30]. Thought to act as
fracture initiation sites, TCP phases also deplete the solid solution
strengthening elements in the y matrix leading to a marked reduction
in rupture life [8]. Minimization of TCP phase formation can be
achieved through compositional adjustments that reduce the total
amount (wt%) of TCP phases expected at equilibrium.

* Achieve a heat treatment window of at least 25 oC

o The local enrichment of elements in segregated materials can lead to
defect or deleterious TCP phase formation. The ability to solution
heat treat an alloy to reduce or eliminate chemical segregation, can be
limited by the size of the heat treatment window. A heat treatment
window of at least 25 oC is sufficiently large for adequate solutioning
of the y' precipitate at temperatures that will not risk incipient melting.
Elemental adjustments, that increase the solidus temperature and
depress the y' solvus, can result in the optimization of the heat
treatment window.









* Minimize elemental segregation during solidification

o During solidification, partitioning of elements to the either the dendrite
solid or the interdendritic liquid results in material segregation. A
reduction of elemental segregation can be achieved through
compositional adjustments to decrease elemental partitioning of
specific elements. Decreasing elemental segregation may reduce
homogenization processing costs (lower temperatures and shorter
times) and may reduce an alloy's propensity towards TCP phase
formation.

Alloy Development Model Base Chemistry

The identification of potential IGT alloy compositions started with the

definition of an initial baseline alloy. First and second generation commercial and

experimental alloys, including CMSX-4, CMSX-11B, CMSX-11C, SC-16, and

PWA 1483 (Table 3-1), served as guides in defining a simplified baseline

chemistry designated Model (Table 3-2).

Table 3-1. Nominal composition in wt% of commercial/experimental Ni-base
superalloys
Ni Cr Co Mo Re W Al Ti Ta Hf C
CMSX-4 Bal 6.5 9.0 0.6 3.0 6.0 5.6 1.0 6.5 0.10 -
PWA 1483 Bal 12.8 9.0 1.9 0.0 3.8 3.6 4.0 4.0 0.00 0.07
SC-16 Bal 16.0 0.0 3.0 0.0 0.0 3.5 3.5 3.5 0.00 0.00
CMSX-11B Bal 12.5 7.0 0.5 0.0 5.0 3.6 4.2 5.0 0.04 0.00
CMSX-11C Bal 14.9 3.0 0.4 0.0 4.5 3.4 4.2 5.0 0.04 0.00

Table 3-2. Model alloy composition in wt% and at%.
Ni Cr Co Mo Re W Ti Al Ti Ta Hf C
wt % Bal 10.12 11.47 3.02 5.25 8.8 0.1 -
at % Bal 12 12 1 12 3 0.05 -

The baseline Model alloy is a simplified second generation single crystal

alloy intended to exhibit hot corrosion resistance, high strength, and material

castability. The baseline composition contains approximately 3 wt% (1 at%) Re,

for solid solution strengthening and increased creep resistance. In conjunction

with the Re addition, 5.96 wt% (2 at%) W was included for solid solution









strengthening and its synergistic effect on Re strengthening [6]. A total of 15 at%

in y'-formers was used in baseline Model alloy including 8.8 wt% (3 at%) Ta and

5.25 wt% (12 at%) Al, for precipitation hardening. The high Al concentration (>5

wt%) was used to stabilize the y' phase, while increasing oxidation resistance

[91]. A high Cr concentration of 10.12 wt% (12 at%) Cr, was used for hot

corrosion resistance needed in IGT applications. Co in the amount of 11.47 wt%

(12 at%) Co, was included for microstructural stability, the reduction of the

stacking fault energy (YSFE), and for limited solid solution strengthening

[11,44,45]. The addition of Co was also used to reduce the y' solvus

temperature, increasing the heat treatment window of the baseline Model alloy

[11,25,45]. A minimal amount of Hf was included to enhance coated oxidation

life [10].

"Phase I" Alloy Development Modeled Elemental Variations

'Phase I' of alloy development was initiated following the selection of the

baseline Model alloy composition. Microstructural stability, phase transformation

temperatures (liquidus, y' solvus, solidus, MC solvus, a solvus, and p solvus),

and segregation behavior for alloy compositions were predicted using the

thermodynamic equilibrium module in the Java based Materials Program 3.0

(JMatPro). All computational work for the project was conducted at Siemens

Westinghouse Power Corporation (Orlando, Fl).

JMatPro's thermodynamic calculation software uses the Ni-DATA (ver.6)

database for the calculation of phase equilibria in Ni-based superalloys. The full

database contains information on Ni, Al, Co, Cr, Fe, Hf, Mo, Mn, Nb, Re, Ru, Si,

Ta, Ti, W, Zr, B, C, N. The database supplies coefficients that uniquely describe









the thermodynamic properties of the various phases found in Ni-based

superalloys, including Liquid, y, y', NiAI, Ni3Nb, y", rl, Ni4Mo, 5NiMo, a(Cr,Mo,W),

Laves, a, p, R, P, M(C,N), M23(B,C)6, M6C, M7(B,C)3, M2N, M3B, (Fe,Ni..)2B,

(Cr,Mo..)2B, M3B2, MB, Cr5B3 TiB2, Ni3Si(h), Ni3Si2, Cr3Si, and Cr3Ni5Si2. This

database, is founded on select commercial alloys including: CMSX-4, CMSX-10,

CMSX-11 B, Rene N6, and PWA 1484 [38,40]. In order to evaluate the database

using alloy compositions outside those used for program development, a final set

of alloys was included in 'Phase III' of this study.

The alloy compositions of interest were used as the input for the JMatPro

thermodynamic calculations. Gibbs free energy minimization routines were

started at 1500 oC and were performed in the program as the temperature

stepped down in increments of 10 oC to 900 oC. These minimization systems

routinely calculated multi-component, multi-phase equilibria as a function of the

temperature. The thermodynamic model also included stability checking for

miscibility gaps or potential ordering to find phase boundaries [40].

Calculated phase fraction diagrams between 900 oC and 1500 oC allowed

the identification of critical transition temperatures such as the solidus, liquidus, y'

solvus, and a solvus. Given that solution heat treatments above the y' solvus are

used to reduce solidification segregation and to control y' precipitate size and

shape, the difference between the solidus and y' solvus temperatures or heat

treatment window was determined. The liquidus and solidus difference or

melting range was also determined. Partitioning coefficient calculations were









conducted for elements previously reported as partitioning toward the dendrite

core (Cr, Co, W, and Re) and interdendritic regions (Ni, Ta, and Al) [4,45].

The characteristic baseline Model alloy properties were used as a baseline

comparison for the other alloy compositions in 'Phase I'. Once baseline Model

alloy properties were calculated, the influences of elemental variations on

microstructural stability, transformation temperatures (liquidus, y' solvus, solidus,

MC solvus, a solvus, and p solvus), and segregation behavior were explored.

Common nickel-base superalloy elements evaluated in this study were Co, C, Cr,

Mo, W, Al, Ti, Ru, Re, and Ta. Elemental effects were determined using 3 levels

of compositional variations (High, Med, and Low). In addition, the total atomic

concentration (at%) of y' forming elements (sum of Ti, Ta, and Al) was varied

between 13.25 and 16.75 at% using 4 alloys. A total of 23 alloys were selected

for evaluation in 'Phase I'. The compositions of the baseline Model alloy and the

'Phase I' alloys are listed in Table 3-3. The compositional ranges considered and

the Ni-base superalloy properties affected by the elemental additions (also see

section 1.2.2) are listed below:

* Elemental Variations:

o C Variations: C variations have been shown to affect castability,
defect formation, elemental segregation, and microstructural stability
[16,35,50].

C additions ranging from 10 to 750 ppm wt% C, in Model 3 (0.01 wt% C),
Model 2 (0.05 wt% C), and Model 1 (0.075 wt% C) were compared to the
baseline Model (0 wt% C) alloy.

o Ru Variations: Ru additions have been shown to affect
microstructural stability, liquidus, solid solution strengthening, and
elemental segregation [4,13,28,45].









Ru additions in Model 5 (1.64 wt% (1 at%) Ru),and Model 4 (2.46 wt% (1.5
at%) Ru) were compared to the baseline Model (0 wt% Ru) alloy.

o Cr Variations: Cr variations have been shown to affect hot corrosion
and oxidation resistance, elemental segregation, microstructural
stability, and the y' solvus temperature [24,35,44].

Cr content ranging from 6.75 wt% to 11.82 wt% Cr, in Model 8 (6.75 wt%
Cr), Model 7 (8.44 wt% Cr), and Model 6 (11.82 wt% Cr) were compared to
the baseline Model alloy (10.12 wt% Cr).

o Ti Variations: Ti additions have been shown to affect elemental
segregation, precipitation hardening, liquidus, y' solvus, oxidation
resistance, and hot corrosion resistance [24,35,47].

Ti additions ranging from 0.2 to 0.58 wt% Ti (0.25 to 0.75 at% Ti) in Model
11, Model 10, and Model 9 alloys were compared to the baseline Model
alloy (0 wt% Ti). An intermediate Ti level of 0.39 wt% (0.5 at %) Ti was also
considered. To maintain a constant y' volume fraction, y'-former content
was kept constant (i.e., 15 at% total of Ti, Al, and Ta in baseline Model
alloy). Al reductions in Model 9 and Model 11 and a Ta reduction in Model
10 were used to balance Ti additions.

o Al Variations: Al (and Ta) variations have been shown to affect
elemental segregation, precipitation hardening, solid solution
strengthening, oxidation resistance, hot corrosion resistance, and
castability [12,18,24,35,45].

Al contents ranging from 5.04 to 6 wt% Al (11.5 to 13.7 at% Al) in Model 14,
13, and 12 alloys were compared to the baseline Model alloy (5.25 wt% (12
at%) Al). An intermediate Al level of 5.7 wt% (13 at%) Al was also
considered. To maintain a constant y' volume fraction, y'-former content
was kept constant (i.e., 15 at% total of Ti, Al, and Ta in baseline Model
alloy). Ta substitutions or reductions were used to balance Al variations.

o Re Variations: Re variations have been shown to affect solid solution
strengthening, castability, defect formation, elemental segregation, y'
solvus, liquidus, and microstructural stability [44,51].

Re reductions of 1.5 wt% and 3.02 wt% Re (0.5 to 1 at% Re) were made in
Model 15 (7.46 wt% W) and Model 16 (7.34 wt% Ta, 7.46 wt% W, 5.45 wt%
Al) alloys, respectively, and were compared to the baseline Model alloy
(3.02 wt% (1 at%) Re). The compositions of Model 16 and Model 15 were
adjusted for a Ta/(W+Re) ratio of one to maintain a low incidence of casting
defects, while keeping a constant y'-former content.

o Co Variations: Co variations have been shown to affect
microstructural stability, y' solvus, elemental segregation, and solid
solution strengthening [11,44,45].











Table 3-3. Baseline Model alloy composition and 'Phase I' variant compositions


C Variations
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
1 wt % Bal 10.1 11.5 6.0 8.8 3.02 5.25 0.1 0.08
2 wt % Bal 10.1 11.5 6.0 8.8 3.02 5.25 0.1 0.05
3 wt % Bal 10.1 11.5 6.0 8.8 3.02 5.25 0.1 0.01
Ru Variations
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
4 wt % Bal 10.1 11.5 6.0 8.8 3.02 5.25 0.1 2.46
5 wt % Bal 10.1 11.5 6.0 8.8 3.02 5.25 0.1 1.64
Cr Variations
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
6 wt % Bal 11.8 11.5 6.0 8.8 3.02 5.25 0.1
7wt% Bal 8.4 11.5 6.0 8.8 3.02 5.25 0.1
8 wt% Bal 6.8 11.5 6.0 8.8 3.02 5.25 0.1
Ti (with Al of Ta) Variations
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
9 wt % Bal 10.1 11.5 6.0 8.8 3.02 4.93 0.1 0.58
10wt% Bal 10.1 11.5 6.0= 3.02 5.25 0.1 0.39
11 wt% Bal 10.1 11.5 6.0 8.8 3.02 5.15 0.1 0.2


Al (and Ta) Variations
Ni Cr Co W Ta
12wt% Bal 10.1 11.5 6.0
13wt% Bal 10.1 11.5 6.0
14wt% Bal 10.1 11.5 6.01
Re (with Ta) Variations


Re Al Hf C Ru Mo Ti


High C
Med C
Low C



Med Ru
Low Ru



High Cr
Med Cr
Low Cr



HighTi (Low Al)
Med Ti (Low Ta)
Low Ti (Low Al)



High Al (Low Ta)
Med Al (Low Ta)
Low Al (High Ta)


Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
15wt% Bal 10.1 11.5 7.5 8.8 1.5 5.25 0.1
16 wt % Bal 10.1 11.5 7.5 7.3 0 5.7 0.1
y' (Ta, AI,Ti) Former Variatons
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
17wt% Bal 10.1 11.5 6.0 9.6 3.02 5.58 0.1 0.58
at% Bal 12.0 12.0 2.0 1 12.8 0.1 0.75
18 wt % Bal 10.1 11.5 6.0 6.8 3.02 6 0.1 0.39
at% Bal 12.0 12.0 2.0 1 13.7 0.1 0.5
19 wt % Bal 10.1 11.5 6.0 8.8 3.02 4.82 0.1 0.2
at% Bal 12.0 12.0 2.0 1 11 0.1 0.25
20 wt % Bal 10.1 11.5 6.0 5.9 3.02 4.82 0.1 0.2
at% Bal 12.0 12.0 2.0 1 11 0.1 0.25
Co Variations
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
21 wt % Bal 10.1 12.0 6.0 8.8 3.02 5.25 0.1
22 wt % Bal 10.1 9.0 6.0 8.8 3.02 5.25 0.1
W Variations
Ni Cr Co W Ta Re Al Hf C Ru Mo Ti
23 wt% Bal 10.1 11.5 3 8.8 3.02 5.25 0.1


Med Re (High Ta)
Low Re (Low Ta, High W)



y' former = 16.75 at%
*High Ta, Al
y' former = 16.5 at%
*High Al, Low Ta
y' former = 14.25 at%
lower Al
y' former = 13.25 at%
Lower Ta, Al



High Co
Low Co


Low W (High Mo)









Co content ranging from 9 wt% (9.41 at%) to 12 wt% (12.54 at%) Co in
Model 22 and Model 21 alloys, respectively, were compared to the baseline
Model alloy (11.47 wt% (12 at%) Co).

o W Variatons: W (and Mo) variations were have been shown to affect
solid solution strengthening, incipient melting temperature,
microstructural stability, hot corrosion resistance, castability, and
elemental segregation [12,23,24].

A 1 at% W reduction with a 1 at% Mo substitution in the Model 23 (1 at%
W, 1 at% Mo) alloy was compared to the baseline Model alloy (2 at% W, 0
at% Mo).

o y' Former Variations: In order to evaluate y' volume fraction
variation effects, the total amount of the y'-former content was
evaluated. Variations in precipitation hardener content (Ti, Al, and Ta)
has been shown to affect precipitation hardening, ductility, y' anti-
phase boundary energy (YAPB), elemental segregation, and
microstructural stability [7,34].

y'-former variations from 13.25 to 16.75 at%, in Model 20 ( 2 at% Ta, 11
at% Al, 0.25 at% Ti); Model 19 ( 3 at% Ta, 11 at% Al, 0.25 at% Ti); Model
18 (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti); and Model 17(3.25 at% Ta, 12.75
at% Al, 0.75 at% Ti) were compared to the baseline Model alloy (3 at% Ta,
12 at% Al).

Material properties were calculated for all 23 compositional variants using

the same techniques utilized for the baseline Model alloy. For clarity, the effect

of composition variations on each property were plotted in conjunction with the

baseline Model property values. Transformation temperature changes of 3 oC or

smaller and equilibrium phase amounts changes of 0.5 wt% or smaller, within the

variant concentration ranges used in this study, were considered 'limited' or

'negligible' in effect. Partitioning coefficient trends, in specific, were grouped by

elements previously reported as partitioning to the dendrite core (Cr, Co, W, and

Re) and those previously reported as partitioning to the interdendritic region (Ni,

Ta, and Al). Partitioning coefficient (kcaic) trends evaluated compositional

variation effects on a given element's direction of segregation. Partitioning









coefficient (kcalc) changes of 3% or smaller, within the variant concentration

ranges used in this study, were considered 'limited' or 'negligible' in effect.

"Phase II" Alloy Development Computational Alloy Refinement

The goal of 'Phase II' was to further refine understanding on variation

effects in alloy chemistry for potential SC IGT applications.

Baseline Model A alloy

Compositional adjustments to the baseline Model alloy after 'Phase I'

included:

* A 1 at% Cr addition: The Cr addition was used to improve hot corrosion
resistance [7,45]. Even though an increase in the amount of a phase was
calculated with increasing Cr content at 900 oC, a decrease in Re and W
partitioning was predicted during solidification.


* A 500 ppm C addition: The carbon addition was used to balance the Cr
addition, and was calculated to lower the amount of TCP phase predicted at
900 oC. C additions are also expected to increase low angel boundary
(LAB) tolerance while reducing casting defects [11].


* A 1 at% Ti addition: The Ti addition was calculated to decrease Re
partitioning during solidification.


* A 1 at% Al reduction: The Al reduction was used to improve hot corrosion
resistance and calculated to decrease Re partitioning during solidification
[35]. The Al reduction also helped maintain a y' former content of 15 at%,
balancing the 1 at% Ti addition.


* A 0.5 at% Co reduction: The Co reduction was used to increase in heat
treatment window, predicted to increase with Co reductions.


A maximum 15 at% in y'-former content was maintained in the alloy

composition to prevent an increase in the amount of TCP phases predicted to

result from increasing y'-former content at 900 oC in 'Phase I'. The new baseline









composition identified was designated 'Model A.' The modified baseline

composition is shown below in Table 3-4.

Table 3-4. Baseline Model A alloy composition in wt% and at%.
Composition IAITi y'at%
Ni Al Co Cr Hf Re Ta W Ti C
wt% 54.47 4.82 11.00 11 0.10 3.02 8.80 5.96 0.78 0.05 6.2
at % 57.21 11.00 11.50 13 0.03 1.00 3.00 2.00 1.00 0.26 15

Modeled elemental variations

Material properties for alloy compositions considered in 'Phase II' were

calculated with the same techniques used in 'Phase I,' within the temperature

range of 600 oC and 1500 oC.

Following the calculation of the baseline Model A alloy properties, elemental

variation evaluations were conducted for Ni-base alloying additions previously

shown to influence hot corrosion resistance (Al/Ti ratio), and microstructural

stability (Re, Cr, and y'-former content) [35,40,45]. A total of 16 alloys were

selected for evaluation in 'Phase II' and were grouped with respect to their

characteristic elemental or elemental group variations. The compositions of the

baseline Model A alloy and the 'Phase II' alloys are listed in Table 3-5. The

compositional ranges investigated and the Ni-base superalloy properties affected

by the elemental and elemental group additions (also see section 1.2.2) are listed

below:

* Elemental/Ratio Variations:

o Re Variations: See 'Phase I' elemental variations for Re effects

Three separate variant groups were used to investigate Re effects on
material properties:

Group 1: Re levels of 0 and 1.5 wt% Re (0 and 0.5 at% Re) in Model B
and Model C were used as a comparison to the 3.02 wt% Re (1 at% Re)









content of the baseline Model A alloy. All three alloys contained a
constant AI/Ti Ratio of 6.2 and a total of 15 at% in y'-formers.

Group 2: Four Re levels of 3.02, 2.28, 1.5, and 0 wt% Re (1, 0.75, 0.5,
and 0 at% Re), were incorporated into the Model D, E, F, and G alloys,
respectively. All four alloys contained a constant AI/Ti ratio of 2.54 and a
y'-former content of 14 at%.

Group 3: Four Re levels of 3.02, 2.28, 1.5, and 0 wt% Re (1, 0.75, 0.5, 0
and at% Re) were incorporated into the Model H, I, J, and K alloys,
respectively. All four alloys contained a constant AI/Ti ratio of 1.69 and a
y'-former content of 14 at%.

o AI/Ti Ratio Variations: See 'Phase I' elemental variations for Ti and
Al effects. The AI/Ti ratio used in a Ni-based superalloy has been
shown to have an inverse relationship on oxidation and hot corrosion
resistance. A decreasing AI/Ti ratio increase hot corrosion resistance
and an increasing AI/Ti ratios increase oxidation resistance [35].

Two separate variant groups were used to investigate AI/Ti effects on
material properties:

Group 1: AI/Ti ratios of 3.10 and 1.88 (wt%/wt%) in alloys Model 0 and
P, were compared. Both alloys contain a constant Re content of 3.02
wt% Re (1 at% Re) and a y'- former content of 16 at%.

Group 2: AI/Ti ratio variations from of 4.11 tol.88 (wt%/wt%) in Model L
(4.11), Model M (3.08), and Model N (1.88) were compared. All three
alloys contained a constant Re content of 2.28 wt% Re (0.75 at%) and a
y'-former content of 15.5 -16 at%.


o y'-Former Variations: See 'Phase I' elemental variations for y'-
former effects

Two separate variant groups were used to investigate y'-former effects
on material properties:

Group 1: y'-former variations from 14 to 16 at%, in Model D (2.5 at% Ta,
9.5 at% Al, 2 at% Ti); Model H (2 at% Ta, 9 at% Al, 3 at% Ti); Model 0
(3 at% Ta, 11 at% Al, 2 at% Ti); and Model P (3 at% Ta, 10 at% Al, 3
at% Ti) were compared to the baseline Model A alloy (3 at% Ta, 11 at%
Al). All five alloys contained a constant Re content of 3.02 wt% Re (1
at% Re).







40


Table 3-5. Model A and "Phase II" design alloy's chemical compositions in wt%
and at%.
Composition AllTi y' at%
Re Variations
Group 1 Ni Al Co Cr Hf Re Ta W Ti C
Model A wt% 54.5 4.82 11 11 0.1 3.02 8.8 5.96 0.78 0.05 6.20 15
Model B wt% 55.2 4.82 11 11 0.1 2.28 8.8 5.96 0.78 0.05 6.20 15
Model C wt% 57.5 4.82 11 11 0.1 0 8.8 5.96 0.78 0.05 6.20 15
Group 2 Ni Al Co Cr Hf Re Ta W Ti C
Model D wt% 56.0 3.97 11 11 0.1 3.02 7.4 5.96 1.56 0.05 2.54 14
Model E wt% 56.7 3.97 11 11 0.1 2.28 7.4 5.96 1.56 0.05 2.54 14
Model F wt% 57.5 3.97 11 11 0.1 1.53 7.4 5.96 1.56 0.05 2.54 14
Model G wt% 59.0 3.97 11 11 0.1 0 7.4 5.96 1.56 0.05 2.54 14
Group 3 Ni Al Co Cr Hf Re Ta W Ti C
Model H wt% 56.4 4 11 11 0.1 3.02 6.0 5.96 2.39 0.05 1.69 14
Model I wt% 57.2 4 11 11 0.1 2.28 6.0 5.96 2.39 0.05 1.69 14
Model J wt% 57.9 4 11 11 0.1 1.53 6.0 5.96 2.39 0.05 1.69 14
Model K wt% 59.5 4 11 11 0.1 0 6.0 5.96 2.39 0.05 1.69 14
AIlTi Variations
Group 1 Ni Al Co Cr Hf Re Ta W Ti C
Model 0 wt% 53.7 4.8 11 11 0.1 3.02 8.8 5.96 1.6 0.05 3.10 16
Model P wt% 53.4 4.4 11 11 0.1 3.02 8.8 5.96 2.3 0.05 1.88 16
Jroup2 Ni Al Co Cr Hf Re la W Ii C
Model L wt% 54.8 4.8 11 11 0.1 2.28 8.8 5.96 1.2 0.05 4.11 15.5
Model M wt% 54.4 4.8 11 11 0.1 2.28 8.8 5.96 1.6 0.05 3.08 16
Model N wt% 54.1 4.4 11 11 0.1 2.28 8.8 5.96 2.3 0.05 1.88 16
Cr Variations
(roup 1 Ni Al Co Cr Hf Re la W Ti C
Model A wt% 54.5 4.8 11 11 0.1 3.02 8.8 5.96 0.8 0.05 6.20 15
Model Q wt% 55.7 4.5 11 12 0.1 3.02 6.0 5.96 1.6 0.05 2.82 14
Y'- Former variations
Group 1 Ni Al Co Cr Hf Re Ta W Ti C
Model D wt% 56.0 4.0 11 11 0.1 3.02 7.4 5.96 1.56 0.05 2.54 14
at % 58.2 9.5 12 13 0.03 1 2.5 2 2 0.26
Model H wt% 56.4 4.0 11 11 0.1 3.02 6.0 5.96 2.39 0.05 1.69 14
at % 58.2 9 12 13 0.03 1 2 2 3 0.26
Model A wt% 54.5 4.8 11 11 0.1 3.02 8.8 5.96 0.78 0.05 6.20 15
at % 57.2 11 12 13 0.03 1 3 2 1 0.26
Model 0 wt% 53.7 4.8 11 11 0.1 3.02 8.8 5.96 1.56 0.05 3.10 16
at % 56.2 11 12 13 0.03 1 3 2 2 0.26
Model P wt% 53.4 4.4 11 11 0.1 3.02 8.8 5.96 2.33 0.05 1.88 16
at % 56.2 10 12 13 0.03 1 3 2 3 0.26
Group 2 Ni Al Co Cr Hf Re Ta W Ti C
Model E wt% 56.7 4.0 11 11 0.1 2.28 7.4 5.96 1.56 0.05 2.54 14
at % 58.5 9.5 12 13 0.03 0.75 2.5 2 2 0.26
Model I wt% 57.1 4.0 11 11 0.1 2.31 6.0 5.96 2.39 0.05 1.69 14
at % 58.5 9 12 13 0.03 0.75 2 2 3 0.26
Model B wt% 55.2 4.8 11 11 0.1 2.28 8.8 5.96 0.78 0.05 6.20 15
at % 57.5 11 12 13 0.03 0.75 3 2 1 0.26
Model M wt% 54.4 4.8 11 11 0.1 2.28 8.8 5.96 1.57 0.05 3.08 16
at % 56.5 11 12 13 0.03 0.75 3 2 2 0.26
Model N wt% 54.1 4.4 11 11 0.1 2.28 8.8 5.96 2.34 0.05 1.88 16
at % 56.5 10 12 13 0.03 0.75 3 2 3 0.26









Group 2: y'-former variations from 14 to 16 at%, in Model E (2.5 at% Ta,
9.5 at% Al, 2 at% Ti); Model I (2 at% Ta, 9 at% Al, 3 at% Ti); Model B (3
at% Ta, 11 at% Al, 1 at% Ti); Model M (3 at% Ta, 11 at% Al, 2 at% Ti);
and Model N (3 at% Ta, 10 at% Al, 3 at% Ti) were compared. All five
alloys contain a constant Re content of 2.28 wt% Re (0.75 at% Re).

o Cr Variations: See 'Phase I' elemental variations for Cr effects

One variant group was used to investigate Cr effects on material
properties:

Group 1: Cr variations from 11 to 12 wt% Cr (13 to 14 at% Cr), in Model
Q (14 at% Cr, 10 at% Al, 2 at% Ti) and the baseline Model A alloy (13
at% Cr, 11 at% Al, 1 at% Ti) were compared.

Material properties for all 'Phase II' alloys were calculated using the same

techniques used in 'Phase I' of alloy development. The calculated material

property values were grouped and plotted as a function of composition to

evaluate elemental property trends.

"Phase III" Alloy Development Experimental Validation

Elemental variation trends on microstructural stability, phase transformation

temperatures, and segregation behavior from 'Phase II' were used to determine

compositional adjustments to the baseline Model A alloy (Table 3-6).

Modifications to the baseline Model A composition produced five alloy

compositions for 'Phase III' listed in Table 3-7.

Table 3-6. "Phase III" compositional variants with respect to the baseline Model
A alloy
Baseline Modifications: Additions or Reductions
Composition AlTi Cr Re
High Low High High Low
Model A Alloy 1 Alloy 2 Alloy Alloy 4 Alloy 5
wt% at% wt% at% wt% at% wt% at% wt% at% wt% at%
Cr 11.0 13.0 1.0 1.0 -
a 8.8 3.0 -2.8 -1.0 -2.8 -1.0 -1.4 -0.5 -1.4 -0.5
He 3.0 1.0 -3.0 -1.0 -3.0 -1.0 -3.0 -1.0 -3.02 -1.00
Al 4.8 11.0 -0.5 -1.0 -0.8 -2.0 -0.5 -1.0 -0.7 -0.5 -0.7 -0.5
1 0.78 1 0.39 0.5 1.61 2 0.78 1 0.78 1 0.78 1
y' former (at%) _15.0 -0.5 -1.0 -1.0 -1.0 -1.0
AI/Ti (wt%Iwt%) 6.2 -2.48 -4.51 -3.31 -3.66 -3.66









Table 3-7. "Phase III" alloy compositions and variation groups in wt%
AlTi\ Variations Ni Al Co Cr Hf Re Ta W Ti C Al/Ti Y' at%
Alloy 1 wt% 57.5 4.37 11 11 0.1 0 8.8 5.96 1.17 0.05 3.72 14.5
Alloy 2 wt% 59.5 4.04 11 11 0.1 0 6 5.96 2.39 0.05 1.69 14
Cr Variation Ni Al Co Cr Hf Re Ta W Ti C Al/Ti Y' at%
Alloy 3 wt% 58.8 4.37 11 12 0.1 0 6.1 5.96 1.56 0.05 2.89 14
Re Variations Ni Al Co Cr Hf Re Ta W Ti C Al/Ti Y' at%
Alloy 4 wt% 56.0 4.17 11 11 0.1 3.02 7.4 5.96 1.56 0.05 2.54 14
Alloy 5 wt% 59.0 4.17 11 11 0.1 0 7.4 5.96 1.56 0.05 2.54 14

The general modifications made to the baseline Model A alloy to produce

the five final compositions are describe more fully below:

* General Elemental/Elemental Group Modifications:

o y'-former reductions of 1 to 1.5 at%: achieved through Al and Ta
reductions, were predicted to decrease the amount of TCP phases
present at 600 oC.

o AI/Ti ratio reductions of 2.5 to 4.5: were used to improve hot
corrosion resistance [35]. Decreasing the Al/Ti ratio was achieved
through Al reductions and Ti additions. Al/Ti reductions were also
shown to reduce the marked partitioning of Re and Ta during
solidification.

o Re reductions of 3.02 wt% (1 at%): in all final alloy compositions,
with the exception of Alloy 4, were predicted to decrease the amount
of TCP phases at 600 oC. Re reductions were also predicted to
decrease elemental segregation, in part by avoiding Re's strong
partitioning tendency towards the dendrite core.

o Cr additions of 1 wt% (1 at%): in Alloy 3 was used to improve hot
corrosion resistance [7,45]. Even though an increase in the amount of
a phase was predicted for at 600 oC, a decrease in Re and W
partitioning was also predicted.

Material properties for alloy compositions considered in 'Phase III' were

calculated with the same techniques used in 'Phase I,' within the temperature

range of 600 oC and 1500 oC.

Final alloy compositions incorporated characteristic variations for alloying

elements previously shown to influence hot corrosion and material stability.









Variations in AI/Ti ratio (with a Ta variation), Cr (with Al and Ta variations), and

Re content are seen in Table 3-7 and are listed below [35,40,45].

* Elemental/Elemental Group Variations:

o AI/Ti ratio variations: ranging from 3.72 to 1.69 (wt%/wt%) in Alloy 1
(8.8 wt% Ta) and Alloy 2 (6.02 wt% Ta), respectively, were compared.

o Cr variations: ranging from 12 to 11 wt% Cr were investigated by
comparing Alloy 3 (4.37 wt% Al, 6.05 wt% Ta) and Alloy 5 (4.17 wt%
Al, 7.37 wt% Ta), respectively.

o Re variations: ranging from 3.02 wt% Re to 0 wt% Re in Alloy 4 and
Alloy 5, respectively, were compared.

Calculated material property values were grouped and plotted as a function

of composition to evaluate elemental property trends.

In order to validate some of the predicted properties, small button samples

the 'Phase III' alloy compositions were prepared. The microstructure, phase

transformation temperatures, and elemental segregation behavior of each of the

samples were characterized and compared to the predicted values.

Materials

Although the five final compositions in 'Phase Ill' were designed for single

crystal IGT application, small polycrystalline specimens were used to validate

material properties in this study. Since this investigation focuses on elemental

variation effects on thermodynamic properties, the polycrystalline nature of the

samples should have no effect on the properties of interest.

High purity elements (> 99.5%), in the forms of granules, wire, and powder

were combined and compacted in the appropriate levels to produce the five

'Phase III' compositions. The alloy button specimens were arc melted in a

Centorr Series T Bell Jar 5BJ-2698 Arc Furnace. The arc furnace consisted of a









water-cooled stainless steel vacuum bell jar with a water-cooled copper hearth.

A mechanical vacuum pump was used to evacuate the chamber prior to back-

filling with inert gas. The compacted buttons were arc melted in a 10-1 Pa inert

argon environment to form 100 g alloy button specimens. The arc was

established between the sample and a tungsten electrode and, prior to melting

the 'Phase III' sample alloys, a Ti getter button was melted to remove any 02 or

N2 impurities from the chamber. To ensure chemical homogeneity, each sample

was melted, turned over and then remelted 9 times. The arc melted buttons

produced were approximately 4 cm in diameter and 1cm in thickness. The

button samples were sectioned using an abrasive cut-off wheel into

approximately 2 cm X 1 cm X 1 cm specimens and cleaned in an ultrasonic

Methanol alcohol bath.

Solution Heat Treatment

In order to reduce the segregation in the button samples, a solution heat

treatment was given to three samples from each alloy. The solution heat

treatment was based on the phase transformation temperatures calculated for

the compositions using the JMatPro thermodynamic equilibrium module. A

maximum solution heat treatment temperature of 1250 'C was used. This

maximum heat treatment temperature was designated to be 50 to 60 oC below

the calculated solidus temperatures for all 'Phase III' alloys to reduce the risk of

incipient melting in the segregated as-solidified microstructure.

The solution heat treatment trial was conducted in an Elatec Technology

Corporation Lab Vac 2 vacuum furnace operating at a maximum pressure of 1 x

10 -6 Torr. The vacuum furnace has a graphite hot zone measuring 15.2 cm x









15.2 cm x 38 cm with graphite heating elements and a graphite hearth plate.

Samples of each button were placed in high purity A12O3 rectangular trays to

prevent interaction of the Ni-base alloys and the graphite hearth plate during heat

treatment. Three type C OMEGA thermocouples, W-5%Re vs. W-26% Re, were

used to monitor the sample and furnace temperatures. Sample thermocouples

were maintained within + 3 oC and the over temperature thermocouple stayed

between + 0 to +15 oC, throughout the experiment. The solution heat treatment,

which was based on heat treatments for similar alloys (Table 3-8), lasted 41

hours and included a 1250 oC hold for 32 hours to homogenize the segregated

as-cast structure.

Table 3-8. Heat treatment used for the IGT experimental alloys
Step Time (hr) Rate (C/hr) Temp. (0C)
1 0.17 10 23 -150
2 1.50 10 150 1050
3 1.00 1050
4 0.08 10 1050 1100
5 1.00 1100
6 0.17 10 1100 1200
7 2.00 1200
8 0.17 3 1200 -1225
9 2.00 1225
10 0.42 1 1225 1250
11 32.00 1250
12 0.50 Gas Furnace Cool

At the completion of the heat treatment, the vacuum chamber was filled with

helium gas at 103 KPa for an increased cooling rate. The circulation of the He

gas throughout the chamber by a fan, along with a Cu H20-cooled radiator,

provided an initial cooling rate of 149 oC/min. Once the temperature dropped

below 100 oC, the samples were removed from the furnace. Following the heat

treatment, samples in the as-cast and heat treated condition were used for

characterization.









Differential Thermal Analysis

DTA testing for as-cast and heat treated samples of all 'Phase III' alloy

specimens was conducted at Dirats Laboratories on a 2910 DSC V4.4E unit. A

high purity He environment was used in all testing. Prior to testing the 'Phase III'

alloy samples, the instrument was calibrated using a high purity 200 mg Ni

standard in an A1203 lined platinum cup scanned at a rate of 20 oC/min. DTA

samples used for all five 'Phase III' alloys were approximately 2 cm X 1 cm X 1

cm in size, with masses ranging from 17 to 25 g. In order to insure that as-cast

specimens contained representative regions of all stages of solidification, the

scale of the solidification was compared to the sample size. Primary dendrite

arm spacings (PDAS) of 29 pm, 26 pm, 20 pm, 16 pm, and 26 pm were

measured for Alloys 1,2,3,4, and 5, respectively. Therefore, the sample sizes

were sufficiently large to ensure that both the dendrite core and the interdendritic

regions were tested. The temperature range analyzed in the DTA test was from

about 10000C to 15500C.

To avoid undercooling effects, reaction temperatures were taken solely

from heating curves. Heating curves were also used to avoid any chemical

changes due to specimen interactions with the A1203 in crucibles and the effects

associated with oxidation that could occur during re-solidification.

The recorded DTA results mentioned are plotted to illustrate the

temperature difference (AT) between the experimental specimen temperature

and the reference sample temperature, as well as the rate of change in the

temperature difference derivatee). The differential and derivative curves











produced can then be used to determine exothermic or endothermic phase


changes in a given alloy.

1216.01*C

3- 1343.15"C 1401.921C
3- -0.4
1200.12*C


2- -D.2 |
1207.04*C 1384. CI


10.0
EI5


0- -0.2

1308.263 C 395 94C
01 -0.4
900 1000 1100 1200 1300 1400 1500 1600
EX Up Tempwtura (0C) UVnrmil V3.2B TA Imn

Figure 3-1. DTA temperature difference (AT) vs. specimen temperature curves
for experimental Alloy 2 from 'Phase III' compositions in the heat
treated condition

DTA results were used in this study to identify the liquidus, solidus, and y'


solvus of the as-cast and heat treated samples, if present in the material. Phase


transformation or reaction temperatures were identified as distinct inflection


points in the temperature difference (AT) vs. specimen temperature curves. The


intersection points in the AT vs T plots in this study, as well as the inflection point


in the derivative vs T plots (or "maximum thermal effect") were considered the


temperature at which the reaction occurred. An example of such a plot is shown


in Figure 3-1, where major inflection points are labeled. For example, the solidus


is identified as the inflection at the right of the main endotherm in the temperature


difference curve, which is 1343 'C for Alloy 2 shown above. The liquidus and y'









solvus temperatures inflections in the derivative curve where adjusted with

respect to the Ni standard used for calibration.

Microscopy

The microstructures of the 'Phase III' alloys, in the as-cast and heat treated

conditions, were examined using optical metallography and scanning electron

microscopy (SEM) techniques.

Specimens were prepared by standard metallographic procedures.

Samples approximately 2 cm X 1 cm X 1 cm in size were mounted in bakelite,

exposing the button alloy cross section (cut surface) as seen below in Figure 3-2.

Button Top View Button Side View Mounting Orientation





Figure 3-2 Button alloy sectioning and mounting orientation in metallographic
analysis

Samples were rough ground wet with 180, 240, 320, and 600 grit silicon

carbide papers, and were then polished using 15 pm, 5 pm, 1 pm, and 0.3 pm

alumina particle suspensions. The samples were given a final polish using 0.04

pm colloidal silica. Etched samples were used for optical and SEM investigation

and un-etched samples were used for quantitative and qualitative compositional

analysis. Material specimens were etched for microscopic examination using the

Pratt and Whitney Etch # 17 (100 ml H2O + 100 ml HCI + 100 ml HNO3 + 3 g

MoOs), which dissolves the y' precipitates.

Optical metallographic examination of the 'Phase III' alloy microstructures

was performed on a LECO NEOPHOT 21 Metallograph at magnifications ranging









from 50X to 100X. Solution heat treated samples were analyzed to determine

the degree of homogenization achieved during the heat treatment. The

elimination of the as-solidified dendritic structure in the heat treated samples

indicated that the chemical segregation had been significantly reduced during the

solution heat treatment. On the other hand, the presence of dendritic structure in

the heat treated samples, indicated that incomplete homogenization had

occurred during the solution heat treatment, resulting in some degree of residual

segregation.

A JSM 6400 analytical scanning electron microscope (SEM) was used to

further characterize the microstructure of the 'Phase III' alloys. The SEM was

operated at an accelerating voltage of 15 KV in both the secondary electron (SE)

and backscattered (BSE) imaging modes. The secondary electron mode was

used to determine the as-cast and heat treated microstructure in etched samples.

The backscattered imaging mode was used to provide preliminary estimates on

residual segregation and discrete phase compositions in un-etched samples.

Qualitative chemical analysis was also preformed using a 6506 Oxford Detector

energy dispersive spectrometer (EDS) on samples in the unetched condition.

Segregation

Quantitative analysis of elemental segregation in the as-cast and as-

polished 'Phase III' samples was conducted with the use of the JEOL

Superprobe 733 electron probe micro-analyzer (EMPA)/wavelength dispersive

spectrometer (WDS). A beam size of 0.5-1.0 pm, a beam current of 20 nA, a

beam voltage of 15 KV, and a take-off angle of 40 0 were used for

characterization of all 'Phase III' alloys.









Specific calibration standards were used as references for Ni, Cr, Co, W,

Re, Ta, Al, and Ti. Wavelength dispersive spectroscopy (WDS) was used with

TAP crystals to detect W, Re, and Ta using Ma lines. The TAP crystal using Ka

lines was needed for Al. A LiF crystal was used to detect Ni, Cr, and Co

examining Ka lines. A PET crystal using La lines was needed for Ti. Each

element was counted for 10 sec per point.

Compositions were measured using 17 to 30 point line scans across a

dendritic area, with a spacing of about 1 pm between measurements. Line scans

began and ended in the interdendritic regions of a sample and intersected the

center of a primary dendrite arm. The resulting elemental readings, normalized

to 100 wt%, were plotted versus 1 pm point measures across the line scan. The

variations in composition, from minimum to maximum concentrations, provided

an estimate of elemental segregation (Figure 3-3). Some elements were

observed to segregate to the dendrite core and some elements segregated to the

interdendritic regions.

The degree of segregation was determined by calculating the partitioning

coefficient (k'). The measured compositions of a given element (x) in wt% at the

dendrite core (Cx,core) and at the interdendritic region (Cx,inter) are used to

calculate the partitioning coefficient (kx') as seen below.

kx' = Cx,core/ Cx,inter

A partitioning coefficient (kx') value of one indicates that a given element

exhibits no preference for "segregation" to the dendrite core or to the

interdendritic region during solidification. A solute with a kx' less than unity

partitions to the interdendritic region. In contrast, solutes with a partitioning










coefficient greater than unity segregate to the dendrite core. As an example, a k'

value of 1.25 demonstrates that a given element's concentration in the dendritic

core is 125% of its concentration in the interdendritic region.


Alloy 2 As-Cast Microprobe Segragation

Dendrite Core Interdentritic


1 0 ----------------------_ __--------------------_ _-


Cr
06 -K Ti



2


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Testing Point (urn)

Figure 3-3. EMPA/WDS compositions in normalized wt% versus line scan
measurement points (pm) for experimental Alloy2

The partitioning coefficients used to evaluate modeled and experimental

data in this study, represented segregation from a phase diagram. A graphic

representation of this method for a hypothetical A-B binary system is seen in

Figure 3-4. In Figure 3-4 the composition (Cs) is the composition of the fist solid

to form; the liquid composition at this temperature is represented as (CI).

In order to calculate the experimental partitioning coefficient (kx,exp), the

laboratory tested composition of a given element (x) in wt% at the dendrite core

(Cx,core) was used for the first solid formed and the nominal composition of the

element in the alloy (Cx,reg) is used as the liquid composition. This relationship

can be expressed in the formula below.

kx,exp = Gx,core/ Gx,reg














T cZ





A % Solute N B

Figure 3-4 Schematic representation of solidification occurring in a eutectic
binary phase diagram.

These experimentally determined partitioning coefficients were then

compared to the solidification compositions simulated using the thermodynamic

equilibrium module in JMatPro. JMatPro simulations provide compositional data

for the first solid to form at the liquidus transformation temperature. Although

Figure 3-4 depicts the solid composition for a binary system, JMatPro calculates

this composition for a multicomponent alloy.

The partitioning coefficients (kx,caic) for modeled alloys were calculated, as

follows, by using the predicted elemental compositions of the first solid to form

(Cx,soliid), and the nominal composition of the elements in the alloy (Cx,reg) as the

liquid composition.

kx,calc = Cx,solid/ Cx,reg

Using the partitioning coefficients defined as kx,exp and kx,caic, segregation

trends were plotted as a function of alloy composition. Partitioning coefficient

changes of 3% or smaller, within the variant concentration ranges used in this

study, were considered 'limited' or 'negligible' in effect. For clarity, elements

previously reported to partition to the dendrite core (Cr, Co, W, and Re) or the






53


interdendritic region (Ni, Ta, and Al) were grouped together to facilitate the

evaluation of elemental variation effects [4,45].














CHAPTER 4
RESULTS

Phase I Modeled Elemental Variations

The calculated material property results for the baseline Model alloy and the

'Phase I' elemental variations are given below. For clarity, the baseline Model

alloy's material properties; specifically, the microstructural stability,

transformation temperatures, and segregation behavior properties are discussed

first. These calculated material properties were then used as the baseline for

compositional comparisons. The calculated material properties for the other 23

experimental alloys were presented in groups with respect to their characteristic

elemental variation (C, Ru, Cr, Ti, Al, Re, Co, W, and the total amount of y'

former content in the alloy chemistry).

Baseline Model Alloy

The baseline Model composition is seen below in Table 4-1.

Table 4-1. Baseline Model alloy composition in wt% and at%.
Model Alloy Ni Cr Co Re W Al Ta Hf
wt% Bal 10.12 11.47 3.02 5.96 5.25 8.8 0.1
at % Bal 12 12 1 2 12 3 0.05

Microstructural stability

The phase fraction diagram calculated for the baseline Model alloy under

equilibrium conditions is shown in Figure 4-1. The calculated phase fraction

diagram gives predicted equilibrium phases and their weight fractions within the

temperature range of 900 'C and 1500 oC. These thermodynamic calculations










revealed that the equilibrium phases at temperatures under 1000 oC include y, y',

and o.


Ni-12.OA1-12.OCo-12.OCr-0.05Hf-1.ORe-3.OTa-2.OW at(%)
100
90
80
70

S *LIQUID
NGAMMA
50' GAMMA PRIME
R -- """ ****.SIGMA






900 1000 1100 1200 1300 1400 1500
Temperature(C)

Figure 4-1. Predicted phase fraction diagram for baseline Model alloy calculated
by the JMatPro thermodynamic equilibrium module

According to the calculated phase diagram, the alloy at 900 oC is made up

primarily of the y' precipitate (approximately 58 wt%) and the y matrix

(approximately 35 wt%). A limited amount of the a TCP phase (approximately 6

wt%) is also predicted to be in equilibrium at 900 oC.

Phase transformation temperatures

According to JMatPro equilibrium thermodynamic calculations, the

solidification path suggested for the baseline Model alloy is seen below.

L -* L + y -* y y + y'-* y + y'+ a

The predicted liquidus, solidus, y' solvus, and a solvus temperatures for the

baseline Model alloy were 1364 oC, 1297 oC, 1282 oC, and 1196 oC, respectively

(Figure 4-2).











Ni-12.OA1-12.OCo-12.OCr-0.05Hf-1.ORe-3.OTa-2.OW at(%)
100
90
80
70

S'601 *LIQUID
SoGAMMA
U GAMMAPRIME
0 SIGMA






900 1000 1100 1200 1300 1400 1500
Temperature(C)

Figure 4-2. Predicted phase diagram for the baseline Model alloy calculated by
the JMatPro thermodynamic equilibrium module with identification of
critical phase transformation temperatures

For the baseline Model alloy composition, the melting range was

determined to be 670C. The calculated heat treatment window for the baseline

Model alloy is 15 oC.

Elemental segregation

The thermodynamic equilibrium module in JMatPro was used to calculate

elemental segregation for specific elements in the baseline Model alloy (Table 4-

2).

Table 4-2. Predicted partitioning coefficient values (kx,calc) for the 'Phase I'
baseline Model alloy
I Ta Al Cr Ni W Co Re
K calculated 1 0.40 0.91 1.01 1.04 1.12 1.13 1.45

The segregation behavior calculations of the baseline Model alloy resulted

in partitioning coefficients, k cal, greater than one (core tendencies) for Cr, Ni, W,

Co, and Re. The core segregation predicted for Re was the most significant,









followed by Co and W. Ni and Cr were predicted to show only slight segregation

tendencies towards the core, with k ca,, values close to unity. Elements

calculated to segregate to the interdendritic region, with k ca,, values less than

one, were Ta and Al. Ta exhibited the strongest segregation with a kca,, value of

0.4, followed by Al with a calculated coefficient of 0.91.

Elemental Variation Effects

The effects of elemental variations on the calculated microstructural

stability, transformation temperature, and segregation behavior properties were

discussed in the sections below.

Chromium variation effects

Cr contents ranging from 6.75 wt% to 11.82 wt% Cr, in Model 8 (6.75 wt%

Cr), Model 7 (8.44 wt% Cr), and Model 6 (11.82 wt% Cr) were compared to the

baseline Model alloy (10.12 wt% Cr).

* Microstructural Stability

The thermodynamic calculations revealed that, for alloys with a Cr content

larger than 8.44 wt% (10 at%) Cr, the equilibrium phases at temperatures below

1000 oC included y, y', and a. The alloy with the lowest Cr content (6.75 wt% (10

at%) Cr), was predicted to contain equilibrium phases y, y', a, and p at

temperatures below 1000'C.

Figure 4-3 shows that the amount of a phase predicted at 900 oC, is

strongly influenced by increasing Cr content. The amount of a phase predicted,

increases linearly with increasing Cr content. Within the Cr range evaluated in

this study, at 900 oC, a 1.3 wt% increase in the amount of a phase was predicted

with every 1 wt% Cr increase.










Stability Effect at 900 C
10 -
9-
Sigma
7 -- Mu
55 6
5
4
3
2 ^ y =1.3x 6.7

0
Model 6 7 8 9 10 11 12
Alloy Cr Variations (wt%)
Figure 4-3 Predicted Cr variation effects on TCP equilibrium phase amounts

The Model 8 alloy with a 6.75 wt% (8 at%) Cr content, was predicted to

exhibit small amounts of both the a and p TCP phases (approximately 2 wt%

each) at 900 oC.

* Phase Transformation Temperatures

According to JMatPro thermodynamic equilibrium calculations, the

solidification path suggested for the alloys with a Cr content larger than 8.44 wt%

(10 at%) Cr is seen below.

L L + y y y + y'-* y + y'+a

The alloy with the lowest Cr content (6.75 wt% (8 at%) Cr) was predicted to

exhibit a solidification path, as seen below.

L L + y y y + y'-* y + y'+ y + '+ +

Figure 4-4 shows the predicted Cr variation effects on the liquidus, solidus,

and y' solvus temperatures.

Calculations indicated that Cr additions suppressed all critical

transformation temperatures. Increased Cr content resulted in a nearly linear

decrease in the solidus and y' solvus temperatures, both at the rate of










approximately 48 oC with a 5 wt% Cr increase. The liquidus temperature was

predicted to decrease linearly with increasing Cr content, decreasing 30 oC with a

5 wt% Cr increase.


Phase Transformation Effects Liquidus
A Solidus
1390
1390 --Y solvus

g 1350
1330
1310
1290x + 1390
y =-9.5x +1378
1270
1250
6 7 8 9 10 11 12 Model
Cr Variations (wt%) Alloy
Figure 4-4 Predicted Cr variation effects on phase transformation temperatures

With increasing Cr content, the calculated decrease in the liquidus was less

than the decrease observed for the solidus, resulting in an increase in the melting

range (16 oC with a 5 wt% Cr addition). The increase in Cr content caused the

solidus and y' solvus temperatures to linearly decrease, at similar rates, which

resulted in a nearly constant heat treatment window for all compositions

evaluated. Increasing the Cr content from 6.73 wt% Cr to 11.82 wt% Cr resulted

in a negligible 2 oC increase in heat treatment window.

* Elemental Segregation

Predicted segregation behavior for Cr variants is seen in Figure 4-5. For the

Cr variants and the baseline Model alloy, k ca,, values greater than one (core

tendencies) were calculated for Ni, W, Co, and Re. The core segregation for Re

was the most significant, followed by W, Co, and then Ni. Small Cr variation

effects were predicted for Ni and Co, where only 2% to 2.5% increases in the











kcr,caic and kco,caic resulted from a 5.61 wt% Cr addition. With a 5.61 wt% Cr

increase, Re segregation decreased by a linear 10% decrease in kRe,calc and W

segregation was predicted to decrease by a 8% linear decrease in the kw,caic.

Partitioning Effect -Ni Partitioning Effect -*-Cr
-1- Ta 16 CO
-a- Al -

0e- Re

08
13
0 7 -
12
06



04

03
6 8 10 12 6 8 10 12
wt% Cr Model Alloy wt% Cr Model Alloy :

Figure 4-5. Predicted Cr variation effects on elemental segregation

Elements predicted to segregate to the interdendritic region, with k calc

values less than one, were Ta, Al, and Cr. The partitioning of Ta was the

strongest, followed by Al and then Cr. Cr, which does not exhibit a strong

tendency to partition, was predicted to change from segregating to the

interdendritic region to the dendrite core as Cr content increased (a 6% increase

in kcr,caic with a 5.61 wt% Cr addition). Increasing Cr content was predicted to

have negligible effects on Al segregation, with observed kAI,calc values of

approximately 0.9 for all Cr variations considered. A linear increase in Ta

segregation was predicted with an increasing Cr content (a 17% decrease in the


kTa,calc with a 5.61 wt% Cr addition).

The bar chart (Figure 4-6) below was produced as an additional visual

comparison to gauge predicted elemental effects on partitioning coefficients. The










predicted Cr variation effect was most significant for Ta followed by Re, W, Cr,

Co, and then Ni.


k calc Comparisons for wt %Cr Variations
1.80
m 6.75

1.40 10.12 Model Alloy
1.20 0 11.82
0 1.00
S0.80
0.60
0.40
0.20
0.00
Ta Al Cr Ni W Co Re

Figure 4-6. kcalc comparisons between the baseline Model alloy and Cr variants

Aluminum (and Tantalum) variation effects

Al contents ranging from 5.04 to 6 wt% Al (11.5 to 13.7 at% Al) in Model 14,

13, and 12 alloys were compared to the baseline Model alloy (5.25 wt% (12 at%)

Al). An intermediate Al level of 5.7 wt% (13 at%) Al was also considered. To

maintain a constant y'-former content Ta substitutions or reductions were used to

balance Al variations.

* Microstructural Stability

Within the Al range evaluated in this study (5.04 wt% to 6 wt% Al and 10.3

wt% to 3.82 wt% Ta (1.3 at% to 3.5 at% Ta)), the thermodynamic calculations

predicted the equilibrium phases under 1000 oC to include the y, y', and a

phases.

Calculations predicted a negligible linear reduction in the amount of a

phase expected at 900 oC (Figure 4-7) with increased Al (and reduced Ta)










content. At 900 oC, the amount of a phase was predicted to decrease only 0.26

wt% for an Al increase of 0.96 wt% (2.2 at%) Al (with a 2.2 at% Ta reduction).


Stability Effect at 900 C
6.4
Sigma
6.3

6.2

6.1
y = -0.3x + 8
6 --
Moe .l 8 5 5.2 5.4 5.6 5.8 6 6.2
Alloy : Al Variations (wt%)
Figure 4-7. Predicted Al (and Ta) variation effect on TCP equilibrium phase
amount with respect to Al (wt%) concentration

* Phase Transformation Temperatures

Thermodynamic equilibrium calculations predicted the solidification path

seen below for alloys with Al (and Ta) variations ranging from 5.04 wt% to 6 wt%

Al (and 10.3 wt% to 3.82 wt% Ta).

L L + y y y + y'-* y +y'+a

Figure 4-8 depicts the predicted Al (and Ta) variation effects on the liquidus,

solidus, and y' solvus temperatures.

Increasing Al (decreasing Ta) content was predicted to result in linear

increases in both the liquidus and solidus temperatures. The liquidus and solidus

temperatures were predicted to increase 19 oC and 40 oC, respectively, with a 1

wt% Al addition (and a 6.75 wt% Ta reduction). The addition of Al (and reduction

of Ta) was predicted to decrease the y' solvus temperature. A linear decrease in










the y' solvus was calculated at the rate of 450C for a 1 wt% Al addition (with a

6.75 wt% Ta reduction).


Phase Transformation Effects A Solidus
Liquidus
1400 m Y' solvus
1380
1360
1340 Y= 19x+ 1266
1320
1300 y= 40x+ 1085


1240
y= -46x+1520
1220
5 5.2 5.4 5.6 5.8 6 Model
Al Variations (wt%) Alloy

Figure 4-8. Predicted Al (and Ta) variation effects on phase transformation
temperatures with respect to Al (wt%) concentration.

With increasing Al (and decreasing Ta) content, the predicted rate at which

the liquidus decreased was greater than the rate at which the solidus decreased,

resulting in a decrease of the melting range. Since Al additions (and Ta

reductions) were predicted to linearly increase the solidus and linearly decrease

the y' solvus at similar rates, the heat treatment window was predicted to

increase 83 oC with a 0.96 wt% Al addition (and 6.48 wt% Ta reduction).

* Elemental Segregation

Figure 4-9 shows the predicted Al (and Ta) variation effects on elemental

segregation. Elements calculated as partitioning to the dendrite core were Cr, Ni,

W, Co, and Re. Re exhibited the greatest segregation, followed by Co, W, Ni,

and then Cr. Re and Co segregation was predicted to increase linearly with Al

additions (and Ta reductions). A linear decrease in W and Ni segregation was

predicted with increased Al (and reduced Ta) content. Al additions (and Ta










reductions) were predicted to linearly shift Ni segregation, from the interdendritic

region to the dendritic core. The increase in Al (and decrease of Ta) content had

a negligible effect on calculated Cr segregation within the 0.96 wt% Al (and 6.48

wt% Ta) concentration range considered.


Partitioning Effect --Ni Partitioning Effect -*-Cr
TaCo
1.1 --Ta 1.6- W
1Al Re

0 1.5
0.7
1.4
0.8-
.2 o 1.3
S 0.7 -
0.6-_ 1.2

0.5 1.1-

0.4 1

0.3 0.9
5 5.5 6 5 5.5 6
wt% AI Model Alloy wt% AI Model Alloy i

Figure 4-9. Predicted Al (and Ta) variation effects on elemental segregation with
respect to Al (wt%) concentration.

Ta and Al were predicted to segregate to the interdendritic region, with k cal

values less than one, for all variants considered. Ta was calculated as the

strongest segregating element, followed by Al. Ta segregation was predicted to

increase linearly with Al additions (and Ta reductions). The segregation of Al

was not predicted to change significantly with an increase in Al (and decrease in

Ta) content.

The bar chart below (Figure 4-10) shows the predicted Al (and Ta) variation

effects on partitioning coefficients. Al (and Ta) variation effects were most

significant for Re, followed by W, Ni, and then Co and Ta. The negligible Al (and

Ta) variation effects predicted for Al and Cr are also evident.











k calc Comparisons for wt /oAl Variations
1.60
06
1.40
m 5.7
1.20 m 5.25 Mcobdel Alloy
S5.04
1.00

~ 0.80
0.60
0.40
0.20

0.00
Ta A Cr Ni W Co Re

Figure 4-10. kcalc comparisons between the baseline Model alloy and Al (and
Ta) variants.

Titanium (and Tantalum, Aluminum) variation effects

The effects of Ti additions (with Ta or Al reductions) on material properties

were investigated using four alloy compositions. Ti additions ranging from 0.2 to

0.58 wt% Ti (0.25 to 0.75 at% Ti) in Model 11, Model 10, and Model 9 alloys

were compared to the baseline Model alloy (0 wt% Ti). An intermediate Ti level

of 0.39 wt% (0.5 at %) Ti was also considered. Al reductions in Model 9 and

Model 11 and a Ta reduction in Model 10 were used to balance Ti additions to

keep a constant y'-former content.

* Microstructural Stability

Thermodynamic calculations for all Ti variants predicted that equilibrium

phases below 1000 OC include y, y', and a. Notably, a 0.58 wt% (0.75 at%) Ti

addition was predicted to only increase the amount of a phase by 0.2 wt% at

900'C. The negligible Ti effect on the amount of a phase, predicted at 900 oC, is

seen in Figure 4-11.











Stability Effect at 900 C
6.5
*Sigma
6.45

6.4

6.35

6.3

6.25
0 0.1 0.2 0.3 0.4 0.5 0.6
Alloy : Ti Variations (wt%)

Figure 4-11. Predicted Ti (and Ta or Al) variation effects on TCP equilibrium
phase amount with respect to Ti (wt%) concentration.

Phase Transformation Temperatures

The solidification path predicted for the Model 11, 10, and 9 alloys is seen

below.

L L + y -* y y +y' -* y + y'+


Phase Transformation Effects A Solidus
Liquidus
1380 a Y' solvus


S1340-
S1320-
S1300,-A


1 1260
4 y=-32x+1281
1240 '
0 0.1 0.2 0.3 0.4 0.5 0.6
Model : Ti Variations (wtO/
Allo -
Figure 4-12. Predicted Ti (and Ta or Al) variation effects on phase
transformation temperatures with respect to Ti (wt%) concentrations

Figure 4-12 depicts the predicted Ti (with Al or Ta reduction) variation

effects on the liquidus, solidus, and y'solvus temperatures.









Ti additions (with Al or Ta reductions) resulted in negligible effects on phase

transformation temperatures with the exception of the y' solvus. A 0.58 wt%

(0.75 at %) Ti addition with a (0.75 at%) Al reduction were predicted to result in

negligible reductions in the liquidus and solidus temperatures. A smaller Ti

addition of 0.39 wt% (0.5 at%) Ti with a (0.5 at%) Ta reduction was predicted to

increase the solidus 6C. The reversed phase transformation trends, and the 6

'C solidus increase, were attributed to the Al and Ta reductions. Regardless of

whether Ti additions had been balanced by Al or Ta reductions, calculations

predicted a clear linear decrease in the y' solvus temperature with Ti additions.

For all three Ti additions, the y' solvus decreased with increasing Ti content,

decreasing approximately 6 oC within the 0.58 wt% (0.75 at%) Ti range analyzed.

The calculated decrease in the y' solvus temperature, with increasing Ti

content, resulted in a considerable increase in the heat treatment widow. With a

0.58 wt% (0.75 at %) Ti addition and (0.75 at%) Al reduction, the calculated heat

treatment window increased from 15 oC to 28 oC. A 0.39 wt% (0.5 at%) Ti

addition with a (0.5 at%) Ta reduction resulted in an increased heat treatment

window of 24 oC.

* Elemental Segregation

Segregation behavior trends calculated for alloys with Ti variations (and Al

reductions in Model 11 and 9 or Ta reductions in Model 10) are seen in Figure 4-

13. For all alloys considered (the baseline Model, Model 11, Model 10, and

Model 9 alloys), the elements Cr, Ni, W, Co, and Re were predicted to partition to

the dendrite core. Re was the most heavily segregated element, followed by W,







68


Co, Ni, and then Cr. Regardless of whether Ti additions were balanced by either

Ta or Al reductions, negligible elemental variation effects were predicted for Cr,

Co, Ta, Ni, W, and Al. When compared to the baseline Model alloy, Re

segregation decreased for all alloys with Ti additions. A small 1% decrease in

the kRe,calc was calculated with a 0.39 wt% (0.5 at%) Ti addition (and a 0.5 at%

Ta reduction) but an approximately 30% decrease was predicted with a 0.58 wt%

(0.75 at %) Ti addition (and a 0.75 at% Al reduction).

Elements predicted as segregating to the interdendritic region were Ta, Ti,

and Al. Ta was calculated to segregate the strongest, followed by Ti and Al. A

0.39 wt% Ti addition was predicted decrease kTi,calc by 4%.


Partitioning Effect


* Ni


* Ta
SAl
*1i


1.5

1.4

1.3

1.2

1.1

1

0.9


Partitioning Effect


A
*



m


iCr
iCo
iW
Re


4
*


0 0.2 0.4 0.6 0 0.2 0.4 0.6
Wt% Ti Model Alloy i wt% Ti Model Alloy i

Figure 4-13. Predicted Ti (and Ta or Al) variation effects on elemental
segregation with respect to Ti (wt%) concentration.

The bar chart (Figure 4-14) below shows the predicted partitioning

coefficients for the Ti variants considered with respect to their Ti content. The

predicted decrease in Re partitioning with increased Ti content is seen with Al


.
"










reductions in Model 9 (0.2 wt% Ti) and Model 11 (0.58 wt%), or with Ta

reductions in Model 10 (0.39 wt% Ti).


k calc Comparisons for wt %Ti Variations
1.60
0 0.58
1.40 0.39
1.20 0.2
1.00
0 0.80 -
0.60
0.40
0.20 --
0.00
Ti Ta Al Cr Ni W Co Re

Figure 4-14. kcalc comparisons between the baseline Model alloy and Ti
variants

Rhenium (and Tantalum, Tungsten) variation effects

Re (with Ta, Al, and W) variation effects on material properties were

evaluated using three alloy compositions. Re reductions to the baseline Model

composition of 1.5 wt% and 3.02 wt% Re (0.5 to 1 at% Re) were made in Model

15 (7.46 wt% W) and Model 16 (7.34 wt% Ta, 7.46 wt% W, 5.45 wt% Al) alloys,

respectively, and were compared to the baseline Model alloy (3.02 wt% (1 at%)

Re).

* Microstructural Stability

Thermodynamic calculations predicted that for alloys with a Re content of

3.02 wt% (1 at%) Re, the equilibrium phases at temperatures below 1000 OC

include y, y', and a. The alloy with the low Re content (1.5 wt% (0.5 at%) Re),

was predicted to contain equilibrium phases y, y', and p at temperatures below

10000C.










No linear relationships on the amount of specific TCP phases (with respect

to Re content) were observed. Even though no linear relationships exist, Figure

4-15 shows a strong Re effect predicted on the amount of TCP phases predicted

at 900 oC. Despite W, Ta, and Al variations in the Model 16 alloy and W

increases in the Model 15 alloy; TCP phase amounts were predicted to decrease

for both alloys when compared to baseline Model alloy. This is observed when

comparing the Model 16 alloy (0 wt% Re) to the baseline Model alloy (3.02 wt%

Re), which are predicted to contain 0 wt% and 6 wt% in TCP phases at 900 oC,

respectively.


Stability Effect at 900 C

Sigma 1 6.3 wt% Sigma
E Mu
5 0 5wt% Mu


5




0
0 0.5 1 1.5 2 2.5 3 Model
Re Variations (wtO/o) Alloy :

Figure 4-15. Predicted Re (with Ta, Al, and W) variation effects on TCP
equilibrium phase amounts with respect to Re concentration (weight
percent).

Phase Transformation Temperatures

JMatPro thermodynamic equilibrium calculations predicted that the baseline

Model alloy with a Re content of 3.02 wt% (1 at%) Re would follow the

solidification path seen below.

L L + y -* y y + y'-* y + y' +o










The Model 15 alloy with a low Re content of 1.5 wt% (0.5 at%) Re was

predicted to solidify as seen below.

L L + y -* y y + y' y + y'+ a* y + + y + P

Thermodynamic equilibrium calculations predicted that the Model 16 alloy,

with a 0 wt% Re (0 at%) Re content, would solidify as seen below.

L -*L+y- y + y'

Figure 4-16 below depicts the liquidus, solidus, and y'solvus temperature

trends predicted with respect to Re alloy variations (with W, Al, and Ta variations

in Model 16 and W increases in Model 15).


Phase Transformation Effects io uds
E Y' solvus
1370

Q 1350 -

1330

1310

1 1290
y=-3x+1289
1270
Model 1 2 3
Alloy Re Variations (wt/o)
Figure 4-16. Predicted Re (with Ta, Al, and W) variation effects on phase
transformation temperatures with respect to Re (wt%) concentration.

A 1.5 wt% (0.5 at%) Re reduction with a 0.5 at% W addition, resulted in a

calculated 4 oC decrease in the liquidus. Removing Re from the alloy chemistry

(while increasing Al 0.5 at%, increasing W 0.5 at%, and reducing Ta 0.5 at%)

decreased the liquidus 4 oC and increased the y' solvus 7 oC. These

thermodynamic calculations predicted that Re reductions in Model 15 (with W

increases) and Model 16 (with Al, W increases and Ta reductions) produced a

linear increase in the y' solvus temperature.










The linearly decreasing y' solvus temperature with increasing Re content

results in a decrease of the heat treatment window. A 1.5 wt% (0.5 at%) Re

reduction with a 0.5 at% W increase, was predicted to decrease the heat

treatment window 4 'C. Removing Re from the alloy chemistry (while increasing

Al 0.5 at%, increasing W0.5 at%, and reducing Ta 0.5 at%) decreased the heat

treatment window 8 oC.

* Elemental Segregation

Predicted elemental variation effects on segregation for the baseline Model,

Model 15, and Model 16 alloys with respect to increasing Re concentration are

seen in Figure 4-17.
Partitioning Effect --Ni Partitioning Effect -4-Cr
1 .1Ta Co
Al -W
1 -- Re
1.4
0.9
0.8 1.3

0 o
o 0.7 g 1.2
0.6 1

0.5
0.4 1
0.3- 0.9
0 1 2 3 4 0 1 2 3 4
wt% Re Model Alloy i wt% Re Model Alloy :
Figure 4-17. Predicted Re (with Ta, Al, and W) variation effects on elemental
segregation with respect to Re (wt%) concentration.

Elements predicted as segregating to the dendrite core were Cr, Ni, W, Co,

and Re. Re exhibited the most severe segregation, followed by W, Co, Ni, and

then Cr. A 3.02 wt% (1 at%) Re reduction (with a 0.5 at% Al increase, a 0.5 at%

W increase, and a 0.5 at% Ta reduction) resulted in a negligeble 2% decrease in










kNi,calc, a limited 2.5% increase in kw,calc, and a more notable 3.5 to 4% increase

in kco,caic and kcr,caic, respectively.

Ta and Al were predicted to segregate to the interdendritic region. Ta was

the most segregated, followed by Al. Partitioning coefficients for Ta and Al went

relatively unchanged in comparisons between the baseline Model, Model 15, and

Model 16 alloys.

The bar chart (Figure 4-18) below shows predicted elemental partitioning

coefficients for the Re variants considered with respect to their Re content.

Extensive Re partitioning was avoided when Re was omitted from alloy

compositions.


k calc Comparisons for wt %Re Variations

O 3.02 Model Alloy r A
14 : 1.5 s
N0
1.20-

1.00 -

2 0.80

0.60

0.40

0.20
Ta Al Cr Ni W Co Re

Figure 4-18. kcalc comparisons between baseline Model alloy and Re variants

Carbon variation effects

C variation effects on material properties were evaluated by comparing the

Model 3 (0.01 wt% C), Model 2 (0.05 wt% C), Model 1 (0.075 wt% C), and

baseline Model (0 wt% C) alloys.










* Microstructural Stability

Thermodynamic calculations for all C variants revealed that equilibrium

phases below 1000 oC include y, y', a, and M23C6. The formation of M23C6

carbides was predicted for all C variants.

C additions were predicted to decrease the amount of a phase expected at

900 'C. The linear relationship predicted between C content and the amount of

TCP phases at 900 oC (Figure 4-19), expected a 10 wt% reduction in the amount

of a phase with a 1 wt% C increase. A 0.68 wt% reduction in the amount of a

phase at 900 oC was calculated for the maximum 0.08 wt% C addition used in

this study.


Stability Effect at 900 C
6.5
Sigma
6.25

11y -x + 6
5.75

5.5 -

5.25
Model 0.02 0.04 0.06 0.08
Alloy C Variations (wt%)
Figure 4-19. Predicted C variation effect on TCP equilibrium phase amount

* Phase Transformation Temperatures

The predicted solidification path for the C containing alloys is seen below.

L -* L + y -* L + y + MC -* y + y'+ MC -y + Y'+ M23C6 -* Y + Y' + +

M23C6

Figure 4-20, depicts the predicted C variation effects on the liquidus,

solidus, and y' solvus temperatures.










C additions resulted in a predicted linear decrease of the liquidus and y'

solvus temperatures. A 0.075 wt% C addition was predicted to linearly decrease

the liquidus and y' solvus by approximately 6 oC and 9 oC, respectively. A linear

increase of the solidus was predicted with increasing C content. Calculations

predicted a 9 oC increase in the solidus with a 0.075 wt% C addition.


Phase Transformation Effects A Solidus
1370
1360 Liquidus
1350 -Y'solvus
1350
1340
1330 -
1320
| 1310
1300
E 1290
1280
1270
Model 0.02 0.04 0.06 0.08
Alloy C Variations (wtO/o%)
Figure 4-20 Predicted C variation effects on phase transformation temperatures

The predicted linear decrease in the liquidus and linear increase in the

solidus with increasing C content results in a decrease in the solidification range

(15 oC with a 0.075 wt% C addition). The predicted inverse C effects on the

solidus and y' solvus temperatures, resulted in the increase of the heat treatment

window with increasing C content. A 0.075 wt% C addition was predicted to

increase the heat treatment window 18 oC.

* Elemental Segregation

The predicted C variation effects on segregation behavior in this study are

seen below in Figure 4-21. Elements predicted to have k calc values greater than

one or who tend to partition to the dendrite core, were Ni, W, Co, and Re. Re

segregation was the strongest, followed by Co, W, and then Ni. Within the 0.075










wt% C range considered, C variations were predicted to have a no effect on Co,

W, and Ni segregation. Negligible C variation effects were calculated for Re.


Figure 4-21. Predicted C variation effects on elemental segregation.


1.40

1.20

1.00

0.80

0.60
0.40

0.20
0.00


C Ta Al Cr Ni W Co Re


Figure 4-22. kcalc comparisons between the baseline Model alloy and C variants

Ta, Al, Cr, and C were predicted to segregate to the interdendritic region.

The partitioning of C was the strongest, followed by Ta, Al and then Cr. A small

linear decrease in Al segregation was predicted for C increases (a 3.3% increase

in kAI,cali with a 0.075 wt% C addition). Negligible C variation effects were










calculated for C, Ta and Cr. Figure 4-22 shows the predicted C variation effects

on partitioning coefficients. Predicted C variation effects were most significant for

Al.

Cobalt variation effects

Co effects on material properties were evaluated using three alloy

compositions. Co content ranging from 9 wt% (9.41 at%) to 12 wt% (12.54 at%)

Co in Model 22 and Model 21 alloys, respectively, were compared to the baseline

Model alloy (11.47 wt% (12 at%) Co).

* Microstructural Stability

Thermodynamic calculations for all Co variants predicted that equilibrium

phases below 1000 oC include y, y', and o.

Co additions had no effect on the amount of a phase predicted at 900 oC.

Within the 3 wt% (3 at%) Co range evaluated in this study, calculations at 900 oC

predicted a phase amounts nearly identical to those of the baseline Model alloy

(Figure 4-23).


Stability Effect at 900 C
6.32

6.3 *Sigma

6.28

6.26-
Y 6.24 y= 0.03x+ 6
6.24-

6.22--

6.2
Model 8.5 9.5 10.5 11.5 12.5
Alloy : Co Variations (wtYo%)
Figure 4-23. Predicted Co variation effect on TCP equilibrium phase amount










* Phase Transformation Temperatures

Co variants considered in this study were predicted to follow the

solidification path seen below.

L L + y -* y y + y' -* y + y'+

Figure 4-24 depicts the predicted Co variation effects on the liquidus,

solidus, aned y' solvus temperatures.

A Solidus
Phase Transformation Effects A oid
Liquidus
1380 -- Y' solvus
1360 :
0 y=-1x+1374
1340
1320
y=-4x+1341
1300
S 1280 -
Y1 4x+1233
1260
9 9.5 10 10.5 11 11.5 12
Model :
Alloy : Co Variations (wt%)
Figure 4-24. Predicted Co variation effects on phase transformation
temperatures.

Calculations predicted a linear decrease in the solidus and a linear increase

in the y' solvus; both at the rate of 4 oC with a 1 wt% Co addition. Co variation

effects on the liquidus temperature were considered negligible.

Even though the solidus was predicted to decrease with increasing Co

content, only a negligible increase in the melting range was observed

(approximately 2 oC with a 3 wt% Co addition). The combined effect of an

increased y' solvus and decreased solidus with increasing Co content, was

predicted to decrease the heat treatment window. Decreasing the Co content

from 11.47 to 9 wt% was predicted to increase the heat treatment window 19 oC.










* Elemental Segregation

Predicted Co effects on segregation behavior are seen in Figure 4-25.

Partitioning Effect -4-Cr Partitioning Effect Cr
1.5 --- Co 1.5 Co
--- W --W
1.4 -- Re 1.4- Re

1.3- 1.3

1.2 1.2
o o
1.1 1.1

1 1 1

0.9 0.9--
8.5 9.5 10.5 11.5 12.5 8.5 9.5 10.5 1 .5 12.5
wt%Co Model Alloy i wt%Co Model Alloy :
Figure 4-25. Predicted Co variation effects on elemental segregation

Ni, W, Co, Cr, and Re were predicted to partition to the dendrite core with k

caic values greater than one. Re was the most segregated element, followed by

W, Co, Ni, and then Cr. Co variations were predicted to have negligible effects

on W, Re, Co, Ni, and Cr segregation, within the 3 wt% Co range considered.

Elements calculated to have kca,,ic values less than one or who partition to

the interdendritic region were Ta and Al. Ta segregated to the greatest extent,

followed by Al. The largest Co variation effect was predicted for Ta, which

linearly increased segregation towards the interdendritic region, with increased

Co content (a 4% decrease in kTa,caIc with a 3 wt% Co addition). No significant

Co variation effects were calculated for Al.

Figure 4-26, below, compares predicted elemental partitioning coefficients

for the Co variants considered. Predicted Co effects were most significant for Ta.

The negligible Co variation effects predicted for Re, Al, and Ni are also observed.












1.60
m9
1.40 11.47 Model Alloy

1.20 _E12

1.00

j 0.80

0.60

0.40

0.20

0.00
Ta Al Cr Ni W Co Re

Figure 4-26. kcalc comparisons between Model alloy and Co variants

Ruthenium variation effects

Ru variation effects on material properties were evaluated using three alloy

compositions. Ru additions in Model 5 (1.64 wt% (1 at%) Ru and Model 4 (2.46

wt% (1.5 at%) Ru) were compared to the baseline Model (0 wt% Ru) alloy.

* Microstructural Stability

Calculations for the Ru variants considered in this study predicted that the


y, y', and o equilibrium phases would be present below 1000 oC.


Stability Effect at 900 C
6.8
Sigma
6.7

6.6

6.5
y = 0.2x + 6.3
6.4

6.3

6.2

Model .0 0.5 1 1.5 2 2.5
Alloy : Ru Variations (wto)
Figure 4-27. Predicted Ru variation effect on TCP equilibrium phase amount










Ru additions were calculated to have a minimal effect on TCP phase

amounts. At 900 oC, the amount of a phase was predicted to only increase (0.42

wt% a) with a 2.46 wt% (1.5 at %) Ru addition (Figure 4-27).

* Phase Transformation Temperatures

Thermodynamic equilibrium calculations for all the Ru variants considered

in this study, predicted the solidification path seen below.

L L + y y y + y'-* y + y'+a

Predicted Ru variation effects on the liquidus, solidus, and y' solvus

temperatures are seen in Figure 4-28.


Phase Transformation Effects A Solidus
1370 Liquidus
1360 5-- Y' solve s
1360 Y
1350 -y=-0.6x+1364
S1340
S1330
S1320
S1310
S1300 Y y=-3x+1997
1 1290 A
1280 y=-3x+12
1270
Model 0 0.5 1 1.5 2 2.5
Alloy Ru Variations (wt/o%)
Figure 4-28. Predicted Ru variation effects on phase transformation
temperatures

Ru additions showed a negligible effect on the liquidus temperature.

Calculations also predicted linear decreases of the solidus and the y' solvus, at

rates of approximately 7 oC and 8 oC, respectively, for a 2.5 wt% Ru addition.

Decreases in the solidus and nearly constant liquidus predictions, resulted

in a small increase of the solidification range. Increasing the Ru concentration by

2.46 wt% Ru was predicted to increase the solidification range 5 oC. Since Ru










additions linearly decrease the solidus and the y' solvus at similar rates, the heat

treating window was predicted to remain relatively constant.

* Elemental Segregation

Predicted Ru variation effects on segregation behavior are seen below in

Figure 4-29.


Partitioning Effect -*-Ni Partitioning Effect -* Cr
S15--CO
1.2 --- Ta 1
1.1 -A-A -*-Re
1 -- Ru

0.9 1 13
0.8
.2 8 12 |
m 0.7
0.6 1
0.5
0.4 1
0.3
0.2- 12
0 1 2 3
Wt%Ru Model Alloy wt% Ru Model Alloy i

Figure 4-29 Predicted Ru variation effects on elemental segregation

Elements calculated to have k ca,, values greater than one, which tend to

partition to the dendrite core, were Re, Ni, W, Co, and Cr. Re was predicted to

exhibit the most severe segregation, followed by Co, W, Ni, and then Cr. Ru

variations were predicted to have a negligible effect on Re, Ni, Cr, and Co

segregation, within the 2.46 wt% Ru range considered. A linear decrease in W

segregation was predicted with increasing Ru content (a 4.5% decrease in kw,calc

was predicted with a 2.46 wt% Ru addition).

Ru, Ta, and Al were predicted to partition to the interdendritic region. The

most significant segregation was predicted for Ta, followed by Al, and then Ru.

No significant Ru variation effects were calculated for Ru,Ta, or Al.




Full Text

PAGE 1

COMPUTATIONAL DESIGN OF NI CKEL BASED SUPERALLOYS FOR INDUSTRIAL GAS TURBINE COMPONENTS By ALMA STEPHANIE TAPIA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Alma Stephanie Tapia

PAGE 3

iii ACKNOWLEDGMENTS The author would like to thank and to acknowledge Dr. Gerhard Fuchs, Dr. Reza Abbaschian, Dr. Robert DeHoff, and Dr. Hans Jurgen Seifert for their support and guidance in this project. Special thanks go out to Allister James and David Hunt of Siemens Westinghouse Po wer Generation (SWPC) in Orlando, FL, who dedicated resources and time to ma ke this project possible, and to the high temperature alloys group. Additi onal thanks go to Wayne Acree and the staff of the Major Analytical Instrument Center (MAIC) at t he University of Florida. This material is based on work suppor ted by the Department of Energy.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDG MENTS.......................................................................................iii LIST OF TABLES.................................................................................................ix LIST OF FI GURES...............................................................................................xi ABSTRACT.......................................................................................................xvii CHAPTER 1 INTRODUC TION............................................................................................1 2 LITERATURE SEARCH.................................................................................5 Microstructure................................................................................................5 The Matrix.............................................................................................5 The Phase............................................................................................5 The / Mismat ch....................................................................................7 Carbi des..................................................................................................7 Phase Instabilities..8 Topologically close packed (TCP ) phases........................................8 Secondary reaction zones (S RZs)....................................................9 Deleterious phase c onsiderat ions ...................................................10 Chemical Com positio n.................................................................................11 Strengthening Methods ................................................................................11 Solid Solution Strengtheni ng.................................................................11 Precipitation Hardeni ng.........................................................................12 Alloying El ements..................................................................................12 Cobalt..............................................................................................13 Carbon............................................................................................13 Rutheni um.......................................................................................13 Rhenium..........................................................................................13 Chromi um........................................................................................14 Aluminum and Titanium...................................................................14 Tantalum and Tungsten..................................................................16 Molybdenu m....................................................................................16 Hafniu m...........................................................................................17 Casting and Pr ocessi ng...............................................................................17

PAGE 5

v Casting Concerns and Defect Form ation...............................................17 Solutionizing/Homogeniza tion Heat Tr eatments....................................18 Predictive Methods.......................................................................................19 PHACOMP............................................................................................20 New PHAC OMP....................................................................................22 NASA Rene N6 M odel...........................................................................23 Secondary Reaction Zone (SRZ) Model................................................24 CALPHAD..............................................................................................24 3 DESIGN AND EXPERIM ENTAL PROC EDURE..........................................28 Alloy De sign.................................................................................................28 Alloy Development Model Base Chem istry.......................................30 Phase I Alloy Development Modeled Elemental Variations............31 Phase II Alloy Development Co mputational Alloy Refinement........37 Baseline Model A allo y....................................................................37 Modeled elemental variati ons..........................................................38 Phase III Alloy Development Experimental Validation....................41 Material s......................................................................................................43 Solution Heat Treatm ent..............................................................................44 Differential The rmal Anal ysis........................................................................46 Microsc opy...................................................................................................48 Segregatio n..................................................................................................49 4 RESULT S....................................................................................................54 Phase I Modeled Elem ental Vari ations .....................................................54 Baseline Model Alloy.............................................................................54 Microstructura l stabilit y....................................................................54 Phase transformation temperat ures................................................55 Elemental s egregatio n.....................................................................56 Elemental Variat ion Effe cts....................................................................57 Chromium variat ion effe cts..............................................................57 Aluminum (and Tantalum) variation e ffects.....................................61 Titanium (and Tantalum, Alum inum) variatio n effects.....................65 Rhenium (and Tantalum, Tungst en) variation effects......................69 Carbon variati on effects..................................................................73 Cobalt variati on effect s....................................................................77 Ruthenium variat ion effe cts.............................................................80 Tungsten (and Molybdenum) variation e ffects................................83 Gamma prime former (Tantal um, Aluminum, and Titanium) variation e ffects............................................................................87 Temperature Range Comparisons for P hase I of Alloy Development.91 Elemental Variation Trend Summary for Phase I................................93 Phase II Computational Alloy Re finement .................................................95 Microstructura l Stabili ty..........................................................................95 Phase Transformati on Temperat ures....................................................95

PAGE 6

vi Elemental S egregatio n..........................................................................96 Elemental Variat ion Effe cts....................................................................96 Rhenium variati on effects................................................................96 Chromium (Aluminum and Titani um) variation effects...................101 Gamma prime former (Tantal um, Aluminum, and Titanium) variation e ffects..........................................................................105 Al/Ti Ratio variat ion effect s............................................................110 Elemental Variation Trend Su mmary for Phase II.............................114 Phase III Experimental Validatio n...........................................................114 Microstructural Stability ........................................................................116 Experimetal results: microstr uctural characte rization....................116 Computational results: JMatPro equilibrium phase predictions.....129 Experimental to computat ional compar isons.................................131 Elemental variati on effect s............................................................131 Phase Transformation Temperatur es..................................................133 Experimental results: DTA resu lts.................................................133 Computational results: JMatPro phase transformation temperatur es..............................................................................140 Experimental to computat ional compar isons.................................141 Temperature range comparisons for Phase III of alloy developm ent..............................................................................144 Elemental variati on effect s............................................................145 Elemental Segr egation........................................................................149 Experimental results: EMPA/ WDS microprobe analysis................149 Computational results: JMatPro solidification pr edictions..............149 Experimental to computat ional compar isons.................................150 Elemental variati on effect s............................................................152 5 DISCUSSI ON.............................................................................................157 Microstructural Stability ..............................................................................159 Elemental Variati on Effects..................................................................159 Carbon variation effects................................................................159 Cobalt variati on effect s..................................................................160 Ruthenium variati on effect s...........................................................160 Tungsten (and Molybdenum) va riation effe cts..............................161 Gamma prime former (Tantal um, Aluminum, and Titanium) variation e ffects..........................................................................161 Aluminum (and Tantalum) variation e ffects...................................163 Titanium (and Tantalum, Alumi num) variation effects...................164 Al/Ti ratio variat ion effect s.............................................................165 Chromium variati on effect s............................................................166 Rhenium variati on effect s..............................................................168 Experimental to Computational Materi al Microstructure Comparisons170 Phase Format ion.................................................................................171 Phase Transformation Temperatur es.........................................................173 Elemental Variati on Effects..................................................................174

PAGE 7

vii Carbon variation effects................................................................174 Cobalt variati on effect s..................................................................175 Ruthenium variati on effect s...........................................................175 Tungsten (and Molybdenum) va riation effe cts..............................176 Gamma prime former (Tantal um, Aluminum, and Titanium) variation e ffects..........................................................................177 Aluminum (and Tantalum) variation e ffects...................................179 Titanium (and Tantalum, Alumi num) variation effects...................180 Al/Ti Ratio variat ion effect s............................................................181 Chromium variati on effect s............................................................183 Rhenium variati on effect s..............................................................185 Experimental to Computational Te mperature Range Comparisons.....186 Segregatio n................................................................................................188 Elemental Variati on Effects..................................................................189 Carbon variation effects................................................................189 Cobalt variati on effect s..................................................................190 Ruthenium variati on effect s...........................................................190 Tungsten (and Molybdenum) va riation effe cts..............................191 Gamma prime former (Tantal um, Aluminum, and Titanium) variation e ffects..........................................................................192 Aluminum (and Tantalum) variation e ffects...................................194 Titanium (and Tantalum, Alumi num) variation effects...................195 Al/Ti Ratio variat ion effect s............................................................195 Chromium (Aluminum and Titani um) variation effects...................197 Rhenium (and Tantalum, Tungst en) variation effects....................199 Experimental to Computational Pa rtitioning Com parisons...................202 Compositional Refinement .........................................................................204 Compositional M odificati ons................................................................205 Alloy Compar isons............................................................................... 206 Future Deve lopment ............................................................................210 6 CONCLUSION S.........................................................................................214 Microstructural Stability ..............................................................................214 Phase Transformation Temperatur es.........................................................214 Segregatio n................................................................................................214 Elemental Variati on Effect s........................................................................214 Future Deve lopment ...................................................................................215 7 FUTURE WORK........................................................................................216 Continued Computatio nal Modelin g...........................................................216 Microstructural Stabili ty Evaluat ions...........................................................216 Further Development of Alloy 1..................................................................217 LIST OF REFE RENCES..................................................................................218

PAGE 8

viii BIOGRAPHICAL SKETCH ...............................................................................224

PAGE 9

ix LIST OF TABLES Table page 3-1 Nominal composition in wt% of commercial/experimental Ni-base superallo ys...............................................................................................30 3-2 Model alloy composit ion in wt% and at%. ................................................30 3-3 Baseline Model alloy composition and Phase I variant compositions.....35 3-4 Baseline Model A alloy com position in wt % and at%...............................38 3-5 Model A and Phase II design alloys chemical compositions in wt% and at%....................................................................................................40 3-6 Phase III compositional variants with respect to the baseline Model A alloy .........................................................................................................41 3-7 Phase III alloy compositions and variation gro ups in wt%.....................42 3-8 Heat treatment used for t he IGT experiment al allo ys...............................45 4-1 Baseline Model alloy com position in wt % and at%..................................54 4-2 Predicted partitioning coefficient values (kx,calc) for the Phase I baseline Model alloy................................................................................56 4-3 Predicted elemental variation effect s on microstructural stability, phase transformation temperatures and elemental s egregatio n.........................94 4-4 Baseline Model A alloy co mposition in wt% and at %...............................95 4-5 Predicted elemental variation effe cts on microstructural stability, phase transformation temperatures and elemental segr egation.......................115 4-6 Phase III alloy compositions and va riation groups in wt% and at%.....116 4-7 Predicted phase transformation te mperatures (C) and ranges for Phase III a lloys..................................................................................... 141 4-8 Predicted and experimental phase transformation temperatures (C) and ranges, for Phase III allo ys............................................................141

PAGE 10

x 4-9 Phase III phase transformation temperature deviations from experimental values (calculated experimental) and modeling error ((deviance/experi mental)* 100)...............................................................142 4-10 Experimental partitioning coefficient values (kx,exp) for as-cast Phase III allo ys.................................................................................................149 4-11 Predicted partitioning coefficient val ues (kx,calc) for as-cast Phase III alloys ......................................................................................................149 4-12 Partitioning coefficient deviati ons (calculated experimental) and modeling error ((deviance/experiment al)*100) for as cast Phase III alloys. .....................................................................................................150 6-1 Alloy 1 com position ...............................................................................215 7-1 Recommendations in approximat e wt%................................................ 217

PAGE 11

xi LIST OF FIGURES Figure page 1-1. Mitsubishi 701 Ga s Turbine Engine...............................................................1 2-1. Typical FCC L12 crystal st ructure.............................................................6 2-2. Transmission electron mi crograph showing cuboidal particles in a matrix for a Ni-9.7Al-1.7Ti-17. 1Cr-6.3Co-2.3W at% alloy............................7 2-3. TEM image of a plate in a SC Ni-bas ed superalloy (SCA)........................9 2-4. Particle diameter vs. hardness for Ni-22Cr-2.8Ti-3.1 Al wt% alloy..............12 2-5. Phase fraction diagram for SAF 2507 Duplex st ainless steel.....................26 3-1. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 2 from Phase III compositions in the heat treated condition....................................................................................................47 3-2 Button alloy sectioning and mounting orientation in metallographic analysis48 3-3. EMPA/WDS compositions in normalized wt% versus line scan measurement points (m) fo r experimental Allo y2.....................................51 3-4 Schematic representat ion of solidification occu rring in a eutectic binary phase diagr am...........................................................................................52 4-1. Predicted phase fraction diag ram for baseline Model al loy........................55 4-2. Predicted phase diagram fo r the baseline M odel allo y...............................56 4-3 Predicted Cr variation effects on TCP equilibrium phase amounts..............58 4-4 Predicted Cr variation effects on phase transformation temperatures..........59 4-5. Predicted Cr variation effe cts on elemental segregation............................60 4-6. kcalc comparisons between the baseline Model alloy and Cr variants.......61 4-7. Predicted Al (and Ta) variation e ffect on TCP equilibrium phase amount with respect to Al (w t%) concent ration.......................................................62

PAGE 12

xii 4-8. Predicted Al (and Ta) variation effects on phase transformation temperatures with respect to Al (wt%) c oncentrati on.................................63 4-9. Predicted Al (and Ta) variation effects on elemental segregation with respect to Al (wt% ) concentra tion..............................................................64 4-10. kcalc comparisons between the baseline Model alloy and Al (and Ta) variant s......................................................................................................65 4-11. Predicted Ti (and Ta or Al) va riation effects on TCP equilibrium phase amount with respect to Ti (wt%) conc entrati on..........................................66 4-12. Predicted Ti (and Ta or Al) variation effects on phase transformation temperatures with respect to Ti (wt%) c oncentrati ons...............................66 4-13. Predicted Ti (and Ta or Al) vari ation effects on elemental segregation with respect to Ti (w t%) concent ration.......................................................68 4-14. kcalc comparisons between the baseline Model alloy and Ti variants......69 4-15. Predicted Re (with Ta, Al, and W) variation effects on TCP equilibrium phase amounts with respect to Re concentration (weight percent)............70 4-16. Predicted Re (with Ta, Al, and W) variation effects on phase transformation temperatures with respec t to Re (wt%) concentration........71 4-17. Predicted Re (with Ta, Al, and W) variation effects on elemental segregation with respect to Re (wt%) concen tration..................................72 4-18. kcalc comparisons between base line Model alloy and Re variants..........73 4-19. Predicted C variation effect on TCP equilibriu m phase amount...............74 4-20 Predicted C variation effects on phase transformation temperatures.........75 4-21. Predicted C variation effect s on elemental s egregati on............................76 4-22. kcalc comparisons between the baseline Model alloy and C variants......76 4-23. Predicted Co variation effe ct on TCP equilibrium phase amount.............77 4-24. Predicted Co variation effect s on phase transformation temperatures.....78 4-25. Predicted Co variation effe cts on elemental segregat ion..........................79 4-26. kcalc comparisons between Model alloy and Co vari ants........................80 4-27. Predicted Ru variation effe ct on TCP equilibrium phase amount.............80

PAGE 13

xiii 4-28. Predicted Ru variation effect s on phase transformation temperatures.....81 4-29 Predicted Ru variation effect s on elemental s egregation...........................82 4-30 kcalc comparisons between Model alloy and Ru variants.........................83 4-31. Predicted W (and Mo) variation effect on TCP equili brium phase amount with respect to W (w t%) concent ration.......................................................84 4-32. Predicted W (and Mo) vari ation effects on phase transformation temperatures with respect to W (wt%) conc entrati on.................................84 4-33. Predicted W (and Mo) variati on effects on elemental segregation with respect to W (wt%) concentrati on..............................................................85 4-34. kcalc comparisons between Model alloy and W (and Mo) variants..........86 4-35. Predicted -former variation effects on TC P equilibrium phase amounts.88 4-36. Predicted -former variation effects on phase transformation temperatur es..............................................................................................89 4-37. Predicted -former variation effects on elemental s egregatio n................90 4-38. kcalc comparisons between Model alloy and -former variants..............91 4-39. Calculated heat treat ment window (solidus solvus) vs. melting range (liquidus solidus) for the baseline M odel composition, the compositional variants in Phase I of alloy development, and selective 1st and 2nd generation commercial and ex perimental alloys........................................92 4-40. Predicted Re variation effe cts on TCP equilibrium phase amounts with respect to Re (wt% ) concentra tion ............................................................98 4-41. Predicted Re variation effe cts on phase transformation temperatures with respect to Re (w t%) concentra tion ................................................... 100 4-42. Predicted Re variation effects on elemental segregation with respect to Re (wt%)..................................................................................................102 4-43 Predicted Cr (with Ti and Al) vari ation effects on TCP equilibrium phase amounts with respect to Cr (wt%) c ontent................................................103 4-44 Predicted Cr (with Ti and Al) vari ation effects on phase transformation temperatures with respect to Cr (wt%) content........................................104 4-45. Predicted Cr (and Ti and Al) va riation effects on elemental segregation with respect to Cr (wt%) cont ent..............................................................105

PAGE 14

xiv 4-46. Predicted -former variation effects on TC P equilibrium phase amounts.106 4-47. Predicted -former variation effects on phase transformation temperatur es ...........................................................................................108 4-48. Predicted -former variation effects on elemental segregation .............109 4-49. Predicted elemental variation effe cts on the amount of TCP equilibrium phases with respect to Al/Ti ra tio.............................................................111 4-50. Predicted elemental variat ion effects on phase transformation temperatures with respec t to Al/Ti ratio....................................................112 4-51 Predicted elemental variation effe cts on elemental segregation for Al/Ti variants with respect to Al/Ti ra tio............................................................113 4-52. Alloy 1 in the as-cast condi tion................................................................117 4-53. Microstructural characteristics for Alloy 1 in the heat treated condition..118 4-54. Alloy 2 in the as-cast condi tion................................................................119 4-55. Alloy 2 in the heat treated c ondition.......................................................120 4-56. Alloy 3 microstructure in the as-cast condition. .......................................123 4-57. Alloy 3 microstructure in the heat treat ed conditi on................................123 4-58 Alloy 4 microstructure in the as-cast condition ........................................125 4-59. Material microstructure for A lloy 4 in the heat tr eated condition.............126 4-60. Alloy 5 in the as-cast condi tion................................................................127 4-61. Alloy 5 material microstructu re in the heat tr eated conditi on..................128 4-62. Predicted Al/Ti ratio (and Ta) va riation effects on TCP equilibrium phase amounts with respect to Al/Ti ra tio...........................................................131 4-63. Predicted Cr (with Al and Ta) va riation effects on TCP equilibrium phase amounts with respect to Cr cont ent..........................................................132 4-64. Predicted Re variation effects on the amount of TCP equilibrium phases133 4-65. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 1 in the as-cast c ondition ...........................................134 4-66. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 1 in t he heat treated c ondition....................................134

PAGE 15

xv 4-67. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 2 in the as-cast c ondition ...........................................135 4-68. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 2 in t he heat treated c ondition....................................135 4-69. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 3 in the as-cast c ondition ...........................................136 4-70. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 3 in t he heat treated c ondition....................................136 4-71. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 4 in the as-cast c ondition ...........................................137 4-72. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 4 in t he heat treated c ondition....................................137 4-73. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 5 in the as-cast c ondition ...........................................138 4-74. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 5 in t he heat treated c ondition....................................138 4-75. Comparison between experimental and calculated phase transformation temperatures for Phas e III allo ys............................................................142 4-76. Comparison between experimental and calculated melting ranges and heat treatment windows for Phase III alloys ...........................................143 4-77. Heat treatment window vs. melting range for the baseline Model composition, the base line Model A compositio n, the compositional variants in Phase III, and selective 1st and 2nd generation commercial and experimental alloys. ..........................................................................144 4-78. Predicted and experimental phase transformation trends for Al/Ti ratio variants with respect to Al/Ti ra tio............................................................146 4-79. Predicted and experimental phase transformation trends for Cr variants with respect to Cr (w t%) concentra tion.................................................... 147 4-80. Predicted and experimental phase transformation trends for Re variants with respect to Re (w t%) concentra tion.................................................... 148 4-81. Comparison between experimental and predicted partitioning coefficient values (kexp vs kcal) fo r Phase III alloys ................................................151 4-82. Experimental and predi cted partitioning coefficien t values for elements.152

PAGE 16

xvi 4-83. Predicted and experimental segregat ion trends for Al/Ti variants with respect to Al/Ti (wit h Ta) variat ions..........................................................153 4-84 Predicted and experimental segregat ion trends for Cr variants with respect to Cr (wt%) concentrati on............................................................154 4-85. Predicted and experimental segr egation trends for Re variants with respect to Re (wt% ) concentra tion...........................................................156

PAGE 17

xvii Abstract of Thesis Pres ented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Master of Science COMPUTATIONAL DESIGN OF NICKEL BASED SUPERALLOYS FOR INDUSTRIAL GAS TURBINE COMPONENTS By Alma Stephanie Tapia May 2006 Chair: Gerhard E. Fuchs Major Department: Material s Science and Engineering Ni-based superalloys play an essential role in the advancement of power technology. Recent initiatives to incr ease efficiency and decrease emissions in power generating industrial gas turbines (IGTs) can only be met by increasing turbine inlet temperatures and turbine component temperature capabilities. To reach these target operati ng temperatures, traditional IGT processing will need to transition to directional solidification pr ocessing, for single crystal turbine blade production. The successful use of single-crystal allo ys in IGT applications is contingent upon overcoming processing problems such as defect formation, and maintaining microstructural stability once in service. Elemental segregation resulting from the casting process, in particular, is link ed to defect formation and the formation of deleterious phases over the extended lifet ime of the IGT component. To avoid the formation of these deleterious phases solutionizing heat treatments between

PAGE 18

xviii the solvus and the solidus temperatures are used to reduce or eliminate segregation from the as-cast materials. To investigate how typical Ni-based s uperalloy elemental additions affect microstructural stability, phase transfo rmation temperatures, and material segregation behavior, a baseline alloy co mposition was used as the foundation from which two iterations of element al variation effect evaluations were conducted. Elemental trends were assessed using a thermodynamic equilibrium module in the 3.0 Java-based Material s Properties Program (JMatPro). Property trends from the Phase I and Phase II studies were used to redefine the baseline alloys chemistry. Compositional modifications resulted in five experimental alloy compositions t hat were manufactured and experimentally tested as a comparison to theoretical results. The JMatPro thermodynamic equili brium module was evaluated and elemental relationships were assessed. Conclusions about elemental effects on microstructural stability, phase transfo rmation temperatures, and material segregation were drawn, which may cont ribute to the development of better alloys for single crystal IGT use in the future.

PAGE 19

1 CHAPTER 1 INTRODUCTION Over the last several decades, Ni-based superalloys have played an essential role in the advancement of pow er technology. Characterized by their high structural, surface, and property st ability, Ni-based superalloys are widely used in the highest temperature component s of power generating industrial gas turbines (IGTs), including turbine di scs, turbochargers, blades, and vanes [3]. The need to increase power output (efficiency) and decrease emissions in power-generating industrial gas turbines (IGTs), can only be met by increasing turbine inlet temperatures and turbi ne component temperature capabilities. Land-based industrial gas turbines (IGT s) operate at inlet temperatures rapidly approaching 1500 C for service lif etimes up to, or in excess, of 10,000 h [45]. IGTs, such as the Mitsubishi 701 seen below (Figure 1-1), are exposed to high temperatures and corrosive environments for a significant portion of their lives, making their components susceptible to hot corrosion (or sulfidation)[1,65]. Figure 1-1. Mitsubishi 701 Gas Turbine Engine

PAGE 20

2 Hot corrosion can be described as the accelerated surface attack of components due to condensed alkali metal salts, such as Na2SO4. When a gas turbine ingests air from the atmosphere to mix it with injected fuel for burning, combustion gases remain that may be c ontaminated with corrosi ve impurities. IGTs consume more air than fuel, with air-tofuel consumption ratios of up to 50 to 1; thus, even a small am ount of sodium chloride and sodium sulfate in the atmosphere can react with residual sulfur in the fuel, leading to severe corrosion problems [51]. Sodium chloride will react with sulfur to form Na2SO4 as shown below. 2NaCl + SO2 + 02 Na2SO4 + Cl2 Currently, the baseline alloy used for IG T applications is IN738. IN738 is a polycrystalline material that exhibits good hot corrosion resistance but does not meet the increasing temper ature demands of the power industry. To increase material temperature capabilit ies for IGTs, the use of si ngle crystal (SC) turbine blades will be required. Processing problem s, such as castab ility, that come from fabricating inherently large IGT components must be factored into the selection of an appropriate SC IGT alloy. Some of the single crystal alloys presently considered for IGT applicati ons include CMSX-4 and PWA 1483. PWA 1483 demonstrates an acceptable le vel of hot corrosion resistance and castability for IGT applicat ions but exhibits limited strength in comparison to other single crystal alloys (i.e., CMSX4). CMSX-4, developed for aerospace applications, is a second -generation, si ngle crystal alloy that exhibits high strength at elevated tem peratures but demonstrat es poor hot corrosion

PAGE 21

3 resistance and castability. Other alloys such as CMSX-11B and CMSX-11C are first-generation, experimental alloys designed, specifically, for single crystal IGT use. Both CMSX-11B and CMSX-11C demonstrate extremely good blends of hot corrosion and oxidation resistance but are prone to recrystallization and freckle formation (along with CMSX-4) during processing [10,34]. SC-16, a firstgeneration single crystal alloy developed by Onera, is not commonly used in the United States and may exhibit po or hot corrosion resistance. The successful use of single-crystal allo ys in IGT applications is contingent upon overcoming processing problems such as defect formation, and maintaining microstructural stability once in service. The increased number of elemental additions in a Ni-based superalloy and t he complex interactions of these additions reveal a need to investi gate elemental variation effects on microstructural stability, phase transfo rmation temperatures, and material segregation behavior. The present work uses a design approach aimed toward the development of a set of alloys for industrial gas tu rbine application. In the hopes of better understanding elemental variation effe cts on the aforementioned material properties, a baseline alloy co mposition named the baseline Model alloy (based on CMSX-4 and PWA 1483) was used as the foundation from which two iterations of elemental variation effe ct evaluations were conducted (Phase I and Phase II). The thermodynamic equi librium module in the 3.0 Java-based Materials Properties Program (JMatPro) was utilize to evaluated Phase I and Phase II theoretical property trends and determine chemistry modifications to

PAGE 22

4 the baseline Model alloy. Five variant alloy compositions were tailored using JMatPro modeling techniques and were labo ratory tested for validation purposes. To address the effects of additions previously shown to influence hot corrosion and material stability, final co mpositions incorporated char acteristic variations of Al/Ti ratio (with Ta variation), Cr (with Al and Ta variations), and Re content for comparison [35, 40, 45]. This study evaluates the computati onal capabilities of the JMatPro thermodynamic equilibrium module to predict material properties related to defect formation and microstructura l stability. Through computational techniques this work also contributes to a better under standing of elemental variation effects on microstructural stability, phase transfo rmation temperatures, and material segregation behavior to facilitate the devel opment of better alloys for future single crystal IGT use.

PAGE 23

5 CHAPTER 2 LITERATURE SEARCH This chapter will provide an overview of the microstructure, chemical composition, casting, and processing of Ni-based superalloys, as well as a discussion of the current methods utilizing empirical and computational models to predict deleterious phase formation and material properties. Microstructure A basic Ni-base superalloy consists, mainly, of a two-phase equilibrium microstructure: the gamma ( ) nickel-chromium matrix and the gamma-prime ( ') precipitate. Carbon additions can lead to the formation of carbides and certain service/heat treatment conditions may result in the formation of deleterious TCP phases. The Matrix The continuous gamma matrix ( ) is a solid solution FCC nickel based austenitic phase, strengthened by high perce ntages of Co, Cr, Mo, W, Ti, and Al [44,51]. The Phase The phase is an intermetallic compound t hat provides strength to the Nibase superalloy [44]. The phase precipitates coherently out of the matrix with an FCC L12 ordered superlattice structure to become the materials major precipitate (Figure 2-1) [44,51].

PAGE 24

6 Figure 2-1. Typical FCC L12 crystal structure This L12 structure is of the Cu3Au-ty pe, where Ni atoms o ccupy the centers of the cube faces and Al typically resides in the cube corners [13]. Ti, Nb, and Ta also contribute to precipitation and can substitu te for up to approximately 50% of the . The binary atomic arrangement has the chemical formula Ni3Al, Ni3Ti, or Ni3(Al,Ti), which mainly consists of Al, Ti, Nb, or Ta [51]. The forming elements, Ti, Nb, and Ta, also increase the anti-phase boundary energy ( APB) [7,35,45]. The phase is, in large part, responsib le for the elevated-temperature strength in a Ni-based superallo y; since the strength of actually increases with increasing temperatur e [13]. The total former content in most Ni-base superalloys is usually maintained at about 12-15 at%. The attractive properties of / superalloys has resulted in a continuous increase in the volume fraction. Recent alloys may contain over 60% and can approach 75% in some Ni-based superalloys [45]. However, increasing volume fractions, must be balanced with modification of the -matrix composition since a c oncentration of the refractory elements in the matrix, can lead to deleterio us phase formation [34].

PAGE 25

7 The / Mismatch morphology is affected by lattice mi smatch, strain energy, and interfacial energy. The lattice mismatch, a result of the differences in and lattice parameters (a and a respectively) result in an in terfacial misfit energy [13]. The unconstrained lattice misfit parameter (d) is defined below. d = (a a ) / a In most superalloys, the precipitate is in tension while the matrix is under compression, leading to a small misfi t. This small misfit results in a cuboidal precipitate morphol ogy and helps ensure a low / interfacial energy [13,33]. The transmission electron micrograph below (Figure 2-2) depicts a typical cuboidal morphology [45,51]. Figure 2-2. Transmission electr on micrograph showing cuboidal particles in a matrix for a Ni-9.7Al-1.7Ti-17.1Cr-6.3Co-2.3W at% alloy Carbides Primary carbides, or MC carbides, form as discrete FCC particles during the solidification of an alloy, and are typically observed throughout the material [44,45]. MC carbides form due to interactions between carbon and reactive or Gamma prime ( ) precipitate Gamma ( ) matrix

PAGE 26

8 refractory metals, such as Ti, Ta, Hf, or Ta apart from other elements like Cr, Mo, W, and Nb [45]. The formation of carbi des in the material matrix consumes refractory elements that contribute to solid solution strengthening or formation, which could also promote phase inst ability during service [45]. Carbide morphologies range from cubic to script, but carbides are most commonly seen as large blocky or spherical particles in superalloys. Primary FCC close-packed carbides are some of the most stable compounds found in nature [45]. Secondary carbides, of the M23C6 type, form through the decomposition of MC type, primary carbides. The degener ations of MCs occur during lower temperature heat treatment s and service in the 760980C range in alloys containing moderate to high amounts of Cr, apart from W and Mo. Phase Instabilities Topologically close packed (TCP) phases Deleterious topologically close packed (TCP) phases can result from microstructural/chemical instabilities in nickel-based superalloys, during the heat treatment or service lifet ime of a component [54]. TCP phases exist in many forms, but typically appear in the sigma ( ), miu ( ), or Laves form. The phases have characteristic close-packed atom planes stacked in the sequence ABCABC which are parallel to the {111} planes of the matrix [45]. An example of a TCP phase; identified by Strunz, in an experimental nickel-base superalloy, is seen below (Figure 2-3) [48].

PAGE 27

9 The chemical formula (Cr,Mo)x(Ni,Co)y has been reported for the phase, where x and y vary from 1 to 7. In general, TCP p hases are predominately made up of refractory elements, and the phase is characteristically dominated by Mo and Co [45]. Accordingly, TCP phase forma tion results in the depletion of solid solution strengthening elements such as W, Mo, Cr, Co and Re from the matrix. The depletion of these st rengthening elements may produce a marked reduction in rupture life at high temper atures [8]. The intrinsically brittle nature of the topologically close-packed (TCP) phases reduces the ductility of an alloy. The physical hardness and, many times, plate-like morphology of TCP phases also provide a source for crack initiation and propagation, leading to material failure. Figure 2-3. TEM image of a plate in a SC Ni-based superalloy (SCA) Secondary reaction zones (SRZs) The occurrence of secondary reaction zones was noted by Walston et al. in Rene N6 [44,61,62]. Secondary reaction zones (SRZs) are regions within a material that contain and P phase needles. These regions are referred to as cellular colonies and can form in dendrit e cores and along low angle boundaries, common in single crystal castings [44].

PAGE 28

10 SRZs are thought to form in areas of local elemental enrichment due to either coating processes or material segregation (resulting from casting processes), apart from factors such as st rain energy and misfit strains [44,61]. These cellular colonies have been obs erved in superalloys containing high concentrations of refractory elements, demonstrating the highest affinity for Rebearing alloys [25,62]. The presence of SRZs beneath material coatings can eventually affect the ruptur e strength of a material and induce premature failures when crack initiation occurs at SRZ interfaces [25,62]. Deleterious phase considerations Overall, deleterious phases tend to nucleate in materials with excessive additions of refractory elements or in areas that are enriched with high concentrations of refractory elements [ 51]. Consequently, careful chemistry control to balance alloy composition and effective homogeniza tion treatments are necessary to minimize regions of localized elemental enrichment and, subsequently, prevent deleterious phase formation. To restrict microstructural instabili ties, limits have been introduced to the concentrations of solid solution strengthener s [25,34]. Typical Re bearing alloys, such as CMSX-4, CMSX-10, and Rene N6 ex hibit deleterious phase formation of either the TCP or SRZ type [1,25,62]. To improve long-tem stability, Re bearing alloys have, in good measure, reduced Cr and Ti concentrations (elements that can provide the hot corrosion resistance vital to industrial gas turbine applications) [35,44]. Studies co nducted with Rene N6, showed that microstructure stability could be improved by decreasing the level of Re in a material and by introducing Mo in order to keep a comparable total amount of

PAGE 29

11 strengthening refractory elements [44,61]. It is evident, therefore, that modifications to alloy composition as a means to improve material stability can be carefully balanced with spec ific IGT material needs. Chemical Composition The properties of Ni-base superallo ys are strongly dependent upon the chemical composition of the given materi al. The high solubility of Ni for a wide variety of alloying elements is largely due to its partially filled third electron shell. Typical Ni-based superalloy compositions can contain up to 12 to 13 different elements, without sacrifici ng microstructural stability [1,2]. However, the interaction all of these solutes ra ises the challenge of balancing alloy compositions to obtain specific desired material properties. Strengthening Methods A key alloying effect is the solid solution strengthening and precipitation hardening of a material. Solid Solution Strengthening Solid solution strengthening arises from so lute interactions with dislocations in the materials matrix. These solu tes strengthen the material by introducing atomic diameter differences, elastic intera ctions, modulus interactions, electrical interactions, and short-range/long-range order interactions [44,45]. The lowering of the stacking fault energy with alloying additions also increases resistance to cross slip and dislocation motion, thus, increasing material strength. Solid solution strengtheners include Re, W, Mo Cr, Co, Ti and Al [45].

PAGE 30

12 Precipitation Hardening Precipitation hardening of a material re sults from dislocation interactions with coherent particles within a material ma trix. Material strength increases with particle size, due to an increased amount of dislocation cutting that occurs in the larger coherent pr ecipitates in the matrix. Once a crit ical particle size is reached; precipitates be come incoherent with the matrix. At this critical stage, dislocations begin to bypass precipitat es, resulting in decreased material strength. The relationship between strength and precipitate size was shown in a study by Mitchell, for a Ni-22Cr-2.8Ti -3.1Al wt% alloy (Figure 2-4)[45]. Figure 2-4. Particle diameter vs. hardness for Ni-22Cr-2.8Ti-3.1Al wt% alloy Alloying Elements Common superalloy elemental additions (C o, C, Cr, Mo, W, Al, Ti, Ru, Re, and Ta) and some of their key characteri stic influences on Ni-based superalloys are mentioned below.

PAGE 31

13 Cobalt Cobalt concentrations of approximat ely 2-15 wt% are used in most Nibased superalloys [44]. Co additions st abilize the material microstructure, reduce the solvus temperature, and reduc e the stacking fault energy ( SFE) [11,44]. Co is also reported to part ition to the dendrite core, and provides a limited amount of solid solu tion strengthening [44,45]. Carbon Minor additions of carbon can be used in Ni-based superalloys. A study of Rene N4 showed that a 0.05 wt% C addition yielded increased rupture strength at high temperatures [35]. An in creased tolerance for grain boundary misorientations at low angle boundaries (LABs) was also attributed to C additions [16,35]. C additions have also been s hown to decrease refr actory element (W, Re) partitioning to the dendrite core, improv ing microstructural stability [50]. Ruthenium Ruthenium additions of approximately 0 to 9 at% in Ni-based superalloys, stabilize the materials microstructure and provide the material with solid solution strengthening [13,28,45]. Ru additions also increase the liquidus temperature and tends to partition to the dendrite core [45,51]. Rhenium Rhenium is a strong solid solution strengthener that improves creep strength, and increases t he temperature capabilities of a material [51]. The amount of Re used in an alloy categorizes it as a first, second, or third generation alloy, containing 0 wt% Re, 3 wt% Re, or 6 wt% Re, respectively. Rhenium additions are expected to increase the density and liquidus te mperature of a Ni-

PAGE 32

14 base superalloy [51]. The use of this refr actory element may result in convective instabilities during solidification. Re al so partitions to the dendrite core, which can lead to the development of deleterious phases in the dendritic region [44,51]. The high temperature strength supplied by Re additi ons must be balanced with the concern for decreasing microstructu ral stability. Refractory elements, including Cr and Ti concentrations (addi tions that increase hot corrosion resistance), have been reduced in more re cent alloys to compensate for Re additions [51]. Chromium Chromium concentrations typically r ange from 10-20 wt% for industrial gas turbine applications. Cr additions improve hot corrosi on and oxidation resistance due to the formation of a protective Cr2O3 rich oxide scale [35,44]. The Cr oxide hinders diffusion and effectively stops environm ental reaction with the bulk alloy. Cr has also been reported to reduce the solvus temperature and the antiphase boundary energy ( APB) of the phase [9]. Cr tends to partition to the dendrite core and may promote delete rious phase formation [24]. To improve microstructural stability, 2nd and 3rd generation alloys have notably reduced Cr concentrations [51]. Th e low Cr concentrations in the higher generation aero-alloys could result in hot corrosion concerns for IGT components that require an extensiv e amount of hot corrosi on resistance [44]. Aluminum and Titanium Aluminum Aluminum concentrations typically used in most Ni-base superalloys range between approximately 3-6 wt %. Aluminum is a low density addition to Ni-based

PAGE 33

15 superalloys, which acts as a primary former, improves material castability, and partitions to the inderdendritic region [12, 45]. Al additions are also considered essential for oxidation resistance [35]. The Al solute contributes to the development of an Al2O3 scale, which increases its ox idation resistance at high temperatures [45]. Titanium Titanium concentrations in most Nibased superalloys can range from 0 to 5 wt%. Titanium is also a low density addi tion to Ni-based superalloys, which acts as a former, strengthens the phase, and increases the anti-phase boundary energy ( APB) [7,45]. Ti partitions to the interdendritic region and generally decreases the oxidation resi stance and increases the hot corrosion resistance of the alloy [24,35]. Al/Ti Ratio Al/Ti ratio is used to illustrate the infl uence of Al and Ti on the oxidation and corrosion resistance of an alloy. Ross and OHara, reported that the Al/Ti ratios in Rene N4 had a significant impact on oxidation and hot corrosion resistance [35]. The study on Rene N4 showed that decreasing Al/Ti ratios, increased hot corrosion resistance, but decreased oxi dation resistance [35]. Consequently, using lower Al/Ti ratios (moderate Al c oncentrations with increased Ti additions) can provide IGT components with increased resistance to hot corrosion attack. Recent Ti reductions in 2nd and 3rd generat ion alloys could then result in the degradation of hot corrosion pr operties that are so impor tant for IGT applications [51].

PAGE 34

16 Tantalum and Tungsten Tantalum Ta concentrations of approximately 4 to 12 wt% are used in many Ni-based superalloys [44]. Ta is a former and acts as a strong solid solution strengthener [45]. Ta additions increase the anti-phase boundary energy ( APB) and tend to partition to the interdendrit ic region [24,53]. Ta has also been reported to improve allo y castability [31]. Tungsten Tungsten additions of approximat ely 5-8 wt%, strengthen the matrix through solid solution strengthening and syner gistic effects with Re strengthening mechanisms [6]. The use of W in superallo ys is reported to increase the incipient melting point, decrease microstructura l stability, and increase hot corrosion susceptibility [45]. W par titions to the dendrite core and decreases material castability [6,45]. Ta/W Ratio The Ta/W ratio is also used to eval uate an alloys castability. Increased Ta/W ratios (from increased Ta or r educed W concentrations), are reported to decrease the incidence of casting defec ts, caused by convective instabilities during processing [34,38]. Molybdenum Molybdenum additions of 0-3 wt%, ar e used to increase solid solution strengthening of the matrix [6,15]. Mo decreases microstructural stability and has been reported to partition to the dendrite core [15,16,21].

PAGE 35

17 Hafnium Hafnium additions of 0-0.2 wt% are us ed to increase oxide scale adherence to the metal substrate. Minimal Hf additions enhance coated oxidation life by diffusing into a metals surface oxide [10,45]. Casting and Processing Casting Concerns and Defect Formation The first solids to form during solidi fication are gamma dendrites [54]. Solute and solvent fluxes during dendrit e growth cause solvent/solute buildups that are unable to redistribute completely before solidification is complete. The supersaturation of the liquid with segregating elements, resu lts in the formation of secondary solidification consti tuents, such as MC carbides, in the interdendritic regions [54]. Elemental build ups in t he dendritic solid and in terdendritic liquid result in material segregation. The degree of segregation in a material is, typically, measured with the use of elemental partitioning coefficients (k) The partitioning coefficient for a given element (x) is expressed as the ratio of interdendritic to dendritic composition; as seen below. kx = Cx,core/ Cx,inter A partitioning coefficient (k) value of one indicates that no partitioning is present for a given element. More succi nctly, an equal amount of the element was measured in both the dendrite core and the interdendritic region, exhibiting no preference or segregation during solidification. The ratio also allows the direction of segregation to be determi ned. Partitioning coefficients (kx) less than

PAGE 36

18 unity, indicate that an element partitions to the interdendritic region. Solutes with partitioning coefficients gr eater than unity segregate to the dendrite core. Elements such as Ni, Ta, and Al, have been previously reported to segregate toward interdendritic regi ons. Cr, Co, W, and Re have been previously reported to segregate to the dendrite core [4,45]. The loss of control over temperatur e gradients in conventional casting techniques, can also lead to convex solidifica tion interfaces that result in material segregation and defect formati on [54,64]. Small castings such as those used in the aero engine field, can more easily ma intain steep temperature gradients as compared to large IGT components. The diff iculty in maintaining these gradients in large single crystal blades, results in a high propensity towards material segregation and defect formation. There are several defects associated with the casting of single crystal structures, su ch as low angle boundaries, slivers, and freckles [13,54]. Defect formation and solute partitioning may be controlled through a combination of alloy design and care ful control of the casting process. Solutionizing/Homogeniza tion Heat Treatments Elemental segregation, resu lting from the casting process, is typically reduced or eliminated by solu tionizing/homogenization heat treatments. Solution heat treatments are conducted at temper atures high enough to dissolve the phase and homogenize the alloy for the, subsequent, re-precipitation of uniform precipitates in the material. Solutionizing heat treatments are limi ted to a temperature range between the solvus and the solidus, called the heat treatment window. Complete homogenization is dependant on both the te mperature and time of the heat

PAGE 37

19 treatment. The ability to homogenize cast st ructures may be severely restricted by low or even negative heat treatment wi ndows; which can result in incipient melting [3]. A study conducted for an experimental Ni-base superalloy (Re3) demonstrated that solution heat treatment s were unable to fully homogenize the material due to minimized heat treatment times. The Re3 solutionizing heat treatments were reduced in time, to avoid the risk of incipient melting, due to the materials narrow heat treatment windo w [33]. Inadequate solutionizing can result in areas of residual elemental enrichment which can then lead to deleterious phase formation. Predictive Methods The development of new Ni-based s uperalloys for the modern gas turbine have primarily been the result of trial and error processes. The task of identifying new alloys that provide increasing temperature capabilities yet balance good microstructural stability is becoming increas ingly difficult. Given the high degree of complexity in nickel based superalloy ch emistry, it is of no surprise that a major area of concern and att ention has been microstructu ral stability [75]. This was evident in the microstructural stability difficulties faced for 720Li, a high strength nickel-base superalloy used for turb ine disk applications. This turbine disk alloy, in the powder or conventi onally processed cast and wrought form, suffers rapid precip itation of the TCP phase above 650C. Property degradation concerns encour aged C.J. Small and N. Saunders to investigate new alloy compositions based on 720 Li and Waspaloy [46]. With the need for a tool that can gui de initial alloy-chem istry selection, semi-empirical and computational m odels have been developed to predict

PAGE 38

20 deleterious phase formation and/or simulate material properties based solely on the alloy composition with some success. Existing tools for design applications will be described in more detail below. PHACOMP Early attempts to predict TCP phase formation were based on the PHAse COMPutation (PHACOMP) method. This conventional calculation tool relates chemical composition to electron valenc e theory to predict the formation of deleterious phases in an alloy. PHACOMP relies on the importance of el ectronic interactions (the unpaired d-electrons or electron vacancies) for each element in an alloy. The average electron vacancy concentration (Nv) of the solid solution matrix is calculated on the premise that closely pack ed phase instabilities (such as ) are electronic compounds. In the PHACOM P technique, the alloy matrix composition is calculated by subtracting the normal pr ecipitation phases (carbides, borides, ) from the total composit ion before the average electron-hole concentration (Nv) is computed as follows.[4] Nv = fi (nv)i where fi is the atomic fraction of an element (i) in the matrix and (nv)i is its corresponding electron-hole number. A ty pical equation to calculate an alloys electron-hole weighted av erage is shown below [45]. Nv = 4.66 (Cr + Mo) + 3.66(Mn) + 2. 66(Fe) + 1.71(Co) + 0.61(Ni) The calculated Nv value is compared to some critical value that determines whether the alloy is prone to sigma-phase precipitation. An average electron

PAGE 39

21 hole number Nv above the threshold would indicate that an alloy is sigma-prone, while a number below the critical val ue would deem the material sigma safe and suitably stable for practical applications [71]. The critical value approximates 2.5 for individual alloys but is not necessa rily a fixed value common to all metals. The models oversimplified nature calcul ates the matrix composition based on assumptions of the amount, type, and co mposition of carbides, borides, and intermetallic compounds that are expected to precipitate during processing and service [45]. Dreshfield and Ashbrook studied a wide range of cast and wrought Inconel alloys using the PHACOMP met hod. Although electron vacancy number calculations indicated a tendency to form the sigma phase, no deleterious phases were observed in any of the examined materials [8]. It is evident that the accuracy of PHACOMP predictions depends on the validity of the assumptions made for estimating the composition of the matrix. Work conducted by Milhalis on a variety of alloys used experimentally determined matrix compositions in PHACOMP analysis, resulting in more accu rate predictions of material phase stability [14]. Difficulties with usi ng the PHACOMP method are described by Murphy et. al. who concluded that PHAC OMP calculations are not accurate unless the compositions of the precip itating phases are available [45,62]. The PHACOMP method lacks the ability to handle the true complexity associated with topologically close-pack ed (TCP) phase formation, in addition to entirely omitting the development of ot her deleterious phases such as and Laves. As a result, PHACOMP techniques do not apply to Re containing alloys, which can contain , and P type TCP phases. Furthermore, PHACOMP is not

PAGE 40

22 capable of providing any details on stability temperatur e ranges or phase boundaries. New PHACOMP Several improvements to the PHAC OMP technique were suggested to increase the accuracy of phase instability calculations. A new technique, known as the new phase computation method (New PHACOMP), was developed by Morinaga et al., to predict the formati on of deleterious phases, such as the phase and phases, in nickel-based s uperalloys [58]. The New PHACOMP method takes into account atomic size factors (atomic radius) in electronic structure calculations (e lectro-negativity). New PHACOMP uses the molecular orbital method to obtain two alloying par ameters used to predict deleterious phase formation. One param eter is the d-orbital energy level of alloying transition metal elements (M) in a base metal (X), known as the Md level. The other alloying parameter is a measure of the strengt h of the covalent bonds between M and X atoms, k nown as the bond order (Bo) [59]. For Ni-based superalloys, the average dorbital energy level (Md-value) is calculated for alloying transition elements in the matrix. Just as the PHACOMP method defines a threshold value, the New PHACOMP method defines a critical Md value, above which instability occurs. Similar problems to those in the PHACOMP method arise in the New PHACOMP method. The oversimplif ied nature of t he New PHACOMP calculations also does not take in to account solute interactions.

PAGE 41

23 NASA Rene N6 Model Work conducted by Frank Ritzert et al. attempted to describe the occurrence of TCP ( and P) phases in Ni-based superalloys, paying particular attention to the potential synergistic e ffects of alloying el ements on deleterious phase formation [60]. In general, refractory metal content in a Ni-base superalloy is thought to contribute to the formation of TCP phas es. On the premise that certain elements, or combinations of alloying el ements, are more pot ent than others in forming TCP phases, a regression m odel was developed on a design-ofexperiments (DOE) methodology for the Ni-based superalloy Rene N6. The regression model developed for Rene N6 calculated both the linear and pairwise interactive effects of Al, Co, Cr, Mo, Re, Ta, and W on final TCP phase content [60]. The resulting relationship to predi ct TCP phase volume fraction (in terms of atomic percent) is seen below [60]. (vol% TCP)1/2 = 16.344782 1.019587(Al) 2.624322(Cr) 3.821997(Mo) + 1.109575(Re) 3.207295(Ta) + 6.462984(W) 2.271803(Co) + 0.052884(Al*Co) + 0.214059(Al*Cr) + 0.300698(Al*Mo) + 0.80011(Co*Re) + 0.257108(Cr*Mo) 5.081598(Re*W) + 1.824441(Ta*W) The confidence interval around the predicted value for Rene N6 in this study was approximately 95%. To simplify the models application, trace elements in the alloys were omitted from calculation. Ti was al so excluded in the model, since it was assumed to behave similar to Ta. An attempt to apply the Rene N6 model to Rene N5 resulted in inflated TCP content pr edictions, indicating that the model is not applicable to all 2nd and 3rd generation Ni-base supera lloys [60]. It is,

PAGE 42

24 therefore, reasonable to c onclude that the usefulness of this relationship is limited to alloys that lie near Rene N6 design parameters. Secondary Reaction Zone (SRZ) Model More extensive studies by Walston et al. on the occurrence of deleterious phases in Rene N6, also indicated that seco ndary reaction zone (SRZ) formation was related to an alloys chemical compos ition [61,62]. Re content played a key role in predicting SRZ formation, due to its extensive segregation during the casting process. Statistical analysis of SRZ formation, measured by quantitative metallographic techniques, produced the fo llowing empirical expression for the relationship between alloy chemistry (i n atomic percent) and the linear % SRZ [62]. [SRZ(%)]1/2 = 13.88 (%Re) + 4.10(%W) 7.07( %Cr) 2.94(%Mo) 0.33(%Co) + 12.13 It is interesting to note that the SRZ empirical correlation is based solely on experimental observations, with no fundament al scientific basis or foundation. Although the relationship was successfully used to minimize SRZ formation in Rene N6; it is exclusively applicable to alloys within the alloys limited composition ranges [61,62]. CALPHAD Computer aided thermodynamic phase diagram calculations (CALPHAD) have been recently used, to predict phase stability in multi-component systems, including Ni-base superalloys. Two main CALPHAD models are the substitutional and the multiple sublattice models. These models predi ct the properties of higher-order systems

PAGE 43

25 from lower-component systems, assuming that higher order interactions are small in comparison to thos e that arise from the binar y terms. Both of these models are broadly represented by the equation below, where G is the free energy of the phase in its pure form, Gideal is the ideal mixing term corresponding to entropy, and Gxs is the excess free energy of mixing of components [13,24]. Once the thermodynamics of the phases are defined, the phase equilibria can be calculated by using Gibbs free energy minimizing routi nes for the multicomponent system, where ni is the number of moles and Gi is the Gibbs energy of phase i. When the minimum Gibbs energy at a given state is achieved, chemical potential, n, of each component, n, is the same in all phases and are related to the Gibbs energy by the equation below. Thermodynamic/mathematical CA LPHAD models require the use coefficients that uniquely describe the proper ties of the various phases in a Nibased superalloy. Coefficients for mu lti-component systems are kept in databases which are proprietary or bas ed on open literature, and are accessed by software packages such as Thermo-Calc or JMatPro [37].

PAGE 44

26 Phase equilibria calculat ions predict temperatur e and chemistry variation effects on phase amounts. The CALPHAD method was used to simulate new turbine disk superalloys, similar to 720 Li and Waspaloy, to increase microstructural stability and crack propagatio n resistance. Through the use of CALHAD calculations, TCP phase formati on was minimized as compared to 720Li [46]. The Java-based Materials Properties (JMatPro) software was developed to facilitate material evaluation by the ca lculation of phase equi libria in complex material systems [38]. JMatPro's therm odynamic calculation software uses the Ni-DATA (ver.6) database for the calculati on of phase equilibria in all types of Nibased superalloys. Figure 2-5. Phase fraction diagram for SAF 2507 Duplex stainless steel In the thermodynamic calculation m odule, Gibbs free energy minimization routines are performed using CALPHAD methods. These minimization systems routinely calculate multi-component, mu lti-phase equilibria as a function of composition or temperatur e. The thermodynamic m odel also includes stability

PAGE 45

27 checking for miscibility gaps or potential or dering to find phase boundaries. The thermodynamic calculations provide a phase fraction diagram for a given alloy chemistry under equilibrium An example of one such phase fraction diagram calculated by N.Saunders and X.Li for SAF 2507 Duplex stainless steel is seen in Figure 2-5 [40]. Once equilibrium information is obtained through conventional thermodynamic methods, JMatPro uses physically based models to correlate equilibrium results to material proper ties. The more generalized software package, JMatPro, allows the calculation of a wide range of material properties. The available material modeling components of this program include: material properties related to thermodynamic calcul ations, solidification, thermo-physical properties, phase transformations, and mec hanical properties, using incorporated theoretical models and proprietary pr operty databases to make quantitative calculations [46]. This software may prov e significant to futu re material design but requires validation between thermodyna mic calculations and final materials properties.

PAGE 46

28 CHAPTER 3 DESIGN AND EXPERIMENTAL PROCEDURE Alloy Design The rapid increase of inle t temperatures in industrial gas turbines (IGTs) produces a corresponding need for advancement in heat resistant and high strength materials [51]. The higher temperature requirements for IGT applications can only be met through the use of single crystal (SC) superalloys. Over the last 30 years, the development of single cr ystal (SC) superalloys has found wide spread use in aircraft jet engines; however, few alloys have been tailored specifically for IGT applications [14,45]. IGT operational environments and fuel im purities make sulfidation or hot corrosion attack an area of major concern. Processing challenges typical of large SC IGT components include convecti ve instabilities during the casting process that lead to defect formation. Additionally, elemental segregation that occurs during solidification requires costly homogenization heat treatments. Incomplete homogenization resulting fr om inadequate heat treatments could even lead to deleterious (TCP) phase formation. In order to develop new SC IGT alloys, a high strength aero-engine composition could be modified to facilit ate the production of larger components that are microstructurally stable. Desired microstructural stability, phase transformation temperatures, and material segregation characteristics could be obtained through the understanding and subsequent control of alloying elements.

PAGE 47

29 In this study, computational techniques were used to identify compositions that could be used for IGT applications. The alloy compositions should reflect the strength of common second generation superalloys such as CMSX-4, and the hot corrosion resistance and castability of first generation superalloys, such as PWA 1483. Two iterations of elemental variation effects on microstructural stability, phase transformation temperatures and material segregation behavior properties were investigated using t he JMatPro thermodynamic equilibrium module. Calculated element al variation trends were used to identify final compositions for experimental validation. Compositional adjustments were sele cted to meet three key property targets outlined in this study: Minimize deleterious topologically close packed (TCP) phase formation o Deleterious TCP phases can result from microstructural/chemical instabilities in nickel-based super alloys during the casting, heat treatment, or service lif etime of a component [30] Thought to act as fracture initiation sites, TCP phases also deplete the solid solution strengthening elements in the matrix leading to a marked reduction in rupture life [8]. Minimiza tion of TCP phase formation can be achieved through compositional adju stments that reduce the total amount (wt%) of TCP phases expected at equilibrium. Achieve a heat treatment window of at least 25 C o The local enrichment of elements in segregated materials can lead to defect or deleterious TCP phase fo rmation. The ability to solution heat treat an alloy to reduce or elim inate chemical segregation, can be limited by the size of the heat tr eatment window. A heat treatment window of at least 25 C is suffi ciently large for adequate solutioning of the precipitate at temperatures that will not risk incipient melting. Elemental adjustments, that incr ease the solidus temperature and depress the solvus, can result in the optimization of the heat treatment window.

PAGE 48

30 Minimize elemental segregation during solidification o During solidification, partitioning of elements to the either the dendrite solid or the interdendritic liquid resu lts in material segregation. A reduction of elemental segregat ion can be achieved through compositional adjustments to decr ease elemental partitioning of specific elements. Decreasing elemental segregation may reduce homogenization processing costs (low er temperatures and shorter times) and may reduce an alloys propensity towards TCP phase formation. Alloy Development Model Base Chemistry The identification of potential IGT a lloy compositions started with the definition of an initial base line alloy. First and second generation commercial and experimental alloys, including CMSX4, CMSX-11B, CMSX-11C, SC-16, and PWA 1483 (Table 3-1), served as guides in defining a simplified baseline chemistry designated Model (Table 3-2). Table 3-1. Nominal composition in wt % of commercial/experimental Ni-base superalloys Ni Cr Co MoRe W Al Ti Ta Hf C CMSX-4 Bal 6.5 9.00.6 3.06.05.61.06.5 0.10 PWA 1483 Bal 12.89.01.9 0.03.83.64.04.0 0.00 0.07 SC-16 Bal 16.00.03.0 0.00.03.53.53.5 0.00 0.00 CMSX-11B Bal 12.57.00.5 0.05.03.64.25.0 0.04 0.00 CMSX-11C Bal 14.93.00.4 0.04.53.44.25.0 0.04 0.00 Table 3-2. Model alloy composition in wt% and at%. NiCrCoMoReWTiAlTiTaHfC wt %Bal10.1211.473.025.258.80.1 at %Bal121211230.05 The baseline Model alloy is a simplif ied second generation single crystal alloy intended to exhibit hot corrosion resistance, high strength, and material castability. The baseline composition c ontains approximately 3 wt% (1 at%) Re, for solid solution strengthening and increas ed creep resistance. In conjunction with the Re addition, 5.96 wt% (2 at%) W was included for solid solution

PAGE 49

31 strengthening and its synergistic effect on Re strengthening [6]. A total of 15 at% in -formers was used in baseline Model allo y including 8.8 wt% (3 at%) Ta and 5.25 wt% (12 at%) Al, for precipitation har dening. The high Al concentration (>5 wt%) was used to stabilize the phase, while increasing oxidation resistance [91]. A high Cr concentration of 10. 12 wt% (12 at%) Cr, was used for hot corrosion resistance needed in IGT applications. Co in the amount of 11.47 wt% (12 at%) Co, was included for microstruc tural stability, the reduction of the stacking fault energy ( SFE), and for limited solid solution strengthening [11,44,45]. The addition of Co was also used to reduce the solvus temperature, increasing t he heat treatment window of the baseline Model alloy [11,25,45]. A minimal am ount of Hf was included to enhance coated oxidation life [10]. Phase I Alloy Development M odeled Elemental Variations Phase I of alloy development was init iated following the selection of the baseline Model alloy composition. Micr ostructural stability, phase transformation temperatures (liquidus, solvus, solidus, MC solvus, solvus, and solvus), and segregation behavior for alloy com positions were predicted using the thermodynamic equilibrium module in t he Java based Materials Program 3.0 (JMatPro). All computational work fo r the project was conducted at Siemens Westinghouse Power Corpor ation (Orlando, Fl). JMatPro's thermodynamic calculation software uses the Ni-DATA (ver.6) database for the calculation of phase equilibr ia in Ni-based superalloys. The full database contains information on Ni, Al, Co Cr, Fe, Hf, Mo, Mn Nb, Re, Ru, Si, Ta, Ti, W, Zr, B, C, N. The database supplies coefficients that uniquely describe

PAGE 50

32 the thermodynamic properties of t he various phases found in Ni-based superalloys, including Liquid, , NiAl, Ni3Nb, , Ni4Mo, NiMo, (Cr,Mo,W), Laves, , R, P, M(C,N), M23(B,C)6, M6C, M7(B,C)3, M2N, M3B, (Fe,Ni..)2B, (Cr,Mo..)2B, M3B2, MB, Cr5B3 TiB2, Ni3Si(h), Ni3Si2, Cr3Si, and Cr3Ni5Si2. This database, is founded on select commercial alloys including: CMSX-4, CMSX-10, CMSX-11B, Rene N6, and PWA 1484 [38,40]. In order to evaluate the database using alloy compositions outside those used for program development, a final set of alloys was included in Phase III of this study. The alloy compositions of interest we re used as the input for the JMatPro thermodynamic calculations. Gibbs fr ee energy minimization routines were started at 1500 C and were performed in the program as the temperature stepped down in increments of 10 C to 900 C. These minimization systems routinely calculated multi-co mponent, multi-phase equilibr ia as a function of the temperature. The thermodynamic model also included stability checking for miscibility gaps or potential orderi ng to find phase boundaries [40]. Calculated phase fraction diagrams between 900 C and 1500 C allowed the identification of critical transition temperatures such as the solidus, liquidus, solvus, and solvus. Given that solu tion heat treatments above the solvus are used to reduce solidification segregation and to control precipitate size and shape, the difference between the solidus and solvus temperatures or heat treatment window was determined. The liquidus and solidus difference or melting range was also determined. Partit ioning coefficient calculations were

PAGE 51

33 conducted for elements previously repor ted as partitioning toward the dendrite core (Cr, Co, W, and Re) and interdendritic regions (Ni, Ta, and Al) [4,45]. The characteristic baseline Model allo y properties were used as a baseline comparison for the other a lloy compositions in Phase I. Once baseline Model alloy properties were calculated, the influences of elemental variations on microstructural stability, transfo rmation temperatures (liquidus, solvus, solidus, MC solvus, solvus, and solvus), and segregat ion behavior were explored. Common nickel-base superalloy elements eval uated in this study were Co, C, Cr, Mo, W, Al, Ti, Ru, Re, and Ta. Elemental effects were determined using 3 levels of compositional variations (High, Med, and Low). In addition, the total atomic concentration (at%) of forming elements (sum of Ti, Ta, and Al) was varied between 13.25 and 16.75 at% using 4 alloys. A total of 23 alloys were selected for evaluation in Phase I. The composit ions of the baseline Model alloy and the Phase I alloys are listed in Table 3-3. The compositional ranges considered and the Ni-base superalloy properties affected by the elemental additions (also see section 1.2.2) are listed below: Elemental Variations: o C Variations: C variations have been shown to affect castability, defect formation, elemental segregat ion, and microstructural stability [16,35,50]. C additions ranging from 10 to 750 ppm wt% C, in Model 3 (0.01 wt% C), Model 2 (0.05 wt% C), and Model 1 (0 .075 wt% C) were compared to the baseline Model (0 wt% C) alloy. o Ru Variations: Ru additions have been shown to affect microstructural stability, liquidus solid solution strengthening, and elemental segregation [4,13,28,45].

PAGE 52

34 Ru additions in Model 5 (1.64 wt% (1 at%) Ru),and Model 4 (2.46 wt% (1.5 at%) Ru) were compared to the ba seline Model (0 wt% Ru) alloy. o Cr Variations: Cr variations have been s hown to affect hot corrosion and oxidation resistance, elemental segregation, microstructural stability, and the solvus temperature [24,35,44]. Cr content ranging from 6.75 wt% to 11.82 wt% Cr, in Model 8 (6.75 wt% Cr), Model 7 (8.44 wt% Cr), and Model 6 (11.82 wt% Cr) were compared to the baseline Model al loy (10.12 wt% Cr). o Ti Variations: Ti additions have been shown to affect elemental segregation, precipitati on hardening, liquidus, solvus, oxidation resistance, and hot corrosion resistance [24,35,47]. Ti additions ranging from 0.2 to 0.58 wt% Ti (0.25 to 0.75 at% Ti) in Model 11, Model 10, and Model 9 alloys were compared to the baseline Model alloy (0 wt% Ti). An intermediate Ti le vel of 0.39 wt% (0.5 at %) Ti was also considered. To maintain a constant volume fraction, -former content was kept constant (i.e., 15 at% tota l of Ti, Al, and Ta in baseline Model alloy). Al reductions in Model 9 and Model 11 and a Ta reduction in Model 10 were used to balance Ti additions. o Al Variations: Al (and Ta) variations have been shown to affect elemental segregation, precipit ation hardening, solid solution strengthening, oxid ation resistance, hot corrosion resistance, and castability [12,18,24,35,45]. Al contents ranging from 5.04 to 6 wt% Al (11.5 to 13.7 at% Al) in Model 14, 13, and 12 alloys were compared to the baseline Model alloy (5.25 wt% (12 at%) Al). An intermediate Al leve l of 5.7 wt% (13 at%) Al was also considered. To maintain a constant volume fraction, -former content was kept constant (i.e., 15 at% tota l of Ti, Al, and Ta in baseline Model alloy). Ta substitutions or reductions were used to balance Al variations. o Re Variations: Re variations have been shown to affect solid solution strengthening, castability, defect fo rmation, elemental segregation, solvus, liquidus, and microstruc tural stability [44,51]. Re reductions of 1.5 wt% and 3.02 wt% Re (0.5 to 1 at% Re) were made in Model 15 (7.46 wt% W) and Model 16 (7.34 wt% Ta, 7.46 wt% W, 5.45 wt% Al) alloys, respectively, and were compared to the baseline Model alloy (3.02 wt% (1 at%) Re). The compositions of Model 16 and Model 15 were adjusted for a Ta/(W+Re) ratio of one to maintain a low incidence of casting defects, while keeping a constant -former content. o Co Variations: Co variations have been shown to affect microstructural stability, solvus, elemental segregation, and solid solution strengtheni ng [11,44,45].

PAGE 53

35 Table 3-3. Baseline Model alloy composit ion and Phase I variant compositions C VariationsNiCrCoWTaReAlHfCRuMoTi 1 wt %Bal10.111.56.08.83.025.250.10.08High C2 wt %Bal10.111.56.08.83.025.250.10.05Med C3 wt %Bal10.111.56.08.83.025.250.10.01Low CRu Variations NiCrCoWTaReAlHfCRuMoTi 4 wt %Bal10.111.56.08.83.025.250.12.46Med Ru5 wt %Bal10.111.56.08.83.025.250.11.64Low RuCr Variations NiCrCoWTaReAlHfCRuMoTi 6 wt %Bal11.811.56.08.83.025.250.1High Cr7 wt %Bal8.411.56.08.83.025.250.1Med Cr8 wt %Bal6.811.56.08.83.025.250.1Low CrTi (with Al of Ta) Variations NiCrCoWTaReAlHfCRuMoTi 9 wt %Bal10.111.56.08.83.024.930.10.58HighTi (Low Al)10 wt %Bal10.111.56.07.33.025.250.10.39Med Ti (Low Ta)11 wt %Bal10.111.56.08.83.025.150.10.2Low Ti (Low Al)Al (and Ta) Variations NiCrCoWTaReAlHfCRuMoTi 12 wt %Bal10.111.56.03.83.0260.1High Al (Low Ta)13 wt %Bal10.111.56.05.93.025.70.1Med Al (Low Ta)14 wt %Bal10.111.56.0103.025.040.1Low Al (High Ta) Re (with Ta) Variations NiCrCoWTaReAlHfCRuMoTi 15 wt %Bal10.111.57.58.81.55.250.1Med Re (High Ta)16 wt %Bal10.111.57.57.305.70.1Low Re (Low Ta, High W) (Ta, Al,Ti) Former Variatons NiCrCoWTaReAlHfCRuMoTi 17 wt %Bal10.111.56.09.63.025.580.10.58 formers = 16.75 at%at %Bal12.012.02.03.3112.80.10.75 *High Ta, Al18 wt %Bal10.111.56.06.83.0260.10.39 formers = 16.5 at%at %Bal12.012.02.02.3113.70.10.5 *High Al, Low Ta19 wt %Bal10.111.56.08.83.024.820.10.2 formers = 14.25 at%at %Bal12.012.02.031110.10.25 lower Al20 wt %Bal10.111.56.05.93.024.820.10.2 formers = 13.25 at%at %Bal12.012.02.021110.10.25 Lower Ta, AlCo Variations NiCrCoWTaReAlHfCRuMoTi 21 wt %Bal10.112.06.08.83.025.250.1High Co22 wt %Bal10.19.06.08.83.025.250.1Low CoW Variations NiCrCoWTaReAlHfCRuMoTi 23 wt %Bal10.111.538.83.025.250.11.56Low W (High Mo)

PAGE 54

36 Co content ranging from 9 wt% (9.41 at%) to 12 wt% (12.54 at%) Co in Model 22 and Model 21 alloys, respectively, were compared to the baseline Model alloy (11.47 wt% (12 at%) Co). o W Variatons: W (and Mo) variations were have been shown to affect solid solution strengthening, inci pient melting temperature, microstructural stability, hot corr osion resistance, castability, and elemental segregat ion [12,23,24]. A 1 at% W reduction with a 1 at% Mo substitution in the Model 23 (1 at% W, 1 at% Mo) alloy was compared to the baseline Model alloy (2 at% W, 0 at% Mo). o Former Variations: In order to evaluate volume fraction variation effects, the total amount of the -former content was evaluated. Variations in precipitat ion hardener content (Ti, Al, and Ta) has been shown to affect precip itation hardening, ductility, antiphase boundary energy ( APB), elemental segregation, and microstructural stability [7,34]. -former variations from 13.25 to 16.75 at%, in Model 20 ( 2 at% Ta, 11 at% Al, 0.25 at% Ti); Model 19 ( 3 at% Ta, 11 at% Al, 0.25 at% Ti); Model 18 (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti); and Model 17 (3.25 at% Ta, 12.75 at% Al, 0.75 at% Ti) were compared to the baseline Model alloy (3 at% Ta, 12 at% Al). Material properties were calculated for all 23 compositional variants using the same techniques utilized for the baselin e Model alloy. For clarity, the effect of composition variations on each property were plotted in conjunction with the baseline Model property values. Transfo rmation temperature changes of 3 C or smaller and equilibrium phase amounts changes of 0.5 wt% or smaller, within the variant concentration ranges used in this study, were considered limited or negligible in effect. Pa rtitioning coefficient trends, in specific, were grouped by elements previously reported as partitioning to the dendrite core (Cr, Co, W, and Re) and those previously reported as parti tioning to the interdendritic region (Ni, Ta, and Al). Partitioning coefficient (kcalc) trends evaluated compositional variation effects on a given elements di rection of segregation. Partitioning

PAGE 55

37 coefficient (kcalc) changes of 3% or smaller, wit hin the variant concentration ranges used in this study, were consider ed limited or negligible in effect. Phase II Alloy Development Co mputational Alloy Refinement The goal of Phase II was to further refine understanding on variation effects in alloy chemistry for potential SC IGT applications. Baseline Model A alloy Compositional adjustments to the baseline Model alloy after Phase I included: A 1 at% Cr addition : The Cr addition was used to improve hot corrosion resistance [7,45]. Even though an increase in the amount of phase was calculated with increasing Cr content at 900 C, a decrease in Re and W partitioning was predicted dur ing solidification. A 500 ppm C addition: The carbon addition was used to balance the Cr addition, and was calculated to lower t he amount of TCP phase predicted at 900 C. C additions are also expect ed to increase low angel boundary (LAB) tolerance while reducing casting defects [11]. A 1 at% Ti addition : The Ti addition was calc ulated to decrease Re partitioning during solidification. A 1 at% Al reduction: The Al reduction was used to improve hot corrosion resistance and calculated to decrease Re partitioning during solidification [35]. The Al reduction also helped maintain a former content of 15 at%, balancing the 1 at% Ti addition. A 0.5 at% Co reduction: The Co reduction was used to increase in heat treatment window, predicted to increase with Co reductions. A maximum 15 at% in -former content was maintained in the alloy composition to prevent an increase in the amount of TCP phases predicted to result from increasing -former content at 900 C in Phase I. The new baseline

PAGE 56

38 composition identified was designat ed Model A. The modified baseline composition is shown below in Table 3-4. Table 3-4. Baseline Model A allo y composition in wt% and at%. Composition Al/Ti at% Ni Al Co Cr Hf Re Ta W Ti C wt % 54.47 4.82 11.00 11 0.10 3.02 8.80 5.96 0.78 0.05 6.2 at % 57.21 11.00 11.50 13 0.03 1.00 3.00 2.00 1.00 0.26 15 Modeled elemental variations Material properties for alloy composit ions considered in Phase II were calculated with the same techniques used in Phase I, within the temperature range of 600 C and 1500 C. Following the calculation of the baseline Model A al loy properties, elemental variation evaluations were conducted for Ni-base alloying additions previously shown to influence hot corrosion resistanc e (Al/Ti ratio), and microstructural stability (Re, Cr, and -former content) [35,40,45]. A total of 16 alloys were selected for evaluation in Phase II and were grouped with respect to their characteristic elemental or elemental gr oup variations. The compositions of the baseline Model A alloy and the Phase II alloys are listed in Table 3-5. The compositional ranges invest igated and the Ni-base supera lloy properties affected by the elemental and elemental group additi ons (also see section 1.2.2) are listed below: Elemental/Ratio Variations: o Re Variations: See Phase I elemental va riations for Re effects Three separate variant groups were used to investigate Re effects on material properties: Group 1 : Re levels of 0 and 1.5 wt% Re (0 and 0.5 at% Re) in Model B and Model C were used as a comparison to the 3.02 wt% Re (1 at% Re)

PAGE 57

39 content of the baseline Model A a lloy. All three alloys contained a constant Al/Ti Ratio of 6.2 and a total of 15 at% in -formers. Group 2 : Four Re levels of 3.02, 2.28, 1.5, and 0 wt% Re (1, 0.75, 0.5, and 0 at% Re), were incorporated into the Model D, E, F, and G alloys, respectively. All four alloys contained a constant Al/Ti ratio of 2.54 and a -former content of 14 at%. Group 3 : Four Re levels of 3. 02, 2.28, 1.5, and 0 wt % Re (1, 0.75, 0.5, 0 and at% Re) were incorporated into the Model H, I, J, and K alloys, respectively. All four alloys contained a constant Al/Ti ratio of 1.69 and a -former content of 14 at%. o Al/Ti Ratio Variations: See Phase I elemental variations for Ti and Al effects. The Al/Ti ratio us ed in a Ni-based superalloy has been shown to have an inverse relations hip on oxidation and hot corrosion resistance. A decreasing Al/Ti ratio increase hot corrosion resistance and an increasing Al/Ti ratios increase oxidation resistance [35]. Two separate variant groups were used to investigate Al/Ti effects on material properties: Group 1 : Al/Ti ratios of 3.10 and 1.88 (w t%/wt%) in alloys Model O and P, were compared. Both alloys cont ain a constant Re content of 3.02 wt% Re (1 at% Re) and a former content of 16 at%. Group 2 : Al/Ti ratio variations from of 4.11 to1.88 (wt%/wt%) in Model L (4.11), Model M (3.08), and Model N (1 .88) were compared. All three alloys contained a constant Re content of 2.28 wt% Re (0.75 at%) and a -former content of 15.5 -16 at%. o -Former Variations: See Phase I elemental variations for former effects Two separate variant groups were used to investigate -former effects on material properties: Group 1 : -former variations from 14 to 16 at%, in Model D (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model H (2 at % Ta, 9 at% Al, 3 at% Ti); Model O (3 at% Ta, 11 at% Al, 2 at% Ti); and Model P (3 at% Ta, 10 at% Al, 3 at% Ti) were compared to the baselin e Model A alloy (3 at% Ta, 11 at% Al). All five alloys contained a cons tant Re content of 3.02 wt% Re (1 at% Re).

PAGE 58

40 Table 3-5. Model A and Phase II design alloys chemical compositions in wt% and at%. CompositionAl/Ti at% Re Variations Group 1 NiAlCoCrHfReTaWTiC Model A wt %54.54.8211110.13.028.85.960.780.056.2015 Model B wt %55.24.8211110.12.288.85.960.780.056.2015 Model C wt %57.54.8211110.108.85.960.780.056.2015 Group 2 NiAlCoCrHfReTaWTiC Model D wt %56.03.9711110.13.027.45.961.560.052.5414 Model E wt %56.73.9711110.12.287.45.961.560.052.5414 Model F wt %57.53.9711110.11.537.45.961.560.052.5414 Model G wt %59.03.9711110.107.45.961.560.052.5414 Group 3 NiAlCoCrHfReTaWTiC Model H wt %56.4411110.13.026.05.962.390.051.6914 Model I wt %57.2411110.12.286.05.962.390.051.6914 Model J wt %57.9411110.11.536.05.962.390.051.6914 Model K wt %59.5411110.106.05.962.390.051.6914 Al/Ti Variations Group 1 NiAlCoCrHfReTaWTiC Model O wt %53.74.811110.13.028.85.961.60.053.1016 Model P wt %53.44.411110.13.028.85.962.30.051.8816 Group 2 NiAlCoC r H f ReTaWTiC Model L wt %54.84.811110.12.288.85.961.20.054.1115.5 Model M wt %54.44.811110.12.288.85.961.60.053.0816 Model N wt %54.14.411110.12.288.85.962.30.051.8816 Cr Variations Group 1 NiAlCoC r H f ReTaWTiC Model A wt %54.54.811110.13.028.85.960.80.056.2015 Model Q wt %55.74.511120.13.026.05.961.60.052.8214 Y'Former Variations Group 1 NiAlCoCrHfReTaWTiC Model D wt %56.04.011110.13.027.45.961.560.052.5414 at %58.29.512130.0312.5220.26 Model H wt %56.44.011110.13.026.05.962.390.051.6914 at %58.2912130.0312230.26 Model A wt %54.54.811110.13.028.85.960.780.056.2015 at %57.21112130.0313210.26 Model O wt %53.74.811110.13.028.85.961.560.053.1016 at %56.21112130.0313220.26 Model P wt %53.44.411110.13.028.85.962.330.051.8816 at %56.21012130.0313230.26 Group 2 NiAlCoC r H f ReTaWTiC Model E wt %56.74.011110.12.287.45.961.560.052.5414 at %58.59.512130.030.752.5220.26 Model I wt %57.14.011110.12.316.05.962.390.051.6914 at %58.5912130.030.752230.26 Model B wt %55.24.811110.12.288.85.960.780.056.2015 at %57.51112130.030.753210.26 Model M wt %54.44.811110.12.288.85.961.570.053.0816 at %56.51112130.030.753220.26 Model N wt %54.14.411110.12.288.85.962.340.051.8816 at %56.51012130.030.753230.26

PAGE 59

41 Group 2 : -former variations from 14 to 16 at%, in Model E (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); M odel I (2 at% Ta, 9 at% Al 3 at% Ti); Model B (3 at% Ta, 11 at% Al, 1 at% Ti); Model M (3 at% Ta, 11 at% Al, 2 at% Ti); and Model N (3 at% Ta, 10 at% Al, 3 at % Ti) were compared. All five alloys contain a constant Re content of 2.28 wt% Re (0.75 at% Re). o Cr Variations: See Phase I elemental variations for Cr effects One variant group was used to invest igate Cr effects on material properties: Group 1 : Cr variations from 11 to 12 wt% Cr (13 to 14 at% Cr), in Model Q (14 at% Cr 10 at% Al, 2 at% Ti ) and the baseline Model A alloy (13 at% Cr, 11 at% Al, 1 at% Ti) were compared. Material properties for all Phase II alloys were calculated using the same techniques used in Phase I of alloy development. The calculated material property values were grouped and plott ed as a function of composition to evaluate elemental property trends. Phase III Alloy Development Experimental Validation Elemental variation trends on microstr uctural stability, phase transformation temperatures, and segregation behavior from Phase II were used to determine compositional adjustments to the baseline Model A alloy (Table 3-6). Modifications to the baseline Model A composition produced five alloy compositions for Phase III listed in Table 3-7. Table 3-6. Phase III compositional va riants with respect to the baseline Model A alloy wt % at % wt % at % wt % at % wt % at % wt % at % wt % at % C r 11.013.0 1.01.0 Ta 8.83.0 -2.8-1.0-2.8-1.0-1.4-0.5-1.4-0.5 Re 3.01.0-3.0-1.0-3.0-1.0-3.0-1.0-3.02-1.00 Al 4 8 11 0 0 5 1 0 0 8 2 0 0 5 1 0 0 7 0 5 0 7 0 5 Ti 0.7810.390.51.6120.7810.7810.781 former (at%) 15.0-0.5-1.0-1.0-1.0-1.0 Al/Ti (wt%/wt%) High BaselineModifications: Additions or Reductions CompositionAl/TiCrRe Low Model A A lloy 1 A lloy 2 A lloy 3Alloy 4 A lloy 5 HighLowHigh -3.66-3.66 6.2-2.48-4.51-3.31

PAGE 60

42 Table 3-7. Phase III alloy compos itions and variation groups in wt% NiAlCoCrHfReTaWTiCAl/Ti Y' at% Alloy 1 wt %57.54.3711110.108.85.961.170.053.7214.5 Alloy 2 wt %59.54.0411110.1065.962.390.051.6914 NiAlCoCrHfReTaWTiCAl/Ti Y' at% Alloy 3 wt %58.84.3711120.106.15.961.560.052.8914 NiAlCoCrHfReTaWTiCAl/Ti Y' at% Alloy 4 wt %56.04.1711110.13.027.45.961.560.052.5414 Alloy 5 wt %59.04.1711110.107.45.961.560.052.5414Re VariationsAl/Ti Variations Cr Variation The general modifications made to the baseline Model A alloy to produce the five final compositions are describe more fully below: General Elemental/Elemental Group Modifications: o -former reductions of 1 to 1.5 at%: achieved through Al and Ta reductions, were predicted to decr ease the amount of TCP phases present at 600 C. o Al/Ti ratio reductions of 2.5 to 4.5: were used to improve hot corrosion resistance [35]. Decreasing the Al/Ti ratio was achieved through Al reductions and Ti additions. Al/Ti reductions were also shown to reduce the marked partitioning of Re and Ta during solidification. o Re reductions of 3.02 wt% (1 at%): in all final alloy compositions, with the exception of Alloy 4, were predicted to decrease the amount of TCP phases at 600 C. Re redu ctions were also predicted to decrease elemental segregation, in part by avoiding Res strong partitioning tendency towards the dendrite core. o Cr additions of 1 wt% (1 at%): in Alloy 3 was used to improve hot corrosion resistance [7,45]. Even thou gh an increase in the amount of phase was predicted for at 600 C, a decrease in Re and W partitioning was also predicted. Material properties for alloy composit ions considered in Phase III were calculated with the same techniques used in Phase I, within the temperature range of 600 C and 1500 C. Final alloy compositions incorporated characteristic variations for alloying elements previously shown to influence hot corrosion and material stability.

PAGE 61

43 Variations in Al/Ti ratio (with a Ta variat ion), Cr (with Al and Ta variations), and Re content are seen in Table 3-7 and are listed below [35,40,45]. Elemental/Elemental Group Variations: o Al/Ti ratio variations: ranging from 3.72 to 1.69 (wt%/wt%) in Alloy 1 (8.8 wt% Ta) and Alloy 2 (6.02 wt% Ta), respectively, were compared. o Cr variations: ranging from 12 to 11 wt% Cr were investigated by comparing Alloy 3 (4.37 wt% Al, 6.05 wt% Ta) and Alloy 5 (4.17 wt% Al, 7.37 wt% Ta), respectively. o Re variations: ranging from 3.02 wt% Re to 0 wt% Re in Alloy 4 and Alloy 5 respectively, were compared. Calculated material property values were grouped and plotted as a function of composition to evaluate elemental property trends. In order to validate some of the predicted propertie s, small button samples the Phase III alloy compositions were prepared. The microstructure, phase transformation temperatures, and elemental segregation behavior of each of the samples were characterized and co mpared to the predicted values. Materials Although the five final compositions in Phase III were designed for single crystal IGT application, sm all polycrystalline specimens were used to validate material properties in this study. Sinc e this investigation focuses on elemental variation effects on thermodynamic proper ties, the polycrysta lline nature of the samples should have no effect on the properties of interest. High purity elements (> 99.5%), in t he forms of granules, wire, and powder were combined and compacted in the appr opriate levels to produce the five Phase III compositions. The alloy button specimens were arc melted in a Centorr Series T Bell Jar 5BJ-2698 Arc Fur nace. The arc furnace consisted of a

PAGE 62

44 water-cooled stainless steel vacuum bell jar with a water-cooled copper hearth. A mechanical vacuum pump was used to evacuate the chamber prior to backfilling with inert gas. T he compacted buttons were arc melted in a 10-1 Pa inert argon environment to form 100 g alloy button specimens. The arc was established between the sample and a tungs ten electrode and, prior to melting the Phase III sample alloys, a Ti ge tter button was melted to remove any O2 or N2 impurities from the cham ber. To ensure chemical homogeneity, each sample was melted, turned over and then remelted 9 times. The arc melted buttons produced were approximately 4 cm in di ameter and 1cm in thickness. The button samples were sectioned usi ng an abrasive cut-off wheel into approximately 2 cm X 1 cm X 1 cm specimens and cleaned in an ultrasonic Methanol alcohol bath. Solution Heat Treatment In order to reduce the s egregation in the button samples, a solution heat treatment was given to three samples from each alloy. The solution heat treatment was based on the phase transformation temper atures calculated for the compositions using the JMatPro t hermodynamic equilibrium module. A maximum solution heat treatment tem perature of 1250 C was used. This maximum heat treatment temperature wa s designated to be 50 to 60 C below the calculated solidus temperatures for all Phase III alloys to reduce the risk of incipient melting in the segregated as-solidified microstructure. The solution heat treatment trial wa s conducted in an Elatec Technology Corporation Lab Vac 2 vacuum furnace ope rating at a maximum pressure of 1 x 10 -6 Torr. The vacuum furnace has a graphite hot zone measuring 15.2 cm x

PAGE 63

45 15.2 cm x 38 cm with graphite heating elements and a gr aphite hearth plate. Samples of each button were placed in high purity Al2O3 rectangular trays to prevent interaction of the Ni-base alloys and the graphite hearth plate during heat treatment. Three type C OMEGA thermocouples, W-5%Re vs. W-26% Re, were used to monitor the sample and furnace temperatures. Sample thermocouples were maintained within + 3 C and the over temperature thermocouple stayed between + 0 to +15 C, throughout the exper iment. The solution heat treatment, which was based on heat treatments for similar alloys (Table 3-8), lasted 41 hours and included a 1250 C hold for 32 hours to homogenize the segregated as-cast structure. Table 3-8. Heat treatment used for the IGT expe rimental alloys Step Time (hr) Rate ( C/hr) Temp. ( C) 1 0.17 10 23 150 2 1.50 10 150 1050 3 1.00 1050 4 0.08 10 1050 1100 5 1.00 1100 6 0.17 10 1100 1200 7 2.00 1200 8 0.17 3 1200 1225 9 2.00 1225 10 0.42 1 1225 1250 11 32.00 1250 12 0.50 Gas Furnace Cool At the completion of the heat treatmen t, the vacuum chamber was filled with helium gas at 103 KPa for an increased cooli ng rate. The circulation of the He gas throughout the chamber by a fan, along with a Cu H20-cooled radiator, provided an initial cooling rate of 149 C/min. Once t he temperature dropped below 100 C, the samples were removed fr om the furnace. Following the heat treatment, samples in the as-cast and heat treated condition were used for characterization.

PAGE 64

46 Differential Thermal Analysis DTA testing for as-cast and heat treated samples of all Phase III alloy specimens was conducted at Dirats Labor atories on a 2910 DSC V4.4E unit. A high purity He environment was used in all te sting. Prior to testing the Phase III alloy samples, the instrument was calibrated using a high purity 200 mg Ni standard in an Al2O3 lined platinum cup scanned at a rate of 20 C/min. DTA samples used for all five Phase III alloys were approximately 2 cm X 1 cm X 1 cm in size, with masses ranging from 17 to 25 g. In order to insure that as-cast specimens contained representative regions of all stages of solidification, the scale of the solidification was compared to the sample size. Primary dendrite arm spacings (PDAS) of 29 m, 26 m, 20 m, 16 m, and 26 m were measured for Alloys 1,2,3,4, and 5, respectively. Ther efore, the sample sizes were sufficiently large to ensure that bot h the dendrite core and the interdendritic regions were tested. T he temperature range analyzed in the DTA test was from about 1000C to 1550C. To avoid undercooling effects, reacti on temperatures were taken solely from heating curves. Heating curves we re also used to avoid any chemical changes due to specimen interactions with the Al2O3 in crucibles and the effects associated with oxidation that could occur during re-solidification. The recorded DTA results mentioned are plotted to illustrate the temperature difference ( T) between the experimental specimen temperature and the reference sample temperature, as well as the rate of change in the temperature difference (derivate). The differential and derivative curves

PAGE 65

47 produced can then be used to determi ne exothermic or endothermic phase changes in a given alloy. Figure 3-1. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 2 from Phase III compositions in the heat treated condition DTA results were used in this study to identify the liquidus, solidus, and solvus of the as-cast and heat treated samp les, if present in the material. Phase transformation or reaction temperatures we re identified as distinct inflection points in the temperature difference ( T) vs. specimen temper ature curves. The intersection points in the T vs T plots in this study, as well as the inflection point in the derivative vs T plots (or maximu m thermal effect) were considered the temperature at which the reac tion occurred. An example of such a plot is shown in Figure 3-1, where major inflection point s are labeled. For ex ample, the solidus is identified as the inflection at the right of the main endotherm in the temperature difference curve, which is 1343 C for Alloy 2 shown above. The liquidus and

PAGE 66

48 solvus temperatures inflections in the derivative curve where adjusted with respect to the Ni standard used for calibration. Microscopy The microstructures of the Phase III alloys, in the as-cast and heat treated conditions, were examined using optic al metallography and scanning electron microscopy (SEM) techniques. Specimens were prepared by standar d metallographic procedures. Samples approximately 2 cm X 1 cm X 1 cm in size were mounted in bakelite, exposing the button alloy cross section (cut surface) as seen below in Figure 3-2. Button Top View Button Side View Mounting Orientation Figure 3-2 Button alloy sectioning and m ounting orientation in metallographic analysis Samples were rough ground wet with 180, 240, 320, and 600 grit silicon carbide papers, and were then polished us ing 15 m, 5 m, 1 m, and 0.3 m alumina particle suspensions. The sample s were given a final polish using 0.04 m colloidal silica. Etched samples were used for optical and SEM investigation and un-etched samples were used for quantit ative and qualitative compositional analysis. Material specimens were etc hed for microscopic examination using the Pratt and Whitney Etch # 17 (100 ml H2O + 100 ml HCl + 100 ml HNO3 + 3 g MoO3), which dissolves the precipitates. Optical metallographic exam ination of the Phase III alloy microstructures was performed on a LECO NEOPHOT 21 Meta llograph at magnifications ranging

PAGE 67

49 from 50X to 100X. Solution heat treat ed samples were analyzed to determine the degree of homogenization achieved during the heat tr eatment. The elimination of the as-solid ified dendritic structure in the heat treated samples indicated that the chemic al segregation had been signif icantly reduced during the solution heat treatment. On the other hand, the presence of dendritic structure in the heat treated samples, indicated that incomplete homogenization had occurred during the solution heat treatment, resulting in some degree of residual segregation. A JSM 6400 analytical scanning electron microscope (SEM) was used to further characterize the microstructure of the Phase III alloys. The SEM was operated at an accelerating voltage of 15 KV in both the secondary electron (SE) and backscattered (BSE) imaging modes. The secondary electron mode was used to determine the as-cast and heat treated microstructure in etched samples. The backscattered imaging mode was used to provide preliminary estimates on residual segregation and discrete phase com positions in un-etched samples. Qualitative chemical analysis was also preformed using a 6506 Oxford Detector energy dispersive spectrometer (EDS) on samples in the unetched condition. Segregation Quantitative analysis of elemental segregation in the as-cast and aspolished Phase III samples was conduc ted with the use of the JEOL Superprobe 733 electron pr obe micro-analyzer (EMPA)/wavelength dispersive spectrometer (WDS). A beam size of 0.5-1.0 m, a beam current of 20 nA, a beam voltage of 15 KV, and a take-off angle of 40 were used for characterization of all Phase III alloys.

PAGE 68

50 Specific calibration standards were used as references for Ni, Cr, Co, W, Re, Ta, Al, and Ti. Wavelength dispersi ve spectroscopy (WDS) was used with TAP crystals to detect W, Re, and Ta using M lines. The TAP crystal using K lines was needed for Al. A LiF crystal was used to detect Ni, Cr, and Co examining K lines. A PET crystal using L lines was needed for Ti. Each element was counted for 10 sec per point. Compositions were measured using 17 to 30 point line scans across a dendritic area, with a spacing of about 1 m between measurements. Line scans began and ended in the interdendritic regions of a sample and intersected the center of a primary dendrit e arm. The resulting elemental readings, normalized to 100 wt%, were plotted versus 1m poi nt measures across the line scan. The variations in composition, from minimu m to maximum concentrations, provided an estimate of elemental segregation (Figure 3-3). Some elements were observed to segregate to the dendrite co re and some elements segregated to the interdendritic regions. The degree of segregation was determi ned by calculating the partitioning coefficient (k). The measur ed compositions of a given element (x) in wt% at the dendrite core (Cx,core) and at the interdendritic region (Cx,inter) are used to calculate the partitioning coefficient (kx) as seen below. kx = Cx,core/ Cx,inter A partitioning coefficient (kx) value of one indicates that a given element exhibits no preference for segregati on to the dendrite core or to the interdendritic region during solidification. A solute with a kx less than unity partitions to the interdendritic region. In contrast, solutes with a partitioning

PAGE 69

51 coefficient greater than unity segregate to the dendrite core. As an example, a k value of 1.25 demonstrates t hat a given elements conc entration in the dendritic core is 125% of its concentration in the interdendritic region. Alloy 2 As-Cast Microprobe Segragation0 2 4 6 8 10 12 12345678910111213141516171819202122232425262728Testing Point (um)Normalized Wt % Co Cr Al Ti Ta W Figure 3-3. EMPA/WDS compositions in normalized wt% versus line scan measurement points (m) for experimental Alloy2 The partitioning coefficients used to evaluate modeled and experimental data in this study, represented segregat ion from a phase diagram. A graphic representation of this method for a hypot hetical A-B binary system is seen in Figure 3-4. In Figure 3-4 the composition (Cs) is the composition of the fist solid to form; the liquid composition at th is temperature is represented as (Cl). In order to calculat e the experimental partitioning coefficient (kx,exp), the laboratory tested composition of a given el ement (x) in wt% at the dendrite core (Cx,core) was used for the first solid formed and the nominal composition of the element in the alloy (Cx,reg) is used as the liquid composition. This relationship can be expressed in the formula below. kx,exp = Cx,core/ Cx,reg Interdentritic Dendrite Core

PAGE 70

52 Figure 3-4 Schematic repr esentation of solidification occurring in a eutectic binary phase diagram. These experimentally determined parti tioning coefficients were then compared to the solidification compositi ons simulated using the thermodynamic equilibrium module in JMatPro. JMatPro simulations pr ovide compositional data for the first solid to form at the liqui dus transformation temperature. Although Figure 3-4 depicts the solid composition for a binary system, JMatPro calculates this composition for a mu lticomponent alloy. The partitioning coefficients (kx,calc) for modeled alloys were calculated, as follows, by using the predicted elemental compositions of the first solid to form (Cx,solid), and the nominal composition of the elements in the alloy (Cx,reg) as the liquid composition. kx,calc = Cx,solid/ Cx,reg Using the partitioning coefficients defined as kx,exp and kx,calc, segregation trends were plotted as a function of allo y composition. Partitioning coefficient changes of 3% or smaller, within the va riant concentration ranges used in this study, were considered limited or neglig ible in effect. Fo r clarity, elements previously reported to parti tion to the dendrite core (Cr, Co, W, and Re) or the T % Solute Cs Cl A B

PAGE 71

53 interdendritic region (Ni, Ta, and Al) were grouped together to facilitate the evaluation of elemental va riation effects [4,45].

PAGE 72

54 CHAPTER 4 RESULTS Phase I Modeled Elemental Variations The calculated material property result s for the baseline Model alloy and the Phase I elemental variations are given below. For clarity, the baseline Model alloys material properties; specifically, the microstructural stability, transformation temperatures, and segregat ion behavior properties are discussed first. These calculated material proper ties were then used as the baseline for compositional comparisons. The calculat ed material properties for the other 23 experimental alloys were presented in group s with respect to their characteristic elemental variation (C, Ru, Cr, Ti, Al Re, Co, W, and the total amount of formers content in the alloy chemistry). Baseline Model Alloy The baseline Model composition is seen below in Table 4-1. Table 4-1. Baseline Model allo y composition in wt% and at%. Model Alloy NiCrCoReWAlTaHf wt %Bal10.1211.473.025.965.258.80.1 at %Bal1212121230.05 Microstructural stability The phase fraction diagram calculated for the baseline Model alloy under equilibrium conditions is shown in Figur e 4-1. The calculated phase fraction diagram gives predicted equilib rium phases and their weig ht fractions within the temperature range of 900 C and 1500 C. These thermodynamic calculations

PAGE 73

55 revealed that the equilibrium phases at temperatures under 1000 C include , and Figure 4-1. Predicted phase fraction diagra m for baseline Model alloy calculated by the JMatPro thermody namic equilibrium module According to the calculated phase diagram, the alloy at 900 C is made up primarily of the precipitate (approxim ately 58 wt%) and the matrix (approximately 35 wt%). A limited amount of the TCP phase (approximately 6 wt%) is also predicted to be in equilibrium at 900 C. Phase transformation temperatures According to JMatPro equilibrium thermodynamic calculations, the solidification path suggested for the baseline Model alloy is seen below. L L + + + + The predicted liquidus, solidus, solvus, and solvus temperatures for the baseline Model alloy were 1364 C, 1297 C, 1282 C, and 1196 C, respectively (Figure 4-2).

PAGE 74

56 Figure 4-2. Predicted phase diagram for the baseline Model al loy calculated by the JMatPro thermodynamic equilibriu m module with identification of critical phase transfo rmation temperatures For the baseline Model alloy com position, the melting range was determined to be 67C. The calculated heat treatment windo w for the baseline Model alloy is 15 C. Elemental segregation The thermodynamic equilibrium module in JMatPro was used to calculate elemental segregation for spec ific elements in the base line Model alloy (Table 42). Table 4-2. Predicted partitioning coeffici ent values (kx,calc) for the Phase I baseline Model alloy Ta Al Cr Ni W Co Re K calculated 0.40 0.91 1.01 1.04 1.12 1.13 1.45 The segregation behavior calculations of the baseline Model alloy resulted in partitioning coefficients, k calc, greater than one (core t endencies) for Cr, Ni, W, Co, and Re. The core segr egation predicted for Re was the most significant, solvus solidus liquidus solvus

PAGE 75

57 followed by Co and W. Ni and Cr were pr edicted to show only slight segregation tendencies towards the core, with k calc values close to unity. Elements calculated to segregate to the interdendritic region, with k calc values less than one, were Ta and Al. Ta exhibite d the strongest segregation with a kcalc value of 0.4, followed by Al with a calculated coefficient of 0.91. Elemental Variation Effects The effects of elemental variations on the calculated microstructural stability, transformation temperature, and segregation behavior properties were discussed in the sections below. Chromium variation effects Cr contents ranging from 6.75 wt% to 11.82 wt% Cr, in Model 8 (6.75 wt% Cr), Model 7 (8.44 wt% C r), and Model 6 (11.82 wt% C r) were compared to the baseline Model alloy (10.12 wt% Cr). Microstructural Stability The thermodynamic calculations reveal ed that, for alloys with a Cr content larger than 8.44 wt% (10 at %) Cr, the equilibrium phases at temperatures below 1000 C included , and The alloy with the lowest Cr content (6.75 wt% (10 at%) Cr), was predicted to contain equilibrium phases , and at temperatures below 1000C. Figure 4-3 shows that the amount of phase predicted at 900 C, is strongly influenced by increasing Cr content. The amount of phase predicted, increases linearly with increas ing Cr content. Within the Cr range evaluated in this study, at 900 C, a 1.3 wt % increase in the amount of phase was predicted with every 1 wt% Cr increase.

PAGE 76

58 Stability Effect at 900 Cy = 1.3x 6.7 0 1 2 3 4 5 6 7 8 9 10 6789101112 Cr Variations (wt%)wt % Sigma Mu Figure 4-3 Predicted Cr variation effe cts on TCP equilibrium phase amounts The Model 8 alloy with a 6.75 wt% (8 at%) Cr content, was predicted to exhibit small amounts of both the and TCP phases (approximately 2 wt% each) at 900 C. Phase Transformation Temperatures According to JMatPro thermodyna mic equilibrium calculations, the solidification path suggested fo r the alloys with a Cr content larger than 8.44 wt% (10 at%) Cr is seen below. L L + + + + The alloy with the lowest Cr content (6.75 wt% (8 at%) Cr) was predicted to exhibit a solidificati on path, as seen below. L L + + + + + + + Figure 4-4 shows the predicted Cr vari ation effects on the liquidus, solidus, and solvus temperatures. Calculations indicated that Cr additions suppressed all critical transformation temperatures. Increased Cr content resulted in a nearly linear decrease in the solidus and solvus temperatures, both at the rate of Model Alloy

PAGE 77

59 approximately 48 C with a 5 wt% Cr in crease. The liquidus temperature was predicted to decrease linearly with increas ing Cr content, decreasing 30 C with a 5 wt% Cr increase. Phase Transformation Effectsy = -6x + 1426 y = -9x + 1390 y = -9.5x + 13781250 1270 1290 1310 1330 1350 1370 1390 14106789101112Cr Variations (wt%)Temperature (C) Liquidus Solidus Y' solvus Figure 4-4 Predicted Cr variation effect s on phase transformation temperatures With increasing Cr content, the calcul ated decrease in the liquidus was less than the decrease observed for the solidus, resulting in an increase in the melting range (16 C with a 5 wt% Cr addition). The increase in Cr content caused the solidus and solvus temperatures to linearly decrease, at similar rates, which resulted in a nearly constant heat tr eatment window for all compositions evaluated. Increasing the Cr content from 6.73 wt% Cr to 11.82 wt% Cr resulted in a negligible 2 C increase in heat treatment window. Elemental Segregation Predicted segregation behavior for Cr variant s is seen in Figure 4-5. For the Cr variants and the base line Model alloy, k calc values greater than one (core tendencies) were calculated for Ni, W, Co and Re. The core segregation for Re was the most significant, followed by W, Co, and then Ni. Small Cr variation effects were predicted for Ni and Co, wher e only 2% to 2.5% increases in the Model Alloy

PAGE 78

60 kCr,calc and kCo,calc resulted from a 5.61 wt% Cr ad dition. With a 5.61 wt% Cr increase, Re segregation decreased by a linear 10% decrease in kRe,calc and W segregation was predicted to decrease by a 8% linear decrease in the kW,calc. Figure 4-5. Predicted Cr variatio n effects on elemental segregation Elements predicted to segregate to the interdendritic region, with k calc values less than one, were Ta, Al, and Cr. The partitioning of Ta was the strongest, followed by Al and then Cr. Cr, which does not exhibit a strong tendency to partition, was predicted to change from segregating to the interdendritic region to the dendrite core as Cr content increased (a 6% increase in kCr,calc with a 5.61 wt% Cr addition). Incr easing Cr content was predicted to have negligible effects on Al segregation, with observed kAl,calc values of approximately 0.9 for all Cr variations considered. A linear increase in Ta segregation was predicted with an increasing Cr content (a 17% decrease in the kTa,calc with a 5.61 wt% Cr addition). The bar chart (Figure 4-6) below was produced as an additional visual comparison to gauge predicted elemental effe cts on partitioning coefficients. The Partitioning Effect0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1681012 wt% Crk calc Ni Ta Al Partitioning Effect0.9 1 1.1 1.2 1.3 1.4 1.5 1.6681012 wt% Crk calc Cr Co W Re Model Alloy Model Alloy

PAGE 79

61 predicted Cr variation effect was most si gnificant for Ta followed by Re, W, Cr, Co, and then Ni. k calc Comparisons for wt %Cr Variations0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 TaAlCrNiWCoRek calc 6.75 8.44 10.12 Model Alloy 11.82 Figure 4-6. kcalc comparisons between t he baseline Model alloy and Cr variants Aluminum (and Tantalum) variation effects Al contents ranging from 5.04 to 6 wt% Al (11.5 to 13.7 at% Al) in Model 14, 13, and 12 alloys were compared to the bas eline Model alloy (5.25 wt% (12 at%) Al). An intermediate Al level of 5.7 wt % (13 at%) Al was also considered. To maintain a constant -former content Ta substitutions or reductions were used to balance Al variations. Microstructural Stability Within the Al range evaluated in this study (5.04 wt% to 6 wt% Al and 10.3 wt% to 3.82 wt% Ta (1.3 at% to 3.5 at % Ta)), the thermodynamic calculations predicted the equilibrium phases under 1000 C to include the , and phases. Calculations predicted a negligible linear reduction in the amount of phase expected at 900 C (Figure 4-7) with increased Al (and reduced Ta)

PAGE 80

62 content. At 900 C, the amount of phase was predicted to decrease only 0.26 wt% for an Al increase of 0.96 wt% (2.2 at %) Al (with a 2.2 at % Ta reduction). Stability Effect at 900 Cy = -0.3x + 8 6 6.1 6.2 6.3 6.4 4.855.25.45.65.866.2 Al Variations (wt%)wt % Sigma Figure 4-7. Predicted Al (and Ta) va riation effect on TCP equilibrium phase amount with respect to Al (wt%) concentration Phase Transformation Temperatures Thermodynamic equilibrium calculations predicted the solidification path seen below for alloys with Al (and Ta) vari ations ranging from 5.04 wt% to 6 wt% Al (and 10.3 wt% to 3.82 wt% Ta). L L + + + + Figure 4-8 depicts the predicted Al (and Ta) variation effects on the liquidus, solidus, and solvus temperatures. Increasing Al (decreasing Ta) content was predicted to result in linear increases in both the liquidus and solidus temperatures. The liquidus and solidus temperatures were predicted to increase 19 C and 40 C, respectively, with a 1 wt% Al addition (and a 6.75 wt% Ta reduction) The addition of Al (and reduction of Ta) was predicted to decrease the solvus temperature. A linear decrease in Model Alloy

PAGE 81

63 the solvus was calculated at the rate of 45C for a 1 wt% Al addition (with a 6.75 wt% Ta reduction). Phase Transformation Effectsy = 40x + 1085 y = 19x + 1266 y = -46x + 1520 1220 1240 1260 1280 1300 1320 1340 1360 1380 1400 55.25.45.65.86Al Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-8. Predicted Al (and Ta) va riation effects on phase transformation temperatures with respect to Al (wt%) concentration. With increasing Al (and decreasing Ta) content, the predicted rate at which the liquidus decreased was gr eater than the rate at which the solidus decreased, resulting in a decrease of the melting range. Since Al additions (and Ta reductions) were predicted to linearly in crease the solidus and linearly decrease the solvus at similar rates, the heat treatment window was predicted to increase 83 C with a 0.96 wt% Al addi tion (and 6.48 wt% Ta reduction). Elemental Segregation Figure 4-9 shows the predicted Al (and Ta) variation effects on elemental segregation. Elements calculated as parti tioning to the dendrite core were Cr, Ni, W, Co, and Re. Re exhibited the greates t segregation, followed by Co, W, Ni, and then Cr. Re and Co segregation was predicted to increase linearly with Al additions (and Ta reductions). A linear decrease in W and Ni segregation was predicted with increased Al (and reduc ed Ta) content. Al additions (and Ta Model Alloy

PAGE 82

64 reductions) were predicted to linearly shi ft Ni segregation, from the interdendritic region to the dendritic core. The increase in Al (and decrease of Ta) content had a negligible effect on calculated Cr segr egation within the 0.96 wt% Al (and 6.48 wt% Ta) concentration range considered. Figure 4-9. Predicted Al (and Ta) variat ion effects on element al segregation with respect to Al (wt%) concentration. Ta and Al were predicted to segregate to the interdendritic region, with k calc values less than one, for all variants c onsidered. Ta was calculated as the strongest segregating element, followed by Al. Ta segr egation was predicted to increase linearly with Al addi tions (and Ta reductions). The segregation of Al was not predicted to change significantly wit h an increase in Al (and decrease in Ta) content. The bar chart below (Figure 4-10) show s the predicted Al (and Ta) variation effects on partitioning coefficients. Al (and Ta) variation effects were most significant for Re, followed by W, Ni, and then Co and Ta. The negligible Al (and Ta) variation effects predicted fo r Al and Cr are also evident. Partitioning Effect 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 55.56 wt% Alk calc Cr Co W Re Partitioning Effect 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 55.56 wt% Alk calc Ni Ta Al Model Alloy Model Alloy

PAGE 83

65 k calc Comparisons for wt %Al Variations0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 TaAlCrNiWCoRek calc 6 5.7 5.25 Model Alloy 5.04 Figure 4-10. kcalc comparisons between the baseline Model alloy and Al (and Ta) variants. Titanium (and Tantalum, Al uminum) variation effects The effects of Ti additions (with Ta or Al reductions) on material properties were investigated using four alloy composit ions. Ti additions ranging from 0.2 to 0.58 wt% Ti (0.25 to 0.75 at% Ti) in Model 11, Model 10, and Model 9 alloys were compared to the baseli ne Model alloy (0 wt% Ti). An intermediate Ti level of 0.39 wt% (0.5 at %) Ti was also c onsidered. Al reductions in Model 9 and Model 11 and a Ta reduction in Model 10 were used to balance Ti additions to keep a constant -former content. Microstructural Stability Thermodynamic calculations for all Ti variants predicted that equilibrium phases below 1000 C include , and Notably, a 0.58 wt% (0.75 at%) Ti addition was predicted to onl y increase the amount of phase by 0.2 wt% at 900C. The negligible Ti effect on the amount of phase, predicted at 900 C, is seen in Figure 4-11.

PAGE 84

66 Stability Effect at 900 C6.25 6.3 6.35 6.4 6.45 6.5 00.10.20.30.40.50.6 Ti Variations (wt%)wt % Sigma Figure 4-11. Predicted Ti (and Ta or Al ) variation effects on TCP equilibrium phase amount with respect to Ti (wt%) concentration. Phase Transformation Temperatures The solidification path predicted for t he Model 11, 10, and 9 alloys is seen below. L L + + + + Phase Transformation Effectsy = -32x + 12811240 1260 1280 1300 1320 1340 1360 1380 00.10.20.30.40.50.6 Ti Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-12. Predicted Ti (and Ta or Al) variation effects on phase transformation temperatures with res pect to Ti (wt%) concentrations Figure 4-12 depicts the predicted Ti ( with Al or Ta reduction) variation effects on the liquidus, solidus, and solvus temperatures. Model Alloy Model Alloy

PAGE 85

67 Ti additions (with Al or Ta reductions) resulted in negligible effects on phase transformation temperatures with the exception of the solvus. A 0.58 wt% (0.75 at %) Ti addition with a (0.75 at%) Al reduction were predicted to result in negligible reductions in the liquidus and solidus temperatures. A smaller Ti addition of 0.39 wt% (0.5 at%) Ti with a (0.5 at%) Ta reducti on was predicted to increase the solidus 6C. The revers ed phase transformation trends, and the 6 C solidus increase, were attributed to the Al and Ta reductions. Regardless of whether Ti additions had been balanced by Al or Ta reductions, calculations predicted a clear linear decrease in the solvus temperature with Ti additions. For all three Ti additions, the solvus decreased with increasing Ti content, decreasing approximately 6 C within the 0.58 wt% (0.75 at%) Ti range analyzed. The calculated decrease in the solvus temperature, with increasing Ti content, resulted in a considerable increas e in the heat treatment widow. With a 0.58 wt% (0.75 at %) Ti addition and (0.75 at%) Al reduction, the calculated heat treatment window increased from 15 C to 28 C. A 0.39 wt% (0.5 at%) Ti addition with a (0.5 at%) Ta reducti on resulted in an increased heat treatment window of 24 C. Elemental Segregation Segregation behavior trends calculated for alloys with Ti variations (and Al reductions in Model 11 and 9 or Ta reducti ons in Model 10) are seen in Figure 413. For all alloys considered (the bas eline Model, Model 11, Model 10, and Model 9 alloys), the elements Cr, Ni, W, Co and Re were predicted to partition to the dendrite core. Re was the most heav ily segregated elemen t, followed by W,

PAGE 86

68 Co, Ni, and then Cr. Regardle ss of whether Ti additions we re balanced by either Ta or Al reductions, negligible elemental variation effects were predicted for Cr, Co, Ta, Ni, W, and Al. When compared to the baseline Model alloy, Re segregation decreased for all alloys with Ti additions. A small 1% decrease in the kRe,calc was calculated with a 0.39 wt% (0.5 at%) Ti addition (and a 0.5 at% Ta reduction) but an approximately 30% decrease was predicted with a 0.58 wt% (0.75 at %) Ti addition (and a 0.75 at% Al reduction). Elements predicted as segregat ing to the interdendritic region were Ta, Ti, and Al. Ta was calculated to segregate t he strongest, followed by Ti and Al. A 0.39 wt% Ti addition was predicted decrease kTi,calc by 4%. Partitioning Effect0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 00.20.40.6 Wt% TiK calc Ni Ta Al Ti Figure 4-13. Predicted Ti (and Ta or Al) variation effects on elemental segregation with respect to Ti (wt%) concentration. The bar chart (Figure 4-14) below shows the predicted partitioning coefficients for the Ti variants considered with respect to their Ti content. The predicted decrease in Re partitioning with increased Ti content is seen with Al Partitioning Effect0.9 1 1.1 1.2 1.3 1.4 1.5 00.20.40.6 wt% Tik calc Cr Co W Re Model Alloy Model Alloy

PAGE 87

69 reductions in Model 9 (0.2 wt% Ti) and Model 11 (0.58 wt%), or with Ta reductions in Model 10 (0.39 wt% Ti). k calc Comparisons for wt %Ti Variations0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 TiTaAlCrNiWCoRek calc 0.58 0.39 0.2 0 Model Alloy Figure 4-14. kcalc comparisons bet ween the baseline Model alloy and Ti variants Rhenium (and Tantalum, Tungsten) variation effects Re (with Ta, Al, and W) variation e ffects on material properties were evaluated using three alloy compositions. Re reductions to the baseline Model composition of 1.5 wt% and 3.02 wt% Re (0 .5 to 1 at% Re) were made in Model 15 (7.46 wt% W) and Model 16 (7.34 wt% Ta, 7.46 wt% W, 5.45 wt% Al) alloys, respectively, and were compared to the ba seline Model alloy (3.02 wt% (1 at%) Re). Microstructural Stability Thermodynamic calculations predicted t hat for alloys with a Re content of 3.02 wt% (1 at%) Re, the equilibrium phases at temperatures below 1000 C include , and The alloy with the low Re c ontent (1.5 wt% (0.5 at%) Re), was predicted to contain equilibrium phases , and at temperatures below 1000C.

PAGE 88

70 No linear relationships on the amount of specific TCP phases (with respect to Re content) were observed. Even though no linear relationships exist, Figure 4-15 shows a strong Re effect predicted on the amount of TCP phases predicted at 900 C. Despite W, Ta, and Al vari ations in the Model 16 alloy and W increases in the Model 15 alloy; TCP phase amounts were predicted to decrease for both alloys when compared to baseline M odel alloy. This is observed when comparing the Model 16 alloy (0 wt% Re) to the baseline Model alloy (3.02 wt% Re), which are predicted to contain 0 wt% and 6 wt% in TCP phases at 900 C, respectively. Stability Effect at 900 C6.3 wt% Sigma 5 wt% Mu0 1 2 3 4 5 6 7 00.511.522.533.5 Re Variations (wt%)wt % Sigma Mu Figure 4-15. Predicted Re (with Ta, Al, and W) variation effects on TCP equilibrium phase amounts with respect to Re concentration (weight percent). Phase Transformation Temperatures JMatPro thermodynamic equilibrium calc ulations predicted that the baseline Model alloy with a Re content of 3. 02 wt% (1 at%) Re would follow the solidification path seen below. L L + + + + Model Alloy

PAGE 89

71 The Model 15 alloy with a low Re cont ent of 1.5 wt% (0 .5 at%) Re was predicted to solidify as seen below. L L + + + + + + + Thermodynamic equilibrium calculations predicted that t he Model 16 alloy, with a 0 wt% Re (0 at%) Re content, would solidify as seen below. L L + + Figure 4-16 below depicts the liquidus, solidus, and solvus temperature trends predicted with respect to Re alloy vari ations (with W, Al, and Ta variations in Model 16 and W increases in Model 15). Phase Transformation Effectsy = -3x + 12891270 1290 1310 1330 1350 1370 0123 Re Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-16. Predicted Re (with Ta, Al, and W) variation effects on phase transformation temperatures with res pect to Re (wt%) concentration. A 1.5 wt% (0.5 at%) Re reduction wit h a 0.5 at% W addition, resulted in a calculated 4 C decrease in the liquidus. Removing Re from the alloy chemistry (while increasing Al 0.5 at%, increasi ng W 0.5 at%, and reducing Ta 0.5 at%) decreased the liquidus 4 C and increased the solvus 7 C. These thermodynamic calculations predicted that Re reductio ns in Model 15 (with W increases) and Model 16 (with Al, W incr eases and Ta reductions) produced a linear increase in the solvus temperature. Model Alloy

PAGE 90

72 The linearly decreasing solvus temperature wit h increasing Re content results in a decrease of the heat treatm ent window. A 1.5 wt% (0.5 at%) Re reduction with a 0.5 at% W increase was predicted to decrease the heat treatment window 4 C. Removing Re from the alloy chemistry (while increasing Al 0.5 at%, increasing W 0. 5 at%, and reducing Ta 0.5 at%) decreased the heat treatment window 8 C. Elemental Segregation Predicted elemental variation effects on segregation for the baseline Model, Model 15, and Model 16 alloys with respec t to increasing Re concentration are seen in Figure 4-17. Figure 4-17. Predicted Re (with Ta, Al, and W) variation effects on elemental segregation with respect to Re (wt%) concentration. Elements predicted as segregat ing to the dendrite core were Cr, Ni, W, Co, and Re. Re exhibited the most severe segregation, followed by W, Co, Ni, and then Cr. A 3.02 wt% (1 at%) Re reduction (with a 0.5 at % Al increase, a 0.5 at% W increase, and a 0.5 at% Ta reduction) resulted in a negligeble 2% decrease in Partitioning Effect 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 01234 wt% Rek calc Ni Ta Al Partitioning Effect 0.9 1 1.1 1.2 1.3 1.4 1.5 01234 wt% Rek calc Cr Co W Re Model Alloy Model Alloy

PAGE 91

73 kNi,calc, a limited 2.5% increase in kW,calc, and a more notable 3.5 to 4% increase in kCo,calc and kCr,calc, respectively. Ta and Al were predicted to segregate to the interdendritic region. Ta was the most segregated, followed by Al. Part itioning coefficients for Ta and Al went relatively unchanged in comparisons between the baseline Model, Model 15, and Model 16 alloys. The bar chart (Figure 4-18) below s hows predicted elemental partitioning coefficients for the Re variants consider ed with respect to their Re content. Extensive Re partitioning was avoided when Re was omitted from alloy compositions. k calc Comparisons for wt %Re Variations0.20 0.40 0.60 0.80 1.00 1.20 1.40 TaAlCrNiWCoRek calc 3.02 Model Alloy 1.5 0 Figure 4-18. kcalc comparisons between baseline Model alloy and Re variants Carbon variation effects C variation effects on material properti es were evaluated by comparing the Model 3 (0.01 wt% C), Model 2 (0. 05 wt% C), Model 1 (0.075 wt% C), and baseline Model (0 wt% C) alloys.

PAGE 92

74 Microstructural Stability Thermodynamic calculations for all C variants revealed that equilibrium phases below 1000 C include , and M23C6. The formation of M23C6 carbides was predicted for all C variants. C additions were predicted to decrease the amount of phase expected at 900 C. The linear relationship predict ed between C content and the amount of TCP phases at 900 C (Figure 4-19), expec ted a 10 wt% reduction in the amount of phase with a 1 wt% C increase. A 0. 68 wt% reduction in the amount of phase at 900 C was calculated for t he maximum 0.08 wt% C addition used in this study. Stability Effect at 900 Cy = -11x + 6 5.25 5.5 5.75 6 6.25 6.5 00.020.040.060.08C Variations (wt%)wt % Sigma Figure 4-19. Predicted C variation effe ct on TCP equilibrium phase amount Phase Transformation Temperatures The predicted solidification path for the C containing alloys is seen below. L L + L + + MC + + MC + + M23C6 + + + M23C6 Figure 4-20, depicts the predicted C variation effects on the liquidus, solidus, and solvus temperatures. Model Alloy

PAGE 93

75 C additions resulted in a predicted linear decrease of the liquidus and solvus temperatures. A 0.075 wt% C additi on was predicted to linearly decrease the liquidus and solvus by approximately 6 C and 9 C, respectively. A linear increase of the solidus was predicted wit h increasing C content. Calculations predicted a 9 C increase in the solidus with a 0.075 wt% C addition. Phase Transformation Effects1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 00.020.040.060.08 C Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-20 Predicted C variation effect s on phase transformation temperatures The predicted linear decrease in the liquidus and linear increase in the solidus with increasing C content results in a decrease in the solidification range (15 C with a 0.075 wt% C addition). The predicted inverse C effects on the solidus and solvus temperatures, resulted in the increase of the heat treatment window with increasing C content. A 0.075 wt% C addition was predicted to increase the heat treatment window 18 C. Elemental Segregation The predicted C variation effects on s egregation behavior in this study are seen below in Figure 4-21. Elements predicted to have k calc values greater than one or who tend to partition to the dendrite core, were Ni, W, Co, and Re. Re segregation was the strongest, followed by Co, W, and t hen Ni. Within the 0.075 Model Alloy

PAGE 94

76 wt% C range considered, C variations were predicted to have a no effect on Co, W, and Ni segregation. Negligible C vari ation effects were calculated for Re. Partitioning Effect0 0.2 0.4 0.6 0.8 1 1.2 00.020.040.060.08 Wt% CK calc Ni Ta Al C Figure 4-21. Predicted C variation effects on elemental segregation. k calc Comparisons for wt %C Variations0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 CTaAlCrNiWCoRek calc 0.075 0.05 0.01 0 Model Alloy Figure 4-22. kcalc comparisons between t he baseline Model alloy and C variants Ta, Al, Cr, and C were predicted to segr egate to the interdendritic region. The partitioning of C was the strongest, follo wed by Ta, Al and then Cr. A small linear decrease in Al segregation was predi cted for C increases (a 3.3% increase in kAl,calc with a 0.075 wt% C addition). Negligible C variation effects were Partitioning Effect 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 00.020.040.060.08 wt% Ck calc Cr Co W Re Model Alloy Model Alloy

PAGE 95

77 calculated for C, Ta and Cr. Figure 422 shows the predicted C variation effects on partitioning coefficients. Predicted C vari ation effects were most significant for Al. Cobalt variation effects Co effects on material properties were evaluated using three alloy compositions. Co content ranging from 9 wt% (9.41 at%) to 12 wt% (12.54 at%) Co in Model 22 and Model 21 alloys, respec tively, were compared to the baseline Model alloy (11.47 wt% (12 at%) Co). Microstructural Stability Thermodynamic calculations for all Co variants predicted that equilibrium phases below 1000 C include , and Co additions had no effect on the amount of phase predicted at 900 C. Within the 3 wt% (3 at%) Co range evaluat ed in this study, calculations at 900 C predicted phase amounts nearly identical to t hose of the baseline Model alloy (Figure 4-23). Stability Effect at 900 Cy = 0.03x + 66.2 6.22 6.24 6.26 6.28 6.3 6.32 8.59.510.511.512.5 Co Variations (wt%)wt % Sigma Figure 4-23. Predicted Co variation e ffect on TCP equilibrium phase amount Model Alloy

PAGE 96

78 Phase Transformation Temperatures Co variants considered in this study were predicted to follow the solidification path seen below. L L + + + + Figure 4-24 depicts the predicted Co variation effects on the liquidus, solidus, aned solvus temperatures. Phase Transformation Effectsy = 4x + 1233 y = -4x + 1341 y = -1x + 13741260 1280 1300 1320 1340 1360 1380 99.51010.51111.512Co Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-24. Predicted Co variati on effects on phase transformation temperatures. Calculations predicted a linear decrease in the solidus and a linear increase in the solvus; both at the rate of 4 C with a 1 wt% Co addition. Co variation effects on the liquidus temperature were considered negligible. Even though the solidus was predict ed to decrease with increasing Co content, only a negligible increase in the melting range was observed (approximately 2 C with a 3 wt% Co addition). T he combined effect of an increased solvus and decreased solidus with increasing Co content, was predicted to decrease the heat treatment window. Decreasing the Co content from 11.47 to 9 wt% was predicted to in crease the heat treatment window 19 C. Model Alloy

PAGE 97

79 Elemental Segregation Predicted Co effects on segregati on behavior are seen in Figure 4-25. Figure 4-25. Predicted Co variation effects on elemental segregation Ni, W, Co, Cr, and Re were predicted to partition to the dendrite core with k calc values greater than one. Re was t he most segregated element, followed by W, Co, Ni, and then Cr. Co variations were predicted to have negligible effects on W, Re, Co, Ni, and Cr segregation, wit hin the 3 wt% Co range considered. Elements calculated to have kcalc values less than one or who partition to the interdendritic region were Ta and Al. Ta segregated to the greatest extent, followed by Al. The largest Co variat ion effect was predicted for Ta, which linearly increased segregation towards the interdendritic region, with increased Co content (a 4% decrease in kTa,calc with a 3 wt% Co addition). No significant Co variation effects were calculated for Al. Figure 4-26, below, compares predict ed elemental partitioning coefficients for the Co variants considered. Predicted Co effects were most significant for Ta. The negligible Co variation effects predi cted for Re, Al, and Ni are also observed. Partitioning Effect 0.9 1 1.1 1.2 1.3 1.4 1.5 8.59.510.511.512.5 wt% Cok calc Cr Co W Re Partitioning Effect 0.9 1 1.1 1.2 1.3 1.4 1.5 8.59.510.511.512.5 wt% Cok calc Cr Co W Re Model Alloy Model Alloy

PAGE 98

80 k calc Comparisons for wt %Co Variations0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60TaAlCrNiWCoRek calc 9 11.47 Model Alloy 12 Figure 4-26. kcalc comparisons betw een Model alloy and Co variants Ruthenium variation effects Ru variation effects on material properti es were evaluated using three alloy compositions. Ru additions in Model 5 (1.64 wt% (1 at%) Ru and Model 4 (2.46 wt% (1.5 at%) Ru) were compared to the baseline Model (0 wt% Ru) alloy. Microstructural Stability Calculations for the Ru variants consi dered in this study predicted that the , and equilibrium phases would be present below 1000 C. Stability Effect at 900 Cy = 0.2x + 6.3 6.2 6.3 6.4 6.5 6.6 6.7 6.8 00.511.522.5 Ru Variations (wt%)wt % Sigma Figure 4-27. Predicted Ru variation e ffect on TCP equilibrium phase amount Model Alloy

PAGE 99

81 Ru additions were calculated to hav e a minimal effect on TCP phase amounts. At 900 C the amount of phase was predicted to only increase (0.42 wt% ) with a 2.46 wt% (1.5 at %) Ru addition (Figure 4-27). Phase Transformation Temperatures Thermodynamic equilibrium calculations for all the Ru variants considered in this study, predicted the solidification path seen below. L L + + + + Predicted Ru variation effects on the liquidus, solidus, and solvus temperatures are seen in Figure 4-28. Phase Transformation Effectsy = -3x + 1297 y = -0.6x + 1364 y = -3x + 12821270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 00.511.522.5 Ru Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-28. Predicted Ru variati on effects on phase transformation temperatures Ru additions showed a negligible effe ct on the liquidus temperature. Calculations also predicted linear decreases of the solidus and the solvus, at rates of approximately 7 C and 8 C, res pectively, for a 2.5 wt% Ru addition. Decreases in the solidus and nearly c onstant liquidus predictions, resulted in a small increase of the so lidification range. Increasi ng the Ru concentration by 2.46 wt% Ru was predicted to increase t he solidification range 5 C. Since Ru Model Alloy

PAGE 100

82 additions linearly decrease the solidus and the solvus at similar rates, the heat treating window was predicted to remain relatively constant. Elemental Segregation Predicted Ru variation effects on s egregation behavior are seen below in Figure 4-29. Partitioning Effect0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 0123Wt% RuK calc Ni Ta Al Ru Figure 4-29 Predicted Ru variatio n effects on elemental segregation Elements calculated to have k calc values greater than one, which tend to partition to the dendrite core, were Re, Ni, W, Co, and Cr. Re was predicted to exhibit the most severe segregation, fo llowed by Co, W, Ni, and then Cr. Ru variations were predicted to have a neglig ible effect on Re, Ni, Cr, and Co segregation, within the 2.46 wt% Ru range considered. A linear decrease in W segregation was predicted wit h increasing Ru content (a 4.5% decrease in kW,calc was predicted with a 2.46 wt% Ru addition). Ru, Ta, and Al were predicted to partiti on to the interdendritic region. The most significant segregation was predict ed for Ta, followed by Al, and then Ru. No significant Ru variation effects were calculated for Ru,Ta, or Al. Partitioning Effect0.9 1 1.1 1.2 1.3 1.4 1.50123 wt% Ruk calc Cr Co W Re Model Alloy Model Alloy

PAGE 101

83 The bar chart (Figure 4-30) below compares predicted elemental partitioning coefficients for the Ru variants considered. Predicted Ru variation effects were most significant for W. k calc Comparisons for wt %Ru Variations0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 RuTaAlCrNiWCoRek calc 0 Model Alloy 1.64 2.46 Figure 4-30 kcalc comparisons betw een Model alloy and Ru variants Tungsten (and Molybdenum) variation effects W variation effects on material properti es were evaluated using two alloy compositions. A 1 at% W r eduction with a 1 at% Mo substitution in the Model 23 (1 at% W, 1 at% Mo) alloy was compared to the baseline Model alloy (2 at% W, 0 at% Mo). Microstructural Stability Thermodynamic calculations for t he Model 23 alloy (the W (and Mo) variant) predicted equilibrium phases under 1000 C to include , and A minimal 0.4 wt% increase in the amount of phase predicted at 900 C, was calculated for a 2.98 wt% (1 at%) W reduction (and 1 at% Mo substitution) at 900C. The limited W (and Mo) effect on TCP phases amounts predicted at 900 C is seen below in Figure 4-31.

PAGE 102

84 Stability Effect at 900 Cy = -0.1x + 7 6.2 6.3 6.4 6.5 6.6 6.7 6.8 2.53.54.55.56.5 W Variations (wt%)wt % Sigma Figure 4-31. Predicted W (and Mo) vari ation effect on TCP equilibrium phase amount with respect to W (wt%) concentration. Phase Transformation Temperatures Thermodynamic equilibrium calculations predicted the solidification path seen below for the Model 23 alloy with a W reduction (and Mo substitution). L L + + + + P + + Figure 4-32 depicts the predicted W ( and Mo) variation effects on the liquidus, solidus, and solvus temperatures. Phase Transformation Effectsy = 1x + 1357 y = 1x + 1288 y = 0.6x + 12781270 1280 1290 1300 1310 1320 1330 1340 1350 1360 13702.53.54.55.56.5W Variations (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-32. Predicted W (and Mo) vari ation effects on phase transformation temperatures with respect to W (wt%) concentration. Model Alloy Model Alloy

PAGE 103

85 Calculations predicted a small linear decrease in the liquidus and solidus with decreasing W (increasing Mo) content. The liq uidus and solidus temperatures were both predicted to decrease approximately 4 C, with a 2.98 wt% (1 at%) W reduction (and a 1 at% Mo substitution). Similar W (and Mo) effects predicted for the solidus and liquidus temperatures yielded a nearly constant solidification range; within the 2.98 wt% W (1.56 wt% Mo) range analyzed. The sm all predicted decrease in the solidus and a nearly constant solvus resulted in a neglig ible decrease of the heat treatment window with W reductions (and Mo substitutions). Elemental Segregation Predicted segregation behavior the W vari ants considered is seen below in Figure 4-33. Partitioning Effect0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.12.53.54.55.56.5 Wt% WK calc Ni Ta Al Mo Figure 4-33. Predicted W (and Mo) vari ation effects on el emental segregation with respect to W (w t%) concentration. Elements calculated as partitioning to t he dendrite core were Cr, Ni, W, Co, and Re. Re segregated to the greatest ex tent, followed by W, Co, Ni and then Partitioning Effect 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 2.53.54.55.56.5 wt% Wk calc Cr Co W Re Model Alloy Model Alloy

PAGE 104

86 Cr. Calculations expected Re and W segregation to increase linearly with W reductions (and Mo substitutions). A 7% linear increases in kRe,calc and a 4.5% linear increase in kW,calc were predicted with a 2. 98 wt% W reduction and 1.56 wt% Mo substitution. With k calc values less than one, Ta, Mo, and Al were predicted to segregate to the interdendritic region. Ta wa s calculated as the strongest segregating element, followed by Mo and Al. A 3.3% linear increase in kAl,calc was predicted with a 2.98 wt% W reduction and 1.56 wt% Mo substitution. A negligible W (and Mo) variation effect was predicted for Ta and Mo partitioning as W content decreased (and Mo content increased). The bar chart below (Figure 4-34) shows predicted W (and Mo) variation effects on partitioning coeffi cients. Increases in s egregation predicted with a 1 at% W reduction (and 1 at% Mo substitution) were mo st significant for Re, followed by W. Al was calculated to decrease in segregation with W reductions (and Mo additions). k calc Comparisons for wt %W Variations0.2 0.4 0.6 0.8 1 1.2 1.4 1.6MoTaAlCrNiWCoRek calc 5.96Model Alloy 2.98 Figure 4-34. kcalc comparisons between Model alloy and W (and Mo) variants

PAGE 105

87 Gamma prime former (Tantalum, Alumi num, and Titanium) variation effects -former variations from 13.25 to 16.75 at%, in Model 20 ( 2 at% Ta, 11 at% Al, 0.25 at% Ti); Model 19 ( 3 at% Ta 11 at% Al, 0.25 at% Ti); Model 18 (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti); and Model 17 (3.25 at% Ta, 12.75 at% Al, 0.75 at% Ti) were compared to the baseline M odel alloy (3 at% Ta, 12 at% Al). Microstructural Stability Thermodynamic calculations for all the variants, with t he exception of Model 17, predicted the , and equilibrium phases at 1000 C. Calculations for the Model 17 alloy predicted the , and as the equilibrium phases at 1000 C. Taking into consideration all -former variants, a linear increase in the amount of phase was predicted with increasing -former content. Comparing Model 17 to Model 20 at 900 C, the amount of phase was predicted to increase approximat ely 5.5 wt% with a -former increase of 3.5 at% (including 1.25 at% Ta, 1.8 at% Al, and 0.5 at% Ti additions). -former effects on the amount of TCP phases, predicted at 900 C, are seen below (Figure 4-35). Model 17 was predicted to contain a small amount of phase, which was not predicted for the baseline Model, Model 20, Model 19, or Model 18 alloys. Comparing variants with similar -former content, Model 18 (16.5 at% -former) and Model 17 (16.75 at% -former) at 900 C, the ph ase was predicted for the alloy with lower Al but higher Ta and Ti content. The presence of the may be attributed to the combined elemental effe cts of the 0.9 at% Al reduction along with the 0.95 at% Ta and 0.25 at% Ti additions.

PAGE 106

88 Stability Effect at 900 C0 2 4 6 8 10 12 1314151617 gamma prime (at%)wt % Sigma Mu Figure 4-35. Predicted -former variation effects on TCP equilibrium phase amounts. Phase Transformation Temperatures The solidification path predicted for the Model 20, Model 19, and Model 18 alloys is seen below. L L + + + + The solidification path predicted for the Model 17 alloy is seen below. L L + + + + + Figure 4-36 depicts predicted -former (including Al, Ta, and Ti) variation effects on the liquidus, solidus, and solvus temperatures. Despite Ti, Ta, and Al variations in the alloys considered; clear trends were predicted for phase transformation temperatures, with increasing -former content. Thermodynamic calcul ations predicted that increased -former content would linearly decrease the solidus and liquidus temperatures. A predicted increase in the solvus can be described by a 2nd order relationship. Linear trends predicted a decrease in the solidus and liquidus at the rates of 13 C and 9.5 C, respectively, with a 1 at% -former increase. With -former increases of Model Allo y

PAGE 107

89 3.5 at% (including 1.25 at % Ta, 1.8 at% Al, and 0.5 at% Ti additions) between Model 17 and Model 20, calculations pr edicted a liquidus decrease of 40 C, a solidus decrease of 58 C, and a solvus increase of 56 C. Phase Transformation Effectsy = -10x + 1507 y = -13x + 15001200 1220 1240 1260 1280 1300 1320 1340 1360 1380 14001314151617gamma prime (at%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-36. Predicted -former variation effe cts on phase transformation temperatures. With increasing -former content, the calculated rate at which the solidus decreased was greater than the rate at which the liquidus decreased, resulting in an increase in the melting range. With a -former increase of 1.75 at% (including 0.25 at% Ta, 0.8 at% Al, and 0.75 at% Ti additions) between Model 17 and the baseline Model alloy, the melti ng range was predicted to increase 4C. Increased -former content was predicted to decrease the solidus and increase the solvus, resulting in a consider able decrease in the heat treatment window. A -7 C heat treatment windo w was calculated for an alloy containing 16.75 at% in -former content (Model 17). Elemental Segregation Predicted -former variation effects on segregation are seen below in Figure 4-37. For the -former variants considered (Model 20, Model 19, baseline Model Alloy

PAGE 108

90 Model, Model 18, and Model 17), the el ements Cr, Ni, W, Co, and Re were predicted to partition to the dendrite core Re was the most heavily segregated element, followed by Co, W, Ni, and then Cr. Partitioning Effect0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1314151617gamma prime (at%)k calc Ni Ta Al Ti Figure 4-37. Predicted -former variation effects on elemental segregation Re, Cr, and Co segregation increased in all alloys with increasing -former content. With a -former increase of 3.5 at% (inc luding 1.25 at% Ta, 1.8 at% Al, and 0.5 at% Ti additions), kRe,calc and kCr,calc, values increased by 5.5% and 6%, respectively. With a -former increase of 1.5 at% (including 0.5 at% Ti and 1.7 at% Al additions with a 0. 7 at% Ta reduction), kRe,calc and kCr,calc values increased by 3.4% and 3.1%, respectively. The influence of -former content on Ni and W segregation could not be clearly establishe d, which may be attributed to specific Al, Ti, or Ta variations in the alloys compositions. Ta, Ti, and Al were predicted to segregate to the interdendritic region. Ta was the most segregated element, follo wed by Ti, and then Al. With a -former increase of 3.5 at% (incl uding 1.25 at% Ta, 1.8 at% Al and 0.5 at% Ti additions), kTi,calc decreased by 9%. No elemental effe cts were predicted for Al or Ta. Partitionin g Effect 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1314151617 gamma prime (at%)k calc Cr Co W Re Model Alloy Model Alloy

PAGE 109

91 kcalc values for the -former variants considered in this study, are seen in Figure 4-38 with respect to their -former content. -former variation effects were most significant for Ti, followed by Cr and then Re. The predicted increase in Re, Cr, and Ti partitioning, with increased -former content, is evident when comparing Model 17 to Model 20. k calc Comparisons for at %y' former Variations 0.35 0.55 0.75 0.95 1.15 1.35 1.55 TiTaAlCrNiWCoRek calc Gamma prime vol = 16.75 at% Increase Ta, Al,Ti Gamma prime vol = 16.5 at% Increase Al and Ti, Lower Ta Model Alloy: Gamma prime vol = 15 at% Gamma prime vol = 14.25 at% *Lower Al, IncreasedTi Gamma prime vol = 13.25 at% Lower Ta and Al, Increase Ti Figure 4-38. kcalc compar isons between Model alloy and -former variants Temperature Range Comparisons fo r Phase I of Alloy Development Figure 4-39 below was produced as an addi tional visual comparison, for the baseline Model alloy and Phase I vari ants, to gauge predicted elemental effects on both the melting range and the heat tr eatment window. In Figure 4-39, compositional variants are i dentified by model number, ma in elemental variation, and a High -H, Medium-M, or Low-L mark er (indicating the interest elements concentration level with respect to the bas eline Model alloy). For example, the Model 20 alloy has a Low former content when compared to the baseline Model alloy (13.25 at % in former content vs 15 at% in former content). The Model 20 alloy is identif ied in Figure 4-39 as 20 -L. The melting ranges and

PAGE 110

92 heat treatment windows calculated for selective 1st and 2nd generation commercial and experimental alloys were also plotted for comparison. Heat treating Window vs. Alloy Melting Range -15 10 35 60 85 110 4050607080Liquidus to Solidus (Celcius)Solidus to Y' solvus (Celcius)1C-H 14Al-L 5Ru-L4Ru-M 6Cr-H 9Ti-H 16Re-L 7Cr-M 19Y'-M 10Ti-M 8Cr-L 22Co-L 2C-M 13Al-M 20Y'-L 12Al-H 11Ti-L Model 23W-L 21Co-H 17Y'-H 3C-L 15Re-M 18Y'-M IN-738 CMSX-4 PWA 1483 CMSX-10 'Phase I' Compositional Variants Experimental and Commercial Figure 4-39. Calculated heat treatment window (solidus solvus) vs. melting range (liquidus solidus) for the baseline Model composition, the compositional variants in Phase I of alloy development, and selective 1st and 2nd generation commercial and experimental alloys. Notably, increases in C, Al, or -former contents, were predicted to decrease the melting range. A predi cted increase in melting range with increased Cr concentration is evident w hen comparing the Model 6 (6Cr-H) and Model 8 (8Cr-L) alloys (where a 17 C melting range increase was calculated for a 3.4 wt% Cr addition). Melting ranges from 51 C to 74 C were calculated for the compositional variants in Phase I of alloy development. Melting ranges fo r the compositional variants were predicted to remain approxim ately + 9 C from the baseline Model alloys melting range (67 C). Melting r anges from 45 C to 85 C are shown for commercial alloys: CMSX-4, CMSX-10, PW A 1483, and IN 738 [11,52,120,121].

PAGE 111

93 Predicted Al (and Ta), Ti, and -former variation effects on the heat treatment window are also seen in Figure 4-39. Notably, t he heat treatment window was predicted to signific antly decrease with increasing -former content. With increasing Al (decreasing Ta) and in creasing Ti content, the heat treatment window was predicted to increase. The predicted increase in heat treatment window with increased Al (reduced Ta) c ontent is easily seen when comparing the Model 12 (12Al-H) and Model 14 ( 14Al-L) alloys (where a 83 C heat treatment window increase was calculat ed for a 0.96 wt% Al addition (and 6.48 wt% Ta reduction)). Heat treatment windows r anging between -7 C to 50 C were calculated for most compositional variants in Phase I, with the exception of the Model 12 and Model 20 alloys. The Model 12 alloy (wit h high Al (and low Ta) concentrations) and the Model 20 alloy (with a low former content) were calculated to have heat treatment windows approximately 20 to 40 C higher than those reported for single crystal alloys considered for IGT use (i.e., CMSX-4 and PWA 1483) [11,121]. Elemental Variation Trend Summary for Phase I The predicted general effects caused by compositional changes to the baseline Model alloy composition in Phas e I are summarized in Table 4-3. Overall calculated elemental variation e ffects on microstructural stability, phase transformation temperatures, and element al segregation are shown for the characteristic elemental ranges analyzed in this study. The predicted elemental effects are listed follows.

PAGE 112

94 Table 4-3 Predicted elemental variation e ffects on microstructural stability, phase transformation temperatures and elemental segregation. indicates an increasing effect, indicates a decreasing effect, indicates a limited or negligible effect, indi cates that the influence of the elemental variation could not be clear ly established, and N/A indicates that the property in question was not applicable to the material system. Characteristic Elemental Variations Elemental Segregation HTBase composition: Model alloy wt%wt% solvusWindow C variations N/A AlIncreasing C from 0 to 750 ppm wt% C Co, W, Ni, Re, Cr Accompanying Compositional Modifications: None Ru Variations Increasing Ru from 0 to 2.46 wt% Ru N/A W Re, Cr, Co, Ni, Ta, Al Accompanying Compositional Modifications: None Cr Variations Increasing Cr from 6.75 wt% to 11.82 wt% Cr N/A Ta W, Re Ni, Co, Al Accompanying Compositional Modifications: None Ti (with Al or Ta) variations Increasing Ti contents of 0, 0.2, 0.39, and 0.58 wt% Ti N/A Re Cr, Co, Ta, Al Accompanying Compositional Modifications: The 0.58 wt% Ti addition is a substitution for Al The 0.39 wt% Ti is a substituion for Ta The 0.2 wt% Ti addition is a substitution for Al Al (and Ta) Variatons Increasing Al from 5.04 to 6 wt% Al N/A Re, Co, Ta W, Ni Al, Cr Accompanying Compositional Modifications: Ta substitution for all Al modifications Re variations Increasing Re contents of 0, 1.5 and 3.02 wt% Re Re Ni, W, Ta, Al Co, Cr Accompanying Compositional Modifications: 1.5 wt% Re alloy has a 1.5 wt% W addition 0 wt% Re alloy has 1.5 wt% W and 0.2 wt% Al additions and a 1.46 wt% Ta reduction (Ta, Al,Ti) Former Variations -former variations from 13.25 to 16.75 at% N/A Re, Cr Co, Ta, Al W Accompanying Compositional Modifications: 13.25 -former: ( 2 at% Ta, 11 at% Al, 0.25 at% Ti) 14.25 -former: ( 3 at% Ta, 11 at% Al, 0.25 at% Ti) 15 -former: (3 at% Ta, 12 at% Al) 16.5 -former: (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti) 16.75 -former:(3.25 at% Ta, 12.75 at% Al, 0.75 at% Ti) Co variations Increasing Co content from 9 wt% to 12 wt% Co N/A Ta W, Re, Co, Ni, Cr, Al Accompanying Compositional Modifications: None W (and Mo) variations Increasing W from 2.98 wt% W to 5.96 wt% W N/A Al Re, W Ta, Al Accompanying Compositional Modifications: 2.98 wt% W alloy has a 1.56 wt% Mo substitution for W MROverall Segregation StabilityPhase Transformation Temps. (L)(S)

PAGE 113

95 Phase II Computational Alloy Refinement Compositional modifications to the bas eline Model alloy based on Phase I material property trends, resulted in a modified baseline composition, assigned the name Model A. The calculated materi al properties for the baseline Model A alloy were discussed first. The calculated material properties for the other 16 experimental alloys were grouped with respec t to their characteristic elemental variations (Al/Ti ratio, Re, Cr (with Ti and Al), and -former content). Baseline Model A Alloy. The composition for the baseline Model A alloy is seen in Table 4-4. Table 4-4. Baseline Model A allo y composition in wt% and at%. CompositionAl/Ti at% NiAlCoCrHfReTaWTiC wt %54.474.8211.00110.103.028.805.960.780.056.2 at %57.2111.0011.50130.031.003.002.001.000.2615 Microstructural Stability Thermodynamic calculations predict ed that the equilibrium phases below 1000 C include , , and M23C6. The predicted phase diagram for the baseline Model A alloy com position calculated an alloy made up primarily of the precipitate (approxim ately 66 wt%) and the matrix (approximately 20 wt%) at 600 C. Limited amounts of the M23C6 carbide (<1 wt%), and the TCP phase (approximately 10 wt%) are predicted to be in equilibrium at 600 C. A small amount of the TCP phase (approximately 3 wt%) is predicted at 600 C. Of the TCP phases, only is predicted at 900 C (approximately 7 wt%). Phase Transformation Temperatures According to JMatPro equilibrium thermodynamic calculations, the solidification path for the baseline Model A alloy is seen below.

PAGE 114

96 L L + L + + MC + + MC + + MC + + + M23C6 + + + M23C6 + + The predicted liquidus, MC solvus, solidus, solvus, solvus, and solvus temperatures for the baseline Model A alloy were 1351 C, 1338 C, 1293 C, 1246 C, and 1208 C, and 887 C, respectively A melting range of 58C was predicted for the baseline Model A allo y composition. The heat treatment window predicted for the Model A alloy was 47 C. Elemental Segregation Elemental partitioning coefficients (kcalc) predicted for the Ni, Ta, Al, Cr, Co, W, and Re elements in the baseline Model A alloy were the same as those shown for the initial baselin e Model alloy (Table 4-2). Elemental Variation Effects The microstructural stability, transf ormation temperatures, and elemental segregation effects from t he 16 compositional variants in Phase I of alloy development are discussed below. Rhenium variation effects Three separate variant groups were used to investigate Re effects on material properties: Group 1: Re levels of 0 and 1.5 wt% Re (0 and 0.5 at% Re) in Model B and Model C were used as a comparison to t he 3.02 wt% Re (1 at % Re) content of the baseline Model A alloy. All three a lloys contained a constant Al/Ti Ratio of 6.2 and a total of 15 at% in -formers. Group 2: Four Re levels of 3.02, 2.28, 1.5, and 0 wt% Re (1, 0.75, 0.5, and 0 at% Re), were incorporated into the M odel D, E, F, and G alloys, respectively.

PAGE 115

97 All four alloys contained a cons tant Al/Ti ratio of 2.54 and a -former content of 14 at%. Group 3: Four Re levels of 3.02, 2.28, 1.5, and 0 wt% Re (1, 0.75, 0.5, 0 and at% Re) were incorporated into the Model H, I, J, and K alloys, respectively. All four alloys contained an Al/Ti ratio of 1.69 and a -former content of 14 at%. Microstructural Stability Thermodynamic calculations predicted , , and M23C6 carbides as equilibrium phases in all Re varian ts at temperatures below 1000C. A increase in the amount of phase at 600 C was predicted for increasing Re content. Regardless of compositi onal differences between variant groups, calculations predicted clear linear and 2nd order increases in the amounts of and phases with increasing Re content, respectively. A 3.02 wt% Re addition in Group 1, Group2, and Group 3 alloys at 600 C, was predicted to increase the amount of phase by 5.03 wt%, 4.42 wt%, and 4.34 wt%, respectively. Increases of 2.41 wt%, 2.79 wt%, and 2.87 wt%, in the amount of phase, were predicted for a 3.02 wt% Re addition in Group 1, Group2, and Group 3 alloys at 600 C, respectively. Predicted Re effects on the amounts of TCP phases predicted at 600 C, are seen in Figure 4-40 for (a) Group 1, (b) Group 2, and (c) Group 3 alloys. Phase Transformation Temperatures JMatPro thermodynamic equilibrium ca lculations predicted that all Re variants with Re contents betw een 1.5 wt% (0.5 at%) Re to 3.02 wt% (1 at%) Re would follow the solidif ication path seen below.

PAGE 116

98 Stability Effects at 600 C Group 10 2 4 6 8 10 12 00.511.522.53 Re (wt%)wt % Sigma Mu Stability Effects at 600 C Gr oup 20 1 2 3 4 5 6 7 00.511.522.53 Re (wt%)wt % Sigma Mu Stability Effects at 600 C Group 30 1 2 3 4 5 6 7 00.511.522.53 Re (wt%)wt % Sigma Mu Figure 4-40. Predicted Re variation e ffects on TCP equilibrium phase amounts with respect to Re (wt%) concentra tion for (a) Group 1, (b) Group 2, and (c) Group 3 variants Model A Alloy (a) (b) (c)

PAGE 117

99 L L + L + + MC + + MC + + M23C6 + + + + M23C6 Compositional variants in all groups with 0 wt% Re (0 at%) Re were predicted to solidify as seen below. L L + L + + MC + + MC + + + MC + + + M23C6 + + + + M23C6 Figure 4-41 below depicts the liquidus, solidus, and solvus temperature trends predicted with respect to Re vari ations in Group 1, Group 2, and Group 3 alloys. Increasing Re concentrations were predicted to linearly decrease the solvus. A 3.02 wt% addition wa s calculated to decrease the solvus 7 C, 9 C, and 7 C in Group 1, Group 2, and Group 3 alloys, respectively. Thermodynamic calculations predicted negligible Re e ffects on the liquidus and solidus. The linear relationship between the solvus temperature and increasing Re content allows for a predicted increase in the heat tr eatment window for all Re variations. Elemental Segregation Figure 4-42 (a) (b) and (c) show t he Re variation effects on elemental segregation for the (a) Group 1, (b) Gr oup 2, and (c) Group 3 variant alloys, respectively. Elements predicted to segr egate to the dendrite core were Ni, W, Co, and Re. Re exhibited the most severe segr egation, followed by W, Co, and then Ni. Re additions resulted in nearly constant kcalc values for Co, Ni, and Re. Increasing Re content was predicted to decrease W segregation (with 5.3 %, 7

PAGE 118

100 Phase Transformation Effects Group 1y = -3x + 1252 1225 1250 1275 1300 1325 1350 1375 0123 Re (wt%)Temperature (C) Solidus Liquidus Y' solvus Phase Transformation Effects Group 2y = -3x + 12171200 1225 1250 1275 1300 1325 1350 1375 0123 Re (wt%)Temperature (C) Solidus Liquidus Y' solvus Phase Transformation Effects Group 3y = -2x + 1203 1175 1200 1225 1250 1275 1300 1325 1350 1375 0123Re (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-41. Predicted Re variati on effects on phase transformation temperatures with respect to Re (wt%) concentration for (a) Group 1, (b) Group 2, and (c) Group 3 alloys Model A Alloy (a) (b) (c)

PAGE 119

101 %, and 7% decreases in kW,calc values for a 3.02 wt% Re addition in Group1, Group 2, and Group 3 variant s, respectively). Ta, Ti, Cr, and Al were predicted to s egregate to the interd endritic region. Ta was the most segregated, followed by Ti Al, and Cr. Partitioning coefficients for Ta and Al went relatively unchanged in a ll Re variants considered. Increasing Re concentrations was predicted to li nearly shift Cr segregation from the dendritic core to the interdendr itic region in all variant groups considered. A 3.02 wt% Re addition was predicted to incr ease Cr segregation by decreasing kCr,calc by 2%, 5%, and 5% in Group 1, Group 2, and Group 3 variants. Increased Re concentrations were predicted to increas e Ti segregation. A 3.02 wt% Re addition was predicted to increase Ti segregation by decreasing kTi,calc by 2.3%, 4.2%, and 4% in Group 1, Group 2, and Group 3 variants. Chromium (Aluminum and Titanium) variation effects One variant group was used to investigat e Cr (Al and Ti) effects on material properties: Group 1: Cr variations from 11 to 12 at % Cr (13 to 14 at% Cr), in Model Q (14 at% Cr 10 at% Al, 2 at% Ti) and the baseline Model A alloy (13 at% Cr, 11 at% Al, 1 at% Ti) were compared. Microstructural Stability Thermodynamic calculations for Cr variants predicted , , and M23C6 carbides as equilibrium phases at temper atures below 1000 C. Small decreases in the amount of and TCP phases at 600 C were predicted for the Cr addition (with a Ti addition and Al reduction) considered in this study. Both and TCP phase amounts were predicted to dec rease approximately 0.6 wt% for a 1

PAGE 120

102 (a) Partitioning Effect Group 10.4 0.5 0.6 0.7 0.8 0.9 1 1.1 0123Re(wt%)K calc Ni Ta Al Ti (b) Partitioning Effect Group 20.4 0.5 0.6 0.7 0.8 0.9 1 1.1 0123Re(wt%)K calc Ni Ta Al Ti (c) Partitioning Effect Group 30.4 0.5 0.6 0.7 0.8 0.9 1 0123Re(wt%)K calc Ni Ta Al Ti Figure 4-42. Predicted Re variation effect s on elemental segregation with respect to Re (wt%) for (a) Group 1, (b) Group 2, and (c) Group 3 variants Partitioning effect Group 1 0.98 1.03 1.08 1.13 1.18 1.23 1.28 1.33 1.38 0123 Re (wt%)K calc Cr Co W Re Model A Alloy Model A Alloy Partitioning effect Group 2 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 0123 Re (wt%)K calc Cr Co W Re Partitioning effect Group 3 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 0123 Re (wt%)K calc Cr Co W Re

PAGE 121

103 wt% (1 at%) Cr increase (with a 1 at % Ti addition and 1 at% Al reduction). Cr (with Ti and Al) effects on predicted and amounts at 600 C are seen below (Figure 4-43). Stability Effects at 600 C 0 2 4 6 8 10 12 1111.211.411.611.812 Cr(wt%)wt % Sigma Mu Figure 4-43 Predicted Cr ( with Ti and Al) variation effects on TCP equilibrium phase amounts with respect to Cr (wt%) content. Phase Transformation Temperatures Thermodynamic solidification path for t he Cr variants considered is seen below. L L + L + + MC + + MC + + + MC + + + M23C6 + + + + M23C6 Figure 4-44 shows the Cr (with Ti and Al) variation effects on the liquidus, solidus, and solvus temperatures. Calculati ons predicted a 48 C decrease in the solvus for the increased Cr content (with a Ti addition and Al reduction) analyzed in this study. Increases in t he liquidus and solidus temperatures were also predicted with increasing Cr content (with a Ti addition and Al reduction). The 1 wt% (1 at%) Cr addition (wit h a 1 at% Ti addition and a 1 at% Al reduction), resulted in 7 C and 16 C in creases calculated for the liquidus and Model A Alloy

PAGE 122

104 solidus temperatures, respectively. Incr easing Cr content (with a Ti addition and Al reduction) was predicted to decreas e the melting range and considerably increase the heat treatment wi ndow. From calculations the 1 wt% (1 at%) Cr increase (with a 1 at% Ti addition and a 1 at% Al reduction) increased the heat treatment window 63 C, while de creasing the melting range 9 C. Phase Transformation Effects y = -47x + 1766 y = 17x + 1109 y = 8x + 12671200 1220 1240 1260 1280 1300 1320 1340 1360 1380 1111.211.411.611.812 Cr (wt%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-44 Predicted Cr (with Ti and Al) variation effects on phase transformation temperatures with re spect to Cr (wt%) content. Elemental Segregation Figure. 4-45 shows the predicted Cr (and Ti and Al) variation effects on segregation behavior. k calc values greater than one were calculated for Ni, W, Co, and Re for the Model Q and the baselin e Model alloy. The core segregation of Re was the most significant, followed by W, Co, and then Ni. Negligible Cr (with Ti and Al) variation effects we re predicted for Ni, Co, and W. Re segregation was predicted to decr ease (a 5.5% decrease in kRe,calc) with a 1 wt% (1 at%) Cr increase (and a 1 at% Ti addition and 1 at% Al reduction). Elements predicted to segregate to the interdendritic region, with k calc values less than one, were Ta, Ti, Al and Cr. Elemental partitioning was Model A Alloy

PAGE 123

105 predicted to be the strongest for Ta, fo llowed by Ti, Al and then Cr. Ti segregation was predicted to incr ease (a 3.8% decrease in kTi,calc) with a 1 wt% (1 at%) Cr increase (and a 1 at% Ti additi on and 1 at% Al reduc tion). Increasing Cr content (with a Ti addition and Al reduction) was calculated to have no significant effect on the segregati on behavior of Al, Cr, or Ta. Partitioning effect 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1111.211.411.611.812Cr(wt%)K calc Ni Ta Al Ti Figure 4-45. Predicted Cr (and Ti and Al) variation effects on elemental segregation with respect to Cr (wt%) content. Gamma prime former (Tantalum, Alumi num, and Titanium) variation effects Two separate variant groups were used to investigate -former effects on material properties: Group 1: -former variations from 14 to 16 at %, in Model D (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model H (2 at% Ta, 9 at% Al, 3 at% Ti); M odel O (3 at% Ta, 11 at% Al, 2 at% Ti); and Model P (3 at% Ta 10 at% Al, 3 at% Ti) were compared to the baseline Model A alloy (3 at% Ta, 11 at% Al). All five alloys contained a constant Re content of 3.02 wt% Re (1 at% Re). Group 2: -former variations from 14 to 16 at %, in Model E (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model I (2 at% Ta, 9 at% Al, 3 at% Ti ); Model B (3 at% Ta, 11 Partitioning effect 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1111.211.411.611.812 Cr(wt%)K calc Cr Co W Re Model A Alloy Model A Alloy

PAGE 124

106 at% Al, 1 at% Ti); Model M (3 at% Ta, 11 at% Al, 2 at% Ti); and Model N (3 at% Ta, 10 at% Al, 3 at% Ti) were compared. All five alloys contained a constant Re content of 2.28 wt% Re (0.75 at% Re). Microstructural Stability Thermodynamic calculations for all variants predicted the , , and M23C6 carbides as equilibrium phases at temperatures below 1000 C. Stability Effects at 600 C 3.02 wt% Re2 3 4 5 6 7 8 9 10 11 12 1414.51515.516 gamma prime (at%)wt % Sigma Mu Stability Effects at 600 C 2.28 wt% Re 2 4 6 8 10 1414.51515.516 gamma prime (at%)wt % Sigma Mu Figure 4-46. Predicted -former variation effects on TCP equilibrium phase amounts for (a) Group 1 and (b) Group 2 variants Negligible -former variation effects were predicted on the amount of phase. In comparing Model O to Model D alloys in Group 1 at 600 C, the amount of phase was predicted to incr ease only 0.2 wt%. Taking into Model A Alloy (a) (b)

PAGE 125

107 consideration all -former variants, a linear increase in the amount of phase was predicted with increasing -former content. At 600C, an approximate 5 wt% increase in the amount of phase was calculated for a 2 at% -former addition, for both Group 1 and Group 2 alloys. -former effects predicted on the amounts of TCP phases at 600 C are seen in Figure 4-46. Phase Transformation Temperatures The solidification path predicted for all -former variants in Group 1 and Group 2, is seen below. L L + L + + MC + + MC + + + MC + + + M23C6 + + + + M23C6 Figure 4-47 shows predicted -former (including Al, Ta, and Ti) variation effects on the liquidus, solidus, and solvus temperatures. Increasing -former content was predicted to linearly decrease the liquidus and solidus temperatures. A 2 at% -former increase within Group 1 allo ys, was predicted to decrease the liquidus 23 C and decrease the solidus 31 C. Increasing -former content was predicted to result in a 2nd order relationship for the solvus temperature, which in turn, was predicted to incr ease and subsequently decrease the solvus. With a 1 at% -former addition between Model E and Model B, the solvus temperature was predicted to incr ease 34 C. An additional 1 at% -former addition between Model B and Model N, was then predicted to decrease the solvus temperature 14 C. With the increasing -former content, the calculated rate at which the solidus decreased was greater than the rate at whic h the liquidus decreased,

PAGE 126

108 resulting in an increase in the melti ng range. With a 2 at% increase in -former content in Group 2 alloys, the melting range was predicted to increase 8 C. Phase Transformation Effects 3.02 wt % Re 1200 1240 1280 1320 1360 1400 1414.51515.516 gamma prime (at%)Temperature (C) Solidus Liquidus Y' solvus Phase Transformation Effects 2.28 wt % Re 1200 1240 1280 1320 1360 1400 1414.51515.516 gamma prime (at%)Temperature (C) Solidus Liquidus Y' solvus Figure 4-47. Predicted -former variation effe cts on phase transformation temperatures for (a) Gr oup 1 and (b) Group 2 variants The 2nd order relationship for the solvus temperature, was predicted to result in considerably larger heat treatment windows at low -former amounts, as compared to heat treatment windows fo r alloys with approximately 15 at% in former content. Calculations for Group 1 alloys predicted heat treatment windows of 100 C, 47 C, and 49 C for alloys with 14 at% (Model D), 15 at% (baseline Model A), and 16 at% (Model P) in -former content. (a) (b)

PAGE 127

109 Elemental Segregation Predicted -former variation effects on segregation are seen in Figure 4-48. (a) Partitioning effect 3.02 wt% Re0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1414.51515.516 gamma prime(at %)K calc Ni Ta Al Ti (b) Partitioning effect 2.28 wt% Re0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1414.51515.516 gamma prime(at %)K calc Ni Ta Al Ti Figure 4-48. Predicted -former variation effects on elemental segregation for (a) Group 1 and (b) Group 2 variants For all -former variants considered; Ni, W, Co, and Re were predicted to partition to the dendrite core. Re was the most heavily segregated element, followed by W, Co, and Ni. Negligible -former effects were predicted for Ni, Co, and W. Re segregation increased in Gr oup 1 and Group 2 alloys with increasing Partitioning effect 3.02 wt% Re 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1414.51515.516 gamma prime (at%)K calc Cr Co W Re Partitioning effect 2.28 wt% Re 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1414.51515.516 gamma prime (at%)K calc Cr Co W Re Model A Alloy Model A Alloy

PAGE 128

110 -former content. With a -former increase of 2 at% between Model D and Model O in Group 1 alloys, kRe,calc, increased 4%. Cr, Ta, Ti, and Al were predicted to s egregate to the interdendritic region. Ta was the most segregated element, fo llowed by Ti, Al, and then Cr. Ti segregation decreased in Group 1 and Group 2 alloys with increasing -former content. With a -former increase of 2 at% bet ween Model D and Model P in Group 1 alloys, kTi,calc, increased 4%. Negligible elem ental effects were predicted for Ta or Cr. The influence of -former variations on Al segregation could not be clearly established from the given information. Al/Ti Ratio variation effects Two separate variant groups were us ed to investigate Al/Ti effects on material properties: Group 1: Al/Ti ratios of 3.10 and 1.88 (wt%/wt%) in alloys Model O and P, respectively, were used as a comparison to the 6.2 Al/Ti ratio of the baseline Model A alloy. All three alloys contained a constant Re cont ent of 3.02 wt% Re (1 at% Re) and a former content of 15 to 16 at%. Group 2: Al/Ti ratio variations from 4. 11 to1.88 (wt%/wt%) in Model L (4.11), Model M (3.08), and Model N (1.88) were co mpared. All three alloys contained a constant Re content of 2. 28 wt% Re (0.75 at%) and a -former content of 15.5 16 at%. Microstructural Stability Thermodynamic calculations for all Al /Ti variants predicted that equilibrium phases below 1000 C include , , and M23C6 carbides.

PAGE 129

111 In Group 1, no clear relationship between Al/Ti ratio and the TCP phase amounts were established; however, in Gr oup 2, the amount of phases were predicted to decrease 1.6 wt% and 1.1%, respectively, with an Al/Ti increase of 2.23 at 600 C. TCP phase amounts predi cted at 600 C with respect to Al/Ti content are seen in Figure 4-49 (a) for Group 1and (b) for Group 2 Al/Ti variants. (a) Stability Effects -2.28 wt% Re at 600C0 2 4 6 8 10 121.522.533.544.5 Al/Ti (wt%/wt%)wt % Sigma Mu Figure 4-49. Predicted elemental vari ation effects on the amount of TCP equilibrium phases for the (a) Gro up 1 and (b) Group 2 Al/Ti variants, with respect to Al/Ti ratio Phase Transformation Temperatures Thermodynamic equilibrium calculations predicted the solidification path, as seen below, for all Al/Ti variants. Stability Effects -3.02 wt% Re at 600C 2 4 6 8 10 121.52.53.54.55.5 Al/Ti (wt%/wt%)wt % Sigma Mu Model A Alloy (b)

PAGE 130

112 L L + L + + MC + + MC + + M23C6 + + + M23C6 Phase Transformation Effects --3.02 wt%Rey = 4x + 1271 y = 3x + 1331 y = 4x + 12251200 1220 1240 1260 1280 1300 1320 1340 1360 1.52.53.54.55.56.5Al/Ti (wt%/wt%)Temperature (C) Solidus Liquidus Y' solvus (b) Figure 4-50. Predicted elemental variat ion effects on phase transformation temperatures for (a) Group 1 and (b ) Group 2 variants with respect to Al/Ti ratio Figure 4-50 (a) for Group 1 and (b) for Group 2, depicts predicted Al/Ti variation effects on the liquidus, solidus, and solvus temperatures. Increasing Al/Ti ratios were predicted to linearly increase the liquidus, solidus, and solvus. The liquidus, solidus and solvus were predicted to linearly increase 12 C, 15 C, and 14 C with an Al/Ti ratio increas e of 2.23 in Group 2 alloys. The predicted linear increases in the liquidus and solvus with increasing Al/Ti ratio, (a) Model A Alloy Phase Transformation Effects --2.28 wt%Rey = 6.7x + 1264 y = 5.6x + 1326 y = 6.2x + 12201200 1220 1240 1260 1280 1300 1320 1340 1360 1.522.533.544.5 Al/Ti (wt%/wt%)Temperature (C) Solidus Liquidus Y' solvus

PAGE 131

113 resulted in a 9 C decrease of the heat treatment window for Group 1 variants, a trend that was not obser ved in Group 2 alloys. Elemental Segregation Figure 4-51 (a) and (b) shows the Al /Ti variation effects on elemental segregation for the (a) Group 1 and (b) Group 2 variants, respectively. (a) Partitioning effect 3.02 wt% Re0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1357 Al/Ti (wt%/wt%)K calc Ni Ta Al Ti (b) Partitioning effect 2.28 wt% Re0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.52.53.54.5Al/Ti (wt%/wt%)K calc Ni Ta Al Ti Figure 4-51 Predicted element al variation effects on elemental segregation for the (a) Group 1 and (b) Group 2 Al/Ti va riants with respect to Al/Ti ratio Elements calculated as partitioning to t he dendrite core were Cr, Ni, W, Co, and Re. Re segregated to the greatest ext ent, followed by Co, W, Ni, and then Partitioning effect 3.02 wt% Re 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1357 Al/Ti (wt%/wt%)K calc Cr Co W Re Partitioning effect 2.28 wt%Re 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1.52.53.54.5 Al/Ti (wt%/wt%)K calc Cr Co W Re Model A Alloy Model A Alloy

PAGE 132

114 Cr. Re segregation was predicted to incr ease linearly with Al/T i increases (a 5% increase in kRe,calc with a 4.32 Al/Ti increase in Group 1 variants, and a 4% increase in kRe,calc with a 2.23 Al/Ti increase in Group 2 variants). A linear decrease in W segregation was predicted with increased Al/Ti ratios for both variant groups considered (a 3.6% decrease in kW,calc with a 4.32 Al/Ti increase in Group 1 variants, and a 3.6% decrease in kW,calc with a 2.23 Al/Ti increase in Group 2 variants). Ta, Ti, and Al were predicted to segregate to the interdendritic region, with k calc values less than one, for all variants considered. Ta was calculated as the strongest segregating element, followed by Ti and Al. Negligible Al/Ti variation effects were predicted for Ta Ti, and Al segregation. No other Al/Ti ratio effects were observed. Elemental Variation Trend Summary for Phase II The general theoretical effects caused by compositional changes to the baseline Model A alloy composition in Pha se II are summarized in Table 4-5. Overall calculated elemental variation e ffects on microstructural stability, phase transformation temperatures, and element al segregation are shown for the characteristic elemental ranges analyzed in this study. Phase III Experimental Validation Final modifications to the baseline Model A alloy, based on Phase II material property trends, produced five P hase III compositions (Table 4-6). The results for Phase III are given below and are grouped by individual material properties (microstructural st ability, phase transformation temperatures, and segregational behavior). Results from laboratory testing and JMatPro

PAGE 133

115 Table 4-5 Predicted elemental variation e ffects on microstructural stability, phase transformation temperatures and elemental segregation. indicates an increasing effect, indicates a decreasing effect, indicates a limited or negligible effect, indi cates that the influence of the elemental variation could not be clearly established. Characteristic Elemental Variations Elemental Segregation HT Base composition: Model A alloy wt%wt% solvusWindow Cr (Al and Ti) Variations Re Increasing Cr from 11 wt% to 12 wt% Cr Ni, Co, W, Al, Cr, Ta GROUP 1 Increasing Cr from 11 wt% to 12 wt% Cr Accompanying Compositional Modifications: 12 wt% Cr alloy has a 1 at% Al reduction and 1at% Ti addition Al/Ti ratio variations Re Increasing Al/Ti ratio from 1.88 to 6.2 Ta, Co, Al, Cr W GROUP 1 Increasing Al/Ti ratio from 1.88 to 6.2 Accompanying Compositional Modifications: All alloys contain 1 at% Re and 15 to 16 at% former content. GROUP 2 Increasing Al/Ti ratio from 1.88 to 4.11 Accompanying Compositional Modifications: All alloys contain 0.75 at% Re and 15.5 -16 at% -former content Re variations Cr Increasing Re from 0 wt% to 3.02 wt% Re W Cr, Co, Ni, W, Re GROUP 1 Increasing Re contents of 0, 1.5, and 3.02 wt% Re Accompanying Compositional Modifications: All alloys contain 6.2 Al/Ti Ratio and 15 at% -former content. GROUP 2 Increasing Re contents of 0, 1.5, 2.28, and 3.02 wt% Re Accompanying Compositional Modifications: All alloys contain 2.54 Al/Ti ratio and 14 at% -former content. GROUP 3 Increasing Re contents of 0, 1.5, 2.28, and 3.02 wt% Re Accompanying Compositional Modifications: All alloys contain 1.69 Al/Ti ratio of and 14 at% -former content. (Ta, Al,Ti) Former Variations to 15 at% to 15 at% Re -former variations from 14 to 16 at% after 15 at% after 15 at% Ni, Co, W, Ta, Cr Al GROUP 1 -former variations from 14 to 16 at% Accompanying Compositional Modifications: 14 -former:(2.5 at% Ta, 9.5 at% Al, 2 at% Ti) 14 -former: (2 at% Ta, 9 at% Al, 3 at% Ti) 15 -former: (3 at% Ta, 11 at% Al) 16 -former: (3 at% Ta, 11 at% Al, 2 at% Ti) 16 -former:(3 at% Ta, 10 at% Al, 3 at% Ti) GROUP 2 -former variations from 14 to 16 at% Accompanying Compositional Modifications: 14 -former:(2.5 at% Ta, 9.5 at% Al, 2 at% Ti) 14 -former: (2 at% Ta, 9 at% Al, 3 at% Ti) 15 -former:(3 at% Ta, 11 at% Al, 1 at% Ti) 16 -former: (3 at% Ta, 11 at% Al, 2 at% Ti) 16 -former:(3 at% Ta, 10 at% Al, 3 at% Ti) Overall Segregation StabilityPhase Transformation Temps. (L)(S)MR

PAGE 134

116 evaluations were presented separately and then together for comparison. Subsequently, elemental variati on effects were then discussed. Table 4-6. Phase III alloy compositions and variation groups in wt% and at%. NiAlCoCrHfReTaWTiCAl/Ti Y' at% Alloy 1 wt %57.544.3711110.108.85.961.170.053.7214.5 at %58.711011.5130.030321.50.26 Alloy 2 wt %59.454.0411110.1065.962.390.051.6914 at %59.21911.5130.0302230.26 NiAlCoCrHfReTaWTiC Alloy 3 wt %58.784.3711120.106.15.961.560.052.8914 at %58.211011.5140.0302220.26 NiAlCoCrHfReTaWTiC Alloy 4 wt %55.974.1711110.13.027.45.961.560.052.5414 at %58.219.511.5130.0312.5220.26 Alloy 5 wt %58.994.1711110.107.45.961.560.052.5414 at %59.219.511.5130.0302.5220.26 Al/Ti variations Increased CrRe Variations Microstructural Stability Experimetal results: micros tructural characterization Material microstructure observations for the as-cast and solution heat treated specimens (Alloy 1, Alloy 2, Alloy 3, Alloy 4, and Alloy 5) were studied in detail and are reported below. Alloy 1: Al/Ti Ratio 3.7 wt%/wt% Material microstructure and observations for Alloy 1 in the as-cast condition are seen in Figure 4-52. Figure 4-52 (a) represents the typical as-cast dendritic microstructure expected of a nickel base superalloy. The matrix with precipitate microstructure is seen in Figure 4-52 (b). Irregular and coarse precipitates appeared throughout the material microstructure (Figure 4-52 (c)) and in + eutectic regions (Figure 4-52 (d)). Primary MC carbides (Figures 452 (e) and (f)) appeared predomi nately in the blocky, cubic morphology phase shown in Figure 4-52 (f). Qualitative analysis using EDS spectra and X-ray dot

PAGE 135

117 mapping indicated that the mo re prevalent carbides were of the (Ti,Ta)C form. The other forms of primary carbides obs erved contained subs tantial levels of both Ta and W (Figure 4-52 (e)). (a) (b) (c) (d) (e) (f) Figure 4-52. Alloy 1 in the as-cast condi tion showing (a) basic dendritic structure (b) + microstructure (c) coarsening (d) + eutectic regions, and (e) Ta, W rich and (f) Ti, Ta rich primary carbides

PAGE 136

118 (a) (b) (c) Figure 4-53. Microstructural characte ristics for Alloy 1 in the heat treated condition showing (a) a fully homogenized structure with (b) + microstructure and (c) primary (Ti,Ta)C carbides Material microstructure and observations for Alloy 1 in the heat treated condition are seen in Figure 4-53. Afte r heat treatment at 1250 C for 32 hrs, Alloy 1 was fully homogenized with no residual dendritic structure, as observed in Figure 4-53 (a). The + material microstructure is seen in Figure 4-53 (b). Coarse precipitates and eutectic regions obs erved in the as-cast state were not observed in the heat treated condition Primary MC carbides appeared only in the blocky morphology observed in Fi gure 4-53 (c). The (Ti,Ta)C primary carbides were rich in Ti and Ta concent rations, confirmed through qualitative EDS spectra and X-ray mapping. The absence of Ta and W rich primary

PAGE 137

119 carbides, and coarse precipitates, in the solution heat treated material indicate their dissolution into the material matrix. (a) (b) (c) (d) (e) (f) Figure 4-54. Alloy 2 in the as-cast conditi on showing (a) basic dendritic structure, (b) + microstructure (c) + eutectic regions, and two forms of primary carbides represented by a (d) SE image of one, and (e) BSE and (e) SE images of another.

PAGE 138

120 (a) (b) (c) (d) Figure 4-55. Alloy 2 in the heat tr eated condition showing (a) a fully homogenized structure with (b, c) + microstructure and (d) primary (Ti,Ta)C carbides Alloy 2: Al/Ti Ratio 1.69 wt%/wt% Material microstructure and observations for Alloy 2 in the as-cast condition are seen in Figure 4-54. A clear dendritic structure was observed for Alloy 2 in the as-cast condition (Figure 4-54 (a)), with a + material microstructure (Figure 4-54 (b)). Extensive + eutectic regions (Figure 4-54 (c)) were observed in the as-cast microstructure. Qualitative EDS s pectra and X-ray dot map analysis were used to identify carbide composition. Figure 4-54 (d) depicts one of the (Ti,Ta)C primary carbides f ound in the materials microstructure. Primary MC carbides in other forms we re observed in limited amounts (Figures 4-54 (e) and (f)). The bright (white) c ontrast of the MC carbide seen in BSE

PAGE 139

121 imaging (Figure 4-54 (e)), indicates its hi gh atomic number cont ent. Qualitative EDS spectra and X-ray mapping identified hi gh concentrations of Ti and W in the MC carbide. Material microstructure and observati ons for Alloy 2 in the heat treated condition are seen in Figure 4-55. After the solutionizi ng heat treatment; Alloy 2 did not exhibit of any remnants of the dendrit ic structure (Figure 4-55 (a) and (c)). Blocky MC carbides were present throughout the material microstructure (Figure 4-55 (d)). Qualitative EDS spectra and X-ray mapping were used to confirm high Ti and Ta concentrations in these (Ti,Ta )C primary carbides. No other carbide forms or eutectic regions we re observed. Alloy 3: Cr Content 12 wt% Material microstructure and observations for Alloy 3 in the as-cast condition are seen in Figure 4-56. Dendritic stru cture (Figure 4-56 (a)) for Alloy 3 was observed in the as-cast condition. The + material microstructure is seen in Figure 4-56 (b) and (c). Coarse precipitates, shown in Figure 4-56 (d), were seen in expansive + eutectic regions. Large blocky shaped MC carbides as illustrated in Figure 4-56 (f) were most common in the as -cast structure. X-ray mapping (Figure 4-56 (h)) and EDS spectra, cl early identified Ti and Ta as main compositional components. Other primary carbides in Alloy 3 (Figure 4-56 (e)) were shown to be high in atomic num ber concentration when analyzed using BSE imaging. X-ray maps indicated conc entrated amounts of W, Ti, and Ta in such carbides.

PAGE 140

122 Material microstructure and observations for Alloy 3 in the heat treated condition are seen in Figure 4-57.Heat treatment of Alloy 3 resulted in an equiaxed and fully homogenized structure, observed in Figure 4-57 (a). The + material microstructure is seen in Figure 4-57 (b). (a) (b) (c) (d) (e) (f)

PAGE 141

123 (g) (h) Figure 4-56. Alloy 3 microstructure in the as-cast condition showing (a) basic dendritic structure (b, c)) + microstructure, (d) + eutectic regions, and (e, f) primary carbides with their respective (g, h) X-ray elemental maps. (a) (b) (c) Figure 4-57. Alloy 3 microstructure in the heat treated condition showing (a) the fully homogenized structure with (b) + microstructure and (c) primary (Ti,Ta)C carbides

PAGE 142

124 Primary MC carbides (Figure 4-57 (c)), observed throughout the microstructure, were of the (Ti,Ta)C type as confirmed through the use of EDS analysis and X-ray elemental mapping. No further observations in the material microstructure indicated the dissolution of the + eutectic regions and other types of MC carbides into the metal matrix. Alloy 4: Re Content 3.02 wt% Material microstructure and observations for Alloy 4 in the as-cast condition are seen in Figure 4-58. The optical micrograph (Figures 4-58 (a)) and SE image (Figures 4-58 (b)) for Alloy 4, identify a clear dendritic structure in the as-cast condition. Marked differences in the material microstruc ture, between the dendrite core and interdendritic microstructure are observed in Figures 4-58 (c). The as-cast material also contained extensive + eutectic nodules (Figure 458 (d)). Primary MC carbides, such as those seen in Figure 4-58 (g) and (h), predominated in the interdendritic regi ons. Qualitative analysis using EDS spectra and X-ray dot maps indicated that the carbides in Fi gure 4-58 (g) were predominately Ti and Ta, while the others co ntain high levels of Ta, Ti, alongside small amounts of W. Material microstructure and observations for Alloy 4 in the heat treated condition are seen in Figure 4-59. Opti cal images in Figures 4-59 (a) and (b) show homogenized and partia lly homogenized regions of Alloy 4 after solution heat treatment. The presence of dendritic structure in the heat treated sample indicates that incomplete homogenization had occurred during the solution heat treatment, resulting in some degree of re sidual segregation. Blocky primary MC carbides such as the one seen in Fi gure 4-59 (d) prevailed throughout the

PAGE 143

125 specimen. EDS spectra and X-ray mapping shown in Figure 4-59 (e) confirmed high Ti, Ta, and C contents in the carbide. (a) (b) (c) (d) (e) (f) Figure 4-58 Alloy 4 microstructure in the as-cast condition showing (a, b) basic dendritic structure, (c) + microstructure, (d) + eutectic regions, and (e) Ti, Ta rich and ( f) Ti, Ta, and W rich primary carbides Dendritic Core Interdendritic Region Dendritic Core Interdendritic Region

PAGE 144

126 (a) (b) (c) (d) (e) Figure 4-59. Material microstructure fo r Alloy 4 in the h eat treated condition showing (a) homogenized regions (b ) partially homogenized regions (c) + microstructure, (d) primary MC carbides, and the carbidess respective (e) elemental X-ray dot map.

PAGE 145

127 (a) (b) (c) (d) (e) (f) Figure 4-60. Alloy 5 in the as-cast condi tion showing (a) basic dendritic structure (b) + microstructure (c) + eutectic regions, and two forms of primary carbides represented by a (d) SE image of one, and the (e) SE and (f) BSE images of another.

PAGE 146

128 (a) (b) (c) Figure 4-61. Alloy 5 material microstruc ture in the heat treated condition showing (a) a fully homogenized structure with (b) a + microstructure and (c) typical Ti and Ta ri ch primary carbides Alloy 5: Re Content 0 wt% Material microstructure and observations for Alloy 5 in the as-cast condition are seen in Figure 4-60. Optical imaging of Alloy 5 in the as-cast condition (Figure 4-60 (a)) shows a dendritic microstructure. Fine precipitates (Figure 460 (b)) were present throughout the specimens matrix. Figure 4-60 (c) exemplifies eutectic + regions made up of irregular and coarse precipitates. Primary MC carbides of many forms were observed, two of which are shown in Figures 4-60 (d) and (e). These carbides appeared predominately to be of the (Ti,Ta)C type and blocky in morphology (Figure 4-60 (d)). Other

PAGE 147

129 forms of primary MC carbides (Fi gure 4-60 (e)) analyzed with BSE imaging (Figure 4-60 (f)), were shown to contai n high atomic number elements in their makeup. Qualitative analysis using EDS spectra and X-ray dot mapping revealed high Ti, Ta, and W contents in these carbi des, while other forms of MC carbides showed high levels of Ti, and W. Material microstructure and observations for Alloy 5 in the heat treated condition are seen in Figure 4-61. Solution heat treatment of Alloy 5 resulted in a fully homogenized and equiaxed micros tructure with no apparent residual dendritic structure as seen in Figure 461 (a) and (b). Primary MC carbides appeared in the blocky form, confirmed as be ing rich in Ti and Ta by qualitative EDS spectra and X-ray mapping. The disso lution of primary carbides containing W concentrations, along with + eutectic zones, occurred after solution heat treatment of Alloy 5. Computational results: JMatPro equilibrium phase predictions Results revealed that below 1000 C, the phases predicted for all Phase III alloys, included , and along with M23C6 carbides that were predicted to degenerate from original MC metal carb ides. Equilibrium phases and their weight fractions, predicted for Phase III variant alloys, are presented below. Alloy 1: Al/Ti Ratio 3.7 wt%/wt% Alloy 1 (3.72 Al/Ti) at 600 C was predicted to contain approximately 65 wt% in the precipitate and approx imately 28 wt% in matrix. TCP phases, (approximately 4 wt%) and (approximately 2 wt%), were also predicted to be in equilibrium at 600 C. No TCP ph ases were expected at 900 C.

PAGE 148

130 Alloy 2: Al/Ti Ratio 1.69 wt%/wt% Alloy 2 (1.69 Al/Ti) was predicted to contain large quantities of the precipitate (approxim ately 62 wt%) and the matrix (approxim ately 33 wt%) at 600 C. Small am ounts of the phase (approximately 3 wt%) and the phase (approximately 1.5 wt%) were also predi cted to be in equilibrium at 600 C. No TCP phases were expected at 900 C. Alloy 3: Cr Content 12 wt% Alloy 3 (12 wt% Cr) was predicted to contain approximately 64 wt% in the precipitate, approxim ately 33 wt% in the matrix, approximately 6 wt% in the TCP phase, and approximately 0.7 wt% in the TCP phase at 600 C. No TCP phases were expected at 900 C. Alloy 4: Re Content 3.02 wt% Alloy 4 (3.02 wt% Re) at 600 C was predicted to contain approximately 60 wt% in the precipitate and approx imately 28 wt% in matrix. A considerable amount of TCP phases, (approximately 7 wt%) and (approximately 4 wt%), were also predicted to be in equilibriu m at 600 C. Of th e TCP phases, only is predicted at 900 C (approximately 6 wt%). Alloy 5: Re Content 0 wt% Alloy 5 (11 wt% Cr) at 600 C, was predicted to contain approximately 63 wt% in the precipitate and approx imately 32 wt% in matrix. A limited amount of TCP phases, (approximately 2.6 wt %) and (approximately 1.3 wt%), were also predicted to be in equilibrium at 600 C.

PAGE 149

131 Experimental to com putational comparisons Laboratory testing revealed that all Phase III alloys produced the , and MC metal carbide phases before and a fter solution heat treating. Thermodynamic calculati ons predicted that the and phases, would be present all Phase III alloys at equilibrium conditions. Thermodynamic equilibrium calculations for all Phas e III alloys predicted that degenerative M23C6 carbides would form from primary MC carbides at M23C6 solvus temperatures of 850 C for Alloy 1, 860 C for Alloy 2, 900 C for Alloy 3, 960C for Alloy 4, and 840 C for Alloy5. Even though they were not observed in the as-cast and solution heat treated specimens, and TCP phases were also predicted to be in the equilibrium structures of all Phase III alloys at temperatures below 600 C. Elemental variation effects Al/Ti Ratio (and Ta) variation effects Al/Ti ratio (and Ta) variation effects on phase stability were investigated using Alloy 1 and Alloy2. Alloy 1 (with a Al /Ti ratio of 3.72) containing 2.8 wt% (1 at%) Ta addition in comparison to Allo y 2 (with a Al/Ti ratio of 1.69). Stability Effect at 600 C [Calculated] y = 0.8902x + 1.0933 y = 0.0926x + 1.3571 0.6 1.1 1.6 2.1 2.6 3.1 3.6 4.1 4.61.522.533.54Al/Ti (wt%/wt%)wt% Sigma Mu Figure 4-62. Predicted Al/Ti ratio (and Ta) variation effects on TCP equilibrium phase amounts with respect to Al/Ti ratio.

PAGE 150

132 Figure 4-62 shows the in crease in the amounts of and TCP phases predicted at 600 C, with respect to in creasing Al/Ti ratio (and Ta addition). Within the Al/Ti range evaluated in this st udy, an increase of 1 wt% in the amount of phase was predicted at 600 C. Negligible Al/Ti ratio effects were observed for phase amounts Chromium (with Al and Ta) variation effects Cr effects on material stability were investigated by comparing Alloy 3 to Alloy 5. Alloy 3 (12 wt% Cr) contains a 0.26 wt% (0.5 at%) Al addition and a 1.32 wt% (0.5 at%) Ta reduction in co mparison to Alloy 5 (11 wt% Cr). Cr additions (with an Al addition and Ta reduction) were predicted to increase the amount of phase, while decreasing the amount of phase at 600 C (Figure 4-63). An increas e of 3 wt% in the amount of phase and a smaller 0.6 wt% decrease in the amount of phase, were predicted with a 1 wt% Cr increase (with a 0.5 at% Al addition and a 0.5 at% Ta reduction) at 600 C. Stability Effect at 600 C [Calculated] y = 3.209x 32.694 y = -0.5952x + 7.8945 0 1 2 3 4 5 610.7510.9511.1511.3511.5511.7511.9512.15Cr (wt%)wt% Sigma Mu Figure 4-63. Predicted Cr ( with Al and Ta) variation effects on TCP equilibrium phase amounts with respect to Cr content.

PAGE 151

133 Rhenium variation effects Re variation effects on material stability were evaluated using Alloy 4 and Alloy 5, with 3.02 wt% (1 at%) Re and 0 wt% (0 at%) Re, respectively. Figure 4-64 shows the strong effect Re has on the amount of TCP phases, predicted at 600 C. The amounts of both and phases were expected to increase with increasing Re content, incr easing 4 wt% and 2 wt%, respectively with a 3.02 wt% (1 at%) Re addition. Stability Effect at 600 C [Calculated] y = 1.3265x + 2.6042 y = 0.8997x + 1.3471 0.4 1.4 2.4 3.4 4.4 5.4 6.4 7.400.511.522.53Re (wt%)wt% Sigma Mu Figure 4-64. Predicted Re variation e ffects on the amount of TCP equilibrium phases Phase Transformation Temperatures Experimental results: DTA results Alloy 1: Al/Ti Ratio 3.7 wt%/wt% The DTA curve for Alloy 1 in the as-cast condition is seen in Figure 4-65. In the heating curve, a strong inflection in t he vicinity of the liquidus temperature, defined the liquidus to be 1362 C, while 132 1 C was identified as the solidus temperature.

PAGE 152

134 Figure 4-65. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 1 in the as-cast condition Figure 4-66. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 1 in the heat treated condition

PAGE 153

135 Figure 4-67. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 2 in the as-cast condition Figure 4-68. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 2 in the heat treated condition

PAGE 154

136 Figure 4-69. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 3 in the as-cast condition Figure 4-70. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 3 in the heat treated condition

PAGE 155

137 Figure 4-71. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 4 in the as-cast condition Figure 4-72. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 4 in the heat treated condition

PAGE 156

138 Figure 4-73. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 5 in the as-cast condition Figure 4-74. DTA temperature difference ( T) vs. specimen temperature curves for experimental Alloy 5 in the heat treated condition

PAGE 157

139 The DTA curve for Alloy 1 in the heat treated condition is seen in Figure 466. For Alloy 1 in the heat treated condition, the heating curve showed an inflection point at 1211 C, representative of the solvus. In the heating curve, a relatively flat region in the main endotherm ended at the solidus temperature 1344 C, followed by a large inflection i dentifying the liquidus temperature to be 1388 C. Alloy 2: Al/Ti Ratio 1.69 wt%/wt% The DTA curve for Alloy 2 in the as-cast condition is seen in Figure 4-67. The DTA heating curve for Alloy 2 in the as-cast condition identified the solidus at 1314 C, and the liquidus at 1364 C. The DTA curve for Alloy 2 in the heat treated condition is seen in Figure 468. The liquidus, solidus, and solvus temperatures were identified at 1388 C, 1341 C, and 1205 C, in the heating curve. Alloy 3: Cr Content 12 wt% The DTA heating curve for Alloy 3 in the as cast condition is seen in Figure 4-69. In the as-cast condi tion, the heating curve for A lloy 3 shows and inflection to the right of the main endotherm, indi cating that the solidus occurrs at 1318C The liquidus was identified to occur at 1365 C. The DTA curve for Alloy 3 in the heat treated condition is seen in Figure 470. For solution heat treated Alloy 3, t he DTA heating curve identified the solidus at 1343 C, and the solvus at 1192 C. A strong in flection to the right of the main endotherm in the heating curve repres ents the liquidus temperature, which was identified at 1387 C.

PAGE 158

140 Alloy 4: Re Content 3.02 wt% The DTA curve for Alloy 4 in the as cast condition is seen in Figure 4-71. The DTA heating curve for Alloy 4, in t he as-cast condition, identified the solidus at 1326 C, and the liquidus at 1377 C. The DTA curve for Alloy 4 in the heat treated condition is seen in Figure 472. The heating curve for Alloy 4 in t he heat treated condition, identified the solvus, the solidus, and the liquidus to be 1191 C, 1354 C, and 1402 C, respectively. Alloy 5: Re Content 0 wt% The DTA curve for Alloy 5 in the as-cast condition is shown in Figure 4-73. Inflections in the heating curve identified the solidus at 1335 C, and the liquidus at 1376 C. The DTA curve for Alloy 5 in the heat tr eaated condition is shown in Figure 4-74. The heating curve for Alloy 5, in the heat treated condi tion, identified the solvus at 1175 C, the solidus at 1359 C, and the liquidus at 1400 C. Computational results: JMatPro phase transformation temperatures According to JMatPro thermodyna mic equilibrium calculations, the solidification path suggest ed for Alloys 1,2, 3, and 5 is shown below. L L + L + + MC + + MC + + M23C6 + + + + M23C6 The predicted solidification path for Re-containing Alloy 4 is seen below. L L + L + + MC + + MC + + + MC + + + + M23C6

PAGE 159

141 The calculated liquidus, MC solvus, solidus, solvus, solvus, and solvus temperatures, melting ranges, and heat treatment windows for the Phase III alloys are seen in Table 4-7. Experimental to com putational comparisons Phase transformation temperatures, melting ranges, and heat treatment windows from JMatPro thermodynamic equ ilibrium calculations and DTA testing of the solutionized Phase III allo ys, are listed below in Table 4-8. Table 4-7. Predicted phase transforma tion temperatures (C) and ranges for Phase III alloys Cr HighLowHighHighLow ModelModel A A lloy 1 A lloy 2 A lloy 3Alloy 4Alloy 5 Calc.Calc.Calc.Calc.Calc.Calc.Calc. Liquidus 1364135113511357135713581357 Solidus 1297129312971309130913061307 MC S o l vus 1338 1335 1338 1337 1340 1338 so l vus 1282 1246 1238 1203 1207 1208 1216 S igma S olvus 1196 7196717431175671 Mu S olvus 1208887764717701849709 Melting Range 67585547485250 Heat Treating Window 1547591071029891 Baseline A l/TiRe Table 4-8. Predicted and experimental phase transformation temperatures (C) and ranges, for Phase III alloys ModelModel A Calc.Calc.Calc.DTACalc.DTACalc.DTACalc.DTACalc.DTA Liquidus 136413511351138813571388135713871358140213571400 Solidus 129712931297134413091341130913431306135413071359 solvus 128212461238121112031205120711921208119112161175 Melting Range 675855444747484452485041 Heat Treating W 1547591331071361021519816391184 Baseline High Alloy 4Alloy 5 Low A lloy 1 A lloy 2 High A lloy 3 High Al/TiCr Low Re Table 4-9 presents the differences and modeling errors between the predicted and tested values of the liquidus, solidus, and solvus transformation temperatures, the melting ranges, and the heat treatment windows for the alloys considered in Phase III.

PAGE 160

142 Even though there were some discr epancies between the experimental and calculated liquidus, solidus, and solvus phase transformation temperatures, modeling errors remained below 4 %. Desp ite the seemingly large errors of the melting range calculations, melting r ange differences between calculated and experimental values ranged between only 0 to 11 C. Table 4-9. Phase III phase transfo rmation temperature deviations from experimental values (calculated experimental) and modeling error ((deviance/experimental)*100) Transformation Temperatures (Predicted vs. Experimental) Alloy CError CError CError CError CError 1 1124%-74-56%-37-3%-47-4%272% 2 00%-29-21%-31-2%-32-2%-20% 3 48%-49-33%-30-2%-34-3%151% 4 49%-65-40%-44-3%-48-4%171% 5 922%-93-51%-43-3%-52-4%414% Solvus Melting RangeHTWLiquidusSolidus Figure 4-75 below, compares th e experimental and predicted phase transformation temperatures for the various experimental alloys. Table 4-9 and Figure 4-75 show, that fo r the most part, JMatPro predicted liquidus and solidus temperatures below experimental values and solvus temperatures above experimental values. Transformation Temperature Comparisons1175 1225 1275 1325 1375 1425 117512251275132513751425Calculated Temperatures (Celcius)Experimental Temperatures (Celcius) Liquidus Solidus Gamma prime solvus Figure 4-75. Comparison between experimental and calculated phase transformation temperatures for Phase III alloys.

PAGE 161

143 Experimental liquidus tem peratures were approximat ely 30 to 40 C higher than the calculated values while experimen tal solidus values were approximately 30 to 50 C higher than the calc ulated values. Experimental solvus temperatures deviated between -2 C to +17 C from JMatPro calculations, with the exception of Alloy 1 and Allo y 5 which deviated 27 C and 41 C, respectively. Alloy 2 demonstrated an ex cellent correlation between calculated and experimental solvus temperature. Figure 4-76 (a) and (b), compare the experimental and calculated melting ranges and heat treatment windows for the five Phase III alloys. (a) (b) Melting Range Comparisons40 42 44 46 48 50 52 54 56 58 60 405060Calculated Melting Range (Celcius)Experimental Melting Range (Celcius) Melting Range 3 1 4 2 5 Heat Treating Window Comparisons50 70 90 110 130 150 170 190 5060708090100110120130140150160Calculated HT Window (Celcius)Experimental HT Window (Celcius) Heat Treating Window 2 1 3 4 5 Figure 4-76. Comparison between exper imental and calculated (a) melting ranges (liquidus solidus) and (b) heat treatment windows (solidus solvus) for Phase III alloys Heat treatment window ca lculations deviated betw een approximately 30 C to 93 C, from the experimentally test ed values. As previously mentioned, melting range calculations deviated between 0 C to 11 C, from the experimentally tested values.

PAGE 162

144 Almost all Phase III alloys demonstra ted an excellent correlation between calculated and experimental melting ranges even though Alloy 5 was predicted to have a melting range 11 C larger than the experimen tal value. The predicted heat treatment window for Alloy 5 was also approximately 74 C lower than the experimentally tested value. Temperature range comparisons for Phase III of alloy development The calculated and experimental me lting ranges and heat treatment windows for the Phase III alloys, the baseline Model composition, the baseline Model A composition, and selective 1st and 2nd generation commercial and experimental alloys, were plott ed for comparison (Figure 4-77). Heat Treating Window vs. Alloy Melting Range 10 30 50 70 90 110 130 150 170 190 283848586878Liquidus to Solidus (Celcius)Solidus to Y' Solvus (Celcius) Calculated Experimental Model Model A 1 3 4 5 3 4 5 2 2 1 PWA 1483 CMSX-10 IN 738 CMSX-4 IN 738 CMSX-4 PWA 1483 CMSX-10 Figure 4-77. Heat trea tment window (solidus solvus) vs. melting range (liquidus solidus) for the baselin e Model composition, the baseline Model A composition, the compositi onal variants in Phase III, and selective 1st and 2nd generation commercial and experimental alloys

PAGE 163

145 Melting ranges from 47 C to 55 C were calculated for the compositional variants in Phase III of alloy development. The experimental melting ranges for all five variants ranged between 41 C to 48. The experimental melting ranges for the Phase III alloys were approx imately 19 C to 26 C lower than the baseline Model alloys predicted melting r ange (67 C), and 10 C to 17 C lower than the baseline Model A alloys predi cted melting range (58 C). Melting ranges for the Phase III variants were comparable to the melting ranges for CMSX-4 and PWA 1483 [11,121]. Heat treatment windows from 91 C to 107 C were calculated for the compositional variants in Phase III. Ex perimental heat treatment windows for all five variants ranged between 133 C to 184. The experimental heat treat ment windows for the Phase III alloys were over 100 C higher than the baseline M odel alloys predicted heat treatment window (15 C), and 86 C to 137 C higher than the baseline Model A alloys predicted heat treatment window (47 C). Experimental heat treatment windows for the Phase III variants greatly sur passed the heat treatment windows for PWA 1483, CMSX-4, and IN 738 [11,121,120]. Elemental variation effects Al/Ti Ratio (with Ta) variation effects Al/Ti ratio (with Ta) variation effe cts on phase transformation temperatures were investigated using Alloy 1 and Alloy2. Alloy 1 (with a Al/Ti ratio of 3.72) contains a 2.8 wt% (1 at%) Ta addition in comparison to Alloy 2 (with a Al/Ti ratio of 1.69). Figure 4-78 below depicts phas e transformation trends for Al/Ti ratio variants with respect to increasing Al/Ti ratio (with a Ta addition).

PAGE 164

146 Calculations predicted a considerable solvus increase of 35 C and a solidus decrease of 12 C, with a 2 (wt% /wt%) Al/Ti ratio increase and a 2.8 wt% (1 at%) Ta addition. Exper imental testing showed that the increased Al/Ti ratio (with a Ta addition), had considerably smaller effects on the solidus and solvus temperatures. The 2 (wt%/wt%) Al/Ti ra tio increase with a 2.8 wt% (1 at%) Ta addition, increased the solvus 6 C but increased the solidus by only 3 C. Figure 4-78. Predicted and experimental pha se transformation trends for Al/Ti ratio variants with respect to Al/Ti ratio. Calculations predicted a melting range increase of 8 C and a large heat treatment window decrease of 48 C, with a 2 (wt%/wt%) Al/Ti ratio increase and a 2.8 wt% (1 at%) Ta addition. Experim ental testing showed that a 2 (wt%/wt%) Al/Ti ratio increase with a 2.8 wt% (1 at%) Ta addition, decreased both the melting range and heat treatment window by only 3 C. Chromium (with Al and Ta) variation effects Cr effects were investigated by compari ng Alloy 3 to Alloy 5. Alloy 3 (with 12 wt% Cr) contains a 0.26 wt% Al addi tion and a 1.32 wt% Ta reduction in comparison to Alloy 5 (11 wt% Cr). Figure 4-79 below depicts phase Phase Transformation Effect [Exp.]y = 1388 y = 3x + 1200 y = 1.5x + 1338 1200 1220 1240 1260 1280 1300 1320 1340 1360 1380 14001.5 2 2.5 3 3.5 4 Al/Ti (wt%/wt%) Solidus Liquidus y' solvus Phase Transformation Effect [Calculated]y = -6.3x + 1320 y = 17.2x + 1173 y = -2.6x + 13611200 1220 1240 1260 1280 1300 1320 1340 1360 1380 1400 1.5 2 2.5 3 3.5 4 Al/Ti (wt%/wt%) Solidus Liquidus y' solvus Temperature (C) Temperature (C)

PAGE 165

147 transformation trends for Cr variants with respect to increasing Cr concentration (with an Al addition and Ta reduction). Calculations predicted a 9 C decrease in the solvus for the increased Cr content (with an Al addition and Ta reduction) analyzed in this study. Experimental testing showed an actual increase in the solvus temperature with increasing Cr content (with an Al addition and Ta r eduction). The Cr increase of 1 wt% Cr (with a 0.26 wt% Al addition and a 1.32 wt% Ta reduction ), resulted in a 17 C increase in solvus. Experimental testing also showed a solidus and liquidus decrease of 16 C and 13 C, res pectively, from the 1 wt% Cr increase (with a 0.26 wt% Al addition and a 1.32 wt% Ta reduction). Figure 4-79. Predicted and experimental phase transformation trends for Cr variants with respect to Cr (wt%) concentration. Increasing Cr content ( with an Al addition and Ta reduction) was predicted to increase the heat treatment window and have a negligible effect on the melting range. In experim ental testing, the 1 wt% Cr increase (with a 0.26 wt% Al addition and a 1.32 wt% Ta reduction) decreased the heat treatment window 33 C, while maintaining a relatively constant melting range. Phase Transformation Effect [Exp.]y = -16x + 1535 y = -13x + 1543 y = 17x + 988 1175 1225 1275 1325 137511 11.2 11.4 11.6 11.8 12 Cr (wt%) Solidus Liquidus y' solvus Temperature (C) Phase Transformation Effect [Calculated]y = -8.7x + 1312 y = 2.2x + 1282 y = -0.25x + 1360 1175 1225 1275 1325 1375 11 11.211.4 11.6 11.812 Cr (wt%) Solidus Liquidus y' solvus Temperature (C)

PAGE 166

148 Rhenium variation effects Re variation effects were evaluated us ing Alloy 4 and Alloy 5, with 3.02 wt% Re and 0 wt% Re, respectively. Figure 480 below depicts Re variation effects on phase transformation temperatures with respect to increasing Re concentration. Phase Transformation Effect [Calculated]y = -2.8x + 1216 y = -0.33x + 1306 y = 0.4x + 1357 1175 1225 1275 1325 1375 1425 0123 Re (wt%)Temperature (C) Solidus Liquidus y' solvus Figure 4-80. Predicted and experimental phase transformation trends for Re variants with respect to Re (wt%) concentration. Re effect calculations between Allo y 5 and Alloy 4, with a 3.02 wt% Re addition, predicted a solvus decrease of 8 C, apar t from negligible effects on the liquidus and solidus temperatures. Ex perimental testing showed that the 3.02 wt% Re addition increased the solvus 16 C and decreased the solidus 5 C. Increasing Re concentrations were predi cted to increase the heat treatment window. An increase in Re concentr ation was predicted to have a negligible effect on the melting range. Experimental testing show ed that the 3.02 wt% Re addition, between Alloy 5 and Alloy 4, dec reased the heat treatment window 21 C, and increased the melting range 7 C. Phase Transformation Effect [Exp]y = 5.3x + 1175 y = -1.65x + 1359 y = 0.7x + 1400 1175 1225 1275 1325 1375 14250 1 2 3 Re (wt%) Solidus Liquidus y' solvus Temperature (C)

PAGE 167

149 Elemental Segregation Experimental results: EMPA/WDS microprobe analysis The experimentally determined kx,exp val ues for the Ni, Ta Ti, Al, Cr, Co, W, and Re elements in Phase III alloy chemistries, are listed in Table 4-10. For all of the experimentally tested Phase III alloys, k calc values greater than one (core tendencies) were observed for Ni, Co, Cr, W, and Re. The core segregation for Re was the most significan t, followed by W, Co, Ni, and then Cr. In general, Ni and Cr showed only slight partitioning tendencies towards the core, with k exp values close to unity. Element s which segregated to the interdendritic region, with k exp values less than one, were Ta, Ti, and Al. The partitioning of Ti was the strongest, followed by Ta and then Al. Table 4-10. Experimental partitioning coe fficient values (kx,exp) for as-cast Phase III alloys k experimentalNiCoC r A lTiTaWRe Alloy 11.021.101.050.880.630.751.28 Alloy 21.021.101.020.800.540.671.49 Alloy 31.031.101.020.790.540.651.34 Alloy 41.001.101.010.730.490.611.231.83 Alloy 51.061.071.020.810.580.771.21 Computational results: JMatPr o solidification predictions Predicted partitioning coefficients (kcalc) for the Ni, Ta, Ti, Al, Cr, Co, W, and Re elements in Phase III alloy chemistr ies are listed below in Table 4-11. Table 4-11. Predicted partitioning coeffici ent values (kx,calc) for as-cast Phase III alloys k calculatedNiCoC r A lTiTaWRe Alloy 11.041.151.030.960.730.431.23 Alloy 21.021.161.020.950.720.421.25 Alloy 31.021.171.020.940.710.401.20 Alloy 41.041.120.981.000.700.411.161.31 Alloy 51.021.151.031.010.730.431.24 For all Phase III alloys, k calc values greater than one (core tendencies) were observed for Ni, Co, W, and Re. T he core segregation for Re was the most

PAGE 168

150 significant, followed by W, Co, and Ni. Elements predicted to segregate to the interdendritic region, with k calc values less than one, were Ta, Ti, and Al. The partitioning of Ta was the strongest, followed by Ti. Calculations predicted only slight partitioning tendencies for Al and Cr, to either the dendrite core or interdendritic region. Experimental to com putational comparisons Table 4-12 presents the differences and the modeling errors between the predicted and experimental elemental partitioning coeffi cients, for the Phase III alloys. Table 4-12. Partitioning coefficient dev iations (calculated experimental) and modeling error ((deviance/experiment al)*100) for as cast Phase III alloys. k mo d e l e d vs k exper i men t a l NiCoCr A lTiTaWRe k Alloy 10.020.05-0.010.080.10-0.32-0.05 k Alloy 20.000.060.000.150.18-0.25-0.24 k Alloy 3-0.010.070.010.150.17-0.25-0.14 k Alloy 40.040.02-0.030.280.22-0.20-0.07-0.52 k Alloy 5-0.040.080.010.200.16-0.340.03 Error Alloy 11%5%-1%9%17%-42%-4% Error Alloy 20%5%0%19%34%-38%-16% Error Alloy 3-1%6%1%19%32%-38%-11% Error Alloy 44%2%-3%38%44%-32%-6%-29% Error Alloy 5-3%8%1%25%27%-45%3% Differences between experimental and pr edicted partitioning coefficients for Al, Ti, Ta, and Re, lead to modeling errors greater than 20%. Differences between kcalc and kexp values ranged from -0.52 for Re to 0.28 for Al. Partitioning coefficient values for Ni and Cr in spec ific, demonstrated an excellent correlation between calculated and experimental va lues. In general, modeling errors calculated for Ni and Cr were less than 4 %. This is seen in Figure 4-81, which compares the experimental and predicted partitioning coefficients for various elements in the Phase III alloys.

PAGE 169

151 Figure 4-81. Comparison between ex perimental and predicted partitioning coefficient values (kexp vs kcal) for Phase III alloys The bar charts in Figure 4-82 serve as an additional visual comparison for the kcalc and kexp values of the elements evaluat ed in this study. The strong segregation tendencies of W and Re to partition to the dendrite core, and the strong segregation tendencies for Ti and Ta to partition the interdendritic areas is evident. The minimal partitioning of Al and Cr is also evident. Calculations produced conservative partitioning coefficient values as compared to experimentally tested coefficien ts, with the excepti on of Ta and, to some extent, Co. Experiment al partitioning coefficient s for W and Re reflected intensified segregation to t he dendrite core as compared to predicted values, with kW,exp and kRe,exp values 7.4% and 40% larger t han calculated values. Calculated kCo,calc and kTa,calc values predicted a higher degree of segregation than what was seen experimentally. Comparisons of Partioning Coefficients0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 0.30.50.70.91.11.31.51.71.9Calculated K ValuesExperimental K Values Ni Co Cr AL Ti Ta W Re (a)

PAGE 170

152 Comparisons of Partioning Coefficients0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 NiCoCrWReElements of IterestPartioning Coefficient (K) 1 2 3 4 5 J Mat Pro Calculated Comparisons of Partioning Coefficients0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 AlTiTaElements of IterestPartioning Coefficient (K) 1 2 3 4 5 J Mat Pro Calculated Figure 4-82. Experimental and predicted partitioning coefficient values for elements typical of the (a) dendrit e core and (b) and interdendritic region Elemental variation effects Al/Ti Ratio (and Ta) variation effects Al/Ti effects on elemental segregation was investigated using Alloy 1 and Alloy2. Alloy 1 (with a Al/Ti ratio of 3. 72) contains a 2.8 wt% (1 at%) Ta addition in comparison to Alloy 2 (with a Al/Ti ra tio of 1.69). Segregation trends for Al/Ti (and Ta) variations are seen with respect to Al/Ti ratio in Figure 4-83. (b) (a)

PAGE 171

153 Calculations predicted that Al/Ti ratio (with Ta) variations would have no effects on elemental segregation, with a negligible decrease in W segregation with increased Al/Ti ratios (and a Ta addition). Experimentally, kW,exp decreased 14% with the 2.2 Al/Ti increase and a 2.8 wt% Ta addition, substantially decreasi ng W segregation. Ex perimental testing also showed that the 2.2 Al/Ti incr ease with a 2.8 wt% Ta addition, increased kAl,exp, kTi,exp and kTa,exp by 10%, 17%, and 12%, respectively. These shifts in experimental coefficient values showed that the Al/Ti increase (and Ta addition) decreased segregation for Al, Ti, and Ta. Figure 4-83. Predicted and exper imental segregation trends for Al/Ti variants with respect to Al/Ti (with Ta) variations Partioning Effect [Calculated]0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.502.503.50 Al/Ti (wt%/wt%)kcalc Ni Ta Al Ti Partioning Effect [Experimental]0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.502.503.50 Al/Ti (wt%/wt%)kexp Ni Ta Al Ti Partitioning Effect [Experimental] 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.502.503.50 Al/Ti (wt%/wt%)k exp Cr Co W Partitioning Effect [Calculated] 1.00 1.10 1.20 1.30 1.40 1.50 1.502.503.50Al/Ti (wt%/wt%) Cr Co W

PAGE 172

154 Chromium (with Al and Ta) variation effects Cr effects were investigated by compari ng Alloy 3 to Alloy 5. Alloy 3 (with 12 wt% Cr) contains a 0.26 wt% Al addi tion and a 1.32 wt% Ta reduction in comparison to Alloy 5 (11 wt% Cr). Segregation trends for Cr (with Al and Ta) variations are seen with respect to Cr concentration in Figure 4-84. Figure 4-84 Predicted and experimental s egregation trends for Cr variants with respect to Cr (wt%) concentration. A 1 wt% Cr addition (with a 0.26 wt% Al addition and a 1.32 wt% Ta reduction), was predicted to decrease W segregation (with a 4% decrease in kW,calc) and increase Ta segregation (with a 7% decrease in kTa,calc). Al was Partioning Effect [Experimental] 0.35 0.45 0.55 0.65 0.75 0.85 0.95 1.05 10.7511.2511.7512.25Cr (wt%)kexp Ni Ta Al Ti Partioning Effect [Calculated]0.35 0.45 0.55 0.65 0.75 0.85 0.95 1.0510.7511.2511.7512.25 Cr (wt%)kcalc Ni Ta Al Ti Partitioning Effect [Experimental] 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35 10.7511.2511.7512.25 Cr (wt%)k exp Cr Co W Partitioning Effect [Calculated] 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35 10.7511.2511.7512.25 Cr (wt%)k calc Cr Co W

PAGE 173

155 predicted to change from segregating to t he dendritic core to the interdendritic region as Cr content increased (with a Al addition and Ta reduction). A 7% decrease in kAl,calc was predicted for the 1wt% Cr addition (with a 0.26 wt% Al addition and 1.32 wt% Ta reduction). Experimentally, the Cr addition (w ith an Al addition and Ta reduction) increased W, Ti, and Ta segrega tion. A 12% increase in kW,exp, a 7% decrease in kTi,exp, and a 17% decrease in kTa,exp, resulted form the 1 wt% Cr addition (with a 0.26 wt% Al addition and a 1.32 wt% Ta reduction). Experimental testing showed negligible Cr (with Al and Ta) vari ation effects on Ni and Co segregation. No other observations were made. Rhenium variation effects Re variation effects were evaluated us ing Alloy 4 and Alloy 5, with 3.02 wt% (1 at%) Re and 0 wt% (0 at%) Re, res pectively. Segregation trends for Re variations are seen in Figure 4-85. Calculations predicted that Re variat ions would have negligible effects on Al, Ta Co segregation. Re variation e ffects were predicted fo r Re, W, Ti, Cr. A 3.02 wt% Re increase was expected to decrease W segregation (a 6.5% decrease in kW,calc) and increase Ti segregation (a 4% decrease in KTi,calc). Cr was expected to change from segregating to the dendritic core to interdendritic region as Re content increased (with a 5% decrease in kCr,calc with a 3.02 wt% Re increase). Experimental results showed an increas e in Ta, Al, and Ti segregation with increasing Re concentration. A 22% decrease in kTa,exp, a 10% decrease in

PAGE 174

156 kAl,exp, and a 15% decrease in kTi,exp, resulted from a 3.02 wt% Re increase. Experimental testing also showed a decr ease in Ni segregation with increasing Re content (a 6% decrease in kNi,exp as a result of a 3.02 wt% Re addition). Figure 4-85. Predicted and experimental s egregation trends for Re variants with respect to Re (w t%) concentration Partioning Effect [Experimental] 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 0123 Re (wt%)kexp Ni Ta Al Ti Partioning Effect [Calculated] 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 0123 Re (wt%)kcalc Ni Ta Al Ti Partioning Effect [Experimental]0.95 1.05 1.15 1.25 1.35 1.45 1.55 1.65 1.75 1.85 0123 Re (wt%)kexp Cr Co W Re Partioning Effect [Calculated] 0.95 1.05 1.15 1.25 1.35 1.45 1.55 1.65 1.75 1.85 0123 Re (wt%)kcalc Cr Co W Re

PAGE 175

157 CHAPTER 5 DISCUSSION The growing demand to increase effi ciency and decrease emissions in advanced gas turbine engines requires the development of higher strength and heat resistant materials [14]. Dram atic improvements in the maximum temperature capabilities of turbines has been achieved through the production of single crystal turbine blades and vanes [ 30,45]. The successful use of singlecrystal alloys in IGT applications is contingent upon overcoming processing problems such as defect formation and main taining microstructural stability once in service [51]. Defect formation is linked to element al segregation in as-cast superalloy microstructures, which results from so lute rejection during the solidification process [16]. Elemental segregation is characterized by concentration gradients between distinct regions throughout the ma terials microstructure [18]. The use of solutionizing heat treatm ents is essential in reducing or eliminating segregatio n from the as-cast material s, but are limited to a temperature range between the solvus and solidus temperatures, referred to as the heat treatment win dow. The inability to fully homogenize a material, due to low or even negative heat treatment wi ndows, results in residual segregation in the material microstructu re. These local enrichments of elements can lead to deleterious phase formation; therefore, maintaining the phase stability of Nibased superalloys at elevated temperatur es is also of primary concern.

PAGE 176

158 The goal of this project was to identif y compositions that could be used for IGT applications. Alloy compositions would exhibit the strength of common second generation superalloys such as CMSX-4, and the hot corrosion resistance and castability of first generat ion superalloys, such as PWA 1483. The number of elemental additions in a Ni-based superalloy and the complex interactions of these additions reveal ed the need to investigate elemental variation effects on microstructural st ability, phase transformation temperatures, and material segregation behavior. In this study, a baseline alloy composition named the Model alloy, was developed from selective experimental (CMSX-11B, CMSX-11C) and industrial (CMSX-4, SC-16, PWA 1484, PWA 1483) first and second generation alloys, including the currently used polycrysta lline and cast baseline alloy for IGT applications (IN738). Iterations of elemental variation e ffect evaluations on microstructural stability, phase transformation temperat ures, and segregation behavior were conducted with the use of the thermodynamic equilibr ium module in the 3.0 JMatPro material program. Elements commonly added to commercial nickelbase superalloys (Co, C, Cr, Mo, W, Al, Ti, Ru, Re, and Ta), and the total atomic concentration (at%) of -forming elements were investig ated in the fist iteration (Phase I) of elemental variation comparis ons. A second iteration (Phase II) of elemental variation comparisons were c onducted to further characterize Al/Ti ratio, Re, Cr, and -former effects on the materi al properties of interest. Calculated elemental variation trends afte r each iteration were used to redefine

PAGE 177

159 the baseline composition. Compositional adju stments were selected to minimize TCP phase formation, achieve a heat tr eatment window of at least 25 C, and minimize elemental segregation during solidification. Five alloys identified for Phase III inco rporated characteristic variations in Al/Ti ratio (with Ta variation), Cr (with Al and Ta variations), and Re content. Phase III alloys samples were prepared, for experimental comparison to predicted properties. A discussion of the predicted and ex perimentally tested elemental and elemental group effects on microstructu ral stability, phase transformation temperatures, and segregation behavior, in a ll phases of this study, is presented below. Microstructural Stability Elemental Variation Effects The effect of TCP phases on alloy pr operties are closely related to TCP quantities. Elemental variation effect s on material stability are seen below. Carbon variation effects C effects on microstructural stability were investigated in Phase I by comparing the Model 3 (0.01 wt% C) Model 2 (0.05 wt% C), Model 1 (0.075 wt% C), and baseline Model (0 wt% C) alloys. Thermodynamic calculations at 900 C, predicted a linear decr ease in the amount of phase with increasing C concentrations (Figure 4-19). Calculations were in agreement with pr evious studies which show that C additions inhibit the precipitation of TCP phases and SRZs; which, in turn, enhances phase stability [50].

PAGE 178

160 Overall, it was shown that JMatPro equilibrium calculations successfully predicted C effects on microstructural stab ility, where microstructural stability increases with C additions. Cobalt variation effects Co effects on microstructural stability were investigated in Phase I by comparing the Model 22 (9 wt% Co), the baseline Model (11.47 wt% Co), and the Model 21 (12 wt% Co) alloys. Thermodynamic calculations at 900 C (Figure 423) predicted a negligible Co effect on phase amounts. In literature, controversy exists r egarding Co effects on microstructural stability. A considerable amount of work has shown that t he nucleation of the phase and other TCP phases is suppressed by increased levels of Co [25,34,50]. Studies conducted by Rae and Reed show that decreased phase formation with Co additions may have a number of possible explanati ons [34]. These explanations include the decrease of the phase thermodynamic stability, or alterations in matrix lattice parameter which suppresses the nucleation of the phase [34]. Conversely, Co additions are also shown to favor the formation of the TCP phase [34,51]. Since the phase is observed to nucleate form the phase, overall material stability can still be expected to improve with Co additions. Comparisons between predict ed results and open literature show that JMatPro was not successful in predicting Co effects on microstructural stability. Ruthenium variation effects Ru effects on microstructural stability were investigated in Phase I by comparing the baseline Model (0 wt% Ru), Model 5 (1.64 wt% Ru) and Model 4

PAGE 179

161 (2.46 wt% Ru) alloys. A limited linear in crease in the amount of phase, was predicted to result from Ru additions at 900 C (Figure 4-27). In contrast to predicted results, Ru additions have been reported to suppress the formation of TCP phases [13]. Conflicts between predicted and reported data show that t he JMatPro calculations were not successful at establishing Ru trends for microstructural stability. Tungsten (and Molybdenum) variation effects W (and Mo) effects on micros tructural stability were investigated in Phase I by comparing the baseline Model alloy (2 at% W, 0 at% Mo) to the Model 23 (1 at% W, 1 at% Mo) alloy. A negligib le linear decrease in the amount of phase was predicted to result from increased W (and decreased Mo) concentrations at 900 C (Figure 4-31). It is generally recognized that supera lloys which contain high levels of refractory elements such as Mo and W are prone to the prec ipitation of TCP phases [34,50]. A study conducted for allo ys similar to CMSX-4, showed that an increase in W content as a substituti on for Mo resulted in a considerable volume fraction increase [34]. Predicted stability trends with respect to W (and Mo) variations were not in agreement with existing literature, sho wing that JMatPro cannot be used to determine W (and Mo) effects on microstructural stability. Gamma prime former (Tantalum, Alumi num, and Titanium) variation effects The effects of -former variations on microstructural stability were investigated in Phase I by comparing -former contents of 13.25 to 16.75 at%, in Model 20 ( 2 at% Ta, 11 at% Al, 0.25 at% Ti); Model 19 ( 3 at% Ta, 11 at% Al,

PAGE 180

162 0.25 at% Ti); baseline Model alloy (3 at% Ta, 12 at% Al); Model 18 (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti); and Model 17 (3.25 at% Ta, 12.75 at% Al, 0.75 at% Ti). Taking into consideration all -former variants in Phase I, a linear increase in TCP phase amounts was predicted at 900C, with increasing -former content (Figure 4-35). -former effects in Phase II were investigated using 2 alloy groups with varying -former contents of 14 to 16 at%. Group 1 compared Model D (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model H (2 at% Ta, 9 at% Al, 3 at% Ti); baseline Model A (3 at% Ta, 11 at% Al); Model O (3 at% Ta, 11 at% Al, 2 at% Ti); and Model P (3 at% Ta, 10 at% Al, 3 at% Ti ) alloys. Group 2 compared Model E (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model I (2 at% Ta, 9 at% Al, 3 at% Ti); Model B (3 at% Ta, 11 at% Al, 1 at% Ti); Model M (3 at% Ta, 11 at% Al, 2 at% Ti); and Model N (3 at% Ta, 10 at% Al, 3 at% Ti). Increased -former contents in both Group 1 and Group 2 variants, were predicted to result in the linear increase of phase amounts expected at 600C (Figure 4-47). In line with the predicted results, previous studies with Ni-base SC superalloys have reported that excessive amounts of the phase renders the matrix prone to TCP phase precipitation [25]. The distri bution of elements between the and phases, tends to partition Al, Ti, and Ta to the phase while solid solution strengtheners such as Re, Mo, and W partition to the matrix [45]. As the volume fraction in a material increases, refractory elements are concentrated in the matrix. Since TCPs are com posed of refractory elements, which easily precipitate from areas with localized elemental enrichment,

PAGE 181

163 increased TCP phase formation may be at tributed to an increase in the phase [13,25,34]. Subsequently increased amounts of -forming elements (Al, Ti, Ta) have been shown to decrease phase stability [13]. Predicted stability trends with respect to -former content were in agreement with existing literature, making JMatPro a viable tool in calculating former effects on microstructural stabilit y. Computational and reported results show that microstructural st ability decreases with increasing -former contents. Aluminum (and Tantalum) variation effects Al (and Ta) effects on microstructural stab ility were investigated in Phase I by comparing the Model 14, the baseli ne Model, Model 13, and Model 12 alloys with Al contents ranging from 5.04 to 6 wt% Al, balanced by Ta variations. Thermodynamic calculations at 900 C (Fig ure 4-7) predicted a negligible Al (and Ta) effect on phase amounts. Ta and Al additions have been reported, in previous studies, to promote the formation of detrimental topologica lly closed packed (TCP) phases that precipitate during elevated temper ature exposure and degrade material properties [8,13]. Both Ta and Al additi ons contribute to the formation of the precipitate and consequently increase a materials volume fraction [25]. As previously mentioned, increas es in the amount of the precipitate result in decreased microstructural stability [13]. Although TCP phase formation is report ed to result from increasing Al content [13], decreasing Ta concentrati ons may be responsible for the decrease in TCP phase amounts, predicted fo r increased Al but decreased Ta concentrations. Due to combined effects of Ta and Al variations in computational

PAGE 182

164 modeling, no definite conclusi ons can be made in this study on the usefulness of JMatPro to predict Al (and Ta) effects on microstructural stability. To fully ascertain the effect of increased Al (and Ta) on microstructural stability, additional testing would be required. Allo y compositions with a single Al or Ta variation could be used to examine elemental influences. Titanium (and Tantalum, Al uminum) variation effects Ti addition effects on microstructural st ability were investigated in Phase I by comparing the baseline Model alloy (0 wt% Ti) to the Model 11 (0.2 wt% Ti) and Model 9 (0.58 wt% Ti) alloys with Al substitutions, along with Model 10 (0.39 wt% Ti) with a Ta substitution. A neg ligible increase in the amounts of TCP phase at 900 C were predicted for Ti additions (with either Al or Ta substitutions) investigated in this study, as seen in Figure 4-11. The limited Ti (and Al, Ta) effect on TCP phase amounts is attri buted to the small 0.58 wt% Ti range analyzed. Work conducted by Dreshfield and As hbrook on IN-100 show that the time to initiate sigma phase formation decreased with Al or Ti additions, subsequently increasing TCP phase formation [8]. Ta has also been reported to increase TCP phase formation [8,13]. Ta, Al, and Ti are all forming elements that stabilize the precipitate and contribute to the increase of a materials volume fraction [45]. Increased volume fractions may lead to t he concentration of refractory elements in the matrix and subsequently intensify TCP formation [13,34]. Despite Ta and Al reductions in the compositions analyzed, JMatPro equilibrium calculations succe ssfully predicted that the add ition of Ti would result in decreased microstructural stability.

PAGE 183

165 Al/Ti ratio variation effects Al/Ti ratio (wt%/wt%) effe cts on microstructural stability in Phase II were investigated using 2 alloy groups with va rying Al/Ti ratios. Group 1 compared Al/Ti ratios from 1.88 to 6.2 in Mo del O, Model P, and the baseline Model A alloys. Group 2 compared Al/Ti ratios fr om 1.88 to 4.11 in Model N, Model M, and Model L alloys. No clear Al/Ti ra tio effects were predicted for the phase, but a decrease in the amount of was predicted with increased Al/Ti ratios at 600 C (Figure 4-50). Al/Ti ratio (with Ta) effects in Phase III were investigated using Alloy 1 and Alloy 2. Alloy 1 (with a Al/Ti ratio of 3.72) contained a 2.8 wt% Ta addition in comparison to Alloy 2 (with a Al/Ti ratio of 1.69). JMatPro TC P phase predictions at 600 C (Figure 4-62) pr edicted that an increase in TCP phase formation would result from an increase in Al/Ti ratio (with a Ta addition). Open literature and previ ous studies report that sigma phase formation increases with high levels of Al, Ti, and Ta [8,13]. Al, Ti, and Ta are all precipitate formers which increase the amount of precipitate in a material. Increased amounts of the phase allow refractory elements to concentrate in the matrix [34]. Localized areas of elemental enrichment are prone to TCP formation, and may decrease microstructura l stability [13,34]. Increasing Al/Ti ratios, based on open literatur e, would result in increased TCP phase formation when Al content was increased, or in decreased TCP phase formation if Ti reductions were used. Increased TCP phase formation was calculated for increasing Al/Ti ratios (with a Ta addition) in Phase III of the present study. Even t hough Ti variations

PAGE 184

166 in this project verified JMatPros ability to use equilibrium calculation to predict Ti effects on microstructural stability, Al and Ta effects have not been fully ascertained. Due to the combined effe cts of Ti, Ta, and Al variations in computational modeling, no def inite conclusions can be made in this study on the usefulness of JMatPro to predi ct Al/Ti or Al/Ti (and Ta) effects on microstructural stability. Additional testing would be requir ed in order to fully determine the effect of increasing Al/Ti ratio on material stab ility, using alloy com positions with single Al/Ti ratio variations. Al/Ti ratios are of special interest wit h respect to hot corrosion resistance. Although in-service conditions have a significant effect on the hot-corrosion process, the alloy composition plays a crit ical role in hot corrosion behavior. Work by Ross and OHara on Ren e N4, showed that decreasing Al/Ti ratios increase hot corrosion resistance of the a lloy system [35]. Consequently, using low Al/Ti ratios in the design of Phas e III compositions could also provide resistance to hot corrosion attack. Chromium variation effects Cr effects on microstructural stability were investigated in Phase I by comparing four alloys with Cr contents r anging from 6.75 wt% to 11.82 wt% Cr, in the Model 8 (6.75 wt% Cr), Model 7 (8 .44 wt% Cr), the bas eline Model (10.12 wt% Cr), and Model 6 (11.82 wt% Cr) allo ys. Thermodynamic calculations at 900 C, predicted a linear increase in TCP phase amounts with increasing Cr content (Figure 4-3). Cr (with Al and Ti) effects in Phase II were investigated by comparing Model Q (14 at% Cr 10 at% Al, 2 at% Ti) to the base line Model A alloy (13 at%

PAGE 185

167 Cr, 11 at% Al, 1 at% Ti). The Cr addition (with a Ti addition and Al reduction) was predicted to produce small decreases in the and TCP phase amounts at 600 C (Figure 4-44). Cr (with Al and Ta) effects in Phase III were investigated by comparing Alloy 3 to Alloy 5. Alloy 3 (12 wt% C r) contained a 0.26 (0.5 at%) Al addition and a 1.32 wt% (0.5 at%) Ta reduction in co mparison to Alloy 5 (11 wt% Cr). The amount of phase predicted at 600 C, increased with a Cr addition (with an Al addition and Ta reduction), according to JMatPro microstability trends (Figure 463). Previous studies have shown that high Cr concentrations in Ni-base superalloys facilitate TCP phase formation [25,34,50,51,52]. Alloys which contain high levels of refractory element s are more prone to the precipitation of TCP phases; Increased Cr content increases this refractory concentration. Increased Cr content may then lead to the supersaturation of the solid solution which may in turn lead to the precipitati on of refractoryrich TC P phases [25]. Cr additions, in general, are also shown to increase TCP volume fraction, allowing the stabilization of the phase [34]. Open literature has also reported decreased microstructural stability with high levels Al, Ti, and Ta, which determine the quantity of the phase and render the matrix more prone to the precipitation of TCP phases [8,13,25]. Cr trends predicted for Phase I were in agreement with existing literature. Discrepancies in Phase II predictions, of Cr trends on microstructural stability, may be attributed to the reduction of Al in Phase II alloy compositions. Despite

PAGE 186

168 Ta and Al variations in Phase III compos itions, JMatPro equilibrium calculations successfully predicted that the addition of Cr would result in decreased microstructural stability. Even though no deleterious phases were seen in the Phase III as-cast or solution heat treated samples, JMatPr o calculations show general agreement with existing literature. Similarities in TCP phase formation trends demonstrate that JMatPro is a viable tool to predict Cr effects on microstr uctural stability. Based on the present work, increased Cr content could result in the eventual formation of deleterious phases, during t he aging heat treatments or service life of a material. Despite decreased material stability due to increased Cr concentrations; Cr additions are used, taking into account the hot corrosion requirements in IGT alloy design. Alloys that incorporate high levels of Cr content report significant improvements in hot corrosion resistance [ 35]. This improvement is mainly due to the formation of a protective Cr rich oxide scale at tem peratures where hot corrosion is active [25]. As a result, low le vels of Cr in a material, as a means to improve long-tem stability, could result in hot corrosion failures. The hot corrosion resistance due to Cr additions must then be balanced with the concern for decreasing stability in the material. Rhenium variation effects The effects of Re (with Ta, Al, and W) variations on microstructural stability were investigated in Phase I by com paring the baseline Model (3.02 wt% Re), Model 15 (1.5 wt% Re, 7.46 wt% W), and Model 16 (0 wt% Re, 7.34 wt% Ta, 7.46 wt% W, 5.45 wt% Al) alloys. Despite W, Ta, and Al variations between alloy

PAGE 187

169 compositions, increased amounts of TCP phases were predicted at 900 C (Figure 4-15), which were attributed to increasing Re concentrations. Re effects in Phase II were investigated using 3 alloy groups with varying Re concentrations from 0 to 3.02 wt% Re. Group 1 compar ed Model B (0 wt% Re), Model C (1.5 wt% Re), and the bas eline Model A alloy (3.02 wt% Re). Group 2 compared Model D (3.02 wt% Re), Model E (2.28 wt% Re), Model F (1.5 wt% Re), and Model G (0 wt% Re). Group 3 compared Model H (3.02 wt% Re), Model I (2.28 wt% Re), Model J (1.5 wt% Re), and Model K (0 wt% Re). Regardless of compositional differences between variant groups, calculations predicted clear increases in and phase amounts with increasing Re content (Figure 4-41). Re effects in Phase III were evaluat ed using Alloy 4 and Alloy 5, with 3.02 wt% Re and 0 wt% Re, respectively. JMat Pro TCP phase predictions at 600 C (Figure 4-64), predicted that increased TCP phase amount s would result from Re additions. Existing literature is in agreement wit h the present work, showing that the precipitation of TCP phases and SRZs incr ease in superalloys which contain high levels of Re [13,25,34,50]. The fo rmation of extended rafts containing the phase were first observed by Darolia et al in a Re containing alloy [6]. In previous studies with Re containing alloys, the planar phase has been shown to act as the nucleation site for other forms of TCP phases [6,34]. Additionally, Re has been reported to contribute to the chemical makeup and stabilization of the phase [13,25,34,50].

PAGE 188

170 Even though no deleterious phases were seen in the Phase III as-cast or solution heat treated samples, JMatPr o calculations show general agreement with existing literature. Similarities in TCP phase formation trends demonstrate that JMatPro is a viable tool to predict Re effects on microstructural stability. It was noted that after solution heat treatment, residual dendritic structure was observed in Alloy 4 (Figure 4-59 (b)), which was not observed in any other Phase III alloy. The residual segregation in the solution heat treated Re containing alloy, points towards local enric hments of the alloying elements in the microstructure, providing a strong driving fo rce for the precipitat ion of deleterious phases [7,13,15,34,50]. Experimental to Computational Mate rial Microstructure Comparisons With respect to elemental trends discu ssed in the sections above, JMatPro was successful in predicting elemental e ffects on microstructural stability in 60% of the cases considered in this st udy. Thermodynamic equilibrium module calculations for Phase I alloys below 1000 C, predicted equilibrium phases that included the , and phases, along with M23C6 carbides for C containing alloys. Calculations for Phase II alloys below 1000 C, predicted equilibrium phases that included , , and M23C6 carbides. Equilibr ium calculations for Phase III alloys below 1000 C, predicted equilibrium phases that included the , and phases, along with M23C6 carbides. Laboratory tested Phase III alloys in the as-cast form revealed a material microstructure which contained and phases along with various forms of primary MC carbides. As-cast st ructures also exhibited coarse precipitates in eutectic + regions.

PAGE 189

171 Different forms of MC carb ides in the as-cast structure were rich in Ti, Ta, or W (Figure 4-56 (g) and (h)). Primary ca rbides of the (Ti, Ta)C type appeared predominately in the blocky, cubic morphol ogy form (Figure 4-56 (f)) common to nickel based superalloys [26,45]. The com positions of the other types of MC carbides were consistent with previ ous studies conducted by Kong and Chen, which measured significant Ta, Mo, and W concentrations in MC carbides of experimental Ni-based SX crystal superallo ys [26]. The substitution of less reactive elements (W and Mo) in primar y carbides have also been observed in Udimet 500 and Rene 77 [7]. After soluti on heat treating, only the , and (Ti,Ta)C carbides remained. It has been reported that Mo and W w eaken the binding forces in MC carbides to such an extent that their degradation may occur during heat treatment [7,26]. The dissolu tion of W rich carbides dur ing solution treatment in this study was in accordance with these pr evious findings. Therefore, it is reasonable to conclude that W concentrati ons in MC carbides decrease the MC binding force, resulting in the dissolu tion of the primary carbides during heat treatment. Phase Formation Predicted equilibrium solidification pat hs suggested for Alloy 1, Alloy 2, Alloy 3, and Alloy 5 are seen below. L L + L + + MC + + MC + + M23C6 + + + + M23C6 The modeled phase transformation process predicted for the Re-containing alloy, Alloy 4, follow ed the progression below.

PAGE 190

172 L L + L + + MC + + MC + + + MC + + + + M23C6 The presence of the , and (Ti,Ta)C carbide phases before and after solution heat treatment validated the fi rst portion of the thermodynamically preferred solidification pat h predicted with JMatPro equi librium calculations. Although calculations predicted TCP and M23C6 phases in the material microstructures of the experimental al loys, these phases were not observed in either the as-cast or solution heat treated alloys. Equilibrium calculations assume that a material system reaches equilibrium at all temperatures even though typical superalloy manufacturing processes and even solution heat treatments do not approach the equilibrium state. The absence of equilibrium secondary ca rbides in material microstructures is attributed to the very stable (Ti,Ta)C carbides that do not easily transform into other forms of carbides (M23C6). Previous studies have reported that Ta additions increase the MC binding forces of (Ti,Ta)C carbides, thus, making them more stable [26]. The increased binding fo rces of the Ta containing MC-carbides would be expected to retard MC carb ide degradation during heat treatment [7,26]. This increase in carbide stab ility was seen in a study conducted for Ta additions in a Mar-M247 type alloy, where carbides did not degenerate or transform into other forms with heat treat ment [45]. It was concluded that Ta concentrations in the (Ti,Ta)C carbides for Phase III alloys contribute to maintaining MC stablility during so lution heat treatment [26].

PAGE 191

173 Topologically close packed (TCP) phases typically form during aging heat treatment processes or during a materia ls service lifetime [13], where prolonged and high temperature heat exposures allow materials to more closely approach equilibrium conditions. Non-equilibrium conditions in initial alloy processing steps probably prevented the formation of the and TCP phases in the laboratory tested alloys. An overall good correlation between the initial steps of the predicted solidification path and present SEM observations were observed. Even though superalloy processing is in the non-equili brium state, it wa s reasonable to conclude that the later portion of t he predicted solidification path can be considered a closer approx imation of a material system in its extended service lifetime. JMatPro equilibriu m calculations may then be considered indicative of a materials propensity to form TCP deleter ious phases, as long as the kinetics of the formation of the TCP phase is taken into consideration. Phase Transformation Temperatures In this study, phase transformation te mperatures for Phase I, II, and III alloy chemistries were obtained using JMatPro 3.0 therm odynamic equilibrium calculations. The phase transformation temperatures, melting ranges, and heat treatment windows predicted for Phase I and Phase II alloys are discussed in detail in Chapter 4. T he transformation temperatures melting ranges, and heat treatment windows predicted for Phase III alloys are listed in Table 4-7. Experimental phase transformation temper atures for Phase III alloys were obtained through DTA testing of solutionized samples. DTA curves for Alloy 1, Alloy 2, Alloy 3, Alloy 4, and Alloy 5, in the solution heat treated conditions, are

PAGE 192

174 presented in Figures 4-66, 4-68, 4-70, 4-72, and 4-74, respectively. The experimentally determined transformati on temperatures, melting ranges, and heat treatment windows are listed in Table 4-8 Elemental Variation Effects Carbon variation effects C effects on phase transformation temperatures (Figure 4-20) were investigated in Phase I by comparing the Model 3 (0.01 wt% C), Model 2 (0.05 wt% C), Model 1 (0.075 wt% C), and the baseline Model (0 wt% C) alloys. C additions were predicted to linearly decrease the liquidus and solvus temperatures. A linear increase of t he solidus was predicted with increasing C content. Additionally, a linear decrease in the melting range and increase in the heat treatment window were predict ed with increasing C content. Previous studies have shown that C additions in a Ni-base superalloy decrease the liquidus, which, in turn, re sults in the decrease of a materials melting range [47,50]. Increased C cont ent has also been shown to increase solvus temperatures [50]. Even though JMatPro calcul ations successfully predicted C effects for only the liquidus temperature, solvus trends were not in line with previous studies in open literature. In general, JMatPro calculations may not be considered a viable tool to calculate C effects on general phase transformation temperatures. To fully ascertain the effect of increased C on solidus temperatures, additional testing would be required.

PAGE 193

175 Cobalt variation effects Co effects on phase transformation te mperatures (Figure 4-24) were investigated in Phase I by compari ng the Model 22 (9 wt% Co), the baseline Model (11.47 wt% Co), and the Model 21 (12 wt% Co) alloys. Calculations predicted that increasing Co concentrati ons would produce a linear decrease in the solidus, a linear increase in the solvus, and a negligible decrease of the liquidus temperature. A c onsiderable decrease in the heat treatment window and a negligible increase in the solidification range were pr edicted for increasing Co content. Co additions have been previously repor ted to produce a decrease in the solidus, a marked decrease in the solvus, and an increase the liquidus [47,51]. The influence of Co additions on these three phase transformation temperatures have been shown to result in the increas e of both the heat treating window and the melting range [47]. JMatPro calculations successfully predi cted Co effects for only the solidus temperature and for the melting range. Predicted solvus and liquidus temperature trends did not agree with existi ng literature. On the whole, JMatPro calculations cannot be considered a dependabl e tool to calculate Co effects on phase transformation temperatures. De spite discrepancies, open literature verified that increased Co content dec reased the solidus and increased the melting range. Ruthenium variation effects Ru effects on phase transformation te mperatures (Figure 4-28) were investigated in Phase I by comparing the baseline Model (0 wt% Ru), Model 5

PAGE 194

176 (1.64 wt% Ru), and Model 4 (2.46 wt% Ru) alloys. Increasing Ru contents were predicted to result in limited linear decreases in the solidus and the solvus. Ru additions were predicted to increase the solidification range and to maintain a relatively constant heat treating window. In previous studies, Ru additions have resulted in an increase the liquidus temperatures of Ni-based s uperalloys, which could, in turn, results in the increase of the melting range [13 ]. Discrepancies between predicted and reported data show that t he JMatPro calculations were not successful at establishing Ru trends for the liquidus te mperature. In or der to validate Ru effects on the solidus and solvus temperatures, additional testing would be required. Tungsten (and Molybdenum) variation effects W (and Mo) effects on phase trans formation temperatures were investigated in Phase I by comparing the baseline Model alloy (2 at% W, 0 at% Mo) to the Model 23 (1 at% W, 1 at% Mo) alloy. W (and Mo) effects predicted for phase transformation temperatures are seen in Figure 4-32. A linear increase in the liquidus and solidus were predict ed with increasing W (decreasing Mo) content. Previous studies have shown that increased W concentrations raise the solvus, liquidus, and solidus temperatures [13,47,51]. Existing literature also shows that the solvus temperature increases with increasing Mo concentrations [47]. JMatPro equilibrium calculations succe ssfully predicted W effects for the liquidus and solidus temperatures. The negligible W (and Mo) effect predicted

PAGE 195

177 for the solvus may be attributed to the Mo reduction, which has been shown to decrease the solvus. Similarities between predicted and reported trends make JMatPro a viable tool in calculating W (and Mo) tr ends on phase transformation temperatures. Computati onal and reported results s how that the liquidus and solidus temperatures increase with W additions (and Mo reductions). Gamma prime former (Tantalum, Alumi num, and Titanium) variation effects The effects of -former variations on phase tr ansformation temperatures were investigated in Phase I by comparing -former contents of 13.25 to 16.75 at%, in Model 20 ( 2 at% Ta, 11 at% Al 0.25 at% Ti); Model 19 ( 3 at% Ta, 11 at% Al, 0.25 at% Ti); baseline Model allo y (3 at% Ta, 12 at% Al); Model 18 (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti); and Model 17 (3.25 at% Ta, 12.75 at% Al, 0.75 at% Ti). Phase I -former effects predicted for phase transformation temperatures are seen in Figure 4-36. Despite Ti, Ta, and Al variations in the alloys considered; increased -former content was predicted to linearly decrease the solidus and liquidus temperatur es but linearly increase the solvus temperature. Increasing -former content was predicted to re sult in a small increase of the melting range and a considerable decr ease in the heat treatment window. -former effects in Phase II were investigated using 2 alloy groups with varying -former contents of 14 to 16 at%. Group 1 compared Model D (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model H (2 at % Ta, 9 at% Al, 3 at% Ti); baseline Model A (3 at% Ta, 11 at% Al); Model O (3 at % Ta, 11 at% Al, 2 at% Ti); and Model P (3 at% Ta, 10 at% Al, 3 at% Ti) alloys. Group 2 compared Model E (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model I (2 at% Ta, 9 at% Al, 3 at % Ti); Model B (3 at% Ta,

PAGE 196

178 11 at% Al, 1 at% Ti); Model M (3 at% Ta, 11 at% Al, 2 at% Ti); and Model N (3 at% Ta, 10 at% Al, 3 at% Ti). Phase II -former effects predicted for phase transformation temperatures are seen in Figure 4-48. Increased -former contents in both Group 1 and Group 2 variants, were predicted to decrease the liquidus and solidus temperatures. A 2nd order relationship predicted for the solvus temperature, expected an increase and subs equent decrease of the solvus with increasing -former content. Increasing the -former content was predicted to increase the melting range. Large heat tr eatment windows were predicted for lower amounts of -formers, which are expected to na rrow at compositions nearing 15 at% former content, before increas ing slightly at higher -former amounts. Existing literature shows that incr eased hardener contents produce a strong decrease in the solidus, decrease in the liquidus, and increase in the solvus [47,51]. The liquidus has been shown to dec rease to the greates t extent with Ti additions, followed by Al, and then Ta additions [47]. Consequently, increased -former content in Ni-base superallo ys sharply decrease the heat treating window [47]. JMatPro equilibrium calculat ions successfully predicted -former effects for all of the liquidus, solidus, and solvus temperatures, making JMatPro a viable tool in calculating -former trends on all phase tr ansformation temperatures. Computational and reported result s show that an increase in -former content produces a decrease in the liquidus, a decre ase in the solidus, an increase in the solvus, and a marked decrease in the heat treating window.

PAGE 197

179 Aluminum (and Tantalum) variation effects Al (and Ta) effects on phase transformati on temperatures (Figure 4-8) were investigated in Phase I by comparing the Model 14, the baseline Model, Model 13, and Model 12 alloys with Al contents ranging from 5.04 to 6 wt% Al, balanced by Ta variations. Calculations for increasing Al (decreasing Ta) content predicted linear increases in both the liquidus and solidus temperatures while decreasing the solvus temperature. Incr easing Al (and decreasing Ta) content, was predicted to decrease the melting range and considerably increase the heat treating window. Al and Ta additions are both r eported to decrease the liquidus temperatures. This decrease in liquidus temperature is most notable for Al additions followed by Ta [47]. Previous studies have shown that Al additions increase the solvus and decrease the liquidus in different Ni-based superalloys [47]. Increased Ta concentrations in IN -100, showed that the rate at which the solvus increased was greater than the rate at which the solidus decreased, resulting in a significant decrease in the heat treatment window [47,51]. Despite Al variations in the co mpositions considered, JMatPro computational modeling successfully predicted Ta effects on all phase transformation temperatures. Decreasing Ta concentrations in the Al (and Ta) variants considered, may be responsible for increased liquidus, solidus, and heat treating windows, apart from decreased -solvus and melting ranges. To fully ascertain the effect of increased Al (and Ta) on phase transformation temperatures, additional testing would be required

PAGE 198

180 Titanium (and Tantalum, Al uminum) variation effects Ti addition effects on phase transformati on temperatures were investigated in Phase I by comparing the baseline M odel alloy (0 wt% Ti) to the Model 11 (0.2 wt% Ti) and Model 9 (0.58 wt% Ti) alloys with Al substitutions, along with Model 10 (0.39 wt% Ti) with a Ta substitution. Phase I Ti addition (with either Al or Ta reduction) effects predicted for phase transformation temperatures are seen in Figure 4-12. A Ti addition, with an Al reduction, was predicted to result in a negligible decreases in the liquidus and solidus. Ti additions with Ta reductions were predicted to increase the solidus and produce a negligib le increase in the liquidus. The limited Ti (and Al, Ta) effects on phase transformation temperat ures were attributed to the small 0.58 wt% Ti range analyzed. Regardl ess of whether Ti additions had been balanced by Al or Ta reductions, calculat ions predicted a linear decrease in the solvus temperature with Ti additions. A considerable increase in the heat treatment window was predicted for increasing Ti content. Ti additions have been report ed to strongly increase the solvus, depress the liquidus, and depress the solidus temperat ures [47]. Accord ingly, Ti additions lead to the narrowing of both the heat tr eatment window and melting range [47]. Previous studies have shown t hat Al reductions decrease the solvus [47]. Existing literature shows t hat Ta reductions lower the solvus, lower the solidus, and increase the liquidus temperatures [ 47]. Reported decreases in the liquidus temperature with alloying addit ions are the largest for Ti additions, followed by Al, and then Ta additions [47].

PAGE 199

181 Trends calculated for Ti additions (wit h Al reductions) were in agreement with reported Ti effects for liquidus and so lidus temperatures. Predicted trends for Ti additions (with Ta reductions) we re in agreement with reported Ta effects for solvus and liquidus. Al and Ta reducti ons in the compositions considered may be responsible for the decrease in the solvus. Due to the combined effects of Ti, Ta, and Al variations in computational modeling, no definite conclusions can be made in this st udy, on the usefulness of JMatPro in predicting Ti (and Ta) or Ti ( and Al) effects on phase transformation temperatures. To fully ascertain the effe ct of Ti, Ta, or Al variations on phase transformation temperatures, additiona l testing would be required. Al/Ti Ratio variation effects Al/Ti ratio (wt%/wt%) effects on phas e transformation temperatures in Phase II were investigated using 2 alloy groups with varying Al/Ti ratios. Group 1 compared Al/Ti ratios from 1.88 to 6. 2 in Model O, Model P, and the baseline Model A alloys. Group 2 compared Al/Ti rati os from 1.88 to 4.11 in Model N, Model M, and Model L alloys. Phase II Al/Ti ratio effects predicted for phase transformation temperatures are seen in Fi gure 4-13. Increasing Al/Ti ratios were predicted to linearly increase the liquidus, solidus, and solvus. The melting range and heat treatm ent window were predicted to remain relatively constant with increasing Al/Ti ratio. Al/Ti ratio (with Ta) effects in Phase III were investigated using Alloy 1 and Alloy 2. Alloy 1 (with an Al/Ti ratio of 3.72) contained a 2.8 wt% Ta addition in comparison to Alloy 2 (with an Al/Ti ratio of 1.69). Figure 4-78 depicts phase transformation temperature trends with respec t to increasing Al/Ti ratio (with a Ta

PAGE 200

182 addition). Calculations predicted a solidus decrease (12 C) and a marked solvus increase (35 C) with the Al/Ti ra tio increase (with a Ta addition). The melting range was predicted to increase 8 C and the heat treatment window was predicted to decrease 48 C, with a 2 (wt% /wt%) Al/Ti ratio increase with a 2.8 wt% (1 at%) Ta addition. Experimental testing for the Phase III alloys showed that the 2 (wt%/wt%) Al/Ti ratio increas e with a 2.8 wt% (1 at%) Ta addition, increased the solvus 6 C. Increasing Al and Ti concentrations in different Ni-based alloys are reported to increase the solvus [47]. Large decreas es reported in the liquidus temperature result from Ti additions, fo llowed by smaller decreases from Al additions [47]. Accordingly, the melt ing ranges of well known commercial superalloys are shown to decrease with in creasing Al and Ti content [47]. Despite Ta additions in Phase III co mpositions, JMatPr o computational modeling successfully predicted increasing solvus temperatures with increased Al/Ti ratios in all variants considered. Apart from the solvus temperature, a generally poor correlation between calcul ated, reported, and experimental phase transformation temperature trends were obser ved for increased Al/Ti ratio or Al/Ti ratio (and Ta additions). Since both Al and Ti are reported to produce similar trends in phase transformation temperatures, no definite conc lusions could be made in this study on the usefulness of JMat Pro to predict Al/Ti effects on phase transformation temperatures. To fully determine t he effect of Al/Ti ratios on phase transformation temperatures, additi onal testing would be required.

PAGE 201

183 Chromium variation effects Cr effects on phase transformation tem peratures were investigated in Phase I by comparing four alloys with Cr contents ranging from 6.75 wt% to 11.82 wt% Cr, in the Model 8 (6.75 wt% Cr), Model 7 (8.44 wt% Cr), the baseline Model (10.12 wt% Cr), and M odel 6 (11.82 wt% Cr) alloys. Phase I Cr effects predicted for phase transformation temper atures are seen in Figure 4-4. Considerable linear decreases in the solidus, liquidus, and solvus were predicted for increasing Cr content. Increas ing Cr content was predicted to result in an increase in the melting range and main tain a nearly constant heat treatment window. Cr (with Al and Ti) effects in Phase II were investigated by comparing Model Q (14 at% Cr 10 at% Al, 2 at% Ti) to the base line Model A alloy (13 at% Cr, 11 at% Al, 1 at% Ti). Phase II Cr (with Ti and Al) effects predicted for phase transformation temperatures are seen in Figu re 4-45. Increased Cr content (with a Ti addition and Al reduction) was pr edicted to increase in the liquidus and solidus temperatures, while decreasing the solvus. Increasing Cr content (with a Ti addition and Al reduction) was pr edicted to decrease the melting range and considerably increase the heat treatment window. Cr (with Al and Ta) effects in Phase III were investigated by comparing Alloy 3 to Alloy 5. Alloy 3 (12 wt% C r) contained a 0.26 (0.5 at%) Al addition and a 1.32 wt% (0.5 at%) Ta reduction in com parison to Alloy 5 (11 wt% Cr). Figure 4-79 depicts phase transformation trends fo r Phase III alloys with respect to increasing Cr concentration (with an Al addition and Ta reduction) Calculations

PAGE 202

184 predicted a decrease in the solvus and increase in the heat treatment window with increasing Cr content (with an Al addition and Ta reduction). Experimental testing for Phase III allo ys; however, showed an increase in the solvus temperature with increasing Cr content (with an Al addition and Ta reduction). A 17 C solvus increase resulted from a 1 wt% Cr addition, a 0.26 wt% Al addition, and a 1.32 wt% Ta reducti on. Experimental testing also showed decreases in the solidus, liquidus, and heat treatment window with increased Cr content (with an Al additi on and a Ta reduction). It has been reported that Cr can be added to a Ni-base superalloy to decrease the solvus and sligthly decrease the solidus, resulting in an increase in the heat treatment window [5 1]. Existing lite rature shows that Al, Ti, and Ta additions produce a strong decrease in the solidus, a decrease in the liquidus, and an increase in the solvus [47,51]. In Phase II compositions, Cr (with Ti and Al) temperature trends were all in agreement with reported Al r eduction effects. Decreased Al concentrations may then be responsible for the predicted te mperature trends in Phase II. In Phase III, discrepancies were observed between predicted, experimental, and reported Cr (Al and Ta) effects on phase transformation temperatures. Due to the combined effect of Cr, Al, and Ta variations in Phase III, no clear determination was made on the usefulness of JMatPr o to predict Cr (Al and Ta) trends on phase transformation temperatures. Despite inconsistencies in Phase II and Phase III results, predicted Phase I Cr trends were in agreement with repor ted Cr effects on phase transformation

PAGE 203

185 temperatures for the solidus, solvus, and heat treatment window. It is, thus, reasonable to conclude that additional elem ental variations in Phase II and Phase III compositions did not allow clear elemental trends to be established. Based on Phase I results, it is also reasonable to conclude that JMatPro is a useful tool in predicting Cr trends on phase transformation temperatures for the solidus, and solvus temperatures. Rhenium variation effects The effects of Re (with Ta, Al, and W) variations on phase transformation temperatures were investigated in Phas e I by comparing the baseline Model (3.02 wt% Re), Model 15 (1.5 wt% Re, 7.46 wt% W), and Model 16 (0 wt% Re, 7.34 wt% Ta, 7.46 wt% W, 5.45 wt% Al) alloys. Re (with Ta, Al, and W) variation effects on phase transformation temperatures are seen in Figure 4-16. Despite Ta, Al, and W variations between alloy co mpositions, predicted decreases in the solvus and small increases in the liq uidus were attributed to increasing Re concentrations. No conclusions could be made on Re (with Ta, Al, and W) effects on the melting range; however, in creasing Re content did result in a predicted increase in the heat treatment window. Re effects in Phase II were investigated using 3 alloy groups with varying Re concentrations from 0 to 3.02 wt% Re. Group 1 compar ed Model B (0 wt% Re), Model C (1.5 wt% Re), and the bas eline Model A alloy (3.02 wt% Re). Group 2 compared Model D (3.02 wt% Re), Model E (2.28 wt% Re), Model F (1.5 wt% Re), and Model G (0 wt% Re). Group 3 compared Model H (3.02 wt% Re), Model I (2.28 wt% Re), Model J (1.5 wt% Re), and Model K (0 wt% Re). Phase II Re effects predicted for phase transfo rmation temperature are seen in Figure

PAGE 204

186 4-42. Negligible Re effects were pr edicted for the liquidus temperature. A decrease in the solvus and increase in the heat treatment window were predicted, as a result of in creased Re concentrations. Re effects in Phase III were evaluat ed using Alloy 4 and Alloy 5, with 3.02 wt% Re and 0 wt% Re, respectively. Phase III Re effects predicted for phase transformation temperatures are seen in Figu re 4-80. According to calculations, increasing Re concentrations would result in a decrease in the solvus, and an increase in the materia ls heat treatment window Experimental testing of Phase III a lloys showed that the 3.02 wt% Re addition increased the solvus 16 C, decreased t he heat treatment window 21 C, and increased the melting range 7 C. Existing literature reports that Re additions increase the liquidus temperature of a multic omponent superalloy [13]. In contrast to predicted results in all phases of this study, Re additions were shown to increase the solvus and decrease the heat treatment window, when compared to experimental data. On t he other hand, predicted Re trends on the liquidus temperature were in line with reported effects. Inconsistencies between the predicted and experimental data show t hat the JMatPro calculations were not successful at establishing Re phase transformation temperature trends. Experimental to Computational Temperature Range Comparisons With respect to the elemental trends discussed in the sections above, JMatPro was successful in predicting el emental effects on phase transformation temperatures in only 40% of the case s considered in this study. JMatPro modeling for Phase III alloys, in general, predicted liquidus and solidus temperatures below experimental values (Figure 4-75). Experimentally tested

PAGE 205

187 liquidus temperatures were approximatel y 30 to 40 C higher than predicted values, with modeling errors of 2 to 3%. Experimentally tested solidus temperatures were 30 C to 50 C above predicted values with modeling errors of 3 to 4%. In general, JMatPro modeling for Phase III alloys predicted solvus temperatures 15 C to 41 C above experi mental values (Fi gure 4-75), with the exception of Alloy 2. Alloy 2 demonstr ated an excellent correlation between its calculated and experimental solvus, with a predicted value only 2 C below its experimental temperature. Overall comparisons between experimental and predicted solvus temperatures for all Phase III alloys lead to modeling errors from 0 to to 4%. Melting ranges predicted for Phase III alloys deviated only 0 C to 11 C, from experimentally tested values. De spite discrepancies in the solidus and liquidus predictions, melting range m agnitudes showed an excellent correlation between modeling and experimental values This can be attributed to the consistent deviations between predicted and experimental values for both the solidus and liquidus temperatures. Although almost all of the alloys de monstrated an excellent correlation between predicted and experimental melt ing range magnitudes, predicted heat treatment windows deviated 30 C to 93 C from experimentally tested values. This high degree of deviation from the ca lculated to experimental heat treatment windows is attributed to the combined e ffect of the conservative solidus and elevated solvus temperatures predict ed with equilibrium calculations.

PAGE 206

188 It is interesting to note that the ab sence of segregational issues at the liquidus should lead to a more accurate prediction of the liquidus temperature with equilibrium calculations. Despite equilibrium-like co nditions at the liquidus, JMatPro predicted liquidus temperatures 30-40 C belo w experimental values. Work conducted by Sponseller showed that variations in heating rate of 10 C/min in DTA testing resulted in trans formation temperature deviations of less than 10 C [47]. The large differences in the liquidus and solidus temperatures between experimental and JMat Pro thermodynamic equilibrium calculations may be then attributed to the specific coe fficients (proprietary or found in open literature) used in the t hermodynamic model to descr ibe the properties of the various phases in a Ni-based superalloy, and not to heating rate effects of laboratory DTA testing. Although there was a good correlation between predicted and experimental melting range magnitudes, this relationshi p is attributed to the consistent calculation of conservative liquidus and solidus temperatures. Discrepancies between the predicted and experimental phase transformation temperatures show that JMatPro equilibrium calculations are not a viable tool in assessing phase transformation temperatures. Segregation Partitioning coefficients, to gauge elem ental segregation in a material, were obtained for all alloys, considering key el ements (Ni, Ta, Al, Cr, Co, W, and Re) in each alloys chemical composition. The segregation behavior of the Phase I and Phase II alloys was determined using calculated partitioning coefficients (kx,calc). The calculated par titioning coefficients (kx,calc) for Phase I and Phase II

PAGE 207

189 alloy chemistries, are discussed in detail in Chapter 4. T he segregation behavior for the Ni, Ta, Ti, Al, Cr, Co, W, and Re elements in Phase III alloys were deduced using calculated (kx,calc) and experimentally (kx,exp) determined partitioning coefficients listed in Tables 4-11. and 4-10, respectively. kx,exp values were determined for as-cast Phase III alloys. Elemental Variation Effects Carbon variation effects C effects on segregation were investi gated in Phase I by comparing the Model 3 (0.01 wt% C), Model 2 (0. 05 wt% C), Model 1 (0.075 wt% C), and baseline Model (0 wt% C) alloys. Phas e I C effects predicted on segregation behavior are seen in Figure 4-21. Within the C range consider ed, calculations predicted a small C effect on Al. C addi tions were predicted to decrease Al segregation to the inte rdendritic region. Carbon additions have been previously reported to enhance the chemical homogeneity of a superalloys dendritic mi crostructure by decreasing the degree of refractory element segregation [50]. Previous work with single crystal alloys have shown that C additions lead to the reduction of W segregation, in particular [50]. Overall, predicted and reported results suggests that carbon additions improve homogeneity by decreasing the degr ee of elemental segregation [50]. Eventhough computational and reported results demons trate that C additions decrease overall material segregation, re spective results correspond to differing elements (elements typical of interdendr itic vs. dendritic partitioning).

PAGE 208

190 Inconsistencies in specific elemental tr ends show that JMatPr o is not a useful tool in predicting general material segregation trends with C additions. Cobalt variation effects Co effects on segregation were investi gated in Phase I by comparing the Model 22 (9 wt% Co), the baseline Model (11.47 wt% Co), and the Model 21 (12 wt% Co) alloys. Predicted Phase I Co effects on segregation behavior are seen in Figure 4-25. Predicted Co variation effects were significant for Ta. Ta segregation was predicted to linearly in crease with increasing Co content. Previous studies have shown that Co additions decrease the segregation of the heavy elements Re, W, and Ta [12, 38,43]. Ta segregation trends with respect to Co variations were not in agreement with existing literature, showing that JMatPro cannot be used to determine Co effects on elemental segregation. Ruthenium variation effects Ru effects on segregation were investi gated in Phase I by comparing the baseline Model (0 wt% Ru), Model 5 (1. 64 wt% Ru), and Model 4 (2.46 wt% Ru) alloys. Predicted Phase I Ru variati on effects on segregation behavior are seen in Figure 4-29. Predicted Ru variation e ffects were significant for W. A linear decrease in W segregation was predict ed with increasing Ru content. In a study conducted for a Ni-base super alloy similar to CMSX-4; Re, W, Ta, and Al segregation (to either the dendritic or interdendritic regions) decreased with Ru additions [13]. In contrast, studies conducted on high refractory Ni-base alloys have shown that material segregation increases with large Ru additions [29,38 ]. Work conducted by Cald well showed no observable Ru effects on elemental segregation with an almost 2 wt% Ru addition; however,

PAGE 209

191 a marked increase in segregation was obs erved with larger Ru additions of approximately 3 % Ru [4]. JMatPro calculations in this study pr edict an overall decrease in material segregation with Ru additions. This c ould lead to extended solution heat treatment times and/or other segregation issues, due to localized concentrations of elements, unless the alloy is properly heat treated. Controversy in existing literature does not permit Ru trends on material segregation to be clearly established. Consequently, inconsist encies between predict ed and reported Ru trends do not allow definite conclusions to be made on the usefulness of JMatPro to predict Ru segregation trends. To fully ascertain the effect of increased Ru concentrations on elemental segregation, addi tional testing would be required. Tungsten (and Molybdenum) variation effects W (and Mo) effects on segregation we re investigated in Phase I by comparing the baseline Model alloy (2 at % W, 0 at% Mo) to the Model 23 (1 at% W, 1 at% Mo) alloy. Predicted Phase I segregation behavior for the W variants considered, is seen in Figure 4-33. De creases in segregation predicted with a 1 at% W addition (and 1 at% Mo reduction) we re most significant for Re, followed by W. Al was calculated to increase in segregation with W additions (and Mo reductions). Existing literature shows that W additi ons result in the increase of the overall elemental segregation in an alloy. The increase in segregation caused by the increasing W content may result in ex tended solution heat treatment times. Work reported by Fela on this same a lloy indicated simila r results. [12]

PAGE 210

192 Overall elemental segregation has been previously reported to increase with Mo reductions [4]. Mo reductions have been shown to significantly increase Al elemental partitioning, in specific alongside small Cr and Co partitioning decreases [4]. The increase in segregat ion caused by the r eduction of Mo may result in a increase in solution heat treat ment times. Previous reports by Fela indicated similar results [12]. Despite W additions in the compositi ons analyzed, JMatPro equilibrium calculations successfully predicted that the Mo reduction would result in increased Al segregation. According to existing literature, W additions and Mo reductions result in an overall increase of material segregation. However, computational results, pr edicted that decreases in Re and W segregation would result from W additions and Mo r eductions. The discrepancies between predicted and reported data s how that JMatPro calculat ions were not successful at establishing W (and Mo) trends for elemental segregation. Gamma prime former (Tantalum, Alumi num, and Titanium) variation effects The effects of -former variations on segregat ion were investigated in Phase I by comparing -former contents of 13.25 to 16.75 at%, in Model 20 ( 2 at% Ta, 11 at% Al, 0.25 at% Ti); Model 19 ( 3 at% Ta, 11 at% Al, 0.25 at% Ti); baseline Model alloy (3 at% Ta, 12 at% Al ); Model 18 (2.3 at% Ta, 13.7 at% Al, 0.5 at% Ti); and Model 17 (3.25 at% Ta, 12.75 at% Al, 0.75 at% Ti). Predicted Phase I -former effects on segregation are seen in Figure 4-37. Predicted former variation effects were most signific ant for Re, followed by Cr. Calculations predicted that an increase in -former content would result in increased Re and Cr partitioning.

PAGE 211

193 -former effects in Phase II were investigated using 2 alloy groups with varying -former contents of 14 to 16 at%. Group 1 compared Model D (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model H (2 at % Ta, 9 at% Al, 3 at% Ti); baseline Model A (3 at% Ta, 11 at% Al); Model O (3 at % Ta, 11 at% Al, 2 at% Ti); and Model P (3 at% Ta, 10 at% Al, 3 at% Ti) alloys. Group 2 compared Model E (2.5 at% Ta, 9.5 at% Al, 2 at% Ti); Model I (2 at% Ta, 9 at% Al, 3 at % Ti); Model B (3 at% Ta, 11 at% Al, 1 at% Ti); Model M (3 at% Ta, 11 at% Al, 2 at% Ti); and Model N (3 at% Ta, 10 at% Al, 3 at% Ti). Predicted Phase II -former effects on segregation are seen in Figur e 4-48. Increases in -former content in both Group 1 and Group 2 variants, were predict ed to increase Re segregation. Limited -former effects were predicted fo r Ni, Co, W, Ta, and Cr. No conclusions could be made on the -former variation effects on Al. In a previous study analyzing el emental segregation in Ni-based superalloys, Al reductions with Ta substi tutions and Ti additions allowed a 1 at% increase of the -former content of the allo y [4]. The increase in -former content resulted in increased Re, Co, Cr, Ti and Ta segregation. Overall, it was reported that the increased -former content, using an Al reduction with a Ta substitution and Ti addition, increased segregation [4]. A good correlation between calculated and reported segregation trends showed increases in Re and Cr segregation with increasing -former content. It is then reasonable to conclude that JM atPro is a viable t ool in predicting former effects on elemental segregation. Comparisons to open literature in this

PAGE 212

194 study verify that increased -former content results in an increase of material segregation. Aluminum (and Tantalum) variation effects Al (and Ta) effects on segregation we re investigated in Phase I by comparing the Model 14, the baseline M odel, Model 13, and Model 12 alloys with Al contents ranging from 5.04 to 6 wt% Al, balanced by Ta variations. Predicted Al (and Ta) effects on element al segregation are seen in Figure 4-9. Al (and Ta) variation effects were most significant for Re, followed by W, Ni, and then Co and Ta. Re, Co, and Ta segregation was pr edicted to increase linearly with Al additions (and Ta reductions). Increased Al (and reduced Ta) contents predicted linear decreases in W and Ni segregation. Increased Al (and decreased Ta) contents predicted a small shift in Ni segrega tion from the inter dendritic region to the dendritic core. Negligible Al (and Ta) variation effects were predicted for Al and Cr. Work conducted by Caldwell showed an overall increase in material segregation due to an increase in Al with a reduction in Ta [4]. A 1 at% Al substitution for a 1 at% Ta reduction, increased Re segregation to the greatest extent followed by W and Co. Ta and Al segregation were also shown to increase with an Al addition and Ta reduction. The reported Al (and Ta) effects on Re, Co and Ta in Caldwells work were in agreement with the present work, whic h also showed an increase in material segregation. Similarities between the predicted and reported results make JMatPro a useful tool in predicting Al (and Ta) effect s on elemental segregation.

PAGE 213

195 Titanium (and Tantalum, Al uminum) variation effects Ti addition effects on segregation we re investigated in Phase I by comparing the baseline Model alloy (0 wt % Ti) to the Model 11 (0.2 wt% Ti) and Model 9 (0.58 wt% Ti) alloys with Al s ubstitutions, along with Model 10 (0.39 wt% Ti) with a Ta substitution. Segregation beh avior trends calculated for Ti variants (with Al or Ta reductions) are seen in Figure 4-13. Regardless of whether Ti additions were balanced by either Ta or Al reductions, no elemental variation effects were predicted for Cr, Co, Ta, Ni, W, and Al. A small decrease in Re segregation was predicted wit h increased Ti content and Ta reductions, while a marked decrease in Re segregation was pr edicted for increasing Ti content, with Al reductions. The calculated results from this study are in agreement with previous work that analyzed Ta and Al substitutions for Ti in a Ni-based superalloy [4]. Substituting 1 at% Ti for a 1 at% Al r eduction has reportedly produced a slight decrease in Re segregation while increas ing W, Co, and Cr segregation [4]. It has also been shown that Re segr egation decreased when 1 at% Ti was substituted for a 1 at% Ta reduction [4]. JMatPro showed an excellent correl ation between predicted and reported Re segregation trends for Ti (and Al) and Ti (and Ta) variations. JMatPro can then be considered a useful tool in calcul ating elemental segregation for Ti (and Al) and Ti (and Ta) variations. Al/Ti Ratio variation effects Al/Ti ratio (wt%/wt%) effects on segregat ion in Phase II were investigated using 2 alloy groups with varying Al/Ti rati os. Group 1 compared Al/Ti ratios from

PAGE 214

196 1.88 to 6.2 in Model O, Model P, and the baseline Model A alloys. Group 2 compared Al/Ti ratios from 1.88 to 4.11 in Model N, Model M, and Model L alloys. Predicted Al/Ti ratio effects on elemental segregation are seen in Figure 4-51. Al/Ti ratio effects were most signific ant for Re, followed by W and then by negligible effects for Co and Al. Re segregation was predi cted to increase linearly with Al/Ti increases. A linear decrease in W segregation was predicted with increased Al/Ti ratios Al/Ti ratio (and Ta) effects on material segregation in Phase III were investigated using Alloy 1 and Alloy 2. Alloy 1 (with a Al/Ti ratio of 3.72) contains a 2.8 wt% (1 at%) Ta addition in comparison to Alloy 2 (with a Al/Ti ratio of 1.69). Segregation trends for Al/Ti (and Ta) variations are seen in (Figure 483). Equilibrium calculations for Phase III alloys predicted that Al/Ti ratio (and Ta) variations would have little to no effect on elemental segregation, with a negligible decrease in W segregation with Al/Ti increases (and a Ta addition). Experimental testing of Phase III alloys did, in fact, confirm that an Al/Ti increase and a Ta addition would result in decreased W segregation. However, experimental testing also showed that increasing Al/Ti ratio with a Ta addition would result in decreased segregation for Al, Ti, and Ta. Segregation trends observed in the pres ent work were in agreement with a study conducted by Caldwell, which invest igated the effect of substituting 1 at% Ti for 1 at% Al in a third generation Ni-bas ed superalloy. Despite differences in Ta variations, this reduction in Al and in crease in Ti content decreased the Al/Ti ratio of the material. The decreased Al/T i ratio in the material resulted in an

PAGE 215

197 increase in W segregation, followed by small increases in Al segregation, and small decreases in Ta segregation [4]. The segregation behavior of Re was also shown to decrease when Al/Ti ratios were decreased [4]. Excellent correlation was seen between calculated, experimental, and reported trends for W, Re, and Al segr egation with increasing Al/Ti ratio or increasing Al/Ti ratio (and a Ta addition). It is then reasonable to conclude that JMatPro is a viable tool in predicting Al/Ti ratio effects on elemental segregation for refractory elements. In addition, computational, reported, and experimental results show that increased Al/Ti ratios ( with a Ta addition) result in a decrease in W segregation and increase in Re segregation. Chromium (Aluminum and Titanium) variation effects Cr effects on segregation were investi gated in Phase I by comparing four alloys with Cr contents ranging from 6. 75 wt% to 11.82 wt% Cr, in the Model 8 (6.75 wt% Cr), Model 7 (8. 44 wt% Cr), the baseline M odel (10.12 wt% Cr), and Model 6 (11.82 wt% Cr) alloys. Predicted Phase I Cr effects on segregation are seen in Figure 4-5. Predicted Cr effects were most significant for Re, followed by W, Ta, and Cr. Negligible Cr effects we re predicted for Co and Ni. Re and W segregation were predicted to decrease linearly with incr easing Cr content. An increase in Ta segregation was predicted with increasing Cr content. Increased Cr content also predicted a shift in Cr s egregation, from the in terdendritic region to the dendrite core. Cr (with Al and Ti) effects in Phase II were investigated by comparing Model Q (14 at% Cr 10 at% Al, 2 at% Ti) to the base line Model A alloy (13 at% Cr, 11 at% Al, 1 at% Ti). Predicted Phase II Cr (with Ti and Al) effects on

PAGE 216

198 segregation are seen in Figure 4-46. Cr (with Ti and Al) variations were predicted to have limited effects on element al segregation, with the exception of Re, which was expected to decrease in s egregation with increased Cr (with a Ti addition and Al reduction). Cr (with Al and Ta) effects in Phase III were investigated by comparing Alloy 3 to Alloy 5. Alloy 3 (with 12 wt% Cr) contains a 0.26 (0.5 at%) Al addition and a 1.32 wt% (0.5 at%) Ta reduction, in comparison to Allo y 5 (11 wt% Cr). Segregation trends between the Cr va riants are seen in Figure 4-84. Calculations for Phase III alloys predict ed the greatest Cr (Al and Ta) variation effect on Al, followed by W and then Ta. Al segregation was expected to shift from the dendritic core to the interdendritic region as Cr content increased (with an Al addition and Ta reduction) Additionally, increases in Cr content (with an Al addition and Ta reduction) were ex pected to decrease W segregation, and increase Ta segregation. In experimental testing of Phase III alloys, the largest Cr (Al and Ta) effects were determined for W, followed by Ta, Ti. Experiment al testing showed that an increase in Cr content (with an Al addition and a Ta reduction), increased W, Ti, and Ta segregation. It is interesting to note that Cr addit ions in a previous study, conducted on high refractory content single crystal Ni-based superalloys, decreased material segregation [4]. Cr additions were show n to increase Ta, Co, and Ni segregation but decreased Re, W, and Al segregation [4]. This same study showed that a small increase in W segregation was predict ed for Ti additions with Al reductions.

PAGE 217

199 Al substitutions for Ta were also report ed to increase the overall segregation in the alloy by increasing Re, W, Co Ta, Al, and Ni segregation [4,12]. Phase I Cr effects were in close agreement with existi ng literature with regards to Re, W, and Ta partitioning. Phase II Cr effects were in agreement with reported Re segregation trends. S egregation trends for Cr variants in Phase III correlated to experimental trends with the exception of W partitioning trends. Contradicting experimental and calculated trends for W segregation in Phase III Cr variants, may be attributed to Al reductions and Ta additions [4]. An overall good correlation between pr edicted, experimental, and reported trends was observed for Cr and Cr (Al and Ti) effects. Di screpancies between predicted and experimental Cr (Al and Ta) effects were only seen for W segregation, which was attributed to t he combined influence of Cr, Al, and Ta variations in Phase III. It is reasonable to conclude that JMatPro is a useful program to predict general segregation trends with respect to Cr variations. Moreover, Cr additions may prove a vi able method to decr ease segregation by reducing the partitioning of t he heavy elements Re and W. Rhenium (and Tantalum, Tungsten) variation effects The effects of Re (with Ta, Al, and W) variations on microstructural stability were investigated in Phase I by com paring the baseline Model (3.02 wt% Re), Model 15 (1.5 wt% Re, 7.46 wt% W), and Model 16 (0 wt% Re, 7.34 wt% Ta, 7.46 wt% W, 5.45 wt% Al) alloys. Segr egation behavior trends calculated for Re variants (with Ta, Al, and W variations) are seen in Figure 417. A Re addition with a W reduction was predicted to result in a small increase in Re segregation.

PAGE 218

200 Re and Ta additions with Al and W reducti ons were predicted to decrease Co and Cr segregation. Re effects in Phase II were investigated using 3 alloy groups with varying Re concentrations from 0 to 3.02 wt% Re. Group 1 compar ed Model B (0 wt% Re), Model C (1.5 wt% Re), and the bas eline Model A alloy (3.02 wt% Re). Group 2 compared Model D (3.02 wt% Re), Model E (2.28 wt% Re), Model F (1.5 wt% Re), and Model G (0 wt% Re). Group 3 compared Model H (3.02 wt% Re), Model I (2.28 wt% Re), Model J (1.5 wt % Re), and Model K (0 wt% Re). Re effects predicted in Phase II are seen in Figure 4-42. In creasing Re content was predicted to decrease W segregation. Increasing Re concentrations were predicted to linearly shift Cr segregati on from the dendritic core to the interdendritic region. Re effects in Phase III were evaluated using Alloy 4 and Alloy 5, with 3.02 wt% (1 at%) Re and 0 wt% (0 at%) Re, re spectively. Segregation trends for Re variations are seen in Figure 4-85. Calcul ated Re variation effects were the most notable for Re and W, followed by Cr, and then Ti. Re additions were predicted to result in a high kRe,calc value in Alloy 4 (reflecting intensified segregation to the dendrite core). Re additions were predi cted to increase Ti segregation and decrease W segregation. Cr was expected to shift segregation directions from the dendritic core to the interdendritic region as Re content increased. Increasing Re content in experimental test ing of Phase III alloys resulted in significant changes to the segregation of Re, followed by typical interdendritic elements Ta, Ti, and Al, and then Ni. Increasing Re content resulted in the

PAGE 219

201 increase of Re, Ta, Al, and Ti segr egation, and produced a decrease in Ni segregation. Previous studies have reported in creased elemental segregation during dendritic solidification, when Re content is increased in an alloy [11,13,44,50]. In a study conducted by Cadwell, Re additi ons resulted in increased segregation for all elements in an alloy. On the whole, predicted, experim ental, and reported data showed an increase in overall material segregation with increasing Re content. Calculated and experimental Re effects, showed good correlation for Re and Ti segregation trends, even though discrepancies in Cr segr egation trends were observed. It is reasonable to conclude that JMatPro equilibr ium calculations allow the prediction of overall Re effects on material segregation. Even though this addition enhances tem perature capabilities in advanced single crystal superalloys, increased elemental segregation may result in increased defect formation [64]. The over all increase in elemental partitioning with Re additions, could also result in t he formation of undesirable phases during elevated temperature exposures [25]. Intensified material segregation in Re containing alloys may lead to the a need for higher temperature or longer tim ed solution heat treatments to achieve fully homogenized structures and avoid the formation of deleterious phases. This was the case in the present work, whic h exemplified residual dendritic structure in the Re-containing Alloy 4 after soluti on heat treatment (Figure 4-59 (b)), not observed in any other Phase III alloy. The residual dendritic structure identifies

PAGE 220

202 some degree of residual segregation in t he Re containing alloy. The partially homogenized structure also alludes to sl owed diffusion rates during the solution heat treating process, as compared to a ll other Phase III alloys. The Alloy 4 composition will require a higher temper ature or longer time solution heat treatment to reach full homogenization. Careful consideration must then be taken in balancing the temperature capabi lities and segregation of Re containing alloys. Experimental to Computationa l Partitioning Comparisons With respect to elemental trends discussed in the se ctions above, JMat Pro was successful in predicting elemental s egregation effects in 70% of the cases considered in this study. Partitioning c oefficients calculated for all Phase I and Phase II alloys predicted Ni, W, Co, and Re to have k calc values greater than one, which tend to partition to the dendr ite core. Re segregation was the strongest, followed by Co, W, and then Ni. Ta and Al were predicted to segregate to the interdendritic region. The partitioning of Ta was the strongest, followed by Al. Cr was predicted to s egregate to the dendritic core, with the exception of alloys with high levels of Re, Cr, or low levels of -former, where Cr was predicted to segregate to the interdendritic region. Calculations for the Phase III alloys predicted the partitioning of the Ni, Co, W, and Re elements towards the dendrite co re. The core segregation for Re was the most significant, followed by W, Co and Ni. Ta and Ti were predicted to segregate to the inter dendritic region, with k calc values less than one. Ta was predicted to segregate to the greate st extent, followed by Ti. kAl,calc and kCr,calc

PAGE 221

203 values remained near unity, favori ng neither the dendrite core nor the interdendritic region for the different alloy compositions considered (Figure 4-82). Experimental testing for all Phase II I alloys verified core partitioning tendencies for Ni, Co, Cr, W, and Re, with kexp values greater than one. The core segregation for Re was the most signi ficant, followed by W, and Co. Ni and Cr showed only slight tendencies to s egregate to the dendrite core, with kexp values that were 0.01-0.06 from unity. Experimental testing also showed that Ta, Ti, and Al segregated to the in terdendritic region, with kexp values less than one. Ti showed strong partitioning tendencies, followed by Ta and then Al. Predicted and calculated par titioning coefficients fo r the alloying elements produced almost identical trends with res pect to partitioning direction. In addition, the order in which alloying el ements increased in segregation was nearly identical for both experimental and calculated data, with the exception of Ti. The results in the calculated and expe rimental form were consistent with previous segregation studies conducted, which resulted in Co, Cr, W, and Re partitioning to the dendrite core [4,9,12,24,25] and Ta, Al, and Ti partitioning to the interdendritic regions [12,24,49]. Differences between kcalc and kexp values for the Phase III alloys ranged from -0.52 for Re to 0.28 for Al, result ing in modeling errors of over 20% for specific alloying elements. Large mode ling errors, such as these, stemmed mainly from conservative kcalc predictions, which expected lower degrees of segregation for Ti, W, and Re, as compared to what was measured

PAGE 222

204 experimentally. In contrast, calculated kCo,calc and kTa,calc values expected higher degrees of Co and Ta segregation, as compared to their respective kCo,exp and kTa,exp experimental values. Ni and Cr, with kcalc and kexp values near unity, demonstrated an excellent correlation bet ween calculated and experimental values. The segregation predictions for element al partitioning direction, either towards the dendrite core or the interdendr itic region, proved to be qualitatively consistent with the modeled values. The agreement between predicted and experimental resulted makes JMatPro a practical tool to determine general partitioning directions for alloying elem ents. Moreover, in the compositions analyzed in this study, Ni, Co W, and Re showed an over all tendency to partition to the dendrite core, while Ta, Ti, and Al t ended towards the interdendritic region. Overall, J Mat Pro was succe ssful in predicting elemental partitioning direction in 100% of the cases considered in this study. Compositional Refinement Property trends predicted in Phase I aided the refinement of the baseline Model composition. Element al variation trends predicted in Phase II studies also contributed to defining five Phase III alloy compositions, from the modified baseline Model alloy (Model A). Specific compositional adjustments to the baseline Model and Model A alloys were selected to meet key ma terial properties. Compositions modifications were to minimize the amount of TCP phases expected at equilibrium, decrease elemental segregat ion, and achieve a heat treatment window of 25 C. For homogenization treatm ents, a heat treatm ent window of at

PAGE 223

205 least 25 C was deemed sufficiently large, to allow the solutionizing of the precipitate at temperatur es that will not risk incipient melting [25]. Compositional Modifications Compositional adjustments to the baseline Model allo y included a 1 at% Cr addition used to improve hot corrosion resistance [7,45]. Although microstructural stability was shown to dec rease with Cr additions, Cr additions were also shown to decrease material segregation. A 500 ppm C addition was used to increase microstructural stability (balancing the Cr addition) and is also expected to increase low angel boundary (LAB) tolerance while reducing casting defects [11]. The 0.5 at% Co reducti on was used to increase the alloys heat treatment window, which was predicted to increase with Co reductions. A 1 at% Ti addition and 1 at% Al reduction were used to improve hot corrosion resistance and to decrease Re partitioning during soli dification [35]. The Al reduction also helped maintain a former content of 15 at%, balancing the 1 at% Ti addition. Compositional adjustments to the baseline Model A alloy (Table 3-7), used to improve upon the Phase III allo ys material properties included -former reductions of 1 to 1.5 at% to in crease microstructural stability. -former reduction were achieved through Al and Ta reductions. Al/Ti ratio reductions of 2.5 to 4.5 were used to improve hot co rrosion resistance, considered crucial in IGT applications [35]. Decreasing the Al /Ti ratios were achieved through Al reductions and Ti additions. Al/Ti reductions were also shown to reduce the marked partitioning of Re and Ta during soli dification. Re reductions of 3.02 wt% in all final alloy compositions, with the ex ception of Alloy 4, were used to improve microstructural stability. Re reductions were also used to avoid Res strong

PAGE 224

206 partitioning tendency towards the dendrite co re. A 1 wt% Cr addition to Alloy 3 was used to provide hot corrosion resistance and was also predicted to decrease Re and W partitioning. Alloy Comparisons The initial baseline Model allo y was predicted to contain a precipitate amount of 58%, and was predicted to contain a limited amount of phase (6.3 %) at 900 C. The refractory rich phase predicted in the baseline Model alloy at equilibrium conditions, may precipitate during the heat treatment or service lifetime of the material [54]. phase formation, predicted in the baseline Model alloy, may be attributed to the high Cr ( 10 wt%) content along with Re (3.02 wt%) additions, which decrease microstructural stability and favor phase formation [34]. CMSX-4, in comparison to the basel ine Model alloy, also contains a 3.02 wt% Re content and is reported to prec ipitate Re-rich TCP phases [25,34]. The baseline Model alloy compositi on was predicted to have a melting range of 67C and a heat treatment window of 15 C. The predicted heat treatment window for the bas eline Model alloy was markedly low in comparison to heat treatment windows, of approximately 55C, for aero-engine alloys such as CMSX-4 and heat treatment wi ndows going up to 62C for PWA 1483. The small 15 C heat treatment window would make it difficult to heat treat the baseline alloy composition. The segregation behavior of the baseline Model alloy predicted k calc values greater than one for Cr, Ni, W, Co, and Re The core segregation for Re was the most significant, followed by Co, W, Ni, and then Cr. Elem ents predicted to segregate to the inter dendritic region, with k calc values less than one, were Ta

PAGE 225

207 and Al. The partitioning of Ta was the strongest, followed by Al. The baseline Model alloys elemental partitioning di rections, predicted for all elements considered in this study with the exc eption of Ni and Cr, are similar to the elemental partitioning directi ons observed in CMSX-4 [54]. The modified baseline (Model A) a lloy was predicted to contain a precipitate amount of 59% and 66 wt% at 900 C and 600 C, respectively. The amount of TCP phase predicted at 900 C and 600 C was approximately 10 wt% and 7 wt%, respectively. The amount of phase predicted in the baseline Model A alloy increased with compositional modifications. A small amount of the TCP phase (approximately 3 wt%) was al so predicted at 600 C. An increase in the predicted amount of and the increased Cr cont ent in the baseline Model A alloy, may explain the decrease in micr ostructural stability. phase formation predicted in this alloy, has also been observed in the Ni-based superalloy CMSX4, being considered for IGT application [26]. The heat treatment window predicted for the baseline Model A alloy was 47 C, which is comparable to that of the common polycrystalline IGT alloy IN 738 (48 C). The alloy modifi cations incorporated, were predicted to increase the baselines heat treatment window 32 C, me eting the 25 C target temperature criteria outlined in this study. The me lting range calculated for the baseline Model A alloy was 58C. Despite compositional modifications, no changes in elemental segregation behavior were s een between the baseline Model and the modified baseline Model A alloy.

PAGE 226

208 The Phase III variants were predicted to contain precipitate amounts between 60 wt% and 64 wt% at 600 C, respectively. The precipitate amounts for all Phase III alloys were approx imately 2 wt% to 6 wt% lower than the phase amounts predicted in the baseline Model A alloy. The amount of phase predicted at 600 C ranged between 3 wt % and 7 wt%, and the amount of phase ranged between approximately 1 wt % and 4 wt%. The total amount of TCP phases in Phase III alloys decreased between 3 wt% to 9 wt% as compared to the baseline Model A alloy at 600 C. It is notable that only the Re containing alloy (Alloy 4) was predicted to contain TCP phases at 900 C. A substantial decrease in the TCP phase am ounts predicted for all Phase III alloys alludes to an increase in microstructural st ability for the modified compositions as compared to the baseline Model and M odel A alloys. The increase in microstructural stability predicted for the P hase III alloys, may be attributed to the Re and -former reductions as compared to the baseline Model A alloy, even though a high Cr content was kept between 11 to 12 wt%. This is also the case for first generation expe rimental IGT alloys CMSX-11B and CMSX-11C, which contain 12.5 wt% Cr and 14. 9 wr% Cr, respectively, but exhibit desirable longterm lives due to a lower tendency for phasia l instability [10]. Accordingly, the absence of Re in CMSX-11C and -11B pe rmits microstructural stability at temperatures where Re containing a lloys tend to form TCP phases [10]. The experimental melting ranges for Phas e III variants ranged between 41 C to 48. The experimental melting ranges for the Phase III alloys were approximately 19 C to 26 C lower t han melting range calculated for the

PAGE 227

209 baseline Model alloy (67 C), and 10 C to 17 C lower than the melting rage predicted for the baseline Model A alloy ( 58 C). Melting ranges for the Phase III variants were comparable to the melting ranges for CMSX-4 and PWA 1483 [7,54]. Experimental heat treatment windows fo r all five variants ranged between 133 C to 184. Experimental heat treatment windows for the Phase III variants greatly surpassed the heat treatment windows for PWA 1483, CMSX-4, and IN 738 [7,54]. Overall, experimental heat tr eatment windows for Phase III alloys, considerably exceeded the 25 C target temperature criteria outlined in this study, and the initial 15 C heat treat ment window predict ed for the baseline Model alloy. The segregational behavior of the Phase III alloys predicted kexp greater than one for Cr, Ni, W, Co, and Re. The core segregation for Re was the most significant, followed by W, Co, Cr, and t hen Ni. Elements predicted to segregate to the interdendritic region, with k exp values less than one, were Ta, Ti, and Al. The partitioning of Ti was the stronges t, followed by Ta and then Al. The partitioning of elements to dendrite core or interdendritic regions in Phase III alloys are similar to the elemental partitioni ng directions observed in the baseline Model, and baseline Model A al loys. It was predicted that decreased Re and Al segregation, and increased W s egregation would result from modifications to the Phase III alloys. In reality, exper imentation showed that Re, W, and Al segregation increased, while Ta and Co segregation decreased, for Phase III alloys, as compared to the predicted s egregation behavior in the baseline Model

PAGE 228

210 and baseline Model A alloys. Compared to CMSX-4, P hase III alloys displayed a large reduction in W segregation, follow ed by Re, Ta, Cr and Ni [10]. Only Co and Al, of the elements considered in this work, displayed small increased segregation in the Phase III allo ys as compared to CMSX-4 [10]. Future Development Notably, all five Phase III alloys met and surpassed the microstructural stability and heat-treating criteria set fo rth in this study. Wide heat treatment windows in all Phase III alloys were larger than what is needed for adequate solution heat treatment. Of the Phase III alloys, more sizable heat treatment windows for Alloy 1 and Alloy 2 were 133 C and 136 C, respectively. Excellent microstructural stability properties were predicted for Alloy 1 and Alloy 2. Calculations at 600 C predicted that Alloy 1 and Alloy 2 would contain phase amounts of 3 wt% and 4 wt%, respectively The amount of phase predicted for both alloys was approximatel y 1.5 wt%. Total TCP phases amounts predicted for the two alloys, were 4.5 wt% to 5.5 wt % lower than TCP phase amount predicted for baseline Model A alloy (10 wt%) at 600 C. With res pect to segregation behavior, both Alloy 1 and Alloy 2 exhibited similar elemental partitioning trends for only Ni and Co. Decreases in KTi,exp, KTa,,exp and KAl,exp of 17 %, 12%, 10%, respectively, along with 16% and 3.1% increases in KW,exp and KCr,exp were observed in Alloy 2 as compared to Alloy 1. The partitioning coefficient differenes between the two alloys, indicate an overall increase in elemental segregation for Alloy 2. From these obs ervations, Alloy 1 was identified as a possible composition that is solution heat treatable, and may exhibit the microstructural stability and segregational properties.

PAGE 229

211 Alloy 1 has a high -former content (14.5 at%) as compared to many experimental or industrial IGT alloys such as CMSX-4 (13.1 at%), CMSX-11C (12.6 at%), CMSX-11B (12.9 at%), IN 738 (9.4 at%) or PWA 1483 (11.6 at%). The increased -former content in Alloy 1 indica tes that the mate rial may exhibit high strength at elev ated temperatures. The 11 wt% Cr content in Alloy 1 prov ides a high level of hot corrosion resistance, crucial in IGT applications. Ta king into consideration microstructural stability, the Alloy 1 compos ition is compared to that of experimental first generation IGT alloys CMSX-11B and CM SX-11C. CMSX-11B and CMSX-11C contain Cr concentrations substantially larger than those in Alloy 1, and still exhibit excellent microstructural stability as compared to 2nd and 3rd generation alloys [10]. It is evident that care ful chemistry control can balance alloy composition to maximize microstructural st ability but still consider specific IGT needs. Since, in most cases, a minimu m 9.5 to 10 wt% Cr content is needed for hot corrosion resistance [10,35,50], Alloy 1 is expected to show improved hot corrosion resistance as compared to CMSX-4 (6.4 wt% Cr). An optimum level of hot corrosion resistance is also obtained with low Al/Ti rati os (wt%/wt%). Therefore, the 4.72 Al/Ti ra tio in Alloy 1, indicates and additional increase in hot corrosion resistance as compared to t he baseline Model alloy (6.2) and CMSX-4 (5.6). The high Ta concentration in Alloy 1 (8.8 wt% Ta) facilitates material strength by increasing the APB energy of the phase and improves castability, as compared to the lower Ta levels found in CMSX-4 (6.5 wt%) and PWA 1483

PAGE 230

212 (4 wt% Ta) [50]. In addi tion, the castability of alloys like PWA 1483 are in part attributed to carbon additions, which indi cate that Alloy 1 may also exhibit improved castability with a C content of 0.05 wt%, while simultaneously improving microstructural stability [10,35,50]. Overa ll, The simplified first generation SC Alloy 1 compos ition may exhibit hot corrosion resistance, high strength, and material castability needed for IGT application. To insure a high strength material, in the continuing development of Alloy 1, modifications to the existing composition can be explored. additions to Alloy 1 (with a predicted phase amount of 65 wt%), to further increase material strength are limited, since excessive amounts of the phase render the matrix prone to TCP phase precipitat ion [25]. An alternativ e approach to increasing strength is to increase the levels of solid solution strengtheners in the matrix. Alloy 1 contains a moderate level of W for solid solution strengthening, similar to that of CMSX-4, but does not contain any Re or Mo. To investigate solid solution strengt hening additions, increased levels of W could be introduced to the Alloy 1 composit ion. W additions for strength must be carefully balanced with a concern for microstructural stability, since these additions also increase the total amount of refractory elements. Another approach to increase alloy strength could be the re-introduction of Re to 1.5 wt% or 3.02 wt%. Even though Re reductions have been used in this work to increase microstructural stability and decrease elemental segregation, Re additions are recognized as one of the most effective ways to strengthen single crystal Ni-base supearlloys, and should be in vestigated. Re, however, requires a

PAGE 231

213 minimum of 4.5 wt% W to achieve a maxi mum strengthening effe ct [6]. The 5.96 wt% W content in Alloy 1 surpasses this requirement. Furthermore, relatively small Mo addi tions could be incorporated into Alloy 1 modifications for comparison to allo ys with W additions, since Mo is a less effective solid solution strengthener and is detrimental to hot corrosion resistance [44,61]. Prudent consideration of alloy composit ions can lead to the attainment of acceptable alloy stability while maintaini ng a high enough total refractory element content to achieve the desired alloy str ength. To restrict microstructural instabilities, limits can be introduced to the total concentration of solid solution strengtheners. Accordingly, increasing the strength of Allo y 1 with increased solid solution strengtheners will require a balanced approach.

PAGE 232

214 CHAPTER 6 CONCLUSIONS Elemental variation effects were invest igated in the pres ent study with the use of the JMatPro thermody namic module for calculating material properties. The conclusions are summarized as follows: Microstructural Stability 1. JMatPro microstructural stability modeling produces predictions that strictly apply to equilibrium conditions. The results can be applied to real nonequilibrium cases, as long as the kineti cs of the formation of the TCP phase is taken into consideration. 2. W concentrations in primary carbi des contribute to MC degradation and dissolution during heat treatment. 3. Ta concentrations in (Ti,Ta)C carb ides contribute to maintaining MC stablility during solution heat treatment. Phase Transformation Temperatures 1. JMatPro equilibrium calculations are not a viable tool in assessing accurate phase transformation temperatures for specific alloy compositions. 2. By applying equilibrium JM atPro calculations to IGT alloy design, five new IGT alloys were developed. All five allo ys met the 25 C target heat treatment window criteria Segregation 1. JMatPro proved to be a viable tool to determine qualitatively information on general elemental partitioning directions; either towards the dendrite core or the interdendritic region. 2. In Phase III compositions analyzed in this study, Ni, Co, W, and Re showed an overall tendency to partition to the dendrite core, while Ta, Ti, and Al tended towards the interdendritic region. Elemental Variation Effects 1. JMatPro equilibrium calculations successfully predicted general elemental trends in microstructural stability properties for C, Ti, Cr, Re, and -former

PAGE 233

215 variations. It was also shown that JMat Pro cannot be considered a viable tool to calculate Co, Ru, or W (Mo) variation e ffects on phase microstructural stability. a. C additions increase microstructural stability b. Increased Ti, Cr, Re, and -former contents decrease microstructural stability 2. JMatPro can be used in predictin g general Ta, Cr, W (and Mo), and former variation effects on phase transform ation temperatures. JMatPro cannot be considered a viable tool to calculate C, Co, Ru, or Re variation effects on phase transformation temperatures. a. Ta concentrations increase the liquidus, solidus and decrease the -solvus. Trends result in an increased heat treating window and decreased melting range. b. Cr additions decrease the solidus and solvus. c. Increasing W concentrations with corresponding Mo reductions increase the liquidus, and solidus temperatures d. Increasing -former content decreases t he liquidus, decreases the solidus, and increase the solvus. Increasing -former content decreases the heat treating window. 3. JMatPro equilibrium calculations successfully predicted general elemental trends in material segregation for -former, Al (and Ta), Ti (and Al), Ti (and Ta), Al/Ti ratio, Cr, and Re variations. JMat Pro calculations were not successful at establishing C, Co, or W (ad Mo ) trends for elemental segregation a. Increasing -former content increases elemental segregation b. Al reductions with Ta substitutions increase material segregation c. Ti additions used to substitu te for Al reductions, decrease Re segregation while increasing W and Cr segregation. d. Ti additions used as a substitu te for Ta reductions, increase W segregation and decreas e Ni segregation. e. Increased Al/Ti ratios (wit h a Ta additions) decrease W segregation, increas e Re segregation. f. Cr additions decrease segregatio n by decreasing the segregation of the heavy elements Re and W. g. Re additions increase elemental segregation. Future Development From observations, Alloy 1 was identifi ed as a possible IGT composition for further development that may exhibit mi crostructural stability and segregational properties needed for IGT application (Table 6-1). Table 6-1. Alloy 1 composition NiAlCoCrHfReTaWTiCAl/Ti Y' at% Alloy 1 wt %57.544.3711110.108.85.961.170.053.7214.5 at %58.711011.5130.030321.50.26

PAGE 234

216 CHAPTER 7 FUTURE WORK Continued Computational Modeling In this study the JMatPro thermody namic equilibrium module was used to determined phase equilibria using Gibbs energy minimization routines. For comparison, non-equilibrium simulations should be conducted us ing the JMatPro Scheil-Gulliver solidification module. T he Scheil-Gulliver solidification module performs modifications to the composit ion of the liquid to account for nonequilibrium conditions that occur during so lidification, but still does not account for back diffusion occurring in the so lid and assumes that the liquid phase undergoes complete mixing. Microstructural Stability Evaluations Equilibrium phase calculations of the alloy compositions in this study predicted the presence of the TCP (Topol ogically Close Packed) phases at probable IGT operating temperatures (i.e., Alloy 4). Even though elemental variation calculations for microstructu ral stability were compared to existing literature, elemental effect s on stability could be fully ascertained with additional laboratory testing. Extended isothermal temperature expos ures for Phase III as-cast and solution heat treated SC superalloys with incorporat ed elemental variations, could be used to investigate material propensity to precipitate TCP phases. Subsequent exposures at elevated tem peratures could be used to reveal

PAGE 235

217 deleterious phase precipitation. Microstr ucture evolution observed during various stages of long time exposure, could be characterized with basic structure analysis. Alloys could be subjected to a long-term exposure at 1000 C to simulate operating condition for turbine blade applications. The microstructure could then be compared at 500 h, 750 h, and 1000 h. Further Development of Alloy 1 The effect of solid solution strengtheni ng additions suggested for Alloy 1, in this work can be characterized to det ermine elemental impact on material strength and the formation of deleterious second phases (e.g. TCP). The results can then be used to either modify existi ng alloy composition and/or develop new alloys with improved microstructural stabili ty. Based on modification suggestions in this work, the following systems ar e recommended for examination (Table 71). Table 7-1 Recommendations in approximate wt%. NiAlCoCrHfReTaWTiCMoAl/Ti Y' at% Alloy 1 wt %57.544.3711110.108.85.961.170.053.73514.5 at %58.711011.5130.030321.50.26NiAlCoCrHfReTaWTiCAl/Ti Y' at% DA1 wt %57.54.3711110.108.87.001.170.053.7214.5 DA2 wt %57.54.3711110.108.88.001.170.053.7214.5 NiAlCoCrHfReTaWTiCAl/Ti Y' at% DA3 wt %57.54.3711110.11.58.85.961.170.053.7214.5 DA4 wt %57.54.3711110.13.028.85.961.170.053.7214.5 NiAlCoCrHfReTaWTiCAl/Ti Y' at% DA5 wt %57.54.3711110.108.85.961.170.050.53.7214.5 DA6 wt %57.54.3711110.108.85.961.170.0513.7214.5 W Variations Re VariationMo Variations

PAGE 236

218 LIST OF REFERENCES 1. Acharya, M.V. Fuchs, G.E., The Effect of Long-term Thermal Exposures on the Microstructure and Properties of CMSX-10 Single Crystal Ni-base Superalloys. Materials Science and Engineeri ng A. 2004 381: p. 143-153. 2. Broz, P., Svoboda, M., Bursik, J., Theoretical and Experimental Study of the Influence of Cr on the + Phase Field Boundary in Ni-Al-Cr system Materials Science and Engineer ing A. 2002, 325: p. 59-65. 3. Bursik, J., Broz, P., Picha, R., Microstructural and Phase Equilibria Study in the Ni-Al-Cr-W System at 1173 and 1273 K Intermetallics. 2003, 11: p. 483-490. 4. Caldwell, E., A New Method for the Modeling of Elemental Segregation Behavior and Partitioning in Single Crystal Nickel Base Superalloys Masters Thesis, University of Florida, FL., 2004. 5. Cox. D.C., Roebuck, B., Rae, C.M.F., Recrystallisation of Single Crystal Superalloy CMSX-4 Materials Science and Technology. 2003, 19: p. 440446. 6. Darolia R., Lahrman D., and Field R., Formation of Topologically Closed Packed Phases in Nickel Base Single Crystal Superalloys Superalloys 1988, TMS, Warrendale, PA. 1988: p. 255-264. 7. Davis J.R (Ed.), Heat-Resistant Materials ASM Specialty Handbook. ASM International. Material s Park. 1997: p. 225-333. 8. Dreshfield, R.L., Ashbrook R.L., Sigma Phase Formation and its Effect on Stress-Rupture Properties of IN-100 National Aeronautics and Space Administration Lewis Research Center, NASA TN D-5185, 1969: p. 9-25 9. Duhl D. and Cetel A., Advanced High Strength Single Crystal Superalloy Composition United States Patent Number 4719080, 1988. 10. Erickson, G.L. The Development of the CMSX-11B and CMSX-11C Alloys for Industrial Gas Turbine Application Superalloys. 1996: p. 45-62. 11. Erickson G., Single Crystal Nickel-Based Superalloy United States Patent Number 5366695, 1994.

PAGE 237

219 12. Fela, F.J., Influence of Chemical Compositio n Variations on the Elemental Solidification Partitioning in Nicke l-Base Single Crystal Superalloys Masters Thesis, University of Florida, FL, 2000. 13. Feng, Q., Nandy, T.K., Tin, S., Pollock, T.M., Solidification of HighRefractory Ruthenium-C ontaining Superalloys Acta Materialia. 2003, 51: p. 269-284. 14. Fuchs, G.E., A Siemens-Westinghouse Proposal : Modified Single Crystal Alloy PWA 1483 for IGT Applications University of Florida, Materials Science & Engineerin g Department, 2002. 15. Fuchs, G.E., Solution Heat Treatment Resp onse of a Third Generation Single Crystal Ni-base Superalloy Materials Science and Engineering A. 2001, 300: p. 52-60 16. Fuchs, G.E., Chin, H.A., Pratt & Whitney, An NSF/GOALI Proposal: Stability of Ni-Base Superalloys Single Crystals University of Florida, West Palm Beach. 17. Fuchs, G.E., Boutwell, B.A., Calculating Solidification and Transformation in As-Cast CMSX-10 The Journal of the Minera ls, Metals, and Materials Society. 2002, 54: p. 45. 18. Fuchs G., and Boutwell B., Modeling of the Partitioning and Phase Transformation Temperature of and As-Cast Third Generation Single Cyrstal Ni-base Superalloy Materials Science A.. 2002, 333: p. 72-79. 19. Fullagar K.P.L., Broomf ield, R.W., Hulands, M., Aero Engine Test Experience with CMSX-4 Alloy Si ngle-Crystal Turbine Blades Journal of Engineering for Gas Turbines an d Power. 1996, 118: p. 380-388. 20. Gallardo, J.M., Rodri guez, J.A., Herrera, E.J., Failure of Gas Turbine Blades WEAR. 2002, 252: p. 264-268. 21. Grosdidier T., Hazotte A., and Simon A., Precipitation and Dissolution Processes in Single Crystal Nickel-Based Superalloys Materials Science and Engineeri ng A. 1998, 256: p.183-196. 22. Harris K. and Erickson G., Single Crystal Alloy Technology United States Patent Number 4643782. 1987. 23. Hopgood A., Nicholls A., Smith G., and Martin J., Effects of Heat Treatment on Phase Chemistry and Microstructure of Single Crystal Superalloys Materials Science and Technol ogy. 1988, 4: p. 146-152.

PAGE 238

220 24. Kearsey R., Beddoes J., Jones P., and Au P., Compositional Design Consideration for Microsegregation in Single Crystal Superalloy Systems Intermetallics. 2004: p. 50 25. Khan, T., Caron, P., Evolution Ni-based Superalloys for Single Crystal Gas Turbine Blade Applications Aerospace Science and Technology. 1999, 3: p. 513-523. 26. Kong, Y.H., Chen, Q.Z., Effect of Minor Additions on the Formation of TCP Phases in Modified RR2086 SX Superalloys Materials Science and Engineering A. 2004, 356: p. 135-143. 27. Matan, N., Winand, H.M.A., Carter, P., A Coupled Thermody namic/ Kinetic Model for Diffusional Processes in Superalloys Acta Matter. 1998, 46: p. 4587-4600. 28. Miller, M.K., Babu, S.S., Burke, M.G., Comparison of the Phase Compositions in Alloy 718 Measur ed by Atom Probe Tomography and Predicted by Thermodynamic Calculations Materials Science and Engineering A. 2002, 327: p. 84-88. 29. Mousavi Anijdan, S.H., Bahrami A., A New Method in Prediction of TCP Phases Formation in Superalloys Materials Science and Engineering A. 2005, 396: p. 138-142. 30. Pollock T., The Growth and Elevated Temp erature Stability of High Refractory Nickelbase Single Crystals Materials Science and Engineering. B. 1995: p. 255-266. 31. Pollock T.M., The Growth and Elevated Tem perature Stability of High Refractory Nickelbase Single Crystals Materials Science and Engineering B. 1995, 32: p. 255-266. 32. Pollock, T.M., Murphy, W.H ., Goldman, E.H., Uram, D.L., Grain Defect Formation During Directional Solidifica tion of Nickel Base Single Crystals Superalloys, TMS,Warrendal e, PA. 1992: p. 125-134 33. Qiu, Y., Effect of the Al and Mo on the Mismatch and Morphology in Nibased Superalloys Scripta Meterialia. 1995, 33,12: p.1961-1968. 34. Rae, C.M.F., Reed, R.C., The Precipitation of Topologically Close-Packed Phases in RheniumContaining Superalloys Acta Materialia. 2001, 49: p. 4113-4125. 35. Ross, E.W., OHara, K.S., Rene N4: A First Generation Single Crystal Turbine Airfoil Alloy with Improved Oxidation Resistance, Low Angle Boundary Strength and Superior Long Time Rupture Strength Superalloys, TMS,Warrendale, PA. 1996: p. 19-25.

PAGE 239

221 36. Saunders, N., Phase Diagram Calculati ons for Ni-based Superalloys Superalloys, TMS,Warrendal e, PA. 1996: p. 101-110. 37. Saunders, N., Fahrma nn, M., Small, C.J., The Application of Calphad Calculations to Ni-Based Superalloys Superalloys, TMS,Warrendale, PA, 2000: p. 803. 38. Saunders, N., Guo, Z., Li, X., Modeling the Material Properties and Behaviour of Ni-based Superalloys Submitted to Superalloys 2004. 39. Saunders, N., Guo, Z., Li, X., Using JMatPro to Model Materials Properties and Behaviour Schille. The Journal of the Mi nerals, Metals, and Materials Society. 2003: p. 65. 40. Saunders, N., Li, X., Miodownik, A.P., Computer Modelling of Materials properties in Materials Desi gn Approaches and Experiences TMS, Warrendale, PA, 2001: p. 185. 41. Saunders, N., Li, X., Miodownik, J-Ph., Modelling of the Thermo-physical and Physical Properties Revelant to Solidification Superalloys, TMS, Warrendale, PA, 2003: p. 669. 42. Saunders, N., Miodownik, A.P., Modelling of the Thermo-physical and Physical Properties for Solidif ication of Ni-based Superalloys Liquid Metal Processing. Nancy, France. September 21-24, 2003: p. 1-5 43. Schulze, C., Feller-Kniepmeier, M., Phase Compositions and Lattice Mistif in CMSX-11B Partition Coefficien ts in Single Crystal Nickelbase Superalloys Scripta Materialia. 2001, 44: p. 731-736. 44. Sims C., Hagel W.C. The Superalloys. John Wiley & Sons. New York, N.Y.. 1972: p. 279-504. 45. Sims, C.T., Stoloff, N.S. and Hage l, W.C. Superalloys II. John Wiley & Sons. New York, N.Y: 1987: p. 98-449. 46. Small, C.J., Saunders, N., The Application of CALPHAD Techniques in the Development of a New Gas-Turbine Disk Alloy Materials Research Society Bulletin, April 1999: p. 22. 47. Sponseller, D.L., Differential Thermal Analysis of Nickel-Base Superalloys Superalloys, TMS,Warrendal e, PA, 1996: p. 259-270 48. Strunz, P., Gilles, R., Mukheriji, D., Microstructural Characterization of Single-Crystal Nickel-Base Superallo ys by Small-Angle Neutron Scattering Materials Structure. 1999, 6,2: p.91-95.

PAGE 240

222 49. Tawancy H., Abbas M., and Al-Mana A., Thermal Stability of Advanced NiBase Superalloys Journal of Materials Science. 1994, Vol. 29, No. 9: p.2445-2458. 50. Tin, S., Pollock, T.M., Phase Instabilities and Carbon Additions in SingleCrystal Nickel-base Superalloys Materials Science and Engineering A. 2003, 348: p. 111-121. 51. Tien J.K., Caulfield T.. Superalloys, Supercomposites and Superceramics. Academic Press. New York N.Y. 1989: p. 120-153. 52. Viswanathan R., Damage Mechanisms and Life Asessement of HighTemperature Components ASM International, Materi als Park, 1989: p. 415. 53. Walston W., Ross E., Pollock T., OHara K., and Murphy W., Nickel-Base Superalloy and Article wit h High Temperature St rength and Improved Stability United States Pat ent Number 5455120, 1995. 54. Wilson, B.C., Hickman, J.A., Fuchs, G.E., The Effect of Solution Heat Treatment on a Single-Cr ystal Ni-Based Superalloy The Journal of the Minerals, Metals, and Materials Society. 2003, 55,3: p. 35. 55. Wukusick C., Buchakjin L., and Darolia R., Heat Treatment for NickelBased Superalloy United States Patent Number 5100484, 1992. 56. Zhao, S., Xie, X., Smit h, G. D., Patel, S.J., Microstructural Stability and Mechanical Properties of a New Nickel Based Superalloy Materials Science and Engineering A. 2003, 355: p. 96-105. 57. Dreshfield, R.L., Ashbrook, R.L., Further Observations on the Formation of Sigma in a Nickel-Ba se Superalloy (IN-100) National Aeronautics and Space Administration Lewis Research Center, NASA TN D-5185, 1969. p. 1-30 58. Morinaga, M., Murata, Y., Yukawa, H. Alloy Design Based on Molecular Orbital Method. Materials Science Forum. 2005, 475-479: p: 3099. 59. Moringa, M., Yoshinori M., Recent Progress in Molecular Orbital Approach to Alloy Design Materials Science Foru m. 2004, 449-452: p. 37-42. 60. Ritzert, F. The Effect of Alloying on T opologically Close Packed Phase Instability in Advanced Nickel-Base Superalloy Rene N6 National Aeronautics and Space Administration Lewis Research Center, NASA TN D-5185, 1998: p. 1-13 61. Walston, W. S., Schaeffer, J. C., Murphy, W. H., A New Type of Microstructural Instability in Superalloys SRZ Superalloys 1996, TMS, 1996: p. 9-18

PAGE 241

223 62. OHara K. S. and E. Ross, Rene N6: Third Generation Single Crystal Superalloy Superalloys 1996, Superalloys, TMS,Warrendale, PA, 1996: p. 27-34. 63. Dieter, G.E., Mechanical Metallurgy Third ed .. McGraw-Hill. Boston. 1986. 64. Giamei, A.F. and B.H. Kear, On the Nature of Freckles in Nickel Base Superalloys Metallurgical Transactions. August 1970, 1: p. 2185-2192. 65. SPG Media Limited, http://www.powertechnoloogy.com/projects/sannicolas, last accessed October 10, 2005.

PAGE 242

224 BIOGRAPHICAL SKETCH Alma Stephanie Tapia was born July 1981 in El Paso, Texas. Alma Stephanie graduated from J.M. Hanks High School in May 1999. After graduation she enrolled at the University of Texas at El Paso with a full Monroe Presidential Academic Scholarship. In spring of 2003, Alma Stephanie graduated summa cum laude with a Bachelor of Science in metallurgical and materials engineering, represented the college of engineering as the banner bearer, and was named one of the Top 10 Graduates of the University of Texas at El Paso. Upon receiving her degree, she participated in an internship program with ExxonMobil Upstream in Houston, Te xas. Alma Stephanie enrolled at the University of Florida in August 2003, to pursue a graduate degree in materials science and engineering. In May of 2004, Alma Stephanie participated in an internship program with G eneral Electric Nuclear F uel in Wilmington, North Carolina. Alma Stephanie is currently scheduled to graduate with a Master of Science in Spring of 2006 and will begin wo rking for the National Aeronautics and Space Administration at the Johnson Space Center in Houston, Texas.