<%BANNER%>

Computational Design of Nickel Based Superalloys for Industrial Gas Turbine Components

Permanent Link: http://ufdc.ufl.edu/UFE0013180/00001

Material Information

Title: Computational Design of Nickel Based Superalloys for Industrial Gas Turbine Components
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013180:00001

Permanent Link: http://ufdc.ufl.edu/UFE0013180/00001

Material Information

Title: Computational Design of Nickel Based Superalloys for Industrial Gas Turbine Components
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013180:00001

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.