HIGH STRENGTH, DUCTILE WIDE GAP BRAZE JOINTS

FOR STATIONARY TURBINE COMPONENT REPAIRS

by

Warren Martin Andre Miglietti

Submitted in partial fulfillment of the requirements for the degree

Philosophiae Doctor

in the Faculty of Engineering, the Built Environment and Information Technology

University of Pretoria

April 2008
 ABSTRACT

Wide cracks in land-based Ni- or Co-base superalloy turbine components are difficult to
repair successfully using conventional welding or brazing techniques. This project examined
the feasibility of liquid phase diffusion brazing using novel Ni- and Co-base braze alloys
containing Hf or Zr as melt point depressant for the repair of wide cracks in turbine
components. An optimized braze cycle was developed and the joints were evaluated using
various metallographic techniques and mechanical tests (elevated temperature tensile tests,
creep rupture tests and low cycle fatigue tests). Microstructural examination revealed the
presence of Hf- or Zr-rich intermetallic phases (most likely Ni7Hf2 or Ni5Zr) in Ni-base braze
joints. These intermetallic compounds were, however, observed to be significantly softer than
the boride phases routinely found in commercially available braze alloys with boron as melt
point depressant. As a result, the novel wide gap brazed joints displayed excellent mechanical
properties (ranging from 80% to 100% of the base metal’s properties). The low cycle fatigue
properties of wide gap braze joints performed using a combination of MarM247 superalloy
powder and Ni-Cr-Hf or Ni-Cr-Zr braze filler metals were found to be superior to those of the
widely used Ni-Cr-B braze filler metals. Wide gap braze repair of FSX-414 Co-base
superalloy using novel MarM509/MarM509B and MarM509/Co-Hf braze alloys resulted in
high temperature tensile properties equivalent to those of weld repairs in the same parent
material (using Nozzalloy filler metal). The creep rupture and low cycle fatigue (LCF)
properties of the braze joints were superior to those of welds performed using MarM918 filler
metal.

i
 TABLE OF CONTENTS

CHAPTER 1 – LITERATURE SURVEY

THE APPLICATION OF BRAZING, WIDE GAP BRAZING AND LIQUID PHASE
SINTERING IN THE REPAIR OF NICKEL-BASE SUPERALLOY COMPONENTS FOR
GAS TURBINE ENGINES
1.1 Introduction p. 1
1.2 The brazing of superalloys and the influence of brazing parameters on p. 3
joint quality
1.3 Characterization of the eutectic phases that form in brazed joints p. 14
1.4 Repair of turbine components using brazing p. 20
1.5 Wide gap diffusion brazing p. 23
1.6 Sintering of Ni-base superalloys using braze filler metals p. 32
CHAPTER 2 – OBJECTIVES
LIQUID PHASE DIFFUSION BONDING OF NICKEL-BASE SUPERALLOY
COMPONENTS USING NOVEL BRAZE FILLER METALS
2.1 Background p. 37
2.2 Objectives p. 37
CHAPTER 3 – EXPERIMENT 1
MICROSTRUCTURAL EXAMINATION OF NOVEL BINARY EUTECTIC Ni-Hf AND Ni-
Zr BRAZE ALLOYS
3.1 Introduction p. 39
3.2 Experimental procedure p. 40
3.3 Results and discussion p. 41
3.3.1 Optimization of the braze temperature p. 41
3.3.2 Microstructures of the Ni-Hf and Ni-Zr joints after brazing at p. 42
1230C for 40 minutes
3.3.3 Microstructures of the Ni-Hf and Ni-Zr joints after brazing at p. 46
1230C for 18 hours
3.4 Conclusions p. 53
CHAPTER 4 – EXPERIMENT 2
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL PROPERTIES
OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf AND Ni-Zr BRAZE
ALLOYS AFTER SHORT PROCESSING TIMES (40 MINUTES)
4.1 Introduction p. 54
4.2 Experimental procedure p. 54
4.3 Results and discussion p. 55
4.3.1 Microstructural investigation p. 55
4.3.2 Tensile test results p. 63
4.4 Conclusions p. 66
CHAPTER 5 - EXPERIMENT 3
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL PROPERTIES
OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf AND Ni-Zr BRAZE
ALLOYS AFTER LONGER PROCESSING TIMES (4 HOURS)

ii
5.1 Introduction p. 68
5.2 Experimental procedure p. 68
5.3 Results and discussion p. 69
5.3.1 Microstructural investigation p. 69
5.3.2 Tensile test results p. 73
5.4 Conclusions p. 78
CHAPTER 6 - EXPERIMENT 4
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL PROPERTIES
OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf AND Ni-Zr BRAZE
ALLOYS AFTER EXTENDED BRAZING TIMES (12 HOURS)
6.1 Introduction p. 79
6.2 Experimental procedure p. 79
6.3 Results and discussion p. 80
6.3.1 Microstructural investigation p. 80
6.3.2 Creep rupture test results p. 88
6.4 Conclusions p. 93
CHAPTER 7 - EXPERIMENT 5
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL PROPERTIES
OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf AND Ni-Zr BRAZE
ALLOYS AFTER BRAZING FOR 12 HOURS, FOLLOWED BY SOLUTION ANNEALING
AND HOT ISOSTATIC PRESSING
7.1 Introduction p. 95
7.2 Experimental procedure p. 95
7.3 Results and discussion p. 96
7.3.1 Microstructural investigation p. 96
7.3.2 Tensile test results p. 100
7.3.3 Creep rupture test properties p. 106
7.4 Conclusions p. 107
CHAPTER 8 - EXPERIMENT 6
CHARACTERIZATION OF THE LOW CYCLE FATIGUE PROPERTIES OF LIQUID
PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf AND Ni-Zr BRAZE ALLOYS
AFTER BRAZING FOR 24 HOURS, FOLLOWED BY SOLUTION ANNEALING, HOT
ISOSTATIC PRESSING AT 1200C AND A FULL AGING TREATMENT
8.1 Introduction p. 109
8.2 Experimental procedure p. 109
8.3 LCF test results p. 111
8.4 Conclusions p. 112
CHAPTER 9 - EXPERIMENT 7
CHARACTERIZATION OF THE MECHANICAL PROPERTIES OF LIQUID PHASE
DIFFUSION BONDS PRODUCED BY MIXING A SOLID SOLUTION-STRENGTHENED
NICKEL-BASE SUPERALLOY POWDER WITH THE EUTECTIC Ni-Hf AND Ni-Zr
BRAZE ALLOYS
9.1 Introduction p. 114
9.2 Experimental procedure p. 114
9.3 Results and discussion p. 115
9.3.1 Tensile test results p. 115
9.3.2 Creep rupture test results p. 119
9.4 Conclusions p. 120

iii
CHAPTER 10 - EXPERIMENT 8
CHARACTERIZATION OF THE MICROSTRUCTURE AND PROPERTIES OF HYPO-
EUTECTIC Ni-Cr-Hf AND Ni-Cr-Hf-B ALLOYS
10.1 Introduction p. 122
10.2 Experimental procedure p. 122
10.3 Results and discussion p. 123
10.3.1 Ternary Ni-Cr-Hf alloys p. 123
10.3.2 Quaternary Ni-Cr-Hf-B alloys p. 125
10.3.3 Microhardness measurements p. 125
10.4 Conclusions p. 126
CHAPTER 11 - EXPERIMENT 9
MICROPROBE ANALYSIS OF INTERMETALLIC PHASES
11.1 Introduction p. 128
11.2 Experimental procedure p. 128
11.3 Results and discussion p. 128
11.3.1 The MarM247/Ni-Hf braze joint produced using Ni-7Cr-31Hf p. 128
braze filler (as-brazed condition
11.3.2 The MarM247/Ni-Zr braze joint produced using Ni-7Cr-13Zr p. 136
braze filler (as-brazed condition
11.3.3 The MarM247/Ni-Zr braze joint produced using Ni-7Cr-13Zr p. 143
braze filler (after an extended diffusion cycle)
11.3.4 Repeat: The MarM247/Ni-Zr braze joint produced using Ni-7Cr- p. 152
13Zr braze filler (after an extended diffusion cycle)
11.4 Conclusions p. 155
CHAPTER 12 - EXPERIMENT 10
DEVELOPMENT OF COMPLEX Ni-Cr-Hf BRAZE ALLOYS
12.1 Introduction p. 157
12.2 Experimental procedure p. 157
12.3 Results and discussion p. 157
12.3.1 Alloy PV9020 (Ni-8.5Cr-25Hf) p. 157
12.3.2 Alloy PV9023 (Ni-7.4Cr-4.7W-1.4Ti-2.3Al-0.03C-25.5Hf) p. 159
12.3.3 Alloy PV9025 (Ni-8.4Cr-5.4W-1.6Ti-6.1Al-14.5Hf) p. 159
12.3.4 Alloy PV9024 (Ni-8.3Cr-5.6Co-5.3W-4.3Al-14.2Hf) p. 162
12.3.5 Alloy PV9026 (Ni-7.9Cr-5.4Co-5W-1.5Ti-4.1Al-0.04C-20.3Hf) p. 163
12.4 Conclusions p. 164
CHAPTER 13 - EXPERIMENT 11
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL PROPERTIES
OF ADH BRAZE JOINTS IN FSX-414 COBALT-BASE SUPERALLOY NOZZLE
SEGMENTS USING NOVEL BRAZE ALLOYS
13.1 Introduction p. 165
13.2 Experimental procedure p. 167
13.3 Results and discussion p. 170
13.3.1 Mechanical test results p. 170
13.3.2 Metallurgical examination p. 175
13.4 Conclusions p. 178

iv
CHAPTER 14 – FINAL CONCLUSIONS p. 180

CHAPTER 15 – RECOMMENDATIONS FOR FUTURE WORK p. 183

REFERENCES p.184

APPENDIX A
MIGLIETTI, W.M., CURTIS, R., HALL, B., AND LAZARIN, C.M. “LIQUID PHASE p. 187
DIFFUSION BOND REPAIR OF WESTINGHAUS 501F, ROW 3 VANES”. PROCEEDINGS
OF THE ASME TURBO EXPO CONFERENCE 2000. 2000. MÜNICH, GERMANY.

APPENDIX B
MIGLIETTI, W.M., KEARNEY, J., AND PABON, L. “LIQUID PHASE DIFFUSION BOND
REPAIR OF SIEMENS V84.2, ROW 2 VANES AND ALSTOM TORNADO 2ND STAGE p. 203
STATOR SEGMENTS”. PROCEEDINGS OF THE ASME TURBO EXPO CONFERENCE
2001. 2001. NEW ORLEANS, USA.

APPENDIX C
MIGLIETTI, W.M.A. “WIDE GAP DIFFUSION BRAZE REPAIR OF CO-BASED
INDUSTRIAL TURBINE VANES”. PROCEEDINGS OF THE INTERNATIONAL BRAZING p. 217
& SOLDERING CONFERENCE. 2-5 APRIL 2000. ALBUQUERQUE, NEW MEXICO, USA.
PP. 476-485.

APPENDIX D
MIGLIETTI, W.M.A. “WIDE GAP DIFFUSION BRAZE REPAIR OF NI-BASED
INDUSTRIAL TURBINE VANES”. PROCEEDINGS OF THE 6TH INTERNATIONAL p. 228
CONFERENCE ON BRAZING, HIGH TEMPERATURE BRAZING AND DIFFUSION
WELDING. DVS BERICHTE, NO. 212. 8-10 MAY 2001. AACHEN, GERMANY. PP. 107-
112.

v
 LIST OF FIGURES

Figure 1: The cuboidal and spheroidal structure of primary and secondary ’ precipitates (10% oxalic
acid etch, magnification: 10000X)
Figure 2: Intergranular and intragranular carbide phases (10% oxalic acid etch, magnification:
1000X)
Figure 3: Centreline cracking in a joint brazed with BNi-4 (with a composition of Ni-2B-3.5Si) [5]
Figure 4: Typical joint microstructures in four filler metals. A: BNi-2; B: BNi-3; C: BNi-4; and D:
NK10. Magnification: 200X [5]
Figure 5: Variation in microstructure with decreasing gap width. Magnification: 200X [5]
Figure 6: Effect of holding time on the brazed joint microstructure. A: 10 minutes; B: 40 minutes; C:
160 minutes; and D: 640 minutes. Magnification: 200X [5]
Figure 7: Effect of brazing temperature on joint microstructure. A: 1100oC; B: 1150oC; C: 1200oC;
and D: 1250oC. Magnification: A, B and C = 200X, and D = 100X [5]
Figure 8: Microstructure of brazed joints with a 0.05 mm gap. A: 1075oC for 10 minutes; B: 1175oC
for 10 minutes; C: 1075oC for 30 minutes; and D: 1075oC for 90 minutes [6]
Figure 9: The effect of brazing time and brazing temperature on joint shear strength (brazing foil
BNi-1) [7]
Figure 10: Effect of brazing time and brazing temperature on the joint shear strength (brazing foil BNi-
2) [7]
Figure 11: Effect of brazing time and brazing temperature on the joint shear strength (brazing foil BNi-
3) [7]
Figure 12: Influence of joint gap on (a) tensile strength, and (b) percentage elongation of copper
brazed alloy steel joints [8]
Figure 13: Influence of gap width on the tensile strength (TS) and proof stress (PS) of copper brazed
mild steel joints. MS1 and MS2 are the yield points of mild steel [8]
Figure 14: The influence of gap width on the fracture toughness, as characterized by c (the crack tip
opening displacement), for copper brazed mild steel joints [8]
Figure 15: Dependence of impact strength on joint clearance for Inconel 625 brazed with Au-6 (Ni-
20.5Au-5.6Cr-2.3Fe-3.4Si-2.3B) filler metal [9]
Figure 16: Fatigue behavior of brazed joints (tested at room temperature, R = 0.1).
 brazing at 1190o C for 5 minutes, followed by aging at 710oC for 16 hours.
o o o
 brazing at 1190 C for 5 minutes and 1100 C for 10 hours, followed by aging at 710 C for

16 hours [10]
Figure 17: Hardness profiles across brazed joints in Inconel 625
Figure 18: Typical microstructure of a joint brazed with BNi-5 [11]
Figure 19: (a) Microstructure of a joint brazed with Nicrobraz 150 at 1125oC; and (b) a magnified view
showing the details of the ternary eutectic component [11]
Figure 20: (a) The constituents and component phases in a joint brazed with BNi-4; and (b) details of
the ternary eutectic and nickel silicide precipitates in the joint centreline [11]
Figure 21: (a) Microstructure of a joint brazed with BNi-1 at 1125oC; and (b) a magnified view
showing the details of the eutectic phases [11]
Figure 22: Brazed joint of BNi-2 with 4 phases present. Magnification: 200X (reduced 50% on
reproduction) [12]
Figure 23: Brazed joint of BNi-4 with 3 phases present. Magnification: 200X (reduced 50% on
reproduction) [12]
Figure 24: Brazed joint of BNi-4 with 2 phases present. Magnification: 200X (reduced 50% on
reproduction) [12]

vi
Figure 25: Brazed joint of BNi-4 with 1 phase present. Magnification: 200X (reduced 50% on
reproduction) [12]
Figure 26: Boride particles in the steel parent metal (BNi-4 braze). Magnification: 500X (reduced
50% on reproduction) [12]
Figure 27: Boride particles in the steel parent metal (BNi-4 braze). Magnification: 2000X (reduced
50% on reproduction) [12]
Figure 28: Brazed joints in two nickel-base superalloys produced with BNi-3 filler metal at 1040oC for
3 hours [14]
Figure 29: Cracks in new and repaired vanes after engine testing [17]
Figure 30: Tensile properties (UTS: ultimate tensile strength, and 0.2% YS: 0.2% proof stress) of
activated diffusion bonded butt joints in Rene 80, compared with those of the base metal
[21]
Figure 31: Creep rupture properties of activated diffusion bonded joints in Rene 80, compared with
those of the base metal and conventional brazed joints [21]
Figure 32: High cycle fatigue strengths of activated diffusion bonded butt joints in Rene 80 at 760oC
[21]
Figure 33: In the ADH repair process, a superalloy is “cast” into the crack using a lower melting point
bonding alloy [22]
Figure 34: Creep rupture curves for different width ADH joints in Rene 80 base metal [22]
Figure 35: Tensile strength data at 870oC (1600oF) for the parent metal and brazed joints in Rene 80
base metal. (‘Oxidised’ refers to a simulated engine run; ‘Strip’ refers to chemical stripping
simulating the removal of the aluminide coating; and ‘FIC’ refers to fluoride ion cleaning)
[22]
Figure 36: Tensile strength at elevated temperature for various ratios of superalloy powder to braze
powder [24]
Figure 37: Creep rupture tests performed at 850oC (base material is In738 and the wide gap braze is a
proprietary material referred to as Elniment 100) [25]
Figure 38: Elniment 10 wide gap braze, as-brazed without subsequent heat treatment. The borides
have an oblong shape [25]
Figure 39: Elniment 10 wide gap braze, as-brazed with annealing heat treatment. The borides appear
more rounded [25]
Figure 40: Percentage porosity as a function of braze alloy content [26]
Figure 41: SEM micrographs of a brazed joint produced from a mixture containing 40 wt.% Nicrobraz
150: (a) blocky boride phase (magnification: approximately 270X), and (b) Ni-Ni3B
eutectic (N). The arrow highlights the blocky boride phase precipitated within the metal
powder particles (magnification: approximately 1800X) [26]
Figure 42: SEM micrographs of a brazed joint produced from a mixture containing 40 wt.% DF4B: (a)
blocky boride phase (magnification: approximately 350X), and (b) darker ’ phase. The
arrow highlights the blocky boride phase precipitated within the metal powder particles
(magnification: approximately 1750X) [26]
Figure 43: SEM micrographs of a brazed joint produced from a mixture containing 40 wt.% BRB: (a)
blocky boride phase (magnification: approximately 225X), and (b) darker ’ phase. The
arrow highlights the blocky boride phase precipitated within the metal powder particles
(magnification: approximately 1325X) [26]
Figure 44: Braze joint of 1.25 mm clearance using sinter filler metal S1 and braze filler metal B1 (as
shown in Table 7) [28]
Figure 45: Tensile strength as a function of temperature for the base metal, and for sintered and brazed
joints with 0.5 mm clearance [27]
Figure 46: Room temperature tensile strength as a function of joint clearance for sintered and brazed
joints and for the joints containing no sinter powder [27]

vii
Figure 47: Larson Miller plot for wide gap joints produced by the Liburdi solid state sintering process
[28]
Figure 48: Liquid phase diffusion sintering: general description of diffusion solidification [29]
Figure 49: Binary Ni-Hf phase diagram [2]
Figure 50: Binary Ni-Zr phase diagram [2]
Figure 51: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 50X
Figure 52: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 100X
Figure 53: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 200X
Figure 54: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 50X
Figure 55: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 100X
Figure 56: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 200X
Figure 57: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 40 minutes. Magnification: 500X
Figure 58: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 50X
Figure 59: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 50X
Figure 60: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 100X
Figure 61: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 200X
Figure 62: Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 500X
Figure 63: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 50X
Figure 64: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 100X
Figure 65: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 200X
Figure 66: Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at 1230C
for 18 hours. Magnification: 500X
Figure 67: SEM micrograph of the Ni-Hf braze, showing the phases identified as  (labelled “grain
boundary particle”) and Ni7Hf2 (labelled “grain boundary film”)
Figure 68: SEM micrograph of the Ni-Zr braze, showing the phases identified as  (labelled “grain
boundary particle”) and Ni5Zr (labelled “grain boundary film”)
Figure 69: Configuration of tensile and creep rupture specimens: Dg = 4.6 mm; Lg = 18.3 mm; Lr =
21.8 mm; Lo = 46.5 mm; Rf = 3.2 mm; Lt = 9.5 mm; and Dt = 8 mm
Figure 70: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 50X
Figure 71: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 100X
Figure 72: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 200X

viii
Figure 73: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 200X
Figure 74: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 500X
Figure 75: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 500X
Figure 76: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 50X
Figure 77: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 100X
Figure 78: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 200X
Figure 79: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 200X
Figure 80: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 500X
Figure 81: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 40
minutes. Magnification: 500X
Figure 82: SEM micrograph of the Ni-Hf braze, showing the phases identified as  (labelled “grain
boundary particle”) and Ni7Hf2 (labelled “grain boundary film”)
Figure 83: SEM micrograph of the Ni-Zr braze, showing the phases identified as  (labelled “grain
boundary particle”) and Ni5Zr (labelled “grain boundary film”)
Figure 84: The tensile strength (UTS) and yield strength (YS) of the Ni-Hf and Ni-Zr braze joints and
the MarM247 parent metal as a function of the test temperature (each data point shown
represents the average of three tests)
Figure 85: The ductility of the Ni-Hf and Ni-Zr braze joints and the MarM247 parent metal expressed
as % elongation and % reduction in area (RA) as a function of the test temperature (each
data point shown represents the average of three tests)
Figure 86: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 50X
Figure 87: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 100X
Figure 88: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 200X
Figure 89: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 500X
Figure 90: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 50X
Figure 91: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 100X
Figure 92: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 200X
Figure 93: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 4
hours. Magnification: 500X
Figure 94: SEM micrograph of the Ni-Hf braze, showing the phases identified as  (labelled “grain
boundary particle”) and Ni7Hf2 (labelled “grain boundary film”)
Figure 95: SEM micrograph of the Ni-Zr braze, showing the phases identified as  (labelled “grain
boundary particle”) and Ni5Zr (labelled “grain boundary film”)

ix
Figure 96: The tensile strength (UTS) of the Ni-Hf and Ni-Zr braze joints and the MarM247 parent
metal as a function of the test temperature and braze time (each data point shown represents
the average of three tests)
Figure 97: The yield stress of the Ni-Hf and Ni-Zr braze joints and the MarM247 parent metal as a
function of the test temperature and braze time (each data point shown represents the
average of three tests)
Figure 98: The % elongation measured for the Ni-Hf and Ni-Zr braze joints and the MarM247 parent
metal as a function of the test temperature and braze time (each data point shown represents
the average of three tests)
Figure 99: The % reduction in area measured for the Ni-Hf and Ni-Zr braze joints and the MarM247
parent metal as a function of the test temperature and braze time (each data point shown
represents the average of three tests)
Figure 100: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 50X
Figure 101: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 100X
Figure 102: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 200X
Figure 103: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 500X
Figure 104: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 50X
Figure 105: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 100X
Figure 106: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 200X
Figure 107: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and quenching.
Magnification: 500X
Figure 108: SEM micrograph of the Ni-Hf braze, showing the phases identified as  and Ni7Hf2
Figure 109: SEM micrograph of the Ni-Hf braze, showing islands of intergranular Ni7Hf2 intermetallic
compound
Figure 110: SEM micrograph of the Ni-Zr braze, showing the phases identified as  and Ni5Zr
Figure 111: SEM micrograph of the Ni-Zr braze, showing islands of intergranular Ni5Zr intermetallic
compound
Figure 112: Larson-Miller plot at 845C for creep rupture tests performed at three applied stress levels:
345 MPa (50 ksi), 276 MPa (45 ksi) and 228 MPa (33 ksi) (where: T is temperature (F)
and t is time (hours))
Figure 113: Larson-Miller plot at 900C for creep rupture tests performed at three applied stress levels:
242 MPa (35 ksi), 186 MPa (27 ksi) and 152 MPa (22 ksi) (where: T is temperature (F)
and t is time (hours))

x
Figure 114: Larson-Miller plot at 980C for creep rupture tests performed at three applied stress levels:
138 MPa (20 ksi), 104 MPa (15 ksi) and 76 MPa (11 ksi) (where: T is temperature (F) and
t is time (hours))
Figure 115: Larson-Miller plot for the In738 base metal, the MarM247/Ni-Zr joint and the
MarM247/Ni-Hf braze joints in the solution annealed condition
Figure 116: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 50X
Figure 117: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 100X
Figure 118: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 200X
Figure 119: Ni-Hf braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 500X
Figure 120: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 50X
Figure 121: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 100X
Figure 122: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 200X
Figure 123: Ni-Zr braze dispersed between MarM247 powder particles after brazing at 1230C for 12
hours, followed by solution heat treatment at 1230C for 4 hours and a HIP cycle at 1080C
for 4 hours. Magnification: 500X
Figure 124: SEM micrograph of the Ni-Hf braze, showing the intergranular Ni7Hf2 intermetallic phase
and ’ precipitates within the MarM247 particles
Figure 125: Higher magnification SEM micrograph of the Ni-Hf braze, showing cuboidal ’ precipitates
within the MarM247 powder particles
Figure 126: SEM micrograph of the Ni-Zr braze, showing intergranular Ni5Zr intermetallic compound
Figure 127: The tensile strength of the MarM247 base metal, and the MarM247/Ni-Hf and
MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at 1230C, and after
brazing for 12 hours at 1230C, followed by solution heat treatment and a HIP cycle
Figure 128: The yield stress of the MarM247 base metal, and the MarM247/Ni-Hf and MarM247/Ni-Zr
braze joints after brazing for 40 minutes and 4 hours at 1230C, and after brazing for 12
hours at 1230C, followed by solution heat treatment and a HIP cycle
Figure 129: The percentage elongation of the MarM247 base metal, and the MarM247/Ni-Hf and
MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at 1230C, and after
brazing for 12 hours at 1230C, followed by solution heat treatment and a HIP cycle
Figure 130: The percentage reduction in area of the MarM247 base metal, and the MarM247/Ni-Hf and
MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at 1230C, and after
brazing for 12 hours at 1230C, followed by solution heat treatment and a HIP cycle
Figure 131: Larson-Miller plot for the In738 base metal and the MarM247/Ni-Hf and MarM247/Ni-Zr
joints after brazing for 12 hours at 1230C, followed by solution heat treatment (SHT) and a
HIP cycle
Figure 132: In738 plate with a 1.5 mm wide groove machined in the center

xi
Figure 133: Enlarged view of the 1.5 mm wide groove machined in the centre of the In738 plate
samples
Figure 134: LCF specimen with the braze joint in the centre of the gauge length
Figure 135: Low cycle fatigue properties of the In738 base metal, the In738/BRB braze joint, and the
MarM247/Ni-Zr and MarM247/Ni-Hf braze joints
Figure 136: Tensile strength of Haynes 230, In738, MarM247, the MarM247/Ni-Zr braze joint and the
Haynes 230/Ni-Zr braze joint as a function of test temperature
Figure 137: Yield stress of Haynes 230, In738, MarM247, the MarM247/Ni-Zr braze joint and the
Haynes 230/Ni-Zr braze joint as a function of test temperature
Figure 138: Percentage elongation of In738, MarM247, the MarM247/Ni-Zr braze joint and the Haynes
230/Ni-Zr braze joint as a function of test temperature
Figure 139: Creep rupture properties of In738, Haynes 230, the MarM247/Ni-Zr braze joint and the
Haynes 230/Ni-Zr braze joint
Figure 140: BSE image of the Ni-13Cr-15 Hf alloy showing primary  dendrites, and a fine eutectic
component consisting of  phase and the Ni7Hf2 intermetallic compound
Figure 141: BSE image of the Ni-13Cr-20 Hf alloy showing primary  dendrites, and a fine eutectic
component consisting of  phase and the Ni7Hf2 intermetallic compound
Figure 142: BSE image of the Ni-13Cr-25 Hf alloy showing primary Ni7Hf2, and a fine eutectic
component consisting of  phase and the Ni7Hf2 intermetallic compound
Figure 143: BSE image of the Ni-13Cr-25 Hf-1B alloy showing the Ni7Hf2 phase (white), Ni-rich 
(dark gray), chromium borides (black) and a possible ternary eutectic component consisting
of  phase, Cr boride, and the Ni7Hf2 intermetallic compound
Figure 144: Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-31Hf joint in
the as-brazed condition
Figure 145: (a) Secondary electron image of the Ni-Cr-Hf alloy in the as-brazed condition
(b) Enlarged view of the secondary electron image of the Ni-Cr-Hf braze shown in Figure
145(a), highlighting the location of three spot chemical analyses of the intermetallic
compound within the braze
Figure 146: EMPA maps displaying the distribution of Al, C, Ni and Cr within the MarM247/Ni-Cr-Hf
braze microstructure (as-brazed condition)
Figure 147: EMPA maps displaying the distribution of Zr, Hf and Si within the MarM247/Ni-Cr-Hf
braze microstructure (as-brazed condition)
Figure 148: EMPA maps displaying the distribution of Ru, Ta, Ti and W within the MarM247/Ni-Cr-Hf
braze microstructure (as-brazed condition)
Figure 149: EMPA maps displaying the distribution of Fe and Mo within the MarM247/Ni-Cr-Hf braze
microstructure (as-brazed condition)
Figure 150: Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-13Zr joint in the
as-brazed condition (adjacent to the interface)
Figure 151: (a) Secondary electron image of the Ni-Cr-Zr alloy in the as-brazed condition
(b) Enlarged view of the secondary electron image of the Ni-Cr-Zr braze, shown in Figure
151(a), highlighting the location of four spot chemical analyses of the intermetallic
compound within the braze
Figure 152: EMPA maps displaying the distribution of Al, C, Ni and Cr within the MarM247/Ni-Cr-Zr
braze microstructure (as-brazed condition)
Figure 153: EMPA maps displaying the distribution of Zr and Hf within the MarM247/Ni-Cr-Zr braze
microstructure (as-brazed condition)
Figure 154: EMPA maps displaying the distribution of Si, Ta, Ti and W within the MarM247/Ni-Cr-Zr
braze microstructure (as-brazed condition)
Figure 155: EMPA maps displaying the distribution of Fe and Mo within the MarM247/Ni-Cr-Zr braze
microstructure (as-brazed condition)

xii
Figure 156: Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-13Zr joint after
an extended diffusion cycle (adjacent to the interface)
Figure 157: Secondary electron image of the Ni-Cr-Zr alloy (after an extended brazing cycle).Enlarged
view of the secondary electron image of the Ni-Cr-Zr braze, shown in Figure 157(a),
highlighting the location of four spot chemical analyses of the intermetallic compound
within the braze
Figure 158: EMPA maps displaying the distribution of Al, C, Ni and Cr within the MarM247/Ni-Cr-Zr
braze microstructure (after an extended diffusion cycle)
Figure 159: EMPA maps displaying the distribution of Zr and Hf within the MarM247/Ni-Cr-Zr braze
microstructure (after an extended diffusion cycle)
Figure 160: EMPA maps displaying the distribution of Si, Ta, Ti and W within the MarM247/Ni-Cr-Zr
braze microstructure (after an extended brazing cycle)
Figure 161: EMPA maps displaying the distribution of Fe and Mo within the MarM247/Ni-Cr-Zr braze
microstructure (as-brazed condition)
Figure 162: Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-13Zr joint after
an extended diffusion cycle (adjacent to the interface)
Figure 163: Backscatter electron microprobe images of the Ni-8.5Cr-25Hf alloy in the as-cast condition
(PV9020)
Figure 164: Backscatter electron microprobe images of the Ni-7.4Cr-4.7W-1.4Ti-2.3Al-0.03C-25.5Hf
alloy in the as-cast condition (PV9023). (All compositions given as atomic %)
Figure 165: Backscatter electron microprobe images of the Ni-8.4Cr-5.4W-1.6Ti-6.1Al-14.5Hf alloy in
the as-cast condition (PV9025). (All compositions given as atomic %)
Figure 166: Backscatter electron microprobe images of the Ni-8.3Cr-5.6Co-5.3W-4.3Al-14.2Hf alloy in
the as-cast condition (PV9024). (All compositions given as atomic %)
Figure 167: Backscatter electron microprobe images of the Ni-7.9Cr-5.4Co-5W-1.5Ti-4.1Al-0.04C-
20.3Hf alloy in the as-cast condition (PV9026). (All compositions given as atomic %)
Figure 168: Schematic illustration of the first stage nozzle segment of a Frame7FA+e IGT engine
Figure 169: An example of craze-cracks and large individual thermal fatigue cracks on the outer
sidewall of a first stage nozzle segment
Figure 170: The binary Co-Hf phase diagram [2]
Figure 171: The FSX-414 test plate with a 1.5 mm wide groove machined in the centre
Figure 172: Enlarged view of the machined groove in the braze test plate
Figure 173: Photographs of: (a) the MarM509/Co-Hf ADH test plate in the as-brazed condition; and (b)
the MarM509/MarM509B ADH test plate in the as-brazed condition
Figure 174: Mechanical test sample machined from the braze test plates. The 1.5 mm wide braze joint
is located in the centre of the gauge length
Figure 175: Larson-Miller plot of FSX-414 base metal, the MarM509/MarM509B ADH braze joints,
and the Nozzalloy weld
Figure 176: Graphical representation of the LCF data determined for FSX-414 base metal, FSX-414
welded with MarM918 or Nozzalloy filler metal, FSX-414 ADH brazed with
MarM509/MarM509B, and FSX-414 ADH brazed with MarM509/Co-Hf
Figure 177: (a) Chromium oxides on the crack surface; and (b) oxide reduction on the sidewalls of the
crack after H2 cleaning
Figure 178: Craze-cracks filled with the MarM509/MarM509B braze alloy
Figure 179: Individual crack filled with the MarM509/MarM509B braze alloy
Figure 180: (a) Braze flow in a 7.5 mm deep crack; and (b) braze flow in an 11 mm deep crack
Figure 181: Photomicrograph of the interface between the braze crack repair and the base metal
Figure 182: Microstructure of the fully diffused braze repair of craze-cracks on a nozzle segment
Figure 183: Microstructure of the fully diffused braze repair of craze-cracks on a nozzle segment

xiii
 LIST OF TABLES

Table 1: Nominal chemical compositions (wt.%, balance Ni) of some Ni-base superalloys
Table 2: Influence of heat treatment on the tensile strength (1000C) of reacting brazed joints
Table 3: Nominal compositions of powders and base metal (wt.%, balance Ni)
Table 4: Chemical compositions of phases in specimens containing 40 wt.% Nicrobraz 150 braze
filler metal
Table 5: Chemical compositions of phases in specimens containing 40 wt.% DF4B braze filler
metal
Table 6: Chemical compositions of phases in specimens containing 40 wt.% BRB braze filler
metal
Table 7: Chemical compositions of the base metal and sinter powders (S1-S4) used in the
Chasteen and Metzger investigation
Table 8: Nominal chemical composition of the In738 parent metal used during the course of this
investigation (wt.%, balance nickel)
Table 9: Measured chemical compositions (wt.%) of the Ni-Hf and Ni-Zr braze alloys
Table 10: Nominal chemical composition of the MarM247 powder used in this investigation (wt.%,
balance Ni)
Table 11: Joint tensile properties measured at 21oC
Table 12: Joint tensile properties measured at 540oC
Table 13: Joint tensile properties measured at 650oC
Table 14: Joint tensile properties measured at 760oC
Table 15: Joint tensile properties measured at 870oC
Table 16: Joint tensile properties measured at 980oC
Table 17: Joint tensile properties measured at 21oC
Table 18: Joint tensile properties measured at 540oC
Table 19: Joint tensile properties measured at 650oC
Table 20: Joint tensile properties measured at 760oC
Table 21: Joint tensile properties measured at 870oC
Table 22: Joint tensile properties measured at 980oC
Table 23: Creep rupture test conditions
Table 24: Creep rupture properties at 845oC of the In738 base metal, and LPDB joints produced
with MarM247 superalloy powder and Ni-Hf or Ni-Zr braze alloy. The test samples
were produced with a joint gap of 1.5 mm to simulate a worst-case crack repair scenario
Table 25: Creep rupture properties at 900oC of the In738 base metal, and LPDB joints produced
with MarM247 superalloy powder and Ni-Hf or Ni-Zr braze alloy. The test samples
were produced with a joint gap of 1.5 mm to simulate a worst-case crack repair scenario
Table 26: Creep rupture properties at 980oC of the In738 base metal, and LPDB joints :
produced with MarM247 superalloy powder and Ni-Hf or Ni-Zr braze alloy. The test
samples were produced with a joint gap of 1.5 mm to simulate a worst-case crack repair
scenario
Table 27: Joint tensile properties measured at 21oC
Table 28: Joint tensile properties measured at 540oC
Table 29: Joint tensile properties measured at 870oC
Table 30: Joint tensile properties measured at 980oC
Table 31: The influence of a HIP cycle on the creep rupture properties of the In738 base metal and
the MarM247/Ni-Hf and MarM247/Ni-Zr joints (1.5 mm joint gap)

xiv
Table 32: Nominal chemical composition of the Haynes 230 powder used in this investigation
(wt.%, balance Ni)
Table 33: Creep rupture test conditions
Table 34: Joint tensile properties measured at 21oC
Table 35: Joint tensile properties measured at 95oC
Table 36: Joint tensile properties measured at 315oC
Table 37: Joint tensile properties measured at 540oC
Table 38: Joint tensile properties measured at 650oC
Table 39: Joint tensile properties measured at 870oC
Table 40: Joint tensile properties measured at 980oC
Table 41: Creep rupture properties of In738, Haynes 230 and the MarM247/Ni-Zr and Haynes
230/Ni-Zr joints (1.5 mm joint gap)
Table 42: Chemical composition (wt.%) of the eutectic component observed in the ternary Ni-Cr-
Hf alloys
Table 43: Measured liquidus and solidus temperatures of the ternary Ni-Cr-Hf alloys
Table 44: Measured microhardness values of various phases observed in the ternary and quaternary
alloys
Table 45: The chemical compositions of various phases within the ternary and quaternary alloys
(all compositions given as wt.%)
Table 46: The results of the EMPA scan of the MarM247/Ni-7Cr-31Hf braze joint in the as-brazed
condition. (All compositions given in wt.%)
Table 47: Quantitative EPMA data (taken at 15 keV and 80 nA with a focused spot beam) at three
locations (indicated in Figure 145(b)) within the intermetallic phase. (Percentage by
weight)
Table 48: The results of the EMPA scan of the MarM247/Ni-7Cr-13Zr braze joint in the as-brazed
condition. (All compositions given in wt.%)
Table 49: Quantitative EPMA data (taken at 15 keV and 80 nA with a focused spot beam) at four
locations within the intermetallic phase (indicated in Figure 151(b)). (Percentage by
weight)
Table 50: The results of the EMPA scan of the MarM247/Ni-7Cr-13Zr braze joint after an
extended diffusion cycle. (All compositions given in wt.%)
Table 51: Quantitative EPMA data (taken at 15 keV and 80 nA with a focused spot beam) at four
locations within the intermetallic phase (indicated in Figure 157(b)). (Percentage by
weight)
Table 52: The results of the second EMPA scan of the MarM247/Ni-7Cr-13Zr braze joint after an
extended diffusion cycle. (All compositions given in wt.%)
Table 53: Chemical compositions (wt.%) of the novel “near-eutectic” alloys examined during the
course of this investigation
Table 54: EMPA results of the individual phases within the eutectic component observed in alloy
PV9020 (wt.%)
Table 55: Nominal chemical compositions of the alloys considered in this investigation (wt.%)
Table 56: Tensile properties at 870C of FSX-414 base metal, FSX-414 welded with Nozzalloy,
FSX-414 brazed with MarM509/MarM509B, and FSX-414 brazed with MarM509/Co-
Hf. The test samples were produced with a joint gap of 1.5 mm to simulate a worst-case
crack repair scenario
Table 57: Low cycle fatigue (LCF) data for FSX-414 base metal, FSX-414 welded with MarM918,
FSX 414 welded with Nozzalloy, FSX-414 ADH brazed with MarM509/MarM509B,
and FSX-414 ADH brazed with MarM509/Co-Hf

xv
 ACKNOWLEDGEMENTS

Thank you to my wife Portia who had to endure me working many weekends to undertake the
experiments and write up the thesis. Thank you to Kevin Pelletier and Jan Williams of Dirats
Laboratories for the Scanning Electron Microscope analyses and tensile/stress rupture testing,
respectively. In addition thanks to the people at Metcut and Westmoreland for the low cycle
fatigue testing. Thanks to Joe Foroughi of the M&P lab and Louis Peluso of GRC for help
with the Electron Microprobe analyses and a tremendous thanks to Ann Ritter of GRC for the
help in developing and analyzing the phases that formed in the hypo-eutectic alloys and for
help in developing and characterizing the microstructures of novel “near-eutectic” braze
alloys in the as-cast condition

Thank you also to Prof Lugscheider of the RWTH Institute in Aachen, Germany, for
introducing me to the concept of using Hf as a melt point depressant in Ni-base braze filler
alloys, and similarly thank you to Dr Khorunov of the Paton Institute in Kiev, Ukraine, for
introducing me to the concept of using Zr as a melt point depressant in Ni-base braze filler
alloys.

xvi
 CHAPTER 1 - LITERATURE SURVEY
THE APPLICATION OF BRAZING, WIDE GAP BRAZING AND LIQUID PHASE
SINTERING IN THE REPAIR OF NICKEL-BASE SUPERALLOY COMPONENTS
FOR GAS TURBINE ENGINES

1.1) Introduction

A gas turbine engine, be it for an aircraft application or for power generation, typically ingests
air from the atmosphere. The air is then compressed by the compressor and mixed with fuel.
The mixture is combusted in a combustion chamber, usually referred to as the combustor can
or the burner can, following which the hot gas is expanded through the turbine section,
producing thrust. In essence, thermal energy is converted to mechanical energy.

Components within the turbine section of a gas turbine engine are generally identified as
‘blades’ and ‘vanes’, with ‘blades’ referring to rotating components and ‘vanes’ to stationary
components. Some original engine manufacturers (OEM’s), such as General Electric, refer to
blades as ‘buckets’ and vanes as ‘nozzles’, while a number of European OEM’s, such as ABB
and Alstom, further confuse matters by referring to vanes as ‘stator blades’.

Blades and vanes are not only exposed to high service temperatures (typically 1000oC to
1200oC), but also experience severe stresses and strains, oxidation and corrosive attack during
service. As a result of the high temperatures that the turbine section is exposed to during
service, the blades and vanes are usually manufactured from materials referred to as
superalloys. By definition a superalloy is a material that can operate at 75% and higher of its
melting point, while maintaining adequate mechanical strength. Steels have little mechanical
strength within the temperature range mentioned above and are therefore not used in the
manufacture of turbine components.

Materials used in the production of turbine components include iron-base superalloys, cobalt-
base superalloys and nickel-base superalloys. In the early days of engine manufacture, iron-
base alloys were widely utilized, but nowadays the majority of blades and vanes are produced
from nickel- and cobalt-base superalloys. These alloys exhibit outstanding mechanical
strength over the entire operating temperature range in a gas turbine engine and maintain good
mechanical integrity, including high creep rupture strength, over long periods of time (up to
20000 hours). The nickel- and cobalt-base superalloys have close packed face centred cubic
(FCC) austenitic crystal structures. Of significance in determining the properties of
superalloys is the solubility of alloying elements in the austenite matrix and the precipitation
of intermetallic compounds (such as gamma prime, ’) and various carbide phases. The
cobalt-base superalloys are generally solid solution strengthened, whereas the nickel-base
alloys used in engines today are typically ’ strengthened. Gamma prime, ’, is a face centred
cubic intermetallic compound of composition Ni3(Al,Ti). Primary ’ exhibits a cuboidal
structure, whereas secondary ’ has a more spheroidal morphology (as demonstrated in
Figure 1). Although the ’ particle size varies depending on the heat treatment that the alloy
receives, the typical size of the primary ’ particles is 0.5 m to 0.8 m, and that of the
secondary ’ phase is between 0.1 m and 0.2 m. The two main heat treatments performed
in industry are the solution heat treatment, which is performed above the ’ solvus
temperature, and the aging heat treatment, which induces the precipitation of the cuboidal and

1
spheroidal ’ precipitates shown in Figure 1. Carbide phases may form intergranularly or
intragranularly, as shown in Figure 2. Intergranular carbides prevent grain boundary sliding.

Figure 1 – The cuboidal and spheroidal structure of primary and secondary ’ precipitates
(10% oxalic acid etch, magnification: 10000X).

Intergranular
carbides

Intragranular
carbides

Figure 2 - Intergranular and intragranular carbide phases (10% oxalic acid etch,
magnification: 1000X).

2
Table 1 displays the typical chemical compositions and commercial trade names of
superalloys used in the casting of blades and vanes. These alloys usually consist of at least
fourteen major alloying elements, in addition to some minor trace elements. Aluminium and
titanium are added to form the ’ phase, while chromium is added for corrosion resistance and
to promote the formation of carbides. Other carbide formers include molybdenum, tungsten,
vanadium, niobium, tantalum and titanium. Nickel, cobalt, iron, molybdenum, tungsten and
chromium may be found in the austenite () matrix of the alloy.

Table 1 - Nominal chemical compositions (wt.%, balance Ni) of some Ni-base superalloys.
Alloy C Co Cr Mo Al W Ta Nb Ti Zr Fe Hf B
In738 0.17 8.5 16.0 1.7 3.4 2.6 1.7 2.0 3.4 0.1 0 0 0.001
In939 0.15 19.0 22.5 0 1.9 2.0 1.4 0 3.7 0.1 0 0 0.001
Rene 80 0.17 9.5 14.0 4.0 3.0 4.0 0 0 5.0 0.03 0 0 0.001
MarM247 0.15 10.0 8.25 0.7 5.5 10.0 3.0 0 1.0 0.05 0.5 1.5 0.001

1.2) The brazing of superalloys and the influence of brazing parameters on joint quality

Fusion welding served as the main fabrication technique for joining these materials for many
years. The solid solution family of nickel-base superalloys has been extensively welded using
arc welding processes (mainly gas tungsten arc welding and plasma arc welding) and high
energy density welding processes (such as electron beam welding). The solid solution alloys,
however, do not contain significant amounts of Al and Ti, hence welding problems are
seldom encountered. With the addition of Al and Ti (see Table 1 above), the ’ strengthened
alloys become unweldable due to heat-affected zone (HAZ) microfissuring and fusion zone
(FZ) cracking (including liquation cracking, ductility dip cracking and strain age cracking).

As a result of these welding-related problems, designers and welding engineers examined

brazing as an alternative method of joining these “difficult-to-weld” superalloys. Brazing is
defined as a process that produces bonding of materials at the brazing temperature in the
presence of a filler metal with a liquidus above 450oC and below the solidus of the parent
metal. Unlike welding, the parent metal does not melt. The filler metal is distributed between
the closely fitted mating surfaces of the joint through capillary action.

The brazing parameters have been shown to have a significant influence on the properties of
the brazed joint. Five main factors, listed below and considered in more detail in the
remainder of this chapter, affect the quality of a brazed joint. These factors are:

 the braze gap and joint design,

 the braze filler metal and its flow characteristics,
 the braze temperature and the holding time at temperature,
 the surface preparation and cleanliness of the joint, and
 the rate and source of heating (for example, induction brazing versus furnace brazing).

In the late 1960’s, OEM’s mainly used gold-base braze filler metals to join airfoils to stator
rings to form nozzles. Au/Ni and Au/Cu alloys were used because Au flows well into joints
at elevated temperatures. The braze temperatures for these filler metals varied from 900oC to
1100oC, depending on the alloying elements present. The Au-based filler metals resulted in

3
very ductile joints, but with the price of Au increasing, OEM’s started replacing Au with less
expensive precious metals such as palladium.

Regardless of the type of precious metal used, however, the high prices on the world market
prompted inventors to develop Ni-base braze filler metals with alternative melt point
depressants. In the early 1970’s, three main melt point depressants were introduced, namely
boron, silicon and phosphorus. It should be noted that the melting point of pure Ni is 1455oC.
In 1976, Sheward and Bell [1] evaluated several braze filler metals doped with P for joining
nuclear parts and components. Boron was considered undesirable in this application because
of its high neutron capture cross section and the fact that it transmutes to He. During long
term exposure this results in voids and possibly swelling within the joints. Shear test results,
corrosion test results and oxidation test results revealed that acceptable braze joints can be
achieved on using a braze filler metal with composition Ni-25Cr-10P. The liquidus
temperature was in the range of 927oC to 950oC. According to Lee and Nash, quoted in the
ASM Handbook [2], a eutectic point exists in the Ni-P system at 11% P and a temperature of
870oC. Many commercial manufacturers of braze filler metals have subsequently
successfully produced braze filler metals with P contents ranging from 7% to 14%.

It is important to note that one of the most important factors in achieving a high strength joint
is the joint gap. When brazing Ni-base superalloys, studies by Sulzer Metco (a popular
supplier and manufacturer of Ni-base filler metals) have shown that the joint gap should be
between 0.03 mm and 0.25 mm. Because of the difficulty in maintaining a joint gap of close
tolerance, many manufacturers and fabricators intensely dislike the brazing process. If the
recommended joint gap is exceeded, brittle centreline eutectic phases can form. In joints
produced using Ni/P filler metal, wide gaps promote the formation of brittle Ni3P and Ni5P2
phases. In joints with large clearances, soaking for long periods of time at elevated
temperature promotes the diffusion of the melt point depressant, such as P, into the parent
metal, preventing the formation of brittle phases. This has the distinct disadvantage that braze
times on the order of several hours are required to achieve this, which may be uneconomical.
Nevertheless, diffusing melt point depressants into the parent metal is often performed when
high strength joints are required in turbine components. This process is referred to as
diffusion brazing, and will be considered in more detail later.

Ni-base filler metals with Si as melt point depressant display a eutectic point at 30% Si
according to Nash and Nash, quoted in the ASM Handbook [2]. Braze filler metals with 30%
Si, however, result in brittle joints, as Si has a large atomic diameter and is not readily
diffused into the parent metal. Hypoeutectic compositions containing between 8 and 12% Si
are commercially manufactured and produce sound, reliable joints in the gas turbine industry.
Ni4Si and Ni3Si silicide phases have, however, been found in joints with large clearances,
resulting in severe embrittlement.

Similarly, for Ni-base filler metals with B as melt point depressant, Liao and Spear [2]
reported a eutectic point in the Ni-B system at 11% B, with a eutectic temperature of
1018oC. However, braze filler metals containing 11% B also result in brittle joints.
Nevertheless, the small atomic diameter of B allows rapid diffusion into the parent metal, and
joints with gaps as wide as 0.3 mm to 0.6 mm have been brazed successfully using filler
metals with B as melt point depressant. A hypoeutectic composition of 3.6% B,
corresponding to a melting temperature of 1100oC, is widely used, because it is easier to
diffuse a small amount of B into the parent metal. Where filler metals containing 3.6% B are
used, the joint is almost indistinguishable from the parent metal after brazing. Joints

4
produced with filler metals containing between 2% and 5% B are commercially manufactured
and produce sound, reliable joints in the gas turbine industry. Nevertheless, if the joint gap is
too wide, a variety of boride phases can form. Some of the brittle phases found to date
include Ni3B, Ni2B and NiB.

In the majority of applications, braze filler metals do not contain a single melt point
depressant, but a combination. The most popular Ni-base braze filler metals contain both B
and Si. The Aerospace Materials Specification (AMS) introduced four Ni-base braze filler
metals containing both B and Si. These are:

a) AMS 4776, with a composition Ni-14Cr-3.3B-4.5Si-4Fe and a liquidus temperature of

1077oC,
b) AMS 4777, with a composition Ni-7Cr-3.2B-4.5Si-3Fe and a liquidus temperature of
1000oC,
c) AMS 4778, with a composition Ni-3.2B-4.5Si and a liquidus temperature of 1038oC, and
d) AMS 4779, with a composition Ni-2B-3.5Si and a liquidus temperature of 1066oC.

These filler metals are the most popular braze filler metals used in the aircraft industry and are
produced by many suppliers throughout the world.

During the course of an investigation into Ni-base braze filler metals containing Si and P as
melt point depressants, Lugscheider and Kim [3] considered four braze filler metals with the
following compositions:

1) Ni-24Cr-15Fe-11.5Si-0.6P 2) Ni-24Cr-15Fe-8.5Si-3.0P
3) Ni-24Cr-15Fe-7.5Si-0.6P 4) Ni-24Cr-15Fe-11.5Si-1.8P

Brazing experiments were conducted to enable the brazing parameters (particularly

temperature and time) to be optimized with regards to a suitable joint microstructure, and to
determine the maximum allowable joint gap. Mechanical tests and corrosion tests revealed
that the joints possessed high tensile strengths and excellent corrosion resistance.

Lugscheider and Partz [4] examined the effect of two filler metal compositions on the
maximum allowable joint gap during the brazing of Ni-base superalloys. BNi-2 has a
composition of Ni-7Cr-3.2B-4.5Si-3Fe and contains Si and B as melt point depressants, and
BNi-5 has a composition of Ni-19Cr-10Si and contains only Si as melt point depressant. The
authors reported that when BNi-2 was used as filler metal, the maximum allowable joint gap
ranged between 0.03 mm and 0.04 mm for a brazing time of 10 minutes at a temperature of
between 1010oC and 1120oC. Increasing the brazing time to 60 minutes, however, increased
the maximum allowable joint gap to between 0.05 mm and 0.1 mm. When BNi-5 was used as
filler metal, the maximum joint gap was between 0.018 mm and 0.022 mm for a brazing time
of 10 minutes at a temperature of 1150oC to 1210oC. Increasing the brazing time to 60
minutes increased the maximum allowable joint gap to between 0.035 mm and 0.05 mm. The
authors also reported that brazing for either 10 minutes or 60 minutes at temperatures between
980oC and 1100oC reduced the maximum joint gap to 0.01 mm. The narrow gap required for
producing quality joints limits the wider application of braze filler metals containing P as melt
point depressant. In this author’s opinion, the most popular melt point depressants used for
joining Ni-base superalloy turbine components are those containing B or Si, or a combination
of these two elements, as melt point depressants.

5
In a similar study, Johnson [5] examined the hard, brittle centreline eutectic phases in a joint
brazed with BNi-4 filler metal (with a composition of Ni-2B-3.5Si), shown in Figure 3. In
Figure 3A, the wide centreline eutectic component contains multiple cracks. Even the
narrow, but continuous, centreline eutectic phase evident in Figure 3B promotes crack
propagation. The centreline phases are probably a combination of boride and silicide phases,
although this was not stated in the document.

Centreline eutectic
phase

Figure 3 – Centreline cracking in a joint brazed with BNi-4 (with a composition of Ni-2B-
3.5Si) [5].

Figure 4 shows the different microstructures that exist within braze joints produced using
four braze filler metals with compositions shown below. The brittle centreline phases are
clearly visible.

BNi-2: Ni-7Cr-3.2B-4.5Si-3Fe BNi-3: Ni-3.2B-4.5Si

BNi-4: Ni-2B-3.5Si NK10: Ni-15Cr-1.5Fe-3.6B

Figure 4 – Typical joint microstructures in four filler metals. A: BNi-2; B: BNi-3; C: BNi-4;
and D: NK10. Magnification: 200X [5].

Figure 5 depicts the manner in which the joint microstructures vary as a function of the joint
gap. In Figure 5E, the joint gap is wide and a continuous eutectic centreline phase is evident.

6
When the joint gap is narrow, however, as shown in Figure 5H, only isolated eutectic phases
are present. Continuous centreline eutectic phases facilitate crack propagation.

Figure 5 – Variation in microstructure with decreasing gap width. Magnification: 200X [5].

As shown in Figure 5, the joint gap needs to be controlled in order to ensure brazed joints
with good mechanical properties. It is, however, often difficult to control the joint gap in
practice. In the event of a wide gap, a longer brazing time can be used to remove continuous
centreline phases. This is illustrated in Figure 6 for brazing times between 10 minutes and
640 minutes. As the brazing time increases, the centreline eutectic component shrinks, and
would probably become discontinuous at holding times exceeding 640 minutes.

10 minutes 40 minutes

160 minutes 640 minutes

Figure 6 – Effect of holding time on the brazed joint microstructure. A: 10 minutes; B: 40

minutes; C: 160 minutes; and D: 640 minutes. Magnification: 200X [5].

In the event of a wide joint gap, temperature can also be used to render a braze joint free of
continuous centreline phases. Figure 7 shows the effect of increasing temperature on the
joint microstructure. The microstructures shown in the first three micrographs (A, B and C in
Figure 7) are comparable with those in Figure 6 showing the effect of increasing braze time.

7
At 1250oC, however, the microstructure changes considerably. As a result of isothermal
diffusion, elements from the parent metal also diffuse into the joint at the high braze
temperature and a new type of eutectic forms in the joint. This new eutectic is believed to be
less brittle than the more typical boride or silicide centreline phases.

1100C 1150C

1200C 1250C

Figure 7 – Effect of brazing temperature on joint microstructure. A: 1100oC; B: 1150oC; C:

1200oC; and D: 1250oC. Magnification: A, B and C = 200X, and D = 100X [5].

Sakamoto et al [6] performed a similar investigation in 1989, and found that increasing the
time at the brazing temperature or increasing the brazing temperature can result in joints
without eutectic phases. Figure 8A displays the characteristic continuous brittle eutectic
centreline component in a 0.05 mm wide joint, held for 10 minutes at 1075oC. On increasing
the braze temperature to 1175oC, a braze joint free of eutectic phases was obtained, as seen in
Figure 8B. Similarly, Figure 8C shows irregular centreline eutectic phases in a 0.05 mm
wide joint, held for 30 minutes at 1075oC. On increasing the holding time to 90 minutes at
1075oC, a braze joint free of eutectic phases was obtained (Figure 8D).

It is evident from the published research described above that the type of braze filler metal,
the joint gap, the time at the brazing temperature, and the brazing temperature are important
variables associated with the braze process.

These variables also play a major role in determining the mechanical properties of the braze
joint. Miyazawa and Ariga [7] reported the effect of brazing time and temperature on the
shear strength of brazed joints produced using BNi-1 braze foil (Ni-14Cr-3.3B-4.5Si-3Fe).
As demonstrated in Figure 9, the shear strength of the joint increased with an increase in
braze time and temperature. After 60 minutes at 1275oC the brazed joint displayed the
highest shear strength (320 MPa).

8
Similarly, when using BNi-2 braze foil (Ni-7Cr-3.2B-4.5Si-3Fe), the shear strength of the
joint increased with an increase in braze time and temperature, as shown in Figure 10. After
60 minutes at 1275oC the brazed joint displayed the highest shear strength (320 MPa). When
joining steel to nickel, it is therefore more desirable to use BNi-2, rather than BNi-1, because
the same shear strength can be achieved at a lower brazing temperature (1150oC compared to
1275oC).

Figure 8 – Microstructure of brazed joints with a 0.05 mm gap. A: 1075oC for 10 minutes; B:
1175oC for 10 minutes; C: 1075oC for 30 minutes; and D: 1075oC for 90 minutes [6].

400

300
Shear strength (MPa)

200

100

1175°C
1225°C
1275°C
0
0 30 60
Brazing time (minutes)

Figure 9 – The effect of brazing time and brazing temperature on joint shear strength (brazing
foil BNi-1) [7].

9
The same trend is evident on using BNi-3 braze foil (Ni-3.2B-4.5Si), Figure 11. After 60
minutes at 1190oC the brazed joint displayed the highest shear strength of 250 MPa, 70 MPa
less than that achieved when using BNi-1 or BNi-2 braze filler metals. The lower shear
strength can be attributed to the absence of Cr and Fe in the BNi-3 filler metal. The joint
therefore relies on diffusion of Cr from the nickel-base parent metal for solid solution
strengthening, and a brazing time of 60 minutes is too short to allow sufficient Cr diffusion
from the parent metal into the joint.
400

300
Shear strength (MPa)

200

100

1050°C
1100°C
1150°C
0
0 30 60
Brazing time (minutes)

Figure 10 - Effect of brazing time and brazing temperature on the joint shear strength
(brazing foil BNi-2) [7].

400

300
Shear strength (MPa)

200

100

1090°C
1140°C
1190°C
0
0 30 60
Brazing time (minutes)

Figure 11 - Effect of brazing time and brazing temperature on the joint shear strength
(brazing foil BNi-3) [7].

10
In a similar study undertaken by Kavishe and Baker [8], an increase in tensile strength with
decreasing gap width was reported for copper brazed alloy steel joints, as shown in Figure
12(a). Figure 12(b) demonstrates a corresponding increase in the percentage elongation with
an increase in gap width.

Figure 12 – Influence of joint gap on (a) tensile strength, and (b) percentage elongation of
copper brazed alloy steel joints [8].

These authors also reported that the proof stress of the joints was slightly lower than the yield
strength of mild steel for narrow gaps, but that the tensile strength was significantly greater
than the yield strength of mild steel, as seen in Figure 13. The fracture toughness of the steel
brazed joints decreased with a decrease in joint gap, as shown in Figure 14. This is
associated with the increase in joint strength.

450

400

TS
Stress (MPa)

350

300
MS2
MS1
250
PS

200

0.0 0.2 0.4 0.6 0.8 1.0 1.2

Gap width (mm)

Figure 13 – Influence of gap width on the tensile strength (TS) and proof stress (PS) of
copper brazed mild steel joints. MS1 and MS2 are the yield points of mild steel [8].

11
 180

160

140

c (µm)
120

100

80

60
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Gap width (mm)

Figure 14 – The influence of gap width on the fracture toughness, as characterized by c (the
crack tip opening displacement), for copper brazed mild steel joints [8].

Lugscheider and Krappitz [9] studied the influence of gap width on the impact strength of
Inconel 625 joints brazed with Au-6 filler metal. These authors reported a maximum notch
impact energy at a joint gap of 12 m, as illustrated in Figure 15.
20
maximum brazing clearance

15
Impact strength (J)

10

Brazing cycle: 1060°C for 10 minutes

1000°C for 60 minutes
0
0 25 50 75 100
Brazing clearance (µm)
Figure 15 – Dependence of impact strength on joint clearance for Inconel 625 brazed with
Au-6 (Ni-20.5Au-5.6Cr-2.3Fe-3.4Si-2.3B) filler metal [9].

Draugelates and Hartmann [10] studied the fatigue properties of a Nimonic 80 Ni-base
superalloy, brazed with BNi-5 (Ni-19Cr-10Si-0.1C) filler metal. They found that the braze
cycle, particularly braze temperature and time, had a significant influence on the fatigue
properties of the joint. As seen in Figure 16, brazing at 1190o C for 5 minutes, followed by

12
soaking at 1100oC for 8 hours and aging at 710oC for 16 hours, resulted in a 40% increase in
fatigue life, when compared to brazing at 1190o C for 5 minutes, followed by aging at 710oC
for 16 hours. The increase in fatigue life was attributed to the removal of brittle Cr-silicide
phases from the joint during the extended heat treatment. Since Si has a large atomic
diameter, extended times at temperature are required to diffuse the silicon into the parent
metal and to reduce the occurrence of brittle phases.
600

500
S (MPa)

400

300

200
3 4 5 6
10 10 10 10
N
Figure 16 – Fatigue behavior of brazed joints (tested at room temperature, R = 0.1).
 brazing at 1190o C for 5 minutes, followed by aging at 710oC for 16 hours.
o o o
 brazing at 1190 C for 5 minutes and 1100 C for 10 hours, followed by aging at 710 C for

16 hours [10].

The preceding discussion illustrates that four important parameters need to be controlled in
order to achieve high-strength joints free of brittle centreline eutectic phases. These
parameters are:

 brazing time,
 brazing temperature,
 gap width/joint gap, and
 braze filler metal.

It is also apparent from the preceding discussion that metallurgical reactions which take place
during the braze cycle affect the properties of the joints. During the braze cycle the melt point
depressants in the nickel-base braze filler metal (such as P, B and/or Si) diffuse into the parent
metal, thereby increasing the remelt temperature of the brazed joint and making the joint
capable of higher temperature service. The diffusion of elements from the braze filler metal
also affects the ductility of the completed joint. During the braze cycle a limited amount of
parent metal may dissolve in the liquid braze filler metal, which increases the melt liquidus
temperature. Diffusion of alloying elements from the parent metal into the joint has a similar
effect.

13
As mentioned earlier, the main melt point depressants in use today are B, Si and P.
Consideration of the alloy phase diagrams [2], however, suggests that Zn and Cd may also act
as melt point depressants. Due to the high vapour pressures of these elements, however, they
are unsuitable for use in vacuum brazing. Many Ag- and Cu-base braze filler metals are
doped with Zn or Cd, but use of these fillers is restricted to flame brazing, induction brazing
or inert atmosphere brazing.

1.3) Characterization of the eutectic phases that form in brazed joints

As stated earlier, Ni-base braze filler metals typically contain melt point depressants, such as
B, Si and P. These elements tend to form hard, brittle structures in the joint. If the melt point
depressants diffuse into the parent metal during the braze cycle, brittle phases may also form
adjacent to the joint, depending on the composition of the parent metal. Figure 17 displays
hardness profiles across brazed joints in Inconel 625 (a Ni-base superalloy) for three different
filler metals, BNi-2, BNi-5 and Au-6. The highest hardness was observed in the joint brazed
with BNi-5, which contains Si as melt point depressant and forms brittle silicide phases even
after an extended brazing cycle at 1150oC for 10 minutes, followed by holding at 1100oC for
60 minutes. As a result of this braze cycle, the hardness of the parent metal was increased to a
level exceeding that of the brazed region. The joint with the next highest hardness was BNi-
2, containing both Si and B as melt point depressants. Brittle silicide and boride phases
formed even after brazing at 1065oC for 10 minutes, followed by holding at 1000oC for 60
minutes. The joint with the lowest hardness was Au-6, containing 20.55% Au, as well as B
and Si. Au is known for its ductility and softness. The B and Si in Au-6 still resulted in the
formation of some silicide and boride phases after brazing at 10600C for 60 minutes, but the
joint hardness was considerably lower.
brazed zone
600
Au-6
BNi-2
550 BNi-5

500
Hardness (HV 0.05 )

450

400

350

300

250

200
-150 -120 -90 -60 -30 0 30 60 90 120 150
Distance from centre of brazed joint (µm)

Figure 17 – Hardness profiles across brazed joints in Inconel 625 [7].

Tung, Lim and Lai [11] examined four braze filler metals, BNi-5 (Ni-19Cr-10.2Si), Nicrobraz
150 (Ni-15Cr-3.5B), BNi-4 (Ni-3.5Si-1.5B) and BNi-1 (Ni-14Cr-4.5Si-4.5Fe-3B), and
concluded that the braze melt point depressants tend to form hard and brittle intermetallic
compounds with nickel, and that the presence of these compounds is detrimental to the
mechanical properties of the brazed joints. An incentive therefore exists to control the

14
microstructures that develop in brazed joints and to develop brazing alloys that are less
sensitive to the formation of detrimental phases. Figure 18 shows the typical microstructure
of a 0.5 mm wide joint brazed with BNi-5. The only intermetallic compound observed in this
braze was nickel silicide, which appears white in the micrograph. It solidified as a component
phase of the binary eutectic with -nickel, the latter appearing as the dark phase in Figure 18.
The microstructure of a joint brazed with Nicrobraz 150 is shown in Figure 19(a). The phase
marked G1 in this micrograph is nickel boride, while the phase marked G2 is the nickel-rich 
solid solution. In the centreline region of the joint a ternary eutectic of -nickel, nickel boride
and chromium boride was present, as shown in Figure 19(b).

Figure 18 – Typical microstructure of a joint brazed with BNi-5 [11].

Figure 19 - (a) Microstructure of a joint brazed with Nicrobraz 150 at 1125oC; and (b) a
magnified view showing the details of the ternary eutectic component [11].

Typical microstructures of joints brazed with BNi-4 filler metal are shown in Figure 20. The
various phases formed in this joint, shown in Figure 20(a), include -nickel solid solution
layers (marked M*) found adjacent to the joint mating surfaces, -nickel nodules (marked M),
bulky nickel borides (marked R), binary eutectics of -nickel and nickel boride, and ternary
eutectics of -nickel, nickel boride and nickel silicide (marked X). The binary eutectic of
normal morphology (marked N) formed adjacent to the layers of single-phase -nickel.
Figure 20(b) shows details of the ternary eutectic (X). The phases designated X1, X2 and
X3, are nickel silicide, -nickel and nickel boride, respectively. Nickel silicide (marked m),
which precipitates in the solid state, was found in the outer regions of the -nickel nodules,
especially those in the centreline region of the joint.

15
Figure 20 - (a) The constituents and component phases in a joint brazed with BNi-4; and (b)
details of the ternary eutectic and nickel silicide precipitates in the joint centreline [11].

The microstructure resulting from the solidification of the molten filler metal in the joint
brazed with BNi-1 was found to be extremely complex. Three intermetallic compounds
formed in this braze, namely nickel boride, nickel silicide and chromium boride. The typical
microstructure of a BNi-1 joint is shown in Figure 21(a). The phases marked L1a, L1b and
L1c are nickel borides, L2a and L2b are -nickel, L3 is nickel silicide, and L4 is chromium
boride. Three eutectic constituents were identified: a binary eutectic of -nickel and nickel
boride (L2a/L1a), a binary eutectic of -nickel and chromium boride (L2a/L4), and a ternary
eutectic of -nickel, nickel boride and nickel silicide (L2a/L1c/L3). Details of the ternary
eutectic are given in Figure 21(b). It is interesting to note that in the presence of silicon, the
formation of the ternary eutectic of -nickel, nickel boride and chromium boride was
suppressed, in that chromium boride solidified with -nickel as a binary eutectic. The fine
precipitates in -nickel (L2a) are nickel silicide particles, which formed in the solid state on
cooling to room temperature.

Figure 21 - (a) Microstructure of a joint brazed with BNi-1 at 1125oC; and (b) a magnified
view showing the details of the eutectic phases [11].

It can therefore be concluded that the intermetallic compounds formed in brazed joints using
nickel-base filler metals containing chromium, boron and silicon, are nickel boride, nickel
silicide and chromium boride. These intermetallic compounds form as constituents of binary
or ternary eutectics with -nickel. The binary eutectics are -nickel/nickel boride, -
nickel/nickel silicide and -nickel/chromium boride, whereas the ternary eutectics are -
nickel/nickel boride/chromium boride and -nickel/nickel boride/nickel silicide.

16
Johnson [12] performed microprobe analyses to identify the phases formed on brazing ferritic
steel using various gap widths. The braze filler metals evaluated were the popular BNi-2 (Ni-
7Cr-3.2B-4.5Si-3Fe) and BNi-4 (Ni-3.5Si-1.5B). The phases identified were -nickel
(containing Si, but no B), nickel boride, nickel silicide and chromium boride. As the joint gap
narrowed, the number of phases was reduced, as shown in Figures 22 to 25. The first phase
to disappear was nickel boride, followed by nickel silicide, and lastly chromium boride.
Finally only -nickel remained, as seen in Figure 25.

Figure 22 – Brazed joint of BNi-2 with Figure 23 – Brazed joint of BNi-4 with
4 phases present. Magnification: 200X 3 phases present. Magnification: 200X
(reduced 50% on reproduction) [12]. (reduced 50% on reproduction) [12].

Figure 24 – Brazed joint of BNi-4 with Figure 25 – Brazed joint of BNi-4 with
2 phases present. Magnification: 200X 1 phase present. Magnification: 200X
(reduced 50% on reproduction) [12]. (reduced 50% on reproduction) [12].

While the braze filler metal was enriched in elements from the steel parent metal through
erosion and diffusion, filler metal alloying elements were also entering the parent metal. In
BNi-4, nickel and silicon entered by bulk diffusion into the parent metal to a distance of about
15 m, but boron (which diffuses along the grain boundaries) migrated to a maximum
distance of 140 m. In BNi-2, the corresponding diffusion distances were 8 m and 60 m.
It is evident in Figures 26 and 27 that the B formed almost continuous stringers of iron
boride along the grain boundaries.

17
 Figure 26 – Boride particles in the steel Figure 27 – Boride particles in the
parent metal (BNi-4 braze). steel parent metal (BNi-4 braze).
Magnification: 500X (reduced 50% on Magnification: 2000X (reduced 50%
reproduction) [12]. on reproduction) [12].

The implication of Johnson’s research is that in order to make serviceable brazed joints, the
braze filler metal should be ductile and contain no regions of weakness or brittleness. With
the nickel-base filler metals investigated, this ideal was realized in narrow joints less than 40
to 50 m in width, containing only -nickel. Wider joint gaps allowed the formation of
boride and silicide eutectics along the centreline of the joint. As these phases have high
hardness values, in the region of 900 HV (hardness on the Vickers scale), they possess only
limited ductility and may crack if high tensile stresses are generated along or across the joint.
These hard particles may also rupture to form fatigue crack nuclei. While a long brazing time
facilitated the diffusion of B into the steel parent metal to some extent, such an extended high
temperature soak is generally not a viable option if liquation of the grain boundaries is to be
avoided.

Gale and Wallach [13] studied Ni-Cr joints brazed with BNi-3 (Ni-3.2B-4.5Si) and observed
Ni23B6 within the joints. This phase formed intragranularly in the proeutectic nickel phase
immediately adjacent to the eutectic mixture. These boride phases generally formed as
platelets along <001>Ni directions, although cuboidal morphologies were also observed. The
origin of Ni23B6 precipitation remains uncertain, especially as this phase does not appear on
the available phase diagrams for the Ni-Si-B system. Nevertheless, Ni23B6 formation occurred
reproducibly in samples held for 3 minutes at 1065oC, but not in those subjected to longer
soaking times. The Ni3Si phase formed throughout the joint within the nickel component of
the filler metal eutectic, and had a cuboidal structure aligned along the <100> directions. The
most pronounced precipitation in the eutectic phase occurred in samples not subjected to long
holding times.

Gale and Wallach concluded that:

 for Ni and Ni-5Cr substrates, Ni3B formed at the original substrate/filler interface and
M23B6 formed in the bulk substrate,
 for Ni-10Cr substrates, M23B6 formed throughout the substrate,
 for Ni-15Cr substrates, M23B6 and Cr5B3 formed throughout the substrate, and
 for Ni-20Cr substrates, M23B6, Cr5B3 and Ni3Si formed at the original substrate/filler
interface, while M23B6 and Cr5B3 formed throughout the substrate.

18
Similarly, Lasalmonie [14] examined the phases precipitated during the brazing of DS 247
(Ni-9.5Co-8Cr-3.3Ta-0.5Mo-9.5W-5.5Al-0.6Ti-0.07C) and ATGP 3 (Ni-15Co-15Cr-5.2Mo-
4.4Al-3.5Ti-0.06C) using 120 m thick BNi-3 (Ni-4.5Si-3.2B) braze filler metal. Brazing
was performed at 1040oC for 3 hours. The phases present in the structure (shown in Figure
28) were reported as:

 M23B6 with a hardness in the region of 1000 HV (with M consisting of Ni, Cr, Co, Al and
Ti),
 Cr5B3, with a hardness of 3000 HV,
 W2B,
 Ta-Hf carbides, and
 MB12 (with M consisting of Cr, Mo and W).

(a) DS 247 (b) ATGP 3

Figure 28 - Brazed joints in two nickel-base superalloys produced with BNi-3 filler metal at
1040oC for 3 hours [14].

Savage and Kane [15] attributed a number of field failures observed during the joining of
Nickel 201 tubing to Nickel 201 plate stock using BNi-3 filler metal to the formation of nickel
silicide and nickel boride phases in the joint. The braze cycle of 1093oC for 30 minutes
resulted in rapid grain boundary diffusion of B from the braze joint into the base metal. This
sensitized the grain boundaries, causing the tubing to undergo intergranular corrosion. The
authors reported that quenching, as opposed to slow cooling from the braze temperature,
changed the boride morphology in the joint. After quenching the boride phase displayed a
needle-like morphology and no intergranular corrosion attack was observed. Failure was
primarily as a result of crevice corrosion involving attack of this needle-like boride phase,
with the corrosion rate determined by the joint gap. The quench rate from the brazing
temperature is therefore also an important parameter that needs to be controlled, since a
different boride or silicide morphology can affect the mechanical and corrosion properties of
the brazed joints.

Tung, Lim and Lai [16] evaluated the intermetallic phases that form in Ni-base brazed joints,
and reported the presence of bulky nickel boride and nickel silicide phases in a solid solution
area referred to as -nickel. The compositions and hardness values of the phases observed are
as follows:

19
 -nickel, consisting of Ni and 3.0 to 4.8% Si, with a hardness between 108 and 188 HV,
 nickel boride, consisting of Ni and 0.2 to 0.3% Si, with a hardness between 776 and 906
HV, and
 nickel silicide, consisting of Ni and 8.0 to 13.1% Si (the hardness was not measured).

The relative nickel and silicon weight percentages listed above were obtained by means of
standardless EDS analysis. Since the EDS system employed could not measure B contents
accurately, the B concentrations were not reported. Despite this deficiency, the information
given above was sufficient to distinguish the different phases. The general direction of
solidification was from the base metal towards the centreline of the joint. The first phase to
form was primary -nickel, solidifying in the form of solid solution layers next to the joint
faying surfaces. This caused the melt to become enriched in B. As solidification proceeded,
eutectics of nickel boride solidified. Higher brazing temperatures and faster cooling rates
facilitated the formation of binary -nickel/nickel boride eutectics. Higher cooling rates also
promoted the formation of ternary eutectics of -nickel, nickel boride and nickel silicide in the
last remaining liquid. On further cooling to room temperature, solid state precipitation of
nickel silicide occurred within the -nickel nodules, especially those formed in the latter
stages of solidification near the centreline of the joint. Bands of nickel boride precipitates
were observed in the base metal adjacent to the joint interface in joints brazed at 1090oC,
below the eutectic temperature of the binary Ni-B system.

1.4) Repair of turbine components using brazing

Degradation of engine components during service may result from thermal cracking and
distortion, oxidation, sulphidation, corrosion, erosion, fretting, wear, foreign object damage
(FOD, i.e. objects ingested into the engine) and creep. Degraded components are replaced not
only because of their loss of structural integrity, but because there is a loss of operating
efficiency in the engine. In many instances, especially in the land-based industry (turbines
used to generate electrical power) it is more cost effective to carry out repairs to return the
components to serviceable condition for a fraction of the cost of a new component. Typically,
repair costs are approximately 10 to 50% of the cost of a new component. This is a very
lucrative business and there are numerous companies throughout the world whose primary
business is to repair turbine components.

As shown in §1.2 and §1.3, the major variables which need to be controlled in order to
achieve a high quality brazed joint are gap width, braze alloy/filler metal, braze temperature
and time, cooling rate and cleanliness. When it comes to braze repair of turbine components,
however, controlling the joint gap is near impossible. This is because components have
predefined and precise crack widths after service exposure. In the early era of crack repair
using brazing, only cracks with widths between 0.025 mm and 0.254 mm were repaired.
Once the combustion products were removed by processes such as fluoride ion cleaning (a
process that uses a combination of H2 gas and HF gas to reduce the aluminum oxide and
titanium oxide present in the crack), capillary action would draw the molten braze filler metal
into the crack. Braze filler metals containing B and/or Si as melt point depressants were
initially evaluated for crack repair, but as a result of the brittle silicide and boride phases that
form in the repaired joint, the braze joint would re-crack when returned to service. Since Si
has a large atomic diameter and do not readily diffuse into the parent metal, investigators
concluded that only braze filler metals with B as melt point depressant could be used for crack
repair. This encouraged the introduction of a number of new filler metals into the market.
Some of these braze alloys are listed below:

20
 DF3 - (Ni-20Cr-20Co-3Ta-3B-0.05La)
 DF4B – (Ni-14Cr-10Co-2.5Ta-3.5Al-2.7B-0.02Y)
 DF5 – (Ni-13Cr-3Ta-4Al-2.7B-0.02Y)
 DF6 – (Ni-20Cr-3Ta-2.8B-0.04Y)
 BRB – (Ni-14Cr-9.5Co-4Al-2.5B)
 775 – (Ni-15Cr-3.5B)
 S57B – (Co-10Ni-21Cr-5Ta-2.5Al-3B-3Si-0.02Y)

Lecomte-Martens [17] from the repair company C.R.M in Belgium investigated the use of
DF4B (with the composition shown above) in the braze repair of In792 nickel-base superalloy
vanes, as well as the use of S57B in the braze repair of H188 cobalt-base superalloy vanes.
In792 has a composition of Ni-12.7Cr-9Co-2Mo-4.2W-4.1Ti-3.4Al-4.2Ta-1.0Hf-0.05Zr-
0.015B-0.14C, and H188 consists of Co-22Cr-22Ni-14W-1.5Fe-0.08La-0.08C. The brazing
temperature for the DF4B filler metal was 1200oC and the brazing time 15 minutes, followed
by a 3 hour diffusion cycle at 10570C (to diffuse the B away from the joint). The brazing
temperature for the S57B filler metal was 1220oC and the brazing time 15 minutes, followed
by a 3 hour diffusion cycle at 10400C (to diffuse the B and Si away from the joint). For both
joints, 18 hour and 30 hour diffusion cycles were also investigated. The joints produced with
S57B filler metal displayed evidence of a Co-Ni-Cr solid solution, Cr carbides, W carbides,
Ta carbides, Cr borides and Cr-Co-W silicide phases. The joints produced with DF4B filler
metal contained a Ni-Co-Cr solid solution, Ta-Ti-Hf carbides, Cr carbides and Cr borides.
Fatigue tests were performed on actual braze repaired vanes, as opposed to test specimens.
The first engine test showed that new vanes cracked after 441200 cycles, whereas the braze
repaired vanes cracked after 324200 cycles. The second engine test showed that new vanes
cracked after 409200 cycles, whereas the braze repaired vanes cracked after 270300 cycles.
In the third and fourth engine tests, new vanes cracked after 397600 and 417325 cycles,
respectively, whereas the braze repaired vanes cracked after 240100 and 280250 cycles. It
was concluded that, on average, the number of cycles required to fracture a braze repaired
vane is about 2/3 of that required for a new vane. This was acceptable since the price of a
repaired vane is typically 10 to 50% of that of a new vane, and hence the braze repair was still
economical. The total number of cracks found on the new vanes was 40, the total crack
length was 287 mm and the average crack length was 7.2 mm. For the braze repaired vane,
39 cracks were found in total, the total crack length was 218 mm and the average crack length
was 5.6 mm. These results suggest that, although cracks initiated earlier in a braze repaired
vane, the average crack length was less than in a new vane. The crack propagation rate
therefore appeared to be slower in the brazed joint than in the parent metal. Figure 29
compares cracks detected in a new vane and in a repaired vane after engine testing. An
important conclusion reached by the author was that cracks wider than 0.5 mm were not
completely filled by the molten braze filler metal and should be repaired using wide gap
brazing. The topic of wide gap diffusion brazing is considered in §1.5.

Elder, Thamburaj and Patnaik [18] investigated the repair of narrow cracks in MA 754 (Ni-
19.7Cr-0.35Al-0.45Ti-0.37Fe-0.05C-0.57Y2O3-0.39O2) using vacuum brazing with fourteen
different braze alloys. The braze filler metals contained either B, or B and Si, as melt point
depressants. Microprobe analyses revealed the presence of Cr boride and Ni silicide in brazed
joints containing both B and Si as melt point depressants. Joints containing only B as melt
point depressant contained Cr-W boride phases. Narrow cracks (0.038 mm in this
investigation) resulted in joints free of brittle phases. As the crack width approached 0.762
mm, brittle intermetallic phases formed, reducing joint ductility to 1.3% (compared with the
parent metal ductility of 4%). The authors also concluded that these brittle intermetallic

21
phases act as preferential crack initiation sites, and recommended the use of B as the sole melt
point depressant in repair brazing. For joints approaching 0.762 mm in width, wide gap
brazing was recommended.

Figure 29 – Cracks in new and repaired vanes after engine testing [17].

It is important to note that when cracks form during engine operation, combustion products
are often entrapped within the cracks. The crack surfaces also tend to be covered in Al and Ti
oxides. These must be removed before cracks are repaired. As mentioned earlier, cleanliness
is important in achieving high quality brazed joints. If the crack is dirty and filled with
contaminants, capillary action will not draw the molten braze filler metal into the joint. In
order to clean a crack prior to vacuum brazing, a process called FIC (fluoride ion cleaning)
was developed to reduce the oxide layers in the crack and to remove dirt and other
contaminants. The process is performed at elevated temperature and uses a mixture of HF
and H2 gas. Rose, Price and Thyssen [19], employed by Chromalloy, developed a braze
repair process using FIC and a braze filler metal with a composition of Ni-12Cr-4.5Mo-6Al-
4Ta-0.6Ti-0.1Zr-0.05C-2.5B. Without the boron alloying addition, the composition of the
braze filler metal matches that of In713 nickel-base superalloy. This publication contained
the first mention in open literature of a new braze filler metal composition developed by a
repair company. The braze temperature and time were not documented and are considered
proprietary. An interesting observation reported was that the joints contained a certain
percentage of porosity, and that the authors relied on the use of a process referred to as
HIP’ing to close the pores. HIP’ing is a process performed at elevated temperature under
high isostatic pressure, applied using Ar gas, to plastically close the pores. This repair
process using FIC, a proprietary braze filler metal and HIP’ing, was used to repair cracks in
the saddle and blade areas of a Garrett GTCP 85 APU component that provides electrical
power and compressed air for commercial aircraft such as Boeing 727 and 737 and
McDonnell-Douglas DC-9 and MD-80. To qualify the braze repair, tensile tests and fatigue
tests were performed at room temperature. After achieving acceptable mechanical test results,
the authors performed thermal fatigue tests, and reported satisfactory results.

Lee, McMurray and Miller [20] patented a novel technique, referred to as M-Fill, to repair
cracks in stationary components. This process does not rely on capillary action to draw the
molten metal into the joint area. The crack is mechanically removed as per a welding process,
then the surface of the routed out area is Ni-plated. A braze paste is placed in the groove and
the component is processed in a vacuum furnace at 954oC for 1 hour, followed by soaking at
1065oC for 15 minutes and 982oC for 3 hours to diffuse the B and Si into the parent metal.

22
The braze filler metal contained Ni, 8-10% Cr, 7-9% W, 2.5-4% Fe, 3-4.5% Si, 2-3% B and
0.06% C. Although boride and silicide phases formed, they were inconsequential because the
cracks were in low stress areas and hence the braze repair was considered non-structural. The
authors showed that the braze joint could be readily coated with an aluminide coating for
corrosion and oxidation protection. The joint remained intact during engine testing. This
repair is still applied in the TF40 and T55 Allied Signal engines used in helicopters.

1.5) Wide Gap Diffusion Brazing

As mentioned earlier, controlling the joint gap is very important in producing good quality,
high-strength joints free of brittle eutectic and intermetallic phases. It is fairly easy to control
joint gaps in manufacturing assemblies, but when it comes to repair, the size of the crack
cannot be controlled. In the author’s experience, cracks in aircraft turbine vanes range from
0.05 mm to 1 mm in width, whereas cracks in industrial turbine engines can range from 1 mm
to 6 mm in width. To repair these types of defects, the wide gap diffusion brazing process
was invented.

GE, an OEM, invented and patented a process called activated diffusion brazing (ADB). It is
also referred to as activated diffusion bonding or activated diffusion healing (ADH). Hoppin
and Berry [21] defined this process as a joining method for high strength nickel-base
superalloys, involving a combination of the best current knowledge of conventional brazing
and solid state diffusion bonding technology. The process combines the manufacturing ease
of brazing with the high joint strengths achievable during solid state diffusion bonding. ADB
involves vacuum furnace brazing with an ultra high-strength bonding alloy, nearly identical in
chemical composition to the base metal being joined. Diffusion and aging heat treatments
enable the joint to achieve maximum mechanical strength at the operating temperature of the
engine component. The ultra high-strength bonding alloy is a mixture of braze powder and
superalloy powder in a 60/40 ratio.

Figure 30 shows the tensile strength properties of 0.25 mm wide ADB joints in Rene 80 base
metal at 650oC and 870oC, and indicates that the tensile properties of the joints are equivalent
to those of the base metal. Results of creep rupture tests on Rene 80 brazed with AMS 4782
(Ni-19Cr-10Si), AMS 4786 (Au-22Ni-8Pd) and the ultra high-strength bonding alloy can be
seen in Figure 31. Typical creep rupture strengths of 80 to 90% of the parent metal’s creep
rupture properties are attained, and the creep rupture properties of the ADB joints are two
orders of magnitude greater than those achievable with conventional braze filler metals such
as AMS 4782. Fatigue tests were performed at 760oC for the ADB joints and the resultant
high cycle fatigue curve is shown in Figure 32. The high cycle fatigue properties are
approximately 70% of those of the Rene 80 base metal. The elevated temperature tensile,
creep rupture and high cycle fatigue tests therefore demonstrated that ADB joint strengths of
70 to 90% of those of the base metal can be achieved, and the authors concluded that the ADB
process is suitable for replacing fusion welding in the crack repair of aircraft gas turbine
components (Rene 80 base metal displays a tendency towards cracking during fusion
welding).

23
 160
Rene 80 parent metal
Butt joints
140

120

100
Stress (ksi)

80

60

40

20

0
UTS 0.2% YS UTS 0.2% YS
650°C 870°C

Figure 30 - Tensile properties (UTS: ultimate tensile strength, and 0.2% YS: 0.2% proof
stress) of activated diffusion bonded butt joints in Rene 80, compared with those of the base
metal [21].

80

Rene 80

60
Stress (ksi)

ADB joints
40

20
AMS 4782

AMS 4786
0
50 52 54 56 58 60 62
Parameter = T (25 + log t) x 10-3

Figure 31 - Creep rupture properties of activated diffusion bonded joints in Rene 80,
compared with those of the base metal and conventional brazed joints [21].

24
 80

Rene 80 base metal

Maximum stress (ksi) 60

40

20

ADB joints
ABD joints - no failure

0
4 5 6 7
10 10 10 10
Cycles to failure

Figure 32 - High cycle fatigue strengths of activated diffusion bonded butt joints in Rene 80
at 760oC [21].

The ADB (or ADH) process was also described by Demo and Ferrigno [22]. As shown in
Figure 33, the ADH process uses a mixture of superalloy powder and braze powder called a
low-melting matrix. The superalloy powder in this low-melting matrix has the same
composition as the substrate material. The low-melting braze powder used by Demo and
Ferrigno consisted of the elements Ni-Cr-Co-Al-Ta-B, but the relative percentages of the
elements were not documented. Brazing was performed at 1205oC for 30 minutes, followed
by an undisclosed diffusion cycle and aging treatment. Figure 34 shows the tensile properties
measured at 870oC for various joint gaps ranging from 0.075 mm to 1 mm. The joint tensile
strength was approximately 80% of the parent metal tensile strength, and the tensile strength
did not appear to vary significantly over the range of joint gaps tested. The creep rupture
properties, however, varied significantly with joint gap, as shown in Figure 35. Joints with
gaps of 0.075 mm achieved approximately 90% of the creep rupture properties of the parent
metal, whereas joint gaps of 0.5 mm and 1.0 mm resulted in only 60% and 50% of the parent
metal creep rupture properties, respectively.

Figure 33 - In the ADH repair process, a superalloy is “cast” into the crack using a lower
melting point bonding alloy [22].

25
 100
Tensile strength
PARENT METAL
0.2% yield strength
90
ADH JOINTS
80

Stress (ksi) 70

60

50

40

30

20

10

0 Control Oxidize Oxidize Strip 0.075 mm 0.25 mm 0.5 mm 1 mm

+ strip + FIC wide gap wide gap wide gap wide gap

Figure 34 - Tensile strength data at 870oC (1600oF) for the parent metal and brazed joints in
Rene 80 base metal. (‘Oxidised’ refers to a simulated engine run; ‘Strip’ refers to chemical
stripping simulating the removal of the aluminide coating; and ‘FIC’ refers to fluoride ion
cleaning) [22].
38
Note: Inconel 625 60% 90% 100%
36 80%
@ 7.4 ksi: life 10 hours / 925°C
@ 4.0 ksi: life 100 hours / 925°C
34
Slope from
32 design curve
Stress (ksi)

30

28

26

24
Parent metal - Control
Parent metal - Oxidized + strip
22 Parent metal - Oxidized + strip + FIC
ADH joint - 0.075 mm gap
ADH joint - 0.25 mm gap
20 ADH joint - 0.5 mm gap Life Life
10 hours 100 hours
ADH joint - 1 mm gap

18
52 53 54 55 56 57 58 59
Parameter = T (25 + log t)
Figure 35 - Creep rupture curves for different width ADH joints in Rene 80 base metal [22].

Young [23] evaluated the ADH process by determining the nil strength temperature, defined
as the temperature at which the bonded joint is incapable of sustaining a constant load of 0.86
MPa (125 psi), for two ADH alloys. An ADH joint consisting of 50% D-15 braze alloy and
50% X-40 cobalt-base superalloy powder attained a room temperature tensile strength of 82.4
ksi (569 MPa) and a nil strength temperature of 1260C, whereas an ADH mixture consisting
of 80% H-33 braze alloy and 20% X-40 cobalt-base superalloy powder had a room
temperature tensile strength of 658 MPa (95.4 ksi) and a nil strength temperature of 1232C.
This compares well with the X-40 parent metal, which has a room temperature tensile strength

26
of 794 MPa (115 ksi) and a nil strength temperature of 1329C. For ADH processing of
nickel-base superalloys such as Rene 80, a 50% D-15 braze alloy and 50% Rene 80 powder
mixture was evaluated. The braze cycle consisted of holding at 1204oC for 30 minutes,
followed by a boron diffusion cycle at 1093oC for 2 hours, 1149oC for 2 hours, and 1190oC
for 2 hours. The first aging cycle was performed at 1093oC for 4 hours, followed by a second
aging cycle at 843oC for 24 hours. The 1 mm wide joints attained 80% of the Rene 80 parent
metal tensile strength measured at 870C. Creep rupture testing at 925C demonstrated a
10% reduction in rupture life for a joint gap of 0.025 mm, a 20% reduction in rupture life for a
0.075 mm gap, and a 30% reduction in rupture life for a 0.25 mm gap. A 1 mm gap resulted
in a 40% reduction in creep rupture life. It is evident that a wider joint gap results in a
reduction in mechanical properties. Despite this, the ADH process has been used successfully
for the past two decades to repair cracks in aircraft turbine blades, and the FAA (Federal
Aviation Authority) has never reported an engine failure as a result of an inappropriate ADH
braze application. This shows that, although the mechanical properties are not equivalent to
the base metal properties, ADH braze joints can perform reliably in an engine. For this
reason, the process has replaced conventional gas tungsten arc welding (GTAW) with In625
filler metal in this application.

Knotek and Lugscheider [24] mixed stainless steel powder and Ni-base braze filler metal of
composition Ni-10Cr-4.2Si-6.3Fe-3B in three different ratios: 45:55, 50:50 and 55:45, and
performed brazing at a temperature of 1000oC with a joint gap of 0.1 mm. The results of
tensile tests carried out at 20oC, 800oC, 1000oC and 1200oC are shown in Figure 36. The last
two test temperatures are higher than the solidus temperature of the braze alloy in the mixture.
As shown in Figure 36, all three mixtures had similar tensile strengths at elevated
temperature. At room temperature the tensile strengths ranged from 480 MPa to 540 MPa,
and at 1000oC the tensile strengths were in the region of 70 MPa. These tensile strengths are
higher than those of the popular Ni-Cr-B-Si braze filler metals such as AMS 4777 and AMS
4776. No significant change in the tensile strength values at 1000oC was observed after a 1
hour diffusion heat treatment, as shown in Table 2.

600
Ultimate tensile strength (MPa)

400

200

0
0 200 400 600 800 1000 1200
Temperature (°C)

Figure 36 - Tensile strength at elevated temperature for various ratios of superalloy powder to
braze powder [24].

27
 Table 2 – Influence of heat treatment on the tensile strength (1000C) of reacting brazed
joints [24].
Component Composition after reaction Tensile strength at 1000C
(wt.%) (wt.%, balance Ni) As-brazed Brazed + 1 hr @ 800C
A B Cr B Si Fe MPa ksi MPa ksi
55 45 8.25 1.65 2.30 2.00 74.6 10.8 66.7 9.7
50 50 7.50 1.50 2.10 1.85 68.7 10.0 68.6 10.0
45 55 6.75 1.35 1.90 1.70 64.8 9.4 67.6 9.8

Van Esch [25] from Elbar, a repair vendor in Holland, developed a proprietary wide gap braze
process to repair blades from the Industrial Gas Turbine engine referred to as GE Frame 5.
The compositions of the two alloys used, the mix ratio and the braze temperature and time
were not documented. Since the wide gap braze process was used to repair blades, not vanes,
the butt joints were not tested. Instead, samples with gauge diameters consisting of 20%
braze build-up and 80% parent metal were tested. Creep rupture tests were performed using
In738 nickel-base superalloy parent material. Figure 37 shows the creep rupture properties of
the joints. The joints displayed creep rupture lives ranging from 62% to 100% of that of the
In738 alloy. Because of the comparatively low amount of boron in the wide gap braze joint,
the borides could be rounded and distributed more evenly by an annealing heat treatment, as
seen in Figures 38 and 39. This improved the joint ductility.
3
10

In738LC
scatterband
average +/- 20%
Stress (MPa)

2
10

Measured in high temperature corrosive environment

1 Measured in air
10
-1 0 1 2 3 4 5
10 10 10 10 10 10 10
Time (hours)

Figure 37 - Creep rupture tests performed at 850oC (base material is In738 and the wide gap
braze is a proprietary material referred to as Elniment 100) [25].

Su, Chou, Wu and Lih [26] evaluated the phases that formed in ADB joints in In738 base
metal. Three braze alloys, namely Nicrobraz 150, BRB and DF4B, were evaluated. The
chemical compositions of these alloys are given in Table 3.
Table 3 - Nominal compositions of powders and base metal (wt.%, balance Ni).
Alloy Cr Co Al Ti W Mo Ta Nb B C Zr Y
In738 16.0 8.4 3.4 3.4 2.7 1.7 1.7 0.8 0.01 0.1 0.03 -
Nicrobraz 150 15.0 - - - - - - - 3.50 0.1 - -
BRB 13.5 9.5 4.0 - - - - - 2.50 N/A - -
DF4B 14.0 10.0 3.5 - - - 2.5 - 2.70 N/A - 0.02

28
 Figure 38 - Elniment 10 wide gap braze, Figure 39 - Elniment 10 wide gap braze,
as-brazed without subsequent heat as-brazed with annealing heat treatment.
treatment. The borides have an oblong The borides appear more rounded [25].
shape [25].

Brazing mixtures, consisting of brazing alloy and In738 superalloy powder, were prepared in
weight percentage ratios of 30:70, 40:60 and 50:50. Microstructural observations were made
using optical microscopy, scanning electron microscopy (SEM) and electron probe x-ray
microanalysis (EPMA). The percentage porosity detected in the ADB joints is shown in
Figure 40 as a function of the filler metal ratio for the BRB and DF4B braze alloys. It is
evident that joints containing 30% BRB or 30% DF4B mixed with In738 powder exhibited
very high porosity levels after brazing at temperatures of 1180oC and 1190oC. Powder mixes
containing 40% or 50% DF4B brazing alloy exhibited no porosity. The interstitial voids were
small and randomly distributed in the 40% and 50% joints containing BRB as braze alloy.
16
DF4B (1190°C)
DF4B (1180°C)
BRB (1190°C)
BRB (1180°C)
12
Porosity (area %)

0
30 40 50
Filler metal ratio (wt.%)

Figure 40 - Percentage porosity as a function of braze alloy content [26].

Figure 41(a) shows the microstructure of a 40% Nicrobraz 150 specimen after ADB
processing. No pores are evident. The microstructure at high magnification is shown in
Figure 41(b). The matrix and secondary (blocky) phases are more evident, and island-like
blocky boride phases are present in the In738 powder particles. Table 4 shows the chemical

29
compositions of the two phases observed. Quantitatively, the blocky phase has a higher
concentration of light elements (boron and carbon) than the matrix. The results of the
quantitative analysis indicate that the blocky phase is Cr(B,C) with low levels of Co, Al, Ni,
Ta and Ti in solution, while comparatively rich in Mo and W. The matrix of the Nicrobraz
150 braze alloy is predominantly Ni3B with reduced amounts of B and C, but with high
concentrations of Co, Ti and Ta. Cr boride is thermodynamically stable at all the
temperatures evaluated, and its chemistry does not change once it has nucleated. The
compound present probably has the stoichiometry CrB. Figure 41(b) also shows a Ni-Ni3B
eutectic precipitated in the braze alloy matrix (designated ‘N’). This phase contains higher
amounts of Co and Ti, derived from the superalloy powder by diffusion. The chain-like
structure of the CrB, Ni3B and Ni-Ni3B eutectic reduces the ductility and the strength of the
joint.

(a) (b)
Figure 41 – SEM micrographs of a brazed joint produced from a mixture containing 40 wt.%
Nicrobraz 150: (a) blocky boride phase (magnification: approximately 270X), and (b) Ni-
Ni3B eutectic (N). The arrow highlights the blocky boride phase precipitated within the metal
powder particles (magnification: approximately 1800X) [26].

Table 4 – Chemical compositions of phases in specimens containing 40 wt.% Nicrobraz 150

braze filler metal [26].
Matrix Blocky phase
Element
wt.% at.% wt.% at.%
Ni 76.248 67.473 4.753 2.957
Cr 4.789 4.783 68.858 48.349
Co 5.833 5.140 1.917 1.188
Al 0.594 1.144 0.006 0.008
Ti 5.117 5.548 0.257 0.196
Ta 2.276 0.654 0 0
Mo 0.248 0.134 5.978 2.275
Nb 1.394 0.779 0.038 0.015
W 0.366 0.103 4.467 0.887
B 1.447 6.946 7.135 24.083
C 1.687 7.295 6.591 20.041

Figure 42(a) shows the microstructure of a joint containing 40% DF4B and 60% In738
processed at 1190oC. The initial spherical shape of the superalloy powder has disappeared
and the blocky phase is still present. However, the matrix is not clearly identifiable, in
contrast to Figure 41(a). As seen in Figure 42(b), the blocky boride phase is present within
the superalloy powder particles. Darker matrix phases are also present alongside the blocky
phase. The results of compositional analyses are shown in Table 5. The blocky boride phase

30
contains higher concentrations of Mo and W, compared to the Nicrobraz 150 joints. The
composition of the irregular lathy dark-etching phase alongside the blocky boride phase
revealed it to be ’ phase, containing Al and Ti, which had diffused from the superalloy
powder. The results shown in Figure 42 and Table 5 indicate that within the superalloy
powder particles, the gamma prime (’) phase had coarsened to form a lathy structure in the
boron diffusion direction. Within the DF4B powder, a dark matrix structure and blocky
boride phase formed.

(a) (b)
Figure 42 – SEM micrographs of a brazed joint produced from a mixture containing 40 wt.%
DF4B: (a) blocky boride phase (magnification: approximately 350X), and (b) darker ’ phase.
The arrow highlights the blocky boride phase precipitated within the metal powder particles
(magnification: approximately 1750X) [26].

Table 5 – Chemical compositions of phases in specimens containing 40 wt.% DF4B braze

filler metal [26].
Matrix Blocky phase
Element
wt.% at.% wt.% at.%
Ni 73.337 67.269 4.673 3.373
Cr 3.474 3.596 70.230 57.209
Co 6.064 5.539 3.725 2.678
Al 6.844 13.656 0.018 0.028
Ti 4.548 5.111 0.528 0.467
Ta 3.899 1.160 0.684 0.160
Mo 0.256 0.144 5.816 2.568
Nb 0.545 0.316 0.108 0.049
W 0.339 0.099 5.777 1.331
B 0.008 0.040 6.081 23.811
C 0.685 3.070 2.360 8.325

Figure 43 shows the microstructure of a joint consisting of 40% BRB filler metal and 60%
In738 powder processed at 1190oC. The blocky boride phase and the darker lathy phase are
similar to those observed within the DF4B joint. The compositions of the darker phase and
the blocky phase can be found in Table 6.

Subsequent tensile investigations of these ADB butt joints showed that the ultimate tensile
strength (UTS) of the DF4B/In738 joints attained 85% of the room temperature UTS of the
In738 base metal, and 100% of the UTS of the In738 base metal at 980oC. For the Nicrobraz
150/In738 joints, the UTS was 40% of the room temperature UTS of the In738 base metal.
The reduction in mechanical strength was attributed to the presence of the Ni-Ni3B eutectic
phase within the joint. This eutectic not observed in the DF4B/In738 or BRB/In738 joints.

31
 (a) (b)
Figure 43 – SEM micrographs of a brazed joint produced from a mixture containing 40 wt.%
BRB: (a) blocky boride phase (magnification: approximately 225X), and (b) darker ’ phase.
The arrow highlights the blocky boride phase precipitated within the metal powder particles
(magnification: approximately 1325X) [26].

Table 6 – Chemical compositions of phases in specimens containing 40 wt.% BRB braze

filler metal [26].
Matrix Blocky phase
Element
wt.% at.% wt.% at.%
Ni 6.724 5.500 73.889 68.778
Cr 67.700 62.494 4.724 4.962
Co 3.028 2.467 6.183 5.731
Al 0.185 0.329 6.511 13.184
Ti 0.445 0.446 4.352 4.964
Ta 0.113 0.030 2.372 0.716
Mo 6.373 3.189 0.308 0.175
Nb 0.073 0.338 0.825 0.485
W 9.893 2.583 0.657 0.195
B 3.099 13.751 0 0
C 2.283 9.126 0.178 0.810

1.6) Sintering of Ni-base superalloys using braze filler metals

As illustrated in §1.5, mixing of braze filler metal powder and base metal powder does not
result in adequate joint strength when the joint gap approaches 1 mm in width. The results
considered above also indicate a significant loss in creep rupture life when the ADB process is
used for joints gaps approaching 1 mm in width. In an aircraft engine, stationary components
such as vanes seldom contain cracks approaching 1 mm in width, and the ADB process has
therefore been used successfully during the last two decades to repair aircraft engine
components. There are many profitable companies throughout the world which offer a
service repairing aircraft engine components. South African Airways spent $50 million
dollars per annum in the 1990’s on the repair of aircraft engine components at repair
companies located in Europe and the USA.

With the tremendous growth in global energy consumption, however, the repair of land-based
power generation gas turbines has become a lucrative business. A typical land-based turbine
blade is ten times the length of an aircraft turbine blade and a land-based vane weighs more
than 20 kg as opposed to an aircraft vane that weighs approximately 5 kg. Thermal fatigue

32
 cracks in land-based turbine vanes can be up to 20 cm long, 5 mm wide and 6 mm deep. The
ADB process and conventional narrow gap brazing processes cannot be used to repair these
large defects successfully. An incentive therefore exists for the development of new joining
technologies suitable for repairing wide cracks in land-based turbine components.

In 1979, Chasteen and Metzger [27] developed a sintering process during which a large gap is
filled with a metal powder. The powder is then sintered in the gap. Simply stated, sintering is
a process involving the compacting and agglomeration of loose powder particles, usually at
elevated temperature. The authors sintered gaps varying from 0.025 mm to 2 mm in width in
Hastelloy X base metal using four different powders: Hastelloy C (fine and coarse particle
size), Rene 80 and a 79%Ni-20%Cr-1%Si powder. The compositions of the powders are
given in Table 7.

Table 7 – Chemical compositions of the base metal and sinter powders (S1-S4) used in the
Chasteen and Metzger investigation [27].
Melting
Common Nominal chemical composition (wt.%, balance Ni)
Alloy temperature (F) Mesh size
designation
Cr Fe Co Mo W Si B Al Ti Solidus Liquidus
Base metal Hastelloy X 22 18 2 9 1 - - - - 2350 2470 -
Sinter filler
S1 - 20 - - - - 1 - - - 2500 2575 -325
S2 Hastelloy C 16 6 2 16 4 - - - - 2320 2380 -150/+325
S3 Hastelloy C 16 6 2 16 4 - - - - 2320 2380 -325
S4 Rene 80 14 - 10 4 4 - - 3 5 - - -100
Braze filler
B1 - 7 3 - - 6 4.5 3 - - 1790 1900 -

The powders were sintered at a temperature of 1129oC for 30 minutes. After sintering, the
joints displayed an average of 33% porosity by volume. By adding a braze filler metal,
referred to as B1 (see Table 7), the porosity could be minimized. The sintering and
subsequent brazing process resulted in the joint microstructure shown in Figure 44. A few
voids are evident in the micrograph and this is believed to be due to the low temperature used
during the brazing cycle (1129oC for 2 hours).

Figure 44 - Braze joint of 1.25 mm clearance using sinter filler metal S1 and braze filler
metal B1 (as shown in Table 7) [28].

Figure 45 shows the tensile strength variation as a function of temperature for 0.5 mm joints
with sinter powder, joints with no sinter powder, and the base metal. The joints with sinter

33
powder displayed tensile strengths corresponding to approximately 70% of the base metal
strength at temperatures below 650C, but more than 90% of the base metal strength at
temperatures greater than 760C. The brazed joints without sinter powder displayed very low
tensile strength values (about 20% of the base metal tensile strength).

110
Base metal (published)
100 Base metal (tested)
No sinter filler
90
Tensile strength (ksi)

S1
80 S2

70
60
50
40
30
20
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Test temperature (°F)
Figure 45 - Tensile strength as a function of temperature for the base metal, and for sintered
and brazed joints with 0.5 mm clearance [27].

Figure 46 shows the variation in tensile strength with joint clearance. The results indicate
that at low joint clearances, joints brazed without sinter powder had joints strengths
comparable with those of joints containing sinter powder and braze powder. At joint
clearances greater than 0.15 mm, the joint strength (with no sinter powder) decreased rapidly.
Joints containing S2 powder displayed higher strength values at all joint clearances tested
than joints with S1 sinter powder. The strength of joints brazed with S1 sinter powder
remained fairly constant at about 60 ksi (414 MPa) (54% of the base metal strength), while
the curve for the S2 joints displayed a slight decrease in strength with increasing clearance.
100
90
80
Tensile strength (ksi)

70
60
50
40
30
20
No sinter filler
10 S1
S2
0
0 10 20 30 40 50 60 70 80
Joint clearance (mil)

Figure 46 – Room temperature tensile strength as a function of joint clearance for sintered
and brazed joints and for the joints containing no sinter powder [27].

34
The method developed by Chasteen and Metzger for filling wide gaps therefore consists of
two steps. The first step involves sintering a nickel-base alloy of high melting point (called
the sinter filler metal) in powder form in the joint. Light sintering results in a porous, but not
friable, metallic mass in the joint. The second step involves filling the powder interstices by
brazing with a nickel-base filler metal (called the braze filler metal) to complete the joint.
The results of this investigation encouraged researchers to optimize the process for use in
repairing wide cracks in gas turbine components.

In 1992, Liburdi et al [28] patented a similar two step process. The first step involved
sintering in the solid state at elevated temperatures and for extended times (e.g. 1200oC for 4
hours), resulting in a dense structure. The braze filler metal was therefore only required to
close a small number of pores, leading to superior mechanical properties (as shown in Figure
47). On average, the creep rupture properties were at least 75% of the base metal properties
for a 1.5 mm wide gap (such as those typically found in gas turbine components). Whereas
Chasteen and Metzger used conventional, commercially available braze filler metals, Liburdi
et al used braze filler metal especially designed to match the composition of the base metal
being repaired, except for the addition of both B and Si as melt point depressants.
110
Butt samples HIP'd
In738
Butt specimens not HIP'd
100 Overlay specimens (2/3 area)
Rene 77
Rene 80 Butt specimens

90

80

70
Stress (ksi)

60

50

40

30

20

10
21 22 23 24 25 26 27 28
-3
P = T (20 + log t) x 10
Figure 47 - Larson Miller plot for wide gap joints produced by the Liburdi solid state
sintering process [28].

In 1983, Jahnke and Demny [29] developed a process referred to as liquid phase diffusion
sintering. They claimed it had potential for closing wide gaps, and for developing improved
mechanical strength compared to conventional wide gap brazed joints, such as those produced
through the ADH process. The joints produced with this method generally have high thermal
fatigue resistance and good hot corrosion properties. The steps involved in this process are
listed below, and schematically illustrated in Figure 48.

35
1) Melting of the filler metal on heating above the solidus temperature.
2) Rearrangement of solid grains by sliding on the liquid films separating them.
3) Densification.
4) Melting of the parent metal (powder particles) and dilution until the concentration of the
liquid reaches CL (see Figure 48).
5) Diffusion of the melt point depressants (boron and silicon) into the solid, resulting in
isothermal solidification.
6) Recession of the solid-liquid interface and a reduction in the liquid zone until the
solubility limit in the base metal powder particles is reached.

Figure 48 - Liquid phase diffusion sintering: general description of diffusion solidification

[29].

In a subsequent development, Miglietti (US patent number US6520401(B1)) developed a

liquid phase diffusion bonding process, described in more detail in four technical papers [30-
33] supplied in Appendices A to D. During this bonding process, a crack is filled with
MarM247 nickel-base superalloy powder. This powder has a solidus temperature in excess of
1315oC, well above the recommended brazing temperature range. The MarM247 powder is
then covered with a Ni-base superalloy braze powder or paste. The braze powder or paste is
doped with either B, Zr or Hf, or a mixture of all three elements, to lower its melting point.
Adding Zr or Hf as melt point depressant to braze filler metals for joining Ni-base superalloys
is novel and the results are considered in more detail in subsequent chapters. The braze filler
metal is then heated to the brazing temperature of between 1220oC to 1280oC. The braze
powder melts to form a liquid, which is carried into the spaces between the MarM247 powder
particles to form a solid diffusion bond that joins the powder particles. Thermal processing
consists of heating to a brazing temperature between 1220oC to 1280oC and holding at this
temperature for a maximum of 30 to 45 minutes. The temperature is then lowered to
approximately 1200oC to diffuse the melt point depressants into the parent metal.

The remaining chapters of this document describe the results of a series of experiments aimed
at evaluating the potential of novel Ni-Hf and Ni-Zr braze filler metals for joining Ni-base
superalloys using the liquid phase diffusion bonding process.

36
 CHAPTER 2 - OBJECTIVES
LIQUID PHASE DIFFUSION BONDING OF NICKEL-BASE SUPERALLOY
COMPONENTS USING NOVEL BRAZE FILLER METALS

2.1) Background

Turbine vanes or nozzles operating in power generation engines often develop wide cracks
during service. When the engine is overhauled, or when cracks propagate to a specific length
during service, a decision has to be made whether new replacement parts should be
purchased, or whether the damaged components should be repaired at a fraction of the cost of
a new part. On the basis of economic considerations, the majority of users opt for repair.

Many ’-strengthened nickel-base superalloys are difficult to weld successfully and brazing
techniques have gained prominence in the turbine repair industry in the last decades. Since
conventional brazing techniques rely on capillary action to draw the molten braze filler metal
into the joint gap, these processes are only suitable for repairing narrow cracks (less than 0.25
mm in width). In order to overcome this limitation, wide gap brazing techniques have been
developed for repairing cracks up to 1 mm in width. At present these techniques are widely
used in applications involving the repair of turbine components for aircraft engines, where
wide cracks (more than 1 mm in width) are seldom encountered.

Conventional wide gap brazing techniques, however, find limited application in the repair of
land-based turbine components, such as those found in power generation engines, where
cracks widths often exceed 1 mm. When wide gap brazing techniques are applied to repair
cracks wider than 1 mm, the repaired areas normally achieve only 60 to 70% of the
mechanical properties of the base metal, and display a tendency for premature degradation
and cracking in service. A need has therefore been identified for a reliable technique suitable
for repairing the wide cracks often observed in land-based turbine engine components.

In the past most repair vendors operated independently from the original engine
manufacturers (OEM’s). As a result of the lucrative nature of the repair market, all the
OEM’s (including General Electric, Pratt & Whitney (P&W), Siemens and Alstom) have
since opened their own repair shops or acquired some of the independent repair vendors. For
the end user the most important deciding factor when choosing a repair vendor is often not the
cost involved, but the technology utilized during the repair. In order to remain competitive in
the repair business, repair vendors therefore have to develop, implement and market new
repair techniques for refurbishing degraded turbine components on an ongoing basis.

2.2) Objectives of this investigation

The majority of commercially available Ni-base braze filler metals contain B and/or Si as
melt point depressants, and display a tendency for forming brittle intermetallic phases (such
as borides and silicides) within the braze joint when used to repair wide gaps. The objective
of this investigation was to capitalize on the need for a reliable repair technique for wide
cracks in land-based turbine components by developing novel braze filler metals which ensure
high strength joints with good ductility when used in conjunction with the liquid sintering
process to repair wide cracks. The investigation was designed to:

37
 evaluate the use of two novel melt point depressants, namely Hf and Zr, in Ni- and Co-
base braze filler metals, and
 optimize the liquid phase diffusion bonding process in combination with these novel
braze filler metals with the aim of producing high strength joints with good ductility and
mechanical properties at elevated temperature approaching those of the base metal.

The results of this investigation are considered in more detail in the remainder of this
document.

38
 CHAPTER 3 - EXPERIMENT 1
MICROSTRUCTURAL EXAMINATION OF NOVEL BINARY EUTECTIC Ni-Hf
AND Ni-Zr BRAZE ALLOYS

This chapter describes the results of a series of preliminary experiments aimed at determining
whether hafnium and zirconium can be used as melt point depressants in novel Ni-base braze
filler metals.

3.1) Introduction

Examination of binary phase diagrams with nickel as solvent revealed that hafnium and
zirconium may be suitable for use as melt point depressants in novel braze filler metals for
repairing Ni-base superalloys. Nash and Nash, as quoted in ASM Handbook [2], reported the
existence of an invariant eutectic reaction in the Ni-Hf binary system at a composition of 30.5
wt.% Hf and 69.5 wt.% Ni (see Figure 49). The eutectic temperature was quoted as 1190oC.
According to Nash and Jayanth, also quoted in ASM Handbook [2], the Ni-Zr binary system
contains a eutectic point located at 13 wt.% Zr and 87 wt.% Ni (see Figure 50). The eutectic
temperature was quoted as 1170oC. In order to determine whether Hf and Zr can be used as
melt point depressants in Ni-base braze filler metals, nickel powder was mixed with hafnium
powder or zirconium powder to make up the simple binary eutectic compositions obtained
from the phase diagrams.

Figure 49 – Binary Ni-Hf phase diagram [2].

The objectives of Experiment 1 were therefore:

 to verify that eutectic Ni-Hf and Ni-Zr braze alloys melt and flow in the same way as the
braze alloys currently used in the aerospace and land-based gas turbine repair industries,
 to determine the optimum braze temperatures for both alloys, and
 to examine the microstructures of the braze joints after relatively short brazing times (40
minutes), and to compare the joint microstructures with those obtained after extended
times (18 hours) at the braze temperature.

39
 Figure 50 – Binary Ni-Zr phase diagram [2].

3.2) Experimental procedure

Inconel 738 (abbreviated as In738 throughout the remainder of this document), with the
nominal chemical composition shown in Table 8, was used as parent metal during the course
of this investigation. Flat plate was obtained by sectioning slabs of material from the root of
scrap blades using Electric Discharge Machining (EDM). The root of a turbine blade is not
exposed to appreciable temperatures or stresses in service, and material removed from this
area is often viewed as virgin-like as-cast material. The In738 material was prepared for
brazing by grinding to a thickness of 2 mm. The plates were mechanically cleaned by grit
blasting with 220 grit silicon carbide media, and then wiped with acetone to remove any
surface residue from the grit blasting process.

Table 8 – Nominal chemical composition of the In738 parent metal used during the course of
this investigation (wt.%, balance nickel).
B C Co Cr Mo Al W Ta Nb Ti Zr
0.001 0.17 8.5 16 1.7 3.4 2.6 1.7 2.0 3.4 0.1

The novel Ni-Hf braze filler metal was prepared by mixing 69.5 wt.% Ni and 30.5 wt.% Hf
(the eutectic composition) in powder form. The powder mixture was then mixed with a
binder to form a braze paste. The Ni-Zr braze paste was produced in a similar way by mixing
87 wt.% Ni powder and 13 wt.% Zr powder with a binder. Following evaporation of the
acetone from the In738 plates, a layer of Ni-Hf or Ni-Zr braze paste was applied over the
In738 plate or between two In738 plates (like a filling in a sandwich). The samples were
dried for an hour before being placed in a laboratory vacuum furnace with a maximum
temperature capability of 1260oC. It should be noted that the vacuum furnace used in this
investigation did not have any quench facilities. The samples were therefore slow cooled to
room temperature in the furnace.

40
The vacuum braze cycle used for both braze alloys was as follows:

1) Ramp up to a temperature of 450oC at a minimum rate of 9oC/minute.

2) Hold at 450oC for 20 minutes to allow the binder to burn off.
3) Ramp up to a temperature of 1150oC at a minimum rate of 9oC/minute.
4) Hold at 1150oC for 20 minutes to allow the samples to stabilize at this temperature.
5) Ramp up to a temperature of 1190oC, 1214oC or 1238oC at a minimum rate of
9oC/minute to allow the Ni-Hf braze alloy to melt; or ramp up to a temperature of
1170oC, 1204oC or 1238oC to allow the Ni-Zr alloy to melt.
6) Hold at the braze temperature for 40 minutes or 18 hours.
7) Furnace cool to room temperature.

After exposing the samples to the vacuum brazing cycle described above, the brazed samples
were sectioned and mounted in resin using conventional metallographic practices. The
mounted samples were polished, etched with Marble’s reagent to reveal the microstructure
and examined using optical and scanning electron microscopy (SEM) techniques. A Nikon
Epiphot 200 optical microscope with a maximum magnification of 1000X, and a JOEL-
Oxford scanning electron microscope equipped with ZAF software correction, were used.
The acceleration voltage was 20 kV, the % dead time approximately 50 seconds, the sample
tilt angle 0o, and the total acquisition time 150 seconds.

3.3) Results and discussion

3.3.1 Optimization of the braze temperature:

Neither the Ni-Hf, nor the Ni-Zr braze, melted at their published eutectic temperatures of
1190oC and 1170oC, respectively. This could be as a result of experimental error, such as
incorrect weighing of the Ni-Hf or Ni-Zr powders (unlikely), insufficient mixing (resulting in
segregated braze powders), or inaccurate furnace calibration. It may also be due to
insufficient thermodynamic driving force for melting as a result of the low levels of superheat
above the equilibrium eutectic temperatures.

Since neither alloy melted at their published eutectic temperatures, the braze temperatures
were raised to 1214oC for the Ni-Hf alloy and 1204oC for the Ni-Zr alloy. Partial melting
occurred, suggesting that the braze temperatures used were between the solidus and liquidus
temperatures of the filler metals. Since eutectic alloys have congruent melting points, this
implies that the compositions of the alloys deviated from the eutectic compositions given on
the phase diagrams in Figures 49 and 50.

Complete melting was eventually achieved at a furnace temperature of 1238oC for both braze
alloys. This is 48oC above the equilibrium eutectic temperature for the Ni-Hf alloy, and 68oC
above the published Ni-Zr eutectic temperature. It should be noted that three furnace
thermocouples were used to control the temperature within the vacuum furnace. All three
thermocouples were within 3oC of the specified temperature of 1238oC during brazing. Two
additional load thermocouples placed in contact with the In738 plate gave readings of 1229oC
and 1232oC. The real temperature at which the Ni-Hf and Ni-Zr alloys successfully melted
and flowed was therefore closer to 1230oC  2oC. This is 40oC higher than the published
eutectic temperature of the Ni-Hf alloy and 60oC higher than the eutectic temperature of the
Ni-Zr alloy.

41
In order to clarify this discrepancy between the published and measured melting temperatures,
the solidus and liquidus temperatures of both braze alloys were determined using Differential
Thermal Analysis (DTA), and the chemical compositions were measured using ICP-OE
(Inductively Coupled Plasma-Optical Emission) techniques. The results are summarized in
Table 9.

Table 9 – Measured chemical compositions (wt.%) of the Ni-Hf and Ni-Zr braze alloys.
Alloy Ni Hf Zr Co Cr Cu Fe P Mn Si C
Ni-Hf braze 69.995 28.87 0.70 0.02 0.04 0.01 0.14 0.007 - 0.20 0.018
Ni-Zr braze 86.212 - 13.26 0.02 0.05 0.01 0.18 - 0.01 0.24 0.018

The measured solidus temperature of the Ni-Hf braze alloy was 1176oC and the liquidus
temperature 1208oC. The solidus temperature of the Ni-Zr braze was 1163oC and the liquidus
temperature 1206oC.

These results indicate that the measured liquidus temperature of the Ni-Hf braze alloy was
approximately 18oC higher than the published eutectic temperature, whereas the liquidus of
the Ni-Zr alloy was 36oC higher than the equilibrium eutectic temperature. Since the load
thermocouples in contact with the In738 plate measured a temperature of 1230oC  2oC, the
actual brazing temperature was about 22oC to 24oC higher than the measured liquidus
temperatures determined by DTA for the Ni-Hf and Ni-Zr braze alloys used in this
investigation. Brazing at a temperature  25oC higher than the melting point of the braze
alloy is within acceptable braze practice.

The observed differences between the published and measured liquidus temperatures can
probably be attributed to deviations in chemical composition from the eutectic compositions
specified by the phase diagrams (as indicated in Table 9). The conclusion was drawn that the
Ni, Hf and Zr powders used to make up the braze alloys were not completely pure, leading to
contamination of the braze alloys by various residual elements. These elements shifted the
compositions of the braze powders away from the eutectic points, leading to higher melting
temperatures and incongruent melting points.

3.3.2 Microstructures of the Ni-Hf and Ni-Zr joints after brazing at 1230C for 40 minutes:

Optical micrographs of the Ni-Hf brazed joint after 40 minutes at 1230oC are shown in
Figures 51 to 53 at various magnifications. Figure 51 indicates that the Ni-Hf braze alloy
melts and flows adequately when processed at 1230oC for 40 minutes, and that a satisfactory
bond forms with the In738 substrate material. The braze microstructure, shown in Figures 52
and 53, consists of a mixture of coarse and fine dendrites (probably ), and a second phase
that is most likely the Ni5Hf or Ni7Hf2 intermetallic compound.

Figures 54 to 57 display optical micrographs at various magnifications of the Ni-Zr brazed

joint after 40 minutes at 1230oC. Like the Ni-Hf braze alloy, the Ni-Zr braze melts and flows
satisfactorily when processed at 1230oC for 40 minutes, and an acceptable bond forms with
the In738 substrate material (see Figures 54 and 55). The braze microstructure, shown in
Figures 55 to 57, consists of  dendrites, a Ni5Zr intermetallic phase, and a /’ flower-shaped
eutectic component.

42
 Braze

Figure 51 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 50X.

Eutectic component
containing Ni5Hf or
Ni7Hf2

Fine 
dendrites

Coarse 
dendrites

Figure 52 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 100X.

43
 Eutectic
component
containing Ni5Hf
or Ni7Hf2

Coarse 
dendrites

Figure 53 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 200X.

Fillet area

Figure 54 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 50X.

44
 Braze joint

Figure 55 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 100X.

Figure 56 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 200X.

45
 Possible eutectic
component
containing Ni5Zr

Possible
-’ eutectic
component

Figure 57 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 40 minutes. Magnification: 500X.

3.3.3 Microstructures of the Ni-Hf and Ni-Zr joints after brazing at 1230C for 18 hours:

Optical micrographs of the Ni-Hf braze joint after 18 hours at 1230oC are shown in Figures
58 to 62 at various magnifications. A brazed joint between two In738 plates is shown in
Figures 58 and 59. An acceptable fillet area, indicating good braze flow, is evident and an
acceptable bond forms with the In738 substrate material. The braze microstructure, shown in
Figures 58 to 61, consists of mixture of coarse and fine  dendrites, and a /’ flower-shaped
eutectic component that is often observed in Ni-based superalloys. The formation of the /’
eutectic component is probably promoted by the high solubility of Hf in ’ (Hf is a good ’
former). Higher levels of the /’ eutectic component most likely suppress the formation of
Hf-rich intermetallic phases in the joint. Nevertheless, the Ni5Hf or Ni7Hf2 intermetallic
phase is evident, as shown in Figure 62.

A Ni-Zr brazed joint between two In738 plates is shown at various magnifications in Figures
63 to 66 after brazing at 1230C for 18 hours. Adequate bonding of the braze alloy to the
In738 substrate is evident from Figures 63 and 64. The braze microstructure, shown in
Figures 63 to 66, consists of a mixture of coarse and fine  dendrites and a Ni5Zr intermetallic
phase.

46
Figure 58 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 50X.

Fillet area

Figure 59 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 50X.

47
Figure 60 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 100X.

Similar in
appearance to the
-’ eutectic
component often
observed in Ni-base
superalloys

Figure 61 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 200X.

48
 Possible eutectic
component
containing Ni5Hf
or Ni7Hf2

Figure 62 – Optical micrograph of a Ni-Hf brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 500X.

Figure 63 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 50X.

49
 Possible
 dendrites

Figure 64 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 100X.

Possible eutectic
component
containing Ni5Zr

Possible
 dendrites

Figure 65 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 200X.

50
 Possible Ni5Zr
phase

Figure 66 – Optical micrograph of a Ni-Zr brazed joint in In738 parent metal after brazing at
1230C for 18 hours. Magnification: 500X.

A SEM micrograph of the Ni-Hf joint brazed at 1230oC for 18 hours is shown in Figure 67.
SEM-EDS analysis was performed on the two phases highlighted by the arrows in Figure 67.
One phase was arbitrarily labelled “grain boundary particle”. The energy dispersive x-ray
analysis revealed a composition of 78.72Ni-5.97Co-4.47Cr-4.00Al-2.82W-2.53Fe-0.53Mo-
0.38Ti (wt.%) for this phase, identifying it as the  phase. The second phase was arbitrarily
designated “grain boundary film”, and had a composition of 49.21Ni-45.74Hf-2.32W-1.94Co-
0.37Cr-0.28Al-0.14Ti. The binary Ni-Hf phase diagram, shown in Figure 49, indicates that
the Ni7Hf2 intermetallic phase contains 46.5% Hf, whereas the Ni5Hf phase has a nominal Hf
content of 37.9% Hf. The “grain boundary film” phase was therefore provisionally identified
as the Ni7Hf2 intermetallic compound. The SEM-EDS analysis also indicates that small
amounts of W, Co, Cr, Al and Ti are soluble in this phase. The presence of coarse  dendrites
in the Ni-Hf braze alloy suggests that the alloy may be slightly hypoeutectic, resulting in the
formation of a proeutectic  phase on solidification. The existence of Ni7Hf2 in the braze
alloy was, however, unexpected, since the phase diagram (Figure 49) predicts a -Ni5Hf
eutectic component. It is not clear why the formation of the Ni5Hf intermetallic compound
was suppressed on solidification. The cooling rate after brazing may have been too fast to
achieve true equilibrium, or the published phase diagram may be incorrect. In-depth
investigation into the cause of this deviation from the predictions of the phase diagram falls
outside the scope of this experiment and further speculation is therefore not justified at this
time.

Figure 68 displays a SEM micrograph of the Ni-Zr joint brazed at 1230oC for 18 hours. Two
phases, highlighted by the arrows in Figure 68, were analyzed. The phase arbitrarily labelled
“grain boundary particle” had a composition of 78.82Ni-9.00Zr-3.08Co-2.97Cr-3.05Al-
1.24Fe-1.57Mo-0.26Ti (wt.%) and was identified as the  phase. This braze alloy also
appears to be slightly hypoeutectic, forming a proeutectic  phase on solidification. The

51
second phase, arbitrarily labelled “grain boundary film”, had a composition of 71.24Ni-
24.91Zr-1.45W-1.47Co-0.57Cr-0.38Mo. According to the Ni-Zr phase diagram shown in
Figure 50, the Ni5Zr phase contains between 21.32 and 25.95% Zr, suggesting that the phase
labelled “grain boundary film” is the Ni5Zr intermetallic compound that forms as part of the -
Ni5Zr eutectic component. Small amounts of W, Co, Cr and Mo are soluble in this phase.

Figure 67 – SEM micrograph of the Ni-Hf braze, showing the phases identified as  (labelled
“grain boundary particle”) and Ni7Hf2 (labelled “grain boundary film”).

Figure 68 – SEM micrograph of the Ni-Zr braze, showing the phases identified as  (labelled
“grain boundary particle”) and Ni5Zr (labelled “grain boundary film”).

52
3.4) Conclusions

 The eutectic Ni-Hf and Ni-Zr braze alloys did not melt at the equilibrium eutectic
temperatures predicted by the binary phase diagrams. This was attributed to a shift in
chemical composition away from the eutectic point due to contamination of the powders
used to mix the braze alloys. The brazing temperature for both alloys was optimized at
1230oC  2oC. This temperature is approximately 25C higher than the measured
liquidus temperatures of the alloys.

 The Ni-Hf and Ni-Zr braze alloys exhibited good flow and wetting of the substrate at a
brazing temperature of 1230oC. Narrow gaps in In738 parent metal can be repaired
successfully using either of the novel braze alloys. The joints produced at 1230oC
appeared to consist of:
1) coarse and fine  dendrites,
2) a flower-shaped -’ eutectic component (in the Ni-Hf joints), and
3) a -Ni7Hf2 or -Ni5Zr eutectic component.

 The mechanical properties of the joints were not evaluated during the course of
Experiment 1. Based on the chemistry (simple binary eutectic Ni-Hf or Ni-Zr alloys), it
is reasonable to assume that wide joints with clearances greater than 0.10 mm will not
exhibit strength levels approaching those of the base metal. In order to determine
whether high strength joints can be produced using the novel braze alloys, liquid phase
diffusion bonds were produced using the Ni-Hf and Ni-Zr braze powders in combination
with Ni-base superalloy powder within the joint. The results are presented in Chapter 4.

53
 CHAPTER 4 - EXPERIMENT 2
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL
PROPERTIES OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf
AND Ni-Zr BRAZE ALLOYS AFTER SHORT PROCESSING TIMES (40 MINUTES)

This chapter examines the microstructure and mechanical properties of liquid phase diffusion
bonds, produced by mixing the novel eutectic Ni-Hf or Ni-Zr braze alloys with MarM247 Ni-
base superalloy powder, after short processing times (40 minutes) at 1230C.

4.1) Introduction

As shown in the preceding chapter, simple binary eutectic Ni-Hf and Ni-Zr braze alloys can
be used to successfully join In738 Ni-base superalloy. The joint gap must, however, be
narrow (less than 0.15 mm in width) to allow capillary action to draw the filler metal into the
joint. Such narrow joint gaps are usually only found in the manufacturing industry. In the
repair industry, especially in IGT (industrial gas turbine) engine components, cracks are
frequently wider than 0.15 mm and the simple eutectic Ni-Hf and Ni-Zr brazes will most
likely not deliver adequate mechanical properties if used independently. In order to improve
the mechanical properties of the In738 braze joints, the Ni-Hf or Ni-Zr eutectic braze alloy
was mixed with MarM247 Ni-base superalloy powder, and processed using the liquid phase
diffusion bonding (LPDB) process described in §1.6 of the literature survey.

The objectives of Experiment 2 were therefore:

 to demonstrate that the novel Ni-Hf and Ni-Zr braze alloys can be used in conjunction
with MarM247 Ni-base superalloy powder to successfully join In738 using the liquid
phase diffusion bonding (LPDB) process, and
 to examine the microstructure and mechanical properties of the LPDB joints after short
processing times (40 minutes).

4.2) Experimental procedure

In738 plate material was produced and prepared as described in §3.2. A mixture of coarse
(+325 mesh) and fine (–325 mesh) MarM247 Ni-base superalloy powder with a nominal
composition shown in Table 10 was mixed with braze binder to form a paste. MarM247
powder, as opposed to In738, was selected because MarM247 is recognized as the equiaxed
Ni-base superalloy with the highest mechanical integrity. The MarM247 paste was applied to
the In738 plate over an area of approximately 2500 mm2 and allowed to dry. A layer of
eutectic Ni-Hf or Ni-Zr braze alloy in paste form (produced as described in §3.2) was then
applied over the dry MarM247 powder. The samples were dried for one hour, and placed into
a laboratory vacuum furnace.

Table 10 – Nominal chemical composition of the MarM247 powder used in this investigation
(wt.%, balance Ni).
B C Co Cr Hf Mo Al W Ta Ti Zr Fe
0.001 0.15 10.0 8.25 1.50 0.70 5.50 10.0 3.0 1.0 0.05 0.5

54
The vacuum braze cycle used for both braze alloys was as follows:

1) Ramp up to a temperature of 450oC at a minimum rate of 9oC/minute.

2) Hold at 450oC for 20 minutes to allow the binder to burn off.
3) Ramp up to a temperature of 1150oC at a minimum rate of 9oC/minute.
4) Hold at 1150oC for 20 minutes to allow the samples to stabilize at this temperature.
5) Ramp up to a temperature of 1230oC at a minimum rate of 9oC/minute to melt the Ni-Hf
and Ni-Zr braze alloys and to allow the melt to infiltrate the MarM247 powder.
6) Hold at 1230oC for 40 minutes.
7) Furnace cool to room temperature.

After exposing the In738 plates to the vacuum LPDB cycle described above, samples were
sectioned and mounted using conventional metallographic practices. The mounted samples
were polished and etched with Marble’s reagent to reveal the microstructure. Optical and
scanning electron microscopy techniques were then used to characterize the microstructures.

Tensile tests were performed at room temperature (21oC) and at temperatures of 540oC,
650oC, 760oC, 870oC and 980oC. In order to produce the tensile specimens, MarM247 plate
material was joined using the procedure described earlier. The tensile test samples were aged
at 870oC for 20 hours, as is typical during a normal braze repair cycle. The tensile test
specimens were prepared in a butt joint configuration, shown schematically in Figure 69.
The LPDB joint was 1.5 mm wide to simulate the typical crack widths found in Industrial Gas
Turbine (IGT) components, and was located in the centre of the sample gauge length.
Lt Lr

Dt Dg
Rf
Lg

Lo

Figure 69 – Configuration of tensile and creep rupture specimens: Dg = 4.6 mm; Lg =

18.3 mm; Lr = 21.8 mm; Lo = 46.5 mm; Rf = 3.2 mm; Lt = 9.5 mm; and Dt = 8 mm.

4.3) Results and discussion

4.3.1 Microstructural investigation:

The microstructure of the LPDB joint produced with Ni-Hf braze filler metal and MarM247
powder is shown in Figures 70 to 75. The equiaxed, light-etching particles are the original
MarM247 powder particles, and the darker areas consist of eutectic Ni-Hf braze alloy. The
variation in MarM247 particle size, evident in Figures 71 and 72, is due to the mixture of
coarse and fine MarM247 powders used in producing the superalloy powder. Such a mixture
results in a denser, less porous structure, provided the braze temperature is high enough and
the Ni-Hf braze alloy has sufficient time to flow.

Figures 72 and 73 indicate that the Ni-Hf joint is composed of MarM247 powder particles,
interspersed with a thick layer consisting of  dendrites and an intermetallic Ni5Hf or Ni7Hf2

55
phase. It is expected that the mechanical properties of this joint will fall between those of the
MarM247 superalloy and the Ni-Hf braze alloy. The dark etching component in Figures 74
and 75 most likely consists of proeutectic , and the -Ni5Hf or -Ni7Hf2 eutectic component.
The  dendrites and the Ni5Hf or Ni7Hf2 phase surround each MarM247 powder particle,
resembling a film or coating around the particles. If this film is brittle, cracks can propagate
around each powder particle and connect to form what is referred to in industry as craze-
cracked areas.

Liquid Phase
Diffusion Bond

In738 base
metal

Figure 70 – Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 50X.

Fine MarM247
powder
particles

Coarse
MarM247
powder
particles

Figure 71 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 100X.

56
 Fine MarM247
powder
particles

Original
Ni-Hf
braze alloy

Coarse
MarM247
powder
particles

Figure 72 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 200X.

 dendrites

Ni-Hf
intermetallic
compound

Figure 73 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 200X.

57
  dendrites

Ni5Hf or Ni7Hf2
intermetallic
compound

Figure 74 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 500X.

Ni5Hf or Ni7Hf2
intermetallic
compound

Figure 75 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 500X.

58
The microstructure of the LPDB joint produced using MarM247 powder and Ni-Zr braze
paste is shown in Figures 76 to 81 at various magnifications. The original MarM247 powder
particles are evident as the lighter, more equiaxed component, while the darker regions consist
of eutectic Ni-Zr braze alloy. As explained earlier, the variation in MarM247 particle size is
due to the mixture of coarse and fine powders in the joint. This ensure a denser, less porous
structure, provided the braze temperature is high enough and the Ni-Zr braze alloy has
sufficient time to flow during the braze cycle.

Figures 78 to 81 illustrate that the joint is composed of individual MarM247 powder

particles, separated by a layer consisting of  dendrites and Ni5Zr intermetallic compound
(darker regions). Since this layer surrounds each MarM247 particle, its properties will most
likely influence the properties of the joint. A brittle film will promote the propagation of
cracks and may lead to the development of craze-cracked areas in service.

Figure 82 displays a SEM micrograph of the MarM247/Ni-Hf joint after brazing at 1230oC
for 40 minutes. SEM-EDS analysis was performed on two phases, highlighted by the arrows
in Figure 82. One phase was arbitrarily labelled “grain boundary particle”, and was shown to
consist of 78.06Ni-6.68Co-3.74Cr-4.22Al-2.42W-3.01Fe-0.79Mo-0.81Ti-0.27Zr (wt.%). On
the basis of this composition, the “grain boundary particle” phase was identified as . The
second phase was arbitrarily labelled “grain boundary film”, and contained 50.31Ni-43.93Hf-
2.32W-2.50Co-0.40Cr-0.39Al-0.14Ti (wt.%). According to the Ni-Hf phase diagram (shown
in Figure 49), the Ni7Hf2 intermetallic phase contains approximately 46.5% Hf, whereas the
Ni5Hf phase contains about 37.9% Hf. On the basis of the SEM-EDS analysis, the “grain
boundary film” phase was therefore provisionally identified as the Ni7Hf2 intermetallic phase,
with some W, Co, Cr, Al and Ti in solution.

Liquid Phase
Diffusion
Bond

In738 base metal

Figure 76 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 50X.

59
 Original Ni-Zr
braze alloy

Original
MarM247 powder

Figure 77 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 100X.

Original fine
MarM247 powder

Original coarse
MarM247 powder

Original Ni-Zr
braze alloy

Figure 78 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 200X.

60
 Layer of  and Ni-
Zr intermetallic
compound

Figure 79 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 200X.

 dendrites

Ni5Zr
intermetallic
compound

Figure 80 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 500X.

61
  dendrites

Ni5Zr
intermetallic
compound

Figure 81 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 40 minutes. Magnification: 500X.

Figure 82 – SEM micrograph of the Ni-Hf braze, showing the phases identified as  (labelled
“grain boundary particle”) and Ni7Hf2 (labelled “grain boundary film”).

62
Figure 83 displays a SEM micrograph of the MarM247/Ni-Zr joint brazed at 1230oC for 40
minutes. Two phases, highlighted by the arrows in Figure 83, were analyzed using the SEM-
EDS technique. The phase arbitrarily labelled “grain boundary particle” was identified as 
dendrites (with a composition of 74.48Ni-8.76Zr-5.64Co-3.21Cr-3.27Al-2.26W-1.86Mo-
0.52Ti). The second phase, arbitrarily labelled “grain boundary film”, had a composition of
69.49Ni-24.24Zr-2.03W-2.99Co-0.62Cr-0.41Mo-0.21Ti (wt.%). This phase was identified as
the Ni5Zr intermetallic phase, with small amounts of W, Co, Cr and Mo in solution.

Figure 83 – SEM micrograph of the Ni-Zr braze, showing the phases identified as  (labelled
“grain boundary particle”) and Ni5Zr (labelled “grain boundary film”).

4.3.2 Tensile test results:

The results of tensile tests performed at various temperatures are shown in Tables 11 to 16,
and displayed graphically in Figures 84 and 85. The values shown represent the average of
three tests.
Table 11 – Joint tensile properties measured at 21oC.
Tensile properties Ni-Hf joints Ni-Zr joints MarM247 base metal
Tensile Strength (UTS) 323 MPa (46.8 ksi) 557 MPa (80.7 ksi) 960 MPa (139 ksi)
Yield Strength (YS) 239 MPa (34.7 ksi) 505 MPa (73.2 ksi) 800 MPa (116 ksi)
Elongation (%) 4.7 3.0 7.9
Reduction in area (RA) % 6.3 4.4 10.0

63
 Table 12 – Joint tensile properties measured at 540oC.
Tensile properties Ni-Hf joints Ni-Zr joints MarM247 base metal
Tensile Strength (UTS) 328 MPa (47.6 ksi) 619 MPa (89.7 ksi) 1014 MPa (147 ksi)
Yield Strength (YS) 243 MPa (35.2 ksi) 539 MPa (78.1 ksi) 801 MPa (116 ksi)
Elongation (%) 3.8 2.7 7.8
Reduction in area (RA) % 4.2 3.2 9.9

Table 13 – Joint tensile properties measured at 650oC.

Tensile properties Ni-Hf joints Ni-Zr joints MarM247 base metal
Tensile Strength (UTS) 343 MPa (49.7 ksi) 562 MPa (81.4 ksi) 1040 MPa (151 ksi)
Yield Strength (YS) 265 MPa (38.4 ksi) 513 MPa (74.4 ksi) 805 MPa (117 ksi)
Elongation (%) 2.8 2.2 7.0
Reduction in area (RA) % 3.6 2.9 9.7

Table 14 – Joint tensile properties measured at 760oC.

Tensile properties Ni-Hf joints Ni-Zr joints MarM247 base metal
Tensile Strength (UTS) 333 MPa (48.3 ksi) 593 MPa (85.9 ksi) 1000 MPa (145 ksi)
Yield Strength (YS) 282 MPa (40.8 ksi) 553 MPa (80.1 ksi) 815 MPa (118 ksi)
Elongation (%) 2.5 1.2 6.0
Reduction in area (RA) % 3.2 5.9 8.4

Table 15 – Joint tensile properties measured at 870oC.

Tensile properties Ni-Hf joints Ni-Zr joints MarM247 base metal
Tensile Strength (UTS) 289 MPa (41.9 ksi) 444 MPa (64.4 ksi) 790 MPa (115 ksi)
Yield Strength (YS) 211 MPa (30.5 ksi) 414 MPa (60.0 ksi) 650 MPa (94 ksi)
Elongation (%) 2.0 1.0 5.0
Reduction in area (RA) % 3.0 2.4 7.7

Table 16 – Joint tensile properties measured at 980oC.

Tensile properties Ni-Hf joints Ni-Zr joints MarM247 base metal
Tensile Strength (UTS) 195 MPa (28.3 ksi) 261 MPa (37.8 ksi) 506 MPa (73 ksi)
Yield Strength (YS) 141 MPa (20.4 ksi) 245 MPa (35.5 ksi) 312 MPa (45 ksi)
Elongation (%) 1.8 0.8 4.7
Reduction in area (RA) % 2.9 1.9 7.5

As shown in Table 11, the room temperature tensile and yield strengths of the Ni-Hf LPDB
joints were 34% and 30%, respectively, of those of the MarM247 base metal, whereas the
tensile and yield strengths of the Ni-Zr joints were 58% and 63%, respectively, of those of the
MarM247 base metal. The Ni-Hf joints therefore exhibited considerably lower strength than
the Ni-Zr joints after short processing times (40 minutes). The Ni-Hf joints, however,
exhibited approximately 60% of the ductility of the MarM247 base metal, whereas the Ni-Zr
joints only displayed about 40% of the base metal’s ductility. It should be noted that, after

64
short processing times, the ductility of the Ni-Zr and Ni-Hf joints was significantly higher
than the ductility of similar joints containing boron as melt point depressant. Boron-
containing joints typically exhibit ductility values in the range of 0.5 to 1.0% after short
processing times [35-37]. This observation provides initial evidence that, if higher strength
joints can be achieved by increasing the processing time and similar levels of ductility are
obtained, the use of Zr or Hf as melt point depressant in Ni-base braze alloys (instead of B)
holds considerable potential for crack repair in the IGT industry.

1100

1000

900

800

700
Stress (MPa)

Ni-Hf braze - UTS

600 Ni-Zr braze - UTS
MarM247 - UTS
500 Ni-Hf braze - YS
Ni-Zr braze - YS
400 MarM247 - YS

300

200

100

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 84 – The tensile strength (UTS) and yield strength (YS) of the Ni-Hf and Ni-Zr braze
joints and the MarM247 parent metal as a function of the test temperature (each data point
shown represents the average of three tests).

11

10

7
Ductility (%)

Ni-Hf braze - Elongation

6 Ni-Zr braze - Elongation
MarM247 - Elongation
5 Ni-Hf braze - RA
Ni-Zr braze - RA
4 MarM247 - RA

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 85 – The ductility of the Ni-Hf and Ni-Zr braze joints and the MarM247 parent metal
expressed as % elongation and % reduction in area (RA) as a function of the test temperature
(each data point shown represents the average of three tests).

65
As shown in Table 12, the tensile and yield strengths of the Ni-Hf joints at 540oC were 32%
and 30%, respectively, of the MarM247 base metal properties, whereas the tensile and yield
strengths of the Ni-Zr joints were 55% and 67%, respectively, of those of the MarM247 base
metal. The same trend is evident at 650oC, shown in Table 13, which indicates that the
tensile and yield strengths of the Ni-Hf joints were approximately 33% of the MarM247 base
metal properties, and those of the Ni-Zr joints 54% and 64%, respectively, of the MarM247
strength levels. The Ni-Hf joints were therefore considerably weaker than the Ni-Zr joints at
540oC and 650oC. As was the case at room temperature, however, the Ni-Hf joints displayed
higher ductility than the Ni-Zr joints. At 540C, the Ni-Hf joint had approximately 50% of
the MarM247 ductility, compared to about 35% for the Ni-Zr joint. At 650C, the ductility
changed to 40% and 32% of the MarM247 ductility for the Ni-Hf and the Ni-Zr joints,
respectively.

At higher temperatures, the tensile and yield strengths of the Ni-Hf LPDB joints increased
slightly relative to those of the base metal, as shown in Figure 84 and Tables 14 to 16 for test
temperatures of 760oC, 870oC and 980oC, respectively. The Ni-Hf joint tensile strength
increased from 33% of the base metal strength at 760oC, to 37% at 870oC and 39% at 980oC.
This trend was not observed for the Ni-Zr joints, with the tensile strength changing from 59%
of the base metal strength, to 56% and finally to 52% as the test temperature increased. The
yield strength of the Ni-Hf joints changed from 35% of the base metal yield strength at
760C, to 33% at 870C and 45% at 980C. Corresponding yield strength values for the Ni-
Zr joint were at 68%, 64% and 79% of the base metal yield strength at 760C, 870C and
980C, respectively. After short processing times (40 minutes), the Ni-Hf joints therefore
remained considerably weaker than the Ni-Zr joints as the test temperature increased. The
ductility of the Ni-Hf joints was, however, higher than that of the Ni-Zr joints, regardless of
test temperature. Ductility values for the Ni-Hf joints changed from 42% of the base metal
ductility at 760C, to 40% at 870C and 38% at 980C. Corresponding ductility values for
the Ni-Zr joints were approximately 20% of the base metal ductility at 760C and 870C, and
about 17% at 980C.

4.4) Conclusions

 A relatively dense, low porosity LPDB joint, with a microstructure consisting of

MarM247 powder particles surrounded by Ni-Hf or Ni-Zr braze alloy, formed during
processing at 1230oC for 40 minutes. The braze microstructure consisted of  dendrites
and Ni5Zr or Ni7Hf2 intermetallic compound in the form of thick layers surrounding each
MarM247 powder particle.

 At all test temperatures (room temperature and at elevated temperatures of 540oC, 650oC,
760oC, 870oC and 980oC), the tensile and yield strengths of the Ni-Zr joints exceeded
those of the Ni-Hf joints after short processing times (40 minutes). The tensile and yield
strengths of the Ni-Hf joints were typically only about 30 to 40% of those of the
MarM247 base metal. For the Ni-Zr joints, the tensile and yield strengths were in the
range of 52 to 62% of the MarM247 base metal properties.

 The Ni-Hf joints, however, displayed superior ductility compared to the Ni-Zr joints.
Depending on the test temperature, the Ni-Hf joints exhibited ductility on the order of 38
to 60% of that of the MarM247 base metal, whereas the Ni-Zr joints had ductility values
in the region of 17 to 40% of the ductility of the MarM247 base metal.

66
 The ductility of the novel braze alloys containing Hf or Zr as melt point depressant was
significantly higher than the ductility of similar joints containing boron as melt point
depressant. Boron-containing joints typically exhibit ductility values in the range of 0.5
to 1.0% after short processing times.

 The results described in this chapter were obtained after short processing times of 40
minutes. Longer brazing cycles are frequently used to diffuse melt point depressants into
the parent metal and to reduce the occurrence of brittle eutectic phases in the braze joint
(as described in Chapter 1). In order to determine whether such an extended brazing
cycle will reduce or even eliminate the eutectic films of  and Ni5Zr or Ni7Hf2
intermetallic compound around the MarM247 powder particles, the brazing time was
increased to 4 hours at 1230oC. The results obtained after extended brazing cycle are
described in Chapter 5.

67
 CHAPTER 5 - EXPERIMENT 3
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL
PROPERTIES OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf
AND Ni-Zr BRAZE ALLOYS AFTER LONGER PROCESSING TIMES (4 HOURS)

5.1) Introduction

The results described in Chapter 4 revealed that a reasonably dense, low porosity LPDB joint,
consisting of Ni-Hf or Ni-Zr braze filler metal and MarM247 Ni-base superalloy powder, can
be produced using a short brazing cycle (40 minutes at 1230oC). The braze joints consisted of
the original MarM247 powder particles encircled by a layer of braze alloy containing  phase
and a Ni5Zr or Ni7Hf2 intermetallic compound.

Experiment 3, described in this chapter, aimed at determining whether a longer brazing time
(4 hours at 1230oC) leads to an improvement in the mechanical properties of the brazed joint
due to diffusion of the melt point depressant (Hf or Zr) from the braze alloy into the parent
metal. If diffusion of the melt point depressant is promoted by the longer brazing cycle, the
layers of eutectic braze alloy around the MarM247 powder particles may be reduced in
thickness or even eliminated.

5.2) Experimental procedure

In order to study the influence of a longer brazing time, LPDB samples joining In738 parent
material were prepared using the procedure described in §4.2. The vacuum brazing cycle,
shown below, differed from that used in Experiment 2 (Chapter 4) only in the extended
brazing time (4 hours).

1) Ramp up to a temperature of 450oC at a minimum rate of 9oC/minute.

2) Hold at 450oC for 20 minutes to allow the binder to burn off.
3) Ramp up to a temperature of 1150oC at a minimum rate of 9oC/minute.
4) Hold at 1150oC for 20 minutes to allow the samples to stabilize at this temperature.
5) Ramp up to a temperature of 1230oC at a minimum rate of 9oC/minute to melt the Ni-Hf
and Ni-Zr braze alloys and to allow the melt to infiltrate the MarM247 powder.
6) Hold at 1230oC for 4 hours.
7) Furnace cool to room temperature.

After exposing the samples to the vacuum LPDB cycle described above, the samples were
sectioned, mounted and polished using conventional metallographic practices. The polished
metallographic samples were etched with Marble’s reagent, and characterized using optical
and scanning electron microscopy techniques.

In order to characterize the mechanical properties of the joints, tensile tests were performed at
room temperature, and at temperatures of 540oC, 650oC, 760oC, 870oC and 980oC. The
tensile samples were prepared by joining MarM247 plate material using the LPDB process
and the vacuum brazing cycle described above. After brazing, the tensile test samples were
aged at 870oC for 20 hours. The dimensions of the tensile test samples used in this
investigation were previously given in Figure 69. The LPDB joint was 1.5 mm wide to
simulate typical crack widths found in Industrial Gas Turbine (IGT) components, and was
located in the centre of the sample gauge length.

68
5.3) Results and discussion

5.3.1 Microstructural investigation:

The microstructure of the Ni-Hf LPDB joint processed at 1230C for 4 hours is shown in
Figures 86 to 89. The joint consists of the original MarM247 powder particles, separated by
layers of Ni-Hf braze alloy. It is evident from Figures 87 and 88 that the layers of  and
Ni7Hf2 intermetallic compound between the MarM247 powder particles are finer and less
continuous (compared with those shown in Figures 72 and 73 for a brazing time of 40
minutes). This supports the premise stated in §5.1 that an extended brazing cycle promotes
the diffusion of the melt point depressant (Hf) into the parent metal and minimizes the
occurrence of intermetallic-containing eutectic component in the braze. This should lead to
an improvement in mechanical properties.

The microstructure of the LPDB joint produced using MarM247 powder and Ni-Zr braze
paste is shown in Figures 90 to 93 at various magnifications. The original MarM247 powder
particles are evident as the lighter, more equiaxed component, while the darker regions consist
of Ni-Zr braze alloy. The braze alloy contains  dendrites and Ni5Zr intermetallic compound,
but the layers of braze metal around each MarM247 particle appear less continuous and
considerably finer than those observed after processing for 40 minutes at 1230C (Figures 79
to 81).

The islands of braze alloy retained between the MarM247 particles resemble the intergranular
carbides often observed in Ni-base superalloys (see Figure 2). The structures shown in
Figures 86 to 93 can be viewed as consisting of a mixture of fine and coarse-grained material,
dispersed with intergranular  and Ni7Hf2 or Ni5Zr compounds. It is hoped that the mixture of
coarse and fine grains within the braze joint will ensure a good combination of high tensile
strength (promoted by a fine grain size) and good creep rupture strength (promoted by a
coarse grain structure).

Original
MarM247
powder particles

Figure 86 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 50X.

69
 Fine
MarM247
powder
particles

Coarse
MarM247
powder
particles

Figure 87 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 100X.

Original Ni-
Hf braze
alloy

Original
MarM247
powder
particles

Figure 88 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 200X.

70
  dendrites

Ni7Hf2
intermetallic
compound

Figure 89 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 500X.

LPDB joint

In738 base
metal

Figure 90 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 50X.

71
 Coarse
MarM247
powder
particles

Fine MarM247
powder
particles

Figure 91 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 100X.

 dendrites
within the
original Ni-Zr
braze

MarM247
powder
particles

Figure 92 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 200X.

72
  phase within the
original Ni-Zr
braze

Ni5Zr intermetallic
compound

Figure 93 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 4 hours. Magnification: 500X.

Figure 94 displays a SEM micrograph of the MarM247/Ni-Hf joint brazed at 1230oC for 4
hours. Two phases, highlighted by the arrows in Figure 94, were analyzed using the SEM-
EDS technique. The phase arbitrarily labelled “grain boundary particle” was identified as 
dendrites (with a composition of 75.27Ni-6.26Co-6.88Cr-3.67Al-2.29W-2.22Fe-1.93Mo-
0.73Ti-0.74Zr) (wt. %). The second phase, arbitrarily labelled “grain boundary film”, had a
composition of 49.50Ni-46.56Hf-1.36W-2.01Co-0.38Cr-0.19Ti (wt.%). This phase was
provisionally identified as the Ni7Hf2 intermetallic phase, with small amounts of W, Co, Cr
and Ti in solution.

A SEM micrograph of the MarM247/Ni-Zr joint after brazing at 1230oC for 4 hours is shown
in Figure 95. SEM-EDS analysis revealed that the phase arbitrarily labelled “grain boundary
particle” consisted of 79.54Ni-5.90Co-5.34Cr-3.98Al-1.81W-11.11Fe-0.85Mo-0.64Ti-0.83Zr
(wt.%). On the basis of this composition, the “grain boundary particle” phase was identified
as . The second phase was arbitrarily labelled “grain boundary film”, and consisted of
71.35Ni-24.47Zr-0.70W-2.28Co-0.61Cr-0.16Mo-0.43Al (wt.%). This phase was identified as
the Ni5Zr intermetallic phase, with some W, Co, Cr, Mo and Al in solution.

5.3.2 Tensile test results:

The results of tensile tests performed at various temperatures are shown in Tables 17 to 22
and presented graphically in Figures 96 to 99. The values shown represent the average of
three tests.

73
Figure 94 – SEM micrograph of the Ni-Hf braze, showing the phases identified as  (labelled
“grain boundary particle”) and Ni7Hf2 (labelled “grain boundary film”).

Figure 95 – SEM micrograph of the Ni-Zr braze, showing the phases identified as  (labelled
“grain boundary particle”) and Ni5Zr (labelled “grain boundary film”).

74
 Table 17 – Joint tensile properties measured at 21oC.
Ni-Hf joints Ni-Hf joints Ni-Zr joints Ni-Zr joints MarM247
Tensile properties
1230oC for 40 min. 1230oC for 4 hours 1230oC for 40 min. 1230oC for 4 hours Base metal
Tensile strength 323 MPa (46.8 ksi) 344 MPa (49.8 ksi) 557 MPa (80.7 ksi) 564 MPa (81.7 ksi) 960 MPa (139 ksi)
Yield strength 239 MPa (34.7 ksi) 302 MPa (43.8 ksi) 505 MPa (73.2 ksi) 511 MPa (74.0 ksi) 800 MPa (116 ksi)
Elongation (%) 4.7 2.7 3.0 3.0 7.9
Reduction in area (%) 6.3 3.9 4.4 4.5 10.0

Table 18 – Joint tensile properties measured at 540oC.

Ni-Hf joints Ni-Hf joints Ni-Zr joints Ni-Zr joints MarM247
Tensile properties
1230oC for 40 min. 1230oC for 4 hours 1230oC for 40 min. 1230oC for 4 hours Base metal
Tensile strength 328 MPa (47.6 ksi) 406 MPa (58.9 ksi) 619 MPa (89.7 ksi) 625 MPa (90.6 ksi) 1014 MPa (147 ksi)
Yield strength 243 MPa (35.2 ksi) 329 MPa (47.7 ksi) 539 MPa (78.1 ksi) 556 MPa (80.6 ksi) 801 MPa (116 ksi)
Elongation (%) 3.8 3.0 2.7 2.8 7.9
Reduction in area (%) 4.2 3.8 3.2 3.4 10.0

Table 19 – Joint tensile properties measured at 650oC.

Ni-Hf joints Ni-Hf joints Ni-Zr joints Ni-Zr joints MarM247
Tensile properties
1230oC for 40 min. 1230oC for 4 hours 1230oC for 40 min. 1230oC for 4 hours Base metal
Tensile strength 343 MPa (49.7 ksi) 408 MPa (59.1 ksi) 562 MPa (81.4 ksi) 569 MPa (82.5 ksi) 1040 MPa (151 ksi)
Yield strength 265 MPa (38.4 ksi) 308 MPa (44.7 ksi) 513 MPa (74.4 ksi) 501 MPa (72.6 ksi) 805 MPa (117 ksi)
Elongation (%) 2.8 3.1 2.2 2.3 7.0
Reduction in area (%) 3.6 3.8 2.9 2.9 9.7

Table 20 – Joint tensile properties measured at 760oC.

Ni-Hf joints Ni-Hf joints Ni-Zr joints Ni-Zr joints MarM247
Tensile properties
1230oC for 40 min. 1230oC for 4 hours 1230oC for 40 min. 1230oC for 4 hours Base metal
Tensile strength 333 MPa (48.3 ksi) 464 MPa (67.2 ksi) 593 MPa (85.9 ksi) 602 MPa (87.3 ksi) 1000 MPa (145 ksi)
Yield strength 282 MPa (40.8 ksi) 380 MPa (55.1 ksi) 553 MPa (80.1 ksi) 524 MPa (75.9 ksi) 815 MPa (118 ksi)
Elongation (%) 2.5 2.9 1.2 2.2 6.0
Reduction in area (%) 3.2 3.0 5.9 4.3 8.4

Table 21 – Joint tensile properties measured at 870oC.

Ni-Hf joints Ni-Hf joints Ni-Zr joints Ni-Zr joints MarM247
Tensile properties
1230oC for 40 min. 1230oC for 4 hours 1230oC for 40 min. 1230oC for 4 hours Base metal
Tensile strength 289 MPa (41.9 ksi) 342 MPa (49.6 ksi) 444 MPa (64.4 ksi) 480 MPa (69.6 ksi) 790 MPa (115 ksi)
Yield strength 211 MPa (30.5 ksi) 277 MPa (40.1 ksi) 414 MPa (60.0 ksi) 404 MPa (58.6 ksi) 650 MPa (94 ksi)
Elongation (%) 2.0 2.2 1.0 2.1 5.0
Reduction in area (%) 3.0 3.2 2.4 3.1 7.7

75
 Table 22 – Joint tensile properties measured at 980oC.
Ni-Hf joints Ni-Hf joints Ni-Zr joints Ni-Zr joints MarM247
Tensile properties
1230oC for 40 min. 1230oC for 4 hours 1230oC for 40 min. 1230oC for 4 hours Base metal
Tensile strength 195 MPa (28.3 ksi) 266 MPa (38.6 ksi) 261 MPa (37.8 ksi) 308 MPa (44.7 ksi) 506 MPa (73 ksi)
Yield strength 141 MPa (20.4 ksi) 206 MPa (29.9 ksi) 245 MPa (35.5 ksi) 214 MPa (31.0 ksi) 312 MPa (45 ksi)
Elongation (%) 1.8 2.7 0.8 3.1 4.7
Reduction in area (%) 2.9 3.1 1.9 6.2 7.5

1100

1000

900

800

700
Stress (MPa)

Ni-Hf - 40 min @ 1230°C

600 Ni-Hf - 4 hrs @ 1230°C
MarM247
500 Ni-Zr - 40 min @ 1230°C
Ni-Zr - 4 hrs@ 1230°C
400

300

200

100

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 96 – The tensile strength (UTS) of the Ni-Hf and Ni-Zr braze joints and the MarM247
parent metal as a function of the test temperature and braze time (each data point shown
represents the average of three tests).

1100

1000

900

800

700
Stress (MPa)

Ni-Hf - 40 min @ 1230°C

600 Ni-Hf - 4 hrs @ 1230°C
MarM247
500 Ni-Zr - 40 min @ 1230°C
Ni-Zr - 4 hrs@ 1230°C
400

300

200

100

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 97 – The yield stress of the Ni-Hf and Ni-Zr braze joints and the MarM247 parent
metal as a function of the test temperature and braze time (each data point shown represents
the average of three tests).

76
 11

10

% Elongation
Ni-Hf - 40 min @ 1230°C
6 Ni-Hf - 4 hrs @ 1230°C
MarM247
5 Ni-Zr - 40 min @ 1230°C
Ni-Zr - 4 hrs@ 1230°C
4

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 98 – The % elongation measured for the Ni-Hf and Ni-Zr braze joints and the
MarM247 parent metal as a function of the test temperature and braze time (each data point
shown represents the average of three tests).

11

10

8
% Reduction in area

7
Ni-Hf - 40 min @ 1230°C
6 Ni-Hf - 4 hrs @ 1230°C
MarM247
5 Ni-Zr - 40 min @ 1230°C
Ni-Zr - 4 hrs@ 1230°C
4

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 99 – The % reduction in area measured for the Ni-Hf and Ni-Zr braze joints and the
MarM247 parent metal as a function of the test temperature and braze time (each data point
shown represents the average of three tests).

As shown in Figures 96 and 97, the tensile and yield strengths of the Ni-Hf braze alloy were
improved considerably over the entire range of tensile test temperatures by increasing the
brazing time. The measured joint tensile strength values ranged from approximately 36% of
the base metal tensile strength at room temperature, to about 53% at a test temperature of
980C. The measured yield strength values ranged from about 38% (at room temperature) to
56% of the base metal yield strength (at 980°C). This suggests that the longer braze cycle
facilitated the diffusion of the melt point depressant into the parent metal (as suggested by
Figures 87 to 89), reducing the amount of brittle intermetallic compound in the joint and
allowing the mechanical properties of the braze joint to approach those of the MarM247

77
superalloy powder. The longer braze time reduced the ductility of the joint at lower test
temperatures, but at higher temperatures the ductility was improved (Figures 98 and 99).

The tensile and yield strength values of the Ni-Zr braze joint exceeded those of the Ni-Hf
joint at all test temperatures, attaining levels of approximately 60% of the base metal tensile
strength, and up to 70% of the base metal yield strength after brazing for 4 hours (Figures 96
and 97). The longer braze time, however, did not improve the joint strength to any
significant extent. A significant improvement in ductility was, however, evident after 4 hours
at 1230C, particularly at higher test temperatures. The measured joint ductility ranged from
about 38% of the base metal ductility at room temperature, to approximately 74% at 980C
(Figures 98 and 99).

5.4) Conclusions

 A relatively dense, low porosity LPDB joint, with a microstructure consisting of

MarM247 powder particles surrounded by Ni-Hf or Ni-Zr braze alloy, formed during
processing at 1230oC for 4 hours. The braze microstructure consisted of  dendrites and
Ni5Zr or Ni7Hf2 intermetallic compound.

 Brazing at 1230C for 4 hours, compared to 40 minutes, decreased the amount of

Ni5Zr/Ni7Hf2 intermetallic compound, and reduced the thickness of the layers of braze
alloy surrounding the MarM247 powder particles. The regions of braze alloy between
the powder particles appeared less continuous after 4 hours, and in many areas only
isolated  dendrites and Ni5Zr/Ni7Hf2 phases existed between the superalloy particles.

 At all test temperatures, the tensile and yield strength values of the Ni-Zr joints exceeded
those of the Ni-Hf joints. The measured Ni-Hf joint strength ranged from approximately
36% of the base metal tensile strength and 38% of the yield strength at room temperature,
to about 53% of the tensile strength and 56% of the yield strength at a test temperature of
980C. The Ni-Zr joint attained strength levels of approximately 60% of the base metal
tensile strength, and up to 70% of the base metal yield strength after a braze time of 4
hours.

 A longer braze cycle (4 hours compared to 40 minutes at 1230C) significantly improved

the strength of the Ni-Hf joint, but had little effect on the strength of the Ni-Zr joint.
Both joints exhibited a significant improvement in ductility after the extended braze
cycle, particularly at higher test temperatures.

 The ductility of the novel braze alloys containing Hf or Zr as melt point depressant was
significantly higher than the ductility of similar joints containing boron as melt point
depressant. Boron-containing joints typically exhibit ductility values in the range of 0.5
to 1.0% after short processing times.

 The results described in this chapter were obtained after processing times of 4 hours. In
order to determine whether an even longer brazing cycle will further reduce or even
eliminate the eutectic films of  and Ni5Zr or Ni7Hf2 intermetallic compound around the
MarM247 powder particles, the brazing time was increased to 12 hours at 1230oC. The
results of this experiment are described in Chapter 6.

78
 CHAPTER 6 - EXPERIMENT 4
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL
PROPERTIES OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf
AND Ni-Zr BRAZE ALLOYS AFTER EXTENDED BRAZING TIMES (12 HOURS)

6.1) Introduction

The results described in Chapter 5 revealed that a reasonably dense, low porosity LPDB joint,
consisting of Ni-Hf or Ni-Zr braze filler metal and MarM247 Ni-base superalloy powder, can
be produced by brazing for 4 hours at 1230oC. After processing the braze joints consisted of
the original MarM247 powder particles surrounded by a layer of  phase and Ni5Zr or Ni7Hf2
intermetallic compound. Increasing the braze time from 40 minutes to 4 hours reduced the
thickness of the braze layer around the MarM247 powder particles and increased the tensile
and yield strengths of the Ni-Hf braze joint at all test temperatures. The strength of the Ni-Zr
joint was not improved significantly by extending the braze cycle, but the ductility of both
braze joints was improved, particularly at higher test temperatures. These results suggest that
the extended braze cycle promoted the diffusion of the melt point depressants from the braze
metal into the parent plate, partially decomposing the brittle eutectic phases within the braze
joints, and resulting in improved mechanical properties.

Since extending the braze cycle from 40 minutes to 4 hours yielded promising results,
Experiment 4 aimed at determining whether an even longer braze cycle (12 hours at 1230C,
followed by solution annealing at the same temperature for an additional 4 hours) promotes
further decomposition of the layers of  and Ni7Hf2/Ni5Zr surrounding the MarM247 powder
particles. Creep rupture tests were also performed at various stress levels at three elevated
temperatures in order to obtain preliminary creep/stress rupture properties for the
MarM247/Ni-Zr and MarM247/Ni-Hf braze formulations.

6.2) Experimental procedure

Liquid phase diffusion bonding was used to join In738 parent material using the eutectic Ni-
Hf and Ni-Zr braze alloys described in earlier chapters. The braze samples were prepared
using the procedure presented in §4.2. The vacuum brazing cycle, shown below, differed
from those described in Chapters 4 and 5 only in the extended brazing time (12 hours) used.

1) Ramp up to a temperature of 450oC at a minimum rate of 9oC/minute.

2) Hold at 450oC for 20 minutes to allow the binder to burn off.
3) Ramp up to a temperature of 1150oC at a minimum rate of 9oC/minute.
4) Hold at 1150oC for 20 minutes to allow the samples to stabilize at this temperature.
5) Ramp up to a temperature of 1230oC at a minimum rate of 9oC/minute to melt the Ni-Hf
and Ni-Zr braze alloys and to allow the melt to infiltrate the MarM247 powder.
6) Hold at 1230oC for 12 hours.
7) Furnace cool to room temperature.

After joining the samples using the LPDB cycle described above, the brazed samples were
placed in a production vacuum furnace equipped with quenching facilities in order to solution
anneal the joints. Gas quenching, using 2 bar argon pressure, was utilized on completion of

79
the solution annealing cycle. A conventional vacuum solution heat treatment cycle, similar to
that applied in industry to MarM247 Siemens Westinghouse blades operating in W501F
turbine engines, was applied. The solution heat treatment used in this experiment was as
follows:

1) Ramp up to a temperature of 1230oC at a minimum rate of 14oC per minute.

2) Hold at 1230oC for 4 hours.
3) Quench to room temperature at a rate of 50C per minute.

After exposing the plates to the vacuum LPDB and solution heat treatment cycles described
above, samples were sectioned, mounted and polished using conventional metallographic
techniques. The polished metallographic samples were then etched with Marble’s reagent,
and characterized using optical and scanning electron microscopy.

In order to obtain preliminary creep/stress rupture properties for the novel braze formulations,
creep rupture tests were performed at three temperatures: 845oC, 900oC and 980oC. Samples
were subjected to three different applied stress levels at each test temperature, as indicated in
Table 23. The creep rupture test samples were prepared by joining In738 plate material using
the LPDB process and the vacuum brazing and solution heat treatment cycles described
above. The dimensions of the test specimens used in this investigation were identical to those
given in Figure 69. The joint was located in the centre of the gauge length and the joint gap
was maintained at 1.5 mm to simulate typical crack widths found in Industrial Gas Turbine
(IGT) components.

Table 23 - Creep rupture test conditions.

Test temperature Stress level #1 Stress level #2 Stress level #3
845oC 345 MPa (50 ksi) 276 MPa (40 ksi) 228 MPa (33 ksi)
900oC 242 MPa (35 ksi) 186 MPa (27 ksi) 152 MPa (22 ksi)
980oC 138 MPa (20 ksi) 104 MPa (15 ksi) 76 MPa (11 ksi)

The metallographic examination and creep rupture tests were performed on samples in the
solution annealed condition. In industry, In738 turbine components normally receive an
aging heat treatment after solution annealing – typically at a temperature of approximately
1120oC for 2 hours, followed by an additional 24 hours at 845oC. Not aging the samples used
in this experiment was an oversight on the author’s part due to a misunderstanding. Even
though the samples were not aged, the results presented in this chapter still allow a
comparison between the Ni-Zr and Ni-Hf braze compositions. It must be borne in mind,
however, that since the samples were tested in the solution heat treated condition without any
aging, the material is likely to be softer and to exhibit enhanced ductility and reduced creep
rupture strength compared to fully aged specimens.

6.3) Results and discussion

6.3.1 Microstructural investigation:

The microstructure of the Ni-Hf LPDB joint processed at 1230C for 12 hours, followed by
solution annealing, is shown in Figures 100 to 103. The joint consists of the original
MarM247 powder particles, surrounded by layers of Ni-Hf braze alloy. This microstructure
resembles the microstructure of the Ni-Hf joint processed at 1230C for 4 hours, shown in

80
Figures 86 to 89. The layers of  and Ni7Hf2 between the MarM247 powder particles appear
finer and less continuous (compared with those shown in Figures 72 and 73 for a brazing
time of 40 minutes), supporting the conclusion that an extended brazing cycle promotes the
diffusion of the melt point depressant (Hf) into the parent metal and minimizes the amount of
brittle intermetallic compound in the braze.

The microstructure of the LPDB joint produced with MarM247 powder and Ni-Zr braze paste
is shown in Figures 104 to 107 at various magnifications. The original MarM247 powder
particles are evident as the lighter, more equiaxed component, while the darker regions consist
of retained eutectic Ni-Zr braze alloy. The microstructure of the MarM247/Ni-Zr joint
processed at 1230C for 12 hours, followed by solution heat treatment, resembles the
microstructure shown in Figures 90 to 93 for a braze cycle of 4 hours. The braze alloy
contains  dendrites and Ni5Zr intermetallic compound, but the layers of braze metal around
each MarM247 particle are less continuous and significantly finer than those observed after
processing for 40 minutes at 1230C (Figures 79 to 81). The islands of braze alloy retained
between the MarM247 powder particles resemble the intergranular carbides often observed in
Ni-base superalloys (see Figure 2). The structures shown in Figures 100 to 107 can be
viewed as consisting of a mixture of fine and coarse-grained superalloy material, dispersed
with intergranular  and Ni7Hf2/Ni5Zr intermetallic compound. It is hoped that the mixture of
coarse and fine grains within the braze joint will ensure a good combination of high tensile
strength (promoted by a fine grain size) and good creep rupture strength (promoted by a
coarse grain structure).

Original MarM247
powder particles

Original
Ni-Hf braze
alloy

Figure 100 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 50X.

81
 Original coarse
MarM247
powder

Original fine
MarM247
powder

Isolated
microporosity

Original
Ni-Hf braze
alloy

Figure 101 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 100X.

Ni7Hf2
intermetallic
compound

 phase surrounded
by Ni7Hf2
intermetallic
compound

Figure 102 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 200X.

82
 Ni7Hf2
intermetallic
compound

Figure 103 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 500X.

Original
MarM247
powder

Isolated
microporosity

Figure 104 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 50X.

83
 Original coarse
MarM247
powder

Original fine
MarM247
powder

Intergranular
Ni5Zr
intermetallic
phases

Figure 105 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 100X.

Original
MarM247
powder

Intergranular
Ni5Zr
intermetallic
compound

 phase
surrounded by
intermetallic
compound

Isolated
microporosity

Figure 106 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 200X.

84
  phase
surrounded by
intermetallic
compound

Ni5Zr
intermetallic
compound

Figure 107 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and
quenching. Magnification: 500X.

SEM micrographs of the Ni-Hf joint, processed at 1230C for 12 hours, followed by solution
heat treatment at 1230C for 4 hours, are shown in Figures 108 and 109. Two phases within
the original braze alloy were analyzed, indicated by the arrows in Figure 108. As indicated
on the micrograph, one of these phases was identified as , consisting of 73.14Ni-6.33Co-
7.37Cr-2.73Al-6.10W-1.49Mo-0.65Ti (wt.%). The phase surrounding the  particles was
identified as Ni7Hf2, with a composition of 52.96Ni-43.90Hf-2.18Co-0.46Cr-0.16Ti-0.34W
(wt.%). The more angular intergranular phase, highlighted by the arrows in Figure 109, was
shown to consist of 53.01Ni-43.09Hf-2.43Co-0.82Cr-0.20Ti-0.45W (wt.%), and was
therefore also identified as Ni7Hf2.

Figures 110 and 111 display SEM micrographs of the Ni-Zr joint after processing at 1230C
for 12 hours, followed by solution annealing at 1230C for 4 hours. Two of the intergranular
phases in Figure 110 were analyzed, as indicated by the arrows. One of these phases,
consisting of 75.46Ni-5.27Co-5.23Cr-3.42Al-7.17W-1.16Fe-0.92Mo-0.70Ti-0.68Zr (wt%.),
was identified as  phase. The second phase had a composition of 73.63Ni-23.75Zr-2.09Co-
0.53Al (wt.%), and was identified as Ni5Zr. The intergranular phase highlighted by the
arrows in Figure 111 had a similar composition (70.78Ni-25.44Zr-2.27Co-0.81Cr-0.44Al-
0.25W), and was therefore also identified as Ni5Zr intermetallic compound.

85
  phase surrounded
by Ni7Hf2
intermetallic
compound

Ni7Hf2
intermetallic
compound

Figure 108 – SEM micrograph of the Ni-Hf braze, showing the phases identified as  and
Ni7Hf2.

Ni7Hf2
intermetallic
phase

Figure 109 – SEM micrograph of the Ni-Hf braze, showing islands of intergranular Ni7Hf2
intermetallic compound.

86
  phase
surrounded by
Ni5Zr
intermetallic
compound

Ni5Zr
intermetallic
compound

Figure 110 – SEM micrograph of the Ni-Zr braze, showing the phases identified as  and
Ni5Zr.

Ni5Zr
intermetallic
compound

Figure 111 – SEM micrograph of the Ni-Zr braze, showing islands of intergranular Ni5Zr
intermetallic compound.

87
 6.3.2 Creep rupture test results:

The results of creep rupture tests performed at three different stress levels at a temperature of
845oC are documented in Table 24, and displayed graphically in the form of a Larson-Miller
plot in Figure 112. In Table 24, the creep rupture properties of the MarM247/Ni-Hf and
MarM247/Ni-Zr joints are compared with those of the In738 base metal in the solution
annealed condition.

Table 24 - Creep rupture properties at 845oC of the In738 base metal, and LPDB joints produced with
MarM247 superalloy powder and Ni-Hf or Ni-Zr braze alloy. The test samples were produced with a
joint gap of 1.5 mm to simulate a worst-case crack repair scenario.
Failure Hours to % Reduction
Specimen identification % Elongation
location failure in area
APPLIED STRESS 345 MPa (50 ksi)
BM-1T (In738 as-cast base metal) Base metal 30.0 5.0 7.0
BM-2T (In738 as-cast base metal) Base metal 24.7 6.0 7.5
Average for the In738 as-cast base metal 27.4 5.5 7.3
738-1 (In738 brazed with MarM247/Ni-Zr) Joint 20.5 3.6 4.6
738-2 (In738 brazed with MarM247/Ni-Zr) Joint 21.8 3.0 4.0
738-3 (In738 brazed with MarM247/Ni-Zr) Joint 20.9 3.7 4.7
Average for the MarM247/Ni-Zr joint 21.1 3.4 4.4
738-A (In738 brazed with MarM247/Ni-Hf) Joint 20.6 3.7 5.8
738-B (In738 brazed with MarM247/Ni-Hf) Joint 18.2 3.0 4.5
738-C (In738 brazed with MarM247/Ni-Hf) Joint 17.0 4.5 6.5
Average for the MarM247/Ni-Hf joint 18.6 3.7 5.6
APPLIED STRESS 276 MPa (40 ksi)
BM-3T (In738 as-cast base metal) Base metal 100.0 6.2 8.7
BM-4T (In738 as-cast base metal) Base metal 137.8 6.1 8.4
Average for the In738 as-cast base metal 118.9 6.2 8.6
738-4 (In738 brazed with MarM247/Ni-Zr) Joint 75.2 5.7 8.6
738-5 (In738 brazed with MarM247/Ni-Zr) Joint 78.8 3.9 7.9
738-6 (In738 brazed with MarM247/Ni-Zr) Joint 70.1 4.9 7.2
Average for the MarM247/Ni-Zr joint 74.7 4.8 7.9
738-D (In738 brazed with MarM247/Ni-Hf) Joint 73.2 4.9 7.6
738-E (In738 brazed with MarM247/Ni-Hf) Joint 70.9 5.9 8.9
738-F (In738 brazed with MarM247/Ni-Hf) Joint 68.9 5.2 8.4
Average for the MarM247/Ni-Hf joint 71.0 5.3 8.3
APPLIED STRESS 228 MPa (33 ksi)
BM-5T (In738 as-cast base metal) Base metal 200.0 6.6 8.8
BM-6T (In738 as-cast base metal) Base metal 344.6 6.0 8.6
Average for the In738 as-cast base metal 272.3 6.3 8.7
738-7 (In738 brazed with MarM247/Ni-Zr) Joint 253.6 4.2 6.0
738-8 (In738 brazed with MarM247/Ni-Zr) Joint 240.0 4.4 6.3
738-9 (In738 brazed with MarM247/Ni-Zr) Joint 232.2 4.9 6.7
Average for the MarM247/Ni-Zr joint 241.9 4.5 6.3
738-G (In738 brazed with MarM247/Ni-Hf) Joint 209.8 5.3 7.5
738-H (In738 brazed with MarM247/Ni-Hf) Joint 220.8 4.9 6.9
738-I (In738 brazed with MarM247/Ni-Hf) Joint 214.3 5.2 7.0
Average for the MarM247/Ni-Hf joint 215.0 5.1 7.1

88
As shown in Table 24, the creep life to failure of the braze joints at 845C was well below
that of the In738 base metal at all three applied stress levels. At an applied stress level of 345
MPa, the MarM247/Ni-Zr and the MarM247/Ni-Hf braze joints achieved average creep lives
to failure of approximately 77% and 68%, respectively, of that of the In738 base metal. At an
applied stress level of 276 MPa, the creep life to failure of the joints was 63% and 60% of that
of the base metal for the MarM247/Ni-Zr and MarM247/Ni-Hf joints, respectively. At 228
MPa, these values were 89% (Ni-Zr joints) and 79% (Ni-Hf joints) of the creep life
determined for the In738 base metal. The Larson-Miller plot shown in Figure 112 confirms
that, compared to the MarM247/Ni-Hf joints, the MarM247/Ni-Zr braze joints displayed
superior creep rupture properties at 845C at all applied stress levels evaluated. In spite of the
lower creep rupture properties of the MarM247/Ni-Hf braze joints, the ductility of these joints
was marginally higher than that of the MarM247/Ni-Zr joints (as shown in Table 24).

380
TEST TEMPERATURE: 845°C (1550°F)
360

340

320
Stress (MPa)

300

280

260

240
In738 parent metal
MarM247/Ni-Zr
220
MarM247/Ni-Hf

200
42.5 43.0 43.5 44.0 44.5 45.0 45.5
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 112 – Larson-Miller plot at 845C for creep rupture tests performed at three applied
stress levels: 345 MPa (50 ksi), 276 MPa (45 ksi) and 228 MPa (33 ksi) (where: T is
temperature (F) and t is time (hours)).

The creep rupture properties of the In738 base metal and the MarM247/Ni-Hf and
MarM247/Ni-Zr joints at 900C are shown in Table 25, and displayed graphically in the form
of a Larson-Miller plot in Figure 113. The creep rupture lives of both braze joints were well
below that of the parent metal at all three stress levels tested. At an applied stress of 242
MPa, the creep rupture life of the MarM247/Ni-Zr braze joint was 69% of that of the In738
base metal, whereas that of the MarM247/Ni-Hf braze joint was 64% of the base metal creep
rupture life. At an applied stress of 186 MPa, the MarM247/Ni-Zr joint achieved 71%, and
the MarM247/Ni-Hf joint 65%, of the creep rupture life of the In738 base metal. At 152
MPa, the creep rupture lives of the braze joints were 77% of the life of the base metal for the
Ni-Zr braze, and 66% for the Ni-Hf braze. The Larson-Miller plot shown in Figure 113
confirms that the MarM247/Ni-Zr braze joints outperformed the MarM247/Ni-Hf braze joints
during creep rupture testing at 900C. In spite of the lower creep rupture properties of the
MarM247/Ni-Hf braze joints, the ductility of these joints was superior to that of the
MarM247/Ni-Zr braze joints.

89
 Table 25 - Creep rupture properties at 900oC of the In738 base metal, and LPDB joints produced with
MarM247 superalloy powder and Ni-Hf or Ni-Zr braze alloy. The test samples were produced with a
joint gap of 1.5 mm to simulate a worst-case crack repair scenario.
Failure Hours to % Reduction
Specimen identification % Elongation
location failure in area
APPLIED STRESS 242 MPa (35 ksi)
BM-7T (In738 as-cast base metal) Base metal 30.0 5.5 7.8
BM-8T (In738 as-cast base metal) Base metal 29.5 5.9 7.9
Average for the In738 as-cast base metal 29.8 5.7 7.9
738-10 (In738 brazed with MarM247/Ni-Zr) Joint 22.9 3.7 4.4
738-11 (In738 brazed with MarM247/Ni-Zr) Joint 20.3 4.2 5.6
738-12 (In738 brazed with MarM247/Ni-Zr) Joint 18.2 4.5 6.0
Average for the MarM247/Ni-Zr joint 20.5 4.1 5.3
738-J (In738 brazed with MarM247/Ni-Hf) Joint 17.1 4.5 5.9
738-K (In738 brazed with MarM247/Ni-Hf) Joint 19.4 4.6 6.1
738-L (In738 brazed with MarM247/Ni-Hf) Joint 20.5 4.7 6.2
Average for the MarM247/Ni-Hf joint 19.0 4.6 6.1
APPLIED STRESS 186 MPa (27 ksi)
BM-9T (In738 as-cast base metal) Base metal 100.0 7.0 10.0
BM-10T (In738 as-cast base metal) Base metal 109.1 6.9 9.9
Average for the In738 as-cast base metal 104.6 7.0 9.9
738-13 (In738 brazed with MarM247/Ni-Zr) Joint 75.5 4.2 7.3
738-14 (In738 brazed with MarM247/Ni-Zr) Joint 72.0 4.0 6.8
738-15 (In738 brazed with MarM247/Ni-Zr) Joint 76.2 4.6 7.7
Average for the MarM247/Ni-Zr joint 74.6 4.3 7.3
738-M (In738 brazed with MarM247/Ni-Hf) Joint 70.3 5.0 6.8
738-N (In738 brazed with MarM247/Ni-Hf) Joint 69.2 5.6 8.0
738-O (In738 brazed with MarM247/Ni-Hf) Joint 64.2 5.8 8.2
Average for the MarM247/Ni-Hf joint 67.9 5.5 7.7
APPLIED STRESS 152 MPa (22 ksi)
BM-11T (In738 as-cast base metal) Base metal 200.0 8.5 10.8
BM-12T (In738 as-cast base metal) Base metal 404.2 8.0 10.0
Average for the In738 as-cast base metal 302.1 8.3 10.4
738-16 (In738 brazed with MarM247/Ni-Zr) Joint 242.2 4.8 7.6
738-17 (In738 brazed with MarM247/Ni-Zr) Joint 238.6 4.4 6.4
738-18 (In738 brazed with MarM247/Ni-Zr) Joint 217.4 4.0 6.1
Average for the MarM247/Ni-Zr joint 232.7 4.6 6.7
738-G (In738 brazed with MarM247/Ni-Hf) Joint 205.9 5.0 7.7
738-H (In738 brazed with MarM247/Ni-Hf) Joint 198.4 5.5 7.8
738-I (In738 brazed with MarM247/Ni-Hf) Joint 192.8 5.7 8.0
Average for the MarM247/Ni-Hf joint 199.0 5.4 7.8

The creep rupture properties of the In738 base metal, and the MarM247/Ni-Zr and
MarM247/Ni-Hf joints measured at 980C are displayed in Table 26, and shown graphically
in the form of a Larson-Miller plot in Figure 114. At an applied stress level of 138 MPa, the
creep rupture life of the MarM247/Ni-Zr joint was 79% of the creep rupture life of the base
metal, and that of the MarM247/Ni-Hf joint 64%. These values decreased to 70% and 61% of
the creep rupture life of the base metal, respectively, for the MarM247/Ni-Zr and

90
MarM247/Ni-Hf joints at an applied stress level of 104 MPa. The creep rupture life of the Ni-
Zr joint at an applied stress of 76 MPa was 81% of that of the In738 base metal, whereas that
of the Ni-Hf braze joint achieved 78% of creep rupture life of the base metal. The Larson-
Miller plot shown in Figure 114 indicates that the creep rupture properties of the
MarM247/Ni-Zr braze joints were superior to those of MarM247/Ni-Hf brazed joints. The
MarM247/Ni-Hf joints, however, displayed superior ductility at all stress levels. The average
percentage elongation and percentage reduction in area of the MarM247/Ni-Hf brazed joints
were 73% and 70% of those of the base metal, respectively, whereas the MarM247/Ni-Zr
brazed joints displayed approximately 60% of the ductility of the base metal.

260
TEST TEMPERATURE: 900°C (1650°F)
240

220
Stress (MPa)

200

180

160

In738 parent metal

140 MarM247/Ni-Zr
MarM247/Ni-Hf

120
44.5 45.0 45.5 46.0 46.5 47.0 47.5 48.0 48.5
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 113 –Larson-Miller plot at 900C for creep rupture tests performed at three applied
stress levels: 242 MPa (35 ksi), 186 MPa (27 ksi) and 152 MPa (22 ksi) (where: T is
temperature (F) and t is time (hours)).
150
TEST TEMPERATURE: 980°C (1800°F)
140

130
Stress (MPa)

120

110

100

90
In738 parent metal
80 MarM247/Ni-Zr
MarM247/Ni-Hf

70
47.0 47.5 48.0 48.5 49.0 49.5 50.0 50.5 51.0 51.5
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 114 –Larson-Miller plot at 980C for creep rupture tests performed at three applied
stress levels: 138 MPa (20 ksi), 104 MPa (15 ksi) and 76 MPa (11 ksi) (where: T is
temperature (F) and t is time (hours)).

91
 Table 26 - Creep rupture properties at 980oC of the In738 base metal, and LPDB joints produced with
MarM247 superalloy powder and Ni-Hf or Ni-Zr braze alloy. The test samples were produced with a
joint gap of 1.5 mm to simulate a worst-case crack repair scenario.
Failure Hours to % Reduction
Specimen identification % Elongation
location failure in area
APPLIED STRESS 138 MPa (20 ksi)
BM-13T (In738 as-cast base metal) Base metal 25.0 8.0 11.6
BM-14T (In738 as-cast base metal) Base metal 17.3 8.2 11.9
Average for the In738 as-cast base metal 21.2 8.1 11.8
738-19 (In738 brazed with MarM247/Ni-Zr) Joint 15.9 5.6 7.5
738-20 (In738 brazed with MarM247/Ni-Zr) Joint 16.3 5.2 7.0
738-21 (In738 brazed with MarM247/Ni-Zr) Joint 17.3 4.5 6.9
Average for the MarM247/Ni-Zr joint 16.5 5.1 7.1
738-S (In738 brazed with MarM247/Ni-Hf) Joint 15.3 6.1 8.8
738-T (In738 brazed with MarM247/Ni-Hf) Joint 14.1 6.7 9.3
738-U (In738 brazed with MarM247/Ni-Hf) Joint 11.3 6.9 9.9
Average for the MarM247/Ni-Hf joint 13.6 6.6 9.3
APPLIED STRESS 104 MPa (15 ksi)
BM-15T (In738 as-cast base metal) Base metal 100.0 8.1 11.9
BM-16T (In738 as-cast base metal) Base metal 108.5 8.2 12.1
Average for the In738 as-cast base metal 104.3 8.2 12.0
738-22 (In738 brazed with MarM247/Ni-Zr) Joint 73.3 4.3 7.0
738-23 (In738 brazed with MarM247/Ni-Zr) Joint 74.2 4.0 6.8
738-24 (In738 brazed with MarM247/Ni-Zr) Joint 71.4 4.9 7.4
Average for the MarM247/Ni-Zr joint 73.0 4.4 7.1
738-V (In738 brazed with MarM247/Ni-Hf) Joint 61.5 5.7 7.6
738-W (In738 brazed with MarM247/Ni-Hf) Joint 66.6 5.1 7.0
738-X (In738 brazed with MarM247/Ni-Hf) Joint 62.5 5.6 7.4
Average for the MarM247/Ni-Hf joint 63.5 5.5 7.3
APPLIED STRESS 76 MPa (11 ksi)
BM-17T (In738 as-cast base metal) Base metal 200.0 8.8 12.9
BM-18T (In738 as-cast base metal) Base metal 332.8 8.6 12.4
Average for the In738 as-cast base metal 266.4 8.7 12.7
738-25 (In738 brazed with MarM247/Ni-Zr) Joint 214.7 5.9 8.2
738-26 (In738 brazed with MarM247/Ni-Zr) Joint 218.8 5.0 7.0
738-27 (In738 brazed with MarM247/Ni-Zr) Joint 216.1 5.4 7.3
Average for the MarM247/Ni-Zr joint 216.5 5.4 7.5
738-Y (In738 brazed with MarM247/Ni-Hf) Joint 202.6 6.2 8.9
738-Z (In738 brazed with MarM247/Ni-Hf) Joint 214.8 6.0 8.6
738-ZZ (In738 brazed with MarM247/Ni-Hf) Joint 208.0 6.1 8.7
Average for the MarM247/Ni-Hf joint 208.5 6.1 8.7

The Larson-Miller plots shown in Figures 112 to 114 were constructed individually for each
of the three test temperatures. In order to compare the creep rupture test data at all three test
temperatures, the data shown in Tables 24 to 26 were combined into a single Larson-Miller
plot, shown in Figure 115. Based on this figure, the three materials were ranked in
descending order of creep rupture properties: In738 base metal, the MarM247/Ni-Zr braze
joint, and the MarM247/Ni-Hf joint. It must be emphasized, however, that linear regression

92
methods revealed no statistically relevant differences between the Larson-Miller plots of the
MarM247/Ni-Zr and MarM247/Ni-Hf braze joints.

350
y = -4.8537x + 254.75
300

250 y = -4.8655x + 254.06

Stress (MPa)

200
y = -4.8883x + 254.05

150

100

In738 parent metal

50 MarM247/Ni-Zr
MarM247/Ni-Hf

0
40.0 42.0 44.0 46.0 48.0 50.0 52.0
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 115 – Larson-Miller plot for the In738 base metal, the MarM247/Ni-Zr joint and the
MarM247/Ni-Hf braze joints in the solution annealed condition.

6.4) Conclusions

 A relatively dense, low porosity LPDB joint, with a microstructure consisting of

MarM247 powder particles surrounded by Ni-Hf or Ni-Zr braze alloy, formed during
processing at 1230oC for 12 hours, followed by solution heat treatment at 1230C for 4
hours. The braze microstructure consisted of  dendrites, surrounded by Ni5Zr or Ni7Hf2
intermetallic compound, interspersed with more continuous regions of Ni5Zr or Ni7Hf2
intermetallic compound.

 The extended brazing cycle further reduced the amount of  and Ni5Zr/Ni7Hf2
intermetallic compound, and decreased the thickness of the layers of braze alloy
surrounding the MarM247 powder particles.

 Creep rupture tests performed at 845C, 900C and 980C at various stress levels
revealed that the MarM247/Ni-Zr brazed joints have superior creep rupture properties
compared with the MarM247/Ni-Hf joints. A combined Larson Miller plot, however,
showed little statistically relevant difference between the creep rupture properties of the
two braze filler metals in the solution annealed condition.

 The creep rupture properties described in this chapter were measured in the solution
annealed condition and should improve following a suitable aging heat treatment. In the
solution heat treated condition, the creep rupture properties of the Ni-Hf and Ni-Zr joints
achieved between 65% and 75% of the creep rupture properties of the In738 base metal.
Based on past experience with brazed joints using a Ni-B braze alloy, the aging treatment
is expected to improve the creep rupture properties by between 15% and 20%.

93
 The ductility of the Ni-Hf braze joints was higher than that of the Ni-Zr joints. The joint
ductility is expected to decrease marginally following aging heat treatment.

 The results presented in this chapter were obtained after processing times of 12 hours,
followed by solution heat treatment for an additional 4 hours. Promising joint
microstructures and creep rupture properties were obtained, but isolated microporosity
was observed in the Ni-Zr and Ni-Hf joints. In order to improve the joint quality, Hot
Isostatic Pressing (HIP’ing) was used in an attempt to close the micropores that form
during the braze cycle. Chapter 7 describes the influence of such a HIP cycle on the
microstructure and amount of microporosity in MarM247/Ni-Hf and MarM247/Ni-Zr
joints after brazing for 12 hours at 1230C, followed by a solution annealing treatment.

94
 CHAPTER 7 - EXPERIMENT 5
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL
PROPERTIES OF LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf
AND Ni-Zr BRAZE ALLOYS AFTER BRAZING FOR 12 HOURS, FOLLOWED BY
SOLUTION ANNEALING AND HOT ISOSTATIC PRESSING

7.1) Introduction

The results considered in Chapter 6 revealed that a braze microstructure consisting of 

dendrites, surrounded by Ni5Zr or Ni7Hf2 intermetallic compound, formed during an extended
braze cycle of 12 hours at 1230C, followed by solution annealing at the same temperature for
4 hours. These joints displayed promising creep rupture properties, but contained a
significant volume fraction of microporosity.

In order to determine whether the properties of the braze joints could be improved by
reducing the amount of microporosity, a Hot Isostatic Pressing (HIP) cycle was introduced
after solution annealing. During the HIP cycle, the joint was subjected to pressurized argon
gas at elevated temperature. The high isostatic gas pressure is known to induce plastic
deformation in the material, causing the collapse of any micropores. The success of the HIP
cycle was assessed by examining the joint microstructures and by performing elevated
temperature tensile and creep rupture tests.

7.2) Experimental procedure

The liquid phase diffusion bonding technique was used to join In738 parent material using the
novel eutectic Ni-Hf and Ni-Zr braze alloys. The In738 plate material and the experimental
braze pastes were prepared using the techniques presented in §4.2, and braze samples were
produced using the vacuum braze cycle described in §6.2. After joining the brazed samples
were placed in a production vacuum furnace equipped with quenching facilities, and solution
annealed at 1230C for 4 hours, followed by gas quenching at a rate of 50C per minute.

The solution heat treatment was followed by a HIP cycle at an isostatic pressure of 104 MPa
(15 000 psi), using the following procedure:

1) Ramp up to a temperature of 1080oC at a minimum rate of 11oC per minute.

2) Hold at 1080oC for 4 hours.
3) Quench to room temperature at a rate of 50C per minute.

After exposing the plates to the vacuum LPDB, solution heat treatment and HIP cycles
described above, samples were sectioned, mounted and polished using conventional
metallographic techniques. The polished metallographic samples were then etched with
Marble’s reagent, and characterized using optical and scanning electron microscopy
techniques.

In order to characterize the tensile properties of the braze joints after solution annealing and
HIP’ing, tensile tests were performed at room temperature, 540oC, 870oC and 980oC using the
tensile test procedure described in §4.2. Creep rupture tests were performed (as described in

95
§6.2) using the following combinations of temperature and applied stress: 345 MPa at 845oC;
242 MPa at 900oC; and 138 MPa at 980oC.

7.3) Results and discussion

7.3.1 Microstructural investigation:

The microstructure of the Ni-Hf LPDB joint processed at 1230C for 12 hours, followed by
solution annealing and HIP’ing, is shown in Figures 116 to 119 at various magnifications.
The joint consists of MarM247 powder particles, surrounded by layers of Ni-Hf braze alloy.
Isolated  particles are evident within the Ni-Hf braze, but the braze microstructure between
the MarM247 powder particles appears to be dominated by Ni7Hf2 intermetallic compound.

The microstructure of the LPDB joint produced with MarM247 powder and Ni-Zr braze paste
is shown in Figures 120 to 123 at various magnifications. The original MarM247 powder
particles are evident as the lighter, more equiaxed component, while the darker regions consist
of eutectic Ni-Zr braze alloy. The braze alloy contains isolated  dendrites and Ni5Zr
intermetallic compound.

Isolated micropores were observed in both the Ni-Hf and Ni-Zr joints, suggesting that the HIP
cycle was not effective in eliminating microporosity from the braze joints. A higher
temperature may be required during the HIP treatment to close the pores formed during the
brazing cycle.

Original
MarM247
powder

Isolated
microporosity

Figure 116 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 50X.

96
 Original coarse
MarM247
powder

Original fine
MarM247
powder

Original
Ni-Hf braze
alloy

Figure 117 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 100X.

MarM247
grains

Intergranular
intermetallic
compound

 phase
surrounded by
intermetallic
compound

Figure 118 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 200X.

97
  phase

Ni7Hf2
intermetallic
compound

Figure 119 - Ni-Hf braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 500X.

Original
MarM247
powder

Isolated
pore

Figure 120 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 50X.

98
 Original coarse
MarM247
powder

Original fine
MarM247
powder

Original
Ni-Zr braze
alloy

Figure 121 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 100X.

MarM247
grains

 phase
surrounded by
intermetallic
compound

Intergranular
intermetallic
compound

Figure 122 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 200X.

99
 Ni5Zr
intermetallic
compound

 phase

Figure 123 - Ni-Zr braze dispersed between MarM247 powder particles after brazing at
1230C for 12 hours, followed by solution heat treatment at 1230C for 4 hours and a HIP
cycle at 1080C for 4 hours. Magnification: 500X.

SEM micrographs of the Ni-Hf joint, processed at 1230C for 12 hours, followed by solution
heat treatment and HIP’ing, are shown in Figures 124 and 125. The intergranular phase
highlighted by the arrow in Figure 124 was shown to have a composition of 52.21Ni-
44.72Hf-2.11Co-0.53Cr-0.17Ti-0.26W (wt. %), and was provisionally identified as the Ni7Hf2
intermetallic compound. The micrograph shown in Figure 125 indicates that a fine, cuboidal
gamma prime (’) phase had precipitated within the MarM247  matrix. The cuboidal and
spheroidal ’ precipitates in MarM247 generally form during primary and secondary aging
cycles at 1080oC for 4 hours, followed by soaking at 870oC for 24 hours. The HIP cycle used
in this investigation therefore overlapped with the primary aging treatment for MarM247,
resulting in some ’ precipitation within the  phase.

Figure 126 displays a SEM micrograph of the Ni-Zr joint after processing at 1230C for 12
hours, followed by solution annealing and HIP’ing. The phase identified by the arrow had a
composition of 72.18Ni-24.43Zr-2.18Co-0.64Cr-0.45Al-0.28W (wt. %), and was identified as
Ni5Zr. Although not evident in Figure 126, cuboidal ’ precipitates were observed in the
MarM247 powder particles at higher magnification.

7.3.2 Tensile test results:

The results of tensile tests performed at various temperatures are shown in Tables 27 to 30,
and presented graphically in Figures 127 to 130. The values shown represent the average of
three tests at each temperature.

100
 Ni7Hf2
intermetallic
phase

Figure 124 – SEM micrograph of the Ni-Hf braze, showing the intergranular Ni7Hf2
intermetallic phase and ’ precipitates within the MarM247 particles.

’
precipitates

Figure 125 – Higher magnification SEM micrograph of the Ni-Hf braze, showing cuboidal ’
precipitates within the MarM247 powder particles.

101
 Ni5Zr
intermetallic
compound

Figure 126 – SEM micrograph of the Ni-Zr braze, showing intergranular Ni5Zr intermetallic
compound.

Table 27 – Joint tensile properties measured at 21oC.

Ni-Hf joints Ni-Zr joints
Ni-Hf joints Ni-Zr joints Ni-Hf joints Ni-Zr joints
1230oC for 1230oC for MarM247
Tensile properties 1230oC for 40 1230oC for 40 1230oC for 4 1230oC for 4
12 hours + 12 hours + base metal
minutes minutes hours hours
SHT + HIP SHT + HIP
Tensile strength 323 MPa 557 MPa 344 MPa 564 MPa 713 MPa 736 MPa 960 MPa
Yield strength 239 MPa 505 MPa 302 MPa 511 MPa 644 MPa 622 MPa 800 MPa
Elongation (%) 4.7 3.0 2.7 3.0 2.5 3.7 7.9
Reduction in area (%) 6.3 4.4 3.9 4.5 5.6 8.0 10.0

SHT: Solution heat treatment.

Table 28 – Joint tensile properties measured at 540oC.

Ni-Hf joints Ni-Zr joints
Ni-Hf joints Ni-Zr joints Ni-Hf joints Ni-Zr joints
1230oC for 1230oC for MarM247
Tensile properties 1230oC for 40 1230oC for 40 1230oC for 4 1230oC for 4
12 hours + 12 hours + base metal
minutes minutes hours hours
SHT + HIP SHT + HIP
Tensile strength 328 MPa 619 MPa 406 MPa 625 MPa 740 MPa 999 MPa 1014 MPa
Yield strength 243 MPa 539 MPa 329 MPa 556 MPa 647 MPa 707 MPa 801 MPa
Elongation (%) 3.8 2.7 3.0 2.8 2.7 3.7 7.8
Reduction in area (%) 4.2 3.2 3.8 3.4 5.3 7.9 9.9

SHT: Solution heat treatment.

102
 Table 29 – Joint tensile properties measured at 870oC.
Ni-Hf joints Ni-Zr joints
Ni-Hf joints Ni-Zr joints Ni-Hf joints Ni-Zr joints
1230oC for 1230oC for MarM247
Tensile properties 1230oC for 40 1230oC for 40 1230oC for 4 1230oC for 4
12 hours + 12 hours + base metal
minutes minutes hours hours
SHT + HIP SHT + HIP
Tensile strength 289 MPa 444 MPa 342 MPa 480 MPa 544 MPa 743 MPa 790 MPa
Yield strength 211 MPa 414 MPa 277 MPa 404 MPa 458 MPa 650 MPa 650 MPa
Elongation (%) 2.0 1.0 2.2 2.1 2.4 3.4 5.0
Reduction in area (%) 3.0 2.4 3.2 3.1 4.2 6.0 7.7

SHT: Solution heat treatment.

Table 30 – Joint tensile properties measured at 980oC.

Ni-Hf joints Ni-Zr joints
Ni-Hf joints Ni-Zr joints Ni-Hf joints Ni-Zr joints
1230oC for 1230oC for MarM247
Tensile properties 1230oC for 40 1230oC for 40 1230oC for 4 1230oC for 4
12 hours + 12 hours + base metal
minutes minutes hours hours
SHT + HIP SHT + HIP
Tensile strength 195 MPa 261 MPa 266 MPa 308 MPa 282 MPa 311 MPa 506 MPa
Yield strength 141 MPa 245 MPa 206 MPa 214 MPa 223 MPa 239 MPa 312 MPa
Elongation (%) 1.8 0.8 2.7 3.1 4.3 4.8 4.7
Reduction in area (%) 2.9 1.9 3.1 6.2 5.7 2.9 7.5

SHT: Solution heat treatment.

The tensile strength and yield stress of the MarM247 base metal, and the MarM247/Ni-Hf and
MarM247/Ni-Zr joints in various conditions, are shown in Figures 127 and 128 as a function
of test temperature. It is evident that the HIP cycle improved the tensile and yield strengths of
both joints at all test temperatures, with the strength values approaching those of the base
metal at elevated temperatures. The MarM247/Ni-Zr joints displayed higher tensile and yield
strengths than the Ni-Hf joints at almost all test temperatures. With an increase in test
temperature, the tensile strength of the MarM247/Ni-Zr joints after the HIP cycle varied from
98% of the base metal tensile strength at 540C, to 94% at 870C, and 62% at 980C.
Measured yield strength values ranged from 88% of the base metal yield stress at 540C, to
100% at 870C, and 77% at 980C for the Ni-Zr joints following the HIP cycle.
Corresponding strength measurements for the MarM247/Ni-Hf joints ranged from 73% and
81% of the base metal properties (for tensile strength and yield stress, respectively) at 540C,
to 69% and 71% at 870C, and 56% and 72% at 980C. The increase in strength after the HIP
cycle can most likely be attributed to precipitation hardening linked to the formation of fine ’
precipitates in the MarM247 powder particles during the HIP treatment.

As shown in Figures 129 and 130, the HIP cycle generally resulted in an improvement in the
ductility of the braze joints, particularly at elevated temperatures. The MarM247/Ni-Hf braze
joints displayed approximately 45% of the ductility of the base metal at 540C, increasing to
51% at 870C, and finally to 84% at 980C. The ductility of the MarM247/Ni-Zr joints
ranged from approximately 64% of the base metal ductility at 540C, to 73% at 870C, and
almost 100% at 980C (calculation based on the percentage elongation). The increase in
ductility measured at elevated test temperatures is probably associated with the observed
decrease in strength.

103
 1100
1000
900
800
Stress (MPa)

700
600
500
400
MarM247/Ni-Hf: 40 min @ 1230°C
300 MarM247/Ni-Hf: 4 hrs @ 1230°C
MarM247/Ni-Hf: 12 hrs @ 1230°C + SHT + HIP
200 MarM247/Ni-Zr: 40 min @ 1230°C
MarM247/Ni-Zr: 4 hrs @ 1230°C
100 MarM247/Ni-Zr: 12 hrs @ 1230°C + SHT + HIP
Parent metal
0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 127 – The tensile strength of the MarM247 base metal, and the MarM247/Ni-Hf and
MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at 1230C, and after
brazing for 12 hours at 1230C, followed by solution heat treatment and a HIP cycle.

1100
1000
900
800
Stress (MPa)

700
600
500
400
MarM247/Ni-Hf: 40 min @ 1230°C
300 MarM247/Ni-Hf: 4 hrs @ 1230°C
MarM247/Ni-Hf: 12 hrs @ 1230°C + SHT + HIP
200 MarM247/Ni-Zr: 40 min @ 1230°C
MarM247/Ni-Zr: 4 hrs @ 1230°C
100 MarM247/Ni-Zr: 12 hrs @ 1230°C + SHT + HIP
Parent metal
0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 128 – The yield stress of the MarM247 base metal, and the MarM247/Ni-Hf and
MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at 1230C, and after
brazing for 12 hours at 1230C, followed by solution heat treatment and a HIP cycle.

104
 9

6
% Elongation

3 MarM247/Ni-Hf: 40 min @ 1230°C

MarM247/Ni-Hf: 4 hrs @ 1230°C
2 MarM247/Ni-Hf: 12 hrs @ 1230°C + SHT + HIP
MarM247/Ni-Zr: 40 min @ 1230°C
MarM247/Ni-Zr: 4 hrs @ 1230°C
1 MarM247/Ni-Zr: 12 hrs @ 1230°C + SHT + HIP
Parent metal
0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)
Figure 129 – The percentage elongation of the MarM247 base metal, and the MarM247/Ni-
Hf and MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at 1230C, and
after brazing for 12 hours at 1230C, followed by solution heat treatment and a HIP cycle.

11
10
9
% Reduction in area

8
7
6
5
4 MarM247/Ni-Hf: 40 min @ 1230°C
MarM247/Ni-Hf: 4 hrs @ 1230°C
MarM247/Ni-Hf: 12 hrs @ 1230°C + SHT + HIP
3 MarM247/Ni-Zr: 40 min @ 1230°C
MarM247/Ni-Zr: 4 hrs @ 1230°C
2 MarM247/Ni-Zr: 12 hrs @ 1230°C + SHT + HIP
Parent metal
1
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 130 – The percentage reduction in area of the MarM247 base metal, and the
MarM247/Ni-Hf and MarM247/Ni-Zr braze joints after brazing for 40 minutes and 4 hours at
1230C, and after brazing for 12 hours at 1230C, followed by solution heat treatment and a
HIP cycle.

105
7.3.3 Creep rupture test properties:

The results of creep rupture tests performed at various temperatures and applied stress levels
are shown in Table 31 and presented graphically in the form of a Larson-Miller plot in
Figure 131.

Table 31 – The influence of a HIP cycle on the creep rupture properties of the In738 base
metal and the MarM247/Ni-Hf and MarM247/Ni-Zr joints (1.5 mm joint gap).
Ni-Hf Ni-Hf Ni-Zr Ni-Zr
1230C for 1230C for 1230C for 1230C for In738 base
Test condition Property
12 hours + 12 hours + 12 hours + 12 hours + metal
SHT SHT + HIP SHT SHT + HIP
Creep rupture life 18.6 hours 20.2 hours 21.2 hours 24.4 hours 27.4 hours
345 MPa at 845C % Elongation 3.7 3.8 3.4 3.9 5.5
% Reduction in area 5.6 5.6 4.4 4.7 7.3
Creep rupture life 19.0 hours 19.4 hours 20.5 hours 22.1 hours 29.8 hours
242 MPa at 900C % Elongation 4.6 4.3 4.1 4.2 5.7
% Reduction in area 6.1 5.8 5.3 6.7 7.9
Creep rupture life 13.6 hours 14.8 hours 16.5 hours 17.7 hours 21.2 hours
138 MPa at 980C % Elongation 6.6 7.0 5.1 8.0 8.1
% Reduction in area 9.3 24.7 7.1 23.5 11.8

400

y = -5.9482x + 303.48
350 y = -5.8371x + 299.19
y = -5.8374x + 300.80
y = -5.8901x + 301.17
300 y = -5.9026x + 302.72
Stress (MPa)

250

200
In738 parent metal
MarM247/Ni-Hf (1230°C for 12 hours + SHT)
150 MarM247/Ni-Hf (1230°C for 12 hours + SHT + HIP)
MarM247/Ni-Zr (1230°C for 12 hours + SHT)
MarM247/Ni-Zr (1230°C for 12 hours + SHT + HIP)

100
42 43 44 45 46 47 48 49
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 131 – Larson-Miller plot for the In738 base metal and the MarM247/Ni-Hf and
MarM247/Ni-Zr joints after brazing for 12 hours at 1230C, followed by solution heat
treatment (SHT) and a HIP cycle.

The results shown in Table 31 and in Figure 131 illustrate that the HIP cycle resulted in an
improvement in the creep rupture properties of both the Ni-Hf and Ni-Zr braze joints. The
best creep rupture properties were obtained in the MarM247/Ni-Zr joints after brazing at
1230C for 12 hours, followed by a solution heat treatment at 1230C for 4 hours and a HIP

106
cycle at 1080C for 4 hours. Depending on the test temperature, these joints displayed creep
rupture lives between 74% and 89% of that of the base metal. Since the joints were tested in
the partially age hardened condition, the creep rupture properties are expected to improve
further once the full aging heat treatment is employed.

Although linear regression methods revealed very little statistically relevant difference
between the lines representing the braze joints in Figure 131, the Larson-Miller plot can be
used to rank the braze joints in order of descending creep rupture properties:

 MarM247/Ni-Zr joint brazed at 1230C for 12 hours, followed by solution heat treatment
and a HIP cycle.
 MarM247/Ni-Zr joint brazed at 1230C for 12 hours, followed by solution heat
treatment.
 MarM247/Ni-Hf joint brazed at 1230C for 12 hours, followed by solution heat treatment
and a HIP cycle.
 MarM247/Ni-Hf joint brazed at 1230C for 12 hours, followed by solution heat
treatment.

It is important to note that the improved creep rupture properties displayed by the braze joints
after the HIP cycle cannot be attributed to the closure of micropores in the microstructure.
The higher creep rupture properties are most likely due to the precipitation of fine ’
precipitates within the MarM247 powder particles during the HIP cycle.

Although equivalent base metal creep properties were not achieved, excellent joint ductility
was obtained. For example, during testing at an applied stress level of 138 MPa at 980C,
both the Ni-Zr and Ni-Hf braze joints displayed ductility values in the region of 87% of that
of the base metal. This implies that the presence of the Ni7Hf2 or Ni5Zr intermetallic
compound within the braze alloy did not embrittle the joints to the same extent as the boride
phases observed in Ni-B braze alloys. In boride-containing braze joints, the ductility values
rarely exceed 30% of the ductility of the base metal in the case of wide joint gaps.

7.4) Conclusions

 Brazing for 12 hours at 1230C, followed by solution heat treatment at 1230C for 4
hours and a HIP cycle at 1080C for 4 hours, resulted in joint microstructures consisting
of MarM247 powder particles, surrounded by eutectic Ni-Hf or Ni-Zr braze alloy
containing  phase and Ni7Hf2 or Ni5Zr intermetallic compound. The extended braze
cycle and the solution annealing treatment reduced the fraction of  phase and the amount
of intermetallic compound between the MarM247 particles, resulting in a less continuous
film of braze alloy surrounding the MarM247 grains. The HIP cycle induced the
precipitation of fine ’ precipitates within the MarM247 powder particles.

 The HIP cycle was not effective in eliminating microporosity from the joint. The HIP
temperature may have been too low to close the micropores formed during brazing.

 The HIP cycle significantly improved the tensile and yield strength values of both joints
at all test temperatures, with the strength values approaching those of the base metal at
elevated temperatures. The MarM247/Ni-Zr joints displayed higher tensile and yield
strengths than the Ni-Hf joints at almost all test temperatures. The increase in strength

107
 after the HIP cycle can most likely be attributed to precipitation hardening linked to the
formation of fine ’ precipitates in the MarM247 powder particles during the HIP
treatment. The HIP cycle generally resulted in an improvement in the ductility of the
braze joints, particularly at elevated temperatures. The increase in ductility measured at
elevated test temperatures is probably associated with the observed decrease in strength.

 The HIP treatment resulted in an improvement in the creep rupture properties of both the
Ni-Hf and Ni-Zr braze joints. The best creep rupture properties were obtained in the
MarM247/Ni-Zr joints. Depending on the test temperature, these joints displayed creep
rupture properties between 74% and 89% of those of the base metal. Since the joints
were tested in the partially age hardened condition, the creep rupture properties are
expected to improve further once the full aging heat treatment is employed. Although
equivalent base metal creep properties were not achieved, excellent joint ductility was
obtained, suggesting that the presence of the Ni7Hf2 or Ni5Zr intermetallic compound
within the braze alloy does not embrittle the joints to the same extent as the boride phases
observed in Ni-B braze alloys.

 The majority of cracks which form in turbine components during service initiate under
low cycle fatigue conditions. The low cycle fatigue properties of braze repaired turbine
components therefore ultimately determine whether a braze alloy can be commercialized
successfully. Experiment 6, described in Chapter 8, aimed at determining the low cycle
fatigue (LCF) properties of the novel braze alloys. In order to obtain representative LCF
results, the thermal treatment of the braze joints was optimized based on the results of
earlier experiments. The braze time was increased to 24 hours (to promote diffusion of
the melt point depressants into the parent metal), the HIP temperature was raised to
1200C (to eliminate microporosity) and the joints were subjected to the complete
primary and secondary aging cycles to precipitate ’ in the MarM247 powder particles.

108
 CHAPTER 8 - EXPERIMENT 6
CHARACTERIZATION OF THE LOW CYCLE FATIGUE PROPERTIES OF
LIQUID PHASE DIFFUSION BONDS USING EUTECTIC Ni-Hf AND Ni-Zr BRAZE
ALLOYS AFTER BRAZING FOR 24 HOURS, FOLLOWED BY SOLUTION
ANNEALING, HOT ISOSTATIC PRESSING AT 1200C AND A FULL AGING
TREATMENT

8.1) Introduction

The results of Experiment 5, described in Chapter 7, revealed that brazing at 1230C for 12
hours, followed by solution annealing at 1230C for 4 hours and a HIP treatment at 1080C
for 4 hours, resulted in a microstructure consisting of MarM247 powder particles, interspersed
with Ni-Hf or Ni-Zr braze alloy. The layers of braze alloy surrounding the MarM247 powder
particles appeared less continuous than those observed during Experiment 4, containing a
lower volume fraction of  phase and Ni5Zr or Ni7Hf2 intermetallic compound. The HIP cycle
used in Experiment 5 overlapped with the primary aging cycle for MarM247, resulting in the
precipitation of fine ’ particles within the MarM247 grains. The braze joints displayed
promising tensile and creep rupture properties, despite being only partially age hardened.

The aim of Experiment 6 was to determine the low cycle fatigue (LCF) properties of the braze
joints. Low cycle fatigue testing is essential since cracks in nozzle segments almost always
initiate under LCF conditions, and nozzle segments rarely fail under tensile overload or creep
rupture conditions. Tensile and creep rupture tests provide the approval authority with an
indication of the potential of novel braze alloys, but ultimately it is the low cycle fatigue test
results which determine whether a braze alloy will be commercialized. In order to obtain
representative LCF results, the thermal treatment of the braze joints was optimized based on
the results of earlier experiments. The braze cycle was extended to 24 hours at 1230C to
facilitate diffusion of the melt point depressant (Hf or Zr) into the parent metal, followed by
solution annealing at the same temperature for 4 hours. The braze joints were then subjected
to a 4 hour HIP treatment. Since the HIP cycle used in Experiment 5 was not effective in
eliminating microporosity, the HIP temperature was increased to 1200C. After the HIP
treatment, the specimens were subjected to the full MarM247 aging treatment designed to
precipitate a mixture of coarse and fine ’ particles.

8.2) Experimental procedure

During the course of this investigation, In738 plates with dimensions of approximately 240
mm long, 125 mm wide and 15 mm thick were prepared. A 1.5 mm wide groove was
machined along the length of the plate, as shown in Figures 132 and 133. After machining,
the groove was grit blasted with 220 grit silicon carbide media, and wiped with acetone to
remove any grit blasting residue.

MarM247 paste was prepared from a mixture of coarse and fine MarM247 powder (as
described in §4.2) and used to fill the groove machined in each In738 plate. After drying, a
layer of eutectic Ni-Hf braze alloy, Ni-Zr braze alloy or commercially available BRB braze
alloy in paste form was applied over the MarM247 powder. As described in Chapter 1, BRB
is a B-containing braze of nominal composition Ni-13.5Cr-7.5Co-4Al-2.5B. The plates were
dried, and subjected to the following vacuum braze cycle:

109
1. Ramp up to a temperature of 450oC at a minimum rate of 9oC/minute.
2. Hold at 450oC for 20 minutes to allow the binder to burn off.
3. Ramp up to a temperature of 1150oC at a minimum rate of 9oC/minute.
4. Hold at 1150oC for 20 minutes to allow the samples to stabilize at this temperature.
5. Ramp up to a temperature of 1230oC at a minimum rate of 9oC/minute to melt the Ni-Hf
and Ni-Zr braze alloys and to allow the melt to infiltrate the MarM247 powder.
6. Hold at 1230oC for 24 hours.
7. Furnace cool to room temperature.

Figure 132 – In738 plate with a 1.5 mm wide groove machined in the center.

Figure 133 – Enlarged view of the 1.5 mm wide groove machined in the centre of the In738
plate samples.

The brazed plates were transferred to a production furnace and solution annealed at 1230C
for 4 hours, followed by gas quenching. The solution annealed samples were then subjected
to a revised HIP cycle at a pressure of 104 MPa (15000 psi), as described below:

1. Ramp up to a temperature of 1200C at a minimum rate of 11C/minute.

2. Hold at 1200oC for 4 hours.
3. Gas quench as rapidly as possible using argon gas at 2 bar pressure.

Following the HIP cycle, the plates were subjected to the primary and secondary aging heat
treatments used in industry to precipitate a combination of fine and coarse ’ precipitates in
MarM247. The primary aging heat treatment involves holding the material at 1080C for 4
hours, followed by gas quenching. This aging cycle is followed by the secondary aging heat

110
treatment, during which the samples are held at 870oC for 20 hours, followed by gas
quenching.

After thermal processing, 12.7 mm diameter low cycle fatigue test specimens were machined
from the plates in such a way that each braze joint formed a full butt joint in the centre of the
gauge length. The ends of the samples were threaded and the diameter of the gauge length
reduced to 6.35 mm (as shown in Figure 134).

Figure 134 - LCF specimen with the braze joint in the centre of the gauge length.

Low cycle fatigue tests were performed at a test temperature of 870oC under strain control
conditions. The tests were performed at a constant strain rate of 0.01 s–1 with a symmetrical
triangular waveform. The total strain was varied from 0.3% to 2% and the test frequency
from 1 Hz to 0.33 Hz. A constant A-ratio (defined below) of +1 was maintained during
testing.

A-ratio = (maximum strain – minimum strain) / (maximum strain + minimum strain)

8.3) LCF test results:

The results of the LCF tests for the In738 base metal, and the MarM247/Ni-Hf, MarM247/Ni-
Zr and In738/BRB braze joints, are displayed graphically in Figure 135.

The In738 base metal displayed the best LCF properties, whereas the In738/BRB braze joint
had the lowest LCF properties. BRB is a B-containing braze filler metal with a known
tendency to form brittle borides in wide gap brazed joints. These boride phases are hard and
brittle, and have a detrimental effect on the LCF properties. The Ni-Hf braze joint displayed
slightly higher LCF life than the Ni-Zr braze joint.

The LCF graph shown in Figure 135 can be divided into three regions:

 Below a strain of 0.4%, the BRB braze joint exhibited 70% of the fatigue life of the base
metal, compared to 85% for the MarM247/Ni-Zr braze, and 96% for the MarM247/Ni-Hf
braze joint. Different repair authorities normally have different acceptance and rejection
criteria, but it is safe to say that if a braze joint displays 70% or more of the base metal’s
LCF life, the braze filler can be utilized for production braze repairs of cracks in
stationary nozzle segments. Unfortunately, only a small number of cracks are normally
located in regions of nozzles where the strain is less than 0.4%.

 Between 0.4% and 0.8% strain, the BRB braze joint exhibited only 11% of the base
metal’s life, whereas the MarM247/Ni-Zr braze displayed 70% and the MarM247/Ni-Hf
braze 74% of the base metal’s low cycle fatigue life. The majority of cracks in nozzle

111
 segments occur in this strain range, suggesting that both the Ni-Zr and Ni-Hf braze filler
metals are likely to outperform the B-containing braze alloy in service.

2.0
y = 1.7753x-0.1701
1.8 y = 2.9352x-0.2225
y = 3.1437x-0.2264

1.6
-0.2504
1.4 y = 4.187x
Percentage strain

1.2

1.0

0.8

0.6

0.4 In738 base metal

In738/BRB braze joint
0.2 MarM247/Ni-Zr
MarM247/Ni-Hf
0.0
0 1 2 3 4 5
10 10 10 10 10 10
Cycles to failure

Figure 135 – Low cycle fatigue properties of the In738 base metal, the In738/BRB braze
joint, and the MarM247/Ni-Zr and MarM247/Ni-Hf braze joints.

 Between 0.8% and 2% strain, representing the region of a nozzle segment where one or
two cracks typically form during service, none of the braze joints evaluated displayed
good LCF life. At a strain level of 1.5%, the BRB braze joint failed after only 1 cycle,
whereas the Ni-Zr and Ni-Hf braze joints failed after 30 and 40 cycles, respectively. The
base metal failed after 80 cycles at this strain level, indicating that none of the three braze
filler metals evaluated approach the LCF properties of the base metal at high strains.
Nevertheless, the Ni-Zr and Ni-Hf braze joints outperformed the BRB braze joint by an
order of magnitude, suggesting improved performance and easier acceptance compared
to the commercially available B-containing BRB braze filler metal.

8.4) Conclusions

 Since cracks in nozzle segments generally initiate under low cycle fatigue conditions, a
braze joint should ideally display LCF properties equivalent or better than those of the
base metal. Since the original cracks in a nozzle segment form in the base metal, a
repaired area with inferior LCF properties compared to those of the base metal will result
in earlier crack initiation and faster crack propagation.

 The measured LCF properties of the B-containing BRB braze were vastly inferior to
those of the base metal, with the In738/BRB braze joint displaying a low cycle fatigue
life of only 1 cycle at an applied strain of 1.5%, compared to 80 cycles for the base metal
at the same level of strain. Although the Ni-Zr and Ni-Hf braze joints displayed superior
LCF properties at high strains compared to the BRB joints, the LCF properties of the

112
 novel braze alloys at high strains were only 27% and 50%, respectively, of the base
metal’s LCF properties.

 At intermediate strain levels (0.4% to 0.8% strain), the Ni-Zr and Ni-Hf braze joints
displayed 70% and 74%, respectively, of the LCF properties of the base metal. At these
strain levels, the novel alloys outperformed the BRB braze alloy that displayed only 11%
of the base metal’s LCF properties.

 At low strain rates (0.2% to 0.4% strain), the Ni-Zr and Ni-Hf braze joints displayed LCF
properties equivalent to those of the base metal, whereas the BRB joint exhibited
approximately 70% of the base metal’s LCF properties.

 Although the Ni-Zr or Ni-Hf braze alloys did not display LCF properties equivalent to
those of the base metal at high strain values, the novel braze filler metals outperformed
the commercially available B-containing BRB braze alloy. The new alloys should
therefore have a distinct advantage over the B-containing braze filler metals in the repair
of nozzle segments in industry.

 The good ductility values and promising LCF properties of LPDB joints produced using
the novel braze alloys suggested that it may be possible to further improve the joint
ductility by combining the braze paste with a ductile solid solution strengthened Ni-base
superalloy within the joint, rather than the ’-strengthened MarM247 powder evaluated in
previous chapters. The results of such a change in joint matrix alloy are presented in
Chapter 9.

113
 CHAPTER 9 - EXPERIMENT 7
CHARACTERIZATION OF THE MECHANICAL PROPERTIES OF LIQUID
PHASE DIFFUSION BONDS PRODUCED BY MIXING A SOLID SOLUTION
STRENGTHENED NICKEL-BASE SUPERALLOY POWDER WITH THE
EUTECTIC Ni-Hf AND Ni-Zr BRAZE ALLOYS

9.1) Introduction

During the preceding experiments, described in Chapters 4 to 8, a precipitation hardenable, ’-

strengthened Ni-base superalloy powder, known as MarM247, was mixed with the eutectic
Ni-Zr or Ni-Hf braze alloy within the joint in an attempt to produce high-strength braze joints
with good ductility. As a possible alternative, a ductile solid solution strengthened superalloy
powder, such as Haynes 230, may be substituted for the MarM247 powder in an attempt to
improve the ductility of the joint.

Experiment 9 therefore aimed to compare the mechanical properties (and in particular the
ductility) of MarM247/Ni-Zr joints with Haynes 230/Ni-Zr joints in order to determine
whether a braze joint with a solid solution strengthened matrix (in lieu of a ’-strengthened
matrix) displays a higher level of ductility.

9.2) Experimental procedure

In738 plate material was produced and prepared as described in §3.2. A mixture of coarse
and fine Haynes 230 Ni-base superalloy powder with a nominal composition shown in Table
32 was mixed with braze binder to form a paste. The Haynes 230 paste was applied to the
In738 plate over an area of approximately 2500 mm2 and allowed to dry. A layer of eutectic
Ni-Zr braze alloy in paste form (produced as described in §3.2) was then applied over the dry
Haynes 230 powder. The samples were dried for one hour, and placed in a laboratory vacuum
furnace.

Table 32 – Nominal chemical composition of the Haynes 230 powder used in this
investigation (wt.%, balance Ni).
B C Co Cr Mn Mo Al W La Si Fe
0.015 0.1 5.0 22.0 0.5 2.0 0.3 14.0 0.02 0.4 3.0

The vacuum braze cycle used was as follows:

1) Ramp up to a temperature of 450oC at a minimum rate of 9oC/minute.

2) Hold at 450oC for 20 minutes to allow the binder to burn off.
3) Ramp up to a temperature of 1150oC at a minimum rate of 9oC/minute.
4) Hold at 1150oC for 20 minutes to allow the samples to stabilize at this temperature.
5) Ramp up to a temperature of 1230oC at a minimum rate of 9oC/minute to melt the Ni-Zr
braze alloy and to allow the melt to infiltrate the Haynes 230 powder particles.
6) Hold at 1230oC for 12 hours.
7) Furnace cool to room temperature.

114
 After brazing, the samples were transferred to a production vacuum furnace equipped with
quenching facilities in order to solution anneal the joint. Gas quenching, using 2 bar argon
pressure, was utilized on completion of the solution annealing cycle. The solution heat
treatment used in this experiment was as follows:

1) Ramp up to a temperature of 1230oC at a minimum rate of 14oC per minute.

2) Hold at 1230oC for 4 hours.
3) Quench to room temperature at a minimum rate of 50C per minute.

The solution heat treatment was followed by a HIP cycle at an isostatic pressure of 104 MPa
(15 000 psi), using the following procedure:

1) Ramp up to a temperature of 1080oC at a minimum rate of 11oC per minute.

2) Hold at 1080oC for 4 hours.
3) Quench to room temperature at a rate of 50C per minute.

In order to characterize the tensile properties of the braze joint after solution annealing and
HIP’ing, tensile tests were performed at room temperature, 95oC, 315oC, 540oC, 650oC,
870oC and 980oC using the tensile test procedure described in §4.2. Creep rupture tests were
performed (as described in §6.2) using the combinations of test temperature and applied stress
shown in Table 33. These temperature/stress combinations were selected because the creep
rupture properties of the standard Haynes 230 alloy under these test conditions have been
documented.

Table 33 – Creep rupture test conditions.

Test temperature Applied stress
650C 385 MPa (56 ksi)
760C 185 MPa (27 ksi)
870C 95 MPa (14 ksi)
980C 41 MPa (6.0 ksi)
1040C 24 MPa (3.5 ksi)
1095C 14 MPa (2.1 ksi)

9.3) Results and discussion

9.3.1 Tensile test results:

The results of tensile tests performed at various temperatures are shown in Tables 34 to 40,
and are presented graphically in Figures 136 to 138. The values shown represent the average
of three tests at each temperature.
Table 34 – Joint tensile properties measured at 21oC.
MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength 736 MPa (107 ksi) 737 MPa (107 ksi) 860 MPa (125 ksi) 821 MPa (119 ksi) 960 MPa (139 ksi)
Yield strength 622 MPa (90 ksi) 718 MPa (104 ksi) 395 MPa (57 ksi) 683 MPa (99 ksi) 800 MPa (116 ksi)
Elongation (%) 3.7 1.3 50 4.0 7.9

115
 Table 35 – Joint tensile properties measured at 95oC.
MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength Not tested 739 MPa (107 ksi) 835 MPa (121 ksi) 883 MPa (128 ksi) 1014 MPa (147 ksi)
Yield strength Not tested 726 MPa (105 ksi) 359 MPa (52 ksi) 752 MPa (109 ksi) 801 MPa (116 ksi)
Elongation (%) Not tested 2.4 51 5.0 7.8

Table 36 – Joint tensile properties measured at 315oC.

MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength Not tested 533 MPa (77 ksi) 787 MPa (114 ksi) 856 MPa (124 ksi) 1001 MPa (145 ksi)
Yield strength Not tested 520 MPa (75 ksi) 324 MPa (47 ksi) 725 MPa (105 ksi) 807 MPa (117 ksi)
Elongation (%) Not tested 2.0 52 5.0 7.7

Table 37 – Joint tensile properties measured at 540oC.

MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength 999 MPa (145 ksi) 638 MPa (92 ksi) 705 MPa (103 ksi) 876 MPa (127 ksi) 1014 MPa (147 ksi)
Yield strength 707 MPa (103 ksi) 624 MPa (90 ksi) 275 MPa (40 ksi) 745 MPa (108 ksi) 807 MPa (117 ksi)
Elongation (%) 3.7 1.9 53 5.1 7.8

Table 38 – Joint tensile properties measured at 650oC.

MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength Not tested 511 MPa (74 ksi) 675 MPa (98 ksi) 828 MPa (120 ksi) 1021 MPa (148 ksi)
Yield strength Not tested 497 MPa (72 ksi) 275 MPa (40 ksi) 690 MPa (100 ksi) 814 MPa (118 ksi)
Elongation (%) Not tested 1.8 55 5.5 7.2

Table 39 – Joint tensile properties measured at 870oC.

MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength 743 MPa (108 ksi) 555 MPa (80 ksi) 435 MPa (63 ksi) 690 MPa (100 ksi) 790 MPa (115 ksi)
Yield strength 650 MPa (94 ksi) 541 MPa (78 ksi) 255 MPa (37 ksi) 137 MPa (76 ksi) 650 MPa (94 ksi)
Elongation (%) 3.4 1.5 65 7.5 5.0

Table 40 – Joint tensile properties measured at 980oC.

MarM247/Ni-Zr joint Haynes 230/Ni-Zr joint
Tensile
1230oC for 12 hours + 1230oC for 12 hours + Haynes 230 In738 MarM247
properties
SHT + HIP SHT + HIP
Tensile strength 311 MPa (45 ksi) 267 MPa (39 ksi) 240 MPa (35 ksi) 414 MPa (60 ksi) 506 MPa (73 ksi)
Yield strength 240 MPa (35 ksi) 189 MPa (27 ksi) 145 MPa (21 ksi) 283 MPa (41 ksi) 312 MPa (45 ksi)
Elongation (%) 4.8 1.8 83 8.0 4.7

116
 1100

1000

900

800
Stress (MPa)

700

600

500
Haynes 230
400 In738
MarM247
300 MarM247/Ni-Zr
Haynes 230/Ni-Zr

200
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 136 – Tensile strength of Haynes 230, In738, MarM247, the MarM247/Ni-Zr braze
joint and the Haynes 230/Ni-Zr braze joint as a function of test temperature.

1000

900

800

700
Stress (MPa)

600

500

400

300
Haynes 230
200 In738
MarM247
MarM247/Ni-Zr
100
Haynes 230/Ni-Zr

0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 137 – Yield stress of Haynes 230, In738, MarM247, the MarM247/Ni-Zr braze joint
and the Haynes 230/Ni-Zr braze joint as a function of test temperature.

As shown Table 34 and Figures 136 and 137, the room temperature tensile and yield
strengths of the Haynes 230/Ni-Zr joint compared well with those of the MarM247/Ni-Zr
joint. The high strength of the Haynes 230/Ni-Zr joint relative to that of the MarM247/Ni-Zr
joint is attributed to it being in its highest strength condition (i.e. solution heat treated),
whereas the MarM247/Ni-Zr joint was in the solution heat treated and partially aged
condition. The MarM247/Ni-Zr joint only achieves its highest mechanical strength after
secondary aging at 870oC.

117
 12
In738
11
MarM247
10 MarM247/Ni-Zr
Haynes 230/Ni-Zr
9
8
% Elongation

7
6
5
4
3
2
1
0
0 100 200 300 400 500 600 700 800 900 1000
Test temperature (°C)

Figure 138 – Percentage elongation of In738, MarM247, the MarM247/Ni-Zr braze joint and
the Haynes 230/Ni-Zr braze joint as a function of test temperature.

As shown in Table 34 and Figure 138, the room temperature ductility (expressed as
percentage elongation) of the Haynes 230/Ni-Zr joint was surprisingly low, given that Haynes
230 material normally exhibits up to 50% elongation at room temperature. In contrast, the
percentage elongation of the MarM247/Ni-Zr joint after solution heat treatment and HIP’ing
(representing partial aging) achieved approximately 63% of the ductility of MarM247 and
75% of the measured ductility of the In738 base metal. Although not examined in more
detail, it is postulated that the low ductility of the Haynes 230/Ni-Zr joint may be due to the
braze alloy disrupting the continuity of the matrix, limiting the ductility of the joint. It is also
possible that the majority of the plastic deformation in the joint was concentrated in the more
ductile Haynes 230 powder particles, resulting in rapid work hardening of the austenitic
matrix and premature failure during testing.

With an increase in the test temperature, the tensile and yield strengths of the MarM247/Ni-Zr
joint increased, reaching peak strength values at a test temperature of approximately 540C.
At peak strength, the tensile and yield strength values of the MarM247/Ni-Zr joint attained
98% and 88%, respectively, of the tensile and yield strengths of MarM247 at the same
temperature, and exceeded those of the In738 base metal. With a further increase in test
temperature above 540C, the tensile and yield strengths of the MarM247/Ni-Zr joint
gradually decreased, following the same general trend as that observed for the MarM247 and
In738 samples. The tensile and yield strengths of the Haynes 230/Ni-Zr joint gradually
decreased with an increase in test temperature, following the same general trend as that of the
Haynes 230 samples.

As shown in Figure 138, the ductility values (expressed as percentage elongation) of both
joints were well below the ductility of the MarM247 and In738 samples at all test
temperatures. The ductility of the MarM247/Ni-Zr joint increased slightly with an increase in
test temperature, probably associated with the observed decrease in strength at higher
temperatures. The measured ductility of the Haynes 230/Ni-Zr joint remained very low and
did not vary to any significant extent with an increase in test temperature. It is postulated that

118
most of the plastic deformation during tensile testing was concentrated in the more ductile
Haynes 230 powder particles, resulting in rapid work hardening and premature failure.

9.3.2 Creep rupture test results:

Creep rupture test results measured at various temperatures and stress levels are shown in
Table 41 and presented graphically in Figure 139 in the form of a Larson-Miller plot. Also
included are the creep rupture properties of In738 base metal and MarM247/Ni-Zr joints
(from Chapter 7). Each value shown represents the average of three tests.

Table 41 – Creep rupture properties of In738, Haynes 230 and the MarM247/Ni-Zr and
Haynes 230/Ni-Zr joints (1.5 mm joint gap).
MarM247/Ni- Haynes 230/Ni-
Test condition Property Zr braze Zr braze In738 Haynes 230
SHT + HIP SHT + HIP
Creep rupture life 110 hours
385 MPa at 650C % Elongation Not tested 1.7 Not tested 100 hours
% Reduction in area 0.7
Creep rupture life 212.8 hours
185 MPa at 760C % Elongation Not tested 0.3 Not tested 100 hours
% Reduction in area 1.7
Creep rupture life 24.4 hours 27.4 hours
345 MPa at 845C % Elongation 3.9 Not tested 5.5 Not tested
% Reduction in area 4.7 7.3
Creep rupture life 114.6 hours
95 MPa at 870C % Elongation Not tested 0.6 Not tested 100 hours
% Reduction in area 0.9
Creep rupture life 22.1 hours 29.8 hours
242 MPa at 900C % Elongation 4.2 Not tested 5.7 Not tested
% Reduction in area 6.7 7.9
Creep rupture life 252.2 hours
41 MPa at 980C % Elongation Not tested 1.7 Not tested 100 hours
% Reduction in area 11.1
Creep rupture life 17.7 hours 21.2 hours
138 MPa at 980C % Elongation 8.0 Not tested 8.1 Not tested
% Reduction in area 23.5 11.8
Creep rupture life 139.3 hours
24 MPa at 1040C % Elongation Not tested 4.1 Not tested 100 hours
% Reduction in area 6.0
Creep rupture life 126.4 hours
14 MPa at 1095C % Elongation Not tested 1.8 Not tested 100 hours
% Reduction in area 18.5

SHT: Solution heat treatment at 1230C for 4 hours.

HIP: Hot isostatic pressing at 1080C for 4 hours.

It is evident from Table 41 and Figure 139 that the creep rupture properties of the Haynes
230/Ni-Zr joint were equivalent to that of Haynes 230 parent material, but inferior to those of
the MarM247/Ni-Zr joint, and well below the creep rupture properties of the In738 base
metal. The measured creep rupture life of the MarM247/Ni-Zr joint varied between 74% and
89% of that of the In738 base metal. It must be emphasized that the creep rupture properties
of the MarM247/Ni-Zr joints are expected to improve on application of the full aging cycle
(the results given in Table 41 and Figure 139 are relevant to the partially aged condition).

119
 400

350

Stress (MPa) 300

250

200

150

100
In738
Haynes 230/Ni-Zr braze
50 Haynes 230
MarM247/Ni-Zr braze
0
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 139 – Creep rupture properties of In738, Haynes 230, the MarM247/Ni-Zr braze joint
and the Haynes 230/Ni-Zr braze joint.

The creep rupture test results shown in Table 41 confirm the observation (described in §9.3.1)
of poor ductility in the case of the Haynes 230/Ni-Zr joints. Both the percentage elongation
and the percentage reduction in area values were very low, especially in the high stress/low
temperature creep regime. At higher test temperatures (such as 1040C and 1095), where the
applied stress levels were lower, the ductility improved considerably. The MarM247/Ni-Zr
joints displayed very good ductility, achieving up to 88% of the percentage elongation and
reduction in area of the base metal (at a temperature of 980C and an applied stress level of
138 MPa). Although equivalent base metal ductility was not obtained, the MarM247/Ni-Zr
joints outperformed the commercially available B-containing brazes, for which elongation and
reduction in area values below 30% are usually obtained in 1.5 mm wide joints. This suggests
that the Ni5Zr intermetallic phase within the Ni-Zr joints does not embrittle the braze joints to
the same extent as the boride phases commonly observed when B-containing braze alloys are
used to fill in wide gaps.

9.4) Conclusions

 Although the solid solution strengthened Haynes 230 Ni-base superalloy displays
excellent ductility over a temperature range from room temperature to 980C (well in
excess of that documented for cast, ’-strengthened alloys), Haynes 230/Ni-Zr braze
joints displayed very poor ductility during testing. This may be due to the property
mismatch between the Haynes 230 powder particles within the joint and the In738 parent
material. During tensile testing, most of the plastic deformation is expected to take place
in the more ductile Haynes 230 material, resulting in rapid work hardening of the
austenitic matrix, and premature failure.

 The room temperature tensile and yield strengths of the Haynes 230/Ni-Zr joint
compared well with those of the MarM247/Ni-Zr joint. With an increase in test

120
 temperature, the tensile and yield strengths of the MarM247/Ni-Zr joint increased,
reaching peak strength values at a test temperature of approximately 540C. At peak
strength, the tensile and yield strength values of the MarM247/Ni-Zr joint attained 98%
and 88%, respectively, of the tensile and yield strengths of MarM247 at the same
temperature, and exceeded those of the In738 base metal. With a further increase in test
temperature above 540C, the tensile and yield strengths of the MarM247/Ni-Zr joint
gradually decreased, following the same general trend as that observed for the MarM247
and In738 samples. The tensile and yield strengths of the Haynes 230/Ni-Zr joint
gradually decreased with an increase in test temperature, following the same trend as
Haynes 230.

 The MarM247/Ni-Zr joints displayed excellent creep rupture properties, achieving

between 74% and 89% of the creep rupture life of the In738 parent material. The creep
rupture properties of the MarM247/Ni-Zr joints are expected to improve further on
application of the full aging cycle. The MarM247/Ni-Zr joints also displayed very good
ductility, achieving up to 88% of the ductility of the base metal. Although equivalent
base metal ductility was not obtained, the MarM247/Ni-Zr joints outperformed the
commercially available B-containing brazes.

 The Haynes 230/Ni-Zr joints displayed creep rupture properties equivalent to those of the
Haynes 230 base metal, but inferior to those of the In738 parent material and the
MarM247/Ni-Zr braze joints. Since the creep rupture properties of solid solution
strengthened alloys, such as Haynes 230, are inferior to those of ’-strengthened alloys,
the poor creep rupture properties of the Haynes 230/Ni-Zr joints were not unexpected.
The ductility values of the Haynes 230/Ni-Zr joint were, however, disappointingly low,
especially in the high stress/low temperature creep regime. These low ductility values
are attributed to the property mismatch between the braze joints and the parent material.

 Since the use of a more ductile matrix alloy in the braze joint did not result in any
significant improvement in mechanical properties, the next chapter describes preliminary
experiments aimed at improving the performance of the braze joints through the addition
of chromium and boron to hypo-eutectic Ni-Hf braze alloys.

121
 CHAPTER 10 - EXPERIMENT 8
CHARACTERIZATION OF THE MICROSTRUCTURE AND PROPERTIES OF
HYPO-EUTECTIC Ni-Cr-Hf AND Ni-Cr-Hf-B ALLOYS

10.1) Introduction

The experiments described in earlier chapters focused on simple binary eutectic Ni-Zr or Ni-
Hf braze alloys. Since the joint ductility could not be improved by using a more ductile solid
solution strengthened matrix alloy within the joint (as described in Chapter 9), an attempt was
made to improve the performance of the novel braze filler metals by producing a series of
ternary and quaternary hypo-eutectic Ni-Hf alloys containing chromium and boron.

The aim of this experiment was therefore to develop ternary and quaternary hypo-eutectic
braze alloys containing chromium and boron, and to characterize the microstructures of these
alloys using microprobe analysis techniques.

10.2) Experimental procedure

Four off-eutectic experimental alloys were produced by melting various metal powders in a
crucible at a temperature of 1500oC. The alloy compositions (in wt.%) are given below:

 Ni-13Cr-15Hf
 Ni-13Cr-20Hf
 Ni-13Cr-25Hf
 Ni-13Cr-25Hf-1B

After cooling, the alloy samples were mounted in resin and prepared using conventional
metallographic techniques before being subjected to microstructural analysis using
microprobe techniques. The mounted samples were carbon coated using a sputtering
technique and electrically grounded with copper tape in order to minimize charging of the
specimens by the microprobe electron beam. Calibration was performed using polished pure
element standards.

Analysis conditions were established for either a 15 or 20 keV focused electron beam,
adjusted to provide 35 nA of specimen current on pure nickel. Desired locations for analysis
were selected by generating a backscattered electron (BSE) image before placing the
instrument in "spot mode", during which the focused beam was directed to locations within
the area shown in the BSE image. The EDS (energy dispersive x-ray spectroscopy) peak
identifications were verified using a wavelength dispersive x-ray spectrometer (WDS), tuned
to the elements of interest. X-ray intensity data were collected in triplicate at each phase
location of interest. Measured x-ray counts were then converted to elemental weight
percentages using commercially available algorithms.

A preliminary characterization of the mechanical properties of the alloys was performed by

measuring the hardness of each specimen using a micro-Vickers hardness tester.

122
10.3) Results and discussion

10.3.1 Ternary Ni-Cr-Hf alloys:

The microstructure of the Ni-13Cr-15Hf alloy after cooling is shown in Figure 140. The
coarse dendritic phase (dark gray areas) was shown to have a composition of Ni-19.2Cr-1.6Hf
(wt.%), identifying it as the Ni-rich  phase. The white areas consisted of 53.7Ni-1.2Cr-
45.1Hf (wt.%) and are most likely islands of Ni7Hf2 intermetallic compound within the
eutectic component. The fine eutectic component had a composition of Ni-9.0Cr-25.8Hf.
This microstructure suggests that the alloy was hypo-eutectic, forming a primary Ni-rich 
phase on solidification. On reaching the eutectic temperature, the remaining liquid
transformed to the eutectic mixture of  and the Ni7Hf2 intermetallic compound. The majority
of the chromium present in this ternary alloy apparently partitioned to the  phase on
solidification.

Eutectic
component

Figure 140 - BSE image of the Ni-13Cr-15 Hf alloy showing primary  dendrites, and a fine
eutectic component consisting of  phase and the Ni7Hf2 intermetallic compound.

The microstructure of the Ni-13Cr-20Hf alloy, shown in Figure 141, is very similar to that of
the alloy containing 15% Hf. The dark-etching phase was identified as Ni-rich  dendrites
(consisting of Ni-22.4Cr-1.4Hf), the white phase appeared to be the Ni7Hf2 intermetallic
compound (consisting of 53.8Ni-1.3Cr-44.9Hf), whereas the eutectic component had a
composition of Ni-10.5Cr-25.2Hf (all compositions given as wt.%). This alloy also appeared
to be hypo-eutectic, although at 20% Hf it is located closer to the eutectic point, resulting in
less primary  phase.

Figure 142 shows the microstructure of the Ni-13Cr-25Hf alloy. The white areas were
identified as Ni7Hf2 intermetallic compound, with a nominal composition of 53.5Ni-0.8Cr-
45.7Hf. The fine eutectic component consisted of Ni7Hf2 intermetallic compound and the Ni-
rich  phase (dark gray regions). The eutectic component had a nominal composition of Ni-

123
15.1Cr-21.8Hf (all compositions given as wt.%). This microstructure suggests that the alloy
was hypereutectic, forming a primary Ni7Hf2 intermetallic phase on solidification. On
reaching the eutectic temperature, the remaining liquid transformed to the eutectic mixture of
Ni-rich  phase and Ni7Hf2 intermetallic compound. The majority of the chromium present in
the alloy apparently partitioned to the  phase on solidification.

Eutectic
component

Figure 141 - BSE image of the Ni-13Cr-20 Hf alloy showing primary  dendrites, and a fine
eutectic component consisting of  phase and the Ni7Hf2 intermetallic compound.

Eutectic
component

Figure 142 - BSE image of the Ni-13Cr-25 Hf alloy showing primary Ni7Hf2, and a fine
eutectic component consisting of  phase and the Ni7Hf2 intermetallic compound.

124
Table 42 summarizes the measured chemical compositions of the eutectic components
formed in each alloy on solidification. The binary Ni-Hf phase diagram shown in Figure 49
indicates that the eutectic point in the binary Ni-Hf system is located at 30.5 wt.% Hf and a
temperature of 1190C. The addition of chromium therefore appears to shift the eutectic
composition to lower Hf contents. The measured solidus and liquidus temperatures of the
three ternary alloys are shown in Table 43. An increase in Hf content decreased the liquidus
and increased in the solidus temperatures of the alloys. As expected, the solidus and liquidus
temperatures of the ternary alloys were well above the Ni-Hf eutectic temperature of 1190C.

Table 42 – Chemical composition (wt.%) of the eutectic component observed in the ternary
Ni-Cr-Hf alloys.
Alloy Eutectic formed Ni Cr Hf
Ni-13Cr-15Hf Ni-rich  + Ni7Hf2 65.2 9.0 25.8
Ni-13Cr-20Hf Ni-rich  + Ni7Hf2 64.3 10.5 25.2
Ni-13Cr-25Hf Ni-rich  + Ni7Hf2 63.1 15.1 21.8

Table 43 – Measured liquidus and solidus temperatures of the ternary Ni-Cr-Hf alloys.
Solidus Liquidus
Alloy
temperature temperature
Ni-13Cr-15Hf 1205C 1337C
Ni-13Cr-20Hf 1220C 1304C
Ni-13Cr-25Hf 1231C 1291C

10.3.2 Quaternary Ni-Cr-Hf-B alloy:

The microstructure of the quaternary Ni-13Cr-25Hf-1B alloy is shown in Figure 143. The
addition of 1% B to the alloy resulted in the formation of a significant volume fraction of
brittle borides (black phase). These boride particles were identified as chromium borides,
with a composition of Ni-88.3Cr-0.0Hf-13.6B. The white areas were identified as the Ni7Hf2
intermetallic phase (with a composition of Ni-0.7Cr-46.1Hf-0.3B), and the dark gray areas
were shown to be the Ni-rich  phase (containing Ni-25.9Cr-1.2Hf-0.2B). Analysis of the
eutectic regions yielded compositions of Ni-11.0Cr-20.3Hf-2.0B (coarser eutectic component)
and Ni-10.7Cr-16.9Hf-2.7B (fine eutectic regions). (All compositions given in wt.%).

The sequence of phase transformations on solidification of this alloy was probably:

L1  Ni7Hf2 + L2
L2  Ni-rich  + L
L (eutectic)  Ni-rich  + Ni7Hf2

10.3.3 Microhardness measurements:

In order to characterize the hardness of the phases within the braze alloys, Vickers
microhardness tests were performed. The measured hardness values are given in Table 44.
These results indicate that the chromium boride phase was significantly harder than the Hf-
bearing intermetallic phase. This is consistent with the high ductility values obtained during
earlier testing of Ni-Hf joints, compared with those typically found in B-containing brazes.

125
 The high hardness of the eutectic component observed in the B-containing alloy suggests that
the chromium boride phase participated in the eutectic reaction, possibly resulting in the
formation of a ternary Cr boride//Ni7Hf2 eutectic component.

Eutectic
component

Figure 143 - BSE image of the Ni-13Cr-25 Hf-1B alloy showing the Ni7Hf2 phase (white),
Ni-rich  (dark gray), chromium borides (black) and a possible ternary eutectic component
consisting of  phase, Cr boride, and the Ni7Hf2 intermetallic compound.

Table 44 – Measured microhardness values of various phases observed in the ternary and
quaternary alloys.
Microhardness, Hv
 phase Eutectic constituent Boride phase Other
Alloy Standard Standard Standard Standard
Average Average Average Average
deviation deviation deviation deviation
Ni-13Cr-15Hf 242 19 294 56 N/A N/A N/A N/A
Ni-13Cr-20Hf 309 19 322 72 N/A N/A N/A N/A
Ni-13Cr-25Hf 398 32 324 30 N/A N/A N/A N/A
Ni-13Cr-25Hf-1B 326 93 746 108 1150 240 350 43

Table 45 summarizes the microprobe results, and displays the chemical compositions of the
microstructural constituents observed within the ternary and quaternary alloys.

10.4) Conclusions

 The ternary Ni-13Cr-15Hf and Ni-13Cr-20Hf alloys appeared to be hypo-eutectic, and

consisted of primary Ni-rich  dendrites (with approximately 20% Cr in solution), and a
eutectic component consisting of  phase and Ni7Hf2 intermetallic compound. Increasing
the Hf content from 15% to 20% reduced the amount of primary  phase. The Ni-13Cr-
25Hf alloy appeared to be hypereutectic, with a primary Ni7Hf2 phase and a eutectic

126
 component consisting of  and Ni7Hf2. Since the eutectic point in binary Ni-Hf alloys is
located at approximately 30.5% Hf, the addition of chromium apparently shifted the
eutectic composition to lower Hf contents.

Table 45 – The chemical compositions of various phases within the ternary and quaternary
alloys (all compositions given as wt.%).
Weight %
Alloy Microstructural constituent Transformation
Ni Cr Hf B
Ni-rich  Primary 79 19.2 1.6 -
Ni7Hf2 Eutectic 55.5 1.2 45.1 -
Ni-13Cr-15Hf
 Ni Hf
Eutectic component  7 2 Eutectic 65.2 9 25.8 -
 
Ni-rich  Primary 75.9 22.4 1.4 -
Ni7Hf2 Eutectic 55.9 1.3 44.9 -
Ni-13Cr-20Hf
 Ni Hf
Eutectic component  7 2 Eutectic 64.3 10.5 25.2 -
 
Ni7Hf2 Primary 54.9 0.8 45.7 -
Ni-13Cr-25Hf  Ni Hf
Eutectic component  7 2 Eutectic 63.1 15.1 21.8 -
 
Ni7Hf2 Primary 54.9 0.7 46.1 0.3
Ni-rich  Primary 73.2 25.9 1.2 0.2
Cr boride Primary 2.6 88.3 0.0 13.6
Ni-13Cr-25Hf-1B
 Ni 7 Hf 2
 65.9 11.0 20.3 2.0
Eutectic component   Eutectic
Cr boride 70.2 10.7 16.9 2.7


 An increase in the Hf content of the experimental alloys lowered the liquidus

temperature, but increased the solidus temperature. Since single-crystal Ni-base
superalloys can be processed at temperatures up to 1310oC, only the Ni-13Cr-25Hf alloy
holds any potential for production brazing. The liquidus temperatures of the Ni-13Cr-
20Hf and Ni-13Cr-15Hf alloys exceeded 1300C.

 The quaternary Ni-13Cr-25Hf-1B alloy consisted of a coarse Ni7Hf2 phase, islands of Ni-
rich  (with Cr in solution), a ternary eutectic component (Ni7Hf2,  and Cr boride) and a
significant volume fraction of chromium boride particles. The boride particles were
shown to be considerably harder than the Ni7Hf2 intermetallic phase. This difference in
hardness may account for the high ductility and good LCF properties of the Ni-Hf braze
joints, compared with those of braze alloys containing B as a melt point depressant.

127
 CHAPTER 11 - EXPERIMENT 9
MICROPROBE ANALYSIS OF INTERMETALLIC PHASES

11.1) Introduction

Previous chapters revealed a discrepancy between the equilibrium intermetallic phase

predicted by the Ni-Hf binary phase diagram (i.e. Ni5Hf) for the novel braze alloys, and the
intermetallic phase provisionally identified in the Ni-Hf braze joints on the basis of SEM-
EDS and preliminary microprobe analyses (Ni7Hf2). The SEM-EDS technique is, at best,
semi-quantitative, and an attempt was therefore made to determine the chemical make-up of
the intermetallic phases in modified Ni-Hf and Ni-Zr braze joints using more detailed electron
microprobe analysis (EMPA) techniques. This experiment also aimed at assessing the effect
of chromium additions to the Ni-Hf and Ni-Zr braze alloys on the microstructure of the wide
gap braze joints.

11.2) Experimental procedure

MarM247/Ni-Hf and MarM247/Ni-Zr braze joints in In738 base metal were produced using
two experimental ternary braze alloys with compositions of Ni-7Cr-31Hf and Ni-7Cr-13Zr
(wt.%). Brazing was performed using the procedure described in §4.2. The joints were
examined, either in the as-brazed condition (without applying an extended diffusion cycle), or
after a diffusion treatment to facilitate the diffusion of the melt point depressant into the
parent metal.

The brazed samples were mounted in resin and prepared using conventional metallographic
techniques before being subjected to microstructural examination using microprobe analysis
techniques. The mounted samples were carbon coated using a sputtering technique and
electrically grounded with copper tape to minimize charging of the specimens by the
microprobe electron beam. Calibration was performed using polished pure element standards.

Analysis conditions were established for either a 15 or 20 keV focused electron beam,
adjusted to provide 35 nA of specimen current on pure nickel. Desired locations for analysis
were selected by generating a backscatter electron (BSE) image before placing the instrument
in "spot mode", during which the focused beam was directed to selected locations within the
area shown in the BSE image. The BSE images were annotated to highlight the features
subjected to elemental characterization by energy dispersive x-ray spectroscopy (EDS). The
EDS peak identifications were verified using a wavelength dispersive x-ray spectrometer
(WDS), tuned to the elements of interest. X-ray intensity data were collected in triplicate at
each selected location. Measured x-ray counts were then converted to elemental weight
percentages using commercially available algorithms.

11.3) Results and discussion

11.3.1 The MarM247/Ni-Hf braze joint produced using Ni-7Cr-31Hf braze filler (as-brazed
condition):

Secondary and backscatter electron images of the MarM247/Ni-Hf joint produced using the
experimental Ni-7Cr-31Hf braze filler metal are shown in Figure 144 in the as-brazed

128
 condition. The arrow indicates the location and direction of the 5 m stepped EPMA line
scan. The results of the EMPA scan are shown in Table 46.

(a) (b)

Figure 144 – Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-
31Hf joint in the as-brazed condition.

Table 46 – The results of the EMPA scan of the MarM247/Ni-7Cr-31Hf braze joint in the as-
brazed condition. (All compositions given in wt.%).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
74.213 64.158 11.067 4.40 1.71 68.70 9.30 0.30 0.58 3.40 1.04 7.15 2.65 0.01
74.218 64.158 11.067 4.48 1.72 69.15 9.18 0.28 0.58 3.16 1.03 7.14 2.54 0.00
74.223 64.158 11.067 4.48 1.66 68.69 9.25 0.30 0.58 3.43 1.09 7.11 2.59 0.00
74.228 64.158 11.067 4.45 1.64 68.40 9.13 0.30 0.60 3.42 1.03 7.12 2.83 0.00
74.233 64.158 11.066 4.54 1.69 68.41 8.99 0.34 0.57 3.33 1.06 7.08 2.86 0.00
74.238 64.158 11.066 4.55 1.67 68.13 9.07 0.31 0.58 3.23 1.00 7.03 2.66 0.00
74.243 64.158 11.066 4.74 1.64 68.00 8.72 0.30 0.60 3.05 1.00 6.96 2.78 0.01
74.248 64.158 11.068 0.40 27.68 56.16 0.72 9.57 0.09 3.08 0.13 2.98 0.00 0.00
74.253 64.158 11.068 0.63 26.18 56.43 1.41 8.81 0.14 2.61 0.23 3.28 0.00 0.00
74.258 64.158 11.068 4.68 1.69 68.03 8.77 0.31 0.59 3.09 1.07 7.02 2.55 0.00
74.263 64.158 11.068 4.52 1.66 68.11 8.98 0.31 0.61 3.13 1.06 7.16 2.76 0.00
74.268 64.158 11.068 4.50 1.63 67.83 9.03 0.31 0.61 2.97 1.07 7.17 2.75 0.00
74.273 64.158 11.068 4.22 13.37 58.87 7.35 1.30 0.48 5.22 0.86 5.69 1.15 0.01
74.278 64.158 11.066 4.56 1.64 67.57 8.92 0.31 0.62 2.87 1.04 7.17 2.85 0.00
74.283 64.158 11.068 4.64 1.64 68.17 8.84 0.32 0.61 3.00 1.04 7.16 2.66 0.00
74.288 64.158 11.068 4.59 1.60 67.68 8.87 0.32 0.65 2.92 1.01 7.17 2.66 0.00
74.293 64.158 11.066 4.54 1.76 67.77 8.66 0.32 0.60 3.00 1.02 7.12 2.67 0.00
74.298 64.158 11.068 0.88 24.80 57.35 1.81 8.08 0.17 2.92 0.26 3.48 0.00 0.00
74.303 64.158 11.068 0.29 28.34 55.23 0.59 9.36 0.08 2.78 0.16 2.96 0.00 0.00
74.308 64.158 11.068 0.29 28.36 55.77 0.54 9.44 0.06 2.91 0.17 2.96 0.00 0.01
74.313 64.158 11.068 0.29 28.45 55.56 0.57 9.36 0.07 2.76 0.14 2.90 0.00 0.00
74.318 64.158 11.066 3.04 11.30 63.34 5.83 3.30 0.40 2.91 0.76 5.58 1.32 0.01
74.323 64.158 11.068 4.54 1.70 68.45 8.56 0.35 0.56 2.92 1.03 7.17 2.26 0.00
74.328 64.158 11.068 4.53 1.59 68.39 8.69 0.33 0.60 2.81 1.04 7.23 2.55 0.01
74.333 64.158 11.068 4.55 1.63 68.18 8.71 0.29 0.59 2.69 1.08 7.19 2.64 0.01
74.338 64.158 11.067 4.61 1.74 68.19 8.58 0.31 0.56 3.08 1.04 7.05 2.38 0.00
74.343 64.158 11.067 4.47 1.61 68.41 8.86 0.26 0.60 3.07 1.02 7.20 2.47 0.00
74.348 64.158 11.067 4.60 1.77 68.11 8.67 0.36 0.61 3.22 1.03 7.18 2.23 0.00
74.353 64.158 11.067 4.55 1.74 68.08 8.72 0.29 0.59 3.29 1.03 7.16 2.66 0.01
74.358 64.158 11.067 4.52 1.66 68.61 8.67 0.34 0.57 3.24 1.06 7.23 2.35 0.00
74.363 64.158 11.067 4.64 1.73 67.96 8.58 0.32 0.60 3.22 1.05 7.17 2.37 0.00
74.368 64.158 11.067 5.68 3.56 68.87 7.46 0.63 0.51 3.34 0.85 6.36 1.07 0.02
74.373 64.158 11.066 4.81 1.59 67.81 8.58 0.32 0.59 3.09 1.02 7.10 2.43 0.00

129
 Table 46 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
74.378 64.158 11.066 4.55 1.71 67.21 8.65 0.29 0.58 3.12 1.05 7.17 2.79 0.02
74.383 64.158 11.066 4.60 1.70 67.51 8.60 0.32 0.59 3.36 1.05 7.08 2.43 0.00
74.388 64.158 11.066 4.68 1.75 67.72 8.76 0.37 0.58 3.20 1.02 7.11 2.29 0.00
74.393 64.158 11.066 4.40 3.13 66.37 9.22 0.60 0.61 2.46 1.05 7.33 1.89 0.01
74.398 64.158 11.066 4.72 1.62 68.42 8.81 0.34 0.57 3.12 1.02 7.14 2.38 0.00
74.403 64.158 11.066 4.88 1.61 68.42 8.69 0.35 0.55 3.01 1.04 7.07 2.05 0.00
74.408 64.158 11.068 4.88 1.71 68.67 8.73 0.35 0.54 3.30 1.03 7.10 2.12 0.01
74.413 64.158 11.068 4.87 2.40 68.59 8.63 0.37 0.59 3.03 0.99 6.97 2.11 0.02
74.418 64.158 11.066 4.87 1.71 68.70 8.70 0.35 0.57 3.41 1.00 7.00 1.65 0.00
74.423 64.158 11.068 3.53 9.56 65.11 6.64 3.03 0.44 2.68 0.78 5.90 0.94 0.00
74.428 64.158 11.066 0.32 28.28 55.10 0.69 9.34 0.09 2.99 0.15 2.90 0.00 0.01
74.433 64.158 11.066 0.34 28.32 55.65 0.68 9.21 0.05 2.53 0.15 2.83 0.00 0.00
74.438 64.158 11.066 0.37 26.67 56.91 0.92 9.85 0.11 1.75 0.17 3.07 0.00 0.01
74.443 64.158 11.068 3.03 16.92 62.22 1.44 9.70 0.13 2.14 0.25 3.20 0.00 0.00
74.448 64.158 11.068 0.32 29.43 55.77 0.58 8.15 0.06 2.83 0.12 2.91 0.00 0.00
74.453 64.158 11.068 0.97 45.68 34.07 2.84 5.77 0.22 9.53 0.30 2.80 0.00 0.01
74.458 64.158 11.068 5.33 1.66 69.66 8.51 0.40 0.55 2.95 0.93 6.84 1.39 0.01
74.463 64.158 11.068 5.32 1.62 69.86 8.50 0.35 0.59 2.74 0.98 6.84 1.53 0.01
74.468 64.158 11.066 5.43 15.67 60.91 4.16 3.19 0.30 4.50 0.46 4.10 0.25 0.01
74.473 64.158 11.066 1.63 17.51 58.98 2.26 8.42 0.17 2.10 0.25 3.54 0.00 0.00
74.478 64.158 11.003 2.20 3.35 37.04 7.31 0.76 0.24 5.20 0.73 2.91 1.37 0.00
74.483 64.158 11.067 5.81 5.41 68.75 7.04 0.80 0.46 2.92 0.80 5.89 0.70 0.00
74.488 64.158 11.067 5.06 1.60 69.21 8.59 0.35 0.57 2.61 0.95 6.86 2.51 0.00
74.493 64.158 11.065 5.48 2.09 70.60 8.20 0.49 0.50 2.81 0.91 6.51 1.46 0.01
74.498 64.158 11.067 1.86 16.95 62.21 3.76 8.91 0.25 1.53 0.43 4.36 0.00 0.01
74.503 64.158 11.067 7.86 6.11 71.63 3.87 1.42 0.27 3.64 0.45 4.10 0.29 0.02
74.508 64.158 11.067 0.67 43.17 37.81 1.04 7.89 0.09 9.41 0.15 2.04 0.00 0.01
74.513 64.158 11.066 0.29 34.25 50.08 0.58 8.76 0.06 4.81 0.12 2.54 0.00 0.01
74.518 64.158 11.066 0.35 29.23 55.59 0.74 8.58 0.09 3.02 0.18 2.93 0.00 0.01
74.523 64.158 11.066 0.29 27.62 55.05 0.76 9.13 0.08 2.91 0.16 2.90 0.00 0.00
74.528 64.158 11.064 4.46 1.79 67.89 8.56 0.43 0.53 3.30 1.05 6.85 3.06 0.00
74.533 64.158 11.064 4.69 1.57 68.33 8.52 0.33 0.53 3.02 1.05 6.85 2.77 0.01
74.538 64.158 11.062 3.00 23.04 50.75 6.71 2.44 0.44 6.71 0.73 4.93 0.19 0.00
74.543 64.158 11.064 2.55 27.62 45.92 6.57 2.10 0.36 7.45 0.65 4.84 0.00 0.01
74.548 64.158 11.066 0.29 27.04 55.63 0.63 9.62 0.07 2.61 0.17 2.92 0.00 0.00
74.553 64.158 11.064 0.29 28.52 54.53 0.57 9.16 0.05 2.63 0.16 2.87 0.00 0.00
74.558 64.158 11.064 0.28 28.40 54.96 0.56 9.31 0.07 2.40 0.12 2.83 0.00 0.00
74.563 64.158 11.062 0.28 26.58 53.92 0.62 9.28 0.05 2.32 0.16 2.83 0.00 0.01
74.568 64.158 11.062 0.28 27.48 53.96 0.69 9.39 0.07 3.07 0.14 2.80 0.00 0.02
74.573 64.158 11.066 3.49 21.00 53.49 6.96 0.88 0.46 6.58 0.76 5.35 0.58 0.00
74.578 64.158 11.066 4.49 1.63 68.45 9.00 0.35 0.56 2.82 1.04 7.10 2.25 0.00
74.583 64.158 11.064 4.45 1.72 67.62 8.97 0.35 0.56 3.14 1.01 6.94 2.74 0.00
74.588 64.158 11.066 4.45 1.69 68.66 9.09 0.32 0.55 3.06 1.02 7.01 2.31 0.01
74.593 64.158 11.064 4.30 1.64 67.67 9.16 0.35 0.56 2.99 1.04 6.95 2.76 0.00
74.598 64.158 11.068 4.99 4.13 65.68 9.99 0.74 0.62 2.80 1.13 7.50 2.95 0.00
74.603 64.158 11.066 4.36 1.65 67.84 9.01 0.33 0.57 3.15 1.04 7.08 2.76 0.00
74.608 64.158 11.068 4.37 1.65 68.21 8.98 0.33 0.58 2.86 1.03 7.17 2.89 0.01
74.613 64.158 11.066 4.34 1.66 67.18 8.97 0.32 0.55 3.04 1.02 7.09 3.13 0.01
74.618 64.158 11.068 4.37 1.70 67.58 8.88 0.33 0.59 2.87 1.04 7.15 3.11 0.00
74.623 64.158 11.065 4.34 1.70 67.05 8.88 0.36 0.54 2.72 1.04 7.09 3.04 0.02
74.628 64.158 11.065 4.27 1.60 66.69 9.10 0.31 0.52 2.82 1.05 7.06 3.24 0.01
74.633 64.158 11.067 4.36 1.64 67.42 8.98 0.33 0.57 2.67 1.03 7.17 2.90 0.00
74.638 64.158 11.067 4.36 1.65 67.84 9.01 0.33 0.57 3.15 1.04 7.08 2.76 0.00
74.643 64.158 11.065 4.37 1.65 68.21 8.98 0.33 0.58 2.86 1.03 7.17 2.89 0.01
74.648 64.158 11.065 4.34 1.66 67.18 8.97 0.32 0.55 3.04 1.02 7.09 3.13 0.01
74.653 64.158 11.067 4.37 3.05 67.15 8.49 0.75 0.50 3.01 0.97 6.95 2.38 0.00
74.658 64.158 11.066 0.29 28.76 54.29 0.67 8.10 0.08 2.91 0.13 2.93 0.00 0.00
74.663 64.158 11.066 0.34 29.19 54.49 1.02 8.40 0.09 2.97 0.20 3.05 0.00 0.00
74.668 64.158 11.066 4.00 5.23 64.01 8.40 0.69 0.45 3.95 1.00 6.80 2.53 0.01
74.673 64.158 11.066 1.01 23.28 57.27 2.47 7.48 0.15 3.32 0.35 3.83 0.00 0.01

130
 Table 46 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
74.678 64.158 11.064 3.92 5.53 62.65 8.56 0.50 0.42 4.78 1.00 6.87 3.25 0.00
74.683 64.158 11.066 3.17 16.43 54.78 7.06 1.27 0.36 6.96 0.81 5.90 2.15 0.01
74.688 64.158 11.066 3.94 6.32 62.13 8.40 0.67 0.41 4.26 0.97 6.93 3.36 0.01
74.693 64.158 11.066 4.17 1.61 66.17 9.04 0.32 0.40 3.35 1.07 7.44 3.60 0.00
74.698 64.158 11.066 4.17 1.63 66.09 9.14 0.30 0.40 3.38 1.07 7.55 3.69 0.00
74.703 64.158 11.066 4.13 1.61 65.62 9.28 0.27 0.41 3.20 1.08 7.57 3.74 0.01
74.708 64.158 11.066 3.69 6.97 60.67 8.77 0.41 0.38 4.89 0.97 7.01 2.93 0.01
74.713 64.158 11.066 4.19 1.66 65.97 9.19 0.29 0.39 3.29 1.06 7.56 3.50 0.00
74.718 64.158 11.066 3.70 9.69 59.90 8.38 0.72 0.37 5.49 0.99 6.76 2.70 0.00
74.723 64.158 11.066 4.21 1.68 65.99 9.07 0.29 0.40 3.44 1.10 7.56 3.39 0.00
74.728 64.158 11.066 0.82 21.49 56.85 2.30 6.85 0.14 3.42 0.27 3.97 0.00 0.01
74.733 64.158 11.066 4.16 1.61 66.51 9.31 0.32 0.38 3.34 1.01 7.69 3.25 0.00
74.738 64.158 11.066 4.24 1.71 66.33 9.23 0.30 0.36 3.48 1.03 7.52 3.20 0.00
74.743 64.158 11.066 4.19 1.64 66.16 9.36 0.29 0.37 3.51 1.09 7.63 3.40 0.00
74.748 64.158 11.066 4.16 1.62 66.64 9.42 0.29 0.38 3.46 1.08 7.71 3.60 0.00
74.753 64.158 11.066 4.26 1.70 66.61 9.33 0.32 0.35 3.76 1.06 7.55 3.20 0.00
74.758 64.158 11.066 4.19 1.69 66.65 9.38 0.27 0.37 3.63 1.09 7.70 3.27 0.00
74.763 64.158 11.065 4.16 1.69 67.08 9.43 0.29 0.37 3.65 1.11 7.67 2.93 0.00
74.768 64.158 11.065 4.26 1.74 66.54 9.20 0.30 0.37 3.57 1.00 7.60 2.90 0.00
74.773 64.158 11.065 3.67 7.10 62.84 7.89 1.43 0.31 4.24 0.95 6.75 2.47 0.00
74.778 64.158 11.065 2.13 7.05 60.67 6.54 2.10 0.26 3.50 0.58 6.08 1.85 0.02
74.783 64.158 11.063 4.10 1.70 65.52 9.27 0.30 0.35 3.83 1.05 7.39 3.69 0.00
74.788 64.158 11.069 4.03 3.81 65.13 9.14 0.32 0.37 4.32 1.05 7.56 3.44 0.00
74.793 64.158 11.067 4.17 1.67 66.26 9.40 0.29 0.37 3.54 1.09 7.81 3.69 0.00
74.798 64.158 11.064 4.11 1.72 65.84 9.40 0.28 0.35 3.90 1.09 7.73 3.64 0.00
74.803 64.158 11.064 3.45 10.86 58.52 8.52 0.45 0.34 6.31 0.98 6.84 3.03 0.01
74.808 64.158 11.017 1.73 2.65 40.95 9.97 0.38 0.23 6.43 1.01 4.53 5.29 0.00
74.813 64.158 11.066 4.21 1.76 66.03 9.32 0.27 0.35 3.57 1.06 7.74 3.56 0.01

The line scan results shown in Table 46 revealed the presence of three possible Hf-rich
intermetallic phases, with a phase containing between 24% and 28% Hf being most prevalent.
Other Hf-rich phases detected during the scan contained between 11 and 13% Hf, and 34 to
45% Hf. In order to perform a more detailed analysis of the intermetallic phases observed in
the braze joint, three spot chemical analyses were performed within the intermetallic
compound in the braze joint. Two backscatter electron images of the MarM247/Ni-Hf braze
joint are shown in Figure 145, with the location of the three phase analyses highlighted. The
results of the microprobe phase analyses, shown in Table 47, suggest that the intermetallic
phase within the braze alloy contained between 31 and 36% Hf (wt.%). The binary Ni-Hf
phase diagram, shown in Figure 49, indicates that the Ni5Hf intermetallic phase contains
approximately 38 wt.% Hf, whereas the Ni7Hf2 intermetallic compound contains about 46
wt.% Hf. The results shown in Table 47 therefore suggest that the intermetallic phase in the
braze joint is Ni5Hf, rather than Ni7Hf2. It is, however, interesting to note that the Hf-rich
phase identified in Table 47, as well as the phase containing between 24% and 28% Hf in
Table 46, contained high levels of Zr (between 7.4% and 9.6%), probably present as a
contaminant that segregated preferentially to the intermetallic phase on solidification, and
approximately 3% Ta, which may be present as a result of dilution with the MarM247 powder
particles during brazing. The presence of Zr and Ta may shift the composition of the Hf-rich
intermetallic phase away from the equilibrium composition predicted by the phase diagram.
If it is assumed that the intermetallic compound contains between 7.4% and 9.6% Zr in
solution, as well as 3% Ta, its composition may well approach that of the Ni7Hf2 (or rather
Ni7(Hf,Zr,Ta)2) intermetallic compound.

131
In order to study the distribution of various elements within the Ni-Cr-Hf braze
microstructure, EMPA maps were constructed. These maps are shown in Figure 146 (for Al,
C, Ni and Cr), Figure 147 (for Zr, Hf and Si), Figure 148 (for Ru, Ta, Ti and W) and Figure
149 (for Fe and Mo).

Figure 145(a) – Secondary electron

image of the Ni-Cr-Hf alloy in the as-
brazed condition.

Figure 145(b) – Enlarged view of the

secondary electron image of the Ni-Cr-Hf
braze shown in Figure 145(a), highlighting
the location of three spot chemical analyses
of the intermetallic compound within the braze.

Table 47 - Quantitative EPMA data (taken at 15 keV and 80 nA with a focused spot beam) at
three locations (indicated in Figure 145(b)) within the intermetallic phase. (Percentage by
weight).
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti Si
Phase 1 0.29 31.53 55.16 0.84 8.43 0.06 3.35 0.15 3.37 0.07 0.00 0.21
Phase 2 0.29 31.09 55.37 0.71 8.54 0.05 3.39 0.18 3.10 0.08 0.01 0.25
Phase 3 0.29 36.34 53.02 0.62 5.30 0.05 3.91 0.15 3.62 0.13 0.00 0.21

132
 Figure 146 – EMPA maps displaying the distribution of Al, C, Ni and Cr within the
MarM247/Ni-Cr-Hf braze microstructure (as-brazed condition).

The EMPA results shown in Figure 146 indicate that Al and Cr partitioned strongly to the
MarM247 powder particles during brazing. Given that the braze alloy contains 7% Cr and
that a diffusion cycle was not applied, the low Cr concentration measured in the braze alloy
between the MarM247 particles was surprising. Ni and C appeared to be fairly uniformly
distributed between the braze alloy and the matrix, with a slightly higher Ni concentration
measured in the MarM247 particles.

133
 Figure 147 – EMPA maps displaying the distribution of Zr, Hf and Si within the
MarM247/Ni-Cr-Hf braze microstructure (as-brazed condition).

The results shown in Figure 147 suggest that both Hf and Zr partitioned strongly to the braze
alloy, which confirms the results of the line and spot scans shown in Tables 46 and 47. Zr
was not intentionally added to the braze alloy, and is almost completely absent from the
MarM247 matrix particles (MarM247 has a nominal Zr content of 0.05 wt.%). A Zr content
of 0.7 wt.% was measured in the binary Ni-Hf braze alloy as a result of contamination (see
Table 9), but this is not considered high enough to cause the high Zr levels shown in Tables
46 and 47, and Figure 147. It is possible that the Cr powder used to make up the ternary Ni-
Cr-Hf braze alloy contained a significant amount of Zr as contaminant. The Hf EMPA map
confirms that the intermetallic phase within the braze joint is Hf-rich.

134
 Figure 148 – EMPA maps displaying the distribution of Ru, Ta, Ti and W within the
MarM247/Ni-Cr-Hf braze microstructure (as-brazed condition).

As shown in Figure 148, Ru, Ta and Ti were distributed fairly evenly between the braze alloy
and the matrix. Tungsten, however, partitioned strongly to the MarM247 powder particles,
suggesting low solubility within the intermetallic phase. The W appeared to be distributed
evenly throughout the matrix particles, implying adequate solubility and little driving force
for the precipitation of W-rich carbide particles or Laves phase.

The EMPA maps for Fe and Mo, shown in Figure 149, display preferential partitioning to the
matrix particles, but due to the low levels of these elements within the MarM247 particles and
the braze alloy, partitioning was weak.

135
Figure 149 – EMPA maps displaying the distribution of Fe and Mo within the MarM247/Ni-
Cr-Hf braze microstructure (as-brazed condition).

11.3.2 The MarM247/Ni-Zr braze joint produced using Ni-7Cr-13Zr braze filler (as-brazed
condition):

Secondary and backscatter electron images of the MarM247/Ni-Zr joint produced using the
experimental Ni-7Cr-13Zr braze filler metal are shown in Figure 150 in the as-brazed
condition. The arrow indicates the location and direction of the 2 m stepped EPMA line
scan. The results of the EMPA scan are shown in Table 48.

The line scan results shown in Table 48 revealed the presence of a Zr-rich intermetallic
phase, containing between 23% and 28% Zr. In order to perform a more detailed analysis of
the intermetallic phase observed in the braze joint, four spot chemical analyses were
performed within the intermetallic compound in the braze joint. Two backscatter electron
images of the MarM247/Ni-Zr braze joint are shown in Figure 151, with the location of the
four spot analyses highlighted. The results of the microprobe phase analyses, shown in Table
49, suggest that the intermetallic phase within the braze alloy contained between 27 and 29%
Zr. The binary Ni-Zr phase diagram, shown in Figure 50, indicates that the Ni7Zr2
intermetallic phase contains approximately 30 wt.% Zr, whereas the Ni5Zr intermetallic
compound contains between 21 and 26 wt.% Zr. The results shown in Table 49 are therefore
inconclusive, suggesting that the intermetallic phase in the braze joint could be either Ni5Zr or
Ni7Zr2. Phase 3, shown in Figure 145(b), is most likely the Ni-rich  phase (with a
composition shown in Table 49).

136
 (a) (b)

Figure 150 – Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-
13Zr joint in the as-brazed condition (adjacent to the interface).

Table 48 – The results of the EMPA scan of the MarM247/Ni-7Cr-13Zr braze joint in the as-
brazed condition. (All compositions given in wt.%).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
76.613 65.076 11.043 4.45 2.01 64.25 7.85 0.21 0.62 4.39 1.34 8.94 3.84 0.00
76.615 65.076 11.043 4.46 2.22 64.32 7.71 0.25 0.66 4.31 1.30 8.91 4.08 0.00
76.617 65.076 11.043 4.37 2.30 64.17 7.81 0.29 0.67 4.26 1.32 8.99 3.75 0.00
76.619 65.076 11.043 4.46 2.00 64.08 7.77 0.24 0.67 4.48 1.37 8.76 4.02 0.00
76.621 65.076 11.043 4.48 2.06 63.94 7.71 0.27 0.68 4.41 1.31 8.75 3.86 0.00
76.623 65.076 11.045 4.40 3.48 63.34 7.48 0.55 0.70 5.15 1.32 8.60 3.73 0.00
76.625 65.076 11.043 4.14 4.85 60.39 7.44 0.89 0.69 6.02 1.30 8.34 3.61 0.00
76.627 65.076 11.043 4.32 2.58 63.11 7.80 0.36 0.75 4.23 1.43 8.81 3.91 0.00
76.629 65.076 11.043 4.54 2.04 64.22 7.74 0.24 0.72 4.02 1.33 8.61 4.03 0.00
76.631 65.076 11.043 4.41 1.95 63.21 7.94 0.22 0.81 3.95 1.37 8.76 4.19 0.00
76.633 65.076 11.043 4.46 2.06 63.96 7.94 0.27 0.84 4.09 1.38 8.71 4.02 0.00
76.635 65.076 11.043 4.41 1.87 64.25 7.99 0.24 0.85 4.21 1.36 8.84 3.82 0.00
76.637 65.076 11.045 4.18 4.88 61.14 7.61 0.59 0.83 5.57 1.28 8.40 3.56 0.01
76.639 65.076 11.043 4.44 2.25 63.61 7.78 0.28 0.83 4.32 1.34 8.71 3.94 0.00
76.641 65.076 11.045 4.52 1.99 64.14 7.80 0.25 0.86 3.86 1.38 8.84 4.04 0.00
76.643 65.076 11.043 4.47 2.01 63.15 7.73 0.24 0.88 4.07 1.35 8.66 3.87 0.00
76.645 65.076 11.043 4.41 1.91 63.34 7.82 0.23 0.89 3.77 1.42 8.84 3.87 0.00
76.647 65.076 11.043 4.36 1.91 63.75 7.84 0.24 0.91 4.09 1.35 8.81 4.07 0.00
76.649 65.076 11.043 4.45 2.00 63.77 7.74 0.24 0.88 4.10 1.40 8.75 4.27 0.00
76.651 65.076 11.043 4.40 1.92 63.55 7.82 0.23 0.94 3.90 1.41 8.82 4.35 0.00
76.653 65.076 11.043 4.34 1.91 63.37 7.79 0.28 0.93 3.87 1.44 8.77 4.13 0.00
76.655 65.076 11.043 4.32 1.85 63.62 7.81 0.25 0.93 3.75 1.43 8.87 3.96 0.00
76.657 65.076 11.043 4.40 1.90 63.30 7.76 0.21 0.98 3.79 1.35 8.81 4.19 0.01
76.659 65.076 11.043 4.36 1.83 63.90 7.83 0.20 1.01 3.50 1.44 8.89 3.90 0.00
76.661 65.076 11.043 4.37 1.87 64.21 7.83 0.24 1.01 3.55 1.50 8.77 4.05 0.00
76.663 65.076 11.043 4.49 2.05 64.23 7.72 0.22 0.97 3.77 1.37 8.70 4.23 0.01
76.665 65.076 11.043 4.43 2.01 64.16 7.70 0.25 1.01 3.74 1.41 8.65 3.67 0.01
76.667 65.076 11.043 4.42 1.96 64.26 7.74 0.22 1.00 3.72 1.42 8.72 3.97 0.00
76.669 65.076 11.043 4.48 2.00 64.48 7.69 0.24 0.97 3.64 1.45 8.57 3.74 0.00
76.671 65.076 11.043 4.45 1.96 64.21 7.77 0.26 1.01 3.62 1.42 8.74 3.54 0.00
76.673 65.076 11.042 4.36 1.95 64.01 7.80 0.19 1.00 3.79 1.42 8.72 3.87 0.00
76.675 65.076 11.042 4.38 1.95 63.79 7.81 0.22 0.98 3.77 1.40 8.71 3.90 0.00
76.677 65.076 11.043 4.45 2.00 64.28 7.74 0.22 1.00 3.65 1.41 8.73 3.71 0.00
76.679 65.076 11.043 4.40 1.96 64.36 7.73 0.24 1.00 3.92 1.39 8.73 3.69 0.00
76.681 65.076 11.043 4.38 1.97 64.19 7.70 0.20 1.04 3.63 1.40 8.68 3.56 0.00
76.683 65.076 11.043 4.41 1.97 64.20 7.68 0.23 1.04 3.79 1.44 8.73 3.90 0.02
76.685 65.076 11.043 4.38 1.94 64.42 7.59 0.24 1.03 3.85 1.44 8.78 3.88 0.00

137
 Table 48 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
76.687 65.076 11.043 4.42 1.94 64.33 7.63 0.20 1.03 3.80 1.44 8.68 3.90 0.00
76.689 65.076 11.043 4.36 1.92 63.96 7.62 0.23 1.06 3.99 1.43 8.71 3.66 0.01
76.691 65.076 11.043 4.29 1.81 63.56 7.64 0.18 1.03 3.93 1.48 8.79 3.64 0.00
76.693 65.076 11.043 4.41 1.94 64.08 7.56 0.24 1.09 3.94 1.43 8.71 3.55 0.01
76.695 65.076 11.043 4.38 1.92 64.03 7.59 0.22 1.07 3.93 1.44 8.69 3.90 0.00
76.697 65.076 11.043 4.37 1.90 64.16 7.63 0.22 1.11 3.69 1.46 8.66 3.80 0.00
76.699 65.076 11.043 4.42 1.91 64.26 7.64 0.20 1.10 3.76 1.54 8.73 3.64 0.00
76.701 65.076 11.043 4.39 1.94 64.11 7.67 0.25 1.10 3.88 1.43 8.63 3.54 0.00
76.703 65.076 11.043 4.35 1.88 64.24 7.79 0.23 1.15 3.54 1.50 8.71 3.89 0.00
76.705 65.076 11.043 4.38 1.90 64.52 7.82 0.28 1.12 3.65 1.52 8.69 3.55 0.00
76.707 65.076 11.043 4.38 1.89 64.51 7.86 0.21 1.18 3.72 1.48 8.67 3.31 0.00
76.709 65.076 11.043 4.44 1.92 64.62 7.93 0.22 1.19 3.53 1.52 8.68 3.24 0.00
76.711 65.076 11.043 4.40 1.89 64.75 7.95 0.25 1.15 3.78 1.51 8.64 3.44 0.00
76.713 65.076 11.043 4.43 1.99 64.31 7.80 0.23 1.21 3.69 1.50 8.55 3.08 0.01
76.715 65.076 11.043 4.34 1.85 64.00 7.98 0.21 1.25 3.87 1.57 8.63 3.22 0.00
76.717 65.076 11.043 4.33 1.86 64.46 7.97 0.26 1.28 3.59 1.56 8.68 3.13 0.00
76.719 65.076 11.043 4.41 2.00 64.22 7.87 0.24 1.27 3.90 1.54 8.49 3.11 0.00
76.721 65.076 11.043 4.36 1.91 64.30 7.96 0.24 1.34 3.99 1.56 8.57 3.32 0.00
76.723 65.076 11.043 4.33 1.85 64.47 7.95 0.22 1.35 4.07 1.61 8.58 3.13 0.00
76.725 65.076 11.043 4.33 2.81 63.11 7.74 0.34 1.39 4.51 1.61 8.35 3.06 0.00
76.727 65.076 11.042 4.26 1.81 63.33 8.02 0.24 1.48 4.19 1.65 8.42 3.04 0.02
76.729 65.076 11.043 3.78 7.14 59.37 7.68 0.63 1.49 5.70 1.56 7.92 2.94 0.00
76.731 65.076 11.043 3.86 7.12 59.30 7.69 0.64 1.56 5.50 1.58 8.04 2.71 0.00
76.733 65.076 11.043 4.21 1.74 63.93 8.38 0.23 1.85 3.78 1.90 8.51 3.31 0.00
76.735 65.076 11.043 4.13 1.70 63.82 8.47 0.23 1.98 3.83 1.88 8.48 3.34 0.00
76.737 65.076 11.043 4.00 1.52 63.46 8.79 0.20 2.18 3.54 2.07 8.50 3.67 0.00
76.739 65.076 11.043 3.95 1.41 63.11 9.04 0.16 2.41 3.38 2.20 8.34 3.50 0.00
76.741 65.076 11.043 3.92 1.38 63.12 9.12 0.16 2.56 3.26 2.24 8.34 3.46 0.00
76.743 65.076 11.043 3.87 1.31 63.07 9.30 0.14 2.71 3.17 2.37 8.38 3.49 0.00
76.745 65.076 11.043 3.83 1.25 63.10 9.42 0.14 2.93 2.99 2.51 8.34 3.08 0.00
76.747 65.076 11.043 3.74 1.17 62.72 9.68 0.16 3.16 2.86 2.62 8.30 3.47 0.00
76.749 65.076 11.043 3.64 1.08 62.55 9.76 0.15 3.34 2.98 2.74 8.36 3.20 0.00
76.751 65.076 11.043 3.61 1.02 61.95 10.00 0.17 3.60 2.99 2.91 8.06 3.07 0.00
76.753 65.076 11.043 3.51 0.93 62.06 10.23 0.12 3.83 2.93 3.05 8.00 3.04 0.00
76.755 65.076 11.043 3.41 0.82 62.00 10.45 0.14 4.10 2.81 3.19 8.09 2.90 0.00
76.757 65.076 11.043 3.31 0.72 61.47 10.67 0.10 4.34 2.73 3.41 8.03 2.92 0.00
76.759 65.076 11.043 3.26 0.65 60.91 10.86 0.07 4.61 2.69 3.51 7.89 3.14 0.00
76.761 65.076 11.043 3.17 0.51 60.73 11.10 0.06 4.87 2.50 3.64 7.81 2.65 0.00
76.763 65.076 11.043 3.09 0.87 60.18 11.17 0.11 5.03 2.31 3.77 7.76 3.09 0.00
76.765 65.076 11.043 1.83 0.07 41.03 8.09 28.44 3.80 7.28 2.90 5.16 0.98 0.02
76.767 65.076 11.043 2.71 4.42 56.84 11.29 0.57 5.32 2.77 3.97 7.34 2.40 0.00
76.769 65.076 11.043 2.81 0.49 59.27 12.00 0.07 5.83 2.14 4.19 7.54 2.77 0.00
76.771 65.076 11.043 2.71 0.24 58.61 12.20 0.03 6.24 1.79 4.38 7.42 2.77 0.00
76.773 65.076 11.042 2.57 0.82 57.49 12.50 0.12 6.35 1.85 4.53 7.20 2.65 0.00
76.775 65.076 11.042 0.90 8.45 35.71 6.05 26.87 3.31 9.97 2.49 3.14 0.00 0.04
76.777 65.076 11.042 1.13 9.42 33.34 7.90 25.31 4.43 8.72 3.25 3.87 0.00 0.01
76.779 65.076 11.043 0.95 8.54 40.64 7.75 23.83 4.67 6.23 3.38 3.63 0.00 0.03
76.781 65.076 11.042 1.93 3.62 54.13 13.61 0.33 8.16 1.66 5.48 6.27 1.81 0.00
76.783 65.076 11.043 1.62 3.47 47.74 12.44 11.30 7.78 2.23 5.10 5.47 0.72 0.01
76.785 65.076 11.043 1.82 2.55 54.10 14.38 0.20 9.27 1.41 5.87 6.14 1.66 0.00
76.787 65.076 11.043 1.81 0.05 55.28 15.05 0.00 10.03 1.07 6.12 6.06 1.75 0.00
76.789 65.076 11.043 1.71 0.03 54.97 15.52 0.00 10.57 1.14 6.23 5.85 1.56 0.00
76.791 65.076 11.043 1.62 0.02 54.92 15.64 0.02 10.85 1.03 6.49 5.75 1.73 0.00
76.793 65.076 11.043 1.58 0.03 55.03 16.00 0.00 11.24 1.07 6.52 5.55 1.24 0.00
76.795 65.076 11.043 1.47 0.02 54.09 16.20 0.00 11.79 0.79 6.62 5.23 1.17 0.00
76.797 65.076 11.043 1.40 0.02 53.18 16.49 0.00 12.21 0.83 6.85 4.94 1.33 0.00
76.799 65.076 11.043 1.32 0.01 53.32 16.65 0.00 12.66 0.77 7.04 4.74 1.18 0.00
76.801 65.076 11.043 1.25 0.02 52.98 16.84 0.01 13.03 0.81 7.06 4.55 1.27 0.00
76.803 65.076 11.043 1.19 0.03 52.58 17.03 0.01 13.32 0.70 7.10 4.37 1.27 0.00
76.805 65.076 11.043 1.15 0.02 52.71 17.27 0.00 13.52 0.69 7.22 4.27 1.10 0.00

138
 Table 48 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
76.807 65.076 11.043 1.09 0.00 52.21 17.50 0.00 13.70 0.56 7.40 4.05 1.02 0.00
76.809 65.076 11.043 1.03 0.02 51.56 17.81 0.00 14.30 0.52 7.54 3.89 0.97 0.00
76.811 65.076 11.043 0.98 0.03 51.29 18.06 0.00 14.54 0.51 7.73 3.70 1.13 0.00
76.813 65.076 11.043 0.94 0.00 51.20 18.40 0.02 14.71 0.58 7.77 3.51 1.17 0.00

Table 49 - Quantitative EPMA data (taken at 15 keV and 80 nA with a focused spot beam) at
four locations within the intermetallic phase (indicated in Figure 151(b)). (Percentage by
weight).
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti Si
Phase 1 0.20 0.0 66.81 1.34 27.54 0.12 1.32 0.19 2.31 0.03 0.62 0.26
Phase 2 0.22 0.0 66.64 1.15 27.83 0.08 1.11 0.21 2.23 0.00 0.64 0.28
Phase 3 3.83 0.0 74.46 9.70 0.82 0.49 1.00 0.95 4.39 0.35 3.61 0.18
Phase 4 0.65 0.0 70.43 6.80 9.74 0.26 0.66 0.81 4.42 0.02 2.07 0.15

In order to study the distribution of various elements within the Ni-Cr-Zr braze
microstructure, EMPA maps were constructed. These maps are shown in Figure 152 (for Al,
C, Ni and Cr), Figure 153 (for Zr, Hf and Si), Figure 154 (for Ru, Ta, Ti and W) and Figure
155 (for Fe and Mo).

Figure 151(a) – Secondary electron

image of the Ni-Cr-Zr alloy in the as-
brazed condition.

Figure 151(b) – Enlarged view of the

secondary electron image of the Ni-Cr-Zr
braze, shown in Figure 151(a), highlighting
the location of four spot chemical analyses
of the intermetallic compound within the braze.

139
 Figure 152 – EMPA maps displaying the distribution of Al, C, Ni and Cr within the
MarM247/Ni-Cr-Zr braze microstructure (as-brazed condition).

The EMPA results shown in Figure 152 indicate that in this braze Al and Cr also partitioned
strongly to the MarM247 powder particles during brazing. Given that the braze alloy contains
7% Cr and that a diffusion cycle was not applied, the low Cr concentration measured in the
braze alloy between the MarM247 particles was again surprising. Ni and C appeared to be
fairly uniformly distributed between the braze alloy and the matrix, with a slightly higher Ni
concentration measured in the MarM247 particles.

140
 Figure 153 – EMPA maps displaying the distribution of Zr and Hf within the MarM247/Ni-
Cr-Zr braze microstructure (as-brazed condition).

The results shown in Figure 153 suggest that both Hf and Zr partitioned strongly to the braze
alloy, which substantiates the results of the line and spot scans shown in Tables 48 and 49.
The Zr EMPA map confirms that the intermetallic phase within the braze joint was Zr-rich,
whereas Zr was almost completely absent from the MarM247 matrix particles (MarM247 has
a nominal Zr content of 0.05 wt.%). Hf was not intentionally added to the braze alloy, and the
binary eutectic Ni-Zr alloy analyzed in Chapter 3 (see Table 9) did not contain any Hf. The
Hf detected in the Ni-Cr-Zr alloy may therefore be derived from the MarM247 powder
particles (MarM247 has a nominal Hf content of 1.5 wt.%), or from the Cr powder used to
make up the ternary alloy.

141
 Figure 154 – EMPA maps displaying the distribution of Si, Ta, Ti and W within the
MarM247/Ni-Cr-Zr braze microstructure (as-brazed condition).

As shown in Figure 154, Si, Ta and Ti were distributed fairly evenly between the braze alloy
and the matrix. The Si apparently partitioned weakly to the intermetallic phase (suggesting
limited solubility within the intermetallic compound), whereas Ti partitioned preferentially to
the MarM247 powder particles. Tungsten, however, partitioned strongly to the MarM247
powder particles, suggesting low solubility within the intermetallic phase. The W appeared to
be distributed evenly throughout the matrix particles, implying adequate solubility and little
driving force for the precipitation of W-rich carbide particles or Laves phase.

The EMPA maps for Fe and Mo, shown in Figure 155, display weak partitioning to the
matrix particles, but due to the low levels of these elements within the MarM247 particles and
the braze alloy, this tendency was limited.

142
Figure 155 – EMPA maps displaying the distribution of Fe and Mo within the MarM247/Ni-
Cr-Zr braze microstructure (as-brazed condition).

11.3.3 The MarM247/Ni-Zr braze joint produced using Ni-7Cr-13Zr braze filler (after an
extended diffusion cycle):

Secondary and backscatter electron images of the MarM247/Ni-Zr joint produced using the
experimental Ni-7Cr-13Zr braze filler metal are shown in Figure 156 after an extended
diffusion cycle at 1230C for 12 hours. The arrow indicates the location and direction of the
2 m stepped EPMA line scan. The results of the EMPA scan are shown in Table 50.

The line scan results shown in Table 50 revealed the presence of Zr-rich intermetallic phases
containing between 20% and 29% Zr, and occasionally between 10% and 18% Zr. In order to
perform a more detailed analysis of the intermetallic phase observed in the braze joint, four
spot chemical analyses were performed within the intermetallic compound in the braze joint.
Two backscatter electron images of the MarM247/Ni-Zr braze joint are shown in Figure 157,
with the location of the spot analyses highlighted. The results of the microprobe phase
analyses, shown in Table 51, suggest that the intermetallic phase within the braze alloy
(phases 1, 2 and 3) contained between 27 and 28% Zr, as well as some Hf, Ta and Co. Phase
4, shown in Table 51, was most likely the Ni-rich  phase, with some Cr, Co and W in
solution. The binary Ni-Zr phase diagram, shown in Figure 50, indicates that the Ni7Zr2
intermetallic phase contains approximately 30 wt.% Zr, whereas the Ni5Zr intermetallic
compound contains approximately between 21 and 26 wt.% Zr. The results shown in Table
51 are therefore inconclusive, suggesting that the intermetallic phase in the braze joint could
be either Ni5Zr or Ni7Zr2. As a result of this uncertainty, another line scan was performed.
The results of this line scan are described in §11.3.4.

143
 (a) (b)

Figure 156 – Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-
13Zr joint after an extended diffusion cycle (adjacent to the interface).

Table 50 – The results of the EMPA scan of the MarM247/Ni-7Cr-13Zr braze joint after an
extended diffusion cycle. (All compositions given in wt.%).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
47.333 62.92 11.122 2.62 0.05 71.83 9.93 0.70 0.56 1.68 0.50 4.77 6.11 0.46
47.335 62.92 11.122 2.61 0.07 71.38 9.89 0.70 0.53 1.67 0.52 4.82 5.97 0.46
47.338 62.92 11.122 2.62 0.03 71.76 9.89 0.72 0.57 1.56 0.49 4.75 6.00 0.46
47.34 62.92 11.12 2.59 0.03 70.71 9.93 0.72 0.54 1.90 0.49 4.75 6.05 0.46
47.341 62.92 11.12 2.59 0.04 70.62 9.86 0.74 0.57 1.67 0.49 4.69 6.13 0.46
47.343 62.92 11.12 2.57 0.05 70.84 9.83 0.72 0.55 1.51 0.50 4.76 6.09 0.45
47.346 62.92 11.12 2.61 0.04 71.22 9.78 0.76 0.56 1.44 0.47 4.79 5.77 0.48
47.348 62.92 11.12 2.58 0.07 71.05 9.86 0.76 0.56 1.70 0.50 4.81 5.98 0.46
47.35 62.92 11.12 2.55 0.07 71.08 9.76 0.72 0.54 1.72 0.47 4.75 6.09 0.46
47.352 62.92 11.122 2.61 0.04 71.34 9.70 0.79 0.56 1.69 0.45 4.72 5.83 0.48
47.354 62.92 11.122 2.62 0.08 71.67 9.67 0.69 0.55 1.79 0.45 4.76 5.74 0.43
47.356 62.92 11.12 2.59 0.08 71.61 9.64 0.72 0.54 1.71 0.51 4.78 5.93 0.46
47.358 62.92 11.122 2.61 0.03 71.99 9.54 0.65 0.54 1.77 0.47 4.71 6.05 0.46
47.36 62.92 11.122 1.29 0.97 68.22 5.20 14.72 0.33 1.69 0.27 3.29 2.88 0.28
47.362 62.92 11.122 0.20 1.69 65.25 1.02 26.72 0.07 1.55 0.09 1.97 0.75 0.11
47.364 62.92 11.122 0.16 1.71 64.84 0.82 27.12 0.06 1.56 0.07 1.94 0.59 0.09
47.366 62.92 11.122 0.17 1.78 65.10 0.73 26.92 0.07 1.53 0.06 1.88 0.21 0.10
47.368 62.92 11.122 0.17 1.76 65.53 0.71 27.04 0.05 1.47 0.07 1.93 0.25 0.10
47.37 62.92 11.12 0.16 1.79 65.22 0.65 27.06 0.06 1.57 0.09 1.92 0.46 0.10
47.372 62.92 11.12 0.17 1.76 64.78 0.69 27.16 0.07 1.54 0.05 1.89 0.52 0.10
47.374 62.92 11.122 0.16 1.79 65.09 0.68 27.27 0.07 1.36 0.07 1.95 0.39 0.11
47.376 62.92 11.12 0.17 1.78 64.38 0.72 27.23 0.07 1.49 0.07 1.85 0.54 0.09
47.378 62.92 11.122 0.16 1.81 65.08 0.69 27.23 0.08 1.49 0.08 1.90 0.58 0.11
47.38 62.92 11.122 1.03 1.22 67.75 3.65 18.44 0.22 1.59 0.23 2.79 2.10 0.21
47.382 62.92 11.12 2.57 0.09 71.37 8.95 1.72 0.49 1.47 0.46 4.67 5.59 0.45
47.384 62.92 11.122 2.66 0.01 71.73 9.45 0.66 0.56 1.70 0.49 4.76 5.74 0.42
47.386 62.92 11.12 2.60 0.04 71.10 9.65 0.73 0.55 1.72 0.52 4.75 5.73 0.44
47.388 62.92 11.122 2.63 0.07 71.80 9.52 0.77 0.54 1.80 0.53 4.76 5.87 0.46
47.39 62.92 11.122 2.63 0.04 71.58 9.61 0.74 0.55 1.76 0.51 4.79 5.72 0.44
47.392 62.92 11.12 2.58 0.06 71.02 9.65 0.72 0.52 1.83 0.51 4.76 6.24 0.45
47.394 62.92 11.122 2.61 0.08 72.02 9.63 0.78 0.54 1.85 0.44 4.86 6.11 0.44
47.396 62.92 11.122 2.61 0.06 71.92 9.56 0.73 0.51 1.73 0.49 4.83 5.91 0.45
47.398 62.92 11.122 2.62 0.04 71.41 9.53 0.69 0.52 1.77 0.48 4.87 6.47 0.45
47.4 62.92 11.122 2.62 0.03 71.46 9.63 0.72 0.54 1.73 0.47 4.84 6.18 0.45
47.402 62.92 11.122 2.59 0.05 71.49 9.68 0.75 0.53 1.96 0.50 4.84 6.01 0.44
47.404 62.92 11.122 2.61 0.05 71.66 9.63 0.75 0.54 1.86 0.44 4.86 6.02 0.43
47.406 62.92 11.122 2.59 0.05 71.82 9.61 0.73 0.50 1.78 0.44 4.92 6.31 0.45

144
 Table 50 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
47.408 62.92 11.122 2.59 0.05 71.68 9.70 0.75 0.51 1.88 0.47 4.88 6.10 0.45
47.41 62.92 11.12 2.57 0.05 70.88 9.71 0.75 0.51 1.76 0.45 4.87 6.46 0.45
47.412 62.92 11.122 2.60 0.05 71.21 9.71 0.75 0.50 1.77 0.47 4.90 6.41 0.47
47.414 62.92 11.122 2.57 0.06 70.97 9.66 0.74 0.51 1.70 0.43 4.84 6.70 0.43
47.416 62.92 11.122 2.59 0.06 71.50 9.74 0.73 0.51 1.89 0.46 4.94 6.28 0.45
47.418 62.92 11.122 2.59 0.03 71.51 9.65 0.75 0.52 1.82 0.46 4.94 6.16 0.44
47.42 62.92 11.122 2.60 0.05 71.74 9.62 0.70 0.48 1.72 0.44 4.91 6.23 0.45
47.422 62.92 11.12 2.48 0.05 70.20 9.63 0.72 0.49 1.96 0.52 4.83 6.13 0.44
47.424 62.92 11.12 2.55 0.05 70.49 9.56 0.71 0.48 1.81 0.49 4.86 6.18 0.46
47.426 62.92 11.122 2.58 0.05 71.49 9.60 0.70 0.50 1.94 0.46 4.91 6.07 0.44
47.428 62.92 11.12 2.53 0.05 70.61 9.56 0.68 0.49 1.97 0.46 4.89 6.48 0.45
47.43 62.92 11.122 2.55 0.06 71.19 9.47 0.72 0.46 2.03 0.47 4.99 6.33 0.44
47.432 62.92 11.12 2.36 1.77 67.34 9.20 2.77 0.47 2.27 0.48 4.65 6.64 0.44
47.434 62.92 11.12 2.51 0.05 70.29 9.61 0.76 0.49 2.15 0.51 4.91 6.65 0.44
47.436 62.92 11.12 2.52 0.05 70.50 9.60 0.77 0.47 2.04 0.47 4.90 6.70 0.43
47.438 62.92 11.122 2.55 0.10 71.45 9.56 0.67 0.49 2.06 0.48 4.85 6.53 0.44
47.44 62.92 11.122 2.57 0.06 71.73 9.57 0.68 0.48 1.81 0.45 5.00 6.66 0.44
47.442 62.92 11.122 2.59 0.06 71.42 9.71 0.77 0.49 2.03 0.47 4.90 6.54 0.46
47.444 62.92 11.122 2.56 0.07 70.87 9.69 0.83 0.48 2.05 0.46 4.97 6.63 0.43
47.446 62.92 11.12 2.23 0.70 64.10 8.98 6.50 0.42 3.50 0.45 4.42 6.40 0.44
47.448 62.92 11.122 2.54 0.21 69.71 9.59 1.74 0.50 2.44 0.49 4.87 6.72 0.43
47.45 62.92 11.122 2.56 0.03 71.11 9.77 0.69 0.50 1.78 0.50 4.90 6.64 0.45
47.452 62.92 11.12 2.53 0.04 71.08 9.86 0.72 0.51 1.92 0.50 4.89 7.11 0.44
47.454 62.92 11.12 2.54 0.03 70.91 9.83 0.74 0.50 1.89 0.47 4.90 7.04 0.46
47.456 62.92 11.12 2.53 0.06 70.46 9.86 0.76 0.50 1.96 0.47 4.84 6.85 0.46
47.458 62.92 11.122 2.58 0.07 71.35 9.79 0.71 0.53 1.72 0.50 4.87 6.44 0.45
47.46 62.92 11.12 2.57 0.04 70.46 9.82 0.67 0.52 2.13 0.50 4.79 6.39 0.44
47.462 62.92 11.122 2.59 0.05 71.32 9.84 0.73 0.55 1.85 0.47 4.89 6.26 0.45
47.464 62.92 11.122 2.58 0.06 71.64 9.83 0.75 0.54 1.95 0.47 4.92 6.07 0.44
47.466 62.92 11.122 2.62 0.04 71.78 9.93 0.71 0.53 1.87 0.47 4.87 5.76 0.44
47.468 62.92 11.12 2.54 0.06 70.81 10.04 0.69 0.55 2.06 0.47 4.85 6.32 0.44
47.47 62.92 11.12 2.56 0.05 70.93 10.05 0.77 0.55 2.25 0.49 4.82 6.20 0.47
47.472 62.92 11.122 2.62 0.03 71.40 9.88 0.70 0.59 2.04 0.49 4.79 5.99 0.44
47.474 62.92 11.122 2.83 0.04 72.11 9.65 0.61 0.57 1.79 0.49 4.77 5.69 0.49
47.476 62.92 11.12 2.17 0.45 71.35 7.02 9.44 0.44 1.35 0.30 3.75 2.28 0.45
47.478 62.92 11.122 2.87 0.04 72.69 9.62 0.60 0.53 1.69 0.46 4.78 5.38 0.49
47.48 62.92 11.12 2.62 0.05 70.86 9.89 0.72 0.58 1.82 0.48 4.89 6.28 0.46
47.482 62.92 11.122 2.60 0.03 71.48 9.79 0.75 0.58 1.94 0.47 4.88 6.20 0.43
47.484 62.92 11.122 2.57 0.04 71.12 9.77 0.73 0.57 1.85 0.49 4.83 6.20 0.45
47.486 62.92 11.12 2.52 0.06 70.49 9.83 0.71 0.57 1.80 0.49 4.84 6.73 0.46
47.488 62.92 11.122 2.58 0.17 71.20 9.68 0.73 0.55 1.90 0.48 4.85 6.44 0.45
47.49 62.92 11.12 2.52 0.38 70.46 9.64 0.92 0.52 2.15 0.51 4.85 6.60 0.46
47.492 62.92 11.122 2.55 0.07 71.46 9.64 0.79 0.51 2.17 0.48 4.94 6.31 0.44
47.494 62.92 11.12 2.51 0.05 70.14 9.78 0.72 0.53 1.99 0.49 4.90 6.86 0.46
47.496 62.92 11.12 2.54 0.04 70.88 9.78 0.70 0.53 2.00 0.46 4.90 7.00 0.44
47.498 62.92 11.12 2.53 0.05 70.78 9.77 0.70 0.51 1.78 0.49 4.87 7.16 0.44
47.5 62.92 11.122 2.57 0.04 71.07 9.62 0.72 0.53 1.72 0.48 4.92 7.03 0.44
47.502 62.92 11.12 2.54 0.03 70.78 9.63 0.69 0.51 1.66 0.52 4.98 7.42 0.46
47.504 62.92 11.122 2.57 0.05 71.01 9.58 0.71 0.51 1.65 0.50 5.04 7.24 0.46
47.506 62.92 11.12 2.54 0.04 70.16 9.65 0.69 0.49 1.69 0.47 4.95 7.35 0.44
47.508 62.92 11.12 2.50 0.07 70.61 9.57 0.75 0.52 1.71 0.48 5.01 7.25 0.45
47.51 62.92 11.12 2.49 0.04 70.57 9.60 0.66 0.50 1.55 0.47 4.98 6.98 0.44
47.512 62.92 11.122 2.44 0.16 69.53 9.42 1.53 0.52 1.94 0.48 4.94 6.96 0.45
47.514 62.92 11.12 2.50 0.02 70.72 9.58 0.68 0.49 1.66 0.49 5.01 7.21 0.43
47.516 62.92 11.12 2.49 0.15 68.94 9.48 1.24 0.50 2.02 0.46 4.96 7.22 0.44
47.518 62.92 11.12 2.51 0.08 69.86 9.57 0.84 0.50 1.83 0.45 4.97 7.12 0.45
47.52 62.92 11.12 2.53 0.05 70.54 9.64 0.76 0.50 1.94 0.52 4.97 7.13 0.45
47.522 62.92 11.12 2.53 0.04 70.39 9.58 0.62 0.50 1.88 0.50 5.03 7.32 0.42
47.524 62.92 11.12 2.47 0.05 70.26 9.67 0.60 0.53 1.73 0.48 4.98 7.17 0.44
47.526 62.92 11.12 2.50 0.05 70.67 9.57 0.71 0.53 1.69 0.51 5.03 7.26 0.44

145
 Table 50 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
47.528 62.92 11.12 2.51 0.04 70.12 9.55 0.67 0.50 1.81 0.49 5.04 7.50 0.44
47.53 62.92 11.12 2.49 0.06 69.85 9.51 0.69 0.50 1.82 0.48 4.89 7.46 0.45
47.532 62.92 11.12 2.52 0.03 70.29 9.48 0.75 0.53 1.78 0.50 4.95 7.46 0.47
47.534 62.92 11.12 2.53 0.06 70.57 9.37 0.79 0.49 1.79 0.50 4.98 7.33 0.45
47.536 62.92 11.12 2.51 0.05 70.34 9.44 0.70 0.53 1.71 0.51 4.92 7.59 0.45
47.538 62.92 11.122 2.56 0.06 71.21 9.50 0.80 0.54 1.46 0.51 4.92 6.94 0.44
47.54 62.92 11.12 2.53 0.07 70.98 9.59 0.65 0.54 1.65 0.49 4.91 7.05 0.45
47.542 62.92 11.12 2.53 0.03 70.23 9.66 0.70 0.56 1.60 0.47 4.90 6.86 0.45
47.544 62.92 11.12 2.55 0.06 70.39 9.64 0.73 0.58 1.81 0.45 4.87 6.62 0.44
47.546 62.92 11.12 2.55 0.03 71.00 9.70 0.76 0.56 1.66 0.52 4.89 6.48 0.44
47.548 62.92 11.12 2.56 0.04 71.19 9.84 0.70 0.60 1.65 0.49 4.88 6.44 0.43
47.55 62.92 11.122 2.63 0.01 71.46 9.74 0.63 0.59 1.58 0.51 4.88 6.36 0.46
47.552 62.92 11.12 2.78 0.04 72.40 9.63 0.52 0.56 1.50 0.45 4.80 5.62 0.50
47.554 62.92 11.12 3.04 0.04 73.72 9.70 0.70 0.58 1.32 0.48 4.69 3.41 0.54
47.556 62.92 11.12 2.81 0.01 72.05 9.70 0.68 0.60 1.53 0.46 4.77 4.94 0.50
47.558 62.92 11.12 2.63 0.04 71.79 9.72 0.67 0.58 1.59 0.51 4.81 5.99 0.48
47.56 62.92 11.12 2.58 0.05 72.18 9.79 0.73 0.58 1.63 0.52 4.86 6.09 0.44
47.562 62.92 11.12 2.56 0.04 71.97 9.67 0.72 0.63 1.65 0.51 4.85 6.62 0.45
47.564 62.92 11.122 2.56 0.12 71.41 9.45 1.51 0.53 1.75 0.47 4.78 6.34 0.45
47.566 62.92 11.122 2.57 0.06 72.23 9.49 0.68 0.59 1.71 0.52 4.80 6.31 0.44
47.568 62.92 11.122 2.54 0.13 71.01 9.31 1.62 0.56 1.58 0.52 4.78 6.58 0.44
47.57 62.92 11.122 2.59 0.04 72.32 9.40 0.73 0.56 1.60 0.49 4.89 6.80 0.45
47.572 62.92 11.12 2.55 0.07 72.27 9.50 0.71 0.56 1.67 0.51 4.92 7.00 0.44
47.575 62.92 11.122 2.56 0.07 71.82 9.36 0.77 0.57 1.60 0.51 4.86 7.13 0.44
47.576 62.92 11.12 2.52 0.06 71.41 9.55 0.73 0.59 1.71 0.53 4.81 7.12 0.45
47.578 62.92 11.12 2.54 0.07 71.55 9.65 0.70 0.59 1.83 0.47 4.90 7.06 0.44
47.581 62.92 11.122 2.58 0.05 72.01 9.49 0.71 0.53 1.43 0.51 4.87 6.77 0.45
47.583 62.92 11.12 2.54 0.04 71.98 9.67 0.70 0.56 1.49 0.53 4.94 6.62 0.44
47.584 62.92 11.122 2.59 0.04 72.58 9.67 0.73 0.60 1.53 0.52 4.93 6.52 0.44
47.586 62.92 11.12 2.57 0.06 71.81 9.75 0.73 0.56 1.72 0.53 4.88 6.34 0.44
47.589 62.92 11.12 2.51 0.03 71.69 9.70 0.73 0.56 1.87 0.54 4.85 6.48 0.45
47.59 62.92 11.12 2.53 0.06 71.59 9.75 0.72 0.59 1.72 0.48 4.87 6.30 0.44
47.592 62.92 11.12 2.55 0.04 71.92 9.71 0.70 0.58 1.80 0.49 4.85 6.63 0.45
47.595 62.92 11.122 2.59 0.07 72.38 9.59 0.73 0.59 1.89 0.50 4.87 5.89 0.44
47.597 62.92 11.122 2.59 0.06 72.64 9.61 0.67 0.57 1.69 0.48 4.86 6.31 0.47
47.599 62.92 11.12 2.55 0.07 72.30 9.71 0.71 0.60 1.75 0.46 4.87 6.43 0.45
47.601 62.92 11.12 2.56 0.05 72.04 9.63 0.69 0.60 1.93 0.50 4.84 6.27 0.45
47.603 62.92 11.12 2.59 0.05 71.86 9.66 0.70 0.59 1.89 0.48 4.84 6.18 0.44
47.605 62.92 11.12 2.61 0.03 72.12 9.66 0.72 0.58 1.71 0.50 4.76 6.32 0.46
47.607 62.92 11.12 2.67 0.03 72.16 9.46 0.70 0.60 1.84 0.51 4.82 6.03 0.46
47.609 62.92 11.12 2.94 0.05 73.96 9.42 0.68 0.60 1.53 0.49 4.74 4.29 0.53
47.611 62.92 11.122 2.89 0.00 74.09 9.34 0.54 0.60 1.78 0.50 4.77 5.02 0.50
47.613 62.92 11.12 2.61 0.05 72.24 9.62 0.62 0.63 1.85 0.50 4.82 6.00 0.46
47.615 62.92 11.122 2.59 0.06 72.34 9.66 0.71 0.68 1.93 0.55 4.79 5.97 0.42
47.617 62.92 11.12 2.54 0.05 71.93 9.80 0.69 0.69 1.79 0.55 4.78 6.14 0.44
47.619 62.92 11.122 2.58 0.05 72.55 9.75 0.75 0.75 1.83 0.54 4.78 5.99 0.43
47.621 62.92 11.122 2.56 0.04 72.55 9.75 0.71 0.76 1.68 0.57 4.80 6.20 0.43
47.623 62.92 11.122 2.58 0.03 73.00 9.73 0.68 0.79 1.56 0.63 4.84 5.74 0.43
47.625 62.92 11.122 2.49 0.11 72.13 9.91 0.94 0.82 1.77 0.66 4.80 5.78 0.42
47.627 62.92 11.122 2.52 0.03 72.23 9.88 0.65 0.84 1.84 0.66 4.80 5.73 0.42
47.629 62.92 11.122 2.20 0.85 65.33 9.14 6.58 0.82 3.24 0.67 4.42 5.25 0.41
47.631 62.92 11.12 2.37 0.28 69.64 9.63 2.21 0.94 2.09 0.71 4.71 6.06 0.41
47.633 62.92 11.12 2.43 0.04 71.41 9.76 0.64 1.02 1.71 0.77 4.75 5.85 0.42
47.635 62.92 11.12 2.39 0.04 71.52 9.85 0.60 1.11 1.77 0.83 4.75 6.36 0.41
47.637 62.92 11.122 2.39 0.02 71.73 9.76 0.63 1.19 1.74 0.90 4.82 6.03 0.40
47.639 62.92 11.122 2.35 0.07 70.54 9.76 1.02 1.29 1.78 0.93 4.67 5.56 0.36
47.641 62.92 11.122 2.35 0.08 70.80 9.80 0.94 1.37 1.76 1.01 4.67 5.59 0.38
47.643 62.92 11.122 2.32 0.02 71.37 9.95 0.59 1.52 1.85 1.04 4.73 6.19 0.38
47.645 62.92 11.122 2.28 0.04 71.07 9.95 0.53 1.56 1.70 1.12 4.68 6.19 0.39
47.647 62.92 11.122 2.26 0.05 70.75 10.12 0.56 1.73 1.67 1.16 4.74 5.97 0.38

146
 Table 50 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
47.649 62.92 11.122 2.23 0.03 70.37 10.16 0.57 1.85 1.45 1.25 4.68 6.13 0.34
47.651 62.92 11.122 2.17 0.03 69.91 10.27 0.53 1.95 1.83 1.36 4.65 5.98 0.35
47.653 62.92 11.122 2.13 0.03 69.66 10.41 0.45 2.15 1.59 1.46 4.61 6.15 0.36
47.655 62.92 11.122 2.06 0.14 68.70 10.41 1.81 2.30 1.67 1.52 4.51 5.73 0.33
47.657 62.92 11.122 2.05 0.02 69.91 10.83 0.43 2.62 1.28 1.77 4.55 5.95 0.31
47.659 62.92 11.122 2.01 0.00 69.86 11.13 0.43 2.82 1.36 1.84 4.59 5.33 0.32
47.661 62.92 11.122 2.00 0.03 69.28 11.27 0.43 3.01 1.30 1.97 4.53 5.38 0.29
47.663 62.92 11.122 1.93 0.00 68.83 11.41 0.37 3.17 1.35 2.09 4.45 5.33 0.29
47.665 62.92 11.122 1.90 0.00 68.45 11.53 0.33 3.36 1.40 2.18 4.37 5.02 0.28
47.667 62.92 11.122 1.83 0.01 68.75 11.70 0.28 3.56 1.26 2.34 4.33 4.99 0.27
47.669 62.92 11.122 1.79 0.02 68.58 11.74 0.28 3.84 1.40 2.47 4.40 4.86 0.25
47.671 62.92 11.122 1.75 0.00 68.48 11.92 0.19 4.11 1.48 2.63 4.32 4.56 0.24
47.673 62.92 11.12 1.52 0.43 63.72 11.60 5.46 4.07 1.67 2.64 3.99 4.04 0.25
47.675 62.92 11.12 0.96 1.88 47.72 9.21 23.85 3.62 2.96 2.34 3.00 2.84 0.33
47.677 62.92 11.122 1.06 1.74 51.00 9.87 20.56 4.08 2.20 2.61 3.15 3.14 0.40
47.679 62.92 11.122 0.99 1.66 50.60 10.18 20.07 4.48 2.03 2.90 3.13 3.29 0.40
47.681 62.92 11.122 1.36 0.00 65.94 12.92 0.34 5.85 1.35 3.62 4.01 4.10 0.18
47.683 62.92 11.12 1.30 0.02 64.65 13.10 0.33 6.15 1.27 3.84 3.94 4.03 0.17
47.685 62.92 11.122 1.26 0.00 64.76 13.35 0.10 6.63 1.24 4.04 3.96 3.94 0.15
47.687 62.92 11.122 1.21 0.03 63.69 13.63 0.24 6.82 1.16 4.17 3.84 3.67 0.15
47.689 62.92 11.122 1.09 0.40 60.40 13.37 3.19 6.98 1.46 4.18 3.66 3.20 0.19
47.691 62.92 11.12 1.06 0.22 60.67 13.85 2.05 7.32 0.90 4.43 3.65 3.41 0.18
47.693 62.92 11.122 0.87 1.37 53.20 12.39 10.50 6.73 1.83 4.07 3.15 2.67 0.34
47.695 62.92 11.122 0.80 1.73 51.11 12.36 12.66 7.03 1.83 4.20 3.02 2.58 0.36
47.697 62.92 11.12 0.90 0.43 58.27 14.44 3.23 8.38 0.92 4.93 3.42 3.03 0.24
47.699 62.92 11.12 0.87 0.07 59.55 15.03 0.80 9.04 0.71 5.33 3.38 2.67 0.12
47.701 62.92 11.12 0.85 0.01 59.91 15.41 0.10 9.44 0.52 5.46 3.38 2.48 0.10
47.703 62.92 11.122 0.81 0.00 59.78 15.53 0.17 9.80 0.65 5.50 3.33 2.19 0.10
47.705 62.92 11.12 0.78 0.00 59.52 15.85 0.01 9.94 0.68 5.67 3.25 1.87 0.08
47.707 62.92 11.12 0.74 0.00 59.02 15.93 0.02 10.30 0.58 5.86 3.14 1.90 0.09
47.709 62.92 11.122 0.72 0.00 58.68 16.07 0.00 10.84 0.73 5.95 3.08 1.72 0.06
47.711 62.92 11.122 0.67 0.00 57.97 16.20 0.02 11.17 0.65 6.06 3.00 1.59 0.07
47.713 62.92 11.12 0.63 0.00 57.60 16.39 0.01 11.34 0.74 6.17 2.93 1.53 0.06
47.715 62.92 11.122 0.60 0.01 57.29 16.53 0.00 11.80 0.67 6.41 2.81 1.37 0.05
47.717 62.92 11.12 0.56 0.00 56.56 16.79 0.01 12.11 0.72 6.40 2.73 1.70 0.06
47.719 62.92 11.122 0.55 0.00 56.47 16.90 0.00 12.41 0.77 6.58 2.61 1.55 0.04
47.721 62.92 11.122 0.52 0.00 55.53 16.94 0.00 12.80 0.62 6.69 2.53 1.50 0.04
47.723 62.92 11.122 0.49 0.00 55.37 16.98 0.01 12.98 0.44 6.70 2.48 1.36 0.04
47.725 62.92 11.122 0.47 0.00 55.51 16.98 0.00 13.17 0.57 6.86 2.41 1.35 0.03
47.727 62.92 11.122 0.45 0.01 54.70 17.13 0.02 13.44 0.45 6.95 2.35 1.28 0.02
47.729 62.92 11.12 0.42 0.01 54.12 17.23 0.02 13.57 0.30 6.97 2.23 1.30 0.02
47.731 62.92 11.122 0.44 0.00 54.30 17.22 0.01 13.75 0.41 6.95 2.13 1.23 0.02
47.733 62.92 11.12 0.40 0.00 53.15 17.62 0.01 14.16 0.34 7.17 2.09 1.27 0.02
47.735 62.92 11.118 0.36 0.00 52.29 17.92 0.03 14.33 0.36 7.35 1.99 1.04 0.01
47.737 62.92 11.116 0.34 0.00 51.72 18.34 0.01 14.48 0.47 7.41 1.91 1.17 0.02
47.739 62.92 11.118 0.33 0.00 51.83 18.50 0.01 14.83 0.41 7.58 1.84 1.20 0.01
47.741 62.92 11.116 0.32 0.00 51.16 18.69 0.03 15.02 0.57 7.54 1.78 0.94 0.01
47.743 62.92 11.12 0.32 0.01 51.89 18.44 0.01 15.38 0.30 7.65 1.82 0.83 0.01
47.745 62.92 11.12 0.30 0.00 51.65 18.87 0.00 15.46 0.32 7.83 1.71 0.86 0.00
47.747 62.92 11.12 0.29 0.00 50.87 18.79 0.00 15.81 0.22 7.82 1.64 0.87 0.01
47.749 62.92 11.12 0.27 0.00 50.58 18.82 0.01 16.00 0.31 7.85 1.63 0.73 0.00
47.751 62.92 11.122 0.26 0.00 50.84 18.81 0.00 16.13 0.44 7.90 1.59 0.95 0.01
47.753 62.92 11.12 0.24 0.00 50.47 19.01 0.00 16.25 0.38 7.94 1.55 0.34 0.01
47.755 62.92 11.122 0.24 0.00 50.36 19.07 0.00 16.48 0.57 8.06 1.57 0.64 0.00
47.757 62.92 11.122 0.23 0.00 50.17 18.92 0.00 16.59 0.48 8.21 1.48 0.51 0.01
47.759 62.92 11.122 0.22 0.00 49.50 18.92 0.00 16.69 0.42 8.12 1.46 0.54 0.01
47.761 62.92 11.122 0.21 0.00 49.52 19.18 0.00 16.90 0.62 8.19 1.44 0.58 0.00
47.763 62.92 11.122 0.21 0.00 49.53 19.26 0.00 16.80 0.62 8.21 1.44 0.73 0.00
47.765 62.92 11.122 0.20 0.00 49.41 19.31 0.00 17.08 0.53 8.27 1.40 0.47 0.01
47.767 62.92 11.12 0.19 0.00 48.90 19.61 0.02 17.10 0.69 8.28 1.39 0.66 0.00

147
 Table 50 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
47.769 62.92 11.122 0.19 0.00 49.01 19.67 0.02 17.18 0.61 8.27 1.36 0.75 0.00
47.771 62.92 11.122 0.18 0.00 48.63 19.96 0.01 17.28 0.48 8.37 1.36 0.78 0.00
47.773 62.92 11.122 0.19 0.00 48.29 19.94 0.03 17.50 0.57 8.48 1.36 0.67 0.00
47.775 62.92 11.12 0.18 0.01 48.03 20.10 0.00 17.38 0.47 8.54 1.36 0.70 0.00
47.777 62.92 11.122 0.18 0.01 48.33 20.13 0.00 17.40 0.52 8.53 1.32 0.80 0.00
47.779 62.92 11.122 0.17 0.00 47.98 20.33 0.00 17.56 0.57 8.52 1.29 0.94 0.00
47.781 62.92 11.122 0.17 0.00 47.70 20.49 0.02 17.63 0.45 8.49 1.31 0.85 0.00
47.783 62.92 11.122 0.17 0.01 47.60 20.26 0.02 17.58 0.45 8.53 1.27 0.69 0.00

Figure 157(a) – Secondary electron

image of the Ni-Cr-Zr alloy (after an
extended brazing cycle).

Figure 157(b) – Enlarged view of the

secondary electron image of the Ni-Cr-Zr
braze, shown in Figure 157(a), highlighting
the location of four spot chemical analyses
of the intermetallic compound within the braze.

Table 51 - Quantitative EPMA data (taken at 15 keV and 80 nA with a focused spot beam) at
four locations within the intermetallic phase (indicated in Figure 157(b)). (Percentage by
weight).
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti Si
Phase 1 0.18 1.60 66.77 0.60 27.87 0.06 1.39 0.05 1.83 0.35 0.08 0.16
Phase 2 0.19 1.43 66.83 0.77 27.97 0.08 1.33 0.07 1.81 0.07 0.14 0.24
Phase 3 0.17 1.71 66.22 0.61 27.65 0.06 1.81 0.09 1.85 0.45 0.10 0.17
Phase 4 3.16 0.04 75.05 9.13 0.63 0.54 1.58 0.42 4.53 3.35 0.55 0.19

148
In order to study the distribution of various elements within the Ni-Cr-Hf braze
microstructure, EMPA maps were constructed. These maps are shown in Figure 158 (for Al,
C, Ni and Cr), Figure 159 (for Zr, Hf and Si), Figure 160 (for Ru, Ta, Ti and W) and Figure
161 (for Fe and Mo).

Figure 158 – EMPA maps displaying the distribution of Al, C, Ni and Cr within the
MarM247/Ni-Cr-Zr braze microstructure (after an extended diffusion cycle).

The EMPA results shown in Figure 158 indicate that Al and Cr partitioned strongly to the
MarM247 powder particles and to one of the eutectic phases (most likely the Ni-rich  phase)
during the brazing and diffusion cycles. Both elements appeared to have limited solubility
within the intermetallic phase in the eutectic component. The majority of the carbon within
the structure was concentrated within small particles, probably carbides or pores saturated
with C during the carbon coating process. Ni appeared to be fairly uniformly distributed
between the braze alloy and the matrix.

149
 Figure 159 – EMPA maps displaying the distribution of Zr and Hf within the MarM247/Ni-
Cr-Zr braze microstructure (after an extended diffusion cycle).

The results shown in Figure 159 suggest that Zr partitioned strongly to the intermetallic phase
within the braze alloy, which substantiates the results of the line and spot scans shown in
Tables 50 and 51. The Zr EMPA map confirms that the intermetallic phase within the braze
joint was Zr-rich, whereas Zr was almost completely absent from the MarM247 matrix
particles. The Hf detected within the intermetallic compound in the Ni-Cr-Zr alloy after
brazing (Figure 153) was not evident after the extended diffusion cycle, suggesting that the
Hf diffused from the braze alloy into the MarM247 particles during heat treatment.

150
 Figure 160 – EMPA maps displaying the distribution of Si, Ta, Ti and W within the
MarM247/Ni-Cr-Zr braze microstructure (after an extended brazing cycle).

As shown in Figure 160, Si, Ta and Ti were distributed fairly evenly between the braze alloy
and the matrix. Si apparently partitioned weakly to the intermetallic phase (suggesting
limited solubility within the intermetallic compound), whereas Ti partitioned preferentially to
the MarM247 powder particles. Tungsten, however, partitioned strongly to the MarM247
powder particles, suggesting low solubility within the intermetallic phase. The W appeared to
be distributed evenly through the matrix phase.

The EMPA maps for Fe and Mo, shown in Figure 161, display weak partitioning to the
matrix particles, but due to the low levels of these elements within the MarM247 particles and
the braze alloy, this tendency was weak.

151
Figure 161 – EMPA maps displaying the distribution of Fe and Mo within the MarM247/Ni-
Cr-Zr braze microstructure (as-brazed condition).

11.3.4 Repeat: The MarM247/Ni-Zr braze joint produced using Ni-7Cr-13Zr braze filler
(after an extended diffusion cycle):

Secondary and backscatter electron images of the MarM247/Ni-Zr joint produced using the
experimental Ni-7Cr-13Zr braze filler metal are shown in Figure 162 after an extended
diffusion cycle. The arrow highlights the location and direction of the 2 m stepped EPMA
line scan. The results of the EMPA scan are shown in Table 52.

(a) (b)

Figure 162 – Secondary (a) and backscatter (b) electron images of the MarM247/Ni-7Cr-
13Zr joint after an extended diffusion cycle (adjacent to the interface).

152
 Table 52 – The results of the second EMPA scan of the MarM247/Ni-7Cr-13Zr braze joint
after an extended diffusion cycle. (All compositions given in wt.%).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
7.431 70.131 11.255 2.18 0.00 69.20 13.26 0.58 0.51 1.58 1.51 5.67 2.11 1.86
7.433 70.131 11.255 2.19 0.01 69.33 13.29 0.60 0.53 1.53 1.48 5.59 2.12 1.87
7.435 70.131 11.255 2.23 0.00 69.32 13.10 0.59 0.56 1.55 1.50 5.54 2.34 1.91
7.437 70.131 11.255 2.09 0.00 68.91 11.79 3.46 0.46 1.58 1.33 5.20 1.85 1.88
7.439 70.131 11.255 1.05 0.01 66.61 5.95 17.02 0.27 1.42 0.71 3.54 0.68 1.27
7.441 70.131 11.255 1.30 0.00 67.61 6.59 14.95 0.28 1.44 0.74 3.78 0.91 1.38
7.443 70.131 11.255 1.30 0.00 67.38 7.03 14.15 0.31 1.26 0.86 3.91 0.84 1.42
7.445 70.131 11.255 0.95 0.00 66.98 5.24 18.23 0.25 1.37 0.63 3.36 0.63 1.17
7.447 70.131 11.255 1.06 0.00 67.79 5.53 17.26 0.24 1.45 0.69 3.50 0.76 1.25
7.449 70.131 11.255 1.07 0.00 67.26 5.56 17.39 0.26 1.65 0.70 3.41 0.84 1.22
7.451 70.131 11.255 1.07 0.00 66.95 5.92 16.81 0.27 1.52 0.74 3.55 0.58 1.27
7.453 70.131 11.255 1.57 0.00 68.01 8.08 11.77 0.36 1.40 0.93 4.20 1.12 1.53
7.455 70.131 11.255 1.58 0.00 67.85 8.12 11.70 0.34 1.53 0.94 4.14 0.85 1.52
7.457 70.131 11.255 1.30 0.00 67.89 6.92 14.22 0.30 1.46 0.83 3.88 0.83 1.38
7.459 70.131 11.255 1.05 0.00 67.72 5.64 17.06 0.25 1.49 0.67 3.49 0.39 1.22
7.461 70.131 11.255 1.00 0.01 67.36 4.94 18.11 0.23 1.41 0.65 3.38 0.76 1.14
7.463 70.131 11.255 1.41 0.00 67.60 7.11 13.29 0.31 1.24 0.86 3.96 0.88 1.44
7.465 70.131 11.255 1.52 0.02 67.97 8.18 10.93 0.38 1.42 0.98 4.24 1.08 1.55
7.467 70.131 11.255 0.69 0.00 66.72 3.79 21.20 0.19 1.57 0.55 2.95 0.85 1.00
7.469 70.131 11.255 1.55 0.00 68.11 7.74 11.47 0.36 1.32 0.99 4.20 1.37 1.54
7.471 70.131 11.255 2.32 0.01 69.87 12.24 1.20 0.51 1.33 1.44 5.42 1.98 1.95
7.473 70.131 11.255 2.27 0.01 69.81 12.58 0.57 0.53 1.47 1.49 5.46 2.13 1.91
7.475 70.131 11.255 2.22 0.01 69.35 12.89 0.60 0.54 1.43 1.47 5.45 2.43 1.88
7.477 70.131 11.255 2.21 0.01 69.15 13.04 0.62 0.54 1.50 1.52 5.60 2.27 1.84
7.479 70.131 11.255 2.20 0.00 69.56 13.12 0.59 0.53 1.26 1.51 5.64 2.14 1.84
7.481 70.131 11.255 2.22 0.00 69.29 13.12 0.58 0.52 1.45 1.50 5.59 2.23 1.85
7.483 70.131 11.255 2.21 0.03 70.13 13.20 0.59 0.51 1.28 1.49 5.67 2.23 1.85
7.485 70.131 11.255 2.24 0.01 70.30 13.18 0.57 0.53 1.30 1.48 5.71 2.00 1.86
7.487 70.131 11.255 2.23 0.00 69.98 13.26 0.59 0.54 1.16 1.48 5.69 2.05 1.85
7.489 70.131 11.255 2.23 0.00 69.46 13.29 0.58 0.56 1.25 1.54 5.75 1.93 1.80
7.491 70.131 11.255 2.22 0.00 69.70 13.22 0.59 0.51 1.34 1.49 5.71 2.17 1.85
7.493 70.131 11.255 2.22 0.00 69.99 13.13 0.57 0.51 1.25 1.49 5.71 2.16 1.84
7.495 70.131 11.255 2.23 0.00 69.97 13.07 0.58 0.57 1.51 1.55 5.76 2.30 1.84
7.497 70.131 11.255 2.23 0.00 69.84 12.93 0.52 0.53 1.24 1.53 5.77 2.17 1.85
7.499 70.131 11.255 2.21 0.00 69.72 12.84 0.56 0.56 1.25 1.51 5.76 2.15 1.84
7.501 70.131 11.255 2.23 0.00 69.60 12.89 0.57 0.53 1.06 1.53 5.70 2.29 1.85
7.503 70.131 11.255 2.23 0.00 70.00 12.86 0.55 0.53 1.35 1.48 5.74 2.15 1.82
7.505 70.131 11.255 2.23 0.02 70.03 12.83 0.57 0.54 1.49 1.50 5.73 2.10 1.83
7.507 70.131 11.255 2.24 0.00 70.11 12.65 0.60 0.51 1.29 1.49 5.68 2.02 1.83
7.509 70.131 11.255 2.23 0.00 70.00 12.72 0.59 0.51 1.41 1.53 5.71 2.13 1.85
7.511 70.131 11.255 2.25 0.00 70.08 12.79 0.61 0.53 1.39 1.49 5.71 2.11 1.88
7.513 70.131 11.255 2.26 0.00 69.53 12.77 0.57 0.52 1.27 1.51 5.55 2.33 1.87
7.515 70.131 11.255 2.30 0.00 69.88 12.09 1.47 0.50 1.24 1.36 5.43 1.87 1.98
7.517 70.131 11.255 0.90 0.00 39.93 5.86 37.02 0.25 5.56 0.75 2.73 0.61 1.44
7.519 70.131 11.255 1.92 0.03 66.82 10.97 7.59 0.46 1.47 1.30 4.89 1.43 1.77
7.521 70.131 11.255 2.25 0.02 69.54 13.11 0.60 0.55 1.43 1.51 5.60 2.19 1.93
7.523 70.131 11.255 2.24 0.01 69.97 13.17 0.58 0.54 1.35 1.56 5.62 1.82 1.86
7.525 70.131 11.255 2.25 0.02 69.52 13.23 0.64 0.53 1.17 1.53 5.63 2.10 1.84
7.527 70.131 11.255 2.23 0.00 69.58 13.17 0.52 0.51 1.39 1.51 5.68 2.04 1.82
7.529 70.131 11.255 2.23 0.00 69.93 13.23 0.59 0.55 1.38 1.50 5.75 1.81 1.83
7.531 70.131 11.255 2.24 0.00 69.92 13.18 0.59 0.51 1.41 1.50 5.77 2.08 1.84
7.533 70.131 11.255 2.21 0.00 69.64 13.11 0.60 0.54 1.38 1.49 5.77 1.96 1.84
7.535 70.131 11.255 2.25 0.02 69.46 13.15 0.58 0.51 1.48 1.54 5.87 1.84 1.84
7.537 70.131 11.255 2.29 0.00 69.13 13.09 0.54 0.51 1.74 1.51 5.76 1.65 1.84
7.539 70.131 11.255 2.17 0.04 65.92 12.74 0.50 0.49 1.55 1.27 5.49 1.71 1.71
7.541 70.131 11.255 2.11 0.01 66.78 12.76 0.56 0.51 1.73 1.35 5.55 1.85 1.75
7.543 70.131 11.255 2.16 0.00 68.70 13.06 0.94 0.51 1.67 1.50 5.68 1.98 1.84
7.545 70.131 11.255 2.21 0.00 69.66 13.27 0.54 0.51 1.48 1.50 5.87 2.02 1.85
7.547 70.131 11.255 2.21 0.00 69.70 13.47 0.58 0.52 1.49 1.51 5.90 2.13 1.86

153
 Table 52 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
7.549 70.131 11.255 2.22 0.01 69.54 13.37 0.59 0.51 1.41 1.53 5.86 1.99 1.87
7.551 70.131 11.255 2.22 0.00 69.15 13.41 0.58 0.52 1.56 1.54 5.84 1.88 1.86
7.553 70.131 11.252 2.19 0.00 68.86 13.56 0.56 0.51 1.48 1.58 5.84 1.85 1.84
7.555 70.131 11.255 2.21 0.00 69.48 13.45 0.56 0.53 1.52 1.58 5.88 2.18 1.82
7.557 70.131 11.255 2.20 0.00 69.67 13.55 0.54 0.50 1.54 1.51 5.89 2.14 1.85
7.559 70.131 11.255 2.21 0.00 69.37 13.57 0.56 0.51 1.36 1.54 5.96 2.19 1.85
7.561 70.131 11.255 2.21 0.02 69.59 13.56 0.55 0.52 1.48 1.56 5.90 2.22 1.87
7.563 70.131 11.255 2.22 0.00 69.11 13.62 0.53 0.50 1.38 1.57 5.86 2.03 1.83
7.565 70.131 11.255 2.23 0.00 69.26 13.54 0.55 0.52 1.58 1.54 5.89 1.87 1.87
7.567 70.131 11.255 2.22 0.00 69.61 13.72 0.53 0.49 1.69 1.54 5.93 1.79 1.83
7.569 70.131 11.255 2.20 0.00 69.50 13.50 0.55 0.53 1.61 1.57 5.93 2.09 1.83
7.571 70.131 11.255 2.23 0.02 69.38 13.59 0.54 0.53 1.57 1.54 5.92 1.98 1.81
7.573 70.131 11.255 2.22 0.00 69.32 13.51 0.56 0.51 1.62 1.53 5.87 2.16 1.82
7.575 70.131 11.255 2.23 0.00 69.45 13.44 0.51 0.50 1.50 1.50 5.88 1.93 1.86
7.577 70.131 11.255 2.24 0.00 69.61 13.50 0.56 0.52 1.50 1.56 5.82 1.86 1.83
7.579 70.131 11.255 2.21 0.02 69.94 13.50 0.56 0.54 1.46 1.51 5.87 2.33 1.84
7.581 70.131 11.255 2.22 0.00 70.09 13.40 0.57 0.55 1.70 1.46 5.76 2.06 1.82
7.583 70.131 11.255 2.22 0.00 69.80 13.44 0.52 0.51 1.60 1.51 5.72 1.96 1.86
7.585 70.131 11.255 2.22 0.00 70.03 13.28 0.56 0.55 1.37 1.52 5.69 2.12 1.84
7.587 70.131 11.255 2.21 0.00 69.73 13.42 0.55 0.54 1.50 1.54 5.74 2.53 1.83
7.589 70.131 11.255 2.22 0.00 69.89 13.29 0.57 0.54 1.07 1.57 5.61 2.43 1.83
7.591 70.131 11.255 2.24 0.00 69.93 13.34 0.55 0.55 1.37 1.49 5.63 2.07 1.88
7.593 70.131 11.255 2.23 0.00 70.27 13.34 0.56 0.57 1.45 1.58 5.65 1.92 1.84
7.595 70.131 11.255 2.22 0.00 70.06 13.32 0.55 0.57 1.34 1.52 5.56 2.19 1.85
7.597 70.131 11.255 2.09 0.01 69.85 12.24 3.08 0.52 1.32 1.44 5.21 1.85 1.85
7.599 70.131 11.255 2.19 0.00 70.21 12.83 1.71 0.53 1.45 1.40 5.35 1.87 1.90
7.601 70.131 11.255 2.25 0.00 70.33 13.39 0.59 0.59 1.50 1.51 5.50 1.75 1.89
7.603 70.131 11.255 2.25 0.00 70.38 13.41 0.58 0.57 1.46 1.47 5.59 1.83 1.85
7.605 70.131 11.255 2.23 0.00 70.32 13.60 0.57 0.57 1.40 1.49 5.51 2.25 1.84
7.607 70.131 11.255 2.25 0.00 70.42 13.39 0.59 0.58 1.53 1.51 5.63 1.74 1.86
7.609 70.131 11.255 2.22 0.01 70.14 13.42 0.52 0.59 1.29 1.51 5.55 2.06 1.85
7.611 70.131 11.255 2.23 0.00 70.12 13.31 0.55 0.58 1.44 1.52 5.61 2.17 1.82
7.613 70.131 11.255 2.21 0.01 70.17 13.37 0.57 0.60 1.39 1.50 5.47 2.05 1.85
7.615 70.131 11.255 2.22 0.00 69.97 13.17 0.57 0.64 1.67 1.54 5.54 2.20 1.84
7.617 70.131 11.255 2.22 0.01 70.34 13.20 0.52 0.65 1.35 1.55 5.59 2.26 1.81
7.619 70.131 11.255 2.19 0.00 69.74 13.22 0.57 0.67 1.52 1.59 5.57 2.03 1.80
7.621 70.131 11.255 2.21 0.00 69.90 13.16 0.55 0.70 1.44 1.60 5.57 2.23 1.79
7.623 70.131 11.255 2.20 0.00 69.75 13.22 0.50 0.77 1.46 1.62 5.63 2.51 1.78
7.625 70.131 11.255 2.18 0.00 69.29 13.43 0.52 0.87 1.48 1.71 5.68 2.40 1.76
7.627 70.131 11.255 2.17 0.00 69.55 13.41 0.55 0.94 1.17 1.72 5.75 2.05 1.71
7.629 70.131 11.255 2.13 0.01 69.17 13.54 0.51 1.08 1.12 1.73 5.70 2.48 1.69
7.631 70.131 11.255 2.12 0.00 68.82 13.62 0.46 1.17 1.29 1.83 5.80 2.58 1.63
7.633 70.131 11.255 2.07 0.00 68.36 13.83 0.44 1.29 1.10 1.96 5.81 2.53 1.61
7.635 70.131 11.255 2.05 0.00 68.08 13.95 0.39 1.44 1.08 2.05 5.81 2.13 1.54
7.637 70.131 11.255 2.04 0.00 67.86 14.10 0.40 1.62 1.04 2.09 5.89 2.42 1.53
7.639 70.131 11.255 2.00 0.00 67.59 14.10 0.32 1.76 1.24 2.20 5.87 2.16 1.48
7.641 70.131 11.252 1.96 0.00 67.14 14.24 0.33 1.88 1.14 2.28 5.85 2.35 1.46
7.643 70.131 11.255 1.97 0.00 67.27 14.15 0.33 2.05 1.07 2.40 5.85 2.30 1.43
7.645 70.131 11.255 1.93 0.01 66.73 14.35 0.28 2.16 1.09 2.44 5.82 2.20 1.38
7.647 70.131 11.255 1.91 0.00 66.30 14.41 0.24 2.40 1.08 2.60 5.84 2.25 1.33
7.649 70.131 11.255 1.87 0.00 66.22 14.63 0.24 2.64 1.00 2.75 5.93 2.05 1.27
7.652 70.131 11.255 1.85 0.00 66.12 14.72 0.23 2.77 0.93 2.81 5.85 1.91 1.20
7.654 70.131 11.255 1.82 0.00 65.55 14.77 0.21 3.04 1.08 2.87 5.83 2.17 1.18
7.655 70.131 11.255 1.45 0.00 55.42 12.92 10.59 2.88 2.90 2.67 4.89 1.57 1.63
7.657 70.131 11.255 1.26 0.01 51.72 12.41 13.76 2.97 2.88 2.69 4.65 1.59 1.65
7.66 70.131 11.255 1.68 0.01 64.13 15.24 0.13 3.86 1.14 3.30 5.73 1.77 1.01
7.661 70.131 11.255 1.62 0.00 62.95 14.98 1.42 4.00 1.19 3.40 5.63 2.11 1.06
7.663 70.131 11.255 1.39 0.00 56.23 13.64 8.52 3.82 1.99 3.18 4.92 1.54 1.73
7.666 70.131 11.255 1.55 0.00 62.67 15.42 0.66 4.53 1.09 3.62 5.57 2.06 0.94
7.668 70.131 11.255 1.55 0.00 62.77 15.90 0.07 4.92 1.00 3.83 5.59 1.90 0.81

154
 Table 52 – (continued).
X Y Z
Al Hf Ni Cr Zr Fe Ta Mo Co W Ti
(mm) (mm) (mm)
7.669 70.131 11.255 1.47 0.00 62.37 15.96 0.08 5.18 1.09 3.86 5.48 1.91 0.77
7.672 70.131 11.252 1.43 0.00 61.43 16.29 0.08 5.48 0.78 4.11 5.43 2.08 0.71
7.674 70.131 11.252 1.39 0.00 61.25 16.45 0.04 5.80 0.76 4.22 5.37 2.12 0.69
7.676 70.131 11.255 1.33 0.00 61.32 16.51 0.15 6.20 0.97 4.36 5.26 1.72 1.04
7.678 70.131 11.255 1.28 0.01 60.53 16.69 0.12 6.59 0.81 4.53 5.17 1.67 0.93
7.68 70.131 11.255 1.25 0.00 60.40 17.02 0.01 6.96 0.73 4.70 5.13 1.41 0.71
7.682 70.131 11.255 1.18 0.02 58.17 16.79 0.29 6.92 1.04 4.71 4.90 1.49 2.04
7.684 70.131 11.255 1.09 0.05 54.96 16.05 0.73 6.96 1.43 4.65 4.65 1.35 4.74
7.686 70.131 11.255 0.88 0.18 46.92 14.11 1.79 6.29 2.10 4.22 3.90 1.21 11.56
7.688 70.131 11.255 0.92 0.14 49.69 15.23 1.21 6.98 2.08 4.63 4.09 1.26 8.26
7.69 70.131 11.255 1.06 0.00 57.96 17.33 0.13 8.35 0.81 5.39 4.63 1.10 0.94
7.692 70.131 11.252 0.86 0.12 50.06 15.78 1.21 7.79 1.49 5.00 4.02 1.13 7.39
7.694 70.131 11.252 0.88 0.07 52.64 16.84 0.76 8.45 1.06 5.19 4.13 0.96 5.49
7.696 70.131 11.255 0.79 0.13 47.59 15.51 1.50 8.11 1.43 4.89 3.65 1.13 9.98
7.698 70.131 11.255 0.90 0.00 56.24 18.30 0.03 10.05 0.76 5.98 4.11 0.93 0.26
7.7 70.131 11.252 0.85 0.00 56.06 18.38 0.00 10.32 0.85 6.06 4.05 1.11 0.20
7.702 70.131 11.255 0.83 0.00 56.23 18.40 0.08 10.67 0.80 6.28 3.97 1.31 0.22
7.704 70.131 11.252 0.80 0.00 55.11 18.69 0.04 10.83 0.62 6.42 3.77 1.50 0.28
7.706 70.131 11.255 0.77 0.00 55.24 18.74 0.00 11.33 0.71 6.55 3.68 1.47 0.13
7.708 70.131 11.252 0.73 0.00 54.73 19.09 0.00 11.46 0.69 6.76 3.49 1.09 0.12
7.71 70.131 11.255 0.70 0.00 54.72 19.21 0.00 12.08 0.54 6.83 3.36 1.19 0.11
7.712 70.131 11.255 0.68 0.01 54.07 19.19 0.02 12.33 0.58 6.97 3.23 1.30 0.11
7.714 70.131 11.255 0.65 0.01 54.06 19.32 0.00 12.54 0.62 6.99 3.13 1.11 0.14
7.716 70.131 11.252 0.61 0.00 53.25 19.58 0.01 12.92 0.63 7.11 2.88 1.30 0.08
7.718 70.131 11.255 0.58 0.00 52.97 19.67 0.01 13.33 0.59 7.24 2.81 1.09 0.08
7.72 70.131 11.255 0.57 0.00 52.94 19.77 0.02 13.55 0.47 7.30 2.69 1.04 0.06
7.722 70.131 11.252 0.53 0.00 52.30 19.84 0.01 13.67 0.44 7.35 2.60 1.07 0.07
7.724 70.131 11.255 0.53 0.00 52.42 19.69 0.02 14.06 0.41 7.41 2.46 0.90 0.08
7.726 70.131 11.255 0.50 0.00 52.12 19.82 0.00 14.21 0.33 7.50 2.38 1.16 0.05
7.728 70.131 11.255 0.48 0.01 51.69 19.89 0.02 14.45 0.60 7.57 2.26 0.98 0.05
7.73 70.131 11.255 0.47 0.01 51.41 19.97 0.00 14.73 0.42 7.64 2.15 0.97 0.05
7.732 70.131 11.255 0.45 0.00 51.49 20.22 0.02 14.94 0.31 7.73 2.10 1.11 0.04
7.734 70.131 11.255 0.41 0.00 51.05 20.24 0.00 15.22 0.30 7.73 1.97 0.96 0.04
7.736 70.131 11.255 0.39 0.00 50.51 20.27 0.00 15.50 0.47 7.91 1.86 1.22 0.04
7.738 70.131 11.255 0.38 0.01 50.34 20.41 0.00 15.79 0.55 8.04 1.83 0.66 0.04
7.74 70.131 11.255 0.37 0.00 50.07 20.50 0.00 15.89 0.50 8.05 1.79 0.87 0.02

The line scan results shown in Table 52 revealed the presence of Zr-rich intermetallic phases
containing between 11% and 22% Zr, and occasionally containing approximately 37% Zr.
The binary Ni-Zr phase diagram, shown in Figure 50, indicates that the Ni7Zr2 intermetallic
phase contains approximately 30 wt.% Zr, whereas the Ni5Zr intermetallic compound contains
approximately between 21 and 26 wt.% Zr. The results shown in Table 52 therefore suggest
that the intermetallic phase in the braze joint was Ni5Zr, rather than Ni7Zr2.

11.4) Conclusions

 The MarM247/Ni-7Cr-31Hf braze mixture in the as-brazed condition (prior to diffusion

treatment) apparently contained a Hf-rich intermetallic phase with between 31 and 36%
Hf. The binary Ni-Hf phase diagram predicts that the Ni7Hf2 phase contains
approximately 46 wt.% Hf, whereas the Ni5Hf phase contains approximately 38 wt.% Hf.
The presence of significant amounts of Zr and Ta in solution in the intermetallic phase,
however, complicates identification, shifting its composition away from the equilibrium
composition predicted by the phase diagram. If it is assumed that the intermetallic
compound contains between 7.4% and 9.6% Zr in solution, as well as 3% Ta, its

155
 composition may well approach that of the Ni7Hf2 intermetallic compound, rather than
that of the Ni5Hf phase.

 The ternary MarM247/Ni-7Cr-13Zr braze mixture in the as-brazed condition (i.e. no

diffusion heat treatment) was shown to contain an intermetallic compound with a Zr
content ranging between 27 and 29%. The binary Ni-Zr phase diagram indicates that the
Ni7Zr2 intermetallic phase contains approximately 30 wt.% Zr, whereas the Ni5Zr
intermetallic compound contains approximately between 21 and 26 wt.% Zr. This
suggests that the intermetallic phase in the braze joint could be either Ni5Zr or Ni7Zr2.

 The MarM247/Ni-7Cr-13Zr braze mixture contained a Zr-rich intermetallic phase with a

Zr content of approximately 37 wt.% after an extended braze cycle. This suggests that
the intermetallic phase in the braze joint is Ni5Zr, rather than Ni7Zr2.

156
 CHAPTER 12 – EXPERIMENT 10
DEVELOPMENT OF COMPLEX Ni-Cr-Hf BRAZE ALLOYS

12.1) Introduction

During the course of Experiment 8 (described in Chapter 10), the microstructures of a number
of ternary Ni-Cr-Hf alloys, containing 13% Cr and between 15% and 25% Hf, were
examined. The compositions of the eutectic component within these alloys were found to
deviate from the predicted eutectic point in the binary Ni-Hf system (located at 30.5% Hf).
The composition of the eutectic constituent observed in the Ni-13Cr-15Hf braze alloy was
65.2Ni-9.0Cr-25.8Hf (wt.%), the eutectic within the Ni-13Cr-20Hf braze alloy had a
composition of 64.3Ni-10.5Cr-25.2Hf, whereas the eutectic component observed in the Ni-
13Cr-25Hf braze alloy consisted of 63.1Ni-15.1Cr-21.8Hf (wt.%). The Hf content of the
eutectic component in these alloys therefore ranged from approximately 21% Hf to 26% Hf.

This experiment aimed at developing and characterizing the microstructures of novel “near-
eutectic” braze alloys in the as-cast condition, with or without additional alloying elements
(Co, W, C, Ti and/or Al). An attempt was also made to identify the intermetallic phases and
eutectic components observed in each alloy.

12.2) Experimental procedure

Five novel “near-eutectic” alloys, with chemical compositions given in Table 53, were melted
at 1500oC, and allowed to cool to form an ingot. In addition to Ni, Cr and Hf, deliberate
additions of Co (for high temperature strength and oxidation resistance), W and C (for high
temperature strength and to promote the formation of carbides), as well as Ti and Al (to
facilitate the formation of ’ precipitates) were made to some of these alloys. The resultant
as-cast microstructure of each alloy was analyzed using microprobe spot analysis techniques
(as described in §11.2).

Table 53 – Chemical compositions (wt.%) of the novel “near-eutectic” alloys examined

during the course of this investigation.
PV Heat Cr Co W Ti Al C Hf Ni
9020 8.5 - - - - - 25 66.5
9023 7.4 - 4.7 1.4 2.3 0.03 25.5 58.7
9025 8.4 - 5.4 1.6 6.1 - 14.5 64.0
9024 8.3 5.6 5.3 - 4.3 - 14.2 62.4
9026 7.9 5.4 5.0 1.5 4.1 0.04 20.3 55.8

12.3) Results and discussion

12.3.1 Alloy PV9020 (Ni-8.5Cr-25Hf):

Figure 163 displays photomicrographs at increasing magnification of the as-cast structure of

alloy PV9020, with a chemical composition of Ni-8.5Cr-25Hf (wt.%).

157
 (a) (b)

VFE
FE

(c)

Figure 163 – Backscatter electron microprobe images of the Ni-8.5Cr-25Hf alloy in the as-
cast condition (PV9020).

This alloy appears to be hypereutectic, with a needle-shaped primary Hf-rich intermetallic

phase (light etching phase), consisting of 71.5Ni-1.5Cr-4.6Co-22.4Hf (at.%). This
composition (given as atomic percentage) suggests that the intermetallic phase in Figure 163
could be Ni3Hf (a non-equilibrium phase) or Ni7Hf2 (more likely). The fine binary eutectic
component evident in the micrographs was shown to consist of 76.3Ni-14.2Cr-9.4Hf
(analyzed at the location labelled “VFE” in Figure 163(c)) and 76.5Ni-13.3Cr-10.1Hf
(analyzed at the location labelled “FE” in Figure 163(c)). An attempt was made to analyze
the individual phases within the eutectic component (see Table 54), but the results were
inconsistent (probably due to the fine morphology of the eutectic component). It is assumed
that the light phase within the eutectic component is the non-equilibrium Ni3Hf or Ni7Hf2
intermetallic phase, and the dark component Ni-rich  phase. This alloy had a solidus
temperature of 1195C and a liquidus temperature of 1281C (T = 86C). This solidification
temperature range makes the Ni-8.5Cr-25Hf alloy unsuitable for use in the repair of equiaxed
Ni-base superalloys, such as In738 and MarM247. Single crystal Ni-base superalloys,
however, can be processed at temperatures up to 1310oC, so this braze alloy may be suitable
for use in repairing these materials.

158
Table 54 – EMPA results of the individual phases within the eutectic component observed in
alloy PV9020 (wt.%).
Ni Cr Hf
Dark eutectic phase (Analysis 1) 84.0 11.8 4.1
Dark eutectic phase (Analysis 2) 78.9 6.5 14.5
Light eutectic phase (Analysis 1) 92.7 2.9 4.3
Light eutectic phase (Analysis 2) 70.4 11.2 18.4

12.3.2 Alloy PV9023 (Ni-7.4Cr-4.7W-1.4Ti-2.3Al-0.03C-25.5Hf):

Figures 164(a) and (b) display photomicrographs at increasing magnification of the as-cast
structure of alloy PV9023, with a chemical composition of Ni-7.4Cr-4.7W-1.4Ti-2.3Al-
0.03C-25.5Hf (wt.%). The as-cast microstructure is considerably more complex than that of
alloy PV9020, probably as a result of the deliberate addition of W, Al, Ti and C to this alloy.
The white needle-shaped Hf-rich phase was shown to consist of 52.4Ni-1.1Cr-0.2W-0.4Ti-
0.4Al-46.2Hf (at.%), suggesting a non-equilibrium NiHf or Ni11Hf9 intermetallic compound.
The light to medium gray phase consisted of 61.0Ni-1.9Cr-2.1W-3.4Ti-3.2Al-28.9Hf (at.%)
and 63.3Ni-2.8Cr-3.7W-3.8Ti-3.6Al-23.1Hf (at.%), suggesting a possible non-equilibrium
Ni3Hf or Ni7Hf2 intermetallic compound. The darkest phase contained 61.4Ni-21.2Cr-5.4W-
2.0Ti-5.4Al-4.6Hf (at.%) and appears to be the Ni-rich  phase with significant amounts of
Cr, Al and W in solution. An attempt was made to analyze the individual phases within the
eutectic component, yielding compositions of 61.7Ni-17.1Cr-11.3W-1.9Ti-3.8Al-4.2Hf for
the dark eutectic phase (Ni-rich  phase) and 52.6Ni-1.9Cr-0.4W-0.6Ti-0.8Al-44.4Hf for the
light eutectic phase (possibly the non-equilibrium NiHf or Ni11Hf9 intermetallic compound).
This alloy had a solidus temperature of 1140C and a liquidus temperature of 1282C (T =
142C). This solidification temperature range makes the Ni-7.4Cr-4.7W-1.4Ti-2.3Al-0.03C-
25.5Hf alloy unsuitable for use in the repair of equiaxed Ni-base superalloy materials, such as
In738 and MarM247, but this braze alloy may be suitable for use in repairing single crystal
Ni-base superalloys.

12.3.3 Alloy PV9025 (Ni-8.4Cr-5.4W-1.6Ti-6.1Al-14.5Hf):

Figures 165(a) and (b) display photomicrographs of the as-cast structure of alloy PV9025,
with a chemical composition of Ni-8.4Cr-5.4W-1.6Ti-6.1Al-14.5Hf (wt.%). This alloy
appears to be hypo-eutectic, with primary Ni-rich  dendrites consisting of Ni-7.1Cr-16.5Al-
1.8W-2.6Ti-2.9Hf (at.%). The fine binary eutectic component between the dendrites was
shown to consist of 61.2Ni-15.0Cr-13.3Al-0.6W-1.3Ti-8.6Hf (at.%). An attempt was made to
analyze the individual phases within the eutectic component, yielding compositions of
56.4Ni-19.4Cr-17.6Al-0.7W-2.4Ti-3.5Hf for the dark eutectic phase (Ni-rich  phase) and
64.4Ni-10.3Cr-6.8Al-1.1Ti-17.4Hf for the light eutectic phase (possibly the Ni7Hf2
intermetallic compound). Figure 165(b) also reveals the presence of a number of W-rich
particles, probably (W,Cr) carbides or Laves phase. The solidus and liquidus temperatures of
this alloy were 1049oC and 1119oC, respectively (T = 70C). Both the solidus and liquidus
temperatures were therefore lower than those of the simple eutectic Ni-Hf braze alloy
described in Chapter 3. This alloy shows great promise for the repair of equiaxed,
directionally solidified and single crystal Ni-base superalloys.

159
 (a)

52.4Ni-1.1Cr-0.2W-0.4Ti-
0.4Al-46.2Hf

63.3Ni-2.8Cr-3.7W-3.8Ti-
3.6Al-23.1Hf

61.4Ni-21.2Cr-5.4W-2.0Ti-
5.4Al-4.6Hf

61.0Ni-1.9Cr-2.1W-3.4Ti-
3.2Al-28.9Hf

(b)

Figure 164 – Backscatter electron microprobe images of the Ni-7.4Cr-4.7W-1.4Ti-2.3Al-

0.03C-25.5Hf alloy in the as-cast condition (PV9023). (All compositions given as atomic %).

160
 Light phase in eutectic:
64.3Ni-10.3Cr-6.8Al-1.1Ti- 70.3Ni-2.9Cr-4.0Al-
17.4Hf
0.5Ti-22.2Hf

Dark phase in eutectic:

56.3Ni-19.4Cr-0.7W-
17.6Al-2.4Ti-3.5Hf

Dendrite core:
69.1Ni-7.1Cr-16.5Al-
Fine eutectic:
1.8W-2.6Ti-2.9Hf
61.1Ni-15.0Cr-0.6W-
13.3Al-1.3Ti-8.6Hf Edge of dendrite:
69.9Ni-4.9Cr-16.1Al-
0.6W-2.8Ti-5.6Hf

(a)

1.7Ni-28.1Cr-69.9W-0.2Hf

(b)

Figure 165 – Backscatter electron microprobe images of the Ni-8.4Cr-5.4W-1.6Ti-6.1Al-

14.5Hf alloy in the as-cast condition (PV9025). (All compositions given as atomic %).

161
 12.3.4 Alloy PV9024 (Ni-8.3Cr-5.6Co-5.3W-4.3Al-14.2Hf):

Figures 166(a) and (b) display photomicrographs of two different areas in the as-cast
structure of alloy PV9024, with a chemical composition of Ni-8.3Cr-5.6Co-5.3W-4.3Al-
14.2Hf (wt.%).

70.5Ni-2.7Cr-4.4Co-2.4Al-
19.9Hf

66.5Ni-9.6Cr-5.6Co-1.4W-
14.3Al-2.6Hf

Fine eutectic:
67.5Ni-6.9Cr-5.1Co-0.7W-
7.3Al-11.4Hf
65.0Ni-13.1Cr-7.3Co-10.7Al-
3.1W-0.7Hf

68.5Ni-5.7Cr-4.3Co-15.5Al-
1.5W-4.4Hf

(a)

64.6Ni-12.5Cr-7.3Co-10.9Al-
3.0W-0.7Hf

Fine eutectic:
69.3Ni-8.3Cr-5.6Co-5.6Al
1.0W-10.0Hf

69.5Ni-5.5Cr-4.3Co-15.7Al-
1.6W-4.3Hf

(b)

Figure 166 – Backscatter electron microprobe images of the Ni-8.3Cr-5.6Co-5.3W-4.3Al-

14.2Hf alloy in the as-cast condition (PV9024). (All compositions given as atomic %).

This alloy also appears to be hypo-eutectic, with dark gray primary Ni-rich  dendrites
consisting of Ni-13.1Cr-7.3Co-10.7Al-3.1W-0.7Hf (at.%) in Figure 166(a), and Ni-17.5Cr-
7.4Co-1.0W-11.5Al-0.7Hf (at.%) in Figure 166(b). The fine binary eutectic component
between the dendrites was shown to consist of 68.6Ni-6.9Cr-5.1Co-7.3Al-0.7W-11.4Hf

162
 (Figure 166(a)), and 69.3Ni-8.3Cr-5.6Co-1.0W-5.6Al-10.0Hf (Figure 166(b)).
Compositions of 70.6Ni-2.7Cr-4.4Co-2.4Al-19.9Hf (Figure 166(a)) and 74.5Ni-1.8Cr-4.5Co-
1.4Al-21.8Hf (Figure 166(b)) were determined for the light eutectic phase (possibly the
Ni7Hf2 intermetallic compound). The solidus and liquidus temperatures of this alloy were
1185oC and 1334oC, respectively (T = 149C). The liquidus temperature of this alloy was
therefore too high to be of any practical use in the repair of the Ni-base superalloys utilized in
production today, and the alloy is of no practical value in the short term.

12.3.5 Alloy PV9026 (Ni-7.9Cr-5.4Co-5W-1.5Ti-4.1Al-0.04C-20.3Hf):

Figure 167 displays a photomicrograph of the as-cast structure of alloy PV9026, with a
chemical composition of Ni-7.9Cr-5.4Co-5W-1.5Ti-4.1Al-0.04C-20.3Hf (wt.%). This alloy
appears to be hypo-eutectic, with dark gray primary Ni-rich  dendrites consisting of Ni-
9.8Cr-6.2Co-12.9Al-2.2W-2.7Ti-2.9Hf (at.%). The fine binary eutectic component between
the dendrites was shown to consist of Ni-9.9Cr-6.4Co-7.8Al-1.5W-1.6Ti-10.2Hf (at.%). The
coarse white phase was identified as the non-equilibrium Ni3Hf or Ni7Hf2 intermetallic
compound, consisting of 66.1Ni-2.4Cr-5.1Co-3.6Al-0.8Ti-22.0Hf (at.%). The composition of
a number of white rosette-shaped particles (1.5Ni-11.4Cr-0.3Co-0.2Al-87.5W-0.1Ti-0.2Hf)
suggests the presence of Laves phase or (W,Cr) carbides. Laves phase is not uncommon in
some unstable Ni-base superalloys. Smaller dark particles observed in the microstructure
were identified as Ni3Al (’) with a composition of 45.5Ni-6.2Cr-4.3Co-29.1Al-13.7Ti-1.2Hf.
The solidus and liquidus temperatures of this alloy were 1034oC and 1210oC, respectively (T
= 176C). Although the solidus and liquidus temperatures were lower than those of the
simple eutectic Ni-Hf braze alloy described in Chapter 3, the potential of this alloy for the
repair of equiaxed, directionally solidified and single crystal Ni-base superalloys may be
limited by the presence of brittle Laves phase within the as-cast microstructure.

65.7Ni-2.4Cr-5.1Co-3.6Al-
0.8Ti-22.0Hf

61.1Ni-9.8Cr-6.2Co-2.2W-
2.7Ti-12.9Al-2.9Hf

51.6Ni-6.2Cr-4.3Co-13.7Ti- 47.5Ni-23.7Cr-5.7Co-1.2W-
29.1Al-1.2Hf 3.7Ti-12.3Al-1.8Hf

1.5Ni-11.4Cr-0.3Co-0.2Al- Fine eutectic:

87.5W-0.1Ti-0.2Hf
62.3Ni-9.9Cr-6.4Co-1.5W-
1.6Ti-7.8Al-10.2Hf

Figure 167 – Backscatter electron microprobe images of the Ni-7.9Cr-5.4Co-5W-1.5Ti-

4.1Al-0.04C-20.3Hf alloy in the as-cast condition (PV9026). (All compositions given as
atomic %).

163
12.4) Conclusions

 An alloy containing 66.5Ni-8.5Cr-25Hf (wt.%) appeared to be hypereutectic, with a

needle-shaped primary Hf-rich intermetallic phase identified as non-equilibrium Ni3Hf or
Ni7Hf2, and a fine eutectic component consisting of Ni-rich  and the Ni3Hf or Ni7Hf2
intermetallic phase. This alloy had a solidus temperature of 1195C and a liquidus
temperature of 1281C (T = 86C). This solidification temperature range makes the Ni-
8.5Cr-25Hf alloy unsuitable for repairing equiaxed Ni-base superalloys, such as In738
and MarM247. Single crystal Ni-base superalloys, however, can be processed at
temperatures up to 1310oC, so this braze alloy may be suitable for use in repairing these
materials.

 An alloy with a chemical composition of Ni-7.4Cr-4.7W-1.4Ti-2.3Al-0.03C-25.5Hf

(wt.%) displayed a complex microstructure consisting of a needle-shaped Hf-rich phase
(identified as the non-equilibrium NiHf or Ni11Hf9 intermetallic compound), a non-
equilibrium Ni3Hf or Ni7Hf2 intermetallic compound, Ni-rich  phase with a significant
amount of Cr in solution, and a eutectic component consisting of Ni-rich  and the non-
equilibrium NiHf or Ni11Hf9 intermetallic compound. This alloy had a solidus
temperature of 1140C and a liquidus temperature of 1282C (T = 142C). This
solidification temperature range makes this material unsuitable for repairing equiaxed Ni-
base superalloy materials, such as In738 and MarM247.

 Alloy PV9025, with a chemical composition of Ni-8.4Cr-5.4W-1.6Ti-6.1Al-14.5Hf

(wt.%), appeared to be hypo-eutectic, with primary Ni-rich  dendrites and a fine binary
eutectic component consisting of Ni-rich  phase and Ni7Hf2 intermetallic compound. A
number of W-rich particles, probably (W,Cr) carbides or Laves phase, were also
observed. The solidus and liquidus temperatures of this alloy were 1049oC and 1119oC,
respectively (T = 70C). Both the solidus and liquidus temperatures were lower than
those of the simple eutectic Ni-Hf braze alloy, and the alloy demonstrates great potential
for the repair of equiaxed, directionally solidified and single crystal Ni-base superalloys.

 An alloy consisting of Ni-8.3Cr-5.6Co-5.3W-4.3Al-14.2Hf (wt.%) also appeared to be

hypo-eutectic, with primary Ni-rich  dendrites and a fine eutectic component consisting
of Ni-rich  and the Ni7Hf2 intermetallic compound. The solidus and liquidus
temperatures of this alloy were 1185oC and 1334oC, respectively (T = 149C). The
liquidus temperature of this alloy was therefore too high to be of any practical use in the
repair of the Ni-base superalloys utilized in production today, and it is of no practical
value as a braze alloy in the short term.

 The as-cast structure of alloy PV9026, with a chemical composition of Ni-7.9Cr-5.4Co-

5W-1.5Ti-4.1Al-0.04C-20.3Hf (wt.%), appeared to be hypo-eutectic, with primary Ni-
rich  dendrites and a fine eutectic component between the dendrites. A coarse phase,
(identified as the non-equilibrium Ni3Hf or Ni7Hf2 intermetallic compound), a number of
smaller rosette-shaped particles (identified as Laves phase), and small Ni3Al particles
were also observed. The solidus and liquidus temperatures of this alloy were 1034oC and
1210oC (T = 176C), respectively. Although the solidus and liquidus temperatures were
lower than those of the simple eutectic Ni-Hf braze alloy, the potential of this alloy for
the repair of equiaxed, directionally solidified and single crystal Ni-base superalloys may
be limited by the presence of brittle Laves phase within the as-cast microstructure.

164
 CHAPTER 13 - EXPERIMENT 11
CHARACTERIZATION OF THE MICROSTRUCTURE AND MECHANICAL
PROPERTIES OF ADH BRAZE JOINTS IN FSX-414 COBALT-BASE
SUPERALLOY NOZZLE SEGMENTS USING NOVEL BRAZE ALLOYS

13.1) Introduction

The preceding chapters focused on the use of binary and ternary Ni-Zr and Ni-Hf braze alloys
for the repair of turbine components cast from Ni-base superalloys, such as Inconel 738 and
MarM247. The promising mechanical properties of the novel braze joints in Ni-base
superalloy components prompted an investigation into the use of novel braze alloys for
repairing Co-base superalloy nozzle segments.

GE Energy Services, an OEM, currently produce a range of Industrial Gas Turbine (IGT)
engines, including models such as the Frame7FA, Frame 7FA+ and Frame 7FA+e engines.
All these engines contain first stage nozzle segments cast from FSX-414 cobalt-base
superalloy. The chemical composition of FSX-414 is given in Table 55. This alloy is a solid
solution strengthened alloy, in contrast to the ’ strengthened Ni-based superalloys (such as
Inconel 738 and MarM247) considered in earlier chapters. A schematic illustration of a
Frame7FA+e first stage nozzle segment with two airfoils is shown in Figure 168.

Table 55 – Nominal chemical compositions of the alloys considered in this investigation

(wt.%).
Alloy Co Cr Ni W Fe Ta Zr Si Mn Ti Al Ce C B
FSX-414 Bal 29.0 10.0 7.5 1.0 - - - - - - - 0.25 0.01
Nozzalloy Bal 29.0 20.0 - - 7.0 0.01 - - - - 0.06 0.05 -
X-40 Bal 25.5 10.0 7.5 - - - 0.7 0.7 - - - 0.5 -
D15 10.3 15.3 Bal - - 3.5 - - - - 3.0 - - 2.5
MarM509 Bal 23.5 10.0 7.0 - 3.5 0.5 - - 0.2 - - 0.6 -
MarM509B Bal. 23.5 10.0 7.0 - 3.5 0.5 0.2 - 0.2 - - 0.6 2.5
MarM918 Bal 20.0 20.0 - - 7.5 0.1 - - - - - 0.05 -

Figure 168 - Schematic illustration of the first stage nozzle segment of a Frame7FA+e IGT
engine.

165
Nozzle segments in IGT engines tend to degrade during engine service, with degradation
modes ranging from craze cracking, oxidation, corrosion and creep, to foreign object damage.
Examples of characteristic cracks in a first stage nozzle segment are shown in Figure 169.
The cracks in this nozzle segment were originally repaired using manual arc welding with
processes such as GTAW (gas tungsten arc welding) or PAW (plasma arc welding), and a Co-
base filler metal known as Nozzalloy. The chemical composition of Nozzalloy is given in
Table 55.

Thermal Craze-cracks
fatigue cracks

Figure 169 - An example of craze-cracks and large individual thermal fatigue cracks on the
outer sidewall of a first stage nozzle segment.

Although as a solid solution strengthened alloy, FSX-414 is readily weldable, the extent of
cracking displayed in Figure 169 resulted in several hours of weld time. Extensive welding
caused distortion of the nozzle segments to the extent that key dimensions were outside
allowable tolerance limits. This resulted in the nozzle segments not fitting properly on return
to the engine. Since brazing in a vacuum furnace involves uniform heating and cooling,
resulting in minimal distortion, a vacuum braze repair was developed. The patented wide gap
braze process known as ADH (activated diffusion healing), described in §1.5, was used. The
ADH repair involved mixing equal parts (by weight) X-40 cobalt-base superalloy powder and
a ’ strengthened Ni-base braze alloy known as D15 (containing 2.5% B as melt point
depressant). The nominal compositions of X-40 and D15 are given in Table 55. The X-
40/D15 mixture resulted in a hybrid braze mix with poor mechanical properties, and the joint
cracked rapidly in service. As a result of the poor mechanical integrity of the braze joints,
customers refused to have their nozzle segments ADH braze repaired, and insisted that the
nozzle segments be welded.

The poor service performance of the X-40/D15 ADH braze repair prompted further
investigation with the following objectives:

 To examine an alternative ADH braze mixture consisting of MarM509 Co-base

superalloy powder and MarM509B braze alloy for crack repair of FSX-414 nozzle
segments, with the aim of obtaining mechanical integrity equivalent or superior to the
properties of Nozzalloy welds.
 To develop and test a novel Co-base braze alloy containing Hf instead of B as melt point
depressant for the crack repair of FSX-414 nozzle segments.

166
Three properties are considered critical with respect to the quality of the braze joints: elevated
temperature tensile properties, creep rupture properties and low cycle fatigue life. If a newly
developed braze repair procedure ensures mechanical properties equivalent to those of a weld
repair, the improved braze procedure would most likely be preferred by repair customers due
to the lower risk of distortion.

In view of the stated requirements for an acceptable braze procedure, the X-40/D15 ADH
braze mixture was changed to a MarM509/MarM509B braze. The chemical compositions of
MarM509 and MarM509B are shown in Table 55. MarM509 is a Co-base superalloy with an
excellent combination of mechanical properties, including high tensile strength, fatigue
strength and creep rupture strength. MarM509B is a Co-base braze alloy with a nominal
chemical composition virtually identical to that of MarM509, with 2.5 wt.% B added as melt
point depressant. Certification documents for the MarM509B braze alloy assign a solidus
temperature of 1121C and a liquidus temperature of 1160C to the alloy. Si is present as a
trace element in MarM509B and is not added intentionally to lower the melting point.

The promising results described in earlier chapters for the binary and ternary eutectic Ni-Hf
braze alloys encouraged the development of a novel Co-base braze alloy containing Hf as
melt point depressant. The binary Co-Hf phase diagram [2], shown in Figure 170, illustrates
that a eutectic point exists at a composition of 27.2%Hf and 72.8%Co (by weight), and a
temperature of 1230C. This investigation therefore included an evaluation of a novel ADH
braze joint consisting of a mixture of MarM509 powder and the binary eutectic Co-Hf braze
alloy.

Figure 170 – The binary Co-Hf phase diagram [2].

13.2) Experimental procedure

Plates of FSX-414 with dimensions 248 mm long, 127 mm wide and 19 mm thick were
investment cast for this experiment. A 1.5 mm wide groove was machined in the center of
each plate, as illustrated in Figure 171. This gap width was selected on the basis of an
investigation into the crack widths observed in 96 nozzle segments (four engine sets, with 24

167
nozzle segments in each engine set). The data revealed a mean crack width of 0.64 mm and a
maximum crack width of 1.5 mm. The approval authority required that the new ADH process
be capable of repairing up to the maximum crack width, i.e. 1.5 mm. The gap was only
machined partially, leaving 0.5 mm of parent metal intact at the base of the groove (as shown
in Figure 172). This was done to maintain the 1.5 mm gap width during brazing without the
need for special fixturing. Since the mechanical test samples were machined to a gauge
diameter of 6.4 mm, the 0.5 mm parent metal retained at the base of the groove was machined
away after brazing and therefore did not affect the test results in any way.

Figure 171 - The FSX-414 test plate with a 1.5 mm wide groove machined in the centre.

Figure 172 - Enlarged view of the machined groove in the braze test plate.

Two sets of braze samples were produced. The first set of samples was produced by filling
the groove with a mixture of MarM509 powder and MarM509B braze powder (with
compositions shown in Table 55). The powders were mixed in a ratio of 60% MarM509 and
40% MarM509B (by weight). The second set of samples was produced by filling the groove
with a mixture of MarM509 powder and eutectic Co-Hf powder (containing 27.2% Hf) in a
ratio of 60% MarM509 powder and 40% braze alloy. The plates were then processed in a
vacuum furnace using the following braze cycle:

1) Evacuate the vacuum furnace to a pressure of 5x10-4 torr or lower.

2) Ramp up to a temperature of 450oC14oC at a rate between 8oC/minute and 14oC/minute.
3) Hold at 450oC for a minimum of 40 minutes to allow the binder to burn off. Proceed to
the next step once the vacuum level had returned to 1x10-3 torr or lower.

168
4) Ramp up to a temperature of 1093oC14oC at a rate between 8oC/minute and
14oC/minute.
5) Hold at 1093oC for a minimum of 40 minutes to allow the samples to stabilize at this
temperature.
6) Ramp up to the melting temperature of 1245oC14oC at a rate between 8oC/minute and
14oC/minute.
7) Hold at 1245oC for a minimum of 30 minutes.
8) Furnace cool under vacuum to 1150oC14oC.
9) Hold at 1150oC for a minimum of 240 minutes.
10) Ramp up to a temperature of 1177oC14oC at a rate between 8oC/minute and
14oC/minute.
11) Hold at 1177oC for a minimum of 240 minutes.
12) Ramp up to a temperature of 1205oC14oC at a rate between 8oC/minute and
14oC/minute.
13) Hold at 1205oC for a minimum of 240 minutes.
14) Gas quench with argon or helium gas to a safe temperature (about 90C) before
unloading the test plates.

A melting temperature of 1245C was selected as it is above the liquidus temperatures of both
the MarM509B ADH braze alloy and the novel Co-Hf braze alloy. The diffusion
temperatures (1150C, 1177C and 1205C) were selected to be below the liquidus
temperatures of both braze alloys, but high enough to facilitate rapid diffusion of the melt
point depressants, i.e. B or Hf, into the surrounding superalloy powder particles.

Photographs of the MarM509/Co-Hf and the MarM509/MarM509B ADH joints are shown in
Figures 173(a) and (b). The appearance of the joints suggests that the MarM509/MarM509B
ADH braze alloy was more fluid than the MarM509/Co-Hf braze at the melting temperature.
This is attributed to the amount of superheating above the liquidus temperature of the braze
alloy, with the MarM509/MarM509B ADH braze being processed at a temperature 68C
higher than its liquidus temperature, whereas the MarM509/Co-Hf ADH braze was processed
at a temperature only 16C higher than its eutectic temperature.

(a) (b)

Figure 173 - Photographs of: (a) the MarM509/Co-Hf ADH test plate in the as-brazed
condition; and (b) the MarM509/MarM509B ADH test plate in the as-brazed condition.

In order to compare the mechanical properties of the novel ADH braze joints with those of the
welds currently preferred by repair customers, MarM509 plates were fusion welded using a

169
double V-weld preparation in the downhand position. The GTAW (gas tungsten arc welding)
process was used with the following weld parameters:

 Welding current: 60 to 80 A
 Arc voltage: 10 to 15 V
 Welding speed: Manual
 Tungsten electrode: 2% Thoriated tungsten (classification: EWTh-2), 1.6 mm or 2.4
mm diameter
 Polarity: DCEN (direct current electrode negative)
 Filler wire diameter: 0.9 mm to 1.6 mm
 Shielding gas: 95% Ar + 5% H2 (by volume)
 Preheat temperature: 10C minimum
 Interpass temperature: 150C maximum
 Welding technique: Stringer bead
 Gas cup size: 9.5 mm to 19 mm
 Filler metal: Nozzalloy or MarM918 (nominal chemical compositions are shown in
Table 55).

Tensile tests, creep rupture tests and low cycle fatigue (LCF) tests were performed at various
temperatures. An example of a mechanical test sample machined from the test plates is
shown in Figure 174. The elevated temperature tensile tests were performed at 870C.
Creep rupture tests were performed at 760oC, 815oC, 870oC, 980oC, 1040oC and 1090oC at
various applied stress levels. The LCF tests were performed at 870oC at various strain levels.
Since the success of this investigation hinged mainly on the LCF properties of the joints, the
Approval Repair Authority requested that the following LCF curves be generated:

a) FSX-414 base metal welded with MarM918 filler metal. This weld filler metal is used to
salvage virgin castings with casting defects, such as hot tears or macroporosity. The
chemical composition of MarM918 is given in Table 55.
b) FSX-414 welded with Nozzalloy filler metal. This weld filler metal is used to repair
nozzle segments degraded during service.
c) FSX-414 brazed with the new MarM509/MarM509B wide gap braze mixture.
d) FSX-414 brazed with the new MarM509/Co-Hf ADH braze mixture.

6.4 mm

Figure 174 - Mechanical test sample machined from the braze test plates. The 1.5 mm wide
braze joint is located in the centre of the gauge length.

13.3) Results and discussion

13.3.1 Mechanical test results:

The results of tensile tests performed at 870C are shown in Table 56 for the FSX-414 base
metal, FSX-414 welded with Nozzalloy, FSX-414 ADH brazed with MarM509/MarM509B,
and FSX-414 ADH brazed with the novel Co-Hf braze alloy.

170
 Table 56 - Tensile properties at 870C of FSX-414 base metal, FSX-414 welded with
Nozzalloy, FSX-414 brazed with MarM509/MarM509B, and FSX-414 brazed with
MarM509/Co-Hf. The test samples were produced with a joint gap of 1.5 mm to simulate a
worst-case crack repair scenario.
Failure Tensile Yield stress, % Reduction
Specimen identification % Elongation
location strength, MPa MPa in area
BLT-1T (FSX-414 base metal) Base metal 291 168 32 32
BLT-2T (FSX-414 base metal) Base metal 304 177 31 31
BLT-3T (FSX-414 base metal) Base metal 296 167 44 44
BLT-4T (FSX-414 base metal) Base metal 283 164 36 36
Average for the FSX-414 base metal - 295 169 35.8 35.8
NOZ-1 (FSX-414 welded with Nozzalloy) HAZ* 278 175 44 53
NOZ-7 (FSX-414 welded with Nozzalloy) Base metal 274 179 46 45
NOZ-10 (FSX-414 welded with Nozzalloy) Base metal 275 177 42 52
NOZ-16 (FSX-414 welded with Nozzalloy) Base metal 275 178 38 43
Average for the FSX-414/Nozzalloy weld 275 177 42.5 48.3
509-1 (FSX-414 brazed with MarM509/MarM509B) Base metal 277 190 48 46
509-7 (FSX-414 brazed with MarM509/MarM509B) Base metal 282 189 28 44
509-10 (FSX-414 brazed with MarM509/MarM509B) Base metal 268 178 30 38
509-16 (FSX-414 brazed with MarM509/MarM509B) Base metal 284 179 29 48
Average for the MarM509/MarM509B ADH joint 277 184 33.8 44.0
ENH-1 (FSX-414 brazed with MarM509/Co-Hf) Base metal 272 175 28 51
ENH-7 (FSX-414 brazed with MarM509/Co-Hf) Base metal 268 175 50 59
ENH-10 (FSX-414 brazed with MarM509/Co-Hf) Base metal 285 194 25 56
ENH-16 (FSX-414 brazed with MarM509/Co-Hf) ADH joint 275 190 4.3 4.3
Average for the MarM509/Co-Hf ADH joint 275 183 34.3 55.3

* HAZ: Heat-affected zone.

It is evident from the tensile test results shown in Table 56 that the FSX-414 samples fusion
welded with Nozzalloy filler metal performed very well, with three of the four samples tested
failing in the parent metal and the fourth failing in the heat-affected zone. The tensile
strength, yield strength, elongation and reduction in area values were equivalent or even
superior to those of the base metal.

FSX-414 brazed with the MarM509/MarM509B ADH material yielded similar results, with
all the tensile samples failing in the base metal. The tensile test results shown in Table 56
therefore represent the properties of the base metal, rather than those of the braze joint. It can,
however, be concluded that the tensile properties of the MarM509/MarM509B braze joints
were at least equivalent to those of the base metal.

The FSX-414 samples brazed with the MarM509/Co-Hf braze alloy yielded promising results,
with three of the four samples failing in the base metal. It can be concluded that the tensile
properties of these joints were at least equivalent to those of the base metal. The fourth
sample failed in the braze joint, but displayed tensile and yield strength values equivalent to
those of the base metal (as shown in Table 56). The elongation and reduction in area values
of the fourth sample were, however, very low when compared with those of the base metal
(4.3% elongation and reduction in area, compared with 35%). Examination of the fractured
sample revealed porosity on the fracture surface. It is assumed that the low ductility values
measured during this tensile test were related to the presence of porosity in the braze joint.

171
A Larson-Miller plot comparing the creep rupture properties of the FSX-414 base metal,
FSX-414 welded with Nozzalloy filler metal, and FSX-414 ADH brazed with
MarM509/MarM509B, is shown in Figure 175. The Larson-Miller plot indicates that the
MarM509/MarM509B braze samples outperformed the Nozzalloy welds. For example, at a
Larson-Miller number of 49 (the dotted line in Figure 175), the MarM509/MarM509B
samples exhibited 90% of the base metal’s creep rupture properties, whereas the Nozzalloy
samples displayed 75% of the base metal’s properties. At higher test temperatures the creep
rupture properties of the MarM509/MarM509B braze joints approached those of the base
metal. On average the MarM509/MarM509B samples demonstrated 90% of the base metal’s
creep rupture properties, whereas the Nozzalloy samples displayed 72%.

225
760°C

200

175 815°C

150
Stress (MPa)

125 870°C

100
980°C
75

50
1040°C
FSX-414
1090°C
25 MarM509/MarM509B ADH joint
Nozzalloy

0
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
Larson-Miller parameter = (460+T)(20+log t)/1000

Figure 175 – Larson-Miller plot of FSX-414 base metal, the MarM509/MarM509B ADH
braze joints, and the Nozzalloy weld.

The results of the low cycle fatigue (LCF) tests are shown in Table 57, and displayed
graphically in Figure 176.

Although MarM918 filler metal is approved for salvaging virgin castings, all except one of
the FSX-414 samples welded with MarM918 failed in the weld metal or heat-affected zone.
The MarM918 welds displayed the poorest fatigue life of all the samples evaluated. The
Nozzalloy welds performed marginally better, with four of the twelve samples failing in the
weld or heat-affected zone. The remaining eight Nozzalloy samples failed in the base metal.

The MarM509/MarM509B braze joints displayed consistently higher LCF properties than the
welded joints, with four of the twelve samples failing at the interface between the braze joint
and the base metal. The remaining eight samples failed in the base metal.

172
Table 57 – Low cycle fatigue (LCF) data for FSX-414 base metal, FSX-414 welded with MarM918,
FSX 414 welded with Nozzalloy, FSX-414 ADH brazed with MarM509/MarM509B, and FSX-414
ADH brazed with MarM509/Co-Hf.

Cycles to Cycles to Failure

Sample Strain
initiation failure location
0.60 1298 1460 Base metal
0.60 1598 2099 Base metal
0.50 2978 3563 Base metal
0.50 3206 3773 Base metal
0.46 6177 6375 Base metal
0.40 4422 5787 Base metal
FSX-414 base metal
0.40 7126 8765 Base metal
0.35 8584 9087 Base metal
0.35 11085 12972 Base metal
0.33 14290 17141 Base metal
0.30 22736 24712 Base metal
0.30 49953 51066 Base metal
0.60 814 1013 Weld
0.60 1212 1398 Weld
0.50 1669 2039 HAZ*
0.50 2153 2955 HAZ*
0.45 1301 2056 HAZ*
0.40 4194 4539 Weld
FSX-414 welded with MarM918
0.40 3876 5302 Weld
0.40 4885 5312 Weld
0.35 8308 8768 Weld
0.35 12006 12310 HAZ*
0.30 10058 12310 Weld
0.30 20999 26395 Base metal
0.60 1920 1957 Base metal
0.60 1871 2649 Base metal
0.50 3036 3618 HAZ*
0.50 2655 3294 Base metal
0.45 2009 2357 Base metal
0.40 2316 2924 Weld
FSX-414 welded with Nozzalloy
0.40 4989 6583 HAZ*
0.40 6895 7163 Base metal
0.35 6569 7372 HAZ*
0.35 9947 11526 Base metal
0.30 18243 20239 Base metal
0.30 20206 22163 Base metal
0.69 989 1007 Interface**
0.60 1156 1238 Interface**
0.60 876 1360 Base metal
0.50 1854 1912 Interface**
0.50 2641 3010 Base metal
0.45 3893 4939 Base metal
FSX-414 ADH brazed with MarM509/MarM509B
0.44 4890 5289 Base metal
0.40 3735 5189 Base metal
0.40 5173 5556 Base metal
0.35 8036 9188 Interface**
0.35 11911 14986 Base metal
0.30 26421 30221 Base metal

173
 Table 57 – continued.
Cycles to Cycles to Failure
Sample Strain
initiation failure location
0.70 1195 1588 Base metal
0.60 1627 1970 Base metal
0.60 2711 2927 Base metal
0.50 2134 2248 Base metal
0.50 3589 3914 Base metal
0.45 3501 3898 Base metal
FSX-414 ADH brazed with MarM509/Co-Hf
0.45 3092 3906 Base metal
0.40 4975 5771 Base metal
0.40 5455 6849 Base metal
0.35 16584 17726 Base metal
0.35 23185 25365 Base metal
0.30 40911 41754 Base metal

* HAZ: Heat-affected zone

** Interface: Joint line between the base metal and the ADH braze

0.8
FSX-414 base metal
FSX-414 welded with MarM918
FSX-414 welded with Nozzalloy
0.7 FSX-414 ADH brazed with MarM509/MarM509B
FSX-414 ADH brazed with MarM509/Co-Hf

0.6
Strain

0.5

0.4

0.3

0.2
1000 10000 100000
Cycles to failure
Figure 176 – Graphical representation of the LCF data determined for FSX-414 base metal,
FSX-414 welded with MarM918 or Nozzalloy filler metal, FSX-414 ADH brazed with
MarM509/MarM509B, and FSX-414 ADH brazed with MarM509/Co-Hf.

Figure 176 demonstrates that the MarM509/Co-Hf braze joint outperformed all the other
repairs during LCF testing. It is also important to note that all the MarM509/Co-Hf samples
failed in the base metal, and that the ADH braze joint displayed LCF properties superior to
those even of the FSX-414 base metal. No publication in open literature has ever reported
consistent base metal failure for braze joints. This resulted in consensus in both the aircraft
and industrial gas turbine repair industries that braze joints are inherently brittle, and fail
rapidly under conditions of applied strain. Joints produced using braze alloys with B as melt

174
point depressant have been shown to fail consistently in the braze joint under LCF test
conditions.

Statistical analysis of the LCF results revealed that FSX-414 welded with MarM918 suffered
a 38% reduction in LCF life, compared to the base metal. FSX-414 welded with Nozzalloy
displayed a 27% reduction in LCF life (for the four samples that failed in the weld metal or
heat-affected zone), whereas the four FSX-414 samples brazed with MarM509/MarM509B
that failed at the joint interface suffered a 10% reduction in LCF life. All the FSX-414
samples brazed with MarM509/Co-Hf displayed LCF properties superior to those of the base
metal.

13.3.2 Metallurgical examination:

Nozzle segments (especially first stage nozzle segments) endure the highest temperatures in a
gas turbine engine and are exposed to fluctuating thermal strains which may result in thermal
fatigue cracking. During engine service these thermal fatigue cracks can manifest as
individual cracks or clusters of cracks, with the latter usually referred to as craze-cracks
(shown earlier in Figure 169).

Thermal fatigue cracks develop surface oxide layers during service which must be removed
prior to ADH brazing. These oxide layers are usually comprised of Cr-oxides, Ni-oxides or a
combination of these compounds. The hydrogen cleaning process effectively reduces these
oxides within cracks in Co-base superalloys. Simply stated, the hydrogen cleaning process
pulses H2 into a furnace at elevated temperature (around 1200C) for a time of between 8 and
12 hours. The surface oxides are reduced at elevated temperature in this hydrogen-containing
environment. Figures 177(a) and (b) compare the sidewalls of a crack before and after
hydrogen cleaning, and illustrate the effectiveness of the cleaning process in removing surface
oxide layers.

(a) (b)

Figure 177 – (a) Chromium oxides on the crack surface; and (b) oxide reduction on the
sidewalls of the crack after H2 cleaning.

Figure 178 displays a macrograph of craze-cracks successfully repaired using the

MarM509/MarM509B braze alloy. This photograph illustrates that more than eleven small
cracks had been filled with braze alloy. Figure 179 shows an individual crack, also filled

175
with the MarM509/MarM509B braze. Some porosity is evident, but it amounts to less than
10% (by volume), which is the acceptance limit.

Figure 178 - Craze-cracks filled with the MarM509/MarM509B braze alloy.

Figure 179 - Individual crack filled with the MarM509/MarM509B braze alloy.

Figures 180(a) and (b) demonstrate the ability of the MarM509/MarM509B braze alloy to
repair deep cracks which had propagated from the concave to the convex surface of an airfoil
during service. The structure is undiffused, i.e. the repair process had only progressed to the
melting stage (1245C for 30 minutes). As shown in Figure 181, good bonding of the braze
filler metal to the sidewalls of the crack was obtained.

(a) (b)

Figure 180 – (a) Braze flow in a 7.5 mm deep crack; and (b) braze flow in an 11 mm deep
crack.

176
 Braze FSX-414
joint base metal

Figure 181 – Photomicrograph of the interface between the braze crack repair and the base
metal.

Figures 182 and 183 display micrographs of braze joints used for craze-crack repair. These
joints have undergone the full braze procedure, including the diffusion treatments. An
equiaxed, homogenous structure, resembling that of the parent metal, is evident.

Figure 182 - Microstructure of the fully diffused braze repair of craze-cracks on a nozzle
segment.

177
 Figure 183 - Microstructure of the fully diffused braze repair of craze-cracks on a nozzle
segment.

13.4) Conclusions

 This experiment examined the use of alternative ADH braze mixtures consisting of
MarM509 Co-base superalloy powder and MarM509B braze alloy, or MarM509 powder
and a eutectic Co-Hf braze alloy, for crack repair in FSX-414 Co-base superalloy nozzle
segments. The aim of the investigation was to obtain mechanical integrity equivalent or
superior to the properties of the fusion welds currently preferred by repair customers.

 Both the eutectic Co-Hf braze alloy (consisting of 27.2% Hf and 72.8% Co by weight)
and the MarM509B braze were effective in repairing cracks in FSX-414 Co-base
superalloy nozzle segments. Since nozzle segments have been shown to contain cracks
up to 1.5 mm in width, the Co-Hf and MarM509B braze alloys had to be mixed with
MarM509 superalloy powder to facilitate a wide gap ADH braze repair.

 The hydrogen cleaning process was effective in reducing the oxide layers typically found
in thermal fatigue cracks in FSX-414 nozzle segments. Although the MarM509/Co-Hf
braze alloy proved to be more sluggish than the MarM509/MarM509B alloy during
brazing, the braze alloys filled craze-cracks and individual cracks, and bonded effectively
to the sidewalls of the cracks.

 ADH braze repair of FSX-414 using the novel MarM509/MarM509B and MarM509/Co-
Hf braze alloys resulted in high temperature tensile properties equivalent to those of weld
repairs in the same parent material (using Nozzalloy filler metal). The creep rupture and
low cycle fatigue (LCF) properties of the braze joints were superior to those of welds
performed using Nozzalloy or MarM918 filler metal.

178
 FSX-414 samples braze repaired using the novel MarM509/Co-Hf braze alloy
consistently failed in the base metal under low cycle fatigue conditions, implying that the
LCF properties of the braze joint were superior to those of the base metal. This suggests
that failure is likely to occur preferentially in the base metal if cracks in FSX-414 nozzle
segments are ADH braze repaired using this braze alloy. A braze alloy consisting of a
mixture of 60% MarM509 powder (by weight) and 40% eutectic Co-Hf powder therefore
shows potential for crack repair of nozzle segments in industry.

 Statistical analysis of the LCF results revealed that FSX-414 welded with MarM918
suffered a 38% reduction in LCF life compared to the base metal. FSX-414 welded with
Nozzalloy displayed a 27% reduction in LCF life (for four samples that failed in the weld
metal or heat-affected zone), whereas the four FSX-414 samples brazed with
MarM509/MarM509B that failed at the joint interface suffered a 10% reduction in LCF
life. All the FSX-414 samples brazed with MarM509/Co-Hf displayed LCF properties
superior to those of the base metal.

 Microstructural examination of brazed and fully diffused MarM509/MarM509B crack

repairs revealed equiaxed braze microstructures with few boride particles.

179
 CHAPTER 14 – FINAL CONCLUSIONS

This investigation was designed to:

 evaluate the use of two novel melt point depressants, namely Hf and Zr, in Ni- and Co-
base braze filler metals, and
 optimize the liquid phase diffusion bonding process in conjunction with these novel
braze filler metals with the aim of producing high strength joints with good ductility and
mechanical properties at elevated temperature approaching those of the base metal.

This chapter describes the most important results obtained and the major conclusions drawn
on completion of this project.

14.1 Liquid phase diffusion bonding of In738 Ni-base superalloy using binary eutectic
Ni-Hf and Ni-Zr braze alloys

During the course of this investigation, the merit of two novel melt point depressants, namely
Hf and Zr, was evaluated and the liquid phase bonding process optimized for the repair of
wide cracks in Ni-base superalloy turbine components. The following conclusions were
drawn:

 Hf and Zr were found to act as effective melt point depressants in Ni-base braze alloys.
Binary eutectic Ni-Hf and Ni-Zr braze alloys melt and flow satisfactorily when processed
at a melting temperature of 1230C. In the as-brazed condition, the braze alloys were
found to consist of a nickel-rich  phase and an intermetallic compound (provisionally
identified as Ni2Hf7 in the Ni-Hf braze alloy and Ni5Zr in the Ni-Zr braze).
 In order to repair wide cracks in land-based turbine components, the binary Ni-Hf or Ni-
Zr braze alloys need to be combined with a ’ strengthened Ni-base superalloy powder
(such as MarM247) within the joint to form a liquid phase diffusion bond on brazing.
 A vacuum braze cycle consisting of melting at 1230C, followed by an extended
diffusion cycle of 24 hours at the same temperature, solution annealing at 1230C, hot
isostatic pressing at 1200C and a full primary and secondary ageing cycle to precipitate
’ within the MarM247 particles, is recommended for obtaining optimal mechanical
properties in the novel wide gap braze joints.
 Although the braze joints did not exhibit elevated temperature strength and creep rupture
properties equivalent to those of the base metal, the novel wide gap braze joints
consistently outperformed the commercially available B-containing braze alloys. The
ductility of the novel braze alloys containing Hf or Zr as melt point depressant was
significantly higher than the ductility of similar joints containing boron as melt point
depressant.
 The Ni-Zr or Ni-Hf braze braze joints did not display LCF properties equivalent to those
of the base metal, but the novel braze filler metals outperformed the commercially
available B-containing brazes. Within the intermediate strain range where most cracks
occur in nozzle segments (0.4% to 0.8% strain), the Ni-Zr and Ni-Hf braze joints
displayed 70% and 74%, respectively, of the LCF properties of the base metal, compared
to 11% for a B-containing braze alloy.

180
14.2 Liquid phase diffusion bonding of In738 Ni-base superalloy using more complex Ni-
Hf and Ni-Zr braze alloys alloyed with chromium

In an attempt to further improve the performance of the novel braze filler metals, a series of
more complex braze formulations containing chromium, in addition to various other
deliberately added alloying elements, was produced. The microstructures of these
formulations were examined in the as-cast, as-brazed or fully diffused condition, and electron
microprobe analysis techniques were used to identify the phases observed in each alloy. The
following conclusions were drawn:

 The addition of Cr to the novel Ni-Hf braze alloy shifted the eutectic composition to a
lower Hf content.
 The boride particles which form in B-containing braze alloys were found to be
considerably harder than the Ni7Hf2 intermetallic phase that forms in the novel Ni-Cr-Hf
braze alloys. This difference in hardness may account for the high ductility and good
LCF properties of the Ni-Hf braze joints, compared to those of B-containing braze alloys.
 Electron microprobe analysis of braze joints in complex alloys containing Hf or Zr as
melt point depressants revealed that the intermetallic phases which form in the joints are
most likely Ni5Zr in Zr-containing brazes, and the nonequilibrium Ni7Hf2 phase in Hf-
containing braze alloys. Dilution and contamination of the braze mixtures during
production complicated identification of the intermetallic compounds.
 The high solidus and liquidus temperatures of many of the more complex alloys make
them unsuitable for use in the repair of equiaxed Ni-base superalloy materials, such as
In738 and MarM247. An alloy with a composition of Ni-8.4Cr-5.4W-1.6Ti-6.1Al-
14.5Hf (wt.%), however, demonstrated great potential for the repair of equiaxed,
directionally solidified and single crystal Ni-base superalloys.

14.3 Liquid phase diffusion bonding of FSX-414 Co-base superalloy using two novel
ADH braze mixtures

The poor service performance of ADH braze joints currently used for refurbishing FSX-414
Co-base superalloy nozzle segments prompted an investigation into the use of alternative
ADH braze mixtures in this application. The investigation examined ADH mixtures
consisting of MarM509 Co-base superalloy powder and commercially available MarM509B
braze alloy, and MarM509 powder and a eutectic Co-Hf braze alloy for FSX-414 crack repair,
with the aim of obtaining mechanical integrity equivalent or superior to the properties of weld
repairs. The following conclusions were drawn based on the results of the investigation:

 Both the eutectic Co-Hf braze alloy (consisting of 27.2% Hf and 72.8% Co by weight)
and the MarM509B braze were effective in repairing cracks in FSX-414 Co-base
superalloy nozzle segments. Since nozzle segments have been shown to contain cracks
up to 1.5 mm in width, the Co-Hf and MarM509B braze alloys were mixed with
MarM509 superalloy powder to facilitate a wide gap ADH braze repair.
 ADH braze repair of FSX-414 using the novel MarM509/MarM509B and MarM509/Co-
Hf braze alloys resulted in high temperature tensile properties equivalent to those of weld
repairs in the same parent material (using Nozzalloy filler metal). The creep rupture and
low cycle fatigue (LCF) properties of the braze joints were superior to those of welds
performed using Nozzalloy or MarM918 filler metal.

181
 Statistical analysis of the LCF results revealed that FSX-414 welded with MarM918
suffered a 38% reduction in LCF life compared to the base metal. FSX-414 welded with
Nozzalloy displayed a 27% reduction in LCF life, whereas four FSX-414 samples brazed
with MarM509/MarM509B failed at the joint interface and suffered a 10% reduction in
LCF life.
 FSX-414 samples braze repaired using the novel MarM509/Co-Hf braze alloy
consistently failed in the base metal under low cycle fatigue conditions, implying that the
LCF properties of the braze joint were superior to those of the base metal. This suggests
that failure is likely to occur preferentially in the base metal if cracks in FSX-414 nozzle
segments are ADH braze repaired using this braze alloy. A braze alloy consisting of a
mixture of 60% MarM509 powder (by weight) and 40% eutectic Co-Hf powder therefore
has considerable potential for crack repair of nozzle segments in industry.

182
 CHAPTER 15 – RECOMMENDATIONS FOR FUTURE WORK

 The results of SEM-EDS and electron microprobe analyses of the braze joints suggested
that the intermetallic compound observed in the binary and more complex Ni-Hf joints
was the nonequilibrium Ni7Hf2 phase, whereas the intermetallic phase observed in the
Ni-Zr joints was most likely Ni5Zr. In order to confirm the chemical composition and
structure of the intermetallic phases observed in the braze joints, transmission electron
microscopy work is recommended.

 Experiment 10, reported in Chapter 12, demonstrated that a complex Ni-base braze alloy
with a composition of Ni-8.4Cr-5.4W-1.6Ti-6.1Al-14.5Hf (wt.%) (referred to as alloy
PV9025) had solidus and liquidus temperatures below those of the simple eutectic Ni-Hf
braze alloy evaluated in earlier experiments. This braze alloy therefore holds promise for
the repair of equiaxed, directionally solidified and single crystal Ni-base superalloys.
Since the microstructure of this complex Ni-base alloy was evaluated in the as-cast
condition, it is recommended that the alloy be mixed with Ni-base superalloy powder and
used to produce a wide gap liquid phase diffusion bond in a ’ strengthened Ni-base
superalloy, such as In738. The properties of the braze samples should then be evaluated
in the fully heat treated and aged condition. Such an evaluation should include
examination of the joint microstructure using optical microscopy, scanning electron
microscopy and electron microprobe techniques, and full mechanical testing (including
elevated temperature tensile tests, creep rupture tests and low cycle fatigue tests).

 The results of Experiment 11, described in Chapter 13, revealed that ADH braze repair of
wide cracks in FSX-414 cobalt-base superalloy nozzle segments using novel
MarM509/MarM509B or MarM509/Co-Hf ADH braze mixtures resulted in excellent
joint mechanical properties. The braze joints displayed tensile properties at elevated
temperature equivalent to those of the base metal, and creep rupture properties superior
to those of weld repairs in the same base metal. The low cycle fatigue properties of the
MarM509/Co-Hf joints were shown to be at least equivalent to those of the FSX-414
base metal, with failure occurring in the base metal of all the joints evaluated. In order to
further characterize these novel ADH braze alloys, it is recommended that the
microstructures of MarM509/MarM509B and MarM509/Co-Hf joints be characterized
using optical microscopy, scanning electron microscopy and electron microprobe
analysis techniques.

 FSX-414, the Co-base superalloy currently used by GE Energy Services for casting
nozzle segments, is recognized as having mechanical properties inferior to those of the
Co-base superalloys used by a number of other original engine manufacturers (OEM’s)
for the same application. Siemens Westinghouse, an OEM, uses ECY 768 Co-base
superalloy for nozzle segment casting, whereas Pratt & Whitney, Siemens and Alstom,
also OEM’s, use MarM509. Future work should evaluate the mechanical integrity of
ADH braze joints produced using the novel MarM509/MarM509B and MarM509/Co-Hf
ADH braze mixtures in these higher strength Co-base superalloys.

183
 CHAPTER 16 – REFERENCES

[1] Sheward, G.E., and Bell, G.R. “Development and evaluation of a Ni-Cr-P brazing filler
metal”. Welding Journal, vol. 55, no. 10. October 1976. pp. 285s-292s.

[2] ASM Handbook: Volume 3 - Alloy Phase Diagrams. ASM International, Metals Park,
OH. 1992. pp. 2.127 (Ni-Cd binary phase diagram), 2.143 (Co-Hf binary phase
diagram), 2.241 (Ni-Hf binary phase diagram), 2.321 (Ni-Zn binary phase diagram) and
2.322 (Ni-Zr binary phase diagram).

[3] Lugscheider, E., and Kim D.-S. “New low-melting nickel-based high-temperature
brazing alloys”. Schweissen und Schneiden, vol. 43, no. 4. April 1991. pp. 222-226.

[4] Lugscheider, E., and Partz, K.-D. “High-temperature brazing of stainless steel with
nickel-base filler metals BNi-2, BNi-5 and BNi-7”. Welding Journal, vol. 62, no. 6.
June 1983. pp. 160s-165s.

[5] Johnson, C. “The use of TETIG diagrams in high-temperature brazing”. Welding

Journal, vol. 60, no. 10. October 1981. pp. 185s-193s.

[6] Sakamoto, A., Fujiwara, C., Hattori, T, and Sakai, S. “Optimizing processing variables
in high-temperature brazing with nickel-based filler metals”. Welding Journal, vol. 68,
no. 3. March 1989. pp. 63-71.

[7] Miyazawa, Y., and Ariga, T. “A study of the brazeability of nickel-based brazing filler
metal foil for joining nickel base metal to mild steel base metal”. Welding Journal, vol.
71, no. 7. July 1993. pp. 294s-300s.

[8] Kavishe, F.P.L., and Baker, T.J. “Influence of joint gap width on strength and fracture
toughness of copper brazed steels”. Materials Science and Technology, vol. 6, no. 2.
February 1990. pp. 176-181.

[9] Lugscheider, E., and Krappitz, H. “The influence of brazing conditions on the impact
strength of high-temperature brazed joints”. Welding Journal, vol. 65, no. 10. October
1986. pp. 261s-267s.

[10] Draugelates, U., and Hartmann, K.-H. “Behaviour of brazed nickel alloy under cyclic
and thermal load”. Welding Journal, vol. 57, no. 10. October 1978. pp. 298s-302s.

[11] Tung, S.K., Lim, L.C., and Lai, M.O. “Solidification phenomena in nickel base brazes
containing boron and silicon”. Scripta Materialia, vol. 34, no. 5. March 1996. pp. 763-
769.

[12] Johnson, R. “Microstructural aspects of brazing a ferritic steel with two Ni-Si-B braze
filler metals”. Welding Journal, vol. 57, no. 4. April 1978. pp. 93s-102s.

[13] Gale, W.F., and Wallach, E.R. “Influence of isothermal solidification on microstructural
development in Ni-Si-B filler metals”. Materials Science and Technology, vol. 7, no. 12.
December 1991. pp. 1143-1148.

184
[14] Lasalmonie, A. “Diffusion brazing of nickel based superalloys: some basic aspects”.
Annales de Chimie France, vol. 12, no. 3. 1987. pp. 247-257.

[15] Savage, E.I., and Kane, J.J. “Microstructural characterisation of nickel braze joints as a
function of thermal exposure”. Welding Journal, vol. 63, no. 10. October 1984. pp.
316s-323s.

[16] Tung, S.K., Lim, L.C., and Lai, M.O. “Microstructural evolution and control in BNi-4
brazed joints of nickel 270”. Scripta Metallurgica et Materialia, vol. 33, no. 8. October
1995. pp. 1253-1259.

[17] Lecomte-Martens. Private communication (repair company C.R.M., Belgium).

[18] Elder, J.E., Thamburaj, R., and Patnaik, P.C. “Braze repair of MA754 aero gas turbine
engine nozzles”. Proceedings of the ASME Gas Turbine and Aeroengine Congress and
Exposition (Paper No. 89-GT-235). 4-8 June 1989. Toronto, Canada.

[19] Rose, D.P., Price, G., and Thyssen, J. “Development and evaluation of a repair for
saddle cracks in a radial flow turbine wheel”. Proceedings of the Conference on Life
Assessment & Repair: Technology for Combustion Turbine Hot Section Components.
17-19 April 1990. Phoenix, Arizona, United States of America. pp. 271-286.

[20] Lee, J.W., McMurray, H.J., and Miller, J.A. “Development of a new brazing technique
for repair of turbine engine components”. Welding Journal, vol. 64, no. 10. October
1985. pp. 18-21.

[21] Hoppin, G.S., and Berry, T.F. “Activated Diffusion Bonding”. Welding Journal, vol.
49, no. 11. November 1970. pp. 505s-509s.

[22] Demo, W.A., and Ferrigno, S.J. “Brazing method helps repair aircraft gas-turbine
nozzles”. Advanced Materials & Processes, vol. 131, no. 3. March 1992. pp. 43-45.

[23] Young, W.R. “Turbine airfoil repair”. Proceedings of the Conference on Materials
Synergisms. October 1978. Kiamesha Lake, New York, United States of America. pp.
924-938.

[24] Knotek, O., and Lugscheider, E. “Brazing filler metals based on reacting Ni-Cr-B-Si
alloys”. Welding Journal, vol. 55, no. 10. October 1976. pp. 314s-318s.

[25] Van Esch, H. “Braze repair techniques for gas turbine blades and vanes”.
Turbomachinery International, vol. 27, no. 7. September 1986. pp. 29-32.

[26] Su, C.Y., Chou, C.P., Wu, B.C., and Lih, W.C. “Microstructural characterisation of
transient liquid phase zone of activated diffusion brazed nickel base superalloy”.
Materials Science and Technology, vol. 15, no. 3. March 1999. pp. 316-322.

[27] Chasteen, J.W., and Metzger, G.E. “Brazing of Hastelloy X with wide clearance butt
joints”. Welding Journal, vol. 58, no. 4. April 1979. pp. 111s-117s.

185
[28] Liburdi, J., Lowden, P., and Ellison, K. “Powder metallurgy repair technique.” US
patent number US5156321. 27 August 1990 (application date).

[29] Jahnke, B., and Demny, J. “Microstructural investigations of nickel-based repair coating
processed by liquid phase diffusion sintering”. Thin Solid Films, vol. 110, no. 3.
December 1983. pp. 225-235.

[30] Miglietti, W.M., Curtis, R., Hall, B., and Lazarin, C.M. “Liquid phase diffusion bond
repair of Westinghaus 501F, row 3 vanes”. Proceedings of the ASME Turbo Expo
Conference 2000. 2000. Münich, Germany.

[31] Miglietti, W.M., Kearney, J., and Pabon, L. “Liquid phase diffusion bond repair of
Siemens V84.2, row 2 vanes and Alstom Tornado 2nd stage stator segments”.
Proceedings of the ASME Turbo Expo Conference 2001. 2001. New Orleans, United
States of America.

[32] Miglietti, W.M.A. “Wide gap diffusion braze repair of Co-based industrial turbine
vanes”. Proceedings of the International Brazing & Soldering Conference. 2-5 April
2000. Albuquerque, New Mexico, United States of America. pp. 476-485.

[33] Miglietti, W.M.A. “Wide gap diffusion braze repair of Ni-based industrial turbine
vanes”. Proceedings of the 6th International Conference on Brazing, High Temperature
Brazing and Diffusion Welding. DVS Berichte, no. 212. 8-10 May 2001. Aachen,
Germany. pp. 107-112.

186
 APPENDIX A
MIGLIETTI, W.M., CURTIS, R., HALL, B., AND LAZARIN, C.M. “LIQUID PHASE
DIFFUSION BOND REPAIR OF WESTINGHAUS 501F, ROW 3 VANES”.
PROCEEDINGS OF THE ASME TURBO EXPO CONFERENCE 2000. 2000. MÜNICH,
GERMANY.

187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
 APPENDIX B
MIGLIETTI, W.M., KEARNEY, J., AND PABON, L. “LIQUID PHASE DIFFUSION
BOND REPAIR OF SIEMENS V84.2, ROW 2 VANES AND ALSTOM TORNADO 2ND
STAGE STATOR SEGMENTS”. PROCEEDINGS OF THE ASME TURBO EXPO
CONFERENCE 2001. 2001. NEW ORLEANS, USA.

203
204
205
206
207
208
209
210
211
212
213
214
215
216
 APPENDIX C
MIGLIETTI, W.M.A. “WIDE GAP DIFFUSION BRAZE REPAIR OF CO-BASED
INDUSTRIAL TURBINE VANES”. PROCEEDINGS OF THE INTERNATIONAL
BRAZING & SOLDERING CONFERENCE. 2-5 APRIL 2000. ALBUQUERQUE, NEW
MEXICO, USA. PP. 476-485.

217
218
219
220
221
222
223
224
225
226
227
 APPENDIX D
MIGLIETTI, W.M.A. “WIDE GAP DIFFUSION BRAZE REPAIR OF NI-BASED
INDUSTRIAL TURBINE VANES”. PROCEEDINGS OF THE 6TH INTERNATIONAL
CONFERENCE ON BRAZING, HIGH TEMPERATURE BRAZING AND DIFFUSION
WELDING. DVS BERICHTE, NO. 212. 8-10 MAY 2001. AACHEN, GERMANY. PP.
107-112.

228
229
230
231
232
233
234