University of Southern California Dissertations and Theses

High temperature creep behaviors of additively manufactured IN625

(USC Thesis Other)

High temperature creep behaviors of additively manufactured IN625

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

High Temperature Creep Behaviors of Additively Manufactured IN625

by

Kwangtae Son

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)

May 2021

ii

DEDICATION

Whole-heartedly, and gratefully so,
I dedicate this work to my parents,

Jae-Young and Seo-Hyun

모든 마음과 감사를 담아
이 책을 제 부모님께 헌정합니다

손재영, 신서현 님께 헌정됨

iii

ACKNOWLEDGEMENTS

In the Summer of 2016, I met Dr. Michael Kassner in a special class he delivered at Inha University.
It was the first step for me to begin my Ph.D. life at University of Southern California under his
guidance. He, as a Professor in Materials Science, guided me to find proper ways whenever I was
unable to decide the flow of my research project. One important guidance I received from him was
how to write a good-quality research paper. His advice on my papers submitted to journals has
given me expert insight to judge whether the content of papers sounds scientifically reasonable.
The support and encouragement I received from Dr. Kassner made the difference in my Ph.D.
completion. I was always inspired to push building a strong confidence in my ability as an engineer.
It also gives me great pleasure to extend my gratitude to Dr. Thien Phan, working at National
Institutes of Standards and Technologies. I’d like to thank for him with his cooperation to fabricate
samples for my study.

Next, I’d like to thank my other committee members for their acceptance to guide my thesis. I am
also extremely grateful for the time set aside to hear my defense. My three additional committee
members are Dr. Edward Goo, an Associate Professor in Materials Science; Dr. Alejandra Uranga,
an Assistant Professor in Aerospace and Mechanical Engineering; and Dr. Lessa Grunenfelder, a
Senior Lecturer in Materials Science.

As a natural Korean, everything I experienced in the United States was interesting and meaningful.
I believe those experience I gained during the duration of my Ph.D. will be precious spiritual assets
to overcome adversities I will meet during my lifetime.
iv

이어서, 제 박사과정 중 아낌없는 정신적인 지원과 변함없는 믿 음을 주신 부모님께 이
책을 헌정하고자 합니다. 이 면을 빌어서 자 랑스러운 여동생 세희에게도 고맙다고 전하고
싶습니다. 우리 가족 모두 지금까지 처럼 건강하고 행복하게 지 냈으면 합니다. 그리고
나랑 즐거움과 추억을 많이 공유한 내 친구 헌주에게 정말로 고맙 고 앞으로도 지금까지
처럼 지내자. 지금 목표한 바를 반드시 이루 길 바란다.

아울러, 제가 연구자의 길을 선택하도록 도와주신 인하대학교 현승균 교수님께도 감사를
드리고 싶습니다. 석사과정 중 지도를 해주신 덕택으로 먼 타국에서 박사과정까지 마칠
수 있었습니다. 한국 방문연구 시 아낌없는 지원을 해주신 덕분 에 예상보다 훨씬 빨리
박사과정을 마칠 수 있 었습니다.
제가 연구의 걸음마를 시작할 때 연구자로써 구실할 수 있게 가르쳐 준 지운이 형에게도
감사의 말을 전하고 싶습니다. 상욱이형과 현우형도 같이 보낸 시간만큼 서로에게 좋은
영감과 의지가 되어서 고맙습니다. 가끔 내가 어려운 부탁을 할 때 군말없이 도와주었던
명훈이, 경서, 창희, 재철이도 정말로 고맙다. 친정에서 도움이 없 었더라면 박사 과정을
마치는데 훨씬 많은 시 일이 소요되었을 겁니다. 다시한번 모두에게 감사드립니다.

한국 방문 연구 간 실 질적으로 제 연구를 지 원해주신 이기안 인하 대 교수님께도 감사의
인사를 전달하고자 합 니다. 교수님께서 제 장기 한국 방문을 승 인하지 않으셨다면, 제
박사 졸업은 훨씬 요원 한 일이 되었을 겁니다. 제 방문 연구에서 행 정일을 도맡아 해결해
주었던 규식이 형과 태훈씨에게도 감사드립니다. 가끔씩 연구실에 갈 때마다 반갑게
이야기를 나눠준 민 석이, 영균이, 태현이, 성준이 등 이기안 교수님 연구실 모든
인원들에게 감사드립니다.

또한, 제가 불쑥 부탁 을 해도 가능한 한 최 대로 제 연구를 도와 준 남석이형, 상훈이형,
재진이 형, 성규 씨 등 한국 생산기술 연구원 분들께도 정말로 감사드립니다.

석사과정 때부터 수많 은 술자리와 해프닝(?) 을 같이 했던 경수형, 광선이형, 승백이형,
규현이형, 신우형, 봉중이, 태희, 승현이, 규호 등 경합금 멤버들도 건승하길 바랍니다.
영우와 윤호 등 박현순 교수님 방 친구들도 모든 게 잘 풀리길 바 란다.

v

TABLE OF CONTENTS
DEDICATION .............................................................................................................................................. ii
ACKNOWLEDGEMENTS ......................................................................................................................... iii
LIST OF TABLES ..................................................................................................................................... viii
LIST OF FIGURES ..................................................................................................................................... ix
ABSTRACT ............................................................................................................................................... xiv
1. INTRODUCTION ................................................................................................................................ 1
1.1. Ni-base Superalloys ...................................................................................................................... 1
1.1.1. Inconel 625 (IN625) .............................................................................................................. 3
1.1.1.1. Secondary Phases in IN625 ........................................................................................... 3
1.1.1.2. Solid Strengthening in IN625 ....................................................................................... 8
1.1.1.3. Precipitation Hardening in IN625 ............................................................................... 10
1.1.2. Embrittlement Mechanisms in Ni based superalloys .......................................................... 14
1.1.2.1. Oxygen Induced Embrittlement .................................................................................. 14
1.1.2.2. Sulfur Induced Embrittlement ..................................................................................... 16
1.1.2.3. Strain Aging Effect ..................................................................................................... 17
1.2. Metal Additive Nanufacturing .................................................................................................... 18
1.2.1. Laser Powder Bed Fusion (LPBF) ...................................................................................... 19
1.2.2. Processing Parameters in LPBF .......................................................................................... 22
1.2.2.1. Laser Energy Density .................................................................................................. 22
1.2.2.2. Effects of Other Processing Parameters ...................................................................... 23
1.2.3. Mechanical Properties of AM IN625 Compared to Wrought IN625 .................................. 24
1.2.4. Microstructure Evolution of AM IN625 ............................................................................. 27
1.3. High Temperature Creep of IN625 ............................................................................................. 29
1.3.1. Creep Deformation .............................................................................................................. 29
1.3.2. Creep Deformation Mechanisms ........................................................................................ 31
1.3.3. Literature Reviews on The Creep Behavior of IN625 ........................................................ 35
2. EXPERIMENTAL METHODS ......................................................................................................... 43
2.1. LPBF Sample Fabrication Process .............................................................................................. 43
2.1.1. Powder Characterization .......................................................................................................... 43
2.2. High Temperature Creep Test ..................................................................................................... 52
vi

2.2.1. Determination of Temperature and Applied Stress Conditions .......................................... 52
2.2.2. Long Term Cyclic Heat Treatment (LHT) at 650 ˚C ............................................................... 55
2.2.3. Creep Testing ........................................................................................................................... 58
2.2.3.2. Data Acquisition System ................................................................................................... 61
2.2.3.3. Other Conditions in The Creep testing System ................................................................. 63
2.2.3.4. Creep Testing Procedure ................................................................................................... 65
2.3. Chemical Composition Analysis ................................................................................................. 67
2.3.1. Optical Emission Spectrometer ........................................................................................... 69
2.3.2. Spark Spectrometer ............................................................................................................. 70
2.3.3. C/S Analyzer ....................................................................................................................... 71
2.3.4. Oxygen and Nitrogen Determination .................................................................................. 73
2.4. Microstructural Characterization Methods ................................................................................. 75
2.4.1. Sample Sectioning............................................................................................................... 75
2.4.2. Sample Mechanical Polishing ............................................................................................. 76
2.4.3. Optical Micrograph ............................................................................................................. 80
2.4.4. Scanning Electron Microscopy (SEM) ............................................................................... 80
2.4.5. Electron Backscattered Diffraction (EBSD) ....................................................................... 83
2.4.6. Transmission Electron Microscope (TEM) ......................................................................... 87
2.4.7. Secondary Phase and Porosity Quantification Method ....................................................... 90
2.5. Other Material Characterization Methods ................................................................................... 93
2.5.1. Electron Probe Microanalysis – Wavelength Dispersion X-ray Spectroscopy ................... 93
2.5.2. Nano-Secondary Ion Mass Spectroscopy............................................................................ 96
2.5.3. Atom Probe Tomography .................................................................................................. 100
2.5.4. X-ray Diffraction Analysis ................................................................................................ 102
3. Part Ⅰ. HIGH TEMPERATURE CREEP BEHAVIORS OF AM IN625 ....................................... 104
3.1. Creep Curves of Non-LHT AM IN625 and Wrought IN625 .................................................... 104
3.2. Creep Curves of LHT AM IN625 and Wrought IN625 ............................................................ 110
3.3. Creep Strength, Ductility and Mechanism of AM IN625 ......................................................... 115
3.3.1. High Temperature Creep Mechanism of AM IN625 ........................................................ 120
3.3.2. Extra Creep Strength of AM IN625 .................................................................................. 123
4. Part Ⅱ. MICROSTRUCTURE EVOLUTION OF CREEP TESTED AM IN625 .......................... 131
4.1. Microstructure of Non-creep Deformed AM IN625 and Wrought IN625 ................................ 131
4.1.1. Grain Size Evolution of Non-creep Deformed AM IN625 and Wrought IN625 Samples over
the 1 Year LHT ................................................................................................................................ 132
vii

4.1.2. Grain Boundary Misorientation Distribution of Non-creep Deformed IN625 Samples over
the 1 Year LHT ................................................................................................................................ 134
4.1.3. Microstructure Observation of Non-creep Deformed AM IN625 and Wrought IN625 by
Using SEM and TEM ....................................................................................................................... 136
4.1.4. Secondary Phase Identification of As-HIPed AM IN625 and As-solution Annealed Wrought
IN625 141
4.2. Microstructure of Creep-deformed AM IN625 and Wrought IN625 at 650 ˚C and 800 ˚C ..... 146
4.2.1. Microstructural Observation of Creep Deformed Non-LHT SCT AM IN625 and Wrought
IN625 at 650 ˚C and 800 ˚C ............................................................................................................. 146
4.2.2. Microstructural Observations of Creep Deformed LHT AM IN625 and Wrought IN625 at
650 ˚C and 800 ˚C ............................................................................................................................ 154
4.2.2.1. Microstructural Observations of Creep Deformed 6 Months LHT AM IN625 and
Wrought IN625 at 650 ˚C and 800 ˚C .......................................................................................... 154
4.2.2.2. Microstructural Observation of Creep Deformed 1 Year LHT AM IN625 and Wrought
IN625 Samples ............................................................................................................................. 165
4.3. Precipitation Behavior in Both AM IN625 and Wrought IN625 Before/After Creep Tests ..... 172
4.3.1. Matrix γ’’ and δ Volume Fraction Evolution and γ’’ Coarsening Observation ................ 172
4.3.2. Intergranular Precipitation of Secondary Precipitates ....................................................... 175
4.3.3. Discussions on the Expedited Precipitation Behavior in AM IN625 ................................ 178
5. Part Ⅲ. HIGH TEMPERATURE EMBRITTLEMENTS OF AM IN625 ....................................... 181
5.1. Fracture Analysis for Creep Tested AM IN625 and Wrought IN625 Specimens..................... 182
5.1.1. Fractographs of Ruptured AM IN625 and Wrought IN625 Samples ............................... 183
5.1.2. EPMA-WDS Analysis on Secondary Cracks of Ruptured AM IN625 and Wrought IN625
Samples 186
5.1.3. EBSD Mapping for Secondary Cracks of Creep Tested AM IN625 ................................ 191
5.2. Chemical Characterization for Grain Boundary Regions in AM IN625 Specimens ................ 193
5.2.1. Nano-SIMS Results for Both AM IN625 and Wrought IN625 in Non-LHT or 6 Months LHT
Conditions ........................................................................................................................................ 194
5.2.2. APT Analysis for Randomly Oriented Grain Boundaries in AM IN625 Samples ........... 198
5.2.3. Discussions on the Embrittlement by Oxygen .................................................................. 201
5.2.4. Discussions on Embrittlement by Sulfur........................................................................... 203
5.3. Other Possibilities Causing the High Temperature Embrittlement in AM IN625 .................... 206
5.3.1. Porosity Measurement for Non-LHT AM IN625 and Wrought IN625 Samples .............. 207
5.3.2. Misfit Strain Effect on the Experimental IN625 Alloys ................................................... 209
6. CONCLUSIONS .............................................................................................................................. 214
REFERENCES ................................................................................................................................. 218
viii

LIST OF TABLES
Table 1-1. Types of the precipitates in alloy IN625 and the detail [14]. ...................................................... 6

Table 1-2. Solid strengthening constants for alloying elements in Nickel [10]. .......................................... 9

Table 1-3. Values of the stress exponent n, the grain size exponent p and diffusivity D corresponding to
creep mechanisms [98]. .............................................................................................................................. 33

Table 1-4. Specimen creep test conditions from references. ...................................................................... 36

Table 1-5. Chemical composition of reference materials. .......................................................................... 37

Table 2-1. Parameter values of LPBF process of this study....................................................................... 48

Table 2-2. The chemical composition data of specimens of this study in wt% ......................................... 68

Table 2-3. Parameter settings for each step of the automatic polishing ..................................................... 79

Table 3-1. The measured extra strength and ductility changes for creep deformed IN625 alloys in this study.
SCT wrought IN625 was used as reference to assess the extra strength and ductility changes. Only 650 ˚C,
658 MPa and 800 ˚C, 192 MPa creep data were used. .............................................................................. 119

Table 3-2. The measured strength and γ’’ (650 ˚C) or δ (800 ˚C) hardening calculation for creep deformed
IN IN625 samples, SCT wrought IN625 was used as reference to assess the extra strength and ductility
changes. Unlike Table 3-1, all available creep data was used for measuring extra creep strength. .......... 126

Table 3-3. Calculated (by the XLPA method) and measured (by TEM observations) values of dislocation
densities of three different non-creep tested samples. .............................................................................. 130

Table 4-1. Summary of phase analysis, new phases and microstructures formed after creep tests were
shaded in grey. The types of secondary phases precipitated before and after creep tests are mostly the same
in those 6 months and 1 year long-term heat treated wrought IN625 and AM IN625, except for one
exception of the remaining γ ’ ’ in the 800 ˚C creep tested 1 year LHT wrought IN625 (marked in yellow).
.................................................................................................................................................................. 171

ix

LIST OF FIGURES
Figure 1-1. (a) Content distribution of materials in an aircraft (b) A schematic structure of a typical aircraft
engine and temperature profile in the engine [7]. ......................................................................................... 2

Figure 1-2. The Time-Temperature-Transformation Diagram for precipitates at elevated temperatures in
wrought IN625 [13,14,17]. ........................................................................................................................... 7

Figure 1-3. A schematic of LPBF machine [75]. ....................................................................................... 21

Figure 1-4. The mechanical properties comparison between solution annealed wrought IN625 and AM
HIPed IN625 at various temperatures (a) yield strength and ultimate tensile strength (b) ductility. .......... 26

Figure 1-5. A typical creep curve at a constant stress condition. ............................................................... 30

Figure 1-6. Dependence of Young’s modulus on temperature in IN625. .................................................. 39

Figure 1-7. The plot between 𝜀𝜀𝜀𝜀 𝜀𝜀 and 𝜎𝜎 𝜀𝜀 𝜀𝜀 / 𝐸𝐸 of solution treated wrought IN625 and PM IN625 over the
given temperature range. ............................................................................................................................. 40

Figure 2-1. (a) SEM image of IN625 powders for LPBF fabrication (b) The particle size distribution of
powder particles in this study. .................................................................................................................... 45

Figure 2-2. (a) EOSINT M290 LPBF machine (b) as-fabricated LPBF IN625 coupons........................... 49

Figure 2-3. The scanning strategy of LPBF process in this study. ............................................................. 50

Figure 2-4. Temperature and pressure history of HIP and Solution Annealing treatment (S.A.). ............. 51

Figure 2-5. Temperature and air pressure conditions for several aircraft engine parts, T and S represent
temperature and air pressure, respectively [111-113]. ................................................................................ 54

Figure 2-6. A heating profile of the longtime heat treatment cycles for 6 months and 1 year in this study.
.................................................................................................................................................................... 57

Figure 2-7 . The setup of Arcweld creep testing machine used in this study (a) The front view of the testing
machine(b) a schematic of the creep testing machine. ................................................................................ 59

Figure 2-8. The design of the creep test sample used in this study. ........................................................... 64

Figure 2-9. The high temperature creep test procedure of this study. ........................................................ 66

Figure 2-10. An example of the intergranular misorientation determination next to a secondary crack (a)
the EBSD image, (b) misorientation profile along the yellow line in (a). .................................................. 86

Figure 2-11. Examples of image conversion from various microscopy sources (a) an original TEM image
source for γ’’ phase quantification, (b) the binary converted drawing from (a), (c) an original TEM source
for dislocation density measurement, (d) the binary image converted from (c), (e) an original OM image
source, (f) the dark field OM image of (c), (g) an original SEM image source for grain boundary occupation
x

frequency, (h) the binary image of total grain boundary length from (g), (i) the binary image of grain
boundary area occupied by secondary particles from (g). .......................................................................... 92

Figure 2-12. (a) a schematic of WDS analysis principle, (b) X-ray spectra from WDS and EDS analyses
[132]. ........................................................................................................................................................... 95

Figure 2-13. The mass profiles of ions in the Nano-SIMS to show how two similar mass-to-charge ratio
ions are separated (a)
32
S
2-
and
32
O 2
2-
ions, (b)
43
BO 2
2-
and
43
AlO
-
ions. ..................................................... 98

Figure 2-14. SEM and EBSD grain boundary maps of 650 ˚C SCT AM IN625, which shows precipitates
prefer to form at random boundaries and how to specify random grain boundaries (non-special) using
EBSD. ......................................................................................................................................................... 99

Figure 3-1. Several representative creep curves for non-LHT samples at 650 ˚C (a) AM HIPed IN625 at
294 MPa, (b) AM HIPed IN625 at 466 MPa, (c) wrought IN625 at 294 MPa, (d) wrought IN625 at 466
MPa. The number on the bottom right is the minimum creep rate (= steady-state creep rate). ................ 106

Figure 3-2. Several representative creep curves for non-LHT samples at 800 ˚C (a) AM HIPed IN625 at 65
MPa, (b) AM HIPed IN625 at 192 MPa, (c) wrought IN625 at 65 MPa, (d) wrought IN625 at 192 MPa,
The blue dash line in (d) is an extrapolated curve and the number on the bottom right is the minimum creep
rate (= steady-state creep rate). The X mark suggests where a rupture occurred. ................................... 107

Figure 3-3. Several representative creep curves for LHT samples at 650 ˚C (a) 6 months LHT AM HIPed
IN625 at 294 MPa, (b) 6 months LHT AM HIPed IN625 at 658 MPa, (c) 1 year LHT AM HIPed IN625 at
658 MPa, (d) 6 months LHT wrought IN625 at 658 MPa, (e) 1 year LHT wrought IN625 at 658 MPa. The
number on the bottom right is the minimum creep rate (= steady-state creep rate), the blue dash line in (e)
is an extrapolated curve, X mark suggests where a rupture occurred or would have occurred. ............... 112

Figure 3-4. Several representative creep curves for LHT samples at 800 ˚C (a) 6 months LHT AM HIPed
IN625 at 65 MPa, (b) 6 months LHT AM HIPed IN625 at 192 MPa, (c) 1 year LHT AM HIPed IN625 at
192 MPa, (d) 6 months LHT wrought IN625 at 192 MPa, (e) 1 year LHT wrought IN625 at 192 MPa. The
number on the bottom right is the minimum creep rate (= steady-state creep rate), the blue dash line in (e)
is an extrapolated curve. The X mark suggests where a rupture occurred or was anticipated to occur. ... 113

Figure 3-5. (a) minimum (= steady-state) creep rate versus the modulus compensated (normalized) stress.
(b) creep ductility at 650 ˚C (c) creep ductility at 800 ˚C. The ductility at the two test temperatures is
indicated by solid circle dots for SCT samples, X marks for 6 months LHT samples and + marks for 1 year
LHT samples. ............................................................................................................................................ 116

Figure 4-1. Grain size histograms of AM IN625 and wrought IN625 alloys over the 1 year LHT presented
in two different directions of each sample. Building direction is along the z axis in Fig. 2-3. ................. 133

Figure 4-2. Grain boundary misorientation histograms of AM IN625 and wrought IN625 alloys over the 1
year LHT presented in two different directions of each sample. Building direction is along the z axis in Fig.
2-3. ............................................................................................................................................................ 135

Figure 4-3. SEM images of non-creep deformed IN625 samples (a) as-HIPed AM IN625, (b) as-solution
treated wrought IN625, (c) 6 months LHT AM HIPed IN625, (d) 6 months LHT wrought IN625 after
xi

solution treatment, (e) 1 year LHT AM HIPed IN625, (f) 1 year LHT wrought IN625 after solution
treatment. .................................................................................................................................................. 137

Figure 4-4. TEM observations after 6 months and 1 year LHT of AM and wrought IN625 samples (a)
matrix region of 6 months LHT AM IN625, (b) grain boundary region of 6 months AM IN625, (c) matrix
region of 1 year LHT AM IN625, (d) grain boundary region of 1 year LHT AM IN625, (e) matrix region
of 6 months LHT wrought IN625, (f) grain boundary region of 6 months LHT wrought IN625, (g) matrix
region of 1 year LHT wrought IN625, (h) grain boundary region of 1 year LHT wrought IN625 (i) Scanning
TEM of 6 months LHT wrought IN625 with the EDS point spectra of the M 6C carbide, (j) Scanning TEM
image of 6 months LHT AM IN625 with EDS point spectra of the M 23C 6 carbide; (i) and (j) show how
M 23C 6 and M 6C were identified using the EDS analysis [42-44]. The specimens were undeformed. ...... 139

Figure 4-5. TEM analyses of AM IN625 alloys (a) a TEM sample of AM HIPed IN625, (b) a TEM sample
of AM as-solution annealed IN625, (c) a magnified image of particles in (a), (d) EDS line profile for the
particles in (c), (e) one other magnified image of particles in (b), (f) SAED pattern of the red circle area in
(e) and its key. ........................................................................................................................................... 142

Figure 4-6. EDS mapping result for MC carbides in as-solution annealed wrought IN625. ................... 145

Figure 4-7. SEM images of SCT AM IN625 and wrought IN625 (a) AM HIPed IN625 at 650 ˚C and 416
MPa, (b) AM HIPed IN625 at 800 ˚C and 66 MPa, (c) wrought IN625 at 650 ˚C and 416 MPa, (d) wrought
IN625 at 800 ˚C and 66 MPa. ................................................................................................................... 147

Figure 4-8. TEM observations of AM IN625 and wrought IN625 samples after SCT at 650 ˚C, 416 MPa
(a) matrix area of AM HIPed IN625, (b) grain boundary area of AM HIPed IN625, (c) matrix area of
wrought IN625, (d) grain boundary area of wrought IN625. .................................................................... 149

Figure 4-9. TEM observations of AM IN625 and wrought IN625 samples after SCT at 800 ˚C, 104 MPa
(a) matrix region of AM HIPed IN625, (b) grain boundary area of AM HIPed IN625, (c) matrix region of
wrought IN625, (d) grain boundary area of wrought IN625. .................................................................... 150

Figure 4-10. STEM analysis of IN625 alloys after the 24 hours creep test at 800 ˚C, 104 MPa with EDS
element maps (a) AM HIPed IN625 sample (b) wrought IN625 sample, the red squares indicate the mapping
area where the EDS mapping was conducted. .......................................................................................... 153

Figure 4-11. SEM images of 6 months LHT AM IN625 and wrought IN625 after creep tests (a) AM HIPed
IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c) wrought IN625 at 650 ˚C
and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa; the magnification of (a) is equal to (c) and, the
magnification of (b) is equal to (d). Grain boundaries in all samples are occupied with δ, M23C6 and M6C.
Matrix δ is evident in those 800 ˚C creep tested samples. ........................................................................ 155

Figure 4-12. TEM of 6 months LHT IN625 samples after creep tests at 650 ˚C, 658 MPa (a) matrix region
of AM IN625 (b) intergranular region of AM IN625 (c) matrix region of wrought IN625 (d) intergranular
region of wrought IN625. The microstructure is generally same between AM IN625 and wrought IN625
except for the volume fraction of γ’’. ....................................................................................................... 158

Figure 4-13. TEM of 6 months LHT IN625 samples after creep tests at 800 ˚C, 192 MPa (a) matrix region
of AM IN625 (b) intergranular region of AM IN625 (c) matrix region of wrought IN625 (d) intergranular
xii

region of wrought IN625. Large δ phase formation was observed in both AM IN625 and wrought IN625.
Only wrought IN625 shows the sub-grain boundary formations. ............................................................. 159

Figure 4-14. STEM analysis of 6 months LHT IN625 alloys after the creep tests at 650˚C, 658 MPa with
EDS element maps (a) AM HIPed IN625 sample (b) wrought IN625 sample, the red squares indicate the
mapping area where the EDS mapping was conducted. ........................................................................... 160

Figure 4-15. STEM analysis of 6 months LHT IN625 alloys after creep tests at 800 ˚C, 192 MPa with EDS
element maps (a) AM HIPed IN625 sample (b) wrought IN625 sample, the red squares indicate the mapping
area where the EDS mapping was conducted. .......................................................................................... 161

Figure 4-16. SEM images of 1 year LHT AM IN625 and wrought IN625 after creep tests (a) AM HIPed
IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c) wrought IN625 at 650 ˚C
and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa; the magnification is the same for all images in
this figure. Grain boundaries in all samples are occupied with δ, M23C6 and M6C. Matrix δ is evident in
those 800 ˚C creep tested samples. ........................................................................................................... 166

Figure 4-17. TEM of 1 year LHT AM IN625 and wrought IN625 after creep testing (a) AM HIPed IN625
at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c) wrought IN625 at 650 ˚C and
658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa. ............................................................................... 167

Figure 4-18. STEM results for 1 year LHT IN625 after creep testing with EDS spectra data (a) AM HIPed
IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c) wrought IN625 at 650 ˚C
and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa. The EDS analysis showed the presence of M 23C 6
and M 6C carbides on grain boundary area in those suggested creep tested samples of AM IN625 and
wrought IN625. ......................................................................................................................................... 168

Figure 4-19. Matrix secondary phase evolution in AM IN625 and wrought IN625 (a) volume fraction
development of γ’’ and δ under various thermal conditions (b) the coarsening behavior of γ’’ at 650 ˚C
under various conditions with the Nb content change in matrix; C.T. = creep tested. The volume fraction of
secondary phases in AM IN625 is generally larger than wrought IN625 in all thermomechanical conditions.
The coarsening of γ’’ is also faster in AM IN625 at 650 ˚C. .................................................................... 173

Figure 4-20. Grain boundary particle frequency content evolution in AM IN625 and wrought IN625 under
various testing conditions. ........................................................................................................................ 176

Figure 5-1. SEM fractographs of failed IN625 samples after creep tests [(a),(b),(c),(d)] SCT creep tested
samples, [(e),(f),(g),(h)] 6 months LHT samples, and [(i),(j),(k),(l)] 1 year LHT samples. The information
on the top-right of each figure gives the sample type whether AM IN625 or wrought IN625 with its creep
testing conditions. R.T. on the bottom-right is the rupture time. The window in each image is the SEM
image of the same sample under a lower magnification. The scale bars for the images are shown in (c).
.................................................................................................................................................................. 184

Figure 5-2. EPMA-WDS analysis for the secondary cracks in creep tested AM IN625 samples (a) SCT
sample at 650 ˚C and 658 MPa, (b) SCT sample at 800 ˚C and 192 MPa, (c) 6 months LHT sample at 650
˚C and 658 MPa, (d) 6 months LHT sample at 800 ˚C and 192 MPa. ...................................................... 187

xiii

Figure 5-3. EPMA-WDS analysis for the secondary cracks in creep tested wrought IN625 samples (a) SCT
sample at 650 ˚C and 658 MPa, (b) SCT sample at 800 ˚C and 192 MPa, (c) 6 months LHT sample at 650
˚C and 658 MPa, (d) 6 months LHT sample at 800 ˚C and 192 MPa. ...................................................... 188

Figure 5-4. The EBSD maps for secondary cracks in creep deformed AM IN625 samples (a) 650 ˚C and
658 MPa for 2.6 h (b), 800 ˚C and 192 MPa for 6.4 h (c) 6 months LHT and creep tested at 650 ˚C, 658
MPa for 20 h (d) 6 months LHT and creep tested at 800 ˚C, 192 MPa for 11.4 h, and (e) the misorientation
profile of grain boundaries where secondary cracks formed. Grain boundaries are shown in white lines and
the black regions are cracked. All cracks in this figure propagated along the perpendicular direction to the
loading direction and all of them are intergranular. .................................................................................. 192

Figure 5-5. Nano-SIMS results for non-LHT IN625 alloys (a) as-HIPed IN625, (b) as-solution annealed
IN625. All samples were undeformed. Lighter regions in each Nano-SIMS figure indicate higher
concentration of ions.
27
Al
16
O signal represents the position of Al2O3 particles and sulfur is highly
concentrated at where Al 2O 3 particles are located in the AM IN625 sample. No strong sulfur segregation is
not observed in the wrought IN625........................................................................................................... 195

Figure 5-6. Nano-SIMS results for 6 months LHT IN625 alloys (a) 6 months LHT AM IN625, (b) 6 months
wrought IN625. All samples were undeformed. Lighter regions in each Nano-SIMS figure indicate higher
concentration of ions.
27
Al
16
O signal represents the position of Al2O3 particles and sulfur is highly
concentrated at where Al 2O 3 particles are located in the AM IN625 sample. No strong sulfur segregation is
not observed in the wrought IN625........................................................................................................... 196

Figure 5-7. The chemical distribution plot along given distances from the grain boundaries and γ’’ interface
in APT samples: (a) the creep deformed AM IN625 sample at 294 MPa and 650 ˚C within 24 h (b) the
creep deformed 6 months LHT AM IN625 sample at 658 MPa and 650 ˚C (c) concentration profile near γ
’ ’ particles in (b). The shaded area are grain boundary regions or γ ’’ precipitates. .................................. 199

Figure 5-8. The grain boundary character maps for several non-creep deformed AM IN625 and wrought
IN625 samples. CSL (Σ3) and RHAB fraction of each sample is given in the title. ................................ 205

Figure 5-9. Optical micrographs of IN625 samples (a) bright field image of as-solution annealed wrought
IN625, (b) dark field image of (a), (c) bright field image of as-solution annealed (non-HIPed) AM IN625,
(d) dark field image of (c), (e) bright field image of as-HIPed AM IN625, (f) dark field image of (e). .. 208

Figure 5-10. XRD analysis for several IN625 samples (a) matrix lattice parameter change in both AM
IN625 and wrought IN625 over the 1 year LHT, (b) XRD line profiles for four LHT IN625 samples: 6
months LHT AM IN625, 1 year LHT AM IN625, 6 months LHT wrought IN625, and 1 year LHT wrought
IN625. ΔV is the volume change. ............................................................................................................. 211

xiv

ABSTRACT

Inconel (IN625) is a type of Ni-based superalloys which has been widely used in the aerospace
field due to its high tensile, creep, and rupture strength. Service temperatures of this alloy range
from cryogenic to approximately 1000 ˚C thanks to its outstanding fatigue and thermal-fatigue
strength with excellent oxidation resistance.

Additive Manufacturing (AM) is the formalized term for what is popularly called 3D printing. AM
has gained attention in the aerospace industry as it provides the possibility for economical and
complicated-configuration engine parts with fewer joining steps, and greater geometric freedom.
Among various AM technologies, laser powder bed fusion (LPBF) has been widely chosen for
metal part fabrication. Metal powder-particles are fully melted and fused by a high energy beam.
Subsequent hot isostatic pressing (HIP) is typically conducted to reduce porosity after the LPBF.

The main interest of this study is to verify that AM IN625 can replace the conventional IN625 for
high temperature applications. High temperature creep tests of additively manufactured (AM)
nickel-based superalloy 625 (IN625) and wrought IN625 were conducted at 650 ˚C and 800 ˚C
over the stress range of 65 MPa to 658 MPa. Thermal treatments were conducted for both AM and
wrought IN625 samples prior to creep testing: either solution heat-treated or hot isostatically
pressed and, additionally, long-term cyclic heat-treatments (LHT) at 650 ˚C for 6 months and 1
year. The purpose of long-term cyclic heat-treatments was to simulate a near-operational
environment of one real application of IN625 over a target worktime (= 1 year).
xv

AM IN625 showed equal or even higher creep strength than wrought IN625 for all creep tests.
However, AM IN625 exhibited poor ductility compared to wrought IN625 under all creep testing
conditions, and the ductility of AM IN625 additionally decreased after the LHT. Some creep
ductility recovery in HIPed AM IN625 was observed, however, it was still much inferior to the
ductility of wrought IN625. Microstructural observations revealed the intergranular precipitation
rate and types of the precipitates of AM IN625 was generally comparable with that of wrought
IN625, so that the intergranular precipitation cannot explain the poor ductility of AM IN625.
Both AM and wrought IN625 obtained some additional strength after the LHT. The amount of
extra strength in the alloys was generally proportional to the increase in matrix volume fraction of
γ’’ phase (650 ˚C) and δ phase (800 ˚C).
The creep analysis suggests that dislocation climb is the rate controlling mechanism for creep of
both AM IN625 and wrought IN625 in the temperature range of 650 ˚C to 800 ˚C.

Fractography observations and electron backscatter diffraction analysis revealed that brittle
intergranular fracture was evident in AM IN625 samples. Atomic probe tomography (APT) and
nano-Secondary ion mass spectrometer (Nano-SIMS) indicated no remarkable intergranular
impurity (particularly sulfur and oxygen) segregation at grain boundaries of AM IN625 alloys in
both as-HIPed and the 6 months LHT states. However, the Nano-SIMS results clarified that sulfur
segregated to the Al2O3

/matrix interfaces the AM IN625, and the sulfur segregation level at Al2O3
interface increased after the 6 months LHT.
Cracking can occur from these interfaces and repeated sulfur diffusion to the crack tip is the
foremost possibility to explain the poor ductility of AM IN625 within the temperature range tested.
1

1. INTRODUCTION
1.1. Ni-base Superalloys
Superalloys are special high temperature materials and typically used in gas turbine engines such
as an aircraft and power plants. Fe, Ni and Co superalloys have been used for this application due
to their excellent resistance to mechanical and chemical degradation. Among these superalloys,
Ni-based superalloys have emerged as the materials of choice for aircraft engine parts with its
outstanding resistance to loading under static, fatigue and creep environments [1-4]. When cost is
considered first, Fe based ferritic steels can be used, however, ferritic steels are not preferred for
extreme temperature applications such as above of 700˚C [5]. On the other hand, Ti alloys become
the first choice when weight-to-strength ratio is the most important. However, despite its good
weight-to-strength ratio, the application of Ti alloys to aircraft engine parts are also restricted due
to their temperature limit of ≈ 700 ˚C [6].
Figure 1-1a shows a schematic of the material usage distribution in an aircraft engine. It indicates
Ni-based superalloys has been portioned around 40 % in aircraft engines since 60’s. Most of Ni-
based alloy components are used in the back section of an aircraft engine.
Fig. 1-1b delineates a temperature profile in an aircraft engine. Fig. 1-1b shows that the back parts
of the aircraft engine are typically under high temperature condition (> 500 ˚C). As some parts in
the area requires higher temperatures above 700 ˚C, Fe and Ti cannot be placed in the region. Thus,
Ni-based superalloys have been considered as non-replaceable materials for harsh environment
applications like the turbine portions in an aircraft.
2

Figure 1-1. (a) Content distribution of materials in an aircraft (b) A schematic structure of a typical
aircraft engine and temperature profile in the engine [7].

3

1.1.1. Inconel 625 (IN625)
Among various Ni-based superalloys, Inconel 625 (IN625) is valuable to investigate due to its
diverse commercial applications in the aviation industry [8]. Specifically, since it was invented in
the mid 1950’s, IN625 has been utilized as combustion liners, engine exhaust systems and turbine
shroud rings of an aircraft due to its applicable temperature range from cryogenic to roughly 1000
˚C [8,9]. High portion of chromium in this alloy grants not only a thick protective oxide layer
which is stable even at elevated temperature. Some solid strengthening by Cr occurs as well [9,10].
As a representative Ni-Cr-Mo alloy, 8-10 % Mo prevents IN625 from experiencing pitting and
crevice corrosion. Approximately 4 % of Nb in this alloy contributes to be be resistant to
intergranular corrosion [7]. In addition, this material can be coated with Thermal Barrier Coatings
such as Co-Cr-Al-Y based metallic coating or ceramic coatings, which increases the service
temperature of this alloy.

1.1.1.1.Secondary Phases in IN625
The presence of niobium in this material leads to precipitation hardening by the formation of
metastable Ni3Nb (γ”) phase over the temperature of 537˚C [11-14]. This bct precipitate has a
stacking sequence abcdef which is coherent with the FCC matrix of IN625 [11]. The coarsening
of γ’’ is typically faster under higher temperature condition (> 650 ˚C) and longer exposed time to
the elevated temperatures [12]. The stable orthohombic Ni3Nb (δ) phase is typically formed on
grain boundaries at higher temperatures (more than 760 ˚C [13]) and under longer exposure time
to the elevated temperatures (> 650 ˚C) by the solid phase transformation from γ’’ [14]. This
4

acicular δ phase has a relatively small dispersion strengthening effect compared to γ’’ because its
incoherency to Ni matrix, which generally degrades the ductility of IN625 [11,13].

Aside from the two Ni3Nb type precipitates (γ’’ and δ), some carbides are also formed in IN625
under various thermal conditions. First, primary MC type carbide is generally formed in the highest
temperature regime (>1038 ˚C) or during the solidification and this carbide is rich in Ti and Nb
[4,14]. For the nucleation of MC type carbide (M = Metallic elements, C = carbon), ceramic
inclusions, such as TiN, Al2O3, MgO and CaO, provide nucleation sites for MC carbide during the
solidification process [15,16]. MC usually exhibits coarse, blocky or globular morphology (at grain
boundaries), or fine, script microstructure (inside grains) [4,14]. Additionally, this carbide
primarily precipitates on grain boundaries during the solidification process [14].
Between 816 ˚C and 1038 ˚C [4,17], M6C carbide primarily precipitates at grain boundaries [14].
This type carbide is typically rich in Mo and Si [13,18,19]. The morphology of M6C is generally
blocky at grain boundaries. Similar to M6C, M23C6 type carbides prefer to be formed at grain
boundaries and the decomposition of grain boundary MC generally occurs with the formation of
both M6C and M23C6 carbides [14]. Most secondary carbides including M6C and M23C6 are
transformed from initial MC (formed in the molten state) by sequential reactions in the solid
solution state in Ni matrix of superalloys following breakdown of the MC [4]. However, the
presence of MC is not the prerequisite for precipitations of other secondary carbides in Ni base
superalloys. For example, as reported by Li et al [20], grain boundary Cr segregation precedes the
precipitation of Cr-rich M23C6 carbides. M23C6 also has a high preference to form at grain
boundaries as MC and M6C carbides [4,14,21]. It has been reported that the continuous
precipitation of hard carbides at grain boundaries, such as MC, M6C and M23C6, can harm the
5

ductility of IN625 by providing preferential sites for crack nucleation and propagation sites
[4,14,21].
Additionally, all carbides in this section (MC, M6C and M23C6) commonly have a cubic structure,
which has a potential to be coherent to the Ni matrix [14]. Hence, those carbides remain coherent
to Ni matrix when its size is less or equal to 1.5 µm and loses the coherency beyond the size of 1.5
µm [22].
Lastly, Pt2Mo type hexagonal structured laves phase is one other representative secondary phase
in IN625. Similar to M6C carbide, the formation of laves phase in IN625 is highly influenced by
the content of Mo and Si, and the higher Mo and Si levels in IN625 lead to the more precipitation
of laves phase in the matrix [14]. This similarity in the content between M6C and laves is
presumably because the laves phase forms at the surface of M6C carbides in the prolonged thermal
exposure at temperatures between 760 ˚C – 870 ˚C in IN625 after the solid solution treatment [18].
Laves phase can also precipitate during the solidification with MC carbide, however, laves phase
tend to be dissolved into Ni matrix during a solid solution annealing treatment of IN625 [14]. This
acicular phase is primarily located at grain boundaries and provides favorable sites for crack
nucleation and propagation, so that embrittles IN625 [4,14].
Those precipitates introduced in this section are listed in Table 1-1. In addition to that, Figure 1-2
gives the Time-Temperature-Transformation (TTT) diagram for the secondary particles of IN625,
obtained from several sources [13,14,17].
6

Table 1-1. Types of the precipitates in alloy IN625 and the detail [14].
Phase Structure Lattice parameter(nm) Typical Composition
MC Cubic, Fm3m a = 0.43
Matrix blocky MC:
(Ti0.07Cr0.04Fe0.02Ni0.09Nb0.75Mo0.03)C
(Ti0.53Cr0.03Ni0.04Nb0.39Mo0.01)C

Grain Boundary MC:
(Ti0.15Cr0.04Fe0.01Ni0.08Nb0.67Mo0.01)C
M6C Cubic, Fd3m a = 1.13 (Cr0.21Fe0.02Ni0.37Nb0.08Mo0.24Si0.08)6C
M23C6 Cubic, Fm3m a = 1.08 (Cr0.85Fe0.01Ni0.07Mo0.07)23C6
γ” bct, I4/mmm a = 0.36, c = 0.74 Ni3(Nb>0.5 (Ti,Al)<0.5)
δ
Orthohombic,
Pmmm
a = 0.51, b= 0.42, c =
0.45
Ni3Nb
Laves
Hexagonal,
P63/mmc
a = 0.47, c = 0.77 (Cr0.31Fe0.08Ni0.41)2(Si0.17Ti0.01Nb0.19Mo0.63)

7

Figure 1-2. The Time-Temperature-Transformation Diagram for precipitates at elevated
temperatures in wrought IN625 [13,14,17].

8

1.1.1.2.Solid Strengthening in IN625
IN625 has various solutes at either interstitial or substitutional sites in Ni matrix. The difference
between lattice parameters of the host Ni atoms and solute atoms generates a localized lattice
distortion around the solute atom, which can interact with dislocations and generate a subsequent
solid solution strengthening effects [23].

In the case of nickel-based superalloys, the difference in atomic diameter between nickel matrix
and a solute atom needs to be less than 15% to maintain stable solubility [24]. Within the stable
solubility range, the strengthening effect generally goes up as the size difference between a solute
and the host Ni atom increases. In the case that two different solutes have comparable atomic
diameter deviations from the host, one solute provides more strengthening when the separation in
the periodic table between Ni and the solute is greater [25]. The amount of stacking fault drop in
Ni-based superalloys increases when the separation in the periodic table between an alloying
element and the Ni host is larger. Hence, the dislocation dissociation by the decrease in stacking
fault energy can explain some of the solid strengthening effect in Ni-based superalloys [24-26].

Table 1-2 lists the predicted flow stress increment in binary Ni alloys by dissolving various
elements, which helps to approximate the solid strengthening effect on complex Ni-based
superalloys [10].

9

Table 1-2. Solid strengthening constants for alloying elements in Nickel [10].
Alloying elements
Strengthening Constant
(MPa·at%
-1
)
Al 225
Si 275
Zn 386
Ga 310
Ge 332
In 985
Sn 1225
Sb 960
Ti 775
V 408
Zr 2359
Hf 1401
Nb 1183
Ta 1191
Cr 337
Mo 1015
W 977
Mn 448
Fe 153
Ru 1068
Co 39.4
Rh 520
Cu 86.7
C 1061
Pd 492

10

1.1.1.3.Precipitation Hardening in IN625
Aside from the pure solute effect for the matrix strengthening, some elements such as Nb, Al and
Ti contribute to the additional strength of Ni base superalloys by forming precipitates like γ’ and
γ’’ [11-14,17]. Basically, the precipitation hardening by those secondary particles is obtained from
the interaction between the precipitate and dislocations in Ni matrix. In the case of precipitation
hardening of IN625, the main strengthening precipitate is γ’’ [11,12] and the strengthening
mechanism can be categorized into the following four: coherency strengthening [27], anti-phase
boundary (APB) shearing [28], mechanical twinning [29], and dispersion hardening [30].

Coherency strengthening originates from the lattice misfit between the Ni matrix and a coherent
secondary phase. Due to the acceptable lattice parameter mismatch of the precipitate to that of Ni
matrix, an associated stress field is formed around the coherent precipitates, which interacts with
dislocation and hinders the dislocation motion. Hence, there is an increase in strength [27]. The
coarsening of a coherent precipitate leads to the gradual decrease in coherency, therefore the
gradual loss of coherency strengthening. The loss of lattice-misfit strain field generates misfit
dislocations can be introduced along the longer axis of a coarsened plate type precipitate like γ’’
as γ’’ grows in Ni matrix [31]. Under an ideal coherent condition (or acceptable coherent strain
level), the precipitate hardening by coherency strengthening due to γ’’ can be calculated by the
following equation [27]:

∆ 𝜎𝜎 𝑐𝑐𝑐𝑐 ℎ
= 1.7 𝐺𝐺 | 𝜖𝜖 |
3/ 2
�
4 𝑅𝑅 𝑓𝑓 𝛾𝛾 ′′(1 − 𝑓𝑓 ′
)
𝐴𝐴 2
𝑏𝑏 �
1/ 2

(1-1)

11

where ∆ 𝜎𝜎 𝑐𝑐𝑐𝑐 ℎ
is the coherency strengthening, G is the shear modulus of IN625, 𝜖𝜖 is the lattice misfit
along the c axis of γ’’, R is the radius of γ’’, 𝑓𝑓 𝛾𝛾 ′′ is the volume fraction of γ’’, f’ is the fraction of
γ’’ which has the c axis parallel to the Burger’s vector of the Ni matrix and A is the aspect ratio
of γ’’.

The second mechanism for the γ’’ hardening is anti-phase boundary shearing (or order
strengthening). This strengthening requires a shearing of precipitate due to substantial stress
localization ahead of the precipitate. Shearing of an intermetallic like γ’’ leads to the formation of
stacking fault inside the intermetallic, i.e. the formation of anti-phase boundary. In other words,
an anti-phase boundary can be created on the slip plane of the precipitate by matrix dislocations.
As the activation of energetically unfavorable anti-phase boundaries in γ’’ requires very high stress
ahead of precipitates, this anti-phase formation contributes to the extra strength of precipitation-
hardened alloys. In Ni based superalloys, matrix dislocations usually dissociate into partial
dislocations at the precipitate interface, so that slip on the matrix dislocation partial pair restores
perfect order on the slip plane of the precipitate. This precludes the presence of the unstable
precipitate stacking fault after the shearing [32]. For example, in the case of γ’’, a pair of a/2 [11
�
0]
type dislocation slip is required to restore the order of γ’’ when its c-axis is parallel to the [001]
direction in the Ni matrix [28]. Moreover, the reference proposes an equation for anti-phase
boundary strengthening by γ’’ as [28]:

∆ 𝜎𝜎 𝐴𝐴𝐴𝐴 𝐴𝐴 =
𝛤𝛤 𝐴𝐴𝐴𝐴 𝐴𝐴 4 𝑏𝑏 � �
4 𝛤𝛤 𝐴𝐴𝐴𝐴 𝐴𝐴 𝑓𝑓 𝛾𝛾 ``
𝜋𝜋𝜋𝜋
�
√6 𝑅𝑅 2
3 𝐴𝐴 �
1/ 2
�
1/ 2
� 1 + 𝛽𝛽 1/ 2
� − 𝑓𝑓 𝛾𝛾 ′′(1 + 𝛽𝛽 ) � (1-2)

12

where ∆ 𝜎𝜎 𝐴𝐴𝐴𝐴 𝐴𝐴 is the extra strength by APB shearing, 𝛤𝛤 𝐴𝐴𝐴𝐴 𝐴𝐴 is the APB energy of γ’’ (roughly 296
mJ/m
2
[27]), T is line tension of matrix dislocation (approximated as Gb
2
/2) and 𝛽𝛽 is a constant
(typically 1/3 [28]).

The third precipitate hardening mechanism is mechanical twinning. The generation of mechanical
twins in Ni-base superalloys can be rationalized by the increase in the internal friction due to large
amount of precipitates. Even though several different mechanical twinning nucleation models have
been proposed in the precipitate hardened Ni superalloys, all mechanisms require a high internal
friction of Ni matrix as the a/3<112> twinning partial needs high strain level at precipitate
interfaces to be activated [33]. Once mechanical twins are formed in the matrix, the mechanical
twinning contributes to the increase in the material strength. This effect can be explained by both
a reduction of the effective slip length (Hall-Petch effect) and the formation of sessile dialocations
in the twinned region due to the interaction between the matrix dislocation with propagating twins
(Basinski mechanism) [34]. In the case of γ’’, the additional strength due to the mechanical
twinning can be measured by using an equation [29]:

∆ 𝜎𝜎 𝑡𝑡 𝑡𝑡𝑡𝑡𝑡𝑡 = 𝐺𝐺 �
15 𝑏𝑏 𝜎𝜎 𝑇𝑇 16 𝜋𝜋 𝐺𝐺 �
1/ 2
�
𝑓𝑓 𝛾𝛾 ′′
𝐴𝐴 2 𝑅𝑅 2
�
𝜋𝜋 2
�
1/ 2
�
1/ 4

(1-3)

In Equation 1-3, ∆ 𝜎𝜎 𝑡𝑡 𝑡𝑡𝑡𝑡𝑡𝑡 suggests the obtained strength from the mechanical twinning effect, 𝜎𝜎 𝑇𝑇 is
the critical stress for twin nucleation (typically, G/ 𝜎𝜎 𝑇𝑇 ≈ 440 [29]).

13

The last representative particle strengthening in Ni based superalloys is dispersion hardening (or
Orowan strengthening). In this mechanism, secondary phases only act like obstacles for dislocation
movement. In other words, the distribution of secondary particles in Ni matrix reduces the
dislocation slip length or restricts the dislocation motion. In this case, a dislocation climb over a
secondary particle is necessary for an extension of dislocation movement. Not only a coherent
phase like γ’’, but also incoherent secondary phases such as carbides, δ and laves can exhibit some
strengthening through the dispersion hardening effect. This mechanism becomes dominant at
higher temperatures as the internal friction of the material decreases by increasing the temperature
[33]. For the calculation of dispersion hardening effect, the shape of the precipitates must be
considered. Kelly [30] suggested an equation for the dispersion hardening effect calculation with
the consideration on the shape effect of precipitates as:

∆ 𝜎𝜎 𝑑𝑑 𝑡𝑡 𝑑𝑑 =
0.85 𝐺𝐺𝑏𝑏 2 𝜋𝜋 (1 − 𝜐𝜐 )
1/ 2
𝐶𝐶 2 𝑅𝑅 �1 −
𝜋𝜋 2 𝐴𝐴 𝐶𝐶 �
𝑙𝑙𝑙𝑙 �
4 𝑅𝑅 𝜋𝜋 𝑟𝑟 0
�
(1-4)

In the above equation, ν is the Poisson’s ratio, C is the shape factor and r0 is the dislocation core
radius (simply taken as b). Also, C of γ’’ is equal to:

𝐶𝐶 = � 𝑓𝑓 𝛾𝛾 ′′
𝐴𝐴 + (
2
𝜋𝜋 −
𝜋𝜋 2 𝐴𝐴 ) 𝑓𝑓 𝛾𝛾 ′′
𝐴𝐴 (1-5)

14

1.1.2. Embrittlement Mechanisms in Ni based superalloys
As discussed in the previous section, some grain boundary precipitation of continuous secondary
phases can decrease the ductility of Ni based superalloys. This is because those hard precipitates
provide preferential sites for crack nucleation and propagation sites. Additionally, as the
precipitation process consumes some solid strengthening element like Cr in the case of M23C6
carbide precipitation [20], Cr is depleted in the vicinity of M23C6 precipitated grain boundaries,
which forms “soft” area near the grain boundary. As the result, the softened grain boundary area
can be more vulnerable for cracking under stressed conditions with the hard M23C6 continuous
grain boundary film [4,14,21]. This mechanism may be applied to the case of MC and M6C as
well.

Aside from the effect of grain boundary secondary particles, several other mechanisms have been
investigated to interpret the embrittlement of Ni based superalloys. Grain boundary segregation of
some harmful elements such as oxygen [35-51] and sulfur [52-59] can be responsible for
embrittlement. In addition to that, the strain aging model has possibly been accepted as an
interesting model to explain the poor ductility in Ni based superalloys [60,61].

1.1.2.1.Oxygen Induced Embrittlement
Numerous studies [35-51] have suggested that oxygen induced embrittlement was observed in Ni-
based superalloys at elevated temperatures. For wrought Ni-based superalloys, oxygen is supplied
from the air. Under this atmosphere, oxygen penetrates from air toward the interior of superalloys
along grain boundaries at elevated temperatures [35]. One previous study [36] also reported that
15

the single-edge notch bending life of Inconel 718 at 650 ˚C depended on the oxygen partial
pressure of the air. According to their result, the fracture time became shorter from ≈ 100 hours to
≈ 1 hour as the oxygen partial pressure increased from 0.1 Pa to 50 Pa. The author explained this
phenomenon by a mechanism that grain boundary monatomic oxygen atoms are absorbed at the
crack tip due to the high local triaxial stress fields just ahead of the tip. The high oxygen content
at the crack tip enables the crack propagation.

An excessive amount of oxygen in Ni-base alloy can activate the formation oxide inclusions during
the solidification process [63]. The presence of non-metallic inclusion at grain boundaries can drop
the ductility of Ni alloys by the decohesion induced void formation around the non-metallic
inclusions. In a dislocation creep dominant regime, a substantial dislocation density at the interface
between the Ni matrix and non-metallic particles acts like the source for void formation and
coalescence. This decohesion process depends on the size of non-metallic inclusions. T. Denda et
al. [37] conducted low cycle fatigue tests for vacuum induction melted Inconel 718 and their result
showed that the fatigue life at 538 ˚C of a Ni superalloy decreased by micro-sized Al2O3 or TiN in
the matrix. Specifically, the number of cycles to fracture dropped in one order as the inclusion size
was varied from 10 µm to > 100 µm. The dependence of inclusion size on the embrittlement has
been reported in several previous studies as well [38-40]. Ceramic inclusions such as Al2O3, SiO2
and HfO2 provided the nucleation sites for voids and their size varied between 10 – 100 µm in
these studies. On the other hand, Ca and Mg are the deoxidizer in Ni based superalloys and collect
excessive oxygen generated during the solidification process [41,42]. Additionally, as CaO and
MgO typically act like nuclei of MC carbides [16], CaO and MgO are possibly less harmful than
other unfavorable oxides suggested above.
16

Another oxygen effect on ductility of Ni-based superalloys stems from the oxidation of grain
boundary carbides or intermetallics. Grain boundary particles such as MC, M26C6 and Ni3Nb can
form brittle surface oxide films through the oxidation reaction with the diffusing oxygen atoms
along grain boundaries. NbO, TiO, Cr2O3 or NiO films are interface oxide films obtained from this
reaction, and those films can accelerate the intergranular crack growth due to their incoherency to
the Ni matrix [43-45]. This oxidation reaction leads to another detrimental effect on the ductility
of Ni-based superalloys. As the carbon of the carbide interface can react with oxygen and, CO and
CO2 can be formed by this reaction. The subsequent increase in internal pressure due to the
formation of CO and CO2 involves the expedited crack growth during mechanical tests at elevated
temperatures [46-49].

1.1.2.2.Sulfur Induced Embrittlement
While “high” sulfur containing IN625 (> 30 ppm) has almost equal strength and ductility to the
“low” wrought IN625 (< 20 ppm) at ambient temperature [52], sulfur segregation to the grain
boundaries can lead to embrittlement at elevated temperatures (mostly over 538 ˚C for IN625) in
Ni based superalloys [53-59]. Typically, sulfur tends to segregate at grain boundaries of Ni base
superalloys [53]. Most sulfur induced embrittlement occurs by intergranular fracture in Ni
superalloys. An important point is that Mg appears to have a beneficial effect [54] for this case
and is often (or usually) added to wrought Ni-based superalloys to react with the sulfur before the
sulfur has an opportunity to diffuse to and embrittle grain boundaries. Additionally, some previous
studies [41,42] have suggested that Ca, Hf and Zr have a similar effect to Mg for the alleviation of
sulfur induced embrittlement in Ni superalloys.
17

The first mechanism of sulfur embrittlement is the formation of continuous eutectic sulfide film of
Ni3S2 at the grain boundaries of Ni alloys [55]. As sulfur is not favorable to be alloyed in the Ni
matrix, S atoms tends to be segregated on grain boundaries when they are added in Ni superalloys
[10]. The grain boundary S, then, tends to form Ni-Ni3S2 eutectic films on grain boundaries of a
Ni alloy. This compound film has a low melting temperature of 535 ˚C and some partial melting
of this eutectic film can cause an intermediate temperature embrittlement of Ni alloys [55].

Aside from that detrimental chemical reaction, other research [56] reported that the presence of S
on grain boundaries of Ni alloys can affect grain boundary energy subsequent embrittlement. One
other mechanism for sulfur embrittlement is originated from the grain boundary diffusion of
monatomic sulfur atoms and the subsequent grain boundary decohesion by the sulfur [53,57,58].
Kart et al. [59] suggested a model for sulfur induced intergranular embrittlement in Ni matrix. In
this mechanism, the grain boundary decohesion is related to the repulsive force between sulfur
atoms at Ni grain boundaries. The sulfur atoms surrounded by Ni atoms tend to be separated from
each other by ~3.3 Å, whereas both Ni-Ni and Ni-S distances at grain boundaries are typically ~
2.5 Å [59]. Therefore, segregation of sulfur at an intergranular crack tip can bulge the crack tip as
S-S pair is apart more than both Ni-Ni and Ni-S pairs.

1.1.2.3.Strain Aging Effect
In welded Ni superalloys, residual strains generated during the restrained cooling can cause high
temperature embrittlement with matrix hardening precipitates such as γ’ or γ’’ [ 61,62].
Specifically, a large population of dislocation can be formed due to the faster cooling process in
18

welding and a highly dense residual strain in the welded samples mostly is mostly transferred to
grain boundary during elevated temperature service. It is also related to the grain boundary
precipitates as those precipitates (mostly carbides) typically leave solute depletion zone in the
vicinity of grain boundaries, hence a matrix precipitate (γ’ or γ’’) depletion zone near the grain
boundary. In this phenomenon, γ’ or γ’’ matrix hardening, with grain boundary region softening
due to solute depletion, is the main reason why the residual strain becomes concentrated on grain
boundaries rather than uniformly distributed [62]. The strain localization introduces the early crack
nucleation and propagation at grain boundaries or precipitate interfaces in Ni-base superalloys
[61,62].

1.2. Metal Additive Nanufacturing
Additive Manufacturing (AM) is the formalized term for what is popularly called 3D Printing and
it has been attracting attention from academic and industrial areas. This technology is a group of
processes that join raw materials to build objects from designs drafted using computer software
such as Computer Aided Design (CAD), as opposed to subtractive manufacturing methodologies,
such as conventional machining [64]. The final built object of AM usually has a layer-by-layer as
the fabrication is done by consecutive additions of layers [65]. The main goal of AM is to achieve
cost and time saving in manufacturing processes. Compared to conventional production processes
of metals, AM technology has several advantages in metal manufacturing processes because the
design of products can be easily adapted or changed, and it produces less waste during the entire
process than conventional machining [66,67]. Another benefit of AM is that controlling of
microstructure is easier in AM than in conventional manufacturing because the built
19

microstructure of AM products is mostly decided by the processing parameters such as laser
powder, laser speed and precursor feeding rate [68,69].

1.2.1. Laser Powder Bed Fusion (LPBF)
Many types of metal additive manufacturing processes can be categorized according to the heat
source and how the precursor is supplied [69-71]. According to the ASTM F2792 – 12a standard
[64], seven major types of AM methodologies have been developed. The following lists those as
major AM technologies: (1) Powder Bed Fusion; (2) Directed Energy Deposition; (3) Materials
Jetting (4) Binder Jetting (5) Material Extrusion (6) Vat Photopolymerization (7) Sheet
Lamination. Among these technologies, (1) Powder Bed Fusion; (2) Directed Energy Deposition;
(4) Binder Jetting; (7) Sheet Lamination are applicable to fabricate metal pieces. Laser powder bed
fusion (LPBF) is one of the processes under Powder Bed Fusion.

Generally, material supply via wire feed or powder is conducted in metal AM processes, whereby
selected regions (heat exposed regions) are sintered or melted differently depending on beam
power and velocity [72]. Among many different AM methods, LPBF can encompass most metal
raw materials including intermetallics such as TiAl and refractory metals such as Ni and high-
entropy alloys with a reasonable fabrication speed [73]. The manufacturing process of LPBF is
basically separated in three steps: (1) preparation, (2) the LPBF process for metallic powders and
(3) post-treatment [74]. The first step is related to the preparation for manufacturing works
including 3D CAD designing and some data conversion to adjust the design to the LPBF machine.
The next step is the sequential deposition of layers through melting and sintering (sintering is a
20

solid-state process) of metal powders with a laser source. The net shaped parts are directly
fabricated in this stage and process parameters like laser power, beam speed and others determine
some material properties of the fabricated parts such as porosity, microstructure and residual stress
[74]. Therefore, it is recommended to conduct post treatment to attenuate negative consequences
generated during the LPBF process.

In the LPBF, metal powders (typically between 20 µm and 100 µm) are melted by laser heat
sources to create solid layers. In this process, metal powders are melted quickly as the powders are
exposed to a rapidly moving heat source, then solidifying of the melt pool occurs with producing
very strong metallic bonds directly between the metal powders [69]. Now, Figure 1-3 shows a
schematic of LPBF machine.
In Fig. 1-4, the recoater arm distributes metal powders to the powder bed for every metal layer.
After completion of the laser operation on a layer, another powder layer spreads over the previous
layer by the recoater arm. All components in Figure 1-4 are inside a chamber of the LPBF machine
and this chamber is conserved under an inert gas environment with nitrogen (N2) or argon (Ar).
The primary purpose of this inert gas environment is to prevent undesirable oxidation of the
building product in the chamber during fabrication processes. The building part is typically placed
on a stainless-steel substrate while being fabricated and it needs to be separated from the substrate
after the completion of LPBF processes [75].
21

Figure 1-3. A schematic of LPBF machine [75].

22

1.2.2. Processing Parameters in LPBF
Since the microstructure of a sample fabricated by the LPBF method is strongly affected by process
parameters [76], a proper setting of the process parameters is required to guarantee favorable
mechanical properties of LPBF products. Laser power, laser energy input, laser exposure time and
scan speed are process parameters of a typical LPBF machine which determine the microstructure,
porosity level and mechanical behavior of the fabricated sample [75].

1.2.2.1.Laser Energy Density
As previously explained, laser powder (P), laser energy input, laser exposure time and scanning
speed (s) are important parameters in a LPBF process. Moreover, mean powder size (w), hatch
distance (h) and scanning pattern should be considered to obtain final products of favorable quality.
Laser energy density (L) is a concept introduced to explain effects of parameters suggested above
in a simple way, which has the unit of J/mm
3
and is described as follows in Equation 1-6 [77]:

𝐿𝐿 =
𝑃𝑃 𝜀𝜀 ∙ ℎ ∙ 𝑤𝑤
(1-6)

High value of L indicates small values of s, h or w; or high laser power condition, that involves
longer overheat for building materials. Since this prolonged exposure on heat can involve
unexpected microstructure evolution like grain growth and grain geometry change, a careful
control of L value should be considered [78-80]. A previous study [81] reported that the L value
has a relationship with the porosity of a building part. 23.4 % of porosity was measured when L
value was around the minimum at 180 J/mm
3
, however, the porosity decreased a lot to 1.6 % by
23

increasing L value to around 330 J/mm
3
. Therefore, it is concluded a higher laser energy density
results in a lower porosity level in LPBF products. It is considered that high fluidity of the melt
pool due to high L value was effective to restrict formation of pores during the fabrication process.
In addition, surface roughness was improved with conditions of high L values [81].

1.2.2.2.Effects of Other Processing Parameters
Other processing parameters can affect properties of LPBF samples as well. Temperature of the
LPBF chamber, powder size (and morphology) and the chamber gas environment are those
parameters of concern.
First, the temperature inside the chamber can vary between room temperature (RT) to 300 ˚C.
Generally, metal powders are easily melted under higher chamber temperature conditions. In other
words, the high temperature condition has a similar effect to large L levels (low porosity). In
addition, a temperature gradient between the most upper layer of the building metal part and the
stainless steel substrate, where the fabrication is ongoing, influences on the microstructural
evolution such as dendritic grain size and porosity levels [76]. Another study [82] also reported
that a drop in residual stress of the fabricated metal part is expected with proper control of
temperature conditions inside the LPBF chamber components.

Several powder properties like mean powder particle size and the size distribution are another
important factor to build products. The typical range in mean powder size relies between 20-100
µm in the LPBF process. Since the surface quality of each layer of the building parts depends on
the particle size distribution, a homogeneous size distribution of powders is required to secure
24

acceptable building part quality throughout the LPBF sample. Powder morphology needs to be
considered as well. Metal powders in a LPBF machine are recommended to be spherical to avoid
uneven and rough layer surfaces caused by non-uniform distribution of powders on the building
layer due to an irregular shape due to their poor fluidity [76].

The gas environment inside chamber is typically filled with protective gas such as Ar, He and N2.
A primary reason of securing this inert atmosphere in the chamber is to preclude oxidation
phenomena of building metal parts inside the chamber. For example, decarburization of Ni alloy
samples caused by formation of gas like CO and CO2 with a high level of oxygen inside the
chamber can be harmful to mechanical properties of the building sample. Other negative effects
of high oxygen level in the chambers are oxidation of some susceptible elements such as Ti, Al,
Mn, Cr, Si in a Ni alloy and pore formation during the fabrication process due to gas formation
such as CO and CO2. Hence, the chamber needs to be protected with inert gas to avoid these
unfavorable results [76].

1.2.3. Mechanical Properties of AM IN625 Compared to Wrought IN625
The tensile strength of AM IN625 alloy is almost the same as that of conventional wrought and
cast IN625 alloys at room temperature and 760 ˚C [74,83]. Furthermore, according to a previous
report by the authors [52], AM IN625 alloys have equal or even higher creep strength at elevated
temperatures of 650 ˚C and 800 ˚C.
Earlier works [74,83] suggested that the tensile elongation (not creep tests) of AM HIPed IN625
alloy dropped significantly at 760 ˚C. This result agreed well with our early findings which suggest
25

the high temperature creep ductilities of AM were poorer than wrought IN625 at 650 ˚C and 800
˚C, whereas room temperature ductility of AM IN625 was equal to the wrought [52]. Murr et al.
[84] reported that the tensile ductility of AM HIPed IN625 remained comparable with that of
wrought IN625 at 538 ˚C. Those results appear consistent with the observation of embrittlement
in the AM IN625 alloy. As discussed in section 1.1.2., the elevated temperature embrittlement (>
600 ˚C) of Ni base superalloys has been explained by several mechanisms: monotonic oxygen
induced grain boundary embrittlement (section 1.1.2.1.), sulfur induced intergranular fracture
(section 1.1.2.2.), continuous secondary phase precipitation at grain boundaries (section 1.1.2.),
and porosity.

Figure 1-4 presents the summarized data obtained from references [8,52,74,84-86]. As shown in
Fig. 1-4a, the mechanical strength of AM HIPed IN625 is fairly comparable with that of wrought
IN625 in the range of room temperature to 760 ˚C. However, in the case of ductility data (Fig. 1-
4b), even though AM HIPed IN625 shows somewhat competitive ductility to wrought IN625 up
to 538 ˚C, AM HIPed IN625 tends to be more brittle than wrought IN625 beyond the temperature
of 538 ˚C. Therefore, the high temperature embrittlement mechanism of AM HIPed IN625 needs
to be investigated.
26

Figure 1-4. The mechanical properties comparison between solution annealed wrought IN625 and
AM HIPed IN625 at various temperatures (a) yield strength and ultimate tensile strength (b)
ductility.
27

1.2.4. Microstructure Evolution of AM IN625
As the microstructure is an essential parameter which determines the properties of a material, the
microstructure evolution of AM IN625 during the fabrication process deserves focus. Some detail
of microstructural changes following post treatments after the LPBF fabrication is reported by
Kreitcberg et al. [74]. Immediately after the layer deposition process for an AM IN625 part
building, no eutectic phases were distinguished in Ni matrix. However, some other researches
[87,88] have reported that precipitation such as γ’, γ’’, Laves and MC carbide were often observed
in the as-built AM IN625. Interestingly, Yang et al. [89] found the same result as Kreitcerg et al.
(no secondary phases in the as-built state). Therefore, it is determined that the chemical
composition of IN625 powders and processing parameter of the LPBF significantly affected the
microstructural evolution of AM IN625 in the as-built state. Additionally, a strong <100>
orientation texture occurred along the building direction due to the favorable heat flow direction
during the building step [74]. Despite the favorable condition for <100> texture evolution, in the
as-built state, the <100> texture exhibits fine grain structure due to impurity segregation and heat
dissipation disorder at the molten pool boundaries during the fabrication process [87].

In general, a stress relief treatment on a stainless steel substrate of AM IN625 is conducted before
cutting the built part from the substrate to prevent warping due to residual stress. The stress relief
temperature condition for nickel alloy varies within the range from 650 ˚C to 870 ˚C [74]. As
expected from Fig. 1-2, the TTT diagram, the temperature condition for the stress relief treatment
can substantially influence on the secondary phase precipitation. One previous study [74] observed
the formation of M6C carbide and δ phase. However, the calculated thermodynamic prediction by
Stoudt et al. [90] did not predict the formation of M6C carbide after the stress relief treatment,
28

instead, they predicted the presence of MC carbide and, they confirmed the prediction using SEM
and XRD analysis. Additionally, one other result reported by Lass et al. [91] is consistent with the
observation of Stoudt et al. [90]. They found MC carbide after the stress relief treatment of AM
IN625. Additionally, the research by Lass et al. [91] suggests an ideal temperature condition for
stress relief treatment as 800 ˚C after 1 hour treatment due to the absence of δ phase.

As the AM IN625 part typically involves relatively high porosity in the as-fabricated part, hot
isostatic pressing (HIP) treatment is required to restore the mechanical properties of AM IN625.
Typically, non-HIPed AM IN625 part fabricated by the LPBF process can ranged from 0.1 % to
2.7 % or even higher depending on the LPBF processing parameter settings [87]. The
recommended HIP condition for IN625 is at temperatures between 1120 ˚C – 1240 ˚C under the
gas pressure between 100 MPa – 165 MPa for 3 – 4 hours. The reduction of microporosity in the
AM IN625 is expected to be negligible after the HIP treatment on the AM IN625 [74]. In addition
to that, weakening of <100> orientation texture along the Z direction can be obtained by the HIP
treatment. The representative phase transformation occurring in the HIP process is the dissolution
of secondary phases except for MC carbide [74].

HIP treatment should be the previous process right before a service. For the AM HIPed IN625, the
microstructure evolution after long-term service has not been investigated in spite of its importance.

29

1.3. High Temperature Creep of IN625
‘Creep’ is the term given to the slow deformation and slow fracture under stress at elevated
temperatures. Creep testing is often conducted using samples for tensile or compressive
deformation under a constant elevated temperature. The creep rate is defined as a change in strain
versus time and the creep rate is used to evaluate materials for high temperature applications under
load such as gas turbines and jet engines.

1.3.1. Creep Deformation
At elevated temperatures, increase in thermal energy leads to the raise in vibration frequency of
atoms and this activates the movement and concentration of vacancies leading to enhanced
dislocation climb. At a fixed stress or load, whether it is higher or lower than the conventional (e.g.
10
-3
s
-1
) yield strength of materials, the applied load overcomes or at least reduces the amount of
thermal energy barrier for dislocation climb, which allows deformation.
Although creep deformation is a thermally activated mechanism, it can actually occur over the
entire temperature range of metallic solids. The temperature range for creep is often greater than
0.5 Tm, where Tm is the absolute melting temperature, however, a wider temperature band between
certainly 0.3 and 0.9 Tm, but even lower, can be accepted as the general creep activation region
[92,93]. A typical strain versus time creep curve is suggested in Figure 1-5:
30

Figure 1-5. A typical creep curve at a constant stress condition.

31

In Fig. 1-5, three regions are evident. The first stage (Stage Ⅰ) is the primary creep region, where
hardening occurs and the creep rate, 𝜀𝜀 ̇ = d 𝜀𝜀 𝑝𝑝 /𝑑𝑑𝑑𝑑 , is decreasing with increasing timet. While the
variation in 𝜀𝜀 ̇ occurs typically as a gradual drop in value, 𝜀𝜀 ̇ sometimes increases during primary
creep due to the interaction between dislocations and obstacles such as solute drag [92] and also
sigmoidal creep. This change in creep rate continues until a steady-state creep rate ( 𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
) is reached.
The second regime (Stage Ⅱ) in Fig. 1-5 is termed secondary creep or steady-state creep, where
the creep rate is maintained at a constant of 𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
over the entire time in the regime. Secondary creep
regime has a particular importance as it indicates the microstructure has attained a state of dynamic
equilibrium, i.e. constant structure. After the material undergoes secondary creep for a sufficient
time, the creep rate starts to increase and the material eventually fractures at a certain strain or
time. This stage is termed tertiary creep or Stage Ⅲ. This transition occurs due to cavitation,
cracking and of the material [92].

1.3.2. Creep Deformation Mechanisms
In steady-state, dynamic hardening due to deformation is balanced with the formation of constant
structure by dynamic softening such as dynamic recovery. As large strains can be accumulated
during steady-state, it is particularly important to interpret the creep behaviors in steady-state.
Often, the fracture times is predictable based on steady-state behavior (e.g. the Monkman-Grant
relation). The steady-state creep behavior of commercially pure material and alloys has been
suggested as a combination of stress, temperature T, self-diffusion coefficient D, stacking fault
energy, and shear modulus G [94]. It is also customary that 𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
is identified as an Arrhenius type
equation 𝜀𝜀 ̇ ∝ 𝜎𝜎 𝑡𝑡 ∙ exp ( − 𝑄𝑄 𝑐𝑐 /𝑘𝑘 𝐴𝐴 𝜋𝜋 ), where 𝑄𝑄 𝑐𝑐 is the activation energy for creep, n is stress
32

exponent and kB is Boltzmann constant. Based on this concept, a constitutive relation that links
𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
, grain size and T was suggested by Mukherjee et al. [95-97]:

𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
= 𝐴𝐴 0
∙
𝐷𝐷 𝐺𝐺𝑏𝑏 𝑘𝑘 𝐴𝐴 𝜋𝜋 ( 𝑏𝑏 /𝑑𝑑 )
𝑝𝑝 ( 𝜎𝜎 𝑑𝑑𝑑𝑑
/𝐺𝐺 )
𝑡𝑡 (1-7)

In Equation 1-7, A0 is a dimensionless constant, D is the appropriate diffusivity (lattice or grain
boundary or pipe diffusion), b is the Burger’s vector, d is the grain size, p is the grain size exponent,
σss is the steady-state stress, and G is the shear modulus. In this equation, the stress exponent n is
often expressed as the reciprocal constant of strain sensitivity m and it is determined to be about
3.5-7 for pure metals, ceramics and many alloys over relatively wide range of experimental
variables such as temperature and strain rate [92]. In addition to the stress exponent n, other
parameters in Equation 1-7 like p and D are correlated to the creep mechanism and Table 1-3 lists
those parameter values depending on creep mechanisms [98]:

33

Table 1-3. Values of the stress exponent n, the grain size exponent p and diffusivity D
corresponding to creep mechanisms [98].
Creep mechanism n p D
Harper-Dorn Creep 1 0 Dl
Dislocation climb at T < 0.5 Tm >3.5 - 7
0
Dp
Dislocation climb at T > 0.5 Tm 3.5 - 7 Dl
Grain boundary sliding 2 2 - 3 Dgb or Dl
Grain boundary diffusion creep
(Coble creep)
1 3 Dgb
Lattice diffusion creep
(Nabarro-Herring creep)
1 2 Dl

34

From Table 1-3, grain boundary effects on creep properties become negligible when the creep is
governed by dislocation climb mechanisms. Thus, from this point of view, an equation for steady-
state five-power-law (PL) is suggested from Equation 1-2 [92]:

𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
= 𝐴𝐴 1
∙ exp [ 𝑄𝑄 𝑐𝑐 /𝑘𝑘 𝐴𝐴 𝜋𝜋 ]( 𝜎𝜎 𝑑𝑑𝑑𝑑
/𝐸𝐸 )
𝑡𝑡 (1-8)

where A1 is a constant (that may include the stacking fault energy), σss is the constant stress at
steady-state and E is Young’s modulus. For this equation, Shear modulus G can replace E. Since
the creep activation energy, Qc,represents the energy barrier level to initiate or sustain the creep
deformation, a proper calculation for this value is required to understand creep behaviors. Equation
1-8 implies that the value of Qc can be a simple description of the change in steady state creep-rate
for a given substructure (or strength) at dynamic equilibrium s, at a constant stress σss with changes
in temperature. Therefore, Qc can be calculated using the following equation [92]:

𝑄𝑄 𝑐𝑐 = − 𝑘𝑘 𝐴𝐴 [ 𝛿𝛿 ( 𝑙𝑙𝑙𝑙 𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
)/𝛿𝛿 (1/𝜋𝜋 )]
𝜎𝜎 𝑠𝑠𝑠𝑠
/ 𝐸𝐸 , 𝑑𝑑 (1-9)

In five-power-law creep, as dislocation climb is dominant, either the activation energy of pipe
diffusion Qp or that of lattice self-diffusion Ql seem to be essentially equal to the value of Qc.
Generally, T > 0.5 Tm, Ql has been found to replace Qc for a large class of materials. In addition,
Qc generally decreases as n increases [92].
Equation 1-8 and Equation 1-9 are applicable for the case of n < 3 with other dominant creep
mechanisms which are also climb-controlled. In the power law creep equations, n generally raises
above 3.5-7 as the applied stress increases into the power law to power law breakdown (PLB)
regime. In PLB regime, steady-state creep rate shows an observed exponential dependency on the
35

applied stress [99]. Moreover, a hyperbolic sine function can describe the transition from power
to PLB [92].

1.3.3. Literature Reviews on The Creep Behavior of IN625
According to the ASTM B443 standard [100], the application area of IN625 is divided into 2
categories:
(1) Grade 1 (Annealed) – Material is normally employed in service temperature up to 593 ˚C
These products are generally annealed in the range of 871 – 982 ˚C and exhibit a fine grain
structure.
(2) Grade 2 (Solution Annealed) – Material is normally employed in service temperatures above
593˚C when resistance to creep and rupture is required. These products are annealed between 1120
– 1180 ˚C and exhibit a coarse grain structure.
Since the main interest of this study is about creep behavior of IN625 at elevated temperatures, the
literature conducted for solution annealed samples are primarily chosen to review. The following
Table 1-4 and Table 1-5 show the specification of data from previous studies:

36

Table 1-4. Specimen creep test conditions from references.
S: Solution heat treated, 30min-1hr at 1100-1150℃; M: Mill annealed, 5 minutes at 950-1050℃

Ref # Heat treatment Air condition Grain size (μm) Product
[8] S Air >100 -
[85] - - - Sheet
[86] S Air <10 Cryomilled
[101] M+S Air >100 Sheet
[102] - - - -
[103] S Air or He >45 Arc melting
37

Table 1-5. Chemical composition of reference materials.
Ref #
Elements
Ni Cr Fe Mo
Nb +
Ta
C Mn Si Al Ti Co S
ASTM Bal. 20-23 5.0 8-10
3.15-
4.15
0.1 0.5 0.5 0.4 0.4 1.0 0.015
[8] ASTM standard
[85] Bal. 22.12 3.30 9.13 3.47 0.04 0.04 0.2 0.27 0.25 - 0.006
[86] Bal. 20.08 4.32 8.94 3.55 0.08 0.13 0.15 0.18 0.08 0.19 0.001
[101] Not suggested
[102] ASTM standard
[103] Bal. 22.09 2.54 8.81 3.53 0.053 0.24 0.28 0.24 0.13 0.06 0.002
The given ASTM standard compositions are maximum allowable quantities.

38

The first five references [8,85,86,101-103] provide data for commercial IN625, whereas the other
reference [86] gives a dataset for a powder metallurgy (PM) sample. It is considered that a
comparison in creep properties between commercial IN625 alloy and PM IN625 alloy is
worthwhile as AM IN625 is also fabricated from powders. Since the chemical composition of
reference data satisfy the ASTM standard and those values are comparable to each other, any
significant effect from solid solution strengthening on mechanical property is not expected among
those samples. Grain size is another factor which affects the mechanical behaviors of materials
and generally the smaller grain size, the higher the material strength. Thus, it is assumed that the
grain sizes of [85] and [102] are in the grain size range suggested in Table. 1-4.
As suggested in Equation 1-8, elastic modulus E should be determined before the plot between
steady-state creep rate and stress values is drafted. Elastic modulus of IN625 was calculated by
measured data from [8]. Also, from the data of the elastic moduli in Fig. 1-6, steady-state creep
rate data in references are arranged as the plot between log ( 𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
) and log (σ/E) as shown in Figure
1-7.

39

Figure 1-6. Dependence of Young’s modulus on temperature in IN625.

40

Figure 1-7. The plot between 𝜀𝜀 ̇ 𝑑𝑑𝑑𝑑
and 𝜎𝜎 𝑑𝑑𝑑𝑑
/𝐸𝐸 of solution treated wrought IN625 and PM IN625
over the given temperature range.

41

The stress exponent (n) of wrought IN625 decreases as temperature increases from 13.25 at 650
˚C to 4.00 at 950 ˚C, and it may reach to about 3 at 1100 ˚C. This suggests that power law
breakdown is dominant in high temperature creep of wrought IN625 at least from 650 ˚C to 800
˚C, where n lies between 8 and 13. Beyond 800 ˚C, dislocation creep by dislocation climb becomes
important for the creep behavior of wrought IN625. However, in the case of powder IN625, the
creep strength was higher than wrought IN625 at 650 ˚C and n values lie between 4.4 to 5 in
temperature range of 538 ˚C – 650 ˚C. As suggested in Table 1-4, the average grain size of PM
IN625 is much smaller than wrought IN625.

Observation of the PLB region at intermediate temperatures ( ≤ 700 ˚C) was also reported by de
Oliveira et al. [105]. The author of the study argued that the creep activation energy 407 kJ/mol of
their work was near the dislocation core diffusion activation energy. It suggests that dislocation
core diffusion is the dominant creep mechanism for the wrought IN625 alloy at 650 ˚C. While the
typical n value for dislocation core diffusion lies between 5 – 7 (Table 1-3), according to their
argument, this unusually high n value( ≥ 7) is probably due to the strong dislocation drag effect by
solutes, such as Nb, Mo, and Ti, in IN625 [105-107]. One previous study [107] more specifically
expanded this argument by the “dislocation arrest theory” in the dynamic strain aging phenomenon
occurring in IN625. According to this theory, at the given temperature range, solute atom
atmospheres form on dislocation forests, and a mutual interaction between mobile dislocations and
forest dislocations occurs through solute exchange between dislocations by core diffusion. This
interaction contributes to the slow strain rate, i.e. the steep stress exponent. As this interaction
essentially requires the absence of secondary matrix particles, acting like solute sink, this
mechanism will be unacceptable if Ni matrix has some uniformly distributed secondary phases
42

[107]. As a representative solute strengthening alloy, IN625, is presumably appropriate for being
applied to this hypothesis.

However, the calculated creep activation energy (Qc) from Fig. 1-7 is 275 kJ/mol in the
temperature range between 650 ˚C – 800 ˚C (calculated at 263 MPa), which is quite different from
the value 407 kJ/mol from Olivera (core diffusion energy, 407 kJ/mol). This value is rather closer
to the lattice self-diffusion energy of Ni, 287 kJ/mol [86], than the core diffusion activation energy.
Thus, according to the calculated Qc value in this paper, dislocation creep should be the main
mechanism for creep of wrought IN625 at the given temperature range. This different Qc values
between the calculated value from Fig. 1-7 and previous work by de Oliveira et al. [105] does
probably stem from the difference in the stress for the Qc calculation. 500 MPa was used for Qc
calculation of the reference, which is almost double the calculated Qc value in this section.
43

2. EXPERIMENTAL METHODS
The following section includes an overview of the sample preparation of IN625 alloys in this study,
the creep test procedures and micrographic analyses for the samples. Specifically, in addition to
the detail of the LPBF fabrication process, characterization methods for microstructural analysis,
including phase analysis and chemical composition, will be described.

2.1. LPBF Sample Fabrication Process
LPBF was chosen for the fabrication of AM IN625 sample preparation in this study. As highlighted
in Section 2.2.1, LPBF is one of AM technologies applicable to metal sample building. Between
wire feeding method and powder feeding method, the latter method was utilized for the sample
fabrication of this study.

2.1.1. Powder Characterization
EOS (EOS GmbH, Munich, Germany) art. No. 9011-0022 IN625 powder was used for the AM
sample preparation in this study. Secondary Electron Microscopy (SEM) image of powders are
shown in Figure 2-1, which displays a general spherical morphology. The powder size distribution
measurement was conducted with Malvern (Malvern Panalytical, Malvern, United Kingdom)
Mastersizer 2000 particle size analyzer. This analysis was conducted by an operator who is an
expert at Inha University (Incheon, South Korea). The measured value of powder size distribution
and mean particle size are given in Fig. 2-1b. The average diameter of powder was determined 37
µm and, the minimum diameter and the maximum diameter of powders was measured as 11.5 µm
44

and 104.7 µm, respectively. 44.57% of powders have the diameter under the average and 55.43%
of powders have the diameter above the average.

45

Figure 2-1. (a) SEM image of IN625 powders for LPBF fabrication (b) The particle size
distribution of powder particles in this study.
46

2.1.2. LPBF Processes for Sample Fabrication
An EOSINT M290 (EOS GmbH, Munich, Germany) LPBF machine equipped with a 400W Yb-
fiber laser source was utilized to fabricate AM IN625 specimens. Figure 2-2 (a) is a photo of the
LPBF machine. The specimen fabrication was performed at the National Institution of Standard
and Technologies (NIST). Wrought IN625 samples were purchased from Rolled Alloys, Inc. to
compare the high temperature creep behaviors of AM IN625 alloys with conventional wrought
IN625 alloys. The details of LPBF fabrication parameters are given in Table 2-1 where stripe
scanning indicates the scanning for main body of the coupon and contour scanning represents the
border scanning outside the main body. Every border scanning follows each stripe scanning. The
purpose of contour scanning is to reduce surface defects of scanned layer and to obtain surface
smoothness of the samples.

12.5 mm × 12.5 mm × 102 mm coupons for creep tests were fabricated using LPBF process on a
250 mm × 250 mm substrate. As shown in Figure 2-2 (b), those coupons were still bonded with
the substrate even after the fabrication processes and it needed to be separated from the substrate.
The gas environment of the machine chamber was provided by injecting 99.995 % purity Ar gas
and the oxygen level was maintained at less than 0.1 % O2. As discussed earlier in Section 1.2.2.2,
the control of chamber gas environment can significantly affect the quality of the built sample.
The substrate temperature was held at 80°C to reduce residual stresses in the sample [82].

The scanning orientation was rotated by 180° from the previous scanning direction, which is
referred to as the “bi-directional method”. The scanning direction of the next layer was rotated by
67.5° from the previous layer. Sun et al. reported that this method was effective in reducing the
47

texture on XY plane for Ni-based superalloys [108]. A schematic image displaying the scanning
method in this study is suggested in Figure 2-3.

As the built AM sample had a residual stress and it was typically warped if any post heat treatment
was conducted before separation of the samples from the substrate. The coupons were stress
relieved at 800 °C for 1 hour. This heat treatment prevents warping of the samples. Two different
final heat treatments were conducted after cutting the samples from the substrate: the first heat
treatment was solution annealing at 1100 °C for 1 hour in air with a heating rate of 20 ˚C/min and,
second (in selected cases), the specimens were hot isostatic pressing (HIP) at 1175 °C and 150
MPa for 3 hours under an Ar gas environment. The HIP process was conducted by Quintus, Ohio
and the HIP process history is given in Figure 2-4 with the thermal history of solution annealing
heat treatment. For the solution annealing treatment, a Vulcan D 130 Burnout furnace was used.
Those heat treatment conditions are expected to attenuate the Nb segregation in pre-dendritic
regions of LPBF IN625. Zhang et al. [109] reported that Nb segregation of LPBF IN625 alloy
decreased from 4.41 wt% to 4.00 wt% after a solute solution treatment. In the case of wrought
samples, the same solution annealing heat treatment at 1100 °C for 1 hour in air was conducted
before the creep tests.

48

Table 2-1. Parameter values of LPBF process of this study
Parameters
Beam type
Stripe Contour
Beam Speed 960 mm·s
-1
300 mm·s
-1

Beam Power 285 W 138 W
Beam thickness 10 mm 0.04 mm
Beam offset 0.015 mm 0.012 mm
Hatch distance 0.04 mm 0.04 mm

49

Figure 2-2. (a) EOSINT M290 LPBF machine (b) as-fabricated LPBF IN625 coupons.

50

Figure 2-3. The scanning strategy of LPBF process in this study.

51

Figure 2-4. Temperature and pressure history of HIP and Solution Annealing treatment (S.A.).

52

2.2. High Temperature Creep Test
Creep tests at elevated temperatures helps to determine whether a material is applicable at given
temperature conditions for industrial purposes. As shown earlier in Section 1.3, a creep test has
two main variables to set; applied stress and temperature. For the verification if AM IN625 alloys
satisfy with the industrial requirements in the aerospace field, a careful consideration how to set
those two variables was performed. The following sub-section will illustrate how to decide those
two experimental variables. The configuration of a creep testing machine used in this study is also
illustrated specifically in this section.
As the main target of this creep test is to clarify the possibility that AM IN625 alloys can replace
commercial wrought IN625 alloys for an aircraft engines; a longtime service under the condition
of the aircraft engine needs to be assessed [110]. Therefore, in addition to the creep test for either
as-HIPed or as-solution annealed samples (as-HIPed and as-solution annealed suggest the samples
were HIPed or solution treated without any long term cyclic heat treatments), other creep tests of
6 months or 1 year long-term cyclic heat-treated (LHT) samples at service temperatures were
conducted as well.

2.2.1. Determination of Temperature and Applied Stress Conditions
From Section 1.1.1 and a reference [8], the representative applications of IN625 for an aircraft
engine at elevated temperatures are listed such as: combustion liners, spray bars, engine exhaust
systems and turbine shroud rings. As suggested in Figure 1-1b, those aircraft engine parts beyond
the combustion chamber undergo high temperature conditions of over 650 ˚C. All the parts listed
above are placed in this region.
53

A reference [111] provided some specific temperature and air pressure conditions for this region
in a turbofan engine, JT9D-7 produced by Pratt & Whitney. According to the reference, the typical
temperature range required in those parts is between 504 ˚C and 1273 ˚C under air pressure from
0.15 to 2.31 MPa. Among the suggested application in an aircraft engine, IN625 is typically used
in fuel injector, combustion liner, turbine shroud ring, and exhaust systems as illustrated in Figure
2-5. For the application as combustion outer liners, another reference [112] indicates that
temperatures around 800 ˚C are required for the part. In the case of spray bars and engine exhaust
systems, relatively low temperatures under 560 ˚C are the target condition [110,111]. For turbine
shroud rings which are attached to the turbine chamber wall, at least 650 ˚C under 0.57 MPa is
suggested for operation [111,113]. Fig. 2-5 provides the summary of temperature and pressure
conditions for IN625 aircraft engine parts. Now, based on the literature searches, two main
temperature conditions of 650 ˚C and 800 ˚C are chosen for the high temperature creep tests for
IN625 alloys in this study.

As the suggested maximum requirement of air pressure, 2.31 MPa [111], is far lower than the yield
strength of IN625 at the given maximum temperature ( ≈ 76 MPa at 1273 ˚C), a new approach to
derive the applied stress conditions for this study was required. Therefore, it would be reasonable
if the creep tests are conducted on the identical experimental conditions to reference data of
wrought IN625. From two references [8,96], those compensated stress conditions are provided
such as: -2.4 to -2.8 for 650 ˚C and -2.9 to -3.5 for 800 ˚C. In other words, the applied stress range
of this study is determined between 65 to 658 MPa.

54

Figure 2-5. Temperature and air pressure conditions for several aircraft engine parts, T and S
represent temperature and air pressure, respectively [111-113].

55

2.2.2. Long Term Cyclic Heat Treatment (LHT) at 650 ˚C
As a commercial aircraft is designed to operate for a long time, the aircraft engine parts are
expected to work for a scheduled lifespan. In this study, as it will be illustrated, 6 months and 1
year are set to the target service time.

The typical aircraft engine operation time before an overhaul is expected 16000 hours [114]. A
previous study [115] has reported that the elongation of IN625 dropped after ≈ 1 year heat exposure
at elevated temperatures above 650 ˚C with some increase in strength due to phase transformation.
The elongation slightly dropped from 1 year to 2 years without a change in the strength. However,
the ductility drop during this time interval was much smaller than the drop in ductility between the
non-long-time heat treated state and 1 year heat exposed samples to 650 ˚C. Therefore, it is
suggested that AM IN625 alloy can replace wrought IN625 alloy in long term applications for
aircraft engines if the mechanical properties of 1 year heat-treated AM IN625 is comparable with
those of wrought IN625. The other time condition of 6 months was chosen to observe
microstructure and mechanical properties changes at the half point of the target time.
For the experimental temperature of a longtime heat treatment, 650 ˚C was selected. This
temperature covers the application temperature range of combustion liners and turbine shroud
rings as shown in Fig. 2-5. In order to simulate an aircraft engine environment, a cyclic thermal
heat treatment profile was derived from an aircraft engine testing standard [116]. Figure 2-6
suggests the heating profile for the longtime heat treatment. In this profile, the total flight time was
set to 24 hours and the cruising time, when an aircraft is operated at the highest elevation, was
determined by a study reported in an article referring to the world longest non-stop flights [117].
The heating and cooling rates of the profile follow the condition suggested in the standard [116].
56

A Laboratory 9/12/C450 muffle furnace (Nabertherm Inc., Lilienthal, Germany) was utilized to
conduct this long term cyclic thermal cycle. This furnace provides a temperature capacity up to
1200 ˚C with its maximum heating rate of 13 ˚C/min. The controller of this furnace has a
functionality to operate a repeating thermal cycle which is appropriate for this study. A VCD
software (Nabertherm Inc., Germany) gives convenience for users to control the furnace with a
computer. Additionally, the thermal cycle repetition from the Nabertherm heater is only viable
using this software. A cooling fan offered an air cooling to the furnace through a front hole in the
heater after the heating steps.

57

Figure 2-6. A heating profile of the longtime heat treatment cycles for 6 months and 1 year in
this study.

58

2.2.3. Creep Testing
A creep tester, made by Arcweld Manufacturing Company, is used for creep tests at elevated
temperatures. The front view and a schematic of this creep tester are given in Figure 2-7. An ideal
creep testing system installation is required to obtain a reliable quality for creep test data. This
section illustrates the configuration of the creep testing system of this study. As one part of the
effort to secure creep data quality, all devices which consume electric powers for operation are
connected to external batteries to prevent any unexpected disturbance, such as power outrages
during creep tests.

As suggested in Fig. 2-7b, this creep testing machine applies a load on the sample by a lever system
which has the distance ratio of 20:1 between the distance from input to fulcrum and the distance
from output to fulcrum. Thus, by the lever rule, this creep tester applies a load to the output which
is 20 times higher value to the applied weight on the input. Aside from this loading system, there
are two other main components in the creep testing machine: the furnace and the data acquisition
system with a linear variable distance transducer (LVDT). The following sub-sections will
illustrate the details of those components.

59

Figure 2-7 . The setup of Arcweld creep testing machine used in this study (a) The front view of
the testing machine(b) a schematic of the creep testing machine.

60

2.2.3.1. Furnace System
An induction furnace 3210 97-1370 (Applied Test Systems Inc., Butler, U.S.A.) is used to obtain
elevated temperature conditions for creep tests. This furnace has a maximum temperature capacity
of 1100 ˚C without any external cooling system. As the ends on both sides of this furnace have 30
cm holes, these two ends need to be appropriately closed with specially designed ceramic end caps
to prevent heat loss during creep tests.
The equipped three zone heating system of this furnace enables a homogeneous temperature. These
three heating zones are controlled by a combination of Eurotherm 808 temperature indicators and
Eurotherm 832 thermal controllers. Each zone has a K-type thermocouple to measure the
temperature. The temperature of each zone is measured by the temperature indicator. An operator
can enter a setting temperature by using Eurotherm 808 module. This module then controls the
Eurotherm 832 power controller to adjust the furnace to reach to the sett temperature. This furnace
system provides heating rates of 25 – 30 ˚C/min until it reaches the target temperature. However,
the absence of external cooling system does not allow the heater to have any faster cooling rate.

In addition to the three thermocouples for the three heating zones in the furnace, one other K-type
thermocouple is equipped with the furnace. The function of this extra thermocouple is to measure
the real temperature of the creep testing sample. Thus, the end tip of this thermocouple touches
directly on the surface of the sample to check if the sample is within the target temperature. An
Omega engineering 831A thermometer is used to measure the specimen temperature.

61

2.2.3.2. Data Acquisition System
A proper installation of a data acquisition system is required for creep tests to obtain reliable creep
data. Generally, a data acquisition system for a creep test has three main components: a
displacement reader, a signal conditioner and a data acquisition module.

First, the displacement reader measures the current length of the sample, or gauge length. An
LVDT is one type of this displacement reader. It converts the physical value of measured gauge
length into an electrical signal of AC or DC voltages. An Omega Engineering GP 911-5 AC LVDT
was used in this study. This LVDT can measure the physical displacement up to 0.4 inches. The
electrical signal obtained from a LVDT typically shows strong fluctuation due to some interference
such as insensible ground vibration, surrounding electric or magnetic fields from other electronic
devices and unstable power input. The noise from the unstable power input can be suppressed by
connecting the LVDT to a power conditioner, not to a normal electric outlet. In this data acquisition
system, an Emerson Sola HD 23-22-112-2 powder conditioner, which regulates the output voltage
to have ± 1% noise at maximum, was used to provide a clean electrical power to the LVDT.

Despite applying a power conditioner, the LVDT usually generates relatively small noise which
cannot be attenuated by the power conditioner. This noise level would not be a problem for
macrolevel measurements, but for microlevel measurement ( ≤ 10
-6
m). A signal conditioner is the
module to regulate the small noise. This device is equipped with a complicated electrical circuit.
This electrical circuit accepts the input signal from the LVDT and this signal is regulated when it
passes some large impedances in the signal conditioner, and the regulated signal is properly
amplified and sent to a recording module. An Omega Engineering LDX-3A signal conditioner is
62

utilized in the creep tests. This device is also connected to the power conditioner to prevent any
electrical noise from an irregular voltage input.

The regulated electrical signal is now transmitted from the signal conditioner to a data acquisition
module. This data acquisition module collects and records the electric signal data from the LVDT.
The module then transmits the signal data to a personal computer. The actual electric signal
conversion into a physical length is conducted in the computer. In order to secure high quality data
measurement, a computer software-based signal regulation is also recommended. A data
acquisition data module as Omega Engineering iNet-601, which is used in this study, can record
the dataset directly into a computer. This module has a good empirical stability in the data
recording frequency range between 0.1s
-1
to 10000s
-1
.

Thermal fluctuation due to heat generation of electronic devices such as the signal conditioner and
the data acquisition module can disturb the quality of gage length measurements. Prevention of
this potential disturbance is viable by using a C&W CW 3000 industrial water chiller. Several
winded copper cooling channels surround those possible heat sources (electrical devices) and the
distilled water chiller circulates the cooling channels to keep temperature of these devices uniform.
The reason for using distilled water is to prevent the blockage of water flow due to the formation
of chemical compounds by a reaction between impurities in the tap water and the copper channels.
Additionally, as one more effort to reduce the data noise, all the electronic devices in use are
electrically grounded to reduce some possible disturbances from their internal electric noises.

63

2.2.3.3. Other Conditions in The Creep testing System
The creep test samples of this study are designed in accordance with ASTM E8 standard [118] and
it is presented in Figure 2-8. Both ends of the creep specimen have screw threads. The gauge length
of the specimens is 32 mm in length with 6.35 mm diameter.

As the creep tests are conducted at elevated temperatures at 650 ˚C and 800 ˚C, some components
of the loading system inside the furnace are required to have high oxidation resistance and
acceptable strength at temperature. Thus, Ni-based superalloys are used for those components. The
extensometer, which moves to measure displacements of the creep testing sample, is made of
Inconel X-750 alloy. However, a set of extensometer inserts that directly contact the specimen
during creep tests is fabricated using direct-aged Inconel 718 due to its excellent hardness at
elevated temperatures. The higher hardness of the insert than the creep testing sample can forestall
a slippage of the extensometer and this is significantly important. Any slippage indicates the
detachment of extensometer from the sample and the whole data acquisition system cannot collect
reliable creep testing data once it occurs.

The grip to hold the screw part of the specimen is made of Hastelloy X to prevent the specimen
from getting stuck when the specimen needs to be separated from the grip after a creep test.
Hastelloy X is a representative Ni-based superalloy with excellent oxidation resistance. Boron
nitride is also applied to the screw part as lubricant. Other Inconel rods shown in Fig. 2-7b are
made of Inconel 713.
64

Figure 2-8. The design of the creep test sample used in this study.

65

2.2.3.4. Creep Testing Procedure
A creep test of this study includes three main stages: the heating phase, the creep testing phase
under an applied stress and the cooling phase. Figure 2-9 shows the detailed steps of a creep test.

In the heating phase, a creep sample is heated to the target temperature with a heating rate of 25 –
30 ˚C/min. The furnace then is sustained for 30 minutes after it reaches to the set temperature to
obtain an equilibrated temperature throughout the whole heating area of the furnace. An actual
creep test is started after the 30 minutes of the temperature homogenization step by applying a load
on the sample as illustrated in Fig. 2-7b. As the TTT diagram in Figure 1-2 suggests, it is expected
that precipitations of secondary phases can be avoided if creep tests are restricted within 24 hours.
Hence, short-term creep tests for either as-HIPed or as-solution annealed (SCT), were performed
within 24 hours to minimize the possible effect of secondary phases on the creep behavior. For
LHT samples, creep experiments were conducted up to rupture, under the applied stress of 658
MPa for 650 ˚C and 192 MPa for 800 ˚C, to compare the creep ductility of AM IN625 and wrought
IN625 after LHT. Thus, the typical time limit of each creep test was 24 hours except for unexpected
cases such as an early rupture and the creep tests of LHT samples. Finally, upon the end of a creep
test cycle, the applied load is removed and the furnace is turned off to cool down.

66

Figure 2-9. The high temperature creep test procedure of this study.

67

2.3. Chemical Composition Analysis
As previously mentioned in Section 2.1.1, IN625 is imbued its strength, high temperature oxidation
resistance and other characteristics from alloying elements in its matrix. According to the ASTM
standard of this alloy [100], typically 12 different elements except for Ni are included in this alloy.
As discussed in section 1.1.1, some of those elements (Cr, Mo and Nb) are regarded particularly
important to understand the strength of IN625. In addition to that, section 1.1.2 illustrates that
some minor elements such as oxygen and sulfur must be considered to investigate high temperature
embrittlement of IN625. Therefore, careful chemical composition analyses are required to interpret
the mechanical properties of IN625 samples in this study. The following is the list of composition
analysis methods used in this study and elements measured by each method are given.

Optical Emission Spectrometer (ICP-OES): Mg, Cr, Fe, Mo, Nb, Mn, Si, Al, Ti, Co, S
Optical Emission Spectrometer (ICP-MS): Zr, Hf
Spark spectrometer: Cr, Fe, Mo, Nb, Mn, Si, Al, Ti, Co, S, P, Mg, Cu, V, W, B, Pb, Zr
C/S analyzer : C, S
Oxygen and Nitrogen determinator: O, N

Table 2-2 presents the result of chemical composition analysis for Powder, bulk AM IN625, and
bulk wrought IN625.
Specific details for the techniques above are discussed in the following sub-sections.

68

Table 2-2. The chemical composition data of specimens of this study in wt%
Specimen
Elements
Ni Cr Fe Mo
Nb +
Ta
C Mn Si Al Ti Co W
Powder Bal. 19.95 0.65 7.72 3.70 0.008 0.04 0.20 0.43 0.32 0.18
AM IN625 Bal. 20.72 <2.0 8.98 4.15 0.014 0.05 0.07 0.32 0.37 0.22 0.17
Wrought
IN625
Bal. 22.22 4.28 8.29 3.54 0.065 0.35 0.22 0.14 0.23 0.05 0.26

O N S P B Pb Mg Zr Hf

Powder 0.020 0.011 0.0034

AM IN625 0.020 0.013 0.0033 <0.001 0.0028 0.002 <0.001 N/D N/D

Wrought
IN625
0.002 0.019 0.0017 0.002 0.0033 <0.001 0.008 <0.001 N/D

N/D: Not detected

69

2.3.1. Optical Emission Spectrometer
The inductively coupled plasma (ICP) method is a powerful source for multi-element analysis in
a sample with its low detection limit of less than ppb level. A wide linear dynamic range and
precise detection result. There are two representative categories in ICP; one is optical emission
spectroscopy (OES) and the other is mass spectroscopy (MS). The advantage of ICP-MS method
against ICP-OES is its low detection limit ( ≤ ppb). In the case of ICP-OES method, it is cheaper
than ICP-MS and more appropriate for high-matrix elements with its detection limit ( ≈ ppm) [119].
In this study, the OES method is utilized to measure chemical composition of powder of AM IN625
and, Mg and Ca content level in bulk AM IN625 and wrought IN625. Zr and Hf in bulk AM IN625
and wrought IN625 is detected using ICP-MS method.

The term ICP represents a partially ionized gas generated in a quartz torch using a power supply.
Samples for a measurement are generally placed at the center of the plasma which is typically Ar
ions. The reaction between elements in the sample and the Ar plasma atmosphere excites those
elements to be ions (M →M + e
-
+ hν) under the temperature condition between 3000 – 9000 K.
This ionization emits photons which have their identical frequencies depending on the elements
and a monochromator or polychromator monitor those emitted photons either sequentially or
simultaneously. As the detected signal intensity and its wavelength vary with both the type of ions
in the plasma and the fraction of those excited ions of interest, the presence of elements and their
fraction in the sample can be determined by analyzing the detected signal from the sample [119].

An Optima 7300DV (Perkin-Elmer, Waltham, U.S.A) is the ICP-OES machine used for the
analysis. The typical procedure of this machine requires preparation of a sample in a liquid state
70

before introducing the sample to the machine. IN625 needs to be dissolved in an acid such as
nitric acid, hydrochloric acid or hydrofluoric acid to be a liquid sample. One negative aspect of
IN625 as an ICP sample is that, due to its complex chemical composition, it requires all of three
acids listed above to be fully dissolved in a liquid mixture. In addition, IN625 samples from the
bulk has a slow rate to be solved in those acids. Thus, a powder sample is the proper form to
conduct this analysis because of the large specific area of powders, as a result, only AM IN625
powder before LPBF fabrication was analyzed using this method at Inha University. In order to
analyze the content level of minor elements in bulk AM IN625 and wrought IN625, the samples
were first cut to ≈ 100 µm thick chips and then sliced to be ≈ 2 mm × 2 mm size.

2.3.2. Spark Spectrometer
Despite these advantages of the ICP-OES and ICP-MS techniques, they require a very complex
preparation for analysis, and other methods are considered to obtain the entire chemical
information of matrix elements in bulk IN625 samples. The spark spectrometer is one of the
methods. Since this method is extensively used in the metal industry, it is suitable for the analysis
of bulk IN625 samples [120]. The basic principle of the spark spectrometer is almost identical to
that of ICP method. The sample introduced in a spark spectrometer and is vaporized or excited by
the supplied energy. There are two main sources of energy: electrical and non-electrical. Lasers,
chemical flames and thermal sources are the non-electrical energy sources whereas electrical
sources which includes ICP, arc, spark from a voltage discharge. Thus, the ICP method can be a
category of the spark spectrometer. However, while the ICP method requests all elements to be
ionized in a chemical solution, the spark spectrometer can analyze a solid-state sample.
71

Specifically, it uses electrical discharge to create sparks on the solid sample where an electrode is
located. The high energy sparks are the result of avalanche break down, a form of electron
avalanche between the sample electrode and a counter electrode. The heat generated by sparks
ionizes the chamber gas, typically pure Ar, and ablates elements from the sample. It disperses
photons possessing characteristic wavelengths and detectors equipped with the spectrometer
measures the light signal and then analyzes the type and the quantity of elements [120].

An QSN 750-Ⅱ (OBLF GmbH, Witten, Germany) spark spectrometer is used in this study to obtain
the composition of matrix elements in both AM IN625 and wrought IN625. The composition
analyses were conducted at Korea Institute of Industrial Technology (KITECH). This machine
provides relatively quantitative composition data from a solid sample and the measurement is
typically completed in 30 seconds. However, this method does not guarantee a low detection of
ppm. Probably, the detection limit of this method is about 10 – 100 ppm, and this is one
disadvantage of this apparatus against the ICP method. In this study the spark spectrometer
requires 10 mm diameter wide samples with a proper thickness of more than 1 mm. In order to
guarantee the reliability of the measurement data, a flat and well-polished surface up to the 1200
grit is prepared.

2.3.3. C/S Analyzer
The acquisition of a precise quantity of carbon and sulfur are significantly important to Ni-based
superalloys. As previously illustrated in section 1.1.1, IN625 is a complex alloy capable of
possessing diverse types of secondary phases which includes carbides such as MC, M6C and
72

M23C6. These carbides can influence mechanical properties of this alloy, as will be discussed later.
The formation of carbides is strongly influenced by the content of carbon in the alloy. Previous
studies [50,121] reported that carbide formation was reduced as the carbon concentration decreases
in IN625 and Inconel 718. Therefore, precision is required to measure the carbon content in both
AM IN625 and wrought IN625 alloys.
Sulfur has another importance, particularly with regard to the high temperature embrittlement.
Sulfur is one main element which causes a serious embrittlement in Ni-based superalloys as
discussed in section 1.1.2. As it will be discussed more later, the embrittlement from sulfur is due
to grain boundary segregation of this impurity. Since a minimal amount of sulfur in the superalloy
can be detrimental to the ductility of the material, sulfur content control and measurement are
especially important to obtain desirable mechanical properties. Hence, a highly quantitative
analysis for sulfur is required to interpret the mechanical properties of IN625.

It is very desirable that sulfur can be detected by using ICP methods, however, the detection of
carbon and sulfur is not recommended in the ICP analysis. The C/S analyzer has been known as
the most reliable method to obtain a quantitative analysis for carbon and sulfur [122]. The basic
principal of a C/S analyzer is a mass measurement. At the very first stage of this technique, a
sample for measurement is introduced to an induction furnace which has the temperature ability
above 2000 ˚C. The furnace chamber, then, is filled with high purity oxygen (typically >99.5%
pure) and then heated up to melt the sample. The carbon or sulfur elements in the sample react
with oxygen molecules of the gas environment and these reactions produce gases such as CO, CO2,
SO2, and SO3. The generated gases pass through a dust filter and moisture absorber for purification.
In the next step a set of infrared cells measure the weight of each chemical compounds and that
73

measurement provides the sulfur and carbon contents data [122]. This method allows to obtain
quantitative compositions of sulfur and carbon under ppm level.

An CS-800 carbon sulfur analyzer (ELTRA GmbH, Haan, Germany) at Korea Advanced Institute
of Science and Technology (KAIST: Daejeon, South Korea) was chosen for the analysis. The
analysis was conducted at least three times per one sample to ensure reliability. The sample size is
required to weigh between 100 to 200 mg and the surface of samples was ground to remove any
excessive oxidation layer. The polishing grit was typically 400. The gas environment in this
analysis was kept under the 99.999 % pure oxygen.

2.3.4. Oxygen and Nitrogen Determination
A LPBF process can be categorized as a type of powder metallurgy. In powder metallurgy process,
a high oxygen content was typically reported [86]. The oxygen concentration value of a PM sample
in the reference was 40 times higher than that of a typical wrought Ni superalloy sample [51]. In
addition, as the consequence of high oxygen content, the degradation in high temperature ductility
was observed in Inconel 718 [51]. Therefore, a precise determination for oxygen concentration in
samples is necessary to clarify if the oxygen effect on the mechanical behavior of a Ni-based
superalloys. However, as with sulfur and carbon, oxygen cannot be detected using ICP and spark
spectrometer, so that oxygen and nitrogen determination was used for the purpose.

The operation sequence of an oxygen/nitrogen determination is similar to that of the C/S analyzer.
One of differences between two methodologies is the type of crucible of the induction furnace.
74

While a C/S analyzer uses ceramic materials, the oxygen/nitrogen determination uses a high purity
graphite crucible. For the operation, a weighed sample is placed on the graphite crucible and is
inserted for the oxygen/nitrogen determination. The sample is then melted under helium gas stream
by the inter gas fusion principle at temperatures sufficient to release oxygen and nitrogen [123].
The oxygen in the sample, whatever form present, reacts with the carbon in the graphite crucible
to form CO and CO2. The nitrogen contained in the sample generates molecular nitrogen upon a
full evaporation. The released gases first enter the infrared absorption module and pass through
CO and CO2 detectors. By these detectors, the presence of oxygen in the sample firstly observed
by detecting CO or CO2 in the passing gas. In next step the gases pass through heated copper oxide
catalyst to convert CO to CO2 and to remove some residual hydrogens in the gas. The purified
gases are then injected again to the infrared absorption CO2 detector for final oxygen measurement.
In the case of nitrogen, thermal conductivity of the gas is measured to determine the nitrogen
content. Upon the stage of the purification with the heated copper oxide catalyst, CO2 and nitrogen
are separated from each other and the nitrogen gas flows to the thermal conductivity measurement
cell for nitrogen detection [123].

A TC-800 oxygen/nitrogen determinater (Leco Co., St. Joseph, U.S.A) at Research Institute for
Industrial Science & Technology (RIST: Pohang, South Korea) was used to measure the oxygen
and nitrogen impurity levels in the samples. The specimen for this technique needs to be a
cylindrical sample of 4 mm diameter × 10 mm height. In order to avoid a measurement error due
to excessive oxide layers of the sample surface, samples were mechanically grinded first and a
subsequent chemical pickling was conducted to clean the surface. The solution composition of
75

chemical pickling is 88 ml HNO3 (60% purity), 8 ml HF (50% purity), and 665 ml distilled water
at 50 ˚C.

2.4. Microstructural Characterization Methods
As mechanical properties of IN625 are given by microstructure characteristics, multiple
characterization methods are leveraged to investigate the features of the experimental IN625
alloys. Microstructural observation, phase fraction analysis, sample preparation processes and
methods are representative topics for this part. These various techniques are discussed in the
following section.

2.4.1. Sample Sectioning
Sample sectioning is the first stage of all microstructural characterization methods. Since few
machines for microstructure analysis accept non-sectioned samples after creep tests, a proper
sample sectioning is required to conduct some subsequent analyses. Computer numerical control
milling, which is performed to machine creep testing samples, would be one of choices for this
purpose. One another possible choice is electrical discharge machining (EDM). An EDM machine
can section any material with a high precision of 0.001 mm if the material is electrically
conductive. One advantage of EDM is that the machining speed is not affected by the hardness of
materials while a computer numerical control milling requires replacement of the cutting tip
depending on hardness of the material.

76

Typically, 80 % Cu – 20 % Zn brass wires were used for sectioning using an EDM. Brass wire of
EDM has a thickness range between 0.1 – 0.3 mm. This thin brass wire touches and passes through
the specimen during a sectioning process. Both the cutting wire and the sample are electrically
charged and a discharge from these counter-charged materials partially melts the sample and the
consecutive partial melting of sample is the basic principle of EDM operation. This indicates that
the melting-affected surface (roughly ≈ 50 µm deep) should be removed to investigate the
“reliable” microstructure for this study. An EDM has processing parameters such as wire voltage
(V), sample voltage (SV), and wire feeding speed (SF). These parameters also conclude the sample
cutting speed and the depth of melting-affected region. Therefore, a careful control of these
parameters is required.

An A360 EDM machine (Sodick, Yokohama, Japan) was used to cut the samples of this study.
The sample for cutting should be sunk thoroughly into a deionized water chamber. This machine
is capable of sectioning various metals such as aluminum, magnesium, iron, copper and nickel.
The processing parameters to section IN625 were followed: V: 8, SV: 15, and SF: 25-30. This
setting offered the maximum available sample cutting rate ( ≈ 1 – 1.5 mm/min). In a situation that
the sample needed to be thinned to under 500 µm, the wire voltage value was set to 0 to avoid
warping the sample. However, there is a relatively low cutting rate at this setting ( ≈ 0.1 – 0.2
mm/min).

2.4.2. Sample Mechanical Polishing
Generally, most microstructural observations require a high-quality polishing for smooth and
scratch-free sample surface. Thus, a proper polishing method should be conducted to obtain
77

reliable microstructure observation data. There are two different methods for sample polishing.
One is the polishing by hand and the other is the automatic polishing by using an auto-polishing
machine.

A GLP Korea polishing wheel was used to grind the sample by hand using SiC sandpapers. Starting
from 400 grit,SiC paper, the polishing was followed using 600, 800 grit papers down to 1200 grit
(~ 2.5 µm median particle). A subsequent polycrystalline 1 µm diamond paste polishing was
conducted with a Struers Velcloth polishing cloth after the set of sandpaper polishing process. For
finishing the polishing process, either 0.02 or 0.04 µm colloidal SiO2 solution with Struers Final
A polishing cloth was used. A 0.05 µm Al2O3 particle liquid can replace the final step for colloidal
SiO2 polishing. However, as the SiO2 polishing solution partially reveals microstructure of Ni-
superalloy due to high pH value of the solution, the SiO2 final polishing is highly recommended.
The rotating speed of the polishing wheel was kept around 150 rpm for all hand polishing
processes. An automatic vibratory polishing is a desirable alternative for hand polishing,
particularly for the final stage. Since the quality of sample surface is significantly sensitive to the
result of final polishing process, applying a well-regulated load on the sample is ideal to prevent
scratches or minor deformation of the sample. Buehler Vibromet 2 was the machine for vibratory
polishing in this study. During the vibratory polishing, MasterMet 2 collidal silica suspension was
used under the power gauge setting of 50 over 2 – 3 hours.
As previously mentioned, applying an unregulated load to the sample would be detrimental to
microstructure observation as it can cause scratches and micro-deformation on the sample surface.
An automatic polisher avoids those expected problems with the hand polishing method. An Allied
METPREP 4
TM
auto polisher was used for most sample preparation in this study. The detail of
78

parameter setting for machine polishing is suggested in Table 2-2. The grinding papers or polishing
cloths of every step of machine polishing were the same as those of the hand polishing method. In
this automatic polishing method, samples should be mounted in a 25 mm graphite disk for grinding
with the automatic polishing method. Buehler SilmliMet 1000 was used to mount samples on the
25 mm graphite disk.

79

Table 2-3. Parameter settings for each step of the automatic polishing
Polishing step
Parameters
RPM Force (lbs) Mode Time
SiC sandpaper (400 – 1200 grit) 150 3 Compressor 3 min
Polycrystalline diamond (1 µm) 130 2 Compressor 10 min
Colloidal silica (0.04 µm) 120 1 Contrast < 8min

80

2.4.3. Optical Micrograph
As the term “microstructure” indicates, the microstructural observations are conducted under
micro- or nano- scale levels. Optical micrographs are one of the tools to approach microscale
observation. This micrograph uses a set of optical lenses to magnify images from the sample.
Typically, it supports the magnification range from 20x to 5000x depending on the optical
micrograph model.
Optical microscopy techniques are applied when the sample preparation quality needs to be
checked. Before conducting other further analyses such as scanning electron microscopy,
transmission electron microscopy, and electron probe microanalysis, optical microscopy
observations are conducted to check if scratches are present or the grain structure is revealed as
intended. Pore density of LPBF samples was measured using this method as well. MX53M Optical
Microscope (Olympus, Tokyo, Japan) was utilized for optical microstructural observation. The
details of porosity measurements will be shown in a next section (section 2.4.7).

2.4.4. Scanning Electron Microscopy (SEM)
The Scanning Electron Microscope (SEM) is one of the most widely used instruments applicable
to the examination and characterization of the microstructure morphology and chemical
composition of the sample surface. While optical microscopy has the limit of resolution of ~ 200
nm, SEM has resolution of 0.5 – 4 nm due to the difference in their illumination sources. Since the
resolution of optical microscopy depends on the wavelength of the illumination source (photon),
the electron source enables SEM to have finer resolution than optical microscopy [124]. Image
acquisition in the SEM is the output of collecting signals produced from the interaction between
81

the electron beam and specimen surface. The incident electron beam is scattered into various
signals, including backscattered electrons, characteristic x-rays, Auger electrons, and secondary
electrons [124].

Among those listed signals in the previous paragraph, secondary electron emission signals are
collected to construct the surface morphology image of specimen. As these electrons have low
energy around 3 – 5 eV, they can only travel within a few nanometers of the sample surface. Thus,
secondary electrons accurately indicate the position of atoms on the surface and provide high
resolution topographic images [124]. Another notable signal is backscattered electrons (BSE). In
contrast to secondary electrons, backscattered electrons are generated by the elastic collision
between the incident electron beam and material atoms. The intensity of BSE signal depends on
the crystallographic orientation or chemical composition, so it gives information of grain structure
with contrast variation in the image. One disadvantage of using BSE to construct SEM images is
that the lateral resolution of BSE image is considerably worse than that of a secondary electron
image due to the high energy of BSE (~ 50 eV) [124]. Characteristic x-rays are the other important
class of collected signals in the SEM machine. X-rays are emitted as the incident electron beam
ionizes electrons on the shell of the material atoms. Since the ionization energy of shell electrons
is differs with each element, analyzing the energy of emitted x-ray photon provides chemical
information of the surface. This is the basis of the energy dispersive x-ray spectroscopy (EDS)
technique [124].
There are three main controllable operation parameters in the SEM: electron voltage, current
density and working distance [125]. At a high value of electron voltage, the accelerated electrons
increase the number of larger energy signals such as BSE and characteristic x-rays. Thus, typically,
82

the higher electron voltage, the less secondary electron-based SEM image quality. However, it is
expected to obtain better results in BSE mode or EDS analysis at the expense of SEM image
resolution. Current density relates to the number of incident electrons to specimen surface. The
SEM image becomes brighter at higher current densities. SEM image resolution tends to be
degraded at larger current density condition due to interference from other signals. In the case of
working distance, smaller working distance is generally appropriate for better SEM image qualities
whereas larger working distance is effective to obtain good results from BSE or characteristic x-
ray-based analyses (such as EDS). This is due to the fact that secondary electron can escape from
the sample surface much shorter than BSE or characteristic x-rays.

In this study, a JSM-7001F-LV SEM (JEOL Ltd., Tokyo, Japan) was utilized to obtain SEM-based
data. Depending on the purpose of examination, different operating parameter settings were
adjusted. The working distance was fixed at 10 mm and the probe current value was typically
under 10 µA if an SEM image needed to be obtained. 10 kV electron voltage was good for sample
surface morphology observation (SEM image). However, a 20 kV setting was recommended for
EDS chemical composition analysis. The sample preparation for SEM was conducted after
chemical etching of the sample surface. A Struers Lectropol-5 electrochemical polisher was used
to reveal the microstructure of the samples. One condition for this is 5 - 10 V, 20 ˚C, flow rate 5 –
10 with 10% H3PO5 water solution for 10 s. However, sometimes, this method was not effective
to obtain a good etching quality, particularly if the total fraction of secondary particles is over ~
15%. In this case, 20% HClO4 in ethanol solution under 20 V for 10 – 20 s replaced with the first
method.

83

2.4.5. Electron Backscattered Diffraction (EBSD)
As the BSE signal collection can exhibit information on crystallographic orientation of grains on
the sample surface, it implies a potential that the grain orientation of each grain can be identified
by collecting the corresponding BSE signal. Therefore, the microstructure of the surface can be
easily analyzed using the BSE signal. Due to this property, the Electron Backscattered Diffraction
(EBSD) technique can determine crystal structure of crystalline solids [124]. Average grain size,
misorientation between adjacent grains, and texture distribution of a sample can be easily
determined using EBSD. Additionally, subgrain boundary misorientation can be measured using
EBSD. The typical precision value of the measurable misorientation is between 0.5 ˚ to 2.0 ˚ [126],
however, several up-to-date EBSD cameras support a better precision less than 0.05 ˚ [127].

Typically, the EBSD camera is equipped with a SEM machine. Therefore, the EBSD setting shares
the three important operation parameters with the SEM. First, the optimal accelerated voltage value
was set to 20 kV for IN625 samples in this study. This value may be able to be increased up to
30kV. When it comes to the probe current density, 14 µA is the maximum value available for the
JSM-7001F-LV SEM machine, so that this value was used for taking EBSD images. The working
distance was 15 mm and, most importantly, the EBSD sample must be tilted by 70 ˚ from the
bottom during the image collection process.
As EBSD data includes not only grain size, but also grain misorientation and subgrain
misorientation, the sample surface must be perfectly scratch-free to avoid any possible EBSD data
distortion. Most of EBSD samples in this study were prepared through the automatic polishing
process described in section 2.4.2 to secure the best surface quality. In some cases when an
84

acceptable sample quality was not obtained from the automatic polishing process, an additional
step was followed after the polishing. An IM4000 (Hitachi, For EBSD, Tokyo) ion milling
machine was used to enhance the sample flatness in this case. An operator of Seoul National
University helped with the operation of the ion milling machine. The acceptable quality of EBSD
data was determined by the absence of scratches and the value of confidence index over 0.6. Two
different EBSD imaging condition were used in this study. For general grain structure observation,
a 700 µm × 350 µm area was selected with a step size of 1 µm. The average grain size and
misorientation profile of non-creep deformed AM IN625 and wrought IN625 samples was
investigated using this EBSD technique.
In order to conduct the secondary crack analysis, a 80 µm × 80 µm with area, including at least
secondary crack, was analyzed by EBSD with a step size of 0.2 µm. The main purpose of this
EBSD analysis was to clarify whether the crack was intergranular or intragranular. In addition to
that, the misorientation distribution of intergranular secondary cracks was also determined from
EBSD analysis. Figure 2-10 illustrates how to measure the misorientation between grains next to
a secondary crack of a creep tested AM IN625 sample. In this figure, the distance profile in Fig.
2-10b was obtained along the yellow line from the starting point to the end point in Fig. 2-10a. In
Fig. 2-10b, the point-to-origin profile means the measured misorientation between the origin
(=starting point) and each data point corresponding to the distance values in the x axis. Hence,
from Fig. 2-10, two values of intergranular misorientation next to a secondary crack were
determined to be ≈ 48 ˚ and ≈ 26 ˚ as shown in Fig. 2-10b.
Four different AM IN625 samples were analyzed using EBSD for the secondary crack analysis:
SCT AM IN625 at 650 ˚C and 658 MPa, SCT AM IN625 at 800 ˚C and 192 MPa, 6 months LHT
85

AM IN625 at 650 ˚C and 658 MPa, and 6 months LHT AM IN625 at 800 ˚C and 192 MPa. Totally,
37 cracks were examined to derive the misorientation profile.

86

Figure 2-10. An example of the intergranular misorientation determination next to a secondary
crack (a) the EBSD image, (b) misorientation profile along the yellow line in (a).

87

2.4.6. Transmission Electron Microscope (TEM)
Transmission electron microscopy (TEM) is an attractive instrument to reveal nano-scale structure
and internal fine structure in solids. The spatial resolution of TEM a image is higher than SEM
image due to its relatively large electron energy value, typically 200 – 300 kV than that of SEM
(~ 30kV maximum in general). As the De Broglie’s mass wave equation indicates, the high
electron energy causes electrons to have short wavelength. Thus, this electrons with smaller
wavelength enables TEM images to be finer than SEM images [124,128].

The imaging methodology is another difference between TEM and SEM. As mentioned earlier,
the SEM image is constructed by collecting scattered electron signals from the interaction between
the incident beam and specimen. Despite an identical illumination source, TEM displays an image
in a way similar to OM. While SEM uses deflected electrons from sample surface, TEM construct
an image by gathering electrons which penetrate the thin sample. As the contrast of TEM image
depends on the number electron penetrating a certain part of sample, TEM image may represent
the grain orientation or different phases in a sample. This is because, the number of transmitting
electrons varies with the crystallographic orientation and chemical composition [124,128,129].
One advantage of using TEM is to obtain crystallographic information from specific positions in
thin specimens. The injected electron beam produces a dot pattern with diffractions by the
crystallographic structure. Since this selected-area diffraction (SAED) pattern contains the
crystallographic plane where diffraction occurs and the lattice parameter data, analyzing the SAED
pattern verifies the phase where electron diffraction occurs. From Bragg’s law with the small angle
approximation, an equation is suggested to show the relationship between the incident electron
beam and the observed phase [129]:
88

𝜆𝜆 𝐿𝐿 = 𝑑𝑑 ℎ 𝑘𝑘𝑘𝑘
𝑅𝑅 (2-1)

where λ is wavelength of the incident electron, L is camera length, 𝑑𝑑 ℎ 𝑘𝑘𝑘𝑘
is lattice plane spacing
and R = measured diffraction distance between two adjacent diffraction spots. By using this
equation, lattice parameter spacing can be calculated.

As the electron source implies, the signals used in SEM method can be collected in TEM to obtain
experimental data as well. The combined TEM with scanning electron method is called a scanning
transmission electron microscope (STEM). In the STEM mode, the image is constructed by
collecting scattered electron signals with a probe. In addition to that, as it is applied to the SEM
technique, characteristic x-ray signals provide the chemical content information of the STEM
specimen. In other words, EDS analysis can be conducted under the STEM mode. The resolution
of STEM technique is affected by the incident beam diameter which is generated by the probe-
forming lens. The finer beam diameter gives the higher resolution in both scanned image and
analytic result. The innate error in the incident beam system of STEM or TEM do not allow those
instruments to exhibit their theoretical performances. Abbreviation-corrected (Cs-corrected) TEM
or STEM are the improved machines which are adjusted to overcome the incident beam system
problem [128].

In this study, a JEM 2100F HR FE-TEM (JEOL LTD., Tokyo, Japan) was used to observe the
sub-micron microstructure of samples and to conduct phase analyses using an SAED pattern. A
JEM 2100F Cs-corrected Analytic STEM (JEOL LTD., Tokyo, Japan) was utilized to obtain
chemical composition data at sites of interest for STEM based experiments. A 200 kV electron
89

beam was applied to both TEM machines. HR FE-TEM analysis was conducted with an operator
of National Nano Fab Center (Daejeon, South Korea) and, in the case of Cs-corrected Analytic
STEM, an operator of Korea Advance Nano Fab Center (Suwon, South Korea) helped with this
study.
Generally, sample preparation for TEM analysis was carried out by a twin-jet polisher. A Struers
Tenupol-5 electrochemical jet polisher was used for TEM sample preparation. Before inserting
sample to the machine, the sample was cut to ≈ 300 µm thick with the Sodick EDM machine, and
grinded up to 80 – 100 µm which is suitable for electro-jet polishing. The last step just before the
jet polishing was to punch the sample to make it a 3 mm disc. A Gatan 659 disc puncher was used
for this purpose. The typical electropolishing condition for IN625 alloys was 20 V, flow rate of 42
– 45 at temperatures between -30 and -40 ˚C. For some samples which have larger secondary
particles and high fractions of them, the electropolishing condition is changed into 20 V, flow rate
of 40 at temperatures between -60 ˚C and -50 ˚C. For the former condition, 20 % HClO4 solution
in ethanol was used and 10 % HClO4 in ethanol liquid was applied for the latter condition.
Precision ion milling (PIPS) is an alternative for the electro-jet polishing. In this method, specimen
becomes thinner up to be available for TEM observation using Ar ion physical etching. PIPS
samples need to be thinner less than 10 µm before the ion milling. Typically, a dimple grinder is
used to make the PIPS sample thin. In this study, specimens are chemically grinded with the
electrochemical polisher. In next step, the grinded sample was placed on the PIPS machine. Then,
the ionized Ar beam etched the center part of the sample. The injection angle of the Ar beam was
set to 4˚ under 4 – 5 eV condition (current density usually between 20 to 30 µA). In this study, a
Gatan 691 precision ion polisher system was used.

90

2.4.7. Secondary Phase and Porosity Quantification Method
As previously mentioned in Section 2.1.1, the mechanical behavior of IN625 essentially partly
depends on the quantity of secondary phases. Hence, secondary phase fraction is one important
factor to interpret the mechanical strength of IN625. Generally, the more secondary phases are
present in the matrix of IN625, the stronger the material. As one derivative from PM technologies,
LPBF involves a consideration on porosity. Pore density and morphology can be a influence the
mechanical properties of LPBF materials. Therefore, an appropriate method to measure porosity
should be proposed.

In this study, a classical method based on image analysis for porosity or second phase fraction
characterization was used. The volume fraction of γ’’ and δ was measured by using either SEM or
TEM images as shown in Figure 2-11. The secondary phase fraction of phases that reside on grain
boundaries was measured by dividing the grain boundary length having secondary particles by the
total grain boundary length (grain boundary particle frequency) as shown in Fig. 2-11g to i. As it
will be shown later in section 3.2.3, random grain boundaries (non-CSL boundaries) were the
primarily precipitation sites for second phases. Likewise, dislocation density measurements of
non-creep deformed samples were conducted using TEM images of as-HIPed AM IN625, as-
solution annealed AM IN625 (non-HIPed) and as-solution annealed wrought IN625. Fig. 2-11c
and d illustrates how to determine the dislocation density from TEM images. The thickness of
TEM samples was approximated as 150 nm for the dislocation density measurement [130]. All
measured experimental values in this paper, including the secondary particle volume fractions,
91

grain boundary particle frequency and dislocation density, were determined by averaging at least
5 different collected data.
OM observation was performed in order to determine the porosity of the samples in this study. For
OM images, dark field mode was used to acquire black & white binary drawings as shown in Fig.
2-11e and f. At least 10 images were collected to determine the average porosity value of the
following samples: as-HIPed AM IN625, as-solution annealead AM IN625 (non-HIPed), and as-
solution annealed wrought IN625.
The image conversion process was conducted using Adobe Photoshop CC and the fraction
calculation and other related image analyses were performed with NIH ImageJ freeware which
calculates the fraction of black regions over an entire white image or vice versa. The uncertainty
range of those experimental values (secondary phase fraction, dislocation density, grain boundary
occupation frequency, and porosity) were given by the standard deviation of the average values.

92

Figure 2-11. Examples of image conversion from various microscopy sources (a) an original TEM
image source for γ’’ phase quantification, (b) the binary converted drawing from (a), (c) an
original TEM source for dislocation density measurement, (d) the binary image converted from
(c), (e) an original OM image source, (f) the dark field OM image of (c), (g) an original SEM
image source for grain boundary occupation frequency, (h) the binary image of total grain
boundary length from (g), (i) the binary image of grain boundary area occupied by secondary
particles from (g).

93

2.5. Other Material Characterization Methods
In this section, several extra material characterization methods used in this study will be
introduced. Those methods appear awkward to be simply categorized into either chemical
composition analysis or microstructural characterization methods. Electron probe microanalysis,
nano-secondary ion mass spectroscopy, atom probe tomography, and x-ray diffraction are those
methods.

2.5.1. Electron Probe Microanalysis – Wavelength Dispersion X-ray Spectroscopy
Electron probe microanalysis (EPMA) can be simply defined as one type of SEM technique.
However, these two methodologies differ from each other due to their main application objective.
Despite their identical illumination source from electrons, EPMA delivers excellent chemical
information on the sample surface whist SEM concentrates on the resolution in images of specimen
surface. This suggests, EPMA has advantages to identify secondary phases and visualization of
impurity segregations against SEM. Moreover, EPMA can be interpreted as a “high-current mode”
SEM, so the higher beam current translates into higher BSE density or x-ray counts after the
electron beam interacts with specimen. In other words, the increases in the amount of BSE and
characteristic x-ray due to the high current density enable EPMA to produce a large contrast
variation in BSE images and to boost characteristic x-ray signals reaching the detector, hence the
enhancement in chemical information quality [131].
Wavelength dispersion x-ray spectroscopy (WDS) is a quantitative composition analysis at a spot
size as small as a micron level. WDS shares the identical analytic source as with EDS, which is
characteristic x-rays generated from specimen surface. The principal difference between those two
94

methods is how the characteristic x-rays are analyzed. While EDS measures the energy of
generated x-rays due to the interaction between incident electron and elements of specimen
surface, WDS uses the wavelength of those generated x-rays to determine the element. Figure 2-
12a shows a schematic of the principle of WDS analysis. Once x-rays escape from the sample,
those x-rays reach an analytical crystal which has a geometry for reflecting only the x-rays of
selected wavelength. The geometry of this analytical crystal can be changed by rotating the crystal.
This implies that an analytical crystal can measure only one element per each step of WDS analysis
whereas EDS delivers a total composition at the area of interest at once. However, as shown in
Fig. 2-12b, WDS provides excellent resolution in composition characterization due to its superior
selectivity in the analytic process. Typically, 4 – 5 analytical crystals (= detectors) are equipped in
an EPMA machine for the WDS analysis. It suggests that 4 – 5 elements can be measured per each
step of typical WDS analysis [131,132].

A JXA-8500F FEG-EPMA (JEOL Ltd., Tokyo, Japan) was used to conduct WDS analyses in this
study. An operator oat the Korea Institute of Science & Technology (KIST: Seoul, South Korea)
supported the operation of EPMA for this study. The samples for this analysis were prepared using
the automatic polishing machine with 0.04 µm colloidal silica finish. Those samples were mounted
in 25 mm conductive graphite mountings due to the sample dimension requirement of the EPMA
machine.

95

Figure 2-12. (a) a schematic of WDS analysis principle, (b) X-ray spectra from WDS and EDS
analyses [132].

96

2.5.2. Nano-Secondary Ion Mass Spectroscopy
One disadvantage of WDS analysis in the previous section is that the available resolution of this
method cannot be less than 1 µm. As illustrated in section 2.4.4, the electron beam of both SEM
and EPMA has the size at least around 1 µm, and the spatial resolution of WDS should be limited
within 1 µm.

Secondary ion mass spectroscopy (SIMS) analysis can suggest a solution for this case. The basic
principle of SIMS stems from an ion probe which uses finely focused primary ion beam (Cs
+
ion
in the case of Nano-SIMS). This ion beam erodes the target and generates secondary ions from the
sample surface. Due to the size of primary ions, the typical analytic resolution of Nano-SIMS can
be down to 50 nm [133]. These sputtered ion particles are then referred to as “secondary ions”.
The identification of these secondary ions can be done in two different ways: time-of-flight and
magnetic mass spectrometer. Between the two, the magnetic mass spectrometer is used in the
Nano-SIMS method. In this technique, the separation of these ejected secondary ions of different
mass-to-charge ratio are physically initiated by the Lorentz force as a magnetic field is applied on
them [133]. One advantage using this method is a better resolution than time-of-flight method to
separate two similar ions. For example, the
32
S
2-
ion can be easily separated from
32
O2
2-
using this
method, while the separation is difficult in the time-of-flight method. As shown in Figure 2-13,
32
S
2-
and
32
O2
2-
ions were successfully separated from each other in the Nano-SIMS machine, so
that this method is one of the best methods to clarify the sulfur segregation in AM IN625 samples
of this study. Additionally, the AlO
-
ion indicates the presence of Al2O3 inclusion in the AM IN625
sample and as it is shown in Fig. 2-13b, the separation of AlO
-
ion from BO2
-
was successful
enough to be reliable.
97

A Nano-SIMS machine (Cameca Nano-SIMS-50L) was used to assess several element
distributions at and in the vicinity of random (not CSL) grain boundaries in both AM IN625 and
wrought IN625. Random grain boundaries for this analysis and atomic probe tomography (APT)
were picked by using EBSD. Specifically, after putting intentional scratches using a knife to select
an area of interest, an EBSD analysis was conducted for the area. Figure 2-14 shows how an area
of interest was selected and random grain boundaries were found. This figure also suggests that
those random grain boundaries of creep tested AM IN625 and wrought IN625 samples were
decorated with secondary particles, while the precipitation only occurred sparsely on special
boundaries.
Four different samples were prepared for the Nano-SIMS analysis: as-HIPed AM IN625; as-
solution annealed wrought IN625; 6 months LHT AM IN625 (not creep tested); 6 months LHT
wrought IN625. The Nano-SIMS tests were conducted at the California Institute of Technology.
The negative ion signal collection mode was selected to detect carbon, oxygen, sulfur, and Al2O3
in the samples through
12
C,
16
O (or
18
O),
32
S and
27
Al
16
O signals.
11
B
16
O and
52
Cr
24
C2 signals were
additionally collected in the 6 months LHT samples to assess the boron distribution (
11
B
16
O) and
M23C6 carbide with the
52
Cr
12
C2 signal. All signals in the Nano-SIMS analysis were normalized
by the matrix element (
58
Ni). All samples were ion-sputtered at least for 2 minutes before the
Nano-SIMS analysis in order to clean the sample surface. The current condition for the ion-
sputtering was 300 pA, and the Nano-SIMS analysis was conducted at 3.5 pA. Dr. Yunbin Guan
operated the Nano-SIMS machine at Caltech to help this study.
98

Figure 2-13. The mass profiles of ions in the Nano-SIMS to show how two similar mass-to-charge
ratio ions are separated (a)
32
S
2-
and
32
O2
2-
ions, (b)
43
BO2
2-
and
43
AlO
-
ions.

99

Figure 2-14. SEM and EBSD grain boundary maps of 650 ˚C SCT AM IN625, which shows
precipitates prefer to form at random boundaries and how to specify random grain boundaries
(non-special) using EBSD.

100

2.5.3. Atom Probe Tomography
As most high temperature embrittlement of Ni-based superalloys exhibits intergranular fracture, it
is very important to investigate the impurities at grain boundaries in AM IN625. However, even
by the spatial resolution of 50 nm in Nano-SIMS might not be sufficient to observe some small
impurity segregation at grain boundaries. In order to secure sub-nanometer level spatial resolution
at grain boundaries (not observable using Nano-SIMS), atom probe tomography (APT) was used.

Typically, the APT delivers a spatial resolution of < 1 nm and, in the best case, 0.1 nm is possible
[134]. The APT is a combination of a mass spectrometer and a field ion microscope [135]. In the
view of field ion microscope, a high positive voltage is first applied on a sharply pointed needle
with a quasi-hemispherical end (the tip radius is ≈ 50 nm). This strong voltage applied on the
sample requires cryogenic condition in the range of 15 – 110 K (or less than this level) to produce
a strong electric field. The intensity of this electric filed is typically over 30 – 50 V ⋅nm
-1
to be high
enough to allow layer-by-layer erosion of atoms at the sample surface [134,135]. In other words,
the surface atoms are field ionized and evaporated from the sample as the consequence of applying
a high electric field. These evaporated ions now travel towards a phosphor screen placed typically
50 mm ahead of the sharp needle. In the convention filed ion microscope, gas atoms around the
needle sample are ionized and detected by the phosphor screen [135]. As those ions have different
mass and charge depending on species, the kinetic energy of ions varies with their mass-to-charge
ratio. Simply speaking, a substantially charged light element reaches the detector more quickly
than a heavy element with low charge [135]. Therefore, the identification of elements in a
reasonably high spatial resolution (< 1 nm) is viable by using APT through this time-of-flight mass
spectrometer. However, one disadvantage of this time-of-flight method against the magnetic mass
101

spectrometer of Nano-SIMS is the low mass resolution to separate two similar ions. For example,
as the typical ∆mass/mass available in the APT is ≈ 1/1000 [135], this mass resolution cannot
separate effectively
32
S and
32
O2 in Fig. 2-13a as it requires a mass resolution of at least ≈ 1/2000.
This low mass resolution can be partially overcome by considering the natural quantity ratio
between isotopes. For example, if an overlapped peak by
32
S
+
and
32
O2
+
ions is found, the first
thing to do will be the check of
34
S
+
and
36
O2
+
peaks. The level of
32
S
+
can be expected from the
intensity of
34
S
+
peak based on the natural abundance ratio between
32
S and
34
S. Similarly, the
32
O2
+
intensity can be calculated by using the natural quantity ratio between
16
O and
18
O. This
suggests that the absence of
34
S
+
peak directly indicates the absence of sulfur in the sample. Hence,
despite of the unfavorable mass resolution to separate O and S, the APT can determine the
quantities of sulfur and oxygen in a target throughout the calculation process.

Two creep-tested AM IN625 samples were chosen for the APT analysis of this study: short-term
creep tested at 650 ˚C and 416 MPa within 24 hours (as-HIPed before the test) and the other was
6 months LHT and creep tested at 650 ˚C and 658 MPa. Random grain boundaries were
investigated and those grain boundaries were selected through the same method as illustrated in
Fig. 2-14. A Cameca LEAF 5000 APT machine at Northwestern University collected the APT data
signals for analysis. The APT samples were prepared using a focused ion beam sample-prep
machine. Both APT sample preparation and the computer-aided APT data interpretation were
conducted by Prof. Dieter Isheim of Northwestern University.

102

2.5.4. X-ray Diffraction Analysis
X-ray diffraction (XRD) is one of the most universal experiment to investigate crystallographic
information of a sample. The main principle of this study is originated from Bragg’s law:

𝜆𝜆 𝑥𝑥 − 𝑟𝑟 𝑟𝑟𝑟𝑟
= 2 ∙ 𝑑𝑑 ℎ 𝑘𝑘𝑘𝑘
∙ 𝜀𝜀 𝑠𝑠 𝑙𝑙 𝑠𝑠 (2-1)

In equation 2-1, λx-ray represents the wavelength of incident x-ray beam and θ is the glancing angle
(the angle between the incident beam and the normal line to the sample surface). The generated
beam from x-ray source travels towards the sample of interest and the incident beam is diffracted
when the glancing angle satisfies the diffraction condition suggested in equation 2-1. The detector
in the XRD machine senses the diffraction and records the diffracted beam intensity. The beam
intensity recording is conducted by adjusting glancing angle throughout a XRD experiment for the
smallest setting θ to the largest θ setting. The θ values where high beam intensity is measured
indicate the interplanar spacing, dhkl, of crystalline phases in the sample. Hence, the XRD results
gives not only which phases are present in the sample, but also the lattice parameters of those
phases.

Two different XRD machines were used in this study. 11-BM high resolution powder diffraction
at Argonne national laboratory (IL, U.S.A) was first used to obtain XRD data for 4 different
samples: as-HIPed AM IN625, SCT AM IN625 at 650 ˚C within 24 hours, as-solution annealed
wrought IN625, SCT wrought IN625 at 650 ˚C at 650 ˚C within 24 hours. This machine requires
the sample size 1 mm width × 8 mm height × < 100 µm thickness and to be secured in a hollow
plastic cylinder. A glue was injected using a syringe to secure the sample at the center of the plastic
cylinder. The machine is equipped with an x-ray source having ≈ 0.4 Å wavelength. The nominal
103

2θ step size was 0.001 ˚ and the time interval per each step was set to 0.1 s. The actual XRD
experiments were conducted by an operator at Advance Photon Source of Argonne national
laboratory through a mail-in service.
Smartlab XRD machine (Rigaku, Tokyo, Japan) was one other XRD machine used in study. 4
different samples were analyzed using this machine: 6 months LHT AM IN625, 1 year LHT AM
IN625, 6 months LHT wrought IN625, and 1 year LHT wrought IN625 (all samples were non-
creep deformed). λ ≈ 1.54 Å . The nominal 2θ step size was 0.01 ˚ and the time interval per step
was set to 1 s. An operator of Incheon university helped with this analysis.

The XRD data from 11BM was used to calculate dislocation density of as-solution annealed
wrought IN625, as-HIPed AM IN625 and as-solution annealed AM IN625. X-ray diffraction line
profile analysis (XLPA) method was conducted for this purpose. A reference LaB6 standard
powder data from the Advanced Photon Source (APS at Argonne national laboratory) was used.
A former graduate student at Inha university, Jae-Chul Lee, entirely conducted this analysis.

104

3. Part Ⅰ. HIGH TEMPERATURE CREEP BEHAVIORS OF AM IN625
In this part, high temperature creep behavior data of both AM IN625 and wrought IN625 will be
presented. The thermal treatments conducted on samples were simply divided into three
conditions: as-HIPed or as-solution annealed, 6 months LHT, and 1 year LHT.

A plot between log [ 𝜀𝜀 ̇ 𝑚𝑚 𝑡𝑡𝑡𝑡 (minimum creep rate = steady-state creep rate)] and log [σ/E] will deliver
some information of the creep strength of AM IN625 and wrought IN625 at 650 ˚C and 800 ˚C.
The plot also suggests some changes in creep strength of both AM IN625 and wrought IN625 after
6 months or 1 year LHT. In addition to that, high temperature ductility data of AM IN625 and
wrought IN625 will be reported at the two temperatures. The effect of 6 months or 1 year LHT on
the creep ductility is also shown in the ductility plots.

Lastly, the high temperature creep mechanism is investigated by analyzing the log [ 𝜀𝜀 ̇ 𝑚𝑚 𝑡𝑡𝑡𝑡 ] and log
[σ/E] plot and calculating the creep activation energy. Solid-solution strengthening was firstly
considered to explain the extra creep strength of AM IN625. The measured extra creep strength of
AM IN625 is also compared to several calculated extra strength data from anti-phase boundary
shearing, mechanical twinning, and dispersion hardening.

3.1. Creep Curves of Non-LHT AM IN625 and Wrought IN625
High temperature creep tests for non-LHTed IN625 samples in this study were conducted at 650
˚C and 800 ˚C. Stress condition of creep tests was between 65 – 658 MPa. More specifically, the
applied stresses for 800 ˚C tests were in the range of 65 MPa – 192 MPa and those of 650 ˚C creep
105

tests were between 294 MPa to 658 MPa. Figure 3-1 and Figure 3-2 propose several creep curves
of non-LHTed AM IN625 and wrought IN IN625. The temperature condition of creep tests in Fig.
3-1 was 650 ˚C and 800 ˚C was the temperature of creep tests in Fig. 3-2. In Fig. 3-1 and Fig. 3-2.
The x and y axes are set to the same scale for as-HIPed AM IN625 and wrought IN625 for a
convenient comparison.

106

Figure 3-1. Several representative creep curves for non-LHT samples at 650 ˚C (a) AM HIPed
IN625 at 294 MPa, (b) AM HIPed IN625 at 466 MPa, (c) wrought IN625 at 294 MPa, (d) wrought
IN625 at 466 MPa. The number on the bottom right is the minimum creep rate (= steady-state
creep rate).

107

Figure 3-2. Several representative creep curves for non-LHT samples at 800 ˚C (a) AM HIPed
IN625 at 65 MPa, (b) AM HIPed IN625 at 192 MPa, (c) wrought IN625 at 65 MPa, (d) wrought
IN625 at 192 MPa, The blue dash line in (d) is an extrapolated curve and the number on the bottom
right is the minimum creep rate (= steady-state creep rate). The X mark suggests where a rupture
occurred.

108

From the creep curve data above (Figs. 3-1 and 3-2), at the two temperatures, minimum creep rates
of AM HIPed IN625 samples were generally lower than those of wrought IN625 while the creep
ductility of wrought IN625 was larger than AM HIPed IN625 at all conditions.

At the temperature of 650 ˚C, there were a large increase in plastic strain in the early stage of creep
deformation for the all creep curves of Fig. 3-1. In other words, the plastic strain obtained within
first 1 hour has roughly ≥ 50 % plastic strain in the total strain over 24 hours. The degree of this
“elastic deformation-like” strain was measured highly in AM IN625 samples when it compared to
wrought IN625 samples. As the yield strength of IN625 at 650 ˚C is ≈ 260 MPa [8], those stress
conditions used in this study exceeded this value for IN625 alloy. One interesting aspect in Fig. 3-
1 is the difference in the creep curve between the two alloys. While wrought IN625 samples seems
to undergo a primary creep state throughout the entire deformation, it appears that AM HIPed
IN625 samples entered the secondary creep regime in a few hours after launching the tests.

One clear difference of graphs between Fig. 3-1 and 3-2 is the absence of large plastic strain in the
primary creep region of Fig. 3-2. As the yield strength of IN625 does not noticeably drop from
650 ˚C to 800 ˚C [8], the low applied stress conditions at 800 ˚C appears a reason for the difference
between the two figures. According to a reference [8], the yield strength of IN625 at 800 ˚C is ≈
250 MPa.
At 800 ˚C and 65 MPa (Fig. 3-2a and 3-2c), a similar primary creep region to those 650 ˚C creep
tests were shown within 2 – 3 hours. After the primary creep region, both AM IN625 and wrought
IN625 entered into a steady-state until ≈ 11 hours (AM IN625) and ≈ 7 hours (wrought IN625)
after starting those creep tests. After that time, as shown in Fig. 3-2a and 3-2c, an abrupt jump in
109

the creep rate was observed and the increased strain rate was conserved until the creep tests were
finished. Several speculations can be suggested for this strange creep behavior:

1. Weakening in the solid solution strengthening effect due to secondary phase formations,
2. Activation of secondary slip system
3. Transformation of primary coherent precipitate (γ’’) into other precipitates (δ).
4. Malfunction of extensometer

For the speculation 1, as Nb and Mo are the most powerful solid strengthening elements for IN625
[10], some decreases in matrix concentration of Nb and Mo can drop the material strength of
IN625. At 800 ˚C, the precipitation of δ phase can cause a drop in matrix Nb content [14].
However, in the case of Inconel 718, it was reported that the volume fraction of δ phase, at
temperatures close to 800 ˚C, gradually increased over a long time (> 100 hours) with a gradual
decrease in the matrix Nb level [136]. This result implies that the creep rate must have gradually
increased, rather than sudden jumps.
Typically, activation of secondary slip system requires a high resistance for dislocation slip on the
primary slip plane. Therefore, the creep rate jump should have shown in Fig. 3-2b and 3-2d (higher
applied stress than Fig. 3-2a and 3-2c) if this mechanism led to this phenomenon. Additionally, as
the yield strength at 800 ˚C is ≈ 250 MPa [8], due to its small stress 65 MPa, it is believed that the
applied stress level in Figs. 3-2a and 3-2c seems not sufficient to activate the secondary slip system.
As intragranular δ phase formation requires the pre-existing intragranular γ’’ particles [136], the
transformation of coherent γ’’ to incoherent δ phase typically degraded the yield strength of IN625
over a 1000 hours aging at temperature more than ≈ 670 ˚C [11]. In other words, the metastable
110

γ’’ behaves like a premature phase of stable δ phase above the temperature. However, this γ’’ to δ
transformation does not occur at once and the yield strength degraded continuously up to 1000
hours, so that, similar to the case of speculation 1, this mechanism cannot explain the strain rate
jump as well.
Hence, none of the suggested speculations are applicable to understand the strange strain jumps
shown in Fig. 3-2a and 3-2c. As the main purpose of this study is to compare the mechanical
properties of AM IN625 to wrought IN625, any deeper investigation to this phenomenon was not
conducted in this study.

3.2. Creep Curves of LHT AM IN625 and Wrought IN625
In this section, creep curves of long-term heat-treated (LHT) samples are delivered. Two long-
term heat treatments were conducted; 6 months and 1 year. For both 6 months and 1 year LHT of
wrought IN625, only one stress condition was conducted per the two temperatures of 650 ˚C and
800 ˚C (658 MPa for 650 ˚C and 192 MPa for 800 ˚C).
However, for 650 ˚C creep testing data of both 6 months and 1 year LHT AM IN625, in addition
to 658 MPa (AM HIPed then LHT), 466 MPa and 416 MPa tests (AM non-HIPed, solution treated,
then LHT) were conducted. One more creep test result at 294 MPa (AM HIPed then LHT) was
added in the 6 months LHT AM IN625. In the case of 800 ˚C creep testing, three different stresses
(65 MPa, 104 MPa, 192 MPa) were applied to the 6 months LHT AM HIPed IN625 samples.
However, only 192 MPa data is available for 800 ˚C creep test of 1 year LHT AM HIPed IN625.

111

Figure 3-3 and 3-4 give several creep curves for those LHT AM IN625 and LHT wrought IN625
samples at different temperatures of 650 ˚C (Fig. 3-3) and 800 ˚C (Fig. 3-4).

112

Figure 3-3. Several representative creep curves for LHT samples at 650 ˚C (a) 6 months LHT AM
HIPed IN625 at 294 MPa, (b) 6 months LHT AM HIPed IN625 at 658 MPa, (c) 1 year LHT AM
HIPed IN625 at 658 MPa, (d) 6 months LHT wrought IN625 at 658 MPa, (e) 1 year LHT wrought
IN625 at 658 MPa. The number on the bottom right is the minimum creep rate (= steady-state
creep rate), the blue dash line in (e) is an extrapolated curve, X mark suggests where a rupture
occurred or would have occurred.
113

Figure 3-4. Several representative creep curves for LHT samples at 800 ˚C (a) 6 months LHT AM
HIPed IN625 at 65 MPa, (b) 6 months LHT AM HIPed IN625 at 192 MPa, (c) 1 year LHT AM
HIPed IN625 at 192 MPa, (d) 6 months LHT wrought IN625 at 192 MPa, (e) 1 year LHT wrought
IN625 at 192 MPa. The number on the bottom right is the minimum creep rate (= steady-state
creep rate), the blue dash line in (e) is an extrapolated curve. The X mark suggests where a rupture
occurred or was anticipated to occur.
114

When Fig. 3-3 is compared to Fig. 3-1 (non-LHT samples), for 650 ˚C creep tests, one remarkable
difference is the amount of plastic strain in the primary creep region drastically decreased after
both 6 months and 1 year LHT even at 658 MPa. One possible explanation for this is the long time
thermal exposure effect on the yield strength. Some previous papers [138,139] reported that a 500
hours heat treatment on wrought IN625 at 650 ˚C was effective to increase the yield strength.
Those paper observed that the yield strength of aged wrought IN625 was almost double to that of
as-solution annealed wrought IN625. Additionally, some drops in minimum creep rates for the
samples in Fig. 3-3 compared to the data in Fig. 3-1 also support that IN625 alloys had been
strengthened during 6 months and 1 year LHT.
In those 650 ˚C, 658 MPa creep tests, 6 months LHT AM HIPed IN625 was still stronger (lower
minimum creep rate) than 6 months wrought IN625. However, the creep strength of wrought
IN625 became almost identical to that of AM HIPed IN625 after the 1 year LHT. Despite some
ductility drop in the wrought IN625 sample between 6 months and 1 year LHTs, this ductility is
much larger than AM HIPed IN625 in both 6 months and 1 year LHT conditions.

Unlike the case of Fig. 3-2a and 3-2c, Fig. 3-4a (6 months LHT AM HIPed IN625) does not show
any noticeable change in creep rate over the 24 hours creep tests at 800 ˚C, 192 MPa. Similar to
the results in Fig. 3-3, even at 800 ˚C, 192 MPa creep tests, the minimum creep rate of both 6
months and 1 year LHT AM HIPed IN625 was lower than wrought IN625 samples of the same
thermal history conditions. The minimum creep rates of the two wrought IN625 samples in Fig. 3-
4d and 3-4e did show almost negligible decrease when those compared to Fig. 3-2d. Also, rather
than increasing the extra strengthening, the creep strength of 1 year LHT wrought IN625 appears
115

slightly degraded as 6 months LHT wrought IN625 exhibited a lower creep rate than the 1 year
LHT wrought IN625 sample as shown in Fig. 3-4d and 3-4e.
However, when ductility is concerned, some significant embrittlement of AM HIPed IN625 was
observed. For example, the ductility of 6 months LHT AM HIPed IN625 was roughly 8 % of that
of 6 months LHT wrought IN625. Not only when compared to the wrought IN625 samples in Fig.
3-4d and 3-4e, but also when compared to Fig. 3-2 (non-LHT AM HIPed IN625), both 6 months
and 1 year LHT AM HIPed IN625 samples exhibited poorer ductility than the non-LHT condition
in the 800 ˚C creep test.

3.3. Creep Strength, Ductility and Mechanism of AM IN625
Figure 3-5 shows a summary of all creep data conducted in this study. This figure is a set of a log
[ 𝜀𝜀 ̇ 𝑚𝑚 𝑡𝑡𝑡𝑡 ] vs log [σ/E] plot and the applied stress – ductility plots at both 650 ˚C and 800 ˚C.

116

Figure 3-5. (a) minimum (= steady-state) creep rate versus the modulus compensated (normalized)
stress. (b) creep ductility at 650 ˚C (c) creep ductility at 800 ˚C. The ductility at the two test
temperatures is indicated by solid circle dots for SCT samples, X marks for 6 months LHT samples
and + marks for 1 year LHT samples.

117

The results of Fig. 3-5a suggest, first, that the creep strength of the AM IN625 is always at least
equal to the creep strength of the wrought alloy at both elevated temperatures. In fact, the AM
IN625 may have slightly superior creep strength. The strength of those AM samples seems to be
improved more after 6 months and 1 year heat exposure at a service temperature. This is an
important result in that AM IN625 over a period of 1 year at 650 ˚C can replace the wrought alloy
at least in terms of strength.

Fig. 3-5b and 3-5c also suggest that high temperature creep ductilities of AM IN625 samples were
substantially inferior to those of wrought specimens. It is also evident that HIPed specimens have
much improved ductility over the non-HIPed IN625. However, it still appears that the AM alloy
(both HIPed and non-HIPed) have inferior ductility compared to the wrought alloy. This is evident
after both 24 hours, and, perhaps, more substantially, after six months and 1 year under the LHT.
The AM IN625 alloys after both 6 months and 1 year appear “embrittled”.
Fig. 3-5b shows the 650 ˚C creep ductility. At 658 MPa, when only AM HIPed IN625 data were
compared, the SCT sample showed larger uniaxial strain than the LHT sample. There was no
noticeable difference in ductility between the 6 months and 1 year LHT samples. In the case of
wrought IN625, at 658 MPa and 650 ˚C, while the 6 months LHT did not degrade the ductility, a
drop in ductility of one specimen at 650 ˚C was observed after the 1 year LHT. No drop in ductility
was observed after 1 year LHT and tested at 800 ˚C. In several previous studies for wrought IN625
[8,40,41], there were drops at ambient temperature tensile elongation after 500 and 8000 hours
aging at 650 ˚C. This is not fully consistent with the observations of ductility at 650 ˚C and 800
˚C. It was proposed that grain boundary δ precipitation led to the ductility drop in these other
studies. As will be discussed later, this appears unlikely. The 800 ˚C creep ductility for AM HIPed
118

IN625 alloy is reported in Fig. 3-5c. The ductility difference between SCT and LHT AM IN625
samples (both 6 months and 1 year) at 800 ˚C, 192 MPa was ≈ 0.06, which is smaller than the gap
at 650 ˚C, 658 MPa (≈ 0.2). In the case of wrought IN625, no ductility drop was observed in the
800 ˚C, 192 MPa creep even after 1 year LHT.

Table 3-1 lists extra creep strength and rupture time values, and ductility differences in creep tested
SCT AM IN625, 6 months (and 1 year) LHT AM and wrought IN625 samples compared to “the
SCT wrought IN625 data”. In Table 3-1, the extra creep strengths of all AM IN625 alloys are
typically higher than those of wrought IN625 at the same creep testing conditions. In both AM
IN625 and wrought IN625, at 650 ˚C and 800 ˚C, all 6 months and 1 year LHTed samples showed
more strength than SCT AM IN625 and wrought IN625.
Table 3-1 also proposes that the rupture times of wrought IN625 are mostly longer than AM IN625
alloys at the same creep conditions. When compared to SCT sample rupture times, the LHT over
both 6 months and 1 year was effective on increasing or at least conserving the rupture time for
both AM IN625 and wrought IN625. In the two LHT AM samples, at the 800 ˚C creep test, some
drop in the rupture time was observed in the 1 year LHT AM IN625 compared to the 6 months
LHT AM IN625. However, at 650 the rupture time was almost the same between those 1 year and
6 months LHT AM IN625 alloys. Similarly, the 800 ˚C creep rupture ductility decreased in the
LHT wrought IN625 over the interval between 6 months and 1 year. Moreover, unlike the case of
AM IN625, the 650 ˚C creep rupture time decreased in the 1 year LHT wrought IN625 when
compared to the 6 months LHT wrought IN625.

119

Table 3-1. The measured extra strength and ductility changes for creep deformed IN625 alloys in
this study. SCT wrought IN625 was used as reference to assess the extra strength and ductility
changes. Only 650 ˚C, 658 MPa and 800 ˚C, 192 MPa creep data were used.
Creep condition
Rupture
time
(hours)
Measured
extra strength
(MPa)
Ductility changes
AM solution
annealed
AM HIPed
AM
IN625
650 ˚C short-term creep 2.6 10 -0.49 -0.38
800 ˚C short-term creep 6.4 41 - -0.59
6 months LHT + 650 ˚C creep 19.9 240 - -0.61
6 months LHT + 800 ˚C creep 11.4 63 - -0.64
1 year LHT + 650 ˚C creep 19.6 312 - -0.59
1 year LHT + 800 ˚C creep 6.6 47 - -0.64
Wrought
IN625
650 ˚C short-term creep 16.1 Used as reference
800 ˚C short-term creep 17.5 Used as reference
6 months LHT + 650 ˚C creep 25.7 94 +0.03
6 months LHT + 800 ˚C creep 43.3 38 +0.06
1 year LHT + 650 ˚C creep 16.6 312 -0.28
1 year LHT + 800 ˚C creep 17.7 14 +0.11

120

3.3.1. High Temperature Creep Mechanism of AM IN625
In Fig. 3-5a, for the 24 hours creep data at 650 ˚C, the n value of AM IN625 was 12.1, and that of
wrought IN625 was measured to be roughly 11 for 650 ˚C. These n values are typically considered
to suggest creep in the power-law breakdown regime [92]. The n value of SCT AM IN625 was
measured to be 5.5 at 800 ˚C. The n value for the as-solution annealed wrought IN625 alloy was
6.8 in the same creep regime. These values lie between 4 to 7, which is consistent with five power-
law creep in a class M alloys (pure metal behavior), where the dislocation climb is regarded as the
rate-controlling process. Unlike the case of 650 ˚C creep tests, the 800 ˚C creep tests of AM IN625
showed a negligible change in n value after the 6 months LHT.

The green line in Fig. 3-5a indicates the other possibility to derive the n value of SCT AM IN625
alloy at 650 ˚C. According to the green line, a transition of n value from 4.8 to 18.6 was observed
at log(σ/E) = -2.6. This consideration is inspired by a database [8] which observed the n value of
a non-solution annealed IN625 alloy was 4.2 in the log(σ/E) range from -2.8 to -2.6, then the n
value changed into 24.7 on the region of log (σ/E) > -2.6 at 650 ˚C. Additionally, a report [102]
suggested that the stress exponent of long-time crept IN625 (< 3600 hours) was 4.7 in the log (σ/E)
range between -2.85 to -2.55 at 650 ˚C. Interestingly, this n value transition is similar to that of a
precipitate hardening Ni-based superalloy, Inconel 718. According to Molins et al. [46], a stress
exponent transition occurred at 500 MPa, and specifically, it changed from 5.9 to 13.6 in the 650
˚C creep test of Inconel 718. This n value transition of other precipitation-hardened Ni superalloys
at intermediate temperatures has been reported in other studies as well [140-145], and one paper
[145] suggests that the main creep mechanism in the lower stress regime is the dislocation climb
over coherent precipitates such as γ’ and γ’’. This explanation sounds reasonable as the stress
121

exponent of solution strengthened or pure Ni matrix, in power law creep-regime, is typically
around 4 – 5 [54,55], which satisfies the n value of 4.8 in the present study and other values in
[8,102] under the lower log(σ/E) range less than -2.6. In the higher stress regime, either Orowan
bowing or precipitate shearing are believed as the dominant creep mechanism [140-145].
Specifically, Ni et al. [146] reported that the deformation mechanism of precipitation-hardened
superalloys showed dislocation climb ahead of precipitates in the n value range 4 to 6, and
precipitate cutting becomes dominate when n value is over 8.
However, one important result in Fig. 3-5a is that this transition is not shown for 6 months LHT
AM IN625 samples in 650 ˚C creep tests. As expected from Fig. 1-2 (the TTT graph) and other
studies about IN625 [13,14,138,139], γ’’ precipitation should have largely occurred over the 6
months LHT, hence more precipitate hardened material behavior should be exhibited in the 6
months LHT AM IN625 than SCT AM IN625. This suggests that the n value transition due to the
precipitation hardening behavior seems hard to be introduced to the SCT AM IN625 data.
Therefore, it is difficult to conclude that a transition in n value is applied to the creep data of SCT
AM IN625 samples at 650 ˚C, instead, it seems reasonable to determine the n value is constant as
12.1 over the experimental stress regime of 650 ˚C SCT AM IN625.

In accordance with the determination above, several previous studies have reported that the n value
range of wrought IN625 is between 10– 13 over the normalized stress range between -2.9 to -2.4
at 650 ˚C [8,105,147]. Thus, the n value of the as-solution annealed wrought IN625 in Fig. 3
(approx. 11) is within the reported n range. Similarly, AM IN625 exhibited the n value of 12.1.
Typically, this high n value suggests power-law breakdown (PLB) in single phase metals and
alloys. At 800 ˚C, the steady-state stress exponent of AM IN625 and wrought alloys was measured
122

between 5 – 7 (Fig. 3a). As Purohit and Burke [101] reported, the stress exponent at 800 ˚C of
solution treated wrought IN625 was 6.65 over the normalized stress range from -3.3 to -3.0, which
agrees well with the data in this study. The range of stress exponent is indicative of the five power-
law regime (n ≈ 4 – 7). The liquidus temperature for IN625 is 1350 ˚C [8], so that the two testing
temperatures are 0.57 and 0.66 of the liquidus temperature which are in the region for five power-
law creep and power-law breakdown (PLB) [148]. Thus, the creep mechanism is also the same for
wrought and AM.

In this study, in the experimental temperature range of 650 ˚C to 800 ˚C, the calculated creep
activation energies (Qc) were 273 kJ/mol for the SCT AM IN625 and 284 kJ/mol for the SCT
wrought IN625 (calculated at 263 MPa). The calculated Qc of 6 months LHT AM IN625 was 257
kJ/mol. The Qc of SCT wrought IN625 was also close to the lattice self-diffusion energy of Ni
(287 kJ/mol [69]). The Qc value for reference wrought IN625 alloys in the temperature between
650 ˚C – 800 ˚C was 275 kJ/mol [8,101] and this value is almost identical to that of SCT AM
IN625. According to Sherby et al [149], the creep of Class Ⅱ solid solution alloys, such as Ni-
20Cr, is governed by dislocation climb. Hence, the Qc value of SCT wrought IN625, close to the
self-diffusion energy, verifies well that wrought IN625 exhibits the typical creep behavior of Class
Ⅱ solid solution alloy. Even though those Qc values of SCT AM IN625 and 6 months LHT AM
IN625 are somewhat deviated from Ni self-diffusion energy, it is determined those values are
within the acceptable range as Ni-20Cr alloys exhibited their lattice diffusion energy range of 260
– 290 kJ/mol depending on their purity level [150].
123

Thus, the main creep mechanism of AM IN625 and wrought IN625 over the present stress region
appears to be lattice diffusion-controlled dislocation climb in both the power-law creep (800 ˚C)
and power-law breakdown creep (650 ˚C).

3.3.2. Extra Creep Strength of AM IN625
As clearly shown in Table 3-1, AM IN625 exhibits more than or at least equal strength to wrought
IN625 at all creep testing conditions.

Generally, strengthening of a material can be obtained through four different main mechanisms:
solid strengthening, secondary phase hardening, grain size hardening and dislocation hardening.
As IN625 is a representative solid solution hardening alloy [10,11,151], the solid solution
strengthening should be considered first, however, it seems this mechanism does not need to
discussed in that the solute content difference between the two experimental IN625 alloys is not
sufficient to exhibit a significant difference by the solute effects. From the solid strengthening
effect data of Table 1-1, the calculated solid strengthening effect of AM IN625 alloy with higher
content of Mo, Nb and other elements is ≈ 5 MPa at room temperature. Thus, the higher creep
resistance of AM IN625 alloy cannot be readily explained with solid solution effects.

Dislocation interaction with second phase particles can enhance the strength through Orowan
bowing (both coherent and incoherent particles) or particle shearing mechanisms (only coherent
particles) [140-145]. As discussed in the previous section (section 1.1.1.3.) and as the sufficiently
large volume fractions of either γ’’ or δ (will be shown later in section 4.3.1) indicate, particle
124

strengthening is strongly believed as the main reason for the extra strength of AM IN625 alloy in
this study. Specifically, an increased γ’’ fraction involves an increased hardening effect in nickel-
based superalloys. Some previous studies have reported that the thermally exposed wrought IN625
to 650 ˚C (for 500 hours) gained additional ultimate tensile strength (from 150 MPa to 250 MPa)
at 650 ˚C when compared to the solution annealed state [138,139]. This γ’’ precipitated explains
the reported extra tensile strength of AM IN625 and wrought IN625 at 650 ˚C in Table 3-1 (from
100 MPa to 312 MPa) as the reference SCT wrought IN625 (in Table 2) possessed negligible
amounts of matrix precipitates. Therefore, the increase in extra strength of both AM IN625 and
wrought IN625 under prolonged thermal exposure to 650 ˚C can be explained by the gradual
increase in γ’’ precipitate amounts over the l year LHT as shown in Figure 4-19a. As discussed in
the previous section 1.1.1.3, coherency strengthening, order strengthening, mechanical twinning,
and Orowan strengthening can contribute to the extra creep strength of both AM IN625 and
wrought IN625 at 650 ˚C. In 800 ˚C creep tests, the coherent γ’’ is replaced by δ and it only
provides a dispersion hardening effect [13,14], so that relatively smaller strengthening was
observed in 800 ˚C creep tests than in 650 ˚C.
As it will be suggested later in section 4.1.3, as-HIPed or as-solution annealed AM IN625 include
only Al2O3 without MC carbides in their matrices. In contrast, only MC carbides are present in the
matrix of as-solution annealed wrought IN625. Similar to the case of δ, Al 2O3 in AM IN625 may
provide some dispersion hardening effects in the AM IN625 at both 650 ˚C and 800 ˚C. However,
as will be suggested later in section 4.3.1, the volume fraction of alumina in as-HIPed AM IN625
(≈ 0.4 %) is smaller than the volume fraction of MC carbide in as-solution annealed wrought IN625
(≈ 1.2 %). No MC carbides were present in AM IN625 and Al2O3 was not observed in wrought
IN625. This suggests that the extra strength of AM IN625 was primarily obtained from γ’’ and δ
125

phases, otherwise wrought IN625 should show higher creep strength with ≈ 1.2 % MC carbides
than AM IN625 with ≈ 0.4% Al2O3.
Table. 3-2 proposes the calculated extra strengthening by either γ’’ (650 ˚C) or δ (800 ˚C) phases
by three different mechanisms (anti-phase boundary shearing, mechanical twinning, and
dispersion hardening). Equation 1-2 – 1-4 were used for the calculation. The values of extra
strengthening in Table 3-1 were derived from all available creep data in Fig. 3-5a.

126

Table 3-2. The measured strength and γ’’ (650 ˚C) or δ (800 ˚C) hardening calculation for creep
deformed IN IN625 samples, SCT wrought IN625 was used as reference to assess the extra
strength and ductility changes. Unlike Table 3-1, all available creep data was used for measuring
extra creep strength.

Creep condition
Measured
extra strength
(MPa)
Calculated extra strength (MPa)
Δ σAPB Δ σtwin Δσdis
AM
IN625
650 ˚C SCT 71 ± 41 73 258 629
800 ˚C SCT 41 ± 23 - - 68
6 months LHT + 650 ˚C creep 132 ± 86 268 129 451
6 months LHT + 800 ˚C creep 62 ± 43 - - 134
1 year LHT + 650 ˚C creep 167 ± 125 345 105 470
1 year LHT + 800 ˚C creep 47 - -

Wrought
IN625
650 ˚C SCT Used as reference
800 ˚C SCT Used as reference
6 months LHT + 650 ˚C creep 94 180 137 235
6 months LHT + 800 ˚C creep 38 - - 92
1 year LHT + 650 ˚C creep 312 241 134 240
1 year LHT + 800 ˚C creep 14 - -

Δ σAPB =extra strength due to anti-phase boundary strengthening (order strengthening), Δσtwin =
extra strength due to mechanical twinning formation, Δσdis = extra strength due to dispersion
hardening (Orowan strengthening)

127

In Table 3-2, coherent strain hardening (equation 1-1) was not derived due to the ambiguity to
determine f’, the fraction of γ’’ or δ parallel to the burger’s vector. Additionally, the strength
calculation was not conducted for 1 year LHT creep tested AM IN625 and wrought IN625 due to
unacceptable quality of microstructure observation for those samples.

According to the strengthening calculation in Table 3-2, either anti-phase boundary shearing ( σAPB)
or mechanical twinning strengthening ( σtwin) mechanisms appear adequate to explain the extra
strength of creep tested AM IN625 samples at 650 ˚C.
In the case of AM IN625 creep tested samples at 800 ˚C, while the additional strength of the SCT
IN625 sample maybe related to the dispersion hardening effect ( σdis), however, for the 6 months
LHT AM IN625 at 800 ˚C, the measured extra strength shows large discrepancy with the measured
value of dispersion hardening effect using Eq. 1-4. It seems the suggested morphology factor
calculation in Eq. 1-5 may not be fully applicable to the plate shape of δ phase.
Any of calculated strength of wrought IN625 does not agree with the measured extra strength. This
is presumably because the number of creep data used for the calculation is not sufficient to obtain
a reliable strengthening value and uncertainty range at temperatures of 650 ˚C and 800 ˚C. As it is
shown in Fig. 3-5a, only one data point was given per each LHT wrought IN625 creep testing
condition.

One other possible candidate for the strengthening would be the grain size effect and texture effect
(Taylor factor). However, as it will be shown later in section 4.1.1, the grain size of as-HIPed AM
IN625 was measured to be 14.7 µm and that of as-solution annealed wrought IN625 was 11.9 µm.
This grain size difference between the two IN625 samples does not support a grain size
128

strengthening effect. This suggests that wrought IN625 should be stronger than AM IN625 if only
the grain size effect is main strengthening mechanism in this study. As it is not the case, it is
regarded that grain size effect is not an important factor to interpret the creep strength of AM
IN625 in this study. Additionally, the calculated Taylor factor using EBSD results of experimental
alloys were: wrought-solution annealed: 3.16: AM-HIPed: 3.02: AM-solution annealed: 2.95.
Similar to the case of grain size effect, wrought IN625 is expected to be slightly stronger than AM
IN625 with the Taylor factor effect. Therefore, both grain size and texture effects do not appear a
proper approach to the understanding the extra creep strength of AM IN625.

The last possibility which can influence the strength of materials is the dislocation density. This
effect has been popularly described by the Taylor equation [151]. Generally, the higher dislocation
density, the higher strength of the material due to the mutual interaction between dislocations
during plastic deformation, which hinders dislocation movements. In this study, the dislocation
densities of as-fabricated IN625 samples were measured by both the XLPA method and TEM
observation. The measured dislocation density of wrought IN625 was initially at least 3 times
higher than those of AM IN625 samples in Table 3-3. Thus, it is hard to suggest that the dislocation
hardening effect can explain the superior creep strength of AM IN625 alloys against wrought
IN625 in this study.

In summary, among four investigated strengthening mechanisms in this section, precipitation
hardening effect due to γ’’ (650 ˚C) or δ (800 ˚C) seems reasonable to interpret the extra creep
strength of AM IN625. Some discrepancies between the measured strengthening and the
strengthening calculation at 800 ˚C were also observed for AM IN625. The additional strength of
129

LHT IN625 wrought IN625 does not match with any precipitation hardening mechanism. This
may be due to the insufficient number in creep data points of LHT IN625 wrought IN625.
130

Table 3-3. Calculated (by the XLPA method) and measured (by TEM observations) values of
dislocation densities of three different non-creep tested samples.
Samples
Method
The XLPA method TEM observation
As-solution annealed
wrought IN625
1.00 × 10
13
·m
-2

± 3.20 × 10
11
·m
-2

1.05 × 10
13
·m
-2

± 3.26 × 10
12
·m
-2

As-HIPed AM IN625
9.93 × 10
11
·m
-2

± 9.72 × 10
9
·m
-2

1.38 × 10
12
·m
-2

± 4.07 × 10
11
·m
-2

As-solution annealed
non-HIPed AM IN625
3.61 × 10
12
·m
-2

± 2.70 × 10
11
·m
-2

2.48 × 10
12
·m
-2

± 1.19 × 10
12
·m
-2

131

4. Part Ⅱ. MICROSTRUCTURE EVOLUTION OF CREEP TESTED AM
IN625
As the mechanical properties of materials are essentially related to microstructure of materials, the
observation of microstructure evolution of AM IN625 before and after a creep testing is important
to interpret the creep behaviors of AM IN625.
In order to respond to this necessity, the grain size evolution of both AM IN625 and wrought
IN625 over the 1 year LHT is presented first. The identification of secondary phases in matrix is
also conducted using EDS spectra and SAED pattern analyses. Additionally, the γ’’ or δ volume
fractions of creep tested IN625 samples in this study are conducted as well.

4.1. Microstructure of Non-creep Deformed AM IN625 and Wrought IN625
In order to identify which are changed after the creep tests, the information of initial non creep-
deformed microstructure must be attained. Grain size, secondary phases, and misorientation profile
(or fraction of coincident site lattice boundaries) are the three representative items which can
significantly influence the mechanical properties of materials.

In this section, EBSD results for non-creep tested IN625 samples are used to derive the grain size
and misorientation histograms of non-creep tested IN625 samples. The identification of secondary
phases in matrix of those samples were conducted by using a TEM machine equipped with EDS
spectra and SAED modules. Some general observations of microstructures of those samples are
also delivered by TEM and SEM images.

132

4.1.1. Grain Size Evolution of Non-creep Deformed AM IN625 and Wrought IN625 Samples
over the 1 Year LHT
The grain size evolution over the 1 year heat treatment is suggested in Figure 4-1. The total average
grain size was determined by the cubic root of Dav(parallel) × Dav(perpendicular) ×
Dav(perpendicular), where Dav stands for the average grain diameter along the given direction in
Fig. 4-1. Each grain size profile in Fig. 4-1 is a statistic obtained from ≥ 741 grains. The grain size
of AM HIPed IN625 was conserved over the 1 year LHT, in other words, no remarkable change
in grain size was not observed in AM HIPed IN625 from the as-HIPed state to the 1 year LHT
state. In wrought IN625, however, the average grain size increased during the first 6 months cyclic
heat exposure to 650 ˚C. Similar to the case of AM IN625, the increased grain size of 6 months
LHT wrought IN625 was conserved until finishing the 1 year LHT.
The grain size of AM IN625 is slightly larger than wrought IN625 in all thermal conditions in Fig.
4-1, however, this grain size difference is not believed to lead to any significant effects on
mechanical properties.

133

Figure 4-1. Grain size histograms of AM IN625 and wrought IN625 alloys over the 1 year LHT
presented in two different directions of each sample. Building direction is along the z axis in Fig.
2-3.
134

One interesting aspect in the grain size evolution over 1 year is that the grain size of wrought
IN625 increased slightly over 6 months, but no average grain size change in AM HIPed IN625
over the same period. Even though the typical grain growth temperature of wrought IN625 has
been proposed as high temperatures over 900 ˚C [13], Moore et al. observed a grain growth in
IN625 over 3000 hours isothermal heat treatment at 650 ˚C [68].

Moore et al. [68] also argued, the grain boundary precipitation can restrict the grain boundary
migration, hence, the limitation of grain boundary growth in AM HIPed IN625 is probably related
to precipitation of grain boundary particles. This will be discussed later in section 4.3.2.

4.1.2. Grain Boundary Misorientation Distribution of Non-creep Deformed IN625 Samples over
the 1 Year LHT
The grain boundary misorientation characteristic is one determinant factor for high temperature
ductility of metallic materials [153]. Specifically, a low fraction of coincident site lattice (CSL)
boundary or a high fraction of random boundaries (= non-CSL) causes early nucleation or
expedited propagation of cracks upon applied load conditions.
In addition to that, as non-CSL boundaries provide preferential sites for intergranular precipitation
of brittle particles like M23C6 in an Inconel 690 alloy [154], and subsequent grain boundary
embrittlement [4,14,21], the grain boundary misorientation profile has a particular importance, so
that the misorientation profile of non-creep deformed IN625 samples are given in Figure 4-2.
135

Figure 4-2. Grain boundary misorientation histograms of AM IN625 and wrought IN625 alloys
over the 1 year LHT presented in two different directions of each sample. Building direction is
along the z axis in Fig. 2-3.

136

In Fig. 4-2, the general trends in misorientation profile evolution over the 1 year LHT is similar
between AM HIPed IN625 and wrought IN625. Additionally, the misorientation distribution does
not depend on the direction of IN625 samples in Fig. 4-2.
Both alloys have the highest fraction of Σ3 CSL boundary ( ≈ 60˚) and small fraction ( ≤ 0.1) in low
angle boundaries (< 15˚) until the 6 months LHT. However, after the 1 year heat treatment, the
portion of low angle grain boundaries significantly increased by ≈ 0.35 (AM HIPed IN625) and
by ≈ 0.28 (wrought IN625). As it will be discussed in section 5.3.2, some misfit dislocation
generation would be one possibility for this phenomenon.

In general, over the 1 year LHT, no remarkable difference in the fraction of CSL boundaries
between AM HIPed IN625 and wrought IN625 is observed.

4.1.3. Microstructure Observation of Non-creep Deformed AM IN625 and Wrought IN625 by
Using SEM and TEM
SEM and TEM observations for non-creep tested AM IN625 and wrought IN625 were conducted
to investigate any meaningful microstructural characteristic of AM IN625 in comparison with
wrought IN625. The samples were either non-LHT or LHT. The LHT temperature condition was
650 ˚C. Those non-LHT samples were either as-HIPed (AM IN625, 1150 ˚C) or as-solution
annealed (wrought IN625, 1100 ˚C).

Fig. 4-3 gives the SEM images of non-creep deformed AM IN625 and wrought IN625.
137

Figure 4-3. SEM images of non-creep deformed IN625 samples (a) as-HIPed AM IN625, (b) as-
solution treated wrought IN625, (c) 6 months LHT AM HIPed IN625, (d) 6 months LHT wrought
IN625 after solution treatment, (e) 1 year LHT AM HIPed IN625, (f) 1 year LHT wrought IN625
after solution treatment.
138

Figure 4-3 shows the SEM images of non-deformed microstructures of AM IN625 as-HIPed and
wrought IN625 as-solution annealed (Fig. 4-3a and 4-3b), after the 6 months LHT (Fig. 4c and d)
and the 1 year LHT (Fig. 4e and f). White particles in the AM HIPed IN625 matrix of Fig. 4-3a
were either Al2O3 or TiN inclusions. The identification method of secondary particles will be
delivered later in section 4.1.4. Moreover, as shown in Fig. 4-3a, the TiN and Al2O3 were uniformly
distributed in the AM IN625 matrix. However, MC carbides were not present in AM IN625
possibly due to the insufficient carbon content. As shown in Table 2-2, the carbon content of
wrought IN625 is ≈ 5 times higher than that of AM IN625.
In contrast, MC carbides, such as (Nb,Ti)C or NbC, were found in the wrought IN625 sample as
shown in Fig. 4-3b, however, no Al2O3 and TiN were observed in the wrought due to a much lower
oxygen concentration in in the wrought. Additionally, large MC carbides (> a few µm) were still
observed even after the 1 year LHT at 650 ˚C in wrought IN625.
Fig. 4-3c and 4-3d show that grain boundaries of AM HIPed and wrought IN625 were largely
occupied by particles after the 6 months LHT. The acicular phases at grain boundaries in Ni based
superalloys indicate the presence of δ phase [10-14,24]. Fig. 4-3c and 4-3d show that δ phase
precipitated at grain boundaries during the 6 months LHT of both AM and wrought IN625 alloys.
Fig. 4-3e and 4-3f exhibit the non-creep deformed microstructures of AM HIPed and wrought
IN625 samples after the 1 year LHT. When Fig. 4-3e and 4-3f are compared to the 6 months LHT
samples (Fig. 4c and 4d), no remarkable change in microstructure was found under at least through
the SEM observations.

Now, Figure 4-4 presents the TEM data of non-creep tested IN625 samples.
139

Figure 4-4. TEM observations after 6 months and 1 year LHT of AM and wrought IN625 samples
(a) matrix region of 6 months LHT AM IN625, (b) grain boundary region of 6 months AM IN625,
(c) matrix region of 1 year LHT AM IN625, (d) grain boundary region of 1 year LHT AM IN625,
(e) matrix region of 6 months LHT wrought IN625, (f) grain boundary region of 6 months LHT
wrought IN625, (g) matrix region of 1 year LHT wrought IN625, (h) grain boundary region of 1
year LHT wrought IN625 (i) Scanning TEM of 6 months LHT wrought IN625 with the EDS point
spectra of the M6C carbide, (j) Scanning TEM image of 6 months LHT AM IN625 with EDS point
spectra of the M23C6 carbide; (i) and (j) show how M23C6 and M6C were identified using the EDS
analysis [42-44]. The specimens were undeformed.

140

In Fig. 4-4, TEM images of 6 months and 1 year LHT AM IN625 and wrought IN625 are presented
and the sub-micron area of those samples are observable with some secondary phases such as γ’’,
δ, M23C6, M6C and TiN are identified using SAED pattern indexing and EDS spectra.

Fig. 4-4a shows that the matrix of AM HIPed IN625 is filled with γ’’ precipitates after 6 months
LHT. In Fig. 4-4b, the grain boundary of the same sample was decorated with δ, M23C6, M6C and
TiN. The presence of TiN particles simultaneously suggests the presence of Al2O3 as TiN typically
nucleates and grows at the Al2O3 interface during the solidification processes.
Similar to Fig. 4-4a, Fig. 4-4c exhibits large γ’’ precipitation in the grain interior of the 1 year LHT
AM IN625. In addition, some subgrain walls were observed in the 1 year LHT AM IN625 sample.
The grain boundary region of the same sample was occupied by M23C6 and M6C type carbides as
shown in Fig. 4-4d. This is generally consistent with the grain boundary region of 6 months LHT
AM IN625 sample (Fig. 4-4b).
As shown in Fig. 4-4e and 4-4g, γ’’ had formed in the matrix of both 6 months and 1 year LHT
wrought IN625, similar to the LHT AM IN625 samples. The grain boundary region of both the 6
months and 1 year LHT wrought samples were also decorated with M23C6, M6C and δ phases (Fig.
4-4f and 4-4h). Despite the δ absence in Fig. 4-4f, (wrought 6 months LHT) some grain boundary
δ was found in the SEM image (Fig. 4-3d).
Moreover, as shown in Fig. 4-4i, some particles at the grain boundaries of Fig. 4-4f are enriched
with Mo and Si, and it is identified as the M6C carbide. Fig. 4-4j also gives the EDS spectra of
M23C6 carbide (from Fig. 4-4b) which has strong Cr peaks. However, this M6C carbide
precipitation was unexpected in the both LHT AM and wrought specimen at 650 ˚C over 1 year
according to the general TTT diagram of IN625 [14]. Guo et al [155] observed an expedited M6C
141

precipitation in wrought 718 by adding 0.32 wt% Si after the direct aging. The TTT diagram is
based on IN625 alloy of < 0.1 wt% Si content. Perhaps the higher Si content in the IN625 alloys
of this study (≈ 0.2 wt% for both wrought IN625 and AM IN625) can explain the early M6C
precipitation over 6 months and 1 year LHT [14,155].

To sum up, in all thermal conditions suggested in Fig. 4-3 and 4-2, AM IN625 samples have
ceramic inclusions Al2O3 and TiN without MC carbides, whereas those ceramic inclusions are not
found in wrought IN625 samples. However, MC carbides are present in all wrought IN625
samples, which is not observed in AM IN625 samples. Additionally, after both 6 months and 1
year LHT in AM IN625 and wrought IN625, matrix is filled mostly with γ’’ and grain boundaries
are occupied with δ, M23C6 and M6C in both AM IN625 and wrought IN625.

4.1.4. Secondary Phase Identification of As-HIPed AM IN625 and As-solution Annealed
Wrought IN625
As some noticeable quantity of matrix ceramic inclusions such as Al2O3 and TiN in AM IN625
alloys has rarely reported previously [74,87-91], a careful confirmation for the observation of those
ceramic inclusions is required. Hence, the phase identification of Al2O3 and TiN is separately
delivered in this section, rather than simply given concurrently with carbides, γ’’, and δ in the
previous section.

The following Figure 4-5 shows TEM images, SAED pattern analysis, and EDS spectra for those
ceramic inclusions.
142

Figure 4-5. TEM analyses of AM IN625 alloys (a) a TEM sample of AM HIPed IN625, (b) a
TEM sample of AM as-solution annealed IN625, (c) a magnified image of particles in (a), (d)
EDS line profile for the particles in (c), (e) one other magnified image of particles in (b), (f) SAED
pattern of the red circle area in (e) and its key.
143

Fig. 4-5 illustrates the TEM images for AM alloys before creep tests. Fig. 4-5c indicates that two
particles are present in the as-HIPed AM IN625 sample. An EDS line spectra analysis was
conducted along the yellow line in Fig. 4-5c. In Fig 4-5d, EDS line spectra shows that Al2O3 and
TiN inclusions were initially present in the as-HIPed AM IN625 specimen.
One other magnified TEM image of Fig. 4-5b is shown in Fig. 4-5e. A SAED pattern analysis was
conducted for the red circle in Fig. 4-5e and the result is shown in Fig. 4-5f. From the key of SAED
pattern in Fig. 4-5e, the presence of Al2O3 and TiN inclusions is confirmed in the non-HIPed (as-
solution annealed) AM IN625 specimen as well. The ceramic particles tend to be distributed
homogeneously on the matrix of AM IN625 samples, rather than populating at grain boundaries.
TiN particles appear to be present as well and “attached” to the Al 2O3. Those inclusions were
smaller than 1 µm with an average size of 0.14 ± 0.11 µm.

The presence TiN on the surface of Al2O3 proposes that these ceramic inclusions are formed during
the solidification processes such as casting or powder atomization [28]. Several studies have
reported that Al2O3 inclusions are commonly formed during the atomization process due to the
reaction between Al in Ni superalloys and O from the gas environment if the O partial pressure in
the gas is not carefully controlled [50]. In the case of Inconel 718, the minimum composition
requirement at the melting temperature to avoid forming those ceramic inclusions are N < 5ppm
and O < 2ppm [63]. As Table 2-2 indicates, the oxygen content of IN625 powders used in this
study is almost the same as that of the bulk AM IN625. It is determined that the oxide formation
occurred during the powder fabrication process, so that the LPBF fabrication process may not
contribute to the oxide particle formation.

144

The large specific area of powders seems to provide preferable sites to form ceramic inclusions
(Al2O3 and TiN). Gao et al. [156] suggested that the atomized powder size is an important factor
to determine the oxygen content of atomized powders. According to the reference, the oxygen
content of Ni superalloy powders increased from 28 to 71 ppm as the powder size decreased from
147 to 38 µm. However, in the case of wrought IN625, as only the surface of molten liquid is
exposed to air during the casting process, the formation of ceramic inclusions tended to be gathered
on the surface of cast materials [157]. Additionally, this surface part is cut to guarantee the product
quality of Ni superalloys.

Unlike the case of AM IN625, the precipitation of MC carbide in wrought IN625 is considered as
a typical phenomenon upon the solution treatment as illustrated in section 1.1.1.1. Aside from MC
carbides, no other secondary precipitates were found in the as-solution annealed wrought IN625.

The EDS analysis result for MC carbides in the as-solution annealed wrought IN625 is given in
Figure 4-6.
145

Figure 4-6. EDS mapping result for MC carbides in as-solution annealed wrought IN625.

146

4.2. Microstructure of Creep-deformed AM IN625 and Wrought IN625 at 650 ˚C and 800
˚C
The undeformed microstructure of IN625 samples was investigated in the previous section. In this
section, for the comparison between the un-creep and creep deformed samples, the microstructures
of creep deformed specimens are observed. The same instruments used in the previous section,
SEM and TEM were chosen for the microstructure observation of non-creep deformed samples.
Phase identification was also conducted by applying the same methods illustrated in the previous
section. some relative data such as SAED pattern key and EDS spectra data are included.

4.2.1. Microstructural Observation of Creep Deformed Non-LHT SCT AM IN625 and Wrought
IN625 at 650 ˚C and 800 ˚C
Some changes in microstructure of SCT AM IN625 and wrought IN625 are compared in this
section. The creep tests are conducted within 24 hours and 650 ˚C and 800 ˚C were selected for
the creep tests.

As it is shown in Fig. 3-1, several SCT creep testing showed rupture in some shorter times than 24
hours. Thus, in order to keep the 24 hours for microstructure change, the target specimens for
microstructure observation did not fail within 24 hours in those creep tests.

The SEM observation results of microstructure revealing of SCT AM IN625 and wrought IN625
(at both 650 ˚C and 800 ˚C) are shown in Figure 4-7.
147

Figure 4-7. SEM images of SCT AM IN625 and wrought IN625 (a) AM HIPed IN625 at 650 ˚C
and 416 MPa, (b) AM HIPed IN625 at 800 ˚C and 66 MPa, (c) wrought IN625 at 650 ˚C and 416
MPa, (d) wrought IN625 at 800 ˚C and 66 MPa.

148

In Fig. 4-7a, some continuous particle precipitation at a grain boundary was observed. Similarly,
this grain boundary decoration by secondary phases is also evident in Fig. 4-7b as well. Thus, over
the 24 hours creep tests at both 650 ˚C and 800 ˚C, grain boundary precipitation occurred in the
AM IN625 samples. However, one difference between Fig. 4-7a and 4-7b is the acicular phase
which grows towards the grain interior direction from grain boundaries. This acicular phase is δ
phase [10-14,24]. Additionally, while only M23C6 precipitation occurred at grain boundaries of
SCT AM IN625 at 650 ˚C, two types of carbides (M23C6 and M6C) are precipitated at grain
boundaries of SCT AM IN625 at 800 ˚C.
However, in wrought IN625, as Fig. 4-7c presents, no remarkable microstructural evolution was
observed after the creep test at 650 ˚C in comparison with Fig. 4-3b (non-creep deformed wrought
IN625). In the case of 800 ˚C creep tested wrought IN625 (Fig. 4-7b), similar to the microstructure
in Fig. 4-7b, some large grain boundary precipitation of M23C6 and M6C carbides are found. In
other words, the continuous grain boundary precipitation occurred only in the 800 ˚C SCT wrought
IN625, not in the 650 ˚C wrought IN625.

The evidence of phase identification for those suggested secondary phases in Fig. 4-7 is followed
with Figure 4-8 and Figure 4.9. Fig. 4-8 is a set of TEM observation data of the same samples in
Fig. 4-7a and 4-7c which are those 650 ˚C SCT AM IN625 and wrought IN625, respectively. Fig.
4-9 displays the TEM results for those creep tested SCT AM IN625 and wrought IN625 at 800 ˚C.

149

Figure 4-8. TEM observations of AM IN625 and wrought IN625 samples after SCT at 650 ˚C,
416 MPa (a) matrix area of AM HIPed IN625, (b) grain boundary area of AM HIPed IN625, (c)
matrix area of wrought IN625, (d) grain boundary area of wrought IN625.

150

Figure 4-9. TEM observations of AM IN625 and wrought IN625 samples after SCT at 800 ˚C,
104 MPa (a) matrix region of AM HIPed IN625, (b) grain boundary area of AM HIPed IN625, (c)
matrix region of wrought IN625, (d) grain boundary area of wrought IN625.
151

According to the TTT diagram of IN625 in Fig. 1-2, it was expected to observe only additional γ’’
precipitation in IN625 after 24 hours at 650 ˚C. However, this expectation was not correct at least
for the AM IN625 specimen. As shown in Fig. 4-8a, in accordance with the expectation from Fig.
1-2, γ’’ is precipitated in the matrix of 650 ˚C SCT AM IN625. An unexpected M23C6 carbide
formation is also found at the grain boundary as shown in Fig. 4-8b, which suggests those grain
boundary particles in Fig. 4-7a are M23C6 carbides.
In Fig. 4-8c, γ’’ precipitation is also found in the wrought IN625 after 24 hours 650 ˚C creep test.
However, as suggested in Fig. 4-7c, M23C6 carbides are not formed yet, instead, some large (> 100
nm) MC carbides are decorating grain boundaries of SCT wrought IN625 at 800 ˚C.

Additionally, when Fig. 4-8a and 4-8c are compared, the amount of γ’’ was much higher in AM
IN625 than wrought IN625 after the 24 hours creep test at 650 ˚C. No MC carbides are found in
the 650 ˚C creep tested AM IN625 even after 24 hours from the initial as-HIPed state.

The TTT diagram in Fig. 1-2 predicts M23C6 and δ phase formations after 24 hours heat exposure
at 800 ˚C. Similar to the result of γ’’ precipitation in Fig. 4-8a, δ phase was largely formed in the
AM HIPed IN625. The windowed SAED pattern analysis in Fig. 4-9a evinces those acicular
phases in the 800 ˚C SCT AM IN625 are δ phases. This acicular orthohombic Ni3Nb phase is
generally nucleated on grain boundaries and has a smaller strengthening effect than γ’’ phase
[11,13]. Additionally, as shown in Fig. 4-9a, those acicular δ phases seemed to contribute to
dislocation pinning, which implies some material strengthening. Suave et al. [14] suggested the
increase in hardness of IN625 alloy due to the precipitation of δ phase, and it may support the idea
of dispersion hardening effect of δ phase. Specifically, δ acts as a hindrance of dislocation motion
152

during plastic deformation process. The accumulation of dislocations at the interface between δ
and Ni matrix can cause an additional strengthening effect [60]. Despite of its preference to
nucleate on grain boundaries, δ phases are formed uniformly in the deformed matrix of AM IN625
during the 24 hours creep tests at 800 ˚C as presented in Figure 4-10a.
δ phase was also found at grain boundaries of the 800 ˚C SCT AM IN625 as shown in Fig. 8b.
Except for the δ phase, other grain boundary particles were identified as either M23C6 or M6C
carbides using TEM and SEM EDS analyses as it is shown in Fig. 4-10. In this EDS analysis,
based on the phase information in Table 1-1, Cr, Mo rich particles were identified as M23C6
carbides, whereas Mo, Si rich particles were identified as M6C carbides.

Interestingly, similar to the case of M23C6 carbide in the 650 ˚C creep data, only the AM IN625
includes δ phase in its matrix after the 24 hours creep tests at 800 ˚C. In other words, no δ phase
is precipitated in the wrought IN625 alloy under the same creep testing condition. However, two
carbides, M23C6 and M6C, were found at grain boundaries of the crept wrought IN625 specimen.
The identification of carbides in the samples was conducted using EDS analysis (Fig. 4-10b) as
well. In Fig. 4-9c, dislocation pinning by some matrix MC carbides occurs, which appears
comparable with δ phase strengthening in Fig. 8a. This suggests that the volume fraction of δ phase
in Fig. 4-9a and that of MC carbides in Fig. 4-9c need to be compared to understand the extra creep
strength of AM IN625 presented in Table 3-2.

Fig. 4-10a and 4-10b explains how those secondary phases in Fig. 4-9 are identified using EDS
analysis.
153

Figure 4-10. STEM analysis of IN625 alloys after the 24 hours creep test at 800˚C, 104 MPa with
EDS element maps (a) AM HIPed IN625 sample (b) wrought IN625 sample, the red squares
indicate the mapping area where the EDS mapping was conducted.

154

4.2.2. Microstructural Observations of Creep Deformed LHT AM IN625 and Wrought IN625 at
650 ˚C and 800 ˚C
In this part, creep deformed microstructure of long-term heat-treated (LHT) IN625 samples are
studied. Two long-term heat treatments are considered: 6 months and 1 year. AM-HIPed IN625
and wrought IN625 alloys analyzed in this portion.

The result and discussion on microstructural observations of both 6 months and 1 year LHT and
creep tested at 650 ˚C and 800 ˚C samples will be discussed. SEM and TEM are mainly utilized.

One difference from the specimens in section 4.2.1 is that the observed samples in this section
were ruptured. In other words, most specimens were creep deformed less than 24 hours as shown
in Table 3-1.

4.2.2.1.Microstructural Observations of Creep Deformed 6 Months LHT AM IN625 and Wrought
IN625 at 650 ˚C and 800 ˚C
This section delivers the observation results for 6 months LHT AM IN625 and wrought IN625
after creep tests at 650 ˚C and 800 ˚C. The stress conditions for the two temperatures were 658
MPa at 650 ˚C and 192 MPa at 800 ˚C.
Figure. 4-11 suggests the SEM observation results for those creep tested IN625 samples.
155

Figure 4-11. SEM images of 6 months LHT AM IN625 and wrought IN625 after creep tests (a)
AM HIPed IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c)
wrought IN625 at 650 ˚C and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa; the
magnification of (a) is equal to (c) and, the magnification of (b) is equal to (d). Grain boundaries
in all samples are occupied with δ, M23C6 and M6C. Matrix δ is evident in those 800 ˚C creep
tested samples.

156

In general, in comparison with Fig. 4-3b and 4-3c, no additional secondary phase formation is
observed in Fig. 4-11 as the consequence of high temperature creep deformation. In all images in
Fig. 4-11, δ, M23C6 and M6C carbides are mostly located at grain boundaries. γ’’ phase is
embedded typically in the intergranular region. Some intragranular δ phases are observed in Fig.
4-11a, 4-11b and 4-11d.

One remarkable difference in Fig. 4-11a from the undeformed microstructure (Fig. 4-3c) is the
intragranular δ transformation of matrix γ’’ occurred solely in the 6 months LHT AM IN625 alloy.
In contrast, no δ transformation was observed in Fig. 4-11c, which suggests the intragranular δ
transformation occurred in the LHT AM IN625 sample at a high stress. Xing et al. [158] suggested
that a stress induced transformation from γ’’ to δ occurs in IN625 by shearing γ’’ and the
subsequent coarsening of the precipitate. This mechanism can accelerate the δ phase
transformation when compared to the isothermal treatment condition without any stress on the
sample. Specifically, stacking fault formation inside γ’’ due to the particle shearing acts like a
nucleus of δ phase, and this δ nucleus develops along three basic close-packed direction of the bct
γ’’ phase unit cell [158].
One derivable question at this point is why the intragranular δ transformation is not exhibited in
the 6 months LHT wrought IN625 as shown in Fig. 4-11c. One possible explanation of this result
is the size of γ’’. As the growth kinetics of γ’’ and the subsequent δ phase is dependent on the
initial γ’’ size before the thermal process [158], which is, in this case, the creep test at 650 ˚C and
658 MPa. One other possibility comes from the γ’’ fraction. As the γ’’ to δ phase transformation
is a solid phase transformation, not a simple resolution and subsequent precipitation process, this
can be interpreted as a process of connection of γ’’ phase along their close packed orientation. This
157

may suggest the fraction of initial γ’’ may have an important effect on the transformation kinetics.
Therefore, the fraction of γ’’ in the 6 months LHT wrought IN625 is presumably not sufficient for
the δ phase, whereas the γ’’ fraction in the 6 months LHT AM IN625 is sufficient to initiate the δ
transformation at 650 ˚C. As it will be shown in the later section 4.3.1, both particle size and
fraction of γ’’ were larger in AM IN625 than wrought IN625 after 6 months and 1 year LHTs.

The SEM images of 800 ˚C creep tested 6 months LHT IN625 alloys have a clear difference in
their microstructure as displayed in Fig. 4-11b and 4-11d, specifically for the δ size. While most
of γ’’ seems to be transformed into acicular and small δ phases in the wrought IN625 (Fig. 4-11d),
very long (> 10 µm) δ phases are present in the AM IN625 (Fig. 4-11b). This result seems to be
due to the longer creep time in the wrought IN625 than the time in the AM IN625.
One paper [159] reported the partial dissolution and breakage of δ phase in Inconel 718 under
prolonged isothermal exposure to some elevated temperatures above 900 ˚C. Specifically, the δ
phase coarsening occurred first and subsequent stress localization at δ interfaces causes partial
dissolution of delta phases under the high temperature deformation. Next, this partial dissolution
of δ phase breaks the coarsened δ phase and forms several small δ phases. Hence, according to this
explanation, the large δ phases in Fig. 4-11b is considered as one earlier microstructure before to
be the final microstructure in Fig. 4-11d.

For further analysis, TEM and STEM were conducted as shown in the following figures:

158

Figure 4-12. TEM of 6 months LHT IN625 samples after creep tests at 650 ˚C, 658 MPa (a) matrix
region of AM IN625 (b) intergranular region of AM IN625 (c) matrix region of wrought IN625
(d) intergranular region of wrought IN625. The microstructure is generally same between AM
IN625 and wrought IN625 except for the volume fraction of γ’’.

159

Figure 4-13. TEM of 6 months LHT IN625 samples after creep tests at 800 ˚C, 192 MPa (a) matrix
region of AM IN625 (b) intergranular region of AM IN625 (c) matrix region of wrought IN625
(d) intergranular region of wrought IN625. Large δ phase formation was observed in both AM
IN625 and wrought IN625. Only wrought IN625 shows the sub-grain boundary formations.

160

Figure 4-14. STEM analysis of 6 months LHT IN625 alloys after the creep tests at 650˚C, 658
MPa with EDS element maps (a) AM HIPed IN625 sample (b) wrought IN625 sample, the red
squares indicate the mapping area where the EDS mapping was conducted.

161

Figure 4-15. STEM analysis of 6 months LHT IN625 alloys after creep tests at 800˚C, 192 MPa
with EDS element maps (a) AM HIPed IN625 sample (b) wrought IN625 sample, the red squares
indicate the mapping area where the EDS mapping was conducted.

162

One remarkable observation in Figure 4-12 is the presence of matrix nano-twins in both AM IN625
and wrought IN625 alloys (Fig. 4-12b and 4-12d). As the formation of matrix nano-twins requires
high internal friction around coherent matrix γ’’ particles [33], the γ’’ populated matrix in Fig. 4-
12a and 4-12c may be favorable for mechanical twinning at 650 ˚C, 658 MPa for both 6 months
LHT AM IN625 and wrought IN625. It has been reported that mechanical nano-twinning is
generally observed in γ’’ bearing Inconel 718 [21] and γ’ strengthened high Al containing Ni
superalloys [160,161], at intermediate temperatures between 400 ˚C – 800 ˚C under substantial
stress (> 500 MPa).
More specifically, in Ni-based superalloys, it has been reported that shearing of precipitates,
mainly γ’ and γ’’, occurred by a/2<110> matrix dislocation and it produces anti-phase boundaries
or complex stacking faults in the sheared precipitates. The formation of such defects in the particle
nucleates derivative partial dislocations like a/6<112> Shockley partials on adjacent {111} matrix
planes. Those generated Shockley partial dislocations then travel on (111) matrix planes, which
form matrix twins in Ni-based superalloys [160]. Thus, the experimental condition is considered
enough to nucleate the matrix nano-twins between secondary phases and Ni matrix.
It is deduced that the evolution of twinning may have a role in the ductility improvement. However,
the relationship between increase in twinning density and strain has not yet been established
although the twinning density increased with strain [162].

Figure 4-14 suggests M23C6 and M6C carbides are present on grain boundaries of the 6 months
LHT AM IN625 and wrought IN625 alloys. The TiN ceramic inclusion is still observed in the
matrix of AM IN625 alloy, according to Figure 4-14a, and the Ni-Nb rich globular δ is present on
the grain boundaries as well. The plate-like δ phase has a crystallographic orientation relationship
163

with Ni matrix such as {111}γ//(010)δ, <11
�
0>γ//[110]δ [163]. The distortion of acicular δ phases
leads to the degradation of this crystallographic relationship between δ and the matrix, which
provides a driving force for the dissolution of δ phase. By the consequence of distortion of plate-
like δ phases, the spheroidization of δ phase occurred due to some dissolved breakages of the
deformed plate-like δ phases [1 64]. Therefore, the presence of globular δ phase suggests that
considerable dislocation accumulation occurred for the δ spheroidization at the grain boundary
during the 650 ˚C creep of 6 months LHT AM IN625.
However, as this δ spheroidization is not found in the 650 ˚C creep tested 6 months LHT wrought
IN625 (Fig. 4-12d and 4-14b), the grain boundary δ spheroidization may be beneficial to secure
more high temperature ductility of IN625 alloys. As previously shown, the creep ductility of
wrought IN625 is much higher than AM IN625 for all creep testing conditions.
In Fig. 4-13a and 4-13c, the δ phase is now largely precipitated in matrix after 800 ˚C creep tests
of the 6 months LHT samples. As shown in Fig. 4-13b, some remaining γ’’ precipitates were still
observed in the 6 months LHT AM IN625 after creep tests at 800 ˚C and 192 MPa. Table 3-1
suggests that ≈ 11 hours were taken in this creep testing condition, so that the γ’’ to δ
transformation had not been completed within 11 hours in the AM IN625 sample under the 800
˚C creep test. However, as shown in Fig. 4-13c, it was determined that most of the γ’’ phase had
been transformed into δ within ≈ 43 hours 800 ˚C for creep tests of 6 months LHT wrought IN625.
As wrought IN625 obtains ≈ 15% additional room temperature hardness with δ precipitation, but
≈ 42% extra hardness with γ’’ bearing matrix [106], the progress in δ transformation seems related
to the lower creep strength of wrought IN625 than AM IN625 after the 6 months LHT and
following 800 ˚C creep test.
164

Another difference is the morphology of δ phase at grain boundaries. Fig. 4-13b clearly shows that
acicular type δ phase is primarily decorating grain boundaries of the AM HIPed sample. However,
the wrought sample (Fig. 6h) evinces the presence of some spherical δ phase with subgrain
formation around grain boundaries. The spherical morphology of δ phase is more clearly observed
in Fig. 4-15b. Gan et al. [159] observed the breakage of lamellar δ with matrix dynamic
recrystallization near δ phases and subsequent δ spheroidization in a 968 ˚C compression test of
Inconel 718 at 0.5 strain. Their finding suggests that dislocation accumulation on the long acicular
δ phase caused partial dissolution of δ phase due to the stress localization by the accumulated
dislocations and this dislocation will nucleates recrystallization on the δ interface. That observation
by Gan et al is somewhat similar to Fig. 4-13c and 4-13d, exhibiting subgrain formation around δ
phases. Therefore, the shorter δ phases present in Fig. 4-11d (wrought IN625) than AM IN625
sample of Fig. 4-11b may be due to the lower strain of AM IN625 in the 800 ˚C creep testing.

Figure 4-13c and 4-13d suggests some dynamic recovery or subgrain boundary formation occurred
in the 800 ˚C creep tested LHT wrought IN625 alloy. This feature was not found in the 650 ˚C
creep tested LHT wrought IN625 sample as presented in Fig. 4-12c and 4-12d. Hence, the pure
effect by plastic strain appears not a proper explanation on the formation of subgrain boundaries
in the creep deformed 6 months LHT wrought IN625. The plastic strain of 6 months wrought
IN625 exceeded 0.5 at the two experimental temperatures. Dynamic softening mechanisms such
as recovery and recrystallization were activated in a Ni-based superalloy over 750 ˚C [165,166].
This effectively explains why the subgrain structure is only observed in the 800 ˚C creep testing
of 6 months LHT wrought IN625, but not in the 650 ˚C creep test.

165

4.2.2.2.Microstructural Observation of Creep Deformed 1 Year LHT AM IN625 and Wrought
IN625 Samples
In section 4.1.3, it is proposed that the microstructure of non-creep deformed 1 year LHT AM
IN625 and wrought IN625 have almost the same as the microstructure of non-creep deformed 6
months LHT AM IN625 and wrought IN625 samples. Similarly, the observed microstructure of
the creep deformed IN625 alloys do not show any noticeable difference from the observation
results in the previous section 4.2.2.1.
The subgrain boundary formation in the 1 year LHT AM IN625 (non-deformed) is one interesting
microstructural difference from the microstructure of the 6 months LHT and non-creep tested AM
IN625 as illustrated in section 4.1.3. However, when only the AM IN625 samples are concerned,
as present in Table 3-1, there was no remarkable differences in mechanical properties or creep
strength and ductility between the two LHT AM IN625 samples.

The following figures show the observation results for the creep tested 1 year LHT IN625 alloys:
166

Figure 4-16. SEM images of 1 year LHT AM IN625 and wrought IN625 after creep tests (a) AM
HIPed IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c) wrought
IN625 at 650 ˚C and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa; the magnification is the
same for all images in this figure. Grain boundaries in all samples are occupied with δ, M23C6
and M6C. Matrix δ is evident in those 800 ˚C creep tested samples.
167

Figure 4-17. TEM of 1 year LHT AM IN625 and wrought IN625 after creep testing (a) AM HIPed
IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c) wrought IN625
at 650 ˚C and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa.
168

Figure 4-18. STEM results for 1 year LHT IN625 after creep testing with EDS spectra data (a)
AM HIPed IN625 at 650 ˚C and 658 MPa, (b) AM HIPed IN625 at 800 ˚C and 192 MPa, (c)
wrought IN625 at 650 ˚C and 658 MPa, (d) wrought IN625 at 800 ˚C and 66 MPa. The EDS
analysis showed the presence of M23C6 and M6C carbides on grain boundary area in those
suggested creep tested samples of AM IN625 and wrought IN625.
169

When only focused on the type of precipitates, not their amounts, the results in Figure 4-16 and 4-
17 are generally coincident with those results in Fig. 4-12 and 4-13 (creep-tested 6 months LHT
AM IN625 and wrought IN625).
Similar to the result in Fig. 4-12c, in Fig. 4-16c, the matrix δ particles was not observed in the 1
year LHT wrought IN625 sample even after the 1 year cyclic heat treatment at 650 ˚C, followed
by a 650 ˚C 658 MPa creep test. However, as shown in Fig. 4-16a, in the case of 1 year LHT AM
IN625, after the 650 ˚C creep test, some matrix δ phases are found as like the case in the 6 months
LHT AM IN625 sample (Fig. 4-12a). As it will be shown in the later section 4.3.2, both γ’’ size
and volume fraction was higher in the 1 year LHT AM IN625 than the 1 year LHT wrought IN625
at 650 ˚C. This presumably explains why the matrix γ’’ → δ transformation did not occur at 650
˚C in the 1 year LHT wrought IN625.
Additionally, the matrix nano-twins are still formed in the matrix of both 1 year LHT AM IN625
and wrought IN625 after creep test at 650 ˚C, 658 MPa as shown in Fig. 4-17a and 4-17c.
In the 800 ˚C creep data, δ transformation is observed in the two 1 year LHT IN625 alloys as
displayed in Fig. 4-17b and 4-17d. Subgrain formation near grain boundaries is evident in the 1
year LHT wrought IN625 after 800 ˚C, 192 MPa creep test, which is another common point with
the 6 months LHT wrought IN625 result in Fig. 4-13d.

However, the size of matrix δ phase in Fig. 4-16b is not comparable with the matrix δ size in the
6 months LHT and creep deformed sample in the same condition (Fig. 4-12b). As suggsted in
Table 3-1, the total creep testing time of the 1 year LHT AM LHT IN625 was roughly a half of
that of the 6 months LHT IN625 for the 650 ˚C, 658 MPa creep test. The shorter thermal exposure
170

time may be not sufficient to lengthen those δ phases in the 1 year LHT AM IN625 sample during
the 800 ˚C heat exposure condition.
Moreover, in Fig. 4-17d, some remaining γ’’ phases are found in the matrix of 800 ˚C creep-
deformed 1 year LHT wrought IN625. It suggests the δ tranformation had not been finished in the
wrought IN625 specimen during 800 ˚C and 192 MPa creep test, which is different from the result
of Fig. 4-13d (6 months LHT wrought IN625) indicating the completion of δ transformation during
the same creep test condition. This discrepency between the two creep-tested LHT wrought
samples are attributed to a shorter rupture time (17.7 hours) of the 1 year LHT wrought IN625
when compared to the 6 months LHT wrought IN625 (43.3 hours) under the same creep testing
condition. In other words, the sample in Fig. 4-13d had sufficient time for the δ transformation,
while the creep tesing time appeared insufficient for the δ transformation to the sample in Fig. 4-
17d.
Now, the following Table 4-1 provides the summary of phase and microstructure observations that
have been reported in sections 4.1 and 4.2.

171

Table 4-1. Summary of phase analysis, new phases and microstructures formed after creep tests
were shaded in grey. The types of secondary phases precipitated before and after creep tests are
mostly the same in those 6 months and 1 year long-term heat treated wrought IN625 and AM
IN625, except for one exception of the remaining γ’’ in the 800 ˚C creep tested 1 year LHT wrought
IN625 (marked in yellow).

as-HIPed or as-solution annealed
(creep tested within 24 hours)

6 months and 1 year long-term heat treated
Non-creep
deformed
After creep
deformed at
650 ˚C
After creep
deformed at
800 ˚C
Non-creep
deformed
After creep
deformed at
650 ˚C
After creep
deformed at
800 ˚C
AM
IN625
Al
2
O
3

TiN
Al
2
O
3

TiN
γ’’
M
23
C
6

Al
2
O
3

TiN
δ
M
23
C
6

M
6
C
Al
2
O
3

TiN
γ’’
δ
M
23
C
6

M
6
C
Al
2
O
3

TiN
γ’’
δ
M
23
C
6

M
6
C
nano-twins
Al
2
O
3

TiN
γ’’
δ
M
23
C
6

M
6
C
Wrought
IN625
MC
MC
γ’’
MC
M
23
C
6

M
6
C
MC
γ’’
δ
M
23
C
6

M
6
C
MC
γ’’
δ
M
23
C
6

M
6
C
nano-twins
MC
γ’’
δ
M
23
C
6

M
6
C
Sub-grains
Phase identification was conducted by using EDS-spectra analysis (SEM, TEM) and SAED pattern
analysis (TEM) as it has been shown from Figs. 4-5 to 4-18.

172

4.3. Precipitation Behavior in Both AM IN625 and Wrought IN625 Before/After Creep
Tests
As discussed previously, the secondary phases can influence the mechanical properties of IN625
alloys in this study. For example, large matrix precipitation of γ’’ (at 650 ˚C) or δ (at 800 ˚C) can
enhance the creep strength of both AM IN625 and wrought IN625. On the other hand, some
unfavorably large intergranular precipitation of δ, M23C6 and M6C can degrade the high
temperature ductility of experimental alloy by providing preferential sites for crack nucleation and
propagation.
According to those perspectives, it is considered important to quantify the amount of precipitates
in the matrix and at grain boundaries to interpret the mechanical creep properties of AM IN625.
In other words, as shown in Table 3-1, AM IN625 exhibits the improved high temperature creep
strength as compared to wrought IN625, whereas AM IN625 is embrittled in all creep testing
conditions in this study. Some careful investigation on the precipitation behavior of both AM
IN625 and wrought IN625 before/after creep tests is anticipated to provide an explanation of the
mechanical properties of AM IN625.

4.3.1. Matrix γ’’ and δ Volume Fraction Evolution and γ’’ Coarsening Observation
As discussed in section 1.1.1.3, the volume fraction of matrix secondary phases such as γ’’ and δ
is one determinant factor for the precipitate hardening effect and subsequent extra creep strength
of AM IN625. Figure 4-19 gives the volume fraction evolution of those particles.
173

Figure 4-19. Matrix secondary phase evolution in AM IN625 and wrought IN625 (a) volume
fraction development of γ’’ and δ under various thermal conditions (b) the coarsening behavior of
γ’’ at 650 ˚C under various conditions with the Nb content change in matrix; C.T. = creep tested.
The volume fraction of secondary phases in AM IN625 is generally larger than wrought IN625 in
all thermomechanical conditions. The coarsening of γ’’ is also faster in AM IN625 at 650 ˚C.
174

When only the as-HIPed or as-solution annealed states are compared, Fig. 4-19a suggests that the
initial secondary particle fraction of MC carbides in wrought IN625 (≈ 1.2 %) was slightly higher
than the volume fraction of ceramic inclusions in AM IN625 (≈ 0.4 %).
As shown in Figs. 4-4 and 4-8, for AM IN625 and wrought IN625 samples, the γ’’ is dominantly
present at 650 ˚C, whereas δ largely precipitates in the grain interior region over at least 11 hours
thermal exposure at 800 ˚C. In Fig. 4-19a, the volume fraction of phases, such as γ’’ and δ, in AM
IN625 and wrought IN625 under different creep testing conditions are given. The volume fraction
of γ’’ gradually increased in both AM IN625 and wrought IN625 from SCT to 1 year LHT in 650
˚C creep tests. Moreover, the amount of γ’’ is higher in AM IN625 than wrought IN625 in all 650
˚C creep tests of Fig. 4 -19a. As γ’’ coarsening precedes δ precipitation under long-term heat
exposure at 650 ˚C [15], this expedited γ’’ precipitation may partially explain the ductility drop of
AM IN625 after the LHT. Köhler [167] found some ductility loss of high Nb containing wrought
IN625 (> 3.6%) in the temperature range from 600 ˚C to 700 ˚C due to the fast precipitation of
Ni3Nb phases at grain boundaries (mostly δ). However, the SEM image of non-creep deformed 6
months LHT AM IN625 at 650 ˚C (Fig. 4-3c) does not show a noticeably larger δ precipitation at
grain boundaries when it compared to the wrought IN625 of the same thermal history (Fig. 4-3d).
Even for 800 ˚C creep tests, AM IN625 still exhibits higher δ volume fraction than wrought IN625
in both SCT and the creep-deformed after 6 months LHT.

The γ’’ size coarsening trend and the matrix Nb content variation with respect to the heat treatment
time are exhibited in Fig. 4-19b. As shown in Fig. 4-19b, the γ’’ size of AM IN625 was smaller
than wrought IN625 after the short term creep test at 650 ˚C, however, γ’’ coarsening in AM IN625
occurred more rapidly over the 1 year heat treatment time at 650 ˚C. Additionally, in both IN625
175

alloys, decrease in matrix Nb content was observed over the LHT, which occurred concurrently
with the coarsening of γ’’. The matrix Nb content was measured by using the EDS point spectra
method in a SEM machine and ten EDS point data were collected to obtain one datapoint in Fig.
4-19b. Corresponding to the γ’’ coarsening rate, the amount of Nb content drop in AM IN625 from
as-HIPed state to the 1 year LHT state is larger than the Nb wt% drop in wrought IN625 over the
1 year LHT.
As previously discussed, the absence of γ’’ → δ transformation in 650 ˚C creep tests for both 6
months and 1 year LHT wrought IN625 is probably attributed to the insufficient γ’’ coarsening in
comparison of those LHT AM IN625 samples. Fig. 4-19b suggests that the γ’’ average size of the
1 year LHT wrought IN625 is still smaller than that of the 6 months LHT AM IN625.

4.3.2. Intergranular Precipitation of Secondary Precipitates
The large precipitation in the matrix generally benefits the strength of a material. In contrast, an
excessive intergranular precipitation of some hard particles such as intermetallic and carbides can
degrade the elevated temperature ductility of Ni-based superalloys [4,14].
Additionally, as discussed previously in section 4.1.1, the fast-intergranular precipitation can
inhibit the grain growth in IN625 at high temperatures.
Therefore, the determination of grain boundary precipitation behavior of AM IN625 and wrought
IN625 is essentially important to interpret the poor ductility of AM IN625 in this study.

Now, the grain boundary occupation frequency data plot is illustrated in Figure 4-20.
176

Figure 4-20. Grain boundary particle frequency content evolution in AM IN625 and wrought
IN625 under various testing conditions.

177

The grain boundary particles were mostly formed on random grain boundaries (non-CSL
boundaries) as shown in Fig. 2-14.

At 650 ˚C, ≈ 30 % of random grain boundary area in AM IN625 were occupied by precipitates
within 24 hours of creep testing, whereas negligible grain boundary precipitation occurred in the
wrought IN625 at the 24 hours thermomechanical condition. This is probably due to the absence
of primary grain boundary MC carbides in as-HIPed AM IN625. Typically, the grain boundary
precipitation of secondary phases drops the interfacial energy of grain boundaries [168]. Therefore,
as-HIPed AM IN625 may have higher grain boundary energy than as-solution annealed wrought
IN625, which may activate the faster intergranular M23C6 precipitation in the AM IN625 at 650˚C.
Additionally, this rapid M23C6 precipitation at 650 ˚C appears to explain partially the grain size
conservation of AM IN625 over the 1 year LHT as shown in Fig. 4-1. As Moore et al. [152] argued,
the grain boundary precipitation can restrict the grain boundary migration, hence the limitation of
grain boundary growth in AM IN625 due to the early precipitation of grain boundary M23C6
carbides within 24 hours.

However, the LHT wrought IN625 showed comparable precipitation quantities at grain boundaries
with the LHT AM IN625 after 6 months. Therefore, the 650 ˚C creep ductility drop of 6 months
LHT AM IN625 is difficult to explain by pure grain boundary particle effects; the wrought IN625
should have had much less precipitates if particles were primarily responsible for the drop in
ductility. At 800 ˚C, the grain boundary precipitation amount of both AM IN625 and wrought
IN625 were almost identical not only after 24 hours short term creep, but also those following the
178

creep tests after both 6 months and 1 year LHT. Hence, at 800 ˚C, the grain boundary particle
populations also cannot explain the high temperature embrittlement of AM IN625.

4.3.3. Discussions on the Expedited Precipitation Behavior in AM IN625
As discussed so far, the early precipitation behavior of AM IN625 significantly contributed to
several material properties of AM IN625, such as mechanical strengthening with γ’ and δ
precipitation and the inhibition of grain growth due to the faster intergranular M23C6 precipitation.
This expedited precipitation of the AM IN625 material has been reported in other studies
[88,90,91,109]. In most cases, this fast precipitation behavior was observed in non-fully
homogenized (or non-HIPed) AM IN625 alloys [88,90,91,109].

Anderson et al. [169] proposed three determinant factors for precipitation of secondary phases in
Ni-base superalloys: grain size, dislocation density, and solute micro-segregation in the matrix.
Specifically, the smaller grain size is, the faster precipitation is observed.
As dislocations are highly dense, the highly strained region by dense dislocations can provide a
favorable site for nucleation.
Lastly, as secondary phases typically have completely different chemical composition from the
matrix, the clustering of solutes, which are the main component of the phases, is required to
nucleate secondary phases. According to Anderson et al. [169], among those three possibilities,
the most important factor to form Ni3Nb particles, such as γ’’ or δ, is the micro-segregation level
of Nb for Ni-based superalloys.
179

As shown previously, the grain size of wrought IN625 is smaller than AM IN625 before creep
tests over the entire LHT history, and the initial dislocation density of wrought IN625 was higher
than AM IN625 as well. Thus, it seems that the micro-segregation level of Nb may explain the
expedited precipitation behavior in AM IN625. However, Zhang et al. [109] found that Nb
segregation level in AM IN625 decreased from 4.41 wt% to 4.00 wt% with dissolution of pre-
dendritic structure after a heat treatment at 1100 ˚C for 1 hour. As the HIP treatment of this study
was conducted at 1175 ˚C for 3 hours, this HIP treatment should be sufficient to dissolve fully the
Nb content into the matrix.
However, as the conclusion of Zhang et al. [109] was derived from EDS observations, it casts
doubt as to whether this result can represent the complete removal of Nb micro-segregation due to
the limited resolution of the EDS analysis in the SEM [124]. Based on a computational calculation,
Connétable et al. [170] revealed that Nb in Ni matrix prefers to form a cluster, forming of at least
3 Nb atoms, rather than existing individually in the matrix. In addition to that, the micro-
segregation level of Nb in a highly alloyed Ni base superalloy decreased continuously up to 50
hours at 1160 ˚C, and the temperature is close to the HIP temperature of this study [171]. Thus,
the higher segregation level of AM IN625 material may be not completely relieved by the HIP
treatment at 1175 ˚C over 3 hours.

According to Table 2-2, the Nb content in the bulk AM IN625 was measured to be 4.15 wt%, and
this value is relatively larger than 3.54 wt% in the bulk wrought IN625. This bulk composition gap
between the two alloys is slightly different from the EDS result in Fig. 4-19b, which shows the Nb
content in AM IN625 is higher than that of wrought IN625 by ~ 0.9 wt%. This suggests that AM
IN625 contains more Nb than wrought IN625. Thus, despite the possibility of full dissolution of
180

Nb in the matrix of as-HIPed AM IN625, as the precipitation of Ni3Nb is highly influenced by
both the diffusivity and concentration of Nb in Ni matrix [169], the higher matrix Nb concentration
in AM IN625 may partially explain the early precipitation behavior. According to Mu et al. [171],
≈ 27 % of δ was obtained in an IN625 alloy of 3.8 Nb wt%, whilst only ≈ 17 % of δ was observed
in 3.15 Nb wt% IN625 over 1000 hours aging at 700 ˚C. This δ volume fraction difference was
primarily due to 0.65 wt% Nb content gap, which may support the argument that the early
precipitation of Ni3Nb phases in the AM HIPed IN625 is possibly explained by a higher Nb level.

In summary, the early precipitation of secondary phases in this study is mainly attributed to the
higher matrix Nb content in the AM IN625 than the wrought IN625 sample. The incomplete
homogenization of Nb micro-segregation is proposed as one possibility. More careful investigation
on this phenomenon needs to be continued further.
181

5. Part Ⅲ. HIGH TEMPERATURE EMBRITTLEMENTS OF AM IN625
This work confirmed a dramatic drop in ductility for elevated temperatures that includes our testing
temperatures, which generally agrees with the ductility versus temperature plot of AM IN625
alloys in Fig. 1-4b. As shown so far, at both 650 ˚C and 800 ˚C, AM IN625 exhibits inferior
ductility to wrought IN625 under all creep testing conditions in this study. In other words, AM
IN625 is significantly embrittled at least at some high temperatures. As typical application
temperature of IN625 is higher or equal to 650 ˚C (Fig. 2-5), the high temperature embrittlement
problem of AM IN625 indicates that AM IN625 maybe not appropriate for the full replacement
for the wrought IN625 in many applications.

In order mitigate this difficulty in application of AM IN625, a careful investigation on the high
temperature embrittlement of AM IN625 is required.
As discussed previously in section 1.1.2, three main mechanisms can be suggested for the high
temperature embrittlement for AM IN625: impurity segregation at grain boundaries, intergranular
precipitation of hard particles, and strain aging effect. Among the three mechanisms, the
intergranular precipitation of secondary phases has already been rejected according to the results
in section 4.3.2. In addition to that, porosity of the AM IN625 can be a detrimental factor for the
embrittlement. Therefore, those three possibilities (impurity segregation, porosity, and strain
aging) will be investigated in this part to investigate the high temperature embrittlement of AM
IN625 and to propose a possible solution for the embrittlement problem.

182

5.1. Fracture Analysis for Creep Tested AM IN625 and Wrought IN625 Specimens
Fracture analysis includes the observation of the ruptured surfaces of creep tested IN625 specimens
and chemical composition analysis on the secondary cracks Additionally, EBSD data for
secondary cracks were collected to clarify if the fracture mode in AM IN625 samples was
intergranular.

The purpose of taking fractography is to investigate the high temperature embrittlement of AM
IN625 alloys by specifying the fracture mode. In general, observing fractography of ruptured
samples helps researchers to understand the aspect of failure development ( ≅ fracture mode) of
samples.
SEM images of the fracture surfaces of creep tested AM IN625 and wrought IN625 were obtained
to explore any differences in the high temperature fracture mode between the two IN625 alloys.

The chemical composition analysis using WDS for the failed surface can provide some information
on the secondary phases which might accelerate the crack nucleation and propagation. As the
ruptured IN625 samples in this study was exposed to the furnace environment during the cooling
phase of creep tests, the possibility that the failed surface had already been contaminated is not
negligible.
Therefore, rather than conducting the WDS directly on the fracture surface, the perpendicular
section to the fracture plane was mechanically polished and some secondary cracks in the vicinity
of the failure were examined to obtain the chemical data of secondary cracks.

183

5.1.1. Fractographs of Ruptured AM IN625 and Wrought IN625 Samples
The following Figure 5-1 shows the failure surface of creep tested AM IN625 and wrought IN625
samples under the two creep testing conditions: 650 ˚C, 658 MPa; 800 ˚C, 192 MPa.
Those ruptured specimens include all the three pre-heat treatment conditions in this study, i.e.
SCT, 6 months LHT, and 1 year LHT conditions.
In order to provide the entire morphology of the failure surface with a magnified image, each figure
in Fig. 5-1 includes two images under different magnification. The higher magnification image is
the larger image and lower magnification image is the smaller image (window).

184

Figure 5-1. SEM fractographs of failed IN625 samples after creep tests [(a),(b),(c),(d)] SCT creep
tested samples, [(e),(f),(g),(h)] 6 months LHT samples, and [(i),(j),(k),(l)] 1 year LHT samples.
The information on the top-right of each figure gives the sample type whether AM IN625 or
wrought IN625 with its creep testing conditions. R.T. on the bottom-right is the rupture time. The
window in each image is the SEM image of the same sample under a lower magnification. The
scale bars for the images are shown in (c).

185

In Fig. 5-1, all fractographies of AM IN625 samples exhibit faceted morphologies with sharp
edges, which suggests intergranular brittle fracture occurred in all AM IN625 samples under all
creep testing conditions in Fig. 5-1. Some small, globular particles are found at the fracture surface
of AM IN625 samples, however, the identification result of those small particles using SEM-EDS
was not reliable due to the surface roughness of the sample.
Those web-like structure on the surface facets (looks like wrinkles on the facets) in Fig. 5-1a, 5-
1b, 5-1e, 5-1f, 5-1i, and 5-1j is presumably small intergranular dimples [173]. The presence of
small and dense small intergranular dimples on the surface suggests that the coalescence of those
small dimples did not occur in those creep tested AM IN625 samples. In other words, AM IN625
did not allow the formation of large dimples at grain boundaries through the coalescence of those
small intergranular cavities during the high temperature creep tests.

In contrast, those fractographies of wrought IN625 samples show large dimple formations on the
failed surface after the creep tests. The depth and diameter of dimples of creep ruptured wrought
IN625 are weigh larger than those of AM IN625 samples after their failure, which suggests the
grain boundary strength of wrought IN625 was sufficiently higher to allow merging cavities at
grain boundaries without propagating intergranular cracks. Additionally, unlike the case of AM
IN625, the fracture surface of wrought IN625 does not exhibit the faceted morphology, rather, the
“elongated” morphology along the tensile load direction is well observed. Thus, the fracture mode
of wrought IN625 is not the intergranular brittle fracture. In fact, a typical ductile fracture mode is
shown on the fracture surface of creep deformed wrought IN625 samples in Fig. 5-1. Some large
particles located inside those dimples are identified as MC carbides.

186

5.1.2. EPMA-WDS Analysis on Secondary Cracks of Ruptured AM IN625 and Wrought IN625
Samples
As clearly shown in Fig. 5-1, AM IN625 exhibits a general intergranular embrittlement behavior
in all creep testing conditions. Several previous papers [4,14,21] proposed one possibility that
secondary phases at grain boundaries can lead to the intergranular embrittlement in Ni-based
superalloys.
However, as already shown in Fig. 4-20, the total quantity of intergranular secondary phases
cannot explain the poor ductility of AM IN625. Therefore, rather than focusing on the quantity,
identifying the type of secondary phases located at grain boundaries of both AM IN625 and
wrought IN625 may provide a more analysis for the poor ductility of AM IN625.

Figure 5-1 and Figure 5-2 give the WDS analysis results for secondary cracks in AM IN625 and
wrought IN625 samples, respectively, under the two creep testing conditions: 650 ˚C, 658 MPa;
800 ˚C, 192 MPa.
Those ruptured specimens include all the two pre-heat treatment conditions in this study, i.e. SCT,
6 months LHT conditions.

187

Figure 5-2. EPMA-WDS analysis for the secondary cracks in creep tested AM IN625 samples (a)
SCT sample at 650 ˚C and 658 MPa, (b) SCT sample at 800 ˚C and 192 MPa, (c) 6 months LHT
sample at 650 ˚C and 658 MPa, (d) 6 months LHT sample at 800 ˚C and 192 MPa.
188

Figure 5-3. EPMA-WDS analysis for the secondary cracks in creep tested wrought IN625 samples
(a) SCT sample at 650 ˚C and 658 MPa, (b) SCT sample at 800 ˚C and 192 MPa, (c) 6 months
LHT sample at 650 ˚C and 658 MPa, (d) 6 months LHT sample at 800 ˚C and 192 MPa.
189

The concurrent strong signals of both Al and O in the EPMA analyses of Fig. 5-2 suggest the
presence of Al2O3 particles. Similarly, the simultaneous Ti and N strong signals in Fig. 5-2b and
Fig. 5-2c indicate the TiN ceramic inclusion. Hence, the result in Fig. 5-1 show that Al2O3 or TiN
inclusions are present at the secondary crack of AM IN625 samples.
One strange observation in Figs. 5-2c and 5-2d is the continuous oxygen signal along the secondary
crack and it does not appear exactly coexisting with Al signals unlike Figs. 5-2a and 5-2b. From
the continuous Si in Fig. 5-2c, it is determined that this unusual oxygen signal possibly stems from
colloidal silica contamination which occurred during the sample preparation process.
It appeared from Fig. 5-2 that, in a couple of cases, the intergranular cracks may have originated
from the alumina particles. The volume fraction of Al2O3 in the AM alloy is about 0.4 %. In Table
2-2, the oxygen content of our AM samples were 200 ppm and this value was ten times higher
than that of wrought 625 alloy in this study. The calculated Al2O3 fraction from the oxygen content
in AM IN625 is ≈ 0.36, which reasonably agrees with the measured value of ≈ 0.4 %.
A previous study [51] reported that oxygen content over 140 ppm in powder metallurgy Inconel
718 contributed to a loss of ductility at elevated temperature while oxygen content seemed to do
not affect on the room temperature ductility. Additionally, some metal oxide particles (presumably
Al2O3 and TiO2) were found in the previous study generally at prior particle boundaries where the
embrittlement occurred. However, those prior particle boundaries were not observed in AM IN625
samples in this study.
All secondary cracks in Fig. 5-2 propagate perpendicular to the load direction.

In Fig. 5-3a and 5-3b, the absence of Al signals in the creep tested wrought IN625 specimens
indicates the absence of Al2O3 inclusions and this result is coincident effectively with the
190

microstructure observation result in section 4.1 and 4.2. The strong simultaneous Nb, Ti, and C
signals in those samples represent the presence of MC type carbides. EPMA-WDS results in Fig.
5-3 suggest that some MC carbides are located at the secondary cracks, implying those large
particles inside large dimples in Fig. 5-1 are MC carbides.
One interesting aspect in Figs. 5-3a, 5-3b and 5-3d is the strong oxygen level at MC carbides. Si
signal in Fig. 5-3d implies that those strong oxygen signals may be originated from the colloidal
silica contamination similar to the cases in Figs. 5-2c and 5-2d. However, as shown in Fig. 5-3a,
the oxygen signal is collected exactly at the identical sites to where MC carbides are present. As it
has been reported previously [43-45], this unusual oxygen signal at MC carbides is probably due
to the oxidation of metallic carbides in wrought IN625 samples. The surface oxidation of MC (M
is mostly Nb and Ti) or M23C6 (M is mostly Cr) carbides can produce NbO, TiO2 or Cr2O3 oxide
films on the surface of those carbides in Ni-based superalloys under elevated temperature
conditions [43-45]. The high Cr signal in Figs. 5-3b and 5-3d, and the high Nb signal in Figs. 5-
3a, 5-3b and 5-3d are concurrently detected with high oxygen signals in those figures.
Interestingly, unlike those secondary cracks of AM IN625 samples (Fig. 5-2), most secondary
cracks in Fig. 5-3 are elongated along the loading direction. In other words, the crack in wrought
IN625 alloys propagated along the loading direction even though the maximum stress is applied
perpendicular to the loading direction upon a tensile load [174]. This strange cracking direction in
Fig. 5-3 is due to necking during the tensile creep tests in wrought IN625 samples. The radial load
component, activated during the necking, is perpendicular to the load direction and it can lead to
the crack (or cavity) growth parallel to the loading direction [174].

191

Eight secondary cracks in AM IN625 alloys were examined using the WDS method and it was
confirmed that 6 cracks had Al2O3 inclusions at the crack surface. Even though the Al2O3 signal
was not collected in two cracks, this is may be due to the small size of most Al2O3 inclusions in
AM IN625 (≈ 0.14 µm); the WDS typically has the resolution of only ≈ 1 µm [124]. Thus, it was
determined that careful investigation to the role of Al2O3 in the elevated temperature embrittlement
of AM IN625 and chemical characterization of grain boundaries needed to be conducted.

5.1.3. EBSD Mapping for Secondary Cracks of Creep Tested AM IN625
In those previous sections 5.1.1 and 5.1.2, the fracture mode of AM IN625 alloys has been
categorized as intergranular fracture without any clear evidence. One ideal method to clarify
whether the secondary cracks propagated along grain boundaries or not is the EBSD. As illustrated
in section 2.4.5, the orientation information of some areas in the vicinity of secondary cracks is
easily collected by conducting the EBSD analysis on the cracked sample. By using the orientation
data, the fracture mode can be readily defined.

Figure 5-4 provides the EBSD results for several secondary cracks in AM IN625 alloys after
rupture.

192

Figure 5-4. The EBSD maps for secondary cracks in creep deformed AM IN625 samples (a) 650
˚C and 658 MPa for 2.6 h (b), 800 ˚C and 192 MPa for 6.4 h (c) 6 months LHT and creep tested at
650 ˚C, 658 MPa for 20 h (d) 6 months LHT and creep tested at 800 ˚C, 192 MPa for 11.4 h, and
(e) the misorientation profile of grain boundaries where secondary cracks formed. Grain
boundaries are shown in white lines and the black regions are cracked. All cracks in this figure
propagated along the perpendicular direction to the loading direction and all of them are
intergranular.
193

In EBSD maps of Fig. 5-4, grain boundaries are shown as white lines and cracks are illustrated in
black. 37 cracks, in total, including data in Fig. 5-4, were examined by EBSD and all cracks were
intergranular.

Fig. 5-4e shows the measured misorientation profile of grain boundaries where secondary cracks
are located. The misorientation profile indicates that > 80% of grain boundaries where cracks had
nucleated or propagated were randomly oriented grain boundaries (non-special Σ3 CSL boundary:
Σ3 is a grain boundary rotated about 60 ˚ about [111] direction). Thus, the fracture of AM IN625
samples mostly occurred along random grain boundaries during the high temperature creep testing.
It has been reported that special boundaries, mainly Σ3 boundaries, are not favorable crack paths
due to their lower grain boundary energy, whereas random grain boundaries are the favorable path
for crack propagation [175-178].

5.2. Chemical Characterization for Grain Boundary Regions in AM IN625 Specimens
From the results in the previous section 5.1, the high temperature intergranular embrittlement is
evident in AM IN625 alloys of this work. Specifically, most crack propagation occurred along
randomly oriented grain boundaries and some Al2O3 inclusions located at those grain boundaries
appear closely related to the embrittlement of AM IN625 samples.
However, as previously discussed in section 1.1.2, impurity segregation, such as sulfur and
oxygen, at grain boundaries generally causes the high temperature embrittlement in Ni-based
superalloys. Therefore, the impurity behavior at randomly oriented grain boundaries must be
conducted to elucidate the role of Al2O3 inclusions at random grain boundaries.
194

Due to the high reliability in identifying species on the sample surface, EPMA-WDS may be
chosen first to detect impurities at grain boundaries. However, its poor spatial detecting resolution
( ≈ 1 µm) restricts the application of EPMA-WDS analysis to the minor impurity observation at
grain boundaries [124]. Instead, Nano-SIMS and APT were used to characterize the chemical
information in the vicinity of random grain boundaries of AM IN625 samples.

5.2.1. Nano-SIMS Results for Both AM IN625 and Wrought IN625 in Non-LHT or 6 Months
LHT Conditions
The purpose of Nano-SIMS analysis is to investigate the chemical information of several grain
boundaries with better resolution than the WDS. Typically, a spatial resolution of 50 nm is
available in the Nano-SIMS machine when it uses Cs
+
ions as the implanting source [135].

Nano-SIMS data were collected for as-HIPed AM IN625, as-solution annealed wrought IN625, 6
months LHT AM IN625, and 6 months LHT wrought IN625. The randomly oriented grain
boundaries in those samples were selected to obtain the Nano-SIMS data.
The Nano-SIMS results are given in the following Figures 5-5 and 5-6.

195

Figure 5-5. Nano-SIMS results for non-LHT IN625 alloys (a) as-HIPed IN625, (b) as-solution
annealed IN625. All samples were undeformed. Lighter regions in each Nano-SIMS figure
indicate higher concentration of ions.
27
Al
16
O signal represents the position of Al2O3 particles and
sulfur is highly concentrated at where Al2O3 particles are located in the AM IN625 sample. No
strong sulfur segregation is not observed in the wrought IN625.

196

Figure 5-6. Nano-SIMS results for 6 months LHT IN625 alloys (a) 6 months LHT AM IN625, (b)
6 months wrought IN625. All samples were undeformed. Lighter regions in each Nano-SIMS
figure indicate higher concentration of ions.
27
Al
16
O signal represents the position of Al2O3
particles and sulfur is highly concentrated at where Al2O3 particles are located in the AM IN625
sample. No strong sulfur segregation is not observed in the wrought IN625.
197

In those Nano-SIMS data of Figs. 5-5 and 5-6, At least two grain boundaries, including at least
one random grain boundary, were examined in each specimen. No element segregation at grain
boundaries was observed in either as-HIPed AM IN625 or as-solution annealed wrought IN625.
However, some strong
12
C (for carbon) and
11
B
16
O (for boron) at grain boundaries of the 6 months
LHT AM and wrought samples are evident as shown in Figs. 5-5b and 5-6b. The grain boundary
boron segregation under prolonged thermal exposures at 650 ˚C has been reported in cast IN625
and wrought Inconel 600 [179,180], and it may have improved the high temperature ductility (at
700 ˚C) of the IN625 by a factor of 1.3 [179]. An increase in carbon content from 0.003 wt% to
0.055 wt% showed a negligible effect on the creep ductility of a powder metallurgy Ni base
superalloy tested at 732 ˚C [181]. Thus, both boron and carbon segregation cannot explain the
ductility drop in the 6 months thermally exposed AM IN625.

The
27
Al
16
O signal in both Figs. 5-5 and 5-6 indicates the location of Al2O3 inclusions in the
alloys. Both as-HIPed and 6 months LHT AM IN625 samples show that Al2O3 uniformly dispersed
in the Ni matrix and grain boundaries. The absence of
27
Al
16
O signal in Figs. 5-5b and 5-6b indicate
that alumina is present only in AM IN625 alloy in this study. The composition of oxygen in AM
IN625 is a factor of 10 higher than the wrought alloy. Even though some
16
O or
18
O signals were
collected at Al2O3, this signal appears to primarily originate from Al2O3 itself rather than
monatomic oxygen segregation at the Al2O3. Interestingly, Al2O3 particles emitted strong
32
S
signals, which suggests sulfur is segregated at the alumina/matrix interfaces. While some small
32
S signals were collected in wrought IN625 (Figs. 5-5b and 5-6d), the intensities of those signals
were much lower than in Figs. 5-5a and 5-6a for AM alloys. The sulfur segregation level of each
sample was firstly estimated by averaging the
32
S/
58
Ni peak values in the Nano-SIMS data (mostly
198

at Al2O3 in AM IN625 alloys and carbon segregated regions in wrought IN625). The measured
normalized sulfur segregation level (
32
S/
58
Ni) was 1.31 ± 0.97 for as-HIPed AM IN625 and 0.18
± 0.11 for as-solution annealed wrought IN625. It was 3.08 ± 2.69 for the 6 months LHT AM
IN625 and 0.13 ± 0.03 for the 6 months LHT wrought IN625. From these values, the average
sulfur segregation level (at Al2O3 for AM IN625 and at carbon-segregated regions for wrought
IN625) of each sample can be calculated as 721 ± 534 ppm for as-HIPed AM IN625; 99 ± 61 ppm
for as-solution annealed wrought IN625; 1694 ± 1480 ppm for the 6 months LHT AM IN625 and
72 ± 17 ppm for wrought IN625. Thus, the sulfur level at Al2O3 of AM IN625 increased over the
6 months thermal treatment at 650 ˚C. In contrast, no remarkable sulfur level change occurred in
the wrought IN625 over the same heat treatment.

The results of Nano-SIMS data in this work indicate the high sulfur segregation at Al2O3 interface,
rather than randomly oriented grain boundaries in AM IN625 alloys. Some sulfur segregation were
also observed at the carbide interface of wrought IN625, however, the intensity of sulfur
segregation in the wrought IN625 samples is about 7 to 23 times lower than the sulfur segregation
level in the AM IN625 alloys.

5.2.2. APT Analysis for Randomly Oriented Grain Boundaries in AM IN625 Samples
In order to focus on the chemical information on random grain boundaries of creep tested AM
IN625 samples, APT analysis for two creep tested AM IN625 alloys were conducted as shown in
Figure 5-7.
199

Figure 5-7. The chemical distribution plot along given distances from the grain boundaries and
γ’’ interface in APT samples: (a) the creep deformed AM IN625 sample at 294 MPa and 650 ˚C
within 24 h (b) the creep deformed 6 months LHT AM IN625 sample at 658 MPa and 650 ˚C (c)
concentration profile near γ’’ particles in (b). The shaded area are grain boundary regions or γ’’
precipitates.

200

As the fracture of AM IN625 is intergranular, an analysis of grain boundary chemical
characteristics was required. APT was chosen to obtain the chemical information on a random
grain boundary in both SCT AM IN625 (at 294 MPa and 650 ˚C) and creep-tested 6 months LHT
AM IN625 (at 658 MPa and 650 ˚C). Figure 5-7a presents the element profiles in the vicinity of
the grain boundary in the SCT AM IN625 sample. The element profiles of the LHT AM IN625 is
given in Fig. 5-7b and c shows the concentration profile of elements near the γ’’ precipitate. The
γ’’ chemical composition (Fig. 5-7c) was investigated in order to understand the intergranular
oxygen content decrease in the 6 months LHT + creep tested AM IN625 (Fig. 5-7b) compared to
the SCT AM IN625 at 650 ˚C (Fig. 5-7a).

Fig. 5-7a exhibits not only the segregation of major metallic elements such as Cr, Mo and Si, but
also minor segregations of several non-metallic elements (C, B, P and O) on the grain boundary.
Cr and C co-segregation is typically considered as the indicative of the precipitation of Cr-rich
M23C6 carbide [20]. M23C6 precipitation due to Cr-C co-segregation did not embrittle a Ni-Cr-W
alloy over an aging from 30 mins to 5 h at 700 ˚C [182]. A high concentration of Mo at grain
boundary also did not significantly influence the rupture life of Inconel 718 at 650 ˚C as well [58].
Similar to the effect of boron [179], phosphorous segregation can improve cohesion of grain
boundary, which increased the stress rupture life of a Ni base superalloy [64]. Hence P is not
expected to be a basis for the embrittlement of AM IN625. As it has been widely reported, oxygen
can lead to the grain boundary embrittlement of Ni base superalloys [35-51]. However, the
approximate number of measured oxygen atoms in the grain boundary area was 341 atoms, which
covers only 0.06 % of a monolayer of the grain boundary.

201

Fig. 5-7b reveals Nb and B segregation at the grain boundary of 6 months LHT AM IN625 sample
after a creep test at 658 MPa and 650 ˚C. Similar to the effect of M o at grain boundaries, Nb
segregation did not degrade the stress rupture life of Inconel 718 [58]. Fig. 5-7b also suggests that
the slight oxygen segregation at grain boundary in the 650 ˚C SCT AM IN625 has almost
disappeared in the AM IN625 sample during the 6 months LHT. Instead, somewhat high oxygen
concentration was observed in matrix γ’’ precipitates as shown in Fig. 5-7c. Thus, the oxygen
effect is not likely to explain the creep ductility degradation in AM IN625 alloys both before and
after 6 months LHT.

5.2.3. Discussions on the Embrittlement by Oxygen
According to previous research [35-51], oxide particle induced embrittlement [37-40] and oxygen
atom induced embrittlement [35,36] are the two main mechanisms for oxygen embrittlement in
several Ni based superalloys.

Ceramic inclusions such as Al2O3 and TiN are present in AM IN625 alloys in this study (Fig. 4-5)
that are not present in the wrought alloy. However, at ambient temperature, where wrought and
AM have identical ductility, the Al2O3 is benign. Although embrittlement by monotonic oxygen
and sulfur occurs over a limited range of elevated temperatures, it is unclear how this would cause
selective embrittlement at certain elevated temperatures.

Additionally, the monotonic oxygen at grain boundaries can contribute to the embrittlement of Ni
alloys [35,36]. According to Pfaendtner et al. [35] monotonic oxygen segregated at grain
202

boundaries is responsible for embrittlement of Inconel 718 over a restricted elevated temperature
range. Grain boundary monatomic oxygen atoms are absorbed at the crack tip due to the high local
triaxial stress fields just ahead of the tip. However, as shown in Figs. 5-5 and 5-6, negligible
oxygen segregation at random grain boundaries is present even after 6 months LHT at 650 ˚C,
since Fig. 5-7b suggests no significant oxygen segregation in a creep-tested 6 months LHT AM
IN625. Moreover, the APT analyses in Fig. 5-7 exhibit a slight oxygen segregation at grain
boundaries. After short-term creep test of as-HIPed AM IN625, the amount of oxygen atoms at
the boundary can cover only 0.06 % of a monolayer at the grain boundary. Therefore, the oxygen
atoms at grain boundaries do not appear to have a role in the poor elevated-temperature ductility
of AM IN625 unlike Inconel 718.

Additionally, for AM IN625, the calculated maximum volume fraction of ceramic inclusions
according to the given oxygen and nitrogen levels in Table 2-2 was 0.0036, which is within the
range of measured ceramic inclusion fraction, 0.0041 ± 0.0009. Thus, it appears most oxygen
atoms in AM IN625 lead to the formation of Al2O3 particles rather than segregation at the grain
boundaries. Also, even after the 6 months LHT in air, which may allow a significant oxygen
penetration from air to grain boundaries, the wrought IN625 sample did not exhibit any drop in
ductility for both 650 ˚C and 800 ˚C creep tests as shown in Figs. 3-5b and 3-5c. This result casts
doubt on whether the grain boundary oxygen contributes to the ductility drop in AM IN625. The
case of 650 ˚C creep tested AM IN625 with an LHT of 1 year showed a similar ductility to the 6
months LHT condition. The ductility of wrought IN625 dropped after the 1 year LHT (and scatter
may be partially responsible), however it was still larger than that of AM IN625. The oxygen (from
air) grain boundary penetration appears to be ineffective in decreasing the ductility of AM IN625.
203

5.2.4. Discussions on Embrittlement by Sulfur
The effects of sulfur on the embrittlement of various Ni-based superalloys has been reported in
several studies [53-59]. Typically, sulfur tends to segregate at grain boundaries and this leads to
intergranular brittle fracture in Ni and Ni-based superalloys. However, the addition of some strong
sulfur getters such as Mg can prevent the sulfur-induced embrittlement of Ni-based superalloys
[54]. This may partially explain the higher creep ductility of wrought IN625 in this study as
wrought IN625 contains more Mg than AM IN625 according to Table 2-2.

There is significant sulfur segregation at Al2O3/matrix interface as confirmed by the Nano-SIMS
results (Figs. 5-5 and 5-6). One study [56] calculated the segregation energy of sulfur at both grain
boundary and free surface in the Ni matrix. The calculation revealed that sulfur is likely to be
segregated at free surfaces rather than grain boundaries. The incoherent interface between Al2O3
and Ni matrix may act like a free surface more than a coherent interface, which may rationalize
the strong sulfur segregation at Al2O3 particles illustrated in Figs. 5-5 and 5-6. In a previous study
by Meier et al, it was proposed sulfur is segregated at Al2O3 – Ni interfaces of PWA 14xx Ni
superalloys due to weak adherence between Al2O3 and Ni [184].

One question driven by this proposition is why the intergranular fracture, rather than transgranular
mode, was primarily present in Fig. 5-4 in spite of the uniform distribution of alumina inclusions.
First, the diffusion of sulfur at grain boundaries is typically faster than in the matrix so
embrittlement would be faster at the grain boundaries [53,185]. Additionally, T. Denda et al. [37]
204

explored the size effect of Al2O3 and TiN on the low cycle fatigue life of Inconel 718 sintered
powder at 538 ˚C. Their data revealed that the fatigue life of Inconel 718 significantly dropped
when Al2O3 and TiN, were larger than 10 µm. Sub-micron sized Al2O3 and TiN inclusions had no
impact on the fatigue life. This suggests that the size of Al2O3 inclusion in AM IN625 (0.14 ± 0.11
µm) appears not large enough to sustain sufficient strain for crack nucleation at Al2O3 particles.
Thus, rather than the pure Al2O3 inclusion effect, faster sulfur diffusion from Al2O3 to intergranular
crack tips is considered important for the poor ductility of AM IN625.
As shown in Fig. 5-4e, intergranular crack formation occurred mostly along random high angle
grain boundaries (RHAB). In Ni based alloys, as the RHAB provide a favorable site for impurity
diffusion, like sulfur, more than low-CSL boundaries [186], the large fraction of low-CSL
boundaries was effective on improving sulfur segregation-induced intergranular brittleness in
ultrafine grained polycrystalline Ni [187]. In order to check the contribution of RHAB frequency
to the intergranular embrittlement in AM IN625, the random grain boundary structures of several
non-creep tested AM IN625 and wrought IN625 are evident in Figure 5-8. As the RHAB fractions
of non-creep tested AM IN625 samples is not higher than those of non-creep tested wrought IN625
even after the 6 months LHT, it is determined that the grain boundary character of AM IN625
cannot simply explain the intergranular embrittlement by itself. Instead, the sulfur diffusion along
RHABs from intergranular Al2O3 inclusions seems reasonable to interpret the intergranular
fracture along RHABs observed in Fig. 5-4 (AM IN625 samples).
205

Figure 5-8. The grain boundary character maps for several non-creep deformed AM IN625 and
wrought IN625 samples. CSL (Σ3) and RHAB fraction of each sample is given in the title.

206

In summary, the sulfur causes cracking at the matrix interface. The sulfur diffuses from the
intergranular Al2O3 inclusions to the crack tip along mostly RHABs due to the high triaxial stress
state ahead of the crack tip. There is a repeating process of cracking, sulfur diffusion to the triaxial
zone ahead of the crack and new cracking. The AM IN625 embrittlement could be eliminated if
the sulfur concentration is reduced to 0.0015 wt% and Mg is added to 0.005 wt.% as it is the case
for wrought alloys. The Mg ties up the sulfur without and reduces the possibility of embrittlement
[54]. The wrought has this amount of Mg, while the AM powder we utilized in this study had none.
Additionally, as Zr and Hf can act like a scavenger for sulfur during the solidification process of
Ni-based superalloys [53], adding trace amount of Zr and Hf may enhance the high temperature
ductility of AM IN625.

5.3. Other Possibilities Causing the High Temperature Embrittlement in AM IN625
From the experimental works in section 5.1 and 5.2, the sulfur effect is determined as the foremost
possibility which led to the embrittlement in AM IN625 alloys.

Yet, there are two more possibilities which can contribute to the high temperature poor ductility
of AM IN625. Porosity and strain-aging effect are the two possibilities.
Porosity was measured using an OM machine and misfit strain effect was evaluated using XRD
technique.

207

5.3.1. Porosity Measurement for Non-LHT AM IN625 and Wrought IN625 Samples
Figure 5-9 shows the optical micrographs for porosity measurements of AM IN625 samples in this
study. The white dots in dark field images represent the pores in the samples.
The as-solution annealed, non-HIPed, AM IN625 sample has a porosity of 0.19 ± 0.08 %, the as-
HIPed AM IN625 alloy has a porosity of 0.06 ± 0.01 % while the wrought has < 0.02 % porosity.
Had the AM HIPed poroity been reduced to the level of that in the wrought alloy, it is possible that
the ductility would have improved. R. Haynes has reported that the ductility of sintered steel can
decrease by 5 % to 8 % in the case of steel when the porosities of samples are different by ~0.1%
[188]. The measured average pore diameter in Fig. 5-9a is 0.47 ± 0.24 µm. The average pore
diameters of AM IN625 alloys are 0.73 ± 0.37 µm for the as-HIPed sample and 1.13 ± 1.47 µm
for the non-HIPed (as-solution treated) sample. Dubensky et al. reported that the pore size is
inversely proportional to the ductility of aluminum alloys [189].
Therefore, the improved ductility of AM HIPed IN625 in comparison of that of AM non-HIPed
IN625 in Fig. 3-5b and 3-5c can be explained with the porosity decrease after the HIP process.
However, the porosity values of AM HIPed IN625 alloy and wrought IN625 alloy in this study are
almost comparable and some references [74,83,190] suggested that room temperature ductility of
AM HIPed Inconel alloys is identical to that of wrought samples while high temperature ductility
of AM IN625 samples are typically inferior to that of wrought alloys.

In contrast, non-HIPed AM IN625 has a poorer ductility even at room temperature than wrought
alloy [191]. Thus, the significant drop in ductility of AM-HIPed IN625 samples when it compared
to wrought IN625 sample appears to difficult to rationalize with the porosity effect.
208

Figure 5-9. Optical micrographs of IN625 samples (a) bright field image of as-solution annealed
wrought IN625, (b) dark field image of (a), (c) bright field image of as-solution annealed (non-
HIPed) AM IN625, (d) dark field image of (c), (e) bright field image of as-HIPed AM IN625, (f)
dark field image of (e).
209

5.3.2. Misfit Strain Effect on the Experimental IN625 Alloys
One interesting aspect in the microstructure of 1 year LHT AM IN625 (Fig. 4-4a) is the formation
of subgrain boundaries by pure heat exposure to 650 ˚C without any stress on the sample. In the
case of a γ’ strengthened Inconel 713C, Singh et al. [192] observed some generation of misfit
dislocations during the aging treatment at 1000 ˚C. Coherency loss in a plate type precipitate (like
γ’’) is typically accepted as the source for misfit dislocation generation on the interface of the
precipitate [31]. Several studies [193-195] have suggested that the coherency loss in γ’’ generally
occurred at the its critical size around 80 nm to 120 nm, which agrees well with the γ’’ size of 1
year LHT AM IN625 in this study ( ≈ 145 nm, Fig. 4-19b). As these misfit strain generation seems
similar to the residual strain after welding of Ni superalloys, the misfit strain generated during the
1 year LHT might occur a misfit strain induced embrittlement. This proposed mechanism is
generally similar to the strain-aging effect on the welded Ni superalloys.

In welded Ni superalloys, residual strains generated during the restrained cooling can cause high
temperature embrittlement with matrix hardening precipitates such as γ’ or γ’’ [61,62].
Specifically, the strain age-cracking generally occurs because residual strain in the welded samples
is mostly transferred to grain boundary upon elevated temperature service. In this phenomenon, γ’
or γ’’ matrix hardening is the main reason why the residual strain becomes concentrated on grain
boundaries rather than uniformly distributed in matrix [62]. The strain localization leads to the
early crack nucleation and propagation at grain boundaries or intergranular precipitate interfaces
in Ni-base superalloys [61,62]. However, this traditional strain-age embrittlement concept seems
hard to be directly applied to this study as all the experimental samples are thought to be “fully
stress-relieved”, by either HIP or solution annealing, before creep tests.
210

Strain-age embrittlement requires the sufficient precipitation of matrix strengthening γ’ or γ’’ and
the concurrent presence of stresses on the material having residual stress [61]. It suggests that a
residual-stress induced embrittlement in a precipitation hardened material can be defined as strain-
age embrittlement. The difference between the general strain-aging effect and the misfit strain
effect (in this study) is the source of residual strain. The general strain-aging obtains the residual
strain from the fast cooling and resultant thermal shock in welding, whereas the misfit strain occurs
under an isothermal exposure and subsequent precipitate coherency loss.
As illustrated in Table 3-2 and Fig. 4-19a, the precipitation hardening behavior of AM IN625
samples was observed when γ’’ volume fraction was more than 14 %. While the AM IN625
showed 14 % γ’’ precipitation after 24 hours at 650 ˚C, the wrought IN625 precipitated only ≈ 0.7
% of γ’’ in the same heat exposal condition (Fig. 4-19a). However, after the 6 months LHT, the
amount of γ’’ in both the AM IN625 and the wrought IN625 provided some substantial
precipitation hardening effects as shown in Table 3-2.

Fig. 5-10 gives XRD data analysis results for several AM IN625 and wrought IN625 samples.
211

Figure 5-10. XRD analysis for several IN625 samples (a) matrix lattice parameter change in both
AM IN625 and wrought IN625 over the 1 year LHT, (b) XRD line profiles for four LHT IN625
samples: 6 months LHT AM IN625, 1 year LHT AM IN625, 6 months LHT wrought IN625, and
1 year LHT wrought IN625. ΔV is the volume change.
212

In Fig. 5-10a, the lattice contraction in both AM IN625 and wrought IN625 occurred over the 1
year LHT. The volume contraction observed in Fig. 5-10a may be related to the misfit dislocation
generation during the 1 year LHT of AM IN625. However, as shown in Fig. 4-19b, this gradual
decrease in lattice parameters of both AM IN625 and wrought IN625 appears primarily due to the
Nb content drop in γ matrix with γ’’ precipitation.
Zhang et al. [196] suggested that the misfit strain due to γ’’ coarsening can be measured by the
gap between those interplanar spacings of {200}γ and (004)γ’’. The calculated total misfit strain
in the 1 year LHT AM IN625 is ≈ 0.03 along the c-axis of γ’’. However, as this amount of strain
includes both elastic and plastic strain, the plastic strain needs to be evaluated from the calculate
0.03 strain.

Coherency stress near the matrix/precipitate interface can punch dislocation loops around the
precipitate [195]. In order to punch the dislocation loops, a stress criterion should be satisfied so
that the shear stress by coherency enables to exceed the critical resolved shear stress (CRSS) of
the matrix. Weatherly et al. [197] proposed the maximum stress applied on the slip planes of a
plate-type precipitate is about 3/2· µ 𝜀𝜀 3 3
𝑇𝑇 . µ is the shear modulus and 𝜀𝜀 3 3
𝑇𝑇 is the transformation strain
from coherent state to semi-(or in-)coherent state. Based on this idea, now the stress criterion for
punching dislocation loops around a plate-type precipitate can be suggested:

𝑀𝑀𝑀𝑀 𝑑𝑑 𝑟𝑟 𝑠𝑠 𝑀𝑀 𝐶𝐶 𝑅𝑅 𝐶𝐶𝐶𝐶 = 3/2 ∙ 𝜇𝜇 ∙ 𝜀𝜀 3 3
𝑇𝑇 (5-1)

Now, the Ni matrix CRSS at 650 ˚C is µ/30.8 [195]. Therefore, the 𝜀𝜀 3 3
𝑇𝑇 is calculated 0.022.
213

Hence, the effective plastic misfit strain originated from the γ’’ coarsening is ≈ 0.008. The misfit
dislocation density can be estimated from this effective plastic misfit strain. The misfit dislocation
density on a plate-like precipitate interface can be calibrated by the following equation [198]:

𝜌𝜌 𝑀𝑀𝑀𝑀
= 𝜀𝜀 𝑒𝑒 𝑒𝑒 𝑒𝑒 /(| 𝑏𝑏 | ∙ 𝑐𝑐𝑐𝑐 𝜀𝜀 𝑠𝑠 ) (5-2)

where ρMD is the misfit dislocation density, 𝜀𝜀 𝑒𝑒 𝑒𝑒 𝑒𝑒 is the effective plastic misfit strain, and b is the
Burger’s vector, and 𝑠𝑠 is the angle between Burger’s vector and the precipitate interface. The 𝑐𝑐𝑐𝑐 𝜀𝜀 𝑠𝑠
of γ’’ precipitate in Ni matrix is 1/ √2 and b is ≈ 0.254 nm, so that the calculated ρMD is ≈ 0.04/nm.
This value suggests at least one dislocation is placed on every 25 nm of the precipitate diameter.
The linear density of γ’’, which is large enough to be at least semi-coherent (> 91 nm [195]), in
the 1 year LHT AM IN625 sample was measured to be 0.0004 ± 0.0001 /nm.
From the values of ρMD and the γ’’ linear density in the 1 year LHT AM IN625, the misfit
dislocation density was calibrated to be 1.65 ± 0.71×10
13
·m
-2
.

The generated misfit dislocation density appears not high enough to occur a residual strain-induced
embrittlement. As it is shown in Table 3-3, the dislocation density of as-solution treated wrought
IN625 was comparable with this value and any embrittlement was observed in the SCT wrought
IN625 samples. Therefore, the misfit strain effect cannot explain the embrittlement of AM IN625
in this study.

214

6. CONCLUSIONS
In this study, high temperature creep tests for additively manufactured Inconel 625 (AM IN625)
and wrought IN625 were conducted at 650 ˚C and 800 ˚C in the stress ranged between 65 MPa –
658 MPa. The AM IN625 samples were fabricated by the laser powder bed fusion method. Three
different heat treatments were applied to the alloys before creep tests: solution annealing of both
wrought and AM alloys, hot isostatic pressing (HIP) of AM alloys, 6 months and 12 months cyclic
exposure (continuous thermal cycling per Fig 1) at 650 ˚C for the two alloys.

(1) Overall, the high temperature creep strength of AM IN625 was equal or even stronger than
wrought IN625 over the whole range of temperature exposures, times and stress conditions. The
creep ductility of AM IN625 was inferior to that of wrought IN625 in all creep tests of this work.
After the cyclic long-term heat-treatments at 650 ˚C (LHT) over 6 months and 1 year, both AM
IN625 and wrought IN625 obtained some extra creep strength. The increased volume fraction of
γ’’ and δ phases in the LHT IN625 samples was proportional to creep strength improvement under
different testing conditions. In the case of creep ductility, AM IN625 exhibited further decreases
in the high temperature creep ductility after both 6 months and 1 year LHT. However, in the case
of wrought IN625, the creep ductility was conserved even after those LHT processes except for
one creep condition (650 ˚C, 658 MPa creep test of 1 year LHT wrought IN625).

(2) The calculated Qc of AM IN625 was 273 kJ/mol and wrought IN625 was of 284 kJ/mol before
the LHT. The Qc of AM IN625 decreased to 257 kJ/mol after the 6 months LHT (negligible
difference considering the many sources of error). As those Qc values are close to being equivalent
215

to Ni lattice self-diffusion, the creep mechanism in this alloy appears to be controlled by
dislocation climb at the two temperatures of this study.

(3) The TEM observation for non-LHT and non-creep deformed AM IN625 revealed the presence
of Al2O3 and TiN inclusions in as-HIPed and as-solution annealed AM IN625 samples without
other secondary phases. In the as-solution annealed wrought IN625, MC carbides were found and
no Al2O3 and TiN inclusions were confirmed. Upon high temperature creep tests, more and faster
matrix γ’’ (or δ) precipitation occurred in AM IN625 at 650 ˚C (or at 800 ˚C). The secondary phase
volume fraction (γ’’ and δ) in AM IN625 had remained higher than that in wrought IN625 over
the 1 year LHT and after subsequent creep tests at 650 ˚C and 800 ˚C. Aside from γ’’ and δ, M 23C6
and M6C carbides were formed at grain boundaries in both AM IN625 and wrought IN625 in all
two 650 ˚C and 800 ˚C thermal testing conditions except for 24 hours short term creep tests at 650
˚C. The grain boundary precipitation rate was similar between AM IN625 and wrought IN625 in
800 ˚C all creep test conditions. At 650 ˚C, the intergranular M23C6 was formed more rapidly in
AM IN625 than wrought IN625 in the 24 hours creep test, then the amount of intergranular
precipitation became comparable between the two IN625 alloys in all creep tested specimens after
6 months LHT.
Those LHT IN625 alloys (both AM IN625 and wrought IN625) showed the activation of
mechanical nano-twinning after the 650 ˚C, 658 MPa. In the 800 ˚C and 192 MPa creep test, only
LHT wrought IN625 exhibited the subgrain formation typically at grain boundaries.

(4) The creep ductility recovery in AM IN625 after the HIP is considered due to the lower porosity
in the HIPed AM IN625 ( ≈ 0.06 %) than the non-HIPed AM 625 ( ≈ 0.19%). However, the ductility
216

of AM HIPed IN625 was still poorer than wrought IN625. Additionally, some misfit dislocation
density increase was estimated over the 1 year LHT in the AM IN625, however, ductility drop in
AM IN625 was not explainable with the misfit strain induced embrittlement.
The fractography analysis on AM IN625 and wrought IN625 revealed that the intergranular brittle
fracture was typical in AM IN625, whereas the dimple formation induced ductile fracture occurred
in wrought IN625 in all examined creep testing conditions. EPMA-WDS analysis on secondary
cracks also indicated the presence of ceramic inclusions (Al2O3 and TiN) at the secondary cracks
in AM IN625 samples. However, some MC carbides were located at the secondary cracks in
wrought IN625 specimens as well.
The EBSD analysis on secondary cracks indicated that > 80 % of intergranular cracks propagated
along randomly oriented (non-CSL) grain boundaries. The APT results clarified that the oxygen
level at a randomly oriented grain boundary was too low to cause embrittlement in one 650 ˚C 24
hours creep tested AM IN625 sample. One more APT analysis on a randomly oriented grain
boundary of 6 months LHT and 650 ˚C creep tested AM IN625 additionally revealed that the
intergranular oxygen level rather dropped when compared to the 24 hours creep-deformed sample.
Nano-SIMS analysis showed high sulfur segregation at Al2O3 in both as-HIPed and 6 months LHT
AM alloys before creep tests. The sulfur segregation level at Al2O3 particles increased in the AM
IN625 alloy after the 6 months LHT. No sulfur and oxygen segregation at grain boundaries of the
AM IN625 alloys was observed.
Based on these results, intergranular sulfur diffusion (from Al2O3 at grain boundaries at high
temperatures (650 ˚C and 800 ˚C) is the most plausible explanation for the embrittlement in AM
IN625 in this study.

217

5) The embrittlement could be eliminated by having the S concentration being lowered to typical
values in the wrought alloy) e.g. < 15 ppm. Additionally, the addition of Mg as with the wrought
to AM alloys would tie up the sulfur and assist with eliminating the embrittlement.

218

REFERENCES
1. W. Betteridge, S.W.S. Shaw, Development of superalloys, Mater. Sci. Technol. 3 (1987)
682-694.
2. C.T. Sims, N.S. Stoloff, W.C. Hagel, Superalloys Ⅱ: High-Temperature Materials for
Aerospace and Industrial Power, second ed., John Wiley & Sons, New York, USA (1987).
3. D.E. Laughlin, K. Hono, Physical metallurgy, fifth ed., Elservier Ltd., Amsterdam, The
Netharlands (2015).
4. M.J. Donachie, S. Donachie, Superalloys: a technical guide, second ed., ASM
International, Materials Park (2002).
5. G. Smith and L. Shoemaker, Advanced nickel alloy, Advanced Materials and Processes
162 (2004) 23-26.
6. J.C. Williams and E.A. Starke, Progress in structural materials for aerospace systems, Acta.
Mater. 51 (2003) 5775-5799.
7. I. Dul, Application and processing of nickel alloys in the aviation industry, Welding
international 27 (2013) 48-56.
8. Special Metals Corporation Website (2013) Inconel-alloy-625.pdf,
http://www.specialmetals.com/assets/documents /alloys/inconel/inconel-alloy-625.pdf.
(accessed 2013).
9. M.J. Brand, Mechanical properties and microstructure evaluation of powder bed fused
Inconel 625 nickel alloy, Master Thesis, New Mexico Institute of Mining and Technology
(2016).
10. H.A. Roth, C.L. Davis and R.C. Thomson, Modeling solid solution strengtheing in nickel
alloys, Metall. Mater Trans 28A (1997) 1329-1335.
11. H.L Eiselstein, D.J. Tillack, The invention and definition of alloy 625, in: E.A. Loria (Ed.),
Superalloys 718, 625 and Various Derivatives, TMS, Warrendale (1991) 1-14.
12. G.F.V. Voort, J.W. Bowman, R.B. Frank, Microstructural charaterization of custom age
625 plus alloy, in: E.A. Loria (ed.), Superalloys 718, 625 and Various Derivatives, TMS,
Warrendale, PA (1994) 489-498.
13. L.M. Suave, J. Comier, P. Villechaise, A. Soula, Z. Hervier, D. Bertheau and J. Laigo,
Microstructural evolutions during thermal aging of alloy 625: impact of temperature and
forming process, Metall Mater Trans 45A (2014) 2963-2982.
219

14. S. Floreen, G.E. Fuchs, W.J. Yang, The metallurgy of alloy 625, in: E.A. Loria (ed.),
Superalloys 718, 625, 706 and Various Derivatives, TMS, Warrendale, PA (1994) 13-37.
15. J. Teimouri, S.R. Hosseini, K. Farmanesh, Effect of magnesium and calcium addition on
carbides characterizations and anisotropy ductile fracture of inconel 718, Metall. Micro.
Anal. 7 (2018) 268-276.
16. X.C. Chen, C.B. Shi, H.J. Guo, F. Wang, H. Ren, D. Feng, Investigation of oxide inclusions
and primary carbonitrides in inconel 718 superalloy refined through elctroslag remelting
process, Metall Mater. Trans 43B (2012) 1596-1607.
17. J.F. Radavich, A. Fort, Effects of long-time exposure in alloy 625 at 1200 ˚F, 1400 ˚F and
1600 ˚F, in: E.A. Loria (ed.), Superalloys 718, 625 and Various Derivatives, TMS,
Warrendale, PA (1994) 635-647.
18. M.G. Burke, W.J. Mills, R. Bajaj, Microstructure and properties of direct-aged alloy 625,
in: E.A. Loria (ed.), Superalloys 718, 625, 706 and Various Derivatives, TMS, Warrendale,
PA (2001) 389-398.
19. H.M. Tawancy, Thermal stability of an Ni-Cr-Mo-Nb alloy, Mater. Charact. 28 (1992)
221-240.
20. H. Li, S. Xia, B. Zhou, W. Liu, C-Cr segregation at grain boundary before the carbide
nucleation in Alloy 690, Mater. Charact. 66 (2012) 68-74.
21. M. Sundararaman, P. Mukhopadhyay, S. Banerjee, Carbide precipitation in nickel base
superalloys 718 and 625 and their effect on mechanical properties, in: E.A. Loria (Ed.),
Superalloys 718, 625, 706 and Various Derivatives, TMS, Warrendale (1997) 367-378.
22. A.K. Sinha, Growth of MC particles on stacking faults and dislocations, Metallography 20
(1987) 37-45.
23. H.U. Rehman, Solid solution strengthening and diffusiion in nickel- and cobalt-based
superalloys, PhD dissertation, FAU University, 2016.
24. A.K. Jena, M.C. Chaturvedi, The role of alloying elements in the design of nickel-base
superalloys, J. Mater. Sci. 19 (1984) 3121-3139.
25. R.M.N. Pelloux and N.J. Grant, solid solutions and second phase strengthening of nickel
alloys at high and low temperatures, Trans. Met. Soc. AIME 218 (1960) 232-237.
26. P.C.J. Gallagher, The influence of alloying, temperature, and related effects on the stacking
fault energy, Met. Trans.1 (1970) 2429-2461.
220

27. J.M. Oblak, D.S. Duvall, D.F. Paulonis, Coherency strengthening in Ni base alloys
hardened by DO22 γ’’ precipitates, Metall. Trans. 5 (1974) 143-153.
28. M. Sundararaman, J.B. Singh, P. Mukhopadhyay, Estimation of order strengthening in
Inconel 718 type alloys containing all γ’’ precipitate variants, Scripta Metall. 29 (1993)
557-562.
29. M. Sundararaman, P. Mukhopadhyay, S. Banerjee, Deformation behavior of γ’’
strengthened Inconel 718, Acta metall. 36 (1988) 847-864.
30. P.M. Kelly, The effect of particle shape on dispersion hardening, Scripta Metall. 6 (1972)
647-656.
31. D.A. Porter, K.E. Easterling, Phase transformation in metals and alloys, second ed.,
Springer, Heidelberg (1992).
32. J.S. Zhang, High temperature deformation and fracture of materials, first ed., Woodhead,
Cambridge, (2010).
33. L. Kovarik, R.R. Unocic, J. Li, P. Sarosi, C. Shen, Y. Wang, M.J. Mills, Microstwinning
and other shearing mechanisms at intermediate temperatures in Ni-based superalloys, Prog.
Mater. Sci. 54 (2009) 839-973.
34. S.R. Kalidindi, A.A. Salem, R.D. Doherty, Role of deformation twinning on strain
hardening in cubic and hexagonal polycrystalline metals, Adv. Eng. Mater. 5 (2003) 229-
232.
35. R.H. Bricknell, D.A. Woodford, The embrittlement of nickel following high temperature
air exposure, Metall. Trans. 12A (1981) 425-433.
36. J.A. Pfaendtner, C.J. Mcmahon jr., Oxygen-induced intergrnular cracking of a ni-base alloy
at elevated temperatures-an example of dynamic embrittlement, Acta mater. 49 (2001)
3369-3377.
37. T. Denda, P.L. Bretz, J.K. Tien, Inclusion size effect on the fatigue crack propagation
mechanism and fracture mechanics of a Superalloy, Metall. Trans. 23A (1992) 519-526.
38. R.G. Menzies, R.H. Bricknell, A.J. Craven, STEM microanalysis of precipitates and their
nuclei in a nickel-base superalloy, Philos. Mag. 41A (1980) 493-508.
39. J.M. Hyzak, I.M. Bernstein, The effect of defects on the fatigue crack initiation process in
two P/M superalloys: Part Ⅰ. Fatigue Origins, Metall. Trans. 13A (1982) 33-43.
221

40. J.M. Hyzak, I.M. Bernstein, The effect of defects on the fatigue crack initiation process in
two P/M superalloys: Part Ⅱ. Surface-subsurface transition, Metall. Trans. 13A (1962) 45-
52.
41. S. Yamaguchi, H. Kobayashi, T. Matsumiya, S. Hayami, Effect of minor elements on hot
workability of nickel-base superalloys, Met. Technol. 6 (1979) 170-175.
42. R.T. Holt, W. Wallace, Impurities and trace elements in nickel-base superalloys, Inter. Met.
Rev. 21 (1976) 1-24.
43. C.F. Miller, G.W. Simmons, R.P. Wei, High temperature oxidation of Nb, NbC and Ni3Nb
and oxygen enhanced crack growth, Scripta Mater. 42 (2000) 227-232.
44. X.J. Pang, D.J. Dwyer, M. Gao, P. Valerio, R.P. Wei, Surface enrichment and grain
boundary segregation of niobium in inconel 718 single- and poly-crystals, Scripta Metall.
Mater. 31. (1994) 345-350.
45. Y.X. Xu, J.T. Lu, W.Y. Li, X.W. Yang, Oxidation behaviour of Nb-rich Ni-Cr-Fe alloys:
Role and effect of carbides precipitates, Corros. Sci. 140 (2018) 252-259.
46. R. Molins, G. Hochstetter, J.C. Chassaigne, E. Andrieu, Oxidation effects on the fatigue
crack growth behaviour of alloy 718 at high temperature, Acta Mater. 45 (1997) 663-674.
47. J.E. King, Fatigue crack propagation in nickel-base superalloys – effects of microstructure,
load ratio, and temperature, Mat. Sci. Tech. 3 (1987) 750-764.
48. R.H. Bricknell, D.A. Woodford, The mechanism of cavity formation during high
temperature oxidation of nickel, Acta. Metall. 30 (1982) 257-264.
49. D.A. Woodford, Gas phase embrittlement and time dependent cracking of nickel based
superalloys, Energy Mater. 1 (2006) 59-79.
50. F.J. Rizzo, S.B. Justus, PM alloy 625m - a high strength modification of alloy 625, in E.A.
Loria (ed.), Superalloys 718, 625, 718 and Various Derivatives, TMS, Warrendale, PA
(1994) 903-911.
51. G.A. Rao, M. Srinivas, D.S. Sarma, Effect of oxygen content of powder on microstructure
and mechanical properties of hot isostatically pressed superalloy Inconel 718, Mater. Sci.
Eng. 435A (2006) 84-99.
52. K.S. Kim, T.H. Kang, M.E. Kassner, K.T. Son, K.A. Lee, High-temperature, high-cycle
fatigue behavior of Inconel 625 alloy manufactured by selective laser melting, Add. Manu.
35 (2020) 101377.
222

53. J.P. Beckman, D.A. Woodford, Intergranular sulfur attack in nickel and nickel-base alloys,
Superalloys 1988, in: S. Reichman, D.N. Duhl, G. Maurer, S. Antolovich, C. Lund (eds.),
TMS, Warrendale (1988) 795-804.
54. K. Banerjee, The role of magnesium in superalloys-a review, Mater. Sci. Appl. 2 (2011)
1243-1255.
55. L. Zhang, G. Schmitz, Y. Meng, R. Chellali, R. Schlesiger, Mechanism of intermediate
temperature embrittlement of Ni and Ni-based superalloys, Crit. Rev. Solid. State. 37
(2012) 181-214.
56. M. Všianská, M. Šob, The effect of segregated sp-impurities on grain-boundary and surface
structure, magnetism and embrittlement in nickel, Prog. Mater. Sci. 56 (2011) 817-840.
57. R.A. Mulford, Grain-boundary embrittlement of Ni and Ni alloys, Treatise Mater. Sci.
Tech. 25 (1983) 1-19.
58. J.X. Dong, X.S. Xie, R.G. Thompson, The influence of sulfur on stress-rupture fracture in
Inconel 718 superalloys, Metall. Mater. Trans. 31A (2000) 2135-2144.
59. H.H. Kart, M. Uludogan, T. Cagin, DFT studies of sulfur induced stress corrosion cracking
in nickel, Comput. Mater. Sci. 44 (2009) 1236-1242.
60. X. Xie, X. Liu, Y. Lu, B. Tang, Z. Xu, J. Dong, K. Ni, The role of phosphorous and sulfur
in Inconel 718, Superalloys 1996, in: R.D. Kissinger, D.J. Deye, D.L. Anton, A.D. Cetel,
M.V. Nathal, T.M. Pollock, D.A. Woodford (eds.), TMS, Warrendale (1996) 599-606.
61. D.A. Metzler, A gleeble®-based method for ranking the strain-age cracking susceptibility
of Ni-based superalloys, Weld. J. 87 (2008) 249s-256s.
62. D.A. Metzler, Strain-age cracking susceptibility of Ni-based superalloys as a function of
strain rate, temperature, and alloy composition, Weld. J. 91 (2012) 163s-168s.
63. S.L. Cockcroft, T. Degawa, A. Mitchell, D.W. Tripp, A. Schmalz, Inclusion precipitation
in superalloys, in: S.D. Antolovich, R.W. Stusrud, R.A. MacKay, D.L. Anton, T. Khan,
R.D. Kissinger, D.L. Klarstrom (eds.), Superalloys 718, 625, 706 and Various Derivatives,
TMS, Warrendale, PA (1992) 577-586.
64. ASTM F2792 – 00 Standard Terminology for Additive Manufacturing Technologies, DOI:
10.1520/F2792-12A.
65. I. Gibson, D. Rosen and B. Stucker, Additive Manufacturing Technologies, second ed.,
Springer, New York, USA (2012).
223

66. C. Qiu, N.J.E. Adkins, M.M. Attallah, Microstructure and tensile properties of selectively
laser-melted and of HIPed laser-melted Ti–6Al–4V, Mater. Sci. Eng. 578A (2013) 230-
239.
67. E.J. Bae, J.H. Kim, W.C. Kim, H.Y. Kim, Bond and fracture strength of metal-ceramic
restorations formed by selective laser sintering, J. Adv. Prosthodont. 6 (2014) 266-271.
68. R. Hague, Rapid manufacturing: an industrial revolution for the digital age, first ed., John
Wiely and Son, New York, USA (2006).
69. J.P Kruth, M.C. Leu, T. Nakagawa, Progress in Additive Manufacturing and Rapid
Prototyping, CIRP Ann. Manuf. Technol. 47 (1998) 525-540.
70. W.E. Frazier, Metal Additive Manufacturing: A Review, J. Mater. Eng. Perform. 23 (2014)
1917-1928.
71. B. Dutta, F.H.S. Froes, The Additive Manufacturing (AM) of Titanium Alloys, Adv. Mater.
Res. 1019 (2014) 19-25.
72. M. Seifi, A. Salem, J. Beuth, O. Harrysson and J.J. Lewandowski, Overview of Materials
Qualification Needs for Metal Additive Manufacturing, JOM 68 (2016) 747-764.
73. J.J. Lewandowski, M. Seifi, Metal Additive Manufacturing: A Review of Mechanical
Properties, Annu. Rev. Mater. Res. 46 (2016) 151-186.
74. A. Kreitcberg, V. Brailovski, S. Turenne, Effect of heat treatment and hot isostatic pressing
on the microstructure and mechanical properties of Inconel 625 processed by laser powder
bed fusion, Mater. Sci. Eng. 689A (2017) 1-10.
75. P.K. Neghlani, Electron beam melting of Alloy 718: Influence of process parameters on
the microstructure, Master Thesis, University West (2016).
76. T. Kurzynowski, E. Chlebus, B. Kuznicka, J. Reiner, Parameters in selective laser melting
for processing metallic powders, Proc. SPIE, 8239: High Power Laeser Materials
Processing, SPIE, Bellingham, WA (2012) 823914
77. J.A. Cherry, H.M. Davies, S. Mehmood, N.P. Lavery, S.G.R. Brown, J. Sienz,
Investigation into the effect of process parameters on microstructural and physical
properties of 316L stainless steel parts by selective laser melting, Int. J. Adv. Manuf.
Technol. 76 (2015) 869-879.
224

78. D.D. Gu, W. Meiners, K. Wissenbach and R. Poprawe, Laser additive manufacturing of
metallic components: materials, processes and mechanisms, Int. Mater. Rev. 57 (2012)
133-164.
79. B. Song, S. Dong, S. Deng, H. Liao and C. Coddet, Microstructure and tensile properties
of iron parts fabricated by selective laser melting, Opt. Lser Technol. 56 (2014) 451-460.
80. D.D. Gu, W. Meiners, Y.C. Hagedorn, K. Wissenbach, R. Poprawe, Bulk-form TiCx/Ti
nanocomposites with controlled nanostructure prepared by a new method: selective laser
melting, J. Phys. Appl. Phys. 43 (2010) 295402.
81. Q. Jia and D. Gu, Selective laser melting additive manufacturing of Inconel 718 superalloy
parts: Densification, microstructure and properties, J. Alloys. Compd. 585 (2014) 713-721.
82. C. Ding, X. Cui, J. Jiao, P. Zhu, Effects of Substrate Preheating Temperatures on the
Microstructure, Properties, and Residual Stress of 12CrNi2 Prepared by Laser Cladding
Deposition Technique, Materials 11 (2018) 2401.
83. A. Kreitcberg, V. Brailovski, S. Turenne, Elevated temperature mechanical behavior of
IN625 alloy processed by laser powder-bed fusion, Mater. Sci. Eng. 700A (2017) 540-553.
84. L.E. Murr, E. Martinez, S.M. Gaytan, D.A. Ramirez, B.I. Machado, P.W. Shindo, J.L.
Martinez, F. Medina, J. Wooten, D. CISCEL, U. Ackelid, R.B. Wicker, Microstructural
architecture, microstructures, and mechanical properties for a nickel-base superalloy
fabricated by electron beam melting, Metall. Mater. Trans. 42A (2011) 3491-3508.
85. O.L. Deel and H. Mindlin, Engineering data on new aerospace structural materials,
Technical Report AFML-TR-71-249, Air Force Mater. Lab., Wright-Patterson Air Force
Base, Ohio (1971).
86. R. Rodriguez, R.W. Haynes, P.B. Berbon, E.J. Lavernia, Tensile and creep behavior of
cryomilled Inco 625, Acta Mater. 51 (2003) 911-929.
87. Z. Tian, C. Zhang, D. Wang, W. Liu, X. Fang, D. Wellmann, Y. Zhao, Y. Tian, A review
on laser powder bed fusion of Inconel 625 nickel-based alloy, Appl. Sci. 10 (2020) 81-94.
88. K. Amato, J. Hernandez, L. Murr, E. Martinez, S. Gaytan, P. Shindo, S. collins,
Comparison of micrstructures and properties for a ni-base superalloy (alloy 625) fabricated
by electron beam melting, J. Mater. Sci. Res. 1 (2012) 3-41.
89. Q. Yang, Y. Wu, F. Sha, Microstructure and mechnical properties of Inconel 625 alloy
manufactured by selective laser melting, Mater. Mech. Eng. 40 (2016) 83-87.
225

90. M.R. Stoudt, E.A. Lass, D.S. Ng, M.E. Williams, F. Zhang, C.E. Campbell, G. Lindwall,
L.E. Levine, The influence of annealing temperature and time on the formation of δ-phase
in additively manufactured inconel 625, Metall. Mater. Trans. 49A (2018) 3028-3037.
91. E.A. Lass, M.R. Stoudt, M.E. Williams, M.B. Katz, L.E. Levine, T.Q. Phan, T.H.
Gnaeupel-Herold, D.S. Ng, Formation of the Ni3Nb δ-phase in stress-relieved inconel 625
produced via laser powder-bed fusion additive manufacturing, Metall. Mater. Trans. 48A
(2017) 5547-5558.
92. M.E. Kassner, "Fundamentals of creep in metals and alloys" 2
nd
ed., Elservier Ltd.
Amsterdam, The Netharlands (2009).
93. D. McLean, The physics of high temperature creep in metals, Rep. Prog. Phys. 29 (1966)
1-31.
94. W. Blum and P. Eisenlohr, F. Breutinger, Understaning creep - a review, Metall. Mater.
Trans 33A (2002) 291-303.
95. J.E. Bird, A.K. Mukherjee, J.E. Dorn, Quantative Relation Between Properties and
Microstructure, Israel Univ. Press, Jerusalem, (1969) 255-282.
96. A.K. Mukherjee, J.E. Bird, J.E. Dorn, The stress/creep rate behavior of precipitation-
hardened alloys, Trans. Am. Soc. Metals 62 (1969) 155-179.
97. A.K. Mukherjee, High temperature creep, in: R.J. Arsenaut (ed.), Treatise on Mater. Sci.
Technol. 6, Academic Press, New York (1975) 163-224.
98. M.S. Kim, High temperature metallic materials, Inha Univ. Press, Incheon (2003).
99. H.J. Frost and M.F. Ashby, Deformation-mechanism maps : the plasticity and creep of
metals and ceramics, Pergamon press, Oxford, United Kingdom (1982).
100. ASTM B443 – 00 Specification for Nickel-Chromium-Molybdenum-Columbium Alloy
(UNS N06625) and Nickel-Chromium-Molybdenum-Silicon Alloy (UNS N06219) Plate,
Sheet, and Strip, DOI: 10.1520/B0443-00R14.
101. A. Purohit, W.F. Burke, Elevated temperature creep behavior of Inconel alloy 625,
Technical Report CONF-841105—31, Argonne National Lab., Illinois, IL (1984).
102. R. M. Goldhoff, Towards the standardization of time-temperature parameters usage in
elevated temperature data analysis, J. Test. Eval. 2 (1974) 387-424.
226

103. S. Shimizu, K. Satoh, S. Matsuda, H. Murase and J. Fujioka, Creep properties of heat-
resistant superalloys for nuclear plants in helium, J. Soc. Mater. Sci. Jpn. 28 (1979) 379-
385.
104. R.W. Kozar, A. Suzuki, W.W. Milligan, J.J Schirra, M.F. Savage and T.M. Pollock,
Strengthening mechanisms in polycrystalline multimodal nickel-base superalloys, Metall
Mater. Trans. 40A (2009) 1588-1603.
105. M.M. de Oliveira, A.A. Couto, G.F.C. Almeida, D.A.P. Reis, N.B. de Lima, R. Baldan,
Mechanical behavior of Inconel 625 at elevated temperatures, Metals 9 (2019) 301-312.
106. V. Shankar, M. Valsan, K.B.S. Rao, S.L. Manman, Effects of temperature and strain rate
on tensile properties and activation energy for dynamic strain aging in alloy 625, Metall.
Mater. Trans. 35A (2004) 3129-3139.
107. A.A. Beese, Z. Wang, A.D. Stoica, D. Ma, Absence of dynamic strain aging in an
additively manufactured nickel-base superalloy, Nat. Commun. 9 (2018) 2083.
108. S. H. Sun, K. Hagihara, T. Nakano, Effect of scanning strategy on texture formation in Ni-
25 at.%Mo alloys fabricated by selective laser melting, Mater. Design 140 (2018) 307-316.
109. F. Zhang, L. E. Levine, A. J. Allen, C. E. Campbell, E. A. Lass, S. Cheruvathur, M. R.
Stoudt, M. E. Williams, Y. Idell, Homogenization kinetics of a nickel-based superalloy
produced by powder bed fusion laser sintering, Scripta Mater. 131 (2017) 98-102.
110. P. M. Sforza, Theory of Aerospace Propulsion, fisrt ed., Elsevier Ltd., Amsterdam, the
Netheland (2012).
111. B.J. Cantwell, Aircraft and Rocket Propulsion, Standford University,
https://web.stanford.edu/~cantwell/AA103_Course_Material/AA283Aircraft_and_Rocket
_Propulsion_BOOK_BJ_Cantwell.pdf (accessed 2019).
112. N. Yun, H. Jeon, K. Kim, D. Lee and H. Cho, Thermal and creep analysis in a gas turbine
combustion liner, in: L. Trilling, D. Perkins, D.D. Dinoysiou, L. Perlovsky, K. Davey, D.
Langdgrebe, M.A. Marino, D.L. Russell, S.H. Collicott, M. Ceccarelli, J.W. Lund (eds.),
Proceedings of the 4th IASME/WSEAS international conference on Energy &
environment, WSEAS, Stevens Point, WI (2009) 315-320.
113. S.B. Lattime, B.M. Steinetz, Turbine Engine Clearance Control Systems: Current Practices
and Future Directions, Technical Report TM-2002-211794, NASA, Cleveland, OH (2002).
227

114. E.C.P. Batalha, Aircraft engines maintenance costs and reliability: an appraisal of the
decision process to remove an engine for a shop visit aiming at minimum maintenance unit
cost, Master Thesis, Universidade Nova de Lisboa (2012).
115. J.F. Radavich, A. Fort, Effects of long-time exposure in alloy 625 at 1200 ˚F, 1400 ˚F and
1600 ˚F, in: E.A. Loria (ed.), Superalloys 718, 625, 706 and Various Derivatives, TMS,
Warrendale (1994) 635-647.
116. Radio Technical Commission for Aeronautics (RTCA), DO-160G section 4: Temperature
and Altitude: SC-135 (2010) http://my.rtca.org.
117. Hopper editors, Hopper, Aug. 2016, Retrieved from https://media.hopper.com/articles/the-
worlds-20-longest-non-stop-flights.
118. ASTM E8 – 16a Standard Test Methods for Tension Testing of Metallic Materials, DOI:
10.1520/E0008_E0008M-16A.
119. J.W. Olesik, Elemental analysis using ICP-OES and ICP/MS, Anal. Chem. 63 (1991) 12-
21.
120. M.P. Mogorosi, The optimization and calibration of spark-optical emission spectroscopy
for the analysis of trace impurities in ultra-pure Pt, Pd and Rh, PhD Dissertation, University
of the Witwatersrand (2013).
121. L.A. Jackman, M.D. Boldy, A.L. Coffey, The influence of reduced carbon on alloy 718,in:
E.A. Loria (ed.), Superalloys 718, 625, 706 and Various Derivatives, TMS, Warrendale,
PA (1991) 261-270.
122. Eltra GmbH Website, https://www.eltra.com/dltmp/www/5b75b08b-c2b4-44cc-85b5-
9c12c3c9c754-3c1d312a7caf/brochure_ELEMENTRAC_CSi_en.pdf, (accessed 2018)
123. Leco Co. Website, Inorgarnic appiclation note, http://www.leco.co.za/wp-
content/uploads/2012/02/TC600_IRON_STEEL_NICKEL_COBALT_ALLOYS_203-
821-184.pdf, (accessed 2007)
124. W. Zhou, R. Apkarian, Z.L. Wang, D. Joy, Fundamentals of Scanning Electron
Microscopy", Scanning Microscopy for Nanotechnology, Springer, New York, USA
(2006).
125. Jeol Co. website, A guide to scanning microscope observation,
https://www.jeol.co.jp/en/applications/pdf/sm/844_en.pdf (accessed 2020).
228

126. EDAXinsight, The importance of the orientation precision performance of electron
backscatter diffraction EBSD, https://www.edax.com/-
/media/ametekedax/files/resources/tips_tricks/theimportanceoftheorientationprecisionperf
ormanceofebsd.pdf (accessed 2020)
127. H. Mansour, J. Goulden, P. Trimby, A. Bewick, Symmertry CMOS-New detector
technology for EBSD, in: M. Boutaous, J. Chevalier, C. Le Bourlot (eds.), Journées
Annuelles SF2M 2017, (2017).
128. P. Goodhew, Aberration-Corrected Analytical Transmission Electron Microscopy, John
Wiley & Sons, New York, USA (2012).
129. D.B. Williams and C.B. Carter, Transmission Electron Microscopy: A Textbook for
Materials Science, Springer, New York, USA (1996).
130. S. Graça, R. Colaço, P.A. Carvalho, R. Vilar, Determination of dislocation density from
hardness measurements in metals, Mater. Lett. 62 (2008) 3812-3814.
131. J.I. Goldstein, D.E. Nebury, J.R. Michael, N.W.M. Ritchie, J.H.J. Scott and D.C. Joy,
Scanning Electron Microscopy and X-Ray Microanlaysis, Springer, New York, USA
(2018)
132. JEOL Ltd. Website, JXA-8230 electron probe microanalyzer,
http://www.avba.co.il/image/users/390742/ftp/my_files/JEOL%20JXA-
8230.pdf?id=29024654, (accessed 2009)
133. P. Hoppe, S. Cohen, A. Meibom, NanoSIMS: technical aspects and applications in
cosmochemistry and biological geochemistry, Geostand. Geoanal. Res. 37 (2013) 111-154.
134. G. Van Tendeloo, D. Van Dyck, S.J. Pennycook, Handbook of Nanoscopy, Ch. 23 Atom
probe tomography: principle and applications, Wiley-VCH, Weinhein, Germany (2012)
793-832.
135. T.F. Kelly, M.K. Miller, Atom probe tomography, Rev. Sci. Instruments 78 (2007) 031101.
136. S. Azadian, L. Wei, R. Warren, Delta phase precipitation in inconel 718, Mater. Chract. 53
(2004) 7-16.
137. H. Yuan, W.C. Liu, Effect of the δ phase on the hot deformation behavior of Inconel 718,
Mater. Sci. Eng. 408A (2005) 281-289.
138. V. Shankar, K.B.S. Rao, S.L. Mannan, Microstructure and mechanical properties of
Inconel IN625 superalloy, J. Nucl. Mater. 288 (2001) 222-232.
229

139. L.M. Suave, D. Bertheau, J. Cormier, P. Villechaise, A. Soula, Z. Hervier, J. Laigo, Impact
of microstructural evolutions during thermal aging of Alloy IN625 on its monotonic
mechanical properties, MATEC Web of Conferences 14 (2014) 21001-p.1-p.6.
140. R. Langerborg, A modified recovery-creep model and its evaluation, Met. Sci. J. 6 (1972)
127-133.
141. R. Langneborg, Bypassing of dislocations past particles by a climb mechanism, Scripta
Met. 7 (1973) 605-613.
142. A. Bauer, S. Neumeier, F. Pyczak, Mathias Göken, Creep strength and microstructure of
polycrystalline γ’-strengthened cobalt-base superalloys, in: E.S. Huron, R.C. Reed, M.V.
Hardy, M.J. Mills, R.E. Montero, P.D. Portella, J. Telesman (eds.), Superalloys 2012: 12
th

International Symposium on Superalloys, TMS, Warrendale (2012) 695-703.
143. B. Reppich, M. Heilmaier, K. Liebig, G. Schumann, K. Stein, T. Woller, Microstructural
modelling of the creep behavior of particle-strengthened superalloys, Steel Res. Int. 61
(1990) 251-257.
144. P.W. Davies, J.P. Dennison, The use of heat-treatment to recover the creep properties of
Nimonic 115 after high-temperature creep, Met. Sci. 9 (1975) 319-323.
145. V.C. Nardone, J.K. Tien, On the creep rate stress dependence of particle strengthened
alloys, Scripta Metall. 20 (1986) 797-802.
146. T.W. Ni, J.X. Dong, Creep behaviors and mechanisms of Inconel718 and Allvac718plus,
Mater. Sci. Eng. 700A (2017) 406-415.
147. F. Garazarolli, K.P. Francke, J. Fischer, Influence of neutron irradiation on tensile and
stress rupture properties of Inconel IN625, J. Nucl. Mater. 39 (1969) 242-246.
148. A. Manonukal, F.P.E. Dunne, D. Knowles, Physically-based model for creep in nickel-
base superalloy C263 both above and below the gamma solvus, Acta Mater. 50 (2002)
2917-2931.
149. O.D. Sherby, E.M. Taleff, Influence of grain size, solute atoms and second-phase particles
on creep behavior of polycrystalline solids, Mater. Sci. Eng. 332A (2002) 89-99.
150. G. Venkataraman, F. Cosandey, Creep behavior of Ni-Cr alloys with trace additions of
cerium, Mater. Sci. Eng. 93 (1987) 175-179.
151. M.E. Kassner, Application of the Taylor equation to five-power-law creep considering the
influence of solutes, Metals 8 (2018) 813-816.
230

152. I.J. Moore, I.J. Taylor, M.W. Tracy, M.G. Burke, E.J. Palmiere, Grain coarsening behavior
of solution annealed alloy 625 between 600-800 ˚C, Mater. Sci. Eng. 682A (2017) 402-
409.
153. T. Watanabe, Grain boundary design and cotrol for high tempertature materials, Mater. Sci.
Eng. 166A (1993) 11-28.
154. T.H. Lee, Y.J. Lee, S.H. Joo, H.H. Nersisyan, K.T. Park, J.H. Lee, Intergranular M23C6
carbide precipitation behavior and its effect on mechanical properties of Inconel 690 tubes,
Metall. Mater. Trans. 46A (2015) 4020-4026.
155. S. Guo, W. Sun, D. Lu, Z. Hu, Effect of minor elements on microstructure and mechanical
properties of In 718 alloy, Superalloys 718, IN625, 706 and Various Derivatives, TMS,
Warrendale (1997) 521-530.
156. Z. Gao, G. Zhang, Z. Li, H. Yuan, W. Xu and Y. Zhang, Mater Sci Forum 2013, 747, 518\
157. R. Soundararajan, Behavior of TiN inclusions and their influenmce in random grain
formation in Ni-based superalloys, Master Thesis, The University of British Columbia
(1998).
158. X. Xing, X. Di, B. Wang, The effect of post-weld heat treatment temperature on the
microstructure on Inconel 625 deposited metal, J. Alloy. Comp. 593 (2014) 110-116.
159. H. Gan, M. Cheng, Z. Guo, Q. Tai, G. Kozhevnikova, H. Song, S. Zhang, Spheroidization
mechanism of lamellar δ phase in Inconel 718 during delta processing, J. Mater. Eng.
Perform. 29 (2020) 3852-3857.
160. T.M. Smith, R.R. Unocic, H. Deutchman, M.J. Mills, Creep deformation mechanism
mapping in nickel base disk superalloys, Mater. High. Temp. 33 (2012) 1-12.
161. D. Barba, E. Alabort, S. Pedrazzini, D.M. Collins, A.J. Wilkinson, P.A.J. Bagot, M.P.
Moody, C. Atkinson, A. Jérusalem, R.C. Reed, On the microtwinning mechanism in a
single crystal superalloy, Acta Mater. 135 (2017) 314-329.
162. G.B. Viswananthan, P.M. Sarosi, M.F. Henry, D.D. Whitis. W.W. Milligan, M.J. Mills,
Investigation of creep deformation mechanisms at intermediate temperatures in René 88
DT, Acta. Mater. 53 (2005) 3041-3057.
163. M. Sundararaman, P. Mukhopadhyay and S. Banerjee, Precipitation of the δ-Ni3Nb phase
in two nickel base superalloys, Metall Trans. 19A. (1988) 453-465.
231

164. H.Y. Zhang, S.H. Zhang, M, Cheng and Z.X. Li, Deformation characteristics of δ phase in
the delta-processed Inconel 718 alloy, Mater. Charact. 61 (2010) 49-53.
165. D.D. Keiser, H.L. Brown, Review of the physical metallurgy of Alloy 718, Technical
Report. E(10-1)-1375, Idaho National Engieering Laboratory, Idaho Falls, ID (1976).
166. Y.C. Lin, J. Deng, Y.Q. Jiang, D. Wen and G. Liu, Effects of initial δ phase on hot tensile
deformation behaviors and fracture characteristics of a typical Ni-based superalloy, Mater.
Sci. Eng 598A. (2014) 251-262.
167. M. Köhler, Effect of the elevated-temperature-precipitation in alloy IN625 on properties
and microstructure, in E.A. Loria (ed.), Superalloys 718, IN625 and Various Derivatives,
TMS, Warrendale (1991) 363-374.
168. G.W. Lorimer, Precipitate Nucleation, Mater. Sci. Forum 3 (1985) 199-210.
169. M.J. Anderson, J. Benson, J.W. Brooks, B. Saunders, H.C. Basoalto, Predicting
precipitation kinetics during the annealing of additive manufactured Inconel 625
components, Integrating Materials and Manufacturing Innovation 8 (2019) 154-166.
170. D. Connétable B. Ter-Ovanessian. E. Andrieu, J. Physic.: Condens. Matter 24 (2012)
095010.
171. P. Xiao-lin, Y. Hai-yan, T.U. Gan-feng, S. Wen-ru, H.U. Zhuang-qi, Segregation and
diffusion behavior of niobium in a highly alloyed nickel-base superalloy, Trans.
Nonferrous Met. Soc. China 21 (2011) 2402-2407.
172. Yi-Qiang Mu, Chang-Shuai Wang, Wen-Long Zhou, Lan-Zhang Zhou, The properties of
cast alloy IN625 in relation to d phase precipitation, Acta. Metall. Sin. 32 (2019) 535-540.
173. V. Shankar, M. Balsan, K.B.S. Rao, S.L. Mannan, Room temperature tensile behavior of
service exposed and thermally aged service exposed alloy 625, Scripta mater. 44 (2001)
2703-2711.
174. J.F. Knott, Fracture resistance in aluminium, in: R. Lumley (ed.), Fundamentals of
Aluminium Metallurgy, first ed., Woodhead Publishing, Cambridge, 2011.
175. M. Bashir, R. Kannan, R. Sandhya, G.A. Harmain, Low cyclic fatigue behavior of alloy
IN625 at ambient and elevated temperatures, in: R. Prakash, R. Suresh Kumar, A. Nagesha,
G. Sasikala, A. Bhaduri (eds.), Structural Integrity Assessment, Springer, Singapore (2020)
653-661.
232

176. C.A. Schuh, R.W. Minich, M. Kumar, Connectivity and percolation in simulated grain-
boundary networks, Philos. Mag. 83 (2003) 711-726.
177. E.M. Lehockey, G. Palumbo, On the creep behavior of grain boundary engineered nickel,
Mater. Sci. Eng. 237A (1997) 168-172.
178. B. Li, S. Tin, The role of deformation temperature and strain on grain boundary engineering
of Inconel 600, Mater. Sci. Eng. 603A (2014) 104-113.
179. F. Yang, J. Hou, S. Gao, C. Wang, L. Zhou, The effects of boron addition on the
microstructure stability and mechanical properties of a Ni-Cr based superalloy, Mater. Sci.
Eng. 715A (2018) 126-136.
180. P. Cacerras and B. Ralph, G.C. Allen, R.K. Wild, Grain boundary segregation in Inconel
600, Surf. Interf. Anal. 12 (1988) 191-197.
181. T.J. Garosshen, T.D. Tillman, G.P. McCarthy, Effects of B, C, and Zr on the structure and
properties of a P/M nickel base superalloy, Metall. Trans. 18A (1987) 69-77.
182. R. Hu, G. Bai, J. Li, J. Zhang, T. Zhang, H. Fu, Precipitation of grain boundary M23C6
and its effect on tensile properties of Ni-Cr-W based superalloy, Mater. Sci. Eng. 548A
(2012) 83-88.
183. M. Wang, J. Du, Q. Deng, Z. Tian, J. Zhu, The effect of phosphorous on the microstructure
and mechanical properties of ATI 718Plus alloy, Mater. Sci. Eng. 626A (2015) 382-389.
184. G.H. Meier, F.S. Pettit, J.L. Smialek, The effects of reactive element additions and sulfur
removal on the adherence of alumina to Ni- and Fe-base alloys, Mater. Corros. 46 (1995)
232-240.
185. M. Allart, F. Christien, R. Le Gall, P. Nowakowski, C.R.M. Grovenor, A multi-technique
investigation of sulfur grain boundary segregation in nickel, Scripta Mater. 68 (2013) 793-
796.
186. G. Palumbo, K.T. Aust, Structure-dependence of intergranular corrosion in high purity
nickel, Acta Metall. 38 (1990) 2343-2352.
187. S. Kobayashi, S. Tsurekawa, T. watanabe, G. Palumbo, Grain boundary engineering for
control of sulfur segregation-induced embrittlement in ultrafine-grained nickel, Scripta
Mater. 62 (2010) 294-297.
188. R. Haynes, A study on the effect of porosity content on the ductility of sintered metals,
Powder Metall. 20 (1977) 17-20.
233

189. E.M. Dubensky, D.A. Koss, Void/pore distributions and ductile fracture, Metall. Trans.
18A (1987) 1887-1895.
190. G.A. Rao, M. Kumar, M. Srinivas, D.S. Sarma, Effect of standard heat treatment on the
microstructue and mechanical properties of hot isostatically pressed superalloy inconel
718, Mater. Sci. Eng. 335A (2003) 114-125.
191. I. Yadroitsev, M. Pavlov, Ph. Bertrand, I. Smurov, Mechanical properties of sample
fabricated by selective laser melting, Proc. 14th Eur. Conf. on ‘Rapid prototyping and
manufacturing' Paris, France (2009) Available at:
https://www.researchgate.net/profile/I_Smurov/publication/268183323_Mechanical_prop
erties_of_samples_fabricated_by_selective_laser_melting/links/548ec6820cf225bf66a62
aa9/Mechanical-properties-of-samples-fabricated-by-selective-laser-melting.pdf
192. A.K. Singh, N. Louat, K. Sadananda, Dislocation network formation and coherency loss
around gamma-prime precipitates in a nickel-base superalloy, Metall. Mater. Trans. 19A
(1988) 2965-1973.
193. C. Slama, C. Servant, G. Cizeron, Aging of the Inconel 718 alloy between 500 and 750 ˚C,
J. Mater. Res. 12 (1997) 2298-2316.
194. R, Cozar, A. Pineau, Morphology of y’ and y” precipitates and thermal stability of inconel
718 type alloys, Metall. Mater. Trans 4B (1973) 47-59.
195. Y. Ji, Y. Lou, M. Qu, J.D. Rowatt, F. Zhang, T.W. Simpson, L. Chen, Predicting coherency
loss of γ'' precipitates in IN718 superalloy, Metall. Mater. Trans. 47A (2016) 3235-3247.
196. R.Y. Zhang, H.L. Qin, Z.N. Bi, J. Li, S. Paul, T.L. Lee, S.Y. Zhang, J. Zhang, H.B. Dong,
Evolution of lattice spacing of gamma double prime precipitates during aging of
polycrystalline Ni-base superalloys: an in situ investigation, Metall. Mater. Trans. 51A
(2020) 574-585.
197. G.C. Weatherly, R.B. Nicholson, An electron microscope investigation of the interfacial
structure of semi-coherent precipitates, Philos. Mag. 17 (1968) 801-831.
198. J.Y. Tsao, Materials fundamentals of molecular beam epitaxy, first ed., Academic Press,
San Diego (1993).

Abstract (if available)

Abstract Inconel (IN625) is a type of Ni-based superalloys which has been widely used in the aerospace field due to its high tensile, creep, and rupture strength. Service temperatures of this alloy range from cryogenic to approximately 1000 ℃ thanks to its outstanding fatigue and thermal-fatigue strength with excellent oxidation resistance. ❧ Additive Manufacturing (AM) is the formalized term for what is popularly called 3D printing. AM has gained attention in the aerospace industry as it provides the possibility for economical and complicated-configuration engine parts with fewer joining steps, and greater geometric freedom. Among various AM technologies, laser powder bed fusion (LPBF) has been widely chosen for metal part fabrication. Metal powder-particles are fully melted and fused by a high energy beam. Subsequent hot isostatic pressing (HIP) is typically conducted to reduce porosity after the LPBF. ❧ The main interest of this study is to verify that AM IN625 can replace the conventional IN625 for high temperature applications. High temperature creep tests of additively manufactured (AM) nickel-based superalloy 625 (IN625) and wrought IN625 were conducted at 650 ℃ and 800 ℃ over the stress range of 65 MPa to 658 MPa. Thermal treatments were conducted for both AM and wrought IN625 samples prior to creep testing: either solution heat-treated or hot isostatically pressed and, additionally, long-term cyclic heat-treatments (LHT) at 650 ℃ for 6 months and 1 year. The purpose of long-term cyclic heat-treatments was to simulate a near-operational environment of one real application of IN625 over a target worktime (= 1 year). ❧ AM IN625 showed equal or even higher creep strength than wrought IN625 for all creep tests. However, AM IN625 exhibited poor ductility compared to wrought IN625 under all creep testing conditions, and the ductility of AM IN625 additionally decreased after the LHT. Some creep ductility recovery in HIPed AM IN625 was observed, however, it was still much inferior to the ductility of wrought IN625. Microstructural observations revealed the intergranular precipitation rate and types of the precipitates of AM IN625 was generally comparable with that of wrought IN625, so that the intergranular precipitation cannot explain the poor ductility of AM IN625. ❧ Both AM and wrought IN625 obtained some additional strength after the LHT. The amount of extra strength in the alloys was generally proportional to the increase in matrix volume fraction of γ’’ phase (650 ℃) and δ phase (800 ℃). ❧ The creep analysis suggests that dislocation climb is the rate controlling mechanism for creep of both AM IN625 and wrought IN625 in the temperature range of 650 ℃ to 800 ℃. ❧ Fractography observations and electron backscatter diffraction analysis revealed that brittle intergranular fracture was evident in AM IN625 samples. Atomic probe tomography (APT) and nano-Secondary ion mass spectrometer (Nano-SIMS) indicated no remarkable intergranular impurity (particularly sulfur and oxygen) segregation at grain boundaries of AM IN625 alloys in both as-HIPed and the 6 months LHT states. However, the Nano-SIMS results clarified that sulfur segregated to the Al₂O₃ /matrix interfaces the AM IN625, and the sulfur segregation level at Al₂O₃ interface increased after the 6 months LHT. ❧ Cracking can occur from these interfaces and repeated sulfur diffusion to the crack tip is the foremost possibility to explain the poor ductility of AM IN625 within the temperature range tested.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Long range internal stresses in cyclically deformed copper single crystals; and, Understanding large-strain softening of aluminum in shear at elevated temperatures

PDF

Mechanical behavior of materials in extreme conditions: a focus on creep plasticity

PDF

Mechanical behavior and deformation of nanotwinned metallic alloys

PDF

Processing and properties of phenylethynyl-terminated PMDA-type asymmetric polyimide and composites

PDF

Properties of a ZK60 magnesium alloy processed by severe plastic deformation

Asset Metadata

Creator Son, Kwangtae (author)

Core Title High temperature creep behaviors of additively manufactured IN625

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Publication Date 01/28/2021

Defense Date 12/04/2020

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag creep,embrittlement,impurity segregation,laser powder bed fusion,nickel-based superalloy,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Kassner, Michael Ernest (committee chair), Goo, Edward Kwock (committee member), Grunenfelder, Lessa Kay (committee member), Uranga, Alejandra (committee member)

Creator Email kwangtas@usc.edu,skwangtae@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-416358

Unique identifier UC11673125

Identifier etd-SonKwangta-9246.pdf (filename),usctheses-c89-416358 (legacy record id)

Legacy Identifier etd-SonKwangta-9246.pdf

Dmrecord 416358

Document Type Dissertation

Format application/pdf (imt)

Rights Son, Kwangtae

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA