Compositional Optimisation of a Ni-based
Superalloy for Additive Repair
Applications
Jonathon F. S. Markanday
Queens’ College
University of Cambridge
Department of Materials Science & Metallurgy
The dissertation is submitted for the degree of
Doctor of Philosophy
April 2022
Declaration
This dissertation is submitted for the degree of Doctor of Philosophy at the University of Cambridge. The work described
herein was carried out in the Department of Materials Science, the Department of Physics and the Department of Earth
Sciences between September 2017 and March 2022 under the supervision of Prof. H. J. Stone. Except where reference is
made to the work of others, the content of this thesis is entirely original. No part of the work described here has been or is
currently being submitted for the purpose of gaining academic qualification at this or any other institute of higher learning.
This dissertation does not exceed 60,000 words in length.
The following manuscripts been accepted:
Markanday, J.F.S., Carpenter, M.A., Jones, N.G., Thompson, R.P., Rhodes, S.E., Heason, C.P., and Stone, H.J., 2021.
Occurrence of a brass texture and elastic anisotropy in laser blown powder processed superalloy IN718. Materials Science and Engineering:
A. 825: p. 141781.
Markanday, J.F.S., Carpenter, M.A., Jones, N.G., Thompson, R.P., Rhodes, S.E., Heason, C.P., and Stone, H.J., 2021.
Research data supporting ”Occurrence of a Brass Texture and Elastic Anisotropy in Laser Blown Powder Processed Superalloy IN718”.
Data in Brief. 39: p. 107570.
The following manuscripts have been submitted:
Markanday, J.F.S, 2022, Overview: Alloy Design and Applications to Additive Manufacturing of Ni-based Alloys. Materials Science and
Technology. Submitted
Markanday, J.F.S., Carpenter, M.A., Jones, N.G., Thompson, R.P., Christofidou, K.A., Fairclough, S.M., Heason, C.P.,
and Stone, H.J., 2021. Effect of NbC Inoculants on the Elastic Properties and Microstructure of Additively Manufactured IN718. Additive
Manufacturing. Submitted
Markanday, J.F.S., Carpenter, M.A., Jones, N.G., Thompson, R.P., Christofidou, K.A., Fairclough, S.M., Heason, C.P.,
and Stone, H.J., 2021. Research data supporting ”Effect of NbC Inoculants on the Elastic Properties and Microstructure of Additively
Manufactured IN718”. Additive Manufacturing. Submitted
Markanday, J.F.S., Conduit, G.A., Conduit, B.D., Christofidou, K.A., Chechik, L., Baxter, G.J., Heason, C.P., and Stone,
H.J., 2021. Design of a Ni-based Superalloy for Laser Repair Applications using Probabilistic Neural Network Identification. Data-Centric
Engineering. Submitted
Jonathon Frederick Shiv Markanday
April 2022
2
Acknowledgements
Firstly, I would like to thank the EPSRC and Rolls-Royce plc. for funding my PhD and the Department of Materials
Science and Metallurgy for providing the laboratories and facilities used for this work.
I would like to give special thanks to Howard Stone for being an exceptional supervisor over the past four years.
He has been a constant source of inspiration and has supported my pursuits both academic and sporting.
I would like to thank Kathy Christofidou for her encouragement and support with my work, especially at the beginning
of my PhD.
I especially thank Dr. H.T. Pang for performing the DSC and TGA measurements on all of my samples, spending
hours with me on TEM attemping to get SADPs and working with me on determining the true texture of my samples.
He has kept me on my toes in the lab and I owe him quite a few pints.
Further thanks goes to Sue Rhodes for arc melting my samples and working tirelessly with me on the preparation
and measurement of my RUS samples.
A massive thanks goes to the Rolls-Royce UTC group for their support, it has been an amazing team to be part of
during my PhD.
Special thanks to my family for all of love and encouragement throughout my education. My mum and Chris have
been my source of discipline and drive. My grandmother for years of love and support. My father for being a source
of fun and for writing support. Joan and Michael for far too many things to be listed, I would not have made it to
Cambridge or through this thesis without them.
Thanks to Queens’ College and the QCBC for being an excellent source of fun and positive distraction. Thank you
for having me as your Woodville Steward and Men’s Captain.
Enormous thanks to my club, the CULRC (now the Cambridge University Boat Club) for an unforgettable experi-
ence during my PhD. The coaches and my fellow trialist have been my family in Cambridge and have had a profound
impact on my life.
Finally, an extra special thanks to my girlfriend Hannah. She has done everything for me, literally everything, over
the last four year and has sacrificed so much to not only get me through my PhD, but also through a pandemic,
numerous injuries and two boatraces.
3
Contents
Contents 4
List of Figures 7
List of Tables 13
1 Introduction 1
2 Literature Review 4
2.1 Ni-Based Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Alloy Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 Additive Manufacturing of Ni-based Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.1 Directed Energy Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Powder Bed Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.3 Effect of Process Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Microstructure and Mechanical Properties of AM Ni-based Alloys . . . . . . . . . . . . . . . . . . . . . . 22
2.4.1 Microstructural Morphology and Anisotropy of AM Ni-Based Alloys . . . . . . . . . . . . . . . . 23
2.4.2 Elemental Segregation and Post-Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.4.3 Cracking of AM Ni-based alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 Alloy Design For Additive Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.5.1 Alloy Design for Crack Mitigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.5.2 Control of Residual Stress and Anisotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.5.3 Application of Inoculants for Microstructural Control in Additive Manufacturing . . . . . . . . . 39
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4
3 Texture and Elastic Anisotropy in Laser Blown Powder Processed Superalloy IN718 42
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.2 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.1 Compositional and Microstructural Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.2 Texture Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3.3 Anisotropic Elastic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4 Modification of Superalloy IN718 for Additive Repair Applications Through Inoculant Addition 66
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.2 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3.1 Compositional and Microstructural Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3.2 Texture and Elastic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5 Optimisation of Superalloy IN718 for Laser Blown Powder Repair Applications 92
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.2.1 Target Alloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.2.2 Neural Network Formalism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.2.3 Model Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.2.4 Optimisation Routine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.2.5 Alloy Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.3.1 Alloy Fabrication and Laser Pass Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.3.2 Microstructural and Phase Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.3.3 Physical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5
6 Conclusions and Future Work 118
6.1 LBP-DED of IN718 for Additive Repair Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.2 Benefits of Inoculant Additions for the LBP-DED of Ni-based Superalloys . . . . . . . . . . . . . . . . . 119
6.3 Applications of Alloy Design for the Optimisation of Ni-based Superalloys for LBP-DED Applications . 120
6.4 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.4.1 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.4.2 Design of in situ Composite Superalloys for LBP-DED Repair Applications . . . . . . . . . . . . 123
Bibliography 129
6
List of Figures
1.1 Large weight reductions can be achieved through the manufacturing of a Blisk, courtesy of Rolls-Royce
plc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1 SEM Images of selected phases that occur in Ni-based alloys. (a) As-deposited DED IN718 with Laves
eutectic identified along sub-grain boundaries. (b) DED IN718 solutioned at 1100 ◦C for 2 hours, with
spherical MC-type carbides identified. (c) Forged IN718 with rod-like delta precipitates along the grain
boundaries. (d) Image from Wilson et al. [45] highlighting the occurrence of TCP phases and carbides
in polycrystalline RR1000 following a 5000-hour heat-treatment at 800 ◦C. . . . . . . . . . . . . . . . . 5
2.2 A materials system flowchart for the design of blast resistant naval hull steels, adapted from Saha and
Olson [61] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 A graph, adapted from Saha and Olson [36], representing the expected contributions of the individual
strengthening mechanisms to achieve the targeted strength goal, equivalent to 389 VHN. . . . . . . . . . 8
2.4 Schematic representation of a neural network used by Conduit et al. [53]. The schematic describes the
relationship between two selected output values (y1 and y2). Given properties (red) are used for the
calculation of the hidden nodes (blue) which yield the predicted properties (green). . . . . . . . . . . . . 10
2.5 Property trade-off diagrams for Ni-based superalloys as presented by Reed, Tao and Warnken [75] (a)
density vs. stability number (Md) (b) density vs. cost, with the change in Re content identified (c) cost
vs. stability number, with regions of Re content and change in Ta and Al content identified. . . . . . . . 12
2.6 A compilation process map for the DED of numerous alloy systems presented by Dass and Moridi [95].
The plot identifies process parameter zones that yield desirable deposition condition, as well as zones
that can lead to complications including keyholing and lack of fusion. . . . . . . . . . . . . . . . . . . . 14
2.7 Directed Energy Deposition Process Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.8 Powder Bed Fusion Process Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
7
2.9 Representations for selected scan strategies (a) Linear (b) Bi-linear (c) Off-set (d) Checkboard Island. . 18
2.10 EBSD IPF-Z map of as-deposited DED IN718 showing the growth direction of columnar grain. The LBP-
DED IN718 sample was fabricated with a bi-linear scan strategy which caused the heat-flow direction to
be altered. The build direction on this image has been indicated. The IN718 sample used in this analysis
was deposited via LBP-DED onto a forged IN718 substrate using standard deposition parameters. . . . 20
2.11 Examples of melt pool morphologies resulting from different process parameters, taken from Scime and
Beuth [94]. (a) desirable (285W, 960mm/s), (b) balling (370W, 1200mm/s), (c) under-melting (100W,
1000mm/s), (d) severe keyholing (250W, 400mm/s), and (e) keyholing porosity (150W, 200mm/s). . . 21
2.12 SEM images of IN718 demonstrating the different morphologies that develop during AM deposition (a)
lower magnification image indicating the (C) columnar zone, (E) equiaxed zone and (S) Substrate zone.
(b) A higher magnification image of the equiaxed zone in (a). The IN718 sample used in this analysis
was deposited using LBP-DED onto a forged IN718 substrate using standard deposition parameters. . . 24
2.13 The measured directional elastic moduli from an As-SLM and a heat treated sample taken from Mun˜oz-
Moreno et al. [31] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.14 SEM analysis of LBP-DED IN718, a backscattered electron image is given in the top left. This is
supported by the Mo, Nb, and Ti elemental distribution maps determined by SEM-EDX. . . . . . . . . 26
2.15 Refinement of the Laves phase distribution in AM IN718 as presented by Chen et al.[43]. The bottom
(a), middle (b) and top (c) regions for the flat-top laser fabricated sample. The bottom (d), middle (e)
and top (f) regions for the Gaussian laser fabricated sample. . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.16 SEM analysis of LBP-DED IN718 deposited onto a forged IN718 substrate, demonstrating the grain
growth as a result of a solution heat treatment. (a) As-deposited condition, (b) heat-treated condition
(1100 ◦C / 2 hours). The IN718 sample used in this analysis was deposited using (LBP-DED) onto
forged IN718 substrate using standard deposition parameters. . . . . . . . . . . . . . . . . . . . . . . . . 28
2.17 A weldability assessment chart adapted from the work of Haafkens and Matthey [147] and Basak and
Das [146] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.18 Examples of liquation and solidification cracking in air-cooled and water-cooled samples of laser-AM
IN718 as presented by Chen et al. [100]. (a) Liquation cracking of the water-cooled sample, (b) liquation
cracking of the air-cooled sample and (c) solidification cracking of the air-cooled sample. . . . . . . . . . 31
2.19 Examples of solid-state cracking in the AM of superalloys CM247LC and IN939, as presented by Tang
et al. [76]. The microstructures were presented in the as-deposited condition in all cases. Features and
region of interest have been identified. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
8
2.20 Crack densities of two variants of Hastelloy X (original and modified Hastelloy X, OHX and MHX
respectively) as presented by Harrison, Todd and Mumtaz [105]. The top graph displays crack densities
in the horizontal build orientation. The bottom graph displays crack densities in the vertical build
orientation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.1 Composite SEM images of the microstructure of LBP-DED IN718 in the planes normal to the build
direction (BD), scanning direction (SD) and transverse direction (TD). The images are presented in a
parallelepiped representation to show the orientational dependence of the microstructural sections . . . 47
3.2 (a) XRD pattern of the As-DED IN718 sample. (b) magnified view of the XRD pattern in the vicinity
of the arrow shown in (a). The reflections from the γ phase, MC carbide and Laves (φ)phases are labelled. 48
3.3 ((a) XRD pattern of the IN718 sample in the heat treatment F condition. (b) magnified view of the XRD
pattern in the vicinity of the arrow shown in (a). The reflections from the gamma phase, MC carbide
and Laves phases are labelled. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.4 SEM analysis of DED IN718 samples in different conditions As-DED, A and B. For each sample, a
backscattered electron image is shown at the top, below which are the Cr, Mo, Nb, and Ti elemental
distribution maps determined by SEM EDX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.5 SEM analysis of DED IN718 samples in different heat-treated states C, D, and F. For each sample, a
backscattered electron image is shown at the top, below which are the Cr, Mo, Nb, and Ti elemental
distribution maps determined by SEM EDX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.6 Left – Inverse pole figure maps with respect to the build direction (BD) and scanning direction (SD)
for LBP-DED RUS samples in the As-DED state and following heat treatments A, B, and C. Right –
Corresponding {001}, {011} and {111} pole figures in the BD plane. The idealised spots for the {011}
<211> Brass component are identified by black spots [216]. . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.7 Left – Inverse pole figure maps with respect to the build direction (BD) and scanning direction (SD) for
LBP-DED RUS samples following heat treatments D, E, and F. Right – Corresponding {001}, {011}
and {111} pole figures in the BD plane. The idealised spots for the {011} <211> Brass component are
identified by black spots [216]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.8 Fibre texture analysis of the 001 and 011 components in the As-DED and F samples. The colour scale
represents the misorientation of grains (maximum 20◦) with respect to the 001 and 011. . . . . . . . . . 56
3.9 (a) Shear anisotropy factors, where [001], [010] and [100] correspond to the cubic shear planes of the
RUS samples. (b) The orientation dependence of the Young’s modulus for samples in the As-DED state
and following heat treatment G with respect to the TD and BD. . . . . . . . . . . . . . . . . . . . . . . 58
9
3.10 (a) - Room temperature RUS spectra for an LBP-DED sample before (red) and after (blue) an extended
ageing at 1100 ◦C (b) - High temperature RUS spectra of an LBP-DED sample collected during heating
(black), during the isothermal hold (red) and during cooling (blue). Circles are shown above individual
resonances to allow tracking through the thermal cycle, though this was not possible when peak broad-
ening became too great to allow resonant modes to be distinguished clearly from the high-temperature
background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.11 (a) - High temperature RUS spectra of an LBP-DED parallelepiped sample collected during heating, a
circle (blue) is shown above the first resonant peaks in the series of interest. An arrow (blue) has been
added as a guide to the eye to allow tracking of the peaks through the temperature range. (b) – Acoustic
loss plotted against temperature for each resonant peak in the series of interest. . . . . . . . . . . . . . . 61
4.1 (a) XRD patterns of the IN718 and IN718-NbC samples in the heat treatment A condition. (b) higher
resolution XRD patterns for the IN718 and IN718-NbC samples over a selected range of 2Θ. The reflec-
tions of the gamma (γ), MC carbide and Laves (φ) phases are labelled. . . . . . . . . . . . . . . . . . . 72
4.2 XRD patterns for LBP-DED IN718 (a) and IN718-NbC (b) samples in the heat treatment B condition.
The reflections of the matrix γ phase and MC type carbide have been labelled. . . . . . . . . . . . . . . 73
4.3 DSC traces for the first heating of LBP-DED IN718 and IN718-NbC samples from room temperature to
1400 ◦C. (b) A magnified section of the trace to highlight the differences in carbide solvus temperature
between the samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.4 SEM analysis of IN718 and IN718-NbC samples in the As-DED condition. Top, a backscattered electron
image. Beneath, elemental distribution maps for Cr, Mo, Nb, and Ti determined by SEM-EDX. . . . . . 76
4.5 SEM-EDS analysis of IN718 and IN718-NbC samples in the heat treatment A condition. An electron
image of the sample is given at the top, below this are elemental distribution maps for Cr, Mo, Nb, and
Ti obtained by SEM-EDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.6 SEM-EDS analysis of IN718 and IN718-NbC samples in the heat treatment B condition. A back-scattered
electron image of the sample is given at the top, below this are elemental distribution maps for Cr, Mo,
Nb, and Ti obtained by SEM-EDS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.7 TEM analysis of LBP-DED IN718-NbC in the heat treatment A condition. An annular dark field electron
image is shown at the top left; accompanied by selected area electron diffraction patterns for the matrix
[001], MC carbide [112] and C36 Laves [1¯100]. Below are elemental distribution maps determined by
STEM-EDX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
10
4.8 TEM analysis of LBP-DED IN718 in the heat treatment A condition. An annular dark field electron
image is shown at the top left; accompanied by selected area electron diffraction patterns for the matrix
[001], MC carbide [112] and C36 Laves [1¯100]. Below are elemental distribution maps determined by
STEM-EDX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.9 Mass gains of LBP-DED IN718 and IN718-NbC samples during isothermal oxidation at 650 ◦C for 200
hours. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.10 Grain size distribution maps for the recrystallised (heat treatment B) IN718 and IN718-NbC samples.
The false colour scale indicates the approximate size of a grain in microns. . . . . . . . . . . . . . . . . . 84
4.11 Inverse pole figure maps with respect to the build direction (BD) texture for the IN718 and IN718-
NbC samples in the As-DED condition. Both the build direction (BD) and scanning direction (SD)
have been identified. Below each IPF map the corresponding cubic pole figures down the BD have been
provided. For both samples, pole figures are included for the identified regions A, B, and C to highlight
the differences in observed texture symmetry. The idealised spots for the 011 <211> Brass component
are identified by black spots in the complete pole figures. . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.12 Inverse pole figure maps with respect to the build direction (BD) and scanning direction (SD) for the
LBP-DED IN718 and IN718-NbC samples in the heat treatment A and B conditions (Left). Correspond-
ing pole figures for 001, 011 and 111 poles in the BD plane of the samples. . . . . . . . . . . . . . . . . . 88
4.13 (a) Anisotropy coefficients, calculated from equations (2), (3) and (4) for the IN718 and IN718-NbC
samples. The [001], [010] and [100] directions correspond to the cubic shear planes of the samples. . . . 89
5.1 Algorithm describing the procedure for accounting for missing data entries for the vector x of the design
variables and properties. The value is computed recursively using n iterations. . . . . . . . . . . . . . . . 99
5.2 Schematic illustration of the neural network framework. The framework illustrates how the predicted
properties (outputs) are calculated from the input properties. The input layer is constructed from the
property database, this layer is used to calculate the hidden nodes (indicator functions) to give the
predicted properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5.3 Cross-validation tests for the properties of phase stability and yield strength. (a) Predicted phase stability
at 650 ◦C against calculated phase stability (CALPHAD). Poor predictions for high Nb and Ta containing
compositions have been circled. (b) Predicted yield strength vs experimental yield strength. For both
plots error bars have provided for the predicted values. Additionally, an idealised line has been added
as an aid to the eye. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
11
5.4 Ashby plot showing the probability of an alloy composition satisfying all of the design criteria when the
properties of phase stability (y – axis) and solidification strain (x – axis) are varied. The black regions
show areas of design space that have a low probability of fulfilling the targets. The lighter shading
indicates an increased likelihood of satisfying all the target criteria. The blue circles show the current
alloy IN718 and the designed alloy AM718R. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.5 Back-scattered electron images of the laser pass heat-affected zone (HAZ) and arc-melted microstructure
of IN718 (a) and AM718R (b) in the precipitation heat-treated condition. For ease of identification the
extent of the HAZ has been identified with a yellow line in both micrographs. . . . . . . . . . . . . . . . 108
5.6 SEM analysis of the IN718 and AM718R HAZs. Top, a secondary electron image. Beneath, elemental
distribution maps for Cr, Fe, Mo and Nb determined by SEM-EDX. . . . . . . . . . . . . . . . . . . . . 111
5.7 High magnification back-scattered electron images of the laser pass HAZs of IN718 (a) and AM718R (b)
in the precipitation heat-treated condition. These are example micrographs of those used for the phase
fraction analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.8 DSC traces for the first heating of IN718 and AM718R (a) samples from room temperature to 1400 ◦C
and the accompanying cooling curves (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.9 (a) XRD patterns for IN718 and AM718R samples in the precipitation heat-treated condition. (b) higher
resolution XRD patterns for the IN718 and AM718R samples over a selected range of 2Θ. Labels have
been added to highlight the reflections of the gamma (γ), MC carbide and Laves (φ) phases. Intensity
has altered to the square root of peak intensity for ease of visualisation of the weaker peaks present. . . 114
5.10 TGA traces for the 200-hour exposure of IN718 and AM718R in air at 650 ◦C. The graphs show the
mass gain with respect to area against the square root of time. . . . . . . . . . . . . . . . . . . . . . . . 116
6.1 Back-scattered electron images of the laser pass heat-affected zone (HAZ) and arc-melted microstructure
of NN7418A (a) and NN7418B (b) in the precipitation heat-treated condition. For ease of identification
the extent of the HAZ has been identified with a yellow line in both micrographs. . . . . . . . . . . . . . 125
6.2 Electron micrographs of the laser pass heat-affected zone (HAZ) NN7418A (a) and NN7418B (b) in the
precipitation heat-treated condition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.3 DSC traces for the first heating of NN7418A, NN7418B and IN718 arc-melted samples from room
temperature to 1400 ◦C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.4 TGA traces for the 100-hour exposure of NN7418A, NN7418B and IN718 in air at 750 ◦C. The graphs
show the mass gain with respect to area against the square root of time. . . . . . . . . . . . . . . . . . . 128
12
List of Tables
3.1 Composition (wt% unless indicated) of the Ni-based alloy used in this study; in all cases the balancing
element is Ni. Details of the composition of conventional forged materials have been given for reference.
The nominal compositions are those quoted from the production facility. The measured compositions
are those determined by EDX, LECO and ICP. The LBP-DED composition is that of the powder prior
to deposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.2 Dimensions and masses of the EDM prepared LBP-DED IN718 parallelepiped sample. . . . . . . . . . . 44
3.3 Heat Treatment protocols used during this study for the LBP-DED samples, the terms ppt and ReX
refer to precipitate and recrystallisation, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.4 Elastic constants and properties for the LBP-DED IN718 parallelepiped samples in their respective
heat treatment (HT) conditions. The sample number has been included in brackets beneath the heat
treatment condition letter. The elastic constants were calculated with sample dimensions for the RD,
TD and BD representing the x (1), y (2) and z (3) axes, respectively. An error of approximately ± 1% is
associated with the C11-C33 and C12-C23 constants and an error of approximately ± 0.1% is associated
with the C44-C66 constants. B – effective bulk modulus, E – effective Young’s modulus, G – effective
shear modulus, RMS - root mean square error. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.5 Anisotropy coefficients A100, A010 and A001 for the LBP-DED IN718 parallelepiped samples in their
respective heat treatment (HT) conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.1 Details of the heat treatments applied to the LBP-DED samples in this study. . . . . . . . . . . . . . . . 68
4.2 Dimensions and masses of the parallelepiped samples LBP-DED IN718 and IN718-NbC samples. . . . . 69
4.3 The measured compositions of LBP-DED IN718 and LBP-DED IN718-NbC, a combination of certified
techniques (ICP-OES for the major elements and B, LECO for C) was used to assess the composition
of each element (wt% unless indicated, Ni balance). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
13
4.4 Phase composition and solvus temperatures predicted for selected phases in IN718 and IN718-NbC.,
calculated with the Thermo-CalcTM software package assuming Scheil solidification and equilibrium to
400 ◦C. The quoted volume fractions were taken from the predictions at 400 ◦C under equilibrium
conditions. All solvus points are given in ◦C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.5 Approximate onset and termination point for thermal events observed in the DSC analysis of IN718 and
IN718-NbC. All temperatures are given in ◦C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.6 Compositional analysis of the phases in the IN718 and IN718-NbC samples obtained from STEM-EDX
point spectral analysis. The given compositions were averaged over five-point spectra. The compositions
have been given in atomic percent (at%). There is a standard error of ± 1% for all measurements. . . . 81
4.7 Elastic constants and average moduli for IN718 and IN718-NbC samples in the As-DED and heat-
treatment B conditions. The sample dimensions for the scanning direction (SD), transverse direction
(TD) and build direction (BD) were used as the x (1), y (2) and z (3) axes respectively in calculations of
the elastic moduli. Full details of the specimen used can be found in Table 4.2. An error of approximately
± 1% is associated with the C11-C33 and C12-C23 constants and an error of approximately ± 0.1% is
associated with the C44-C66 constants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.1 Composition design space selected for this alloy design framework. Elemental concentration ranges are
given in wt %. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.2 Table of properties predicted, and the method used for the prediction. In addition, the range of data
and number of entries used to train the neural network has been provided. The final two columns show
the prediction and targets for each property of the designed alloy. . . . . . . . . . . . . . . . . . . . . . . 96
5.3 The compositional ranges and recommended post-processing conditions for standard IN718 and AM718R.
For the measured composition of AM718R, SEM-EDX was used to assess the composition with a nominal
error of 1% for all measurements. The EDX measurements were taken from the laser pass heat-affected
zone. Carbon and boron have not been included due to the insensitivity of EDX in measuring light
elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.1 Compositions for the COTAC 74 and COTAC 744 alloy systems in wt % with a Ni balance. . . . . . . . 123
6.2 Compositions for the COTAC 74 and COTAC 744 alloy systems in wt % with a Ni balance. . . . . . . . 124
14
Abstract
The use of additive manufacturing (AM) for the fabrication of Ni-based alloys has seen a massive uptake in commercial
institutions. Among their many merits, AM techniques offer a unique route for the repair of a wide range of com-
plex components. However, processing of the Ni-based superalloys used by the aerospace industry through AM has
encountered numerous issues. Certain complications can be resolved through intense post-processing. Although, such
mitigation strategies are not appropriate for repaired material as intense heat-treatments might deteriorate substrate
properties. Therefore, new methods are required to control and optimise the deposition of Ni-based superalloys. The
commercial superalloy IN718 is currently used for certain AM repair applications. However, due to the limitations on
the post-processing that may be tolerated, the complications of an irregular microstructure, an uncontrollable texture
and severe elemental segregation remain unresolved for this system.
In this thesis, the laser-blown-powder directed-energy-deposition (LBP-DED) of IN718 is characterised and novel
methods for compositionally optimising the deposition of this system are investigated. It is shown that a Brass texture
({110} <211>) can be obtained in LBP-DED IN718. This texture is observed to be enhanced through recrystallisa-
tion following selected heat treatments. The evolution of elastic properties from the as-deposited state were studied
as a function of heat treatment time and duration using Resonant Ultrasound Spectroscopy (RUS). To control the
crystallographic texture formed, the addition of niobium carbide (NbC) inoculant particles to the precursor powder
was investigated. It was observed that the inoculants enhanced the formation of the Brass texture component, as
well as leading to the occurrence of alternating regions possessing mirror symmetry. The addition of the inoculant
therefore offers a method of achieving a degree of microstructural and textural control during LBP-DED of IN718 [3].
In the final chapter, a neural network framework is used to design a new Ni-based superalloy that surpasses the
performance of IN718 for LBP-DED repair applications. The compositional design space was based on IN718, although,
W was additionally included, and elemental limits were modified allowing the alloy to approach the composition of ATI
718Plus®. The newly designed alloy was fabricated, and the properties were experimentally investigated. The testing
confirms that this alloy offers advantages for additive repair applications over standard IN718.
Chapter 1
Introduction
Owing to their remarkable high temperature mechanical properties, Ni-based alloys are used extensively in aerospace
and land-based gas turbine applications [1, 2]. In modern superalloys, up to 13 elements have been used to optimise their
resilience in harsh environments whilst maintaining excellent mechanical properties over a wide range of temperatures
[1]. Indeed, through a combination of solid-solution and precipitate strengthening, superalloys are routinely used in
service applications at temperatures in excess of 1000 ◦C [1, 3]. Because of their complexity, careful process control is
required during fabrication and post-processing. This is critical to ensure that the microstructure and properties of these
systems are optimised for service. Improper fabrication conditions or post-processing steps can lead to complications
including cracking, underaged precipitates, occurrence of deleterious phases and severe segregation [1, 4]. To avoid
such complications both the selected processing route and alloy composition require optimisation to yield the best
performance in service.
The utility of superalloys facilitated the expansion of the civil aerospace industry during the 20th century [2] and, as
with many sectors, aerospace is continually striving for greater energy efficiency. However, engineering more efficient
engines that are affordable whilst adhering to strict safety guidelines is a monumental challenge [5]. Two direct methods
for improving engine efficiency are operating at a higher gas stream temperature or the reduction of engine weight [2,
6]. The former method involves the development of new higher performance materials, the uptake of which can take
years [5]. The latter method can be achieved through component weight reductions within the engine, which are often
easier to implement in service. An example of reducing weight is through the fabrication of single-piece components,
such as a bladed turbine discs (Blisks), see Figure 1.1. This eliminates the need for mechanical fasteners and joiners.
1
Additive manufacturing (AM) technologies offer greater design freedom and provide fabrication routes to achieving
component weight reductions [7, 8]. In addition, AM techniques have several promising applications including the
production of near-net-shape components, multi-material joining, and rapid prototyping [7]. There is even the possibility
of using elemental powders to produce unique compositions and graded components [9]. Over the last two decades,
extensive research into the additive manufacturing of alloys has been carried out leading to an uptake of this technology
within industry [10-15]. There are several comprehensive reviews summarising the successes of AM within the field of Ni-
based alloys [16-19]. These reviews also highlight the significant challenges encountered when additively manufacturing
superalloy components, some of which have yet to be resolved. However, additive methods have the potential to
revolutionise component production and development once these challenges are overcome [14].
Figure 1.1: Large weight reductions can be achieved through the manufacturing of a Blisk, courtesy of Rolls-Royce
plc.
There are numerous AM processes through which metallic material can be processed [7, 9, 20]. Each technique has
a unique set of processing conditions which lead to a different final product [16]. However, these methods are still
commonly plagued with complications including cracking, residual stress, elemental segregation, mechanical and mi-
crostructural anisotropy, and microstructural inhomogeneity [7, 21]. These issues have hindered the performance of AM
components when compared to conventional counterparts [21, 22]. Many of these deficiencies arise from the intense
thermal cycling which the material undergoes during production [17, 23]. These processes can be likened to multi-pass
welding techniques. Indeed, similar challenges have been shared between the two fields [24-26].
The geometrical versatility of some DED techniques, such as laser-blown-powder directed-energy-deposition (LBP-
DED), allows for the fabrication and repair of geometrically complex components. The option of component repair
using this technique is particularly attractive for parts with high unit cost such as the bladed disk assemblies, see Figure
1.1 [27]. These components could suffer damage to the leading edges of the blade aerofoils in service, resulting in a
possible reduction in engine performance. The repair of damaged aerofoils therefore offers the potential of substantial
cost savings by mitigating the expense of whole component replacement. It has been shown that DED repair methods
have the potential to be superior to conventional welding-based repair techniques as lower heat inputs allow reductions
in the size of the heat-affected-zone (HAZ) and substrate distortions to be achieved [28, 29]. In addition, the high
2
solidification rates of additive manufacturing techniques can produce components with highly refined microstructures.
This can lead to improved mechanical properties over conventional analogues [7, 10]. These methods also offer the
potential benefit of minimising microstructural changes to the substrate and hence loss in mechanical strength. To
capitalise on these opportunities, such methods are now being evaluated for the repair of a range of Ni-based superalloy
turbine blades [27, 28, 30, 31].
Recently, excellent progress has been made on the optimisation of process parameters and material post-processing;
with many of the complications being alleviated through intense heat treatments. Unfortunately, such regimens are not
applicable to repaired material as the substrate material can be compromised through over-ageing [32, 33]. However,
less attention has been given to the optimisation of specific alloys for AM processes. Instead, the majority of current
research has been focused on optimising the process around specific alloys [21, 34]. To fully realise the potential of
additive manufacturing in the production of superalloys, both the process and the alloy must be optimised. Extensive
discussion in the following literature review examines how alloy augmentation and design could yield an optimised
system for use in additive repair applications.
3
Chapter 2
Literature Review
2.1 Ni-Based Alloys
The extraordinary high-temperature tensile properties of some Ni-based alloys has been attributed to the presence of
ordered precipitates, such as the γ′ or γ′′. The origins of the strength at high temperatures stems from the anomalous
yielding effect, which is widely accepted to be the result of a change in dislocation behaviour at higher temperatures [1,
3]. In addition to high mechanical strength, these systems must also be resistant to the effects of corrosion, oxidation,
and thermo-mechanical fatigue. To meet such demanding requirements in performance, the compositions of these
alloys tend to be highly complex. Through augmentation of the composition, the microstructure of superalloys can
be controlled, leading to enhancement of certain properties. This was demonstrated by the single-crystal family of
superalloys where creep performance at high temperatures was prioritised [1]. As these systems are manufactured
without grain boundaries, the requirement for certain minority elements has been eliminated. A further example is that
of filler metal used in welding applications. These alloys are designed to be more resistant to hot cracking mechanisms;
although the mechanical properties of filler metals might be inferior when compared to the welding substrate. It should
be noted that enhancement of certain performance aspects might cause other properties to deteriorate. To achieve a
higher weldability, the concentration of strengthening and minority elements are carefully controlled [1, 3, 4, 35].
Within the austenitic (γ) matrix of superalloys, atomic ordering can lead to the formation of several stable and
metastable phases. The gamma prime (γ′) and gamma double prime (γ′′) precipitates are formed as result of solute
4
elements (e.g. Al, Nb, Ta, Ti etc.) that promote ordered crystallographic phases within the FCC-γ matrix [1, 4, 36].
In addition to γ′ and γ′′ many other phases exist that can precipitate within the austenitic matrix, including carbides,
borides and topologically closed packed phases (TCPs) [1]. Selected examples of certain phases have been given in
Figure 2.1. Whether these phases are beneficial or detrimental to the mechanical properties of the alloy depends
upon the volume fraction, location, and morphology [37, 38]. In general, the brittle TCP phases are considered highly
detrimental to the properties of superalloys since they deplete the matrix of strengthening elements lowering rupture
strength and ductility [1, 39]. This can also be true for certain carbides and borides, especially in the case of non-MC
type carbides [40, 41]. It is, therefore, critical that the concentration of the strengthening and minority elements are
controlled. These elements have a capacity to promote precipitation of potentially undesired phases in high volume
fractions [1]. This is a particular issue for welding and additive manufacturing techniques. Such processes involve high
peak temperatures and thermal gradients, as well as intense thermal cycling [7, 42]. This often gives rise to re-melting
and re-solidification processes which can exacerbate elemental segregation [43]. In many circumstances, elemental
segregation leads to precipitation of undesired phases which can compromise alloy properties [44].
Figure 2.1: SEM Images of selected phases that occur in Ni-based alloys. (a) As-deposited DED IN718 with Laves
eutectic identified along sub-grain boundaries. (b) DED IN718 solutioned at 1100 ◦C for 2 hours, with spherical MC-
type carbides identified. (c) Forged IN718 with rod-like delta precipitates along the grain boundaries. (d) Image from
Wilson et al. [45] highlighting the occurrence of TCP phases and carbides in polycrystalline RR1000 following a 5000-
hour heat-treatment at 800 ◦C.
5
The extensive research that has been completed on phase stability has greatly improved the understanding of the nu-
merous precipitates which can occur in Ni-based alloys [46, 47]. The design of Ni-based alloys against undesirable phase
formation has also been significantly aided by new PHACOMP— a method based on an electron vacancy concept [48].
In addition, the occurrence of phases within Ni-based alloys can now be simulated for given composition and temper-
ature using computational CALPHAD (CALculation PHAse Diagrams) methods [49]. These modern computational
methods can be used to design highly specialised, theoretical alloy compositions.
2.2 Alloy Design
Traditionally, alloy development relied upon an informed trial and error approach. However, this is time consuming
and can fail to reach the optimal composition. In recent years, automated alloy design has shown significant progress
with the enhancement of computational power and modelling [6, 46]. This has allowed for more accurate predictions of
mechanical and thermodynamic properties. More advanced physical models have also been developed to describe key
mechanical processes such as stress rupture life [50] and creep deformation [51, 52]. With the advances in computational
power the use of artificial intelligence methods, such as neural networks, which require the analysis of vast datasets,
is now possible [53]. This has allowed for a deeper understanding of property-property relationships to be developed.
Similarly, software such as Thermo-CalcTM, which calculates predictions based-off vast databases of experimental and
theoretical results, have given researcher access to reliable phase diagram calculation (CALPHAD) [54-56]. These
data can in turn be used to complement physical models or artificial intelligence frameworks for the for prediction of
mechanical and environmental properties [57-60].
Early alloy design frameworks utilising computational tools have been demonstrated by Saha and Olson. [61, 62] and
Tancret, Bhadeshia & Mackay [63, 64]. Saha and Olson sought to develop steels for naval applications. Their systematic
design strategy, operated by a multilevel approach, linked several factors to final alloy performance using mechanistic
and thermodynamic models. Discrete steps were used in the design process so that specific properties were optimised.
For certain steps, feasibility studies were carried out to verify predictions and design considerations. A schematic
representation of their approach is shown in Figure 2.2.
6
Figure 2.2: A materials system flowchart for the design of blast resistant naval hull steels, adapted from Saha and
Olson [61]
Using this framework, the authors were able to achieve the property objectives for a transformation toughened steel.
The properties of toughness, strength and weldability were targeted in the design process. In the framework, both ther-
modynamic and quantitative models were used to describe transformation toughening mechanisms and strengthening
dispersions. The full explanation of the models used have been given by the authors in their design approach section
[61]. For most modelling calculations, the predictions were supported with appropriate experimental data.
The contribution from each strengthening precipitate was used to determine the optimum concentration. Utilising the
Russell-Brown model, the contribution to strengthening from copper precipitates was described. A graph detailing the
expected hardness contribution of each substructure is given in Figure 2.3. For austenitic transformation toughening the
Olson-Cohen model was used to describe the stability of the dispersed austenite. Through a series of calculations, with
selected thermodynamic data from Thermo-CalcTM, the optimum concentration of Ni for austenitic stabilisation was
determined. Following the individual compositional optimisation for the strengthening and toughening mechanisms, an
overall composition was determined from the respective stages. Further evaluation of alloy composition with respect to
processability and environmental resistance was also undertaken by the authors [61]. Overall, the systematic strategy
was successful in designing a new steel and the evaluation of this system is given in a follow up study [62].
7
Figure 2.3: A graph, adapted from Saha and Olson [36], representing the expected contributions of the individual
strengthening mechanisms to achieve the targeted strength goal, equivalent to 389 VHN.
The alloy design framework developed by Tancret, Bhadeshia and Mackay [63, 64] utilised an alternative methodology.
The authors sought to design an affordable Ni-based superalloy for application in steam power plants operating at high
temperatures. The new alloy system was required to be less expensive than existing superalloys, by the exclusion of
expensive elements, whilst retaining the required high temperature properties such as creep life. To design such an alloy
the authors employed a two-stage approach. The first stage was modelling and design for the mechanical properties.
This was then followed by a second stage modelling the phase formation and solidification thermodynamics. For the
modelling of the mechanical properties in the first stage, Gaussian processes were employed. Such methods work by
performing a non-linear multidimensional regression of an output value as a function of selected input values. The
output values, for example creep life, have several inputs: temperature, thermal history, and composition. The full
explanation of Gaussian processes is given by the authors [63], though these processes are regression tools related to
neural networks [65]. For the Gaussian modelling to be reliable a large property database is required for training. The
authors demonstrated that the Gaussian processes were robust with comprehensive testing and training of the models
for each property.
In the second stage, the Thermo-CalcTM software, alongside a large Ni-based superalloy database developed by Rolls-
Royce plc., was used extensively for the prediction of phase diagrams and solidification behaviour. Additionally, a
special model based on Scheil’s approximation was used for the prediction of segregation during solidification [64]. Such
modelling ensured that the designed system had a stable microstructure during operation and was processable. Careful
consideration was given to the behaviour of carbon and boron within the alloy. These elements are responsible for the
formation of deleterious phases on grain boundaries [1]. It was determined that, if correctly solutioned, no deleterious
8
carbides would form. However, the formation of borides was unclear and experimental verification was required [66].
Another deleterious phase, α–Cr, was supressed by minimising the content of W, Al and Ti. The experimental results
for verification of this designed alloy were given in the third stage [66].
Though the described frameworks are both impressive and were successful, the procedure for searching for the optimum
composition was limited. These frameworks are somewhat ’manual’ due to the separate property-composition predic-
tions. With the recent advances in computational power and optimisation algorithms it is possible to fully automate the
design process [53, 67]. Such computational methods offer the ability to process and correlate massive volumes of data
[65]. Indeed, Tancret et al. [68, 69] has demonstrated the incorporation of genetic algorithms for efficient optimisation
of alloy properties in the design process. These algorithms and related optimisation routines often operate through a
minimisation procedure.
In genetic algorithms, a target with selected inputs will be given a merit criterion. This criterion is optimised to provide
the best target from selected inputs. This can be done as a random-step process where individual inputs are varied in
discrete steps. Only steps that provide an improvement are accepted [65, 67]. This whole process can be operated with
numerous input values. The ’semi-elitist’ genetic algorithm used by Tancret is well described in the author’s work [69].
Though such computational tools are impressive, it is important that the design space for searching is appropriately
limited. If the input space is too vast or data steps too precise, the computational cost and time required to search
become prohibitive [65, 69]. In general, this genetic algorithm is used for the optimisation of a single property within a
set of constraints or characteristics. Several optimisations can be used in a multi-objective optimisation procedure for
the selection of the alloy system [70].
A more recent computational framework that shares similarities with the aforementioned work [68, 69] is that designed
by Conduit [65] and Conduit et al. [46, 53]. The authors utilised neural networks to design an alloy optimised for
directed-energy-deposition (DED). A schematic representation of a neural network of the type used in their work is
given in Figure 2.4. The input layer consists of physical alloy properties and thermodynamic data describing different
alloys systems. Such neural networks rely on large databases for property predictions and multiple models were used in
the framework to calculate certain values. Indicator functions were used to transform these input variables into hidden
node values. Indicator functions then combined these hidden nodes to yield a final output.
To ensure that the optimisation procedure employed by the framework was efficient, Conduit et al. [46] utilised an
approach termed ’simulated annealing’. This methodology automatically defined the step-size used to explore multidi-
mensional design space. The authors used step lengths comparable to the accuracy with which the material could be
9
manufactured. This value is varied by the framework to ensure rapid exploration. For the multi-objective optimisation
Conduit et al. [46, 53, 71] used a single merit index calculated from the contribution of each target property. This index
is used to estimate the likelihood that the material properties will satisfy the overall design criteria. As a likelihood
contribution from each property is used, this tool can efficiently select the optimal compromise between properties and
predict the overall likelihood of meeting all property targets.
Figure 2.4: Schematic representation of a neural network used by Conduit et al. [53]. The schematic describes the
relationship between two selected output values (y1 and y2). Given properties (red) are used for the calculation of the
hidden nodes (blue) which yield the predicted properties (green).
10
This form of machine learning requires a large and reliable database from which to optimise. However, there can be a
problem because data for a single alloy composition manufactured using different techniques might not be comparable.
This can lead to discrepancies in the training database, reducing the reliability and the database must be holistic
and contain comparable data. For additive manufacturing, there is often limited data readily available for Ni-based
alloys systems. This restricts the available training data and, therefore, reduces the reliability of the design procedure.
However, this methodology can also be extended to improve the process parameters used to deposit Ni-based alloys.
Following extensive research, significant volumes of data have been collected with regard to process parameters. Machine
learning techniques can use such data to optimise the process parameter for certain alloys [72, 73]. For example, Kappes
et al. [74] were able to demonstrate how the deposition conditions in laser powder bed fusion could be optimised for
Inconel 718. The development of machine learning presents an opportunity to improve the additive manufacturing
process in addition to alloy design. However, as the current research has been focused on a select few alloys, use of
neural networks alone is potentially unreliable.
Another successful methodology is the Alloys-by-Design framework of Reed, Tao and Warnken [75, 76]. Alloys-by-
Design is a systematic approach for the optimisation of a Ni-based superalloy through a series of rules and models.
The models include key alloy characteristics such as mechanical properties and microstructural stability. The rules are
integrated with these models in order to identify a few optimal compositions. The authors demonstrated that with this
approach an alloy offering an ideal balance of properties can be identified. Their approach is best visualised through the
use of trade-off diagrams that demonstrate the relationship between properties. Example property trade-off diagrams
from the work of Reed, Tao and Warnken [75] are given in Figure 2.5. These plots show the relationship between
properties over a substantial range of alloy compositions. In effective alloy design it is necessary to accommodate
the compromises between properties which are required for specific applications over that of others; for example, the
optimisation of creep resistance against that of tensile strength. This is, of course, dependent upon the intended use of
the alloy system. For such optimisations to be successful, reliable models are required for each property that is being
considered.
Models for the mechanical properties of Ni-based alloys have been described in the work of Tancret et al. [63, 77]
and other authors [52, 78-81]. The tensile and creep properties of superalloys are often prioritised; however, the en-
vironmental resistance and stability of the microstructure are essential. Additionally, the volume fraction, location,
and size of precipitates in Ni-based superalloys must be carefully considered in the alloy design processes to ensure
that the optimal system is selected. For example, in ATI 718Plus®, a small fraction of the delta and eta phases are
used to pin-grain boundaries, adding some strength to the matrix [82, 83]. In high concentrations, these phases can be
embrittling and serve as a path for crack propagation. Using computational methods, the microstructure, with respect
to temperature, can be predicted [55, 84, 85].
11
Figure 2.5: Property trade-off diagrams for Ni-based superalloys as presented by Reed, Tao and Warnken [75] (a)
density vs. stability number (Md) (b) density vs. cost, with the change in Re content identified (c) cost vs. stability
number, with regions of Re content and change in Ta and Al content identified.
12
The work by Conduit et al. [46, 53] and Tang et al. [76] has demonstrated how an alloy can be effectively designed for an
AM technique. Recently, more sophisticated modelling of solidification, microstructure and properties has allowed for
better predictions of alloy performance [86-88]. Using such tools, it might be possible to optimise existing alloy systems
or design completely new compositions which are more amenable to AM methods. In Section 2.3 below, standard
AM techniques and process parameters used for the deposition of metallic materials are introduced. The properties
of additively manufactured Ni-based alloys and current challenges associated with deposition of these system are also
discussed.
2.3 Additive Manufacturing of Ni-based Alloys
Extensive work has been carried out on the control and manipulation of process parameters to deliver a more optimised
component. Several reviews have been published which describe this research in detail [16, 21, 89]. Studies have shown
that process control can alleviate microstructural defects, elemental segregation, and residual thermal stress. However,
control of the production conditions can be difficult and attempts to solve one issue might lead to or exacerbate
other problems [20, 34]. For example, the reduction of heat input to control residual stress can induce porosity due to
insufficient energy for precursor fusion [21, 90-92].
Some studies, such as that by Helmer et al. [93], have sought to produce process maps detailing ranges of processing
parameters in which suitable components can be manufactured. It should be noted, however, that these ranges might
only be suited to a specific alloy and additive manufacturing method [74, 94]. Dass and Moridi [95] collected data on
the AM of numerous alloy systems to construct a DED compilation process map; this is presented in Figure 2.6. It is
clear from these data that different alloy systems require different process parameters to be optimally fabricated using a
specific AM technique. In addition, some complications have been mitigated through proper post-processing of the final
component [21, 30, 96]. It was shown by Ruttert et al. [33] that process-induced defects in EBM-fabricated CMSX-4
can be significantly reduced through hot isostatic pressing (HIPing). As it has been mentioned, intense post-processing
would not be suitable for components repaired through AM techniques [97, 98]. It is, therefore, clear that more research
is required until AM components can fully surpass their conventional analogues [7, 22].
When reviewing the literature, it is noticeable that a significant portion of research has been completed on a select
few alloys — frequently IN718 [87, 99, 100]. Alloys such as IN718, IN625 and Waspaloy are relatively inexpensive and
reputable for their good processability [21, 25, 90, 101]. It is worth mentioning that the cost of producing high-quality
13
Figure 2.6: A compilation process map for the DED of numerous alloy systems presented by Dass and Moridi [95].
The plot identifies process parameter zones that yield desirable deposition condition, as well as zones that can lead to
complications including keyholing and lack of fusion.
precursor (for example, powder) is a significant factor and can limit studies to using conventional compositions. Also,
it has been well reported that the precursor itself has a profound impact on the final quality of the build [102, 103]. A
recent study by Zhong et al. [104] demonstrated that the use of gas atomized (GA) IN718 powder led to builds with
a higher level of porosity than powder produced by the plasma-rotating electrode process (PREP). Interestingly, the
microstructure for the GA builds was finer, with a lower level of Nb segregation and, consequently, a lower volume
fraction of the Laves phase. It was explained that there was a higher cooling rate during deposition when using GA
powder. This is caused by a different laser energy allocation when using GA versus PREP powder.
As research has been focused on a select few alloys, there is limited data to fully understand how alloy composition
responds to different AM techniques. However, significant advancements have been made with respect to the AM
of Ni-based systems [11, 16, 89, 102]. Studies have demonstrated the importance of alloy composition, critically the
concentration of the minority elements, on defect density and microstructure [26, 53, 87, 105]. For example, it was
shown by Yang et al. [106] that vanadium had a significant influence on the porosity and the extent of Nb segregation
in laser clad IN718.
14
Given the extensive literature available within the field, this review will focus on a select few additive manufacturing
techniques and the associated process parameters. Two methods, directed energy deposition (DED) and powder bed
fusion (PBF), have been used extensively for the production of metallic parts [7, 17, 92]. These techniques normally
use a powder or wire feedstock and an energy source to build up parts in a layer-by-layer fashion [17]. Commonly for
DED techniques, the building head (containing a precursor nozzle and laser) can move freely in 3D space, yielding
greater versatility [7, 30]. Both are highly popular and utilise the method of melting or sintering successive powder
layers. Each additive technique has both unique advantages and complications that can limit their application [21].
The processes also share several similarities, during material fabrication, which lead to comparable microstructures
and properties. Thus, it might be possible to design an alloy to suit these general process parameters [102, 103].
2.3.1 Directed Energy Deposition
In DED, an energy source (laser, electron or plasma-arc) is used to melt a metallic feed material (powder or wire)
creating a heat affected zone (HAZ) on the substrate surface around the melt-pool [17], see Figure 2.7. The nature
of this pool can be monitored using a range of sensors which include pyrometers and infrared cameras. Resulting
information is fed into a computer for data collection and feedback control, allowing the dynamics of the melt pool to
be altered or kept constant throughout the deposition process.
Figure 2.7: Directed Energy Deposition Process Schematic
15
There are many build parameters to be controlled and monitored: scanning velocity, scan strategy, material feed-rate,
interlayer dwell time, energy density, atmospheric conditions and blown powder dynamics (in the case of powder feed
systems) [20, 89]. Since the standard setup consists of a static build table with a moveable deposition head, the DED
technique has a greater geometric freedom. This allows for a higher level of control when varying deposition parameters
such as layer thickness and traverse velocity. The DED technique is capable of high build rates and, due to geometric
freedom, is well-suited to complex repair applications [17]. However, residual stresses from intense thermal cycling
and a poor surface finish are common with this method. Thus, post-process machining and heat treatments are often
necessary to eliminate these complications [91, 92]. Furthermore, this technique often requires structures to support a
component during deposition. Though conceptually simple, the process of DED is highly complex with poor control
leading to problems in the final product.
2.3.2 Powder Bed Fusion
Similar to DED, powder bed techniques rely upon the use of an energy source to melt and fuse consecutive layers of a
metallic precursor [20]. A schematic illustration of a generalized PBF technique is given in Figure 2.8. Energy sources
are most commonly laser or electron based and the relevant techniques are subdivided into the selective laser melting
(SLM) and electron beam melting (EBM) methods, respectively. An important difference between SLM and EBM is
the deposition atmosphere. A high vacuum is required for EBM, while the SLM technique can operate under an inert
atmosphere. As both processes require a controlled environment, the maximum size of a component is limited by the
dimensional constraints of the deposition chamber [9].
The general PBF setup consists of a static build platform and an energy source with three-axis movement. A roller
is used to distribute a bed of powder on to the platform, where the layer thickness can be controlled by the amount
of powder used. Following deposition of a single layer the platform is lowered to a height corresponding to the layer
thickness. A component is built up in a layer-by-layer fashion dictated by a 3D computer-aided design (CAD) model
and the chosen scan strategy [20]. In general, the PBF method produces components with a better surface finish and
higher dimensional accuracy than other AM processes. Thus, less post-process machining is required for components
produced by this method. However, as with DED, the residual stresses are typically high [92]. The PBF method
shares similar build parameters to DED including scanning velocity, scan strategy and energy density. Control of these
parameters significantly influences the resulting microstructure and properties of the manufactured component [73].
Many defects, such as porosity, cracking, and irregular microstructure, are common complications encountered from
the AM of Ni-based superalloys [21]. Unsurprisingly, significant research has been completed to understand why these
defects occur and how they can be mitigated.
16
Figure 2.8: Powder Bed Fusion Process Schematic
2.3.3 Effect of Process Parameters
Many studies suggest that careful control over the process parameters can eliminate or alleviate most defects resulting
from deposition [11, 73]. The work of Moussaoui et al. [107] demonstrated that porosity in SLM IN718 builds could be
significantly reduced via control of the energy density. Studies have highlighted that a higher energy density reduces
the porosity as the melting and fusing of layers improves [107]. However, this can lead to increases in other defects,
such as cracking. It was shown by Reed et al. [20] that a higher laser power reduced porosity in LPB CM247LC builds,
although instances of solidification cracking increased. Indeed, their study described, in excellent detail, the different
mechanisms of cracking that can occur in a γ′-strengthened, Ni-based superalloy. The susceptibility to different modes
of cracking, due to the processing parameters, was highlighted and it was noted that both high and low laser powers
increased susceptibility to cracking in LPB CM247LC. However, it must be understood that different alloys do not
necessarily behave in the same way when additively manufactured. A set of process parameters optimised for one alloy
might not be appropriate for another system. Several of the common process parameters in AM are described below
and the general effects these have on Ni-based alloys
17
2.3.3.1 Scan Strategy and Heat Flow
The scan strategy is an essential parameter set in the AM process as it influences microstructural features, residual
stress, and thermal distortions [108]. Several common pattens include the raster, bi-directional and island strategies
(selected strategies given in Figure 2.9). The raster strategy is suitable for almost all component geometries. However, it
can lead to a greater final distortion in the manufactured part [109]. This parameter can control the heat distribution
during fabrication of a component, affecting the grain structure. Specifically, it has been established that the ratio
between the thermal gradient and the growth rate (G/R ratio) influences grain morphology. Studies have shown that
a lower G/R ratio promotes the formation of equiaxed grains, and a higher G/R ratio favours a columnar structure
[26, 110, 111].
Figure 2.9: Representations for selected scan strategies (a) Linear (b) Bi-linear (c) Off-set (d) Checkboard Island.
The scan strategy can control the G/R ratio while also being utilized to encourage growth of a preferred microstructure
[112]. It was shown by Carter et al. [108] that equiaxed growth was favoured by a chessboard scan strategy that promoted
a uniform heat distribution. The scan strategy can also have a significant influence on the degree of anisotropy in a
manufactured component. It has been demonstrated by Popovich et al. [113] that fine control over the microstructure
and texture of SLM IN718 can be achieved with a suitable selection of process parameters that include the scan
strategy. Similar studies by Wan et al. [114] and Geiger et al. [115] have also demonstrated how the scan strategy
influences the texture and properties of a manufactured component. Through manipulation of the G/R ratio and in
situ heat dissipation, an equiaxed microstructure can be produced [93, 100, 116]. However, this can complicate and, thus,
extend the manufacturing process. In general, the columnar microstructure is highly anisotropic with grains primarily
orientated along the <100> cubic axes that were parallel to the direction of heat flow during processing [117]. The
18
direction heat flow is generally along the building axis, with heat being dissipated into the substrate. It is important to
note that the substrate should be significantly large enough and possess a reasonable thermal conductivity to prevent
excessive heat being built up during deposition. However, a lower thermal gradient can be generated through a build
up and retention of heat. This may be desired to control the microstructure and reduce the generation of extreme
thermal stresses during deposition [7, 10, 16, 20]
The importance of heat flow in the determination of the crystallographic texture is described in detail by Ma et al.
[117]. In addition, Dinda et al. [111] discussed how heat-flow mechanics gave rise to a rotated cube texture when a
bidirectional scan strategy was implemented in the DED of IN718. The effect of heat flow is easily understood when
layer-by-layer growth mechanics are considered. Nucleation and growth of dendrites will occur in the direction of local
heat flow and this is normally the build direction in AM processes [111]. Therefore, a degree of texture orientated
parallel to the build direction is to be expected. Nucleation and growth of dendrites will also be dependent upon
crystallographic nature of the underlying substrate, with epitaxial growth from the grains of the substrate occurring
in numerous circumstances. Epitaxial growth of the dendrites from the underlying substrate grains might be desired
and control of the microstructure texture in this manner has been demonstrated [7, 16, 20, 21, 133]. For FCC Ni-based
alloys, the fastest crystal growth direction is along <100> axes and, as a result, preferential growth occurs in <100>-
type grains that have an orientation parallel to the direction of heat flow. This gives rise to the observed cube texture
along the build direction that has been reported in many studies [21, 31, 117, 118]. If the laser power is sufficient, the
scanning direction can affect the angle of local heat flow and thus the growth mechanics. Therefore, the scan strategy
can be used to tailor (or grade) the local texture of a deposited Ni-based alloy [113, 115]. This has been highlighted
in Figure 2.10 which shows the textured growth of columnar grains in LBP-DED IN718, inspired by the work of Wei
et al. [119]. Through modelling the authors examined the heat flow with respect to the deposition strategy so as to
explain observed cubic growth mechanism in AM IN718.
In addition, studies have reported that the G/R ratio affects the segregation of elements to the inter-dendritic regions
and, therefore, the formation of undesired phases. Excessive precipitation of these phases can deteriorate the mechanical
properties as the matrix is depleted of the elements that are used in both precipitate and solid solution strengthening
[4, 17, 35, 82]. These phases are also brittle and, depending upon the morphology and distribution, can allow for crack
propagation. Intense heat treatments can also be used to homogenise the microstructure [32, 33].
Interestingly, it has been reported by Nie et al. [120] that the residual stress can be reduced and the microstructure
refined by the introduction of a static magnetic field during the DED process. The authors stipulated that the in-
troduction of a magnetic field suppressed the convection within the melt pool, altering the heat-flow mechanics. This
19
Figure 2.10: EBSD IPF-Z map of as-deposited DED IN718 showing the growth direction of columnar grain. The
LBP-DED IN718 sample was fabricated with a bi-linear scan strategy which caused the heat-flow direction to be
altered. The build direction on this image has been indicated. The IN718 sample used in this analysis was deposited
via LBP-DED onto a forged IN718 substrate using standard deposition parameters.
suppression led to a reduction in the cooling rate and, consequently, reduced residual stresses. A reduction of 26% in
the residual deformation was measured in the study. It was also stated that the magnetic field disrupted the growth of
dendrites, refining the average dendrite spacing. Furthermore, it has been reported that control of the heat flow can
be used to manipulate the grain boundary orientation and further reduce cracking susceptibility [100, 121]. As with
many process parameters, the ideal scan strategy will depend upon the geometry of the component, the alloy being
deposited and the intended use of the part. In order to determine the optimum strategy, trial and error methods or
numerical modelling must be used [34, 73].
2.3.3.2 Energy Density
The energy density is a general parameter describing the energy that is absorbed by the surface. Variations of this term
have been used in the literature and its calculation has included the use of parameters such as input power, scanning
velocity and aspects of the scan strategy [17, 18, 30]. Several other process parameters can be included in its calculation
and the term can be described in terms of length, area or volume [122]. Since energy can be reflected by the surface,
the true magnitude of this term is often challenging to determine. This is especially important in laser-based systems
20
where beam attenuation can have a significant effect on this parameter [89]. Although this term can be useful, it is
often not comparable between studies if different inputs have been used in the calculation.
For the purposes of this discussion, input power, scanning velocity and hatch spacing are the focus because of their
importance. Studies have demonstrated how control of these parameters not only reduces defect density but can also
be used to tailor the microstructure of a deposited alloy [11, 102, 115]. This is not surprising since heat flow, thermal
gradients and melt-pool dynamics are each dependent upon these parameters [93]. For example, a lower input power
and a higher scanning speed decrease the incident energy on the surface, increasing the cooling rate [7, 123]. This can
result in a more refined microstructure. However, a lower power can lead to complications [124-126]. It was shown by
Carter et al. [20] that a low laser power increases porosity in SLM CM247LC. This is a result of poor melting and
fusion of the alloy precursor during deposition.
In contrast, high energy inputs can lead to complications including key-holing mechanisms during deposition. This
can result in porosity and, in some circumstances, cracking in the manufactured component [20, 94]. Examples of
numerous deposition-induced defects have been shown by Scime and Beuth [94] and are presented here in Figure 2.11.
The authors effectively demonstrated how different process parameter combinations could give rise to either desirable
deposition or defects in the AM of IN718. The very high cooling rates can also hinder the precipitation of strengthening
precipitates during deposition. This must be resolved through post-processing [31]. In circumstances where the input
power is too high, cracking can occur and segregation of elements such as Nb might be exacerbated [17, 18].
Figure 2.11: Examples of melt pool morphologies resulting from different process parameters, taken from Scime and
Beuth [94]. (a) desirable (285W, 960mm/s), (b) balling (370W, 1200mm/s), (c) under-melting (100W, 1000mm/s),
(d) severe keyholing (250W, 400mm/s), and (e) keyholing porosity (150W, 200mm/s).
In addition, the way in which the energy is delivered to the surface during deposition is also critical in determining
the heat distribution within the melt pool. Several studies have demonstrated how the microstructure can be refined
21
and segregation reduced by altering the mode or type of energy source [43, 127, 128]. The work presented by Xiao et
al. [127] demonstrated how the use of a quasi-continuous wave energy source could promote the formation of equiaxed
dendrites in laser-AM IN718. The use of a non-continuous energy source lowers the total energy delivered, although the
peak temperatures reached by the surface are comparable to the continuous wave system. This ensures good melting
of material whilst limiting the build-up of heat. As a result, the solidification conditions are significantly modified. The
cooling and the solidification rate are both increased leading to the refinement of the microstructure. This also limits
the segregation of Nb, suppressing the formation of Laves phase. Similar results have been reported when using pulsed
and non-Gaussian energy sources.
Clearly, the distribution of heat is important. The hatch distance (the separation between adjacent scan tracks centres)
controls the concentration of energy and influences the resulting microstructure. The work of Nadammal et al. [129]
demonstrated that a more textured, columnar microstructure with higher residual stresses was developed when a
shorter hatch length was used in the SLM of IN718. When the hatch length was increased, the thermal gradient was
reduced allowing for a better dissipation of heat through the underlying substrate. Manipulation of the energy density
also affects the melt-pool geometry and can influence the grain morphology [130, 131]. Clearly, microstructural, and
tensile properties can be influenced through process parameter manipulation, though complications persist. Segregation,
residual stress, and microstructural anisotropy have not yet been fully eliminated through process control alone and
their presence has the potential to hinder the performance of additively manufactured Ni-based alloys [21, 43, 132].
2.4 Microstructure and Mechanical Properties of AM Ni-based Alloys
Discussions in the literature have illustrated that the microstructure and properties of additively manufactured Ni-
based alloys are heavily dependent upon the thermal history of the manufactured part. The high level of user control
over the AM process parameters (dictating factors that include thermal gradients, cooling rates, peak temperatures,
and melt pool geometries) have also been discussed [30, 43, 92]. As a result of this control, there is a high variability of
microstructures reported in the literature. Indeed, it is still challenging to accurately predict the resultant microstruc-
ture from a set of process parameters [73, 85, 88].
22
2.4.1 Microstructural Morphology and Anisotropy of AM Ni-Based Alloys
In many circumstances, the accumulation of heat and the occurrence of re-solidification during deposition result in
the epitaxial growth of columnar grains [21, 133]. Commonly, there is also a prominent ’fish scale’ pattern at the
microstructural level that is produced via layer-by-layer scanning of the energy source [109]. It has previously been
discussed how the direction of the heat flow can be altered via the scan strategy, allowing a degree of control over
the microstructural texture [111]. This manipulation can be extended to the deposition of single-crystal Ni-based
superalloys via certain precision AM techniques [11].
As mentioned, the manipulation of the deposition characteristics can be used to deposit equiaxed, columnar or mixed
grain morphologies for simple component geometries [11, 101, 108]. Figure 2.12 presents simple examples of common
morphologies encountered with the AM of IN718. Both the columnar and equiaxed zones have been identified in the
first image, with a higher magnification image of the equiaxed region also provided. The development of these two zones
is common, with the equiaxed zone arising from the initial high solidification rate conditions. The columnar region
results from the continuous retention of heat during the remelting and deposition of consecutive layers. Ideal values
for the thermal gradient and solidification rate in the DED of simple structure Ni-based alloys have been discussed by
Thompson et al. [7]. For complex components, control over the local heat flow is a challenging process since substrate,
environmental and material properties can have complicated effects on heat-flow mechanics [17, 93]. As a result, issues
persist in controlling the microstructure of certain AM Ni-based alloy components. The common columnar texture
can result in a strong texture and resulting orientational dependence of mechanical properties. This might either be
desirable or unwanted depending upon the intended application.
The orientational dependence of elastic properties, in SLM CM247LC, has been studied by Mun˜oz-Moreno et al. [31]
demonstrating that intense heat treatments can alleviate the anisotropy. Another study by Ni et al. [134] highlighted
the difference, in mechanical properties, between longitudinal and transverse IN718 samples processed with SLM. It
was observed that the longitudinally built samples displayed a lower tensile strength but a greater elongation. This was
attributed to the anisotropy of the columnar microstructure. An example of the anisotropic elastic properties has been
provided in Figure 2.13, taken from Mun˜oz-Moreno et al. [31]. The directional growth of the columnar grains along the
build direction (BD) translates into anisotropic elastic and mechanical properties in the manufactured samples. The
author used resonant ultrasound spectroscopy (RUS) to measure the elastic moduli for samples in the as-deposited
(as-SLM) condition and heat-treated condition. It is observed that there is a higher degree of elastic anisotropy for the
as-SLM sample. This anisotropy can be alleviated with a high-temperature heat treatment that induces recrystallization
in the sample.
23
Figure 2.12: SEM images of IN718 demonstrating the different morphologies that develop during AM deposition (a)
lower magnification image indicating the (C) columnar zone, (E) equiaxed zone and (S) Substrate zone. (b) A higher
magnification image of the equiaxed zone in (a). The IN718 sample used in this analysis was deposited using LBP-DED
onto a forged IN718 substrate using standard deposition parameters.
24
Figure 2.13: The measured directional elastic moduli from an As-SLM and a heat treated sample taken from Mun˜oz-
Moreno et al. [31]
25
2.4.2 Elemental Segregation and Post-Processing
Elemental segregation during additive manufacturing has been mentioned and is heavily discussed in the literature [43,
135]. A high concentration of elements including Nb, Mo and Ti within the inter-dendritic regions of AM Ni-based alloys
has been heavily reported [17, 26, 30, 33]. An example of the described elemental segregation is given in Figure 2.14.
The occurrence of phases including TCPs, Laves, δ and η is common in the microstructure of as deposited Ni-based
alloys. For example, high-volume fractions of the Laves phase in both AM IN718 and ATI 718Plus® (718Plus) have
been well reported [12, 17]. The presence of this phase can be detrimental to the mechanical properties of Ni-based
alloys. It is therefore critical that the extent of elemental segregation and the precipitation of undesired phases is
controlled [43].
Figure 2.14: SEM analysis of LBP-DED IN718, a backscattered electron image is given in the top left. This is
supported by the Mo, Nb, and Ti elemental distribution maps determined by SEM-EDX.
As mentioned, the precipitation of such phases is directly influenced by the process parameters and the resultant
thermal distribution during deposition. Through modelling, Nie et al. [26] demonstrated how heat flow can affect
the morphology and distribution of the Laves phase in AM IN718. Under high cooling rates, the fraction of eutectic
liquid at the end of solidification is small and isolated in discrete locations between dendrite arms. An interesting
study by Chen et al. [43] described how using a flat-top (rather than a Gaussian) energy-distribution laser reduced
26
Laves formation in DED IN718. The flat-top laser gave rise to a fine particle size and discrete distribution of the
Laves phase in the microstructure. As with the previously described non-continuous energy sources, the solidification
conditions were altered leading to an increase in the cooling and solidification rates [127, 128]. The authors further
detailed how the maximum melt pool temperature is important in influencing elemental segregation. The maximum
melt pool temperature is greater in the Gaussian distribution laser giving rise to a higher cooling rate and yielding a
finer microstructure. However, as there is a higher energy concentration within the centre of the melt pool, segregation
is enhanced. This demonstrates that the cooling rate is not the only significant parameter in controlling the elemental
segregation. Images of the IN718 microstructures from this study produced using either a flat-top or a Gaussian energy
distribution laser are shown in Figure 2.15. It was discussed how the solidification behaviour imparts a resistance to
hot cracking, as cracks cannot easily propagate through a discontinuous distribution of small Laves particles [17, 26,
43].
Figure 2.15: Refinement of the Laves phase distribution in AM IN718 as presented by Chen et al.[43]. The bottom
(a), middle (b) and top (c) regions for the flat-top laser fabricated sample. The bottom (d), middle (e) and top (f)
regions for the Gaussian laser fabricated sample.
The increased Nb enrichment near the Laves particles has led to the unexpected precipitation of high-volume fractions
of the delta and eta phases during conventional precipitate ageing heat treatments in IN718 and 718Plus. Standard
post-processing regimes for these alloys were not designed to cope with severe Nb segregation, which can promote
the precipitation of these phases at lower than expected temperatures [136]. As mentioned, the use of post-processing
techniques, such as HIPing, have been cited as a remedy for many complications [17, 33, 137]. This is highlighted
in Figure 2.16 which presents SEM images of AM IN718 in the as-deposited and heat-treated condition. In the as-
deposited condition, segregation of Nb is observed leading to a significant volume fraction of the Laves phase in the
inter-dendritic regions. The Laves phase is dissolved using a subsequent high-temperature heat treatment, though there
27
is significant grain growth in both the AM material and forged substrate. This would lead to a deterioration in the
mechanical properties of the forged substrate due to over-ageing.
Figure 2.16: SEM analysis of LBP-DED IN718 deposited onto a forged IN718 substrate, demonstrating the grain
growth as a result of a solution heat treatment. (a) As-deposited condition, (b) heat-treated condition (1100 ◦C / 2
hours). The IN718 sample used in this analysis was deposited using (LBP-DED) onto forged IN718 substrate using
standard deposition parameters.
It is also noted in the literature that heat treatments designed for conventionally manufactured Ni-based alloys might
28
not be appropriate for those produced by AM [12]. Oguntuase [136] cautioned that conventional heat treatments were
not appropriate for AM 718Plus and that improper ageing could compromise mechanical properties. Other studies
have highlighted that AM Ni-based alloys displayed poorer mechanical properties than conventional counterparts due
to suboptimal post-processing [21, 96, 138]. These issues are commonly encountered when attempting to process Ni-
based superalloys using AM techniques [7, 21].
In their as-deposited state, precipitates are normally in an under-aged condition and poorly dispersed [136, 139]. Heat
treatments are required to correctly age the material and produce a uniform distribution of fine γ′ or γ′′ precipitates [1].
However, complications such as cracking, due to high residual stresses and the precipitation of undesired phases, can
occur during heat-treatment processes [140, 141]. These issues stem from the intense processing conditions of AM and
are not normally encountered with conventionally manufactured components [43, 135, 142]. Improper heat treatments
can result in thermal stresses not being properly relieved and further precipitation of undesired phases. Furthermore,
such processing can give rise to strain-age cracking occurring due to the rapid precipitation of γ′ [140]. However, with
correct post-processing, the properties of some AM Ni-based alloys are comparable (and, in some cases, superior) to
conventionally manufactured systems [143, 144]. It is clear that new heat-treatment protocols that are specifically
designed to optimise the performance of AM Ni-based alloys are required.
2.4.3 Cracking of AM Ni-based alloys
During the deposition of Ni-based alloys, a number of cracking mechanisms can occur, particularly for systems designed
to contain a high-volume fraction of γ′ precipitates [20]. Common cracking mechanisms observed in Ni-based alloys
include strain-age, ductility-dip, solidification, and liquation [25]. Zhang et al. [145] has described some of the cracking
mechanisms that occur during the DED of IN738. Cracking during AM can also result from process induced mechanisms
such as key-holing [74]. It is well known that Ni-based alloys with a high γ′ content are considered to be ’unweldable’
and, therefore, difficult to process using AM techniques. This is due to a susceptibility to specific forms of cracking,
especially that of strain-age and/or ductility-dip cracking [76, 109, 140, 146]. This processability can be directly related
to the content of Al and Ti of many Ni-based alloys [146, 147]. A weldability assessment chart which considers the
effect Al and Ti content versus that of Cr and Co content is presented in Figure 2.17. Such a chart can be used to
roughly predict the processability of a given Ni-based superalloy using its composition. However, it must be noted that
such assessments should only be used as guides and cracking in so called ’weldable’ systems might still occur due to
incompatible processing conditions [40, 42, 100].
29
Figure 2.17: A weldability assessment chart adapted from the work of Haafkens and Matthey [147] and Basak and
Das [146]
Significant welding and AM research has been completed to mitigate forms of hot cracking that plague these alloy
systems [20, 21, 42, 145]. It has been possible to reduce or eliminate some forms of cracking through altering alloy
composition and process control [17, 131]. Importantly, it has been shown that the concentration of minority elements
directly influences the susceptibility to solidification and liquation cracking mechanisms [40, 148]. Examples of types
of hot cracking commonly encountered in the AM of Ni-based superalloys are given in Figure 2.18. Chen et al. [100]
demonstrated how liquation cracking could be limited by improving the base-cooling when depositing IN718. The
authors explained that solidification cracking is mainly attributed to stresses that concentrate in the trapped liquid at
the final stages of solidification and cause tearing of liquid films within the inter-dendritic regions. These cracks are
often remelted and dissolved during subsequent deposition, though persist in the final layers of the build. However,
the final layers can be mechanically removed without deteriorating the component properties. This is in contrast to
liquation cracks which occur in the HAZ below the melt pool. These cracks cannot be resolved through subsequent
layer deposition and must be mitigated through other means. The authors highlight that susceptibility to liquation
cracks is dependent upon the orientation of grains; with a higher tendency to crack at grain boundaries, which have a
greater misorientation. The authors give a full explanation as to why this occurs [100].
Strain-age and/or ductility-dip cracking is also a common form of cracking encountered when processing Ni-based
superalloys. It is commonly attributed to the stresses generated when precipitating high-volume fractions of the γ′
phase in an uncontrolled manner. For conventional wrought and forged material this form of cracking has been overcome
30
Figure 2.18: Examples of liquation and solidification cracking in air-cooled and water-cooled samples of laser-AM
IN718 as presented by Chen et al. [100]. (a) Liquation cracking of the water-cooled sample, (b) liquation cracking of
the air-cooled sample and (c) solidification cracking of the air-cooled sample.
with careful process control and heat-treatment protocols that regulate the precipitation of the γ′ phase [1, 4] However,
difficulties with controlling the precipitation of the γ′ phase are commonly encountered with welding and AM processes
[25, 140]. In fact, high γ′ containing alloys are commonly deemed as ’unweldable’ due to their sensitivity to strain-age
cracking. Examples of these forms of cracking are given in Figure 2.19 [76]. These cracks are characterised as featuring
no liquation or solidifying regions but are rather ’clean’ and are associated with sharp kinks. A detailed explanation
of the causes of such cracking has been given by the authors [76].
Figure 2.19: Examples of solid-state cracking in the AM of superalloys CM247LC and IN939, as presented by Tang et
al. [76]. The microstructures were presented in the as-deposited condition in all cases. Features and region of interest
have been identified.
31
2.5 Alloy Design For Additive Manufacturing
There are several studies that have sought to optimise Ni-based alloys for AM techniques [53, 76, 106, 149]. Indeed, it
has been shown that an alloy can be designed that should be able to overcome the current complications of cracking,
elemental segregation, residual stress, and anisotropic growth [46]. Primary challenges include the identification of how
composition relates to each of these complications, as well as the generation of models that effectively describe all the
key phenomena occurring during production [88, 150].
Computational models which describe the process environment during AM are being developed. These models include
predictions of heat flow, elemental segregation, and residual stress development during manufacturing [26, 65, 85, 88,
150-152]. In addition, there is research to support the prediction of cracking and solidification mechanisms in Ni-based
alloys [153-157]. Section 2.5.1 below includes discussion of the prospects and ongoing efforts to design an alloy that is
suitable for AM techniques incorporating such methods. Much of the discussion is focussed on reducing susceptibility
to cracking as this is paramount.
2.5.1 Alloy Design for Crack Mitigation
Numerous studies have characterised cracking and defect formation during AM of Ni-based alloys [21, 140]. Through the
control of deposition parameters and the application of correct post-processing, cracking can be significantly reduced or
even eliminated in some systems [21, 33, 158]. This is the case for alloys that are considered weldable, such as Waspaloy,
IN718 and IN625 [25, 42, 109]. However, despite concerted efforts to address the underlying issues, for many superalloys
cracking persists and cannot be effectively alleviated using process control or available post-processing techniques [21].
These systems tend to be the ’unweldable’ high-volume fraction γ′ containing alloys. As previously discussed, the high
γ′ content of these alloys (≥ 30%) increases the susceptibility to several cracking mechanisms, most notably strain-age
cracking [25, 159]. It is often difficult to avoid such reheating during multi-pass welding or AM techniques, leading to
the ’unweldable’ term being applied to these systems [25].
If process control alone is not sufficient, then it must be accepted that these alloys might not be appropriate for AM.
Therefore, the compositions must be altered to fit the process or new alloys designed [76, 160]. Design of new alloys
compatible with AM have been undertaken by Tang et al. [76] and Conduit et al. [53]. The authors of these studies
developed novel Ni-based superalloys processable through AM whilst retaining exemplary high-temperature properties.
32
Throughout their investigation Tang et al. [76] make comparisons of the designed alloys to ’heritage’ systems including
IN939 and CM247LC. Their designed alloy systems were shown to have superior processability when compared to the
legacy alloys, whilst retaining mechanical performance.
The success of alloy design requires an understanding of the process and the numerous defect-forming mechanisms
[1, 20, 161]. Tang et al. [76] specifically sought to understand susceptibility of an alloy system to different cracking
mechanisms. With that understanding, the authors effectively designed against solidification, liquation, and solid-state
cracking. This type of methodology is key in alloy design. In the following subsections (2.5.1.1 to 2.5.1.3) there is
general discussion around the prediction and mitigation of the common forms of cracking.
2.5.1.1 Solidification Cracking
One simple relationship that has been noted in the literature is that of the freezing range to hot cracking susceptibility
[86, 153, 160, 162]. It is generally considered that alloys with larger freezing ranges are more susceptible to forms of
hot cracking. Due to the high alloying addition in superalloys, a large freezing range is quite common [1]. From this
simple observation it could be of considerable interest to investigate alloys systems with narrow solidification ranges
for AM.
Though this relationship offers a route by which improved resistance to solidification cracking can be sought, it is
arguably more appropriate to consider the full solidification behaviour. A simple model for describing solidification
strain was described by Zhang et al. [163] and could be used to predict the susceptibility to hot tearing. The authors
used both the thermal expansion and volume shrinkage coefficients to predict the change of volume on solidification.
This relationship can then be used to predict the strain at points where the molar liquid fraction is between 11% and
0.5%. This has been described as a critical temperature range (CTR) during solidification, where Ni-based alloys are
susceptible to hot cracking [25, 131].
This approach can be taken further by relating the resistance to solidification cracking to the change in volume fraction
of liquid during solidification [76, 164]. This change in solid (or liquid) fraction with temperature was used by Kou et
al. [164] to create a criterion for solidification cracking. This solidification cracking index (SCI) quoted in equation 1 is
an accessible relationship that can be used to describe the solidification behaviour of different alloys. In this expression,
dT is the change in temperature and fs is the fraction of solid solidified.
SCI = |dT/d(f1/2s )| (1)
33
This index was used by Tang et al. [76] to design alloys resistant to solidification cracking. This index was crucial in
describing the behaviour near the final stages of solidification, 0.9 < fs < 0.99. The superalloys CM247LC and IN939
exhibited sharp increases in the SCI near the terminal stages of solidification. In contrast, IN718 had a much lower
value. Interestingly, by removing Hf from CM247LC and IN939 the SCI values for these alloys was decreased. Both Hf
and Zr enrich in the liquid during the final stages of solidification and are reported to give rise to hot cracking [148,
160, 162]. Using the Thermo-CalcTM software [55] it is possible to predict the change in volume and the molar liquid
fraction at specific temperatures. From this, the effects of individual elements on the solidification behaviour near the
final stages of solidification can be assessed.
A study by Harrison, Todd and Mumtaz [105] sought to improve the cracking resistance of Hastelloy X, during selective
laser melting, through compositional modification. The authors aimed to increase the thermal shock resistance of the
alloy by increasing the concentration of solid solution strengthening elements. It was proposed that this would increase
the tensile strength of the alloy and reduce cracking susceptibility. The compositional modification achieved an average
reduction of 65% in crack densities along the vertical section and 57% in the horizontal section (see Figure 2.20).
The modified Hastelloy X (MHX) contained a slightly higher concentration of Co, Mo and W, at the expense of Ni,
compared to the original variant (OMX). Tramp element concentrations were also decreased in the modified variant.
Additionally, it was reported that the modified system exhibited a higher yield strength and a higher ultimate tensile
strength. This supported the authors’ proposed theory that tensile strength is related to the cracking susceptibility
and might be applicable to other alloy systems.
Figure 2.20: Crack densities of two variants of Hastelloy X (original and modified Hastelloy X, OHX and MHX
respectively) as presented by Harrison, Todd and Mumtaz [105]. The top graph displays crack densities in the horizontal
build orientation. The bottom graph displays crack densities in the vertical build orientation.
Thermal resistance was used by Conduit et al. [53] when optimising a Ni-based alloy for DED. A value was calculated
34
using the coefficients of Young’s modulus, 0.2% proof stress, thermal expansivity and electrical resistivity. The electrical
resistivity is correlated with the thermal resistivity and is simpler to determine. The authors proposed that cracking
would be limited if the value of thermal resistance were > 0.04 KΩ−1m−1 and an alloy with desired properties was able
to be predicted. An increased value of thermal resistance is likely to improve the thermal shock resistance, reducing
crack formation during deposition.
2.5.1.2 Solid-State Cracking
For strain-age and/or ductility-dip cracking the volume fraction of γ′ precipitates can be used to predict susceptibility.
The concentration of the γ′ forming elements is directly linked to the volume fraction of the γ′ precipitates. To mitigate
this type of cracking either the content of these elements must be strictly controlled or proper process control and/or
stress relief heat treatments must be applied [25, 158]. Such systems can be readily fabricated through conventional
means due to decades of research and experience. However, the use of AM techniques to fabricate these systems is still
developing and numerous challenges are being encountered. These alloys are especially difficult to process through AM
due to the intense thermal cycling inherent to these techniques. This acts as a form of intrinsic heat treatment and
impairs the ability to control the precipitation of the γ′. Fortunately, modifying alloy composition to avoid this form
of cracking has been well demonstrated in the literature, especially in welding research [25, 42, 140, 153, 165, 166].
Obviously, this form of cracking can be avoided by lowering the γ′ content, though this limits the high-temperature
mechanical properties [1, 25]. There is the option of using γ′′ strengthened alloys. As the solvus temperature of the γ′′ is
lower and the precipitation kinetics sluggish, alloys such as IN718 are resistant to solid-state cracking [25]. However, the
γ′′ phase is metastable and will eventually transform to the stable δ after prolonged exposure to temperatures above
650◦C [38]. The γ′′ phase can be stabilised with compositional modification, though these alloys still have limited
temperature capability when compared to systems such CM247LC [32, 167-169].
Therefore, to avoid this form of cracking, whilst simultaneously retaining a reasonable volume fraction of the γ′, the
kinetics of γ′ must be altered. This can be achieved through supressing of the γ′ solvus [76] or by altering the lattice
parameters of the γ′. The lattice parameters are altered to augment the misfit between the γ and the γ′ in such a way
as to reduce the driving force for precipitation [165, 170, 171]. In 718Plus, the misfit of the γ′ was altered by additions
of Nb to reduce susceptibility to cracking [172, 173]. In these systems, the γ′ precipitation is relatively sluggish [83,
171]. These modifications can be used to design alloys which are less susceptible to cracking during AM [53, 76].
35
In the study by Tang et al. [76] the authors used an empirical relationship to quantify the susceptibility to strain-
age cracking. The strain-age cracking (SAC) index used is stated to be a prediction of cracking tendency for a given
composition. This value was calculated by summing the content of γ′ strengthening elements with each contribution
approximately averaged by its atomic fraction. A strong correlation between this value and the γ′ precipitation driving
force was found at high temperatures, near 1000 ◦C. This correlation was not apparent at lower temperature as the
driving forces for precipitation with high undercoolings is large for all systems [76].
2.5.1.3 Liquation Cracking and the Manipulation of Minority Elements
Liquation cracking is often difficult to predict as there are many aspects which determine susceptibility [40, 106]. This
form of cracking is dependent upon: phases present and their morphology, freezing ranges, grain boundary orientation,
heat-flow mechanics, and thermal stresses. Generally, there are two critical factors: the formation of localised liquid
films and the nature of thermal stresses. It is well-known that liquid films are associated with eutectics and other low-
melting point phases [4, 47, 148, 153]. During AM processes, this form of cracking generally appears in the heat-affected
zone (HAZ) where there is continuous thermal cycling. This can cause local remelting, often of eutectic phases, and
generate stresses [25, 40, 56, 76, 78, 87, 174].
Residual stresses can be reduced through process control, though there are practical limits to this [92]. Therefore, it is
prudent to focus on the formation of liquid films during the final stages of solidification [76, 145, 160]. Tang et al. [76]
highlighted that the melting behaviour is critical. The authors discussed the importance of designing an alloy with a
sharper onset to melting. They identified the γ/γ′ eutectic as the main cause for the liquation cracking, in addition
to carbides, in CM247LC. This knowledge can be used to alter an alloys composition in an effort to avoid forming
susceptible liquid films. However, an understanding of different elemental effects, especially those of minority elements,
is required.
The role of minority elements in Ni-based alloys is complicated when determining cracking susceptibility. For example,
C, B and Zr have been reported to give rise to different forms of cracking in AM IN738 [145], while Yang et al.
[106] demonstrated that V additions could inhibit elemental segregation in laser-clad IN718. Numerous studies have
been published on the effects that minority elements, including B, C, Hf, Si, V and Zr have on the properties and
microstructure of Ni-based alloys [1, 25, 175]. However, it is still a challenge to predict what the optimum concentration
of these elements should be in a specific alloy [1, 166].
36
B, C, and Hf are referred to as grain-boundary strengtheners and can enhance the high temperature creep properties
of a polycrystalline Ni-based alloy. However, the location, type and morphology of carbides and borides are important.
Deleterious carbides and borides can be precipitated in chains along grain boundaries and within the inter-dendritic
regions [25, 45]. During AM, severe segregation can occur and lead to the formation of these undesirable phases. Melting
of these carbides and borides often occurs in the HAZ during AM, leading to the formation of liquid films [3, 35, 40,
76, 78, 121].
Hf is commonly added to increase the transverse ductility in directionally solidified superalloys and can be used to
control hot cracking [161, 176]. The formation of Hf carbides can inhibit grain growth and strengthen grain boundaries.
In addition, Hf can act as a getter to detrimental sulphur impurities [109]. However, Hf is synonymous with low-melting
point phases in Ni-based alloys. Such phases increase the freezing range and raise the susceptibility to the formation
of liquid films. As such, the reduction or removal of Hf can diminish susceptibility to liquation cracking [25, 76, 160].
Minor additions of Si have been shown to improve the stress rupture life of Ni-based alloys [177]. It is also commonly
added to enhance the corrosion and oxidation resistance and has been shown to give rise to higher elongation ratios
[177]. However, Si has been reported to give rise to hot cracking during AM, though it is unclear what the exact
cracking mechanism is [109, 178]. It has been speculated that segregation of Si during AM alters the energy of the
solid-liquid interface. This can lead to unfavourable wetting of grain boundaries by thin liquid films during solidification
and, ultimately, increase susceptibility to cracking [141]. In addition, Si affects the stability and formation of carbides.
Formation of undesired carbides can lead to crack initiation points along grain boundaries [177].
As with Si, Zr additions have been shown to be both beneficial and detrimental to Ni-based alloys depending upon the
circumstances. Under conventional processing conditions, Zr addition has been shown to increase the strength of the
grain boundaries through the refinement of carbides, extending creep life [179]. However, when added in the presence
of B, Zr can increase the susceptibility to hot tearing during casting. Cracking has been attributed to the effects that
these two elements have on the formation of grain boundaries and liquid wetting properties during solidification [162].
The detrimental effect of Zr segregation during AM has also been discussed in the literature [21, 180].
The consequences of adding these minority elements to a Ni-based alloy are dependent upon both the processing
conditions and the concentration of other elements. In order to reduce segregation through alloy design, the non-
equilibrium partitioning coefficients of strongly segregating elements would have to be determined with respect to the
thermal profile of the HAZ [43]. In addition, robust predictions of low-melting point phases through computational
analysis are required [53, 55, 76].
37
2.5.2 Control of Residual Stress and Anisotropy
In addition to cracking, the development of residual stresses and microstructural anisotropy during AM techniques are
often undesirable. These are intrinsic products of the processing; hence, manipulation of the deposition parameters
can be used to control these potential issues [92, 118]. Additionally, mitigation of these complications can be achieved
through post-processing techniques, including heat treatments and HIPing [33]. However, for specific AM applications,
such as the repair and manufacture of highly complex components, it may not be possible to manipulate the process
parameters or apply intense post-processing techniques. Thus, altering the composition of the deposited alloy might
be required for successful application.
In order to reduce the residual stress developed during deposition, the thermal properties of an alloy could be altered
[53]. While the control of thermal properties is a simple idea, it is limited in traditional Ni-based superalloy compo-
sitional space. To vastly alter the thermal resistance, systems beyond that of conventional Ni-based alloys need to be
considered. The work of Francis et al. [181] on steel welds demonstrated that the distribution of residual stress could
be controlled by manipulating the temperature at which martensite forms. In steels, the transformation from austen-
ite into plates of bainite or martensite can be influenced and induced by external stresses. The stresses can bias the
formation of plates in specific crystallographic orientations, leading to an unequal distribution of the crystallographic
variants. This stress-induced phase transformation can counteract the build-up of residual stresses during cooling if the
transformation continues to lower temperatures. For the reduction of residual stress, it might be of interest to design
alloys with stress-induced phase transformations.
While there is no analogous transformation in Ni-based superalloys, nickel-containing high entropy superalloys that
may support such effects are currently under development [182]. Specifically, it is well known that Co can be used to
reduce the stacking fault energy in Ni-based alloys [183]. Using the martensitic transformation in Co-Cr-Mo systems
as a starting point, it might be possible to design a high entropy alloy with an analogous phase transformation [184].
Solid-state transformations could also be utilised to reduce microstructural anisotropy and control the grain structure.
For example, cyclic austenitizing or recrystallization transformations have been used for grain refinement and the
reduction of anisotropy [30, 185, 186].
38
2.5.3 Application of Inoculants for Microstructural Control in Additive Manufacturing
In conventional casting, grain size control has been achieved through the use of inoculants [187, 188]. The potential of
inoculants in AM has recently been studied [189-191], with promising results being reported for Al and Ti alloys [192,
193]. Despite the potential benefits that may be derived from the use of inoculants, limited studies have been performed
to date on their use to control grain growth during the AM of Ni-based alloys [189, 191, 194, 195]. The studies by Tiparti
et al. [195] and Ho et al. [194] highlighted the potential of grain refiners for the AM of Ni-based alloys. The studies added
CoAl2O4 nanoparticles in varying concentrations to gas atomized IN718 powder for deposition through SLM. These
inoculants led to reactions within the melt pool during solidification to form a dispersion of aluminium nano-oxide
particles. In both studies, the particles acted as nucleation sites and led to a refinement of the grain structure in SLM
IN718. However, different results were observed depending on the amount and size of the inoculant added, as well as the
deposition conditions. Ho et al. [194] observed an increase in the formation of equiaxed grains and a general reduction
in the anisotropy. In contrast, Tiparti et al. [195] observed an increase in the anisotropy with the addition of higher
weight fractions of the particles. Though this effect was unintended by the authors it indicates that inoculant additions
might offer a route by which alloy microstructure and texture can be controlled. Indeed, an undesired agglomeration
of the particles was found to occur with increased additions. The reasons behind these observations were not discussed
in detail, but it was believed to be a result of the process parameters.
Recently, the addition of carbide inoculants to SLM IN718 has been investigated [191]. An improvement in the creep
properties was reported, as well as a minor decrease in the anisotropy. However, it remains unclear whether such
benefits can be achieved through the use of carbide inoculants to superalloys fabricated with other AM methods.
Continued investigation of the use of inoculants in the AM of Ni-based superalloys is therefore warranted. This could
be of particular value in repair applications, where post-processing options might be limited and, therefore, improved
microstructural control through the deposition process is beneficial.
39
2.6 Summary
This literature review is split into four sections, describing Ni-based superalloys, their processing by additive manu-
facturing and the applications of alloy design towards mitigating current complications. It is clear that new Ni-based
alloys systems that take advantage of additive manufacturing and the associated thermal-history must be designed
if the full potential of these techniques is to be realised. Compositions which are readily processible whilst retaining
exceptional mechanical properties are essential. There has been significant progress with the development of accurate
and efficient alloy design frameworks over the past two decades [61, 63, 71, 75]. These frameworks take advantage of
multiple models when designing alloys. In addition, there has been extensive development of effective property models
[49-52] and the vast experimental and theoretical databases required for software such as Thermo-CalcTM [55, 196].
This has greatly improved the prediction of microstructural and mechanical properties for alloy compositions.
In addition, there have been notable successes with the optimisation of process parameters and the enhancement
of current alloy systems [11, 191]. Through continued research and development in this field, many challenges faced
by AM of current Ni-based alloys could be overcome. In the not-too-distant future, it could be possible to rapidly
design and fabricate alloys specifically optimised for a single application. For this to be possible though, an improved
understanding of the processes involved during deposition of these alloy systems is required. Specifically, more research
is needed to fully understand component thermal history and the resultant microstructure produced for a deposited
alloy. In addition, for the effective use of machine learning, new experimental and theoretical databases containing data
on the fabrication of alloys through AM techniques would be valuable. These future repositories should contain the
properties of additively manufactured Ni-based alloys that were fabricated using comparable deposition conditions.
It is clear, from this review, that there are significant research opportunities for developing additive manufacturing
techniques and optimising, or augmenting, Ni-based alloys for additive manufacturing processes. To that end, Chapter
3 will present work on the LBP-DED of IN718. The primary goal of this section is to introduce the features of additively
manufactured IN718, an alloy currently used for LBP-DED repair applications. The deposited microstructure of this
system will be discussed, with a specific focus on the anisotropy produced during the deposition process. Interestingly,
a new texture component, not previously reported in the literature for AM IN718, was identified in this study. The
elastic properties were correlated to this microstructural texture and the effect of different heat treatments investigated.
Chapter 4 will present the effects that adding niobium carbide (NbC) inoculants to the IN718 precursor powder have
on the resultant additively manufactured microstructure. It was intended that the addition of the NbC particle would
40
aid in controlling the microstructural, and therefore elastic anisotropy developed during the deposition process. Such
control would allow for a tailoring of the microstructure. In addition, it was hypothesised that the addition of NbC
would enhance the formation of carbides at the expense of the undesired Laves phase.
The final results chapter introduces an alloy design framework that was used to optimise a new Ni-based superalloy,
based around the IN718 system, specifically for LBP-DED repair applications. The design framework was tasked with
determining an alloy composition with a higher phase stability than IN718, whilst retaining the desired performance.
A new composition, AM718R, was identified by the framework. A preliminary analysis of the properties of the mi-
crostructure and properties of this new alloy system is presented.
41
Chapter 3
Texture and Elastic Anisotropy in Laser
Blown Powder Processed Superalloy
IN718
3.1 Introduction
Challenges are encountered with the laser-blown-powder directed-energy-deposition (LBP-DED) of high-performance
alloys such as IN718, where repaired material exhibits a deficit in mechanical properties when compared to the parent
material [30, 197, 198]. This is due to the formation of undesired phases, an irregular microstructure, and pronounced
texture leading to irregular anisotropic elastic properties. To mitigate these issues, deposition parameters and post-
processing treatments must be carefully selected in order to achieve the desired properties in the final product. The
deposited IN718 is typically only subjected to a standard two-stage precipitation heat treatment following deposition
and thus, the repaired material retains undesirable features including deleterious phases, residual anisotropy, and an
irregular grain size [28].
Control of the texture is a particular priority because irregular anisotropic properties can lead to design challenges and
component instabilities when in service. Specifically, the observed anisotropy is not limited to the phase constituents,
42
but to defects as well, and this has led to wide distributions in some mechanical testing results due to an orientation
dependence [30, 31]. Careful control and characterisation of the anisotropy of these materials is therefore essential to
ensure component integrity [118, 199]. It has been shown that anisotropy in IN718 can be reduced and even eliminated
through intensive heat treatments that induce recrystallization of the deposited material [31, 118, 200]. However, certain
post-processing operations may be inappropriate for LBP-DED repaired components if the properties of the substrate
will be compromised [198, 201]. New methods are therefore required to optimise the microstructure of the deposited
material whilst minimising any deleterious effects to the substrate [7, 202].
In this Chapter, the occurrence of a previously unidentified Brass textural component ({110} <211>) in LBP-DED
IN718 samples is evaluated. The material has been characterised in the As-DED state and following heat treatments
in the vicinity of the recrystallisation temperature. This was done to assess the extent to which post-processing might
control this textual component. Microstructural characterisation was performed using scanning electron microscopy
(SEM) and chemical and phase analyses were completed using SEM-EDX and electron back-scattered diffraction
(EBSD). The evolution of bulk elastic anisotropy was also determined using resonant ultrasound spectroscopy (RUS),
with both ex-situ measurements at room temperature and in situ measurements at elevated temperatures to directly
study kinetic effects. These results were also correlated to the local textures obtained via EBSD.
3.2 Experimental Methods
Samples were produced by LBP-DED using IN718 powder of the nominal composition given in Table 3.1. The IN718
powder was prepared via gas atomisation, yielding a powder with a log-normal particle size distribution of between
40-150 µm. Subsequent LBP-DED was carried out using a commercial set-up in line with current industry practices.
A bi-linear raster pattern was used with an overlap of half the laser diameter and a specific energy of 40 J mm−1. The
deposition parameters utilised were selected to improve microstructural uniformity and avoid the linear stacking trends
that are observed as a result of co-linear and co-planar features. Cao et al. [203] and Liu et al. [132] have described
this in detail. Samples were deposited onto forged IN718 substrate coupons that have been sectioned near the base.
The section was prepared to a 1200 gt finish and plasma cleaned prior to deposition. The forged IN718 coupons were
in the fully aged conditions and are representative of conventionally processed aerofoils prepared for repair.
To study the residual elastic anisotropy of IN718 LBP-DED samples and the effect of different heat treatment regimes,
parallelepiped samples of the approximate dimensions 3.5 × 2.5 × 4.0 mm were prepared via EDM. Details of the
43
Table 3.1: Composition (wt% unless indicated) of the Ni-based alloy used in this study; in all cases the balancing
element is Ni. Details of the composition of conventional forged materials have been given for reference. The nominal
compositions are those quoted from the production facility. The measured compositions are those determined by EDX,
LECO and ICP. The LBP-DED composition is that of the powder prior to deposition.
exact masses and dimensions of the parallelepiped samples can be found in Table 3.2. The dimensions correspond
to the scanning direction (SD), the transverse direction (TD) and the build direction (BD) respectively. The parent
LBP-DED specimen from which these samples were machined was a column-like build. Metallographic techniques were
used to prepare the faces of the samples to a finish of 4000 grit. An in-house hand-polishing rig was designed that
limited the z-axis movement of the samples during polishing, ensuring that the opposing faces remained parallel to
each other.
Table 3.2: Dimensions and masses of the EDM prepared LBP-DED IN718 parallelepiped sample.
Prior to heat treatment, the samples were encapsulated in evacuated and argon-backfilled quartz ampoules to minimise
oxidation at high temperature. The specific heat treatment protocols applied are listed in Table 3.3. These heat
treatments were chosen so that the evolution of the microstructure and texture of the samples could be characterised
at temperatures near the recrystallisation point. All samples were air cooled following heat treatment.
Table 3.3: Heat Treatment protocols used during this study for the LBP-DED samples, the terms ppt and ReX refer
to precipitate and recrystallisation, respectively.
To identify the phases present in the AM samples, X-ray diffraction was performed using a Bruker D8 X-ray diffrac-
44
tometer equipped with a Cu-Kα source operated at 40 kV and 40 mA. To suppress the measurement of the Kβ peaks
a 12 µm thick Ni filter was used. In order to improve counting statistics, samples were rotated at 30 RPM during data
acquisition. An angular range (2Θ) of 20◦ to 115◦ was chosen with a step size of 0.05◦ and a dwell time of 2.8 seconds.
For higher resolution scans a dwell time of 5 seconds was used.
The anisotropies of the As-DED and heat-treated (HT) samples were calculated using elastic constants determined
using RUS. A detailed explanation of the RUS technique has been given by Migliori and Sarrao [204]. The experimental
apparatus and methods used for the present study described by McKnight et al. [205, 206]. RUS spectra were collected
at room temperature in the frequency range of 100-1200 kHz, with 50,000 data points in each spectrum. For each
sample, five spectra were collected with the samples mounted in a different orientation during collection so as to ensure
that all the resonances in this frequency range were observed. Analyses of the RUS spectra were completed in the
Wavemetrics IGOR Pro software package; with individual resonance peak frequencies identified using an asymmetric
Lorentzian function. The open-source rectangular parallelepiped resonances (RPR) code [204] was used to analyse the
resonant frequencies from each sample to calculate the elastic stiffness coefficients. Initially the RPR code calculates
the theoretical resonant frequencies using the measured density, dimensions, and estimates of the elastic moduli for
each sample, assuming an orthotropic symmetry. The elastic moduli are then calculated based on the iterative fitting
of the theoretical resonant frequencies to the experimental results. The scanning direction (SD), transverse direction
(TD) and build direction (BD) were set as x, y, and z respectively. A root-mean-square (RMS) error was calculated
for each sample to describe the quality of fit for comparison. For orthotropic samples, an RMS error of < 0.8 % was
accepted as being indicative of a reasonable fit. From the elastic constants, the Poisson’s ratios as well as the Hill
averages for the Young’s and shear moduli were calculated using the open source EIAM code [207]. Elastic anisotropy
coefficients for the samples along the cubic directions were determined following the approach described by Ravindran
et al. [208].
For high-temperature RUS measurements, samples were held lightly between the tips of two horizontal alumina buffer
rods inserted into a Netsch 1600 ◦C furnace [206]. The driving and receiving transducers were attached to the ends
of the buffer rods outside the furnace. Pureshield purity argon, filtered through a Matheson NanoChem Purifilter,
was used to purge the furnace and maintain an inert atmosphere with a flow rate of 0.2 litres / min. Spectra were
collected over a frequency range specific to each sample and contained approximately 4,000 data points. Data were
acquired during isothermal holds in 100 ◦C steps from room temperature to 1000 ◦C, with 400 seconds allowed for
thermal equilibration after each step-in temperature. For the final step to 1100 ◦C only 60 seconds were allowed for
thermal equilibration so that measurements could be taken quickly. Further scans at 1100 ◦C were taken with a gap of
60 seconds and a scan time of approximately 120 seconds. Values of the elastic moduli were determined for each scan
45
using the previously described methodology. A measure of peak acoustic loss was calculated from Q−1 = ∆ff where
Q−1 is the inverse mechanical quality factor and ∆f is the width at half the maximum height of the resonance peaks.
Following the RUS measurements, the parallelepiped samples were prepared for SEM analysis. The samples were
mounted in conductive Bakelite and polished using a standard metallographic protocol to a finish of 0.25 µm, after
which an additional chemical polish was applied using 0.04 µm colloidal silica for 5 minutes. A Zeiss Gemini SEM
450 equipped with an Oxford Instruments Symmetry EBSD detector and an Oxford Instruments X-MaxN 50 detector
was used to perform the chemical and microstructural analyses for all samples. The crystallographic orientation data
were acquired with the Oxford Instruments AZtec system. Post-analysis of the Euler angles was performed through the
Channel 5-HKL software. EBSD orientation maps were acquired with a step size of 2 µm from areas of ∼3 mm2 and
grain or cell boundaries boundary angles were limited to >5◦ for EBSD grain boundary maps. The following parameters
were used for the contouring of the cubic pole figures: a cluster size of 5◦, a half width of 15◦ and a maximum intensity
for the multiples of uniform density (mud) of 6 mud [209]. The orientation of the EBSD diagrams is indicated in the
figure captions. Further image processing and analysis were completed using ImageJ software [210]. Samples were sent
to AMG Analytical Services for minority element analysis; LECO was used for the quantification of carbon and ICP
was used for boron quantification.
3.3 Results
3.3.1 Compositional and Microstructural Analysis
The measured compositions of the IN718 samples are given in Table 3.1. SEM-EDX analysis and minority element
testing were used to confirm the compositions and to identify any deviations from the nominal compositions. Notably,
the concentrations of the minority elements in the As-DED samples were higher than quoted in the nominal powder
composition. This is consistent with impurity pick up through deposition giving rise to higher-than-expected concen-
trations of C and O in the deposited alloy. Thus, a higher volume fraction of carbides is to be expected in the LBP-DED
samples.
Figure 3.1 shows a low magnification SEM composite image of the microstructures seen in the build direction (BD),
scanning direction (SD), and transverse direction (TD) displayed as a parallelepiped representation. A zig-zag mi-
crostructure produced by the bi-linear scan strategy can be observed along the build direction. This is most evident
46
Figure 3.1: Composite SEM images of the microstructure of LBP-DED IN718 in the planes normal to the build
direction (BD), scanning direction (SD) and transverse direction (TD). The images are presented in a parallelepiped
representation to show the orientational dependence of the microstructural sections
on the BD-SD face. The SD-TD face shows clear evidence of the cellular structure with bright contrast of the phases
occurring in the intercellular regions.
Two XRD patterns obtained from the As-DED sample are given in Figure 3.2. Further patterns from the samples
in the heat treatment condition F are given in Figure 3.3. In the first XRD pattern (Figure 3.2a) the γ phase peaks
can be readily identified, consistent with previous reports [43, 211]. However, the superlattice reflections from the γ′
and the γ′′ phases are not readily discernible. This is to be expected given the low scattering contrast from these
reflections. The second XRD pattern (Figure 3.2b) shows data obtained in the vicinity of the feature identified by the
arrow in Figure 3.2a. This indicates the presence of MC type carbides and the Laves phase, with diffraction peaks at
approximately 35◦, 37.5◦ and 41◦ 2Θ [199, 211]. The peaks at approximately 35◦ and 37.5◦ 2Θ can be attributed to
the MC carbide {111} and Laves phase {0110} reflections respectively [43, 212, 213].
The peak at 41◦ 2Θ can be attributed to the MC carbide {200} phase [43, 113, 211]. There was no detection of the
{0220} reflection of the Laves phase, which normally occurs at 45◦ 2Θ. The peaks at 35◦ and 41◦ 2Θ were retained
following homogenisation heat treatment at 1100 ◦ C for 2 hours (Figure 3.3), supporting the assertion that they are
associated with the MC carbide phase [43]. The precipitation heat treatment B had no discernible effect on the XRD
pattern. As such, only X-ray diffraction data from the As-DED sample are presented in Figure 3.2. Given that the
compositional analysis revealed higher than expected carbon concentration, the presence of the MC-type carbides is
unsurprising. It is also noted that greater carbon concentration has been reported to lead to a reduction in the Laves
phase in IN718 as MC-type carbides form in preference [4]. Further evidence for the occurrence of the MC carbide
phases is provided by the SEM analysis. The occurrence of both phases is well reported within the field [28, 43, 110,
113, 211] and, thus, further characterisation was not pursued.
47
Figure 3.2: (a) XRD pattern of the As-DED IN718 sample. (b) magnified view of the XRD pattern in the vicinity of
the arrow shown in (a). The reflections from the γ phase, MC carbide and Laves (φ)phases are labelled.
48
Figure 3.3: ((a) XRD pattern of the IN718 sample in the heat treatment F condition. (b) magnified view of the XRD
pattern in the vicinity of the arrow shown in (a). The reflections from the gamma phase, MC carbide and Laves phases
are labelled.
49
Backscattered electron images of the DED IN718 samples, following heat treatments A, B and F along with associated
SEM EDX elemental distribution maps of Cr, Mo, Nb and Ti, are given in Figure 3.4. Backscattered electron images
and SEM EDX elemental distribution maps of samples in As-DED, heat treatment C and heat treatment D conditions
are given in Figure 3.5. As expected, the γ′ and γ′′ phases cannot be resolved at the low magnification in Figure 3.4
because of the very high cooling rates encountered in additive manufacturing, which results in very fine precipitates
[28, 214].
A typical cored, dendritic microstructure was observed in both the As-DED sample and the sample subjected to
heat treatment A, with inter-dendritic segregation of the refractory elements. The segregation of such elements is
well reported in the literature for additively manufactured Ni-based alloys, including IN718 [28, 215]. From the EDX
elemental maps in Figure 3.4 numerous inter-dendritic precipitates can be identified. EDX point analyses revealed most
of these phases to be rich primarily in Mo, Nb and Ti. Higher Cr content was also identified in several precipitates,
and such phases are likely to be Laves or carbides. As noted above, the XRD results indicate the presence of MC-type
carbides and the Laves phase. This is supported by the lack of Cr in some of the precipitates, as well as the high
content of Mo, Nb and Ti. Such phases are commonly reported in other studies, including the work of Parimi et al.
[110] where carbides, the δ-phase and the Laves phase were shown to be present in the as-fabricated condition.
The inter-dendritic structures are seen to be absent from the microstructure after the relatively short solution period
of 10 minutes at 1080 ◦C in heat treatment B, with only fine precipitates rich in Nb and Mo being observed. These
precipitates are again expected to be carbides, likely MC carbides, as characteristic reflections from this phase could be
identified in the post-homogenisation XRD pattern between 35◦ and 41◦ 2Θ [211], see Figure 3.3. The microstructure
for the sample in condition C is very similar to that of B which is to be expected for the similar heat-treatment
temperature and duration. A small degree of recrystallisation was also observed in Samples B and C, though the
microstructure remained largely dendritic and deformed grains can be identified. The higher temperature of 1100 ◦C
experienced by the samples subjected to heat treatments D and F resulted in a microstructure that comprised deformed
or recrystallised grains, along with a more regular distribution of coarser carbide particles.
50
Figure 3.4: SEM analysis of DED IN718 samples in different conditions As-DED, A and B. For each sample, a
backscattered electron image is shown at the top, below which are the Cr, Mo, Nb, and Ti elemental distribution maps
determined by SEM EDX.
51
Figure 3.5: SEM analysis of DED IN718 samples in different heat-treated states C, D, and F. For each sample, a
backscattered electron image is shown at the top, below which are the Cr, Mo, Nb, and Ti elemental distribution maps
determined by SEM EDX.
52
3.3.2 Texture Analysis
EBSD pole figures and inverse pole figure (IPF) maps showing the microstructural texture of the As-DED state and
samples subjected to heat treatments A, B and C are given in Figure 3.6. The EBSD pole figures and IPF maps for
samples in heat treatments conditions D, E and F are given in Figure 3.7. The IPF maps were generated with respect
to the x axis and describe the texture along the build direction. These are accompanied by pole figures orientated down
the x axis which serve to provide an alternative representation of the measured texture. From these data, a strong
<011> and a weak <001> texture is observed along the building direction. This texture is intensified following heat
treatments B (10 minutes at 1080 ◦C), C (10 minutes at 1090 ◦C), D (10 minutes at 1100 ◦C) and E (30 minutes at 1100
◦C), before diminishing in heat treatment F (60 minutes at 1100 ◦C). The strong {011} texture has been identified as
a Brass component ({011} <211>). The associated pole figures reveal that the texture maxima correspond to that of
the ideal Brass texture component spots. In terms of (Bunge) Euler angles the Brass texture is offset from the normal
cube texture by 35◦ in φ1 and 45◦ in Φ [216, 217].
A fibre texture analysis with respect to the build direction is presented in Figure 3.8. This identifies the misorientation
of grains with respect to a specified texture component along the BD. In this figure, a false colour scale was used
to represent the extent of misorientations relative to the {001} and {011} components. A darker colour indicates a
closer alignment of a grain to a specified texture component. Grey areas correspond to regions with a misorientation
of greater than 20◦ to the specified component. In the As-DED condition there is prominent zig-zag grain structure
with a strong {011} texture along the BD and a weaker {001} texture. Following heat treatment F, an equiaxed grain
structure is observed suggesting that recrystallisation had occurred. Notably, the strong {011} component parallel to
the BD was retained, although the {001} component appeared to have diminished.
3.3.3 Anisotropic Elastic Properties
The elastic stiffness coefficients determined from the RUS data, and the calculated elastic moduli of all the samples,
are presented in Table 3.4. As the resonant modes of the samples have a predominantly shear character, the shear
components exhibit the lowest uncertainty and are therefore the most reliable [31]. The associated uncertainty for the
calculated elastic moduli was consistently below 2.5 %.
Following the approach presented by Ravindran et al. [208], the elastic anisotropy of the samples in the cubic shear-
planes was quantified into three terms, A100, A010 and A001 (Eq. 2 – 4), where an isotropic single crystal would yield
53
Figure 3.6: Left – Inverse pole figure maps with respect to the build direction (BD) and scanning direction (SD)
for LBP-DED RUS samples in the As-DED state and following heat treatments A, B, and C. Right – Corresponding
{001}, {011} and {111} pole figures in the BD plane. The idealised spots for the {011} <211> Brass component are
identified by black spots [216].
54
Figure 3.7: Left – Inverse pole figure maps with respect to the build direction (BD) and scanning direction (SD) for
LBP-DED RUS samples following heat treatments D, E, and F. Right – Corresponding {001}, {011} and {111} pole
figures in the BD plane. The idealised spots for the {011} <211> Brass component are identified by black spots [216].
55
Figure 3.8: Fibre texture analysis of the 001 and 011 components in the As-DED and F samples. The colour scale
represents the misorientation of grains (maximum 20◦) with respect to the 001 and 011.
Table 3.4: Elastic constants and properties for the LBP-DED IN718 parallelepiped samples in their respective heat
treatment (HT) conditions. The sample number has been included in brackets beneath the heat treatment condition
letter. The elastic constants were calculated with sample dimensions for the RD, TD and BD representing the x (1), y
(2) and z (3) axes, respectively. An error of approximately ± 1% is associated with the C11-C33 and C12-C23 constants
and an error of approximately ± 0.1% is associated with the C44-C66 constants. B – effective bulk modulus, E – effective
Young’s modulus, G – effective shear modulus, RMS - root mean square error.
56
values of 1 for all terms. The equations derived by Ravindran et al. [208] have been slightly modified to analyse the
RUS data from the textured polycrystalline samples rather than a single crystal as used in the original study. The
denominator terms for Eq. 3 and 4 were altered to better reflect the anisotropy down the defined cubic shear planes.
The dominator terms in Eq. 3 and 4 have been modified for the RUS analysis of the polycrystalline samples here. These
modifications to the original equations have allowed for description of the anisotropy down the cubic shear planes of
the samples. These values are not intended to provide a full description of the elastic anisotropy of the samples, but
rather serve as a measure for the degree of anisotropy present. The samples were treated as if they were equivalent to
orthorhombic single crystals, exhibiting anisotropic elastic properties.
A100 =
4C44
C22 + C33 − 2C23 (2) A010 =
4C55
C11 + C33 − 2C13 (3) A001 =
4C66
C11 + C22 − 2C12 (4)
Using these equations, elastic anisotropy factors have been calculated and plotted in Figure 3.9a for the samples
described in Table 3.4. The elastic anisotropy factors of the samples following the other heat treatment conditions are
given in Table 3.5. A plot of the orientational dependence of the Young’s modulus of samples in the As-DED state and
following heat treatment G determined from the RUS data is given in Figure 3.9b. These plots show a clear reduction
in the elastic anisotropy along the cube axes after 120 minutes at 1100 ◦C. This reduction is interesting since it was
shown previously that the microstructural texture was not eliminated by ageing at 1100 ◦C, Figures 3.6 and 3.7.
57
Figure 3.9: (a) Shear anisotropy factors, where [001], [010] and [100] correspond to the cubic shear planes of the RUS
samples. (b) The orientation dependence of the Young’s modulus for samples in the As-DED state and following heat
treatment G with respect to the TD and BD.
The results in Figure 3.9a show little change in the elastic anisotropy along the cube shear planes following the
standard precipitation heat treatment (A), but a progressive reduction in elastic anisotropy with increasing time
and temperature (from C to G). The former is expected as the temperatures are too low to initiate recrystallisation
mechanisms. Short duration heat treatments which influence the anisotropy are of particular interest as they may prove
to be less deleterious to the microstructure of the substrate in post-processing of components repaired by LBP-DED.
To investigate the effect of short duration heat treatments around the recrystallisation temperature, RUS was used to
monitor how individual resonant modes responded to heat treatment temperature and duration. Figure 3.10a displays
the room temperature frequencies for two resonant modes of an As-DED sample before and after a heat treatment for
58
Table 3.5: Anisotropy coefficients A100, A010 and A001 for the LBP-DED IN718 parallelepiped samples in their
respective heat treatment (HT) conditions.
approximately 30 minutes at 1100 ◦C. The shift in the two resonant modes is quite clear, though the time scale at
which this transformation takes place is not. To determine the timescale of the peak-shift, the system was set up to
scan a small frequency range every minute. Spectra were collected while the sample was held at 1100 ◦C. The results of
these measurements are given in Figure 3.10b, where the spectra obtained during the heating cycle, isothermal hold at
1100 ◦C, and cooling cycle are given. In the figure the red traces correspond to spectra obtained during the isothermal
hold, while the black and blue traces correspond to the spectra obtained on heating and cooling, respectively. As a
guide to the eye, circles have been added to a single resonance peak, allowing for the tracking of this peak through the
different thermal regimes.
The spectra obtained from the isothermal hold show the progressive variation in the positions of the peaks due to
microstructural evolution. Comparison of the heating and cooling curves enable the difference in the relative peak
positions to be identified. It should be noted that the frequency range shown in Figure 3.10b is different to that in
Figure 3.10a to accommodate thermal softening effects on the resonant peaks at high temperature, shifting them to
lower frequencies [218]. The effect is greater for higher frequencies, causing the resonant modes to move closer together.
Pronounced broadening of resonant peaks also occurs at temperatures near 950 ◦C due to an increase in acoustic loss
Q−1. Broadening from a high acoustic loss complicates peak identification against the high temperature background.
A further complication is the high-temperature enhancement of resonant modes from the alumina holders, which can
interfere with sample resonant modes. This obstructs the identification of the exact frequency of individual resonant
modes at high temperature. Excessive broadening and interference at temperatures greater than 1000 ◦C hindered the
tracking of the circled peak for the final heating spectrum measured at 1080 ◦C. The true position of this peak at 1080
59
◦C was not identifiable but likely lies within the range of 395 – 405 kHz. During the isothermal hold at 1100 ◦C this
peak was observed to reappear and become enhanced in intensity with increasing heat treatment duration.
Figure 3.10: (a) - Room temperature RUS spectra for an LBP-DED sample before (red) and after (blue) an extended
ageing at 1100 ◦C (b) - High temperature RUS spectra of an LBP-DED sample collected during heating (black), during
the isothermal hold (red) and during cooling (blue). Circles are shown above individual resonances to allow tracking
through the thermal cycle, though this was not possible when peak broadening became too great to allow resonant
modes to be distinguished clearly from the high-temperature background.
Sharp increases in the acoustic loss can indicate the start of processes such as the diffusion of elements and the
dissolution of precipitates [218]. A single peak from a high temperature RUS spectrum, analogous to that given in
Figure 3.10b, was chosen for analysis throughout the recorded temperature range, see Figure 3.11a. The acoustic loss
calculated for this peak has been given in Figure 3.11b. The event around 1100K in 3.11b is attributed to the formation
of the γ′′. The event near 1200K is attributed to the dissolution of phases prior to melting of the matrix. It is intended
that this spectroscopy tool can be further developed such that initiation and termination of precipitation events can be
accurately measured. This would allow for a new method of designing heat treatment protocols optimised for specific
60
alloys.
Figure 3.11: (a) - High temperature RUS spectra of an LBP-DED parallelepiped sample collected during heating,
a circle (blue) is shown above the first resonant peaks in the series of interest. An arrow (blue) has been added as
a guide to the eye to allow tracking of the peaks through the temperature range. (b) – Acoustic loss plotted against
temperature for each resonant peak in the series of interest.
61
3.4 Discussion
The steep thermal cycles generated during LBP-DED are expected to influence grain growth, leading to a degree
of texture orientated parallel to the direction of local heat flow. As with other AM methods, this is normally the
building direction [118]. The fastest crystal growth direction for FCC Ni-based alloys is along <001> directions.
Hence, for deposition on a polycrystalline substrate, preferential growth typically occurs in grains which have a <001>
crystallographic direction parallel to the direction of local heat flow [111]. Where there is no local <001> orientation
parallel to the heat flow, there is nucleation of new grains. A competitive growth mechanism occurs, where the <001>
orientation is preferred, and this leads to the formation of a columnar microstructure with a strong crystallographic
orientation [111].
In this study, a strong < 011 > texture has been identified as a Brass component ({011} <211>). The occurrence
of selected grains with a Brass texture component have been observed in the wire and arc additive manufacturing
(WAAM) of filler metal ERNiCrFe-7A [219]. However, to the best of the author’s knowledge a pronounced Brass
texture component has not been previously identified in additively manufactured Ni-based superalloys. Rather, the
closely-related Goss component ({011} <100>) is more commonly observed in DED Ni-based alloys, as is well described
by Ma et al. [117]. Nevertheless, it may be anticipated that under certain deposition conditions the local direction of
heat flow is altered, thereby changing the growth direction of the γ phase. This, in turn, may lead to the development
of different textural components. In the present study, it appears that the deposition conditions have altered the local
heat-flow direction towards the laser scan direction. Specifically, the bi-directional contour pattern has modified the γ
growth direction such that the <100> are aligned at two acute angles across the build direction. Such alignment can be
seen within the microstructures shown in Figure 3.1 and gives rise to the Brass texture observed in the corresponding
EBSD analysis.
Explanations for offsets in the texture from the typical cube component have been presented in the literature [111, 117,
219]. An excellent example has been provided by Dinda et al. [111], where a bidirectional scan strategy gave rise to a
rotated cube texture. To explain the observed textural rotation, the layer-by-layer growth mechanics in a bidirectional
scan strategy must be considered. As the local direction of the thermal gradient changes with the scanning direction
of the laser, the nucleation and growth of the primary dendrites is altered. As described by Dinda et al. [111], the
bidirectional scanning of the laser leads to a change in the growth angle of the dendrites by ± 45◦ between layers. The
microstructure was shown to comprise large <100> dendrites with a directional orientation of 45◦-60◦ to the interface
of the previous layer, the result of epitaxial growth from the secondary dendrites of each previous layer. As the epitaxial
growth occurs only from the <100> grains with a 45◦ orientation, the number of grains is reduced to larger columnar
62
grains in the upper part of the build. It can be seen in Figures 3.6 that the As-DED microstructure in the present
study has a similar layer-by-layer orientational relationship to that described above. These large columnar grains have
been observed in the later stages of the LBP-DED build employed in this study. In this case, the combination of
deposition parameters has altered the growth angle of the dendrites. However, unlike the previous studies, the angle is
± 55◦ between layers. This corresponds to the offset of 35◦ between the typical <100> cube direction and the <211>
direction.
When considering the results presented here, together with the work of Dinda et al. [111] and Ma et al. [117] it is
clear that control of the scan strategy is critical in determining the final texture. Both the thermal gradient (G) and
growth rate (R) have been described in many studies as key factors in the resulting microstructures of AM Ni-based
alloys [110, 129] . In the absence of other factors, columnar growth is expected in systems with a higher G/R ratio
and equiaxed growth is promoted with a lower ratio. However, the effect of other factors such as heat retention and
local fluctuations in the thermal gradient may affect the growth behaviour. Therefore, to mitigate the microstructural
anisotropy of a DED build, control of these thermal factors and careful selection of scan strategy is important [110,
111, 117].
The As-DED pole figures orientated on the SD and TD direction in Figures 3.6 and 3.7 also yield further information
about the texture in the samples. A strong {111} component was observed along the SD, and when considering the
degrees of rotational freedom along the three crystal axes, this will be a {111} component which is orthogonal to
the BD {110} component. The texture along the TD was seen to be less pronounced and more random, though a
preference was observed for {001} and {111} components. The TD is the less constrained direction, which is reasonable
considering the heat-flow governing growth in the other directions. These results provide insights into the orientational
dependence of the elastic properties of the samples, with the SD expected to be the stiffest direction and the BD the
least stiff direction given the intrinsic elastic anisotropy of Ni-based alloys.
Previous studies have shown that intensive high temperature heat treatments, such as hot isostatic pressing (HIPing)
reduce, or in some cases, eliminate texture through recrystallisation [118]. However, the results shown in Figure 3.8,
indicate that the short duration furnace heat treatments considered near the recrystallisation temperature do not
reduce the anisotropy. The work by Messe´ et al. [214] on IN738LC showed that much higher temperatures are required
for complete recrystallisation and concomitant reduction of anisotropy; while recovery was identified as the dominant
process occurring during heat treatments near 1180 ◦C. Interestingly, the work of Kunze et al. [118] showed that heat
treatment at 1180 ◦C did not heavily affect the grain size, morphology, or texture in SLM IN738LC. Such intensive
heat treatments were not the focus of this study, which instead focussed on the microstructural effects at lower heat
63
treatment temperatures as it is critical to preserve the substrate properties in components repaired using LBP-DED.
These results have shown that such heat treatments do not eliminate the microstructural anisotropy. This is likely due
to the position of carbides which sit along the grain boundaries and sub-grain boundaries. At these heat treatment
temperatures, the carbides would not be expected to dissolve. Due to Zener drag, the carbides have the capacity
to restrict the growth of the recrystallised grains towards the original <211> growth directions preferred due to the
deposition strategy. At higher temperature, where the carbides are less stable and the driving forces for recrystallisation
are greater; it is expected that the nucleation and growth of the grains will be less affected by the carbide populations.
This would probably lead to a more equiaxed microstructure, as reported in the studies of Messe´ et al. [214].
A persistent, strong microstructural texture has significant implications for the mechanical properties of an AM com-
ponent [28], with the anisotropy giving rise to different mechanical responses depending on the testing direction. This
creates design challenges as certain directions could be more susceptible to mechanical failure as well as to instabilities
in service [118, 214]. In view of this, knowledge of the orientational dependence of the stiffnesses is critical when de-
signing components. The results obtained in this study also highlight that the heat treatments significantly affect the
elastic properties of the samples, with a general reduction in the residual anisotropy after relatively short treatments.
This suggests that the microstructural symmetry is shifting towards axisymmetric from orthotropic.
A similar shift toward an axisymmetric symmetry was observed by Mun˜oz-Moreno et al. [31], though the microstructural
texture was also shown to be reduced. This behaviour of varying elastic anisotropy with heat treatment might be
rationalised with consideration of the single crystal elastic constants along the <001>, <011> and <111>. It has been
shown that the <001> are low stiffness directions while the <111> have high stiffness in face centred cubic materials
such as Ni-based alloys [220, 221]. It is also known that the stiffness of <011> lies in-between that of the <001> and
<111> and is also close to the intersection of the Voigt and Reuss models of polycrystal stiffness [221, 222]. Therefore,
a strong texture aligning the {011} with a cartesian axis would result in a value of Cii for that axis, close to the average
value expected from a random grain structure. This would explain the apparent reduction in the elastic anisotropy
despite the {011} texture being enhanced through the presented heat treatments. The Young’s moduli along these
directions for IN718 have been presented by Aba-Perea et al. [222]. The reduction in the values of the C11 and C22
following an extended heat treatment might also be a result of the enhancement of the {011} at the expense of {100}
and {111} orientated grains.
Further research is required to support this hypothesis and this phenomenon could prove to be useful in the future
design of components. Specifically, it might be possible to significantly reduce the elastic anisotropy, as observed, with
short duration localised heat treatments, thus improving the performance of repaired components. The potential of
64
optimising such short duration heat treatments is demonstrated by the data presented in Figure 3.9b, which show
that the resonant modes and, therefore, the elastic properties of the sample change within a short period time at high
temperature. Within twelve minutes at 1100 ◦C a significant shift had occurred. This is interesting since only minor
changes in the microstructural texture were observed for the DED sample aged, for a similar length of time. Such data
are useful in the optimisation of heat treatment protocols as they provide the ability to determine the optimum ageing
time. Although further development of this research is required for such optimisations to be realised.
3.5 Conclusions
In this Chapter the microstructure and elastic properties of LBP-DED IN718 have been presented. Using EBSD
mapping it was possible to study the texture under different post-processing conditions. These data were supplemented
by RUS which allowed for the characterisation of the elastic properties. It was observed that the elastic anisotropy did
not fully correlate with the microstructural anisotropy. Key observations and conclusions are:
A Brass component has been identified as a new texture in LBP-DED Ni-based superalloy IN718. This adds to the work
presented by Ma et al. [117] and emphasises the importance of deposition conditions in determining the microstruc-
ture produced. This texture was enhanced by post-processing heat treatments near the quoted recrystallisation start
temperature of IN718. Such results could be used to guide future post-processing conditions for alloys exhibiting this
texture.
The RUS analyses highlighted that in several cases the apparent elastic anisotropy was significantly reduced whilst
the microstructural anisotropy was observed to increase. High temperature RUS showed that the resonant modes and
therefore the elastic constants are sensitive to small duration heat treatments. Due to the sensitivity of the resonant
modes, this technique might prove useful in the future optimisation of both heat treatment duration and temperature.
There is a significant presence of the Laves phase within the microstructure of LBP-DED IN718. This eutectic phase
forms because of the segregation of strengthening elements during solidification. The removal of this phase is possible
using post-processing. However, the use of such post-processing will cause grain growth and affect the physical proper-
ties. Further research is required to determine methods of manipulating the microstructural and phases formed during
the LBP-DED process.
65
Chapter 4
Modification of Superalloy IN718 for
Additive Repair Applications Through
Inoculant Addition
4.1 Introduction
Though AM techniques are highly versatile, challenges remain with controlling the microstructure of deposited alloys
[21]. Alloys fabricated through AM generally have a columnar microstructure reminiscent of directionally solidified
materials, with a high degree of texture [89]. Mechanical performance is dependent upon the microstructural texture
and is therefore anisotropic [185]. If an equiaxed microstructure is desired, then either intense post-processing or
carefully controlled deposition conditions are required [223].
For specific applications, some microstructural anisotropy might be desirable; although, a high level of control over the
anisotropy would be required to avoid component instabilities in service [186]. The continuous intense thermal cycling
that occurs during AM fabrication can lead to additional deformation and recrystallisation mechanisms [89]. These
effects can be particularly problematic in the fabrication of components with irregular geometries due to the variations
66
in local heat-flow. This, in turn, can cause specific regions in the component to have varying heat retention, resulting
in an irregular microstructure and non-uniform properties [21, 89]. Therefore, methods to control the growth of the
microstructure during fabrication need to be developed further. One possible route to controlling the microstructure
during additive manufacturing is through the use of inoculants.
In this Chapter, the effects of NbC inoculants on the elastic properties and microstructure of laser-blown-powder
directed-energy-deposition (LBP-DED) IN718 are investigated. The NbC particles were chosen as this phase is normally
present in the form of MC-type carbides within the microstructure of IN718. This reduces the likelihood of introducing
additional phases which could be detrimental. One weight percent of the NbC inoculant was added to the IN718
powder, this value was chose based on previous observations in the literature (Section 2.5.3). It was hypothesised that
one weight percent of NbC would influence the solidification of the microstructure without leading to the occurrence
of detrimental phases or significant cracking. To evaluate the effect of the NbC inoculants, the microstructure, texture,
and mechanical properties of LBP-DED IN718 samples were evaluated. Comparisons were made between reference
IN718 samples and the inoculant containing samples in the As-DED and heat-treated conditions. Correlations were
determined between the observed microstructural texture and the measured elastic properties. The addition of the NbC
particles increased the volume fraction of MC-type carbides and decreased that of the Laves phase. It was found that
the inoculant-containing samples exhibited a marginally increased hardness and an enhanced Brass texture component
{110} <211>. The occurrence of this textural enhancement is theorised to be a result of the NbC particles restricting
dendritic growth along specific directions during build. The addition of the inoculant therefore offers a method of
achieving a degree of microstructural and textural control during additive manufacturing.
4.2 Experimental Methods
Samples of standard IN718 and NbC modified IN718 (IN718-NbC) were prepared using LBP-DED with deposition
conditions identical to those in Section 3.2. Gas atomisation (GA) was used to produce the IN718 precursor powder.
For the modified samples, one percent by weight of NbC inoculant, with a size distribution of 5-10 µm, was added
via direct mixing under inert atmosphere to the standard (GA) IN718 powder. Compositional analysis of the samples
post-build was performed by AMG Analytical Services. For quantification of carbon content, the LECO method was
used and inductively coupled plasma optical emission spectrometry (ICP-OES) was used for the other elements.
Two post-deposition heat treatments were chosen to assess the impact of the inoculants on the ageing response and
subsequent mechanical properties of the samples. The full details of the heat treatments used are given in Table 4.1.
67
The first heat treatment was a standard two-stage precipitation heat treatment for IN718. A solution heat treatment
was not applied to samples prior to precipitate ageing, and this was omitted so that the effect of the inoculant on the
microstructure and phases could be resolved. The second heat treatment was a recrystallisation heat treatment at 1100
◦C, the reported recrystallisation start temperature for IN718 [224]. A temperature at the start of recrystallisation was
chosen so that any effect of the inoculant on grain structure would be most apparent. Prior to all heat treatments,
the samples were encapsulated in argon-backfilled quartz ampoules. This was done to minimise oxidation at high
temperature. The Thermo-CalcTM [55] software was used for solidification and equilibrium phase predictions using the
TCNI8 database.
Table 4.1: Details of the heat treatments applied to the LBP-DED samples in this study.
Differential scanning calorimetry (DSC) was performed using a Netzsch 404 calorimeter with disc-shaped specimens
having a diameter of 5 mm and a thickness of 1 mm. The specimens were prepared using electro-discharge-machining
(EDM). For all measurements, the samples were placed in an alumina crucible under flowing argon at a rate of ∼50
mL/min. DSC data were acquired during heating to 1400 ◦C and cooling to 450 ◦C, at a rate of 10 ◦C/min with a 10
min isothermal hold between steps to allow the sample to thermally equilibrate.
X-ray diffraction (XRD) was performed using a Bruker D8 diffractometer operated in identical conditions to those
given in Section 3.2. Analysis and fitting of the diffraction data were undertaken using the Bruker DIFFRAC.EVA
software package.
Microstructural imageing and compositional analysis were performed using a Zeiss GeminiSEM 450 scanning electron
microscope (SEM) equipped with both Oxford Instruments X-MaxN 50 energy dispersive X-ray spectroscopy (EDX)
and Symmetry electron back scattered diffraction (EBSD) detectors. Samples were prepared as described in Chapter 3.
Identical protocols to those described in Chapter 3 were used for the operation of the SEM and detectors. Aztec software
was again used for the data analysis [209]. On the EBSD maps and pole figures obtained, the axes that corresponded
to the build direction (BD), scanning direction (SD), and the transverse direction (TD) were labelled. The open-source
ImageJ software [210] was used for further image processing including particle analysis and the determination of phase
area fractions.
68
Transmission electron microscopy (TEM) was performed on 3 mm diameter discs prepared from an LBP-DED sample
using EDM. A twinjet electropolishing rig was used to thin the samples to electron transparency with a solution of 6%
perchloric acid in methanol at –8 ◦C under an applied potential of 20 V. An FEITM Tecnai Osiris TEM operated at 200
kV was used to perform EDX compositional analysis through scanning transmission electron microscopy (STEM-EDX)
and to acquire selected area diffraction patterns (SADP) from the phases present.
Resonant ultrasound spectroscopy (RUS) was used to measure the elastic properties of the samples in the As-DED and
heat-treated conditions. The experimental apparatus and parameters used are identical to those described in Section
3.2. Parallelepipeds of dimensions 4.9 × 3.8 × 3.1 mm were cut from the as-built LBP-DED IN718 specimen using
EDM. The faces of the parallelepiped samples were polished using 4000 grit SiC paper, with care taken to ensure that
opposing faces remained parallel to each other. The exact masses and dimensions of each sample have been provided
in Table 4.2. As with EBSD data, the dimensions are reported with respect to the BD, SD and TD. The positions of
each resonant frequency in a spectrum were identified using an asymmetric Lorentzian function in the Wavemetrics
IGOR Pro software package.
Table 4.2: Dimensions and masses of the parallelepiped samples LBP-DED IN718 and IN718-NbC samples.
The resonant frequencies were analysed using the open-source rectangular parallelepiped resonances (RPR) code [204]
as described in Section 3.2. Anisotropy coefficients were calculated for the samples using an approach similar to that
described by Ravindran et al. [208]. The open source EIAM code [207] was used to calculate the Hill averages for the
Young’s and shear moduli from the fitted elastic constants.
Isothermal oxidation testing at 650 ◦C was carried out to determine the effect of the NbC inoculants on the environ-
mental resistance of IN718. For these tests, samples of LBP-DED IN718 which had undergone heat-treatment A (Table
2) were cut from the builds with dimensions 20 × 10 × 1 mm and the surfaces ground using 4000-grit SiC paper. A
Setaram Setsys Evolution thermogravimetric (TGA) instrument was used to measure the mass gain of samples during
a 200-hour exposure at 650 ◦C. Samples were hung from a microbalance within the instrument with a continuous
stream of air flowing through at 30 mL/min and a pressure of 1 bar. The acquired data were fitted to Eq. 5 (Pierragi
Method) [225] to obtain the parabolic growth rate constants Kp (mg2 cm−4 s−1). In this equation ∆m is the mass gain
in mg, s is the surface area in cm2, t is the exposure time in seconds and mo is a variable to account for the mass gain
associated with the transient oxidation prior to the establishment of parabolic growth.
69
∆m
s
= mo +
√
Kpt (5)
Hardness testing of the samples was performed using a QNESS Q30 A+ tester with a Vickers tip and a starting load of
30 kgf. Samples of IN718 and IN718-NbC following heat-treatment A with dimensions 20 × 10 × 2 mm were cut from
the LBP-DED IN718 builds. The specimens were mounted in Bakelite with the build direction exposed and polished
to a 4000-grit finish using standard metallographic techniques before hardness testing.
4.3 Results and Discussion
4.3.1 Compositional and Microstructural Analysis
The compositions of the standard LBP-DED IN718 and IN718-NbC are given in Table 4.3. The compositions of both
alloys are within the acceptable tolerance ranges quoted for commercially supplied IN718. As expected, the IN718-
NbC sample has a noticeably higher carbon content than the IN718, accompanied by a slight increase in Nb content.
However, this remains within the acceptable ranges for commercial IN718. Consequently, a higher volume fraction of
carbides is expected in the IN718-NbC samples.
Table 4.3: The measured compositions of LBP-DED IN718 and LBP-DED IN718-NbC, a combination of certified
techniques (ICP-OES for the major elements and B, LECO for C) was used to assess the composition of each element
(wt% unless indicated, Ni balance).
The Thermo-CalcTM software package was used to predict the effect of the increased Nb and C content on the phase
composition of the IN718-NbC. Phase predictions were obtained under equilibrium conditions at 400 ◦C and assuming
Scheil solidification. The predicted volume fraction and solvus temperature for key phases are listed in Table 4.4.
Importantly, the compositions used for the modelling of both systems were those given in Table 4.1. It is noted that
these models do not accurately describe the solidification behaviour during the LBP-DED processing. In general, from
the predictions it is expected that the volume fraction and solvus temperature for the MC-type carbide will be increased
in IN718-NbC compared with the IN718.
70
Table 4.4: Phase composition and solvus temperatures predicted for selected phases in IN718 and IN718-NbC.,
calculated with the Thermo-CalcTM software package assuming Scheil solidification and equilibrium to 400 ◦C. The
quoted volume fractions were taken from the predictions at 400 ◦C under equilibrium conditions. All solvus points are
given in ◦C
The phases present in the samples were initially investigated using XRD and XRD patterns from the IN718 and
IN718-NbC samples following heat treatment A are given in Figure 4.1. Note that there were no discernible differences
in the XRD patterns obtained from samples in the As-DED condition. For both samples, the γ phase peaks are
readily identifiable in Figure 4.1a. However, the γ′ and the γ′′ superlattice reflections are not distinguishable from the
primary γ matrix peaks. Such observations for IN718 are consistent with previous reports in the literature [90]. Several
reflections of lower intensity were observed within the patterns. Figure 4.1b shows higher resolution XRD patterns
taken over a narrower 2Θ range where these reflections were observed. The peaks at approximately 35◦ and 37.5◦ 2Θ
may be attributed to the MC carbide {111} and Laves phase {0110} reflections respectively [43, 212, 213]. The peak
at approximately 41◦ 2Θ may be attributed to the MC carbide {200}. The peak at 45◦ 2Θ can primarily be attributed
to the Laves phase {0220} reflection [226, 227]. The final peak identifiable in Figure 4.1a, at approximately 58◦ 2Θ was
attributed to the MC carbide {220} reflection.
Additional spectra were measured for the samples following recrystallisation heat treatment B; see Figure 4.2. The
peaks corresponding to the MC carbide phase were retained along with the γ phase reflections. Further confirmation
of the presence of the identified phases was performed using DSC, SEM and TEM. Due to the strong residual texture
inherent to AM material, quantitative analysis of the phase fractions was not possible from the XRD data.
71
Figure 4.1: (a) XRD patterns of the IN718 and IN718-NbC samples in the heat treatment A condition. (b) higher
resolution XRD patterns for the IN718 and IN718-NbC samples over a selected range of 2Θ. The reflections of the
gamma (γ), MC carbide and Laves (φ) phases are labelled.
72
Figure 4.2: XRD patterns for LBP-DED IN718 (a) and IN718-NbC (b) samples in the heat treatment B condition.
The reflections of the matrix γ phase and MC type carbide have been labelled.
The solvus temperatures for the primary phases in IN718 and IN718-NbC were measured using DSC and compared
to the Thermo-CalcTM predictions. The DSC traces for both samples are given in Figure 4.3. Both samples appear
to follow the well described solidification sequence for IN718; L → L + γ → L + NbC/γ → Laves/γ [87, 228]. The
temperatures corresponding to the approximate onset and termination of the peaks are given in Table 4.5. On first
inspection the thermograms appear to be very similar. However, there are subtle differences, notably the dissolution
event near 1300 ◦C is different in the two samples. This event is attributed to the carbide solvus and occurs at a higher
temperature in the IN718-NbC sample. The carbide solvus temperature was determined to be 1292 ◦C and 1304 ◦C
for the IN718 and IN718-NbC samples respectively. The increased solvus temperature for the carbide in IN718-NbC
is of similar magnitude to the CALPHAD predictions in Table 4.4. Interestingly, both these values are greater than
that reported in the literature for additively manufactured IN718 [87], which may be attributed to the higher carbon
content in both these samples.
The IN718-NbC sample was also found to have a higher liquidus temperature on cooling (1318 ◦C) than IN718 (1310
◦C). However, the solidus temperatures for the samples, were found to be similar, being 1143 ◦C for IN718 and 1145
◦C for IN718-NbC. The approximate solidus point was derived from the L + NbC/γ → Laves/γ reaction and agrees
with the reported literature for IN718 [229]. There are several other peaks in the thermogram that can be identified
and attributed to both dissolution and precipitation events in IN718. The event starting at 1060 ◦C for both samples
is attributed to the δ and Laves phase solvi; the range being comparable to those reported previously [87, 230, 231].
The major event starting at approximately 750 ◦C and ending near 950 ◦C can be attributed to the precipitation of
the γ′′ and δ phases [230, 231]. The minor event near 550 ◦C is associated with the precipitation of the γ′ [230].
73
Figure 4.3: DSC traces for the first heating of LBP-DED IN718 and IN718-NbC samples from room temperature to
1400 ◦C. (b) A magnified section of the trace to highlight the differences in carbide solvus temperature between the
samples.
74
Both samples were analysed with back-scattered electron (BSE) imageing and SEM-EDX compositional analysis to
study the morphology and composition of the phases within the microstructure. SEM-EDX elemental distribution
maps for Cr, Mo, Nb and Ti and the associated BSE images for both IN718 and IN718-NbC in the As-DED condition
are shown in Figure 4.4.
Table 4.5: Approximate onset and termination point for thermal events observed in the DSC analysis of IN718 and
IN718-NbC. All temperatures are given in ◦C.
Additional microscopy of the samples following heat-treatment A and B are provided in Figures 4.5 and 4.6. No
significant differences were observed between the microstructure of the samples in the As-DED condition and following
heat-treatment A. This is not surprising as the two-stage precipitation heat-treatment should not significantly affect
the MC carbides or Laves phases. It was, however, observed that the inter-dendritic precipitates were largely dissolved
after heat-treatment B, leaving only discrete spherical precipitates. This agrees with the XRD data, which indicated the
presence of the Laves phase in the As-DED condition, but not following heat-treatment B, in which only the MC carbide
and matrix peaks remained. Due to the extremely high solidification rates associated with additive manufacturing, the
γ′ and γ′′ precipitates are extremely fine and could not be resolved at this magnification in the SEM. This is consistent
with the previous reports in the literature [28, 189]. In the recrystallised condition following heat treatment B, large,
twinned grains were observed with carbides precipitated in the inter- and intra-granular regions.
The EDX elemental distribution maps (Figures 4.4 and 4.5) indicate a cored dendritic microstructure for the As-
DED and precipitation heat-treated sample, with numerous inter-dendritic precipitates rich in the refractory elements,
consistent with previous reports [28, 43]. The inter-dendritic Laves phase and more spherical MC carbide precipitates
are readily apparent in both samples. Additionally, the carbides are mainly situated along the grain and sub-grain
boundaries, with limited intragranular formation. However, there are several important differences between the samples.
Image analysis was completed on five sets of electron images taken from different sites and samples for IN718 and IN718-
NbC to quantify these differences. Due to the variability in heat flow during additive manufacturing, different locations
in a build can have noticeably different microstructures, as described by Tian et al. [232]. Hence, several sites needed to
be investigated to assess the microstructure and determine the average sizes and distribution of the visible precipitates.
75
Figure 4.4: SEM analysis of IN718 and IN718-NbC samples in the As-DED condition. Top, a backscattered electron
image. Beneath, elemental distribution maps for Cr, Mo, Nb, and Ti determined by SEM-EDX.
76
Figure 4.5: SEM-EDS analysis of IN718 and IN718-NbC samples in the heat treatment A condition. An electron
image of the sample is given at the top, below this are elemental distribution maps for Cr, Mo, Nb, and Ti obtained
by SEM-EDS.
77
From these analyses several observations could be made. For the IN718 sample a combined average area fraction of
3.2 ± 0.3% was determined for the Laves and carbide precipitates. When only measuring the carbide precipitates the
area fraction was found to be 0.37 ± 0.10%. For the IN718-NbC samples the combined precipitate area fraction was
determined to be 2.9 ± 0.3% and 0.58 ± 0.11% for the carbide area fraction. The size of the carbide precipitates was
also found to be slightly larger in the IN718-NbC samples, 700 nm compared to 300 nm in IN718. In contrast, the
size of the Laves phase particles between the two samples was found not to vary significantly. However, the Laves
phase particles were found to have formed more discretely in IN718-NbC. The reduced fraction of the Laves phase was
consistent with the CALPHAD predictions in Table 4.4, which indicated a small decrease in the volume fraction of
the Laves phase on solidification for the IN718-NbC compared to the IN718. This decrease in the Laves phase due to
an increase in carbon has been previously reported [4]. Interestingly, both samples show a similar average grain size
of ∼2500 µm2, suggesting no appreciable refinement of the grain structure has occurred during build. The absence of
microstructural refinement and the effect on the size of the Laves particles have not been reported previously. Finally,
the sub-grain boundaries and inter-dendritic regions were comparable in size for both samples.
The difference in volume fraction of carbides and Laves particles in the IN718-NbC sample may be expected to influence
the mechanical properties. Indeed, a higher volume fraction of carbides is known to lead to an improved tensile strength
[233]. Additionally, the results of Hsu et al. [191] showed that improved creep properties may be achieved through the
inclusion of the carbide inoculants. To gain insight into the effect of the NbC additions on the mechanical properties
Vickers hardness testing was used. The IN718-NbC sample was found to have an average hardness of 450 ± 15 HV. This
is in comparison to the standard IN718 which had an average hardness of 430 ± 10 HV. These results are comparable
to those reported in the literature [28] for LBP-DED IN718. The small change in hardness is to be expected due to the
minor increase in volume fraction of the carbides in the IN718-NbC sample.
78
Figure 4.6: SEM-EDS analysis of IN718 and IN718-NbC samples in the heat treatment B condition. A back-scattered
electron image of the sample is given at the top, below this are elemental distribution maps for Cr, Mo, Nb, and Ti
obtained by SEM-EDS.
79
Figure 4.7: TEM analysis of LBP-DED IN718-NbC in the heat treatment A condition. An annular dark field electron
image is shown at the top left; accompanied by selected area electron diffraction patterns for the matrix [001], MC
carbide [112] and C36 Laves [1¯100]. Below are elemental distribution maps determined by STEM-EDX.
The Laves and carbide precipitates were investigated further using STEM-EDX to probe the chemistry of these phases
80
as well as that of the matrix. STEM-EDX data from the IN718-NbC sample is given in Figure 4.7 and clearly highlights
the presence of the carbides and Laves phases. The [001] SADP from the matrix region includes superlattice reflections
from the γ′ and γ′′, indicating the presence of these phases. No direct orientation relationship was found between the
matrix and the Laves phase, however the [1¯100] for the C36 Laves was observed to be reasonably close to the [001] for
the matrix. For the carbides a γ [100] // carbide [112] relationship was found. The complementary data obtained from
the IN718 sample have been given in Figure 4.8, and it is noted that the morphologies of the phases for both samples
are comparable.
Table 4.6: Compositional analysis of the phases in the IN718 and IN718-NbC samples obtained from STEM-EDX
point spectral analysis. The given compositions were averaged over five-point spectra. The compositions have been
given in atomic percent (at%). There is a standard error of ± 1% for all measurements.
The average compositions of the carbides, Laves phase and matrix for both samples were determined using STEM-EDX
point spectra (Table 4.6). Each composition given is the average of five EDS point spectra taken from the specified
phases. The measured composition of the phases in the IN718 sample are similar to results reported in the literature
[234, 235]. However, the results clearly indicate an influence of the NbC inoculants on the primary phases. Whilst the
composition of the matrix for both alloys is similar, the other phases exhibited more pronounced differences. Most
notably, the composition of the MC carbide phase for the two alloys differs markedly. In the IN718-NbC sample the Nb
content is enhanced when compared to the standard IN718 sample, which may be expected given the NbC additions
to the alloy. Interestingly, the Ti content within the MC carbide phase is also elevated in the IN718-NbC sample. This
might be attributable to Ti being a strong MC carbide former [4] and the increase in carbon is likely to have increased
the segregation of Ti to the carbide phase.
It is reasonable to assume that the difference in composition and the increased carbide content would affect the
oxidation resistance of IN718-NbC [236]. Thus, TGA was used to measure the mass gain after 200 hours at 650 ◦C for
IN718 and IN718-NbC. The results for the TGA analysis have been provided in Figure 4.9. The mass gains recorded
are extremely small, near the mass detection limit for the TGA. Clearly, undulations due to the temperature changes
during the night and day have slightly affected the results. However, the final mass gain values are valid. For the IN718
sample the parabolic rate constant was found to be 9.5×10−6 mg2 cm4 s−1 which is consistent with the literature
for isothermal testing at 650 ◦C [225, 237]. For the IN718-NbC sample the parabolic rate constant was found to
81
Figure 4.8: TEM analysis of LBP-DED IN718 in the heat treatment A condition. An annular dark field electron
image is shown at the top left; accompanied by selected area electron diffraction patterns for the matrix [001], MC
carbide [112] and C36 Laves [1¯100]. Below are elemental distribution maps determined by STEM-EDX.
82
be 8.7×10−6 mg2 cm4 s−1. In addition, the final mass gains of the two samples were also similar. Consequently, the
oxidation behaviour of the two alloys is considered comparable.
Figure 4.9: Mass gains of LBP-DED IN718 and IN718-NbC samples during isothermal oxidation at 650 ◦C for 200
hours.
EBSD analysis of the heat-treated samples indicated that NbC additions also affected the recrystallisation behaviour
of IN718. Both IN718 and IN718-NbC fully recrystallised after two hours at 1100 ◦C. EBSD maps and pole figures
for both these samples can be seen in Figure 4.12. However, the extent of grain growth following recrystallisation was
significantly different between the two samples. Grain size distribution maps derived from the EBSD data for both
samples are given in Figure 4.10 and highlight the difference between the two samples.
For both IN718 and IN718-NbC the recrystallised fraction was greater than 90%. The difference in recrystallised grain
size between the two samples is distinct, with the IN718 samples having a much larger average grain size of 4200 µm2,
compared with 2500 µm2 for the IN718-NbC sample. It is likely that the increased volume fraction of carbides in the
IN718-NbC samples has given rise to enhanced Zener pinning [1, 4]. Similar suggestions have been made in studies
using other inoculants to AM IN718 [189, 194, 195].
83
Figure 4.10: Grain size distribution maps for the recrystallised (heat treatment B) IN718 and IN718-NbC samples.
The false colour scale indicates the approximate size of a grain in microns.
84
4.3.2 Texture and Elastic Analysis
A highly textured columnar microstructure is typically formed along the build direction under standard DED building
conditions. The origin of this texture is well described by Dinda et al. [111] and commonly leads to a pronounced [001]
texture in AM Ni-based alloys. However, other textural components, including the Goss and Brass textures, have been
achieved using different deposition conditions [117, 238]. The deposition conditions used in this study were shown to
give rise to a Brass textural component {011}<211> in Chapter 2. To evaluate the influence of the NbC additions
on the residual texture, EBSD mapping was performed on IN718 and IN718-NbC samples in different heat-treated
conditions. The EBSD analysis for the As-DED samples is shown in Figure 4.11. The accompanying analysis for the
samples in the heat-treatment A and B conditions has been given in Figure 4.12. As can be seen in the inverse pole
figure (IPF) maps in Figure 4.11, a zig-zag columnar microstructure exists in these samples as a result of the bi-linear
raster scan strategy employed during deposition. The laser is scanned up and down the scanning direction (SD), altering
the direction of local heat flow during deposition. This local change in the heat flow causes the columnar grains to
grow in a zig-zag fashion along the build direction (BD). The variation in the columnar growth direction is shaper in
the IN718 sample due to the smaller grains.
For both samples the displayed texture in the IPF maps is with respect to the BD. The dominant grain growth direction
for both samples is [011], however there is more variability in the IN718 sample. For IN718 the columnar grain growth
in this region alternated between [011] and [111] orientated grains. The selective growth mechanism that leads to this
behaviour was described in Chapter 2 and in more detail by Dinda et al. [111]. The presence of the NbC inoculants
has significantly reduced this variability and selectively enhanced the [011] grain growth. The [011] texture component
was identified directly using the associated pole figures in Figure 4.11. The pole figures are orientated with respect to
the BD, so that the observed texture spots are consistent with the IPF maps.
As observed in Chapter 3, a Brass texture was identified for the LBP-DED IN718 samples in the As-DED and precip-
itation heat-treated conditions [238]. Despite the small variation in orientation along the BD, the spots on the cubic
pole figures for the IN718 match reasonably well with the idealised overlaid spots for a Brass texture. However, the
pole figures from the IN718-NbC sample include additional spots that appear to be related by symmetry. To identify
the origin of these additional spots, EBSD pole figures have been calculated from individual columnar grains, identified
as regions A, B, and C in the IPF map from IN718-NbC in Figure 6. The associated pole figures show that for the top
and bottom grains (A and C) one set of spots is observed, whereas for the middle section (B) the reflected set of spots
is observed. The columnar grains in region A and C possess identical [011] growth directions whilst the [011] growth
direction of the columnar grain in region B has been inverted. This inversion is expected to be a result of the selected
85
Figure 4.11: Inverse pole figure maps with respect to the build direction (BD) texture for the IN718 and IN718-NbC
samples in the As-DED condition. Both the build direction (BD) and scanning direction (SD) have been identified.
Below each IPF map the corresponding cubic pole figures down the BD have been provided. For both samples, pole
figures are included for the identified regions A, B, and C to highlight the differences in observed texture symmetry.
The idealised spots for the 011 <211> Brass component are identified by black spots in the complete pole figures.
86
deposition parameters, specifically the bi-linear scan strategy which alters the local heat-flow direction. Interestingly,
the IN718 sample shows contrasting behaviour, with a noticeable difference in columnar grain orientations, as indicated
by the IPF map colouring but have a similar Brass texture orientation in successive regions A, B and C.
Increased anisotropy due to particle addition was observed by Tiparti et al. [195], although grain refinement was not
observed in their study. Similarly, it is conjectured that the higher number density of minority phase particles in
IN718-NbC restricts continued grain growth during building, ensuring that grain orientation becomes more dependent
upon the local heat flow characteristics. However, the grain structure and texture are more homogeneous in the IN718-
NbC sample. For specific applications textural control may be valuable and, as such, the use of inoculants like NbC
could prove to be useful [89, 111]. However, further studies are warranted in order to assess the extent to which the
texture may be controlled by this method. Application of post-deposition heat treatments may be expected to affect
the texture. Indeed, for both samples in the recrystallised condition an equiaxed microstructure was observed (see
Figure 4.12), which is consistent with the literature [31, 194, 195].
The EBSD analysis indicates that the addition of NbC had a strong influence on the residual texture of LBP-DED
IN718-NbC. The presence of a strong microstructural texture would be expected to give rise to directionally dependent
mechanical properties. It is therefore important that the orientational dependence of the mechanical properties is
accounted for during component design. This is to ensure that there are no unforeseen in-service instabilities that arise
due to directional variation of the mechanical properties [186]. To directly investigate this effect, RUS measurements
were performed, and the elastic stiffness coefficients obtained from the IN718 and IN718-NbC samples are given in
Table 4.7. It is worth noting that the shear components have the lowest uncertainty as the resonant modes of the
samples have a predominantly shear character [204].
For all samples the root-mean-square (RMS) errors were below 0.8% indicating reliable fitting of the data. On inspection
it is clear that in the As-DED condition both IN718 and IN718-NbC samples have a high degree of anisotropy and
do not display the isotropic elastic constants observed with equiaxed microstructures [31]. Notably, the results for the
samples following heat-treatment A are similar to those in the As-DED condition. This is to be expected as there should
be no significant change in the elastic anisotropy during precipitation heat-treatment of IN718. However, the elastic
constants for the IN718-NbC are significantly different to the standard IN718. These results are in good agreement with
the EBSD analysis, which showed a strong texture in the As-DED state. The results also show that in the recrystallised
state the elastic properties tend towards cubic symmetry, which is consistent with previous reports [31].
The elastic constants can be used to calculate directional elastic anisotropy [238]. The approach used here follows that
described in Chapter 3 with the elastic anisotropy coefficients in the cubic shear planes A100, A010 and A001 being
87
Figure 4.12: Inverse pole figure maps with respect to the build direction (BD) and scanning direction (SD) for the
LBP-DED IN718 and IN718-NbC samples in the heat treatment A and B conditions (Left). Corresponding pole figures
for 001, 011 and 111 poles in the BD plane of the samples.
88
Table 4.7: Elastic constants and average moduli for IN718 and IN718-NbC samples in the As-DED and heat-treatment
B conditions. The sample dimensions for the scanning direction (SD), transverse direction (TD) and build direction
(BD) were used as the x (1), y (2) and z (3) axes respectively in calculations of the elastic moduli. Full details of the
specimen used can be found in Table 4.2. An error of approximately ± 1% is associated with the C11-C33 and C12-C23
constants and an error of approximately ± 0.1% is associated with the C44-C66 constants.
obtained from equations Eq. 2 – 4. The anisotropy coefficients were obtained for each sample using the data presented
in Table 4.6 and these results are presented in Figure 4.13. The addition of NbC can be seen to have a significant effect
on the anisotropy along the [001] and [010] directions, which correspond to the BD and SD respectively.
Figure 4.13: (a) Anisotropy coefficients, calculated from equations (2), (3) and (4) for the IN718 and IN718-NbC
samples. The [001], [010] and [100] directions correspond to the cubic shear planes of the samples.
It is likely that the addition of the NbC inoculant led to restricted growth along these directions, giving rise to the
enhanced texture and anisotropy. It has been well reported that carbides preferentially precipitate along grain and cell
89
boundaries during the AM of Ni-based superalloys and their presence influences the growth of cells and dendrites [78,
239]. No such growth restrictions were observed in the study by Hsu et al. [191]. This difference could be attributable
to the higher level of additions employed in the present study.
Following recrystallisation heat-treatment B, the anisotropy factors tended towards unity for both samples, albeit
there was a clear degree of residual anisotropy in the IN718-NbC sample. It is likely that the increased volume fraction
of carbides influenced the growth of new grains, resulting in a residual texture and associated elastic anisotropy.
Importantly, the recrystallisation heat treatment eliminated the elastic anisotropy in the IN718 sample, consistent
with findings in the literature [31]. The microstructural control offered by NbC additions may be beneficial for specific
applications of AM Ni-based alloys and further studies with varying NbC content are warranted.
4.4 Conclusions
In the work described in this chapter, NbC inoculants were added to IN718 powder for AM via LBP-DED. The addition
of the inoculants had several notable effects on the LBP-DED deposited IN718 which might prove to be beneficial.
The addition of the inoculants increased the C and Nb concentrations in IN718. This increased the volume fraction of
MC-type carbides and marginally decreased the fraction of the Laves phase. In addition, there was a slight refinement
in the size of the Laves phase. The presence of the additional carbides limited grain growth during recrystallisation
owing to an increased Zener drag pressure, and the additional carbides in the microstructure led to a slight increase in
the hardness of the LBP-DED samples in the precipitation heat-treated condition. The additional Nb and C produced
no significant change in the oxidation properties at 650 ◦C.
The addition of the inoculants enhanced the original Brass texture of IN718 fabricated via LBP-DED and resulted
in alternating textural orientations between successive columnar grains where the scan direction was reversed. The
resultant microstructure was more homogeneous than that of IN718 under the same build conditions, which might be
of value for specific applications. The textural changes were reflected in the elasticity measurements, which showed a
higher degree of anisotropy in the as-built inoculated variant along the building and scanning directions. Importantly,
the use of the NbC inoculant was seen to offer a degree of control over the resultant microstructure, texture, and elastic
properties of LBP-DED IN718.
90
The ability to tailor the microstructure and texture during AM is highly desirable within the field. However, issues still
persist with segregation of the strengthening elements leading to precipitation of undesired phases. In order to address
this issue compositional changes will have to be made. It was shown in this study that additional carbon content
reduces the presence of the Laves phase. However, the content of Mo and Nb continued to drive the precipitation of
this phase. Alloy design is required to optimise the composition such that this phase is reduced without compromising
the desired physical properties. An optimised variant of IN718 could be used in conjunction with inoculants to yield
the greatest results.
91
Chapter 5
Optimisation of Superalloy IN718 for
Laser Blown Powder Repair Applications
5.1 Introduction
The deposition of Ni-based superalloys using additive methods is associated with particular complications [7, 16,
21]. These include micro/macro cracking, high levels of residual stress, microstructural inhomogeneity, and a strong
microstructural texture [7, 21]. Some of these complications can be mitigated through process parameter control and/or
post-processing [16, 93, 102, 214, 238]. Indeed, great success has been achieved with the more processable Ni-based
superalloys, which typically have a low precipitate volume fraction (∼20 - 25 γ′/ γ′′) [21, 76]. Some studies have also
reported success with depositing high γ′ containing superalloys, such as IN738LC (∼55 γ′) [158, 160]. However, many
of the reported successes rely on tight control of the process parameters and the use of intense post-processing methods,
including hot isostatic pressing (HIPing) and complicated heat treatments [33]. For certain applications it might not
be feasible to achieve such fine control or apply certain post-processing steps. This is certainly the case when utilising
AM for certain component repair applications.
In Chapter 2, the challenges associated with the LBP-DED deposition of Ni-based superalloy IN718 were presented.
Currently, only precipitation ageing heat-treatments can be applied to this deposited material when used for repair
92
applications. These heat treatments have a relatively low temperature and, as a result, do not cause over-ageing of the
substrate. However, the microstructure of the material after such a treatment is largely the same as that of material
in the As-DED condition. The pronounced microstructural texture and presence of the undesired Laves phase remain
prominent. In Chapter 3 it was shown that a degree of control over the microstructural texture may be achieved
using inoculants. In addition, it was observed that the volume fraction of the undesired Laves phase could be reduced
via compositional manipulation. However, it is clear that an alloy system specifically design for LBP-DED repair
application is required.
In this chapter, a neural network framework is used to design a new Ni-based superalloy that surpasses the performance
of IN718 for LBP-DED repair applications. The alloy compositional design space investigated by the framework was
based on that of IN718 and 718Plus. IN718 and 718Plus are loosely related alloy systems, sharing similar operating
temperatures and mechanical properties [172, 231]. Indeed, 718Plus was designed as an improved alternative to IN718
with increased γ′ content through altering the element concentrations including Al, Co and Fe as well as including W.
This allows 718Plus to operate at temperatures of approximately 700◦C, whilst maintaining a yield strength around
1100 MPa [172]. Critically, both IN718 and 718Plus are highly processable alloys due to their relatively low fraction
of strengthening precipitates (∼20 - 25 γ′/ γ′′) [4, 172]. There has been reported success for the fabrication of these
alloys using AM techniques [28, 136]. However, both alloys have exhibited pronounced segregation of strengthening
elements to the interdendritic regions and require annealing prior to precipitation heat-treatment to achieve the best
performance. Such intensive post-processing may not be possible in a repair application due to the aforementioned
rationale [238-240]. Whilst IN718 can be used without an annealing heat-treatment, the persistence of the Laves phase
within the microstructure following deposition is undesirable. Therefore, an alloy which possesses similar mechanical
properties and processability of IN718, with reduced occurrence of undesired phases is required. To achieve this, further
alloy design is needed to identify compositions more amenable to AM processing.
To demonstrate the benefits that may be derived from computational alloy design frameworks, with minimal changes
from established alloy compositions, the composition of IN718 was used as the foundation for optimisation. The allowed
compositional ranges of several elements were altered within the machine learning framework. To bring the predicted
alloy in line with 718Plus, the element W was also permitted in the alloy composition, and higher limits allowed on
both Al and Co content. The composition with the highest probability of satisfying target properties including phase
stability, solidification strain, and tensile strength was identified. The alloy was fabricated, and the properties were
experimentally investigated. The properties of the new alloy were benchmarked against IN718 to demonstrate the
performance improvement.
The following section (5.2) will outline the alloy design framework utilised in this study, specifically the neural network
93
tool. The targets for the desired alloy, designated AM718R, will then be specified. Section 5.2 will also present the
optimisation of an alloy based around the IN718 and 718Plus compositions. Experimental results for phase stability,
cracking resistance, hardness, and oxidation resistance will be presented and verify the performance of the alloy. Finally,
to confirm the suitability of the alloy for additive manufacturing a laser pass investigation is used as it offers rapid
assessment without the need for powder production and deposition [241].
5.2 Methodology
The neural network framework used in this study has been described by Conduit et al. [46, 53]. This section (5.2)
includes the key details of this framework and its utilisation during the study. The stated goal of this framework is the
optimisation of an alloy composition most likely to simultaneously fulfil multiple specified targets. For each property
considered in the design framework, the tool uses a predictive model. Crucially, this framework can identify and exploit
both composition-property and property-property relationships. This allows the framework to use a large amount of
data from one property to extrapolate for a second property that has less available data. In addition, the framework
calculates the probability that a designed composition will fulfil a target specification.
As such, multi-dimensional design space is searched for the composition with the highest probability of successfully
fulfilling the specified targets. For this study, the composition of IN718 was used as a foundation for the explored
design space. The composition of 718plus was also considered, giving rise to the inclusion of W in the design space
and a greater upper boundary to Al and Co content. The ranges for several other elements were also altered and
the selected space has been given in Table 5.1. The selected design space was searched using a random walk with a
step size comparable to the accuracy with which the composition could be fabricated, 0.1 wt%, with the possibility
of microsegregation excluded. The neural network framework typically takes ∼ 1 minute to search for an optimal
composition from a set of ∼ 100 design variables. Note that due to the intended AM repair application, any designed
system would be limited to heat-treating at a maximum of 720 ◦C to avoid overageing the substrate.
94
Table 5.1: Composition design space selected for this alloy design framework. Elemental concentration ranges are
given in wt %.
5.2.1 Target Alloy
The targets set for this new Ni-based superalloy were based around wrought IN718, with a specific focus on phase
stability and compatibility with additive manufacturing repair processes, whilst retaining other properties including
strength. The target specifications chosen for this alloy are given in Table 5.2. The phase stability target was set
at > 99 wt %, thereby limiting the fraction of undesired phases such as the Laves to < 1.0 wt %. The target for
percent of strengthening phases was set at ≤ 25 wt %. Both the γ′ and γ′′ strengthening precipitates were allowed to
contribute to the total for the strengthening phases within the γ-matrix. This limit was selected as alloys possessing
greater volume fraction of strengthening phases are often less processable due to cracking susceptibilities [4]. The target
solvus temperature for the strengthening precipitates was set at ≥ 650 ◦C to ensure stability of the precipitates at
the operating temperature. It was noted for γ′′ strengthened systems that the stability of δ phase limits the allowable
service temperature. For the prediction of the fractions of the thermodynamically stable phases, the framework utilised
CALPHAD results calculated with the Thermo-CalcTM TTNI8 database to increase the efficiency and reliability of
some calculations [55].
To ensure the designed alloy was processable through additive manufacturing two key targets were set for freezing
range and solidification strain. These values serve as measures of the susceptibility of the alloy to cracking and are
therefore critical for predicting processability. An equilibrium freezing range of < 280 K and a solidification strain of
< 0.027 were set as targets for these values. Note that the neural network tool used both equilibrium solidification
and Scheil model data in the determination of solidus and liquidus temperatures. It has been reported that a lower
95
Table 5.2: Table of properties predicted, and the method used for the prediction. In addition, the range of data and
number of entries used to train the neural network has been provided. The final two columns show the prediction and
targets for each property of the designed alloy.
96
freezing range contributes to higher resistance to solidification and liquation cracking [76]. However, the contribution
of volume shrinkage during solidification is also an important factor. An equation, from the work of Zhang et al. [162,
163], was used to predict solidification strain, E; where α is the thermal expansion coefficient and β is the volume
shrinkage coefficient of the liquid. Th and Tl represent the boundary temperatures during solidification, both have a
corresponding value for the volume fraction of the liquid fl.
E = α (Th − Tl) + β(1/3) (fl (Th)− fl (Tl)) (6)
α =
1
V1
· V1 − V2
T1 − T2 (7)
β =
V1 − V2
V1
(8)
The data for these coefficients and temperatures were derived from ThermoCalcTM predictions near the solidus and
liquidus points. The volume shrinkage coefficient is proportional to the change in volume upon solidification and,
as a result, the solidification strain. The boundary points Th and Tl used by Zhang et al. [162] were the carbide
solvus temperature and the gamma prime solvus temperature respectively. For this alloy design framework, these
points were altered and now represent the temperature at the points where the molar liquid fractions are 11% and
0.5% respectively. These points describe a common critical temperature range during solidification where Ni-based
superalloys are susceptible to hot cracking [4, 163].
The mechanical property targets chosen were similar to the standard values for wrought IN718. The relationship
between composition and mechanical properties were predicted from a large database of experimental results [46,
53]. The properties of cost and density were also considered important, as any designed alloy needs to be competitive
compared to commercial systems. The predictions for these values was made using the weight fractions and commercial
prices for the elements and binary master alloys. To ensure reliable predictions, the accuracy of the model was confirmed
through 5-fold cross-validation (see model validation, Section 5.2.3).
97
5.2.2 Neural Network Formalism
The design variables for the neural network were limited to the elements: Al, B, C, Co, Cr, Fe, Mo, Nb, Ti and W,
with a Ni balance reducing the number of variables by one, see Table 5.1. A maximum heat-treatment temperature of
720 ◦C was adopted because this is the temperature limit used for the post-processing of repaired IN718 to prevent
substrate over-ageing. Therefore, training data had to correspond to alloy compositions that had been heat-treated
at or below this temperature, limiting the mechanical data for certain alloys. However, the neural network framework
used is capable of identifying the links between properties, allowing for the large dataset to guide the extrapolation
of the incomplete data. This feature of the neural network framework has been described and successfully used by
Conduit et al. [46, 53]. Using this strategy, it is possible for the framework to make accurate predictions without the
use of detailed models describing the mechanistic origins of physical properties.
5.2.2.1 Data Extrapolation
To handle incomplete data, the neural network framework is able to use the information embedded within property-
property relationships. This can allow for interpolation and extrapolation of predictions for sparsely populated datasets.
For example, an accurate model trained on copious phase behaviour data can be used to predict sparse mechanical
properties. Conventional neural networks deal with properties discretely as inputs or outputs of the network, with
outputs requiring all the input values to yield a valid result. However, in the framework used in this study, properties
are treated as both input and output variables with material design variables and properties stored together in a
vector x. To account for a missing value an expectation-maximization algorithm is used [53, 243]. In this algorithm,
preliminary predictions are provided for all missing data points. The network then improves the initial guess through
an iterative process, as shown schematically in Figure 5.1.
For any unknown properties, the missing values are first set to the average of the values present in the data set. With
estimates for all values of the neural network, the function (Eq. 9) is iteratively computed returning the converged
result following the nth iteration.
xn+1 = γxn + (1− γ) f (xn) (9)
98
Figure 5.1: Algorithm describing the procedure for accounting for missing data entries for the vector x of the design
variables and properties. The value is computed recursively using n iterations.
A softening parameter 0 ≤ γ ≤ 1 is used. When γ = 0, the initial guess is ignored for unknown values in the vector
x. These are then determined by applying the function to the unknown values. Oscillations in the predictions are
prevented with γ < 1, here a value of γ = 0.5 was used. Missing values are typically determined over six iterations.
The benefits of this method in dealing with incomplete data are demonstrated in Model Validation, Section 5.2.3.
5.2.2.2 Neural Network Kernel
Several types of neural networks are available including a feedforward [65] deep neural networks that can determine
deep data correlations, and recurrent neural networks that are often used in systems involving the evolution of time
[53]. In this study a feedforward neural network was adopted, comprising a single layer of hidden nodes as shown in
Figure 5.2.
This framework is based upon the formalism designed by Conduit et al. [46, 53, 71] but with further advances. These
include an enhanced algorithm for improved data extrapolation and an improved cross-validation test for the selection
of hidden nodes. This neural network is a linear superposition of hyperbolic tangents (2) and seeks, a function f which
satisfies the fixed-point equation f(x) ≡ x as precisely as possible. x = (x1, . . . , xI) is a vector, here of size I = 24,
comprising the individual design variables and properties. The identity operator is the simple solution to the fixed-point
equation. However, in order to represent data from other components a solution was constructed that is orthogonal to
the identity operator.
f : (x1, . . . , xi, . . . , xI) 7−→ (y1, . . . , yj , . . . , yI) (10)
99
Figure 5.2: Schematic illustration of the neural network framework. The framework illustrates how the predicted
properties (outputs) are calculated from the input properties. The input layer is constructed from the property database,
this layer is used to calculate the hidden nodes (indicator functions) to give the predicted properties.
with
ηhj = tanh
(
I∑
i=1
Aihjxi +Bhj
)
(11)
with
yj =
H∑
h=1
Chjηhj + Dj (12)
The single layer of hidden nodes ηhj have parameters (Aihj ,Bhj ,Chj ,Dj). The individual properties are, defined by yj
where 1 ≤ j ≤ I, are predicted separately. By setting Aihj = 0 the value of yj can be predicted without information
of xi. This allows the network to fully represent the non-linear behaviour through the use of tanh functions. This is
termed an activation function and generally produces predictions of higher quality than similar functions i.e. rectified
linear logic [53]. A five-fold cross-validation test was performed [53, 244, 245] to select three hidden nodes as giving
the best model. Cross-validation testing was also used when training the neural network weights. The given training
dataset was sampled with replacement to form 96 separate datasets containing hundreds of datapoints, with a neural
100
network trained on each with different weights [46, 53, 71], the mean of these was taken as the prediction expectation
value, and their variance as the uncertainty in the predictions that captures both experimental uncertainty in the
underlying data and also in extrapolation [53, 246].
5.2.3 Model Validation
The accuracy of the neural network was verified using cross-validation. The data was randomly separated, with the
network being trained on 80% of the database. The model was then validated against all properties using the remaining
20% of the database. To ensure that there was complete representation of the dataset, cross-validation of the data was
performed four additional times on the remaining data, again randomly selected. I measured the model accuracy using
the coefficient of determination, which is defined as:
R2j = 1−
∑n
i=1
(
xi,j − f j (xi)
)2∑n
i=1 (xi,j − x¯j )2
(13)
Where the subscript I denotes each of the n different materials, subscript j represents the jth material property and
x¯ denotes an average over all materials. Finally, the individual coefficient of determinations was averaged over all m
properties to deliver an overall coefficient of determination.
R2 =
∑m
j=1R
2
j
m
(14)
Different possible hyperparameters were explored, and those that delivered the highest quality model were selected,
which had five hidden nodes. This model delivered a coefficient of determination averaged over all properties of R2 =
0.92. To demonstrate the practical quality of the model, plots of predicted values versus calculated and experimental
data are shown in Figure 5.3.
In Figure 5.3a, the predicted phase stability at 650 ◦C has been plotted against values calculated using the CALPHAD
method with Thermo-CalcTM (which had R2 = 0.94). The plot shows that the predictions closely match the calculated
values at high phase stabilities. However, at lower phase stabilities the model is less reliable due to sparse training data
in that region. This is reasonable as the vast majority of data used in this neural network were from alloy compositions
101
with phase stabilities of greater than 0.6. Specifically, the model is poor at predicting the phase stability of compositions
containing high concentrations of Nb and Ta (circled) that have a higher predicted value of phase stability than that
calculated. To demonstrate the ability of the neural network to predict physical properties, a cross-validation plot of
predicted yield strength versus the experimentally measured values has been provided in Figure 5.3b. The results show
that the model does a good job of predicting the yield strength as a function of alloy composition, with the most
reliable predictions at higher values of yield strength. The three properties with the lowest coefficient of determination
were tensile strength with an R2 = 0.90, γ′ solvus with an R2 = 0.89 and elongation to failure with an R2 = 0.74. On
the other hand, cost and density were both modelled with R2 > 0.995.
Making good predictions is just the first step, for effective material design understanding the robustness of the predic-
tions is vital. Therefore, for both plots (Figure 5.3a and 5.3b) error bars on the predicted values have been provided.
Generally, points closer to the line have smaller error bars as these are more reliable predictions from the model.
Qualitatively, the standard error predicted by the neural network and the actual difference from unseen validation data
were compared, which on average should be unity.
102
Figure 5.3: Cross-validation tests for the properties of phase stability and yield strength. (a) Predicted phase stability
at 650 ◦C against calculated phase stability (CALPHAD). Poor predictions for high Nb and Ta containing compositions
have been circled. (b) Predicted yield strength vs experimental yield strength. For both plots error bars have provided
for the predicted values. Additionally, an idealised line has been added as an aid to the eye.
103
5.2.4 Optimisation Routine
For the selection of design variables that optimise the alloy properties, a single merit index L = Φ
[
Σ−1
(
V⃗ − T⃗
)]
(Eq
15) was calculated from all of the individual properties [46, 53]. This value describes the probability that the properties(
V⃗
)
of the optimised alloy will satisfy the specified targets (T⃗ ). The terms Φ and Σ represent the multivariate
cumulative normal distribution function and covariance matrix respectively [247]. The use of a single index allows the
individual probabilities for the properties to be summed into one probability describing the likelihood of fulfilling the
full specification. Through the maximisation of the merit index, the framework can determine the ideal compromises
between properties. Other methods such as robust design [248] and principal component analysis [249] are unable to
make compromises in this effective way. This ability of the network to determine deep property correlations is not
possible with the traditional linear regression method synonymous with a principal component analysis.
In order to optimise the material properties against the set targets, the framework must vary the design variables.
In other alloy design frameworks this process has the highest computational cost. Examples include the selection
of a few compositions from thousands of predetermined compositions using trade-off diagrams [75] or a Pareto set
[250, 251]. The expense of these methods scales with the number of chosen design variables. A further example is the
use of genetic algorithms [252], as in the work of Tancret [68]. However, the performance of these frameworks with
high dimensional problems can be poor and such algorithms may not reach the optimal solution [253, 254]. For this
framework, the logarithm of the probability of success is maximised. As a result, when searching a region of design
space that is unlikely to meet the targets, the optimisation routine generates a constant gradient favouring the least
optimised property [53, 71].
5.2.5 Alloy Identification
An optimised Ni-based alloy was designed to satisfy the specified targets, with a composition comparable to IN718.
The compositional range of this alloy system, designated AM718R, is given in Table 5.3. For this system, the only
required post-processing is a two-stage precipitation heat-treatment at 720 ◦C and 640 ◦C. The calculated probability
that this concentration will satisfy all of the design criteria is 17%. Compared to IN718, there are several noticeable
differences in composition. The neural network has attempted to make compromises within the composition. To reduce
the likelihood of undesired phases forming, specifically Laves, the Mo and Nb contents have been decreased and C
increased. To compensate for the inevitably reduced strength, the Al concentration has been increased and W has been
104
Table 5.3: The compositional ranges and recommended post-processing conditions for standard IN718 and AM718R.
For the measured composition of AM718R, SEM-EDX was used to assess the composition with a nominal error of 1%
for all measurements. The EDX measurements were taken from the laser pass heat-affected zone. Carbon and boron
have not been included due to the insensitivity of EDX in measuring light elements.
105
added to the system. It is also, worth noting that the addition of W might further serve to destabilise the formation
of the Laves phase.
Such compromises are best visualised through the use of an Ashby plot [255] of the predicted properties, Figure 5.4.
This serves as a representation of the design space and allows for an understanding of the property trade-offs made
during the design process. This plot illustrates the probability of fulfilling a set of target criteria, in this instance phase
stability and solidification strain. For a low solidification strain and high phase stability, the probability of fulfilling all
of the targets is 0 denoting a physical impossibility of success. The alloy identified in this study is in a region with a
relatively high likelihood of success.
Figure 5.4: Ashby plot showing the probability of an alloy composition satisfying all of the design criteria when the
properties of phase stability (y – axis) and solidification strain (x – axis) are varied. The black regions show areas of
design space that have a low probability of fulfilling the targets. The lighter shading indicates an increased likelihood
of satisfying all the target criteria. The blue circles show the current alloy IN718 and the designed alloy AM718R.
106
5.3 Results and Discussion
The physical and thermodynamic properties of the optimised AM718R alloy were assessed using a range of techniques.
The results were compared to conventional IN718. The processability of AM718R was assessed using laser pass testing
[241].
5.3.1 Alloy Fabrication and Laser Pass Testing
Samples of the newly designed alloy AM718R were manufactured using vacuum arc melting from individual elements
and binaries of at least 99.9% purity. The bars were inverted and remelted several times to improve compositional
homogeneity. The manufactured cylindrical bars were machined using electro-discharge machining (EDM) into rect-
angular samples of dimensions 20 × 10 × 5 cm. These rectangular samples were used for assessment of physical and
thermodynamic properties. In an effort to assess the processability of the designed alloy in comparison to existing
alloys, laser pass testing was used. This relatively simple test allows for rapid assessment of the response of an alloy to
processing through additive manufacturing [241]. Prior to laser pass testing, samples were ground to a finish of 1200
grit using standard metallographic techniques.
For the laser pass testing, a laser power of 120 W was used with a scan velocity of 1000 mms−1 and a hatch spacing
of 40µm on a Aconity3D MINI system. A bi-linear raster over an area of 5 × 5 mm was scanned on one face on each
sample. Following the laser pass, the samples underwent a standard two-stage precipitation heat treatment for IN718,
consisting of heating from room temperature to 720 ◦C, 8 hours at 720 ◦C, a furnace cool to 640 ◦C, 8 hours at 640 ◦C
and air cooling to room temperature. Prior to heat treatment, samples were encapsulated in argon-backfilled quartz
ampoules to minimise surface oxidation.
The heat-treated laser pass samples were prepared for SEM analysis on a Zeiss SEM 450 equipped with an Oxford In-
struments X-MaxN 50 detector. Samples were mounted in conductive Bakelite and standard metallographic techniques
were used to prepare the samples to a finish of 0.25 µm. A 10-minute chemical polish was applied to the samples with
a 10:1 solution of 0.04 µm colloidal silica. Post-processing of the data was completed in the Oxford Instruments Aztec
software.
107
Figure 5.5: Back-scattered electron images of the laser pass heat-affected zone (HAZ) and arc-melted microstructure
of IN718 (a) and AM718R (b) in the precipitation heat-treated condition. For ease of identification the extent of the
HAZ has been identified with a yellow line in both micrographs.
108
Micrographs highlighting the laser pass heat affected zones (HAZs) in the AM718R and IN718 samples have been
given in Figure 5.5. The average depth of the laser pass HAZs in the samples is 35 ± 8 µm, measured over 5 mm. For
both samples this zone was inspected for laser induced defects including pores and cracks. In both samples the number
defects found in the 5 mm x 35 µm region was negligible, being ≤ 1 per 100 µm2. While this is expected for the highly
processable IN718 alloy, the low defect density is an encourageing result for AM718R. This system was designed to
have processability comparable to IN718 and a lower volume fraction of undesired phases. For both samples, image
analysis of multiple SEM micrographs was carried out to obtain an estimate of the area fraction of eutectic in the
arc-melted area. Though this area does not have an additively manufactured microstructure, the volume fractions serve
to give an indication of eutectic phase content. For IN718 the approximate area fraction of eutectic is 3.4 ± 0.5%. For
AM718R the area fraction was found to be 1.7 ± 0.3%. This result indicates that the AM718R alloy might contain a
lower volume fraction of eutectic phases, such as the Laves phase, in the additively manufactured condition.
5.3.2 Microstructural and Phase Analysis
Detailed SEM analysis of the laser-pass HAZs was carried out at a higher magnification for the IN718 and AM718R
samples. In Figure 5.6 secondary electron (SE) micrographs have been provided alongside elemental distribution maps
for both samples. In the SE images a fine microstructure of cored dendrites is observed for the two specimens. Such
a microstructure is frequently reported for additively manufactured Ni-based alloys [28, 238]. Indeed, the observed
microstructure for the IN718 sample is similar to that observed for LBP-DED IN718 in the previous chapters [238].
This demonstrates that the laser pass testing could be a viable way of quickly assessing the processability of alloys for
AM techniques. However, the dendrite spacing observed in this zone is extremely fine, as this is only a single laser pass.
A coarsening of the microstructure would be expected in a larger-scale AM deposition due to the heat retention during
fabrication [16]. Despite this, the observed microstructure of the HAZ provides insight into what would be expected.
The microstructure of the AM718R samples is more homogeneous than that of the IN718. This is clear in the Mo and
Nb elemental distribution maps, where the segregation of these elements is reduced in the AM718R sample. Despite
this reduction, interdendritic precipitates still occur in the AM718R sample. The Laves phase is expected to be the
primary constituent of the interdendritic eutectic phase. The occurrence of the Laves phase in additively manufactured
IN718 has been frequently reported in the literature [104, 238]. As the composition of AM718R is comparable to that
of IN718, the occurrence of the Laves phase is therefore to be expected.
109
To determine the area fraction of these phases, image analysis of multiple electron micrographs was performed. The
electron micrographs were taken from random locations within the HAZs of both samples to ensure representative
results. These images are comparable to the secondary electron images in Figure 5.6, and examples are provided in
Figure 5.7. The area fraction of interdendritic precipitates in the IN718 sample is 3.0 ± 0.2% and for the AM718R
sample it is 1.9 ± 0.2%. Both results are comparable to the area fraction determined in the arc-melted microstructure
of both samples. These results indicate that the newly designed AM718R alloy has a higher phase stability than the
standard IN718. The neural network predicted that the AM718R system would have phase stability of 98.5%, and the
experimental results are acceptably close to that prediction.
The thermo-physical characteristics of the phases present were further investigated using differential scanning calorime-
try (DSC) and X-ray diffraction (XRD). Disc specimen of thickness 1.0 mm and 5.0 mm in diameter were prepared
from alloy samples using EDM. For the DSC analysis, a Netzsch 404 calorimeter was used. For the measurements,
identical parameters to those described in Section 4.2 were used. The analysis cycle consisted of an initial heating to
1400 ◦C, cooling to 450 ◦C and reheating to 1400 ◦C before cooling to room temperature.
The DSC thermograms for both samples are given in Figure 5.8. The reported solidification sequence for IN718 is
L → L + γ → L + NbC/γ → Laves/γ [87] and both samples display evidence of this sequence on cooling. Though
these traces appear similar, there are several distinct differences. The on-cooling liquidus temperature of the AM718R
samples, 1332 ◦C, was found to be higher than that of IN718, 1310 ◦C. An approximate freezing range for the alloys has
been determined by analysis of the liquidus points and the events near the termination of solidification. The samples
have comparable solidification ranges, with the IN718 sample having an approximate a range of 178 – 253 K and the
AM718R sample a range of 179 – 258 K. The target for AM718R was set as < 260 K, which has been satisfied. The
results for IN718 are comparable to those reported in the literature [229].
There are similar events for the two samples in the heating thermograms. The carbide dissolution is not easily identifiable
by eye but occurs around 1290 ◦C for IN718 and there is an analogous event at 1305 ◦C for AM718R. This result
is unsurprising given that the addition of W to the composition would increase the solidus and carbide solvus for
the AM718R sample. For IN718 the δ and Laves dissolution event starts near 1070 ◦C, and this agrees with previous
reports in the literature [87, 231]. For AM718R there is an event in the same region, around 1080◦C. Both samples
also display similar events that could be attributed to the precipitation of the γ′′ and δ phases. For IN718 these start
at approximately 760 ◦C and terminate near 905 ◦C. For AM718R there is an event that begins at 770 ◦C and finishes
around 935 ◦C. The final event for IN718 is attributed to the precipitation of the γ′. There is a pronounced peak in a
comparable position for the AM718R sample. The start for this event in IN718 is near 570 ◦C and terminates around
715 ◦C. In contrast, for AM718R this event starts at approximately 590 ◦C and ends near 725 ◦C.
110
Figure 5.6: SEM analysis of the IN718 and AM718R HAZs. Top, a secondary electron image. Beneath, elemental
distribution maps for Cr, Fe, Mo and Nb determined by SEM-EDX.
111
Figure 5.7: High magnification back-scattered electron images of the laser pass HAZs of IN718 (a) and AM718R (b) in
the precipitation heat-treated condition. These are example micrographs of those used for the phase fraction analysis.
112
Figure 5.8: DSC traces for the first heating of IN718 and AM718R (a) samples from room temperature to 1400 ◦C
and the accompanying cooling curves (b).
113
Phase identification was performed using XRD with a Bruker D8 X-ray diffractometer operated with identical pa-
rameters to those described in Section 3.2. The Bruker DIFFRAC.EVA software package was used for analysis of the
diffraction data. The diffraction peaks were matched to their respective phases using the inbuilt database. The XRD
patterns for IN718 and AM718R have been given in Figure 5.9a and are similar. The primary γ matrix peaks are easily
identifiable, however, the γ′ and the γ′′ superlattice reflections are not distinguishable. This observation is consistent
with previous literature regarding IN718 [238]. Several lower intensity peaks were also observed in the pattern given in
Figure 5.9a.
Figure 5.9: (a) XRD patterns for IN718 and AM718R samples in the precipitation heat-treated condition. (b) higher
resolution XRD patterns for the IN718 and AM718R samples over a selected range of 2Θ. Labels have been added to
highlight the reflections of the gamma (γ), MC carbide and Laves (φ) phases. Intensity has altered to the square root
of peak intensity for ease of visualisation of the weaker peaks present.
114
Higher resolution XRD patterns taken over a narrower range of 2Θ range have been given in Figure 5.9b to enhance
these reflections. These peaks have been attributed to the presence of the MC carbide and Laves phases. With the
peaks at approximately 35◦ and 37.5◦ 2Θ being the MC carbide {111} and Laves phase {0110} reflections respectively
[43, 212, 213]. The peak at approximately 41◦ 2Θ can be attributed to the MC carbide {200} reflection. The final
peak at 45◦ 2Θ was identified as the Laves phase {0220} reflection [226, 227]. It is worth noting that the MC carbide
{111} was not observed in the diffraction data from the IN718 sample. Indeed, the peak at 41◦ 2Θ is also reduced
in intensity when compared to AM718R. However, the Laves phase {0110} and {0220} reflections are stronger in the
IN718 samples when compared to AM718R. This is likely to be a result of the reduced presence of the Laves phase in
the AM718R sample.
5.3.3 Physical Properties
Selected physical properties of AM718R were tested and compared with IN718. A critical property is that of oxidation
resistance. To assess the high-temperature oxidation properties, isothermal oxidation testing was performed at 650
◦C. Samples of the dimension 20 × 10 × 1 mm were machined from the alloy bars using EDM and ground to a 4000
gt finish. Oxidation testing of the samples was carried out on a Setaram Instruments Setsys Evolution TGA. Mass
gains for the samples were measured over a 200-hour exposure at 650 ◦C, see Figure 5.10. The TGA parameters and
conditions used for oxidation testing of these samples are identical to those described in Section 4.2. The Pierragi
Method (Eq. 5) [225] was used to fit the data and determine the parabolic rate constants Kp (mg
2 cm−4 s−1).
As the content of Cr was kept comparable to that of IN718 it was expected that the two alloys would have similar
oxidation properties when exposed at the standard operating temperature of 650 ◦C. After 200-hours the mass gain
was similar between the two samples, being 10.5 ± 0.2 µg/cm2 for IN718 and 10.9 ± 0.2 µg/cm2 for AM718R. The
undulations that can be observed in the traces arise as a result of the day and night temperature variations affecting
the buoyancy forces and therefore the measured weight at these extremely small mass gains. A parabolic rate constant
was calculated for both of the samples, excluding the initial transient region. For IN718 the constant was determined
as 1.0 × 10−5 ± 5 × 10−6 mg2 cm−4 s−1 and for AM718R 1.2 × 10−5± 5 × 10−6 mg2 cm−4 s−1. The comparable
oxidation performance of AM718R to IN718 is highly encouraging.
115
Figure 5.10: TGA traces for the 200-hour exposure of IN718 and AM718R in air at 650 ◦C. The graphs show the
mass gain with respect to area against the square root of time.
To gauge the mechanical properties of AM718R compared to IN718 the hardness and elastic modulus were measured at
room temperature using a KLA iNano® series Nanoindenter. The samples used were in the 2-stage precipitation heat-
treated condition and were mounted in Bakelite and polished using the metallographic protocol previously mentioned
to a finish of 0.25 µm. Standard XP testing was carried out using a Berkovich tip with a depth limit of 1 µm, a load
of 50 mN and a peak hold time of 10 seconds. The target strain rate was set at 0.05 s−1 and the allowable drift rate
was set as 0.8 nm s−1. Each test consisted of a matrix of 1 × 20 indents, with a spacing of 50 µm between indents.
Prior to each test, a standard calibration on fused silica was carried out. Nanoindentation was carried out in both the
arc-melted region and the laser-pass HAZ for the IN718 and AM718R samples.
In addition, the hardness and elastic modulus of LBP-DED IN718 were measured for comparison to the laser pass
region on arc-melted IN718. The gas atomized IN718 powder used was identical to that used to fabricate the IN718
samples in Chapter 3 and 4. Subsequent LBP-DED was carried out using an identical setup to produce the previous
LBP-DED sample. A bi-linear raster pattern was used, and the deposition parameters were identical to those from
Chapters 3 and 4. Column like samples were machined from the parent LBP-DED build using EDM. The samples were
aged using a two-stage heat treatment that consisted of two 8-hour isothermal holds at 720 ◦C and 640 ◦C. Samples
were mounted in Bakelite and polished using the metallographic protocol previously mentioned to a finish of 0.25 µm.
Full details of the microstructure and properties of the LBP-DED IN718 have been given in Chapters 3 and 4 [238].
116
The hardness and elastic modulus for LBP-DED IN718 was measured as 6.4 ± 0.3 GPa and 231 ± 7 GPa. These results
are in good agreement with reports in the literature for hardness testing of IN718 to similar depth and measured elastic
modulus [238, 256]. The arc-melted regions in both samples gave highly variable results, which is to be expected with
the length scales over which the microstructure varies. For IN718 the hardness and moduli in this region were measured
as 5.8 ± 1.5 GPa and 241 ± 8 GPa. The results in this region for AM718R were 5.8 ± 1.3 GPa and 243 ± 11 GPa.
The results for the laser pass region in both samples were more comparable to AM718R, with significantly less variability.
For the IN718 the hardness and elastic modulus in the laser pass region were measured as 6.2 ± 0.2 GPa and 221 ±
8 GPa respectively. These results are comparable to that of LBP-DED IN718 and further demonstrate that laser pass
testing can yield results comparable to additively manufactured material. For AM718R the analogous results were 6.3
± 0.2 GPa and 216 ± 7 GPa, a marginal increase in hardness compared to IN718. The results further support the
prediction that the designed AM718R alloy has advantages over conventional IN718.
5.4 Conclusions
A neural network framework was used to design a Ni-based superalloy for LBP-DED repair applications. The com-
position of IN718 was used to confine the compositional design space considered by the neural network tool. A new
Ni-based superalloy composition (AM718R) was identified that fulfilled the targets to surpass the contemporary IN718
phase stability, tensile strength, solidification strain, cost, and density. The properties of phase stability and solidifi-
cation strain were prioritised due to their importance in additive manufacturing. Several properties of AM718R were
experimentally verified using laser pass testing. The microstructure and properties of the laser pass region on the IN718
sample were comparable to that of additively manufactured IN718.
The laser pass microstructure of AM718R was found to be more refined than that of IN718. A higher phase stability
than IN718 was observed, with less Laves phase being formed. In addition, the processability of this new system was
confirmed to be comparable to IN718. The oxidation resistance of AM718R was measured to be comparable to IN718 at
650 ◦C. The room temperature hardness of AM718R was marginally higher than that of IN718. These results support
the assertion that the designed AM718R system possesses improved properties compared to IN718 and may be a better
candidate for LBP-DED repair applications.
117
Chapter 6
Conclusions and Future Work
6.1 LBP-DED of IN718 for Additive Repair Applications
Due to excellent processability and reliable properties, IN718 is currently used for certain additive manufacturing
repair applications [240]. This work has investigated the LBP-DED of IN718 under deposition conditions used for
the repair of aerospace components. Processing this alloy through the LBP-DED method yields a highly defect free
microstructure, though there is considerable segregation of strengthening elements to the interdendritic regions. This
segregation results in the formation of the undesired Laves phase. In addition, there is a high degree of texture that
results from LBP-DED processing and a corresponding irregularity to the grain sizes. This goal of this work was to
study the microstructural and textural evolution of LBP-DED IN718 during post-processing.
Previous studies have identified the cube, {001} <100>, and goss, {011} <100>, components as prominent textures
for Ni-based that occur during additive manufacturing. However, these were observed to a lesser degree in the work
presented in Chapter 3 [111, 114, 117]. A Brass component {011} <211> is observed as the primary texture when
using the deposition parameter from Chapters 3 and 4 [238]. The identification and characterisation of this Brass
texture adds the work presented by Ma et al. [117] on the occurrence of textures in additively manufactured Ni-based
superalloys. The variability in resultant texture emphasises the importance in controlling the deposition conditions
during fabrication. This texture was also observed to be enhanced by post-processing heat treatments near the quoted
recrystallisation start temperature of IN718.
118
The elastic properties of the LBP-DED IN718 in different heat-treated conditions were measured using RUS. The
analyses highlighted that in several cases the apparent elastic anisotropy was significantly reduced following heat treat-
ment near the recrystallisation point. This is in contrast to the observation of the enhancement of the microstructural
texture and is hypothesised to be a result the prominent [011] texture giving rise to pseudo-isotropic elastic behaviour.
Such results could be used to guide future post-processing conditions for alloys exhibiting this texture. In situ high-
temperature RUS was used to study the evolution of the resonant modes during heat treatment. It was observed that
the elastic constants were sensitive to small duration heat treatments. This method could be used for the optimisation
of heat treatment duration and temperature.
It was shown that elimination of the Laves phase and control of the texture can be achieved through high temperature
heat-treatment. However, the use of such intense post-processing can cause grain growth and affect the physical
properties. Therefore, the repaired material is often only subjected to a precipitation heat-treatment that does not
compromise the properties of the substrate. As such, the repaired material retains the highly segregated and anisotropic
microstructure, yielding non-optimum performance in-service. Methods to overcome these complications are required
that do not rely on intense post-processing.
6.2 Benefits of Inoculant Additions for the LBP-DED of Ni-based Su-
peralloys
A method for controlling the microstructure of additively manufactured metallic components is the addition of an
inoculant to the precursor [195, 257]. The inoculants can act as nucleation sites and may lead to refinement of the mi-
crostructure [191, 194]. Critically, the addition of these particles can impart a degree of control over the microstructure.
There has been limited use of these particles in the AM of Ni-based superalloys. However, several studies have reported
notable results [191, 194, 195]. The potential for grain refinement and texture control are the results of greatest inter-
est to the community. As discussed, a strong residual texture and irregular microstructure are current complications
encountered during the AM of Ni-based superalloys. The use of inoculants could prove to be a method of overcoming
these complications.
This work has investigated the use of NbC inoculant in the LBP-DED of Ni-based superalloy IN718. The addition of
these particles had several interesting effects on the microstructure and properties of IN718 in the as deposited and
precipitation heat-treated condition. Most notably was the distinct enhancement of the original Brass texture of IN718
119
fabricated via LBP-DED identified in Chapter 3. This was attributed to the NbC particles restricting the selection
and growth of columnar grains during deposition. The resultant microstructure was observed to be more homogenous
than that of standard IN718, which might be desirable for specific applications. Critically, the addition of these particle
gives a degree of control over the resultant microstructure, texture, and elastic properties of LBP-DED IN718. This
control might be of use for tailoring the properties during deposition.
The addition of the NbC inoculants also led to a reduction in the presence of the undesired Laves phase and a refinement
of individual particles. Control over the formation of this phase in AM IN718 is of great interest to the field. This
result was attributed to the higher carbon content driving the formation of carbides over that of the Laves phase.
The presence of the additional carbides slightly raised the hardness of the IN718 and limited grain growth during
recrystallisation. Notably, the addition of particles did not significantly affect the oxidation properties at 650 ◦C.
The use of an inoculant during the LBP-DED of Ni-based alloys may offer a route to tailoring the microstructure and
texture during deposition. However, the segregation of the strengthening elements Mo and Nb remained prominent,
driving the formation of the Laves phase. High temperature heat-treatments are therefore still required to eliminate
this phase from the microstructure. Such post-processing is incompatible for material used in repair applications as
such treatments might deteriorate the properties of the underlying substrate. It is clear, that the composition of IN718
is not optimised for LBP-DED repair applications. The composition of IN718 must therefore be altered to yield the
best properties when processed in the precipitation heat-treated condition without a high-temperature annealing step.
6.3 Applications of Alloy Design for the Optimisation of Ni-based Su-
peralloys for LBP-DED Applications
The Ni-based superalloy IN718 has many desirable properties, which has led to extensive use of this alloy over the last
fifty years [1]. Alongside impressive mechanical properties, IN718 is highly processable, display excellent environmental
properties and is relatively expensive compared to other alloy systems [1, 3, 25]. However, it has been shown that
there are several complications encountered when using this alloy for LBP-DED repair applications. The segregation
of the strengthening elements is a critical issue and leads to the formation of undesired phases. Currently, this can only
be resolved through intense post-processing methods that are incompatible with repair applications. It is clear that
the composition of IN718 is not optimised for use in AM repair applications. Obviously, a new Ni-based superalloy
specifically designed for LBP-DED repair is therefore required for further progress.
120
Resent advances in machine learning and computational modelling have significantly enhanced the effectiveness of alloy
design [65, 75]. New frameworks have been developed which can efficiently search for an optimal alloy system which can
fulfil a set of target properties [46, 76]. Alloys are being design for specific applications including additive manufacturing
[53] and this work has utilised such a neural network framework to design a Ni-based superalloy specifically for LBP-
DED repair applications.
Due to the many desirable properties of IN718, this composition was used in the present study as the starting point
for the design space considered in the alloy design process. Additionally, the composition of the 718Plus system was
also used for refinement of the design space due to the advantages of this system over that of IN718. The properties of
IN718 were used as benchmarks for this new system to satisfy. Additionally, a high phase stability and low solidification
strain were specified as key targets in the design process. This was done so that the identified alloy would display better
properties than IN718 in the as deposited condition, whilst retaining the desired processability of IN718. Using this
framework, the AM718R composition was identified as a system that might fulfil the specified targets and thereby
surpass the performance of IN718 for LBP-DED repair applications.
The properties of AM718R were experimentally verified, where possible, and compared to IN718. Laser pass testing
confirmed comparable processability to IN718, with a reduction in the occurrence of the undesired Laves phase. In
addition, it was observed that the oxidation resistance at 650 ◦C of AM718R was similar to that IN718 and the
hardness of AM718R was marginally higher. The results are highly encouraging and support the supposition that the
AM718R system should have superior performance to IN718 when used for LBP-DED repair applications. An LBP-
DED deposition trial of AM718R is required to fully investigate the advantages of AM718R over that of IN718.
121
6.4 Future Work
6.4.1 Outlook
The studies reported in this thesis have demonstrated that compositional control using inoculant additions and alloy
design can be used to optimise the LBP-DED of Ni-based superalloy IN718. However, further work is required to
develop the inoculants and verify the properties of the new AM718R system. The proposed next steps are laid out
below:
The mechanical properties of IN718 components fabricated using NbC inoculants will require significant further inves-
tigation. The effect that the presence of additional carbide particles has on the fatigue and creep properties would be
of significant interest. These properties would need to be measured for components with a range of thermal histories.
Further study on the long-term environmental and phase stability properties of NbC containing IN718 is also required.
Additionally, a better understanding of the formation of the final texture during deposition when using these particles
is required. A complete understanding of the amount of inoculant required to yield textural control and inhibit grain
growth is clearly essential.
The addition of inoculants such as NbC to LBP-DED precursor powder is relatively straightforward and a wider
investigation of the use of inoculants is required. For the NbC inoculants, different concentration and particles sizes
should be trialled. This is so other effects can be identified and characterised, in line with those reported in the literature.
Additionally, other potential inoculants such as tungsten carbide and titanium carbide should be considered. These
alternatives may yield better results that NbC. Finally, other alloy systems and additive manufacturing methods should
be considered to see if similar result are observed. Overall, the inoculant addition method might lead to a number of
additional desirable outcomes within the field of additive manufacturing.
This work has demonstrated the potential of computational alloy design for the optimisation of a Ni-based alloy for
specific applications. The AM718R system displays several interesting properties and might outperform IN718 in LBP-
DED repair applications. However, full verification of the properties of this new alloy system is essential if it is ever be
used commercially. The mechanical properties require assessment in components with a range of thermal histories. The
long-term environmental and phase stability properties also stand in need of further research attention and analysis.
122
The alloy design framework used is a powerful tool and can be applied for the design and optimisation of other alloy
systems for additive manufacturing applications. This tool could be valuable in designing other Ni-based alloys such
as high γ′ containing superalloys and purely solid-solution-strengthened systems, although this would require addi-
tional models that predict other defect and crack forming mechanism to be added to the framework. Critical cracking
mechanisms that would need to be accounted for are strain-age and liquation cracking. Additionally, a significantly
large database containing information on the cracking behaviour of numerous alloys fabricated via a range of additive
techniques and processing parameters would be highly desirable. This would allow the neural network framework to
learn the critical relationships between alloy composition and cracking during additive manufacturing.
6.4.2 Design of in situ Composite Superalloys for LBP-DED Repair Applications
The work presented in Chapter 6 on the design of alloy AM718R for LBP-DED repair applications demonstrated the
potential of alloy design. During the research that led to AM718R being developed, other alloy systems were considered.
The use of the NbC inoculants in the LBP-DED of IN718 yielded several interesting results. This work demonstrated
that the LBP-DED technique would lead to the formation of highly refined carbides within the matrix. This observation
could be applied to the fabrication of directionally solidified in situ composite Ni-based alloys which were abandoned in
1980s because of commercially prohibitive production routes. These high carbon containing superalloys relied upon a
combination of highly orientated carbide particles and γ′ precipitates for strength. The carbides were orientated along
the growth direction of the alloy and give the material composite like properties. These alloys also display eutectic
solidification behaviour and, as a result, possess an intrinsically small freezing range. This could be highly beneficial
for additive manufacturing processing as this it might minimise microsegregation and reduce the likelihood of cracking
occurring during solidification [76].
The two systems with the most promising properties reported in previous studies were COTAC 74 and COTAC 744
[258, 259, 260]. Both these alloys displayed impressive mechanical properties, outperforming the conventional Ni-based
superalloys at the time. The compositions of these systems are given in Table 6.1.
Table 6.1: Compositions for the COTAC 74 and COTAC 744 alloy systems in wt % with a Ni balance.
The use of a such a system for LBP-DED repair applications was considered, owing to the promising potential of
these alloys. However, it was recognised that the composition would have to be significantly altered such that it would
123
be compatible with the repair of conventional Ni-based alloys. The alloy design framework described in Chapter 5
was adapted to optimise a new alloy system based on the COTAC compositions and IN718. Two new alloys were
identified that displayed promising predicted mechanical properties whilst fulfilling the targets set for processability.
These systems were named NN7418A and NN7418B, the compositions of these systems are given in Table 6.2.
Some of the preliminary analyses completed on these alloys are briefly presented below with limited discussion. Arc-
melted bars of the alloys were fabricated and subjected to a two-stage precipitation heat-treatment consisting of an
8-hour isothermal hold at 720 ◦C, a furnace cool to 640 ◦C with an additional 8-hour hold and a final air cool to RT.
The results include an SEM analysis of laser pass tested samples, a DSC analysis and oxidation resistance at 750 ◦C.
and highlight the interesting potential for further development of these systems for additive manufacturing.
Table 6.2: Compositions for the COTAC 74 and COTAC 744 alloy systems in wt % with a Ni balance.
SEM analysis was completed on both new alloys using the experimental parameters defined in Section 3.2, where
sample preparation is also described. Low magnification SEM micrographs of the HAZs and arc-melted regions are
presented in Figure 6.1. In the arc-melted zone there are numerous large ’Chinese-script’ carbides, which are expected
in these alloys when fabricated using non-directionally solidified conditions [1, 3]. Within the laser pass HAZs there is
a significant reduction in the presence of these particles. However, there are some residual large particles which were
not remelted during the single laser pass.
Higher magnification micrographs of the HAZ have been presented in Figure 6.2. The results highlight the significant
refinement of the microstructure and precipitates. There remain a small number of large carbides in the HAZ that were
probably not dissolved during the laser pass testing. Therefore, a high laser power would be required in future testing
to ensure a representative area for analysis. However, there are small spherical particles that have precipitated along
the dendrites, likely carbides. However, identification was not possible using an SEM due to the size of these particles
and high-resolution STEM-EDX would be required for identification. If these particles are in fact carbides, then these
images are highly encourageing. It might be possible to control the growth of these particles to yield in situ composite
behaviour [258]. LBP-DED fabrication is required to determine whether these particles will directionally solidify along
the growth axis and give in situ composite properties.
The DSC analysis for NN7418 A and B alloys has been given in Figure 6.3 and include the results for IN718 for
comparison. The parameters used for the DSC analysis are identical to those given in Section 4.2. The results highlight
124
Figure 6.1: Back-scattered electron images of the laser pass heat-affected zone (HAZ) and arc-melted microstructure
of NN7418A (a) and NN7418B (b) in the precipitation heat-treated condition. For ease of identification the extent of
the HAZ has been identified with a yellow line in both micrographs.
125
Figure 6.2: Electron micrographs of the laser pass heat-affected zone (HAZ) NN7418A (a) and NN7418B (b) in the
precipitation heat-treated condition.
126
the significantly narrower freezing range displayed by these alloys. There is also a large event near 1000 ◦C for both of
the NN7418 systems and this likely due to the increased γ′ volume fraction in these alloys. However, significant further
work is required to interpret and confirm these results.
Figure 6.3: DSC traces for the first heating of NN7418A, NN7418B and IN718 arc-melted samples from room tem-
perature to 1400 ◦C.
Oxidation testing was carried out with the same parameters and sample diameters described in Section 4.2, though
at the higher temperature of 750 ◦C and a reduced time of 100 hours. Samples of IN718 were also tested so that
comparisons could be made. The results for the alloys have been presented in Figure 6.4. The results show that these
alloys are more susceptible to oxidation than IN718. However, the mass gains remain relatively low over the 100-hour
exposure. The Kp values for the alloys were calculated using the previously described Pierragi Method. The value for
NN7418A, NN7418B and IN718 are 8.45 × 10−3 ± 1 × 10−5 mg2 cm4 s−1, 12.75 × 10−3 ± 1 × 10−5 mg2 cm4 s−1 and
6.76 × 10−3 ± 1 × 10−5 mg2 cm4 s−1 respectively. Further oxidation and corrosion analysis is required to determine
whether these alloys would be capable of performing adequately in service.
127
Figure 6.4: TGA traces for the 100-hour exposure of NN7418A, NN7418B and IN718 in air at 750 ◦C. The graphs
show the mass gain with respect to area against the square root of time.
The results presented for the NN7418 alloys are highly encouraging for the processing of these systems using additive
manufacturing techniques. Though significant further research is required to fully understand how these systems would
behave and whether it would be possible to control the growth of the carbide particles.
128
Bibliography
[1] Reed, R.C., 2006. The superalloys: fundamentals and applications. Cambridge, UK ; New York: Cambridge Uni-
versity Press. xiv, 372 pages.
[2] Rolls-Royce plc., 2015. The Jet Engine. Chichester, West Sussex: Wiley. 288 pages.
[3] Sims, C.T., Stoloff, N.S., and Hagel, W.C., 1987. Superalloys II. New York: Wiley. xx, 615 pages.
[4] DuPont, J.N., Lippold, J.C., and Kiser, S.D., 2009. Welding metallurgy and weldability of nickel-base alloys.
Hoboken, N.J.: John Wiley & Sons. xiv, 440 pages.
[5] Ward, M.J., Halliday, S.T., and Foden, J., 2011. A readiness level approach to manufacturing technology
development in the aerospace sector: an industrial approach. Proceedings of the Institution of Mechanical Engineers,
Part B: Journal of Engineering Manufacture. 226(3): p. 547-552.
[6] Pollock, T.M., 2016. Alloy design for aircraft engines. Nature Materials. 15(8): p. 809-815.
[7] Thompson, S.M., Bian, L., Shamsaei, N., and Yadollahi, A., 2015. An overview of Direct Laser Deposition
for additive manufacturing; Part I: Transport phenomena, modeling and diagnostics. Additive Manufacturing. 8:
p. 36-62.
[8] Uriondo, A., Esperon-Miguez, M., and Perinpanayagam, S., 2015. The present and future of additive
manufacturing in the aerospace sector: A review of important aspects. Proceedings of the Institution of Mechanical
Engineers, Part G: Journal of Aerospace Engineering. 229: p. 2132 - 2147.
[9] Zhang, Y., Wu, L., Guo, X., Kane, S., Deng, Y., Jung, Y.-G., Lee, J.-H., and Zhang, J., 2018. Additive
Manufacturing of Metallic Materials: A Review. Journal of Materials Engineering and Performance. 27(1): p. 1-13.
[10] Herderick, E.E. 2011. Additive Manufacturing of Metals: A Review. in Materials Science and Technology.
Columbus, Ohio: MS&T.
129
[11] Basak, A., Acharya, R., and Das, S., 2016.Additive Manufacturing of Single-Crystal Superalloy CMSX-4
Through Scanning Laser Epitaxy: Computational Modeling, Experimental Process Development, and Process Pa-
rameter Optimization. Metallurgical and Materials Transactions A. 47(8): p. 3845-3859.
[12] Asala, G., Andersson, J.O., and Ojo, O.A., 2018. Improved dynamic impact behaviour of wire-arc additive
manufactured ATI 718Plus®. Materials Science and Engineering: A. 738: p. 111-124.
[13] Chen, B. and Mazumder, J., 2017. Role of process parameters during additive manufacturing by direct metal
deposition of Inconel 718. Rapid prototyping journal. 23(5): p. 919-929.
[14] Singamneni, S., LV, Y., Hewitt, A., Chalk, R., Thomas, W., and Jordison, D., 2019. Additive Manu-
facturing for the Aircraft Industry: A Review. Journal of Aeronautics & Aerospace Engineering. 8(1).
[15] Blackmore, M.L., 2018. Method of Manufacturing a Component, U. Patent, Editor., Rolls Royce plc.: USA.
[16] Babu, S.S., Raghavan, N., Raplee, J., Foster, S.J., Frederick, C., Haines, M., Dinwiddie, R., Kirka,
M.K., Plotkowski, A., Lee, Y., and Dehoff, R.R., 2018. Additive Manufacturing of Nickel Superalloys:
Opportunities for Innovation and Challenges Related to Qualification. Metallurgical and Materials Transactions
A. 49(9): p. 3764-3780.
[17] Graybill, B., Li, M., Malawey, D., Ma, C., Alvarado-Orozco, J., and Martinez-Franco, E., 2018.
Additive Manufacturing of Nickel-Based Superalloys, in 13th International Manufacturing Science and Engineering.
ASME: College Station.
[18] Kulkarni, A., 2018. Additive Manufacturing of Nickel Based Superalloys. Michigan Technological University:
Michigan, USA.
[19] Gasser, A., Backes, G., Kelbassa, I., Weisheit, A., and Wissenbach, K., 2010. Laser Additive Manufac-
turing. Laser Technik Journal. 7(2): p. 58-63.
[20] Carter, L.N., Attallah, M.M., and Reed, R.C., 2012. Laser Powder Bed Fabrication of Nickel-Base Super-
alloys: Influence of Parameters; Characterisation, Quantification and Mitigation of Cracking, in Superalloys 2012
12th Annual Symposium on Superalloys, E.S. Huron, et al., Editors., TMS (The Minerals, Metals & Materials
Society): Champion, Pennsylvania.
[21] Attallah, M.M., Jennings, R., Wang, X., and Carter, L.N., 2016. Additive manufacturing of Ni-based
superalloys: The outstanding issues. MRS Bulletin. 41(10): p. 758-764.
[22] Kuo, Y.-L., Horikawa, S., and Kakehi, K., 2017. Effects of build direction and heat treatment on creep
properties of Ni-base superalloy built up by additive manufacturing. Scripta Materialia. 129: p. 74-78.
130
[23] Wang, Z., Stoica, A.D., Ma, D., and Beese, A.M., 2018. Stress relaxation in a nickel-base superalloy
at elevated temperatures with in situ neutron diffraction characterization: Application to additive manufacturing.
Materials Science and Engineering: A. 714: p. 75-83.
[24] Duley, W.W., 1999. Laser welding. New York: Wiley. ix, 251 pages.
[25] Lippold, J.C., 2015. Welding metallurgy and weldability. Hoboken, New Jersey: John Wiley & Sons Inc. xvii,
400 pages.
[26] Nie, P., Ojo, O.A., and Li, Z., 2014. Numerical modeling of microstructure evolution during laser additive
manufacturing of a nickel-based superalloy. Acta Materialia. 77: p. 85-95.
[27] Zhu, L., 2009. Arrangement for the repair of the blades of BLISK drums by means of laser deposition welding,
US, Editor., Rolls Royce Deutschland Ltd and Co KG: US.
[28] Jones, J., Whittaker, M., Buckingham, R., Johnston, R., Bache, M., and Clark, D., 2017. Microstruc-
tural characterisation of a nickel alloy processed via blown powder direct laser deposition (DLD). Materials & Design.
117: p. 47-57.
[29] Kong, C.Y., Scudamore, R.J., and Allen, J., 2010. High-rate laser metal deposition of Inconel 718 component
using low heat-input approach. Physics Procedia. 5: p. 379-386.
[30] Sames, W.J., List, F.A., Pannala, S., Dehoff, R.R., and Babu, S.S., 2016. The metallurgy and processing
science of metal additive manufacturing. International Materials Reviews. 61(5): p. 315-360.
[31] Mun˜oz-Moreno, R., Divya, V.D., Driver, S.L., Messe´, O.M.D.M., Illston, T., Baker, S., Carpenter,
M.A., and Stone, H.J., 2016. Effect of heat treatment on the microstructure, texture and elastic anisotropy of
the nickel-based superalloy CM247LC processed by selective laser melting. Materials Science and Engineering: A.
674: p. 529-539.
[32] Blackwell, P.L., 2005. The mechanical and microstructural characteristics of laser-deposited IN718. Journal of
Materials Processing Technology. 170(1): p. 240-246.
[33] Ruttert, B., Ramsperger, M., Mujica Roncery, L., Lopez-Galilea, I., Ko¨rner, C., and Theisen, W.,
2016. Impact of hot isostatic pressing on microstructures of CMSX-4 Ni-base superalloy fabricated by selective
electron beam melting. Materials & Design. 110: p. 720-727.
[34] Carter, L.N., Wang, X., Read, N., Khan, R., Aristizabal, M., Essa, K., and Attallah, M.M., 2016.
Process optimisation of selective laser melting using energy density model for nickel based superalloys. Materials
Science and Technology. 32(7): p. 657-661.
131
[35] Durand-Charre, M., 1997. The Microstructure of superalloys. Amsterdam, The Netherlands: Gordon and Breach
Science Publishers. xiv, 124 pages.
[36] Pollock, T.M. and Tin, S., 2006. Nickel-Based Superalloys for Advanced Turbine Engines: Chemistry, Mi-
crostructure and Properties. Journal of Propulsion and Power. 22(2): p. 361-374.
[37] Schilke, P.W., Pepe, J.J., and Schwant, R.C., 1994. Alloy 706 Metallurgy and Turbine Wheel Application.
in Superalloys 718, 625, 706 and Various Derivatives Pittsburgh: The Minerals, Metals & Materials Society.
[38] Zhang, H.Y., Zhang, S.H., Cheng, M., and Li, Z.X., 2010. Deformation characteristics of δ phase in the
delta-processed Inconel 718 alloy. Materials Characterization. 61(1): p. 49-53.
[39] Rae, C.M.F. and Reed, R.C., 2001. The precipitation of topologically close-packed phases in rhenium-
containing superalloys. Acta Materialia. 49(19): p. 4113-4125.
[40] Chen, W., Chaturvedi, M.C., and Richards, N.L., 2001. Effect of boron segregation at grain boundaries
on heat-affected zone cracking in wrought INCONEL 718. Metallurgical and Materials Transactions A. 32(4): p.
931-939.
[41] Jena, A.K. and Chaturvedi, M.C., 1984. The role of alloying elements in the design of nickel-base superalloys.
Journal of Materials Science. 19(10): p. 3121-3139.
[42] Henderson, M.B., Arrell, D., Larsson, R., Heobel, M., and Marchant, G., 2004. Nickel based superalloy
welding practices for industrial gas turbine applications. Science and Technology of Welding and Joining. 9(1): p.
13-21.
[43] Chen, Y., Guo, Y., Xu, M., Ma, C., Zhang, Q., Wang, L., Yao, J., and Li, Z., 2019. Study on the
element segregation and Laves phase formation in the laser metal deposited IN718 superalloy by flat top laser and
gaussian distribution laser. Materials Science and Engineering: A. 754: p. 339-347.
[44] Radhakrishna, C.H. and Prasad Rao, K., 1997. The formation and control of Laves phase in superalloy 718
welds. Journal of Materials Science. 32(8): p. 1977-1984.
[45] Wilson, A.S., Christofidou, K.A., Evans, A., Hardy, M.C., and Stone, H.J., 2019. Comparison of Meth-
ods for Quantification of Topologically Close-Packed Phases in Ni-Based Superalloys. Metallurgical and Materials
Transactions A. 50(12): p. 5925-5934.
[46] Conduit, B.D., Jones, N.G., Stone, H.J., and Conduit, G.J., 2017. Design of a nickel-base superalloy using
a neural network. Materials & Design. 131: p. 358-365.
132
[47] DuPont, J.N., 1998. A Combined Solubility Product/New PHACOMP Approach for Estimating Temperatures of
Secondary Solidification Reactions in Superalloy Weld Metals Containing Nb and C. Metallurgical and Materials
Transactions A. 29: p. 1449–1456.
[48] Morinaga, M., Yukawa, N., Adachi, H., and Ezak, H., 1984. New PHACOMP and its Application to Alloy
Design, in Superalloys, M. Gell, Editor., The Metallurgical Society of AIME: Warrendale, PA.
[49] Saunders, N., Fahrmann, M., and Small, C.J., 2000. The application of CALPHAD calculations to Ni-based
superalloys, in Superalloys, K.A. Green, T.M. Pollock, and R.D. Kissinger, Editors., TMS: Warrendale, PA p. 803-
811.
[50] MacLachlan, D.W. and Knowles, D.M., 2001. Modelling and prediction of the stress rupture behaviour of
single crystal superalloys. Materials Science and Engineering: A. 302(2): p. 275-285.
[51] Rae, C.M.F. and Reed, R.C., 2007. Primary creep in single crystal superalloys: Origins, mechanisms and
effects. Acta Materialia. 55(3): p. 1067-1081.
[52] Zhu, Z., Basoalto, H., Warnken, N., and Reed, R.C., 2012. A model for the creep deformation behaviour
of nickel-based single crystal superalloys. Acta Materialia. 60(12): p. 4888-4900.
[53] Conduit, B.D., Illston, T., Baker, S., Duggappa, D.V., Harding, S., Stone, H.J., and Conduit, G.J.,
2019. Probabilistic neural network identification of an alloy for direct laser deposition. Materials & Design. 168: p.
107644.
[54] Shafiee, A., Nili-Ahmadabadi, M., Kim, H.-S., and Jahazi, M., 2020. Development and Microstructural
Characterization of a New Wrought High Entropy Superalloy. Metals and Materials International. 26(5): p. 591-602.
[55] Andersson, J.O., Helander, T., Ho¨glund, L., Shi, P., and Sundman, B., 2002. Thermo-Calc & DICTRA,
computational tools for materials science. Calphad. 26(2): p. 273-312.
[56] Tancret, F., 2007. Thermo-Calc and Dictra simulation of constitutional liquation of gamma prime (γ′) during
welding of Ni base superalloys. Computational Materials Science. 41(1): p. 13-19.
[57] Darolia, R., 2019. Development of strong, oxidation and corrosion resistant nickel-based superalloys: critical
review of challenges, progress and prospects. International Materials Reviews. 64(6): p. 355-380.
[58] He, D.-G., Lin, Y.C., Tang, Y., Li, L., Chen, J., Chen, M.-S., and Chen, X.-M., 2019. Influences of
solution cooling on microstructures, mechanical properties and hot corrosion resistance of a nickel-based superalloy.
Materials Science and Engineering: A. 746: p. 372-383.
[59] Hou, J.S., Guo, J.T., Wu, Y.X., Zhou, L.Z., and Ye, H.Q., 2010. Effect of hafnium on creep behavior of
a corrosion resistant nickel base superalloy. Materials Science and Engineering: A. 527(6): p. 1548-1554.
133
[60] Chen, Z., Dong, T., Qu, W., Ru, Y., Zhang, H., Pei, Y., Gong, S., and Li, S., 2019. Influence of Cr
content on hot corrosion and a special tube sealing test of single crystal nickel base superalloy. Corrosion Science.
156: p. 161-170.
[61] Saha, A. and Olson, G.B., 2007. Computer-aided design of transformation toughened blast resistant naval hull
steels: Part I. Journal of Computer-Aided Materials Design. 14(2): p. 177-200.
[62] Saha, A., Jung, J., and Olson, G.B., 2007. Prototype evaluation of transformation toughened blast resistant
naval hull steels: Part II. Journal of Computer-Aided Materials Design. 14(2): p. 201-233.
[63] Tancret, F., Bhadeshia, H.K.D.H., and MacKay, D.J.C., 2003. Design of a creep resistant nickel base
superalloy for power plant applications: Part 1 - Mechanical properties modelling. Materials Science and Technology.
19(3): p. 283-290.
[64] Tancret, F. and Bhadeshia, H.K.D.H., 2003. Design of a creep resistant nickel base superalloy for power
plant applications: Part 2 - Phase diagram and segregation simulation. Materials Science and Technology. 19(3): p.
291-295.
[65] Conduit, B.D., 2012. Probabilistic Alloy Design. PhD Thesis, University of Cambridge, Cambridge.
[66] Tancret, F., Sourmail, T., Yescas, M.A., Evans, R.W., McAleese, C., Singh, L., Smeeton, T., and
Bhadeshia, H.K.D.H., 2003. Design of a creep resistant nickel base superalloy for power plant applications: Part
3 - Experimental results. Materials Science and Technology. 19(3): p. 296-302.
[67] Smith, J.L., 2020. Advances in neural networks and potential for their application to steel metallurgy. Materials
Science and Technology. 36(17): p. 1805-1819.
[68] Tancret, F., 2012. Computational thermodynamics and genetic algorithms to design affordable γ′-strengthened
nickel–iron based superalloys. Modelling and Simulation in Materials Science and Engineering. 20(4): p. 045012.
[69] Tancret, F., 2013. Computational thermodynamics, Gaussian processes and genetic algorithms: combined tools
to design new alloys. Modelling and Simulation in Materials Science and Engineering. 21(4): p. 045013.
[70] Deschamps, A., Tancret, F., Benrabah, I.-E., De Geuser, F., and Van Landeghem, H.P., 2018. Com-
binatorial approaches for the design of metallic alloys. Comptes Rendus Physique. 19(8): p. 737-754.
[71] Conduit, B.D., Jones, N.G., Stone, H.J., and Conduit, G.J., 2018. Probabilistic design of a molybdenum-
base alloy using a neural network. Scripta Materialia. 146: p. 82-86.
[72] Zhang, W., Mehta, A., Desai, P.S., and Higgs, C.F., 2017. Machine Learning Enabled Powder Spreading
Process Map for Metal Additive Manufacturing (AM). in Solid Freeform Fabrication 2017. Pittsburgh, PA: Additive
Manufacturing.
134
[73] Baturynska, I., Semeniuta, O., and Martinsen, K., 2018. Optimization of Process Parameters for Powder
Bed Fusion Additive Manufacturing by Combination of Machine Learning and Finite Element Method: A Conceptual
Framework. Procedia CIRP. 67: p. 227-232.
[74] Kappes, B., Moorthy, S., Drake, D., Geerlings, H., and Stebner, A. 2018. Machine Learning to Optimize
Additive Manufacturing Parameters for Laser Powder Bed Fusion of Inconel 718. in 9th International Symposium
on Superalloy 718 & Derivatives: Energy, Aerospace, and Industrial Applications. Cham: Springer International
Publishing.
[75] Reed, R.C., Tao, T., and Warnken, N., 2009. Alloys-By-Design: Application to nickel-based single crystal
superalloys. Acta Materialia. 57(19): p. 5898-5913.
[76] Tang, Y.T., Panwisawas, C., Ghoussoub, J.N., Gong, Y., Clark, J.W.G., Ne´meth, A.A.N., McCart-
ney, D.G., and Reed, R.C., 2021. Alloys-by-design: Application to new superalloys for additive manufacturing.
Acta Materialia. 202: p. 417-436.
[77] Tancret, F., Bhadeshia, H.K.D.H., and Mackay, D.J.C., 1999. Comparison of Artificial Neural Networks
with Gaussian Processes to Model the Yield Strength of Nickel-base Superalloys. ISIJ International. 39(10): p. 1020-
1026.
[78] Ramakrishnan, A. and Dinda, G.P., 2019. Direct laser metal deposition of Inconel 738. Materials Science and
Engineering: A. 740-741: p. 1-13.
[79] Collins, D.M. and Stone, H.J., 2014. A modelling approach to yield strength optimisation in a nickel-base
superalloy. International Journal of Plasticity. 54: p. 96-112.
[80] Geng, P., Li, W., Zhang, X., Deng, Y., Kou, H., Ma, J., Shao, J., Chen, L., and Wu, X., 2017. A
theoretical model for yield strength anomaly of Ni-base superalloys at elevated temperature. Journal of Alloys and
Compounds. 706: p. 340-343.
[81] Coakley, J., Dye, D., and Basoalto, H., 2011. Creep and creep modelling of a multimodal nickel-base superalloy.
Acta Materialia. 59(3): p. 854-863.
[82] Messe´, O.M., Barnard, J.S., Pickering, E.J., Midgley, P.A., and Rae, C.M.F., 2014. On the precipitation
of delta phase in ALLVAC® 718Plus. Philosophical Magazine. 94(10): p. 1132-1152.
[83] Hassan, B. and Corney, J., 2017. Grain boundary precipitation in Inconel 718 and ATI 718Plus. Materials
Science and Technology. 33(16): p. 1879-1889.
[84] Hillert, M., 2008. Phase equilibria, phase diagrams and phase transformations : their thermodynamic basis. 2nd
ed. Cambridge ; New York: Cambridge University Press. xiv, 510 pages.
135
[85] Keller, T., Lindwall, G., Ghosh, S., Ma, L., Lane, B.M., Zhang, F., Kattner, U.R., Lass, E.A.,
Heigel, J.C., Idell, Y., Williams, M.E., Allen, A.J., Guyer, J.E., and Levine, L.E., 2017. Application of
finite element, phase-field, and CALPHAD-based methods to additive manufacturing of Ni-based superalloys. Acta
Materialia. 139: p. 244-253.
[86] DuPont, J.N., Robino, C.V., and Marder, A.R., 1999. Modelling mushy zones in welds of multicomponent
alloys: implications for solidification cracking. Science and Technology of Welding and Joining. 4(1): p. 1-14.
[87] Chen, Y., Zhang, K., Huang, J., Hosseini, S.R.E., and Li, Z., 2016. Characterization of heat affected zone
liquation cracking in laser additive manufacturing of Inconel 718. Materials & Design. 90: p. 586-594.
[88] Panwisawas, C., Sovani, Y., Anderson, M., Turner, R., Palumbo, N., Saunders, B., Choquet, I.,
Brooks, J.W., and Basoalto, H., 2016. A Multi-Scale Multi-Physics Approach to Modelling of Additive Manu-
facturing in Nickel-Based Superalloys, in Superalloys 2016. John Wiley & Sons: Pennsylvania. p. 1021-1030.
[89] Shamsaei, N., Yadollahi, A., Bian, L., and Thompson, S.M., 2015. An overview of Direct Laser Deposition
for additive manufacturing; Part II: Mechanical behavior, process parameter optimization and control. Additive
Manufacturing. 8: p. 12-35.
[90] Jia, Q. and Gu, D., 2014. Selective laser melting additive manufacturing of Inconel 718 superalloy parts: Den-
sification, microstructure and properties. Journal of Alloys and Compounds. 585: p. 713-721.
[91] Wang, Z., Denlinger, E., Michaleris, P., Stoica, A.D., Ma, D., and Beese, A.M., 2017. Residual stress
mapping in Inconel 625 fabricated through additive manufacturing: Method for neutron diffraction measurements
to validate thermomechanical model predictions. Materials & Design. 113: p. 169-177.
[92] Li, C., Liu, Z.Y., Fang, X.Y., and Guo, Y.B., 2018. Residual Stress in Metal Additive Manufacturing.
Procedia CIRP. 71: p. 348-353.
[93] Helmer, H.E., Ko¨rner, C., and Singer, R.F., 2014. Additive manufacturing of nickel-based superalloy Inconel
718 by selective electron beam melting: Processing window and microstructure. Journal of Materials Research.
29(17): p. 1987-1996.
[94] Scime, L. and Beuth, J., 2019. Melt pool geometry and morphology variability for the Inconel 718 alloy in a
laser powder bed fusion additive manufacturing process. Additive Manufacturing. 29: p. 100830.
[95] Dass, A. and Moridi, A., 2019. State of the Art in Directed Energy Deposition: From Additive Manufacturing
to Materials Design. Coatings. 9(418).
[96] Zhang, D., Niu, W., Cao, X., and Liu, Z., 2015. Effect of standard heat treatment on the microstructure
and mechanical properties of selective laser melting manufactured Inconel 718 superalloy. Materials Science and
Engineering: A. 644: p. 32-40.
136
[97] Ojo, O.A. and Chaturvedi, M.C., 2007. Liquation Microfissuring in the Weld Heat-Affected Zone of an
Overaged Precipitation-Hardened Nickel-Base Superalloy. Metallurgical and Materials Transactions A. 38A.
[98] Stevens, R.A. and Flewitt, P.E.J., 1979. The effects of γ′ precipitate coarsening during isothermal aging and
creep of the nickel-base superalloy IN-738. Materials Science and Engineering. 37(3): p. 237-247.
[99] Bambach, M., Sizova, I., Silze, F., and Schnick, M., 2018. Hot workability and microstructure evolution of
the nickel-based superalloy Inconel 718 produced by laser metal deposition. Journal of Alloys and Compounds. 740:
p. 278-287.
[100] Chen, Y., Lu, F., Zhang, K., Nie, P., Hosseini, S.R.E., Feng, K., and Li, Z., 2016. Dendritic mi-
crostructure and hot cracking of laser additive manufactured Inconel 718 under improved base cooling. Journal of
Alloys and Compounds. 670: p. 312-321.
[101] Dinda, G.P., Dasgupta, A.K., and Mazumder, J., 2009. Laser aided direct metal deposition of Inconel
625 superalloy: Microstructural evolution and thermal stability. Materials Science and Engineering: A. 509(1): p.
98-104.
[102] Sames, W.J., Medina, F., Peter, W.H., Babu, S.S., and Dehoff, R.R., 2014. Effect of Process Control
and Powder Quality on Inconel 718 Produced Using Electron Beam Melting, in 8th International Symposium on
Superalloy 718 and Derivativess: Pittsburgh, PA. p. 409-423.
[103] Minet, K., Saharan, A., Loesser, A., and Raitanen, N., 2019 Superalloys, powders, process monitoring
in additive manufacturing, in Additive Manufacturing for the Aerospace Industry, F. Froes and R. Boyer, Editors.,
Elsevier. p. 163-185.
[104] Zhong, C., Chen, J., Linnenbrink, S., Gasser, A., Sui, S., and Poprawe, R., 2016. A comparative
study of Inconel 718 formed by High Deposition Rate Laser Metal Deposition with GA powder and PREP powder.
Materials & Design. 107: p. 386-392.
[105] Harrison, N.J., Todd, I., and Mumtaz, K., 2015. Reduction of micro-cracking in nickel superalloys processed
by Selective Laser Melting: A fundamental alloy design approach. Acta Materialia. 94: p. 59-68.
[106] Yang, K., Xie, H., Sun, C., Zhao, X., and Li, F., 2019. Influence of Vanadium on the Microstructure of
IN718 Alloy by Laser Cladding. Materials. 12: p. 3839.
[107] Moussaoui, K., Rubio, W., Mousseigne, M., Sultan, T., and Rezai, F., 2018. Effects of Selective Laser
Melting additive manufacturing parameters of Inconel 718 on porosity, microstructure and mechanical properties.
Materials Science and Engineering: A. 735: p. 182-190.
137
[108] Carter, L.N., Martin, C., Withers, P.J., and Attallah, M.M., 2014. The influence of the laser scan strategy
on grain structure and cracking behaviour in SLM powder-bed fabricated nickel superalloy. Journal of Alloys and
Compounds. 615: p. 338-347.
[109] Carter, L.N., 2013. Selective Laser Melting of Nickel Superalloys for High Temperature Applications. PhD
Thesis, University of Birmingham, Birmingham, U.K.
[110] Parimi, L.L., Ravi, G.A., Clark, D., and Attallah, M.M., 2014. Microstructural and texture development
in direct laser fabricated IN718. Materials Characterization. 89: p. 102-111.
[111] Dinda, G.P., Dasgupta, A.K., and Mazumder, J., 2012. Texture control during laser deposition of nickel-
based superalloy. Scripta Materialia. 67(5): p. 503-506.
[112] Dehoff, R.R., Kirka, M.M., Sames, W.J., Bilheux, H., Tremsin, A.S., Lowe, L.E., and Babu,
S.S., 2015. Site specific control of crystallographic grain orientation through electron beam additive manufacturing.
Materials Science and Technology. 31(8): p. 931-938.
[113] Popovich, V.A., Borisov, E.V., Popovich, A.A., Sufiiarov, V.S., Masaylo, D.V., and Alzina, L.,
2017. Functionally graded Inconel 718 processed by additive manufacturing: Crystallographic texture, anisotropy of
microstructure and mechanical properties. Materials & Design. 114: p. 441-449.
[114] Wan, H.Y., Zhou, Z.J., Li, C.P., Chen, G.F., and Zhang, G.P., 2018. Effect of scanning strategy on
grain structure and crystallographic texture of Inconel 718 processed by selective laser melting. Journal of Materials
Science & Technology. 34(10): p. 1799-1804.
[115] Geiger, F., Kunze, K., and Etter, T., 2016. Tailoring the texture of IN738LC processed by selective laser
melting (SLM) by specific scanning strategies. Materials Science and Engineering: A. 661: p. 240-246.
[116] Xuan, W., Ren, Z., and Li, C., 2015. Effect of a high magnetic field on microstructures of Ni-based superalloy
during directional solidification. Journal of Alloys and Compounds. 620: p. 10-17.
[117] Ma, D., Stoica, A.D., Wang, Z., and Beese, A.M., 2017. Crystallographic texture in an additively manu-
factured nickel-base superalloy. Materials Science and Engineering: A. 684: p. 47-53.
[118] Kunze, K., Etter, T., Gra¨sslin, J., and Shklover, V., 2015. Texture, anisotropy in microstructure and me-
chanical properties of IN738LC alloy processed by selective laser melting (SLM). Materials Science and Engineering:
A. 620: p. 213-222.
[119] Wei, H.L., Mazumder, J., and DebRoy, T., 2015. Evolution of solidification texture during additive manu-
facturing. Scientific Reports. 5(1): p. 16446.
138
[120] Nie, J., Chen, C., Shuai, S., Liu, X., Zhao, R., Wang, J., Liao, H., and Ren, Z., 2020. Effect of Static
Magnetic Field on the Evolution of Residual Stress and Microstructure of Laser Remelted Inconel 718 Superalloy.
Journal of Thermal Spray Technology. 29(6): p. 1410-1423.
[121] Wei, X., Xu, M., Wang, Q., Zhang, M., Liu, W., Xu, J., Chen, J., Lu, H., and Yu, C., 2016. Effect of
local texture and precipitation on the ductility dip cracking of ERNiCrFe-7A Ni-based overlay. Materials & Design.
110: p. 90-98.
[122] Francis, L.F., Stadler, B.J.H., and Roberts, C.C., 2016. Materials Processing: A Unified Approach to
Processing of Metals, Ceramics and Polymers. 1-597.
[123] Ma, M., Wang, Z., and Zeng, X., 2015. Effect of energy input on microstructural evolution of direct laser
fabricated IN718 alloy. Materials Characterization. 106: p. 420-427.
[124] Sadowski, M., Ladani, L., Brindley, W., and Romano, J., 2016. Optimizing quality of additively manu-
factured Inconel 718 using powder bed laser melting process. Additive Manufacturing. 11: p. 60-70.
[125] Selcuk, C., 2011. Laser metal deposition for powder metallurgy parts. Powder Metallurgy. 54(2): p. 94-99.
[126] Costa, L. and Vilar, R., 2009. Laser powder deposition. Rapid prototyping journal. 15(4): p. 264-279.
[127] Xiao, H., Li, S., Han, X., Mazumder, J., and Song, L., 2017. Laves phase control of Inconel 718 alloy
using quasi-continuous-wave laser additive manufacturing. Materials & Design. 122: p. 330-339.
[128] Li, S., Xiao, H., Liu, K., Xiao, W., Li, Y., Han, X., Mazumder, J., and Song, L., 2017. Melt-pool
motion, temperature variation and dendritic morphology of Inconel 718 during pulsed- and continuous-wave laser
additive manufacturing: A comparative study. Materials & Design. 119: p. 351-360.
[129] Nadammal, N., Cabeza, S., Mishurova, T., Thiede, T., Kromm, A., Seyfert, C., Farahbod, L.,
Haberland, C., Schneider, J.A., Portella, P.D., and Bruno, G., 2017. Effect of hatch length on the devel-
opment of microstructure, texture and residual stresses in selective laser melted superalloy Inconel 718. Materials
& Design. 134: p. 139-150.
[130] Basak, A. and Das, S., 2016. Epitaxy and Microstructure Evolution in Metal Additive Manufacturing. Annual
Review of Materials Research. 46(1): p. 125-149.
[131] Zhang, J., 2003. Effect of Ti and Ta on hot cracking susceptibility of directionally solidified Ni-based superalloy
IN792. Scripta Materialia. 48(6): p. 677-681.
[132] Liu, F., Lin, X., Yang, G., Song, M., Chen, J., and Huang, W., 2011. Microstructure and residual stress
of laser rapid formed Inconel 718 nickel-base superalloy. Optics & Laser Technology. 43(1): p. 208-213.
139
[133] Ga¨umann, M., Henry, S., Cle´ton, F., Wagnie`re, J.D., and Kurz, W., 1999. Epitaxial laser metal forming:
analysis of microstructure formation. Materials Science and Engineering: A. 271(1): p. 232-241.
[134] Ni, M., Chen, C., Wang, X., Wang, P., Li, R., Zhang, X., and Zhou, K., 2017. Anisotropic tensile
behavior of in situ precipitation strengthened Inconel 718 fabricated by additive manufacturing. Materials Science
and Engineering: A. 701: p. 344-351.
[135] Zhang, Q., Yao, J., and Mazumder, J., 2011. Laser Direct Metal Deposition Technology and Microstructure
and Composition Segregation of Inconel 718 Superalloy. Journal of Iron and Steel Research, International. 18(4):
p. 73-78.
[136] Oguntuase, Q., 2019. Development of Post-Deposition Heat Treatment to Improve Mechanical Properties of
Wire-Arc Additive Manufactured ATI 718Plus Superalloy. Master of Science Thesis, University of Manitoba, Win-
nipeg, Canada.
[137] Wang, F., 2012. Mechanical property study on rapid additive layer manufacture Hastelloy® X alloy by selective
laser melting technology. The International Journal of Advanced Manufacturing Technology. 58(5): p. 545-551.
[138] Forsey, A.N., Das, Y.B., Simm, T.H., Clark, D., Boswell, J., Gungor, S., and Moat, R.J., 2018.
Mechanical property heterogeneity in additively manufactured nickel superalloy. Materials Science and Engineering:
A. 712: p. 681-684.
[139] Chen, J. and Xue, L., 2010. Process-induced microstructural characteristics of laser consolidated IN-738 su-
peralloy. Materials Science and Engineering: A. 527(27): p. 7318-7328.
[140] Boswell, J.H., Clark, D., Li, W., and Attallah, M.M., 2019. Cracking during thermal post-processing of
laser powder bed fabricated CM247LC Ni-superalloy. Materials & Design. 174: p. 107793.
[141] Basak, A., 2019. Additive Manufacturing of High-Gamma Prime Nickel-Based Superalloys through Selective
Laser Melting (SLM), in Solid Freeform Fabrication. Additive Manufacturing.
[142] Denlinger, E.R., Heigel, J.C., Michaleris, P., and Palmer, T.A., 2015. Effect of inter-layer dwell time
on distortion and residual stress in additive manufacturing of titanium and nickel alloys. Journal of Materials
Processing Technology. 215: p. 123-131.
[143] Rickenbacher, L., Etter, T., Ho¨vel, S., and Wegener, K., 2013. High temperature material properties of
IN738LC processed by selective laser melting (SLM) technology. Rapid Prototyping Journal. 19(4): p. 282-290.
[144] Trosch, T., Stro¨ßner, J., Vo¨lkl, R., and Glatzel, U., 2015. Microstructure and mechanical properties of
selective laser melted Inconel 718 compared to forging and casting. Materials Letters. 164.
140
[145] Zhang, X., Chen, H., Xu, L., Xu, J., Ren, X., and Chen, X., 2019. Cracking mechanism and susceptibility
of laser melting deposited Inconel 738 superalloy. Materials & Design. 183: p. 108105.
[146] Basak, A. and Das, S., 2017. Additive Manufacturing of Nickel-Base Superalloy Rene´ N5 through Scanning
Laser Epitaxy (SLE) – Material Processing, Microstructures, and Microhardness Properties. Advanced Engineering
Materials. 19(3): p. 1600690.
[147] Haafkens, M.H. and Matthey, J.H.G., 1982. A new approach to the weldability of nickel-base As-cast and
power metallurgy superalloys. Weld. J. 61(11).
[148] Grodzki, J., Hartmann, N., Rettig, R., Affeldt, E., and Singer, R.F., 2016. Effect of B, Zr, and C
on Hot Tearing of a Directionally Solidified Nickel-Based Superalloy. Metallurgical and Materials Transactions A.
47(6): p. 2914-2926.
[149] Shao, S., Khonsari, M.M., Guo, S., Meng, W.J., and Li, N., 2019. Overview: Additive Manufacturing
Enabled Accelerated Design of Ni-based Alloys for Improved Fatigue Life. Additive Manufacturing. 29: p. 100779.
[150] Meier, C., Penny, R., Zou, Y., Gibbs, J., and Hart, A., 2018. Thermophysical Phenomena in Metal
Additive Manufacturing by Selective Laser Melting: Fundamentals, Modeling, Simulation and Experimentation.
Annual Review of Heat Transfer.
[151] Panwisawas, C., Qiu, C.L., Anderson, M.J., Sovani, Y., Turner, R.P., Attallah, M.M., Brooks,
J.W., and Basoalto, H.C., 2017. Mesoscale modelling of selective laser melting: Thermal fluid dynamics and
microstructural evolution. Computational Materials Science. 126: p. 479-490.
[152] Panwisawas, C., Qiu, C.L., Sovani, Y., Brooks, J.W., Attallah, M.M., and Basoalto, H.C., 2015. On
the role of thermal fluid dynamics into the evolution of porosity during selective laser melting. Scripta Materialia.
105: p. 14-17.
[153] Boellinghaus, T. and Herold, H., 2005. Hot Cracking Phenomena in Welds. Hot Cracking Phenomena in
Welds. Berlin: Springer-Verlag Berlin Heidelberg.
[154] Boellinghaus, T., Herold, H., Cross, C.E., and Lippold, J.C., 2008. Hot Cracking Phenomena in Welds
II. Berlin: Springer-Verlag Berlin Heidelberg. 467.
[155] Lippold, J.C., Boellnghaus, T., and Cross, C.E., 2011. Hot Cracking Phenomena in Welds III. Berlin:
Springer, Berlin, Heidelberg.
[156] Boellinghaus, T., Lippold, J.C., and Cross, C.E., 2016. Cracking Phenomena in Welds IV. Switzerland:
Springer International Publishing.
141
[157] Young, G.A., Capobianco, T.E., Penik, M.A., Morris, B.W., and McGee, J.J., 2008. The Mechanism
of Ductility Dip Cracking in Nickel-Chromium Alloys. Welding Journal. 83: p. 31-43.
[158] Wang, H., Zhang, X., Wang, G.B., Shen, J., Zhang, G.Q., Li, Y.P., and Yan, M., 2019. Selective laser
melting of the hard-to-weld IN738LC superalloy: Efforts to mitigate defects and the resultant microstructural and
mechanical properties. Journal of Alloys and Compounds. 807: p. 151662.
[159] Moosavy, H.N., Aboutalebi, M.-R., and Seyedein, S.H., 2012. An analytical algorithm to predict weld-
ability of precipitation-strengthened nickel-base superalloys. Journal of Materials Processing Technology. 212(11):
p. 2210-2218.
[160] Griffiths, S., Ghasemi Tabasi, H., Ivas, T., Maeder, X., De Luca, A., Zweiacker, K., Wro´bel, R.,
Jhabvala, J., Loge´, R.E., and Leinenbach, C., 2020. Combining alloy and process modification for micro-crack
mitigation in an additively manufactured Ni-base superalloy. Additive Manufacturing. 36: p. 101443.
[161] Zhang, J., 2004. Hot Tearing in Directionally Solidified Ni-Based Superalloys. in Superalloys, K.A. Green, et
al., Editors., TMS (The Minerals, Metals & Materials Society): Champion, Pennsylvania.
[162] Zhang, J. and Singer, R.F., 2004. Effect of Zr and B on castability of Ni-based superalloy IN792. Metallurgical
and Materials Transactions A. 35(4): p. 1337-1342.
[163] Zhang, J. and Singer, R.F., 2002. Hot tearing of nickel-based superalloys during directional solidification. Acta
Materialia. 50(7): p. 1869-1879.
[164] Kou, S., 2015. A criterion for cracking during solidification. Acta Materialia. 88: p. 366-374.
[165] Jacobsson, J., Andersson, J.O., Brederholm, A., and Ha¨nninen, H., 2017. Weldability of superalloys
alloy 718 and ATI® 718Plus— – A study performed by Varestraint testing. Materials Testing. 59(9): p. 769-773.
[166] Dye, D., Hunziker, O., and Reed, R.C., 2001. Numerical analysis of the weldability of superalloys. Acta
Materialia. 49(4): p. 683-697.
[167] Wang, X., Carter, L.N., Pang, B., Attallah, M.M., and Loretto, M.H., 2017. Microstructure and yield
strength of SLM-fabricated CM247LC Ni-Superalloy. Acta Materialia. 128: p. 87-95.
[168] Moat, R.J., Pinkerton, A.J., Li, L., Withers, P.J., and Preuss, M., 2011. Residual stresses in laser direct
metal deposited Waspaloy. Materials Science and Engineering: A. 528(6): p. 2288-2298.
[169] Tien, J.K., Collier, J.P., Bretz, P.L., and Hendrix, B.C., 1990. Raising the High Temperature Limit of
IN718 - Designing Ticolloy. in High temperature materials for power engineering conference. The Netherlands:
Kluwer Academic Publishers.
142
[170] Singh, S. and Andersson, J.O., 2019. Varestraint weldability testing of cast ATI® 718Plus—–a comparison
to cast Alloy 718. Welding in the World. 63(2): p. 389-399.
[171] Kienl, C., 2019. Hot forging of the nickel-base superalloy ATI 718Plus. PhD Thesis, University of Cambridge,
Cambridge.
[172] Kennedy, R.L., 2005. Allvac® 718Plus—, Superalloy for the Next Forty Years, in Superalloys 718, 625, 706 and
Derivatives, E.A. Loria, Editor., TMS.
[173] Agarwal, R.K., 2012. Recent Advances in Aircraft Technology. Rijeka, Croatia: Intech Open.
[174] Cao, X., Rivaux, B., Jahazi, M., Cuddy, J., and Birur, A., 2009. Effect of pre- and post-weld heat
treatment on metallurgical and tensile properties of Inconel 718 alloy butt joints welded using 4 kW Nd:YAG laser.
Journal of Materials Science. 44(17): p. 4557-4571.
[175] Cloots, M., Uggowitzer, P.J., and Wegener, K., 2016. Investigations on the microstructure and crack
formation of IN738LC samples processed by selective laser melting using Gaussian and doughnut profiles. Materials
& Design. 89: p. 770-784.
[176] Amouyal, Y. and Seidman, D.N., 2011. The role of hafnium in the formation of misoriented defects in Ni-
based superalloys: An atom-probe tomographic study. Acta Materialia. 59(9): p. 3321-3333.
[177] Liu, T., Dong, J.S., Xie, G., Wang, L., and Lou, L.H., 2016. Effect of silicon on microstructure and stress
rupture properties of a corrosion resistant Ni-based superalloy during long term thermal exposure. Materials Science
and Engineering: A. 656: p. 75-83.
[178] Engeli, R., Etter, T., Geiger, F., Stankowski, A., and Wegener, K., 2015. Effect of Si on the SLM
processability of IN738LC, in Solid freeform fabrication symposium: Austin, Tx.
[179] Tsai, Y.-L., Wang, S.-F., Bor, H.-Y., and Hsu, Y.-F., 2014. Effects of Zr addition on the microstructure
and mechanical behavior of a fine-grained nickel-based superalloy at elevated temperatures. Materials Science and
Engineering: A. 607: p. 294-301.
[180] Hariharan, A., Lu, L., Risse, J., Kostka, A., Gault, B., Ja¨gle, E.A., and Raabe, D., 2019.
Misorientation-dependent solute enrichment at interfaces and its contribution to defect formation mechanisms dur-
ing laser additive manufacturing of superalloys. Physical Review Materials. 3(12): p. 123602.
[181] Francis, J.A., Stone, H.J., Kundu, S., Bhadeshia, H.K.D.H., Rogge, R.B., Withers, P.J., and
Karlsson, L., 2009. The Effects of Filler Metal Transformation Temperature on Residual Stresses in a High
Strength Steel Weld. Journal of Pressure Vessel Technology. 131(4).
143
[182] Tsao, T.-K., Yeh, A.-C., Kuo, C.-M., Kakehi, K., Murakami, H., Yeh, J.-W., and Jian, S.-R., 2017.
The High Temperature Tensile and Creep Behaviors of High Entropy Superalloy. Scientific Reports. 7(1): p. 12658.
[183] Mahesh, C. and Sondhi, S.K., 2011. First-principle calculation of stacking fault energies in Ni and Ni-Co
alloy. Journal of Applied Physics. 109(10): p. 103525.
[184] Koizumi, Y., Suzuki, S., Yamanaka, K., Lee, B.S., Sato, K., Li, Y., Kurosu, S., Matsumoto, H.,
and Chiba, A., 2013. Strain-induced martensitic transformation near twin boundaries in a biomedical Co–Cr–Mo
alloy with negative stacking fault energy. Acta Materialia. 61(5): p. 1648-1661.
[185] Kok, Y., Tan, X.P., Wang, P., Nai, M.L.S., Loh, N.H., Liu, E., and Tor, S.B., 2018. Anisotropy
and heterogeneity of microstructure and mechanical properties in metal additive manufacturing: A critical review.
Materials & Design. 139: p. 565-586.
[186] Yan, F., Xiong, W., and Faierson, E.J., 2017. Grain Structure Control of Additively Manufactured Metallic
Materials. Materials. 10(11): p. 1260.
[187] Xiong, Y., Ying Wei, X., Du, J., Yang, A., Liu, L., and Zeng, D., 2004. Grain refinement of superalloy
IN718C by the addition of inoculants. Metallurgical and Materials Transactions A. 35(7): p. 2111-2114.
[188] Liu, L., Huang, T., Xiong, Y., Yang, A., Zhao, Z., Zhang, R., and Li, J., 2004. Cast Structure and
Mechanical Properties of Fine Grained Superalloy K4169 by Addition of Refiners. Proceedings of the International
Symposium on Superalloys. 15.
[189] Ho, I.T., Chen, Y.-T., Yeh, A.-C., Chen, C.-P., and Jen, K.-K., 2018. Microstructure evolution induced
by inoculants during the selective laser melting of IN718. Additive Manufacturing. 21: p. 465-471.
[190] Zhang, D., Prasad, A., Bermingham, M.J., Todaro, C.J., Benoit, M.J., Patel, M.N., Qiu, D.,
StJohn, D.H., Qian, M., and Easton, M.A., 2020. Grain Refinement of Alloys in Fusion-Based Additive
Manufacturing Processes. Metallurgical and Materials Transactions A. 51(9): p. 4341-4359.
[191] Hsu, T.-H., Chang, K.-C., Chang, Y.-J., Ho, I.-T., Tin, S., Li, C.-W., Kakehi, K., Chen, C.-P.,
Jen, K.-K., Hsieh, H.-Y., and Yeh, A.-C., 2020. Effect of Carbide Inoculants Additions in IN718 Fabricated
by Selective Laser Melting Process, in Superalloys 2020, S. Tin, Editor., The Minerals, Metals & Materials Series.
[192] Bermingham, M.J., StJohn, D.H., Krynen, J., Tedman-Jones, S., and Dargusch, M.S., 2019. Pro-
moting the columnar to equiaxed transition and grain refinement of titanium alloys during additive manufacturing.
Acta Materialia. 168: p. 261-274.
[193] Wen, X., Wang, Q., Mu, Q., Kang, N., Sui, S., Yang, H., Lin, X., and Huang, W., 2019. Laser
solid forming additive manufacturing TiB2 reinforced 2024Al composite: Microstructure and mechanical properties.
Materials Science and Engineering: A. 745: p. 319-325.
144
[194] Ho, I.T., Hsu, T.H., Chang, Y.J., Li, C.W., Chang, K.C., Tin, S., Kakehi, K., and Yeh, A.C., 2020.
Effects of CoAl2O4 inoculants on microstructure and mechanical properties of IN718 processed by selective laser
melting. Additive Manufacturing. 35: p. 101328.
[195] Tiparti, D., Ho, I.T., Chang, K.-C., Hsu, T.-H., Yeh, A.-C., and Tin, S., 2021. Understanding the
Effects of CoAl2O4 Inoculant Additions on Microstructure in Additively Manufactured Inconel 718 Processed Via
Selective Laser Melting. Metallurgical and Materials Transactions A. 52(6): p. 2630-2641.
[196] Chen, S.L., Daniel, S., Zhang, F., Chang, Y.A., Yan, X.Y., Xie, F.Y., Schmid-Fetzer, R., and
Oates, W.A., 2002. The PANDAT software package and its applications. Calphad. 26(2): p. 175-188.
[197] Lee, J.S., Gu, J.H., Jung, H.M., Kim, E.H., Jung, Y.G., and Lee, J.H., 2014. Directional Solidification
Microstructure Control in CM247LC Superalloy. Materials Today: Proceedings. 1(1): p. 3-10.
[198] Yan, F., Liu, S., Hu, C., Wang, C., and Hu, X., 2017. Liquation cracking behavior and control in the heat
affected zone of GH909 alloy during Nd: YAG laser welding. Journal of Materials Processing Technology. 244: p.
44-50.
[199] O¨zgu¨n, O¨., Gu¨lsoy, H.O¨., Yılmaz, R., and Fındık, F., 2013. Microstructural and mechanical characteriza-
tion of injection molded 718 superalloy powders. Journal of Alloys and Compounds. 576: p. 140-153.
[200] Raghavan, N., Dehoff, R.R., Pannala, S., Simunovic, S., Kirka, M.K., Turner, J., Carlson, N., and
Babu, S.S., 2016. Numerical modeling of heat-transfer and the influence of process parameters on tailoring the
grain morphology of IN718 in electron beam additive manufacturing. Acta Materialia. 112: p. 303-314.
[201] Bian, L., Thompson, S.M., and Shamsaei, N., 2015. Mechanical Properties and Microstructural Features
of Direct Laser-Deposited Ti-6Al-4V. JOM. 67(3): p. 629-638.
[202] Goddard, C.J.L., 2015. A Framework for a Dual Alloy Disc Development. PhD Thesis, University of Cambridge,
Cambridge.
[203] Cao, J., Liu, F., Lin, X., Huang, C., Chen, J., and Huang, W., 2013. Effect of overlap rate on recrystal-
lization behaviors of Laser Solid Formed Inconel 718 superalloy. Optics & Laser Technology. 45: p. 228-235.
[204] Migliori, A. and Sarrao, J., 1997. Resonant Ultrasound Spectroscopy: Applications to Physics, Material Mea-
surements and Nondestructive Evaluation. Wiley.
[205] McKnight, R.E.A., Carpenter, M.A., Darling, T.W., Buckley, A., and Taylor, P.A., 2007. Acoustic
dissipation associated with phase transitions in lawsonite, CaAl2Si2O7(OH)2·H2O. American Mineralogist. 92(10):
p. 1665-1672.
145
[206] McKnight, R.E.A., Moxon, T., Buckley, A., Taylor, P.A., Darling, T.W., and Carpenter, M.A.,
2008. Grain size dependence of elastic anomalies accompanying the α–β phase transition in polycrystalline quartz.
Journal of Physics: Condensed Matter. 20: p. 075229.
[207] Marmier, A., Lethbridge, Z.A.D., Walton, R.I., Smith, C.W., Parker, S.C., and Evans, K.S., 2010.
EIAM: a computer program for the analysis and representation of anisotropic elastic properties. Computer Physics
Communications. 181(12): p. 2102-2115.
[208] Ravindran, P., Fast, L., Korzhavyi, P.A., Johansson, B., Wills, J., and Eriksson, O., 1998. Density
functional theory for calculation of elastic properties of orthorhombic crystals: Application to TiSi2. Journal of
Applied Physics. 84(9): p. 4891-4904.
[209] Oxford-Instruments, 2006. Channel 5-HKL user guide: Denmark.
[210] Rasband, W.S., 1997-2018. ImageJ. U. S. National Institutes of Health: Bethesda, Maryland, USA.
[211] Zupanicˇ, F., Boncˇina, T., Krizˇman, A., and Tichelaar, F.D., 2001. Structure of continuously cast Ni-based
superalloy Inconel 713C. Journal of Alloys and Compounds. 329(1): p. 290-297.
[212] Hosseini, S.A., Abbasi, S.M., and Madar, K.Z., 2018. The Effect of Boron and Zirconium on the Structure
and Tensile Properties of the Cast Nickel-Based Superalloy ATI 718Plus. Journal of Materials Engineering and
Performance. 27(6): p. 2815-2826.
[213] Mostafa, A., Picazo Rubio, I., Brailovski, V., Jahazi, M., and Medraj, M., 2017. Structure, Texture
and Phases in 3D Printed IN718 Alloy Subjected to Homogenization and HIP Treatments. Metals. 7(6): p. 196.
[214] Messe´, O.M.D.M., Mun˜oz-Moreno, R., Illston, T., Baker, S., and Stone, H.J., 2018. Metastable car-
bides and their impact on recrystallisation in IN738LC processed by selective laser melting. Additive Manufacturing.
22: p. 394-404.
[215] Kumara, C., Segerstark, A., Hanning, F., Dixit, N., Joshi, S., Moverare, J., and Nyle´n, P., 2019.
Microstructure Modelling of Laser Metal Powder Directed Energy Deposition of Alloy 718. Additive Manufacturing.
25: p. 357-364.
[216] Buckingham, R.C., Argyrakis, C., Hardy, M.C., and Birosca, S., 2016. The effect of strain distribution
on microstructural developments during forging in a newly developed nickel base superalloy. Materials Science and
Engineering: A. 654: p. 317-328.
[217] Paul, H., Miszczyk, M.M., and Driver, J.H., 2015. Crystallographic aspects of nucleation and growth during
primary recrystallization in stable single crystals of Al and Al-1%Mn alloy. IOP Conference Series: Materials Science
and Engineering. 89: p. 012010.
146
[218] Driver, S.L., 2017. Phase Transitions and Relaxation Phenomena in Alloys as Observed using Reonant Ultra-
sound Spectroscopy. PhD Thesis, University of Cambridge, Cambridge.
[219] Wei, X., Wang, Q., Zhang, M., Yu, C., Chen, J., Chen, J., Lu, H., and Xu, J., 2019. Effect of
microtextures on the intermediate temperature fracture behavior of additive-manufactured Ni-based overlay. Journal
of Alloys and Compounds. 795: p. 103-113.
[220] Stone, H.J., Reed, R.C., and Holden, T.M., 1999. Determination of the plane specific elastic constants of
Waspaloy using neutron diffraction. Scripta Materialia. 40: p. 353-358.
[221] Hauk, V., 1997. Structural and Residual Stress Analysis by Nondestructive Methods. Aachen, Germany: Elsevier.
[222] Aba-Perea, P.E., Pirling, T., Withers, P.J., Kelleher, J., Kabra, S., and Preuss, M., 2016. Determi-
nation of the high temperature elastic properties and diffraction elastic constants of Ni-base superalloys. Materials
& Design. 89: p. 856-863.
[223] Haines, M., Plotkowski, A., Frederick, C.L., Schwalbach, E.J., and Babu, S.S., 2018. A sensitivity
analysis of the columnar-to-equiaxed transition for Ni-based superalloys in electron beam additive manufacturing.
Computational Materials Science. 155: p. 340-349.
[224] Kouraytem, N., Varga, J., Amin-Ahmadi, B., Mirmohammad, H., Chanut, R.A., Spear, A.D.,
and Kingstedt, O.T., 2021. A recrystallization heat-treatment to reduce deformation anisotropy of additively
manufactured Inconel 718. Materials & Design. 198: p. 109228.
[225] Sanviemvongsak, T., Monceau, D., and Macquaire, B., 2018. High temperature oxidation of IN 718
manufactured by laser beam melting and electron beam melting: Effect of surface topography. Corrosion Science.
141: p. 127-145.
[226] Gao, Y., Zhang, D., Cao, M., Chen, R., Feng, Z., Poprawe, R., Schleifenbaum, J.H., and Ziegler,
S., 2019. Effect of δ phase on high temperature mechanical performances of Inconel 718 fabricated with SLM process.
Materials Science and Engineering: A. 767: p. 138327.
[227] Gan, H.Y., Cheng, M., Guo, Z.X., Tai, Q.A., Kozhevnikova, G., Song, H.-W., and Zhang, S.-H.,
2020. Spheroidization Mechanism of Lamellar δ Phase in Inconel 718 during Delta Processing. Journal of Materials
Engineering and Performance. 29(6): p. 3852-3857.
[228] Radhakrishnan, B. and Thompson, R.G., 1989. Solidification of the nickel-base superalloy 718: A phase
diagram approach. Metallurgical Transactions A. 20(12): p. 3.
[229] Thavamani, R., Balusamy, V., Nampoothiri, J., Subramanian, R., and Ravi, K.R., 2018. Mitigation
of hot cracking in Inconel 718 superalloy by ultrasonic vibration during gas tungsten arc welding. Journal of Alloys
and Compounds. 740: p. 870-878.
147
[230] Bassini, E., Marchese, G., and Aversa, A., 2021. Tailoring of the Microstructure of Laser Powder Bed
Fused Inconel 718 Using Solution Annealing and Aging Treatments. Metals. 11(6): p. 921.
[231] Zhang, Y., Yang, L., Lu, W., Wei, D., Meng, T., and Gao, S., 2020. Microstructure and elevated temper-
ature mechanical properties of IN718 alloy fabricated by laser metal deposition. Materials Science and Engineering:
A. 771: p. 138580.
[232] Tian, Y., McAllister, D., Colijn, H., Mills, M., Farson, D., Nordin, M., and Babu, S., 2014. Rational-
ization of Microstructure Heterogeneity in INCONEL 718 Builds Made by the Direct Laser Additive Manufacturing
Process. Metallurgical and Materials Transactions A. 45(10): p. 4470-4483.
[233] Wei, C.-N., Bor, H.-Y., and Chang, L., 2010. The effects of carbon content on the microstructure and
elevated temperature tensile strength of a nickel-base superalloy. Materials Science and Engineering: A. 527(16): p.
3741-3747.
[234] Manikandan, S.G.K., Sivakumar, D., Prasad Rao, K., and Kamaraj, M., 2015. Laves phase in alloy
718 fusion zone – microscopic and calorimetric studies. Materials Characterization. 100: p. 192-206.
[235] Raza, T., Hurtig, K., Asala, G., Andersson, J.O., Svensson, L.-E., and Ojo, O.A., 2019. Influence of
Heat Treatments on Heat Affected Zone Cracking of Gas Tungsten Arc Welded Additive Manufactured Alloy 718.
Metals. 9(8): p. 881.
[236] Chen, J.H., Rogers, P.M., and Little, J.A., 1997. Oxidation behavior of several chromia-forming commercial
nickel-base superalloys. Oxidation of Metals. 47(5): p. 381-410.
[237] Luo, Y.-W., Zhang, B., Song, Z.-M., Li, C.-P., Chen, G.-F., and Zhang, G.-P., 2020. A comparative
investigation of long-term oxidation behavior of selective laser melting–fabricated Inconel 718 at 650 °C. Journal of
Materials Research. 35(15): p. 2036-2045.
[238] Markanday, J.F.S., Carpenter, M.A., Jones, N.G., Thompson, R.P., Rhodes, S.E., Heason, C.P.,
and Stone, H.J., 2021. Occurrence of a brass texture and elastic anisotropy in laser blown powder processed
superalloy IN718. Materials Science and Engineering: A. 825: p. 141781.
[239] Bi, G. and Gasser, A., 2011. Restoration of Nickel-Base Turbine Blade Knife-Edges with Controlled Laser
Aided Additive Manufacturing. Physics Procedia. 12: p. 402-409.
[240] Kumar, L.J. and Nair, C.G.K., 2017. Laser metal deposition repair applications for Inconel 718 alloy. Ma-
terials Today: Proceedings. 4(10): p. 11068-11077.
[241] Zhou N., D.A.D., Forsik S.A.J., Wang T., Colombo G.A., Epler M.E., 2020. Development of a New
Alumina-Forming Crack-Resistant High-γ′ Fraction Ni-Base Superalloy for Additive Manufacturing, in Superalloys
2020, Tin S. et. al., Editor., Springer, Cham.
148
[242] Yan, H.Y., Vorontsov, V.A., and Dye, D., 2014. Effect of alloying on the oxidation behaviour of Co–Al–W
superalloys. Corrosion Science. 83: p. 382-395.
[243] McLachlan, G.J. and Krishnan, T., 2007. The EM Algorithm and Extensions. Wiley.
[244] Faizabadi, M.J., Khalaj, G., Pouraliakbar, H., and Jandaghi, M.R., 2014. Predictions of toughness and
hardness by using chemical composition and tensile properties in microalloyed line pipe steels. Neural Computing
and Applications. 25(7): p. 1993-1999.
[245] Khalaj, G., 2013. Artificial neural network to predict the effects of coating parameters on layer thickness of
chromium carbonitride coating on pre-nitrided steels. Neural Computing and Applications. 23(3): p. 779-786.
[246] apadopoulos, G., Edwards, P.J., and Murray, A.F., 2001. Confidence estimation methods for neural net-
works: A practical comparison. IEEE Transactions on Neural Networks. 12(6): p. 1278-1287.
[247] Wasserman, L., 2004. All of Statistics - A Concise Course in Statistical Inference. Spring Text in Statistics.
New York, NY: Springer, New York.
[248] Backman, D.G., Wei, D.l.Y., Whitis, D.D., Buczek, M.B., Finnigan, P.M., and Gao, D., 2006. ICME
at GE: Accelerating the insertion of new materials and processes. JOM. 58(11): p. 36-41.
[249] Toda-Caraballo, I., Galindo-Nava, E.I., and Rivera-Dı´az-del-Castillo, P.E.J., 2013. Unravelling the
materials genome: Symmetry relationships in alloy properties. Journal of Alloys and Compounds. 566: p. 217-228.
[250] Greeley, J., Jaramillo, T.F., Bonde, J., Chorkendorff, I., and Nørskov, J.K., 2006. Computational
high-throughput screening of electrocatalytic materials for hydrogen evolution. Nature Materials. 5(11): p. 909-913.
[251] Lejaeghere, K., Cottenier, S., and Van-Speybroeck, V., 2013. Ranking the Stars: A Refined Pareto Ap-
proach to Computational Materials Design. Physical Review Letters. 111(7): p. 075501.
[252] Stucke, D.P. and Crespi, V.H., 2003. Predictions of New Crystalline States for Assemblies of Nanoparticles:
Perovskite Analogues and 3-D Arrays of Self-Assembled Nanowires. Nano Letters. 3(9): p. 1183-1186.
[253] Ingber, L. and Rosen, B., 1992. Genetic Algorithms and Very Fast Simulated Reannealing: A comparison.
Mathematical and Computer Modelling. 16(11): p. 87-100.
[254] Mahfoud, S.W. and Goldberg, D.E., 1995. Parallel recombinative simulated annealing: A genetic algorithm.
Parallel Computing. 21(1): p. 1-28.
[255] Ashby, M.F., 1992. Materials Selection in Mechanical Design. 1st ed. Oxford: Butterworth-Heinemann.
[256] Gong, X., Wang, X., Cole, V., Jones, Z., Cooper, K., and Chou, K., 2015. Characterization of Mi-
crostructure and Mechanical Property of Inconel 718 From Selective Laser Melting. in ASME 2015 International
Manufacturing Science and Engineering Conference.
149
[257] McCartney, D.G., 1989. Grain refining of aluminium and its alloys using inoculants. International Materials
Reviews. 34(1): p. 247-260.
[258] Nanni, P., Gesmundo, F., and McLean, M., 1983. Isothermal oxidation of the COTAC 74 in-situ composite
between 800 and 1300°C.
[259] Khan, T., Stohr, J.F., and Bibring, H., 1980. Cotac 744 : An optimized D.S. Composite for Turbine Blades.
Superalloys. p. 531-540.
[260] Dannemann, K., Stoloff, N.S., and Duquette, D.J., 1987. High temperature fatigue of three nickel-base
eutectic composites. Materials Science and Engineering. 95: p. 63-71.
150
151