Proceedings of the ASME 2018

Pressure Vessels and Piping Conference

PVP2018
July 15-20, 2018, Prague, Czech Republic

PVP2018-84090

EFFECT OF CRACK ORIENTATION ON FRACTURE BEHAVIOUR OF WIRE + ARC

ADDITIVELY MANUFACTURED (WAAM) NICKEL-BASE SUPERALLOY

Cui Er Seow Harry Coules Raja Khan

University of Bristol, University of Bristol, TWI Ltd,
National Structural Integrity Bristol, United Kingdom Cambridge, United Kingdom
Research Centre, harry.coules@bristol.ac.uk raja.khan@twi.co.uk
Bristol, United Kingdom
cuier.seow@bristol.ac.uk

ABSTRACT fundamentally different from traditional formative and

In this study, the effect of crack orientation on the fracture subtractive manufacturing methodologies. The former refers to
behaviour of two compact tension C(T) specimens extracted processes which shape raw material using applied pressure and
from a Wire + Arc Additively Manufactured (WAAM) wall made heat, such as forging, bending, casting, injection moulding etc.
from Inconel (IN) 625 nickel-base superalloy was investigated. The latter refers to processes which selectively remove material
Both specimens had different levels of ductile tearing but their using mechanical methods or heat, such as machining, milling,
load vs. crack mouth opening displacement (CMOD) behaviour drilling etc. There are many types of AM processes and several
was relatively similar. The total-and-elastic strain distribution ways in which they can be classified. The classification approach
around a crack tip was measured in both specimens using Digital recommended in the standard ISO/ASTM 52900-17 includes the
Image Correlation (DIC) and neutron diffraction respectively. following categories: material extrusion, material jetting, binder
The results show that the strain distribution and deformation jetting, sheet lamination, vat polymerisation, powder bed fusion
around the crack tip are different in the two directions. In the (PBF) and directed energy deposition (DED). The latter two are
specimen with crack orientation parallel to the build direction, the categories most relevant to metallic AM.
banding was observed in both the total strain maps and the WAAM is a technique which uses a wire feedstock and an
deformation pattern. Neutron diffraction measurements on this electric arc to build a metallic part in successive layers. It is a
specimen also showed non-monotonic elastic strain evolution, DED near-net shape process, which reduces machining
suggesting the occurrence of intergranular load shedding operations and material wastage as compared to the traditional
mechanisms. These were not observed in the specimen with manufacturing route of casting-forging and machining. From a
crack orientation perpendicular to the build direction. Electron practicality and cost viewpoint, electric arc welding torches are
Backscatter Diffraction (EBSD) maps show that the WAAM more frequently available than laser or electron beam equipment,
IN625 material is strongly textured with coarse columnar grains and many manufacturing facilities already have welding
elongated in the build direction. The effect of microstructure has equipment which can be adapted for WAAM. In comparison to
been correlated with the differences in strain distribution in the powder based processes, WAAM is capable of high deposition
two specimens. rates up to 10kg/h [2], which enables large components to be
built within reasonable time. Drawing from a closely related
INTRODUCTION field of welding, materials which are weldable are more likely to
In recent years, there has been a significant interest and be suited to the WAAM process. The main drive is in the cost
growth in additive manufacturing (AM) for its potential to effectiveness of the process, in which case expensive and
improve part geometry, reduce time to market and make it cost difficult to forge and machine materials such as titanium and
effective for low volume production. AM is defined by nickel-base alloys become attractive options.
ISO/ASTM standard 52900-17 [1] as a process of building parts There are several challenges associated with but not unique
from 3D model data, such as computer aided design (CAD) to WAAM, for example: lack of fusion, porosity, distortions and
drawings, by joining materials in a layer-wise fashion. This is the accumulation of residual stresses during component

1 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 manufacturing [3]. In particular, a WAAM part undergoes lines aligned with the part axis, and the latter refers to a path that
complex thermal cycling during deposition, which results in oscillates about the part axis. An illustration of these two tool
epitaxial grain growth, leading to anisotropy and inhomogeneity path strategies is shown in Figure 1. More complex tool paths
in the microstructure. Some researchers have found that cold optimised for the part geometry may also be used.
rolling after every layer or every few layers improves the The material used in this study was deposited onto a thick
microstructure through grain refinement [4]. Further mild steel substrate plate in an oscillating tool path, using IN625
investigations also found that this reduces residual stresses and welding wire of 1.0mm diameter on 15kg 12” spools and a
hence distortions in WAAM components [5]. However, with Fronius CMT advanced 4000R GMAW system with a fully
increasing complexity in part geometry, it may not always be integrated robotic arm. A slice with 47mm x 64mm x 30mm
feasible to roll the layers and ensuring consistency in the rolling dimensions was taken off from the large deposited material and
process becomes a challenge. Rolling also increases total part used during this study (Figure 2).
build time, which reduces the cost effectiveness of the process.
Therefore, it is important to understand behaviour and the
properties of WAAM material and their effect on the structural
integrity of the final part in the as-built condition.
It has been reported in several studies [6], [7] that the
general microstructure of WAAM parts in the as-built condition
is epitaxial in the direction of the thermal gradient, which is
largely parallel to the build direction. Other studies reported that
WAAM tensile [6] and fracture [8] properties show slight
anisotropy (i.e. 3-10 %), which the authors attributed to
anisotropy in microstructure. However, the effect of anisotropy
in microstructure on mechanical properties have not yet been
thoroughly investigated. Figure 1. General types of tool paths used in
WAAM
Hence, this study focusses on determining:
1. Microstructure of WAAM IN625 material.
2. Differences in the fracture behaviour in two
orientations with respect to the build direction.
3. The possible microstructural causes for these
differences with respect to part build orientations.

MATERIAL AND MICROSTRUCTURE

Material Selection
Nickel-base superalloy IN625 has high tensile strength,
good resistance to corrosion, creep and stress rupture at elevated
temperatures [9]. It is widely used in many industrial sectors
including the marine and petroleum industries. IN625 is one of
the most weldable nickel-base alloys due to its low titanium and
aluminium content, making it a suitable material for WAAM.
Figure 2. Photographic image of IN625 WAAM wall
Build Equipment and Strategy
section
There are several types of welding technologies which have
been used with WAAM e.g. gas metal arc welding (GMAW)
processes such as metal inert gas (MIG), cold metal transfer Microstructure Characterisation
(CMT), tungsten inert gas (TIG) and plasma arc [2]. MIG A small piece of IN625 material extracted from the top of
processes benefit from having the wire as the consumable the slice was used for microstructure characterisation. The plane
electrode, which simplifies the tool path. TIG and plasma arc normal to the wall axis was analysed. The piece was cold
processes have an external wire feeder, which introduces wire mounted in ClaroCit acrylic resin, then ground using a series of
feeding direction as an additional complication to the WAAM grit papers starting from 160 to 2500 grit sizes. The ground
process. Of all heat sources, it has been demonstrated that plasma samples were then polished in three steps using 3, 1 and 0.25 µm
wire deposition gives the highest deposition rates with diamond pastes respectively. Because the substrate plate and the
reasonable quality within a defined parameter range [10]. There WAAM wall are of different metals, a two-part etching process
are generally two types of deposition tool path strategies i.e., was used. First, the mild steel substrate was swabbed with a
parallel and oscillating. The former refers to a several parallel solution containing 2% Nital for 5 seconds. Then, the IN625

2 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 WAAM deposited material was etched electrolytically with a microscope. Figure 4 shows two EBSD maps taken in two
solution containing 20% sulphuric acid at 3V for 10 seconds. different planes, i.e. parallel and perpendicular to the build
The microstructure of WAAM IN625 material was found to direction. In both maps, the 001 direction has been aligned with
be dendritic, with secondary phase particles in the inter-dendritic the build direction. High (>15°) and low (>1°) angle grain
region. Optical micrographs shown in Figure 3 suggest that boundaries are shown by thick and thin lines respectively. These
dendritic arms generally grow in the build direction, which is the maps show that the material is strongly textured with columnar
direction of the overall thermal gradient across the build. Some grains elongated in the build direction. Perpendicular to the build
dendrites are at certain inclined angles to the build direction, direction, the grains are relatively less coarse but fairly equiaxed.
possibly due to the more complex thermal field within each build
layer. Similar trends have been reported by Wang et al. [11] who
found that the dendrite growth direction in a single pass IN625
wall is at a 70° incline to the substrate. They attributed this to
partial cooling of the molten pool by previously deposited
material in the same pass, i.e. solidified material on the trailing
edge of the weld pool. This effect, when coupled with the general
heat flow direction i.e. downwards towards the substrate, causes
the overall thermal gradient to shift by a few degrees away from
the general build direction.

A.

Figure 4. EBSD maps of WAAM IN625

FRACTURE EXPERIMENT METHODOLOGY

Specimen geometry and equipment

Two C(T) specimens with dimensions detailed in Figure 5
were extracted from the WAAM IN625 wall section shown in
Figure 2. The specimens have different notch orientations, one
parallel (CT1) and the other perpendicular (CT3) to the build
direction. They were manufactured and fatigue pre-cracked
according to ASTM E1820-17 [12] with a/W of 0.5 as shown
B. schematically in Figure 5.
These specimens were loaded uniaxially using an Instron 50
Figure 3. Optical micrographs of IN625 WAAM kN capacity servo-hydraulic test machine. The CMOD was
material at (a) 100x and (b) 500x measured using a clip gauge with 10 mm gauge length and 4 mm
travel. DIC strain measurements and neutron diffraction
Specimens prepared for EBSD analysis were mounted using measurements were taken from both specimens at the neutron
a 1:1 mixture of ClaroCit powder and conductive filler (i.e. facility Institut Laue-Langevin (ILL), Grenoble, on the
nickel powder). These specimens were ground and polished in instrument SALSA.
the method described earlier. Final polishing was performed
using 0.04 µm colloidal silica suspension.
EBSD maps of the WAAM IN625 specimens were taken
using a Zeiss Sigma field emission gun (FEG) scanning electron

3 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 two primary collimators, a secondary collimator and an area
detector [13]. The stress rig and DIC cameras were secured to
the hexapod positioner. This enabled the sample to be moved
relative to the collimators while under load. The DIC cameras
remained in the same position relative to the specimen
throughout the entire loading regime. This setup is illustrated in
Figure 6.
All measurements were made with an incoming neutron
wavelength of 1.61 Å and a gauge volume of 2x2x2 mm3. As the
WAAM IN625 material was strongly textured, only the {200}
reflection produced observable Bragg peaks in the strain
direction of interest and hence was the reflection used for all
measurements.

Figure 5. Schematic of the C(T) specimen

geometry, showing the neutron measurement
grid and high resolution line scan

Loading Regime
The two C(T) specimens were loaded in displacement
control to a CMOD of 4mm, in several loading steps at a rate of
0.02mm/min. After each loading step, the CMOD was held,
during which a neutron diffraction grid scan was performed. A
table of loading steps and the corresponding CMOD and neutron
grid scan number for each specimen are detailed in Table 1.
Table 1. Loading steps, corresponding CMOD and
neutron grid scan numbers for each specimen

Loading CT1 CMOD CT3 CMOD Neutron

Step (mm) (mm) grid scan
Figure 6. Schematic representation of
0 0 0 1 experimental setup on SALSA
1 0.2 0.2 2
Neutron diffraction measurements were made using a grid
2 0.4 0.6 3
scan, which is a series of measurements made one at a time in an
3 0.6 1.0 4 8-by-10 point grid with points spaced 2.5 mm apart as shown in
4 0.8 1.4 5 Figure 5, after each loading increment. A grid scan was made
5 1.0 1.8 6 before the first loading step (i.e. at step 0 with 0 mm CMOD as
6 1.2 2.4 7 shown in Table 1) and used as a reference measurement. As the
7 1.4 3.6 8 grains in the material are large relative to the gauge volume,
8 1.8 4.0 9 instead of treating the material as a powder, the specimen was
positioned such that each measurement point was wholly within
9 3.6 - 10
a single crystallite. This was done to mitigate against pseudo-
10 4.0 - - strain effects caused by crystallites moving in and out of the
gauge volume. This was done firstly by conducting a high-
resolution line scan at each point on the grid in both directions,
DIC Measurements as shown schematically in Figure 5. The line scans showing
A Dantec Systems Q400 System was used to conduct DIC significant peak shift indicated the measurement points on the
measurements on the C(T) specimens. Two cameras were used grid susceptible to pseudo-strain effects. The corresponding grid
to track the positions of the speckles and captured using Istra 4D points were removed from the measurement grid. Secondly, the
software. Both rigid body motion and strain were measured movement of the specimen during each loading step was
during each loading step. The measurements of strain made accounted for using the DIC measurements. It was observed that
correspond to the total strain of the specimen. in CT1, there were 21 points of the original grid which were
suitable for measurements. In CT3, all 80 points were observed
Neutron Diffraction Measurements to be within single crystallites.
The instrument SALSA at ILL is a monochromatic neutron
powder diffractometer, which comprises a hexapod positioner,

4 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Neutron Data Treatment deformation in the two directions are different. In CT1, where
Peak fitting was done using a Large Array Manipulation the crack orientation is parallel to the build direction (and hence
Program (LAMP). Peaks were fit to a Gaussian distribution and the dendritic grain growth direction), the banding observed
a flat background. The peak positions were used to obtain the around the crack tip suggest some slip plane activation. This is
lattice spacing, using Bragg’s law, at each point for every loading not observed in CT3, but some deformation parallel to the build
step. Bragg’s law is given by direction can still be seen. From these macrographs (Figure 8), it
𝜆 = 2𝑑 𝑠𝑖𝑛𝜃 (1) can be observed that the material has deformed in a ductile
manner.
where 𝜆 and 𝑑 are the incoming neutron wavelength and lattice
spacing respectively and 𝜃 is the diffraction angle. The lattice
spacing from the first grid scan for each CT specimen, 𝑑1 , was CT1 Build Direction CT3
used as the reference value for calculating the applied elastic

Build Direction
strain, 𝜖𝑎𝑝𝑝𝑙𝑖𝑒𝑑 . This is described in the following equation

𝑑 − 𝑑1
𝜖𝑎𝑝𝑝𝑙𝑖𝑒𝑑 = (2)
𝑑1
This means that the values calculated correspond to the
combined effect of residual and applied elastic strain in the
specimen. Figure 8. Crack tip photographic images of both
C(T) specimens after CMOD of 4mm, in the
EXPERIMENT RESULTS AND DISCUSSION unloaded state

Load-CMOD Curves Crack Driving Force Estimation

The Load-CMOD curves in Figure 7 show that both The crack driving force was evaluated at several load
specimens display elastic plastic behaviour and are relatively intervals using the equations of toughness estimation in terms of
ductile. The load paths of both specimens are relatively similar. J-integral in ASTM E1820-17 [12] and these are shown in Figure
The small load drops in the graphs correspond to holds in the 9. These curves show that up to around 8 kN, the elastic
loading regime when the neutron scans were taking place. component of J dominated the behaviour of the specimen. Above
During the hold, each specimen was held at a fixed CMOD, but 8 kN, the plastic component of J dominated. However, the
the load dropped slightly, possibly due to slight relaxation of the specimen was loaded beyond the validity limit defined in ASTM
piston or grips. The drops are unlikely to be caused by any time E1820-17 which is shown by the dashed line in Figure 9. One
dependent behaviour, as the tests were done at room temperature. important assumption used in this evaluation is that the
specimens underwent negligible crack extension and that plastic
collapse was the dominant failure mode.

Figure 7. Load-CMOD curves of CT1 & CT3

Crack tip photo images from the two C(T) specimens as
shown in Figure 8 suggest that the micromechanics of Figure 9: J-load curves for CT1 & CT3

5 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 The tests were ended when the clip gauge reached its travel
limit of 4mm, during which fracture did not occur and the
maximum load was not reached. Therefore, a discrete value of
fracture toughness could not be determined. The maximum value
of J that could be evaluated from the data for CT1 and CT3 are
690.7 N/mm and 692.3 N/mm respectively. The validity limit is
412.5 N/mm. Nevertheless, the results show that WAAM C(T)
specimens have fairly good resistance to fracture.

DIC Results
The total strain field, measured using DIC, in both CT
specimens loaded to a CMOD of 4mm are shown in Figure 10.
CT1 shows the same banding effect as seen in the crack tip
photos, where intense strips of tensile strain occur around the
crack tip. This is not observed in CT3, where the tensile strain
around the crack tip seems to be more evenly spread out. Across
both specimens from the crack tip to the back of the CT
specimen, the transition from tensile to compressive strain is
slightly different. In CT1, this transition is sharper than that of
CT3, which suggests that the mechanisms of load shedding in
the two orientations are different. This could be due to the
direction of grain growth and elongation in the material, which
is parallel to the build direction. Figure 11. Illustration of possible deformation
modes of the grains in different orientations
under loading condition

Neutron Diffraction Results

The applied elastic strain field in CT1 and CT3 are shown
in Figure 12. The black double ended arrow, which represents
the EDM notch and fatigue pre-crack, points in the direction of
crack propagation. Measurement points are represented by the
crosses.
As expected, a region of tension can be observed ahead of
the crack tip, and a region of compression can be observed on
the back face of the CT specimen. From the plot at CMOD = 0.2
mm in Figure 12(a), it can be observed that near the centre of the
Figure 10. DIC strain maps of both C(T) specimens plot (i.e. points reading approx. 1000 με) are in tension at a low
showing range of ±6% strain in the crack load. Observing the same points with increasing CMOD shows
transverse direction that the strain at these points first increase then decrease, some
In CT1, the grains are perpendicular to the loading direction. of which even reach compressive values at CMOD = 3.6 mm.
Under load, the grains pivot about a point causing them to tilt This trend is shown in more detail on graphs of strain vs.
and compress against each other. In this instance, it is likely that CMOD for each measurement point in Figure 13. These graphs
intergranular interactions dominate the behaviour of the material. show the strain evolution from measurement points of the same
In CT3, where the grains are elongated parallel to the loading approximate position in both specimens. For CT1, it can be seen
direction, grains closer to the force are in tension, and the grains from Figure 13(a) that for a few measurement points, the
further away are less affected by the force. Further along the decrease in strain begins from about CMOD = 1 mm. One
specimen the grains are pushed into compression due to the explanation for this trend is that those points are all within a
deformation of the surrounding material. In this case, intra- single crystallite, and the deformation of this crystallite has been
granular mechanisms dominate. An illustration of this is shown inhibited by the deformation of the surrounding crystallites. In
in Figure 11. contrast, this trend is not seen in the plots for CT3. Most points
show monotonically increasing or decreasing strain values,
which is also seen in the graphs of strain vs. CMOD in Figure
13(b).

6 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 A.

BUILD DIRECTION 

B.
BUILD DIRECTION 

Figure 12. Applied elastic strain field in one half of (a) CT1 and (b) CT3 at each CMOD increment.

7 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 A. B.

Figure 13. Evolution of applied strain with increasing CMOD, from (a) CT1 and (b) CT3.

CONCLUSIONS ACKNOWLEDGMENTS
The following conclusions can be drawn based on the results The authors would like to thank Dr Guiyi Wu from TWI Ltd
obtained during this work: for his supervision and contributions to the work presented in
(1) EBSD maps of the material show that the grains are coarse this paper, Mr Adrian Addison from TWI Ltd for providing
and strongly oriented, parallel to the build direction. WAAM IN625 material and to Dr Thilo Pirling of the ILL for
(2) The load-displacement curves of IN625 material deposited assistance with the experiments. This project is jointly funded by
using WAAM indicate that the elastic-plastic behaviour of the Lloyd’s Register Foundation1, the University of Bristol and
the material with crack orientation parallel and TWI Ltd. Access to neutron diffraction facilities was provided
perpendicular to the build direction are relatively similar. by the Institut Laue-Langevin, Grenoble, France under allocation
no. INTER-364.
This may simplify the task of engineering structural
integrity assessment of components made using this 1
A charitable foundation, helping to protect life and property by
material by allowing the assumption of material isotropy. supporting engineering-related education, public engagement
(3) The WAAM C(T) specimens have fairly good resistance to and the application of research. www.lrfoundation.org.uk
fracture. The maximum value of J that could be evaluated
from the data for CT1 and CT3 are 690.7 N/mm and 692.3 REFERENCES
N/mm respectively. [1] ASTM ISO/ASTM 52910-17, “Standard Guidelines for
(4) Total-and-elastic strain measurements around a sharp crack Design for Additive Manufacturing.” ASTM
of this material have shown that there are some differences International, West Conshohocken, PA, 2017.
in micro-scale deformation depending on orientation with [2] S. W. Williams, F. Martina, A. C. Addison, J. Ding, G.
respect to the build direction. Pardal, and P. Colegrove, “Wire + arc additive
manufacturing,” Mater. Sci. Technol., vol. 32, no. 7, pp.
(5) This work has also demonstrated that it is possible to
641–647, 2016.
conduct elastic strain measurements using neutron [3] ASTM F3187-16, “Standard Guide for Directed Energy
diffraction on this coarse grained and strongly textured Deposition of Metals.” ASTM International, West
material via a crystallite tracking method. Conshohocken, PA, 2016.
(6) These results provide a basis for further investigation into [4] F. Martina, P. A. Colegrove, S. W. Williams, and J.
the effects of material anisotropy on the micromechanics of Meyer, “Microstructure of Interpass Rolled Wire + Arc
WAAM samples. Other ways of interpreting the total-and- Additive Manufacturing Ti-6Al-4V Components,”
elastic strain measurements, such as using the two to Metall. Mater. Trans. A Phys. Metall. Mater. Sci., vol. 46,
decouple plastic strain, may shed some light on the no. 12, pp. 6103–6118, Dec. 2015.
deformation mechanisms at the crack tip. [5] P. A. Colegrove et al., “Microstructure and residual
stress improvement in wire and arc additively
manufactured parts through high-pressure rolling,” J.
Mater. Process. Technol., vol. 213, no. 10, pp. 1782–

8 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 1791, 2013. Derivatives, 2005, pp. 409–418.
[6] F. Wang, S. Williams, P. Colegrove, and A. A. [10] F. Martina, J. Mehnen, S. W. Williams, P. Colegrove, and
Antonysamy, “Microstructure and mechanical properties F. Wang, “Investigation of the benefits of plasma
of wire and arc additive manufactured Ti-6Al-4V,” deposition for the additive layer manufacture of Ti-6Al-
Metall. Mater. Trans. A, vol. 44, no. 2, pp. 968–977, 4V,” J. Mater. Process. Technol., vol. 212, no. 6, pp.
2013. 1377–1386, Jun. 2012.
[7] B. Cong, J. Ding, and S. Williams, “Effect of arc mode [11] J. F. Wang, Q. J. Sun, H. Wang, J. P. Liu, and J. C. Feng,
in cold metal transfer process on porosity of additively “Effect of location on microstructure and mechanical
manufactured Al-6.3%Cu alloy,” Int. J. Adv. Manuf. properties of additive layer manufactured Inconel 625
Technol., vol. 76, no. 9–12, pp. 1593–1606, 2015. using gas tungsten arc welding,” 2016.
[8] X. Zhang, F. Martina, J. Ding, X. Wang, and S. W. [12] ASTM E1820-17, “Standard Test Method for
Williams, “Fracture toughness and fatigue crack growth Measurement of Fracture Toughness.” ASTM
rate properties in wire + arc additive manufactured Ti- International, West Conshohocken, PA, 2017.
6Al-4V,” Fatigue Fract. Eng. Mater. Struct., vol. 40, no. [13] D. J. Hughes, G. Bruno, T. Pirling, and P. J. Withers,
5, pp. 790–803, 2017. “First Impressions of SALSA: The New Engineering
[9] L. E. Shoemaker, “Alloys 625 and 725: Trends in Instrument at ILL,” Sci. Rev., vol. 17, no. 3, 2006.
properties and applications,” in 6th International
Symposium on Superalloys 718, 625, 706 and

9 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 05/01/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use