Modelling and Simulation in

Materials Science and Engineering

PAPER • OPEN ACCESS You may also like

- Full-band Monte Carlo analysis of strain
Identification of texture characteristics for effects on carrier transport in GaN
Wataru Miyazaki, Hajime Tanaka and
improved creep behavior of a L-PBF fabricated Nobuya Mori

- Computer simulation of microstructure

IN738 alloy through micromechanical simulations development in powder-bed additive
manufacturing with crystallographic texture
J G Pauza, W A Tayon and A D Rollett
To cite this article: Mahesh R G Prasad et al 2022 Modelling Simul. Mater. Sci. Eng. 30 055007
- Heteroepitaxial growth of diamond films on
3C-SiC/Si substrates with utilization of
antenna-edge microwave plasma CVD for
nucleation
Junya Yaita, Takayuki Iwasaki, Meralys
View the article online for updates and enhancements. Natal et al.

This content was downloaded from IP address 129.45.26.74 on 20/05/2024 at 01:54

 Modelling and Simulation in Materials Science and Engineering
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 (21pp) https://doi.org/10.1088/1361-651X/ac6e7a

Identification of texture characteristics

for improved creep behavior of a L-PBF
fabricated IN738 alloy through
micromechanical simulations
Mahesh R G Prasad1 , Abhishek Biswas1,∗ ,
Napat Vajragupta2 and Alexander Hartmaier1
1
ICAMS, Ruhr Universität Bochum, Universitätsstr. 150, Bochum, Germany
2
VTT, Kivimiehentie 3, Espoo, Finland

E-mail: [email protected], [email protected], [email protected]

and [email protected]

Received 2 November 2021, revised 30 April 2022

Accepted for publication 11 May 2022
Published 30 May 2022

Abstract
Additive manufacturing (AM) of nickel-based superalloys, due to high temper-
ature gradients during the building process, typically promotes epitaxial growth
of columnar grains with strong crystallographic texture in form of a 001 fibre
or a cube texture. Understanding the mutual dependency between AM process
parameters, the resulting microstructure and the effective mechanical proper-
ties of the material is of great importance to accelerate the development of the
manufacturing process. In this work, a multi-scale micromechanical model is
employed to gain deeper insight into the influence of various texture charac-
teristics on the creep behavior of an IN738 superalloy. The creep response is
characterized using a phenomenological crystal plasticity creep model that con-
siders the characteristic γ–γ  microstructure and all active deformation mech-
anisms. The results reveal that the creep strength increases with decreasing
texture intensities and reaches its maximum when the 001 fibre and cube tex-
tures are misaligned to the specimen building direction by 45◦ . The simulations
also predict that the uncommon 111 and 110 fibres offer significantly higher
creep resistance than the typically observed 001 fibre, which provides a fur-
ther incentive to investigate AM processing conditions that can produce these
unique textures in the material. As the intensities and the alignment of 001

∗
Author to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons
Attribution 4.0 licence. Any further distribution of this work must maintain attribution
to the author(s) and the title of the work, journal citation and DOI.

0965-0393/22/055007+21$33.00 © 2022 The Author(s). Published by IOP Publishing Ltd Printed in the UK 1
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

fibre and cube textures can be attributed to the laser energy density and the scan
strategy employed and as the formation of distinct fibre textures depends on the
geometry of the resulting melt pool, the laser powder bed fusion process param-
eters can be optimized to obtain microstructures with features that improve the
creep properties.

Keywords: additive manufacturing, superalloys, crystallographic texture, crys-

tal plasticity, anisotropic creep

(Some figures may appear in colour only in the online journal)

1. Introduction

Unlike the top-down approach in traditional manufacturing processes, additive manufacturing

(AM) employs a bottom-up approach in which the geometry of the component is manufac-
tured layer-by-layer. Due to this building approach, AM can manufacture components with
highly complex geometries. Initially, AM methods were developed for polymers, but the ver-
satility and the flexibility of these methods have led to the development of AM methods also for
metallic components [1, 2]. Depending on the feedstock delivery method, AM processes for
metallic components can be classified into the powder bed fusion (PBF), and the metal depo-
sition methods. The PBF method is more commonly used, wherein a metallic powder layer
is first spread on the build platform and then melted by an energy source following the input
provided by the computer-aided design model. Based on the energy source used, the PBF meth-
ods can be further classified into laser powder bed fusion (L-PBF) and electron beam powder
bed fusion, which use laser or electron beams, respectively. Although the PBF techniques are
being increasingly used for the production of functional parts [3], significant metallurgical dif-
ferences exist between AM-fabricated and conventionally manufactured components in terms
of microstructure, mechanical anisotropy, residual stresses and defects.
L-PBF (also known as selective laser melting) has gained prevalence in aviation and the
energy industries for manufacturing gas-turbine engine components [4]. The materials used
for such cost-intensive components are γ  -precipitation-strengthened nickel-based superalloys
which are conventionally processed by casting [5]. Nickel-based superalloys are a class of
materials with an excellent high-temperature corrosion and oxidation resistance and an out-
standing mechanical strength due to their two-phase microstructure consisting of γ  precipitates
dispersed in the γ matrix phase [6]. A widely used nickel-based superalloy for laser-fabrication
is Inconel-718 [7], due to its low weld cracking susceptibility and good mechanical proper-
ties [8]. Other superalloys studied in the context of L-PBF include IN738LC [9], CM247LC
[10, 11], CMSX486 [12], IN625 [13], Waspaloy [14], and Hastealloy [15], although many
of these γ  -strengthened superalloys are susceptible to the formation of defects such as
micro-cracks and pores.
During L-PBF processing, the selective melting done by the fast-moving laser beam results
in a fast heat-up and rapid solidification of the melt pool. The highly directional heat flow
through the build produces epitaxial growth of columnar grains aligned with the temperature
gradient. As the dendrites grow parallel to the heat flow direction, i.e. perpendicular to the
solidification front, the texture that is produced depends on the melt pool shape [16, 17]. The
elongated grains show a preferred [001] alignment, as the 100 crystal orientations are the easy
growth directions for face-centered cubic (fcc) crystals. The complex solidification process
does not only result in unique microstructures with direction-dependent material properties
2
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

[18] but also induce residual stresses [19, 20]. The formation of such complex microstructures
is influenced by the L-PBF process parameters as well as by the composition of the powder
and the post-heat treatment cycles.
It is well established, that the primary strengthening mechanism in single-crystal nickel-
based superalloys is the precipitation of the ordered γ  phase in the γ matrix. In contrast, due
to their different γ  geometries, the strength of polycrystalline superalloys rather dependents
on the texture, especially when a strong texture co-exists with complex-shaped grains of a par-
ticular orientation. Significant research has been conducted over the last decade to determine
the effect of different AM parameters on the resulting crystallographic texture. Dinda et al [21]
studied the influence of laser parameters on the texture of Inconel 718 and reported that employ-
ing a unidirectional laser scanning pattern results in a fibre texture, whereas a rotated cube
texture is produced by back and forth laser scanning strategy. Other studies on Selective Laser
Melting (SLM) of nickel-based alloys also report a strong 001 fibre texture along the building
direction for CM247LC [10, 22], and IN738LC [9, 18, 23]. However, when subjected to heat
treatment cycles, the materials undergo recrystallization resulting in a reduction of the texture
intensity [23–25].
Most available works in the literature have focused on the influence of AM process parame-
ters on the resulting texture, but the influence of texture on the creep properties in nickel-based
superalloys is less understood. Furthermore, as the root portions of the turbine blade can
undergo significant stresses in directions transverse to the major stress axis, understanding the
effects of anisotropy is particularly important. In this context, the current parametric study is
designed to gain deeper insight into the influence of various texture characteristics on the creep
behavior by examining the γ  -precipitate strengthened nickel-based superalloy IN738LC. The
remainder of the paper is organized as follows. The next section presents a detailed introduc-
tion to the basics of micromechanical modeling in the context of synthetic microstructures. The
capability of the crystal plasticity (CP) model employed to study creep behavior and anisotropy
is validated by comparing it with experimental data. Note that the model itself is detailed in the
appendix A. Section 3 presents the numerical results, where the creep behavior is analyzed for
various applied loads, fibre types and varying intensities, alignments of fibre and cube textures.
Finally, section 4 contains the conclusion of this work with future prospects.

2. Micromechanical modeling

A micromechanical model is developed to accurately capture the deformation behavior in

polycrystalline superalloys by using homogenization techniques. Representative volume ele-
ments (RVEs) with appropriate approximations are used for describing the polycrystalline
microstructure of the L-PBF fabricated superalloy. The RVEs contain complete descriptions
of microstructural features at the mesoscale (grain size, crystallographic texture) and the
microscale (precipitate phase volume fraction, size and lattice misfit). A two-phase superalloy
CP model, originally developed for single-crystal superalloys, is extended to polycrystalline
superalloy to model the creep behavior of the material. In the following sub-sections, different
aspects of the modeling approach such as RVE creation, the homogenization technique used,
material parameters determination and validation are detailed. For the mathematical descrip-
tion of the two-phase superalloy CP model, the reader is kindly referred to the appendix A.
Throughout this work, the abbreviations BD, TD, and ND are used to represent the build-
ing, transverse and normal directions of the L-PBF fabricated specimen, respectively. And the
abbreviation LD is used to represent the direction along which the load is applied. For instance,
LD  BD represents load is applied along the building direction.
3
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 1. (a) Mesoscale RVE used in micromechanical simulations, where each grain is
represented by a single FE mesh element and is color-coded according to the depicted
orientation triangle. Creep simulations are performed at 980 ◦ C and 150 MPa with ten-
sile loads along BD. (b) Comparison of the (100) pole figures between the reference
ODF (left) (obtained with permission from the experimental work [23]) and the RVE
ODF (right). The intensities in the color bar are in multiples of random distribution. The
reference direction used for the color coding is Inverse pole figure (IPF-Z).

2.1. Hierarchical microstructure

The microstructure of AM-fabricated nickel-based superalloys depends significantly on the

AM processing parameters and the subsequent heat treatment cycles they are subjected to.
The resulting microstructure at the mesoscale usually contains elongated grains in the direc-
tion of the build with strong crystallographic textures. And in most superalloys, the matrix
phase (γ) has a fcc crystal structure. In polycrystalline superalloys, the mechanical proper-
ties are not only determined by the grain structure and the crystallographic texture, but also by
secondary phase precipitates (γ  ) which exist at the micro-scale. Hence, to capture the deforma-
tion behavior in superalloys efficiently, the microstructural features on the two scales must be
considered.
The virtual microstructure or RVE used for the micromechanical modeling of the creep
behavior of an IN718 superalloy is shown in figure 1(a). The RVE consists of 512 grains of
which each grain is represented by a single hexahedral finite element of a size equivalent to the
average grain size of 38 μm [23]. The grains in the RVE are colored according to the orientation
triangle depicted in figure 1(a). The color coding helps visualize the orientation of a grain by
showing which crystal direction is parallel to the sample direction. The number of grains in
the RVE is decided by a parametric study, which is performed with RVEs with the number of
grains ranging from 216 to 1000. Creep simulations are performed at 980 ◦ C and 150 MPa
with tensile loads along BD. Since in this work the minimum creep rate is studied, it is used
as the judging criteria. The results of the simulations are listed in table 1, which indicates that
as the number of grains in the RVE increases, the results converge to a minimum creep rate of
2.99 × 10−7 s−1 . Considering this value, the RVE with 512 grains provides the best compromise
between minimum creep rate prediction and computational time.
Polycrystalline metals consist of a non-uniform distribution of crystallographic orientation
distribution [26] which is mathematically represented as the orientation distribution function.
The RVE consists of a smaller number of discrete crystallographic orientations, which pro-
vides computational efficiency as well as it maintains the representativeness of the material
property. The orientations used in the RVE are generated by mimicking the pole figure of the
reference Orientation distribution function (ODF) (figure 1(b) (left)) taken from [23]. The ref-
erence ODF exhibits a crystallographic texture with a preferred 001 orientation in specimen
4
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Table 1. Various RVEs considered for the RVE convergence study. The minimum creep
rate values are calculated from uniaxial tensile creep simulations (980 ◦ C and 150 MPa).
RVE size (μm) Number of grains Minimum creep rate (s−1 )
228 216 3.18 × 10−7
304 512 3.04 × 10−7
342 729 2.99 × 10−7
380 1000 2.98 × 10−7

Figure 2. (Left) SEM image of γ –γ  microstructure obtained after the heat-treatment

process. Reproduced with permission of author from [23]. (Right) Schematic represen-
tation of 1/8th microscale γ –γ  RVE considered at each integration point of an element
in the FE simulation. The volume fractions are: f p = 0.47, f x = f y = f z = 0.177.

BD. The intensities in the color bar are in multiples of the random distribution. Discrete ori-
entations sampled from this ODF using the method described in Biswas et al [27] are used
in micromechanical simulations to simulate the real material response. The pole figure of the
RVE orientation set is shown in figure 1(b) (right).

2.2. Multiscale constitutive model

On the material point level, the characteristic γ –γ  microstructure of a nickel-based super-

alloy, consisting of γ  –precipitates in a γ-matrix is represented by an RVE containing a γ  -
precipitate, and γ-matrix channels along x, y, z directions (see figure 2). This method only
covers the large primary γ  -precipitates and the very small secondary γ  -precipitates within
the channels are treated implicitly as part of the matrix constitutive behavior. In this RVE, vol-
ume fractions of the matrix channels ( f x , f y , f z ) and the precipitate ( f p) satisfy the constraint
f x + f y + f z + f p = 1. Since this RVE is considered as an integration point in the FE method,
the detailed internal stresses and plastic strains are not required. Hence, these quantities are
assumed to be uniform inside each region, which leads to a particularly efficient framework.
To the benefit of computational efficiency, owing to the symmetry of the RVE, only one-eighth
of the RVE (as shown in figure 2) needs to be included in the calculations. For further details
regarding the geometrical and computational aspects please refer to [28].
The constitutive models at the microscale, for the matrix and precipitate phases, are based
on CP theory and are described in detail in the appendix A. To account for the dislocation
5
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

activities in the two phases, 12 octahedral slip systems are considered for both matrix and pre-
cipitate phases (see table 3). The internal stresses due to γ–γ  lattice mismatch and deformation
heterogeneity (as described in [28]), and the Orowan stress required to bow a dislocation line
into the channel between precipitates are implemented in the model. Furthermore, a softening
stress term is introduced to account for the softening effect caused by the climb of dislocations
along the γ –γ  interfaces. However, the material response in the tertiary creep regime, asso-
ciated with fast increasing plastic strains due to many simultaneously occurring deformation
mechanisms (precipitate rafting, shearing and, void nucleation, growth and coalescence), is not
included in the present model.
Based on the γ –γ  microstructure depicted in figure 2, the precipitates are assumed to be
cubic. And the experimentally measured size and volume fraction of the γ  precipitates are
approximately 850 nm and 47%, respectively. Using these values, the matrix channel width
(w) is computed using,
 
1
w=L 1/3
− 1 (1)
ηv
where η v and L are the experimentally measured γ  volume fraction and side length,
respectively.

2.3. Boundary conditions and homogenization

The RVE depicted in figure 1 is subjected to an applied pressure that is held constant for a long
period at high-temperatures, so that the steady-state creep conditions are reached. For realis-
tic deformation behavior of the RVEs, periodic boundary conditions (PBCs) are implemented.
This involves coupling nodes on the opposite surfaces of the RVE to capture deformations and
thus strains. For a detailed description of the implementation of PBC’s, please refer to Boeff
[29] and Smit et al [30]. The stress and strain tensors of the RVE under PBCs are homogenized
by fulfilling the macrohomogeneity condition, which dates back to Hill [31] and Mandel [32].
Nemat-Nasser [33] gives a good overview of averaging theorems in finite deformation plas-
ticity. The RVE is simulated under the uniaxial tensile creep loading conditions using Abaqus
[34] to obtain the creep curves.

2.4. Material parameters and model validation

The model capabilities have been demonstrated by comparing simulated creep curves to exper-
imental results obtained from [23]. Using the multi-scale RVE described earlier, creep sim-
ulations are performed at 980 ◦ C and 150 MPa with tensile loads along BD and TD. The
experimental creep curve corresponding to the 001 fibre texture with loading along BD (see
figure 3) was used as a reference for parameterizing the creep model. The elastic constants of
the alloy (shown in table 2) are temperature-dependent parameters and are obtained from [35].
For nickel-based superalloys, the activation energies (Q110 , Q112 ) required to cause slip in the
different slip systems are obtained from other studies [36, 37]. The Orowan stress arising from
the movement of dislocations through the narrow γ channels is computed using equation (12).
The length of the Burgers vector is taken as 0.254 nm, and the matrix channel width is computed
as the spacing between two precipitates (equation (1)).
The fitted values for the softening stress (s1 ) and the rate of softening (s2 ) are based on
the discrete dislocation dynamic (DDD) simulations reported in the work of Gao et al [38].
The remaining parameters i.e. the reference shear rates (γ̇ 110 112
0 , γ̇ 0 ), the initial slip resistances
110 112 sat
(τ̂ 0 , τ̂ 0 ), the hardening parameters (τ̂ , h0 , p2 ), and the inverse of strain rate sensitivity
(p1 ) are fitted such that good agreement between experimental data and simulation result is
6
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 3. Comparison of creep simulation results and experimental data (source: [23])
for 001 fibre and cube textures loaded along different directions at 980 ◦ C, figure (a)
shows the creep strain vs time (b) shows the creep strain rate vs time. The material
parameters are fitted by using the 001 fibre LD  BD curve, and the creep behaviors
for 001 fibre LD  TD, and cube texture LD  BD are predicted.

Table 2. Material parameters and the corresponding values used for creep simulations.

Parameter Value Unit

c11 200.29 GPa
c12 146.15 GPa
c44 54.14 GPa
γ̇ 110
0 2.0 × 109 s−1
γ̇ 112
0 3.1 × 1015 s−1
Q110 350 kJ mol−1
Q112 500 kJ mol−1
τ̂ 110
0 42 MPa
τ̂ 112
0 225 MPa
p1 3.0 —
p2 0.05 —
τ̂ sat 350 MPa
h0 30 MPa
χαβ coplanar 1.0 —
χαβ non-coplanar 1.4 —
τ oro 46.16 MPa
s1 80 MPa
s2 8 —

obtained for the 001 fibre texture with loading along BD. Once a good agreement with the
experimental data was obtained, the fitted material parameters (summarized in table 2) were
used for the rest of this study. The result demonstrates that the present framework can simulate
the real material response in the primary and secondary creep regimes adequately.
The model does not include deformation mechanisms responsible for fast accumulating
plastic strains in the tertiary creep regime, therefore the tertiary creep prediction is excluded
in this work. To validate the creep model, experimental creep test data from [23] for the 001
7
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

fibre texture specimen loaded along TD is used. From the corresponding simulation result
(figure 3) it is evident that the model can predict the anisotropy in creep behavior very well. Fur-
thermore, by using the obtained material parameters above and by keeping all the microscale
microstructural parameters essentially constant, creep simulations were conducted for an RVE
with cube texture with tensile load along BD. Comparing the experimentally predicted creep
curves to those resulting from simulation, it can be concluded that the model is aptly sensitive
to microstructural changes and can be used for conditions other than those used to identify the
material parameters.
Figure 3(b) shows the strain rate vs strain curves and as explained in Gao et al [38] this
plot can be divided into primary and secondary creep regimes. In the initial primary creep
regime, dislocations with different Burgers vectors are produced, which penetrate the verti-
cal and horizontal γ channels. These dislocation segments accumulate at the γ/γ  interfaces
and the filling of the γ channels with inter-facial dislocation segments renders the motion of
mobile channeling dislocations increasingly difficult and thus increases the creep resistance.
In the secondary creep regime, a stable interfacial dislocation network forms associated with
dislocation climb and diffusion, which leads to a low and constant creep rate. The minimum
creep rate is approximated from this plot by estimating the first point with slope ≈ 0 in the
secondary regime. Comparing the three curves it is evident that the creep model can not only
capture the anisotropic difference between the two loading directions (for the 001 fibre tex-
ture case), but it is also sensitive to changes in crystallographic texture. Thus, in the following
section the influence of various crystallographic texture characteristics on the creep behavior
will be analyzed.

3. Parametric study and numerical results

3.1. Creep anisotropy under iso-thermal loading

The minimum strain rate is often used as a design tool to determine the stress needed to produce
it. In the analysis of creep deformation, the temperature and stress dependence of the minimum
creep rate are used to analyze the plastic deformation processes that control the creep. In this
context, the relationship between the minimum strain rate (), ˙ temperature (T ) and stress (σ)
can be expressed as,
 
n −Q
˙ = Aσ exp (2)
RT
where, A and n are empirical constants. Q is the activation energy for creep, and R is the gas
constant. An Arrhenius term is included in the model to consider the effect of temperature;
however, in the current study iso-thermal creep tests are conducted, hence the exponential
term becomes a constant. The resulting expression is the Norton’s power law ˙ = Aσ n , which
can be linearized (by taking logarithms) to compute A and n.
By performing isothermal creep simulations (at 980 ◦ C) under various applied stresses
(120–180 MPa), a log–log plot depicting the minimum creep rates as a function of applied
stresses is plotted in figure 4. The power-law fitting is done using the non-linear least-squares
fitting method [39] as shown in figure 4. The stress exponent n is determined by power-law
fitting to be 3.2 for BD loading and 3.22 for TD loading; these values are very close to the value
of p1 (see table 2) used in the flow rules (equations (5) and (14)). From figure 4 it is evident
that for the same applied load the minimum creep rate along TD is always higher in magnitude
than that for the BD direction loading. The observed anisotropy in the secondary creep rate, is
consistent with the results reported in the experimental investigation of Fernandez-Zelaia et al
8
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 4. Log–log plot of minimum strain rate as function of applied stress for iso-
thermal creep testing at 980 ◦ C. The power law fits and the corresponding empirical
constants are also shown for the two loading directions BD and TD.

[40]. In this reference the authors have investigated the creep behavior of IN738LC produced by
the electron beam melting process. The authors also use the power-law expression equation (2)
to express the secondary creep rate, but in contrast to the current work both temperature and
applied stress were varied.

3.2. Influence of texture alignment

The influence of the scanning pattern on the resulting crystallographic texture was studied in
[21] for laser metal deposition with IN625 and in [41] for L-PBF with IN718. It was reported
in these studies that, although the laser processing parameters were kept constant, the change
in the scanning pattern altered the dendrite growth from 60◦ in unidirectional cases to 45◦
in bidirectional cases with respect to the horizontal line. As dendrites grow opposite to the
heat flow direction i.e. perpendicular to the solidification front, for the unidirectional scanning
pattern, the growing dendrites form a columnar structure with one of their axis aligned in the
same direction. Depending on this alignment, a rotated fibre texture develops, as only one of
the 100 directions of dendrites is aligned parallel to the heat flow direction and the other two
100 directions remain free. For the bidirectional scanning pattern, owing to the 180◦ change
of the scanning direction between successive layers, the secondary dendrites of the previously
deposited layer become the primary dendrites of the current layer and grow parallel to the
heat flow direction. This results in a rotated cube texture with all three 100 directions of the
dendrites being fixed and the growth direction aligned corresponding to the heat flow direction.
To study the influence of texture alignment on the creep behavior, ten test cases each for
fibre and cube textures are considered in this study. The test cases are generated synthetically
using MTEX [42] by rotating the ODFs obtained from [23]. Figure 5 depicts five exemplary
(100) pole figures of the fibre texture rotated about the ND. And figure 6 depicts five exemplary
9
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 5. (100) pole figures of the ODF (from figure 1(b)) rotated about ND by (a) 10◦ ,
(b) 30◦ , (c) 50◦ , (d) 70◦ , and (e) 90◦ used to study the influence of fibre texture align-
ment in micromechanical simulations. The intensities in the color bar are in multiples of
random distribution.

Figure 6. (100) pole figures of the cube texture ODF rotated about ND by (a) 10◦ ,
(b) 30◦ , (c) 50◦ , (d) 70◦ , and (e) 90◦ used to study the influence of cube texture align-
ment in micromechanical simulations. The intensities in the color bar are in multiples of
random distribution.

(100) pole figures of the cube texture rotated about the ND. Discrete orientations are sampled
from these individual ODFs using the method in Biswas et al [27], and creep simulations
are performed at 980 ◦ C and 150 MPa with tensile loads along BD and TD by keeping all
microscale microstructural features and fitting parameters unchanged. To study the resulting
creep anisotropy, the minimum creep rate values for the loading directions are computed and
plotted as a function of the rotation angle for fibre and cube textures in figures 7(a) and (b),
respectively.
Comparing the two sub-figures it is evident that, with cube texture, the material shows
isotropic behavior when loaded along BD and TD, irrespective of the rotation angle about
ND. This behavior is justified as the material with cube texture resembles a near single crys-
talline structure, and the deformation behavior along crystallographically equivalent directions
is identical. In contrast, creep anisotropy exists for the fibre texture, and the magnitude varies
with the rotation angle defined about ND. For low and high values of rotation angles, creep
anisotropy exists, and the BD has a higher creep strength. And for rotation angles between 20◦
and 50◦ , the magnitude of creep anisotropy is considerably small, and in fact, the TD shows
higher creep strength for the most part within these limits. Comparing figures 5 and 6, it is
obvious that the intensity of the cube texture is substantially higher than that of the consid-
ered fibre texture. And this influence is seen on the magnitude of the resulting minimum creep
rate values. To quantitatively analyze this influence, fibre and cube textures of varying texture
intensities are studied in detail in the following sub-section. Independent of the texture inten-
sity, it can be concluded that a rotated fibre or cube texture offers higher creep strength to the
material.
10
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 7. Plot of minimum creep rate as function of ODF rotation angle about ND for
(a) fibre texture, (b) cube texture. Simulations are performed under creep conditions
980 ◦ C and 150 MPa with tensile loading along both BD and TD.

3.3. Influence of texture intensity

The influence of laser energy density on the microstructure and the mechanical properties dur-
ing AM has been studied by several researchers. Popovich et al [7] reported that within the
high energy density area, a highly inhomogeneous and strongly textured microstructure with
elongated grains along BD was found. In the work of Liu et al [43], it is reported that a strong
fibre texture was induced in the case of high-energy-density, but this became insignificant as
energy density decreased. Recently, Risse [23] studied the influence of pre-heating temperature
on the strength of the resulting crystallographic texture in an L-PBF processed IN738LC alloy.
A strong increase in crystallographic texture intensity was observed when the pre-heating tem-
perature was increased from 500 ◦ C to 1050 ◦ C. This behavior was attributed to the combined
effect of melt pool dimensions and the temperature at the heat-affected zone, which results in
a smaller deviation of the solidification direction from the build direction, resulting in stronger
crystallographic texture intensities. Furthermore, in the work of Divya et al [24], the as-built
CM247LC alloy samples were subjected to heat treatment cycles, and the texture obtained after
the process was observed to be much weaker. The reduction in texture intensity was attributed
to the effect of re-crystallization, wherein fine dislocation cell structures get converted into
sub-grain boundaries by a recovery process [44]. Similar observations were reported in [10]
for CM247LC, [25] for IN718 and in [9, 23] for IN738LC alloys.
To study the influence of the texture intensity on the creep behavior, fibre, and cube textures
of varying intensities are synthetically generated using MTEX. To characterize the sharpness
of the texture, a single dimensionless number called the texture index is used [45]

Texture index = f (g)2 dg, (3)

where f is the orientation distribution function at an orientation g. Theoretically, a texture index

of one represents a random texture (ideally isotropic), and a value of infinity represents a single
crystal material (strongly anisotropic). The (100) pole figures corresponding to the various
fibre and cube test cases generated in this study are depicted in figures 8 and 9, respectively.
Crystallographic orientations are extracted from these ODFs, and creep simulations are done
with tensile loading along the BD. Figure 10 depicts the minimum creep rate values plotted as
a function of the texture index for both fibre and cube textures of varying texture intensities.
11
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 8. (100) pole figures of the synthetically generated 001 fibre ODFs of texture
index (a) 59.3, (b) 33.8, (c) 15.8, (d) 9.5, and (e) 3.1 used to study the influence of fibre
texture intensity. The intensities in the color bar are in multiples of random distribution.

Figure 9. (100) pole figures of the synthetically generated cube texture ODFs of texture
index (a) 59.8, (b) 31.5, (c) 17.4, (d) 8.5, and (e) 3.0 used to study the influence of cube
texture intensity. The intensities in the color bar are in multiples of random distribution.

It is evident from figure 10 that the curves follow a general trend. The minimum creep rate
increases rapidly at first with increasing texture index, but it eventually saturates as the tex-
ture index reaches high values. The texture index is controlled by the half-width of the kernel
function used for generating the ODFs, as the two parameters have an inverse proportional rela-
tionship [46]. Thus, by decreasing the half-width value, similar crystallographic orientations
are sampled from the ODF into the microstructure, which makes it easier to activate the slip sys-
tems to carry plastic deformation. This results in the observed decrease in the material’s creep
resistance with increasing texture index values. In addition, a slight increase in the anisotropic
difference is observed for the 001 fibre texture case as the texture index increases. This how-
ever is not the case for cube texture as the deformation behavior along crystallographically
equivalent directions is identical. Furthermore, it can also be seen that the minimum creep rate
values are similar for very low texture index values, irrespective of the texture type. But the
difference between the curves amplifies in magnitude with increasing texture index values.
This difference can be attributed to the alignment of the dendrites that form the grains, and
consequently, their deformation behavior when subjected to loads (here LD  BD). In cube
texture all three 100 directions of the dendrites are fixed, whereas, in fibre texture, only one
of the 100 directions of dendrites is fixed and the other two 100 directions remain free. This
alignment difference results in a lower value of the average Schmid factor for the fibre texture
case, which consequently reduces the resolved shear stress value making the fibre texture more
creep resistant.

12
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 10. Plot of minimum creep rate as function of texture index for 001 fibre and
cube textures. Simulations are performed under creep conditions 980 ◦ C and 150 MPa
with tensile loading along BD and TD.

3.4. Influence of distinct fibres

During the AM processing of cubic materials, the most common crystallographic texture
formed is the 001 fibre texture aligned along BD [7, 21]. However, in addition to the 001
fibre texture, some studies have reported the formation of 110 [16, 17], and 111 [47, 48]
fibre textures aligned along BD. The formation of these textures has been attributed to the
influence of AM processing parameters on the geometry of the resulting melt pool and the
temperature at the heat-affected zone, which consequently alters the alignment of the 100
oriented grains through changes in the direction of heat flow. Jadhav et al [16] studied the rela-
tionship between laser scan parameters and crystallographic texture during SLM processing
of pure copper. They reported that a change in the melt pool geometry from deep to elliptical
rotates the average temperature gradient, and consequently the strong 100 texture parallel to
BD is transformed into a weak 110 texture parallel to BD. In the work of Sun et al [17], the
crystallographic texture of SLM-processed SS316L was engineered by adjusting the melt pool
geometry through different laser process parameters. They proposed a multi-scan melt pool
strategy that could produce a stable and deep melt pool, which resulted in the formation of a
110 fibre texture aligned along BD. And they demonstrated that this texture assists the acti-
vation of the nano-twinning mechanism during the tensile deformation, which simultaneously
increases the strength and ductility. In the context of nickel-based superalloys, Geiger et al [9]
have reported the formation of 110 texture aligned along BD during the L-PBF processing
of IN738LC alloy.
Recently, L-PBF processing of refractory metals such as molybdenum, tantalum, and tung-
sten has received increased interest, as these metals show great promise as high-temperature
materials for various applications. Although the formation of 001 fibre texture aligned along
13
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 11. (a) Inverse pole figure of all the fibre textures considered to study the creep
behavior. The synthetically generated fibres are aligned parallel to BD. Pole figures for
001 fibre, 110 fibre, and 111 fibre aligned along BD are plotted in subfigures (b),
(c), and (d), respectively.

BD is the preferred one for the AM processing of cubic materials, studies have shown the for-
mation of 111 fibre texture for these pure refractory metals [47–50]. Higashi and Ozaki [48]
reported the formation of 001, 110, and 111 fibre textures aligned along BD as a function
scan speed and volumetric energy density during the SLM-processing of pure molybdenum.
They demonstrated that transitions between the different fibre textures can be correlated to the
changes in the melt pool morphologies, which are influenced by variations in the processing
parameters. In the work of Thijs et al [47], the effect of strong 111 fibre texture on the yield
strength anisotropy was studied for SLM-processed tantalum. The formation of 111 fibre
texture was attributed to the heat flow direction and the 100 preferred growth of tantalum
crystals. And the anisotropy in the yield strength was attributed to the 111 fibre texture, as
slip systems were oriented least favorably when the material was loaded along BD.
To understand the influence of the different fibre textures on the creep behavior, various fibre
ODFs (depicted in figure 11(a)) are synthetically generated by keeping the half-width of the
kernel function a constant (10◦ ). The half-width controls the kernel shape parameter (ψ) and
their interrelationship for the used de la Vallée Poussin kernel can be found in [46]. For each
fibre defined in figure 11(a), discrete orientations are sampled from the corresponding synthet-
ically generated ODF. The pole figures corresponding to 001, 110, and 111 fibre textures
generated in this study are depicted in figures 11(b)–(d), respectively. Similar to the previous
studies, the microscale microstructural features and fitting parameters are kept unchanged and
the creep simulations are performed at 980 ◦ C and 150 MPa with loading along BD and TD.
The minimum creep rate is computed for each fibre loaded along BD (˙min,BD ), and this data is
interpolated by a piecewise cubic interpolant to obtain the contour plot depicted in figure 12(a).
Schmid factor analysis is used to interpret the deformation behavior in the material when sub-
jected to loads along different directions. By defining a stress tensor along BD and with the
known orientation of the grains, the maximum Schmid factor can be computed for each grain
in the RVE. The stress tensor in the specimen coordinate system is rotated into each grain’s
local coordinate system to compute the Schmid factors for all slip systems. Only the highest
14
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Figure 12. (a) Minimum creep rate contours and (b) Schmid factor contours repre-
sented on inverse pole figure plots for illustrating the influence of various fibres on the
creep behavior. The values are computed for load along BD. (c) Comparison of creep
simulation results for the three extreme fibres 001, 011 and 111 loaded along BD.

Schmid factor corresponding to the most active slip system in each grain is considered, and the
corresponding mean value for the fibre is computed. By interpolating the Schmid factor values
the contours are plotted on the inverse pole figure plot as shown in figure 12(b).
During mechanical loading, the plastic deformation will begin once the shear stress in the
slip-plane along the slip-direction reaches the critical value. Thus, a higher Schmid factor
results in lower creep strength as the plastic deformation is carried by active slip systems. The
111 fibre and the fibres in its vicinity exhibit the lowest minimum creep rate (1.5 × 10−7 s−1 )
because these fibres provide the lowest Schmid factor (0.27). In contrast, the 100 fibre and the
fibres in its vicinity exhibit the highest minimum creep rate (5.7 × 10−7 s−1 ) as these fibres pro-
duce the highest Schmid factor (0.46). For all the fibres between these two extremes, a gradual
transition, both in the minimum creep rate and the Schmid factor values can be observed. The
creep curves for the three extreme fibres (001, 011, and 111) are depicted in figure 12(c).
This trend in the creep anisotropy has also been reported by Wollgramm et al [51] and Jácome
et al [52] for high temperature low stress creep testing of single crystal superalloys ERBO1 and
LEK 94, respectively. In both these studies, it is observed that the minimum creep rate for the
[1 1 0] tensile test is lower than that of the [1 0 0] test. Jácome et al attributed the observed
creep anisotropy in the loading directions to the number of slip systems activated, and the
dislocation networks at γ –γ  interfaces which accommodate lattice misfit.
15
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

4. Discussion

The key findings from our present study to understand the influence of crystallographic texture
characteristics (like texture intensity, alignment and fibre orientations) on the creep behavior of
an AM nickel-based superalloy are: (1) independent of the texture type (001 fibre or cube),
a decrease of texture intensity results in an increased creep resistance of the material; (2) the
creep resistance reaches a maximum when the fibre and cube textures are misaligned to the
specimen BD by 45◦; (3) the uncommon 111 and 110 fibres offer higher creep strengths in
comparison to the typical 001 fibre.
The results presented in this work are consistent with several experimental findings reported
in the literature. Many of these studies report significant increase in strength and ductility with
decreasing texture intensities, or with the formation of uncommon fibres like 110 and 111
fibres. This suggests that, while texture characteristics does impact the material strength sig-
nificantly, it is also important to study the high temperature creep resistance of the material in
which its usage occurs.
In the context of creep behavior for the IN738 superalloy, our results provide a first basis for
optimizing L-PBF processing parameters to obtain crystallographic textures that improve the
creep performance. Altering the laser scanning strategy can produce rotated fibre/cube textures.
Varying the laser energy density directly influences the meltpool shape and thus can produce
other atypical fibres. Increasing the post heat-treatment temperatures result in recrystallization
that produce weak textures. Based on our results, such textures offer higher creep resistance at
elevated temperatures. From a future perspective, an in-depth investigation into the relationship
between laser scan parameters and crystallographic texture will shed new light on the AM
process parameter optimization.
The creep anisotropy reported experimentally in the work of Risse [23] could be repro-
duced in this work by considering crystallographic texture as the major contributing factor.
However, grain structures and morphological textures (grains elongated along particular sam-
ple directions) could also contribute to anisotropic mechanical properties. Thus, future analyses
considering realistic grain structures can help further quantify the individual influences of these
microstructural features on the creep anisotropy.

5. Conclusions

This work investigated the influence of texture characteristics on the creep behavior of a L-
PBF IN738 superalloy in a micro-mechanical framework. The distinct individual influences
of texture alignment, intensity and atypical fibres on the creep response and the sensitivity to
variations in these features have been thoroughly investigated. From the results obtained in the
current work, the following conclusions are drawn:
(a) Post manufacturing heat treatments that lead to recrystallization are beneficial in improv-
ing the high temperature creep resistance of the material as they reduce the intensity of
crystallographic texture.
(b) The observed improvement in creep resistance offered by 110 & 111 fibre textures
at elevated temperatures provide further incentive to study the formation of such atypical
fiber textures during the L-PBF process.
(c) The present model can be extended to consider the influence of other microstructural fea-
tures (precipitate characteristics, realistic grain morphologies, pores) on the creep behavior
to optimize the L-PBF and heat treatment processing conditions.

16
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

Table 3. The slip systems for γ matrix and γ  precipitate.


α Mγα Mγα
√ √
1 [011̄] ⊗ (111)/√6 [2̄11] ⊗ (111)/√18
2 [101̄] ⊗ (111)/√6 [12̄1] ⊗ (111)/√18
3 [11̄0] ⊗ (111)/√6 [112̄] ⊗ (111)/√18
4 [011̄] ⊗ (1̄11)/√6 [211] ⊗ (1̄11)/√18
5 [101] ⊗ (1̄11)/√6 [121̄] ⊗ (1̄11)/√18
6 [110] ⊗ (1̄11)/√6 [11̄2] ⊗ (1̄11)/√18
7 [011] ⊗ (11̄1)/√6 [211̄] ⊗ (11̄1)/√18
8 [101̄] ⊗ (11̄1)/√6 [121] ⊗ (11̄1)/√18
9 [110] ⊗ (11̄1)/√6 [1̄12] ⊗ (11̄1)/√18
10 [011] ⊗ (111̄)/√6 [21̄1] ⊗ (111̄)/√18
11 [101] ⊗ (111̄)/√6 [1̄21] ⊗ (111̄)/√18
12 [11̄0] ⊗ (111̄)/ 6 [112] ⊗ (111̄)/ 18

Acknowledgments

AH gratefully acknowledges funding by Deutsche Forschungsgemeinschaft (DFG, German

Research Foundation) under Project-ID 190389738 within TRR 103 ‘From atoms to turbine
blades’.

Data availability statement

All data that support the findings of this study are included within the article (and any
supplementary files).

Appendix A. Constitutive model

The CP model proposed by Gao et al [37] has been used in this work, the fundamental aspects
of which are detailed in the works of Rice, Peirce et al, and Kalidindi [53–55]. In the CP
framework, for each crystal lattice type, a set of slip systems exists along which the slip occurs.
A slip system is defined by the slip direction (dα ) and the slip plane normal (nα ). For nickel-
based superalloys, with fcc lattice, 12 slip systems are identified with three slip directions and
four octahedral slip planes. The slip systems for the γ matrix, and the γ  precipitate are given
in table 3.

A.1. Matrix constitutive model

In the phenomenological CP framework, the shear rate (γ˙α ) is formulated as a function of the
resolved shear stress (τ α ) and the critical resolved shear stress (τ αc ) as:

γ˙α = f (τα , ταc ). (4)

In this framework, the temperature-dependent flow rule for the 110 {111} slip in each matrix
(γ) channel is given by:
   p1
Q110  τα + ταINT 
 sign (τα + ταINT )
γ̇ α = γ̇ 0 · exp −
110 110
(5)
RT  τ̂ slip
α +τ
oro − τ̂ 
soft

17
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

where γ̇ 0 is the reference shear rate, Q is the activation energy for dislocation glide, R is the gas
constant, T is the temperature, p1 is the inverse value of the strain rate sensitivity. τ oro denotes
the Orowan stress, τ̂ slip
α the resistance to slip, τ̂
soft
the softening stress, and ταINT the internal
stress.
A.1.1. Resolved shear stress (τ α ). For each slip system α, the resolved shear stress is obtained
from the externally applied stress as

τα = S · Mα (6)

where Mα is the Schmid tensor. Mα is defined by the slip direction dα and the slip plane normal
nα as Mα = dα ⊗ nα . The second Piola–Kirchhoff stress in the intermediate configuration is
calculated by
  
1
S = C̄ FeT Fe − I (7)
2

where I is the second order unit tensor, and C̄ is the elastic stiffness tensor on the macro-
scopic level. Fe is the elastic part of the deformation gradient (F) obtained from multiplicative
decomposition [56] as: F = Fe Fp .
A.1.2. Internal stress (ταINT ). The γ and γ  -phases both have an fcc lattice structure with a
slightly different lattice parameters [6]. They form a coherent interface, which means that the
crystal lattice planes are continuous across the interface, but a misfit strain exists to accom-
modate the difference in the lattice parameter. To bridge the misfit, both the precipitate and
matrix are strained, causing compressive and tensile misfit stresses. The amount of straining
is dependent on the magnitude of the misfit, elastic moduli of both materials and their relative
sizes. The misfit is defined as,
2(aγ  − aγ )
δ= (8)
(aγ  + aγ )

with aγ and aγ as the lattice parameters of the γ  and γ-phases respectively. The lattice misfit (δ)
between the γ and γ  -phases at 980 ◦ C for this alloy is +0.025% and is obtained from [57]. In
addition to these misfit stresses, due to inhomogeneous deformation in the two phases, plastic
strain gradients also develop. These strain gradients result in additional short-range internal
stresses at the γ –γ  interface. To account for these internal stresses due to lattice mismatch
and deformation heterogeneity, an internal stress term (ταINT ) is introduced in the flow rule
(equation (5)). Before the deformation begins, the misfit strain determines the internal stress
in the RVE. Here, in γ  precipitate mis = δI and in γ matrix channels mis = 0. As the plastic
strain is generated, the misfit strain is combined with the plastic strain as the general Eigen
strain to determine the internal stress in each time step. The new internal stress influences the
plastic strain in the next time step.
Analogous to the resolved shear stress, the short range kinematic hardening resulting from
γ –γ  lattice mismatch and the plastic deformation heterogeneity in the simplified RVE is
given by

ταINT = SINT · Mα (9)

where SINT is the internal stress for each subsection of the RVE (see figure 2) and is given by
a mathematical linear relation,

SINT = C̄  : ¯
 (10)
18
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

where C̄  is an effective stiffness matrix whose values are calculated as detailed in Gao
et al [28].
A.1.3. Slip resistance (τ̂ slip
α ). The influence of any set of slip systems (β) on the hardening
behavior of a slip system (α) is given by:

Ns p2
τ̂ slip
τ̂˙ slip
α = h0 χαβ 1 − βsat |γ̇ β | (11)
τ̂
β =1

where Ns is the number of slip systems, h0 is the initial hardening rate, χαβ is the cross-
hardening matrix, τ̂ sat is the saturation slip resistance due to dislocation density accumulation,
and p2 is a fitting parameter. This term empirically captures the micromechanical interactions
among different slip systems. The parameters h0 , p2 , and τ̂ sat are assumed to be identical for
all fcc slip systems due to the underlying dislocation reactions. The cross-hardening matrix
value is taken as 1.0 for coplanar slip systems α and β, and 1.4 otherwise, which makes the
hardening model anisotropic.
A.1.4. Orowan stress (τ oro ). The stress required to bow a dislocation line into the channel
between two precipitates is the Orowan stress, which is given by,

2 μb
τ oro = (12)
3 w

where μ is the shear modulus, b is the length of the Burgers vector and w is the γ channel width.
This term is introduced in the matrix flow rule as an additional threshold. Dislocations enter
the matrix channels once the effective resolved shear stress (τα + ταINT ) exceeds the Orowan
stress threshold. The slip resistance (τ̂ slip
α ) then determines whether the dislocation can move
any further or not.
A.1.5. Softening stress (τ̂ soft ). To account for the softening effect caused by the climb of dislo-
cations along the γ –γ  interfaces, a softening term τ̂ soft is introduced in equation (5) to counter-
act the hardening. This follows from the DDDs study performed by Gao et al [38] to examine
the creep behavior in nickel-based superalloys. As the softening stress is observed to increase
with increasing strain and to reach a saturation, it is incorporated in a phenomenological
manner,

τ soft = s1 · tanh(s2 · γ−channels ) (13)

where s1 and s2 define the softening stress and the rate of softening.

A.2. Precipitate constitutive model

The γ  -particles do not contain grown-in dislocations like the matrix phase, so deformation
of the precipitate can only occur when a matrix dislocation shears the particle. A matrix dis-
location must overcome a threshold before it can enter the particle, and then the lattice slip
resistance must be exceeded before it can move inside the precipitate. Similar to slip in the
matrix, the flow rule for 112 {111} slip in the precipitate is given by
  p
Q112  τα + ταINT  1
γ̇ 112 = γ̇ 112
· exp − INT
 sign(τα + τα ) (14)
α 0
RT  τ̂ slipα

19
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

where γ̇ α is the shear rate, γ̇ 0 is the reference shear rate, Q is the activation energy for the slip of
dislocation, R is the gas constant, T is the temperature, and p1 is the inverse value of the strain
rate sensitivity. The resolved shear stress (τ α ), the internal stress (ταINT ), and the resistance to
slip (τ̂ slip
α ) in the precipitate flow rule are analogous to those corresponding described in the
matrix constitutive model.

ORCID iDs

Abhishek Biswas https://orcid.org/0000-0001-6984-7144

Napat Vajragupta https://orcid.org/0000-0001-6085-7493
Alexander Hartmaier https://orcid.org/0000-0002-3710-1169

References

[1] Murr L E, Martinez E, Amato K N, Gaytan S M, Hernandez J, Ramirez D A, Shindo P W, Medina

F and Wicker R B 2012 J. Mater. Res. Technol. 1 42–54
[2] Mukherjee T, Zuback J, De A and DebRoy T 2016 Sci. Rep. 6 1–8
[3] Attallah M M, Jennings R, Wang X and Carter L N 2016 MRS Bull. 41 758–64
[4] Fu W, Haberland C, Klapdor E V, Rule D and Piegert S 2017 Streamlined frameworks for advancing
metal based additive manufacturing technologies in gas turbine industry Proc. of the 1st Global
Power and Propulsion Forum GPPF pp 16–8
[5] Pollock T M and Tin S 2006 J. Propul. Power 22 361–74
[6] Reed R C 2008 The Superalloys: Fundamentals and Applications (Cambridge: Cambridge Univer-
sity Press)
[7] Popovich V A, Borisov E V, Popovich A A, Sufiiarov V S, Masaylo D V and Alzina L 2017 Mater.
Des. 114 441–9
[8] Tabernero I, Lamikiz A, Martínez S, Ukar E and Figueras J 2011 Int. J. Mach. Tools Manuf. 51
465–70
[9] Geiger F, Kunze K and Etter T 2016 Mater. Sci. Eng. A 661 240–6
[10] Muñoz-Moreno R, Divya V D, Driver S L, Messé O M D M, Illston T, Baker S, Carpenter M A and
Stone H J 2016 Mater. Sci. Eng. A 674 529–39
[11] Wang X, Carter L N, Pang B, Attallah M M and Loretto M H 2017 Acta Mater. 128 87–95
[12] Carter L N 2013 Selective laser melting of nickel superalloys for high temperature applications PhD
Thesis University of Birmingham
[13] Li S, Wei Q, Shi Y, Zhu Z and Zhang D 2015 J. Mater. Sci. Technol. 31 946–52
[14] Moat R J, Pinkerton A J, Li L, Withers P J and Preuss M 2011 Mater. Sci. Eng. A 528 2288–98
[15] Tomus D, Tian Y, Rometsch P A, Heilmaier M and Wu X 2016 Mater. Sci. Eng. A 667 42–53
[16] Jadhav S D, Dadbakhsh S, Goossens L, Kruth J-P, Van Humbeeck J and Vanmeensel K 2019 J.
Mater. Process. Technol. 270 47–58
[17] Sun Z, Tan X, Tor S B and Chua C K 2018 NPG Asia Mater. 10 127–36
[18] Kunze K, Etter T, Grässlin J and Shklover V 2015 Mater. Sci. Eng. A 620 213–22
[19] Mercelis P and Kruth J P 2006 Rapid Prototyping J. 12 254–65
[20] Vrancken B 2016 Study of residual stresses in selective laser melting PhD Thesis Katholieke
Universiteit Leuven
[21] Dinda G P, Dasgupta A K and Mazumder J 2012 Scr. Mater. 67 503–6
[22] Carter L N, Martin C, Withers P J and Attallah M M 2014 J. Alloys Compd. 615 338–47
[23] Risse J 2019 Additive manufacturing of nickel-base superalloy IN738LC by laser powder bed fusion
PhD Thesis Rheinisch-Westfälische Technische Hochschule Aachen http://publications.rwth-
aachen.de/record/764478
[24] Divya V D, Muñoz-Moreno R, Messé O M D M, Barnard J S, Baker S, Illston T and Stone H J 2016
Mater. Charact. 114 62–74
[25] Sanchez S, Gaspard G, Hyde C J, Ashcroft I A, Ravi G A and Clare A T 2021 Mater. Des. 204
109647

20
Modelling Simul. Mater. Sci. Eng. 30 (2022) 055007 M R G Prasad et al

[26] Kocks U F, Tomé C N and Wenk H R 1998 Texture and Anisotropy: Preferred Orientations in
Polycrystals and Their Effect on Materials Properties (Cambridge: Cambridge University Press)
[27] Biswas A, Vajragupta N, Hielscher R and Hartmaier A 2020 J. Appl. Crystallogr. 53 178–87
[28] Gao S, Gogilan U, Ma A and Hartmaier A 2017 Modelling Simul. Mater. Sci. Eng. 26 025001
[29] Boeff M 2016 Micromechanical modelling of fatigue crack initiation and growth PhD Thesis Ruhr-
Universität Bochum
[30] Smit R J M, Brekelmans W A M and Meijer H E H 1998 Comput. Methods Appl. Mech. Eng. 155
181–92
[31] Hill R 1965 J. Mech. Phys. Solids 13 213–22
[32] Mandel J 1972 Plasticité Classique et Viscoplasticité: Course Held at the Department of Mechanics
of Solids (Berlin: Springer)
[33] Nemat-Nasser S 1999 Mech. Mater. 31 493–523
[34] Systèmes D 2016 Abaqus FEM software packakge (Waltham, MA: Dassault Systèmes)
[35] Bayerlein U and Sockel H 1992 Determination of single crystal elastic constants from DS-and
DR-Ni-based superalloys by a new regression method between 20C and 1200C Superalloys pp
695–704
[36] Ma A, Dye D and Reed R C 2008 Acta Mater. 56 1657–70
[37] Gao S, Wollgramm P, Eggeler G, Ma A, Schreuer J and Hartmaier A 2018 Modelling Simul. Mater.
Sci. Eng. 26 055001
[38] Gao S, Fivel M, Ma A and Hartmaier A 2017 J. Mech. Phys. Solids 102 209–23
[39] Björck Å 1996 Numerical Methods for Least Squares Problems (Philadelphia, PA: SIAM)
[40] Fernandez-Zelaia P, Acevedo O D, Kirka M M, Leonard D, Yoder S and Lee Y 2021 Metall. Mater.
Trans. A 52 574–90
[41] Wei H, Mazumder J and DebRoy T 2015 Sci. Rep. 5 1–7
[42] Bachmann F, Hielscher R and Schaeben H 2010 Texture analysis with MTEX—free and open
source software toolbox Solid State pphenom. 160 63–8
[43] Liu S Y, Li H Q, Qin C X, Zong R and Fang X Y 2020 Mater. Des. 191 108642
[44] Tan J C and Tan M J 2003 Mater. Sci. Eng. A 339 124–32
[45] Bunge H J 2013 Texture Analysis in Materials Science: Mathematical Methods (Amsterdam:
Elsevier)
[46] Schaeben H 1997 Phys. Status Solidi b 200 367–76
[47] Thijs L, Sistiaga M L M, Wauthle R, Xie Q, Kruth J-P and Van Humbeeck J 2013 Acta Mater. 61
4657–68
[48] Higashi M and Ozaki T 2020 Mater. Des. 191 108588
[49] Jones D R, Fensin S J, Ndefru B G, Martinez D T, Trujillo C P and Gray G T III 2018 J. Appl. Phys.
124 225902
[50] Sidambe A T, Tian Y, Prangnell P B and Fox P 2019 Int. J. Refract. Met. Hard Mater. 78 254–63
[51] Wollgramm P, Bürger D, Parsa A B, Neuking K and Eggeler G 2016 Mater. High Temp. 33 346–60
[52] Jácome L A, Nörtershäuser P, Heyer J-K, Lahni A, Frenzel J, Dlouhy A, Somsen C and Eggeler G
2013 Acta Mater. 61 2926–43
[53] Rice J R 1971 J. Mech. Phys. Solids 19 433–55
[54] Peirce D, Asaro R J and Needleman A 1982 Acta Metall. 30 1087–119
[55] Kalidindi S R 1998 J. Mech. Phys. Solids 46 267–90
[56] Lee E H 1969 J. Appl. Mech. 36 1–6
[57] Strunz P, Petrenec M, Davydov V, Polák J and Beran P 2013 Adv. Mater. Sci. Eng. 2013 1–7

21