Author’s Accepted Manuscript

Effect of heat treatment on microstructure and

mechanical properties of Inconel 718 processed by
selective laser melting

E. Chlebus, K. Gruber, B. Kuźnicka, J. Kurzac, T.

Kurzynowski
[Link]

PII: S0921-5093(15)00568-7
DOI: [Link]
Reference: MSA32363
To appear in: Materials Science & Engineering A
Received date: 14 April 2015
Revised date: 12 May 2015
Accepted date: 13 May 2015
Cite this article as: E. Chlebus, K. Gruber, B. Kuźnicka, J. Kurzac and T.
Kurzynowski, Effect of heat treatment on microstructure and mechanical
properties of Inconel 718 processed by selective laser melting, Materials Science
& Engineering A, [Link]
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 Effect of heat treatment on microstructure and mechanical properties

of Inconel 718 processed by selective laser melting

E. Chlebus, K. Gruber*, B. Kuźnicka, J. Kurzac, T. Kurzynowski

Centre for Advanced Manufacturing Technologies / Fraunhofer Project Center, Faculty of

Mechanical Engineering, Wrocław University of Technology,

ul. Łukasiewicza 5, 55-371 Wrocław, Poland

Abstract

Microstructural and mechanical properties of Inconel 718 were determined on the

specimens manufactured by selective laser melting (SLM) of prealloyed powder. High density

(99.8%) cylindrical specimens were built with four orientations (0°, 45°, 45°x45° and 90°) in

relation to the building and scanning directions. Because of directional, dendritic-cellular

grain growth, microstructure of the as-built specimens was characterized by columnar grains

of supersaturated solid solution with internal microsegregation of Nb and Mo, demonstrated

by fractions of Laves eutectic or its divorced form in interdendritic regions. Such a

heterogeneous microstructure is unsuitable for direct post-process ageing and makes the alloy

sensitive to subsolidus liquation during rapid heating to the homogenizing temperature. In

homogenized and aged condition, the alloy received a very good set of mechanical properties

in comparison with the wrought material. In heat-treated condition, like in as-built condition,

weak anisotropy of properties was found, manifested by lower Young's modulus, yield

strength and tensile strength of the specimens extended along the build direction in

comparison to the values for the other variants of the specimens. This is attributed to the fact

*
Corresponding author. Tel.: +48 71 320 41 85; fax:+48 71 320 42 02. E-mail address:

[Link]@[Link] (K. Gruber).

1
that the grains maintained their geometric and crystallographic texture obtained during

solidification.

Keywords: Selective laser melting; Inconel 718; microstructure; mechanical properties

1. Introduction

Selective laser melting (SLM) is an additive manufacturing (AM) technique in that

components are built by controlled melting subsequent powder layers with a focused laser

beam. The process was described in details in [1,2]. The SLM process offers several

advantages in comparison to traditional manufacturing technologies, like reduction of

production steps, high flexibility, high efficiency of material consumption and, the most

important, possibility to manufacture parts with high geometrical complexity and dimensional

accuracy. However, the process-specific short interaction times accompanied by highly

localized heat input result in large thermal gradient and, consequently, in high thermal stresses

during the process [3-5]. On the other side, rapid solidification of thin layers results in

directional grain growth, microsegregation of chemical composition and creation of non-

equilibrium phases [6]. Therefore, post-processing heat treatment should be applied to adapt

properties of the SLM-parts to the requirements of working conditions or at least to reduce

thermal stresses. Another major drawback of the SLM process is hardly avoidable occurrence

of process-induced defects. These pore-like defects result from not optimally selected process

parameters, initial powder contamination and local voids after deposition of a layer [7,8].

Since porosity determines mechanical behavior of the SLM-processed parts, their density

(>99.5%) is accepted as a criterion of optimization of the processing conditions [9].

Inconel 718 (IN718) is a promising candidate for industrial applications of SLM

manufactured parts [9,10]. This is an age-hardenable Ni-Fe-Cr alloy that combines high-
2
temperature strength up to 700 °C with corrosion resistance and excellent fabricability [11,12].

Because of these attributes, IN718 is a commonly used material for parts for aircraft turbine

engines, high-speed airframe parts, high-temperature bolts and fasteners and for nuclear

engineering, as cast, wrought and powder metallurgy products. The studies focused on laser-

melting-processed IN718 are available quite numerously, but they mostly concern thin-walled

builds manufactured using direct laser fabrication (DLF), like e.g. [5,10,13], and few only

concern SLM-fabricated parts [6,14-16]. From here, some inaccuracies and inconsistencies

result at interpretations of the relationship between microstructure of as-built parts and the

process parameters, as well as between initial microstructure and those obtained after different

variants of post-process treatment.

The current study was aimed at clarifying influence of material characteristics and

processing conditions on metallurgical mechanisms and resultant microstructural and

mechanical properties of IN718 processed by SLM, as well as at tailoring the post heat-

treatment parameters to the nature of rapid solidified microstructure, i.e. to microsegregation

degree of the alloying elements and residual stresses left after the process.

2. Materials and methodology

Test specimens were fabricated from IN718 powder using the SLM Realizer II 250 (MPC-

HEK) machine. The process was executed by the continuous wavelength (CW) Ytterbium

fiber laser with a beam spot size of 180 μm and maximum power of 100 W. The process was

conducted under protective atmosphere of high purity argon.

Commercially available IN718 powder was used in this investigation, with chemical

composition given in Table 1. Sieve analysis showed that powder consisted in most of

spherical particles dia. 5 to 50 μm. Fine satellite particles attached to the main-sized particles

were also observed, see Fig. 1. The powder had good flowability.
3
 Table 1. Chemical composition of processed IN718 powder

Element Ni Cr Fe Nb Mo Ti Al C Mn Si
wt% 51.02 18.37 19.5 5.43 3.27 0.91 0.035 0.036 0.13 0,01

Two kinds of specimens were built: cuboid-shaped specimens 10 x 8 x 6 mm and

cylindrical specimens dia. 4 mm with gauge length 19 mm (tensile specimens acc. to ASTM

E8-04). The cuboidal specimens, as control specimens, served for optimum selection of

process parameters to guarantee the builds with maximum density of 99.7 to 99.98%. They

were also used for heat-treatment trials in order to select its optimum version on the grounds

of microstructure examinations and hardness measurements.

Fig. 1. SEM image of processed IN718 powder

The following parameters were used for manufacture of tensile specimens: 100 W laser

power, 50 μm layer thickness, 160 μm distance between scanning lines (hatch spacing), 85.7

mm/s scanning speed. The layers were scanned according to a zigzag pattern that was rotated

90° after each layer. To achieve maximum density of the specimens and to minimize residual

4
stresses, each powder layer was processed by the laser beam twice, so that the rescanning

(remelting) direction was perpendicular to the first scanning (fully melting powder) direction.

The tensile specimens were built in four series, differing with orientation of their axes with

respect to the build direction (Fig. 2): series A (0°) – tension axis parallel to building direction

z, series B (90°) – tension axis normal to building direction z, series C (45°) – tension axis at

45° to building direction z, series D (45°x45°) – tension axis at 45° to building direction z and

also at 45° to both scanning directions x and y. The tensile specimens were CNC finished to

remove residues of support structures and to achieve smooth, notch-free surface. For each

series, five samples were included. Each series was built twice – one for examination in as-

fabricated conditions and one for examination as heat-treated.

Fig. 2. Building strategy of tensile specimens (series A, B, C and D).

The specimens fabricated from IN718 by SLM in the as-fabricated and as heat-treated

conditions were tested (at room temperatures) for their mechanical properties such as yield

strength (σ0.2), tensile strength (σt), Young’s modulus (E), elongation at fracture (εf) and

Vickers microhardness (HV1). Tests were carried out using an Instron 3384 testing machine

5
according to EN 10002-1. Vickers hardness was measured with a Zwick Roel testing

machine.

Heat treatment to adapt the SLM-parts to the requirements of working conditions was

carried-out with use of a Czykol FCF 4/160M furnace. Accepted were four variants of heat

treatment, shown in Table. 2. Individual variants differed from each other in the conditions

preceding the ageing process, i.e. the as-built condition and three conditions after solutioning

at various annealing temperature. Accepted were standard ageing parameters used for

wrought IN718 alloy [10].

Table 2. Heat treatment variants. AC – air cooling; WC – water cooling; FC – furnace cooling.

Variant Solution treatment Age hardening

I As-built
II At 980 °C for 1 h (WC)*
720 °C /8 h (FC 100 °C/h), 620 °C / 10 h (AC)
III At 1040 °C for 1 h (WC)*
IV At 1100 °C for 1 h (WC)*
* Furnace heating up to the annealing temperature.

Both kinds of specimens were prepared in the standard procedure for microstructure

observation. Examinations were carried-out on longitudinal and transverse polished sections,

etched with 87 Glyceregia (standardized acc. to ASTM E 407) consisting of 15 ml HCl, 10 ml

Glycerol and 5 ml HNO3. The etchant was used fresh, 5 to 15 minutes after preparation.

Optical microscope used in microstructure examinations was Olympus LEXT OLS4000

(OM). Scanning electron microscope Zeiss SEM EVO MA25, equipped with an EDS analysis

system, was used for observations of higher magnification images, fracture surfaces and for

analysis of microsegregation.

6
 3. Results and discussion

3.1. Microstructure and properties in as-built condition

Microstructures of as-built specimens, at relatively low magnifications, are shown in Fig. 3

in three mutually perpendicular planes xz, yz, and xy. Visible is laminar material structure and

columnar architecture throughout the builds, whose geometry is determined by the specimen

building strategy, i.e. zig-zag pattern of scanning, hatch spacing and thickness of layers.

Fig. 3. Microstructure of Inconel 718 in as-built condition in three mutually perpendicular planes of
the polished section. Visible is columnar architecture and laminar material structure
in yz and xz sections and laser scanning tracks in section xy.

As can be seen in Fig. 3, the unidirectional columns in the yz plane are slightly deflected

from the build direction following the moving heat source, whereas the columns in the xz-

plane are oriented in a slightly zig-zag fashion in relation to the build direction. Judging from

the etching effects (see Fig. 3), individual columns are created by grains of solid solution γ

elongated in their growth direction, separated by high-angle boundaries. This is in accordance

with the EBSD maps obtained by Tozzi et al. for similar as-built SLM microstructure [14].

Such type of microstructure, similar to the directionally solidified microstructure, is a result of

epitaxial, dendritic grain growth in the direction determined by heat flux direction and, for

FCC structures, by crystallographically favored orientation <001> [6,13,17].

7
 Since effective direction of heat flux is a resultant of horizontal heat flux related to

movement of the heat source and vertical heat flux related to heat transfer into the substrate,

orientation of crystallographically aligned dendrites (or dendritic cells) within the columnar

grains (Fig. 4) and high-angle misorientations between the grains depend on the process

parameters. It should be emphasized that, if the fully columnar grain morphology seems to be

a standard in the case of SLM-builds [6,14-16] irrespective of differences in the ranges of

process parameters, then in the case of direct laser-fabricated builds, columnar or banded

columnar-equiaxed grain microstructure was observed [13].

Apart from grains morphology, differences in thermal stories of IN718 components

affected by process parameters significantly affect residual stresses level, grain size and phase

composition of microstructure of SLM-parts in as-built condition. The alloy IN718 contains

the elements with high susceptibility to segregation, i.e. Nb, Mo and C, from that Nb is

important with regard to its lowest value of partition coefficient (<0.5 [18]) and formation of

NbC, -Ni3Nb phase and Laves-type phases, brittle intermetallic compounds represented as

(Ni, Cr, Fe)2(Nb, Mo,Ti). Rapid solidification of the molten pool during SLM prevents

macrosegregation completely, but does not guarantee lack of microsegregation of these

elements, as it was found in this work and can be seen in Figs. 4a, b. Such a microsegregation

pattern results from rapid dendritic/cellular growth of columnar grains, because at high

cooling rates the primary dendrite shape is altered to a more cell-like form [19,20]. As can be

seen in Fig. 4b, inhomogeneities of chemical composition occur between aligned dendritic

cells and in the region of an overlapping interface between two adjacent laser scanning tracks

or layers.

8
 Fig. 4. Magnified images of microstructure (see Fig. 1) in xz plane in as-built condition, showing
internal dendritic-cellular grain structure; arrows 1 indicate tracks interface: (a) OM image, (b) SEM
image of track interface; arrows 2 indicate segregation traces at dendritic cells tips (see arrows 2).

Antonsson et al. [19] found that solubility of Nb in the primary austenitic  phase increases

with increasing cooling rate. Concentration of Nb controls the solidification sequence together

with interacting alloy elements such as C, so at cooling rates > 103 K/s, characteristic for

SLM, expected can be both complete supersaturation and small quantity of precipitates of

Laves phase and also of NbC carbides. As results from isothermic sections of the Ni-Nb-C

equilibrium diagram [11], NbC carbides and Laves-phase particles can be created in the

interdendritic spaces as products of eutectic reactions when concentration of Nb, Mo and C in

the melt is sufficiently high.

In this work, occurrence of small particles of eutectic mixture (+ Laves-type phase) or its

divorced form and fine MC-type carbides was found in interdendritic regions of all the

examined specimens, see Fig. 5. Spherical shape of carbides indicated their primary formation

in liquid before the Laves eutectic reaction. EDS analysis showed that the carbides are Nb and

Ti-rich. This confirms the statement contained in [19] that TiN particles can nucleate NbC

carbides. Identical results are reported in [13] and [21].

9
 Fig. 5. Interdendritic microsegregation inside γ-phase grains; the arrows indicate:
1 – colonies of eutectic mixture (+ Laves type phase); 2 – divorced eutectics;
3 – MC carbides (black). BSD/SEM.

Results of tensile test and hardness measurements are given in Table 3. It was expected

that directional growth of crystals with privileged orientation <001> would result in clear

dependence of Young's modulus on the angle between the specimen axis and the direction of

solidification. It can be seen that, in spite of large dispersion of the results, average values of

the modulus demonstrate a clear tendency E(0°) < E(90°) < E(45°), identical as that for

directionally solidified nickel-based superalloy IN 738 LC reported in [22]. Considering the

results of Hermann et al. [22], according to that E(45°) to E(0°) for a strongly textured

material reaches 2, the value of 1.28 obtained in the presented research proves the definitely

weaker, but not negligible, <001> texture of the material. This weaker texture is caused by

high-angle misorientation between directions of dendritic-cellular growth of individual

columnar grains (Fig. 4) due to the movement of the heat source affecting local thermal

gradients and crystal growth rates [23]. Since the specimens were processed by the 100-W

low energy laser system at low scanning speed and with 90° rotation of scanning directions

between the layers, the obtained weak crystallographic texture confirms the conclusions

10
resulting from the examinations of IN 718 reported in [10,13,14] and also from the

examinations of other materials, e.g. Al-Si alloys [23] and austenitic steel [24].

As can be seen in Table 3, values of Young’s modulus of the B, C and D specimens are not

significantly different from each other and are close to those specified for IN 718 in as-rolled

condition, while, according to [22], the E(45°) value for directionally solidified IN 738 LC is

over 1.4 times higher than the E(90°) value. This is certainly related to the fact that

microstructures of the SLM specimens built vertically only are very similar to microstructures

of directionally solidified specimens with respect to the shape aspect ratio of columnar grains

and, consequently, to stronger crystallographic texture. The specimens for anisotropy

examinations of tensile properties of directionally solidified materials were cut-out of

unidirectionally solidified plates, while the SLM specimens were built individually, so growth

of columnar grains in parallel to the building direction is restricted by circular (series A) or

elliptical (series C and D) contour of xz sections of the specimens. As a result, large fraction

of grains with small shape aspect ratio in microstructure is the main reason why no significant

difference was found between the E(45°) and E(90°) values of SLM-specimens.

Table 3. Examination results for the specimens of IN718 in as-built condition

SLM
Series As-rolled
Properties (α=90°)
[11]
A B C D [6,15]
Youngs modulus, E
162 ± 18 193 ± 24 200 ± 23 208 ± 48 200-211
[GPa]
Yield strength, 0.2
572 ± 44 643 ± 63 590 ± 15 723 ± 55 830-907 448-727
[MPa]
Tensile strength, σt
904 ± 22 991 ± 62 954 ± 10 1117 ± 45 1120-1148 896-1014
[MPa]
Elongation at fracture,
19 ± 4 13 ± 6 20 ± 1 16 ± 3 19-26 40-46
f [%]
xy-plane: 297 ± 5; xz-plane: 319 ± 10; xy-plane:
Hardness HV1 160-320
yz-plane: 322 ± 10 365

11
 Anisotropy of the microstructure affects fracture mode and tensile properties due to

different orientation of columnar grains in relation to the direction of tension. As can be seen

in Fig. 6, fracture of the specimens series A (as well as B) contained a fibrous zone and a

shear-lip zone, but rupture of the specimens series D (as well as C) occurred on the surface of

maximum shear stresses at approximately 45° in relation to the tensile axis. The differences

are also visible in the surface deformation pattern strictly related to geometry, length and

orientation of columnar grains in relation to the tension direction. It is interesting that the

values of both yield strength and tensile strength of the specimens series D are approximately

1.25 times larger than the values obtained for the specimens series A, just as the values of

Youngs modulus.

The tensile strength values for the A specimens are significantly lower in comparison to

those of the other series, because perpendicular orientation of overlapping interfaces between

layers to the loading direction is conducive to cracking and because strengthening by grain

boundaries (Hall-Petch relation) is relatively weaker in this case. An opposite tendency is

shown by directionally solidified materials where, because no grain boundaries are transverse

to the tension direction, and the axial stresses necessary to activate slip systems in -phase

channels acting in parallel to the solidification direction are higher than those deflected by

45°.

The obtained average values of hardness and tensile strength are lower than those given in

[6,15] for flat SLM specimens built horizontally, but coincide with typical values for the as-

rolled condition [11].

12
 Fig. 6. Deformation relief on lateral surface of the as-built specimens depending on the angle
between the build direction and the tension direction: (a) 0°, (b) 45°.

The elongation values relatively lower than those given in [6,11,15] should be attributed to

microstructure inhomogeneities, in particular to occurrence of Laves phases in interdendritic

spaces and in the regions of layer interfaces. Highly probable is occurrence of secondary

precipitates of /″ phases in the places of higher Nb concentration.

Perhaps, interpretation of the obtained differences between tensile properties for individual

series of specimens requires considering, apart from influence of texture, also influence of

distribution of residual stresses [5] determined by strategy of building the specimens [3,25].

Relatively small dispersion of the results indicates small influence of accidental flat build

defects between the layers, especially in the series A specimens where orientation of the

defects is most unfavorable, transverse to the tension direction [2].

3.2. Microstructure and properties in heat-treated condition

Since components made of IN 718 are mainly designed for working in a strongly corrosive

environment at low and high temperatures, their desirable microstructure is stable γ phase

strengthened with coherent and dispersive precipitates of the phases γ (Ni3(Ti,Al)) and γ

(Ni3Nb). Results and effects of the ageing process depend on initial microstructure of the

alloy, i.e. homogeneity of chemical composition, level of residual stresses related to the SLM
13
process parameters, geometric dimensions of the built part and the build strategy. This is why

microstructure of the as-fabricated IN718 alloy, susceptible to microsegregation of Nb and

Mo and to forming non-equilibrium phases as well as having a significant level of residual

stresses, must be modified by heat treatment tailored to both the as-built condition and to

further service conditions.

Preliminary trials to select optimum conditions of solutioning were carried-out on cuboidal

specimens. Four variants of heat treatment were accepted, differing with the process

conditions, see Table 2. It was considered that the accepted range of solutioning temperature

is close to the temperature of the Laves eutectic reaction that for IN718 alloy begins between

1160 °C and 1075 °C at low (0.25 °C/s) solidification rates [19]. This means that the as-built

microstructure of SLM specimens with metastable Laves eutectic particles is sensitive to

subsolidus liquation at the dissolving particle-matrix interface during rapid heating [26,27].

For this reason, the specimens were slowly furnace-heated to the solution annealing

temperature. Since the partition coefficients of Nb and Mo are negative, it can be supposed

that, during slow heating to the solutioning temperature, dissolving of Laves phases occurs by

back-diffusion of the solute atoms from solute-rich regions into the matrix without local

melting.

Accepted were standard two-stage annealing parameters, recommended for IN718 [11].

Results of heat-treatment were evaluated on the grounds of microstructure observations and

hardness measurements:

1. As-built + aged condition. Direct, post-process ageing resulted in 48-% hardness

increase of the specimens from 312 to 46113 HV1, thus indicating high supersaturation level

of the rapidly solidified solid solution γ. Microstructure maintained its columnar nature and

traces of layered build of the material, see Figs. 7a,b. The effect of irregular precipitation of

the γ/γ phases was marked by intensive etching of interdendritic areas and boundaries of
14
columnar grains. In addition, small quantities of needle-like -Ni3Nb precipitates appeared in

interdendritic regions, as shown in Fig. 8.

Fig. 7. Microstructure aged just after the SLM process. Irregular precipitating of γ/γ phase disclosed
the microsegregation pattern, as well as low- and high-angle boundaries: (a) in xz plane, (b) in xy
plane. OM.

Fig. 8. Magnified SE/SEM image of dendritic-cellular microsegregation inside a columnar grain of -

phase, corresponding to Fig. 7a: Laves-type phase, MC-type carbides and -phase indicated by arrows
1, 2 and 3, resp.

2. Solutioned at 980C/1h and at 1040C/1h + aged conditions. In both cases, solutioning

did not significantly change supersaturation degree of the solution in comparison to the as-

built condition. Ageing after quenching resulted in an increase of hardness by 49% (from 312

to 4658 HV1) after quenching from 980 °C and by 52% (from 312 to 47412 HV1) after

15
quenching from 1040 °C, the increase being not much larger than that after post-processing

ageing. The solution annealing parameters used in both variants appeared insufficient for

complete homogenization of - phase. As can be seen in Figs. 6a and Fig. 6b, residual

particles of Laves phase, identified by EDS analysis (Fig. 7), occur on grain or subgrain

boundaries and along layer interfaces. Occurrence of needle-like  phase is justifiable, since it

can precipitate during annealing within 750 °C to 1000 °C for a time shorter than 100 h [13],

i.e. during both solution and aging annealing. Residual particles of Laves phase are situated in

the places where, in the as-built microstructure, quantity and size of particles of Laves

eutectics were the largest due to Nb concentration, i.e. in overlapping interfaces between

tracks or layers of laser scanning. This results from the fact that, during melting a new powder

layer, surface of the previously solidified layer becomes melted again and thus Nb

concentration increases in the last remaining liquid to be solidified. Slow heating and soaking

at the solution annealing temperature enables progressive dissolving the metastable Laves

particles and reducing the degree of Nb microsegregation by back-diffusion of the solute

atoms from solute-rich regions into the  matrix. Due to low diffusion coefficient of Nb atoms

in Ni, complete dissolution of residual Laves particles requires proper selection of

temperature and time of the annealing.

16
 Fig. 9. Microstructure after solutioning: (a) at 980 °C/1h and aging, (b) at 1040 °C/1h without aging.
Residual particles of Laves phase on grain boundaries (arrows 1) and on the layer interface (arrow 2)
accompanied by -phase precipitates (arrows 3); xz planes. SEM.

Fig. 10. Material solutioned at 980 C/1h and aged. Distribution maps of main alloying elements: (a)
analyzed area; (b) to (d) distribution of Ni, Nb and Mo. EDS analysis.

It should be emphasized that, in the microstructure annealed at 1040C/1h and water-

quenched, observed was the effect of grain boundary migration in form of locally present

recrystallized grains and annealing twins (Fig. 11).

Fig. 11. Microstructure after solutioning at 1040 °C/1h and water quenching (see Fig. 9b). Partially
recrystallized area on xz plane. The recrystallized grains are denoted by R. SEM.

3. Solutioned at 1100C/1h and aged condition. In these annealing conditions,

homogenization of γ solid solution occurred, as can be seen in Figs. 12 a,b. The  grain

boundaries are decorated with MC carbide that pinned-down moving grain boundaries.

17
Increase of hardness after ageing in relation to the as-built condition did not exceed 48%

(from 312 to 463  8 HV1). This heat treatment variant was found optimal from the viewpoint

of satisfying degree of microstructure homogenization and ageing effects.

Fig. 12. Material solutioned at 1100 °C/1h and aged. Grain boundaries of homogenized solid solution
 decorated with MC carbides. Arrows indicate moving grain boundaries pinned-down by carbides.
Visible are also annealing twins: (a) xz plane, (b) xy plane. CM.

When comparing the microstructures in Figs. 7a,b with those in Figs. 12a,b, with regard to

shape and size of grains, one can agree with the authors of [5,6,14] stating that

recrystallization occurs during annealing IN718 at temperature higher than 1100 °C.

Nevertheless, if the term ″recrystallization is used in the commonly accepted meaning [28],

the argument that residual thermal stresses, accumulated by repeated rapid heating and

cooling of molten pool during SLM, are the driving force of recrystallization [5] seems

questionable, especially that partial stress relieving should occur as early as during heating the

alloy to the homogenization temperature. In this research, specimens were manufactured by

double scanning of each layer which should significantly reduce residual stress levels in the

specimens and thus reduce occurrence probability of residual plastic strains.

It seems more probable that the microstructure obtained after solutioning at 1100 °C/1h

and ageing (Fig. 12) is a result of grain boundary migration induced by reduction of grain

18
boundary area itself, discontinuous dissolution of columnar arrayed phases and solute

diffusion along the boundaries [29]. It should be considered that as-built microstructure of the

SLM-ed IN718 alloy is unstable because of high energy accumulated in form of grain

boundaries and, to a lesser degree, of dislocations. Therefore, the driving force of grain

boundary migration is much higher that that in a traditionally manufactured alloy. Moreover,

as can be seen in Fig. 11, the γ/γ and -phase precipitations and Laves phase particles are

easily dissolved at grain boundaries, because diffusion at a migrating grain boundary is much

faster than volume diffusion. On this ground it can be supposed that grain coarsening and

diffusion of Nb atoms at a moving boundary have a decisive effect on mobility of grain

boundaries during solutioning.

At the moment when migration of high-angle grain boundaries became possible, the

process of their straightening and reducing developed surface of columnar grains started,

controlled by the pinning effect of carbides and formation of twin boundaries (see Figs.

11,12). During grain migration are formed annealing twins with density dependent on grain

size, annealing parameters, velocity of grain boundary migration, grain boundary energy and

stacking-fault energy [30]. Nickel has a high stacking-fault energy at room temperature but,

when heated beyond curie temperature (>354 °C), its stacking-fault energy decreases and is

typically about 40 mJ/m2 at 1000 °C [31], so twinning is expected during grain growth in the

IN718 alloy. At the second stage, uniform precipitation of the hardening phases /

occurred. No precipitates were found on grain boundaries, although local presence of -phase

particles can not be excluded.

Tensile specimens of all the series (from A to D) were heat-treated according to the variant

VI regarded as optimum. Results of tensile test and hardness measurements are given in Table

4.

19
 Table 4. Mechanical properties of IN718 specimens in heat-treated condition acc. to variant IV

As-built series SLM As-rolled

Properties (α=90°) and heat-
A B C D [15] treated [11]
Youngs modulus,
163 ± 30 199 ± 15 188 ±19 209 ± 44 201 200-211
E [GPa]
Yield strength, 0.2
1074 ± 42 1159 ± 32 1152 ± 24 1241 ± 68 1102-1161 1048-1138
[MPa]
Tensile strength, σt
1320 ± 6 1377 ± 66 1371 ± 5 1457 ± 55 1280-1358 1317-1348
[MPa]
Elongation at
19 ± 2 8±6 15 ± 5 14 ± 5 10-22 19-21
fracture, f [%]
Hardness HV1 463 ± 8 470 430-460

As can be seen by comparison of the data in Tables 3 and 4, depending on the build

strategy, heat treatment caused strong increase of yield strength by 72 to 95%, of tensile

strength by 30 to 46% and of hardness by 48% in relation to the as-build condition. The

obtained properties are approximate to those published in [15] and typical for the alloy in hot-

worked condition [11]. However, the elongation values are slightly lower, especially for the

B-series specimens. Remarkable is the way how heat treatment influenced mechanical

properties of the specimens series D and series A in comparison to the as-built condition. The

ratio of Youngs modulus values did not change, remaining equal to 1.28, but the ratios of

yield strength and tensile strength values decreased from 1.25 to 1.16 and 1.10, respectively,

mainly due to homogenization of γ-solution and uniform precipitation of the hardening phases

/. In addition, both deformation relief on cylindrical surfaces of the specimens and

orientation of fracture surfaces in relation to the tension direction were similar as for the as-

built specimens shown in Fig. 6. Since, in spite of heat treatment, elongated shape of grains

was maintained (Fig. 13), this proves a significant influence of grain growth direction on

behavior of the product under load depending on the angle between loading direction and

growth direction.

20
 No differences were found between properties, which could result from reduction of

residual stresses dependent on build strategy of the specimens (series from A to D). The

obtained results indicate rather the fact that the applied double-scanning strategy ensured so

low level of residual stresses that no significant effect of them on the examination results was

recorded.

Fig. 13. Orientation of grains and fracture surfaces in the direction: (a) perpendicular to the build
direction (series A); (b) at 45° to the build direction (series D). Longitudinal sections. OM.

Heat treatment reduced the differences between strength of individual series specimens by

homogenization and suppression of intergranular misorientation, but left weak

crystallographic and geometric texture of grains, inherited from the as-fabricated condition.

The obtained results suggest that, in spite of directional solidification, proper selection of

the process parameters enables obtaining a satisfactory set of properties of SLM-fabricated

parts of the alloy IN718 with weak anisotropy in spite of directional solidification.

Microstructure of the SLM- material should be modified by subsequent heat treatment with

parameters determined by the as-built microstructure and the endeavor to achieve mechanical

properties similar to wrought material. Nevertheless, at this designing stage, influence of the

21
build direction in relation to directions of the expected service loads must be taken into

account.

4. Conclusions

In this study, microstructure of SLM-fabricated Inconel 718 in as-built and after heat

treatment conditions was investigated and related to the mechanical performance under tensile

loading. The main findings can be summarized as follows:

1. Manufacturing parts with high density from pre-alloyed IN718 powder by SLM does

not make any problem with the point of view of selecting an optimum set of process

parameters. Using double scanning makes it possible to reach density at the level of

99.8% and to reduce residual stress level significantly.

2. The SLM-fabricated parts in the as-built condition are characterized by textured

microstructure with fine columnar grains across several layers, specific for laminar

directional solidification. Columnar grains, separated from each other by high-angle

boundaries, are characterized by specific substructure created by packages of parallel

arranged dendrite-cells with low-angle misorientation.

3. Dendritic-cellular growth of grains at high cooling rates and susceptibility of Nb and

Mo to microsegregation result in forming non-equilibrium Laves phases in

interdendritic spaces, on grain boundaries and on layer interfaces, as components of

eutectic mixture or its divorced form. In spite of high supersaturation level of γ

solution, such a heterogeneous microstructure should be modified.

4. For the IN718 alloy, heat treatment by solutioning and double ageing is necessary.

Homogenization of the alloy requires higher temperature than that normally used, i.e.

22
 1100 ºC, and slow furnace-heating to avoid local subsolidus liquation of the material

and possible propagation of metastable liquid along grain boundaries. At this

temperature, grain growth is inhibited by presence of niobium carbides.

5. Heat treatment does not eliminate texture completely, because it does not change

completely the elongated shape and crystallographic orientation of solid solution

grains. The largest deflection from average values of mechanical properties towards

their lower values characterizes the specimens built vertically. Yield strength, tensile

strength and hardness values of aged, SLM-fabricated alloy are slightly higher than

those for a wrought material. Elongation values tend to be slightly lower, but are still

on a satisfactory level.

Acknowledgements

This research was financially supported by The National Centre for Research and

Development (No. PBS1/A5/12/2012).

References

[1] T. Kurzynowski, E. Chlebus, B. Kuźnicka, J. Reiner, High Power Laser Materials Processing:
Lasers, Beam Delivery, Diagnostics, and Applications, 823914 (Feb. 9, 2012);
doi:10.1117/12.907292.
[2] E. Chlebus, B. Kuźnicka, T. Kurzynowski, B. Dybała, Materials Characterization 62 (2011) 488.
[3] M. Shiomi, K. Osakada, K. Nakamura, T. Yamashita, F. Abe, CIRP Annals - Manufacturing
Technology (2004) 195.
[4] P. Mercelis, J-P. Kruth, Rapid Prototyping Journal 12 (2006) 254.
[5] F. Liu, X. Lin, G. Yang, M. Song, J. Chen, W. Huang, Optics & Laser Technology 43 (2011) 208.
[6] K.N. Amato, S.M. Gaytan, L.E. Murr, E. Martinez, P.W. Shindo, J. Hernandez, S. Collins, F.
Medina, Acta Materialia 60 (2012) 2229.
[7] B. Vandenbroucke, J-P. Kruth, Rapid Prototyping Journal 13 (2007) 196.
[8] H. Gong, K. Rafi, H. Gu, T. Starr, B. Stucker, Additive Manufacturing 1-4 (2014) 87.

23
[9] D.D. Gu, W. Meiners, K. Wissenbach, R. Poprawe, International Materials Reviews 57 (2012)
133.
[10] X. Zhao, J. Chen, X. Lin, W. Huang, Materials Science and Engineering A, 478 (2008) 119.
[11] ASM Specialty Handbook: Nickel, Cobalt, and Their Alloys, ASM International (2000).
[12] T.M Pollock, S. Tin, Journal of Propulsion and Power, 22 (2006) 36.
[13] LL. Parimi, GA. Ravi, D. Clark, MM. Attallah, Materials Characterization 89 (2014) 102.
[14] P. Tozzi, P. Stella, I. Giovannetti, M. Buccion, International Conference on AM & 3D Printing,
Nottingham (GB), July 2014.
[15] Z. Wang, K. Guan, M. Gao, X. Li, X. Chen, X. Zeng, Journal of Alloys and Compounds 513
(2012) 518.
[16] O. Jia, D. Gu, Journal of Alloys and Compounds 585 (2014) 713.
[17] M.G. Ardakani, N.D. Souza, A. Wagner, B.A. Shollock, M. McLean, Superalloys 2000, Ed. T.M.
Pollock, R.D. Kissinger, R.R. Bowman, K.A. Green, M. McLean, S. Olson, and J.J. Schirra TMS,
Warrendale, PA, (2000) 219.
[18] W. Yang, W. Chen, K-M. Chang, Superalloys 2000, Ed. T.M. Pollock, R.D. Kissinger, R.R.
Bowman, K.A. Green, M. Mclean, S. Olson, J.J. Schirm, Warrendale, The Minerals, Metals and
Materials Society (2000) 75.
[19] T. Antonsson, H. Fredriksson, Metallurgical and Materials Transactions B, 36B (2005) 85.
[20] W. Kurz, D.J. Fisher, Fundamentals of solidification, 4th ed., Trans. Tech. Publications Ltd,
Uetikon-Zuerich, Switzerland 1998.
[21] C.H. Radhkrishna, Journal of Materials Science 32 (1997) 1977.
[22] W. Hermann, H.G. Sockel, I. Han, A. Bertram, Superalloys 1996, ed. R.D. Kissinger, D.J. Deye,
D.L. Anton, A.D. Cetel, M.V. Nathal, T. M. Pollock, D.A. Woodford, Warrendale, The Minerals,
Metals and Materials Society (1996) 229.
[23] L. Thijs, K. Kempen, J-P Kruth, J.V. Humbeeck, Acta Materialia 61 (2013) 1809.
[24] T. Niendorf, S. Leuders, A. Riemer, H.A. Richard, T. Tröster, D. Schwarze, Metallurgical and
Materials Transactions B: Process Metallurgy and Materials Processing Science 44, (2013), 794.
[25] B. Vrancken, V. Cain, R. Knutsen, J.V. Humbeeck, Scripta Materialia 87 (2014) 29.
[26] V.L. Acoff, R.G. Thompson, IBM Journal of Research and Development 44 (2000) 668.
[27] B. Radhakrishnan, R.G Thompson, Metallurgical Transactions A 24A (1993) 2773.
[28] R.D. Doherty, D.A. Hughes, F.J. Humphreys, J.J. Jonas., D.J. Jensen, M.E. Kassner, W.E. King,
T.R. McNelley, H.J. McQueen, A.D. Rollett, Materials Science and Engineering A238 (1997)
219.
[29] I. Manna, S.K. Pabi, W. Gust, International Materials Reviews 46 (2001) 53
[30] S. Mahajan C.S. M.A. Pande Imam., B.B. Rath., Acta Materialia 45 (1997) 2633.
[31] Y.V.R.K. Prasad, S. Sasidhara, Hot Working Guide: A Compendium of Processing Maps (e-Book
Google), ASM International, 1997.

24