You are on page 1of 11

JOM, Vol. 71, No.

4, 2019
https://doi.org/10.1007/s11837-019-03367-1
Ó 2019 The Minerals, Metals & Materials Society

ADDITIVE MANUFACTURING OF COMPOSITES AND COMPLEX MATERIALS

Investigation of the Lithium-Containing Aluminum Copper Alloy


(AA2099) for the Laser Powder Bed Fusion Process [L-PBF]:
Effects of Process Parameters on Cracks, Porosity,
and Microhardness

IRIS RAFFEIS,1,4 FRANK ADJEI-KYEREMEH,1,2 UWE VROOMEN,1


PIYADA SUWANPINIJ,2 SIMON EWALD,3
and ANDREAS BÜHRIG-POLAZCEK1

1.—Foundry Institute, RWTH Aachen University, Aachen, Germany. 2.—The Sirindhorn


International Thai-German Graduate School of Engineering (TGGS-KMUTNB), North Bangkok,
Thailand. 3.—Digital Additive Production, RWTH Aachen University, Aachen, Germany.
4.—e-mail: I.Raffeis@gi.rwth-aachen.de

In order to widen the alloy spectrum for the laser powder bed fusion process,
apart from the popular Al-Si-based alloys, AlSi10Mg and AlSi12, the precip-
itation-hardenable AA2099 wrought alloy has been considered in this work.
The effect of varied laser power, scanning speed, a preheated base plate
(in situ heat treatment) and post-heat treatments on porosity, cracks and
microhardness was observed. The results indicate a successfully printed
crack-free part with very minimal porosity at 90 W laser power and 550 mm/s
scanning speed with microhardness of 72 HV0,1 at a 520°C preheat treatment
temperature. The influence of added titanium aluminide powder in crack re-
moval and grain refinement is also reported in this work.

beam.6 L-PBF printed parts have unique


INTRODUCTION
microstructures owing to rapid melting and solidi-
The laser powder bed fusion [L-PBF] process, fication from highly localized heat.7
which is also called selective laser melting,1 is a Due to their low price, good corrosion resistance,
relatively new additive manufacturing process ductility and formability, as well as its relatively
which has in recent times caught the attention of high specific strength, aluminum alloys are the
many researchers in the materials research com- preferred lightweight material of choice for aero-
munity, owing to its advantages over conventional space, aviation and other related industries.8 The
manufacturing processes such as forming or cast- Al-Si alloys have been the most used and the most
ing.2 Increase in lead time as well as high tooling popular of all the aluminum-based alloys manufac-
cost are some of the known challenges of the tured using the L-PBF process, among them being
conventional manufacturing process.3 The L-PBFs the AlSi10Mg and AlSi12 alloys. It has been sug-
potential not only stems from its wide degree of gested that the relative ease in the processing of the
freedom of design but also from its ability to quickly powder of these alloys is due to the minimal
produce parts of complete densification and complex difference in their liquidus and solidus tempera-
geometry with reduced production processes and tures, their good weldability and their material
efficient material consumption.4 The L-PBF process flowing behavior compared to the wrought Al
is also known to have flexibility in feedstock even alloys.9–12
with composite materials of metals, ceramics and However, in order to expand the spectrum of
plastics.5 During the L-PBF process, a three-dimen- aluminum-based alloys for the L-PBF process,
sional (3D) computer aided design model is printed researchers such as Ahuja et al.13 fabricated and
by melting thin powder, layer upon layer, usually characterized wrought aluminum alloys EN
below 150 lm thickness with the help of a laser AW2219 (AlCu6Mn) and EN AW2618

(Published online February 12, 2019) 1543


1544 Raffeis, Adjei-Kyeremeh, Vroomen, Suwanpinij, Ewald, and Bührig-Polazcek

(AlCu2Mg1.5Ni), and concluded that micro-sized atmosphere. The powder was divided into three: the
powders of these alloys can be formed into fully initial AA2099 base alloy powder (alloy 1), the
dense 3D structures using the L-PBF process. Karg second AA2099 powder (alloy 2), with 99.8% alu-
et al.14 also studied the process conditions on the minum powder added (mixing ratio of alloy 1 to
mechanical behavior of EN AW2219 using L-PBF aluminum powder: 4:1), and the third AA2099
and concluded that it was possible to produce powder (alloy 3), with nitrogen-containing titanium
vertical tensile specimen from EN AW2219 without aluminide powder (Ti48Al48Cr2Nb2) added, (mix-
cracks and with low porosity. Zhang et al.12 inves- ing ratio of alloy 1 to titanium aluminide powder:
tigated the processing and microstructure proper- 99.2: 0.8) (Table I).
ties of an Al-Cu-Mg alloy using the L-PBF process Particle size distribution was measured on the
and reported that 99.8% relative density, micro- AA2099 powder and was found to have the largest
hardness of 111 HV0,1, UTS (ultimate tensile particle size range between 10 lm and 62 lm with
strength) of 402 MPa and a yield stress of an average particle diameter of 35.9 lm. Particle
276 MPa were achieved. shape analysis was carried out and average particle
The processability potential of EN AW7075 using roundness and convexity of 0.935 and 0.995 were
L-PBF has also been investigated.15,16 They found recorded, respectively. The morphology of the gas-
that Si addition produced parts without cracks and atomized powder and the aluminide powder is
that it was possible to achieve a relative density of shown in Fig. 1a and b, respectively. The AA2099
99% with a laser power of more than 300 W. The alloy powder used in this work was oven-dried for
AA6061 wrought aluminum alloy was also investi- 90°C for 24 h before the L-PBF process.
gated using the L-PBF process,17 which confirmed
that the laser beam profile has an effect on the Characterization
width of the melt.
Elemental composition of the argon-atomized
The third generation AA2099 Al-Cu-Li alloy pro-
AA2099 powder was performed using inductively
mises to be a good candidate for the L-PBF process
coupled plasma optical emission spectroscopy. Gas
due to the reduction of the density of the alloy by 3%
atomization was carried out using the patented
and the increase in elastic modulus by 5% when each
NANOVAL process by the Nanoval Company. The
1 wt.% of Li is added, and also for its improved fatigue
Morphologi G3 instrument was used to perform the
and corrosion resistance over the conventional 2xxx
particle size distribution and shape analysis of the
series.18 The AA2099 alloy has been successfully
AA2099 powder. Microstructure characterization
investigated for its thermal stability and specific
was performed using a Zeiss Light Optical Micro-
strength at high temperatures.19,20 Zhang et al.21
scope and a Zeiss Gemini Field Emission Gun
mentioned in their work that it exhibits superior
Scanning Electron Microscope coupled with Oxford
stress corrosion cracking resistance, high transverse
Instruments Inca X-sight EDS detector. The trans-
ductility, and excellent cryogenic properties with
mission electron microscope, Zeiss Libra 200FE,
fewer planar anisotropy when compared to the
was used for the analysis of the titanium nitride in
AA2090. Others have focused on microstructure
the titanium-aluminide powder. The field-emission
investigations and precipitate characterization.19,22
electron microprobe analyzer (EPMA), JEOL JXA-
In spite of all this work carried out on the AA2099
8530F, was used to detect the titanium phases. All
alloy, to the best of our knowledge, no work has been
etched samples were carried out with 0.5%
carried out thus far in investigating it for the L-PBF
hydrofluoric acid for 15 s. Using IMAGEJ software,
process.
the area fraction of the pores and the circularity of
the pores were determined. The microhardness of
METHODS
the samples were determined by taking average
Materials measurements at three different positions, top,
middle and bottom, using a Buehler Microhardness
The AA2099 alloy powder was obtained from a
instrument at 0.1 N load [HV0,1].
pulverized ingot by gas atomization under argon

Table I. Composition of the base alloy (alloy 1), with added Al powder (alloy 2), with added titanium
aluminide powder (alloy 3), and the nominal composition of the titanium aluminide powder

(wt.%) Al Cu Li Zn Mg Mn Zr Si Fe Ti Nb Cr
Initial AA2099 alloy (alloy 1) Bal 2.5 1.49 0.57 0.25 0.15 0.08 0.03 0.08 < 0.001 < 0.001 < 0.001
With Al powder added (alloy 2) Bal 2.11 1.03 0.28 0.17 0.13 0.064 0.29 0.11 0.02 0.03 < 0.001
With titanium aluminide powder Bal 2.81 1.35 0.37 0.20 0.17 0.085 0.05 0.10 0.46 0.04 0.021
added (alloy 3)
Titanium aluminide powder 33.65 app. 800 ppm N 58.83 4.56 2.82
Investigation of the Lithium-Containing Aluminum Copper Alloy (AA2099) for the Laser Powder 1545
Bed Fusion Process [L-PBF]: Effects of Process Parameters on Cracks, Porosity, and Microhardness

Fig. 1. (a) Morphology of the initial gas-atomized AA2099 powder. (b) Morpology of the titanium aluminide powder (inset: STEM-ADF image of
titanium nitride in the powder).

The Aconity3D (Aconity MINI lab unit) Laser The overall goal was to achieve crack- and porosity-
Powder Bed Fusion Equipment was used for the free L-PBF-material with a microhardness of about
printing of all the samples. The size of the samples 70 HV0,1, which was achieved in the first print set
was 10 9 10 9 10 mm3 and 5 9 5 9 10 mm3. The without a heating plate and was so far the maxi-
scanning strategy was an alternating bidirectional mum microhardness observed.
X/Y scanning strategy, which turned about at 90° on For the samples of the first print set carried out
each layer. without a heating plate (in situ heat treatment),
post-heat treatments (PHT) were applied. With the
Design of L-PBF Experiments help of the ThermoCalc equilibrium and the Scheil–
Gulliver diagrams, heat treatment temperatures for
In the first print set, L-PBF samples that were all the print sets were chosen with the goal of
printed without base plate heating showed many finding the most effective precipitates in terms of
cracks. This precipitation-hardening aluminum microhardness (Fig. 2).
alloy contains elements whose solid solubility
decreases with falling aluminum temperature. The L-PBF Process
Rapid cooling leads to a supersaturated solid solu-
tion and results in large residual stress in the as- The first set of samples was printed from alloy 1
printed part,23 which is often relieved through without a heating plate. To improve the microhard-
fracturing when the UTS of the material is locally ness through precipitation strengthening, PHT was
exceeded. Since cracks occured in these samples, carried out at 6 different PHT temperatures: 80°C,
three different paths were followed in order to 220°C, 320°C, 410°C, 450°C and 520°C. The Ther-
reduce cracks and to get an insight into the mocalc simulation at equilibrium (phase diagram in
characteristic melting and solidification behavior Fig. 2a) shows the possible precipitates with their
of this alloy during the L-PBF process. thermal stability, and highlights more than one
type of expected co-existing second phase particles
(a) The cooling rates were reduced through in situ at the chosen temperatures. The main expected
heat treatments with the help of a heated base phases of this multiple phase alloy are as follows:
plate (in situ heat treatment) to reduce the
oversaturation in the samples and to reduce 80°C: AlCuLi (T2), AlLi, Al6Mn
the residual stress that causes the material to 220°C: AlCuLi (T2), Al6Mn
crack. Temperatures between 220°C and 320°C: AlCuLi (T2), Al2CuLi (T1), Al6Mn
520°C were chosen for the heated base plate. 410°C: Al2CuLi (T1), Al6Mn
(b) The copper content was reduced through the 450°C: Al2CuLi (T1), Al6Mn
addition of aluminum powder to the base alloy 520°C: Al6Mn
1 in order to reduce the residual stress as well
as to minimize the copper segregation in the Al2CuLi (T1), which is hexagonal in crystal struc-
melt lines and at the grain boundaries to ture and has a plate-like morphology, is known to be
reduce cracking (alloy 2). the major contributor to the strengthening of the
(c) The possible effect of grain refinement and alloy.24 Other strengthening phases, such as icosa-
microhardness was investigated through add- hedral T2 plates and the tetragonal AlLi phase, both
ing titanium aluminide powder as a heteroge- precipitate at high-angle grain boundaries.25 T2 has
nous nucleant to lower the necessary also reportedly been used for controlling toughness
nucleation energy for grain formation (alloy 3). in the alloy.26 Dispersoids such as orthorhombic
1546 Raffeis, Adjei-Kyeremeh, Vroomen, Suwanpinij, Ewald, and Bührig-Polazcek

Fig. 2. ThermoCalc Simulation: (a) equilibrium phase diagram simulation of the AA2099 alloy composition. (b) Scheil–Gulliver phase diagram
showing the metastable phases present.

Al6Mn are known for inhibiting recrystallization completely dissolved. Again, at least two differently
and also for grain size and texture control.27 They shaped precipitates, sticklike and partly dissolved,
are normally too small to be resolved with light are seen in the SEM micrograph (Fig. 3f).
optical micrographs (LOM). The highest microhardness of 70.7 HV0,1 recorded at
The next four printsets, consisting of 16 samples 80°C PHT temperature constantly decreased with
each, underwent in situ heat treatments at three increasing PHT temperature to 58.8 HV0,1 at 520°C.
different temperatures. In all sets, the layer thick- The decreasing microhardness is believed to be a
ness of 30 lm and the hatching distance of 150 lm result of the growing precipitates and the dissolution
were constant while the scanning speed and the of some of them. Since the kind, size, distribution and
laser power were varied. amount of the precipitates did not improve the micro-
Alloy 1 underwent in situ heat treatment at 220°C hardness, and the goal of improving the hardness
and at 520°C. The temperature of 520°C was chosen beyond 70 HV0,1 was not met, no further investiga-
from the Scheil–Gulliver diagram (Fig. 2b) as the tions were carried out to identify the precipitates.
start temperature of the T1 phase formation.
Alloys 2 and 3 underwent in situ heat treatments
Cracks and Porosity of In Situ Heat-Treated L-
at 320°C. In alloy 2, aluminum powder was added to
PBF Samples
the initial AA2099 base powder (alloy 1). Titanium
aluminide powder was added to alloy 3, as shown in Several reasons can be attributed to the formation
Table I. of hot or cold cracks in L-PBF printed parts. There
Correlations with the microhardness of all the is more than one theory for the cracking discussed
sets of the samples were investigated. in the literature.23,28–31 All of the described mech-
anisms can lead to stress relief through fracturing
RESULTS AND DISCUSSION when the tensile strength of the material exceeds its
local UTS.30 For the prevention of cold cracking,
Microhardness of Post Heat Treated (PHT)
preheated base plates at various temperatures were
L-PBF Samples
used in order to reduce the thermal gradient, the
For exemplary temperatures, LOM and scanning cooling rate and oversaturation.
electron microscope (SEM) images of the 80°C, Low melting phases which form films on grain
450°C and 520°C PHT samples are compared in boundaries and meltlines can lead to hot cracks.32
Fig. 3. For the 80°C PHT sample, cracks, meltlines The chemical composition of the alloy was altered in
and precipitates are shown in Fig. 3a. In the 450°C order to reduce the segregation of the low melting
PHT sample (Fig. 3b and e), two different kind of phases (see ‘‘Design of L-PBF Experiments’’ sec-
coarser precipitates are seen at the grain bound- tion). One method of hot crack prevention which has
aries (GB) and within the grains. In the 520°C PHT been extensively investigated is grain refinement.
sample (Fig. 3c), they seem to have almost Tang31 gives an overview of physical theories. With
Investigation of the Lithium-Containing Aluminum Copper Alloy (AA2099) for the Laser Powder 1547
Bed Fusion Process [L-PBF]: Effects of Process Parameters on Cracks, Porosity, and Microhardness

Fig. 3. LOM micrographs: (a) 80°C PHT precipitates not only at melt lines; (b) 450°C PHT sample showing coarsened precipitates at GB; (c)
520°C PHT sample showing most precipitates have gone into solution; SEM micrographs of (d) 80°C PHT sample showing precipitate growth
which also recorded the highest microhardness; (e) coarsened precipitates in the 450°C PHT sample led to a decline in the microhardness; and
(f) 520°C PHT sample indicating that most precipitates had gone into solution and recording the lowest microhardness.

decreasing grain sizes, the grain boundary count In Situ Heat-Treated L-PBF Samples, Laser
per volume increases. The overall strain can be Parameters
distributed on many grain boundaries and
As shown in Fig. 4, samples from alloy 1 in situ
decreases the possibility of cracking.33 Also, crack
heat-treated at 220°C had both cracks and small
growth will be hindered through a curved crack
spherical gas-induced pores. The cracks are believed
path.
to be hot cracks, as high Cu content was detected at
Porosity formation in the L-PBF process is gen-
the crack locations. Both the gas-induced pores and
erally attributed to several reasons. The small
the hot cracks contributed to making the 220°C
spherical pores may be gas-induced emanating from
samples less dense. Using the image analysis soft-
gas entrapment during melting and rapid solidifi-
ware, IMAGEJ, the least recorded percentage
cation, or from evaporation of the materials/selected
porosity among the 220°C in situ heat-treated
elements during the L-PBF process and carried over
samples was 8.3%, which was printed at a laser
to the as-printed part.34,35 Larger spherical pores
power of 190 W using 550 mm/s scanning speed.
originate from the instability of the molten pool
The highest percentage porosity of 15.8% was
owing to keyhole collapse, while irregular porosity
recorded in the 220°C sample printed using a
usually stems from unmelted powder or when the
scanning speed of 650 mm/s and a laser power of
overlap between the scan tracks is not sufficient.35
190 W.
The unmelted powder from incomplete melting or
When alloy 1 was in situ heat-treated at 520°C,
lack of fusion occurs when there is insufficient
along with varied laser power and scanning speed,
energy from the heat source to melt the powder.36
no cracks were observed and a minimum porosity of
From the volumetric energy density (Ev) equa-
1.2% was achieved. The laser power was also lower
tion, with constant parameters such as the layer
compared to the 220°C in situ heat-treated samples.
thickness (Ds) and the hatching distance (Dys), the
This means that the volumetric energy density (Ev)
interplay between the laser power (pL) and the
was higher during the printing of the 220°C samples
scanning speed (vs) must be adjusted in a suit-
than in the 520°C samples. In Fig. 5a, it was
able manner to achieve the optimum laser power to
observed that, as the intensity of the laser power
minimize pores and lack of fusion:
increased from 110 W and upwards, the pore sizes
pL became spherical and larger. These larger spherical
Ev ¼ ðJ/mm3 Þ ð1Þ
Ds  vs  Dys pores are keyholes resulting from the collapse of an
1548 Raffeis, Adjei-Kyeremeh, Vroomen, Suwanpinij, Ewald, and Bührig-Polazcek

Fig. 4. Alloy 1: 220°C in situ heat-treated samples of alloy 1 showing cracks and pores in all the samples.

unstable melt pool due to high linear energy density speed and 90 W laser power, as seen in Fig. 5a. It
(LED), which means that the input power per unit was possible to achieve minimal porosity and no
speed increases with increasing laser power, result- cracks in the 520°C in situ heat-treated sample. The
ing in an increase in the melt pool.35 following explanation seems plausible: at a low
Irregular pores owing to incomplete melting were scanning speed (550 mm/s), the energy density was
observed from a scanning speed of 550–850 mm/s at sufficient to cause an increase in the temperature of
70 W laser power, which usually happens when the melt pool. This led to a resultant increase in
there is insufficient laser power to cause complete wettability, reduced viscosity and surface tension,
melting. This gives an indication that, even if the which aided the liquid metal to have adequate
scanning speed were to increase up to 850 mm/s, fluidity to bond between the layers and also for the
with a laser power at 70 W, there will still not be liquid phase to flow between the spaces as well as to
sufficient energy for complete melting. back-fill shrinkages.12 In Fig. 5b, the SEM micro-
The dense as-fabricated part with percentage graph shows the uniform distribution of plate-like
porosity of 1.2% was achieved at 550 mm/s scanning precipitates within the matrix and copper-rich
Investigation of the Lithium-Containing Aluminum Copper Alloy (AA2099) for the Laser Powder 1549
Bed Fusion Process [L-PBF]: Effects of Process Parameters on Cracks, Porosity, and Microhardness

Fig. 5. (a) LOM micrographs of in situ heat-treated samples of alloy 1 at 520°C; least porosity at 550 mm/s scanning speed at 90 W laser power.
(b) SEM micrograph of the sample with least porosity. (c) EBSD image: grain structure and orientation; average grain size in equivalent diameter
of 25 lm.

phases at the grain boundaries. The copper content count analysis indicated an average grain size in
was determined by EDS analysis with values the equivalent circle diameter of 25 lm.
between 13 wt.% and 20 wt.% Cu. From the elec- The 320°C in situ heat-treated samples were
tron backscattered diffraction (EBSD) image of the printed from both alloy 2 and alloy 3 with 16
grain structure and orientation (Fig. 5c), grain different laser parameters. As indicated earlier, the
1550 Raffeis, Adjei-Kyeremeh, Vroomen, Suwanpinij, Ewald, and Bührig-Polazcek

significant difference between the two 320°C alloy refinement and increasing the hardness. As in the
compositions was the fact that alloy 2 had added Al 220°C samples, the alloy 2 in situ heat-treated
powder for the purposes of minimizing Cu segrega- samples showed cracks and pores (Fig. 6a). The
tion, while alloy 3 was modified with titanium least percentage porosity recorded was 11.3% which
aluminide powder for the purposes of grain was printed at 550 mm/s scanning speed and 125 W

Fig. 6. (a) LOM micrographs: 320°C in situ heat-treated samples from alloy 2. (b) LOM micrographs: 320°C in situ heat-treated samples from
alloy 3. (c) SEM micrograph: 320°C in situ heat-treated sample from alloy 2. (d) SEM micrograph: 320°C in situ heat-treated sample from alloy 3
showing a refined grain structure sample. (e) EBSD image: 320°C alloy 2 sample with average grain size of 256.81 lm2 and ECD of 18.08 lm. (f)
EBSD image: 320°C alloy 3 sample showing a refined and equiaxed grain structure compared to alloy 2 sample with average grain size of
1.49 lm2 and ECD of 1.28 lm. EPMA analysis images: (g) aluminum matrix with TiAl3 particle; (h) Ti detection in the TiAl3 particle.
Investigation of the Lithium-Containing Aluminum Copper Alloy (AA2099) for the Laser Powder 1551
Bed Fusion Process [L-PBF]: Effects of Process Parameters on Cracks, Porosity, and Microhardness

laser power, while the highest recorded percentage conventionally processed Al-Cu-Li-alloys.26,38,39 In
porosity was 22.5%, printed at 950 mm/s scanning conventionally processed material, the nucleation is
speed and 200 W laser power. In contrast, main- difficult because of the high stacking fault energy of
taining the same L-PBF parameters as well as the the aluminium matrix, so that T1 nucleates via
320°C in situ heat-treatment temperature, samples stacking fault mechanism at dissociated disloca-
printed from alloy 3 were crack-free and with less tions.38 Therefore, the plate-like T1 phase can only
porosity (Fig. 6b). The least and the highest per- grow in 5-atom layers, which leads to the stick-like
centage porosity, which were recorded was 2.3% and appearance.40 Gao et al.41 summarized three differ-
14.03%, were printed at 450 mm/s and 750 mm/s ent theories of nucleation mechanisms for the T1
scanning speeds with 125 W and 200 W laser phase. In addition, depending on the alloy composi-
power, respectively. A closer look at the grain tion, other phases, like the metastable phase Al2Cu
structure of the two samples from both 320°C alloys or Al3Li, can precipitate42 and compete with the T1
using SEM and EBSD shows a completely different phase. The effects of T1 precipitation-promoting
grain structure (Fig. 6). An average grain size (GS) elements and the Cu/Li quotient are well described
and average grain size in equivalent circle diameter by Decreus et al.43 The actual Cu/Li-quotient of 1.7,
(ECD) of 256.81 lm2 and 18.08 lm, respectively, the analysis of the ThermoCalc simulations as an
were recorded in the 320°C alloy 2 sample while orientation, the Scheil–Gulliver-diagram, the evalu-
1.49 lm2 (GS) and 1.38 lm (ECD) were recorded in ation of a ternary equilibrium diagram,44 and the
the 320°C alloy 3 sample. appearance of the needle shaped precipitates, all lead
The choice of the 320°C in situ heat-treatment to the expected main T1 phase in the material. The
temperature and the lowering of the copper content formation temperature of the T1 phase in a ternary
was not sufficient to eliminate cracks in alloy 2. It can system is given by > 635°C, but the exact tempera-
be concluded that the titanium aluminide powder ture is not known.44 In this investigated technical
added to alloy 3 was responsible for the grain size alloy, the ThermoCalc Scheil-diagram, the starting
reduction as well as its transformation into an temperature is about 520°C. Once the T1 phase is
equiaxed grain structure, as highlighted by the formed, it grows rapidly to very large dimensions
EBSD in Fig. 6f. The refined grains contributed to which is explained by a low lattice mismatch of the
the elimination of hot cracks, as grain refinement has interface layer with the {111}Al planes.43
been reported to decrease the hot crack tendency.37 The 520°C in situ heat treatment was more
Films on the grain boundaries (Fig. 6) are copper- effective than the 220°C in situ heat treatment with
rich phases. The smaller grains offered a higher respect to nucleating and growing the T1 phase, as
volume of grain boundary surface so that effective seen by the size, amount and distribution of this
distribution of the copper segregations was possible phase. The dislocation density of L-PBF-parts lead
and oversaturation by alloying elements was to good conditions for nucleating the T1 phase.
reduced, which led to stress reduction inside the Among the 320°C in situ heat-treated samples
grains. Also, complete stress was distributed on printed from the modified alloys, alloy 2 showed the
more grain boundaries, so that the probability of lowest microhardness, of 33.4 HV0,1. The 320°C
cracking decreased.31 in situ heat-treated alloy 3 samples had appreciable
The titanium aluminide powder had a grain microhardness of 58.4 HV0,1; higher than both the
refining effect, although EPMA investigations 220°C and the 320°C alloy 2 printed samples but
showed that TiN-particles did not play a major role lower than the 520°C in situ heat-treated samples.
in the grain-altering effect and the process of Figure 7 shows the compared microhardness graphs
heterogenous nucleation, because very few TiN- of both PHT samples and all the preheat-treated
particles were found in the matrix. Figure 6g and h samples, indicating an increase in microhardness
shows an Al3Ti-particle with 76.7 at.% aluminum from the 220°C in situ heat-treated samples to the
and 22.5 at.% titanium. Detailed investigations will highest recorded microhardness with the 520°C
follow. in situ heat-treated samples. The microhardness
increased from 53.2 HV0,1 at 220°C to 58.4 HV0,1
Microhardness of In Situ Heat-Treated (Pre- with the alloy 3 samples at 320°C, although, at the
heated) Samples same 320°C in situ heat-treatment temperature, the
microhardness declined to 33.4 HV0,1 for the alloy 2
The microhardness recorded in the 220°C in situ
printed samples, while 78.2 HV0,1 was the highest
heat-treated samples was 53.2 HV0,1. Compared to
microhardness recorded at the 520°C in situ heat
the 520°C in situ heat-treated samples, the hard-
temperature (Fig. 7).
ness was lower. From the 520°C in situ heat-treated
As was expected, the decline in the microhardness
samples, the almost completely dense sample
in the 320°C in situ heat-treated alloy 2 samples
printed at 550 mm/s and 90 W laser power, with
was due to the reduced copper content in the matrix
no cracks and the least porosity, recorded a micro-
which led to reduced concentrations of solute atoms
hardness of 72 HV0,1.
to cause solid solution strengthening or to form
The equilibrium phase T1 (Al2CuLi) is responsible
precipitates for precipitation strengthening. As a
for strength and hardness in most pre-deformed
result of the increased in situ heat-treatment
1552 Raffeis, Adjei-Kyeremeh, Vroomen, Suwanpinij, Ewald, and Bührig-Polazcek

Fig. 7. Microhardness graphs of PHT and in situ heat treated samples compared: PHT sample microhardness decreasing with increasing heat-
treatment temperature. In situ heat-treated sample microhardness increased with increasing heat-treatment temperature except for alloy 2
samples which recorded very low microhardness.

temperature of 520°C, uniformly distributed and 3. chemical adjustments, which were chosen with
effective precipitates led to the increased micro- the goal to lower the crack susceptibility and the
hardness. This heat-treatment temperature caused in situ heat- treatment temperature without loss
good conditions for the precipitation kinetics for the of hardness.
nucleation and growth of the T1 phase, even without
prior deformation. In the in situ heat-treated sam-
To obtain information about temperature ranges
ples at 220°C and 320°C, fewer and smaller needles
of anticipated precipitates, an equilibrium phase
were found, which contributed less to the hardness.
diagram was simulated for this particular alloy.
The heat-treatment temperatures were chosen
CONCLUSION
based on the equilibrium diagram, knowing very
The development of this alloy, which is a new well that this could only be a slight clue, because
application for additive manufacturing, took place of the non-equilibrium conditions. For two differ-
under several different aspects. From the beginning, ent in situ temperatures (220°C and 520°C), the
it was clear that this multi-precipitation system is a L-PBF-parameters were investigated and con-
challenge to work with in the field of the generative trolled in order to produce crack-free samples
manufacturing, because, until now, only traditional with low porosity. From the good results of the
manufacturing processes were performed on this 520°C in situ treatment, two major questions were
kind of alloy and no other investigation results are derived:
known. After having ensured that the production of
1. Is it possible to reduce the copper content in
this lithium-containing aluminum alloy in an induc-
order to reduce cracking without loss of hard-
tion oven and the powder production was possible
ness?
and also led to satisfactory results, it was necessary
2. If this is not the case, is it possible to reduce the
to obtain a broad overview about the properties,
in situ temperature and keep the hardness
limitations and opportunities of this alloy.
values with the help of chemical adjustments?
Samples which were produced without in situ
heat treatments showed copper-rich phases on the
meltlines, almost no precipitates in the grains, and The results so far achieved in this work are as
cracks in the matrix. Since cracks can be reduced follows:
through lowering the cooling speed, through grain The alloy AA2099 has successfully been pre-
refinement, and through adjustment of the chemical sented as an alternative alloy for the L-PBF
composition, or a combination of all of them, the process. After varying the laser parameters while
object of investigations was to study the effects of maintaining a constant layer thickness of 30 lm
and a hatching distance of 150 lm, a crack-free
1. post-heat treatment temperatures on hardness dense L-PBF part was produced with minimal
values porosity at 520°C in situ heat-treatment temper-
2. in situ heat-treatment temperatures on poros- ature using a scanning speed of 550 mm/s and
ity, cracks and hardness 90 W laser power.
Investigation of the Lithium-Containing Aluminum Copper Alloy (AA2099) for the Laser Powder 1553
Bed Fusion Process [L-PBF]: Effects of Process Parameters on Cracks, Porosity, and Microhardness

In situ heat treatment at 520°C yielded parts with 11. B. Konrad, S. Ullrich, T. Frick, and M. Schmidt, Phys. Proc.
higher microhardness values compared to those 12A, 393 (2011).
12. H. Zhang, H. Zhu, T. Qi, Z. Hu, and X. Zeng, Mater. Sci.
without preheat treatment and post-heat treat- Eng. A 656, 47 (2016).
ment. The highest microhardness recorded with 13. B. Ahuja, M. Karg, K. Yu, Nagulin, and M. Schmidt, Phys.
the preheat-treated samples was 72.8 HV0,1 at Procedia 56, 135 (2014).
520°C in situ heat-treatment temperature whereas 14. M.C.H. Karg, B. Ahuja, S. Wiesenmeyer, S.V. Kuryntsev,
and M. Schmidt, Micromachines 8, 23 (2017).
that of the PHT samples was 70.7 HV0,1 at 80°C 15. M.L. Montero Sistiaga, R. Mertens, B. Vrancken, X. Wang,
post-heat-treated temperature. This means that B.V. Hoorweder, J.-P. Kruth, and J.V. Humbeeck, J. Mater.
using the L-PBF process with in situ heat treatment Process. Technol. 238, 437 (2016).
of this alloy at a certain temperature, crack-free 16. N. Kaufman, M. Imran, T.M. Wischeropp, C. Emmelmann,
samples with a good microhardess can be achieved S. Siddique, and F. Walther, Phys. Procedia 83, 918 (2016).
17. L.E. Loh, Z.H. Lui, D.G. Zhang, M. Mapar, S.L. Sing, and
without the need of a post-heat treatment. C.K. Chua, Virtual Phys. Prototyp. 9, 11 (2014).
Reduction of copper as major phase forming ele- 18. Y. Wu, X. Wang, J. Li, and Z. Feng, Mater. Sci. Forum. 850,
ment could not prevent cracking and reduced the 575 (2016).
hardness. 19. E. Balducci, L. Ceschini, S. Messieri, S. Wenner, and
R. Holmestad, Mater. Des. 119, 54 (2017).
Grain refinement which resulted in crack elimi- 20. E. Balducci, L. Ceschini, and S. Messieri, JOM (2018).
nation by the addition of titanium-aluminide pow- https://doi.org/10.1007/s11837-018-3006-x.
der has also been successfully presented. However, 21. F. Zhang, J. Shen, X.D. Yan, J.L. Sun, X.L. Sun, and Y.
further adjustment of the process parameters and Yang, Rare Met. 33, 28 (2014).
in situ heat treatment is still needed to eliminate 22. S. Mishraa, V. Kumar Beura, A. Singha, M. Yadava, and N.
Nayan, JOM (2018). https://doi.org/10.1080/02670836.2018.
porosity and also boost hardness. 1510074.
The post-heat-treated samples showed a loss of 23. B. Zhang, Y. Li, and Q. Bai, Chin. J. Mech. Eng. 30, 515 (2017).
microhardness with increasing temperatures. 24. V. Araullo-Peters, B. Gault, F. De Geuser, A. Deschamps,
Further work is still needed to optimize the in situ and J.M. Cairney, Acta Mater. 66, 208 (2014).
25. K.S. Prasad, A.A. Gokhale, A.K. Mukhopadhyay, D. Ban-
preheat-treatment temperature, the AA2099 pow- erjee, and D.B. Goel, Mater. Sci. Forum 1048, 337 (2000).
der and its chemical composition, and the L-PBF 26. R.J. Rioja and J. Lui, Metall. Mater. Trans. A 43A, 3337
parameters. (2012).
The results presented here indicate that this work 27. D. Tsivoulas and P.B. Prangnell, Acta Mater. 77, 16 (2014).
is far from complete; however, a first step has been 28. N.J. Harrison, I. Todd, and K. Mumtaz, Acta Mater. 94, 68
(2015).
taken. This alloying system with its multiple sec- 29. P. Mercelis and J.P. Kruth, Rapid Prototyp. J. 12, 265
ondary phases of high thermal stability offers many (2006).
opportunities to optimize the effect of particle 30. M. Zhong, H. Sun, W. Liu, X. Zhu, and J. He, Scr. Mater. 53,
strengthening. Detailed investigations will follow. 164 (2005).
31. Z. Tang, Heißrissvermeidung beim Schweißen von Alu-
ACKNOWLEDGEMENTS miniumlegierungen mit einem Scheibenlaser, Band 53
(Bremen: Strahltechnik BIAS, 2014), pp. 4–15.
The authors thank ACCESS e.V. for their support 32. G. Schulze, Metallugie des Schweißens, 3rd ed. (New York:
in the analytics. The authors would like to thank the Springer), pp. 276–277, 296–299.
German Research Foundation DFG for the kind 33. J.A. Spittle and A.A. Cushway, Met. Technol. 6, 13 (1983).
34. W.J. Sames, F.A. List, S. Pannala, R.R. Dehoff, and S.S.
support within the Cluster of Excellence ‘‘Integrative Babu, Int. Mater. Rev. 61, 315 (2016).
Production Technology for High-Wage Countries’’. 35. K. Kempen, L. Thijs, J. Van Humbeeck, and J.-P. Kruth,
Mater. Sci. Technol. 31, 917 (2015).
REFERENCES 36. M. Brandt, S. Sun, M. Leary, S. Feih, J. Elambasseril, and
1. J. Trapp, A.M. Rubinchik, G. Guss, and M.J. Mathews, Q. Liu, Adv. Mater. Res. 633, 147 (2013).
Mater. Today 9, 341 (2017). 37. M.A. Easton, M. Qian, A. Prasad, and D.H. StJohn, Curr.
2. J. Suryawanshi, K.G. Prashanth, S. Scudino, J. Ekert, O. Opin. Solid State Mater. Sci. 20, 13 (2015).
Prakash, and U. Ramamurty, Acta Mater. 115, 285 (2016). 38. E. Gumbmann, W. Lefebvre, F. De Geuser, and C. Sigli,
3. L. Ruidi, M. Wang, T. Yuan, B. Song, C. Chen, K. Zhou, and Acta Mater. 115, 114 (2016).
P. Cao, Power Technol. 319, 117 (2017). 39. C. Wang, G. Min, Q. Lu, and Z. Lu, Int. J. Cast Met. Res. 17,
4. X. Mujian, D. Gu, G. Yu, D. Dai, H. Chen, and Q. Shi, Int. J. 266 (2014).
Mach. Tool Manuf. 109, 147 (2016). 40. W. Cassada, G. Shiflet, and E. Starke Jr, J. Phys. C3, 397
5. P.G. Prashanth, S. Scudino, H.J. Klauss, K.B. Surreddi, L. (1987).
Lobber, Z. Wang, A.K. Chaubey, U. Kuhn, and J. Eckert, 41. Z. Gao, J.Z. Liu, J.H. Chen, S.Y. Duan, Z.R. Lui, W.Q. Ming,
Mater. Sci. Eng. A 590, 153 (2014). and C.L. Wu, J. Alloys Compd. 624, 26 (2015).
6. W. Christian, D. Buchbinder, N. Pirch, W. Meiners, K. 42. J. Huang and A. Ardell, Acta Mater. 115, 114 (2016).
Wissenbach, and R. Poprawe, J. Mater. Process. Technol. 43. B. Decreus, A. Deschamps, F. De Geuser, P. Donnadieu, C.
221, 112 (2015). Sigli, and M. Weyland, Acta Mater. 61, 2218 (2013).
7. L. Zheng, Y. Lui, S. Sun, and H. Zhang, Chin. J. Aeronaut. 44. A.B. Nikolay, G.E. Dmitry, and A.A. Aksenkov, Multicom-
28, 564 (2015). ponent Phase Diagram, Applications for Commercial Alu-
8. L. Zhu, N. Li, and P.R.N. Childs, J. Propul. Power 7, 103 minium Alloys, 1st ed. (Moscow: Elsevier, 2005), pp. 257–260.
(2018).
9. E.R. Christopher, D. Bourell, T. Watt, and J. Cohen, Phys. Publisher’s Note Springer Nature remains neutral with re-
Procedia 83, 909 (2016). gard to jurisdictional claims in published maps and institutional
10. D. Koutny, D. Palousek, L. Pantelejev, C. Hoeller, R. Pich- affiliations.
ler, L. Tesicky, and J. Kaiser, Materials 11, 298 (2018).

You might also like