Investigation into spatter behavior during selective laser melting of AISI 316L stainless steel powder

Abstract

During the process of selective laser melting (SLM), spatter is generated with a negative impact on the performance of parts. Two types of spatter have been identified: droplet spatter, produced by the tearing of molten metal and powder spatter, formed when non-molten metallic powder particles around the molten pool are blown away, both arising as a result of the impact of metallic vapor. Single-track experiments were performed in order to observe spatter behaviors by using a high-speed camera. The influence of energy input on the spatter behavior was investigated by employing 316L stainless steel powder. Results indicate that energy input affects the size, scattering state and jetting height of spatter. Energy dispersive spectroscope analyses show that oxygen contents increase in spatter and SLM parts. X-ray diffraction analyses show that diffraction peaks of austenite and ferrite are considerably lower than those in 316L powder owing to the generation of iron oxides (Fe. +. 2Fe2. +. 3O4). Comparative tensile testing results show that although both groups of specimens manufactured with fresh and contaminated powders are mainly characterized by ductile fracture, the tensile properties of the latter are far inferior to those of the former, owing to a greater quantity of inclusions.

Full text

Investigation into spatter behavior during selective laser melting of AISI
316L stainless steel powder
Yang Liu
a
, Yongqiang Yang
a,
⁎, Shuzhen Mai
a
, Di Wang
a
, Changhui Song
b
a
School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China
b
Institute of Sports Medicine, Peking University Third Hospital, Beijing 100191, China
abstractarticle info
Article history:
Received 1 June 2015
Received in revised form 17 August 2015
Accepted 18 August 2015
Available online 22 August 2015
Keywords:
Selective laser melting
Spatter behavior
High-speed photography
Stainless steel powder
Tensile properties
Inclusions
During the process of selective laser melting (SLM), spatter is generated with a negative impact on the perfor-
mance of parts. Two types of spatter have been identified: droplet spatter, produced by the tearing of molten
metal and powder spatter, formed when non-molten metallic powder particles around the molten pool are
blown away, both arising as a result of the impact of metallic vapor. Single-track experiments were performed
in order to observe spatter behaviors by using a high-speed camera. The influence of energy input on the spatter
behavior was investigated by employing 316L stainless steel powder. Results indicate that energy input affects
the size, scattering state and jetting height of spatter. Energy dispersive spectroscope analyses show that oxygen
contents increase in spatter and SLM parts. X-ray diffraction analyses show that diffraction peaks of austenite
and ferrite are considerably lower than those in 316L powder owing to the generation of iron oxides
(Fe + 2Fe2 + 3O4). Comparative tensile testing results show that although both groups of specimens
manufactured with fresh and contaminated powders are mainly characterized by ductile fracture, the tensile
properties of the latter are far inferior to those of the former, owing to a greater quantity of inclusions.
© 2015 Published by Elsevier Ltd.
1. Introduction
Selective laser melting (SLM), a recently developed additive
manufacturing (AM) technique, is capable of directly fabricating
three-dimensional parts with complex structures [1–4]. During this pro-
cess, a high-energy laser beam is irradiated on the powder layer and en-
ergy is absorbed by powder particles through bulk-coupling and
powder-coupling [5], which generates an extremely high temperature
(up to 10
5
°C) and a rapid cooling rate (up to 10
6
–10
8
°C/s) within the
molten pool [6,7]. The interaction between the laser and powder can
be affected by heat, mass, and momentum transfer, all associated with
the consequential problems of balling, pores and thermal cracks. It is
therefore of paramount importance to understand the interaction
between the laser and powder, both to obtain the desired SLM parts
and to counteract these potential defects.
Considerable researches on theSLM process have been conducted by
using AISI 316L stainless steel [1–3,8–12], which possesses excellent
corrosion resistance and mechanical properties. Wu and Yang [8] inves-
tigated the underlying causes of defects such as warp and balling during
the SLM process, and identified corresponding improvements to avoid
these defects. Li et al. [9] studied the influence of process parameters
on densification behavior during the SLM process by using gas- and
water-atomized 316L stainless steel powder. They found that
gas-atomized powder possesses better densification than that of
water-atomized powder owing to its lower oxygen content and higher
packing density. Wang et al. [10,11] studied the impact of energy
input on scanning tracks and established that regular and thin shaped
tracks are the most suitable for SLM. They also found that these tracks
were capable of producing complex parts with overhanging structures.
Yan et al. [12] studied the manufacturability and performance of
advanced and lightweight cellular lattice structures fabricated via SLM
by using 316L stainless steel.
In addition, a few researchers have investigated spatter behavior in
laser processing. Zhang et al. [13–15] concluded that high-pressure
metallic vapor and plasma squeeze out molten metal from the molten
pool, thereby generating spatter. Schweier et al. [16] and Hugger et al.
[17] conducted quantitative analyses of the influence of laser power,
welding speed and pulse frequency on laser welding spatter and
found that laser power and welding speed have the greatest impact
on the degree of spatter, as opposed to the pulse frequency which has
a relatively small impact. Low et al. [18–20] studied the influence of
laser power and pulse, while focusing on the amount and distribution
of spatter during laser drilling. They developed an anti-spatter compos-
ite coating (ASCC) method to reduce the degree of spatter. Guo et al.
[21] developed a X-ray transmission system to study the spatter
forming mechanism of underwater flux-cored wire wet welding. They
divided the spatter into three types: droplet repelled spatter, explosive
spatter and molten pool shock spatter. Dai et al. [22] proposed a masked
laser ablation method to fabricate a micro-dimple array on a substrate
Materials and Design 87 (2015) 797–806
⁎Corresponding author.
E-mail address: meyqyang@sc ut.edu.cn (Y. Yang).
http://dx.doi.org/10.1016/j.matdes.2015.08.086
0264-1275/© 2015 Published by Elsevier Ltd.
Contents lists available at ScienceDirect
Materials and Design
journal homepage: www.elsevier.com/locate/jmad

surface. They investigated the influence of process parameters on the
spatter deposition between the adjacent micro-dimples.
Current literature however reveals a relative paucity of studies on
spatter during the SLM process. Simonelli et al. [23] studied spatter
and oxidation reactionsduring the SLM process, establishing that chem-
ical compositions in spatter particles change significantly compared to
the initial feedstock. Mumtaz and Hopkinson [24] used a pulse shaping
technique to reduce spatter in addition to improving the surface rough-
ness of SLM parts. Although spatter during the SLM process occurs also
owing to the recoil pressure of metallic vapor [14,25–27],theformation
mechanism is dissimilar to that produced during laser welding anddril-
ling. The materials used in SLM are powder, the thermal conduction
within powder as well as heat and mass transfer within the molten
pool are more complex than those within the correspondingsolid mate-
rial, thereby making it more difficult to study the spatter mechanism
during the SLM process. Furthermore, large spatter particles are mixed
with the powder and become inclusions in the parts during the subse-
quent manufacturing, thus affecting the mechanical properties of the
SLM parts [28]. Therefore a better understanding of the spatter mecha-
nism and behavior in the SLM process is necessary in order to reduce
spatter. This study aims to provide an insight into spatter behavior
and its influences on SLM parts, and thereby offering suggestions on
the choice of effective process parameters.
2. Experimental conditions and methods
2.1. Selective laser melting apparatus and materials
The experiments were carried out on an SLM machine DiMetal-100,
as shown in Fig. 1. A 200 W fiber laser (ytterbium-doped, continuous
mode, TEM
00
profile and M
2
b1.1, wavelength 1090 nm) was used.
The building envelop is 100 × 100 × 120 mm. The scanning system
used was a dual axis mirror positioning system and a galvanometer op-
tical scanner, whichdirect the laser beam in the X- and Y-axes through a
f-theta lens. Focusing optics employed a 163 mm focal length lens,
which produces a focused beam spot sizeof approximately 70 μmindi-
ameter. Since the powder is fully melted during the process, protection
of the SLM-processed parts from oxidation is essential, therefore all
metal powder processing took place in an argon or nitrogen atmosphere
with a maximum of 0.05% O
2
. The main technical parameters of
DiMetal-100 are shown in Table 1.
The material used in this study was gas-atomized 316L stainless
steel spherical powder, as shown in Fig. 2a. Fig. 2b shows the particle
size distributions (wt.%): d
10
= 22.5 μm, d
50
=39.02μmandd
80
=
56.04 μm, the average particle size is 42.83 μm, and the relatively densi-
ty is 4.02 g/cm
3
. The chemical compositions of the powdered material
are shown in Table 2.
2.2. Experimental methods
The experimental setup for the high-speed photography is shown in
Fig. 3. A high-speed camera(AOS SPRI-F2) was used to observe the spat-
ter behavior during single-track experiments. To prevent any light scat-
tering in the laser scanning process, which would otherwise saturate
the image, a narrow band-pass interference filter was placed in front
of the camera lens. The photographing direction was perpendicular to
the single-track scanning direction, the distance between the shelter
glass and scanning track was50 mm, and the distance between the shel-
ter glass and the optics filter was 30 mm. The influence of energy input
on spatter behavior was investigated by adjusting thelaser power. Con-
taminated powder was then sieved through a 200-mesh sieve, and the
impurities were collected as a batch. The characterization of thespatter
and its comparison with the fresh powders and SLM part were
Table 1
Main technical parameters of the DiMetal-100.
Item Value Item Value
Wavelength 1090 nm Focus length 163 mm
Max laser power 200 W Building envelop 100 × 100 × 120 mm
Beam coefficient M
2
≤1.1 Scan speed 30–2000 mm/s
Focus beam size 70 μm Layer thickness 20–50 μm
Fig. 2. 316L stainless steel powder. (a) SEM image. (b) Particle distribution.
Fig. 1. Schematic diagram of SLM machine.
Table 2
316L stainless steel powder chemical compositions (mass fraction,%).
Chemical compositions C Cr Ni Mo Si Mn O Fe
Mass fraction (%) 0.03 17.53 12.06 2.16 0.86 0.38 0.13 Bal
798 Y. Liu et al. / Materials and Design 87 (2015) 797–806

conducted by means of a scanning electron microscope (SEM), energy
dispersive spectroscope (EDS) and X-ray diffraction (XRD). The SEM/
EDS analyses were performed on a Carl Zeiss-EVO MA15 SEM machine,
and XRD analyses were performed on a Bruker-D8-ADVANCE XRD ma-
chine. The XRD spectra of the samples were recorded under thefollow-
ing conditions: 2θranges from 10° to 90°, scanning speed 1°/min,
increment 0.02°, and Cu-Kαradiation. To assess the effects of spatter
on the SLM part's mechanical properties, comparative tensile testing
was conducted by using both fresh and contaminated 316L stainless
steel powder.
3. Results and discussion
3.1. Mechanism of spatter formation during the SLM
As shown in Fig. 4a, SLMuses a focused laser beam with a minuscule
diameter to melt metallic powder layer by layer, ultimately forming a
final part possessing effective metallurgical bonding and relatively
high density. The diameter of the laser beam is 30–70 μm; the energy
density can reach up to 10
5
–10
6
W/mm
2
[29]. In addition, the thermal
conductivity of 316L powder is 0.156 ± 0.004 W/mK, while the value
of the corresponding solid material is 15 W/mK [30,31], therefore the
heat in the molten pool is not easily transmitted to the surrounding ma-
terials with the resultant effect that the temperature of the exposed
powder particles exceeds the melting temperature. A further increase
in temperature (to around 2857 °C for Fe) elicits evaporation of the ma-
terial. During this phase a transformation occurs, the rapidly moving
evaporated materials expand and generate recoil pressure on the
molten pool [25,29]. While low recoil pressure facilitates flattening of
the molten pool in SLM, high recoil pressure causes the removal of
molten material by melt expulsion [26,27], thereby creating a metallic
jet. The ejected metallic jets are crushed by the metallic vapor and bro-
ken into micro-droplets during passing through the laser irradiation
field, thereby forming droplet spatter. Furthermore, the non-melted
powders around the molten pool are dispersed owing to the impact of
metallic vapor, but the particle size and shape of the powder are main-
tained, this type of spatter is known as sideways spatter. Fig. 4b shows a
typical spatter behavior generated during an SLM process, the molten
pool radiates intense glare, the oversized metallic jet reaches a height
of 5 cm, and the non-melted powders are emitted at higher velocity.
To some extent, the maximum height of glare reflects the intensity of
spatter.
3.2. Spatter observation
Of all the known factors which occur during the SLM process, energy
input has the most direct influence on spatter behavior. The energy
input can be described as follows [32]:
ψ¼4P
πd2vð1Þ
where Pis laser power, dis laser beam diameter, and vis scanning
speed. When energy input is high (e.g. P=200W,v= 50 mm/s), the
temperature within the molten pool exceeds the vaporization point,
causing the material to evaporate. Moreover, the laser positioning is al-
most perpendicular to the scanning direction, causing the vapor to in-
teract with the laser radiation and thus form a cloud consisting of
small metal particles. While energy input is low (e.g. P=50W,v=
50 mm/s), the temperature within the molten pool can hardly exceed
the vaporization point, and hence the degree of spatter is smaller.
Fig. 5 shows spatter behavior under different energy inputs. The
fixed scanning speed was 50 mm/s, layer thickness was 40 mm, and ni-
trogen gas was used as a shieldinggas. Laser powers were 50 W, 100 W,
150 W and 200 W, thus the energy inputs were 0.26 × 10
6
W/cm
3
,
0.52 × 10
6
W/cm
3
,0.78×10
6
W/cm
3
and 1.04 × 10
6
W/cm
3
respective-
ly. Spatter behavior was found to vary according to the laser power
employed.
As shown in Fig. 5a, when energy input was 0.26 × 10
6
W/cm
3
,spat-
ter behavior in the molten pool was extremely weak, which can be at-
tributed to the fact that temperature in the molten pool does not
exceed vaporization point owing to such a low energy input, and there-
fore almost no metallic vapor is generated, and only a small amount of
molten metal, if any, is torn from the molten pool.
When the energy input was 0.52 × 10
6
W/cm
3
, spatter behavior
tended to intensify. The extent of metallic jet, droplets and sideways
spatter increased, along with the size of jet, as shown in Fig. 5b. This
may be caused by the higher energy input. The temperature in the
pool exceeds the vaporization point, and thus material vaporizes and
forms an intense metallic vapor. The vapor expands outward and gener-
ates a recoil pressureon the molten pool. Therecoil pressurethen brings
about the removal of molten material by melt expulsion [26,27],
Fig. 3. Schematic of experimental setup for the high-speed photography.
Fig. 4. Spatter formation mechanism during SLM process. (a) Schematic diagram of the spatter formation. (b) Typical spatter behavior.
799Y. Liu et al. / Materials and Design 87 (2015) 797–806

thereby forming a larger metallic jet. Consequently, powder around the
pool is dispersed by the intense metallic vapor, resulting in a spatter
behavior of scattering with a maximum jetting height of around 5 cm.
When the energy input is increased to 0.78 × 10
6
W/cm
3
,asshown
in Fig. 5c, spatter behavior consequently enhanced, with a stronger
metallic jet, and droplet and sideways spatter becoming apparent.
Fig. 5. Dynamicprocesses of spatter behavior under differentenergy inputs (fixed speed of 50 mm/s, layer thicknessof 0.04 mm). (a) ψ= 0.26 × 10
6
W/cm
3
.(b)ψ=0.52×10
6
W/cm
3
.
(c) ψ=0.78×10
6
W/cm
3
.(d)ψ=1.04×10
6
W/cm
3
.
800 Y. Liu et al. / Materials and Design 87 (2015) 797–806

Under strong laser radiation, the larger jets are crushed into micro
droplets, or are even vaporized. The scatter of spatter is more evident
at a maximum jetting height of approximately 8 cm.
At an extremely high level of 1.04 × 10
6
W/cm
3
, the energy input
elicited a spatter behavior of a different kind to those discussed previ-
ously, as shown in Fig. 5d. No large size spatter or sideways spatter
occurred, and the aggregation degree was at an extremely high level,
the maximum jetting height can reach up to 12 cm. Furthermore, the
intensity of spatter oscillates periodically, this is because the intense
metallic vapor interacts with the laser radiation [29], absorbs portion
of the incident laser energy and forms a cloud consisting of small
metal particles that absorb and scatter laser radiation [33,34].This
Fig. 6. Four type tracks and corresponding cross sections with different energy inputs. (a) ψ= 0.26 × 10
6
W/cm
3
. (b) ψ= 0.52 × 10
6
W/cm
3
. (c) ψ= 0.78 × 10
6
W/cm
3
. (d) ψ=
1.04 × 10
6
W/cm
3
.
Fig. 7. Spatter particles generated during SLM process. (a) Batch of spatter particles. (b) Particle distribution of spatter particles. (c) SEM microscopy of spatter parti cles.
801Y. Liu et al. / Materials and Design 87 (2015) 797–806

cloud arises due to the condensation of hot metallic vapor, and prevents
most of the laser radiation from reaching the metal powder bed [26,27].
Therefore, vaporization in the molten pool is weakened. This processes
will occur repeatedly during the SLM process.
Fig. 6 shows simultaneously obtained scanning tracks and the corre-
sponding cross sections of the above four cases.When the energy input
was 0.26 × 10
6
W/cm
3
, the laser was not capable of melting sufficient
powder, under these circumstances, irregular and pre-balling shape
track was formed, as shown in Fig. 6a. As the energy input increased
to 0.52 × 10
6
W/cm
3
, regular but occasionally broken shaped track
was formed. The track width was approximatelyequal to the spot diam-
eter, as shown in Fig. 6b. When energy input increased continuously to
0.78 × 10
6
W/cm
3
, regular and thin shape track was formed. The track
width was about 1.5 times the size of the spot diameter, as shown in
Fig. 6c. When energy input increased to an extremely high level of
1.04 × 10
6
W/cm
3
, regular and thick shape track was formed. The
track width was about twice the size of the spot diameter, as shown in
Fig. 6d. We can interpret from this, that as energy input increases, the
scanningtrack becomes more regular in shape and more stable. Howev-
er, a powder-free zone was apparent with an extremely higher energy
input, as illustrated in Fig. 6d. This outcome can be attributed to the
Fig. 8. Comparisons between fresh powder,spatter particle and SLMpart. (a–c) SEM microscopyof fresh powder, spatter particle andSLM part. (d–f) EDS spectraof fresh powder, spatter
particle and SLM part.
Fig. 9. X-ray diffraction patterns of fresh 316L powder and spatter particle.
802 Y. Liu et al. / Materials and Design 87 (2015) 797–806

size of molten pool and the intensity of metallic vapor which increase
with a higher energy input, thus the subsequent intense vapor
causes severe turbulence in the molten pool. As a consequence, the
powder around the molten pool is sucked back into the molten
pool, a phenomena known as the capillary effect.This also demonstrates
why the amount of sideways spatter under an energy input of
1.04 × 10
6
W/cm
3
is relatively smaller.
3.3. Understanding of spatter
Contaminated 316L stainless steel powder (used five times) was
sieved through a 200-mesh sieve (only allowing particles smaller than
75 μm in diameter to pass through). The remnants were collected as a
batch, as shown in Fig. 7a. Fig. 7b shows the particle size distributions
of spatter (wt.%): D10 b46.2 μm, D50 b108.06 μm, D80 b174.48 μm.
The average particle diameter size is 119.70 μm, almost three times
that of the powder particles. Size distribution has a low aggregation de-
gree, owing to the different sources of spatter. This phenomenon con-
forms to the characteristics of the spatter forming process.
Fig. 7c shows a scanning electron microscope (SEM) micrograph of
spatter particles. The particle shape of spatter is mainly spherical,
while a large amount of non-molten powder particles adhere to the sur-
face of the spatters. The reason for this is that spatter particles maintain
a high temperature when falling onto the surface of the powder bed.
Consequently, the 316L powder particles are of small and uniform
size, which can be illustrated in Fig. 2b.
Fig. 8a–c outlines the high magnification SEM micrographs of typical
316L powder, spatter particles and SLM parts. Fig. 8a shows that 316L
powder particles exhibit a flat surface, with only a small amount of
fine particles adhering to its surface. Fig. 8b shows that the size of spat-
ter particle is about 400 μm, the area that is labeled with a red dashed
rectangle on the particle's surface is a bulge owing to the dramatic ex-
pansion of the droplet while passing through the laser radiation field.
Fig. 8c shows the top surface of an SLM part, which has a good connec-
tion between adjacent tracks although spatter particles are embedded
on the surface.
Fig. 8d–f illustrates the results of energy dispersive spectroscope
(EDS) analyses of 316L powder, spatter particles and SLM parts. Chem-
ical compositions within certain regions shown in Fig. 8a–cweredetect-
ed. It is evident that oxygen content increases in the spatter particle and
SLM parts compared with the 316L powder. This is because the molten
metal (both located in the molten pool and spraying upwards) reacts
with the remaining oxygen in the building chamber and generates
iron oxides. Moreover, as spatter particles can come into contact with
more oxygen during the flight process, the oxygen content in spatter
particles is higher than that in SLM parts.
This conclusion can also be illustrated by the XRD patterns of fresh
316L powder and spatter particles, as shown in Fig. 9. In the 316L pow-
der, the austenite (γ-Fe) and delta ferrite (δ-Fe) phases are evident; the
higher diffraction peaks are austenite while the lower diffraction peaks
are ferrite. In the spatter, the diffraction peaks of austenite and ferrite
are considerably lower than those in the 316L powder, and a diffraction
peak of Fe + 2Fe2 + 3O4 is found, owing to the generation of iron ox-
ides. However, the researchers in reference [23] declared that spatter
of 316L is a single austenitic phase. This difference in results can be at-
tributed to the fact that oxygen content in the building chamber in
our study is relatively high. It is concluded that the SLM process is
very sensitive to oxygen content within the building chamber [28].
3.4. Influence of spatter on SLM parts
As shown in Fig. 10a, spatter particles (marked as A andB) are mixed
with metallic powder and are spread onto the previously scanned layer.
The size of particle A is similar to that of the powder layer thickness. It
can be completely melted by the laser beam, thereby merging together
with the SLM part, as shown in Fig. 10b. Particle B is however much
larger than the powder layer thickness and cannot be totally melted by
the laser beam, thereby creating inclusions on the SLM part, as shown
in Fig. 10c. These inclusions not only reduce the relative density of SLM
parts, but also connect badly to the parts, consequently turning into a
source of fracture during tensile testing [35], as shown in Fig. 10d.
Tensile test pieces were designed in accordance with the Chinese
National Standard GB/T 228–2002 [36], as shown in Fig. 11a. They
were then manufactured to the process parameters as shown in
Table 3. The pieces were manufactured to a thickness of 8 mm,
then each piece was cut and machined in order to obtain three speci-
mens, regarded as a group, as shown in Fig. 11b. The tensile properties
(including ultimate tensile strength (UTS), yield strength (σ
0.2
)and
elongation (EL)) were measured on a WDS-20H type tensile testing ma-
chine at room temperature, the moving speed of the crosshead was set
at 0.6 mm/min. The obtained results are shown in Fig. 12 and Table 4.
Fig. 12 demonstrates that both groups of specimens are mainly char-
acterized by ductile fracture. However Table 4 shows that the tension
properties of specimens manufactured with contaminated powder are
far inferior to those manufactured with fresh powder. This discrepancy
can be attributed to the sole difference within the comparative analysis:
the powder. We will explain this difference by combining the stress–
strain curves with fracture morphologies.
Fig. 13a and b shows the fracture morphologies of specimens with
fresh powder, Fig. 13a is the low magnification image (×500), whichre-
veals a small amount of pores evident in the fracture. The dimples in
Fig. 13b(thehighmagnification image (×10,000)), with their irregular
shape and small size, are inhomogeneously distributed, therefore the
specimens are mainly characterized by ductile fracture. Fig. 13c and d
depicts the fracture morphologies of specimens using contaminated
powder. As in Fig. 13b, Fig. 13d also shows a large amount of dimples,
indicating the specimens have ductile fractures. However, Fig. 13c
shows considerably more pores in the fracture compared to Fig. 13a.
These pores act to cause cracks and accelerate crack propagation during
tensile testing [37,38], resulting in a dramatic reduction of mechanical
properties in the specimens.
4. Conclusions
Spatter behavior during the SLM process was investigated in this
study. Two different types of spatter were identified: droplet spatter
Fig. 10. Schematic diagrams of the influences of spatter particle.
803Y. Liu et al. / Materials and Design 87 (2015) 797–806

and powder (sideways) spatter, both of which are generated as a result
of recoil pressure caused by metallic vapor. The dynamic processes of
spatter behavior obtained by different energy inputs were observed by
using a high-speed camera. The characteristics of spatter as well as the
impact of spatter on SLM parts were studied. Several conclusions can
be drawn as a consequence.
1) Four energy inputs (0.26 × 10
6
W/cm
3
,0.52×10
6
W/cm
3
,
0.78 × 10
6
W/cm
3
and 1.04 × 10
6
W/cm
3
) were studied, and it was
found that energy input affects spatter behavior, including spatter
Fig. 11. Tensile testing. (a) Dimensional sizes and manufactured by SLM. (b) The specimens machined into desired shape. (c) The specimens after testing.
Table 3
Process parameters for manufacturing of specimens.
Items Laser power
/W
Scan speed
/mm/s
Scan hatch
/mm
Layer thickness
/mm
Scanning
strategy
Value 150 600 0.08 0.04 X–Y inter-layer
stagger Fig. 12. The stress–strain curves.
804 Y. Liu et al. / Materials and Design 87 (2015) 797–806

size, scattering state and jetting height significantly. The general
trend is that a higher energy input leads to intenser spatter behavior.
2) The particle shape of spatter is mainly spherical. The average particle
size is 119.70 μm, and almostthree times that of thepowder particle.
EDS analyses show thatoxygen content increases in spatter and SLM
part. XRD analyses show that diffraction peaks of austenite and fer-
rite are considerably lower than those exhibited in the case of 316L
powder, and are affected by the generation of iron oxides
(Fe + 2Fe2 + 3O4).
3) Comparative tensile testing results indicate that although both
groups of specimens manufactured with fresh and contaminated
powders are mainly characterized by ductile fracture, the tensile
properties of the latter are far inferior to that of the former as a result
of a larger amount of inclusions.
Acknowledgments
The authors gratefully appreciate the financial support from the Na-
tional Natural Science Foundation of China (NSFC, Nos. 51275179,
51405160), the Special Fund Project for Technology Innovation of Fo-
shan City (No. 2013AH100042).
References
[1] J.P. Kruth, G. Levy, F. Klocke, T.H.C. Child, Consolidation phenomena in laser and
powder-bed based layered manufacturing, CIRP Ann. Manuf. Technol. 56 (2007)
730–759.
[2] D.D. Gu, W. Meiners, K. Wissenbach, R. Poprawe, Laser additive manufacturing of
metallic components: ma terials, processes an d mechanisms, Int. Mater . Rev. 57
(2012) 133–164.
[3] B.C. Zhang, L. Dembinski, C. Coddet, The study of the laser parameters and environ-
ment variables effect on mechanical properties of high compact parts elaborated by
selective laser melting 316L powder, Mater. Sci. Eng. A 584 (2013) 21–31.
[4] B.C. Zhang, H.L. Liao, C. Coddet, Effects of processing parameters on properties of
selective laser melting Mg–9% Al powder mixture, Mater. Des. 34 (2012) 753–758.
[5] P. Fischer, V. Romano, H.P. Weber, N.P. Karapatis, E. Boillat, R. Glardon, Sintering of
commercially pure titanium powder with a Nd:YAG laser source, Acta Mater. 51
(2003) 1651–1662.
[6] D.D. Gu, H.Q. Wang, G.Q. Zhang, Selective laser melting additive manufacturing of
Ti-based nanocom posites: the ro le of nanopowder, Me tall. Mater. Trans. A 45
(2014) 464–476.
[7] M. Das, V.K. Bal la, D. Basu, S. Bose, A. Bandyopa dhyay, Laser processing of SiC
particle-reinforced coating on titanium, Scr. Mater. 63 (2010) 438–441.
[8] W.H. Wu, Y.Q. Yang, Key technology of selective laser melting rapid prototyping
system, Chin J. Mech. Eng. 43 (2007) 175–180.
[9] R.D.Li,Y.S.Shi,Z.G.Wang,L.Wang,J.H.Liu,W.Jiang,Densification behavior of gas
and water atomized 316L stainless steel powder during selective laser melting,
Appl. Surf. Sci. 256 (2010) 4350–4356.
[10] D. Wang,Y.Q. Yang, Z.H. Ye, X.B. Su, Research on the fabricating quality optimization
of ovehanging sur face in SLM process, Int. J. Adv. Manuf. Te chnol. 65 (2013)
1471–1484.
[11] D. Wang, Y.Q. Yang, X.B. Su, Y.H. Chen, Study on energy input and its influence on
single-track, multi-track and multi-laye r in SLM, Int. J. Adv. Manuf. Te chnol. 58
(9–12) (2012) 1189–1199.
[12] C.Z. Yan, L. Hao, A. H ussein, P. Young, D. Raymont, Adv anced lightweight 316L
stainless steel cellular latti ce structures fa bricated via se lective laser m elting,
Mater. Des. 55 (2014) 533–541.
[13] M.J. Zhang, G.Y. Chen, Y. Zhou, S.C. Li, H. Deng, Observation of spatter formation
mechanisms in high-power fiber laser welding of thick plate, Appl. Surf. Sci. 280
(2013) 868–875.
[14] M.J. Zhang, G.Y. Chen, Y. Zhou, S.C.Li, Direct observation of keyholecharacteristicsin
deep penetration laser welding with a 10 kW fiber laser, Opt. Express 21 (2013)
19997–20004.
Table 4
Tensile properties of fresh and contaminated 316L stainless steel powder.
Items UTS
/MPa
σ
0.2
/MPa
EL
/%
Fresh powder
a
664.7 ± 17.5 488.9 ± 16.7 29.3 ± 2.1
Contaminated powder
a
499.8 ± 19.9 436.3 ± 13.8 13.7 ± 2.8
a
Fresh powder refers to metal powder that has never been used before; contaminated
powder refers to metal powder that has been used five times without being sieved.
Fig. 13. The SEM morphology of fractures. (a–b) With fresh powder. (c–d) With contaminated powder.
805Y. Liu et al. / Materials and Design 87 (2015) 797–806

[15] M.J. Zhang, G.Y. Chen, Y. Zhou, S.H. Liao, Optimization of deep penetration laser
welding of thick stai nless steel with a 10 kW fiber laser, Mater. Des. 53 (2014)
568–576.
[16] M. Schweier, J.F. Heins, M.W. Haubold, M.F. Zaeh, Spatter formation in laser welding
with beam oscillation, Phys. Procedia 41 (2013) 20–30.
[17] F. Hugger, K. Hofmann, S. Kohl, M. Dobler, Michael Schmidt, Spatter formation in
laser beam welding using laser beam oscillation, Weld World 59 (2015) 165–172.
[18] D.K.Y. Low, L. Li, A.G. Corfe, Characteristics of spatter formation under the effects of
different process parameters during laser drilling, J. Mater. Process. Technol. 118
(2001) 179–186.
[19] D.K.Y. Low, L. Li, P.J. Byrd, The effects of process parameters on spatter deposition in
laser percussion drilling, Opt. Laser Technol. 32 (2000) 347–354.
[20] D.K.Y. Low, L. Li, P.J. Byrd, Spatter prevention during the laser drilling of selected
aerospace materials, J. Mater. Process. Technol. 139 (2003) 71–76.
[21] N. Guo, C.S.Xu, Y.P. WeiGuo, Du, J.C.Feng, Characterization of spatter in underwater
wet welding by X-ray transmission method, Mater. Des. 85 (15) (2015) 156–161.
[22] F.Z. Dai, D.P. Wen, Y .K. Zhang, J.Z. Lu, X.D. Ren, J.Z. Zhou, Micro-dimp le array
fabricated on surface of Ti6Al4V with a masked laser ablation method in air and
water, Mater. Des. 84 (5) (2015) 178–184.
[23] Simonelli M, Tuck C, Aboulkhair NT, Maskery I, Ashcroft I,Wildman RD, et al. A study
on the laser spatter and the oxidation reactions during selective laser melting of
316L stainless ste el, Al–Si10–Mg, and Ti–6Al–4V. Metal l Mater Trans A doi :
10.1007/s11661-015-2882-8.
[24] K.A. Mumtaz, N. Hopkinson, Selective laser melting of thin wall parts using pulse
shaping, J. Mater. Process. Technol. 210 (2010) 279–287.
[25] A. Matsunawa, J.D. Kim, N. Seto, M. Mizutani, S. Katayama, Dynamics of keyhole and
molten pool in laser welding, J. Laser Appl. 10 (6) (1998) 247–254.
[26] R. Fabbro, M. Hamadou, F. Coste, Metallic vapo r ejection effect on melt pool
dynamics in deep penetration laser welding, J. Laser Appl. 16 (1) (2004) 16–19.
[27] R. Fabbro, S. Slima ni, I. Doudet, F. Co ste, F. Brian, Ex perimental study of the
dynamical coup ling between the induced vapour plume and the melt pool for
Nd-YAG CW laser welding, J. Phys. D. Appl. Phys. 39 (2006) 394–400.
[28] D. Wang, Y.Q. Yang, R.C. Liu, D.M. Xiao, J.F. Sun, Study on the designing rules and
processability of porous structure based on selective laser melting (SLM), J. Mater.
Process. Technol. 213 (2013) 1734–1742.
[29] J.P. Kruth, L. Froyen, J.V. Vae renbergh, P. Me rcelis, M. Rombouts, B. Lauwers,
Selective lase r melting of iron-b ased powder, J. Mater. Process. Tec hnol. 149
(2004) 616–622.
[30] M. Rombouts, J.P . Kruth, L. Froyen, Fundamenta ls of selective laser meltin g of
alloyed steel powders, CIRP Ann. Manuf. Technol. 55 (2006) 187–192.
[31] M. Rombouts, L. Froyen, A.V. Gusarov, E.H. Bentefour, C. Glorieux, Photopyroelectric
measurement of thermal conductivity of metallic powders, J. Appl. Phys. 97 (2005)
024905.
[32] A. Simchi, Direc t laser sinteri ng of metal powders: mechanism , kinetics and
microstructural features, Mater. Sci. Eng. 428 (2) (2006) 148–158.
[33] P.Y. Shcheglov, S.A. Uspenskiy, A.V. Gumenyuk, V.N. Petrovskiy, M. Rethmeier, V.M.
Yermachenko, Plume attenuation of laser radiation during high power fiber laser
welding, Laser Phys. Lett. 8 (6) (2011) 475–480.
[34] P.Y. Shcheglov, A.V . Gumenyuk, I.B. Gornus hkin, M. Rethmeier, V. N. Petrovskiy,
Vapor-plasma plume investigation during high-power fiber laser welding, Laser
Phys. 23 ( 2013) 0 16001.
[35] H.J. Gong, K. Rafi, H.F. Gu, G.D.J. Ram, T. Starr, B. Stucker, Influence of defects on
mechanical properties of Ti–6A l–4V components produced by se lective lase r
melting and electron beam melting, Mater. Des. 86 (5) (2015) 545–554.
[36] GB/T 228-2002, Metallic Materials—Tensile Testing at Ambient Temperature,
Standard Press, Beijing: China, 2002.
[37] Y. Murakami, M. Endo, Effects of defects, inclusions and inhomogeneities on fatigue
strength, Int. J. Fatigue 16 (3) (1994) 163–182.
[38] D. Broek, The role of inclusions inductile fracture andfracture toughness,Eng. Fract.
Mech. 5 (1) (1973) 55–66.
806 Y. Liu et al. / Materials and Design 87 (2015) 797–806