Accepted Manuscript

Title: Residual stress measurements on AISI 316L samples

manufactured by selective laser melting

Author: Thomas Simson Andreas Emmel Anja Dwars Juliane

Böhm

PII: S2214-8604(17)30160-4
DOI: http://dx.doi.org/doi:10.1016/j.addma.2017.07.007
Reference: ADDMA 199

To appear in:

Received date: 11-4-2017

Revised date: 13-5-2017
Accepted date: 18-7-2017

Please cite this article as: Thomas Simson, Andreas Emmel, Anja Dwars,
Juliane Bddotohm, Residual stress measurements on AISI 316L samples
manufactured by selective laser melting, <![CDATA[Additive Manufacturing]]>
(2017), http://dx.doi.org/10.1016/j.addma.2017.07.007

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 proof
before it is published in its final 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.
 Residual stress measurements on AISI 316L samples manufactured by selective
laser melting

Thomas Simsona,∗, Andreas Emmela , Anja Dwarsb , Juliane Böhmb

a OTH Amberg-Weiden, Faculty of Mechanical and Environmental Engineering, BY, DE
b TH Nürnberg, Institute of Electrical Engineering, Precision Engineering, Information Technology, BY, DE

t
ip
cr
Abstract
This paper aims to understand the formation and the effect of residual stress on selective laser melting (SLM) parts.
SLM is a powder bed based additive manufacturing (AM) process and can be compared to a laser welding process.

us
Due to the high temperature gradients and the densification ratio, which are characteristic of this process, residual
stresses occur. The investigation of residual stress is performed using X-ray diffraction (XRD) for samples made of
austenitic stainless steel AISI 316L (EN 1.4404). This research examines residual stress at different depths and at
two outer surfaces. For the measurement of stresses at different depths, the samples’ surface layers were removed
an
by electropolishing. At sufficiently large distances from the top surface, the stresses in the area of the edge layer
initially increase strongly and then decline again. The value and orientation of the resulting main stress components
are dependent on the examined layer. At the top surface, the residual stresses are higher in scan direction than in
perpendicular direction. In contrast, at the lateral surface the maximum main stress is perpendicular to the scan and
M
parallel to the building direction. These two cases can be described very well by the two mechanisms in SLM, namely
the temperature gradient mechanism (TGM) and the cool-down phase. It is also shown that at samples with a relative
structural density of > 99 %, the residual stress values are independent of the applied energy density.
Keywords: selective laser melting (SLM), residual stress, X-ray diffraction (XRD), energy density, AISI 316L
ed

1. Introduction laser beam is deflected and thereby generates the prede-

termined layer geometry.
pt

Additive manufacturing (AM) belongs to the 3D gen- The aim of SLM is to produce almost completely dense
erating processes. The advantages of AM are geomet- parts with mechanical properties comparable to bulk
rical freedom, material flexibility, reduction of process materials [3]. The major influencing factors for this pro-
ce

steps, mass customisation and shortened design to prod- cess are building strategy and process parameters. The
uct time [1]. The process involves the tool-less pro- layered structure of the components provides consider-
duction of samples, prototypes, tools or end-use parts able geometric freedom.
from a digital model, e.g. customised medical parts, In the present study, the SLM samples were fabricated
Ac

tooling inserts with near-surface cooling channels and by using the island scanning strategy. The 3D-CAD
functional components of high geometrical complexity model is divided into single layers with defined layer
[2]. thicknesses, which allows the processing of the individ-
The selective laser melting (SLM) process - as a part of ual component layers independently of each other. To
AM - belongs to the the powder bed fusion technology. generate isotropic properties, the layers are subdivided
During the manufacturing process, a thin layer of pow- into several square subfields and are filled by scanning
der is first applied to a substrate plate. A laser beam then the laser beam in tracks, creating a checkered pattern.
melts the powder of the material and subsequently con- This means the layer plane of the square subfields is
nects it to the underlying layer. With a scanning unit the shifted in the x- and y-direction (Figure 1 a)) and the
superimposed fields are rotated by 90◦ (Figure 1 b)).
The basic procedural principles of SLM and selective
∗ Correspondingauthor laser sintering (SLS) are comparable. However, in the
Email address: [email protected] (Thomas Simson) case of SLS, the connection of the powder particles re-
Preprint submitted to Additive Manufacturing May 13, 2017

Page 1 of 10
 a) b)

t
ip
Figure 1: Exposure strategies a) shift x- and y-direction b) rotation of superimposed fields by 90◦

cr
sults from the melting of a low-melting component (two perature, which leads to high temperature gradients and
component systems) which acts as a binder. Since only surface tension. Due to this instability, the molten pool
one component is used for the SLM, this process is is broken up and reduces its surface area to an energeti-

us
rather comparable to a laser welding process (in particu- cally more suitable state by forming spheres [7].
lar to laser cladding with additional material in powder In addition, a high energy density can lead to pore for-
form). The essential difference, however, is a smaller mation by vaporisation of low melting elements [8] as
melting pool, by which it is possible to generate finer well cavities can be entrapped inside melt pools with
structures and even thinner layers, which directly influ-
ences the shape accuracy [4].
Due to the small melting pools, there are no problems
an
a keyhole geometry by a result of material evaporation
[9].
Hatch spacing also has a major influence on the com-
with shrinking, caves and pores when appropriate laser ponent density. With a hatch spacing equal to or larger
M
parameters are used, but the local solidification of the than the laser beam diameter, no metallurgical connec-
molten regions induces stresses and can cause signifi- tion is produced between the single tracks. If the hatch
cant distortion and cracks. This paper examines pore spacing is too small, a large part of the already solidified
and residual stress formation in the samples relating track is melted and only a small amount of powder is
ed

to the building parameters of the SLM process. The added. Due to the surface tension of the melt, this is at-
residual stresses are determined using the X-ray diffrac- tached to the existing track. After several tracks, the dis-
tion method according to ASTM E915-96/E1426-98. tance between the tracks and the powder has expanded
The investigation is performed on samples made of so far that the melt of the melted powder cannot bridge
austenitic stainless steel AISI 316L (EN 1.4404). this distance. Therefore, a void remains between the so-
pt

lidified structure and the powder [4]. The effects, further

described in subsection 2.2, result in high stresses in the
2. Mechanisms in SLM
component, which may lead to cracks thereby causing
ce

2.1. Densification a reduction in density. Generally, an increase in laser

energy input leads to higher temperatures and thus to a
Basically, SLM is a laser welding process based on lower melt viscosity and surface tension [7]. Therefore,
the heat conduction welding process. The powder is lo- the flow of the molten metal is increased and pores can
Ac

cally melted by the absorption of laser radiation. The be filled.

transition to the liquid phase results in a metallurgical Due to these facts, the coordination of the process pa-
connection between the adjacent track and the previ- rameters must be well understood so that high density
ously scanned layer [5]. components can be produced.
Pore formation may take place in different ways (Fig- Various methods can be used to determine the den-
ure 2). For example, at low energy densities, the pen- sity of SLM components. On the one hand, there is
etration depth of the laser radiation is not sufficient to the ‘Archimedes-method’, in which the density is de-
produce a connection with the lower layer and the en- termined by weighing the component in air and sub-
ergy is too low to melt the powder completely [6]. Thus, sequently in ethanol. Besides, a varnish coating pre-
voids remain between the particles, forming pores. vents the absorption of ethanol by the component dur-
On the other hand, when a sufficiently high energy den- ing the measurement. However, in order to be able to
sity is present which completely melts the existing pow- determine the size and distribution of pores and cracks
der, the increased energy density also increases the tem-
2

Page 2 of 10
 a) b)

t
ip
cr
us
Figure 2: Micrographs showing porosity of SLM-produced samples at lateral surface 50-fold magnification with different energy densities a) 43
J/mm3 b) 143 J/mm3

within a component, analysis by digital image process-

ing is more suitable. For this purpose, metallographic
sections at different planes of the component are pre-
pared. These are used to generate images and to de-
an
shrinkage of adjacent laser melted zones in solidified
material whereby a decrease in thermal stress also
causes a decrease in residual stress.
This behavior can be described by two models, the tem-
termine the pore proportion by digital image analysis. perature gradient mechanism (TGM) and the cool-down
M
This enables a much better quantification of the process mechanism [11]. The TGM is used here to describe
parameters. the stress induction of a single melt track, whereas the
cool-down mechanism describes the behavior of an
2.2. Residual stress entirely melted powder layer. Both mechanisms are
ed

based on the expansion behavior of the material during

Thermal stresses occur when a volume is unable heating or cooling.
to expand or contract without hindrance in response The TGM is already known from the laser forming
to local temperature changes, e.g. during melting by of sheet materials [10]. Figure 3 shows the schematic
pt

laser radiation. Materials in the liquid phase are not principle of TGM. Due to energy insertion by the laser
subject to thermal stresses. Thermal stresses occur from beam, the material expands thermally. This expansion
temperature gradient or from solidification-induced
ce

a) b)
Ac

Figure 3: TGM inducing residual stress based on [10] a) heating b) cooling (plastic ε pl and thermal εth elongation, tensile σtens and compressive
σcomp stress)

Page 3 of 10
 a) b)

t
Figure 4: Cool-down phase inducing residual stress at solid layer connection based on [4] a) heating b) cooling

ip
is limited because of the colder surrounding material, nearly homogeneous across smaller material areas (a
which induces elastic compressive stresses in the heat grain or grain areas) and can lead to macroscopic di-

cr
affected zone. With the exceed of the yield strength mensional changes. Type III residual stresses are in-
at temperature, the resulting stresses are reduced by homogeneous across the smallest material regions (sev-
plastic deformation (Figure 3 a)) or by upsetting of eral atomic distances) and do not lead to deformation.

us
the material in the direction of the energy source. If Type II and Type III residual stresses result from phase
no more heat is introduced, the thermally stretched differences in the material and dislocations in the lat-
area cools down and contracts. During cooling, tensile tice. Figure 5 shows the graphical representation of
stresses are generated in the plastically deformed area, the different residual stresses and their relationship to
which leads to a bending in the opposite direction
(Figure 3 b)).
In the case of the cool-down mechanism, the melted
an
the areas in which they occur. The macroscopic resid-
ual stresses cause a shift of the diffraction angle of the
diffraction reflex as shown in the XRD analysis. The
top layer initially has a higher temperature than the un- software described in section 3 evaluates macroscopic
M
derlying layer (Figure 4 a)). When the melt has cooled residual stresses.
and solidified, the warmer upper layer contracts to a
greater extent than the colder subjacent layer. However,
due to the solid metallurgical connection of both layers,
this contraction is inhibited. As a result, tensile stresses
ed

occur in the upper layer and compressive stresses in the

underlying layers (Figure 4 b)).
The moment of resistance of the already built com-
ponent or the substrate plate at SLM is generally so
pt

high that a plastic deformation such as a deflection of a

single melt track or layer during cooling is prevented.
As a result, the remaining tensile stresses are very
ce

high and the formation of voids and cracks in the com-

ponent may occur when the tensile strength is exceeded. Figure 5: Definition of residual stresses Type I, II, III based on [13]

2.3. Energy density

Ac

Residual stresses can be distinguished into three

types, i.e. Type I (macroscopic stresses), Type II and
Type III (microscopic stresses). Type I is the average of The influencing factors for the SLM can be assigned
the stresses in a volume which contains enough grains to five system elements, namely material, design, equip-
of all material phases to be representative. This means, ment, process parameters and exposure strategy. The
that the stress value is nearly homogeneous across large characteristic process parameters (Figure 6) include the
material areas (several grains) and leads to deforma- laser power PL (in W), the scanning velocity v (in mm/s)
tions of the component. The deviations from the macro- and the spot diameter (in µm). These values in connec-
scopic average value are called micro stresses. The de- tion with the layer thickness d (in mm) and the hatch
viations from the average stress (Type I) within a grain spacing h (in mm) yield the energy density EV (in J/mm3 ).
are defined as Type II, and the position-dependent devi- The energy density specifies the average energy intro-
ations of the stresses from Type II are referred as resid- duced during the exposure of a layer per material vol-
ual stresses Type III [12]. Type II residual stresses are ume.
4

Page 4 of 10
 a) b)

t
ip
cr
Figure 6: Schematic of the selective laser melting process a) relevant process parameters b) melt pool

us
PL
EV = (1)
vs · h · d

The residual stress state during the manufacturing

process, which depends very strongly on heat input, ma-
an
terial parameters and exposure strategy, is essential for
M
shape accuracy, crack-freeness and the application be-
havior. By varying the process parameters to a cer-
tain extent, the same volume energy densities can be
achieved with different parameter values. The process
ed

parameters mentioned in equation (1) must be opti-

mized and adjusted to each other. If this is not the case,
Figure 7: Residual stress measurement positions at the samples (two
pores in the component are produced based on the ef- at the top surface, one at the lateral surface) with marked main stress
fects described in subsection 2.1, which in turn leads to components σ1 and σ2
a drop in strength. However, a certain energy density is
pt

necessary to melt the material completely thus ensuring

that the components are as free of pores as possible. The surface morphology was analyzed by a scanning
ce

electron microscope (SEM) Leo Stereo Scan 440 (Fig-

3. Experimental procedure ure 9).
The relative density was measured on microsections us-
The material used for this investigation is austenitic ing digital image analysis. The examinations of the sec-
tions were performed using the inverse incident light
Ac

stainless steel AISI 316L (EN 1.4404). The powder was

produced by LPW Technology by gas atomization re- microscope Leica DMI5000 M. The pictures were taken
sulting in spherical particles. According to manufac- at 50-fold magnification and were evaluated with the
turer the employed particle size was 15 - 45 µm. All Phase Expert function of the image analysis software
test samples used to determine the residual stress were LAS V4.6. For each sample the ratio of pores was deter-
produced by a Concept Laser Mlab cusing metal laser mined in two polished sections, both in building plane
melting system in nitrogen atmosphere using the island and in building direction (Figure 7). Figure 8 shows
scanning strategy (Figure 7). the relative density of the samples, calculated from the
In table 1 the process parameters of four examined porosity, depending on the energy density.
energy densities (43, 71, 79 and 143 J/mm3 ) are shown. The inserted label (red line) at a relative density of
This is particularly important since the same value of 99 % represents the lower limit of the tolerable porosity.
the energy density can be obtained from several combi- This deviation of 1 % is still acceptable for many ap-
nations of process parameters (Equation (1)). plications, since the pores do not occur coherently but
5

Page 5 of 10
 Nr. Laser Scanning Layer Hatch Island Energy
power velocity thickness spacing size density
(W) (mm/s) (mm) (mm) (mm) (J/mm3 )
1 90 1500 0,025 0,056 5/10 43
2 90 600 0,025 0,084 5/10 71

t
3 90 300 0,045 0,084 5 79

ip
4 90 300 0,025 0,084 5 143
Table 1: Examined process parameters of four different energy densities

cr
100 stress measurements were carried out. The crystalline
structure of the stainless steel was austenitic face cen-

us
Relative density d (%)

tered cubic (fcc). The measurements at the building

95 plane were carried out in two adjacent fields whose scan
vectors stand perpendicular to each other.
XRD measurements were conducted using a Bruker
90
Pores
an
AXS/Siemens D5005 instrument equipped with a
horizontal four-axis goniometer with open Eule-
rian cradle, operating with chromium (Cr) radiation
85
(K α = 2.29103 Å) at 40 kV and 40 mA. To investigate
M
the SLM parts, fcc (220) diffraction peak was used. A
80 2θ angle over 90◦ reduces grazing incidence and re-
0 50 100 150 flection of radiation. Therefore, higher indexed lat-
Energy density EV (J/mm3 ) tice planes should be chosen for residual stress mea-
surement. The scan rate was 5.3̄ ◦ /h in the angular
ed

Figure 8: Relative density with label (red line) at a relative density of range (2θ) 127◦ - 130◦ . The irradiated area of the X-
99 %. Samples with for an energy density of 43 J/mm3 and 143 J/mm3 ray beam was about 4 mm in diameter. For evaluation
are shown (Figure 2)
of all components of the stress tensor at least 6 strain
components have to be determined in independent di-
pt

rections. For this purpose the sample is rotated about

rather sporadically.
the ϕ axes and tilted about ψ axes. An equidistant dis-
It is clearly noticeable that the structural density in-
tribution of the measuring ranges are recommended for
ce

creases with higher energy densities. Beyond an en-

routine measurement. The reliability of the measure-
ergy density of 71 J/mm3 , the resulting density increases
ments can increased by additional strain components.
to > 99 %. All samples manufactured with an energy
The stress measurements were carried out using the
density of 43 J/mm3 have densities of < 95 %. However,
sin2 ψ method at three angles ϕ (0◦ , 45◦ , 90◦ ) and 13
Ac

too high energy densities may also lead to pores within

directions ψ (-60◦ - 60◦ ). The selected parameters en-
the components, because of vaporisation of low melting
ables a high reliability of the measurements, whereby
elements [8].
the measurement time was still acceptable. The diffrac-
The aim of the SLM process is the production of com-
tion reflexes were described using the Pearson VI func-
ponents of maximum possible density. This is an im-
tion. Due to the low penetration depth of the X-rays,
portant factor for the assessment of the process and the
the residual stresses were calculated considering plain
process parameters. Already low residual porosity can
stress conditions. The calculation of the stresses from
be critical for some high strength applications.
the measured lattice strain distributions uses the X-ray
Residual stress measurements were performed at
elastic constants. These are dependent on the mea-
three measuring points on as-fabricated samples - two
sured lattice plane (hkl) and consider the microscopic
measuring points at the building plane (top surface) and
anisotropy. The anisotropy factor which is a measure
one measuring point along the building direction (lateral
for elastic anisotropy of a material is necessary for the
surface) (Figure 7) - which means a total 42 residual
6

Page 6 of 10
 a) b)

t
ip
cr
us
Rz 30 - 69 (µm) Rz 39 - 60 (µm)
Ra 7 - 14 (µm) Ra 7 - 12 (µm)
Figure 9: SEM 100-fold magnification with values of the surface roughness a) top surface b) lateral surface

profile removal
100
0
an
-100
depth (µm)

M
depth (µm)
-200
1 0
-300
2 20 - 50
-400
ed

3 110 - 130
4 200 - 230 -500
-600
5 500 0 5 10 15 20
pt

distance (mm)
Figure 10: Table of removal depth value (left), mapping of removal depth (right)
ce

calculation of the X-ray elastic constants. The used X- ers were removed in the absence of thermal and me-
ray elastic constants are s1 = −1.39 × 10−6 1/MPa and chanical stress using electropolishing. Electropolishing
Ac

1/2 s2 = 6.27 × 10−6 1/MPa with the assumed material was conducted with a Buehler PoliMat 2 instrument, op-
constants E = 190 GPa, ν = 0.3 and an anisotropy factor erating with a stainless steel cathode (EN 1.4571) and
A  1.72 [14]. 10 % sulfuric acid (H2 SO4 ) as electrolyte at 18 V for
Figure 9 shows pictures of the top surface (a)) and lat- various durations (45 s - 720 s). An aperture plate with
eral surface (b)) at 100-fold magnification with a scan- a hole 11 mm in diameter is used, where the electrolyte
ning electron microscope (SEM). It is clearly noticeable contacts the samples surface.
that non-molten powder particles cause a very rough The measurement of the thickness of the removed layers
surface. was conducted by mapping the polishing profiles with
the optical distance sensor CHR 150 N manufactured by
Since the X-rays penetrate the material to a depth of the Precitec Group in a JURCA Optoelektronik GmbH
just a few micrometers, stress can only be measured in measurement setup Nemisis V. Figure 10 (right) shows
low depths. Therefore, further measurements were car- for example the mapping of removal depth of an etched
ried out in which the surface roughness and further lay-
7

Page 7 of 10
sample. Because of the edge effect, the removal by elec- 500
tropolishing at the edges was larger than in the central

Residual stress σ1 (N/mm2 )

section. Therefore, the estimation of the removal depth 400
was carried out in the central section, disregarding the 300
edge area. The values determined are shown in (Fig-
ure 10 left). 200

100

t
ip
4. Results and discussion 0

−100

cr
Figure 11 and 12 show residual stress measurements −200
carried out on eight different samples with removed sur- 0 50 100 150
face roughness (by a depth of about 20 µm). The resid- Energy density EV (J/mm3 )

us
ual stress states are indicated and described with ref-
erence to the energy densities for the top surface (Fig- Figure 11: Residual stress values by different energy densities at the
top surface
ure 11) and the lateral surface (Figure 12).
The main stresses are labeled σi (i = 1, 2, 3); in common
practice the largest value is σ1 and all others are further
declining. The main stress directions are orthogonal to
an 500
Residual stress σ1 (N/mm2 )

each other. For the two measurements conducted at the 400

top surface (see Figure 7), the mean values of the main
stress σ1 with error bars are shown in figure 11. σ1 300
M
follows the scanning direction, i.e. the residual stress
200
values are significantly higher in the scanning direction
than in the perpendicular direction. The stresses in scan- 100
ning direction correspond to the process described by
ed

TGM. The residual stress value determined for the sam- 0

ples with an energy density of 43 J/mm3 is slightly lower
−100
than for samples with an energy density ≥ 71 J/mm3 .
Since components with high porosity have a large num- −200
pt

ber of defects and since there are no fixed connections 0 50 100 150
to the surrounding tracks, the residual stresses can be Energy density EV (J/mm3 )
reduced. Therefore these samples with high porosity
ce

have necessarily low residual stresses. It appears that Figure 12: Residual stress values by different energy densities at the
the residual stress values increase with a higher energy lateral surface
density. From an energy density of ≥ 71 J/mm3 , however,
these residual stress values are similar, considering the Since components with high porosity have a large
Ac

range of error. number of defects, these are not suitable for application.
Figure 12 shows the residual stress measurements at All samples with an energy density of 43 J/mm3 are below
the lateral surface. Here, also the main stress values the limit of tolerable porosity (Figure 8) and therefore
σ1 were plotted. Nevertheless, these have the highest have not been further investigated. Three samples with
value perpendicular to the scan direction, i.e. in building the highest density were selected for further studies, i.e.
direction (Figure 7 z-direction). These stresses corre- two samples with an energy density of 71 J/mm3 and one
spond to the process described for the cool-down phase. sample with 143 J/mm3 .
In contrast to the residual stresses at the top surface Figure 13 exemplifies the depth profiles of the three
(Figure 11), samples with low energy density (43 J/mm3 ) electropolished samples. The points on the graphs de-
also show higher residual stresses. As in figure 11, the pict the determined depths of polishing representing the
determined residual stress values can be considered to distribution of residual stresses. Relative to the distance
be similar for all samples, taking account of the range from the surface, five measurements were carried out on
of error. each lateral surface, see figure 10, of the three samples.
8

Page 8 of 10
a) Island size 5 mm b) Energy density EV 71 J/mm3
300
Residual stress σ1 (N/mm2 )

300

Residual stress σ1 (N/mm2 )

200
200

t
ip
100
100

cr
0 0
0 100 200 300 400 500 0 100 200 300 400 500

us
Distance from lateral surface (µm) Distance from lateral surface (µm)
Energy density EV (J/mm3 ) 71 143 Island size 5 mm 10 mm

Two samples were built with varied scan velocity (600

an
Figure 13: Residual stress depth profiles with different constant process parameters a) island size of 5 mm b) energy density of 71 J/mm3

that the originating residual stresses are strongly de-

mm/s and 300 mm/s) (Figure 13 a)) and one with the energy pendent on the selected process parameters. The ori-
M
density of 71 J/mm3 , where the island size (5 mm and 10 entation of the resulting main stress components is de-
mm) was varied (Figure 13 b)). The first measurement pendent on the examined layer. At the top surface the
was performed on the untreated surface (0 µm). Here residual stresses show higher values in scan direction
the stress measurement mainly reveals the stresses in than in perpendicular direction. However, at the lateral
ed

the powder itself since the uppermost layer is only par- surface the highest main stress is perpendicular to the
tially remelted. scan direction. These two cases can be described very
For the second measurement, the adhesive particles as well by the mechanisms in SLM, namely the TGM and
well as the unevenness caused by the SLM process have the cool-down phase. Adherence of non-fused or partly
pt

been removed (by a depth of about 20 µm). The re- fused powder particles has a strong influence on the
sulting residual stresses were significantly higher than residual stress values, because these will be strongly re-
those at the very surface and increased with increasing duced. For the depth profile of the samples, the stresses
depth. The maximum occurred at about 100 µm below
ce

in the surface area initially rise strongly and then decline

the surface followed by a decrease of the stress level to again.
150 N/mm2 . The experimental results of residual stress measure-
Taking into account the range of error, the residual stress ments on the outer surfaces (top and lateral surface) in-
Ac

distributions as a function of the depth of the different dicate the dependence of the stress values on the struc-
samples may be regarded as similar, because the results tural density. For measurements in the top surface, only
show a comparable tendency. It appears that in the sur- samples with a low energy density (43 J/mm3 ) have lower
face area, the residual stresses initially increase strongly residual stress values. However, since a relative den-
and then decrease again with increasing depth. This sity of > 99 % is a requirement for applications, the
zone is the transition area between the outer and the in- bulk quality of these samples is not suitable for com-
ner surface. ponents. With a relative structural density of > 99 %
all powder particles are completely melted and the sin-
gle laser tracks have a metallurgical connection. Under
5. Conclusion
these conditions the residual stress values are indepen-
Residual stresses were investigated on SLM sam- dent of the applied energy density.
ples made of austenitic stainless steel AISI 316L (EN
1.4404) using X-ray diffraction. The test results show
9

Page 9 of 10
Acknowledgement

This study was supported by the KSB-Stiftung,

“Energie- und ressourceneffiziente Darstellung unikaler
Werkstoff- und Bauteileigenschaften mit Hilfe des ad-
ditiven Lasergenerierens” with grateful acknowledge-
ment.

t
References

ip
[1] J.-P. Kruth, et al. (Eds.), Benchmarking of different SLS/SLM
processes as rapid manufacturing techniques, 2005.

cr
[2] B. Vandenbroucke, J.-P. Kruth, Selective laser melting of bio-
compatible metals for rapid manufacturing of medical parts,
Rapid Prototyping Journal 13 (4) (2007) 196–203. doi:10.
1108/13552540710776142.

us
[3] J.-P. Kruth, et al., Selective laser melting of iron-based powder,
Journal of Materials Processing Technology 149 (1) (2004) 616–
622. doi:10.1016/j.jmatprotec.2003.11.051.
[4] W. Meiners, Direktes selektives Laser-Sintern einkompo-
nentiger metallischer Werkstoffe, Shaker Verlag, Aachen, 1999,
dissertation.
[5] R. Poprawe, Lasertechnik für die Fertigung, Springer-Verlag,
Berlin Heidelberg, 2005.
[6] R. Li, et al., Densification behavior of gas and water atom-
ized 316l stainless steel powder during selective laser melting,
an
Applied Surface Science 256 (13) (2010) 4350–4356. doi:
M
10.1016/j.apsusc.2010.02.030.
[7] A. Simchi, H. Pohl, Effects of laser sintering processing pa-
rameters on the microstructure and densification of iron powder,
Materials Science and Engineering: A 359 (1) (2003) 119–128.
doi:10.1016/S0921-5093(03)00341-1.
ed

[8] J. A. Cherry, et al., Investigation into the effect of process param-

eters on microstructural and physical properties of 316l stainless
steel parts by selective laser melting, Int J Adv Manuf Technol
76 (5) (2015) 869–879. doi:10.1007/s00170-014-6297-2.
[9] W. E. King, et al., Observation of keyhole-mode laser melting
in laser powder-bed fusion additive manufacturing, Journal of
pt

Materials Processing Technology 214 (12) (2014) 2915–2925.

doi:10.1016/j.jmatprotec.2014.06.005.
[10] F. Vollertsen, Laserstrahlumformen, lasergestützte Formgebung:
Verfahren, Mechanismen, Modellierung, Meisenbach Verlag,
ce

Bamberg, 1996.
[11] P. Mercelis, J.-P. Kruth, Residual stresses in selective laser sin-
tering and selective laser melting, Rapid Prototyping Journal
12 (5) (2006) 254–265. doi:10.1108/13552540610707013.
[12] E. Macherauch, H. Wohlfahrt, U. Wolfstieg, Zur zweckmäßigen
Ac

Definition von Eigenspannungen, HTM 28 (3) (1973) 201–211.

[13] P. J. Withers, H. K. D. H. Bhadeshia, Residual
stress part 1–measurement techniques, Materials sci-
ence and Technology 17 (4) (2001) 355–365. doi:
10.1179/026708301101509980.
[14] Bruker Analytical X-ray Systems GmbH, DIFFRACplus
STRESS V 1.0 Residual Stress Determination User’s Manual,
User’s Manual, DOC–M88–E01001 (1997).

10

Page 10 of 10