materials

Article
Laser Polishing of Additive Manufactured 316L
Stainless Steel Synthesized by Selective
Laser Melting
Muhannad A. Obeidi 1,2,3, *, Eanna McCarthy 1,2 , Barry O’Connell 4 , Inam Ul Ahad 1,2,3 and
Dermot Brabazon 1,2,3
1 School of Mechanical & Manufacturing Engineering, Dublin City University, D09 V209 Dublin, Ireland;
eanna.mccarthy@dcu.ie (E.M.); inamUl.Ahad@dcu.ie (I.U.A.); dermot.brabazon@dcu.ie (D.B.)
2 Advanced Processing Technology Research Centre APT, Dublin City University, D09 V209 Dublin, Ireland
3 I-Form Advanced Manufacturing Research Centre, Dublin City University, D09 V209 Dublin, Ireland
4 Nano Research Facility, Dublin City University, D09 V209 Dublin, Ireland; barry.oconnell@dcu.ie
* Correspondence: muhannad.ahmedobeidi@dcu.ie; Tel.: +353-1-700-6058




Received: 6 February 2019; Accepted: 16 March 2019; Published: 26 March 2019


Abstract: One of the established limitations of metal additive manufacturing (AM) methods, such as
selective laser melting (SLM), is the resulting rough surface finish. Laser polishing is one method
that can be used to achieve an improved surface finish on AM printed parts. This study is focused
on the laser surface polishing of AM parts using CO2 laser beam irradiation. Despite the fact that
several researchers have investigated the traditional abrasive polishing method, there is still a lack of
information reporting on the laser surface polishing of metal parts. In this study, AM 316L stainless
steel cylindrical samples were polished using CO2 laser beam irradiation in continuous wave (CW)
working mode. Two design of experiment models were developed for the optimization of the input
processing parameters by statistical analysis of their effect on the resulting roughness. The processing
parameters investigated were the laser beam power, the rotational speed of the sample, the number
of laser scan passes, the laser beam focal position, and the percentage overlap of the laser tracks
between consecutive passes. The characterization of the measured roughness and the modified
layer microstructure was carried out using 3D optical and scanning electron microscopy (SEM).
A maximum reduction of the roughness from 10.4 to 2.7 µm was achieved and no significant change
in the microstructure phase type and micro-hardness was observed.

Keywords: laser polishing; additive manufacturing (AM); metal 3D printing; 316L; stainless steel;
surface processing

1. Introduction
Additive manufacturing (AM) is becoming well established in industry for rapid prototyping and
manufacturing of parts. AM techniques are becoming increasingly popular in manufacturing industries
and are proving very successful in aerospace, automotive, tooling, and biomedical applications.
Metals are processed with different processes, such as direct energy deposition and selective laser
melting (SLM), also known as powder bed fusion (PBF). The heat source used in these techniques
is a laser beam or electron beam. SLM is one of the most important processes used in research and
industry, with significant capability in the manufacture of parts with exceptional properties and
geometric complexity [1–8]. Compared with conventional subtractive manufacturing methods like
Computerized Numerical Control (CNC) milling, turning, and forging, AM can reduce material
wastage and production time for specific component designs [9,10].

Materials 2019, 12, 991; doi:10.3390/ma12060991 www.mdpi.com/journal/materials

Materials 2019, 12, 991 2 of 15

One of the limitations of SLM is the rough finish of the printed parts, requiring manual
post-processing to achieve the required roughness. The final roughness depends on many factors such
as the initial powder particle size, the building layer thickness, the laser beam power, and the laser
scanning speed. The resulting roughness of any AM part is also affected by two main processes—firstly,
the well know phenomenon called waving or the stair effect caused by the building the consecutive
layers; and secondly, the sintering and adherence of the adjacent not fully fused powder particles
on the surface of the produced part, which is called the Balling effect [11,12]. The resulting high
roughness of the AM parts limits the ease of employment of the SLM process for high-end applications
such as in biomedical implants, in which a non-optimized roughness may create bacterial growth or
tissue damage [13], and for high pressure hydraulic valves used in the aerospace industry, where leak
tightness is required between connecting components. In aerospace applications, parts also require
extremely low roughness for better control over the dimensional accuracy, provision of low friction,
and the reduction of surface crack initiation due to vibration and cyclic loads [14].
The polishing of AM parts using a conventional method such as abrasive, mechanical, or electro-
polishing does not provide an ideal solution because of the lack of dimensional accuracy, especially
when selective places require different treatment. Also, electro-polishing has been shown in a number
of cases to have a negative impact on the environment [15–19]. There is a similar consideration of
the negative effect on the environment during the application of the chemical mechanical polishing
technique (CMP), in which chemical and abrasive materials are used [20,21]. AlMangour B. et al. [22]
investigated the improvement of the surface quality and mechanical properties of 17-4 stainless steel
flat samples manufactured by direct metal laser sintering (DMLS). The researcher used the shot peening
technique to reduce the roughness by approximately 70% with a noticeable enhancement in the wear
resistance and hardness to the generation of high compressive stresses and grain refinement.
During recent years, interesting results have been obtained from laser polishing of different metals
and non-metal materials. Laser polishing offers an ecologically-friendly method of achieving good
finishes without compromising dimensional accuracy. It has the potential to be automated, and even
integrated into the production process to allow additive manufacture of useable parts without the
need for labour-intensive post-processing. Laser polishing offers flexibility in producing bespoke high
levels of surface polishing, as well as high dimensional accuracy. Several researchers have investigated
the most significant processing parameters on the final roughness and the resulting chemical and
mechanical properties for different metals and metal alloys [23–31]. These parameters include the
laser beam power, the scanning speed, the laser beam focal position, and the number of scanning
passes. The process melts a controlled, very thin localized region from the raised peaks of the part
surface, which is then relocated to the valleys of the surface profile, thereby reducing the overall
roughness [32,33]. Different surface profiles can be processed once the surface geometry is defined and
entered into the laser machine program. Moreover, selective processing can be carried out at discrete
locations in order to optimize processing speed and the resulting roughness. Without special optics,
however, the laser polishing of the internal surfaces is difficult. Also, in order to achieve symmetrical
and consistent Ra, the different geometries and alignments must be treated in a similar manner in order
not to be influenced unduly by the effect of gravity or the assist gas. Li et al. [33] investigated the laser
polishing of Ti-6Al-4V parts made using the selective laser melting (SLM) technology and achieved
a reduction in the roughness from 6.53 to 0.32 microns. The researcher indicated an improvement
in the surface mechanical properties such as the micro-hardness and the wear resistance. The laser
polishing of different metals has been investigated by several researchers in order to produce a better
understanding of the effect of the input processing parameters and to relate them to the modified
surface characteristics [34,35]. In order to fill the gap of information available for 316L stainless steel,
this paper presents a study of the effect of laser polishing parameters of 316L stainless steel alloy on
the resulting surface properties.
 Materials 2019, 12, x FOR PEER REVIEW 3 of 15
Materials 2019, 12, 991 3 of 15

2.1. Materials
2. Materials
Firstly,and Methods
316L stainless steel samples with cylindrical geometry were additively manufactured
and
2.1. supplied by 3D Systems, High Wycombe, UK, using an EOS M270 3D metal printer made by
Materials
EOS, Munich, Germany. The initial powder particle size distribution was 15–45 µm, the fiber laser
Firstly,
power was316L stainless
set to 195 W steel
withsamples with
a 1070 nm cylindrical ageometry
wavelength, weresize,
0.1 mm spot additively manufactured
and scanning and
speed of 750
supplied by 3D Systems, High Wycombe, UK, using an EOS M270 3D metal printer
mm/s. The samples were printed in the vertical direction along the main axis with a layer thickness made by EOS,
Munich,
of 40 µmGermany. The initial
to the final powder
cylinder particle
diameter of 10size
mm distribution
and lengthwas 15–45
of 60 mm.µm, the 1fiber
Table listslaser power
the chemical
was set to 195 W
composition of with a 1070 nm
the stainless wavelength,
steel powder baseda 0.1onmm spot percentage.
weight size, and scanning speed of 750 mm/s.
The samples were printed in the vertical direction along the main axis with a layer thickness of 40 µm
to the final cylinder diameter of 10
Table 1. The mm and
chemical length of 60
composition mm.
of the Table
316L 1 listssteel
stainless the powder.
chemical composition of
the stainless steel powder based on weight percentage.
Chemical Cr Cu Fe Mn Mo Ni P S Si C O N
Composition Table 1. The chemical composition of the 316L stainless steel powder.
Content 17.42 0.02 66.52 0.60 2.36 12.53 0.01 <0.01 0.51 0.02 0.05 0.06
Chemical
(wt.%) Cr Cu Fe Mn Mo Ni P S Si C O N
Composition
Content (wt.%) 17.42 0.02 66.52 0.60 2.36 12.53 0.01 <0.01 0.51 0.02 0.05 0.06
2.2. Experimental Set-Up
The laser polishing
2.2. Experimental Set-Up experimental set-up is shown in Figure 1. The laser used in this study was a
1.5 kW CO2 laser from Rofin, with a laser beam focus diameter of 0.2 mm. The experimental set-up
andThe
thelaser polishingscanning
overlapping experimental set-up
strategy is shownwere
employed in Figure 1. The
similar laserdescribed
to that used in this bystudy
Obeidi waset aal.
1.5[36,37].
kW CO 2 laser from
Cylindrical Rofin, with
samples wereaemployed
laser beam infocus diameter
this study of 0.2
instead mm.samples,
of flat The experimental
as with thisset-up
set-up,
and the overlapping scanning strategy employed were similar to that described
this allows for a high range of scanning speeds. The samples were rotated by means of a variable by Obeidi et al. [36,37].
Cylindrical
speed DCsamples
motor were
fromemployed in this study
Bodine Electric company,instead of flat samples,
Chicago, IL, USA, as with
whichthisprovided
set-up, this allowsof
a range
forrotational
a high range
speedofofscanning speeds.
0–5000 rpm, whileThe thesamples
samplewereand therotated by means
DC motor of a variable
assembly speedon
were carried DC the
motor from Bodine Electric company, Chicago, IL, USA, which provided a
machine positioning stage, which provided a translational speed range of 0–5000 mm/min. The range of rotational speed of
0–5000 rpm, while
combination theset
of the sample and the
rotational and DC motor assembly
translational speeds were carried
allows the on the laser
fixed machinebeampositioning
to scan the
stage, which provided a translational speed range of 0–5000 mm/min. The
entire sample surface in a spiral track with the ability to define the extent of overlap between combination of the set
each
rotational and translational speeds allows the fixed laser beam to scan the entire
track. Argon gas was continuously supplied in the coaxial direction, in-line with the CO2 laser beam sample surface in
a spiral track with
at a pressure of 50the
kPaability to define
and flow rate ofthe30extent of overlap
liters per between
min in order each track.
to provide inertArgon gas was
gas surrounding
continuously supplied
and avoid oxidation. in the coaxial direction, in-line with the CO 2 laser beam at a pressure of 50 kPa
and flow rate of 30 liters per min in order to provide inert gas surrounding and avoid oxidation.

CO2 Laser
Assist Gas

Translational
Rotational
Movement
Movement

Polished
Surface

Figure 1. Schematic diagram for the CO2 laser polishing scan process of additive manufacturing (AM)
produced stainless steel 316L cylindrical samples.
Figure 1. Schematic diagram for the CO2 laser polishing scan process of additive manufacturing
Carbon dioxide laser was used in this study because it is commonly used in industry owing to the
(AM) produced stainless steel 316L cylindrical samples.
high output power of 0.1 to 50 kW, which can compensate the poor laser–material interaction caused
by the long wavelength.
Carbon Despite
dioxide laser was the
usedfact
in that
this other
studytypes of lasers
because like ND: YAG
it is commonly used(neodymium-doped
in industry owing to
yttrium aluminum
the high garnet)ofand
output power fiber
0.1 to 50 lasers are also
kW, which canexpanding
compensate in the
different industrial applications,
poor laser–material interaction
the CO laser is more reliable for the lower initial cost.
caused by the long wavelength. Despite the fact that other types of lasers like ND: YAG
2
(neodymium-doped yttrium aluminum garnet) and fiber lasers are also expanding in different
industrial applications, the CO2 laser is more reliable for the lower initial cost.
Materials 2019, 12, x FOR PEER REVIEW 4 of 15
Materials 2019, 12, 991 4 of 15

2.3. Method
2.3.AMethod
preliminary test was used to identify the most significant processing parameters and their
levels, with a view of keeping
A preliminary test wasthe process
used just above
to identify the melting
the most pointprocessing
significant in order to parameters
avoid any ablation
and their
of levels,
the material or over melting. Figure 2 shows a schematic diagram for understanding
with a view of keeping the process just above the melting point in order to avoid any ablation the
re-melting process in which the power density and the residence time are adjusted to melt
of the material or over melting. Figure 2 shows a schematic diagram for understanding the re-melting only the
high peaksinofwhich
process the material.
the powerThedensity
roughness peaks,
and the for example,
residence time areatadjusted
position to
(a)melt
in Figure 2, would
only the then of
high peaks
be the
melted
material. The roughness peaks, for example, at position (a) in Figure 2, would then be meltedfor
and the molten material would move and re-solidify within the lower valleys, and
example, at position (b).
the molten material would move and re-solidify within the lower valleys, for example, at position (b).

Rough Maximum
Surface Melting Range (a)

Polished
(b)
Surface
Figure 2. The principle used in the laser polishing process by melting the metal between (a) and
Figure 2. The principle used in the laser polishing process by melting the metal between (a) and (b) [37].
(b) [37].

The experimental
The experimental work
work was divided
was divided into two
into design
two designof of
experiment
experiment (DoE)
(DoE) models
models based
basedononthethe
Box Behnken design. In the first model, the process parameters investigated
Box Behnken design. In the first model, the process parameters investigated were the laser power were the laser power
(W) in in
(W) continuous
continuous mode,
mode,rotational
rotational speed
speed (rpm),
(rpm),and thethe
and number
number of of
repeated
repeated passes. In In
passes. thethe
second
second
model,
model,thethe
optimum
optimumvalues
valuesofofthe
the laser
laser beam power (110
beam power (110W)W)andandthethenumber
number of of repeated
repeated passes
passes (one)
(one)
werewere adopted
adopted fromfrom the first
the first model. model. The rotational
The rotational speedspeed wasexamined
was also also examined
in the in the second
second module,
module,
as wellasaswell
twoasnewtwo parameters,
new parameters, the focal
the focal position
position (below,
(below, on, on,
andand above
above thethe sample
sample surface)
surface) and
andthe percentage overlap of the laser spot in the axial direction, as shown in Figure 3. Table 2 2lists
the percentage overlap of the laser spot in the axial direction, as shown in Figure 3. Table liststhe
theprocessing
processingparameters
parametersand andtheir
theirlevels
levelsinvestigated.
investigated.No Nosignificant
significant effect
effect forfor the
the rotational
rotational speed
speed was
was noted in the results of DoE-1 compared with the other two parameters
noted in the results of DoE-1 compared with the other two parameters in DoE-1. The speed parameter in DoE-1. The speed
parameter
was thuswas thus expanded
expanded to a testingto arange
testing range between
between 20 toin80DoE-2.
20 to 80 rpm rpm in DoE-2.

Table 2. 2.
Table The laser
The processing
laser parameters
processing implemented
parameters implementedfor
fordesign
designofofexperiment
experiment(DoE)
(DoE)model
model11and
and 2,
2, both according to the Box Behnken model.
both according to the Box Behnken model.
DoE-1
DoE-1
Processing parameter Level 1 Level 2 Level 3
Processing parameter Level 1 Level 2 Level 3
Power
Power(W)
(W) 90 90 110110 130130
Rotational speed
Rotational (rpm)
speed (rpm) 30 30 50 50 70 70
No.
No. of of passes
passes 1 1 2 2 3 3
DoE-2
DoE-2
Overlap (%) −20 0 20
Overlap (%) (mm)
Focal position −20
−0.5 (beneath surface) 0
0 (on surface) 20
0.5 (above surface)
Focal position
Rotational (mm)
speed (rpm) −0.5 (beneath20surface) 0 (on surface)
50 0.5 (above surface)
80
Rotational speed (rpm) 20 50 80
The percentage overlap processing parameter here means the percentage of the beam spot
diameter that goes overlap
The percentage over the processing
previous laser track that
parameter has means
here alreadythe
been processedof
percentage [16].
theThree
beampossible
spot
overlapping
diameter scenarios
that goes were
over the applied:
previous lasernegative,
track thatzero, and positive
has already overlapping
been processed laser
[16]. tracks.
Three Here,
possible
the negative
overlapping overlap
scenarios indicates
were applied:unprocessed
negative, zero,gaps
andbetween the consecutive
positive overlapping laserpasses; conversely,
tracks. Here, the
the positive
negative overlap
overlap meansunprocessed
indicates the laser tracks
gapsinterfere
betweenwith each
the other; and passes;
consecutive the zeroconversely,
overlap indicates
the
that the laser tracks just touch each other tangentially—see Figure 3.
positive overlap means the laser tracks interfere with each other; and the zero overlap indicates that
the laser tracks just touch each other tangentially—see Figure 3.
Materials 2019, 12, 991 5 of 15
Materials 2019, 12, x FOR PEER REVIEW 5 of 15

Figure
Figure 3. The
3. The three
three overlappingscenarios
overlapping scenariosused
usedfor
forsurface
surfacelaser
laserscanning
scanningpasses,
passes, determined
determined by
by setting
setting the rate of longitudinal translation of the
the rate of longitudinal translation of the laser beam. laser beam.

The The roughnessprofile

roughness profile was
was measured
measuredbybymeans means of of
non-contact,
non-contact,3D microscope,
3D microscope, from Keyence
from Keyence
VHX2000E 3D digital, Milton Keyense, UK and Bruker Contour GT-X Profilometer (Billerica, MA,
VHX2000E 3D digital, Milton Keyense, UK and Bruker Contour GT-X Profilometer (Billerica, MA,
USA). At least five surface roughness profiles were obtained for each sample in the DoE parallel to
USA). At least five surface roughness profiles were obtained for each sample in the DoE parallel to the
the cylinder axis for a length of 4.75 mm and the average surface roughness was calculated from
cylinder axis for a length of 4.75 mm and the average surface roughness was calculated from these
these profiles along with the 95% confidence intervals (CI). The initial average roughness, Ra, for the
profiles along withwas
as-built samples the measured
95% confidence intervals (CI). The initial average roughness, Ra, for the as-built
at 10.4 µm.
samplesThe wascross-sectional
measured atmicrostructure
10.4 µm. was also investigated for as-received and laser-processed
The additive
metal cross-sectional microstructure
manufactured samples inwas also
order toinvestigated for as-received
compare the effects and processing.
of the sample laser-processed
The metal
samples
additive were sliced using
manufactured samples a diamond
in order todisc cutter of
compare the0.5 mm of
effects thickness
the sampleand processing.
were polished Thebysamples
wereremoving a layerathickness
sliced using diamondofdiscno less thanof0.5
cutter 0.5mmmm in thickness
order to avoid
and the effect
were of the cutting
polished process a layer
by removing
on the microstructure. Silicon carbide papers with different grades of 400, 600, 800, and
thickness of no less than 0.5 mm in order to avoid the effect of the cutting process on the microstructure. 1200 were
used for the grinding with continuous water stream for flushing the loose and abrasive particles,
Silicon carbide papers with different grades of 400, 600, 800, and 1200 were used for the grinding
using a Metkon Forcimat grinder-polisher, Bursa, Turkey. The polishing process was conducted
with continuous water stream for flushing the loose and abrasive particles, using a Metkon Forcimat
using diamond suspension with 9, 6, 3, and 0.05 micron particle size on a Struers Textmet cloth from
grinder-polisher, Bursa, Turkey. The polishing process was conducted using diamond suspension
MetPrep, Coventry, UK. Each polishing grade was applied for three minutes at a rotational speed of
with3009,rpm.
6, 3, and 0.05 micron particle size on a Struers Textmet cloth from MetPrep, Coventry, UK.
Each polishing
The effectgrade
of thewas
laserapplied
process for three
on the minutes
sample at a rotational
hardness speed ofin300
was also measured therpm.
cross-sectional
The effect of the laser process on the sample hardness was also
direction. The aforementioned grinding and polishing process was applied to remove measured innothe cross-sectional
less than a
0.5 mm The
direction. layeraforementioned
in order to avoidgrinding
the effect and
of slicing and toprocess
polishing guarantee
wasthe removal
applied toof any residual
remove no less than
stresses.
a 0.5 mm layerLeitzinWetzlar
order toGermany
avoid the Vickers
effect micro-hardness
of slicing and to tester (Model:the
guarantee 301-252.001,
removal of Wetzlar,
any residual
D-35578,
stresses. Germany
Leitz Wetzlar was used to measure
Germany Vickersthe hardness of thetester
micro-hardness as-built and the301-252.001,
(Model: laser processed samples
Wetzlar, D-35578,
in the cross-sectional direction. The measurements start from the sample surface and move toward
Germany was used to measure the hardness of the as-built and the laser processed samples in the
the center in three parallel lines in order to compare the hardness of the melted and re-solidified
cross-sectional direction. The measurements start from the sample surface and move toward the center
layer with the unprocessed bulk material.
in three parallel lines in order to compare the hardness of the melted and re-solidified layer with the
unprocessed
3. Results bulk material.
The average roughness, Ra, of the as-built metal AM samples was found to be 10.4 µm, and the
3. Results
arithmetical mean height (Sa) was 25 µm. The maximum and the average of maximum height of the
The average
profile Rt and Rz roughness,
are equalRa, of the
in this as-built
case because metal
one AM samples
single roughnesswasprofile
found of to 4.75
be 10.4
mmµm,wasand the
arithmetical
measured meanin five height (Salocations
different ) was 25 and The maximum
µm.averaged (µm). Forandthethe average
as-built of maximum
samples, this value height
was of the
found
profile Rttoand
be 38.77
Rz are µm.
equal in this case because one single roughness profile of 4.75 mm was measured
As mentioned
in five different previously,
locations the roughness
and averaged (µm).profile
For theis constructed from, firstly,
as-built samples, the topography
this value was found to be
caused by the melting of the consecutive layers and, secondly, the adhesion of adjacent and partially
38.77 µm.
melted powder particles at the boundaries. Figure 4 shows a 3D microscope image of the AM sample
As mentioned previously, the roughness profile is constructed from, firstly, the topography caused
after laser polishing of 5 mm length along the main axis of the cylindrical sample. The polished
by sample
the melting of the consecutive layers and, secondly, the adhesion of adjacent and partially melted
was processed with 110 W, 20 rpm, 20% percentage overlapping laser scans (OV), and laser
powder particles
beam focus on the at the boundaries.
surface. The scaleFigure
on the4 z-axis
showsalsoa 3D microscope
includes image of
the curvature of the
the AM sample after
cylindrical
laser polishing
geometry of ofthe5 sample.
mm length along
Figure 5a–ctheshow
mainscanning
axis of the cylindrical
electron sample.
microscopy SEM The polished
images sample
taken by was
processed
means ofwith
(Model:110EVO
W, 20 LSrpm, 20%
15-0723 percentage
from Carl Zeissoverlapping
Ltd, Cambridge, laser scans
UK) (OV), and
microscope, for laser beam focus
the surface
on morphology
the surface.and Thethe cross-sectional
scale on the z-axis view. In includes
also this figure,
thezones (1) and of
curvature (2) the
show the rough geometry
cylindrical surface of of the
the as-built parts, which is mainly caused by the sintering of the adjacent
sample. Figure 5a–c show scanning electron microscopy SEM images taken by means of (Model: powder particles in the EVO
balling effect, which is similar to micro-welding. Zone (3) shows voids in
LS 15-0723 from Carl Zeiss Ltd., Cambridge, UK) microscope, for the surface morphology and thethe build part represented
by unfused powder particles, which could be because of a lack of fusion.
cross-sectional view. In this figure, zones (1) and (2) show the rough surface of the as-built parts, which
is mainly caused by the sintering of the adjacent powder particles in the balling effect, which is similar
to micro-welding. Zone (3) shows voids in the build part represented by unfused powder particles,
which could be because of a lack of fusion.
Materials 2019, 12, x FOR PEER REVIEW 6 of 15

The processed samples exhibit a wide range of variation of the roughness depending on the
Materials 2019, 12, 991
input parameters. The presence of the hard martensitic isolated islands was observed on6 ofthe 15

re-solidified surface
Materials 2019, surrounded
12, x FOR PEER REVIEWby the soft austenite, as can be seen in Figure 6. The measured
6 of 15
micro-hardness was found to be ~234 HV on the as-built and processed samples.
The processed samples exhibit a wide range of variation of the roughness depending on the input
The processed samples exhibit a wide range of variation of the roughness depending on the
parameters. The presence of the hard martensitic isolated islands was observed on the re-solidified
input parameters. The presence of the hard martensitic isolated islands was observed on the
surface surrounded
re-solidified by the
surface soft austenite,
surrounded by the as can
soft be seenasincan
austenite, Figure 6. The
be seen measured
in Figure micro-hardness
6. The measured
was found to be ~234 HV on the as-built and processed samples.
micro-hardness was found to be ~234 HV on the as-built and processed samples.

Figure 4. Three-dimensional (3D) microscope image of 316L SST AM sample after CO2 laser
polishing with 110 W, 20% percentage overlapping laser scans (OV), 20 rpm, and laser beam focused
Figure
on 4. Three-dimensional
theFigure
surface. (3D) (3D)
4. Three-dimensional microscope image
microscope of 316L
image SST AM
of 316L SST sample after CO
AM sample 2 laser
after CO2 polishing
laser
with polishing
110 W, 20%with percentage overlapping
110 W, 20% percentage laser scans
overlapping laser (OV), 20 rpm,
scans (OV), and
20 rpm, laser
and laserbeam focused on
beam focused
No significant change in the micro-hardness of the polished samples was observed despite the
on the surface.
the surface.
fact of the hard martensite formation. This can be explained by the small amount of the hard phase
No andNo significant
significant change
change in the micro-hardness
in the of of
thethe polishedsamples
sampleswas
was observed despite
despite the fact
formed surrounded by the micro-hardness
soft austenite phase, polished
which can absorb theobserved
hardness testerthe probe
of thefact
hardof martensite
the hard martensite formation.
formation. This canThis
be can be explained
explained by theby the small
small amount
amount of of hard
the the hard phase
phase formed
impression and show no alteration in the hardness—see Figure 6. This result agrees well with the
and formed and surrounded
surrounded by the soft by the softphase,
austenite austenite phase,
which canwhich
absorbcan
theabsorb the hardness
hardness tester testerimpression
probe probe
results achievedand
impression by Obeidi
show noetalteration
al. [26] and Kato
in the et al. [38]. Figure 6. This result agrees well with the
hardness—see
and show no alteration in the hardness—see Figure 6. This result agrees well with the results achieved
results achieved by Obeidi et al. [26] and Kato et al. [38].
by Obeidi et al. [26] and Kato et al. [38].

(a)
(a)

(b)

(b)

Figure 5. Cont.
Materials 2019, 12, 991 7 of 15
Materials 2019, 12, x FOR PEER REVIEW 7 of 15
Materials 2019, 12, x FOR PEER REVIEW 7 of 15

(c)
(c)
Figure
Figure 5. 5. (a) Scanning electron microscopy (SEM) micro-image of the the metal
metal AM
AM sample
sample surface
surface
Figure 5. (a) Scanning electron microscopy (SEM) micro-image of the metal AM sample surface
showing
showing (a) (a) the
the surface
surface morphology
morphology with
with the
the fully
fully melted
melted material
material in
in the
the back
back ground,
ground, zones
zones (1)
(1) and
and
showing (a) the surface morphology with the fully melted material in the back ground, zones (1) and
(2) show the BALLING effect and the partially melted and adhered neighboring particles,
(2) show the BALLING effect and the partially melted and adhered neighboring particles, (b) shows (b) shows
(2) show the BALLING effect and the partially melted and adhered neighboring particles, (b) shows
the
the cross-section
cross-sectionSEM SEM micrograph
micrographview of the
view of same surface,
the same zone (3)
surface, shows
zone (3)unmelted powder particles,
shows unmelted powder
the cross-section SEM micrograph view of the same surface, zone (3) shows unmelted powder
and (c) shows
particles, a high
and (c) magnification
shows from the micrograph
a high magnification shown in (b).
from the micrograph shown in (b).
particles, and (c) shows a high magnification from the micrograph shown in (b).

Figure 6.
Figure 6. SEM
SEM image
image of
of AM
AM 316L
316L SST
SST sample
sample showing
showing thethe same
same sample
sample surface
surface before
before (left)
(left) and
and after
after
Figure 6. SEM image of AM 316L SST sample showing the same sample surface before (left) and after
(right) laser
(right) laser polishing with 130 W, 30 rpm, and two passes repetition.
(right) laser polishing
polishing with
with 130
130 W,
W, 30
30 rpm,
rpm, and
and two
two passes
passes repetition.
repetition.

Figure 7a
Figure 7a shows
showsthe theplan
planview
view(top
(topdown)
down)and
andFigure
Figure7b,c
7b,cshow
showthe theside-view
side-view SEM
SEM images
images forfor
a
Figure 7a shows the plan view (top down) and Figure 7b,c show the side-view SEM images for
a sample
sample processed
processed with5050rpm,
with rpm,0%0%overlap,
overlap,oneonepass,
pass,and
and90 90W.W.Figure
Figure7d 7dshows
showsthe theplan
plan view
view and
and
a sample processed with 50 rpm, 0% overlap, one pass, and 90 W. Figure 7d shows the plan view and
Figure 7c,f
Figure 7c,f show
show the the side-view
side-view SEM images for a sample processed at 50 rpm, 0% overlap, one pass,
Figure 7c,f show the side-view SEM images for a sample processed at 50 rpm, 0% overlap, one pass,
and 130
and 130 W.
W.Because
Because of ofthe
theGaussian
Gaussian energy
energydistribution
distribution ofof the
theCOCO2 laser
laser beam
beam inin the
the cross-sectional
cross-sectional
and 130 W. Because of the Gaussian energy distribution of the CO22 laser beam in the cross-sectional
direction, the
direction, the samples
samples showshow unmelted
unmelted gapsgaps between
between thethe laser
laser passes
passes when
when thethe lower
lower power
power level
level of
of
direction, the samples show unmelted gaps between the laser passes when the lower power level of
90 W
90 Wwas
wasemployed,
employed,asasshown showninin Figure
Figure 7a–c.
7a–c. Conversely,
Conversely, when when
the the higher
higher power
power levellevel of W
of 130 130wasW
90 W was employed, as shown in Figure 7a–c. Conversely, when the higher power level of 130 W
was applied,
applied, a cleara clear interference
interference of theofmelt-pools
the melt-pools can
can be be observed—see
observed—see Figure Figure
7d,c,f.7d,c,f.
was applied, a clear interference of the melt-pools can be observed—see Figure 7d,c,f.
In the
In the cross-sectional
cross-sectional direction,
direction, the
the melted
melted and
and re-solidified
re-solidified layer
layer thickness
thickness was found to vary vary
In the cross-sectional direction, the melted and re-solidified layer thickness was found to vary
from 10
from 10 to
to 80
80 µm.
µm. Figure
Figure 8a8a shows
shows thethe peripheral
peripheral external
external edge
edge ofof the
the sample
sample in in which
which most
most ofof the
the
from 10 to 80 µm. Figure 8a shows the peripheral external edge of the sample in which most of the
surface undulations
surface undulationsand andthetheadhered
adheredpowder
powderparticles
particlespresent
present inin Figure
Figure 5b5b were
were removed.
removed. Because
Because of
surface undulations and the adhered powder particles present in Figure 5b were removed. Because
the polishing process, a reduction in the sample diameter from 10.0 ± 0.015 mm to 9.97 ± 0.006 mm
of the polishing process, a reduction in the sample diameter from 10.0 ± 0.015 mm to 9.97 ± 0.006 mm
of the polishing process, a reduction in the sample diameter from 10.0 ± 0.015 mm to 9.97 ± 0.006 mm
was measured.
was measured. The The resulting
resultingreduction
reductionin insample
sampleheight
heightafter
afterpolishing
polishingcompared
comparedwith withbefore
beforecan
canbe
was measured. The resulting reduction in sample height after polishing compared with before can
seen in Figure 8b.

(a) (b) (c)

(a) (b) (c)
was applied, a clear interference of the melt-pools can be observed—see Figure 7d,c,f.
In the cross-sectional direction, the melted and re-solidified layer thickness was found to vary
from 10 to 80 µm. Figure 8a shows the peripheral external edge of the sample in which most of the
surface undulations and the adhered powder particles present in Figure 5b were removed. Because
of the polishing
Materials process, a reduction in the sample diameter from 10.0 ± 0.015 mm to 9.97 ± 0.0068 mm
2019, 12, 991 of 15
was measured. The resulting reduction in sample height after polishing compared with before can

Materials 2019, 12, x FOR PEER REVIEW 8 of 15

(a) (b) (c)
Materials 2019, 12, x FOR PEER REVIEW 8 of 15

(d) (e) (f)

(d) (e) (f)
SEM micrograph
Figure 7. SEM micrographshowing
showingthe
thetop
topand
andside
sideviews
viewsofofAM
AM 316L
316L SST
SST samples
samples polished
polished with
with 50
Figure
50
rpm,rpm, 7.
0%0% SEM micrograph
overlap,
overlap, oneone showing
scanning
scanning the
pass,
pass, top
and
and and side
9090WWinin(a)views
(a)top, of AM
top,(b),
(b),and316L
and(c) SST samples
(c)side-view,
side-view,as polished
as well
well as with 50
as 130 W in
rpm,
(d) 0%(e),
top, overlap,
and (f)one
sidescanning
side view . pass, and 90 W in (a) top, (b), and (c) side-view, as well as 130 W in
view.
(d) top, (e), and (f) side view .

(a) (b)
(a) (b)
Figure 8. SEM micrograph shows (a) the cross-sectional view of the polished sample and (b) the side
Figure 8. SEM
view. This micrograph
sample shows (a)
was processed the 110
with cross-sectional view ofand
W, 20% overlap, the20
polished
rpm, andsample and (b)
the laser the side
beam was
view.
view. This
This sample
samplewas
was processed with
processed
focused on the sample surface. 110
with W,
110 20%
W, overlap,
20% and
overlap, 20
andrpm,
20 and
rpm,the laser
and thebeam was
laser focused
beam was
on the sample
focused on thesurface.
sample surface.
Figure 9 shows an unmelted particle on the polished sample surface, which results from an
Figure 99 shows
Figure an
an unmelted
showssimilar
unmelted particle
particle on the polished sample
sample surface, which results
results from
from an
inconsistent surface to that in Figureon5a the
zonepolished
(2). surface, which an
inconsistent surface similar to that in Figure 5a zone (2).
inconsistent surface similar to that in Figure 5a zone (2).
Figure 10a shows the austenitic microstructure of the bulk material and the build layers boundaries
for the AM sample. Figure 10b shows the modified layer in the polished sample. The microstructure of
316L SST is austenite at room temperature, and the large grains, can grow even through the different
melt-pool boundaries, as shown in area (1) in Figure 10a.
The response surface model (RSM) graphs obtained from DoE-1 show that the lowest roughness
values were always obtained when using the laser beam power of 110 W with one scanning pass—see
Figure 11a. In the DoE-2 test, the laser beam power and the number of passes were kept constants at
110 W and one pass, respectively, and two new parameters employed were the percentage overlap
and the laser beam focal position. A noticeable improvement was found in the produced roughness.
Predominantly, the positive overlap, when applied in combination with the lower rotational speed

Figure 9. SEM image of a remaining unmelted particle due to lack of fusion.

Figure 9. SEM image of a remaining unmelted particle due to lack of fusion.
Figure 10a shows the austenitic microstructure of the bulk material and the build layers
Figure 10athe
boundaries for shows
AM the austenitic
sample. Figure microstructure
10b shows the of the bulk
modified material
layer in the and the build
polished layers
sample. The
boundaries for the AM sample. Figure 10b shows the modified layer in the polished sample. The
 (a) (b)

Figure 8. SEM micrograph shows (a) the cross-sectional view of the polished sample and (b) the side
view.
Materials This
2019, sample was processed with 110 W, 20% overlap, and 20 rpm, and the laser beam was9 of 15
12, 991
focused on the sample surface.

level,Figure
results9inshows
a low Ra. No significant
an unmelted effecton
particle forthe
thepolished
focal position wassurface,
sample observed for the
which rangefrom
results applied.
an
Lower roughness was also found at the lowest level
inconsistent surface similar to that in Figure 5a zone (2).of rotation speed.

Materials 2019, 12, x FOR PEER REVIEW 9 of 15

Materials 2019, 12, x Figure

FOR PEER REVIEW
9. SEM image of a remaining unmelted particle due to lack of fusion. 9 of 15

Figure 10a shows the (a) austenitic microstructure of the bulk material
(b) and the build layers
boundaries for the AM sample. Figure 10b shows the modified layer in the polished sample. The
Figure 10. SEM micrograph of the cross-section of the (a) as-received and (b) re-melted modified
microstructure of 316L SST is austenite at room temperature, and the large grains,
layer of the polished sample processed with 110 W, 20 rpm, 20% OV, and the laser beam can grow
focaleven
through the different
positioned melt-pool
on the sample boundaries, as shown in area (1) in Figure 10a.
surface.

The response surface model (RSM) graphs obtained from DoE-1 show that the lowest
roughness values were always obtained when using the laser beam power of 110 W with one
scanning pass—see Figure 11a. In the DoE-2 test, the laser beam power and the number of passes
were kept constants at 110 W and one pass, respectively, and two new parameters employed were
the percentage overlap and the laser beam focal position. A noticeable improvement was found in
the produced roughness. (a) Predominantly, the positive overlap, when applied (b) in combination with
the lower rotational speed level, results in a low Ra. No significant effect for the focal position was
Figure 10.
10. SEM
SEMmicrograph
micrographofofthe
the cross-section
cross-section of the
of the (a) as-received
(a) as-received and and (b) re-melted modified
(b) re-melted
observed for the range applied. Lower roughness was also found at the lowestmodified
level oflayer
rotation
layer
of the of the polished
polished sample sample processed
processed with 110with
W, 20110
rpm,W,20%
20 OV,
rpm,and20%
theOV,
laserand thefocal
beam laser beam focal
positioned on
speed.
positioned
the sample on the sample surface.
surface.

The response surface model (RSM) graphs obtained from DoE-1 show that the lowest
roughness values were always obtained when using the laser beam power of 110 W with one
scanning pass—see Figure 11a. In the DoE-2 test, the laser beam power and the number of passes
were kept constants at 110 W and one pass, respectively, and two new parameters employed were
the percentage overlap and the laser beam focal position. A noticeable improvement was found in
the produced roughness. Predominantly, the positive overlap, when applied in combination with
the lower rotational speed level, results in a low Ra. No significant effect for the focal position was
observed for the range applied. Lower roughness was also found at the lowest level of rotation
speed.

(a) (b)

Figure 11.
Figure 11. Response
Response surface
surfacemodel
model(RSM)
(RSM)graphs
graphs(a)
(a)related
relatedtotodesign
designofofexperiment
experiment (DoE)-1
(DoE)-1 inin Table
Table 2,
2, showing
showing thethe interaction
interaction effect
effect on of
on Ra Rathe
of the laser
laser beambeam power
power and and the number
the number of passes
of passes at 50atrpm;
50 rpm;
and
andrelated
(b) (b) related to DoE-2,
to DoE-2, the effects
the effects of rotational
of rotational scanning
scanning speedspeed andpercentage
and the the percentage overlap,
overlap, with
with the the
focal
focal position
position set onset
theon the sample
sample surface.
surface.

The
The surface
surface profile
profile was
was investigated
investigated as
as formerly
formerly explained
explained in one
one cut-off
cut-off with
with the
the measured
measured
length of 4.75 mm. Table 3 lists the processing parameters applied with the resulting measured
length of 4.75 mm. Table 3 lists the processing parameters applied with the resulting measured average
average roughness Ra and the average maximum height Rz, which is equal to the maximum height of
the profile in this case of one cut-off profile.
(a) (b)
Table 3. The average roughness, Ra, and the maximum and average height, Rt and Rz, respectively, of
Figure 11. Response
the measured surface
roughness model (RSM) graphs (a) related to design of experiment (DoE)-1 in Table
profile.
2, showing the interaction effect on Ra of the laser beam power and the number of passes at 50 rpm;
Materials 2019, 12, x FOR PEER REVIEW 10 of 15

Materials 2019, 12, 991 10 of 15

1 0 50 0 5.14 34.05
2 0 80 −0.5 5.04 32.57
roughness Ra 3and the average−20
maximum 50
height Rz , −0.5
which is equal to the maximum
6.08
height of the
39.38
profile in this case of one cut-off profile.
4 0 50 0 4.98 27.71
5 average roughness,
Table 3. The 20 Ra , and20 0 average height,
the maximum and 2.74 Rt and Rz , respectively,
47.17 of
the measured
6 roughness 0profile. 20 −0.5 4.16 33.36
7 Sample No.0 OV 50
Rotational 0
Speed F.P 4.99
Ra Rz , Rt 27.71
8 1 20 0 80 50 0 0 4.26
5.14 34.05 39.92
9 2 0 0 50 80 0 − 0.5 5.04
5.77 32.57 30.45
3 −20 50 −0.5 6.08 39.38
10 4 −20 0 80 50 0 0 6.90
4.98 27.71 38.38
11 5 20 20 50 20 −0.5 0 2.74
4.11 47.17 47.44
6 0 20 −0.5 4.16 33.36
12 7 −20 0 50 50 0.5 0 5.44
4.99 27.71 37.81
13 8 20 20 50 80 0.5 0 4.26
3.47 39.92 45.86
9 0 50 0 5.77 30.45
14 10 0 −20 20 80 0.5 0 4.43
6.90 38.38 32.72
15 11 −20 20 20 50 0 − 0.5 4.11
3.99 47.44 48.28
12 −20 50 0.5 5.44 37.81
16 13 0 20 50 50 0 0.5 4.82 3.47 45.86 29.01
17 14 0 0 80 20 0.5 0.5 6.03 4.43 32.72 39.54
15 −20 20 0 3.99 48.28
16 0 50 0 4.82 29.01
where OV is the percentage overlapping laser scans (%); Ra is the average roughness in (µm); Rt and
Rz are the maximum 17 and average 0 of maximum 80 height of the 0.5 profile,
6.03respectively,
39.54 in (µm); the
where OV isspeed
rotational the percentage overlapping
is in (rpm); and F.P laser
is thescans
laser(%); Ra isfocal
beam the average roughness
position in (mm).inRa,
(µm);
Rz,Rtand
and (Rt)
Rz are the
were
maximum and average of maximum height of the profile, respectively, in (µm); the rotational speed is in (rpm);
averaged from five profile measurements.
and F.P is the laser beam focal position in (mm). Ra, Rz, and (Rt) were averaged from five profile measurements.

Figure 12 shows the response

response surface
surface method
method (RSM)
(RSM) plot
plot for
for RRtt and
and RRzz with the overlapping
percentage and the rotational speed corresponding to the data listed in Table 3 when the laser beam
focal is positioned on the sample surface.

Figure 12.
Figure RSM plot
12. RSM plot correlating
correlating the
the R with the
Rzz with the processing
processing parameters
parameters in
in DoE-2.
DoE-2.

Figure 13 shows the optical micrographs extracted by Bruker Contour GT-X for the surface
Figure 13 shows the optical micrographs extracted by Bruker Contour GT-X for the surface
topography of the as-built and a laser polished sample processed with 20% OV, 20 rpm, and the laser
topography of the as-built and a laser polished sample processed with 20% OV, 20 rpm, and the laser
beam focal positioned on the sample surface. The surface analysis of these data shows a noticeable
beam focal positioned on the sample surface. The surface analysis of these data shows a noticeable
reduction in the arithmetical mean height (S ) after laser polishing. The measured value of (S ) was
reduction in the arithmetical mean height (Saa) after laser polishing. The measured value of (Saa) was
reduced from 25 µm for the as-built to 11 µm after laser polishing.
reduced from 25 µm for the as-built to 11 µm after laser polishing.
Materials 2019, 12, x FOR PEER REVIEW 11 of 15
Materials 2019, 12, 991 11 of 15
Materials 2019, 12, x FOR PEER REVIEW 11 of 15

(a) (b)

Figure 13. Three-dimensional

(a) (3D) optical images of the surface topography (b)of (a) as-built and (b)
laser polished samples.
13. Three-dimensional
Figure 13. Three-dimensional(3D)
(3D)optical
optical images
images of of
thethe surface
surface topography
topography of (a)ofas-built
(a) as-built andlaser
and (b) (b)
laser polished
polished
4. Discussion samples.
samples.

4.
4. Discussion
It is mentioned previously that depositing the consecutive metal layers on a curvature or
Discussion
inclination
It of geometrypreviously
will results in increasing the roughness as a result of the Stair aSteps effect in
It is
is mentioned
mentioned previously that that depositing
depositing the the consecutive
consecutive metal
metal layers
layers on on a curvature
curvature or or
addition
inclination to the roughness
of geometry
geometry caused
will resultsby inthe sinteredthe
increasing neighboring
roughnessas partially
asa aresult
resultfused powder particles.
inclination of will results in increasing the roughness ofofthethe Stair
Stair Steps
Steps effect
effect in
Meanwhile,
in addition printing
to the in the vertical
roughness caused direction
by the minimizes
sintered the stair
neighboring steps fused
partially effect powder
on the particles.
surface
addition to the roughness caused by the sintered neighboring partially fused powder particles.
roughness,
Meanwhile, which is simply similardirection
to the waving surface. Figure 15 below shows bothroughness,
printing
Meanwhile, printing
printinginin thethe
vertical
vertical directionminimizes the stair
minimizes thesteps
staireffect
stepson effect
the surface
on the surface
scenarios.
which is simply similar to the similar
wavingto surface. Figuresurface.
15 belowFigure
shows15 both printing scenarios.
roughness, which is simply the waving below shows both printing
InInthis
thisstudy,
study,thetheAM AMsamples
sampleswere werebuilt
builtininthe
thevertical
verticaldirection,
direction,asasshown
shownininFigureFigure14a,
14a,sosothe
the
scenarios.
waving
waving effect
effect on
on the
the surface
surface waswas limited
limited and and thermal
thermal energyenergy
was was
mainly mainly
applied applied
to melt to
themelt the
partially
In this study, the AM samples were built in the vertical direction, as shown in Figure 14a, so the
partially
fused fused particles
particles adhered adhered
on thewas on the surface.
surface. Laser Laser polishing of parts withgeometry,
curved geometry,in
waving effect on the surface limited andpolishing
thermal of parts with
energy curved
was mainly applied to shown
melt the
shown
Figure in Figure
14b, might 14b, might require
requireadhered
a small amounta small amount
of thermal of thermal
energy, energy,
higherofthan higher
that than that in Figure
in Figure
partially fused particles on the surface. Laser polishing parts with curved14a, or may
geometry,
14a, or may
require require multi-passes
multi-passes of laser beam ofscan
laser beam the
because scanprocess
becauseis thethe
in process is in of
micro-scale thepowder
micro-scale of
particles
shown in Figure 14b, might require a small amount of thermal energy, higher than that in Figure
powder
and particles and layer thickness.
14a, layer
or may thickness.
require multi-passes of laser beam scan because the process is in the micro-scale of
powder particles and layer thickness.
Un-fused powder
particles (Balling)
Un-fused powder
particles (Balling)
Printed layers
Stair
Printed layers
steps
Stair
Build platform
steps
(a) (b) Build platform
Figure 14.14.
Figure (a)(a)
Printing
(a) inin
Printing the vertical
the direction
vertical and
direction and(b)
(b)printing
printingcurvature
(b) curvaturegeometry.
geometry.

Theapplied
The Figurethermal
applied 14. (a) Printing
thermal energywas
energy in theadjusted
was vertical direction
adjusted totobebejustand
just (b)
atatthe printing
themelting curvature geometry. in order to
meltingthreshold—firstly,
threshold—firstly, in order to
avoidthe
avoid theloss
lossofof the
the material;
material; and and secondly,
secondly,totoavoid
avoidthe theundesired
undesired overmelting
overmelting of the bulk
of the material
bulk [30].
material
The The applied
presence of thermal
some pores, energy
voids, was
and adjusted
unmelted to be just at in
particles thethe
melting
AM threshold—firstly,
samples, as shown ininFigure
order5bto
[30]. The presence of some pores, voids, and unmelted particles in the AM samples, as shown in
avoid(3),
zone the lossbeofexplained
the material;the andlack secondly, to avoid the undesired overmelting of the bulk material
Figure 5b can
zone (3), can be by explained of
byfusion
the lack andofthe trapping
fusion and ofthegases duringofthe
trapping re-solidification.
gases during the
[30]. The presence
A significant reduction of some pores,
in the mechanical voids, and unmelted particles in the AM samples, as shown in
re-solidification. A significant reductionproperties, due to theproperties,
in the mechanical ease of crackdue initiation,
to the can easebeof expected
crack
Figure 5b
as a resultcan zone
of these(3), can be explained
surfaceasvoids. by the lack of fusion and the trapping of gases during the
initiation, be expected a result of these surface voids.
re-solidification.
Themicrostructure A
microstructure significant reduction in theatmechanical properties, due to the ease of crack
The ofof316L
316LSST SSTisisaustenite
austenite roomtemperature.
at room temperature. ItItwas
was noted
noted thatthe
that themodified
modified
initiation,
layerininthe can be
thepolished expected
polishedsamples as a result
samplesexhibits of
exhibitsthe these
thesame surface voids.
samemicrostructure
microstructurewith withfull
fullchemical
chemicaland andmechanical
mechanical
layer
bondingThe with
microstructure
the bulk of 316L SST
material. The ispresence
austeniteofatsomeroomlamellar
temperature.
grains Itofwas
thenoted
hard that the modified
martensitic laths
bonding with the bulk material. The presence of some lamellar grains of the hard martensitic laths
layer
wasalso in the
alsonoted polished
notedon onthe samples
thesurface
surfacedue exhibits
duetotothe the
thehigh same
highcooling microstructure
coolingratesratesininthis with
thisregion, full chemical
region,enhanced
enhancedby and
bythe mechanical
theflow
flowofofthe
the
was
bonding
assist argon withgasthe
overbulkthe material.
small melt The presence
pools, as of some
shown in lamellar
Figure 15. grains
This of the
figure hardSEM
shows martensitic
micrographslaths
assist argon gas over the small melt pools, as shown in Figure 15. This figure shows SEM
wasthe
for alsotopnoted
view onof the surface
samples no.due to the high cooling rates in this region, enhanced by the flow of the
the
micrographs for the top view of5 and 8 corresponding
samples no. 5 and 8tocorresponding
Table 3 respectively.
to TableThe3 samples exhibit
respectively. The
assist
minimum argon gas
and maximumover the small
amount melt pools, as shown in Figure 15. This figure shows SEM
samples exhibit the minimum and of martensite
maximum formed
amount ofon the surface
martensite after the
formed on laser process.
the surface Image-j
after the
micrographs
software for the
(version top view
V1.52a) was of samples
used to no. 5 the
calculate andtransition
8 corresponding
rate, which to Table
was 3 respectively.
found to be 2% The
and
laser process. Image-j software (version V1.52a) was used to calculate the transition rate, which was
samples
3.1%, exhibit the minimum and maximum amount of martensite formed on the surface after the
respectively.
found to be 2% and 3.1%, respectively.
laser process. Image-j software (version V1.52a) was used to calculate the transition rate, which was
found to be 2% and 3.1%, respectively.
Materials 2019, 12, 991 12 of 15
Materials 2019, 12, x FOR PEER REVIEW 12 of 15

(a) (b)

Figure 15. SEM micrograph showing the martensitic islands on the top view of samples processed
W, beam
with 110 W, beam focused
focused on
on surface,
surface, and
and (a)
(a) 50
50rpm
rpmand
and(b)
(b)80
80rpm.
rpm. .

It
It is
is clear
clear to tosee
seethat
thatthe theamount
amountofof martensitic
martensitic islands
islands is greater
is greater in Figure
in Figure 15b 15bcompared
compared withwiththat
in Figure 15a because of the higher scanning speed and the resultant
that in Figure 15a because of the higher scanning speed and the resultant higher cooling rates. The higher cooling rates. The higher
cooling rate caused
higher cooling rate by the higher
caused by therotational speed in speed
higher rotational the case in of
thethe sample
case of the in sample
Figure 15b allows15b
in Figure for
the increase
allows for the inincrease
the formation
in the of the martensitic
formation laths in thelaths
of the martensitic modified
in thelayer.
modified layer.
In laser polishing, the challenge is the accurate
In laser polishing, the challenge is the accurate and critical adjustment and critical adjustment of of the
the laser
laser beam
beam power,
power,
scanning
scanning speed,speed,and andthe theresidence
residence time,
time,whichwhich in turn
in turn control the applied
control the applied thermal energy.
thermal WhenWhen
energy. these
parameters are under control, one laser scan track was sufficient
these parameters are under control, one laser scan track was sufficient to give the optimum to give the optimum reduction in
roughness. For repeated passes, a re-melting of the processed
reduction in roughness. For repeated passes, a re-melting of the processed surface and increased surface and increased roughening
occurred—see
roughening occurred—seeFigure 11a. Figure 11a.
The
The Gaussianenergy
Gaussian energydistribution
distribution of of
thethe
COCO 2 laser beam
2 laser leaves
beam un-melted
leaves un-melted gaps at the low
gaps at theenergy
low
density zones on both sides of the laser beam track. In order
energy density zones on both sides of the laser beam track. In order to overcome this problem, a to overcome this problem, a positive
overlapping
positive overlappingscanning scanning
of the consecutive scans and/or
of the consecutive scansan and/or
increasean in increase
process powerin processcan resolve
power this. can
An increased overlap, however, results in an increase in the overall
resolve this. An increased overlap, however, results in an increase in the overall processing time. The processing time. The percentage
overlap
percentage control was control
overlap found towas provide
foundmore control more
to provide over the roughness
control over the than control ofthan
roughness lasercontrol
power. of
Figure
laser power. 11a indicates that the laser power range used was well estimated and that the central level
was sufficient to create the required amount of melting. The number
Figure 11a indicates that the laser power range used was well estimated and that the central of process passes should be kept
at one.was
level DoE-2 indicates
sufficient that the
to create thebeam focal amount
required positionof that produces
melting. Thethe minimum
number Ra occured
of process passes when
shouldthe
focus was on the sample surface, in which the highest thermal
be kept at one. DoE-2 indicates that the beam focal position that produces the minimum Ra occured energy was focused on the elevated
peaks
when the andfocusthe adhered
was on particles,
the sample which
surface,leads to the melting
in which the highest of the majority
thermal energyof these peaks, thereby
was focused on the
enabled a more effective surface smoothening. Figure 11b indicates
elevated peaks and the adhered particles, which leads to the melting of the majority of these peaks, that the lower roughness was
reached when 20% overlap was applied with the minimum rotational
thereby enabled a more effective surface smoothening. Figure 11b indicates that the lower roughness speed of 20 rpm. The lower
speed allows for
was reached when the20%
melting of thewas
overlap high metal peaks
applied with the andminimum
the relocating of this speed
rotational moltenofmaterial
20 rpm.toThe the
depressed areas—see Figure 10b.
lower speed allows for the melting of the high metal peaks and the relocating of this molten material
to theThe results listed
depressed areas—seein Table 3 show
Figure 10b.that although the reduction observed on Ra was significant,
thereThewasresults
no improvement
listed in Table in R , Rz , and
3 tshow that Salthough
a . This can the be explained
reduction by twoon
observed main
Ra waseffects. Firstly,
significant,
the waving presented on the re-solidified layer due to the high
there was no improvement in Rt, Rz, and Sa. This can be explained by two main effects. Firstly, convection caused by the argon assist
the
gas
waving presented on the re-solidified layer due to the high convection caused by the argon assistlaser
flow, as well as the high temperature difference between the molten temperature at the gas
incident
flow, as pointwell and the high
as the solidification
temperature front. difference
Secondly, the surfacethe
between topography of the as-built
molten temperature atsamples
the laser is
mainly constructed by isolated peaks—see Figure 13a. The major
incident point and the solidification front. Secondly, the surface topography of the as-built samples part of these peaks can be melted
and relocated
is mainly in the low
constructed by valleys,
isolatedbut this mayFigure
peaks—see result in 13a.high
Theamplitude
major parttopography.
of these peaks can be melted
The initial powder particle size distribution was
and relocated in the low valleys, but this may result in high amplitude topography. between 15 and 45 µm, and the measured
reduction
The initial powder particle size distribution was between 15 and 459.97
in the sample diameter was found to be from 10.0 ± 0.015 mm to µm, ± and
0.006the mmmeasured
(~15 µm
in the radial
reduction in direction),
the samplewhich diameter is less
wasthanfound thetosmaller
be from particle
10.0 ± size.
0.015Thismm indicates
to 9.97 ± 0.006that the mmpolishing
(~15 µm
process was carried out by melting the adhered powder particles on
in the radial direction), which is less than the smaller particle size. This indicates that the polishingthe outer surface of the AM parts,
and that no material loss occurred. No excessive melting occurred
process was carried out by melting the adhered powder particles on the outer surface of the AM in the bulk material. The resulting
roughness
parts, and variation was alsoloss
that no material significantly
occurred. reduced
No excessive from melting
15 µm tooccurred
6 µm. This in is
theanbulk
important
material. benefit
The
resulting roughness variation was also significantly reduced from 15 µm to 6 µm. This is an
important benefit in the control and the adjustment of the dimensional accuracy of the mechanical
Materials 2019, 12, 991 13 of 15

in the control and the adjustment of the dimensional accuracy of the mechanical parts manufactured
using the AM technology, especially when these parts assembled with other parts where the interfacing
area is exposed to interference fit or requires a level of control over fluid containment.

5. Conclusions
In this study, the laser polishing of AM stainless steel cylindrical samples was investigated.
The optical results show that the polishing strategy used was effective, and achieved a reduction of
the average roughness Ra from 10.4 to 2.7 µm. This achievement offers great benefits compared with
conventional chemical, mechanical, and electro-polishing, which involve tool wear, abrasive debris,
and environmentally damaging solvents. Laser polishing is still under investigation and development,
and there is a lack of sufficient data for the optimization of this technique for a broad range of cases.
One of the current difficulties remains in defining the exact level of laser beam fluence required, which
might be locally different because of surface inconsistencies, as shown in Figure 5a zone (2), where a
larger mass would require a higher power density than the surrounding smaller particles or a longer
residence time. This problem can result in leaving residual unmelted particles, as shown in Figure 9.
As a possible solution to overcome this problem, real-time closed loop optical with pyrometer or
image-based monitoring could be used to adjust the fluence during processing.

Author Contributions: M.A.O. conducted the experiment, experimental set-up and work, and the design of the
experiment; M.A.O. and E.M. collaborated on the materials supply, measurement and results, data collection, and
analysis and writing the article; M.A.O and B.O. the surface topography investigations and measurements and
analysis; I.U.A. and D.B. had the supervision role, scientific suggestions, administration, and funding affairs.
Funding: This research is supported in part by a research grant from Science Foundation Ireland (SFI) under Grant
Number 16/RC/3872 and is co-funded under the European Regional Development Fund and by I-Form industry
partners. The roughness measurement and optical imaging were carried out at the Nano Research Facility in
Dublin City University which was funded under the Programme for Research in Third Level Institutions (PRTLI)
Cycle 5. The PRTLI is co-funded through the European Regional Development Fund (ERDF), part of the European
Union Structural Funds Programme 2011–2015.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Bremen, S.; Meiners, W.; Diatlov, A. Selective laser melting. Laser Tech. J. 2012, 9, 33–38. [CrossRef]
2. Toyserkani, E.; Khajepour, A.; Corbin, S. Laser Cladding; CRC Press: Boca Raton, FL, USA, 2005.
3. Frazier, W.E. Metal Additive Manufacturing: A Review. ASM Int. JMEPEG 2014, 23, 1917–1928. [CrossRef]
4. Bourell, D.L.; Leu, M.C.; Rosen, D.W. Roadmap for Additive Manufacturing; University of Texas at Austin:
Austin, TX, USA, 2009.
5. Kempen, K.; Yasa, E.; Thjis, L.; Kruth, J.P.; Van Humbeeck, J. Microstructure and mechanical properties of
Selective Laser Melted, 18Ni-300 steel. Phys. Procedia 2011, 12, 255–263. [CrossRef]
6. Lee, H.; Lim, C.H.J.; Low, M.J.; Tham, N.; Murukeshan, V.; Kim, Y. Lasers in Additive Manufacturing:
A Review. Int. J. Precision Eng. Manuf. Green Technol. 2017, 4, 307–322. [CrossRef]
7. Schmidt, M.; Merklein, M.; Bourell, D.; Dimitrov, D. Laser based additive manufacturing in industry and
academia. Manuf. Technol. 2017, 66, 561–583. [CrossRef]
8. Salman, O.; Gammer, C.; Chaubey, A.K.; Echert, J.; Scudino, S. Effect of heat treatment on microstructure
and mechanical properties of 316L steel synthesized by selective laser melting. Mater. Sci. Eng. A 2019, 748,
205–212. [CrossRef]
9. Marimuthu, S.; Traintaphyllou, A.; Antar, M.; Wimpenny, D.; Morton, H.; Beard, M. Laser polishing of
selective laser melted components. Int. J. Mach. Tools Manuf. 2015, 95, 97–104. [CrossRef]
10. Sims, C.T.; Stoloff, N.S.; Hagel, W.C. Superalloys II, 2nd ed.; Wiley-Blackwell: New York, NY, USA, 1987.
11. Yung, K.C.; Xiao, T.Y.; Choy, H.S.; Wang, W.J.; Cai, Z.S. Laser polishing of additive manufactured CoCr alloy
components with complex surface geometry. J. Mater. Process. Technol. 2018, 262, 53–64. [CrossRef]
12. Strano, G.; Hao, L.; Everson, R.M.; Evans, K.E. Surface roughness analysis, modelling and prediction in
selective laser melting. J. Mater. Process. Technol. 2013, 213, 589–597. [CrossRef]
Materials 2019, 12, 991 14 of 15

13. Gora, W.S.; Tian, Y.; Cabo, A.P.; Ardron, M.; Maier, R.R.; Prangnell, P.; Weston, N.J.; Hand, D.P. Enhancing
surface finish of additively manufactured titanium and cobalt chrome elements using laser based finishing.
Phys. Procedia 2016, 83, 258–263. [CrossRef]
14. Yasa, E.; Kruth, J.-P. Microstructural investigation of Selective Laser Melting, 316L stainless steel parts
exposed to laser re-melting. Procedia Eng. 2011, 19, 389–395. [CrossRef]
15. Li, Y.; Wu, Y.; Zhou, L.; Fujimoto, M. Vibration-assisted dry polishing of fused silica using a fixed-abrasive
polisher. Int. J. Mach. Tools Manuf. 2014, 77, 93–102. [CrossRef]
16. Tsai, M.-Y.; Yang, W.-Z. Combined ultrasonic vibration and chemical mechanical polishing of copper
substrates. Int. J. Mach. Tools Manuf. 2012, 53, 69–76. [CrossRef]
17. Habibzadeh, S.; Li, L.; Shum-Tim, D.; Davis, E.C.; Omanovic, S. Electrochemical polishing as a 316L stainless
steel surface treatment method: Towards the improvement of biocompatibility. Corros. Sci. 2014, 87, 89–100.
[CrossRef]
18. Tong, X.; Wu, X.; Zhang, F.; Ma, G.; Zhang, Y.; Wen, B.; Tian, Y. Mechanism and Parameter Optimization in
Grinding and Polishing of M300 Steel by an Elastic Abrasive. Materials 2019, 12, 340. [CrossRef] [PubMed]
19. Wu, R.; Wang, M.; Xu, J.; Qi, J.; Chu, W.; Fang, Z.; Zhang, J.; Zhou, J.; Qiao, L.; Chai, Z.; et al. Long
Low-Loss-Litium Niobate on Insulator Waveguides with Sub-Nanometer Surface Roughness. Nanomaterials
2018, 8, 910. [CrossRef]
20. Alsaeedi, R. Evaluation of Chemical Mechanical Polishing-Based Surface Modification on 3D Dental Implants
Compared to Alternative Methods. Metals 2018, 11, 2286. [CrossRef] [PubMed]
21. Rinaldi, D.; Montalto, L.; Lebeau, M.; Mengucci, P. Influence of a Surface Finishing Method on Light
Collection Behaviour of PWO Scintillator Crystals. Photonics 2018, 5, 47. [CrossRef]
22. AlMangour, B.; Yang, J. Improving the surface quality and mechanical properties by shot-peening of 17-4
stainless steel fabricated by additive manufacturing. Mater. Des. 2016, 110, 914–924. [CrossRef]
23. Kumstel, J.; Kirsch, B. Polishing titanium-and nickel-based alloys using cw-laser radiation. Phys. Procedia
2013, 41, 355–364. [CrossRef]
24. Lamikiz, A.; Sanchez, J.; de Lacalle, L.L.; Arana, J. Laser polishing of parts built up by selective laser sintering.
Int. J. Mach. Tools Manuf. 2007, 47, 2040–2050. [CrossRef]
25. Dadbakhsh, S.; Hao, L.; Kong, C.Y. Surface finish improvement of LMD samples using laser polishing.
Virtual Phys. Prototype 2010, 5, 215–221. [CrossRef]
26. Obeidi, M.A.; McCarthy, E.; Kailas, L.; Brabazon, D. Laser surface texturing of stainless steel 316L cylindrical
pins for interference fit applications. J. Mater. Process. Technol. 2018, 252, 58–68. [CrossRef]
27. Chikarakara, E.; Naher, S.; Brabazon, D. Spinodal decomposition in AISI 316L stainless steel via high speed
laser remelting. Appl. Surf. Sci. 2014, 302, 318–321. [CrossRef]
28. Chikarakara, E.; Naher, S.; Brabazon, D. Process mapping of laser surface modification of AISI 316L Stainless
Steel for Biomedical Applications. Appl. Phys. A 2010, 101, 367–371. [CrossRef]
29. Ahad, I.U.; Obeidi, M.A.; Budner, B.; Bartnik, A.; Fiedorowicz, H.; Brabazon, D. Surface roughness control
by extreme ultraviolet (EUV) radiation. AIP Conf. Proc. 2017, 1896, 200008.
30. Ukar, E.; Lamikiz, A.; Lopez, L.; Pozo, D.; Arana, J.L. Laser polishing of tool steel with CO2 laser and
high-power diode laser. Int. J. Mach. Tools Manuf. 2010, 50, 115–125. [CrossRef]
31. Willenborg, E. Polieren von Werkzeugstählen mit Laserstrahlung; Shaker: Aachen, Germany, 2005.
32. Reinhart, P. Laser Application Technology; Springer: Berlin/Heidelberg, Germany, 2011.
33. Li, Y.; Wang, B.; Ma, C.; Fang, Z.; Chen, L.; Guan, Y.; Yang, S. Material Characterization, Thermal Analysis,
and Mechanical Performance of a Laser-Polished Ti Alloy Prepared by Selective Laser Melting. Metals 2019,
9, 112. [CrossRef]
34. Kang, J.; Wang, M.; Yue, W.; Fu, Z.; Zhu, L.; She, D.; Wang, C. Tribological Behavior of Titanium Alloy Treated
by Nitriding and Surface Texturing Composite Technology. Materials 2019, 12, 301. [CrossRef] [PubMed]
35. Chun, E.; Sim, A.; Kim, M.; Kang, N. Microstructural Characterization of Surface Softening Behavior for
Cu-Bearing Martensitic Steels after Laser Surface Heat Treatment. Metals 2018, 8, 470. [CrossRef]
36. Obeidi, M.A.; McCarthy, E.; Brabazon, D. Methodology of laser processing for precise control of surface
micro-topology. Surf. Coat. Technol. 2016, 307, 702–712. [CrossRef]
Materials 2019, 12, 991 15 of 15

37. Obeidi, M.A. Laser Processing of Metallic Surfaces for Controlled Micro-Texturing and Metallic Bonding.
Ph.D. Thesis, DCU, Dublin, Ireland, 2018.
38. Kato, H.; Takahashi, M.; Ikeuchi, K. Nanoindentation hardness test for estimation of Vickers hardness. Trans.
JWRI 2006, 35, 57–62.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).