The International Journal of Advanced Manufacturing Technology (2020) 111:2361–2376

https://doi.org/10.1007/s00170-020-06261-2

ORIGINAL ARTICLE

Influence of energy density on selective laser sintering of carbon

fiber-reinforced PA12
Tiago Czelusniak 1 & Fred L. Amorim 1

Received: 8 July 2020 / Accepted: 15 October 2020 / Published online: 27 October 2020
# Springer-Verlag London Ltd., part of Springer Nature 2020

Abstract
This work provides an in-depth investigation of the influence of laser energy density on the mechanical, surface, and dimensional
properties of carbon fiber-reinforced PA12 parts manufactured by selective laser sintering (SLS). A space-filling design of
experiments (DOE) was used to conduct the experimental trials to cover a wide range of laser sintering parameters. The
consolidation mechanism was evaluated by microstructural and crystallization evolution of the samples produced at different
energy densities, supported by mechanical testing, scanning electron microscopy (SEM), X-ray diffraction (XRD), and infrared
spectroscopy (FTIR). Surface morphology was evaluated by profile measurements and SEM images. Laser sintering parameters
had a significant influence on physical and mechanical properties, exhibiting complex and nonlinear behavior. Low energy
density resulted in better dimensional accuracy, whereas intermediate laser energy resulted in the best mechanical properties. A
trade-off could be seen when mechanical or dimensional properties were desired, and optimum energy values were strongly
dependent on the criteria desired. All samples exhibited a brittle fracture behavior, with little plastic deformation present. Fracture
mechanism occurred in the interlayer region or interface between carbon fiber and PA12, depending on the energy density
applied. XRD analysis revealed a decrease in crystal fraction with increasing energy density. FTIR measurement suggested that
polymer degradation at high energy densities could be present by both polymer chain scission and oxygen functional group
decomposition and gas release upon laser heating of carbon fiber, resulting in lower mechanical properties.

Keywords Powder bed fusion . Selective laser sintering . Laser energy density . Polymer composites . Carbon fiber-reinforced
PA12 composite

1 Introduction information for the current market. Much research is being

Selective laser sintering (SLS) is a powder bed fusion additive
manufacturing process that creates parts layer-by-layer by se-
lectively fusing or melting loose powder using the thermal
energy of one or more lasers [1, 2]. SLS can rapidly generate
complex parts without the need for a support structure. The
complex laser-material interaction and consolidation mecha-
nism occurring in SLS limit the range of materials available to
date [3, 4]. By far, polyamide 12 (PA12) is the most commer-
cially used and scientifically researched material due to its
easier processability by SLS [5], accounting for the majority
of the SLS market and providing the most relevant
* Fred L. Amorim
fred.amorim@pucpr.br
1 Previous research has been focusing on improving the me-
Mechanical Engineering Graduate Program, Pontifical Catholic
University of Parana - PUCPR, Av. Imaculada Conceição, 1155
Prado Velho, Curitiba 80215-901, Brazil
conducted in improving polyamide 12 properties to meet the
requirements of more demanding applications. SLS of poly-
mer matrix composite materials is based on the rationale of
combining two or more materials to improve properties
unachievable by the single materials [6]. The most common
method to produce composites by SLS is the combination of
matrix and reinforcement powders. As continuous fibers raise
problems in SLS manufacturing due to the formation of a non-
smooth powder bed, reinforcement powders are generally
used in particulate form or the form of small-sized fibers.
The combination of powders can be made by simply mixing
the matrix and reinforcing powder, using a single composite
powder, or coating reinforcement material with the polymer
matrix, which helps overcome problems associated with non-
uniform mixing of powders, yielding a uniform spread of
composite components.
chanical properties of polymer laser-sintered parts by reinforc-
ing them with micron-sized inorganic fillers [4, 5, 7]; most of
2362
these works focus on reinforcing PA12 or PA11. Aluminum
was added to PA12 through direct mixing to produce compos-
ite parts resulted in accurate parts and with increased mechan-
ical strength compared to pure PA12 [8]. Silicon carbide was
mixed with PA12 to produced composite parts by SLS, and an
investigation on the effect of sintering parameters was per-
formed [9]. A strong relation exists between sintering param-
eters and tensile strength and also a strong sensitivity of the
system to the PA12 content. PA12 was also mixed with lime-
stone for laser sintering processing [10]. The powders were
mechanically mixed at different ratios of filler/matrix, and
studies on the effect of sintering parameters such as pre-
heating and laser power were conducted. Limestone was uni-
formly dispersed in the matrix, exhibiting superior mechanical
properties compared to pure PA12. PA12 reinforced with
glass beads through direct mixing were studied [11], showing
limited mechanical properties. Nanomaterials have also been
combined with PA12 for SLS processing. The literature re-
ports successful combinations of PA12 with nanoclay [12],
nanosized carbon black [13], carbon nanofibers [14], carbon
nanotubes [15, 16], and potassium titanium whiskers [17].
The SLS of carbon fiber-reinforced PA12 composites have
recently received attention from the research community. Yan
et al. [18] prepared carbon fibers by surface modification to
improve interface adhesion with PA12 and further mixing CF
with PA12 by a dissolution-precipitation method. Mechanical
properties were greatly enhanced by the preparation method
when compared to pure PA12. More recently, carbon fiber sur-
face modification via HNO3 treatment was performed to im-
prove adhesion with PA12 via direct mixing [19]. Improved
mechanical properties were obtained after surface modification
of CFs but only when treated in a nitrogen atmosphere. The
porosity of mechanically mixed PA12-CF samples
manufactured in different building directions was studied
employing computed tomography [20]. Highly porous struc-
tures were observed, concentrated between the layers
manufactured, leading to highly anisotropic mechanical proper-
ties. The fracture mechanism of PA12-CF material processed by
SLS was recently studied [21]. Crack growth was found to
initiate at the interface between carbon fiber and PA12.
Most of the research carried out so far in laser sintering of
PA12-CF materials was focused on material aspects, without
taking into consideration the influence of laser sintering param-
eters on quality properties. To the authors’ knowledge, no re-
search has been conducted to investigate the effect of laser
sintering parameters, which influence the energy density deliv-
ered to the powder bed. Therefore, the goal of this research study
is to investigate the influence of energy density on critical quality
criteria such as surface properties, dimensional accuracy, density,
and mechanical properties. To support the study, a space-filling
DOE is applied to conduct the experimental trials, and the laser-
material interaction is evaluated through microstructural and
crystallization analysis of the material processed at different
Int J Adv Manuf Technol (2020) 111:2361–2376
energy density levels, supported by SEM and XRD analysis.
FTIR analysis is performed to investigate polymer degradation
of PA12-CF material processed at different energy densities.
2 Materials and methods
2.1 Materials
The material studied was carbon fiber-reinforced PA12
(PA12-CF), which is a commercially available material (com-
mercially known as CarbonMide, from EOS manufacturer)
for laser sintering applications [22] composed by PA12 and
anthracite black carbon fiber PA12-CF powder. SEM micro-
graphs of the described powder at × 200 and × 1000 magni-
fication are depicted in Fig. 1. As observed from the micro-
graphs, carbon fibers are uniformly dispersed along with the
“potato-shaped” PA12 polymer particles.
2.2 SLS Processing
An EOS P396 laser sintering machine was used to perform the
SLS experiments, equipped with a 70-W continuous-wave
Gaussian CO2 laser (wavelength 10.6 μm) and a laser spot di-
ameter of 0.5 mm. Standard dumbbell-shaped type I (ASTM
Standard D638-02a, 2002 [23]) and rectangular (ASTM
Standard D790-02, 2002 [24]) samples were manufactured. All
samples were positioned in the building platform X-direction (the
same direction as the recoater blade moving direction) and using
an alternate x-y scanning pattern. The parts were placed as close
as possible to the center of the building platform to minimize the
effects of non-uniform powder temperatures on the building plat-
form. To evaluate the influence of laser energy density, the main
influencing laser sintering parameters (laser power, laser scan
speed, scan line spacing) were varied using a Sobol space-
filling DOE methodology, which has the advantage of best de-
sign space coverage and uniformity [25]. 24 laser parameter sets
were defined, with laser power varying from 30 to 45 W, laser
scan speed from 2000 to 5000 mm/s, and scan line spacing from
0.2 to 0.6 mm. The layer thickness was fixed at 150 μm. Figure 2
shows the design space covered by the Sobol space DOE. The
resulting energy density range evaluated is from 0.096 to 0.493
J/mm3, with energy density calculated using the equation:
P
ψ¼ ð1Þ
vhw
where ψ is the energy density in Joules per cubic millimeter, P
is the laser power, v is the laser scan speed, h is the scan line
spacing, and w is the layer thickness. The pre-heating tempera-
ture was kept at 178 °C during all laser sintering process.
Int J Adv Manuf Technol (2020) 111:2361–2376 2363

Fig. 1 Powder images from

scanning electron microscope
(SEM) depicting PA12-CF: a ×
200 and b × 1000

(a) (b)

For each parameter combination, 5 dumbbells and 5 rect- & Tensile strength and nominal strain at Break
angular samples were produced, giving a total of 240 samples & Tensile strength at proportional limit
evaluated. The influence of energy density was evaluated for & Tensile strength at 0.2% offset yield strength
the following output variables: & Elastic modulus

& X-direction dimensional accuracy

& Y-direction dimensional accuracy 2.3 Mechanical analysis
& Z-direction dimensional accuracy
& Apparent density Force displacement measurements were performed in an
& Average surface roughness (Ra) EMIC DL500 mechanical testing machine. Tensile tests were
& Tensile strength and nominal strain at Yield conducted according to ASTM Standard D638-02a [23].

Fig. 2 Laser sintering parameters design space covered by Sobol space-filling DOE
2364
Dumbbell-shaped tensile specimens type I having dimensions
of 165 × 19 × 3.2 mm were tested under a displacement rate of
5 mm/min. For statistical relevance, 5 samples were
manufactured for each parameter set.
2.4 Dimensional, density, and surface roughness
measurements
Dimensional accuracy measurements were conducted accord-
ing to ASTM Standard D5947-11 [26]. Width and thickness
measurements were performed using a micrometer while a
caliper was used to measure the specimen’s length. The
resulting data was used to calculate the volume of the rectan-
gular samples and to further estimate the apparent density of
the samples by measuring their mass with a precision balance.
Surface roughness was measured with a Mahr Perthen
Perthometer S8P with a cut-off length of 8 mm.
2.5 Scanning electron microscopy, X-ray diffractome-
ter, and infrared spectroscopy
A Tescan Vega 3 scanning electron microscope was used to
observe the fractured area of tensile samples and the upper
surface morphology of the samples. The specimens were
gold-sputtered in a Quorum Q150R ES machine. X-ray dif-
fractometer (XRD) analysis was performed using a Shimadzu
XRD000 to identify the phases of PA12-CF processed at dif-
ferent energy densities. The radiation intensity was detected in
the 2θ range of 2–30° with a scan speed of 1°/min. Crystal
fraction in the target area was calculated based on the ratio
between the diffraction peaks area (2θ range of 15–30°) and
the total area of the diffraction spectra. The crystal fraction
calculation was used to perform a comparison of the crystal
fraction for each energy density evaluated. Infrared spectra
(FTIR) of PA12 samples were obtained using a Perkin
Elmer LR64912C spectrophotometer in the range of 500–
4000 cm−1.
3 Results and discussion
3.1 Influence of energy density on dimensional
accuracy
Box whisker plots of dimensional accuracy in X-, Y-, and Z-
directions over energy density are depicted in Fig. 3a–c. High
X-direction accuracy values are observed in all energy density
areas covered, remaining above 99.6% up to 0.381 J/mm3. At
higher energy levels, a higher dispersion can be observed as
well as a slight decrease in the accuracy.
Y-direction accuracy remains above 98% and stable up to
0.212 J/mm3 energy density. Between 0.219 and 0.281
J/mm3, higher dispersion in measured values is observed,
Int J Adv Manuf Technol (2020) 111:2361–2376
although mean values remain at adequate levels (above 98%
on average). An abrupt change occurs at energy densities
equal to or above 0.381 J/mm3, showing much higher disper-
sion of the measured values and significant reduction in aver-
age accuracy to levels below 96%. The overall lower dimen-
sional accuracy in Y-direction compared to X-direction can be
attributed to the better control of the galvanometer mirrors in
the X-direction, which results in more stable laser scanning in
this direction.
Z-direction accuracy measurements are depicted in Fig. 3c.
A clear behavior can be observed from the graphic: between
0.096 and 0.123 J/mm3, accuracy values are above 95% on
average. Further increasing the energy density leads to a step-
wise decrease in Z-direction accuracy from 94% at 0.142
J/mm3 to 88.25% at 0.178 J/mm3. Between 0.178 and 0.281
J/mm3 measurement values remain more or less stable, with
upper and lower hinge oscillating between 85 and 90%. At
energy densities of 0.381J/mm3 and above, Z-direction accu-
racy sharply decreases to average values between 75–80%,
and a higher dispersion between measurements can be
observed.
The laser energy delivered to the powder bed undergoes
radiation absorption by the PA12-CF particles and multiple
transmission to neighboring particles. Both PA12 and CF in
the form of anthracite have good absorption at the CO2 laser
wavelength of 10.6 μm [27, 28]. The addition of carbon fiber
in anthracite form to PA12 improves the thermal conductivity
of the powder bed [29], leading to more heat transferred to the
surrounding and underlying unsintered powder, worsening
dimensional accuracy of the parts. It has been observed by
simulation modeling with PA12 and carbon nanotubes
(CNTs) that the laser heat-affected area is wider and deeper
when laser sintering PA12-CNT compared to PA12 [30]. The
combination of improved heat transfer and higher energy den-
sity leads to a higher heat-affected area, resulting in a larger
volume of consolidated material and dimensional inaccuracy.
3.2 Influence of energy density on surface roughness
Box whisker plots of average surface roughness Ra over ener-
gy density are shown in Fig. 4. No clear tendency can be
observed on surface roughness and energy density, with aver-
age values oscillating between 13 and 17 μm regardless of the
energy density applied. A high variation between measure-
ments at the same energy density can be observed. The high
variation is typical from laser-sintered surfaces, which pro-
duced highly irregular and complex surface topographies.
The presence of short carbon fibers with a high aspect ratio
adds another factor that contributes to the high variation
observed.
Ra values are in good agreement with recently measured
values reported in the literature for laser-sintered surfaces,
Int J Adv Manuf Technol (2020) 111:2361–2376 2365

(a)

(b)

(c)
Fig. 3 Influence of energy density on dimensional accuracy of PA12-CF: a X-direction, b Y-direction, and c Z-direction
2366
Fig. 4 Influence of energy
density on surface roughness
where average Ra values of between 13 and 17 μm were
observed [31, 32].
Laser exposure parameters can influence surface topogra-
phy. In particular, laser scan spacing is an important parameter
that can influence surface roughness. At low scan line spacing
values, the laser scans parallel lines with some degree of over-
lap between the lines. The degree of overlap is a measure
based on the scan line spacing and the laser beam diameter,
which shows how strong is the overlapping between two par-
allel lines. The lower the degree of overlap, the less the inter-
action between laser lines. To illustrate this effect, Fig. 5
shows the effect of overlap degree on average surface rough-
ness Ra as well as the intensity of the laser speed applied
highlighted as a size dimension on the graphic. The overlap
Fig. 5 Influence of degree of
overlap on average surface
roughness
Int J Adv Manuf Technol (2020) 111:2361–2376
degree has no influence on the average values of Ra for SLS
surfaces of PA12-CF. Average surface roughness values os-
cillate between 12 and 16 μm, independently on the scan line
spacing and laser speed applied.
To further evaluate if changes on surface topography could
be observed, Fig. 6a–d shows SLS surface morphology of
PA12-CF samples captured with SEM at different energy den-
sities and overlap degrees. From the micrographs, no differ-
ences can be observed on the surface morphology of samples
processed at different overlap degrees, even for negative over-
lap degrees from Fig. 6c, d (no overlap present), the surface
morphologies are very similar. Scan line spacing plays a small
or even no significant role in surface morphology. Also, the
combination of negative overlap degree and high laser speed
Int J Adv Manuf Technol (2020) 111:2361–2376
Fig. 6 SEM micrographs of
PA12-CF SLS surface
morphology: a × 50, b × 50, c, ×
50, and d × 50
(a) Laser speed: 3000 mm/s; Degree of
overlap: 40 %
(c) Laser speed: 2758 mm/s; Degree of
overlap: -12 %
does not contribute to modifications in the surface morpholo-
gy of PA12-CF samples.
A possible explanation for the surface results observed can
be attributed to the increase in the laser heat-affected area
caused by the presence of carbon fibers on PA12. Due to its
higher thermal conductivity, carbon fibers enhance the heat
transferred to surrounding particles during SLS, increasing the
heat-affected area of the laser. A higher heat-affected area
increases the real overlap degree, promoting a smoothening
of the LS surfaces. Experiments with higher overlap degrees
could be performed to confirm this hypothesis.
3.3 Influence of energy density on PA12 density and
mechanical properties
Box whisker plot of the measured parts' density in function of
energy density is depicted in Fig. 7. Low-density values
around 0.84 g/cm3 are found at an energy density of 0.096
J/mm3. A slight increase in energy density to 0.112 J/mm3
2367
(b) Laser speed: 4206 mm/s; Degree of
overlap: 22 %
(d) Laser speed: 4862 mm/s; Degree of
overlap: -14 %
improves parts density to 0.92 g/cm3. From 0.114 to 0.123
J/mm3, there is a plateau between density values between 0.95
and 0.97 g/cm3.
Further increasing the energy density improves density in a
step-like pattern from 0.98 to 1.03 g/cm3, after which another
plateau is reached until 0.181 J/mm3. The maximum average
density is found at this plateau with an average of 1.03 g/cm3.
From 0.189 to 0.26 J/mm3, density values oscillate between
1.00 and 1.02 g/cm3 on average. Above 0.26 J/mm3, density
starts to decrease to values below 1.00 g/cm3 on average. The
elastic behavior of PA12-CF was evaluated based on force-
displacement measurements with the following properties ob-
tained: proportional limit (MPa), 0.2% offset yield strength
(MPa), and elastic modulus (MPa). Figure 8a shows the pro-
portional limit and 0.2% offset yield strength. Lowest values
of proportional limit (16 MPa) and 0.2% offset yield strength
(22 MPa) are found at the lowest energy density level.
Mechanical properties linearly increase with higher energy
densities up to 0.145 J/mm3, reaching 31 MPa for proportional
2368
Fig. 7 Influence of energy
density on PA12-CF parts’
density
limit and 40 MPa for 0.2% offset yield strength. Maximum
values at 0.19 J/mm3. Further increasing the energy density
reduces the elastic tensile properties of the composite material.
The elastic modulus calculated values in function of energy
density are shown in Fig. 8b. At the lowest energy density of
0.096 J/mm3, the elastic modulus was 2400 MPa, increasing
step-wisely at higher energy values, reaching 3566 MPa at
0.145 J/mm3. The highest elastic modulus of 3669 MPa is
reached, significantly higher than elastic modulus at a slightly
lower energy level, demonstrating the importance of individ-
ual laser exposure parameters on mechanical properties.
Further increasing the energy density above 0.281 J/mm3
leads to an overall elastic modulus reduction, except for the
measuring point at 0.398 J/mm3, which exhibited an elastic
modulus of 3445 MPa.
Figure 9a shows tensile strength at yield and break in func-
tion of energy density. At the lowest energy densities of 0.096
J/mm3, both measured tensile strengths present low values of
31.87 MPa on average. The similar values of yield and break
strength suggest little or no plastic deformation at all, resulting
in brittle fracture. Strength values increase step-wisely with
increasing energy density reaching 65 MPa on average at
0.168 J/mm3, reaching a maximum at 0.2 J/mm3. Further in-
creasing the energy density leads to a reduction in both
strengths at yield and break to values below 63 MPa.
Nominal strains at yield and break in function of energy
density are depicted in Fig. 9b. Both yield and break strain
start with very similar levels of 2.5% at the lowest energy
density level, indicating the brittle fracture behavior in this
area. At higher energy densities, strain values oscillate with
a clear increase gradient. Additionally, yield and break strain
start to achieve different values at 0.142 J/mm3, suggesting
that more plastic deformation is present. The highest strain
values are found at 0.168 J/mm3. At higher energy densities,
both strains at yield and break oscillate between 4.7 and
Int J Adv Manuf Technol (2020) 111:2361–2376
5.75%, with strain at break values slightly above strain at
yield.
SEM micrographs of fractured surfaces of PA12-CF proc-
essed by SLS at different energy densities are depicted in Fig.
10. At 0.096 and 0.114 J/mm3, SEM images (Fig. 10a–d)
reveal the presence of interparticle and interlayer porosity,
with well distinguishable layers. Higher magnification images
show a high content of unmolten PA12 particles and carbon
fibers, which are not surrounded by PA12 matrix, indicating a
reduced level of consolidation.
At a higher energy level of 0.142 J/mm3, fractured surfaces
show a denser structure with PA12 matrix embedding carbon
fibers, indicating better adhesion between the two phases.
Carbon fibers are uniformly dispersed in the polymer matrix.
Porosity can be observed and also some fibers pullout from
the matrix.
Further increasing the energy density to 0.168 J/mm3 (strain
at break was the highest at this energy level) leads to higher
plastic deformation, as observed in Fig. 10e, with elongated
fibrils visible in the direction of the stress applied. The plastic
deformation occurs in the PA12 matrix, and the cracks originate
at the interface between PA12 and carbon fibers, as can be
observed in more detail in Fig. 10h. This fracture mechanism
is well known to composite materials as the interface between
adjoining and different materials result in stress concentration.
Additionally, carbon fibers seem preferentially oriented in the
direction of the tensile force applied, which is in agreement with
the results found by Jansson and Pejryd [20].
Figure 10g, h shows the fractured surfaces of SLS samples
processed with the highest energy level of 0.493 J/mm3. No
significant changes in the microstructure can be seen, with
some level of plastic deformation also occurring. In Fig.
10h, carbon fiber pullout from the PA12 matrix can be ob-
served, reinforcing that the fracture mechanism occurs at the
interface between the carbon fiber and matrix.
Int J Adv Manuf Technol (2020) 111:2361–2376
Fig. 8 Influence of energy
density on elastic behavior for
PA12-CF: a stress and b elastic
modulus
SLS is a highly complex process where a combination of
factors influences porosity formation and density. Material
characteristics such as particle shape and distribution play an
important role in the packing density of the powder bed. PA12
particles have “potato” shape morphology and an adequate
particle size distribution, which favors powder spreading and
flowability during the SLS process. Carbon fibers present in
the PA12 matrix affect the powder spreading over the plat-
form due to their high aspect ratio, causing voids in the packed
powder, leading to porosity after SLS consolidation.
From laser sintering perspective, the very short laser mate-
rial interaction during SLS leads to the primary consolidation
mechanism [20] of PA12-CF powders via particle
2369
(a)
(b)
rearrangement phase during liquid phase sintering (LPS),
where the liquid phase (PA12) flows around the solid particles
(CF) via capillary forces [33].
Laser sintering parameters can significantly affect the rear-
rangement phase during LPS and resulting density. At very
low energy levels, the powder particles do not melt effectively
and in an adequate amount so that PA12 liquid phase can flow
over the solid carbon fiber particles, leading to a higher frac-
tion of unsintered particles and also to interlayer porosity.
Additionally, the lower temperature gradient between the cen-
ter and the edge of the molten pool concentrates the heat near
the center of the laser spot, leading to a narrow molten pool.
All these aspects result in a poor rearrangement phase during
2370
Fig. 9 Influence of energy
density on plastic behavior for
PA12-CF: a stress and b nominal
strain
LPS and poor consolidation of the composite powder and
higher porosity.
Increasing energy density reduces porosity due to the
higher amount of liquid present, improving spreading ki-
netics during the liquid phase sintering rearrangement
phase. Higher temperature gradients are existing, effective-
ly transferring heat from center to the edge of the molten
pool, generating a wider molten pool, leading to improved
LPS particle rearrangement phase. But an optimum level of
energy exists, above which a negative effect on density is
observed. The combination of a higher heat-affected area
Int J Adv Manuf Technol (2020) 111:2361–2376
(a)
(b)
resulting in part volume increase as well as polymer deg-
radation are the leading causes for the density reduction.
The mechanism of polymer degradation and oxidation is
further explained in Sect. 3.4.
The elastic behavior of polymer composites depends on the
homogeneity of filler dispersion on the matrix and the adhe-
sion between filler and polymer matrix. PA12-CF composites
proved to have superior elastic properties compared to pure
PA12, even at the lowest energy density applied elastic mod-
ulus was much higher than pure PA12 (1650 MPa [22]). This
behavior is mainly due to a good dispersion and adhesion of
Int J Adv Manuf Technol (2020) 111:2361–2376
carbon fibers on the PA12 matrix [21]. For polymer compos-
ites, this comes at the cost of lower strain as the reinforcement
phase present causes local fragility to the matrix.
Additionally, the consolidation mechanism by liquid phase
sintering mechanism plays an important role in the homoge-
neity of filler dispersion and adhesion at the matrix-filler in-
terface [34]. At low energy densities, liquid phase sintering is
poor due to the low liquid content flowing around the filler
solid particles. The poor rearrangement phase during LPS
leads to weak adhesion between filler and matrix and lower
elastic properties.
At higher energy density levels, the liquid phase sintering
mechanism is enhanced by the higher liquid content, reducing
the viscosity of the molten pool and promoting an increase in
the capillary forces acting during the rearrangement phase.
(a) Energy Density: 0.096 J/mm3
(c) Energy Density: 0.114 J/mm3
Fig. 10 SEM micrographs of fractured PA12-CF SLS surfaces: a × 150, b × 800, c × 150, d × 800, e × 300, f × 1500, g × 1500, h × 5000, i × 1500, and j
× 5000
2371
Rearrangement phase is therefore improved, contributing to
a better consolidation process and better dispersion of the filler
in the polymer matrix.
Increasing energy density further decreases the elastic ten-
sile properties of PA12-CF. At higher energy densities where
adequate consolidation was achieved and porosity dimin-
ished, recrystallization behavior may play a major role.
Increased energy levels decrease the overall crystal fraction
of the PA12 samples due to the increased molten-
recrystallized phase present in the SLS sample. The reduced
crystal fraction leads to a reduction in elastic properties as a
higher crystallization degree is associated with improved elas-
tic properties [35, 36]. Polymer degradation also has a signif-
icant contribution to decreased elastic properties at very high
energy densities, as further explained in Sect. 3.4.
(b) Energy Density: 0.096 J/mm3
(d) Energy Density: 0.114 J/mm3
2372 Int J Adv Manuf Technol (2020) 111:2361–2376

(e) Energy Density: 0.142 J/mm3 (f) Energy Density: 0.142 J/mm3

(g) Energy Density: 0.168 J/mm3 (h) Energy Density: 0.168 J/mm3

(i) Energy Density: 0.493 J/mm3 (j) Energy Density: 0.493 J/mm3
Fig. 10 (continued)
Int J Adv Manuf Technol (2020) 111:2361–2376
Regarding plastic properties, the interlayer porosity and
large quantity of unmolten particles resulting from the low
energy densities applied in SLS induce crack nucleation in
both the interlayer region and the interface between carbon
fiber and PA12, leading to low yield strength values and very
low elongation at break measured.
Energy increase results in a coherent structure with good
adhesion between fiber and matrix and low porosity, as ob-
served in SEM micrographs. The resulting microstructure is
formed by molten-recrystallized PA12 phase, smaller content
of unmolten particles (particles which did not melt or with
unmolten particle cores), and carbon fibers. As the plastic
deformation occurring in the composite material results from
PA12 elongation of amorphous chains and alignment of la-
mellar crystalline phases, the lower content of unmolten par-
ticles results in lower crystallinity of PA12 (i.e., amorphous
content of PA12 is increased), improving plastic strain.
Compared to PA12, the plastic strain of PA12-CF is much
lower, but such a trend is typical for composites as the fracture
mechanism occurs at the interface between filler and matrix
material, and stress concentrations are induced [21].
3.4 Influence of energy density on crystallization
behavior and degradation
XRD patterns of PA12-CF and PA12 powder and SLS PA12-
CF samples processed at different energy densities are shown
in Fig. 11. PA12-CF powder XRD reveals the existence of
two separate peaks occurring between a 2θ of 20° and 24°.
The first peak at 2θ of 20.92° is associated with the γ crystal
Fig. 11 XRD patterns of PA12,
PA12-CF powder, and LS PA12-
CF samples processed at different
energy densities
2373
form and the second peak at 2θ of 21.9° is associated with the
α crystal form. Compared to PA12 XRD, PA12-CF shows a
broader reflection spectrum in the range between 20° and 25°,
mainly due to the amorphous structure of anthracite present in
the carbon fibers [28]. The estimated PA12-CF crystal fraction
area in the 2θ region between 15° and 30° is 59%.
Laser-sintered samples processed at 0.096 J/mm3 show a
change on the crystal structure, with γ form as main phase
present and α form still present as indicated by a higher inten-
sity at the 2θ location of α. It is known that α form transforms
to γ form when heated to melting point [37] and is associated
with a reduction of unmolten particles during SLS and im-
proved consolidation of the material during the SLS process.
A reduction in the calculated crystal fraction area (55%) is also
observed, compared to the PA12-CF powder.
Increasing the energy density to 0.168 J/mm3 does not
show any significant change in the crystalline structure
compared to 0.096 J/mm3. The estimated crystal fraction
in the targeted area is also 55%. The reason for this can be
associated with the surface position XRD was captured,
which could not detect differences between the surfaces
processed at different energy levels. Only at the highest
energy level of 0.493 J/mm3, there is a significant decrease
in the relative intensity related to the α crystal form and a
reduction in the estimated crystal fraction (52%).
The reduction in crystal fraction at increasingly LS en-
ergy densities is associated with an improvement of the
melting and consolidation of PA12 powder particles.
Improvement of melting conditions results in higher melt-
recrystallized phase fraction, which is less crystalline than
2374
the unmolten particles, decreasing the overall crystallinity
of the material. The lower crystal fraction (i.e., higher
amorphous fraction) observed for samples produced with
higher energy density leads to improved plastic strain.
As observed in SEM micrographs and from nominal
strain measurements performed, there was no improvement
in the nominal strain at break at higher energy levels, with a
trend to slightly reduce strain, which may be associated
with polymer degradation. To investigate the degradation
behavior, Fig. 12 shows the FTIR spectra of PA12-CF LS
samples processed at different energy densities. Infrared
band assignments are given in [38]. The band assignments
of carbon fiber are superposed with band assignments of
PA12, so no significant differences can be explicitly ob-
served regarding carbon fiber spectra, which can be attrib-
uted to the low amount of functional groups and low trans-
mittance of black materials. There is an overall absorption
increase at the assigned wavenumbers towards increasing
energy densities, which can be associated with higher chain
mobility caused by chain scission [39, 40], indicating that
some level of polymer degradation might occur at this laser
energy density.
Additionally, some modifications were observed for
samples produced with 0.281 and 0.493 J/mm3 in the region
of 1700–1750 cm−1, which is related to carbonyl groups
formed during polyamide oxidation [40]. Such changes in
absorption in this region suggest some level of oxidation of
the samples processed at higher energy densities, which
could lead to lower mechanical properties as observed.
Another possible explanation for the effect observed is
based on the decomposition of oxygen groups on the sur-
face of carbon fiber released in the form of CO and CO2
during laser sintering. Upon laser exposure, oxygen groups
released in the form of gas could promote the formation of
Fig. 12 FTIR spectra of SLS
PA12-CF samples processed at
different energy densities
Int J Adv Manuf Technol (2020) 111:2361–2376
pores, creating stress concentration areas in the materials,
and reducing density and mechanical properties. This be-
havior has been observed in previous studies [19], and the
present work shows that it is intensified with increasing
energy density. The porosity increase is corroborated by
density measurements performed (Fig. 7), showing a den-
sity reduction for samples processed at higher energy
levels.
4 Conclusions
This work investigated the influence of the laser energy
density on the mechanical, surface, and dimensional prop-
erties of SLS-produced PA12-CF samples. A wide param-
eter range was studied employing a space-filling DOE
method, and the consolidation mechanism at different en-
ergy levels was studied based on the microstructure and
crystallization behavior of the parts. The following conclu-
sions can be drawn:
& The highest accuracy (above 99.6%) was achieved at
X-direction, which remained stable for energy densi-
ties below 0.381 J/mm 3 . Y-direction accuracy
remained above 98% and stable up to 0.212 J/mm3,
with worsening accuracy at increasing energy densi-
ties. Z-direction accuracy was the lowest measured,
showing the highest accuracy above 95% at the lowest
energy density and decreasing step-wisely towards
higher energy densities.
& No significant differences were observed on surface
roughness depending on energy density applied. Even
at higher scan line spacing and laser speed, the surface
roughness values remained very similar. Surface
Int J Adv Manuf Technol (2020) 111:2361–2376
morphologies were very similar independently of the
overlap degree used. The higher heat affected zone
due to improved thermal conductivity of carbon fiber
present in the system contributed to the observed sur-
face morphologies.
& The highest part density was reached in a mid-range ener-
gy density plateau (between 0.145 and 0.181 J/mm3.
Higher energy densities did not result in denser parts,
which was attributed to volume increase due to increased
heat affected area during SLS and polymer degradation.
& Results show that a compromise exists between part accu-
racy and density, which strongly depends on the individ-
ual laser sintering parameters. Improved density is
achieved in mid-range energy density, whereas low ener-
gy density leads to enhanced part accuracy. Such a trend
has been observed before by Jansson and Pejryd [19].
& Elastic properties are increased with increasing energy
densities, reaching the highest values at 0.19 J/mm3.
Improved rearrangement phase during LPS at higher en-
ergy densities was credited to the higher elastic properties.
At higher energy densities, elastic properties diminished.
A reduction in crystal fraction, as well as polymer degra-
dation, contributed to the observed behavior.
& The highest yield strength was found at 0.19–0.2 J/mm3.
Plastic behavior revealed a brittle fracture of PA12-CF
samples regardless of the energy density used, with small
values of strain at break. Improved nominal strain at break
was achieved with increasing energy densities, which was
attributed to the lower crystal fraction content present with
increasing energy densities.
& SEM of the fractured surfaces showed a highly porous
structure at interlayer regions and between PA12 and CF
particles when very low energy density was applied.
Increasing energy densities promoted better consolidation
of the parts and higher plastic strain, with crack growth
initiating at the interface between PA12 and CF.
& XRD measurements showed a reduction in α phase with
increasing energy densities, which was associated with the
better consolidation of the composites powder during
SLS, reducing the overall crystal fraction of the materials.
The results confirmed the improved plastic strain correla-
tion with crystal fraction reduction assumption for PA12-
CF samples processed at higher energy densities.
& Polymer degradation through chain scission was ob-
served with FTIR measurements of samples processed
at higher energy densities. The reduction of absorption
at assignment bands suggests higher chain mobility of
PA12. Also, carbonyl group modifications were ob-
served at higher energy densities, which were associat-
ed with thermal degradation of oxygen groups present
at the CF surface, released as gas during the SLS pro-
cess, and resulted in higher porosity of the materials and
decreased mechanical properties.
2375
Acknowledgments The authors would like to thank Financiadora de
Estudos e Projetos (FINEP) for financial support.
Data availability The raw/processed data required to reproduce these
findings cannot be shared at this time as the data also forms part of an
ongoing study.
References
1. Kruth JP, Wang X, Laoui T, Froyen L (2003) Lasers and materials
in selective laser sintering. Assem Autom 23:357–371. https://doi.
org/10.1108/01445150310698652
2. ASTM Standard F2792-12a (2012) Standard terminology for addi-
tive manufacturing technologies. West Conshohocken, PA
3. Kruth J-P, Mercelis P, Vaerenbergh J van, Froyen L and Rombouts
M (2005) Binding mechanisms in selective laser sintering and se-
lective laser melting. Rapid Prototyp J 11:26–36. https://doi.org/10.
1108/13552540510573365
4. Wang X, Jiang M, Zhou Z, Gou J, Hui D (2017) 3D printing of
polymer matrix composites: a review and prospective. Compos Part
B 110:442–458. https://doi.org/10.1016/J.COMPOSITESB.2016.
11.034
5. Goodridge RD, Tuck CJ, Hague RJM (2012) Laser sintering of
polyamides and other polymers. Prog Mater Sci 57:229–267.
https://doi.org/10.1016/j.pmatsci.2011.04.001
6. Kumar S, Kruth JP (2010) Composites by rapid prototyping tech-
nology. Mater Des 31:850–856. https://doi.org/10.1016/j.matdes.
2009.07.045
7. Parandoush P, Lin D (2017) A review on additive manufacturing of
polymer-fiber composites. Compos Struct 182:36–53. https://doi.
org/10.1016/J.COMPSTRUCT.2017.08.088
8. Mazzoli A, Moriconi G, Pauri MG (2007) Characterization of an
aluminum-filled polyamide powder for applications in selective
laser sintering. Mater Des 28:993–1000. https://doi.org/10.1016/j.
matdes.2005.11.021
9. Hon KKB, Gill TJ (2003) Selective laser sintering of SiC/
polyamide composites. CIRP Ann Manuf Technol 52:173–176.
https://doi.org/10.1016/S0007-8506(07)60558-7
10. Guo Y, Jiang K, Bourell DL (2014) Preparation and laser sintering
of limestone PA 12 composite. Polym Test 37:210–215. https://doi.
org/10.1016/j.polymertesting.2014.06.002
11. Negi S, Dhiman S, Sharma RK (2015) Determining the effect of
sintering conditions on mechanical properties of laser sintered glass
filled polyamide parts using RSM. Measurement 68:205–218.
https://doi.org/10.1016/j.measurement.2015.02.057
12. Jain PK, Pandey PM, Rao PVM (2009) Selective laser sintering of
clay-reinforced polyamide. Polym Compos 31:732–743. https://
doi.org/10.1002/pc.20854
13. Athreya SR, Kalaitzidou K, Das S (2010) Processing and charac-
terization of a carbon black-filled electrically conductive Nylon-12
nanocomposite produced by selective laser sintering. Mater Sci Eng
A 527:2637–2642. https://doi.org/10.1016/j.msea.2009.12.028
14. Goodridge RD, Shofner ML, Hague RJM, McClelland M, Schlea
MR, Johnson RB, Tuck CJ (2011) Processing of a polyamide-12/
carbon nanofibre composite by laser sintering. Polym Test 30:94–
100. https://doi.org/10.1016/j.polymertesting.2010.10.011
15. Bai J, Goodridge RD, Hague RJM, Song M (2013) Improving the
mechanical properties of laser-sintered polyamide 12 through in-
corporation of carbon nanotubes. Polym Eng Sci 53:1937–1946.
https://doi.org/10.1002/pen.23459
2376
16. Salmoria GV, Paggi RA, Lago A, Beal VE (2011) Microstructural
and mechanical characterization of PA12/MWCNTs nanocompos-
ite manufactured by selective laser sintering. Polym Test 30:611–
615. https://doi.org/10.1016/j.polymertesting.2011.04.007
17. Yang J, Shi Y, Yan C (2010) Selective laser sintering of polyamide
12/potassium titanium whisker composites. J Appl Polym Sci 117:
2196–2204. https://doi.org/10.1002/app.31965
18. Yan C, Hao L, Xu L, Shi Y (2011) Preparation, characterisation and
processing of carbon fibre/polyamide-12 composites for selective
laser sintering. Compos Sci Technol 71:1834–1841. https://doi.org/
10.1016/j.compscitech.2011.08.013
19. Jing W, Hui C, Qiong W, Hongbo L, Zhanjun L (2017) Surface
modification of carbon fibers and the selective laser sintering of
modified carbon fiber/nylon 12 composite powder. Mater Des
116:253–260. https://doi.org/10.1016/j.matdes.2016.12.037
20. Jansson A, Pejryd L (2016) Characterisation of carbon fibre-
reinforced polyamide manufactured by selective laser sintering.
Addit Manuf 9:7–13. https://doi.org/10.1016/j.addma.2015.12.003
21. Liu Y, Zhu L, Zhou L, Li Y (2019) Microstructure and mechanical
properties of reinforced polyamide 12 composites prepared by laser
additive manufacturing. Rapid Prototyp J 25:1127–1134. https://
doi.org/10.1108/RPJ-08-2018-0220
22. EOS (2020) Material Data Center. http://eos.materialdatacenter.
com/eo/en. Accessed 9 Jan 2020
23. ASTM Standard D638-02a (2002) Standard test method for tensile
properties of plastics. In: West Conshohocken
24. ASTM Standard D790-02 (2002) Standard test methods for flexural
properties of unreinforced and reinforced plastics and electrical in-
sulating materials. West Conshohocken
25. Cavazzuti M (2013) Optimization methods: from theory to design.
Springer-Verlag, Berlin Heidelberg
26. ASTM Standard D5947-11 (2012) Standard test methods for phys-
ical dimensions of solid plastics specimens
27. Laumer T, Stichel T, Nagulin K, Schmidt M (2016) Optical analy-
sis of polymer powder materials for selective laser sintering. Polym
Test 56:207–213. https://doi.org/10.1016/j.polymertesting.2016.
10.010
28. Mattos NRD, Oliveira CRD, Camargo LGB, Silva RSRD, Lavall
RL (2018) Azo dye adsorption on anthracite: a view of thermody-
namics, kinetics and cosmotropic effects. Sep Purif Technol 209:
806–814. https://doi.org/10.1016/j.seppur.2018.09.027
29. Herrin JM, Deming D (1996) Thermal conductivity of U.S. coals. J
Geophys Res Solid Earth 101:25381–25386. https://doi.org/10.
1029/96jb01884
30. Bai J, Goodridge RD, Yuan S, Zhou K, Chua CK, Wei J (2015)
Thermal influence of CNT on the polyamide 12 nanocomposite for
Int J Adv Manuf Technol (2020) 111:2361–2376
selective laser sintering. Molecules 20:19041–19050. https://doi.
org/10.3390/molecules201019041
31. Guo J, Bai J, Liu K, Wei J (2018) Surface quality improvement of
selective laser sintered polyamide 12 by precision grinding and
magnetic field-assisted finishing. Mater Des 138:39–45. https://
doi.org/10.1016/j.matdes.2017.10.048
32. Launhardt M, Wörz A, Loderer A, Laumer T, Drummer D,
Hausotte T, Schmidt M (2016) Detecting surface roughness on
SLS parts with various measuring techniques. Polym Test. https://
doi.org/10.1016/j.polymertesting.2016.05.022
33. Anestiev L, Froyen L (1999) Model of the primary rearrangement
processes at liquid phase sintering and selective laser sintering due
to biparticle interactions. J Appl Phys 86:4008. https://doi.org/10.
1063/1.371321
34. Kruth J-P, Levy G, Klocke F, Childs THC (2007) Consolidation
phenomena in laser and powder-bed based layered manufacturing.
CIRP Ann Manuf Technol 56:730–759. https://doi.org/10.1016/j.
cirp.2007.10.004
35. Dupin S, Lame O, Barrès C, Charmeau J-YY (2012)
Microstructural origin of physical and mechanical properties of
polyamide 12 processed by laser sintering. Eur Polym J 48:
1611–1621. https://doi.org/10.1016/j.eurpolymj.2012.06.007
36. Jauffrès D, Lame O, Vigier G, Doré F, Douillard T (2009) Sintering
mechanisms involved in high-velocity compaction of nascent semi-
crystalline polymer powders. Acta Mater 57:2550–2559. https://
doi.org/10.1016/j.actamat.2009.02.012
37. Atkins EDT, Hill MJJ, Veluraja K (1995) Structural and morpho-
logical investigations of nylon 8 chain-folded lamellar crystals.
Polymer 36:35–42. https://doi.org/10.1016/0032-3861(95)90672-
O
38. Rhee S, White JL (2002) Crystal structure and morphology of bi-
axially oriented polyamide 12 films. J Polym Sci B Polym Phys 40:
1189–1200. https://doi.org/10.1002/polb.10181
39. Zhang J, Adams A (2016) Understanding thermal aging of non-
stabilized and stabilized polyamide 12 using 1H solid-state NMR.
Polym Degrad Stab 134:169–178. https://doi.org/10.1016/j.
polymdegradstab.2016.10.006
40. Celina M, Ottesen DK, Gillen KT, Clough RL (1997) FTIR emis-
sion spectroscopy applied to polymer degradation. Polym Degrad
Stab 58:15–31. https://doi.org/10.1016/S0141-3910(96)00218-2
Publisher’s note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.