Materials & Design 83 (2015) 768–776

Contents lists available at ScienceDirect

Materials & Design

journal homepage: www.elsevier.com/locate/matdes

Fused deposition modeling with polypropylene

O.S. Carneiro a,⇑, A.F. Silva b, R. Gomes a
a
IPC/I3N – Institute for Polymers and Composites, Polymer Engineering Department, University of Minho, Campus de Azurém, 4800-058 Guimarães, Portugal
b
CMEMS-UMinho – Center of ElectroMechanical Systems, University of Minho, Campus de Azurém, 4800-058 Guimarães, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: This paper addresses the potential of polypropylene (PP) as a candidate for fused deposition modeling
Received 20 February 2015 (FDM)-based 3D printing technique. The entire filament production chain is evaluated, starting with
Revised 25 April 2015 the PP pellets, filament production by extrusion and test samples printing. This strategy enables a true
Accepted 7 June 2015
comparison between parts printed with parts manufactured by compression molding, using the same
Available online 20 June 2015
grade of raw material. Printed samples were mechanically characterized and the influence of filament
orientation, layer thickness, infill degree and material was assessed. Regarding the latter, two grades of
Keywords:
PP were evaluated: a glass-fiber reinforced and a neat, non-reinforced, one. The results showed the poten-
Fused deposition modeling (FDM)
Polypropylene (PP)
tial of the FDM to compete with conventional techniques, especially for the production of small series of
Glass reinforced polypropylene (GRPP) parts/components; also, it was showed that this technique allows the production of parts with adequate
Filament orientation mechanical performance and, therefore, does not need to be restricted to the production of mockups and
Layer thickness prototypes.
Infill degree Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction The range of applications where this technique can be used is

extensive, ranging from medical applications [10,11] to automotive
Rapid prototyping (RP) represents a new edge on the prototyp- [12] and aeronautics [13]. Moreover, the potential to grow is still
ing process evolution. With the last advances, it is now possible to on its ascending stage and new materials (metallic- [14] or
build physical models quicker and with more complex geometries, wood-based [15], for example) are being constantly delivered to
pushing this type of techniques from printing mockups and the market while the printers are getting less expensive every day.
prototypes models towards printing final products in limited series However, the published studies up to now focused little on the
[1–4]. extrusion processing window used in the production of filaments
Among the different available rapid prototyping techniques, with suitable diameter and homogeneity for FDM. This information
including Stereolithography (SLA), Selective Laser Sintering (SLS) would be essential for the development of new filament materials
[5] or Laminated Object Manufacturing (LOM) [6], to mention just for FDM and for the drive of this technique into the production of
a few, the Fused Deposition Modeling (FDM) [7] technique is the final consumer products. In fact, the optimization of the filaments
trendiest one, although already existed since the 80s. Currently, production could maximize their properties, providing a deeper
FDM is the technique showing the higher potential for product control of the process. The use of commercially available filaments,
manufacturing, with the capability to compete with conventional used in the majority of the published studies, limits them to the
polymer processing techniques [8]. inherent characteristics of the FDM process [16].
Though, the range of filaments commercially available for FDM Moreover, the comparison either at the morphological and
is still somehow limited and costly, which hinders the use of this mechanical level of the parts produced by FDM and by conven-
technology for manufacturing final products. Neither the materials tional techniques would be more authentic if the material used
nor the process have been studied in a systematic manner towards to produce the parts were to be the same in both processes.
functional components production, with adjusted mechanical Literature lacks studies that follow the FDM process from the start,
properties, or with the objective of getting competitive production i.e., from the filament production, allowing an absolute control of
time/cost (for small/medium production series), respectively the processes and a true comparison between competing tech-
[3,8,9]. niques [17].
This paper aims to evaluate the FDM technique in its entire
chain, from the filament production to the printing of final parts,
⇑ Corresponding author.
allowing understanding what are the advantages and limitations
E-mail address: [email protected] (O.S. Carneiro).

http://dx.doi.org/10.1016/j.matdes.2015.06.053
0264-1275/Ó 2015 Elsevier Ltd. All rights reserved.
 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776 769

of FDM towards conventional manufacturing processes. In order to working principle that combines all parameters together, namely
perform such task, two comparisons are performed: at the tech- the filament feeding rate, extrusion width, linear plotting speed
nique level and at the material level. At the technique level, FDM and layer thickness. These parameters are tied together as the
is compared to a conventional polymer processing technique, speed of an FDM system is dependent on the linear plotting speed
namely compression molding, the reference technique for the and feeding rate, being the latter also dependent on the rate at
study. At the material level, commercially available polymer sys- which the liquefier can melt the material and feed it through the
tems, not available for FDM, were selected, which forced their pre- nozzle [20].
vious extrusion into filaments to enable their use both in the FDM Similarly to many other additive manufacturing systems, FDM
printer (in filament form), and in the compression molding tech- starts with a CAD file (generally in .stl format). This file is used to
nique (in pellet shape). In this case, a polypropylene homopolymer create the cross-profile (the slices) that will be printed
(PP), one of the most common plastics used in daily and technical layer-by-layer. The CAD file only gives the layers’ outline, being
products, was selected as the material. Two grades of PP were the software used responsible for defining the infill of each layer
tested: a glass-fiber reinforced (GRPP) and a neat, non-reinforced and setting the plot and path of the nozzle [20]. The final setting
(PP). By selecting a commodity polymer, non-conventional for of the print file defines towards which end of the printing spec-
FDM, it becomes possible to evaluate from the filament production trum one looks: if towards the geometric resolution (fine prints)
stage, how one can use conventional polymer systems in a or if towards the mechanical performance.
FDM-based 3D printer, and to compare the performance of the For a better understanding of the FDM range in terms of print
parts produced with this technique with those produced by a con- quality, Fig. 2 illustrates the influence of the feeding rate and linear
ventional polymer processing technique. Also, by selecting neat movement speed, for a given layer thickness, on the extrusion
and glass-reinforced grades, the reinforcing effect of fibers will width. An increase in the feeding rate leads to an increase of the
be assessed. extrusion width, and similar scenario occurs when the linear plot-
ting speed of the extrusion head, or nozzle, is decreased, for a given
feed rate.
2. Fused Deposition Modeling (FDM) technique
If one looks now to the adjacent paths (Fig. 3), there is a range
comprised by the following extreme scenarios: (i) adjacent paths
The Fused Deposition Modeling process (Fig. 1) is a RP technol-
very close to each other, resulting in an overlap of paths and excess
ogy, originally developed by Stratasys, which horizontally deposits
of deposit filament; and (ii) adjacent paths distant from each other,
molten thermoplastic material, such as ABS and PLA (the two most
resulting in gaps and minimal or non-existing bonding between
common materials used in FDM), extruded from a nozzle head, fab-
them, which weakens the structural integrity of the part. Thus, it
ricating parts layer-by-layer [7,18,19]. A filament is softened and
becomes necessary to optimize the width of the path, by extruding
melted inside the liquefier at a temperature above its melting
more or less material via changing the feed rate/head’s linear plot-
point, and is pushed through the nozzle die by the still solid
ting speed ratio. Although thicker paths lead to a better bonding
upstream filament. As the liquefier moves, the extruded polymer
and thus better mechanical performance, it will most probably fail
is laid down, starting with the object perimeters and then the fill-
to meet the geometrical resolution. For the latter purpose, a thin
ing, for each layer of the object [7,18].
path will ensure shape accuracy in detriment of the mechanical
In FDM, over-hanging features may need to be supported by a
properties.
structure, whose style can be either break-away or water-soluble
FDM is a technique that presents the advantages of being a
support structures [19]. After part printing, the formers are manu-
highly-reliable process, currently requiring a low initial invest-
ally removed by stripping them from the part surface, while the
ment as well as using relatively low-cost materials. It can be oper-
latter are dissolved in a water/solvent solution that does not inter-
ated in office environments, with short built time for parts with
act with the part’s structural material. The water-soluble support-
thin walls, low material waste (limited to the supporting struc-
ing structures, unlike the break-away, can be located in deeply
tures) and enabling the use of different materials or colors in the
recessed regions, and can also be in contact with small features,
same object or layer. On the downside, the materials it uses must
as they are preserved.
have low melting temperature, and if the parts require supports
Although the basic principle behind the FDM is quite straight-
on the overhangs, they will have a poor surface finish and grainy
forward, there is a control model supporting the entire system
appearance, requiring time-consuming hand-working procedures
to improve surface aesthetics.

Filament 3. Experimental work

In this section, the materials, equipment and conditions used in

Filament Pulling the production and mechanical characterization of the samples is
System described in detail.

3.1. Materials and filament extrusion

The materials used in this study were two commercially avail-

able polypropylene homopolymer extrusion grades: ISPLENÒ PP
040 C1E, from Repsol (neat PP) and POLIFOR L6 GF/30
Z NATURALE, from SOFTER, a PP reinforced with 30%, in weight, of
glass fiber (GRPP). Their main characteristics are listed in Table 1.
Object X Polypropylene is used in a variety of application areas such as
Y textile, packaging, furniture, household appliances, medical
devices, automotive and aeronautics, among others. The
Fig. 1. Illustration of the FDM technique [based on RepRap]. homopolymer variant has a high stiffness and the highest service
770 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776

Higher Feeding Rate

or
Lower Plotting
Linear Speed

Nozzle Nozzle

Layer
Thickness

Extrusion
Width

Fig. 2. FDM extrusion width parameter illustration.

Underfilled Area

Overfilled
Area

Tool
Path

Contour
Path

Fig. 3. Illustration of the paths influence when seeking to obtain geometric

precision or mechanical performance.

Fig. 4. General view of the prototype extrusion line used in the production of the
filaments.
Table 1
Main characteristics of the polypropylene grades used.

Property Method PP GRPP Cooling

MFI (g/10 min) ISO 1133 3.0 2.5 Reservoir Support
Oven Pulling Unit
Cylinder
Flexural modulus (MPa) ISO 178 1800 5500
Vicat softening point (°C) ISO 306 153 (10 N) 135 (50 N) Extruder Filament
Processing temperature (°C) – 190–250 190–230

temperature among the commodity thermoplastics. It has also a

relatively high shrinkage coefficient, which negatively affects its
dimensional stability. Its main limitation is the low impact resis-
tance at negative temperatures, characteristic that can be over-
come by copolymerization (with ethylene). Glass reinforced
grades have better thermal and mechanical properties than neat
PP (as can be seen in Table 1), being often used in the production
of semi-structural components for the automotive and aeronautics
industries.
The two grades were supplied in pellets shape and therefore,
their use in the 3D Printer required the previous production of fil-
aments, carried out in a proper extrusion line. The extruder was
equipped with a 2 mm diameter die, especially designed to pro-
duce filaments of circa 1.75 mm diameter, as required by the prin-
Fig. 5. General view of the prototype extrusion line used in the production of the
ter used in this study. The referred extrusion line is illustrated in filaments.
Fig. 4 and the set of their components used in this study is schema-
tized in Fig. 5.
Several extrusion trials, performed in different conditions, were internal voids in the filament); use of the oven at circa 30 °C; and
required to attain the desired circular filaments with circa use of effective pulling promoted by the pulling unit to attain the
1.75 mm diameter. The final extrusion line setup resulted in the required filament diameter. The detailed processing conditions
following (see Fig. 5): use of forced air, to cool down the extrudate adopted for the extrusion of the two materials are depicted in
in the cooling reservoir, instead of water (to avoid the formation of Table 2.
 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776 771

Table 2
Processing conditions used in the filament extrusion of the polypropylene grades
used.

Parameter PP GRPP
Extruder/die temperature (°C) 190-220/220 200-230/230
Extruder screw speed (rpm) 11 6.3
Pulling rolls speed (m/min) 7 5
Oven temperature (°C) 30 30
Distance die-cooling reservoir (cm) 3 4.5

3.2. Production of samples

3.2.1. 3D printer
The 3D printer used in this study was based on an open-source
model known as Prusa i3 (available from RepRap platform [21]).
This is a fully customizable printer that, as any other FDM type
of printers, has as its key elements: the extrusion head, the gantry,
the build surface and the build environment. The full printer is
showed in Fig. 6 and its main features are described next.
Gantry
The XYZ linear movements of the extrusion head and printer Fig. 7. Illustration of the extrusion head.
bed are based on a set of lead-drive (Z-axis) and belt mechanisms
(X- and Z-axes) systems, and on four stepper motors. The bed per-
extrudate is pushed out by the upstream yet solid filament. In
forms only the Y-axis movements while the extrusion head move
order to melt the filament, the nozzle is made of brass and has a
on the X and Z-axes.
heater and a thermistor connected to it. For this study, a nozzle
In order to ensure horizontal leveling of the gantry when per-
with a 0.4 mm die diameter was used.
forming Z-movements, this system uses two motors, one on each
The nozzle holder is built in polyether ether ketone, PEEK. It
side of the gantry, that work in parallel to avoid any unbalance
holds the nozzle and performs the interface between the frame
of this structure.
of the extrusion head (made of acrylonitrile butadiene styrene,
To complete the gantry, a calibration is required before the first
ABS) and the nozzle, working also as heat insulator between both
use to guarantee the leveling of the extrusion head against the
components.
printer’s bed. This is a crucial step as the levelness of the system
Inside the nozzle holder, one can find a polytetrafluoroethylene,
leads to a non-homogeneous stacking of consecutive layers.
PTFE, liner that is used to guide the filament to the nozzle head.
Extrusion head
This tubing has very low friction coefficient and high working tem-
The printer’s extrusion head is based on the EiNSTeiN Compact
perature, which ensures the required guiding of the filament
Extruder variant which includes the material feeding mechanism,
towards the melting site. Also, the tubing enables the nozzle to
extruder and nozzle. Fig. 7 shows an illustration of this system.
work with different filament diameters. In this case, the filaments
The feeding mechanism uses a 50:1 ratio motor to pull the fila-
used had a 1.75 mm diameter.
ment from the coil and to push it towards the nozzle, by trapping
Build surface
the filament between the rotating motor transmission gear and a
The printing was done on top of a glass plate (20  20 mm) that
bearing.
underneath has a heating bed (MK2 model), which has an embed-
The nozzle (type J-head nozzle [21]) is a three piece component:
ded resistor (of circa 1 Ohm). When electric current flows through
the nozzle itself, the nozzle holder and the liner. The aforemen-
the resistor, the energy dissipates as heat, due to the joule effect, to
tioned nozzle is the site where the filament is melted and one looks
the printing surface. In combination with the power supply, the
to have this phase transition close to the nozzle die so that the
heating bed is able to reach 180 °C.
Controller
The electric control of the system is performed via an Arduino
Filament MEGA with a shield to connect with all electrical components
Visual spool
required for the printer. The shield used was a RAMPS 1.4 model
interface specially designed for this type of 3D printers. At the computer
level, two open-source software were used: a slicer tool (Slic3r)
Cooling fan and a printer control tool (Printrun).
Controller
Extrusion
3.2.2. Printing parameters
head
As there is no information about the printing conditions for
polypropylene homopolymer, PP, and especially for the glass rein-
forced one, GRPP, several print runs were first performed to deter-
Heang mine the optimal printing conditions with these two materials. A
Bed systematic strategy was pursued, starting with the production of
thin wall (single filament thick) objects in order to determine the
layer thickness, nozzle temperature and the conditions leading to
a good adhesion between the printed part and the heating bed.
Nozzle
During this process, PP showed a high degree of shrinkage during
cooling (due to its semi-crystalline nature), which led to distortion
Fig. 6. 3D Printer used in the study. and decoupling from the printer’s bed (see Fig. 8).
772 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776

In order to obtain adequate printing conditions, tens of print-

ings were performed varying the nozzle temperature, the heating
bed material and temperature, and the linear plotting speed.
Table 3 resumes the optimal conditions for printing with both
materials.
In addition to the above operating conditions, a few details are
worth being highlighted, namely regarding the printing surface of
the bed. PP showed low adhesion to the typical surfaces used in 3D
printer (namely, glass and blue tape). The solution found to
improve the adhesion of the printing object to the bed was to
use a PP plate, ensuring chemical compatibility. However, the plate
needed to be pre-processed by first scrubbing its surface with a
steel brush and then cleaning it with alcohol. This solution is based
on a similar approach used in thermal welding of PP [22], to pro-
vide a strong adhesion between parts. For the case of neat PP,
the difference between the objects printed with non-optimal and
with the optimal conditions described can be clearly seen when
Fig. 9. PP sample printed with optimal conditions.
comparing the corresponding objects shown in Figs. 8 and 9.

3.2.3. Experimental plan

With the definition of the optimal printing conditions for each
material, the same object (a tensile test specimen) was printed
for analysis and benchmarking in regard to the mechanical proper-
ties. The aforementioned specimens were printed according to the
nominal dimensions from the DIN 53504-S3a standard (Fig. 10).
The same tensile specimen was manufactured via compression
molding to make possible comparing the mechanical performance
of specimens manufactured by both processes.
At the 3D printing process, several specimens were printed with
different infill degrees, deposition orientation and layer thickness.
The infill is one of the key parameters of this FDM process as it Fig. 10. Nominal dimensions according to the DIN 53504-S3a standard (mm) [11].
enables the print of pieces with infill lower than 100%, which in

practical terms reduces the printing time and saves material. At

Table 3
the orientation level, one looks to evaluate the influence of the
Optimal printing conditions for neat polypropylene (PP) and glass reinforced
polypropylene (GRPP). deposition orientation on the mechanical properties. With this pro-
cess, one can choose a given constant orientation for the deposition
Process parameter PP GRPP
or alternate orientation in each layer. The layer thickness is mainly
Bed PP scrubbed PP scrubbed plate or Blue responsible for the geometrical resolution of the part, and is
plate tape (tesaÒ 56250)
directly related to the nozzle diameter. The rule of thumb states
Nozzle temperature (°C) 165 185
Bed temperature (°C) Room Room temperature (PP that the layer thickness should be lower than 80% of the nozzle
temperature Plate) thickness. Table 4 summarizes the evaluated parameters.
80 (Blue tape) It should be mentioned that for the cases where the infill degree
Linear plotting speed for the 8 8 was lower than 100%, two filaments were used, in each layer, to
1st layer (mm/s)
Linear plotting speed for the 60 8
define the external contour of the specimens.
other layers (mm/s) In order to evaluate each parameter in a balanced manner, a
standard print condition was set, being, in this study, a testing
specimen in PP with an infill degree of 100%, an alternate orienta-
tion of 45° and a layer thickness of 0.20 mm.
The infill degree of 100% was chosen as it is the one that best
mimics the compaction of samples produced by compression
molding. The alternate orientation and layer thickness selected
are typical values for 3D FDM printing with a 0.4 mm nozzle diam-
eter. Overall, nine types of testing specimens of six units each were
printed, according to Table 5.

Table 4
Experimental plan summary.

Parameter Value
Infill degree (%) 20, 60 and 100
Orientationa (°) 45, 0, 90, crossed 45 (±45) and crossed 0–90
Layer thickness, t (mm) 0.20 and 0.35
a
Fig. 8. PP sample printed with non-optimal conditions. see Fig. 11.
 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776 773

encompassed the pre-heating of the pellets during 2 min, followed

(a)
by a compression stage of 10 min at 10 TON. Cooling was also per-
formed under pressure until the press plates reached a tempera-
ture of 80 °C. After this stage, the polymer plates were taken
from the mold; (ii) use of a cutter to obtain, from the plates, tensile
test specimens similar to those printed, i.e., with nominal dimen-
sions according to the DIN 53504-S3a standard. This operation
was also carried out in the Moore press, with the plates heated
at 80 °C.

(b) 3.3. Tensile tests

For each sample production condition, six tensile tests were

performed in a universal testing machine, INSTRON 4505, at a
cross-head speed of 50 mm/min and an initial distance of 25 mm
between grips. The tensile modulus and strength were computed.

4. Results and discussion

(c)
In this section, the effect of the printing conditions, the raw
materials and the manufacturing techniques on the mechanical
performance of the samples is assessed.

4.1. Effect of printing orientation

For the study of this parameter, the samples printed with differ-
ent orientations (±45°, 0°, 90°, 45° and 0–90°) and constant layer
(d) thickness (0.20 mm) and infill degree (100%), were compared.
Their mechanical properties are shown in Fig. 12, where it can be
seen that the 0° orientation set of samples is the one showing
the best performance. The remaining samples have similar moduli
(if the corresponding standard deviation values are taken into
account) and, in terms of tensile strength, the crossed 45° (±45°)
samples were the ones with the worst performance.
Having in mind the type of tests performed, the best perfor-
mance of the 0° orientation printed sample sounds obvious. In fact,
Fig. 11. Types of orientation used in 3D printing of the tensile test specimens: (a) in this case the stress applied is aligned with the deposition direc-
45°; (b) crossed 45° (±45); (c) 0°; (d) 90°. tion, minimizing, therefore, problems related with adhesion
between filaments belonging to the same layer (due to the lack
of pressure inherent to the 3D printing process). Thus, in this case
Table 5 the interfaces between neighboring filaments are parallel to the
Conditions used to print the samples.
applied force, being not subjected to tensile forces. The similar per-
Material Orientation (°) Layer thickness (mm) Infill degree (%) formance of the remaining samples suggest, however, that the
PP ±45 0.20 100 cohesion between filaments of the same layer is relatively good
0 0.20 100 since even the sample produced with an orientation of 90° per-
90 0.20 100 forms as well as the others. It should be noted that in this case
45 0.20 100
the interfaces are subjected to the maximum force during the ten-
0–90 0.20 100
±45 0.35 100 sile test since they are perpendicular to the applied load. The most
±45 0.20 20 unexpected result is that obtained for the crossed orientation of
±45 0.20 60 45°. In this case, one would expect a similar performance to that
GRPP ±45 0.20 100 of samples produced with an orientation of 45°. The only plausible
explanation for this decrease in performance is a less effective
Note: reference printing condition in bold.
stacking of the filaments of two consecutive layers.
3.2.4. Compression molding
The selection of compression molding as the reference process 4.2. Effect of the layer thickness
was mainly due to the simplicity of the process (comparable to that
of 3D printing), to the low degree of anisotropy resulting for the The effect of the layer thickness was assessed comparing sam-
moldings, which enables a fair comparison between the two tech- ples produced with layer thicknesses of 0.20 and 0.35 mm, main-
niques, and to the fact that this process is commonly used with taining constant the remaining printing parameters. As can be
thermoplastic composites, such as GRPP, for the production of observed in Fig. 13, the influence of this parameter is not
semi-structural components. remarkable.
The production of the compression molded samples (tensile test The samples produced with the higher thickness layer value
specimens) was carried out in the following sequence: (i) compres- show, however, a slightly higher tensile strength. Most probably,
sion molding of 2 mm thickness plates, using a hot plate hydraulic this is a consequence of the lower number of interfaces between
press Moore (at 230 °C) and a plate mold. This operation filaments, when compared to that corresponding to the reference
774 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776

0º 45º 90º 0-90º ±45º 0º 45º 90º 0-90º ±45º

Filament Orientation Filament Orientation
(a) (b)
Fig. 12. Tensile properties of polypropylene samples printed with different filament orientations: (a) modulus; (b) strength.

0.2 0.35 0.2 0.35

Layer Thickness (mm) Layer Thickness (mm)

(a) (b)
Fig. 13. Tensile properties of polypropylene samples printed with different layer thicknesses: (a) modulus; (b) strength.

ones (produced with 0.20 mm layer thickness), existing in this type potential of the use of fiber reinforced materials for the FDM
of samples. process.

4.3. Effect of the infill degree 4.4.1. 3D printing versus compression molding
For the reference conditions, the mechanical properties of the
The infill degree has a strong impact on the mechanical perfor- samples produced via 3D printing show a decay of circa 30% in
mance of the samples as can be observed in Fig. 14, resulting in dif- relation to those produced by compression molding, as shown in
ferences of more than 250%, in both modulus and strength, when Fig. 16. However, if the best printed samples (produced with 0° ori-
its value varies from 20% to 100%. Furthermore, the effect is linear: entation) are considered, this decay is lower than 20%. As already
above a ‘base property value’, given by the contour filaments that mentioned, we believe that this decay is not promoted by poor
were used to print this particular set of samples, the variation of adhesion between the filaments but due to the existence of voids
modulus and strength as a function of the infill degree was fit by (poor compaction) in the samples.
a linear regression resulting in values of the Pearson correlation Finally, it is worth mentioning that the performance of the GRPP
factor of 0.999 and 0.979, respectively. printed samples is similar to that of the neat PP ones produced by
compression molding. This enables to envisage the use of the 3D
4.4. PP versus GRPP printing technique to produce parts with similar properties to those
of PP parts produced by conventional processing techniques if
In what concerns to the raw materials used, better mechanical proper printing conditions and fiber reinforced grades are employed.
properties were obtained for GRPP (see Fig. 15), that showed
higher values of circa 30% and 40% for the modulus and strength, 4.5. Overall picture
respectively, in regard to PP. Therefore, it can be concluded that
the glass fibers preserved their role (maintaining a minimum crit- Fig. 17 gives a global picture of the study performed, enabling
ical length) even after the extrusion stage (for the filament produc- the direct comparison of all the samples and a clear idea of the rel-
tion) and the 3D printing process. This puts in evidence the ative importance of each parameter.
 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776 775

20 60 100 20 60 100
Infill (%) Infill (%)
(a) (b)
Fig. 14. Tensile properties of polypropylene samples printed with different infill degrees: (a) modulus; (b) strength.

PP GRPP PP GRPP
(a) (b)
Fig. 15. Tensile properties of samples printed with PP and GRPP at the same (reference) conditions: (a) modulus; (b) strength.

PP GRPP PP GRPP PP GRPP PP GRPP

±45º; 0.2mm ±45º; 0.2mm Compression Compression ±45º; 0.2mm ±45º; 0.2mm Compression Compression

(a) (b)
Fig. 16. Tensile properties of polypropylene and glass reinforced polypropylene produced by 3D printing and by compression molding: (a) modulus; (b) strength.
776 O.S. Carneiro et al. / Materials & Design 83 (2015) 768–776

3000 60
Е σ máx.

Ultimate Tensile Strength (MPa)

2500 50

Young Modulus (MPa)

2000 40

1500 30

1000 20

500 10

0 0

Material PP PP PP PP PP PP PP PP GRPP PP GRPP

Orientation (º) ±45 0 90 45 0-90 ±45 ±45 ±45 ±45 - -
Layer thickness (mm) 0.20 0.20 0.20 0.20 0.20 0.35 0.20 0.20 0.20 - -
Infill degree (%) 100 100 100 100 100 100 20 60 100 - -
Technique FDM FDM FDM FDM FDM FDM FDM FDM FDM CM CM

Fig. 17. Tensile properties (modulus and strength) of all the printed (FDM) and compression molded (CM) samples.

5. Conclusions [3] Y. Zhai, D.A. Lados, J.L. LaGoy, Additive manufacturing: making imagination the
major limitation, JOM 66 (2014) 808–816.
[4] D. Bak, Rapid prototyping or rapid production? 3D printing processes move
The approach used in this study enabled the full control over industry towards the latter, Assem. Autom. 23 (2003) 340–345.
the complete process, from the extrusion of the filaments to the [5] J. Kruth, P. Mercelis, J. Van Vaerenbergh, L. Froyen, M. Rombouts, Binding
mechanisms in selective laser sintering and selective laser melting, Rapid
printing of samples, and a fair comparison (using exactly the same
Prototyp. J. 11 (2005) 26–36.
materials) between competing technologies, avoiding many of the [6] B. Mueller, D. Kochan, Laminated object manufacturing for rapid tooling and
issues identified in other studies. patternmaking in foundry industry, Comput. Ind. 39 (1999) 47–53.
[7] M. Too, K. Leong, C. Chua, Z. Du, S.F. Yang, C.M. Cheah, S.L. Ho, Investigation of
The main partial conclusions of this study were the following:
3D non-random porous structures by fused deposition modelling, Int. J. Adv.
(i) given the results obtained with different printing orientations, Manuf. Technol. 19 (2002) 217–223.
the adhesion between adjacent filaments is evident but, as [8] J. Kruth, M. Leu, T. Nakagawa, Progress in additive manufacturing and rapid
expected, the samples are stiffer in the filament direction; (ii) prototyping, CIRP Ann. Technol. 47 (1998) 525–540.
[9] D.E. Cooper, M. Stanford, K.A. Kibble, G.J. Gibbons, Additive manufacturing for
the thickness of the layers has little influence on the mechanical product improvement at Red Bull technology, Mater. Des. 41 (2012) 226–230.
performance of the samples; (iii) the infill degree has a dramatic [10] O.A. Abdelaal, S.M. Darwish, Review of rapid prototyping techniques for tissue
and linear effect on the mechanical properties; (iv) the use of engineering scaffolds fabrication, in: A. Öchsner, L.F. Silva, H. Altenbach (Eds.),
Characterization and Development of Biosystems and Biomaterials, Springer
fibers as reinforcement is also effective in 3D printing; (v) the Berlin Heidelberg, Berlin, 2013, pp. 33–54.
loss in the mechanical performance of the printed samples is [11] E.L. Melgoza, G. Vallicrosa, L. Serenó, J. Ciurana, C.A. Rodríguez, Rapid tooling
circa 20–30%, depending on the printing parameters values used, using 3D printing system for manufacturing of customized tracheal stent,
Rapid Prototyp. J. 20 (2014) 2–12.
when compared to that of samples produced by compression [12] R. Ilardo, C.B. Williams, Design and manufacture of a formula SAE intake
molding; and (vi) the use of enhanced (fiber reinforced) grades system using fused deposition modeling and fiber-reinforced composite
enables to cancel the afore-mentioned decay in properties. materials, Rapid Prototyp. J. 16 (2010) 174–179.
[13] V.K. Vashishtha, Advancement of rapid prototyping in aerospace industry – a
As a final conclusion we believe that there is room to further review, IJEST 3 (2011) 2486–2493.
improve the performance of the printed samples, making this pro- [14] N. Sa’ude, M. Ibrahim, M.H. Ibrahim, Melt flow behavior of metal filled in
cess competitive when compared to the conventional ones, for the polymer matrix for fused deposition modeling (FDM) filament, Appl. Mech.
Mater. 660 (2014) 84–88.
production of small series of parts/components. FDM has, there-
[15] P. Dudek, FDM 3D printing technology in manufacturing composite elements,
fore, the potential to surpass the limitations associated to the Arch. Metall. Mater. 58 (2013) 12–15.
mechanical performance of the produced parts and shall not be [16] D. Drummer, S. Cifuentes-Cuéllar, D. Rietzel, Suitability of PLA/TCP for fused
restricted to the production of mockups and prototypes. deposition modeling, Rapid Prototyp. J. 18 (2012) 500–507.
[17] M. Domingos, F. Chiellini, A. Gloria, L. Ambrosio, P. Bartolo, E. Chiellini, Effect of
process parameters on the morphological and mechanical properties of 3D
Acknowledgements Bioextruded poly(e-caprolactone) scaffolds, Rapid Prototyp. J. 18 (2012) 56–
67.
[18] S.H. Masood, W. Rattanawong, P. Iovenitti, Part build orientations based on
This work is funded by FEDER funds through the COMPETE volumetric error in fused deposition modelling, Int. J. Adv. Manuf. Technol. 16
2020 Program and National Funds through FCT - Portuguese (2000) 162–168.
Foundation for Science and Technology under the projects [19] T. Grimm, Fused Deposition Modeling: A Technology Evaluation, T.A. Grimm &
Associates, Inc., Atlanta, 2002.
UID/CTM/50025/213 and UID/EEA/04436/2013. [20] B. Mueller, Additive Manufacturing Technologies, Springer Science+Business
Media, New York, 2012.
References [21] C. Bell, Maintaining and Troubleshooting Your 3D Printer, Apress, Berkeley,
2014.
[22] P. Barber, J.R. Atkinson, The use of tensile tests to determine the optimum
[1] C.K. Chua, K.F. Leong, C.S. Lim, Rapid Prototyping: Principles and Applications,
conditions for butt fusion welding certain grades of polyethylene, polybutene-
second ed., World Scientific, 2002.
1 and polypropylene pipes, J. Mater. Sci. 9 (1974) 1456–1466.
[2] J. Manyika, M. Chui, J. Bughin, R. Dobbs, P. Bisson, A. Marrs, Disruptive
Technologies: Advances That Will Transform Life, Business, and the Global
Economy, McKinsey & Co., New York, 2013.