You are on page 1of 10

Progress in Additive Manufacturing (2018) 3:205–214

https://doi.org/10.1007/s40964-018-0053-3

FULL RESEARCH ARTICLE

Effect of fused deposition modeling process parameters


on the mechanical properties of a filled polypropylene
Lu Wang1,2 · J. Elliott Sanders1,2 · Douglas J. Gardner1,2 · Yousoo Han1,2

Received: 29 August 2017 / Accepted: 14 May 2018 / Published online: 22 May 2018
© Springer International Publishing AG, part of Springer Nature 2018

Abstract
Few commodity semi-crystalline thermoplastics in neat form are used in fused deposition modeling (FDM) because they
experience dimensional instability (warping) during printing. In this study, a commercially available polypropylene (PP)
filament for FDM processing was used to print tensile and flexural test samples. Three printing parameters were investigated:
(1) extrusion temperature (200, 250 °C), (2) printing speed (45, 90 mm/s) and (3) layer height (0.1, 0.3 mm). Differential
scanning calorimetry (DSC) showed that FDM printed PP samples had fewer α-crystals and more β-crystals than injection-
molded (IM) PP samples. Scanning electron microscopy (SEM) indicated that a higher extrusion temperature, slower printing
speed and smaller layer height facilitated the molecular diffusion at the interfaces and created a smaller neck size within the
printed parts. Density measurements showed that IM PP samples were denser than the FDM PP samples. No differences in
density were found among the FDM PP samples. Compared to the IM PP, the mechanical properties decreased less for the
PP printed at higher extrusion temperature (< 11.5%). The FDM PP parts had similar or even increased flexural modulus
compared to the IM PP.

Keywords 3D printing · Additive manufacturing · Printing parameters · Polypropylene · Morphology · Crystal form ·
Density · Specific mechanical properties

1 Introduction format, the printer builds each layer according to the pro-
grammed pattern [1]. AM exceeds conventional manufactur-
Additive manufacturing (AM), commonly referred to as 3D ing techniques in several aspects: (1) it is able to build parts
printing, is a dynamic processing technology that fabricates with complex geometries which is impossible by traditional
parts layer by layer from the bottom up. Usually an object methods; (2) it does not require additional tooling such as
is drawn using software and stored as a stereolithography dies; and (3) it does not require assembly because products
STL format file. The printer software reduces an object are directly fabricated by the printers [2]. Meanwhile, AM
into slices and converts it into a readable format for the 3D has drawbacks too. Different types of parts require different
printer. After converting the designed part into a readable AM techniques. Techniques that can be used for polymers
include stereolithography, selective laser sintering (SLS),
* Lu Wang selective mask sintering (SMS), laminated object manu-
lu.wang@maine.edu facturing (LOM) and fused layer modeling (FLM) or fused
J. Elliott Sanders deposition modeling (FDM) [1]. During FDM processing
jordan.sanders@maine.edu (Fig. 1), a polymer filament is continuously melted in a
Douglas J. Gardner chamber and extruded through a nozzle. After the first layer
douglasg@maine.edu is built, the platform lowers a certain distance to receive the
Yousoo Han deposition of a second layer. By repeating this process, the
yousoo.han@maine.edu final part can be made. Compared to other rapid prototyping
methods (RP), FDM is cheaper and easier to operate [3].
1
School of Forest Resources, University of Maine, 5755 During the printing process, an interface forms when the
Nutting Hall, Orono, ME 04469‑5755, USA
adjacent laid-down polymer strands come into contact with
2
Advanced Structures and Composites Center, University each other. The interfacial width increases and the interfaces
of Maine, 35 Flagstaff Road, Orono, ME 04469‑5793, USA

13
Vol.:(0123456789)
206 Progress in Additive Manufacturing (2018) 3:205–214

inner structure of samples. Mechanical tests were done to


compare the strength and stiffness properties of the printed
samples and their injection-molded counterparts.

2 Materials and methods

2.1 Materials

PP homopolymer filament (1.75 mm in diameter), Moplen


HP741T, was purchased from Gizmodorks, Temple City,
CA. The HP741T is nucleated, with a modified molecular
weight and high melt flow index (60 g/10 min at 230 °C/2.16
Fig. 1  Configuration of the 3D printer used in this study kg). Its flexural modulus is 1.75 GPa. It features a balance
between rapid cycle time and good mechanical properties.
Food containers are its typical applications. Ash content
gradually disappear as polymer molecular diffusion occurs (~ 30%) of this PP was determined according to ASTM
resulting in an increase in mechanical properties [4]. The D5630-13. Then a metal analysis was done on the ash
interface and pores that are generated in an FDM part are according to “Recommended Soil Testing Procedures for
beneficial for cell reproduction, thus making the technology the Northeastern United States”. The major components of
suitable for medical applications; for example, implants and the identified metals in the ash are listed in Table 1. The
tissue engineering [5]. However, voids lower the mechanical density of this PP filament is around 1200 kg/m3.
properties of the printed parts compared to their injection-
molded counterparts. 2.2 Fused deposition modeling (FDM) of PP
There are commercially available thermoplastics which
can be used for FDM. However, some commodity thermo- The models of testing samples were first drawn in Auto
plastics are not widely used in FDM, for instance, isotactic CAD (2014) according to the dimensions specified in ASTM
polypropylene homopolymer (PP). Neat PP is a very ver- D638-14 (Type I) for tensile tests, and ASTM D790-10 for
satile polymer and is among the most utilized thermoplas- flexural tests. Those files were saved in an STL format for
tics in commercial production of parts [6, 7]. PP shrinks further configuration. The 3D printer used for this study was
rapidly and warps easily during 3D printing, which makes a Makerbot Replicator 2X Experimental 3D Printer (Maker-
it unfavorable for the AM process [8]. Adding PP to the Bot Industries, LLC, NY, USA). Features of this type of 3D
material scope of FDM is desirable and beneficial. A few printer are shown in Fig. 1. Compared to other FDM devices,
papers reported on the production of PP composites by FDM this device has two extrusion heads, enabling the printing
[9–12]. Investigated printing parameters included deposi- of two different filaments within one building process. To
tion orientation, different infill degrees and layer thickness. start a printed part, the STL file was opened in Makerware
Samples were stronger and stiffer when all the filaments software (Version: 3.9.0), centered, laid flat and printed with
were printed along the long axis of the sample. The tensile the right-side nozzle. Printing parameters were changed in
properties increased as the infill density increased. Layer the “settings” section of the Makerware software. Advanced
thickness was found to insignificantly affect the mechanical processing parameters were altered in a customized profile.
properties of samples. To achieve a 0° orientation, “infill orientation offset”, “infill
In this study, we evaluated the effect of layer height (LH), orientation interval”, “solid fill orientation offset” and “solid
extrusion temperature (ET) and printing speed (PS) on the fill orientation interval” were set in the customized profile at
mechanical properties of printed parts using a commercially 90°, 0°, 90° and 0°, respectively. The air gap was changed by
available PP filament. Differential scanning calorimetry varying the number of “grid spacing multiplier” between 0
(DSC) was used to understand how different processing con- and 1. A complete overlap between two adjacent filaments
ditions affected the crystalline form of the samples. Scan- is achieved by choosing 0; while a simple contact occurs
ning electron microscopy (SEM) was applied to study the when 1 is selected. According to the manufacturer, both

Table 1  Major metals contained Elements Ca K Mg Al Fe Zn


in the PP ash and their
percentages Content (%) 26.4 0.02 1.3 0.04 0.03 0.07

13
Progress in Additive Manufacturing (2018) 3:205–214 207

an extrusion temperature range of (210, 230 °C) and (250, direction from the bottom to the top of the sample. Samples
280 °C) should result in good printing quality. Thus, two were sealed in aluminum pans and heated from 25 to 200 °C
temperatures within those ranges were chosen. A platform at a rate of 10°C/min.
temperature of 130 °C was chosen as it is in accordance
with the upper limit of the printer software. The default 2.5 Scanning electron microscopy (SEM)
printing speed was 90 mm/s. A slower printing speed of
45 mm/s was chosen for comparison. Infill density was set Morphological analyses of the FDM samples were carried
to 100% to produce solid samples. The number of shells was out using a TM 3000 SEM (Hitachi High-Technologies Cor-
reduced to 1 to minimize its influence on the mechanical poration, Tokyo, Japan). All samples were cryofractured
properties. The layer width was fixed at 0.4 mm regardless in liquid nitrogen to prepare the surfaces because either
of the layer height. To improve the adhesion between the microtome or hand trimming smeared the surface severely.
first PP layer and platform, a piece of office packing tape Because of the low definition requirement in this research,
(Office ­Depot®, OfficeMax # 24767995) was laid down on no sputter-coating was done before the microscopic observa-
the platform before printing. The total experimental design tions. SEM images were taken at an accelerating voltage of
is shown in Table 2. 15 kV at various magnifications.

2.3 Injection molding 2.6 Density

The PP filaments were pelletized in a Hellweg MDS 120/150 An X-ray densitometer (QMS, Model QDP-01) was
granulator (Hackensack, NJ). The resulting pellets were then used to measure the density profiles of both injection-
processed in an injection molder Model #50 “Minijector” molded and printed samples. Sample dimensions were
with a ram pressure of 17 MPa at 200 °C and held in the 63.5 × 12.5 × 3.2 mm. Density determination by the scan-
molds for 10 s to cool. The mold provides a sample with ning system is based on the relationship between X-ray
dimensions specified in ASTM D638-14 (Type I) and ASTM attenuation and density as expressed in the following equa-
D790-10. The samples were then put into plastic containers tion [13]:
and stored in desiccators to maintain dryness. Before testing,
samples were conditioned in a chamber for at least 40 h at I∕I0 = exp ⋅(𝜇𝜌t), (1)
23 °C ± 2 °C and 50 ±10% RH. where I is the intensity of radiation beam after passing
through the sample, I0 is the intensity of radiation beam
2.4 Differential scanning calorimetry (DSC) before passing through the sample, µ is the material mass
attenuation coefficient (­ m2/kg), ρ is the sample density (kg/
Melting behavior of the processed PP was studied by DSC m3) and t is the sample thickness (m). At least five samples
using a TA Instruments Q 2000 Calorimeter (New Castle, were tested for each group.
Delaware, USA). All measurements were performed under
nitrogen (nitrogen flow = 50 mL/min) to avoid degrada- 2.7 Mechanical properties
tion of PP upon heating. The samples were obtained from
a 0.5 mm-thick slice cut from the cross section of each Tensile and flexural tests were performed according to
specimen. The DSC samples covered the whole thickness ASTM D 638-14 and ASTM D 790-10 to obtain data on

Table 2  Experimental design Processing method Denomination ET/°C PS/(mm/min) LH/mm


and sample nomenclature
Injection molding IM – – –
Fused deposition modeling FDM
250/90/0.3 250 90 0.3
250/90/0.1 250 90 0.1
250/45/0.3 250 45 0.3
250/45/0.1 250 45 0.1
200/90/0.3 200 90 0.3
200/90/0.1 200 90 0.1
200/45/0.3 200 45 0.3
200/45/0.1 200 45 0.1

ET extrusion temperature, PS printing speed, LH layer height

13
208 Progress in Additive Manufacturing (2018) 3:205–214

the tensile and flexural properties. Tests were conducted at 0.7 IM 250/90/0.3 200/90/0.3
room temperature 23 ± 2 °C and 50 ± 10% RH. A universal 0.6 α-crystal
testing machine (Instron 5966) with a 10 kN load cell was 0.5
used for the tests. An extensometer was mounted to measure

Heat Flow (W/g)


0.4
the elongation of the samples. The tensile test speed was set 0.3 β-crystal
at 50 mm/min to break the sample within 5 min. The span
0.2
for the flexural test was 51 mm. The outer fiber strain rate for
0.1
flexural tests was 0.01/min, resulting in a flexural test speed
of 1.4 mm/min. The flexural specimens were positioned 0
90 100 110 120 130 140 150 160 170 180 190
with the bottom facing down. Five replicates of each sample
Endo Up Temperature (°C)
were tested. The average and standard deviation of Young’s
modulus and strengths from tensile and flexural tests were
calculated. Because printed specimens had voids of different Fig. 2  DSC results of IM and FDM PP at scanning rate of 10 °C/min
sizes, to compare the mechanical properties among injec-
tion-molded and printed specimens, their specific mechani- a higher extrusion temperature have less α-crystal character
cal properties were obtained using the following equation: compared to the IM PP, however, displaying a new peak
around 150 °C. This peak is reported to be the β-crystal
Specific mechanical properties = mechanical properties∕density.
form of PP [14]. Factors causing the formation of β-crystals
(2)
include a special nucleating agent and processing strategy
2.8 Statistical analysis
[15, 16]. β-crystals grow faster than α-crystals when crys-
tallized between 100 and 130 °C [11, 14]. In this study,
The density, tensile and flexural properties and specific
the printer build platform was set at 130 °C, which favored
mechanical properties were analyzed using a three-way
β-crystal growth. More detailed description of the formation
analysis of variance (ANOVA) along with a student test at
of β-crystals during FDM manufacture can be found in this
α = 0.05. All the analyses were done using the JMP statisti-
literature [11]. It is important to point out that the percent-
cal analysis program (JMP Statistical Discovery Software
age of β-crystals is much higher in PP processed at a lower
Version 8 2008). A statistical model was used to represent
extrusion temperature. This is because when the extrusion
the properties of PP.
temperature was set at 250 °C, the crystallization temper-
Yijkl = 𝜇 + 𝛼i + 𝛽j + 𝛾k + (𝛼𝛽)ij + (𝛽𝛾)jk + (𝛼𝛾)ik + (𝛼𝛽𝛾)ijk + eijkl , ature for PP during printing exceeded 130 °C which sup-
(3) pressed β-crystal formation. During the DSC measurements,
where i = 1,2; j = 1,2, k = 1,2 and l = 1,2,3,4,5. Yijkl is the possible β-α recrystallization may occur [17, 18]. Also, the
mean of each property; µ is the population mean of each melting enthalpies of β-PP or α-PP were reported with dif-
property. The effects of extrusion temperature, printing ferent values caused by different methods of determination
speed and layer height on each property were represented [18]. Therefore, the β-crystal contents of specimens were
by αi, βj and γk. The effects of interaction between two of determined by X-ray diffraction in a separate study, which
the three factors on each property are represented by (αβ)ij, will be referred to during the discussion of the mechanical
(βγ)jk and (αγ)jk. The three-way interaction is represented by properties of printed samples [11].
(αβγ)ijk. eijkl is the error for this model. After the ANOVA
test was done, a t test was used to investigate whether sta- 3.2 Morphology of FDM PP
tistically significant differences existed between the IM PP
and the FDM PP samples. The micrographs of cryofractured FDM PP surfaces are
shown in Fig. 3. The observed white particles in the SEM
micrographs are attributed to the filler (metal salts) inside
the PP matrix. Each polymer strand laid down is called a
3 Results and discussion “road”. Generally, “necks” are formed by incomplete diffu-
sion among four adjacent roads as shown in Fig. 4. Several
3.1 Crystal forms important observations can be made by comparing groups.
Samples with 0.1 mm layer height have more interfaces, but
The DSC curves of the PP samples are shown in Fig. 2.
a smaller neck size than the samples made with 0.2 mm layer
In each case, at least two endothermic peaks are shown at
heights. As the layer height decreased, the distance between
around 120 and 165 °C, which correspond to the melting
the centers of two adjacent roads (up and down direction)
temperatures of a proprietary ingredient and the α-crystals
decreased. The smaller layer height shortened the distance
in isotactic PP (iPP) [14, 15]. FDM PP samples produced at
required to achieve the same degree of diffusion compared

13
Progress in Additive Manufacturing (2018) 3:205–214 209

Fig. 3  SEM micrographs of


FDM PP printed using different
parameters

to larger layer height (Fig. 4). Also, the total time for print- and smaller neck sizes. However, in some areas, roads of
ing a sample increased if layer height was smaller because smaller layer height do not touch each other. A possible rea-
the sample had longer heat exposure during printing. There- son for that is the printer does not have enough accuracy to
fore, if other parameters are controlled, roads with smaller control layer width when the layer height becomes small
layer height should have better diffusion at the interfaces (0.1 mm). The 250 °C extrusion temperature was able to

13
210 Progress in Additive Manufacturing (2018) 3:205–214

Increasing interfacial width Table 3  Selected values from the ANOVA for densities of PP
Source Degree of Sum of squares F value p value
freedom
Road
Corrected total 39 0.0939
Neck
Model 7 0.031 2.2535
ET 1 0.0148 7.5433 0.0098*
PS 1 0.0027 1.3855 0.2478
Layer height: 0.3 mm
LH 1 0.0081 4.1366 0.0504
Heat
ET × PS 1 0.0024 1.2226 0.2771
Time PS × LH 1 0.0018 0.9275 0.3427
Layer height: 0.1 mm ET × LH 1 0.0000025 0.0013 0.9718
ET × PS × LH 1 0.0011 0.5611 0.4593
Fig. 4  PP molecular diffusion across the road interfaces at different Error 32 0.0629
layer heights
ET extrusion temperature, PS printing speed, LH layer height
*Significant level at α = 0.05
produce denser bottom layers. The first bottom layer was in
contact with the platform that is maintained at 130 °C during
printing. This layer experienced the longest increased tem- that the fillers were not evenly distributed inside the PP. The
perature history and provided enhanced polymer diffusion. A effect of unevenly distributed fillers on the density profiles
250 °C extrusion temperature also increased the interfacial of 3D-printed parts and injection-molded parts is distinct.
width and reduced neck sizes across the whole cross section For the 3D-printed parts, the filament was directly used for
because it enabled the roads to diffuse to a greater extent forming the parts. The density variation of the filament is
[19]. The 45 mm/s printing speed resulted in a larger interfa- reflected in the density of 3D-printed parts as indicated by
cial width and a smaller neck size. When the printing speed
was slower, the total printing time for a sample increased.
Each layer received a longer time of heat exposure. There- 1.25
A
fore, the degree of diffusion was higher compared to the 1.2
AB
AB
B
faster printing speed. B B
Density (kg/m3)

1.15 B
B
3.3 Density profiles 1.1

1.05
As seen in Table 3, among all the printing factors, only 1
extrusion temperature had a significant effect on the density
0.95
of the FDM printed PP. The results in Fig. 5a show that
samples made at 250 °C, except Group 250/45/0.1, were
not significantly different from those printed at 200 °C. The
effect of the printing setting on the density is possibly cov- a Processing methods
ered by the existence of the filler (metal salts) which have a 1.25 A
much higher density than neat PP. As seen in Fig. 5b, even
Group 250/45/0.1 which was the densest printed sample, was 1.2
Density (kg/cm3)

B
less dense than the IM PP. 1.15
Density profiles of the IM and FDM PPs are displayed
1.1
in Fig. 6. Densities of all samples were larger than 1080 kg/
m3. The normal density of PP is around 900 kg/m3. As 1.05
mentioned previously, the presence of the filler (metal
1
salts) increased the density of the samples. These density b IM 250/45/0.1
profiles only represent the portion of parts that encountered Processing methods

the X-ray during the density measurements, so they do not


reflect the overall density profile of each sample. When Fig. 5  Densities of PP from the different FDM processing parameters
determining the density and ash content of the PP filament, (a) and a comparison between the IM PP and a FDM PP (b). “A” and
“B” on top of the bars represent the statistical difference between val-
variations were discovered among replicates. The variations ues. Values with different letters are significantly different at a confi-
in density and ash content values of the filaments indicate dence level of 95%

13
Progress in Additive Manufacturing (2018) 3:205–214 211

IM 250/90/0.3 250/90/0.1 250/45/0.3 250/45/0.1 at 250 °C, regardless of printing speed or layer height,
1400 their density profiles all displayed a similar trend: density
decreased from the bottom to the top of a sample. On the
Density (kg/m3)

1350

1300 contrary, samples printed at 200 °C did not show this trend.
1250 This finding is consistent with the SEM results. At 250 °C,
1200
neck size on the bottom side of the samples was smaller than
those printed at 200 °C. Therefore, the bottom has a higher
1150
density than the top.
1100
0 0.5 1 1.5 2 2.5 3
a
Thickness (mm) 3.4 Mechanical properties
IM 200/90/0.3 200/90/0.1 200/45/0.3 200/45/0.1
The results of the ANOVA test on the mechanical proper-
1350
ties of PP are shown in Table 4. For the tensile properties,
the extrusion temperature × printing speed interaction had
Density (kg/m3)

1300

1250
a significant influence on the mechanical properties. For the
flexural strength, the extrusion temperature × printing speed
1200 interaction and the printing speed × layer height interaction
1150 showed significant effects. Regarding the flexural modulus
of elasticity, only printing speed was a significant processing
1100
0 0.5 1 1.5 2 2.5 3
factor. No other individual factors or interactions showed
b statistical significance.
Thickness (mm)
The mechanical properties of PP samples are shown in
Table 5. The flexural strength of PP, either IM or FDM,
Fig. 6  Density profiles of FDM PP samples made using different pro-
was 1.5–2 times larger than the tensile strengths, which was
cess parameters and compared to IM PP samples
also reported in a previous study on the mechanical behavior
of polymer and polymer composites [20]. According to the
their error bars in Fig. 5a. For the injection-molded parts, the “weakest link” theory, the tensile strength is the strength
filament was first pelletized. Pellets from different sections where the weakest element reaches its limit and breaks [21].
of the filament were hand blended before molding, which The flexural strength is the stress on the surface of a sample
contributed to a more homogenized part density as indicated when it fails. Therefore, the flexural property of a surface
by the error bar in Fig. 5b. Therefore, it was not surprising to affects the whole sample greatly [21]. This phenomenon is
see the density values of some of the 3D-printed parts being especially prominent in FDM manufacturing. As observed
higher than the injection-molded parts as shown in Fig. 6. in the SEM micrographs, the FDM PP sample has dense
Compared to IM PP, which showed consistent density across bottom layers attributed to the annealing which increases the
the thickness of the samples, the density values of FDM properties of the polymer [11]. The flexural modulus of the
PP fluctuated throughout the thickness. For samples printed IM PP was close to the values reported in the material data

Table 4  Selected values from Sources DF Sum of squares p value


the ANOVA for the mechanical
properties of PP TS TM FS FM TS TM FS FM

Corrected total 39 38.98 1.68 75.92 0.80


Model 7 32.57 1.31 65.43 0.31
ET 1 28.06 1.05 44.94 0.048 < 0.0001* < 0.0001* < 0.0001* 0.088
PS 1 0.65 0.0006 1.02 0.094 0.081 0.082 0.0867 0.019*
LH 1 0.02 0.045 14.64 0.035 0.75 0.057 < 0.0001* 0.145
ET × PS 1 0.09 0.034 1.94 0.064 0.51 0.097 0.021* 0.072
PS × LH 1 0.18 0.011 2.21 0.04 0.35 0.34 0.014* 0.118
ET × LH 1 3.19 0.156 0.63 0.022 0.0004* 0.0008* 0.18 0.243
ET × PS × LH 1 0.38 0.014 0.049 0.017 0.18 0.28 0.7 0.307
Error 32 6.41 0.37 10.49 0.49

TS tensile strength, TM tensile modulus, FS flexural strength, FM flexural modulus


*Significant level at α = 0.05

13
212 Progress in Additive Manufacturing (2018) 3:205–214

Table 5  Mechanical properties Processing method Tensile properties Flexural properties


of PP samples
Strength (MPa) Modulus (GPa) Strength (MPa) Modulus (GPa)

IM 17.5 (0.4) – 2.7 (0.04) – 28.4 (0.7) – 1.6 (0.1) –


250/90/0.3 15.5 (0.2) Ba 2.5 (0.07) AB 25.6 (0.5) B 1.6 (0.1) C
250/90/0.1 15.9 (0.7) AB 2.6 (0.1) A 27.1 (0.2) A 1.7 (0.1) ABC
250/45/0.3 15.6 (0.2) B 2.64 (0.05) A 26.9 (0.9) A 1.9 (0.1) A
250/45/0.1 16.4 (0.7) A 2.7 (0.2) A 27.5 (0.8) A 1.8 (0.1) AB
200/90/0.3 14.1 (0.6) D 2.4 (0.06) BC 23.7 (0.3) C 1.6 (0.1) C
200/90/0.1 13.8 (0.1) D 2.2 (0.1) D 25.6 (0.4) B 1.7 (0.09) ABC
200/45/0.3 14.8 (0.3) C 2.3 (0.1) C 24.0 (0.2) C 1.6 (0.1) BC
200/45/0.1 13.8 (0.3) D 2.2 (0.1) D 25.0 (0.6) B 1.7 (0.1) ABC

Values with different letters are significantly different at a confidence level of 95%
a
Capital letters behind each value represent the statistical difference

sheet. The tensile modulus of both PP samples was signifi- almost all the FDM groups exhibited higher flexural modu-
cantly higher than the flexural modulus samples. One thing lus than IM PP. Available literature on flexural properties of
to note is the testing speed for tensile properties was set at FDM polymers is sparse. One study indicated that the flex-
50 mm/min instead of 5 mm/min. This resulted in a higher ural modulus of FDM ABS was close to typical values of IM
tensile modulus value because the polymer is more elastic ABS [23]. In another article, the bottom layer of FDM PLA
than viscous at higher testing speeds. was found to have a higher crystallinity compared to the
The mechanical property changes of FDM PP compared rest of the part [5]. The properly annealed bottom layer may
to injection-molded PP are shown in Fig. 7. Among all the account for the high flexural modulus of FDM PP. Among
three factors studied, extrusion temperature had the most all groups, Group 250/45/0.1 is comparable in mechanical
obvious effect on the mechanical property reduction com- properties to IM PP except that its tensile strength is lower
pared to injection molding. Higher extrusion temperatures by 6% (Fig. 8).
lead to a smaller mechanical property loss. As was previ- The percentage change of the specific mechanical proper-
ously reported, the relative proportions of β-crystals in the ties of IM and FDM PP is shown in Fig. 9. Clearly, the spe-
injection-molded PP, PP printed at 200 °C and PP printed at cific mechanical property loss is smaller than the mechani-
250 °C were around 5, 10 and 75% [11]. The β-crystals and cal property loss. The influence of extrusion temperature
α-crystals of PP have different modes of chain packing [18]. on mechanical properties is discernable. Higher extrusion
More complicated chain orientation was discovered in the temperature leads to less mechanical property loss. How-
β-crystals [18]. Compared to α-PP, β-PP has a lower modu- ever, the influence of extrusion temperature was less when
lus of elasticity and yield strength [18, 22]. The existence the samples were normalized with respect to density. Group
of larger voids and more β-crystals in PP printed at 200 °C 250/90/0.1 showed slightly better overall specific mechanical
make it weaker than PP printed at 250 °C. Interestingly, properties compared to the other groups printed at 250 °C.

Fig. 7  Percentage change in tensile strength tensile modulus


mechanical properties of FDM 20 flexural strength flexural modulus
PP compared to IM PP 15
Change percentage (%)

10

-5

-10

-15

-20

-25
250/90/0.3 250/90/0.1 250/45/0.3 250/45/0.1 200/90/0.3 200/90/0.1 200/45/0.3 200/45/0.3

13
Progress in Additive Manufacturing (2018) 3:205–214 213

tensile modulus flexrual modulus Specific tensile modulus specific flexural modulus

tensile strength flexural strength specific tensile strength specific flexural strength

Specific modulus (106*N·m/kg)

Specific strength (103*N·m/kg)


3 A A 35 3 30
A A 30
A A A
2.5 2.5 25
Modulus (GPa)

Strength (MPa)
A 25
2 A 2 20
A B 20 A A A B
1.5 1.5 15
15
1 1 10
10
0.5 5 0.5 5
0 0 0 0
IM 250/45/0.1 IM 250/90/0.1

Fig. 8  Comparisons of the mechanical properties between IM and Fig. 10  Comparisons of specific mechanical properties between IM
FDM PP (Group 250/45/0.1). “A” and “B” on top of the bars repre- and FDM PP (Group 250/90/0.1). “A” and “B” on top of bars repre-
sent the statistically significant difference. Values with different let- sent the statistical difference. Values with different letters are signifi-
ters are significantly different at a confidence level of 95% cantly different at a confidence level of 95%

Figure 10 shows that Group 250/90/0.1 is stronger than IM smaller layer height resulted in denser internal structures
PP, with a 16% enhancement in specific flexural modulus. with an increased interfacial width. FDM PP parts were
lighter than injection-molded PP. PP printed at 250 °C was
denser than PP printed at 200 °C. Compared to the injection-
4 Conclusions molded PP, FDM PP parts had weaker tensile properties and
flexural strength mostly because the interfaces in printed
In this study, three 3D-printing processing parameters, PP acted as stress-concentrating areas, initiating breakage
extrusion temperature, printing speed and layer height, during mechanical testing. PP printed at 250 °C showed a
were explored to evaluate their influence on the mechanical smaller decrease in mechanical properties compared to IM
properties of FDM PP. The results obtained in this study PP than PP printed at 200 °C. The mechanical properties
clearly indicate that FDM, with proper processing control, of Group 250/45/0.1 were not significantly different from
has the capacity to make parts that are equally strong to those of IM PP, except that its tensile strength was weaker by
their injection-molded counterparts. The specific mechanical 6%. The specific mechanical properties of Group 250/90/0.1
properties of the FDM parts are mostly higher than those of parts revealed no difference from IM PP, except that its spe-
their injection-molded counterparts. FDM has the potential cific flexural modulus was significantly higher by 16%.
to be used for direct manufacturing, instead of prototyping,
especially for lightweight applications.
DSC revealed that FDM PP samples had α and β-crystal
forms, while IM PP only had α-crystals. SEM results showed
that higher extrusion temperature, slower printing speed and

Fig. 9  Change percentages of specific tensile strength specific tensile modulus


the specific mechanical proper- specific flexural strength specific flexural modulus
30.0
ties of FDM PP compared to
IM PP 25.0

20.0
Change percentage (%)

15.0

10.0

5.0

0.0

-5.0

-10.0

-15.0

13
214 Progress in Additive Manufacturing (2018) 3:205–214

Acknowledgements This work was supported in part by the Maine 11. Wang L, Gardner DJ (2017) Effect of fused layer modeling (FLM)
Agricultural and Forest Experiment Station (MAFES) project ME0-M- processing parameters on impact strength of cellular polypropyl-
8-00527-13 and the USDA ARS Forest Products Research Agreement ene. Polymer 113:74–80
58-0202-4-003. The authors thank Joshua Kovach for preparing the 12. Milosevic M, Stoof D, Pickering KL (2017) Characterizing the
CAD drawings of the testing samples. mechanical properties of fused deposition modelling natural fiber
recycled polypropylene composites. J Compos Sci 1:7
13. Jeong GY (2005) Fracture behaviour of wood plastic composite
Compliance with ethical standards (WPC). Dissertation, Louisiana State University
14. Tordjeman P, Robert C, Marin G, Gerard P (2001) The effect of
Conflict of interest On behalf of all the authors, the corresponding au- α, β crystalline structure on the mechanical properties of polypro-
thor states that there is no conflict of interest. pylene. Eur Phys J E Soft Matter 4:459–465
15. Bourbigot S, Garnier L, Revel B, Duquesne S (2013) Characteri-
zation of the morphology of iPP/sPP blends with various compo-
sitions. Express Polym Lett 7:224–237
References 16. Xiao W, Wu P, Feng J, Yao R (2009) Influence of a novel
β-nucleating agent on the structure, morphology, and nonisother-
1. Wendel B, Rietzel D, Kühnlein F, Feulner R, Hülder G, mal crystallization behavior of isotactic polypropylene. J Appl
Schmachtenberg E (2008) Additive processing of polymers. Mac- Polym Sci 111:1076–1085
romol Mater Eng 293:799–809 17. Li JX, Cheung WL, Jia D (1999) A study on the heat of fusion of
2. Vaezi M, Chianrabutra S, Mellor B, Yang S (2013) Multiple mate- β-polypropylene. Polymer 40:1219–1222
rial additive manufacturing –Part 1: a review. Virtual Phys Proto- 18. Papageorgiou DG, Chrissafis K, Bikiaris DN (2015) β-Nucleated
typ 8:19–50 polypropylene: processing, properties and nanocomposites. Polym
3. Korpela J, Kokkari A, Korhonen H, Malin M, Närhi T, Seppälä Rev 55(4):596–629
J (2013) Biodegradable and bioactive porous scaffold structures 19. Sun Q, Rizvi GM, Bellehumeur CT, Gu P (2008) Effect of pro-
prepared using fused deposition modeling. J Biomed Mater Res cessing conditions on the bonding quality of FDM polymer fila-
Part B Appl Biomater 101:610–619 ments. Rapid Prototyp J 14:72–80
4. Kim YH, Wool RP (1983) A theory of healing at a polymer-pol- 20. Landel RF, Nielsen LE (1993) Mechanical properties of polymers
ymer interface. Macromolecules 16:1115–1120 and composites. Marcel Dekker, Inc, NY
5. Drummer D, Cifuentes-Cuellar S, Rietzel D (2012) Suitability 21. Hodgkinson JM (2000) Mechanical testing of advanced fiber com-
of PLA/TCP for fused deposition modeling. Rapid Prototyp J posites. CRC press, Boca Raton, FL
18:500–507 22. Varga J (2002) β-modification of isotactic polypropylene: prepara-
6. Rosato DV, Rosato MG, Schott NR (2001) Plastics technology tion, structure, processing, properties, and application. J Macro-
handbook, volume 1: introduction-properties-fabrication-pro- mol Sci 41(4–6):1121–1171
cesses. Momentum Press, NY 23. Ning F, Cong W, Qiu J, Wei J, Wang S (2015) Additive manufac-
7. Harper CA (2000) Modern plastics handbook. McGraw-Hill, NY turing of carbon fiber reinforced thermoplastic composites using
8. Wang L, Gramlich WM, Gardner DJ, Han Y, Tajvidi M (2018) fused deposition modeling. Compos Part B-Eng 80:369–378
Spray-dried cellulose nanofibril-reinforced polypropylene com-
posites for extrusion-based additive manufacturing: nonisother- Publisher’s Note Springer Nature remains neutral with regard to
mal crystallization kinetics and thermal expansion. J Compos Sci jurisdictional claims in published maps and institutional affiliations.
2(1):7. https​://doi.org/10.3390/jcs20​10007​
9. Carneiro OS, Silva AF, Gomes R (2015) Fused deposition mod-
eling with polypropylene. Mater Design 83:768–776
10. Stoof D, Pickering K (2018) Sustainable composite fused deposi-
tion modelling filament using recycled pre-consumer polypropyl-
ene. Compos Part B 135:110–118

13

You might also like