Journal of

Manufacturing and
Materials Processing

Article
Hybrid Manufacturing and Experimental Testing of
Glass Fiber Enhanced Thermoplastic Composites
Javaid Butt * , Yasasween Hewavidana, Vahaj Mohaghegh, Shabnam Sadeghi-Esfahlani and
Hassan Shirvani
School of Engineering & Built Environment, Anglia Ruskin University, Chelmsford CM1 1SQ, Essex, UK;
yasasween.hewavidana@student.anglia.ac.uk (Y.H.); vahaj.mohaghegh@anglia.ac.uk (V.M.);
shabnam.sadeghi-esfahlani@anglia.ac.uk (S.S.-E.); hassan.shirvani@anglia.ac.uk (H.S.)
* Correspondence: javaid.butt@anglia.ac.uk




Received: 29 October 2019; Accepted: 27 November 2019; Published: 2 December 2019


Abstract: Additive Manufacturing (AM) is gaining enormous attention from academic and industrial
sectors for product development using different materials. Fused Deposition Modelling (FDM) is a
popular AM method that works with thermoplastics. This process offers benefits of customisation
both in terms of hardware and software in the case of desktop-based FDM systems. Enhancement
of mechanical properties for the traditional thermoplastic material is a widely researched area and
various materials have been added to achieve this goal. This paper focuses on the manufacture of
glass fiber reinforced plastic (GFRP) composites using Hybrid Fused Deposition Modelling (HFDM).
Commonly available polylactic acid or polylactide (PLA) material was inter-laced with 0.03 mm
thick glass fiber sheets to manufacture GFRP products followed by tensile testing. This was done
to investigate whether adding more layers increases the tensile strength of the GFRP products or
not. Furthermore, the maximum number of glass fiber layers that can be added to the 4 mm thick
specimen was also identified. This was done to demonstrate that there is an optimal number of glass
fiber layers that can be added as after this optimal number, the tensile strength start to deteriorate.
Microstructural analysis was undertaken after tensile testing followed by ultrasonic testing to assess
the uniformity of the GFRP composites.

Keywords: composite; fused deposition modelling; glass fiber; hybrid process; microstructural
analysis; PLA plastic; tensile test; ultrasonic testing

1. Introduction
Composites have taken a large portion of the manufacturing market due to their properties, most
notably, their high strength to weight ratio. This is further augmented by the possibility of tailoring the
mechanical properties by modifying the process parameters. However, these modifications are mostly
based on trial and error methods and ensuring consistent properties is crucial for wider applications [1–4].
For composites, two or more materials are combined together with the aim of enhancing the overall
properties of the final product. Their mechanical properties can be engineered, preferably by adjusting
the volume or mass fraction of the constituents or completely replacing one constituent by another
suitable one. The study of the bonding interface of these constituents and subsequent fracture
mechanisms are a widely researched area and allow for improving mechanical properties [5–9]. Plastics
are one of the most popular materials that are commonly used for composite development as they are
cost-effective and readily available. In this context, thermoset and thermoplastics are two types that
are commonly used for the manufacture of composites. The molding process is a popular choice for
the manufacture of thermoset composites (materials that undergo a chemical reaction and transform
from a liquid into a solid). This work, however, focuses on additive manufacturing (AM) methods and

J. Manuf. Mater. Process. 2019, 3, 96; doi:10.3390/jmmp3040096 www.mdpi.com/journal/jmmp

J. Manuf. Mater. Process. 2019, 3, 96 2 of 16

more specifically, Fused Deposition Modelling (FDM) that is one of the most common methods for the
manufacture of thermoplastics. These materials are thermo-formable (softening of the thermoplastic
material when heated) and recyclable, that makes them more valuable for industries. FDM is being
used extensively for direct digital manufacturing applications due to the easy access and low-cost of
personal 3D printers that can work with a wide variety of materials. Two commonly used materials
for FDM are acrylonitrile butadiene styrene (ABS) and polylactic acid or polylactide (PLA). Both
have their pros and cons that should be considered before usage. PLA is a bio-degradable material
derived from renewable resources such as corn starch or sugarcane. It is one of the most popular
bioplastics, used for many applications ranging from plastic cups to medical implants. On the other
hand, ABS is also quite popular in the injection molding arena and its applications range from LEGO
toys and electronic housings to bumper parts for automobile use. For these thermoplastics, a high
strength to weight ratio is a major aspect as it allows applications in various industrial sectors, e.g.,
aerospace, automotive and medical. Therefore, efforts have been made to enhance the strength of
these thermoplastic materials while keeping them lightweight. FDM, being an AM method, offers
benefits of design freedom, reduced costs and lead times etc., which further improves its applicability
for engineering applications.
Existing literature shows limited innovation in terms of enhancing the capabilities of the FDM
process but rather focuses on the development of fiber reinforced materials that can be easily printed.
Ning et al. [10] developed carbon fiber filled ABS filaments from plastic pellets and carbon fiber
powder with varying fiber content for FDM process. Carbon fiber reinforced plastic (CFRP) composites
were manufactured to experimentally investigate the effect of adding carbon fiber on the mechanical
properties of the FDM-printed parts. The experimental result showed enhancement of tensile and
flexural properties of the CFRP composites compared to pure ABS. Furthermore, they also showed
that the tensile properties initially increased but then deteriorated drastically highlighting an optimal
carbon fiber content in the ABS filaments. Liu et al. [11] investigated the tensile and flexural properties
of PLA composites with different additives i.e., wood, ceramic, copper, aluminium and glass fiber
manufactured by FDM. They noticed that PLA composites have similar or increased mechanical
properties compared to parent PLA. Coppola et al. [12] studied the effect of adding hemp powder
(diameter between 75 and 180 µm) via melt-compounding into PLA 4043D at three different volume
percentages (1, 3 and 5%). PLA/hemp composites exhibited higher elastic modulus compared to
PLA and showed increase with increase in the hemp powder content. Tao et al. [13] investigated
the development of wood flour (WF)-filled polylactic acid (PLA) composite filaments for FDM. They
concluded that adding WF of 5 wt% has no effect on the melting temperature of the PLA and it also
increased the tensile stress of the WF-PLA composites compared to virgin PLA. Pantanwala et al. [14]
discussed the 3D printing of carbon nanotube-polylactic acid (CNT-PLA) composites with FDM. CNTs
were added at 0.5%, 2.5%, and 5% (w/w) CNT loadings and they showed that the Young’s modulus
increased the by 30% at 5% CNT loading. However, this content of CNT also resulted in reduced
tensile strength and overall toughness of the resulting FDM products. Gkartzou et al. [15] used FDM
for the manufacture and experimental testing of PLA/lignin composites based on different lignin
weight ratios. They observed changes in the mechanical properties with the composites becoming
more and more brittle with increasing lignin content with the 20% wt. lignin product too brittle for
tensile testing. FDM is also quite popular for printing on textiles to manufacture functional composites.
Sanatgar et al. [16] investigated the effect of FDM process parameters (extruder temperature, platform
temperature and printing speed) on the adhesion properties of polymers and nanocomposites for
textiles. They found that PLA and its composites had high adhesion force to PLA fabrics due to their
compatibility. The work done by Eutionnat-Diffo et al. [17] showed the adhesion characteristics of PLA
filaments deposited onto polyethylene terephthalate (PET) textiles through 3D printing. The PLA/PET
composites demonstrated lower deformation compared to virgin PET showcasing better dimensional
stability, higher stiffness and lower flexibility.
J. Manuf. Mater. Process. 2019, 3, 96 3 of 16

This clearly shows the capability of the FDM process in terms of manufacturing plastic-based
composites. However, the literature involves pre-processing to manufacture the filament material
before printing. This aspect can be changed with the process of Hybrid Fused Deposition Modelling
(HFDM) that allows the addition of additive layers that gets adhered to the extruded material.
The process has been explained in detail in our previous work [18] where ABS/copper mesh composites
were manufactured and experimentally tested. We also showed that increasing the number of Cu
layers linearly
J. Manuf.increase the
Mater. Process. tensile
2019, 3, 96 strength of the Cu/ABS products compared to parent 3ABS.of 16 The next

section will provide a breakdown of activities undertaken to manufacture the metal/plastic composites.
(HFDM) that allows the addition of additive layers that gets adhered to the extruded material. The
process has been explained in detail in our previous work [18] where ABS/copper mesh composites
2. Materials
wereand Manufacturing
manufactured Processestested. We also showed that increasing the number of Cu
and experimentally
The layers
processlinearly increase the tensile strength of the Cu/ABS products compared to parent ABS. The
of HFDM was broken down into steps for practicality. A low-cost DIY desktop 3D
next section will provide a breakdown of activities undertaken to manufacture the metal/plastic
printer called Anet A8 3D printer (from Shenzen Anet Technology Company Limited, Hong Kong)
composites.
was used for the manufacture of products based on British and International standards. The printer
had a work2. Materials
envelope andofManufacturing
220 mm × 220 Processes
mm × 250 mm (X × Y × Z) and an extruder nozzle diameter
of 0.4 mm (Figure 1). of
The process PLAHFDM (RSwas PRO filament
broken down intoof 1.75
steps mm diameterA from
for practicality. low-costRSDIY
Components,
desktop 3D Corby,
UK) and printer calledfiber
thin glass Anetsheets
A8 3D printer
(from (from
EasyShenzen Anet Technology
Components, Company UK)
Stoke-on-Trent, Limited,
wereHongusedKong)
for product
was used for the manufacture of products based on British and International standards. The printer
development. The glass fiber sheets had a woven fiber mesh with an aperture of 0.22 mm and fiber
had a work envelope of 220 mm × 220 mm × 250 mm (X × Y × Z) and an extruder nozzle diameter of
diameter 0.4
of mm
0.03(Figure
mm. The 3D (RS
1). PLA CAD PROmodel
filamentof the part
of 1.75 mm todiameter
be builtfromwas RSsent to an openCorby,
Components, source UK)3D printer
slicing software
and thin glass fiber sheets (from Easy Components, Stoke-on-Trent, UK) were used for product part was
package called Ultimaker Cura 4.3.0 that began the build operation. The
built at adevelopment.
speed of 60The mm/sglasswith a layer
fiber sheets hadthickness
a woven fiberof 0.2
mesh mm.
with PLA was printed
an aperture of 0.22 mm atand ◦ C with the
200fiber
diameter of 0.03 mm. ◦The 3D CAD model of the part to be built was sent to
bed temperature set at 60 C as shown in Figure 1. The glass fiber layers were added to the parts byan open source 3D printer
slicing software package called Ultimaker Cura 4.3.0 that began the build operation. The part was
utilising the ‘Pause at height’ option in the CURA 4.3.0 software and then resuming the operation to
built at a speed of 60 mm/s with a layer thickness of 0.2 mm. PLA was printed at 200 °C with the bed
manufacture glass fiber
temperature set atreinforced plastic
60 °C as shown in (GFRP)
Figure 1. products as shown
The glass fiber layers in Figures
were added 2toand 3. Forbyimproved
the parts
adhesionutilising
between the plastic
‘Pause atand theoption
height’ thin inglass fiber 4.3.0
the CURA layers, the number
software of shellsthe
and then resuming were increased
operation to from
manufacture glass fiber reinforced plastic (GFRP) products as shown in
two to three, infill density was set at 100% and the infill pattern is selected as ‘Cubic’ in the Cura Figures 2 and 3. For
improved adhesion between plastic and the thin glass fiber layers, the number of shells were
4.3.0 software. There are other options that can be considered e.g., choosing a different infill pattern
increased from two to three, infill density was set at 100% and the infill pattern is selected as ‘Cubic’
(such as quarter cubic, triangles, concentric, zig-zag), increasing the thickness of the top as well as
in the Cura 4.3.0 software. There are other options that can be considered e.g., choosing a different
bottom layers, layer(such
infill pattern thickness,
as quarterraster angle
cubic, andconcentric,
triangles, build orientation [19,20]. the
zig-zag), increasing Optimisation of top
thickness of the 3D printing
parameters is aascomplex
as well process;
bottom layers, layer therefore, the angle
thickness, raster buildandprocess was kept[19,20].
build orientation simple to ensureofconsistent
Optimisation
results in3D printing manner.
a timely parametersFor is a statistical
complex process; therefore,
accuracy, the build
a total process
of three was kept simple
specimens were to ensurefor all the
tested
consistent results in a timely manner. For statistical accuracy, a total of three specimens were tested
different materials and configurations.
for all the different materials and configurations.

Figure 1. Anet A8 3D printer in operation.

Figure 1. Anet A8 3D printer in operation.
J. Manuf. Mater. Process. 2019, 3, 96 4 of 16
J. Manuf. Mater. Process. 2019, 3, 96 4 of 16

Figure 2.
Figure 2. Manufacture
Manufacture of GFRP composites.
of GFRP composites.

Figure 3.
Figure 3. Glass
Glass fiber
fiber layer
layer atop PLA specimen.
atop PLA specimen.

3.
3. Experimental Methodology

3.1. Tensile Testing

3.1. Tensile Testing
The products for
The products fortensile
tensiletesting
testingwere
weremanufactured
manufactured based
based on on BS EN
BS EN ISO ISO 527-2:2012
527-2:2012 [21]. [21].
The
The dimensions of the specimens are given in Figure 4. Previous work has shown
dimensions of the specimens are given in Figure 4. Previous work has shown the capability of HFDM the capability of
HFDM in producing
in producing products products with higher
with higher strength,
strength, therefore,
therefore, thisistest
this test is being
being conducted
conducted to analyse
to analyse the
the extent of increase in tensile strength of the FDM-printed parts upon the addition
extent of increase in tensile strength of the FDM-printed parts upon the addition of glass fiber layers. of glass fiber
layers.
The GFRPThecomposites
GFRP composites are manufactured
are manufactured by increasing
by increasing number number
of glassoffiber
glasslayers
fiber layers (one,three
(one, two, two,
three
and soand
on).so on).has
This This has
been been
done todone
ensuretothat
ensure that an number
an optimal optimal ofnumber of glass
glass fiber layersfiber
canlayers can be
be identified
identified for the 4 mm thick dog-bone specimen as after this optimal number,
for the 4 mm thick dog-bone specimen as after this optimal number, the tensile strength start the tensile strength start
to
to deteriorate. The overall thickness of the specimens was kept at 4 mm ± 0.2 mm
deteriorate. The overall thickness of the specimens was kept at 4 mm ± 0.2 mm (the precision of the (the precision of the
printer).
printer). Pure
Pure PLA
PLAspecimens
specimenswereweremade
madeaccording
according totothethemethodology
methodology discussed
discussed in Section 2 and
in Section the
2 and
the GFRP were made by pausing the build operation. According to the Cura 4.3.0 software, 20 layers
(layer thickness of 0.2 mm) are required to manufacture the 4 mm thick dog-bone specimen. An
 J. Manuf. Mater. Process. 2019, 3, 96 5 of 16
J. Manuf. Mater. Process. 2019, 3, 96 5 of 16

additional four layers were added to the specimen as ‘raft’ type build plate adhesion was selected
(amounting
GFRP were made to 24
by total
pausing layers). This operation.
the build will keep the specimen
According adhered
to the to thesoftware,
Cura 4.3.0 build plate and will
20 layers (layer
minimize the effect of warping. The first four layers are for the raft and the
thickness of 0.2 mm) are required to manufacture the 4 mm thick dog-bone specimen. An additional remaining 20 build the 4
mm
four thickwere
layers specimen.
addedGFRP to thecomposites
specimen were manufactured
as ‘raft’ with increasing
type build plate adhesion was layers of copper
selected mesh (1, to
(amounting
2, 3 and so on until 15) while keeping the overall thickness to 4 mm ± 0.2 mm. For example, in case of
24 total layers). This will keep the specimen adhered to the build plate and will minimize the effect of
a single layered composite, one glass fiber layer was placed at 50% of the overall thickness of the
warping. The first four layers are for the raft and the remaining 20 build the 4 mm thick specimen.
specimen and for a double layered composite, two glass fiber layers were placed at 40% and 80%. For
GFRP composites were manufactured with increasing layers of copper mesh (1, 2, 3 and so on until
the addition of 15 layers, the same calculations were carried out (relevant layer number of the CAD
15) while keeping the overall thickness to 4 mm ± 0.2 mm. For example, in case of a single layered
model for the fiber layer insertion = (20/16) × 1 = 1.25 so it is considered as layer 1. The layer numbers
composite, one glass fiber layer was placed at 50% of the overall thickness of the specimen and for
are calculated in this manner. Values such as 1.25 and 2.5 were considered as 1 and 2, respectively
a double layered
whereas the values composite,
such astwo 3.75glass
and fiber layersrounded
8.75 were were placed
off toat4 40%
and and 80%. For the
9, respectively. addition
There layer of
15 numbers
layers, the same calculations were carried out (relevant layer number of the CAD
are then used to calculate percentage of product completion. The percentage completion for model for the fiber
layer
the insertion
addition of= a(20/16)
single × 1 = of
layer 1.25 so it
glass is considered
fiber as layer
= (1/20) × 100% = 5%).1. The layer numbers are calculated in
this manner. Values such as 1.25 and 2.5 were considered
The tensile testing was carried out using a Universal Tensile as 1 and 2,Testing
respectively
Machinewhereas the values
at a crosshead
such as 3.75 and 8.75 were rounded off to 4 and 9, respectively. There layer numbers
speed of 1.5 mm/s according to the standard. The glass fiber layers were added without any treatment are then used to
calculate
and thepercentage
programming of product
was notcompletion. The percentage
modified to accommodate thecompletion
0.03 mm thin for layers.
the addition of a single
This effect has
layer
beenofanalysed by=measurements
glass fiber (1/20) × 100%of=thickness
5%). and mass for the GFRP composites in Section 4.1.

Figure 4. Dimensions of the tensile testing dog-bone specimen.

Figure 4. Dimensions of the tensile testing dog-bone specimen.

The tensile testing was carried out using a Universal Tensile Testing Machine at a crosshead speed
3.2. Microstructural Analysis
of 1.5 mm/s according to the standard. The glass fiber layers were added without any treatment and
A JCM-5000 was
the programming NeoScope™ table top
not modified scanning electron
to accommodate themicroscope (JEOL,
0.03 mm thin Boston,
layers. USA)
This was
effect used
has been
to analyse the fractured surfaces of the tensile tested PLA and GFRP products.
analysed by measurements of thickness and mass for the GFRP composites in Section 4.1. This analysis was
conducted to investigate the adhesion quality and porosity of the products. The joining of glass fibers
3.2.using
Microstructural
the adhesionAnalysis
properties of the extruded PLA is an interesting prospect and this test allowed to
investigate how the PLA layers bonded with the glass fibers as well as other PLA layers that had been
A JCM-5000 NeoScope™ table top scanning electron microscope (JEOL, Boston, USA) was used
printed before. The fractured surfaces were carefully cut to size to fit atop the platform of the scanning
to analyse the fractured surfaces of the tensile tested PLA and GFRP products. This analysis was
electron microscope (SEM). They were analysed without any surface treatment to avoid
conducted to investigate the adhesion quality and porosity of the products. The joining of glass fibers
contamination. This test helped in understanding the bond interface between the two materials (PLA
using
and glass fiber) asproperties
the adhesion ofglass
the layers of the extruded PLA is
fibers increase inan
theinteresting
4 mm thickprospect and this test allowed to
specimens.
investigate how the PLA layers bonded with the glass fibers as well as other PLA layers that had been
printed before. The
3.3. Ultrasonic fractured surfaces were carefully cut to size to fit atop the platform of the scanning
Testing
electron microscope (SEM). They were analysed without any surface treatment to avoid contamination.
Non-destructive testing (NDT) techniques are extremely popular in engineering inspection and
This test helped in understanding the bond interface between the two materials (PLA and glass fiber)
evaluation as they assess the integrity of a product without destroying it so that it can be used
as afterwards.
the layers ofUltrasonic
glass fibers increase in the 4 mm thick specimens.
testing is a common NDT method and works on the transmission of ultrasonic
waves through a test piece. For this work, a Proceq PUNDIT® PL-200 device (Proceq, Schwerzenbach,
3.3. Ultrasonic Testing
Switzerland) was used. It comprises two transducers (54 kHz) with a measuring resolution of 0.1 µs.
Non-destructive testing (NDT) techniques are extremely popular in engineering inspection and
evaluation as they assess the integrity of a product without destroying it so that it can be used
J. Manuf. Mater. Process. 2019, 3, 96 6 of 16

afterwards. Ultrasonic testing is a common NDT method and works on the transmission of ultrasonic
waves through a test piece. For this work, a Proceq PUNDIT® PL-200 device (Proceq, Schwerzenbach,
Switzerland)
J. Manuf.
was
Manuf. Mater.
used.
Mater. Process.
It 3,comprises
Process. 2019,
2019, 3, 96
two transducers (54 kHz) with a measuring resolution6 of 0.1 µs.
of 16
16
J. 96 6 of
This is a standard instrument for ultrasonic pulse velocity testing as shown in Figure 5. Couplant
gel was
Thisused
This is on the traducers
is aa standard
standard instrumentasfor
instrument it ultrasonic
for facilitatespulse
ultrasonic the transmission
pulse velocity testing
velocity ofas
testing ultrasonic
as shown in
shown in energy 5.into
Figure 5.
Figure the test
Couplant
Couplant gelpiece.
gel
Therewas
was used
areused
twoon on the
methods traducers as
of receiving
the traducers it facilitates the
the ultrasound
as it facilitates transmission
waveform
the transmission of ultrasonic energy
i.e., reflection
of ultrasonic energyandinto the test
intoattenuation. piece.
the test piece. In the
There
There
former, areone
are
only twotransducer
two methods of
methods ofperforms
receiving the
receiving the ultrasound
ultrasound
sending andwaveform
waveform
receiving i.e.,
i.e., reflection
reflection
whereas and latter,
in and
the attenuation.
attenuation. In the
In the
two transducers
former,
former,
are used. only one
only
Attenuation one transducer
transducer
method was performs
performs the
used inthe sending
thissending
work withand one
and receiving
receiving whereas
whereas
transducer in the
in the latter,
(transmitter) latter, two the
two
sending
transducers are
transducers are used.
used. Attenuation
Attenuation methodmethod waswas used
used in
in this
this work
work with
with one
one transducer
transducer (transmitter)
(transmitter)
signal through the test piece and the other receiving (receiver) the signal. The transmission time (µs)
sending the the signal
signal through the the test
test piece
piece and
and the
the other
other receiving
receiving (receiver) the the signal.
signal. TheThe
was sending
measured for all thethrough
PLA and GFRP products at three places (P1, P2,(receiver)
P3 showing the positions) to
transmission time
transmission time (µs)
(µs) was
was measured
measured for for all
all the
the PLA
PLA and
and GFRP
GFRP products
products at at three
three places
places (P1,
(P1, P2,
P2, P3
P3
ascertain an average
showing the
value to
the positions)
positions) to
pertaining
ascertain an
to the uniformity
an average
average value
of transmission
value pertaining
pertaining to to the
within the test
the uniformity
uniformity of
pieces as shown
of transmission
transmission
showing ascertain
in Figure
within6.the test pieces as shown in Figure 6.
within the test pieces as shown in Figure 6.

Figure
Figure 5.Calibration
5.5.
Figure Calibration of
Calibration of the
the transducers
the transducers before
transducersbefore
beforetesting.
testing.
testing.

Figure 6. Placement of transducers for measurements.

Figure
Figure 6.6.Placement
Placement of
of transducers
transducers for
formeasurements.
measurements.
J. Manuf. Mater. Process. 2019, 3, 96 7 of 16
J. Manuf. Mater. Process. 2019, 3, 96 7 of 16
J. Manuf.
4. Mater. Process.
Experimental 2019, 3,and
Results 96 Discussions 7 of 16

4. Experimental Results and Discussions

4.1. Effect of Glass Fiber Layers on Tensile Properties of the GFRP Composites
4. Experimental Results and Discussions
4.1. Effect of Glass Fiber Layers on Tensile Properties of the GFRP Composites
Force vs displacement curves for the pure PLA and GFRP composites are shown in Figure 7. The
4.1. Effect
curve wasofvs
Force Glass
selected Fiber
from
displacementLayers
the on
curves Tensile
resultsfor ofProperties
thethree of the
pure specimens,
PLA andGFRP Composites
depending
GFRP composites on the maximum
are shown value.7.ItThe
in Figure is
obvious
curveForcethat
was vsthe addition
selected fromofthe
displacement glass fiber
results
curves layers
forof three
the has
pure increased
specimens,
PLA and GFRPthe maximum
depending on load
composites value
the are
maximum that the
shown product
value.
in FigureIt is
7.
can
The curve was selected from the results of three specimens, depending on the maximum value. Itto
sustain.
obvious that The
the addition of one
glass glass
fiber fiber
layerslayer
has (GFRP-1)
increased has
the shown
maximum an increase
load of
value 6%
thatcompared
the product is
pure
obviousPLA.
can sustain. This
that thevalue
The roseof
addition
addition toglass
of 8% and
one glass
fiber10% forlayer
fiber
layers composites
has (GFRP-1)
increasedwiththetwo
has (GFRP-2)
shown
maximum andvalue
an increase
load three
of 6% (GFRP-3)
that comparedglass
the product to
fiber
pure layers
PLA. respectively.
This value roseThis
to 8%linear
and increase
10% for in maximum
composites load
with values
two is consistent
(GFRP-2)
can sustain. The addition of one glass fiber layer (GFRP-1) has shown an increase of 6% compared to and with
three our previous
(GFRP-3) glass
work
pure doneThis
fiber PLA.
layers with ABS rose
respectively.
value plastic
This [18].
to 8%linearThere
and 10%hasforbeen
increase in no difference
maximum
composites loadtwo
with in (GFRP-2)
the ductility
values andofthree
is consistent the
with products upon
our previous
(GFRP-3) glass
the
work addition
done of
with glass
ABS fiber layers
plastic [18]. as evident
There has from
been the
no fracture
difference modes
in theshown in
ductility
fiber layers respectively. This linear increase in maximum load values is consistent with our previous Figure
of the 8. This
products means
upon
that
work glass
donefiber
the addition ofdoes
with ABSnot
glass affect
fiber
plasticlayerstheThere
[18]. brittle
as nature
evident
has from
been of the
nothe pure PLA
fracture
difference plastic
modes
in the and this
shown
ductility is carried
inthe
of Figure over
8. This
products to the
means
upon the
GFRP
addition composites
that glassoffiber
glassdoesasnot
fiber well.
layers All
affectasthethe specimens
brittlefrom
evident nature fractured
theoffracture without
the puremodes
PLA plasticnecking
shownand demonstrating
this is 8.
in Figure carried a brittle
over tothat
This means the
failure.
GFRPfiber
glass composites
does not as well.
affect theAll thenature
brittle specimens
of the fractured
pure PLA without
plastic and necking demonstrating
this is carried over to thea brittle
GFRP
failure.
composites as well. All the specimens fractured without necking demonstrating a brittle failure.

Figure 7. Tensile test results for pure PLA and GFRP composites.
Figure 7. Tensile test results for pure PLA and GFRP composites.
Figure 7. Tensile test results for pure PLA and GFRP composites.

Figure 8. Fracture mode of pure PLA (left) and GFRP-1 composite (right).
Figure 8. Fracture mode of pure PLA (left) and GFRP-1 composite (right).
Figure 8. Fracture mode of pure PLA (left) and GFRP-1 composite (right).
J. Manuf. Mater. Process. 2019, 3, 96 8 of 16
J. Manuf. Mater. Process. 2019, 3, 96 8 of 16

It is
It is evident
evidentin inFigure
Figure77that thatthe
themaximum
maximum load
loadvalue is obtained
value for for
is obtained GFRP-13
GFRP-13 (at 2656 N) and
(at 2656 this
N) and
is an increase of 32.5% compared to pure PLA (at 2005 N). After the addition
this is an increase of 32.5% compared to pure PLA (at 2005 N). After the addition of 13 layers, theof 13 layers, the maximum
load value started
maximum load valueto gostarted
down. At to layer 14, theAt
go down. load value
layer 14,dropped
the loadtovalue
2620 N (30.7% increase
dropped to 2620 compared
N (30.7%
to pure PLA but a reduction of 1.8% compared to GFRP-13) with subsequent
increase compared to pure PLA but a reduction of 1.8% compared to GFRP-13) with subsequent decrease on the addition
of layer 15 to 2552 N (27.3% increase compared to pure PLA but a reduction
decrease on the addition of layer 15 to 2552 N (27.3% increase compared to pure PLA but a reduction of 5.2% compared to
GFRP-13).
of This behavior
5.2% compared is shown
to GFRP-13). in Figure
This behavior 9 where only in
is shown theFigure
specimens withonly
9 where the largest fracture with
the specimens load
values (out of three for each configuration) have been plotted as a percentage difference
the largest fracture load values (out of three for each configuration) have been plotted as a percentage compared to
pure PLA. The
difference figure clearly
compared to pureshows
PLA. Thean upward trend until
figure clearly showslayer
an 13 and then
upward a decrease
trend until layerfrom13 this
andpoint
then
onwards. This indicate that there is an optimal number of layers that can be
a decrease from this point onwards. This indicate that there is an optimal number of layers that can added to the dog-bone
specimen
be added to and
theitdog-bone
has been specimen
identified and as 13it layers.
has been This phenomenon
identified has also
as 13 layers. Thisbeen observed by
phenomenon hasother
also
researchers
been withbyPLA
observed based
other composites
researchers using
with PLAFDM based [10–13]. Anything
composites usingmore
FDM than 13 layers
[10–13]. (of 0.03more
Anything mm
thickness) will result in deterioration of tensile properties.
than 13 layers (of 0.03 mm thickness) will result in deterioration of tensile properties.

35

30
Percentage difference (%)

25

20

15

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number of glass fiber layers

Figure 9. Percentage
Percentage difference
difference of maximum load values between GFRP composites and pure PLA.

values +/−
A graph with the average load values standard deviations
+/− standard deviations asas error bars for the tensile tested
GFRP composites and pure PLA specimens
specimens is shown
shown inin Figure
Figure 10.
10. Such a graph highlights the
variability
variability and uncertainty
uncertainty in
in measured
measured values.
values. Since the programming
programming of the 3D printer has not been
adjusted to accommodate the addition of the glass fiber layers, the resulting GFRP composites should
have aathickness
thicknessvalue
valuegreater than
greater 4.004.00
than mmmm± 0.2± mm. ThisThis
0.2 mm. effecteffect
has been
has analysed in the next
been analysed section.
in the next
section.
J. Manuf. Mater. Process. 2019, 3, 96 9 of 16
J. Manuf. Mater. Process. 2019, 3, 96 9 of 16

2700

2600

2500
Average Load (N)

2400

2300

2200

2100

2000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Number of glass fiber layers

10. Graph
Figure 10. Graph showing
showing average
average load values +/−
load values standard deviations
+/− standard deviations as error bars for the tensile
tensile
tested
tested GFRP
GFRP composites
composites and
and pure
pure PLA
PLA specimens.
specimens.

Effect
Effect of
of Glass
Glass Fiber
Fiber Layers
Layers on on Thickness
Thickness of of the
the GFRP
GFRPComposites
Composites
FDM-printed
FDM-printed parts parts areare anisotropic
anisotropic in in nature
nature due
due to to the
the way
way in in which
which the the layers
layers are
are built
built [22].
[22].
They are also prone to warping [23]. This effect was also present during
They are also prone to warping [23]. This effect was also present during the manufacture of the GFRP the manufacture of the GFRP
composites.
composites. A A Vernier
Vernier caliper
caliper waswas used
used toto measure
measure the the thickness
thickness of of the
the dog-bone
dog-bone specimens
specimens at at three
three
sites
sites i.e., two from the sides and one from the middle. The average value of these measurements was
i.e., two from the sides and one from the middle. The average value of these measurements was
recorded. Figure 11 shows the comparison between the measured values
recorded. Figure 11 shows the comparison between the measured values and theoretical values based and theoretical values based
on
on the
the dimensions
dimensions of of the
the product.
product. This This comparison
comparison showsshows that that the
the thickness
thickness of of the
the GFRP
GFRP composites
composites
did
did not increase based on the theoretical values. At certain instances, there is very little increase
not increase based on the theoretical values. At certain instances, there is very little increase in in
thickness
thickness with
with thethe addition
addition of of another
another glass
glass fiber
fiber layer.
layer. This
This shows
shows that
that thethe semi-molten
semi-molten PLA PLA material
material
extruded
extruded fromfrom the the nozzle
nozzleof ofthethe3D 3Dprinter
printerisisnotnotalways
alwaysdepositing
depositing thethe0.20.2
mm mm thick
thick layer.
layer. As As
thethe0.2
0.2
mm thick PLA layer comes into contact with the 0.03 mm glass fiber layer on the print bed, it
mm thick PLA layer comes into contact with the 0.03 mm glass fiber layer on the print bed, it
encapsulates
encapsulates the the glass
glass fiber
fiber layer
layer andand thus
thus the
the total
total thickness
thicknessdoes doesnot notaddaddup up(0.2
(0.2mmmm++0.03 0.03mm mm=
= 0.23
0.23 mm).
mm). TheThe dimensional
dimensional inaccuracy
inaccuracy cancan be attributed
be attributed to several
to several factors.
factors. Firstly,
Firstly, the accuracy
the accuracy (±0.2
(±0.2 mm) of the Anet A8 3D printer and the prevalence of structural
mm) of the Anet A8 3D printer and the prevalence of structural defects (porosity) properties in the defects (porosity) properties in
the FDM process [24] play a part. Secondly, the programming was
FDM process [24] play a part. Secondly, the programming was not adjusted to accommodate the not adjusted to accommodate the
additional
additional 0.030.03 mmmm thick
thick glass
glass fiber
fiber layers. Third reason
layers. Third reason is is the
the glass
glass fiber
fiber layer
layer itself
itself as
as it is extremely
it is extremely
thin
thin with
with little
little to
to no
no airair gaps. This help
gaps. This help the
the layers
layers toto get
get trapped
trapped in in the
the semi-molten
semi-molten PLA PLA without
without any any
adhesive
adhesive or glue. It is to be noted that the focus here was on analysing the effect of adhesion and
or glue. It is to be noted that the focus here was on analysing the effect of adhesion and notnot
dimensional accuracy. This is a future research area that will include modifications
dimensional accuracy. This is a future research area that will include modifications to the dog-bone’s to the dog-bone’s
g-code
g-code toto accommodate
accommodate the the glass
glass fiber
fiber layers
layers to to ensure
ensure dimensional
dimensionalaccuracy.
accuracy.
J. Manuf. Mater. Process. 2019, 3, 96 10 of 16
J. Manuf. Mater. Process. 2019, 3, 96 10 of 16

4.5
4.45
4.4
4.35
Thickness (mm)

4.3
4.25
4.2
Measured thickness
4.15
Theoretical thickness
4.1
4.05
4
3.95
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Number of fiber layers

Figure 11. Comparison of measured and theoretical thickness values.

It is
is also
also important
importantto tohighlight
highlightthat thatthe themanufacture
manufactureofofGFRP GFRP products
products with
with increasing
increasing layers of
layers
glass fibers
of glass became
fibers became exceedingly
exceedingly difficult afterafter
difficult GFRP-9 GFRP-9 as the
as clearance
the clearance between
between the extruder nozzle
the extruder and
nozzle
the
andprint bed became
the print bed becamesmallersmaller
and smaller. The aimThe
and smaller. of theaim work is towork
of the identify
is tothe maximum
identify the number
maximum of
layers
number thatof can
layers be that
added canbefore the tensile
be added beforestrength
the tensile start to deteriorate.
strength To achieve this
start to deteriorate. in a consistent
To achieve this in
manner
a consistentand manner
keepingand in mind the issue
keeping in mindwith thethe losswith
issue of clearance
the loss of at clearance
GFRP-9, the print bedthe
at GFRP-9, was moved
print bed
downwards
was moved downwardsin the z-direction.
in the Six turns of the
z-direction. Six screws
turns ofunderneath
the screwsthe print bedthe
underneath lowered
print it bylowered
bed 0.2 mm.
This
it by was
0.2 mm.easy Thisto dowasas the experiment
easy to do as the is already
experiment making use of the
is already ‘Pause
making useatof height’ command
the ‘Pause that
at height’
gives
command amplethat time to turn
gives the four
ample time screws
to turnuniformly.
the four screws This practice
uniformly.provided
This enough
practice clearance
providedbetween
enough
the nozzlebetween
clearance and the part being manufactured
the nozzle and the part being to finish without issues.
manufactured to finish without issues.
As expected, the maximum thickness was observed for GFRP-15 (composite with 15 glass fiber
layers) with 4.42 mm compared to the theoretical theoretical 4.45 4.45 mm.mm. The thickness for the GFRP-13 composite
that showed the maximum load value during tensile test (Section 4.1) is 4.37 mm compared to the
theoretical
theoretical valuevalueofof4.39 4.39mm.mm.It It
is is
crucial
crucialto analyse
to analyse the the
maximum
maximum thickness valuevalue
thickness and its
andoverall effect.
its overall
A PLA A
effect. specimen of (say) 4.5
PLA specimen mm thickness
of (say) will show will
4.5 mm thickness a higher
showload value compared
a higher load valuetocompared
a specimen toofa
4specimen
mm thickness. of 4 mm Thisthickness.
is the caseThis
whenisthe thesame casematerial
when the is being
sameextruded
materialout and adhered
is being extruded using
outFDM.
and
In this work,
adhered using there
FDM.areIn glass
thisfiber
work, layers
therepresent
are glass that have
fiber shown
layers to enhance
present that havethe maximum load values.
shown to enhance the
Another
maximum tensile
loadtest was performed
values. based on
Another tensile testthe wasmaximum
performed theoretical
based on thickness of 4.5 mmtheoretical
the maximum (rounded
from 4.45 mm)
thickness of 4.5and mmthe results are
(rounded from shown
4.45 inmm) Figure
and 12. the Itresults
is cleararethat the 4.5
shown inmmFigurethick
12.specimen
It is clearhas a
that
higher load value (2105 N) compared to the 4.1 mm (2005 N) thick
the 4.5 mm thick specimen has a higher load value (2105 N) compared to the 4.1 mm (2005 N) thick pure PLA specimen. However, this
value
pure PLAis stillspecimen.
lower than GFRP-1 (2127.5
However, this value N). is Thisstillshows
lowerthat than the increase
GFRP-1 in thickness
(2127.5 N). This with the addition
shows that the
of glass fiber
increase layers does
in thickness withnottheadversely
additionaffectof glass the tensile properties.
fiber layers does not Thisadversely
is a clear affect
example the of how
tensile
properties.
HFDM can beThis is atoclear
used example of
manufacture how HFDM
products can be used
with tailored to manufacture
properties as increaseproducts with tailored
in the number of glass
properties
fiber layersas increase
help in the number
in increasing of glass
the tensile strengthfiberof layers help in
the GFRP increasing the tensile strength of the
composites.
GFRP composites.
J. Manuf. Mater. Process. 2019, 3, 96 11 of 16
J. Manuf. Mater. Process. 2019, 3, 96 11 of 16

2500

2000

1500
Load (N)

Pure PLA-4.1mm
1000
Pure PLA-4.5mm
GFRP-1
500

0
0 0.5 1 1.5 2 2.5 3 3.5
Displacement (mm)

Figure 12. Tensile test results for pure

pure PLA
PLA and
and GFRP-1
GFRP-1 composite.
composite.

4.2. Results from

4.2. Results from Microstructural
Microstructural Analysis
Analysis
It
It is
is evident
evident from from the the experimental
experimental results
results shown
shown in in Section
Section 4.14.1 that
that the
the maximum
maximum load load values
values
were obtained for GFRP-13 and from this point onwards, the tensile strength
were obtained for GFRP-13 and from this point onwards, the tensile strength started to deteriorate. started to deteriorate. This
is because
This that maximum
is because that maximumfiber content
fiber has beenhas
content reached
beenand any addition
reached and any would
additionresultwould
in degradation
result in
of
degradation of properties. In addition to the optimisation of 3D printing parameters, there arematerial
properties. In addition to the optimisation of 3D printing parameters, there are several other several
related factors that
other material play a factors
related role in ensuring
that play thata high
rolestrength
in ensuring products thatarehigh
manufactured. They include
strength products are
efficiency
manufactured. of theTheyinterface
includestress transferof
efficiency which represents
the interface stressthe transfer
adhesion of therepresents
which internal faces between
the adhesion
glass
of thefibers
internalandfaces
PLA,between
fiber aspect
glassratio,
fibersfiber
andorientation, fiber volume
PLA, fiber aspect fraction
ratio, fiber in the composite
orientation, fiber volumeand
the degree of crystallinity of the PLA [25]. To investigate these factors, the
fraction in the composite and the degree of crystallinity of the PLA [25]. To investigate these factors, fracture interfaces of PLA
and GFRP composites
the fracture interfaces were of PLA observed
and GFRP by JCM-5000
composites NeoScope™
were observed SEM as byshown
JCM-5000 in Figure 13. FirstSEM
NeoScope™ and
foremost, all the tested specimens exhibited a brittle failure i.e., without
as shown in Figure 13. First and foremost, all the tested specimens exhibited a brittle failure i.e., showing any necking or plastic
deformation
without showing beforeany fracture.
necking PLA is a brittle
or plastic material and
deformation the resulting
before fracture. GFRP
PLA iscomposites are alsoand
a brittle material brittle
the
in nature. FDM is a process of adding layers of extruded materials
resulting GFRP composites are also brittle in nature. FDM is a process of adding layers of extruded together and the development
of porositytogether
materials (or voids/micro
and the voids/air
development gaps)ofisporosity
a common (or phenomenon [26]. These
voids/micro voids/air voids
gaps) is ahave
commonbeen
highlighted
phenomenonin[26]. Figure 13. voids
These Voids have
are empty areas between
been highlighted layers where
in Figure 13. Voidsthere areisempty
no connection and a
areas between
higher percentage of voids could lead to significant degradation of
layers where there is no connection and a higher percentage of voids could lead to significant mechanical properties [27]. GFRP-1
shown in Figure
degradation 13b exhibitproperties
of mechanical a neatly packed single layer
[27]. GFRP-1 shown of glass fiber13b
in Figure to resist
exhibitthea tensile
neatly force
packed with the
single
brittle
layer of PLA andfiber
glass voidstohighlighted in its background.
resist the tensile force with the Thebrittle
voids of PLAtheandglassvoids
fiber highlighted
layers increase as
in its
can be seen in Figure 13c, d for GFRP-6 and GFRP-7 respectively. GFRP-13
background. The voids of the glass fiber layers increase as can be seen in Figures 13c, d for GFRP-6 shows the maximum load
value but alsorespectively.
and GFRP-7 shows several voids in
GFRP-13 Figure
shows the13e. However,
maximum load this phenomenon
value but also shows drastically
severalincreased
voids in
for GFRP-14 in Figure 13f as the maximum load value start going
Figure 13e. However, this phenomenon drastically increased for GFRP-14 in Figure 13f as the down from this product onwards.
The composites
maximum load with
value14startandgoing
15 layers
down of from
glass this
fibers undergo
product both interlayer
onwards. (decohesion
The composites withof14adjacent
and 15
layers
layers of ofFDM
glasspart)fibersandundergo
intraleyer (decohesion
both interlayer between adjacent
(decohesion of raster roads)
adjacent fracture
layers due topart)
of FDM the high
and
fiber volume. This is a commonly observed phenomenon with traditional
intraleyer (decohesion between adjacent raster roads) fracture due to the high fiber volume. This is a FDM-printed parts [28] and
when
commonlylayersobserved
of glass phenomenon
fibers are added with then the effect
traditional is enhancedparts
FDM-printed even[28]further
and resulting
when layers in extreme
of glass
deterioration
fibers are added of tensile
then the properties.
effect is enhanced even further resulting in extreme deterioration of tensile
properties.
J. Manuf. Mater. Process. 2019, 3, 96 12 of 16
J. Manuf. Mater. Process. 2019, 3, 96 12 of 16

SEMfracture
Figure 13. SEM fractureinterfaces
interfacesatatx100;
x100;(a)(a) Pure
Pure PLA;
PLA; (b)(b) GFRP-1;
GFRP-1; (c) (c) GFRP-6;
GFRP-6; (d) (d) GFRP-7;
GFRP-7; (e)
(e) GFRP-13;
GFRP-13; (f) (f) GFRP-14.
GFRP-14.

AA closer
closer look
look at
atthe
thesingle
singlelayered
layeredGFRP
GFRPcomposite
composite(GFRP-1)
(GFRP-1) inin
Figure
Figure1414
shows
shows thethe
interaction
interactionof
PLA layers with the glass fibers. It is clear that the fibers are trapped between the PLA
of PLA layers with the glass fibers. It is clear that the fibers are trapped between the PLA layers thatlayers that is
useful because
is useful because the fiber
the matrix
fiber matrixtakes
takesa ahigh
highportion
portionofofthe
theapplied
appliedtensile
tensileload
load resulting
resulting inin enhanced
enhanced
strength
strength compared
comparedtoto pure PLA.
pure ThisThis
PLA. entrapment is required
entrapment to ensure
is required to that the no
ensure thatadditional treatment
the no additional
is needed to ensure cohesion and bonding between PLA and glass fibers. In some
treatment is needed to ensure cohesion and bonding between PLA and glass fibers. In some cases, alkali treatment
cases,
for thetreatment
alkali fibers hasforbeen
theused to enhance
fibers has beenthe bonding
used capability
to enhance [29]. In this
the bonding case, however,
capability the glass
[29]. In this case,
fibers were added without any surface treatment.
however, the glass fibers were added without any surface treatment.
J. Manuf. Mater. Process. 2019, 3, 96 13 of 16
J.
J. Manuf.
Manuf. Mater.
Mater. Process.
Process. 2019,
2019, 3,
3, 96
96 13
13 of
of 16
16

Figure
Figure 14. SEM
14. SEM
Figure14. images
SEM images of
images of GFRP-1
of GFRP-1 at ×200
×200 (left)
at ×200
GFRP-1 at (left) and
and ×900
×900 (right).
×900 (right).
(right).

4.3. Results from

4.3. Results
Results from Ultrasonic
Ultrasonic Testing
Testing
This
This test
test has
has been
been done
done to further evidence
to further evidence the
the effectiveness
effectiveness ofof the
the HFDM
HFDM process
process and
and toto aid
aid the
the
tensile test results so that destructive testing might not be needed. This NDT method
tensile test results so that destructive testing might not be needed. This NDT method can potentially can potentially
help to
help to assess
assess the the bonding
bonding quality
bonding quality of
of the
of the GFRP
the GFRP composites.
GFRP composites.
composites. The The test
The test was
test conducted
was conducted
conducted withwith allall the
the
specimens
specimens
specimens andand three
and three measurements
three measurements were
measurements were taken
were taken based
taken based on
based onthe
on thedescription given
the description
description givenin Section
given in 3.3.
in Section The
3.3.time
Section 3.3. The
The
taken
time by theby
time taken
taken byhigh
thefrequency
the high acoustic
high frequency
frequency sound waves
acoustic
acoustic soundto
sound travelto
waves
waves tofrom onefrom
travel
travel transducer
from one through the
one transducer
transducer test piece
through
through the
the
and
test to the
piece second
and to transducer
the second is measured
transducer is and the
measured results
and are
the shown
results in
are Figure
shown
test piece and to the second transducer is measured and the results are shown in Figure 15. 15.
in Figure 15.

3.5
3.5

33

2.5
2.5

22
(μs)
Time(μs)

GFRP
GFRP composites
composites
Time

1.5
1.5 Pure
Pure PLA-4.1mm
PLA-4.1mm
11 Pure
Pure PLA-4.5mm
PLA-4.5mm

0.5
0.5

00
00 11 22 33 44 55 66 77 88 99 10
10 11
11 12
12 13
13 14
14 15
15 16
16
Number
Number of
of glass
glass fiber
fiber layers
layers

Figure 15.
Figure 15. Results
Results from
from ultrasonic
ultrasonic testing.
testing.

To establish
To base levels of sound travel through
through pure PLA, two specimens
specimens were used i.e., 4.1
4.1 mm
To establish
establish base
base levels
levels of
of sound
sound travel
travel through pure
pure PLA,
PLA, twotwo specimens were
were used
used i.e.,
i.e., 4.1 mm
mm
and 4.5
and mm. The times taken for sound waves to to
travel through the the
specimens werewere
veryvery
similar. It took
and 4.5 mm. The times taken for sound waves to travel through the specimens were very similar. It
4.5 mm. The times taken for sound waves travel through specimens similar. It
2.0 µs,
took 2.02.1
µs, and
µs2.1 µs2.0
and forµs
µs2.0 sound
for waveswaves
sound to travel
to through
travel the 4.1the
through mm thick
4.1 mm specimen.
thick On theOn
specimen. other
the
took 2.0 µs, 2.1 µs and 2.0 µs for sound waves to travel through the 4.1 mm thick specimen. On the
hand,hand,
other it took 2.1 µs, 2.1µs,µs andµs2.2 µs forµs
soundsound
waves to travel through the 4.5 mm specimen. This
other hand, it it took
took 2.1
2.1 µs, 2.1
2.1 µs and
and 2.2
2.2 µs for
for sound waves
waves toto travel
travel through
through the
the 4.5
4.5 mm
mm specimen.
specimen.
shows
This that both the specimens were closely packed and the sound waves travelled quickly through
This shows
shows that
that both
both the
the specimens
specimens werewere closely
closely packed
packed andand thethe sound
sound waves
waves travelled
travelled quickly
quickly
through
through the interconnected solid layers. The reason there are sometimes two values (as seen for pure
the interconnected solid layers. The reason there are sometimes two values (as seen for pure
PLA,
PLA, GFRP-4,
GFRP-4, GFRP-5,
GFRP-5, GFRP-6,
GFRP-6, GFRP-7,
GFRP-7, GFRP-8,
GFRP-8, GFRP-10,
GFRP-10, GFRP-11,
GFRP-11, GFRP-14,
GFRP-14, GFRP-15)
GFRP-15) is is because
because
J. Manuf. Mater. Process. 2019, 3, 96 14 of 16

the interconnected solid layers. The reason there are sometimes two values (as seen for pure PLA,
GFRP-4, GFRP-5, GFRP-6, GFRP-7, GFRP-8, GFRP-10, GFRP-11, GFRP-14, GFRP-15) is because two
measured values were the same. This shows uniformity of the GFRP composites and consistency of
measurements. A trend similar to the one achieved for tensile testing (Figure 9) is observed showing
that ultrasonic testing is an effective tool in assessing the bonding quality for the GFRP composites
made by HFDM.
Typically for a thicker product (like 4.5 mm), the sound waves should take a longer time to travel
from one end to the other because there are more hurdles in the way. However, the average travel
times for 4.1 mm and 4.5 mm thick pure PLA specimens are 2.0 µs and 2.1 µs respectively. Since the
difference is small, these specimens can be sued as reference points. As the number of glass fiber layers
increased for the GFRP composites, the time taken to travel through them also increased up to layer
13. After that, it took less time to travel through composites with layers 14 and 15. This is due to the
increase in porosity and voids in the GFRP composites. As mentioned earlier, porosity is a big issue for
FDM-printed parts and the inclusion of glass fibers enhance this effect further. The voids are generally
random in a FDM-printed part so that when stress is applied, the crack faces a higher resistance where
there are no voids and no resistance at all if it meet a void. The voids/porosity can be controlled by
changing the extruder and print bed temperature [30]. Furthermore, it has been proven by our results
and work done by other researchers [31,32] that products with the lowest porosity result in the highest
mechanical properties. As more and more layers were added to the PLA specimen, the voids increased.
More layers mean more voids as shown in Section 4.2. This increases the statistical probability of the
voids lining-up, thus reducing their randomness. This will lead to fracture of the GFRP at a lower load
value as has been observed for GFRP-14 and GFRP-15. This alignment of voids also resulted in the
sound waves travelling faster as the randomness had decreased, and the sound waves had a much
clearer path to travel as now they face more solid layers than voids. Sound waves travel faster through
solids than air because solid particles are more closely packed. This decrease in randomness of voids
and more solid interaction contribute to the faster travel of sound waves through GFRP composites
with more than 13 fiber layers.

5. Conclusions
In this study, commonly available PLA material was reinforced with 0.03 mm thick glass fiber
layers to manufacture GFRP composites using HFDM. Experimental investigation was conducted to
assess whether adding more glass fiber layers could yield higher tensile strength using standardised
tensile testing. Fracture interfaces of pure PLA and GFRP composites were analysed after tensile testing
using SEM micrographs. The products were also subjected to ultrasonic testing to assess whether this
NDT method is suitable to use for products made by HFDM. The following conclusions were drawn
from this study:
(1) Strategic addition of 0.03 mm thick glass fibers can increase the tensile strength of the PLA products.
The GFRP composites are also characterised by a brittle failure, same as the PLA material.
(2) More glass fiber layers mean more tensile strength. However, their addition causes the thickness
of the products to increase as well. The addition of 13 layers resulted in an increase of 32.5% in
tensile strength compared to pure PLA. However, the addition of more layers led to deterioration
of properties highlighting the optimal fiber content in the product.
(3) Controlling the tensile properties for the manufacture of bespoke and customised GFRP products
is possible. However, careful consideration should be given to the strength required and the
increase in thickness.
(4) The presence of voids is common in FDM-printed parts and this phenomenon increased with the
increase in the number of fiber layers. Until layer 13, the voids/porosity did not affect the upward
trend of tensile strength. However, layers 14 and 15 showed substantially large voids resulting in
lower mean values for maximum loads. SEM micrographs showed the presence of voids and
highlighted their size.
J. Manuf. Mater. Process. 2019, 3, 96 15 of 16

(5) Ultrasonic testing was used to assess the bond quality. It showed a similar trend as the tensile
testing with the peak at 13 layers and downward trend onwards. This shows that ultrasonic
testing can be sued in the future to assess the bond integrity of GFRP composites made by the
process of HFDM.

Author Contributions: J.B. conceptualised the idea, designed the methodology, undertook data curation,
investigation and formal analysis in addition to project administration and manuscript writing; Y.H. undertook
data curation, investigation and wrote the original draft; V.M. undertook data curation and formal analysis; S.S.-E.
acquired resources and worked on the original draft; H.S. reviewed and edited the final manuscript.
Funding: This research did not receive any external funding.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Dudek, P.F. FDM 3D printing technology in manufacturing composite elements. Arch. Metall. Mater. 2013,
58, 1415–1418. [CrossRef]
2. Justo, J.; Távara, L.; García-Guzmán, L.; París, F. Characterization of 3D printed long fibre reinforced
composites. Compos. Struct. 2018, 185, 537–548. [CrossRef]
3. Butt, J.; Mebrahtu, H.; Shirvani, H. Metal Rapid Prototyping Technologies. In Advances in Engineering Research;
Petrova, V.M., Ed.; Nova Science Publishers, Inc.: New York, NY, USA, 2017; Chapter 2; Volume 14, pp. 13–52.
4. Butt, J.; Shirvani, H. Additive, Subtractive, and Hybrid Manufacturing Processes. In Advances in Manufacturing
and Processing of Materials and Structures; CRC Press: Boca Raton, FL, USA, 2018; pp. 187–218.
5. Osman, M.A.; Atia, M.R. Investigation of ABS-rice straw composite feedstock filament for FDM. Rapid
Prototyp. J. 2018, 24, 1067–1075. [CrossRef]
6. Ivanov, E.; Kotsilkova, R.; Xia, H.; Chen, Y.; Donato, R.K.; Donato, K.; Godoy, A.P.; Di Maio, R.; Silvestre, C.;
Cimmino, S.; et al. PLA/graphene/MWCNT composites with improved electrical and thermal properties
suitable for FDM 3D printing applications. Appl. Sci. 2019, 9, 1209. [CrossRef]
7. Butt, J.; Mebrahtu, H.; Shirvani, H. Production of Multiple Material Parts Using a Desktop 3D Printer.
In Advances in Manufacturing Technology XXXI: Proceedings of the 15th International Conference on Manufacturing
Research, Incorporating the 32nd National Conference on Manufacturing Research, University of Greenwich, London,
UK, 5–7 September 2017; IOS Press: Amsterdam, The Netherlands, 2017; Volume 6, pp. 148–153. [CrossRef]
8. Butt, J.; Onimowo, D.A.; Gohrabian, M.; Sharma, T.; Shirvani, H. A desktop 3D printer with dual extruders
to produce customised electronic circuitry. Front. Mech. Eng. 2018, 13, 528–534. [CrossRef]
9. Butt, J. A Novel Additive Manufacturing Process for the Production of Metal Parts. Ph.D. Thesis, Anglia
Ruskin University, Cambridge, UK, 2016.
10. Ning, F.; Cong, W.; Qiu, J.; Wei, J.; Wang, S. Additive manufacturing of carbon fiber reinforced thermoplastic
composites using fused deposition modeling. Compos. Part B Eng. 2015, 80, 369–378. [CrossRef]
11. Liu, Z.; Lei, Q.; Xing, S. Mechanical characteristics of wood, ceramic, metal and carbon fiber-based PLA
composites fabricated by FDM. J. Mater. Res. Technol. 2019, 8, 3741–3751. [CrossRef]
12. Coppola, B.; Garofalo, E.; Di Maio, L.; Scarfato, P.; Incarnato, L. Investigation on the use of PLA/hemp
composites for the fused deposition modelling (FDM) 3D printing. In Proceedings of the AIP Conference
Proceedings, Bali, Indonesia, 11 July 2018; Volume 1981, p. 020086.
13. Tao, Y.; Wang, H.; Li, Z.; Li, P.; Shi, S.Q. Development and application of wood flour-filled polylactic acid
composite filament for 3D printing. Materials 2017, 10, 339. [CrossRef] [PubMed]
14. Patanwala, H.S.; Hong, D.; Vora, S.R.; Bognet, B.; Ma, A.W. The microstructure and mechanical properties of
3D printed carbon nanotube-polylactic acid composites. Polym. Compos. 2018, 39, E1060–E1071. [CrossRef]
15. Gkartzou, E.; Koumoulos, E.P.; Charitidis, C.A. Production and 3D printing processing of bio-based
thermoplastic filament. Manuf. Rev. 2017, 4, 1. [CrossRef]
16. Sanatgar, R.H.; Campagne, C.; Nierstrasz, V. Investigation of the adhesion properties of direct 3D printing of
polymers and nanocomposites on textiles: Effect of FDM printing process parameters. Appl. Surf. Sci. 2017,
403, 551–563. [CrossRef]
J. Manuf. Mater. Process. 2019, 3, 96 16 of 16

17. Eutionnat-Diffo, P.A.; Chen, Y.; Guan, J.; Cayla, A.; Campagne, C.; Zeng, X.; Nierstrasz, V. Stress, strain and
deformation of poly-lactic acid filament deposited onto polyethylene terephthalate woven fabric through 3D
printing process. Sci. Rep. 2019, 9, 1–18. [CrossRef] [PubMed]
18. Butt, J.; Shirvani, H. Experimental analysis of metal/plastic composites made by a new hybrid method.
Addit. Manuf. 2018, 22, 216–222. [CrossRef]
19. Keleş, Ö.; Blevins, C.W.; Bowman, K.J. Effect of build orientation on the mechanical reliability of 3D printed
ABS. Rapid Prototyp. J. 2017, 23, 320–328. [CrossRef]
20. Garg, A.; Bhattacharya, A.; Batish, A. Failure investigation of fused deposition modelling parts fabricated at
different raster angles under tensile and flexural loading. Proc. Inst. Mech. Eng. Part B J. Eng. Manuf. 2017,
231, 2031–2039. [CrossRef]
21. BS EN ISO 527-2:2012. Plastics—Determination of Tensile Properties–Part 2: Test Conditions for Moulding and
Extrusion Plastics; British, European and International Standard: London, UK, 2012.
22. Popescu, D.; Zapciu, A.; Amza, C.; Baciu, F.; Marinescu, R. FDM process parameters influence over the
mechanical properties of polymer specimens: A review. Polym. Test. 2018, 69, 157–166. [CrossRef]
23. Xinhua, L.; Shengpeng, L.; Zhou, L.; Xianhua, Z.; Xiaohu, C.; Zhongbin, W. An investigation on distortion of
PLA thin-plate part in the FDM process. Int. J. Adv. Manuf. Technol. 2015, 79, 1117–1126. [CrossRef]
24. Gordeev, E.G.; Galushko, A.S.; Ananikov, V.P. Improvement of quality of 3D printed objects by elimination
of microscopic structural defects in fused deposition modeling. PLoS ONE 2018, 13, e0198370. [CrossRef]
25. Yang, Z.; Feng, X.; Bi, Y.; Zhou, Z.; Yue, J.; Xu, M. Bleached extruder chemi-mechanical pulp fiber-PLA
composites: Comparison of mechanical, thermal, and rheological properties with those of wood flour-PLA
bio-composites. J. Appl. Polym. Sci. 2016, 133. [CrossRef]
26. Liao, Y.; Liu, C.; Coppola, B.; Barra, G.; Di Maio, L.; Incarnato, L.; Lafdi, K. Effect of Porosity and Crystallinity
on 3D Printed PLA Properties. Polymers 2019, 11, 1487. [CrossRef]
27. Mehdikhani, M.; Gorbatikh, L.; Verpoest, I.; Lomov, S.V. Voids in fiber-reinforced polymer composites:
A review on their formation, characteristics, and effects on mechanical performance. J. Compos. Mater. 2019,
53, 1579–1669. [CrossRef]
28. Young, D.; Wetmore, N.; Czabaj, M. Interlayer fracture toughness of additively manufactured unreinforced
and carbon-fiber-reinforced acrylonitrile butadiene styrene. Addit. Manuf. 2018, 22, 508–515. [CrossRef]
29. Hu, R.; Lim, J.K. Fabrication and mechanical properties of completely biodegradable hemp fiber reinforced
polylactic acid composites. J. Compos. Mater. 2007, 41, 1655–1669. [CrossRef]
30. Spoerk, M.; Gonzalez-Gutierrez, J.; Sapkota, J.; Schuschnigg, S.; Holzer, C. Effect of the printing bed
temperature on the adhesion of parts produced by fused filament fabrication. Plast. Rubber Compos. 2018, 47,
17–24. [CrossRef]
31. Chin Ang, K.; Fai Leong, K.; Kai Chua, C.; Chandrasekaran, M. Investigation of the mechanical properties
and porosity relationships in fused deposition modelling-fabricated porous structures. Rapid Prototyp. J.
2006, 12, 100–105. [CrossRef]
32. Cuan-Urquizo, E.; Barocio, E.; Tejada-Ortigoza, V.; Pipes, R.B.; Rodriguez, C.A.; Roman-Flores, A.
Characterization of the mechanical properties of FFF structures and materials: A review on the experimental,
computational and theoretical approaches. Materials 2019, 12, 895. [CrossRef]

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