Solid Freeform Fabrication 2017: Proceedings of the 28th Annual International

Solid Freeform Fabrication Symposium – An Additive Manufacturing Conference

The application of composite through-thickness assessment

to additively manufactured structures

Isam S. Bitar1,*, Nesma T. Aboulkhair2, Richard Leach1

1
Manufacturing Metrology Team, Faculty of Engineering, University of
Nottingham, NG7 2RD, UK
2
Centre for Additive Manufacturing, Faculty of Engineering, University of
Nottingham, NG7 2RD, UK
*
Corresponding author. E-mail: [email protected]

Abstract
This study looks into the applicability of through-thickness assessment to additive
manufacturing (AM) carbon-fibre reinforced polymers (CFRPs). The study utilised a material
extrusion printer that uses fused filament fabrication and composite filament fabrication
technologies to manufacture functionally-graded polymer and composite polymer parts. The
matrix material of choice was nylon 6. Samples were printed exploring a range of
reinforcement volume content. In summary, this study presents an assessment of the
applicability of through-thickness testing to AM CFRP specimens and provides a performance
comparison between AM composite through-thickness properties and the properties of
equivalent CM CFRP specimens.

Keywords
carbon fibre, through thickness, composite, mechanical behaviour, additive manufacture

1. Introduction
1.1. Composite materials
Composite materials are non-homogenous materials made from two or more constituent
materials that differ in their mechanical and chemical properties and provide desirable tailored
characteristics when combined. Fibre-reinforced polymers (FRPs) are composite materials
made of a polymer matrix, which houses, supports and transfers the bearing load onto a fibre
reinforcement that acts as the load-bearing structure in the composite. The use and manufacture
of FRPs has been widespread since the 1930s, with the mass production of glass fibres
discovered in 1933 [1] and they are today commonly used in industries that include aerospace,
automotive, construction and defence [2].

1.2. Carbon fibre reinforced polymers

Carbon fibre reinforced polymers (CFRPs) use carbon-based (usually graphite) fibres
of 5 µm to 10 µm in diameter, twisted into yarns that are woven or embedded into a polymer
matrix [3]. These fibres provide good tensile properties with respect to their mass, making them
useful for a large number of applications, including aerospace, marine and automotive parts, as
well as end-user products, such as golf clubs and fishing rods.

1.3. Through-thickness testing of CFRPs

The nature of CFRPs and other matrix-based composites, where fibre reinforcement is
applied in a plane-wise fashion, that is, in a layer-by-layer format, means that CFRP structures
exhibit anisotropic properties [4]. This is primarily due to the relatively low structural strength
of the matrix (polymer) material of the composite, as well as, in some cases, a relatively weak

616
fibre-matrix interface between layers, which may constitute a bottle-neck in the transverse (Z-
direction) structural strength of a CFRP structure [2].

Through-thickness testing (also known as Z-direction tensile testing) is a mechanical

property assessment method that applies a tensile force to a test specimen whose axis is
perpendicular to the reinforcement plane of a composite material (Figure 1) [5]. The main
purpose of through-thickness testing is to assess the resistance to lamellar tearing [6], a form
of (initially microscopic) cracks that form parallel to the plate surface of a material and induce
weaknesses in the material’s through-thickness (Z) direction. The process of through-thickness
testing is well-understood in the polymer composite world, and specification standards have
been developed for testing matrix composite materials in the through-thickness direction [5].
Several studies have been conducted on through-thickness performance testing of matrix
composites [7][8][9] and some introduced additional Z-direction reinforcements before testing
[10]. However, these studies focus on the conventionally processed composites, but no
literature is available on through-thickness testing of additively-manufactured (AM)
composites.

Figure 1: Through-thickness tensile test specimen

This study will aim to shed light on the applicability of through-thickness testing on
AM parts as well as to evaluate their performance. This is of particular importance since AM
is the latest process for manufacturing complex, non-homogeneous, micro-structured parts
such as CFRPs. Current composite manufacturing processes, including thermoplastic
compression moulding, filament winding, resin infusion, etc., offer various features,
advantages and disadvantages, which provide variety in approaches and the potential to
optimise production according to a balance of needs (e.g. size, quality, production rate, tooling
costs). AM is the latest composite manufacturing process to be added to the list, which comes
with its own design opportunities and challenges. Therefore, it is important to establish AM’s

617
properties by applying relevant test methods to specimens produced by it, adding to the body
of knowledge on the different properties that this new composite manufacturing process can
bring forth.

1.4. CFRPs in AM
While conventional manufacturing techniques of CFRPs have been well-studied, the
long cycle times and high production costs associated with their production have given AM a
distinct potential advantage over other manufacturing processes, because AM carries the
potential for both reducing time and cost of production as well as increasing design complexity
of the end products [11].

The two main AM techniques that have been studied for the production of CFRPs are
powder bed fusion, e.g. selective laser sintering (SLS) [12][13], and material extrusion, e.g.
fused deposition modelling (FDM) [14][15] – also known as fused filament fabrication – with
the majority of research into polymer composites falling into the latter method [16][17][18].

1.5. Material extrusion of CFRPs

Material extrusion (ME) (Figure 2) is an AM process that makes use of polymer
filament feedstocks by extruding the material in its melted form, through a computer-controlled
liquefier head, into a build platform in a layer by layer fashion. Advanced ME machines can
support multiple extrusion nozzles with multiple materials. For example, in this case, a polymer
matrix material and a separate fibre reinforcement material. Several studies have tested ME
composites compared to composites manufactured using conventional methods and have
shown promising results in tensile strength [15][16]. While several studies into the tensile
properties of AM composite parts have been conducted [14][15][16], these studies focused on
the longitudinal (in-plane) tensile properties, rather than the transverse (through-thickness)
tensile strength.

2. Materials and methods

2.1. Manufacturing equipment and parameters
The experiments utilise an in-house Markforged Mark Two material extrusion machine
for manufacturing the test specimens. The machine uses fused filament fabrication (FFF) and
composite filament fabrication (CFF) technologies to manufacture functional polymer and
composite polymer parts [19]. The machine is capable of printing four types of reinforcement
material: carbon fibre, fibreglass, Kevlar and high strength, high temperature (HSHT)
fibreglass, in addition to two choices of matrix material: nylon and the Markforged Onyx [20].

Test specimens were manufactured from nylon 6 as a matrix material and carbon fibre
as reinforcement. All materials used were Markforged-certified materials and were obtained
from CREAT3D Limited, a certified local distributor.

The print bed was levelled and print nozzle heights adjusted prior to the first print. The
print bed was coated with a layer of Elmer’s washable school glue in order to protect the bed,
facilitate part removal and prevent the part from moving while the print is in progress. A 6 mm-
wide nylon brim was printed under each specimen in order to increase the contact area with the
print bed and prevent thermal curling. The nylon and carbon fibre nozzle temperatures were
set to 275 °C and 245 °C, respectively.

618
 Figure 2: Schematic of the material extrusion process.

Nylon support structures were used during the printing process, which were later
removed through manual prying with pliers. A purge line was printed prior to each print in
order to purge the print head and extruder from any nylon that may have accumulated moisture
from the atmosphere during idle times. The print head was also purged of nylon residue that
may have accumulated on the nylon nozzle while carbon fibre was routed through the part. All
specimens were printed with 100% fill density with a rectangular fill pattern to ensure the
maximum fill density as per manufacturer recommendations. Layer height was fixed by the
manufacturer at 0.125 mm. The specimens were built five at a time in the spacing shown in
Figure 3.

2.2. Specimen design

All test specimens were manufactured to comply with ASTM D7291/D7291M-15
standards on through-thickness tensile testing [5]. The specimens were manufactured in
accordance with recommendations from the National Physical Laboratory (NPL) [21], as
shown in Figure 4. The test piece is a standard Royal Armament Research and Development
Establishment (RARDE) specimen with 25 mm square ends and a 40 mm nominal height. The
reduced gauge section is 12 mm long and has a rectangular cross-section of 10 mm × 16 mm.
The reinforcement was applied in the x-y plane (perpendicular to the height of the specimen)
and was applied uniformly along the entire height of the specimen.

The reinforcement in each layer was laid out in a concentric fashion, where carbon-
fibre rings ran continuously from the perimeter to the centre (Figure 5). This amounted to
twelve rings in base layers and three rings in middle layers.

619
 Figure 3: Specimen spacing in each print

Figure 4: Schematic of through-thickness tensile CFRP specimen in accordance with ASTM

D7291/D7291M-15 standards and NPL recommendations [21]

Figure 5: Per-layer layout of carbon fibre rings (blue) within nylon matrix and supports
(white) at the base (right) and in the gauge section (left)

620
 2.3. Specimen sets and experimental design
Four sample sets were produced for through-thickness tensile testing, each set
consisting of five identical specimens to provide statistical significance and room for printing
errors. The four sets were as follows:

1. Control samples: 100% nylon samples, no composite material deposited. These were
used to benchmark the rest of the samples and provide a comparative baseline.

2. 25% fibre volume fraction (fvf) of carbon fibre (one layer of fibre in every three layers
of nylon), evenly distributed along the through height of the specimen, i.e.
homogeneously reinforced.

3. Half-density samples: approx. 37% fvf of carbon fibre to further test for correlation
between fibre content and tensile strength.

4. Full-density samples: i.e. composite reinforcement present in print every layer, which
amounts to approximately 78% fvf of carbon fibre for maximum possible fibre density.

2.4. Sample preparation

Support structures were removed from the printed samples using pliers. The test
samples were then measured to check machine dimensional accuracy and prepared for tensile
testing in accordance with ASTM D7291/D7291M-15 standards. As part of the preparation, a
number of end tabs were machined from 6082 aluminium alloy bars using a XYZ Proturn SLX
425 lathe. End tabs and specimens were later grit-blasted using 150 µm to 300 µm glass beads
in order to provide a suitable surface for adhesive bonding.

A number of suitable adhesives were tested to determine the best choice for bonding
the samples to end tabs. The Loctite 410 instant adhesive was chosen. Specimens were bonded
to the aluminium alloy end tabs using metal jigs to align the parts. Specimens were then left to
cure for a minimum of twenty-four hours prior to testing.

Following curing, specimens were sprayed with a black and white texture in order to
allow for the video strain gauge to record specimen strain under tensile loading while testing
(Figure 6). A total of four specimens were bonded and prepared for the first round of testing
(one of each set), followed by a second round for the remaining sixteen specimens.

2.5. Tensile testing and analysis

An Instron 5965 Mechanical Test Machine was used to determine the tensile properties
of the specimens. This was then to be used to perform an evaluation on through-thickness
properties and ultimately the applicability of through-thickness testing to AM parts. The strain
under tensile loading was measured using an iMetrum video strain gauge that tracks the
displacement of a set of texture dots painted on the specimens.

621
 Figure 6: Through-thickness tensile specimen with fitted end-tabs and texture spray paint
ready for testing.

The test was conducted at 18 °C and 50% relative humidity. Due to time and resource
restrictions, a test speed of 1mm/min was used as per NPL testing protocols [22] instead of the
ASTM D7291/D7291M-15 0.1 mm/min.

3. Results and discussion

3.1. Manufacturability
A total of five failed and twenty successful specimens were produced during the
printing process in addition to a single-sample test print. The single pure nylon sample was
printed in order to ensure all parameters are set correctly and that no complications during the
print process occurred. Following the successful test print, five Set 1 (0% carbon fibre)
specimens were added to the print job. However, significant distortion was present on all five
samples (Figure 7). The distortion observed on all five samples is believed to be the result of
thermal curling induced by the rapid cooling of nylon layers as they were laid onto the unheated
substrate. Adding a brim to the parts seemed to reduce thermal distortion but failed to eliminate
it. It was later observed that the addition of carbon fibre to the specimens significantly increased
part stiffness and dimensional integrity. As a result, the pure nylon parts were re-printed with
fibre-reinforced bases and a pure nylon gauge (Figure 8). Additionally, prior to each print, the
print substrate was put under running hot tap water until hot in order to reduce the thermal gap
between the printer nozzle and the substrate.

622
 Figure 7: Thermal distortion in original Set 1 specimens.

Figure 8: Set 1 (0% carbon fibre) specimen with reinforced bases.

Thermal distortion was only observed during the five-sample print - but not in the initial
single-sample print - despite both being printed with pure nylon. The reason behind this is
believed to be the five-fold delay in laying each consecutive layer onto the previous ones as
the machine continued to build the next layers on the remaining four specimens, which allowed
more time for the plastic to cool down and deform. The option to programme the machine to
print the samples in succession, rather than simultaneously, was considered. However, as of
the time of writing, the printer firmware was incapable of incorporating such measures.

623
 The production of Set 2 (25% fvf) and Set 3 (37% fvf) samples was successful. Set 4
(78% CF) prints showed carbon fibre splinters protruding through the matrix throughout the
specimen’s height (Figure 9). Repeat prints of Set 4 specimens conducted by the manufacturer
confirmed the observation. The reason for this observation is thought to be excessive fibre
content, where the density of fibre in the part is too great for the printer to print successfully.
This may have caused some of the printed fibre to either tangle with previously-laid fibres or
to dislodge the fibre, causing splinters and possible breaks in the printed fibre. In applications
where maximum density is required, it is recommended that the parts be ‘striped’. That is, to
print several layers of fibre in succession (usually ten) followed by a few (usually two) pure
nylon layers. This is thought to mitigate the effect of fibre splintering. This has not yet been
done, but it is planned for future investigation.

Figure 9: Set 4 specimen with CF splinters

Dimensional measurements were recorded for the gauge sections of each specimen
using a calibrated Clarke® Electronic Digital Calliper. A total of ten repeats were taken of each
measurement and the tolerances were calculated with 95% confidence. As shown in Table 1,
fibre-reinforced specimens consistently showed closer values to the target dimensions of
10 mm and 16 mm of the gauge section than the non-reinforced samples. This is primarily due
to the fact that carbon fibre adds stiffness to the printed parts, thus greatly reducing the effects
of the various factors that may influence the dimensional integrity, such as thermal and
gravitational deformations. Conversely, fibre-reinforced specimens generally exhibit larger
tolerances than non-reinforced samples. This can be attributed to the higher surface roughness
of fibre-reinforced specimens caused by the presence of the fibre within the matrix, thus
causing the matrix to take a rougher outline along the outer surfaces.

624
 Table 1: Dimensional measurements of through-thickness specimens.
Sample Wide Section/mm Narrow Section/mm
Set 1 1 15.83 ± 0.03 9.71 ± 0.03
2 15.85 ± 0.04 9.70 ± 0.05
3 15.76 ± 0.05 9.70 ± 0.03
4 15.73 ± 0.03 9.68 ± 0.02
5 15.83 ± 0.01 9.68 ± 0.03
Set 2 1 15.98 ± 0.05 9.78 ± 0.01
2 15.97 ± 0.05 9.86 ± 0.03
3 15.93 ± 0.03 9.80 ± 0.02
4 15.99 ± 0.06 9.80 ± 0.05
5 15.96 ± 0.05 9.81 ± 0.03
Set 3 1 15.99 ± 0.04 9.81 ± 0.03
2 16.01 ± 0.06 9.83 ± 0.05
3 15.99 ± 0.05 9.80 ± 0.01
4 15.94 ± 0.07 9.83 ± 0.05
5 16.01 ± 0.05 9.81 ± 0.05
Set 4 1 16.02 ± 0.05 9.88 ± 0.07
2 15.96 ± 0.06 9.91 ± 0.04
3 15.99 ± 0.04 9.88 ± 0.05
4 16.02 ± 0.06 9.88 ± 0.07
5 16.04 ± 0.03 9.89 ± 0.06

3.2. Test results

Eight out of twenty specimens failed due to premature adhesive failure. The average
stress value reached at the time the adhesive failed resulting in the tensile test being terminated
was 7 MPa ± 1.5 MPa. Several factors could have played part in the failure of the adhesive.
Firstly, since samples were left to cure for a total of twenty-four hours prior to testing, it is
plausible that the adhesive had not reached its ultimate strength and that further time should
have been allowed for sufficient curing. Secondly, it is possible that the specimens may have
required a coarser surface texture in order to provide a larger surface area for bonding. A good
practice guide by the Composites UK Trade Association on the adhesive bonding of composites
shows that medium grit (120 µm to 200 µm) is ideal for composite surface pre-treatment. The
guide further recommends that cryo-blast and soda-blast be used in treating composites, as
opposed to shot-blasting, which is more suitable for metals [23]. Indeed, upon touch, little
difference in surface roughness was observed following the grit-blasting of the specimens than
the original 3D-printed surface roughness. A more suitable grit will be considered for future
trials. Finally, nylon is notoriously difficult to bond due to the presence of slip additives within
the material, which can migrate to the surface of the material and interfere with proper bonding.
As a result, it may have been difficult for the adhesive to properly bond to the nylon surface,
thus causing premature delamination in the glue and nylon interface.

Successful specimens were as follows:

1. Set 1 (0% fvf): Five out of five specimens successful.

2. Set 2 (25% fvf): Two out of five specimens successful.
3. Set 3 (37% fvf): One out of five specimens successful.
4. Set 4 (78% fvf): Five out of five specimens successful.

625
 3.3. Fracture surfaces
Set 1 specimens
All five Set 1 specimens fractured at the interface between the fibre-reinforced base and
the start of the gauged section, as opposed to the expected typical fracture within the gauge
section itself. This suggests apparent weakness in the adhesion between carbon-fibre-
reinforced layers and unreinforced layers. The cause of this systemic weakness is unknown,
but it is thought that the sudden drop in fibre volume fraction may be inducing certain defects
in the interface within the printed part. The fracture interface itself (Figure 10) exhibited ductile
behaviour in all five specimens. This is due to the ductile nature of the nylon matrix.

Figure 10: The ductile fracture interface of Set 1 specimens

Sets 2 and 3 specimens

The fracture surfaces of successful specimens from Sets 2 and 3 were similar. All
successful specimens belonging to the two sets failed within the gauge section. The fracture
interface was horizontal and suggested a final matrix failure (Figure 11). It was clear that the
fibre had some function in resisting tension, as fibres were observed to stretch and break during
testing.

626
 Figure 11: The fracture interface of Sets 2 and 3. The top photo shows carbon fibre being
stretched during testing, suggesting the fibre assisted in resisting tension

Set 4 specimens
Similarly to Sets 2 and 3, the failure occurred within the gauged section for all
successful specimens. However, the role of carbon fibre providing tensile strength was much
more apparent in Set 4 specimens as can be seen from the stretched and torn fibres embedded
in the fracture surfaces (Figure 12).

Figure 12: The fracture interface of Set 4 specimens. Torn fibres suggest prominent role of
CF in providing tensile strength

627
 3.4. Tensile properties
Table 2 shows the ultimate tensile strength (UTS) values for each tested specimen.

Table 2: UTS values for tested specimens

Max Load UTS Failure
Sample /N /MPa Mode
Set 1 1 1893.74 12.44 Specimen
2 1069.83 7.01 Specimen
3 1148.35 7.54 Specimen
4 908.65 5.91 Specimen
5 495.62 3.23 Glue
Set 2 1 1425.6 9.11 Specimen
2 1337.76 8.54 Specimen
3 1367.35 8.73 Glue
4 1183.73 7.58 Glue
5 1050.23 6.72 Glue
Set 3 1 1354.43 8.64 Specimen
2 1195.11 7.69 Glue
3 1309.49 8.36 Glue
4 1167.52 7.43 Glue
5 1159.33 7.39 Glue
Set 4 1 1101.54 7.01 Specimen
2 1128.63 7.14 Specimen
3 1137.17 7.17 Specimen
4 1176.02 7.43 Specimen
5 1134.31 7.17 Specimen

Figures 13 to 15 illustrate the tensile performance of each set relative to the other
three. These graphs only illustrate specimens that have completed the test successfully.

As shown in Figure 13, the tensile behaviour of Set 1 specimens matched the ductile
fracture patterns observed. While most samples had a fairly consistent elastic behaviour, the
variation in UTS was high. This phenomenon can be explained by the possible presence of
random print defects, which can unpredictably alter the tensile strength of the specimen. These
defects are likely to be due to the sudden change in fvf. More research is required to study the
effects of changes in fvf on part integrity. All specimens exhibited typical ductile behaviour,
including the presence of high and low yield points. The ductile behaviour is consistent with
the expected behaviour of nylon 6.

Figure 14 illustrates the tensile behaviour of the specimens from Sets 2 and 3. They
exhibit consistent elastic behaviour and recorded the highest UTS values. This suggests the
existence of an optimum fvf at which the combined strength of the nylon matrix and carbon
fibre yields the highest tensile strength. Further research is required to determine this optimum
point. The specimens were significantly less ductile than Set 1 specimens in their post-yield
behaviour. This is expected as the carbon fibre provides additional stiffness to the specimens.
However, further tests are required to provide better certainty of the results.

628
 The tensile behaviour of Set 4 specimens is shown in Figure 15. As with other
reinforced specimens, the elastic behaviour of all specimens tested was consistent, as were the
UTS values. This is expected as the carbon fibre provides a more consistent build. However,
the UTS values of Set 4 specimens were significantly lower than those of Sets 2 and 3. This
suggests that the higher fvf came at the expense of matrix material, whose primary function is
to provide interlayer adhesion, thus reducing overall strength.

Figure 13: Tensile stress behaviour of Set 1 specimens

Figure 14: Tensile stress behaviour of Sets 2 and 3 specimens

629
 Figure 15: Tensile stress behaviour of Set 4 specimens

3.5. Future Work

Future work is planned to study the effects of print defects, fvf variations and conduct
more thorough through-thickness tensile testing.

More specimens
More specimens will be manufactured in order to study the effect of the sudden change
in fvf. This may be accompanied with SEM inspection of fracture interfaces to detect possible
print defects in the interface between the last reinforcement layer and the first pure nylon layer.
Different measures will be explored in order to prevent thermal deformation of unreinforced
samples, including the introduction of a thick raft to be filed down in post-process.

Additional sets exploring a range of reinforcement ratios will be printed in order to

determine the optimum fvf that provides the highest UTS values. All sets would have at least
ten specimens in order to provide additional statistical significance as well as allow for failed
parts.

Optimisation of the test

As demonstrated through the trials, despite following proper protocols, specimen
bonding results were erratic, where glue failure occurred on a range of loads which seemed to
follow no particular pattern. This is largely due to the difficult nature of nylon as a low surface
energy material. As such, options will be explored in adjusting and altering the test method in
order to account for inherent difficulties in both the material and the process, including
exploring the options to bond specimens using a lap joint arrangement.

Acknowledgments
The authors would like to thank Dr Graham Sims, Dr Richard Shaw and Dr Peter
Woolliams from the National Physical Laboratory for supporting this work.

630
Funding
This work was funded by EPSRC [Grant EP/L01534X/1], Exova and the National
Physical Laboratory.

4. References

[1] G. Slayter, Method and apparatus for making glass wool. US2133235 A, 1938.

[2] M. Piggott, Load bearing fibre composites, Kluwer Academic Publishers, Dordrecht,
2005.

[3] Zoltek, What is Carbon Fiber?. http://zoltek.com/carbonfiber/, 2017 (accessed 08.06.17).

[4] F. Campbell, Structural composite materials, 1st ed., ASM International, Materials Park,
Ohio, 2010.

[5] ASTM, D7291/D7291M-15 standard test method for through-thickness “flatwise” tensile
strength and elastic modulus of a fiber-reinforced polymer matrix composite material, 1st ed.,
ASTM International, West Conshohocken, PA, 2015.

[6] The Steel Construction Institute (SCI), Guidence Note: Through Thickness Properties,
2015.

[7] S. Vali-shariatpanahi, Determination of through thickness properties for Composite thick

laminate, ICCM-17 17th International Conference on Composite Materials, 2009.

[8] R. Karkkainen, P. Moy, J. Tzeng, Through-thickness property measurement of three-

dimensional textile composites, U.S. Army Research Laboratory, Aberdeen Proving Ground,
MD, 2009.

[9] R. Govender, L. Louca, A. Pullen, A. Fallah, G. Nurick, Determining the through-

thickness properties of thick glass fiber reinforced polymers at high strain rates, Journal of
Composite Materials 46 (2012) 1219-1228.

[10] C. Steeves, N. Fleck, In-plane properties of composite laminates with through-thickness

pin reinforcement. International Journal of Solids and Structures 43 (2006) 3197-3212.

[11] R. Hague, I. Campbell, P. Dickens, Implications on design of rapid manufacturing,

Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical
Engineering Science 217 (2003) 25-30.

[12] W. Zhu, C. Yan, J. Yang, S. Wen, Y. Shi, Parameter optimization for preparing carbon
fiber/epoxy composites by selective laser sintering, Symposium of 2015 Annual International
Solid Freeform Fabrication, 2015, p. 857–865.
[13] C. Yan, L. Hao, L. Xu, Y. Shi, Preparation, characterisation and processing of carbon
fibre/polyamide-12 composites for selective laser sintering, Composites Science and
Technology 71 (2011) 1834-1841.

631
[14] F. Ning, W. Cong, Y . Hu, H. Wang, Additive manufacturing of carbon fiber-reinforced
plastic composites using fused deposition modeling: Effects of process parameters on tensile
properties, Journal of Composite Materials 51 (2016) 451-462.

[15] H. Tekinalp, V. Kunc, G. Velez-Garcia, C. Duty, L. Love, A. Naskar et al., Highly

oriented carbon fiber–polymer composites via additive manufacturing, Composites Science
and Technology 105 (2014) 144-150.

[16] F. van der Klift, Y. Koga, A. Todoroki, M. Ueda, Y. Hirano, R. Matsuzaki, 3D printing
of continuous carbon fibre reinforced thermo-plastic (CFRTP) tensile test specimens. Open
Journal of Composite Materials 06 (2016) 18-27.

[17] F. Ning, W. Cong, J. Wei, S. Wang, M. Zhang, Additive manufacturing of CFRP

composites using fused deposition modeling: effects of carbon fiber content and length.
Volume 1: Processing, 2015.

[18] M. Shofner, K. Lozano, F. Rodríguez-Macías, E. Barrera, Nanofiber-reinforced

polymers prepared by fused deposition modeling. Journal of Applied Polymer Science 89
(2003) 3081-3090.

[19] Markforged, Mark Mark Two Enterprise Kit. https://markforged.com/product/mark-two-

enterprise-kit/#specifications, 2017 (accessed 08.06.17).

[20] Markforged, Markforged Materials. https://markforged.com/materials/, 2017 (accessed

08.06.17).

[21] W. Broughton, M. Gower, M. Lodeiro, R. Shaw, G. Sims, Validation of a test geometry

designed for the measurement of through-thickness tensile and compressive properties of
polymer matrix composites, National Physical Laboratory, Teddington, Middlesex, not
published.

[22] R. Shaw, G. Sims, Through-thickness tensile testing of polymer matrix composites,

National Physical Laboratory. Teddington, Middlesex, 2004.

[23] Composites UK, Adhesive bonding of composites, Composites UK Trade Association,

not dated.

632