Materials Today: Proceedings 56 (2022) 2096–2103

Contents lists available at ScienceDirect

Materials Today: Proceedings

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

The relationship between the thickness and flexural strength

of unidirectional carbon fiber reinforced polymers manufactured
through VARI
Pitso Taele ⇑, Thubalakhe Masango, Velaphi Msomi
Cape Peninsula University of Technology, Faculty of Engineering and the Built Environment, Bellville 7535, South Africa

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

Article history: The three-point bending tests at constant loading rates and constant span to thickness ratio were carried
Available online 16 December 2021 out to investigate the behavior and effect of thickness on unidirectional carbon fiber reinforced polymers
(UD-CFRPs). The specimens were manufactured in different thicknesses through the vacuum assisted
Keywords: resin infusion (VARI) technique. Experiments were conducted at strain rate of 0.01 s 1 with constant
Composites span-to-thickness ratio of 40:1 according to ASTM D7264/D7264M-15 test standard. It was observed
Mechanical properties from the results that the thicker the laminate, the higher the peak force, and less strain the structure
Three-point bending
can withstand. The failure stress and strain curves revealed that the less thick the specimen is, the higher
Flexural Strength
the stress and strain the specimen will endure. Failure modes and microstructures of specimens were
observed to analyze and explain the underlying mechanism of reinforcing effect. It was observed that
the specimens failed due to individual layer delamination and was found to be the main mechanism
for avoiding catastrophic failure and interfacial damage of the sandwich specimen.
Copyright Ó 2021 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the International Confer-
ence on Applied Research and Engineering 2021

1. Introduction and aluminium alloys due to high skilled labour needed to handle
their complex structure to give desired results [4]. CPRPs have a
Carbon fibers were discovered as reinforcements in the early very complex structure, unlike metals they are not isotropic. This
1960s [1]. Ever since then, their use as structural material in the means that their mechanical properties are not the same of all
aerospace, automotive, defense, and sports to name few has directions [5]. In addition to this, there is limited literature review
increased over the past decades because of their favorable proper- research data available which is aimed at defining load capabilities
ties such as low density, higher specific strengths and stiffnesses, of composite structures due to their versatility in manufacturing
superior corrosion resistance as well as improved fatigue [2]. Over processes, fiber orientation, type of precursor and matrix which
the years technology has evolved, coming up with better manufac- affects the loading capability of the structure [6].
turing techniques for fabricating composite structures based on Previous research shows that CFRPS have high modulus (200–
their applications as to improve productivity, reliability and reduce 500 GPa) and high strength (1–2 GPa) [1]. The highest tensile
manufacturing costs [3]. To date, carbon fiber reinforced polymers strength is achieved on unidirectional (UD) carbon fibers when
(CFRPs) are still relatively expensive when compared to conven- the force is applied in the direction of the fibers. For the same vol-
tional materials such as glass fiber reinforced polymers (GFRP) ume of fibers per unit volume of the composite, it cannot match
the stiffness and strength of unidirectional composites [7,8]. Fur-
CFRP, Carbon Fiber Reinforced Polymer; VARI, Vacuum Assited Resin Infusion; thermore, literature reveals that the compressive strength of unidi-
CF, Carbon Fiber; GFRP, Glass Fiber Reinforced Polyers; UD-CFRPS, Unidirectional rectional carbon fiber (UD CFRPs) is a major worry in the design of
Carbon Fiber Reinforced Polymer; TRPCs, Textile Reinfirced Polymer Composites; structural application because of poor performance which is
NDT, Non-Distructive Test; VGCF, Vapor Grown Carbon Fibers; FEA, Finite Element roughly half of its tensile strength [8]. A few compressive failures
Analysis; HDPE, High Density Polyethylene.
⇑ Corresponding author. such as delamination, parting and fiber buckling, have been seen
E-mail address: pitsotaele@gmail.com (P. Taele). in different material frameworks of UD CFRPs [9]. Studies shows

https://doi.org/10.1016/j.matpr.2021.11.436
2214-7853/Copyright Ó 2021 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the International Conference on Applied Research and Engineering 2021
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

that textile-reinforced polymer composites (TRPCs) are favorable to conduct the experiments. The results obtained from experi-
in terms of flexural properties compared to unidirectional fiber- ments will be compared with data available from literature review.
reinforced polymer composites (UD-FRPCs) based on distinct The data published from this work will be beneficial for computa-
mechanical characteristics of textile structures. A study conducted tion fields and simulation modelling to define the properties (Flex-
by Yılmaz and Khan [10] revealed that 3D TRPCs exhibit advanced ural strength) of UD CFRP without lab experiments.
properties by overcoming some limitations such as delamination
and debonding that occur in conventional 2D laminated compos-
ites under flexural loading. CFRPs are known to suffer from delam- 2. Materials and methods
ination and the dangers associated with it are matrix cracks which
join up to produce a fracture surface and shedding load without The specimens were manufactured through vacuum assisted
fibers breaking. It is quite difficult to spot/detect delamination resin infusion (VARI) technique. The standard used to prepare
with visual inspection. Nayak et al. [11] conducted a non- and manufacture the specimens was the ASTM D5687/D5687M-
destructive test (NDT) aimed at detecting small delamination in 20. Even though the general guidelines on this standard do not
CFRPs as to help detect and avoid unanticipated failures which address the processing techniques for preparing specimens
could not be seen due to imperfections. Ultrasonic and X-Ray through resin-transfer molding (RTM) similar to VARI technique,
Diffraction (XRD) tests were found to be more effective in detecting all the work was done in compliance with the standard. The dry
small delamination than X-Ray radiography and thermography Ruban Carbone 50 mm Unidirectional tape was used as the rein-
tests. In the vibration test, it was observed that the natural fre- forcement in the manufacturing of specimens and its properties
quency decreases and the rate of damping increases in delami- are shown in Table 1.
nated carbon fiber composite beam as compared to composite The carbon fiber (CF) tape was cut to length, laid up to the
beam without delamination. desired laminate layers (thickness) and aligned on top of a flat
In an attempt to simplify the prediction of loading capabilities 50 mm thick Polyvinyl acetate (P.V.A) plate after it was thoroughly
of CFRPs, researchers have been heavily invested in obtaining a cleaned followed by applying five coats of ram wax were on inter-
numerical model to define properties of CFRPs using finite element vals of 10 min in-between coats. The minimum tool plate thickness
analysis (FEA) simulation [12]. FEA methods give more accurate as prescribed by ASTM D5687/D5687M-20 standard is 6 mm or
results in a short period of time without the need for physical test- 3 mm with a surface tolerance of 0.05 mm depending on the mate-
ing. However, the numerical model does not always represent the rial used as a tool plate [18]. The specimens were prepared at room
actual physical and geometric aspects of the structure constructed. temperature after the room was cleaned and dusted to avoid con-
Often with composite simulations, there are huge discrepancies tamination. Figs. 1 and 2 shows the mould tool, geometry, and
between the results of the constructed FEA model and the results dimensions of the specimens, respectively.
of modal testing of the actual structure [13]. By minimizing the Peel ply Nylon 085 g/LTX/PS 1000MM from AMT Composites
uncertainties between the FEA model and the experimental input was chosen because it does not contain solvents which might react
forces or the output response, structural parameters such as stiff- with the resin hence tempering the properties of the end result. A
ness and mass can be easily obtained. peel ply was cut to length and carefully placed on top of the fibers
To improve the correlation between numerical predictions and such it covered all the specimens on top the tool and that the fibers
experimentally measured data, the developed FE model needs to remained aligned. A mesh was cut the same way as the peel ply. It
be updated or modified to reduce discrepancies. By minimizing was then carefully placed on top of the peel ply such that it covered
the uncertainties between the FEA model and the experimental all the specimens without mis-aligning the carbon fiber layers. VI1
input forces or the output response, structural properties such as Infusion mesh from Richmond Aerovac was chosen because of its
such as stiffness and mass can be easily obtained. Recent work con- desired properties. The mesh is designed for use in vacuum
ducted using FEA to obtain vibration characteristics and strain assisted resin infusion processes as well as in autoclave. Con-
energies of CFRPs from analytical study reveals that the natural structed from high density polyethylene (HDPE) and manufactured
fiber composites have better damping properties than epoxy mate- in a 3D knitted pattern to allow controlled resin flow across the
rial [14]. A study conducted by Bagha and Bahl [15] revealed that laminate, ideal for use in areas of complex geometry tooling. The
storage modulus of CFRP structure increases with respect to aspect mesh weighs 150 g/m2 with maximum operating temperature of
ratio. However, the loss factor showed a reduction in its perfor- 80 °C.
mance. Also, the loss modulus increases as the fiber aspect ratio High temperature sealant tape C2180 from AMT Composites
increases but sudden drop was found at the higher values of the was applied around the specimens to create a vacuum integral seal
aspect ratio. FEA study on vapor grown carbon fibers (VGCF) to between the vacuum bagging film and the tool surface. Inlet and
predict its mechanical properties for different loadings was con- outlet rubber pipes of the same inner diameter (i.e., 5 mm) were
ducted by Bagha and Bahl [16]. From this study it was observed place on either side of the mould tool to allow the resin to flow
that there is considerable improvement in longitudinal modulus across the entire surface of the tool. The film used to contain the
of the VGCF/pp nanocomposite with small additional amount of desired pressure to be applied on the tool was VACFILMTM 450 V
nanofiber. 50 Mic. This film is suitable for high pressure autoclave cures up
An experimental study on carbon fiber thickness and layer to temperature of 170 °C (338°F). It is also ideal for use with med-
thickness of depositing material in fused deposition modeling
was conducted by Sharma et al. [17]. The results obtained from this
study revealed that increasing thickness of carbon fiber (CF) in Table 1
Properties of the carbon fiber tape.
polylactic acid (PLA) based composite increases the tensile
strength over that of the pure PLA specimen. Moreover, the tensile Type 3558/343/50
strength was found to decrease with an increase in layer thickness Thickness (mm) 0.6
Width (mm) 50
of depositing material in CF-PLA composite. With guidance from Weight (gr/m2) 343
previous research conducted, this paper reports on the perfor- Weave Toile or Cloth
mance of UD-CFRP subjected to bending stress. The type of carbon Contexture WARP: 8 yarns/cm carbon 6 k
fiber, matrix and fabrication method is clearly outlined to provide a WARP: 8 yarns/cm glass EC 7/34 TD
WEFT: 3 double yarn/cm glass EC 7/34 TD
consolidated conclusion. A 200 KN Zwick testing machine was used
2097
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

Fig. 1. (a) Specimen mould tool (50 mm PVC plate), (b) Release agent wax.

Fig. 2. (a) Unidirectional carbon fiber tape cut to length, (b) specimen geometry and dimensions.

ium temperature epoxy and phenolic prepregs. Attention needs to 20 slow hardener A126. The two were mixed and prepared in a
be given on the level of moisture content at the time of use to paper cup at room temperature.
ensure maximum performance. The higher the moisture content At this stage, full vacuum pressure was applied to the system as
the more flexible they become. shown in Fig. 3b. The inlet tube was inserted into the resin cup,
The vacuum film was cut to shape and placed on top of the sea- there after the clamp was released to let the resin flow into the sys-
lant tape such that it was airtight. The inlet pipe was clamped to tem. The applied vacuum pressure dropped from 100 KPa to 70 KPa
close off the mould such that no air was allowed to go through when the resin was injected to dry fibers. Fig. 3(c) above shows the
the mould. The outlet tube was then connected to the vacuum suc- pressure gauge during the infusion of the resin into the system. The
tion pump, there after a full pressure was applied. The mould was reason for the drop in pressure is that there was some air in the
left for 5 min to check if there was a leak somewhere in the system. section of inlet pipe where the clamp was not placed. Fig. 4 sum-
The vacuum was placed 1 m away from the mould (i.e., lay-up site) marizes the process in phases on high level.
to comply with ASTM D5687/D5687M-20 standard maximum dis- The inlet and outlet rubber pipes were clamped just before the
tance of 45 m. Fig. 3 outlines the process. The resin, often time resin inside the paper cup ran out. This was done to avoid air com-
referred to as a matrix, is the material that binds the fibers together ing through to system which may introduce bubbles and affect the
and transfers the load to the fibers. It provides rigidity and shape to strength of the final product. With clampes applied, the system
the structure. The matrix isolates the fibers so that individual fibers was left to cure at room temperature for 24hrs before the speci-
can act separately [19]. mens were released from the mould. Specimens were inspected
The resin used was an epoxy resin Prime 20 LV A147 mixed by visually to identify areas where the resin might not have pene-
weight ratio of 100: 20 resin-to-hardener. The hardener was Prime trated. If the resin is not properly injected into the fibers, that’s a

Fig. 3. (a) Closed inlet pipe, (b) Pressure gauge before the resin was injected, (c) Pressure gauge during resin flow.

2098
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

Fig. 4. (a) CF tape cut-out and layup, (b) Sealant tape and peel ply applied, (c) Resin Infusion. (inlet to outlet)

Fig. 5. (a) Top surface finish (side facing on applied peel ply and mesh), (b) Bottom surface finish (Side facing on top of the tool).

defect which cannot be corrected [20,21]. The end results of the son’s ratio calculations necessary to develop, integrate and imple-
specimens are shown in Fig. 5. ment predictive computational models for CFRPs composites
It is good to note that the geometry of the specimens was kept linking the materials design, moulding process and final perfor-
the same with varying thicknesses. Four specimens of each cate- mance requirements for optimal design and manufacturing of CFRP
gory with 4, 6 and 8 laminate layers were manufactured to form structures [22]. The machine used for the three-point bending tests
a total number of 12 specimens for the three-point bending test. was a servo controlled Universal Testing Machine known as Zwick
The ASTM D7264/D7264M-15 standard recommends at least five 200 kN, with a loading nose and supports of 5 mm radius. The
specimens per test condition unless valid results can be gained crosshead speed of 0.5 mm/min was used in compliance with the
using fewer specimens, such as in the case of a designed experi- ASTM D7264/D7264M-15. A minimum of four tests were con-
ment [21]. The manufactured specimens were then prepared for ducted for each test condition at room temperature. Stacking
the three-point bending test. The aim being to determine the elas- sequence, average values and variation were obtained, analyzed
tic modulus (i.e., flexural modulus), stress–strain behavior, and and recorded. Fig. 6 shows the machine setup and configuration
failure limits of CFRPs manufactured through the VARI technique. before and during the testing.
The standard used to conduct the experiment is ASTM D7264/
D7264M–15. The span-to-thickness ratio selected was 40:1. To 3. Results and discussions
achieve the ratio of 40:1 span-to-thickness ratio for each set of
specimens, the support span length (L) was calculated. Table 2 This section discusses the experimental results after the various
summarizes the span length of each set of test specimen based composite samples were fractured under three-point bending test
on the measured thickness machine. The load at which the complete fracture of the specimen
It is good to note that the radius of the supports was taken into occurred has been accepted as the breakage load. The maximum
consideration as prescribed in the ASTM D7264/D7264M–15 test stress at the outer surface of a flexure test specimen corresponds
standard. The tests were conducted at a strain rate of 0.01 s 1. to the peak applied force prior to flexural failure. The flexural
The reason for this choice is that this specific strain rate can be tied strength is the ability to resist deformation under load for material.
for testing the flexural strength and tensile strength to the same The material deforms significantly but does not break, the load at
loading rate condition. Also, specimen thicknesses and fiber orien- yield is typically measured at 5% deformation/strain of the outer
tation tested are consistent across loading rates for comparison surface and it is reported as the flexural or yield strength [23].
purposes. The strain results are used for chord modulus and Pois- The test beam specimen is under compressive stress at the concave

Table 2
Support span-to-thickness ratios of test specimens.

Specimen Average Thickness (mm) Width (mm) Support span length (mm) Ratio
8-Layers 2.8 50 110 39.29
6-Layers 2 50 80 40
4 Layers 1.2 50 50 41.66

2099
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

Fig. 6. (a) Machine and specimen setup before testing; (b) Specimen deflection during testing.

(inner) surface and tensile stress (outer) at the convex surface as Fig. 9 display an 8-layer specimen with an average thickness of
stated in the D7264/D7264M-15 document. 2.8 mm. Keeping the aspect ratio as close as possible to 40:1, the
The graphical representation of stress–strain data is shown average yield strength was 756 MPa at strain of 0.009 (10.95 mm
from Fig. 7 up to Fig. 9, keeping the aspect ratio as close as possible deflection).
to 40: 1. Fig. 7, represent the stress vs strain curve results for a It can be observed from the results shown from Figs. 7–9 that
4-layer carbon fiber specimen with an average thickness of CFRPs are quite stiff, measuring an average mid-span strain of
1.2 mm. The results showed an average yield strength of 0.014 for all tested specimens. The stress vs strain curves show that
1175 MPa at strain of0.017 (deflection of 9.8 mm). CFRP are not brittle since the failure occur in phases due to delam-
Fig. 8 display a 6-layer specimen with an average thickness of ination. This goes to prove that fibers play an important role
2 mm. Keeping the aspect ratio as close as possible to 40:1, the towards the structural integrity (i.e., mechanical strength) of the
average yield strength was 898 MPa at strain of 0.014 (10.6 mm final product [24]. It is worth noting that the results shown are
deflection). taken from the point of failure at which the first one or two layers

Fig. 7. Stress vs strain curve for 4-layers specimens.

2100
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

Fig 8. Stress vs strain curve for 6-layers specimens.

Fig. 9. Stress vs strain curve for 8-layers specimens.

failed. This means that the specimens could still endure added were calculated from the trend line, so these results are evident
stress and strain on the remaining intact layers had the test that as the thickness increases, the maximum force absorbed by
machine not stopped [8]. Peak force vs laminate layers graph is the structure is increased [17]. Fig. 11 shows the flexural strength
plotted in Fig. 10 to describe the correlation between peak force vs flexural modulus graph in comparison with the laminate layers.
and laminate layers of CFRPs. This can also be translated into aver- The aim of this graph is to describe impact of laminate layers (i.e.,
age peak force vs thickness of CFRP structure. thickness) of specimen on flexural strength and flexural modulus.
From the graph in Fig. 10, the peak force and laminate layers of The flexural strength vs modulus graph shown in Fig. 11 shows
CFRP have a linear relationship. A straight trend line was plotted that the 4-layer specimens have greater strength and stiffness (i.e.,
and can be describe by y = mx + c. where m = 178.29, c = 155.28 Flexural modulus) than the 6- and 8-layer specimens respectively.

2101
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

Fig. 10. Influence of laminate layers (i.e., thickness) of specimen on peak force.

Fig. 11. Influence of laminate layers (i.e., thickness) of specimen on flexural strength and flexural modulus.

Looking at the flexural modulus, the 4-layer specimens were the 4. Conclusion
stiffest, measuring an average flexural modulus of 104GPa fol-
lowed by 8-layers measuring 91GPa and 6-layers measuring CFRPs are taking the world by storm due to their desirable prop-
81GPa. The flexural modulus of a material is a physical property erties. In this paper, the flexural behavior of UD-CFRPs manufac-
that indicates the flexibility of that material. Mechanically, it is tured through the VARI technique was investigated
the relationship between stress and strain in flexural deformation. experimentally. The findings of the study can be summarized as
The higher the flexural modulus, the stiffer the material [25,26]. follows:

2102
P. Taele, T. Masango and V. Msomi Materials Today: Proceedings 56 (2022) 2096–2103

 The stress vs strain curve graphs show that CFRPs are quite stiff References
but not brittlet. It was observed that the cohesive layer behaves
almost like the matrix for the linear elastic deformation prior to [1] P.W.R. Beaumont, C. Soutis, A. Hodzic, The Structural Integrity of Carbon Fiber
Composites, 1 ed., Springer, Manchester, 2017.
damage. This agrees with the study conducted by Bahland [2] J. Lu, W. Li, H. Kang, L. Feng, J. Xu, R. Liu, Pol. Test. 81 (2020) 106267.
Bagha [15]. [3] A.Y. Chen, S. Baehr, A. Turner, Z. Zhang, G.X. Gu, Int. J Lightweight Mater.
 The study reveals that mean flexural strength and strain Manuf. 4 (2021) 468–479.
[4] C. Dong, Comp. Part C 3 (2020) 100047.
increases with decrease in laminate layer thickness, however [5] Y. Kim, Y. Kim, F. Libonati, S. Ryu, Comp. Part B 167 (2019) 736–745.
the dependence of laminate layers (i.e., thickness) on flexural [6] S. Chocron, A.J. Carpenter, N.L. Scott, R.P. Bigger, K. Warren, Int. J. Impact Eng.
modulus cannot be described by a specific trendline. 131 (2019) 39–56.
[7] A. Kotzakolios, Blast Response of Aircraft Structures, Appl. Mech. Vibr. Lab.
 It was observed that the relationship between the peak force
(2011).
and thickness of a CFRP structure is linear (i.e., Peak force [8] N. Hashim, D.L.A. Majid, D.M. Baitab, N. Yadris, R. Zahari, Ency. Mater. Comp. 1
increases with thickness). (2019) 766–773.
 The laminate layer interface behavior needs to be defined in a [9] G.A. Pappas, A. Schlothauer, P. Ermanni, Comp. Sci. Tech. 44 (2021) 108979.
[10] M. Ueda, Y. Tasaki, C. Kawamura, K. Nishida, M. Honda, K. Hattori, T.
more sophisticated way to model the stepwisedelamination Miyananga, T. Sugiyama, Comp. Part C 6 (2021) 100180.
and failure evolution once the complete debonding has [11] N.D. Yilmaz, G.M.A. Khan, Mech. Phys. Test. Biocomp. Fibre-Reinforced Comps.
occurred. Hybrid Comps. (2019) 13–42.
[12] S. Nayak, S. Pattnaik, I. Panigrahi, R. Nayak, S. Parida, Indian J Eng. Mater. Sci.
27 (2020) 789–794.
[13] A.K. Bagha, S. Bahl, Mater. Tod. Proc. 41 (2021) 265–268.
[14] S. Abhishek, B.K. Ashok, S.K. Dinesh, B. Shashi, AIMS Mater. Sci. 8 (2021) 390–
415.
Declaration of Competing Interest [15] M.K. Saini, A.K. Bagha, S. Kumar, S. Bahl, Mater. Today Proc. 42 (2020) 223–
227.
The authors declare that they have no known competing finan- [16] S. Bahl, A.K. Bagha, Mater. Today Proc. 39 (2021) 70–76.
[17] A.K. Bagha, S. Bahl, Mater. Today Proc. 39 (2021) 54–59.
cial interests or personal relationships that could have appeared [18] P. Sharma, A.K. Dhanopia, D. Joshi, Mater. Today Proc. 44 (2021) 4479–4484.
to influence the work reported in this paper. [19] D5687/D5687M-20, Standard Guide for Preparation of Flat Composite Panels
with Processing Guidelines for Specimen Preparation, 2020.
[20] L.K. Cardon, K.J. Ragaert, R. De Santis, A. Gloria, Design and fabrication methods
for biocomposites, in: Biomedical Composites, Elsevier, 2017, pp. 17–36.
Acknowledgements [21] Y.K. Hamidi, C.M. Altan, Compos. Mater II (2) (2018) 95–106.
[22] D7264/D7264M-15, Standard Test Method for Flexural Properties of Polymer
The authors acknowledge the support of this research from the Matrix Composite Materials, 2015.
[23] L.A. Powell, W. Luecke, M. Merzkirch, K. Avery, T. Foecke, SAE. Int. J. Mat.
Cape Peninsula University of Technology. Technical support offered Manuf. (2017) 138–146.
by the technical staff at the AMTL labs. Support from the Depart- [24] D. Zhang, A.M. Waas, C.F. Yen, Int. J. Solid. Struct. 75 (2015) 309–320.
ment of Mechanical Engineering of the University of Cape Town [25] D.K. Jesthi, S. Nayak, S. Saroj, Mater. Today Proc. (2021).
[26] D.K. Rajak, P.H. Wagh, H. Moustabchir, C.I. Pruncu, Forces Mech. 4 (2021)
is also acknowledged. The National Research Foundation (NRF) is 100029.
gratefully acknowledged for funding all the work done to conduct
this paper.

2103