Improved mechanical performance of three-dimensional woven glass/epoxy spacer composites with carbon nanotubes

Abstract

Three-dimensional polymer composites offer various features and design options due to their hollow structure and lightweight. However, to exploit their advantages, it is a must to improve their structural features and mechanical performances including out-plane direction. Although introducing thermoplastic fillers between the plies or multilayered design addresses on this critical issue, the benefits offered by the nanoparticles with superior mechanical properties come a step forward as an another engineering solution. Based on this motivation, the goal of this study is to investigate the impact of multiwalled carbon nanotubes on the mechanical and thermomechanical performances of three-dimensional woven glass/epoxy spacer composites. Therefore, multiwalled carbon nanotubes at various content were introduced into epoxy matrix, and the multiwalled carbon nanotubes–epoxy mixture was infused to three-dimensional woven fabric with the vacuum-assisted resin transfer method. The obtained results indicated that the three-point bending strength and modulus were enhanced up to 25 and 80% for warp direction and enhanced up to 44 and 85% for weft direction with carbon nanotube addition, respectively. Tensile strength developed in the warp direction by 7%, while the strength value in the weft direction did not change. The tensile strain values for warp and weft directions enhanced up to 19 and 12% with carbon nanotube addition, respectively. In addition, thermomechanical analysis has revealed that the glass transition temperature and storage modulus were also improved. Particle dispersion detection with color measurement and scanning electron microscopy analyses revealed the effectiveness of the ultrasonic mixing on the dispersion of carbon nanotubes in the epoxy matrix. The consequences of carbon nanotube addition on microscale morphology were discussed based on the fracture morphologies to nanoscale and microscale toughening mechanisms in the existence of carbon nanotube reinforcement.

Keywords

Characterization glass fiber/epoxy multiwalled carbon nanotube nanocomposite three-dimensional woven spacer composite

Introduction

Fiber-reinforced polymer (FRP) composites have increasingly found usage in automotive, marine, defense, and aerospace industries due to their superior structural properties like lightweight, strong, and multifunctionality compared to the rival metal-based materials. Sandwich composites consist of face sheets and a core where adhesive layers aid to create a sandwich structure. The absence of a physical and mechanical bond between the layers creates an important negativity for them.¹ In addition, delamination formation in sandwich composite is proportional to the adhesive strength between the face sheet and core material. Three-dimensional (3D) woven composites are derived to overcome this drawback in sandwich composites.^2,3 3D spacer woven fabrics are composite reinforced materials whose bottom and top faces are braided to each other with plies aligned in through-thickness direction.⁴ This structure allows forming of channels between the top and bottom plies and creates a lightweight structural material. The weave in the channel direction is called warp and the weave perpendicular to this direction is called weft. 3D composite layers can be produced in various ply thicknesses such as 3, 6, 10, 20 mm, etc. According to the reports that investigated the effect of the ply height on the mechanical properties of 3D woven composites, the general inference is that the increased thickness reduces resistance against compression and bending loadings.^1,5–9 Another critical design feature on the mechanical properties of the 3D woven spacer composite is the weaving direction. It has been reported that the warp direction demonstrated mechanical performance than the weft direction.⁸ When the damage areas were investigated, the detected damage mostly occurred on knitting connections between the plies with bottom and top face sheets.^6,10

In the past, numerous research endeavors focused on improving the composite’s mechanical properties such as strength, rigidity, and toughness by introducing various types of nano-sized particles.^11–19 According to these research studies, mechanical performance of a fiber-reinforced material is primarily influenced by developing an effective fiber–matrix interface, adhesion between the resin–fiber–nanoparticles, and dispersion of the nanoparticles.^19–24 Carbon nanotubes (CNTs) are widely utilized as nanoscale reinforcements due to their superior mechanical properties.^{16,17,20,25–28} On the contrary, CNTs tend to agglomerate due to their high surface areas and aspect ratios which results in impair mechanical properties.^29,30 In the case of CNT-reinforced multi-scale composites, it was declared that CNT addition improves the mechanical properties such as tensile and flexural strength and thermal and thermomechanical properties such as storage modulus and glass transition temperatures.^{17,20–22,31,32} It must be mentioned that the homogenous distribution of nanoparticles in the polymer matrix is an effective strategy to obtain high mechanical and thermomechanical properties. Note that the mechanical performance of multi-scale composites is primarily influenced by obtaining an effective interface between epoxy matrix and nanoparticles without agglomeration.^22,31 However, there are several methods like zeta potential (ζ), dynamic light scattering, transmission electron microscopy (TEM), and color concentration measurement³³ for measuring the particle distribution in the matrix. The color measurement method is one of the facile methods. This method points out the concentration differences of the applied caused by agglomerations or unbalanced particle distributions.³⁴ Because of these prominent features, color measurement has been used in the study for the homogeneity investigation.

In the studies carried out to increase mechanical performance of the 3D woven spacer composites, lamination, filling with thermoplastics (foam, polyurethane, etc.), or hierarchically designed composites were investigated.^2,35–38 These studies examined the structural properties or mechanical performance of 3D composites without particle reinforcement. While there are so many studies that particle reinforcement improves properties, it is remarkable that the studies on the effects of nanoparticles on the structural behavior of 3D composites are limited. In the study, which focused reinforcement of 3D woven spacer composite with nanoparticles, Yildirim et al.³⁴, nano-SiO₂ was incorporated to the 3D woven composite and they declared that tensile and bending mechanical properties and thermal properties were improved. Effects of CNT particles on 3D woven composites should be examined in-depth.

In this regard, the main purpose of this study is to investigate the impact of introducing multiwalled carbon nanotube (MWCNT) reinforcements on mechanical and thermomechanical properties of the 3D woven spacer composites. Therefore, various amounts of MWCNTs were added into the epoxy resin, and multi-scale 3D composites were manufactured by using the partial vacuum infusion method. The impact of MWCNT on the mechanical and thermomechanical properties, tensile test, three-point bending test, thermogravimetric (TG), and dynamical mechanical analysis (DMA) were examined, respectively. Finally, electron microscopy (scanning electron microscopy (SEM)) and surface color measurement have been used in the study for the particle distribution and homogeneity investigation.

Experimental procedure

Materials

The 3D spacer woven fabric was supplied by Parabeam BV from the Netherlands. The woven is produced by using E-glass and has a thickness of 3.3 mm (composite thickness is 3 mm³⁹) and a density of 782 g m⁻². The epoxy resin and curing agent system used as the matrix material were MGS L-160 and MGS H-160, respectively, purchased from Momentive Hexion Inc. The MWCNT used in the study was supplied from Ege Nanotek, Turkey, and technical specifications of the MWCNT particles are shown in Table 1.

Table 1.

Technical specifications of the MWCNT.

Material	Purity, (%)	Density, (true, gcm⁻³)	Specific surface area, (m²)	Length, (μm)	Outer diameter, (nm)	Inner diameter, (nm)
MWCNT	95	2.1	>40	10–20	50	5–15

MWCNT: multiwalled carbon nanotube.

Composite manufacturing

At first step in the manufacturing process, MWCNTs were added to epoxy with various concentrations and then homogenized with ultrasonic mixing process (Vibra-Cell VCX-750). The earlier studies declared that the ultrasonication is an effective way to obtain maximum homogeneity.^{11,15,40–45} Therefore, the mixture of MWCNT and epoxy was stirred using ultrasonic mixer for 40 min at 35% frequency amplitude value. As heat occurs during ultrasonication, the system was cooled down to the 35°C using an iced water bath to control viscosity and to avoid heat-induced chemical decomposition of the resin for all samples. Following ultrasonication process, the curing agent was introduced to the mixture at a 4:1 ratio by weight and mechanically mixed for 3 min. Meanwhile, an aluminum mold, in which vacuum layers are prepared on it (as shown in Figure 1), was heated to 50°C by an external heater just before the infusion process. Thus, according to our previous experiences, it has been ensured to decrease the resin viscosity and improve resin impregnation efficiency.^18,26,34 MWCNT-reinforced epoxy was impregnated to the dry 3D woven by the vacuum-assisted infusion method to achieve fully impregnated fabrics and avoid forming a resin pool.^1,6 After the resin infusion was completed (it takes 8–9 min on average for 400 × 400 mm square plates) and the fabric was completely wetted, the vacuum bag was removed before the curing to provide the crushed plies of the fabric spring back and allowing the material to regain its 3D spacer form. If the vacuum bag was not opened, the composite changes to the 2D structure because of the vacuum effect on the fiber sheet. Finally, the MWCNT-reinforced 3D woven composites were cured for 18 h at room temperature, and then post-cured for 2 h at 80°C in an oven. Following the same steps, 0.3, 0.5, 1, and 2 wt.% MWCNT-reinforced epoxy/3D woven composite materials were remanufactured. Also, neat epoxy/3D woven composite control group were manufactured by using the same process. The manufacturing process is given in Figure 1.³⁴

Figure 1.

Manufacturing process of multiwalled carbon nanotube-reinforced three-dimensional woven composite material.

Specimen preparation and experiments

After the manufacturing process was completed, first, the densities of the neat and MWCNT-reinforced 3D woven composite materials were calculated. The densities were measured according to the Archimedes principle (ASTM D-792) by using Shimadzu AUX320 precision balance. The obtained densities were used to find the specific mechanical properties and void ratios of the materials. Also, dimensional stability measurements were carried out using with Mitutoyo-500 digital caliper on manufactured composites. All the mechanical test specimens with warp and weft directions were prepared from composite plates to examine the effect of fiber orientation on the mechanical performance. Three-point bending tests were carried out on 140 × 30 × 3 mm specimens according to the ASTM D-7264 standard using a UTEST-10 testing machine at ambient temperature, with a distance of 80 mm between the supports and at a speed of 3 mm min⁻¹.⁶ While performing the tensile test of 3D-type hollow materials, the drawing jaws of the test machine may damage the material, and this leads to the failure of testing procedure. To avoid this problem, the tensile specimens were sized to 250 × 40 × 3 mm at first, and then the width of gauge lengths were reduced to 25 mm, according to the standard size. The edges of specimens were filled with liquid plastic paste, with a syringe injection and solidified.³⁴ The liquid plastic paste is a material that becomes solid while it is liquid during application purchased from Permolit Paint Inc. The tensile tests were performed at a drawing speed of 0.5 mm min⁻¹ according to the ASTM D-3039 standard using a ZwickRoell Z100 testing machine at ambient temperature.^7,8 The three-point bending and tensile testing application and test specimens have been given in Figure 2.

Figure 2.

Mechanical test performing; (a) three-pointbending test, (b) tensile test.

Microstructure analysis was performed on damaged region of the bending specimens using FEI Nova NanoSEM 650 SEM device, and Insize ISM-PM200SB optical microscope was used for macro imaging. The DMA were utilized for studying the thermomechanical characteristics of the 3D woven material (Metravib DMA 50). Thermal properties of the 3D woven composite were examined using thermogravimetric analysis (TGA) from 0 to 800°C at a heating rate of 1°C min⁻¹ in air atmosphere using Setaram Labsys Evo TGA device.

There are several methods like zeta potential (ζ), dynamic light scattering, TEM, and color concentration measurement³³ for measuring the particle distribution in composite materials. Emady et al.⁴⁶ investigated the color change in composite surfaces with the addition of particle fillers. According to their hypothesis, the intensity of light reflected from the surface was proportional to the concentration of the material that changed the color in the sample and that was determined by calculation.⁴⁷ The main objective of the applied process is to detect the color hue differences resulting from reinforcement particles. The color change factor, ΔE, is measured by color measurement device and it gives the total chromatic aberration. The obtained results were listed and controlled to see whether they were within the range or not of a database derived from the measured points of surface. The measured parameters and acceptance ranges of the results are given in Table 2.⁴⁸

Table 2.

Comparative color measurement principles.

Color measurement parameter meanings
ΔE	Mean differences between measured materials
ΔL (+)	Partial white	ΔL (−)	Black
Δa (+)	Partial red evolution	Δa (−)	Partial green evolution
Δb (+)	Partial yellow evolution	Δb (−)	Partial blue evolution
Color compare range		Color compare meaning
0–0.25 ΔE		Perfect match
0.25–0.5 ΔE		Can be considered with small difference
0.5–1 ΔE		Can be considered with little difference
1–2 ΔE		Can be considered on some sectors
2–4 ΔE		Can be considered on special applications
4 ΔE		Cannot be considered

It is known that the carbon is used as a black color source in the industry. The used reinforcement material, MWCNTs, in this study causes a color change on composite from transparent to black. Because the black color tone depends to the intensity of carbon, the dispersion stability of MWCNT was investigated by color changes measured using a Mathai Color Striker device on the surface of the 3D woven composite according to the ASTM D-2244 standard. The color measurement process is given in Figure 3.

Figure 3.

Color measurement of multi-walled carbon nanotube reinforced three-dimensional woven composite.

Results and discussions

Bending test results

Figure 4(a) represents bending test results of the MWCNT-reinforced multi-scale 3D composites. As seen, the flexural strengths on warp direction and weft direction for the neat epoxy 3D woven composite are measured as 46.6 MPa and 33.6 MPa, respectively. It is clear that the MWCNT-reinforced 3D woven composites exhibited an increased flexural strength compared to the neat epoxy 3D woven composites. The obtained improvement on strength calculated approximately as 25% and 44% on warp and weft directions, respectively, compared to the neat epoxy 3D woven composite. For the MWCNT-reinforced 3D woven composite materials, the highest flexural strengths on warp direction and weft direction are measured as 58.4 MPa and 48.6 MPa, respectively, with 1wt.% MWCNT addition. According to the previous studies, the nano reinforcement/matrix ratio has an important role on the strength properties of composite materials.^34,49 Composite materials have a threshold value of this ratio. Maximum strength properties are obtained at certain proportional values, which can also be referred to as the saturation point. This saturation point can be as low as 0.1, 0.3, or 0.5wt.% with the filtration effect for multilayered glass or carbon FRP composites. Since 3D woven fibers are produced in one layer, they have higher permeability and lower filtering effect. This shows that the saturation level of the reinforcement material/resin ratio for the 3D woven composites is reached at a 1% MWCNT reinforcement level by weight. In addition, we confidently believe that 1wt.% MWCNT reinforcement ratio leads to the creation of ideal interface between the resin, fiber, and MWCNT. Thus, it improves the strength properties by limiting the growth of nanopores, as well as the propagation of cracks, thanks to the mechanical locking formed by the particles with the epoxy.^42,50–53 It can be said that the multiwall tubular morphology and high specific surface area of the MWCNTs increase the contact areas and help the good adhesion formation between the reinforcement material and resin.³⁴

Figure 4.

Bending performance of the three-dimensional -woven composite, (a) bending strength (b) bending modulus of elasticity (c) deflection.

In addition, the increase in the bending modulus was another evidence of improved load distribution mechanism by good dispersion of MWCNT in the epoxy resin. The obtained ratios of increase for warp and weft directions were calculated as 80 and 85% compared to the neat epoxy 3D woven composite, respectively. The elasticity modulus values were calculated on warp direction as 3 GPa and weft direction as 2 GPa for the neat epoxy 3D woven composite. The highest elasticity modulus on warp direction and weft direction were measured as 5.4 GPa and 3.7 GPa for 1wt.% MWCNT-reinforced 3D woven composite, respectively. Results given in Figure 4(b) were also statistically evaluated and expressed with error bars. In the analyses in which neat epoxy 3D woven composites were taken as a reference, it was seen that the results obtained from reinforced 3D woven composites that exhibit improvements in mechanical properties were in the 95% confidence interval.

There was a decrease measured in the deflection properties of the 3D woven composites at all reinforcement rates of MWCNT. While the deflection values of neat epoxy 3D woven composites were measured as 5.5 mm and 5.9 mm for the warp and weft direction, respectively, the highest decreases were measured as 3.8 mm in the warp direction and 4.4 mm in the weft direction at 2wt.% MWCNT reinforcement level. The changes in deflection with reinforcement rates of MWCNT particles are shown in Figure 4(c).

Tensile test results

The obtained tensile stress–strain curves of the MWCNT-reinforced and neat epoxy 3D woven composites are given in Figure 5. Typical fiber-reinforced polymer composite curves with brittle fracture behavior were obtained. According to the tensile tests, an improvement was obtained in the maximum tensile properties on warp direction. While the neat epoxy 3D woven composite represents a tensile strength of 55.2 MPa in the warp direction, it was improved as 7% and reached 59.3 MPa with the addition of 1wt.% MWCNTs. However, the tensile strength on weft direction remained almost equal with the neat 3D composite. The obtained strength results are given in Figure 6(a). The reason for the enhancement of the mechanical performance is attributed to be the strong interface developed between the particle and the matrix material. It is thought that this robust interfacial formation depends on the very high specific surface area of the MWCNT and the efficiency of the ultrasonic homogeneous distribution of MWCNTs in the matrix. The tensile curve clearly shows, in Figure 5(b), the improvement in strain values is about 12% on weft direction. If we compare the tensile strength values according to the fiber orientation, it is believed that the reason is when the resin flows with the vacuum effect on weft direction (which is positioned at 90° angle to the yarn channels and resin flow direction) MWCNT-reinforced resins cannot progress easily and be exposed to filtration.⁵⁴ In addition, it is thought that every weft weaving lines may have acted as a dam accumulating particles.

Figure 5.

The stress–strain curves of the multi-walled carbon nanotube reinforced and neat epoxy three-dimensional woven composites, (a) warp direction, (b) weft direction.

Figure 6.

Tensile test results, (a) tensile strength, (b) tensile strain.

The strain values have been positively affected with the addition of MWCNT reinforcement. As can be seen in Figure 6(b), when compared with the neat epoxy 3D woven composite, the improvement ratio of the strain for the warp and weft directions are 19% and 12%, respectively. While the strain values are calculated as 2.1% and 2.09% for warp and weft directions, respectively, the highest strain values are calculated as 2.5% and 2.36% for both warp and weft directions with the addition of 1wt.% MWCNT reinforcement, respectively. Similar to the bending test results, it is apparent that the saturation point of the MWCNTs is 1% by weight for the 3D woven fibers. The obtained improvement in tensile properties in warp direction is thought to be associated with the good wetting and the formation of an effective interface between the nanoparticles and the epoxy resin.^42,50–53 On the contrary, the occurred filtration effect, re-agglomeration, and knotting on manufacturing process by the direction of the weaving lines in the weft-directional samples were the cause of the undeveloped strength in this direction.⁵⁴ But this shows a positive effect on the elongation. The improvement ratios of the bending properties are much higher than the tensile properties. While an agglomeration does not adversely affect bending experiments because of the load type, it may cause structural weakness because it disrupts matrix integrity in tensile tests.

Specific mechanical properties

Specific strength (σ_s) and specific modulus of elasticity (E_s) are calculated using the following equations (1) and 2.⁵⁵

σ_{s} = σ / ρ

(1)

E_{s} = E / ρ

(2)

where σ_s is the specific strength (MPa/(mg/m³)), σ is the strength (MPa), ρ is the density (mg/m³), E_s is the specific modulus of elasticity (GPa/(mg/m³)), and the E is the modulus of elasticity (GPa) values of the material. Specific strength and specific modulus of elasticity enable to compare the resultant 3D composites with various structural materials. Densities and the calculated specific properties of the 3D composites are presented in Table 3.

Table 3.

The mechanical properties of three-dimensional woven composites.

Properties	Neat		0.3% MWCNT reinforced		0.5% MWCNT reinforced		1% MWCNT reinforced		2% MWCNT reinforced
Density (mg/m³)	0.001578		0.001608		0.001604		0.001585		0.001610
Direction	Warp	Weft	Warp	Weft	Warp	Weft	Warp	Weft	Warp	Weft
Bending strength (MPa)	46.6	33.6	51	36.6	55.5	36	58.4	48.6	55.3	42.7
Specific bending strength (MPa/(mg/m³))	29.531	21.292	31.716	22.761	34.600	22.443	36.845	30.662	34.347	26.521
Elasticity modulus (GPa)	3	2	4.6	2.7	5.2	2.7	5.4	3.7	5.2	3.5
Specific elasticity modulus (GPa/(mg/m³))	1.901	1.267	2.860	1.679	3.241	1.683	3.406	2.334	3.229	2.173
Tensile strength (MPa)	55.2	51.6	54.5	50.7	58	49	59.3	51	52.6	51.3
Specific tensile strength (MPa/(mg/m³))	34.980	32.699	33.893	31.529	36.159	30.548	37.413	32.176	32.670	31.863

MWCNT: multiwalled carbon nanotube.

The obtained single layer 3D spacer composite consists of a space between the top and bottom face sheets unlike the conventional FRP composites. These structural properties make it an excellent alternative to foams, sandwich cores, and hollow materials rather than laminated composites. Ashby compared engineering materials according to their specific modulus of elasticity (E/ρ) against specific strength (σ/ρ) as given in Figure 7. As seen, 3D composites can be positioned on the foam materials and close to the polymers. Also, MWCNT-reinforced 3D composites can compete with some metals such as lead alloys. In addition, although the calculated values constitute a comparison reference, it is well known that various parameters affect strength and density such as matrix properties, reinforcement, yarn type, manufacturing method, fiber volume fraction, curing conditions, and morphology. We confidently believe that it is possible to improve the obtained mechanical performance by tailoring the parameters given above.

Figure 7.

Comparisons of engineering materials based on their specific modulus of elasticity (E/ρ) against specific strength (σ/ρ).⁵⁵

According to the results, the density slightly increased with the addition of MWCNTs. On the other hand, drastic improvements of the specific strength and specific modulus for 3D multi-scale composites are obtained compared to the neat samples. The improvement rates on specific bending strength are calculated as 24.7 and 43.6% on warp and weft directions, respectively, for 1wt.% MWCNT-reinforced 3D multi-scale composite. Improvement on the specific modulus of elasticity is 79 and 84% on warp and weft directions, respectively. When the tensile properties are compared, the obtained improvement on specific tensile strength is calculated as 6.8% on warp direction for 1wt.% MWCNT-reinforced 3D composite. As stated in the previous chapter, the specific tensile strength in the weft direction is not affected with the MWCNT modification, but a 12% improvement in strain values is obtained. The results showed that all specific strength and specific modulus values constitute an essential reference for comparison with conventional materials.

Dimensional stability

The dimensional stability of the material is crucial for precision industries such as aerospace and defense.⁵⁶ Moreover, dimensional values are directly related to the matrix material, manufacturing method, trend of anisotropy, and the relationship between the matrix and the reinforcement material. To evaluate the effect of manufacturing method, various 3D composite plates were also manufactured by hand layup method and the results compared with the vacuum infusion laminates are given in Figure 8. The samples produced with hand lay-up presents flatness problem as shown in Figure 8. The fact is that the epoxy resin is not evenly distributed by hand and creates resin-rich regions on the fabric surfaces. These regions cause undesirable residual stresses during curing and to deteriorate the stability. According to the thickness measurement, the average thickness of the neat materials manufactured by the vacuum infusion method is 3.05 mm, while the thickness of the samples manufactured by hand lay-up method varies between 3.10 and 3.30 mm. The reason for this difference is not material thickening but surface fluctuations.

Figure 8.

Comparison views of three-dimensional woven composite manufactured by hand lay-up and vacuum infusion method.

As part of our study, the impact of MWCNT introduction on the dimensional stability was also determined. According to the examinations, some differences were obtained in the thickness direction and given in Table 4. The thicknesses of the material increase slightly when the MWCNT amount increases. The obtained results can be considered as in tolerance range.

Table 4.

The thickness measurement results of the three-dimensional woven composites.

Material	Dry fiber thickness (mm)	Expected composite thickness (mm)	Measured composite thickness (mm)	Difference (%)
Neat	3.3	3	3.057	+1.9
0.3% MWCNT reinforced			3.051	+1.7
0.5% MWCNT reinforced			3.104	+3.46
1% MWCNT reinforced			3.110	+3.66
2% MWCNT reinforced			3.152	+5.06

MWCNT: multiwalled carbon nanotube.

Fiber and void volume fraction

The fiber volume fraction (v_f) can be determined experimentally by weighing the composite before and after the test when the resin is removed by ignition loss process according to ASTM D-2584 and can be theoretically calculated by using the following equation (3).⁵⁷

v_{f} = \frac{w_{f} \cdot ρ_{c}}{w_{c} \cdot ρ_{f}}

(3)

where w_f is the weight of the fiber (g), w_c is the weight of the composite (g), ρ_c is the density of the composite (g/cm³), and ρ_f is the density of the fiber (g/cm³). For determining the fiber volume fraction, the density of MWCNTs was added to the glass fiber mass with the following conversion, equation (4)

v_{MWCNT} = \frac{w_{MWCNT} \cdot ρ_{c}}{w_{c} \cdot ρ_{MWCNT}}

(4)

where v_MWCNT is the MWCNT volume fraction, w_MWCNT is the weight of the used MWCNT (g), and ρ_MWCNT is the density of the MWCNT (g/cm³).

Abdalla et al.⁵⁸ reported that in many cases the ignition loss process has shown some limitations when particles (such as MWCNT, graphene, or ceramics) have been added to the composite material. The particles can also burn partially or the resin in them may remain unburned. However, there has been no standard application to separate particles, resin, and fibers from each other of FRP composites.^58,59 Since the TGA process gives the main mass loss of the composite by burning the resin, the v_f was measured experimentally using TGA analysis in our study.³⁴ Fiber volume fraction studies were carried out on the neat 3D composites and 1wt.% of MWCNT-reinforced 3D composites which give the highest mechanical performance.

According to equation (3), v_f results are found as 0.355 (35.5%) for neat and 0.358 (35.8%) for the multi-scale 3D composites. The mass loss after the TGA analysis is represented in Figure 10. The weight ratio of the residues to the composite mass was 54.4% for the neat and 54.2% for multi-scale composites. The experimental v_f results are found as 0.337 (33.7%) for the neat and 0.341 (34.1%) for the multi-scale 3D composites. When the v_f results are compared, a slight increase is obtained. For theoretical calculation, it is assumed that the composite laminate contains no voids. However, during composite manufacturing process, voids form inter- and intra-tow regions.⁶⁰ The difference between the theoretical and experimental v_f values can be attributed to the void content of the composite.⁶¹ The void volume fractions (v_v) are found as 0.018 (1.8%) for the neat and 0.017 (1.7%) for the multi-scale 3D composites with a slight decrease after MWCNT reinforcement.

Particle dispersion investigation

Ultrasonic homogenization is one of the most successful methods to obtain good particle dispersion in liquid thermoset resins.^{11,15,34,40,41–45} This method offers the highest homogeneity opportunity of the particles in the resin before the wetting of fiber reinforcements. When the vacuum infusion manufacturing process was performed, the reinforcement particles were considered homogeneous. Although it is quite important to observe the particle distribution by the aid of microscopic analysis, in some cases, the particle amount, size, or color can limit the visibility of the nanoscale reinforcements embedded into polymer matrix. In this case, different approaches are required to determine the particle distribution. In our previous study, the color change on nano-SiO₂ reinforced 3D woven composites and they found that the SiO₂ particles have changed the color of the resin and composite from transparent to gray. Thus, they obtained important results about the distribution of the nano-SiO₂ on the surface of 3D woven composite.³⁴

The color measurements were performed on at least three different points on the sample surface, including the resin inlet zone to the vacuum bag (first contact region of the resin with woven), middle zone of the vacuum bag, and the outlet zone from the vacuum bag. The ΔE results are expected to be between 0 and 1 to claim that the particles are homogeneously distributed. Obtained ΔE results and all three points in both warp and weft directions are given in Figure 9. In the measurements where the highest ΔE value is 0.99 and the lowest ΔE value is 0.21, all the ΔE results were found in the range of 0–1. It is understood that the numerical magnitude does not make a color tone difference when the highest value here is 0.99 compared to 0.75 in their own series (0.5% MWCNT reinforcement level on warp direction). Similarly, when we compared the lowest result is 0.21 with 0.38 in their own series (0.3% MWCNT reinforcement level on warp direction), the meaning of great similarity should not be deduced. According to the results, all measurement points were considered the same color with small differences and thus prove that the good and balanced MWCNT distribution were obtained on the surface of the 3D woven composites. It is a fact that there is a little amount of filtration as in the production of all the layered materials. The reason why some of the improved mechanical properties in the warp direction are not seen in the direction of the weft is that it is assumed that the effect of filtration is mostly occurred on the inner surface (especially occurred on plies).

Figure 9.

Measured ΔE results of multiwalled carbon nanotubes reinforced three-dimensional woven composites.

It is thought that filtration is considered to be the reason of some strength decreases detected in weft-oriented samples. Obviously, the filtration will occur at the inside of the composite and on the consecutive sequential plies. The absence of color difference in the measurements made from the outer surface is that the MWCNT particles during the vacuum can easily progress on the outer surface by means of the infusion flow net. The strength increase obtained showed that the warp-directional samples did not have any filtration due to the same flow direction as the plies, and the MWCNT could be distributed evenly both on the outer and inner sides. All the findings show that sonication is the most effective way to eliminate any aggregate, manufacturing mistakes, and resin pool formation due to over-wetting, however, achieving good and balanced particle distribution.^{11,15,40–45}

TGA

To investigate the influence of the MWCNT on the thermal properties, TGA was conducted on the 3D woven composites. Since the highest mechanical performances were obtained for the samples reinforced with 1wt.% MWCNT, TGA studies were carried out as a comparation between neat and 1wt.% MWCNT-reinforced samples. The weight loss rates with respect to increasing temperature are given in Figure 10. According to the TGA curves, the weight loss rates of the MWCNT-reinforced 3D woven composite is lower than the neat epoxy composite at all the temperatures. It is clear that the improvement in thermal properties is due to the superior thermal properties of MWCNTs introduced within the 3D composites. While the neat epoxy composites start a constant mass loss at 40°C up to 272°C with more than 5% loss rates. On the other hand, the MWCNT-reinforced 3D woven composites maintain its stability between the 40 and 272°C representing only 1% loss rate. As seen from the TGA graph, the slope of the graphic of the neat epoxy composites is much higher than the MWCNT-reinforced composites for the first 300°C. The decomposition of the neat epoxy composite accelerated after 272°C, for the MWCNT-reinforced composite high loss rate was detected after reaching 321°C. The temperatures at where the decomposition ended was 653°C for the neat epoxy composite and 763°C for the MWCNT-reinforced composite. The results indicate that the thermal stability and decomposition characteristics of 3D woven composite are improved with the addition of MWCNT. In addition, the calculated operating temperature range is increased more than 50°C by the intermolecular bond strength of the MWCNT and epoxy resin on 3D woven composite.

Figure 10.

Thermogravimetric graphic of the neat epoxy and 1wt.% multiwalled carbon nanotube reinforced three-dimensional woven composite.

DMA results

Viscoelastic behaviors of the MWCNT-reinforced 3D woven composite material under thermal and variable load were investigated with DMA studies. DMA studies were performed on 1wt.% MWCNT-reinforced and neat epoxy 3D woven composite samples. The DMA results of the prepared composite samples are given in Figure 11. The obtained findings support the developing of elastic properties obtained from mechanical experiments.

Figure 11.

Dynamical mechanical analysis results of reinforced multiwalled carbon nanotube and neat epoxy three-dimensional woven composites; (a) storage modulus, (b) loss modulus, and (c) tan δ results.

The MWCNT reinforcement improved the storage modulus by approximately 7% in both under and above glass transition temperatures, while the improvement was more significant in the glassy region.²⁰ This can be associated with the formation of an effective interface between the nanoparticles and the epoxy matrix. The fact is that the MWCNTs located in the epoxy affects the alignment capability of the molecule chain and restricts their movement.⁶² Thus, the MWCNT addition improves the ability of the material to transmit the load to which it is exposed and increases its strength by limiting the growth of nanopores, as well as the propagation of cracks. At high temperatures, the material shows more viscous behavior. This affects the load transmission performance and resulted decrease in modulus by reducing the contact between the matrix and reinforcing materials at the reaction of the material given by repeated loads in the thermal environment.⁶³ According to the DMA curve given in Figure 11(a), the storage modulus of the MWCNT-reinforced composite was found to be 4.85 GPa, while the neat epoxy composite was found to be 4.52 GPa for 30°C. After the maximum peak, the storage modulus goes down slowly and the values were found as 4.6 GPa and 4.38 GPa for 1wt.% MWCNT-reinforced and neat epoxy 3D woven composite, respectively, for 50°C. The storage modulus of the neat epoxy composite rapidly decreases, especially after 60°C, while the rate of decrease slowed down in MWCNT-reinforced composites. In addition, higher storage modulus values are still obtained in comparison with all selected reference temperatures between the MWCNT-reinforced and neat epoxy composites.

The loss modules of 3D woven composites were given in Figure 11(b). The loss modulus of the MWCNT-reinforced composite was calculated as 5.07 GPa, while the neat epoxy composite was calculated as 3.84 GPa. The high loss modulus shows the high-energy losses in the material. In addition, the reinforcement material increases the temperature at which the loss modulus maximum values were obtained. Accordingly, if the maximum loss modulus temperature of the neat composite is to be referenced (at 83°C), the loss modulus value of MWCNT-reinforced composite can be read as 3 GPa. This decrease in loss modulus compared to the neat composite for the referenced temperatures means that the elastic responses of reinforced composite have improved. This result supports the conclusion obtained from static mechanical experiments that MWCNT improves the elastic properties of 3D woven composites. In addition, it also supports TGA results in improving thermal stability. It is seen that the module curve goes down to the negative region on the graph in Figure 11(b). The physical shape of the 3D woven composite material is thought to influence this. It is known that the material becomes more viscous at high temperature. The material with a gap between the face sheets cannot show the same elastic reaction in the high temperature experiment. The upper face sheet is more affected by the force and approaching the lower face sheet as deformed while the lower face sheet maintaining the stability. This means that the dimensions of the specimen changes compared to the first definition at the beginning of the test. The changed test condition while the test maintaining is considered the reason for the curve going down to the negative region. Similar trend is valid in the curve of the tan δ.

Peak points of the tan δ curve (Figure 11(c)) give the glass transition temperature of the material. Accordingly, these values were found as 96°C for neat composite and 100°C for MWCNT-reinforced composite and the improvement value calculated approximately 5%. The rising glass transition temperature with the addition of nanoparticles is another proof that the efficient working temperature range of the composite is expanding. In addition, the narrowing of the peaks in the diagnostic curve reveals that the time interval in rubber consistency decreases and material rigidity increases. These results are compatible with the studies of Jiang et al.⁶⁴ indicating that by developing the glass transition temperature of the epoxy nanoparticles, the rubbery thermal zone is narrowed. Similarly, Li et al.²⁰ reported that the storage modulus increased, the thermal decay slowed down, and the glass transition temperature developed in their studies in which CNT and Al₂O₃ reinforced epoxy glass fiber hybrid composites used.

SEM-analysis

In Figure 12(a), a SEM of the “I” shaped piles and the cross section of the 3D woven composite produced in the study are given. The image taken from the damaged sample after mechanical experiments shows how well each fiber and fiber knit group is wetted. The increase in mechanical properties and improvements in thermal properties reveal that the MWCNT additive is homogeneously distributed in the structure and a well-wetted composite is produced by resin. The figure also shows propagated cracks caused by the bending load applied on the matrix material and resulting in damage. It has been found in the images that the beginning of the damage starts from the lower and upper surface connection braids. In the tensile test samples in Figure 12(b), crack caused suddenly by increasing load, that passes through the fibers, causes complete separation by breaking and peeling in the material. It is accepted as a normal result that it shows a different breaking characteristic according to the type of loading applied.

Figure 12.

Scanning electron microscopy images of the fractured samples; (a) “I” shaped piles and damage occurred, (b) damage on matrix material and fibers.

MWCNTs and their distributions on the surface of the 3D woven composite produced in the study are given in Figures 13(a) and (b). In the SEM studies, important findings revealing the homogeneous distribution of MWCNT in the composite structure were obtained. As mentioned in the literature, the contribution of the ultrasonic mixing method is quite large in the production of the structure with these features.^{11,15,40–45} With this application, it is possible to separate the particles into smallest parts and distribute them homogeneously. However, the efficiency of the vacuum infusion method should not be overlooked in order to ensure low porous and good wetting.^1,6,36

Figure 13.

Surface scanning electron microscopy images of three-dimensional woven composite; (a) and (b) distribution of the multiwalled carbon nanotubes.

Conclusion

In this study, MWCNT-reinforced 3D woven spacer composites were manufactured by the partial vacuum infusion method, and mechanical and thermomechanical properties were investigated compared to neat epoxy 3D spacer woven composites. In addition, fracture surfaces of damaged samples after the mechanical test were investigated with scanning electron microscopy, and particle distribution was investigated by color measurement on surfaces. The addition of MWCNT (1wt.%) improved flexural strength of composites by 25% on warp and by 44% on weft directions. In addition, flexural modulus of the reinforced composites improved by 80% on warp and 85% on weft directions. At 1wt.% MWCNT reinforcement level again, while tensile strength on warp direction, 7% improvement was obtained. On the other hand, no improvement was obtained on weft-directed tensile strength, but the tensile curve clearly shows the improvement in elasticity, approximately 12%, on weft-directed samples. The strain values were improved by 19% on warp and by 12% on weft directions. Due to the direction of weaving, the strength in the warp direction is clearly higher than in the weft direction numerically for both neat and MWCNT-reinforced composite materials. However, when the increase rate is compared to each warp- and weft-directed sample, it is determined that the gain in the weft direction is 10% higher than the warp direction. The 3D composite density slightly increased with the addition of MWCNTs, and specific strength and specific modulus decreased compared to the strength and modulus. However, the specific bending strength, specific modulus, and specific tensile strength improve for 1wt.% MWCNT-reinforced 3D composite compared to the neat 3D composite. According to the findings, dimensional stability is mostly affected by the composites manufacturing method. MWCNT addition increases thickness within acceptable rates. Calculated results show that the fiber volume fraction increases and the void volume fraction slightly decreases with the addition of 1wt.% MWCNTs. It was observed that the MWCNT reinforcement improved the thermal decomposition characteristics of the composite by slowing down the decomposition rates below 300°C and increased the operating temperature range more than 50°C by the intermolecular bond strength of the 3D woven composite. The storage modulus exhibited a 7% increase under glassy region, and the glass transition temperature was improved by 4°C with the addition of 1wt.% MWCNT.

Damage mechanisms are typically realized. The knitting region between the piles and face sheets was the first deformed point in the bending loading. With the increased load, the material was completely separated because of pressing on the top surface and tensile on the bottom surface. In the tensile tests, the top and bottom surfaces were first exposed to stretching. With an increased load beyond the elastic region, the composite material was failed with a sudden brittle fracture. SEM images show that the MWCNTs are distributed homogeneously on the surface without agglomeration by using the ultrasonic mixing method and vacuum infusion manufacturing methods. These results are in good agreement with color measurement results, where all the measured point colors are the same. SEM images also prove good wetting and bonding interface between the fiber and matrix material, and this is necessary to obtain improvement in 3D woven composites.

Footnotes

Acknowledgments

This study was supported by the Kütahya Dumlupınar University Scientific Research Project Commission with the 2015/69 acceptance numbered project. The authors also thank the Kütahya Dumlupınar University Advanced Technologies Center and Çanakkale Onsekiz Mart University Science and Technology Application and Research Center for their contributions to the study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Ferhat Yildirim

Mustafa Aydin

References

Zhao

, et al. Experimental study on the compression properties and failure mechanism of 3D integrated woven spacer composites. Mater Des 2014; 56: 50–59.

Hassanzadeh

Hasani

Zarrebini

. Thermoset composites reinforced by innovative 3D spacer weft-knitted fabrics with different cross-section profiles: materials and manufacturing process. Compos A Appl Sci Manuf 2016; 91: 65–76.

Salvatori

Caglar

Michaud

. 3D spacers enhance flow kinetics in resin transfer molding with woven fabrics. Compos A Appl Sci Manuf 2019; 119: 206–216.

Neje

Behera

. Investigation of mechanical performance of 3D woven spacer sandwich composites with different cell geometries. Compos Part B Eng 2019; 160: 306–314.

Fan

Sun

Yang

, et al. Interlocked hierarchical lattice materials reinforced by woven textile sandwich composites. Compos Sci Technol 2013; 87: 142–148.

Zhao

Jiang

, et al. Experimental study on the bending properties and failure mechanism of 3D integrated woven spacer composites at room and cryogenic temperature. Compos Struct 2014; 111: 56–65.

Wang

Zhang

, et al. Effect of structure on the mechanical behaviors of three-dimensional spacer fabric composites. Appl Compos Mater 2009; 16: 1–14.

Sadighi

Hosseini

. Finite element simulation and experimental study on mechanical behavior of 3D woven glass fiber composite sandwich panels. Composites: Part B 2013; 55: 158–166.

Fan

Zhao

Chen

, et al. Ductile deformation mechanisms and designing instructions for integrated woven textile sandwich composites. Compos Sci Technol 2012; 72: 1338–1343.

10.

Fan

Zhou

Yang

, et al. An experiment study on the failure mechanisms of woven textile sandwich panels under quasi-static loading. Composites: Part B 2010; 41: 686–692.

11.

Bilişik

Yolaçan

. Flexural properties of woven e-glass/polyester nano silica composites, XIII. In: International İzmir Textile and Clothing Symposium, Antalya, Turkey, 2–5 April 2014, pp. 229–234.

12.

Nayak

Dash

Ray

. Effect of epoxy modifiers (Al2O3/SiO2/TiO2) on mechanical performance of epoxy/glass fiber hybrid composites. Proced Mater Sci 2014; 6: 1359–1364.

13.

Chandrasekaran

VCS

Advani

Santare

. Role of processing on interlaminar shear strength enhancement of epoxy/glass fiber/multi-walled carbon nanotube hybrid composites. Carbon 2010; 48: 3692–3699.

14.

Gojny

Wichmann

MHG

Fiedler

, et al. Influence of nanomodification on the mechanical and electrical properties of conventional fibre reinforced composites. Composites: Part A 2005; 36: 1525–1535.

15.

Zhou

Pervin

Rangari

, et al. Fabrication and evaluation of carbon nano fiber filled carbon/epoxy composite. Mater Sci Eng A 2006; 426(1–2): 221–228.

16.

Beyakrova

Thostenson

, et al. Multiscale carbon nanotube-carbon fiber reinforcement for advanced epoxy composites. Langmuir 2007; 23: 3970–3974.

17.

Böger

Sumfleth

Hedemann

, et al. Improvement of fatigue life by incorporation of nanoparticles in glass fibre reinforced epoxy. Composites, Part A 2010; 41: 1419–1424.

18.

Eskizeybek

Ulus

Kaybal

, et al. Static and dynamic mechanical responses of CaCO₃ nanoparticle modified epoxy/carbon fiber nanocomposites. Composites, Part B 2018; 140: 223–231.

19.

Gibson

. A review of recent research on mechanics of multifunctional composite materials and structures. Compos Struct 2010; 92(12): 2793–2810.

20.

Dichiara

Zha

, et al. On improvement of mechanical and thermo-mechanical properties of glass fabric/epoxy composites by incorporating CNT–Al₂O₃ hybrids, Compos Sci Technol 2014; 103: 36–43.

21.

Godara

Gorbatikh

Kalinka

, et al. Interfacial shear strength of a glass fiber/epoxy bonding in composites modified with carbon nanotubes. Compos Sci Technol 2010; 70: 1346–1352.

22.

Rahman

Zainuddin

Hosur

, et al. Improvements in mechanical and thermo-mechanical properties of e-glass/epoxy composites using amino functionalized MWCNTs. Compos Struct 2012; 94: 2397–2406.

23.

Yaping

Aibo

Quinghua

, et al. Functionalized effect on carbon nanotube/epoxy nano-composites. Mater Sci Eng A 2006; 435–436: 145–149.

24.

Sahoo

Rana

Cho

, et al. Polymer nano composites based on functionalized carbon nano tubes. Prog Polym Sci 2010; 35: 837–867.

25.

Thostenson

, Ren

Chou

. Advances in the science and technology of carbon nanotube and their composites: a review. Compos Sci Technol 2001; 61: 1899–1912.

26.

Eskizeybek

Avci

Gulce

. Preparation and mechanical properties of carbon nanotube grafted glass fabric/epoxy multi-scale composites. Adv Compos Mater 2017; 26(2): 169–180.

27.

Ustun

Ulus

Karabulut

, et al. Evaluating the effectiveness of nanofillers in filament wound carbon/epoxy multiscale composite pipes. Compos B Eng 2016; 96: 1–6.

28.

Sánchez

Campo

Jiménez-Suárez

, et al. Effect of the carbon nanotube functionalization on flexural properties of multiscale carbon fiber/epoxy composites manufactured by VARIM. Compos B Eng 2013; 45(1): 1613–1619.

29.

Tornabene

Fantuzzi

Bacciocchi

. Linear static response of nanocomposite plates and shells reinforced by agglomerated carbon nanotubes. Compos B Eng 2017; 115: 449–476.

30.

Talo

Krause

Pionteck

, et al. An updated micromechanical model based on morphological characterization of carbon nanotube nanocomposites. Compos B Eng 2017; 115: 70–78.

31.

Salam

MBA

Hosur

Zainuddin

, et al. Improvement in mechanical and thermo-mechanical properties of epoxy composite using two different functionalized multi-walled carbon nanotubes. J Compos Mater 2013; 3: 1–9.

32.

Gojny

Wichmann

MHG

Fiedler

, et al. Influence of different carbon nano tubes on the mechanical properties of epoxy matrix composites – a comparative study. Compos Sci Technol 2005; 65: 2300–2313.

33.

Mattos

Rojas

Magalhães

WLE

. Biogenic SiO2 in colloidal dispersions via ball milling and ultrasonication. Powder Technol 2016; 301: 25–64.

34.

Yildirim

Aydin

Avci

. Mechanical properties of nano-SiO₂ reinforced 3D glass fiber/epoxy composites. Int J Mater Res 2017; 108(4): 308–321.

35.

Fan

Yang

Zhou

. Experimental research of compressive responses of multi-layered woven textile sandwich panels under quasi-static loading. Composites: Part B 2011; 42: 1151–1156.

36.

Vaidya

Uddin

. Impact response of three-dimensional multifunctional sandwich composite. Mater Sci Eng A 2008; 472: 52–58.

37.

Van Vuure

Ivens

Verpoest

. Mechanical properties of composite panels based on woven sandwich fabric performs. Composites: Part A 2000; 31: 671–680.

38.

Zheng

Zhao

Fan

. Energy absorption mechanisms of hierarchical woven lattice composites. Composites: Part B 2012; 43: 1516–1522.

39.

Parabeam 3-d Glass Fabrics . Technical data sheet. https://romar-voss.nl/archief/Composites/kern/pdf/Parabeam%203-D%20glasweefsel%201/ParaGlass/Technical%20Data%20Sheet%20Metric.pdf (accessed 27 September 2020).

40.

Johnsen

Kinloch

Mohammed

, et al. Toughening mechanisms of nano particle nano silica-modified epoxy polymers. Polymer 2007; 48: 530–541.

41.

Huang

, et al. Multiscale carbon nanotube-woven glass fiber reinforced cyanate ester/epoxy composites for enhanced mechanical and thermal properties. Compos Sci Technol 2014; 104: 81–88.

42.

Zhou

Cheng

, et al. Improvement in electrical, thermal and mechanical properties of epoxy by filling carbon nanotube. Express Polym Lett 2008; 2(1): 40–48.

43.

Patra

Mitra

. Interface fracture of sandwich composites: influence of MWCNT sonicated epoxy resin. Compos Sci Technol 2014; 101: 94–101.

44.

Siddiqui

Sham

Tang

, et al. Tensile strength of glass fibres with carbon nanotube–epoxy nanocomposite coating. Composites: Part-A 2009; 40: 1606–1614.

45.

Fernandez

Medina

Pincheira

, et al. The effect of multiwall carbon nanotubes on the in-plane shear behavior of epoxy glass fiber reinforced composites. Composites: Part-B 2013; 55: 421–425.

46.

Emady

Wittman

Koynov

, et al. A simple color concentration measurement technique for powders. Powder Technol 2015; 286: 392–400.

47.

Koynov

, Wang

Redere

, et al.

Measurement of the axial dispersion coefficient of powders in a rotating cylinder: dependence on bulk flow properties powder

Technol 2016; 292: 298–306.

48.

User manual of the color measurement device. 2014, http://mepteknik.com/bg/assets/images/kkilavuz/RENK%20OLCUM%20CIHAZI%20KULLANMA%20KILAVUZU.pdf, (accessed 12 March 2020).

49.

Ülkü

Korkmaz

. Farklı Kompozisyonlardaki Polyesterin Mekanik Özelliklerinin Yapıya Bağlı Olarak Deneysel ve Teorik Olarak İncelenmesi. In: Balıkesir Symposium, Turkey, 1988.

50.

Gojny

Wichmann

MHG

Köpke

, et al. Carbon nanotube-reinforced epoxy-composites: enhanced stiffness and fracture toughness at low nanotube content. Compos Sci Technol 2004; 64: 2363–2371.

51.

Zhou

Pervin

Lewis

, et al. Fabrication and characterization of carbon/epoxy composites mixed with multi-walled carbon nanotubes. Mater Sci Eng A 2007; 475: 157–165.

52.

Liu

Wagner

. Rubbery and glassy epoxy resins reinforced with carbon nanotubes. Compos Sci Technol 2005; 65: 1861–1868.

53.

Nadler

Werner

Mahrholz

, et al. Effect of CNT surface functionalisation on the mechanical properties of multi-walled carbon nanotube/epoxy-composites. Compos Part A 2009; 40: 932–937.

54.

Zhang

Liu

Huo

, et al. Filtration effects of graphene nanoplatelets in resin infusion processes: problems and possible solutions. Compos Sci Technol 2017; 139: 138–145.

55.

Ashby

. Material Selection in Mechanical Design. 2nd ed. Oxford, England: Butterworth-Heinemann, 1999, pp. 43–44.

56.

Wang

, Wang

, Sun

, et al. Study on dimensional stability of SiCp/Al composite, JO - ECCM 2012 - composites at Venice. In: Proceedings of the 15th European Conference on Composite Materials, Venice, January 2012.

57.

Whitacre

Amiri

Ulven

. The effects of corn zein protein coupling agent on mechanical properties of flax fiber reinforced composites. Ind Crops Prod 2015; 77: 232–238.

58.

Abdalla

Megat

Sapuan

, et al. Determination of volume fraction values of filament wound glass and carbon fiber reinforced Composites. J Eng Appl Sci 2008; 4: 7–11.

59.

Binshan

Svenson

Lawrence

. Mass and volume fraction properties of pultruded glass fibre-reinforced composites. Composites 1995; 26: 725–731.

60.

Mehdikhani

Gorbatikh

Verpoest

, et al. Voids in fiber-reinforced polymer composites: a review on their formation, characteristics, and effects on mechanical performance. J Compos Mater 2019; 53(12): 1579–1669.

61.

Daniel

Ishai

. Engineering Mechanics of Composite Materials. 2nd ed.

New York:

Oxford University Press, 2006.

62.

Bao

Tjong

. Mechanical behaviors of polypropylene/carbon nanotube nanocomposites: the effects of loading rate and temperature. Mater Sci Eng A 2008; 485(1–2): 508–516.

63.

Montazeri

Khavandi

Javadpour

, et al. Viscoelastic properties of multi-walled carbon nanotube/epoxy composites using two different curing cycles. Mater Des 2010; 31: 3383–3388.

64.

Jiang

Jin

Park

. Thermo-mechanical behaviors of epoxy resins reinforced with nano-Al₂O₃ particles. J Ind Eng Chem 2012; 18: 594–596.