Effect of MWCNT on the properties of NCF-CFP hybrid composite fabricated via vacuum infusion

Abstract

This work is an attempt to evaluate the effect of multi-wall carbon nanotube (MWCNT) reinforcement on the properties of non-crimp fabric (NCF) carbon fiber-reinforced polymer hybrid composite (CFRP). MWCNT NCF-CFRP plates were produced by vacuum infusion. Epoxy-MWCNT mixtures were prepared by adding a surfactant, Triton-X 100, and ethanol and mixing ultrasonically. The plates were subjected to thermal and electrical conductivity, DSC and FT-IR characterizations, and tensile testing. SEM image analyses evaluated tensile fracture surfaces. Results showed that 0.5 wt% MWCNT-reinforced nanocomposite samples exhibited better tensile properties than pure NCF-CFRP composites and other nanocomposite samples. The 1.5% MWCNT reinforced sample was the best conductor at all temperatures. The temperature increases also increase the conductivity. However, the conductivity value of 0.5% MWCNT reinforced material increased significantly at low frequencies such as 100–10,000 Hz. The heat reaction of the 0.5 wt% MWCNT nanocomposite was also a bit higher than that of 1–1.5 % wt MWCNT nanocomposites, indicating a decrease in the cure.

Keywords

quadriaxial non-crimp fabric carbon fiber reinforced polymer composite multi-wall carbon nanotube tensile properties nanocomposites

Introduction

Carbon fiber-reinforced polymer hybrid composite (CFRP) has found many applications in various sectors such as automotive, aerospace, marine, and so forth, particularly because of its lightweight and high strength in the last decades.^1,2 In addition, reinforcing CFRP with MWCNT has become a popular topic of study due to their substantial contribution to the strength and physical properties of the resulting composite material.^1,3 Besides, the orientations and architecture of the fibers can be tailored for specific strength and modulus requirements.⁴ Researchers have focused on the influence of production methods on the outcome properties of CFRPs.^5–7 There are many different CFRP production methods, each with advantages and disadvantages. For example, common methods are pre-impregnated unidirectional (UD) tapes, woven fabric, and NCF (non-crimp fabric). Composites used in high-performance structural implementation have conventionally been manufactured from unidirectional layers of UD. The UD are thin, unidirectional, equally deployed fibers impregnated with a partly cured thermoset resin.⁸

However, the UD composites also have some disadvantages. Cross-linking of the polymer is deactivated by reducing the temperature; therefore, UD must be stored in freezers. Cross-linking is continued when the temperature is raised. UD need a low temperature to prevent curing, which causes high enhancement storing cost. Autoclave requirement for manufacturing increases the costs of using UD.^9,10 In addition, the in-plane orientation of the reinforcement may also create a material that is rather sensible to delamination cracking under impact loading because of its weak interlaminar fracture toughness. As a result, the post-impact in-plane mechanical properties can be negatively affected.⁸ On the other hand, the non-crimp fabric (NCF) is a recently developed reinforcement element that can combine advantages from both types. Unlike woven fabrics, Non-Crimp Fabrics feature unidirectional layers of fibers that are held together by stitching that goes through the thickness of the material, which means that there is no interlocking of fiber tows.^8,9,11

Overlaying straight tows joined by stitching in Non-Crimp Fabrics (NCFs) offers several benefits over woven fabrics. One significant advantage is that two crimps can be avoided, resulting in improved strength, stiffness, and fatigue life compared to woven fabrics of equivalent weight.^2,8,9 Additionally, manufacturing NCFs is a simple process that uses conventional stitching machinery. The relative absence of two crimps can produce composites with comparable mechanical performance to those made from prepreg. The crashworthiness properties of these materials are the primary factor considered by automotive OEMs when deciding whether to use them for structural applications in lightweight vehicles. NCF reinforcements have excellent processability because they contain dry fibers, resulting in composites with outstanding mechanical properties due to the minimal crimping along the fiber directions.^8,12

CNT enrichment of CFRP has been an important research issue for the last decade, provides significant involvement in the performance of polymer composite, and has been reported.^13–17 However, some important considerations and limitations exist in manufacturing CNT/CFRP composites. One of the most important problems in this concept is that CNTs come together and form bundles, which is most effective because the Van Der Waals bond energy increases the attractiveness between CNTs. Consequently, the CNT’s surface areas can be limited.^18,19 It has also been reported that MWCNTs aggregate can cause weak interfacial between MWCNTs and epoxy resin. However, there are also several solving methods to overcome weak bonding problems. The researchers say these are mainly covalent and noncovalent functionalizations.^20–23 For example, covalent functionalization is an effective process that promotes the homogeneous morphology of carbon nanotubes in an epoxy matrix, thereby increasing the mechanical properties of polymer composites.²³ On the other hand, the noncovalent functionalization process can generate steric forces that are more repulsive and effective than the Van Der Waals Forces of attraction between CNTs²⁴ by making CNT surfaces covered by surfactant molecules.²⁵ Surfactants are utilized to strengthen the interfacial bonds between the epoxy matrix and CNTs by preventing the aggregations of the CNTs,²⁶ improving mechanical properties.

Researchers^27–30 investigated the effect of MWCNT on MWCNT/epoxy composite using 0.2–0.75 wt% MWCNT. Ci et al.³⁰ reported that the increase in ultimate strength, elastic modulus, and impact energy were 41%, 22%, and 45%, respectively. Alawy et al.²⁹ reported that the maximum mechanical property increase was achieved with 0.2% wt COOH-MWCNT. In Ne et al.’s³¹ work, tensile strength and Young’s modulus increased by approximately 14% and 15%, respectively, at 0.6% MWCNT compared to pure epoxy. Shishevan et al.³² used MWCNT and GNP to reinforce epoxy and then dispersed by an ultrasonic mixer (400 W, 24 kHz) for 15 min at 90 W and reported that adding 0.5 wt% MWCNT increased tensile strength.

Ranjbar et al.,²⁷ Chang,³³ Patil et al.,³¹ Shivamurthy et al.,³⁴ and Rahman³⁵ drew attention to the changes in the tensile properties of GFRP to which MWCNT was added between 0 and 3% by weight. According to the neat epoxy, the composite's elastic modulus and tensile strength increased to proximate 21.98%–58.32% by adding 0.34 wt% of MWCNTs into the epoxy at Ranjbar's work.²⁷ The tensile properties of MWCNT epoxy composite increased by 4.4% with the addition of 1% MWCNT but decreased when the 2% wt. MWCNT was added to Chang’s work.³³ Patil's³¹ test indicated that tensile strength increased with 1% MWCNT but decreased by adding 2% and 3% MWCNT. Shivamurthy's test exhibited the highest tensile strength obtained by adding 0.3% wt. of MWCNT.³⁴ Rahman³⁶ also reported similar results in their study.

Chang³³ and Ayatollahi et al.³⁷ evaluated the effects of different MWCNT contents (between 0.1 wt% to 2 wt%) on the tensile behavior of carbon fabric-reinforced epoxy composites. In Chang's work, the tensile strength of MWCNTs–CFRP epoxy composite increased by 34.6% with 0.5 wt% MWCNT decreased when the amount of MWCNT from 0.75 wt% to 2 wt% were added. Ayatollahi³⁷ reported that adding up to 0.5 wt% increased the tensile strength. However, tensile strength deteriorated with MWCNT 0.9 wt%. Kumar et al.,³ Chang, and Ayatollahi et al. studied epoxy nanocomposites containing between 0.25 and 1.0 wt% MWCNTs applied various methods to provide homogeneous dispersion of carbon nanotubes in the polymer matrix and kept the solution temperature during sonication around 45°C, 50°C, and 60°C, respectively. In Kumar’s work, the solution was sonicated at 60% amplitude with, kept open for 10 s, and closed for 10 s for 30 min around the beaker was cooled to control the temperature at 45°C. In Ayatollah's works, it was stated that the mixing vessel was kept in water to maintain its temperature around 50°C. Raza et al.³⁸ researched the effect of the different weight ratios of pristine and functionalized MWCNTs (0.1, 0.5, and 1 wt %) on epoxy composites. MWCNT was added into epoxy and dispersed by a stirrer for 1 h and 6 h by ultrasonication. The maximum tensile strength increase for p-MWCNTs and COOH-MWCNT was 0.1 and 0.5 wt % MWCNT, respectively.

Schulte et al.³⁹ investigated the propagation of diagonal cracks in a glass fiber-reinforced composite of NCF in an epoxy matrix containing 0.3 wt% of MWCNT. Wang⁴⁰ investigated the effect of amounts from 0 to 2% of MWCNT functionalized with carboxyl on flexural properties, interlaminar shear strength, and NCF-CFRP fracture toughness. Pekturk et al.⁴¹ reported that the maximum fatigue life was achieved in the 0.5 wt% MWCNT. Geng et al.⁴² investigated the effects of surfactant-Triton X-100 application of CNT on the mechanical properties of CNT-epoxy nanocomposites. CNT mixed with approximately 0.125–1.25 mg/mL polyethylene glycol test-octyl phenyl ether (Triton X-100) wt/acetone volume ratios were then sonicated at 25°C for 6 h. Epoxy resin was added to the mixtures and sonicated at 60°C for 2 h. Mechanical test results showed that the properties of the composite improved after surfactant treatment. Dehghan⁴³ also declared similar results via mixed with MWCNT.

This study attempts to provide new materials and methods to related areas. Namely, in the open literature, no studies like this study’s concept organized on quadriaxial NCF-CFRP revised by COOH-MWCNT (0.5%, 1.0%, and 1.5%) and vacuum infusion technique. It is important to break Van Der Waals's Forces of attraction between MWCNT to help prevent agglomeration and contribute to the homogeneous distribution in epoxy. As seen from the studies above, using varying amounts of surfactant is possible.

Moreover, according to the open literature above, the solution should be mixed with an ultrasonic sonicator to ensure the best homogeneity. Therefore, sonication was applied to the epoxy and MWCNT solution for 10 h following this issue, and the solution temperature was kept between 30 and 35°C in this study. However, some researchers^{27–31,42,44–46} have not emphasized the sonication temperature; others^3,37,47 reported that the rising sonication process could deteriorate the matrix's MWCNT structure. Because of that, this study has given special consideration to keeping sonication temperature as low as possible. Moreover, the MWCNT's dispersion homogeneity in the matrix of the samples was examined using UV-visible for hours spectroscopy. Consequently, the effect of MWCNT content and surfactant used in quadriaxial NCF CFRP nanocomposites has been investigated by considering the development of tensile, thermal, and electrical properties.

Experimental studies

Material

The epoxy system was EPOSIS 120 with a resin/hardener ratio of 100:60 by weight. The NCF comprises four layers of 3 K carbon fiber tows with directions of −45^o, 90^o, +45^o, and 0^o stitched together. MWCNTs have been functionalized by –COOH groups, with an outside diameter of 10–20 nm, lengths of 5–10 µm, and purity of 96%. A surfactant named Triton X-100 was used for the better dispersionof MWCNT in epoxy resin, a non-ionic surfactant appropriate for epoxy composites. In addition, purification of the MWCNT was performed with ethanol.

The fabrication process for the multi-wall carbon nanotube non-crimp fabric carbon fiber-reinforced polymer

Figure 1 shows the manufacturing process’s schematic flow for MWCNT-reinforced NCFP composites. Four types of samples were processed. One was pure resinous, and the other three were prepared at equal surfactant concentrations and varying concentrations of 0.5, 1.0, and 1.5 wt% MWCNT epoxy resin. As shown in Figure 1, samples were prepared in two main steps. In the first stage, equal amounts of MWCNT and surfactant are put into 200 mL of ethanol, mixed in a beaker, poured into the recycling tank, and sonicated at 80% amplitude using sonotrode with a diameter of 19 in the flow cell for 1 h. Triton X-100, used as a surfactant, has an average molecular weight of 650 g/mol and a CMC of 0.2 mM at 25^o.

An ultrasonic processor UIP2000hdT, Hielscher, was used for sonication. It was aimed to keep the temperature of the solution constant at 30°C–35°C degrees to prevent deterioration of MWCNT from high temperatures. In this cooling process, cold water was circulated in the recycling tank with the help of a flow cell. After mixing, the mixtures prepared for evaporation of ethanol from the solution were processed on a hot plate at 80°C. A 60:100 weight ratio of epoxy hardener was added to the mixture and mixed for 10 min. Ultimately, this mixture was degassed in a vacuum oven for 15 min. As a result, the blend was drawn into a four-axis crimped carbon fiber fabric by vacuum infusion. These dimensions were 300 × 300 mm. These sheets were cured at room temperature for 48 hours, and the manufacturing process was completed. Orientations of the fabric ([0°/+45°/90°/−45°]s quadriaxial non-crimp carbon fiber fabric) and samples are shown in Figure 2.

Characterization of dispersion: UV-Vis spectroscopy

In the 200–1000 nm wavelength range, UV-Vis spectroscopy was applied to detect the homogeneity of MWCNT epoxy mixing. The solution was diluted in acetone. As a measure of the MWCNT dispersion in epoxy, a peak formed at the absorption spectra was used. In opposition to aggregated nanotubes, completely dispersed nanotubes are extremely active in the 200–1200 nm wavelength range.⁴⁸

Fourier transform infrared spectroscopy

Fourier transform infrared (FT-IR) analysis was performed to detect different functional groups in PHB. The FT-IR spectrum was measured between 4000 and 375°cm⁻¹ in these measurements. That was the Bruker Alpha-P FT-IR device equipped with the OPUS software. Measurements were made at room temperature.

Differential scanning calorimetry

Differential scanning calorimetry (DSC) measurements of the materials were made with the Perkin Elmer Jade DSC device loaded with PYRS software under indium reference material, low range (320 mW), nitrogen gas pressure 3 bar, and a temperature range from 30°C to 200°C.

Dielectric impedance spectroscopy

In this study, Impedance spectroscopy measurements were made with a NoVo control Alpha-N High-Resolution Dielectric Analyzer in the 0–1,000,000 Hz range.

Tensile test and fracture surface analysis

Tensile test samples for all groups were prepared from the laminates by cutting using a circular diamond saw with water cooling in the scope of the ASTM D3039 standard,⁴² as seen in Figure 2(a). The tensile tests were performed five times for all group samples. Tests were performed using an INSTRON 5982 machine using a load cell of 100 kN. The crosshead speed was adjusted and set to 1 reproducibility assessment. Additionally, the video extensometer was used to measure the axial strain of the specimens. Figure 3(b) shows the fiber angles in the quasi-isotropic NCF, Machine direction (MD), Bias Direction (BD), and Cross Direction (CD) directions and the fiber direction of the fabric according to these directions. The tensile test coupons were cut in the machine direction (MD), a property of the manufacturing process known as the feed direction of the fabric to the machine. Tensile test samples' fracture surfaces of NCF CFRP composites were coated with gold to achieve electrical conductivity and then inspected using a scanning electron microscope (Carl Zeiss Ultra Plus Gemini FESEM).

Results and discussions

UV-Vis spectra

As explained in the experimental part, to distribute the MWCNTs as homogeneously as possible, they were first added to the ethanol and surfactant mixture and ultrasonicated for 1 h. Then, this mixture was added to the resin and ultrasonicated again for 10 h. These mixtures and times were performed in light of the literature^26,42,44,49 and additional experiments. The resulting results are given in Figures 3 and 4. Figure 3(a) shows the photos of an example (mixing MWCNT, surfactant, and ethanol) of a poorly dispersed suspension after 1 h of ultrasonication. In Figure 3(b), the mixture, which was ultrasonified for 1 h, was added to the resin and ultra-sonificated again for 10 h. Figure 3(b). In other words, the above-described systematic study attempted to distribute MWCNTs homogeneously by reducing their interactions with Van Der Waals bond interactions. Larger numbers of disentangled MWCNTs caused the blackening of the solution as the ultrasonication time increased. The solution color darkens as the MWCNT disperses and absorbs more light.⁵⁰ Triton X-100 covers the surface of MWCNT. The interface between the surfactant and MWCNT becomes saturated and can supply optimum homogeneity of MWCNT dispersion.^26,42,44,49 As a result, the homogeneous distribution of MWCNTs incorporated into the resin added at different rates with the above ultrasonication systematic was investigated. As a result, the dispersion behavior of MWCNT is successfully inspected by UV-Vis spectroscopy, as seen in Figure 4.

Figure 4 shows the epoxy suspensions’ UV-Vis absorption spectra. As can be seen, the spectrum of all solutions was around 330 nm. However, 0.5% MWCNT mixed epoxy resin showed a higher absorbance value than other samples. As mentioned in the literature, the number of individual nanotubes in solution is increased in this condition. MWCNT tends to coalesce due to Van Der Waals's attractive forces, showing that surfactants with steric repulsive power effectively dissolve CNT agglomerates, as mentioned by Geng et al.⁴²

As in this study, the surfactant quantity is enough to surround the carbon nanotubes at lower MWCNT quantities, such as 0.5 wt%.⁴⁵ However, as the concentration of the nanotubes increases, surfactant becomes inadequate for entirely dispersing the agglomerates of MWCNT.⁴⁴ The 1% and 1.5% MWCNT reinforced epoxy revealed an abrupt reduction in the absorbance peak, indicating a higher re-agglomeration ratio. The absorption peak of the 1.5% MWCNT epoxy solution was higher than the 1% MWCNT epoxy solution. As the amount of MWCNTs in the solution increases, the number of individual MWCNTs suspended increases, leading to the rise of absorption peaks.⁵¹

Fourier transform infrared

Figure 5 shows the transmittance (a.u) versus Wave Number (cm⁻¹) result obtained by FT-IR analysis of the samples. The peaks at 2916.5 cm⁻¹, 1610.1 cm⁻¹, 1507.1 cm⁻¹, 1233.7 cm⁻¹, 1028.3 cm⁻¹, and 822.9 cm⁻¹ indicate the existence of epoxy functional groups. The stretch at 1733.7 cm⁻¹ implies¹⁴ the carboxyl group (COOH) band formed during the –COOH modification of MWCNT. In other words, the epoxy containing 0.5%, 1.0%, and 1.5% by weight MWCNTs show C=O stretching at 1733.7 cm⁻¹, indicating that an esterification reaction has occurred between the epoxy resin and the COOH group of surfaces functionalized MWCNTs (during fabrication). A peak shift was observed when MWCNT was added to the epoxy (an O–H strong stretch from 3288 cm⁻¹ to 3313 cm⁻¹ and a C=C moderate stretch from 1417 cm⁻¹ to 1426 cm⁻¹). This band shift indicates a good correlation between the base and the reinforcement elements.

In pure and 0.5% wt samples, a slight decrease in permeability is observed in 3250–3500 cm⁻¹. On the other hand, 1.0 wt% and 1.5 wt% MWCNT samples contained higher COOH, which could react with OH.⁵² Therefore, less OH remained present in these samples. These results confirm that a reaction between epoxy and COOH-functionalized MWCNTs produces a novel epoxy-MWCNTs nanocomposite with improved fracture toughness, as observed experimentally. A peak shift occurred when MWCNT was included in epoxy 1450 cm⁻¹–1454 cm⁻¹ C=C (medium stretch). This band shift indicates a good interaction between the matrix and the reinforcement. As can be understood from these results, the permeability densities obtained decrease proportionally with increased MWCNT added. Consistency and parallelism were obtained between the obtained permeability data.

Additionally, depending on the surfactants used in this study, it is understood from the FTIR result in Figure 5 that there is no residual substance, for example, alcohol. Functionalization of CNTs has been recognized as an effective method to prevent agglomeration of CNTs and improve charge transfer ability across the interface of CNTs and polymer to achieve better dispersion and strong interfacial interactions.⁵³ However, it can be accepted that the absence of residues in the final product indicates the effectiveness of the preferred refrigerated sonication, degassing, and curing processes that did not cause structural deterioration in this study.^42,49 Lui et al.⁵³ reported that functional molecules, such as different inorganic and organic molecules, are concerned with adsorption to the sidewall of CNTs by π–π packing, van der Waals, or charge transfer relations. Among its results, π–π packing had the strongest relationship of delocalized π-bonds between the CNT wall and functional molecules.

Impedance and electrical conductivity

Figure 6 shows electrical conductivity versus frequency (S/cm) and Z'' versus Z' (change and the variation of electrical conductivity and electrical resistance depending on the amount of MWCNT comparatively with the base material). Electrical conductivity and impedance spectroscopy measurements were made at 20°C (a, d), 100°C in (b, e), and 120°C in (c, f), respectively, in Figure 6. As is known, IS measures the electrical impedance or other closely related quantity as a function of frequency. It analyzes electrical polarization processes according to relaxation frequencies or time constants.

As seen in Figure 6(a)–(c), adding the MWCNT starting at 0.5 wt% to the epoxy matrix formed a semicircular curve. These semicircles decrease in diameter as the MWCNT concentration increases, which indicates the composite samples' lower electrical resistance due to the more conductive chains formed in the polymer matrix. Thus, at all temperatures, 0.5% MWCNT and smaller, except for an overall impedance drop with varying amounts of MWCNT, the Nyquist diagram curves change such that the semicircles are distorted to larger Z with increasing MWCNT content.

It was observed in Figure 6(d)–(f) that the 1.5% MWCNT reinforced material was the best conductor at all temperatures. At the same time, it was observed that the increase in temperature also increased the conductivity. At 10 Hz frequency, 1.5% MWCNT doped material shows a conductivity of about 1.1 × 10⁻¹² at 20°C, 1.1 × 10⁻¹⁰ at 100°C, and 1.14 × 10⁻⁸ at 120°C. Furthermore, at high frequencies like 800,000 Hz frequency, 1.5% MWCNT doped material shows a conductivity of about 3.9 × 10⁻⁷ at 20°C and 6.8 × 10⁻⁷ at 100°C while a conductivity of about 7.55 × 10⁻⁷ at 120°C. However, it is observed that the conductivity value of 0.5% MWCNT-doped material increases significantly at low frequencies such as 100–10,000 Hz. MWCNT is naturally an electrical conductor.^35,54 It is also claimed to be superconducting between 20°C and −253°C. MWCNT nano-polymer composites can be effective with a low permeation threshold and lower electrical resistance, with good dispersion of MWCNT.⁵⁵ As seen from Figure 6, electrical conductivity resistance increases or decreases depending on the increase of MWCNT addition amount. As a result, it is seen that the MWCNTs provided by using surfactants can be distributed homogeneously for epoxy. Accordingly, improved electrical and thermal conductivity or conductivity resistance is achieved.^56–58

The electrical conductivity of epoxy nanocomposites increases with increasing MWCNT loading concentration. Chemical functionalized MWCNTs improve the dispersion quality and strengthen interfacial bonding strength with polymer matrix. However, incorporating chemically functionalized MWCNTs in an epoxy matrix decreases the electrical conductivity of bulk epoxy nanocomposite.

For electrostatic charge dissipation, it is necessary to load more than concentration MWCNTs of surfactant additive MWCNTs, as produced MWCNTs and Octadecylated MWCNTs. However, carboxylated MWCNTs require adding more than 1.5 wt% for electrostatic charge dissipation.³⁵

For heterogeneous polymers, a few factors affect the dielectric constant, such as electronic and atomic orientation and interfacial polarization. In addition, the conductivity of the fillers and the interface adhesion between the fillers and the epoxy matrix determined the electric response of the epoxy nanocomposites. Based on the graphs, the epoxy/MWCNT showed a higher dielectric constant than the neat epoxy. That is due to the addition of the MWCNT to the epoxy matrix, which induced the variation of the polarization process. The MWCNT consists of sp2 hybridized carbon atoms exhibiting superior electrical properties. The delocalization of the π electrons in hybridized sp2 makes the electrons free to move when an electric field is applied. As a result, higher conductivity fillers such as MWCNT lead to the formation of a micro capacitor and increase the dielectric constant of the epoxy nanocomposites.⁵⁸

Differential scanning calorimetry

Differential scanning calorimetry is used as a thermoanalytical technique in which the difference in the amount of heat needed to raise the temperature of a material sample and reference is determined as a function of temperature. Figure 7 shows the heat flow (mW) versus temperature graph. As can be seen from this graph, the heat conduction reaches a maximum of about 50°C with the increase in temperature from room temperature, and the heat flow after that does not show a significant change. However it generally indicates a relative increase from 200°C. While the original CFRP without MWCNT containing carbon fiber showed the lowest thermal conductivity, the CFRP sample containing the highest MWCNT (0.5%) showed the highest thermal conductivity. As stated by previous researchers, the addition of relatively conductive MWCNT provided additional conductivity to epoxy CFRP. In other words, without adding MWCNT, pure epoxy makes an exothermic peak at 44.51°C, which the resin can cause. Including 0.5% by weight MWCNT-COOH in the epoxy increases the initial peak at 70°C. That indicates that the COOH groups react with the epoxy groups upon non-isothermal heating, resulting in ester linkage.^52,59

The main exothermic peak started at about 70°C for 0.5% wt MWCNT nanocomposites, while it was marginally slightly lower for 1% wt or 1.5 wt% MWCNT nanocomposites. That shows the catalytic behavior of nanocomposites. In another explanation, the heat of reaction of 0.5 wt% MWCNT nanocomposite was also marginally higher than that for both 1 wt% and 1.5 wt% MWCNT nanocomposites, implying a decrease in the level of curing. Researchers reported that it improves the thermal conductivity of polymer composites with MWCNT and CFRPs.^2,56 Park et al.⁵⁶ also reported a similar result but reported that the thermal conductivity provided by surfactant with the addition of MWCNT decreased relatively compared to the non-use. However, Zakaria et al.⁶⁰ reported that MWCNT reinforced with epoxy increases thermal conductivity not more than graphene reinforcement.

Tensile properties

A typical stress-strain (S-S) curve for the composite specimens containing different percentages of MWCNTs and the neat CFRP is shown in Figure 8. Different MWCNT concentrations have different tensile behaviors from the curves. The relationship between the percentage of MWCNTs and the tensile strength is shown in Figure 9. The Figure shows that the contribution of 0.5% MWCNT increases the tensile strength by 9.2% concerning the neat NCF composite. The homogeneous distribution of nanotubes in solution could be interpreted as the basis for this improvement. Figure 10 represents the relation between the absorption and the tensile strength (Figure 10(a)) and the absorption spectrum according to MWCNTs percentage (Figure 10(b)). The higher absorption level for 0.5% MWCNT indicates that it contains more individual nanotubes than others and has a higher homogeneous distribution. Homogeneously dispersed MWCNT has increased the strength with strong interface adhesion between MWCNT and epoxy.^13,61 Uniform dispersion improves the properties of nanocomposites.⁶² The extra surface is created by homogeneously distributed CNTs between the matrix and fiber, which provides better interfacial adhesion.⁶³ However, the agglomeration possibility increases for the mixtures with 1% and 1.5% MWCNT. The quantity of individual nanotubes decreases, and the absorption peak suddenly drops.

Consequently, an increase in agglomeration lowers the tensile strength of the sample materials. It is reported that the decrease in the number of individual nanotubes indicates agglomeration in the samples. The decrease in tensile strength value of NCF-CFRP composite reinforced with a higher amount of MWCNT than 1% could be attributed to the rise in the agglomeration, lower homogeneity, and poor distribution. The reduction in strength at a higher amount of MWCNT is that agglomeration of MWCNT disrupts load transfer between MWCNTs reinforcement and the epoxy matrix and behaves as defects.^29,49,64 Agglomeration provides a stress distribution that weakens the matrix, reducing the tensile properties.⁶⁰

Figure 11 shows the relationship between failure strain and MWCNT concentration. For the samples having 0.5% MWCNT, the failure strain is higher due to crack formation and propagation delay.^62,65 That indicates that MWCNTs in the matrix contribute to the absorption of strain energy generated during the tensile test.⁶⁶ The addition of MWCNTs improves the failure strain.⁶⁷ The lower failure strain observed for 1% and 1.5% MWCNT samples indicates weak interface bonding caused by agglomeration. A crack forms easily and propagates fast.⁶⁸ It was causing the composite to become more brittle—the fiber and surrounding matrix gap caused local strain differences, leading to discontinuous strains.⁵¹ Agglomeration of MWCNT can cause weak interfacial bonding between MWCNT and matrix; the load transfer from the matrix to the MWCNTs is less for the MWCNT to be fractured under tensile loading.⁶⁹

MWCNTs act as a bridge between matrix and fibers. This property restricts crack growth and increases its strength (Figure 12(a)). A similar phenomenon has also been reported in the literature.^26,64,70 In samples with 0.5% MWCNT, epoxy adherence to the fiber's surface is enhanced (Figure 12(b)), while in samples with 1% and 1.5% MWCNT, the fiber surfaces are smooth (Figure 12(c)–(e)). That is evidence of poorer fiber-matrix interfacial adhesion, which causes weak binding of the fibers to polymer matrix, consequently causing delamination and pulling out of fibers from the matrix. This situation can be seen in Figure 12(c). Agglomerates act as a stress raiser and crack formation, and rapid propagation is promoted, reducing the strength of the material.⁶⁸

Tensile fracture behavior

Figure 12 shows FESEM images of the fracture surface of the MWCNT-added NCF composites sample. From FESEM images, the MWCNTs in the matrix are attached to the fiber. MWCNTs act as a bridge between matrix and fibers. This property restricts the crack growth and increases the strength (Figure 12(a)). A similar phenomenon has also been reported. The fracture surface of the neat CFRP composite (0% MWCNT) is brittle. In comparison, the sample's surface containing 0.5% MWCNT is ductile due to good interfacial bonding (Figure 12(d)–(a)). Good interfacial adhesion helps efficient load transfer from the matrix to the nano-filler.⁶⁴ Figure 12(e), which corresponds to composite with 1.5 wt% MWCNT, shows neat resin and resin with MWCNT on the fracture surface. That indicates that the homogeneity of the MWCNT distribution is inadequate. Figure 12(e) illustrates interfacial debonding between fiber and matrix. MWCNTs are needed to achieve resistance to delamination in composites.

Conclusion

This study investigated the effect of 0.5%, 1%, and 1.5% weight ratios of MWCNTs on tensile, electrical, and thermal characteristics of Non-Crimp Fabric CFRP composites. The hybrid composite material obtained by adding MWCNT to Non-Crip CRP has significantly improved tensile properties and thermal and electrical conductivity. The general results obtained are summarized below.

• The absorbance value of 0.5% MWCNT is higher than other concentrations at 330 nm.

• Pure and 0.5 wt% MWCNT composites showed a slight reduction in permeability within 3250–3500 cm−¹. On the other hand, 1.0% and 1.5% wt MWCNT samples contain more COOH, which can react with OH.

• The FT-IR analysis results substantiate that a reaction between epoxy and COOH-functionalized MWCNTs produces a new epoxy-MWCNT nanocomposite.

• The 1.5% MWCNT reinforced material was the best conductor at all temperatures. At the same time, with the increasing temperature, conductivity increased. At 10,000 Hz frequency, 1.5% MWCNT reinforced material shows a conductivity of about 2 × 10⁻⁷ at 20°C, even though it showed a conductivity of about 6 × 10⁻⁷ at 120°C. However, it is observed that the conductivity value of 0.5% MWCNT-reinforced material increases significantly at low frequencies such as 100–10,000 Hz.

• Tests and analyses performed on the produced samples showed that the optimum COOH-MWCNT amount for the nanocomposite production method used, which improves the tensile properties of the nanocomposite, is 0.5 wt%. Adding surfactant to the epoxy increased the tensile strength by approximately 9.2%.

• Tensile samples fracture surfaces and show how individual COOH-MWCNTs interact with the fibers and form a bridge with the matrix. At the same time, agglomerations cause a weak bond between the fiber and matrix, thus causing delamination. Therefore, brittleness increases.

Footnotes

Acknowledgments

The authors are grateful for the support of the Universities mentioned above and the Companies.

Author contributions

H. Yakut Pektürk: Manufacturing the materials, characterization, and writing. B. Demir: Manufacturing the materials, characterization, and writing. C Bilgi: Characterizations and writing. F. Öz: Manufacturing. N. Ersoy: Manufacturing the materials.

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) disclosed receipt of the following financial support for the research, authorship, and publication of this article: The authors thank Karabuk University Scientific Research Projects Coordinators (KBÜ-BAP) for financially supporting this study with the project numbered KBÜBAP-17-DR-261. Boğaziçi University, Mechanical Eng. Composite Research Laboratory and Kırklareli University Mechanical Eng. Laboratories were used. In addition, METYX Composites and Wax Chemistry supplied carbon fabrics and dispersants.

Consent to participate

Contributions from each of the authors of this article are acknowledged. This article has not been published or uploaded anywhere before. The authors of this article are aware that the article will be uploaded and published in this journal.

ORCID iD

Bilge Demir

Data Availability Statement

The data given in this article are the datasets generated during and/or analyzed during the current study and are available from the corresponding author upon reasonable request. This study uses similar materials with reference .

References

Zakaria

Hilmi

Hamid

ARA

. Review of mechanical properties: materials being subjected by explosive compaction and explosive welding. AIP Conf Proc. 2021; 2347: 020235. doi:10.1063/5.0051931.

Rodrigo

Antonio

Coelho

. Effect of carbon nanotubes addition on the mechanical and thermal properties of epoxy matrices. Mater Res 2008; 11: 347–352.

Kumar

Ghosh

Yadav

, et al. Thermo-mechanical and anti-corrosive properties of MWCNT/epoxy nanocomposite fabricated by innovative dispersion technique. Compos B Eng. 2017; 113: 291–299. doi:10.1016/j.compositesb.2017.01.046.

Haberkern

. Tailor-made reinforcements. Reinforc Plast. 2006; 50: 28–33. doi:10.1016/S0034-3617(06)70974-2.

Stankovich

Dikin

Dommett

GHB

, et al. Graphene-based composite materials. Nature. 2006; 442(7100): 282–286. doi:10.1038/nature04969.

Soutis

. Carbon fiber reinforced plastics in aircraft construction. Mater Sci Eng. 2005; 412: 171–176. doi:10.1016/J.MSEA.2005.08.064.

Bibo

Hogg

Backhouse

, et al. Carbon-fibre non-crimp fabric laminates for cost-effective damage-tolerant structures. Compos Sci Technol. 1998; 58: 129–143. doi:10.1016/S0266-3538(97)00106-1.

Edgren

Asp

. Approximate analytical constitutive model for non-crimp fabric composites. Compos Part A Appl Sci Manuf. 2005; 36: 173–181. doi:10.1016/J.COMPOSITESA.2004.06.007.

Ferreira

LMM

. Study of the behaviour of non-crimp fabric laminates by 3D finite element model. Sevilla: (Tesis Doctoral Inédita). Universidad de Sevilla, 2012. doi:10.17616/R31NJNEG.

10.

Middendorf

Metzner

. Aerospace applications of non-crimp fabric composites. Non-Crimp Fabric Composites. 2011; 441–449e. doi:10.1533/9780857092533.4.441.

11.

de Greef

Gorbatikh

Godara

, et al. The effect of carbon nanotubes on the damage development in carbon fiber/epoxy composites. Carbon N Y. 2011; 49: 4650–4664. doi:10.1016/J.CARBON.2011.06.047.

12.

Dumont

Weimer

. Non-crimp fabric: preforming analysis for helicopter applications. Non-crimp fabric composites. Sawston, Cambridge: Woodhead Publishing, 2011, pp. 449–460. doi:10.1533/9780857092533.4.449.

13.

Gantayat

Rout

Swain

. Mechanical properties of functionalized multiwalled carbon nanotube/epoxy nanocomposites. Mater Today Proc. 2017; 4: 4061–4064. doi:10.1016/J.MATPR.2017.02.308.

14.

Guo

Zhang

Gao

, et al. Significantly improved electrical and interlaminar mechanical properties of carbon fiber laminated composites by using special carbon nanotube pre-dispersion mixture. Compos Part A Appl Sci Manuf. 2017; 95: 294–303. doi:10.1016/J.COMPOSITESA.2017.01.021.

15.

Chakraborty

Plyhm

Barbezat

, et al. Carbon nanotube (CNT)-epoxy nanocomposites: a systematic investigation of CNT dispersion. J Nanoparticle Res. 2011; 13: 6493–6506. doi:10.1007/S11051-011-0552-3/FIGURES/11.

16.

Gojny

Wichmann

MHG

Fiedler

, et al. Influence of different carbon nanotubes on the mechanical properties of epoxy matrix composites – a comparative study. Compos Sci Technol. 2005; 65: 2300–2313. doi:10.1016/J.COMPSCITECH.2005.04.021.

17.

Allaoui

Bai

Cheng

, et al. Mechanical and electrical properties of a MWNT/epoxy composite. Compos Sci Technol. 2002; 62: 1993–1998. doi:10.1016/S0266-3538(02)00129-X.

18.

Yip

Lin

. Effect of multi-walled carbon nanotubes addition on mechanical properties of polymer composites laminate. Polym Polymer Compos. 2011; 19: 131–140. doi:10.1177/0967391111019002-313.

19.

Bao

Guo

Song

, et al. In situ preparation of functionalized graphene oxide/epoxy nanocomposites with effective reinforcements. J Mater Chem. 2011; 21: 13290–13298. doi:10.1039/C1JM11434D.

20.

Spitalsky

Tasis

Papagelis

, et al. Carbon nanotube–polymer composites: chemistry, processing, mechanical and electrical properties. Prog Polym Sci. 2010; 35: 357–401. doi:10.1016/J.PROGPOLYMSCI.2009.09.003.

21.

Guadagno

De Vivo

Di Bartolomeo

, et al. Effect of functionalization on the thermo-mechanical and electrical behavior of multi-wall carbon nanotube/epoxy composites. Carbon N Y. 2011; 49: 1919–1930. doi:10.1016/J.CARBON.2011.01.017.

22.

Abdalla

Dean

Adibempe

, et al. The effect of interfacial chemistry on molecular mobility and morphology of multiwalled carbon nanotubes epoxy nanocomposite. Polymer (Guildf). 2007; 48: 5662–5670. doi:10.1016/J.POLYMER.2007.06.073.

23.

Wang

Liang

Liu

, et al. Effective amino-functionalization of carbon nanotubes for reinforcing epoxy polymer composites. Nanotechnology. 2006; 17: 1551. doi:10.1088/0957-4484/17/6/003.

24.

Vaisman

Marom

Wagner

. Dispersions of surface-modified carbon nanotubes in water-soluble and water-insoluble polymers. Adv Funct Mater. 2006; 16: 357–363. doi:10.1002/ADFM.200500142.

25.

Korayem

Barati

Chen

, et al. Optimizing the degree of carbon nanotube dispersion in a solvent for producing reinforced epoxy matrices. Powder Technol. 2015; 284: 541–550. doi:10.1016/J.POWTEC.2015.07.023.

26.

Dehghan

Al-Mahaidi

Sbarski

, et al. Surfactant-assisted dispersion of MWCNTs in epoxy resin used in CFRP strengthening systems. J Adhes. 2014; 91: 461–480. doi:10.1080/00218464.2014.921159.

27.

Ranjbar

Feli

. Mechanical and low-velocity impact properties of epoxy-composite beams reinforced by MWCNTs. J Compos Mater. 2018; 53(5): 693–705. doi:10.1177/0021998318790049.

28.

Neto

JSS

Banea

Cavalcanti

DKK

, et al. Analysis of mechanical and thermal properties of epoxy multiwalled carbon nanocomposites. J Compos Mater. 2020; 54(30): 4831–4840. doi:10.1177/0021998320939232.

29.

Alawy

Abdlghany

Wafy

, et al. Mechanical analysis of functionalized MWCNTs reinforced biresin CR82 epoxy nanocomposite materials. J Compos Mater. 2021; 55: 125–135. doi:10.1177/0021998320946478.

30.

Bai

. The reinforcement role of carbon nanotubes in epoxy composites with different matrix stiffness. Compos Sci Technol. 2006; 66: 599–603. doi:10.1016/J.COMPSCITECH.2005.05.020.

31.

Patil

Patel

Sharma

. Effect of carbon nanotube on mechanical properties of hybrid polymer matrix nano composites at different weight percentages. Mater Today Proc. 2018; 5: 6401–6405. doi:10.1016/J.MATPR.2017.12.251.

32.

Azimpour-Shishevan

Akbulut

Mohtadi-Bonab

. Synergetic effects of carbon nanotube and graphene addition on thermo-mechanical properties and vibrational behavior of twill carbon fiber reinforced polymer composites. Polym Test. 2020; 90: 106745. doi:10.1016/J.POLYMERTESTING.2020.106745.

33.

Chang

Friedrich

, et al. Evaluation and visualization of the percolating networks in multi-wall carbon nanotube/epoxy composites. J Mater Sci. 2009; 44: 4003–4012. doi:10.1007/S10853-009-3551-3/FIGURES/10.

34.

Shivamurthy

Anandhan

Bhat

, et al. Structure-property relationship of glass fabric/MWCNT/epoxy multi-layered laminates. Compos Commun. 2020; 22: 100460. doi:10.1016/J.COCO.2020.100460.

35.

Sulong

Muhamad

Sahari

, et al. Electrical conductivity behavior of chemically functionalized MWCNTs epoxy nanocomposites. Eur J Sci Res 2009; 29: 13–21.

36.

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. doi:10.1016/J.COMPSTRUCT.2012.03.014.

37.

Ayatollahi

Shadlou

Shokrieh

, et al. Effect of multi-walled carbon nanotube aspect ratio on mechanical and electrical properties of epoxy-based nanocomposites. Polym Test. 2011; 30: 548–556. doi:10.1016/J.POLYMERTESTING.2011.04.008.

38.

Raza

Hussain

Khan

, et al. Properties of lightweight multi-walled carbon nanotubes (MWCNTs) nanocomposites. Int J Lightweight Mater Manuf. 2021; 4: 195–202. doi:10.1016/J.IJLMM.2020.09.003.

39.

Schulte

Chandrasekaran

Viets

, et al. New functions in polymer composites using a nanoparticle-modified matrix. Multifunctionality of polymer composites. Elsevier, 2015, pp. 875–902. doi:10.1016/B978-0-323-26434-1.00030-1.

40.

Wang

LLQ

. Effect of MWCNTs on mechanical properties of carbon fiber non-crimp fabric/epoxy laminates. Acta Mater Compos Sin 2012; 29: 16–22.

41.

Pekturk

Elitas

Goktas

, et al. Evaluation of the effect of MWCNT amount and dispersion on bending fatigue properties of non-crimp CFRP composites. Eng Sci Technol Int J. 2022; 34: 101081. doi:10.1016/J.JESTCH.2021.101081.

42.

Geng

Liu

, et al. Effects of surfactant treatment on mechanical and electrical properties of CNT/epoxy nanocomposites. Compos Part A Appl Sci Manuf. 2008; 39: 1876–1883. doi:10.1016/J.COMPOSITESA.2008.09.009.

43.

Dehghan

Al-Mahaidi

Sbarski

, et al. Effect of fabrication method on thermo-mechanical properties of an epoxy composite. J Adhesion. 2014; 90(5–6): 368–383. doi:10.1080/00218464.2013.842170.

44.

Rastogi

Kaushal

Tripathi

, et al. Comparative study of carbon nanotube dispersion using surfactants. J Colloid Interface Sci. 2008; 328: 421–428. doi:10.1016/J.JCIS.2008.09.015.

45.

Yue

Pircheraghi

Monemian

, et al. Epoxy composites with carbon nanotubes and graphene nanoplatelets – dispersion and synergy effects. Carbon N Y. 2014; 78: 268–278. doi:10.1016/J.CARBON.2014.07.003.

46.

Chang

. An investigation on the dynamic behavior and thermal properties of MWCNTs/FRP laminate composites. J Reinforced Plast Compos. 2010; 29: 3593–3599. doi:10.1177/0731684410379510.

47.

Wang

Ding

, et al. Enhancement of mechanical behavior of resin matrices and fiber reinforced polymer composites by incorporating multi-wall carbon nanotubes. Polym Test. 2021; 96: 107077. doi:10.1016/J.POLYMERTESTING.2021.107077.

48.

Grossiord

Regev

Loos

, et al. Time-dependent study of the exfoliation process of carbon nanotubes in aqueous dispersions by using uv−visible spectroscopy. Anal Chem. 2005; 77: 5135–5139. doi:10.1021/AC050358J.

49.

Cha

Jin

Shim

, et al. Functionalization of carbon nanotubes for fabrication of CNT/epoxy nanocomposites. Mater Des. 2016; 95: 1–8. doi:10.1016/J.MATDES.2016.01.077.

50.

Njuguna

Vanli

Liang

. A review of spectral methods for dispersion characterization of carbon nanotubes in aqueous suspensions. J Spectroscopy. 2015; 2015: 1–11. doi:10.1155/2015/463156.

51.

Gaurav

Singh

. Effect of pristine MWCNTs on the fatigue life of GFRP laminates-an experimental and statistical evaluation. Compos B Eng. 2019; 172: 83–96. doi:10.1016/j.compositesb.2019.05.069.

52.

Zhang

Wang

Yan

, et al. Interfacial morphology and bonding mechanism of explosive weld joints. Chin J Mech Eng. 2021; 34: 1–12. doi:10.1186/S10033-020-00495-7/FIGURES/8.

53.

Liu

Xiao

Liao

, et al. Interfacial design of carbon nanotube-polymer composites: a hybrid system of noncovalent and covalent functionalizations. Nanotechnology. 2007; 18: 165701. doi:10.1088/0957-4484/18/16/165701.

54.

Sanli

Benchirouf

Müller

, et al. Study of the humidity effect on the electrical impedance of MWCNT epoxy nanocomposites, impedance spectroscopy: advanced applications: battery research, bioimpedance, system design. : , 2018, pp. 25–32. DOI: 10.1515/9783110558920-002/MACHINEREADABLECITATION/RIS.

55.

Dresselhaus

Eklund

, et al. Carbon nanotubes. In: The physics of fullerene-based and fullerene-related materials. Dordrecht: Springer, 2000, pp. 331–379. DOI: 10.1007/978-94-011-4038-6_9.

56.

Park

Cheng

, et al. Thermal conductivity of MWCNT/epoxy composites: the effects of length, alignment and functionalization. Carbon N Y. 2012; 50: 2083–2090. doi:10.1016/J.CARBON.2011.12.046.

57.

Eksik

Bartolucci

Gupta

, et al. A novel approach to enhance the thermal conductivity of epoxy nanocomposites using graphene core–shell additives. Carbon N Y. 2016; 101: 239–244. doi:10.1016/J.CARBON.2016.01.095.

58.

Razlan

Helmi

Kudus

, et al. Comparative study of graphene nanoparticle and multiwall carbon nanotube filled epoxy nanocomposites based on mechanical, thermal and dielectric properties. Compos B Eng. 2017; 119: 57–66. doi:10.1016/J.COMPOSITESB.2017.03.023.

59.

Siddiqui

Khan

, et al. Manufacturing and characterization of carbon fibre/epoxy composite prepregs containing carbon nanotubes. Compos Part A Appl Sci Manuf. 2011; 42: 1412–1420. doi:10.1016/j.compositesa.2011.06.005.

60.

Zakaria

Abdul Kudus

Akil

61.

Panchagnula

Kuppan

. Improvement in the mechanical properties of neat GFRPs with multi-walled CNTs. J Mater Res Technol. 2019; 8: 366–376. doi:10.1016/j.jmrt.2018.02.009.

62.

Haesch

Clarkson

Ivens

, et al. Localization of carbon nanotubes in resin rich zones of a woven composite linked to the dispersion state. Nanocomposites. 2015; 1: 204–213. doi:10.1080/20550324.2015.1117306.

63.

Gantayat

Rout

Swain

. Carbon nanomaterial–reinforced epoxy composites: a review. Polym Plast Technol Eng. 2018; 57: 1–16. doi:10.1080/03602559.2017.1298802.

64.

Tariq

Shifa

Baloch

. Mechanical and thermal properties of multi-scale carbon nanotubes–carbon fiber–epoxy composite. Arabian J Sci Eng. 2018; 43: 5937–5948. doi:10.1007/s13369-018-3091-8.

65.

Yang

Lee

Hsieh

, et al. Mechanical property of multiwall carbon nanotube reinforced polymer composites. Polym Polym Compos. 2018; 26: 99–104. doi:10.1177/096739111802600112.

66.

Grimmer

Dharan

CKH

. High-cycle fatigue of hybrid carbon nanotube/glass fiber/polymer composites. J Mater Sci. 2008; 43: 4487–4492. doi:10.1007/s10853-008-2651-9.

67.

Tefera

Bright

Adali

. Flexural and shear properties of CFRP laminates reinforced with functionalized multiwalled CNTs. Nanocomposites. 2021; 7: 141–153. doi:10.1080/20550324.2021.1961507.

68.

Song

Youn

. Influence of dispersion states of carbon nanotubes on physical properties of epoxy nanocomposites. Carbon N Y. 2005; 43: 1378–1385. doi:10.1016/j.carbon.2005.01.007.

69.

Schadler

Giannaris

Ajayan

. Load transfer in carbon nanotube epoxy composites. Appl Phys Lett. 1998; 73: 3842–3844. doi:10.1063/1.122911.

70.

Subhani

Latif

Ahmad

, et al. Mechanical performance of epoxy matrix hybrid nanocomposites containing carbon nanotubes and nanodiamonds. Mater Des. 2015; 87: 436–444. doi:10.1016/j.matdes.2015.08.059.