Carbon fiber-reinforced polymers (CFRPs) are finding increasing applications across various sectors such as aerospace (including rockets, aircraft, and air taxis), energy (wind turbine blades, hydrogen storage tanks, etc.), automotive (particularly in electric vehicles), civil engineering (including pedestrian bridges, CFRP rebars, and traffic poles), naval applications (such as unmanned stealth ships), and sports equipment (such as tennis rackets and bicycles) [1][2][3][4][5][6][7]. The increased applications of CFRPs are due to their exceptional attributes including lightweight, high-specific strength and stiffness, and low thermal expansion [1,3,6].
Among CFRPs, spread tow carbon fiber composites have recently gained significant attention due to their unique properties, including lighter weight and straighter fibers, which enhance the mechanical performance [8,9]. However, like many CFRPs, they face challenges in compressive and shear strength, primarily due to weak interlaminar shear strength (ILSS) and low interfacial shear strength (IFSS) at the fiber/matrix interface [10][11][12][13][14]. To address this, reinforcing the interlaminar region is vital for improving load transfer efficiency between the plies in the laminate, consequently enhancing the overall mechanical properties of CFRPs.
Numerous approaches have been reported to enhance the interlaminar regions of CFRPs. Among these approaches, carbon nanotubes (CNTs) are often used due to their extraordinary physical, chemical, and mechanical properties including a very large specific surface area (up to 1,300 m 2 /g), high modulus (up to 1 TPa), and high strength (up to 150 GPa) [15][16][17]. The prevalent approach reported in the literature is to disperse CNTs in the polymer matrix, which is used to impregnate carbon fibers to make composites improve the matrix dominant properties such as flexural strength/modulus and ILSS of composites. This process usually starts with mechanically mixing CNTs with a polymer matrix, and then using the resin to prepare composites to take advantage of the extraordinary properties of CNTs in CFRPs. A large body of work [18][19][20][21][22][23] has demonstrated interlaminar strength improvement by adding CNTs to the matrix of composites. However, this approach usually leads to agglomeration of CNTs, which prevents their uniform dispersion from manufacturing high-quality composites. For instance, Sanchez et al. [22] introduced CNTs to the matrix of a carbon fiber composite with different CNT loadings by calendaring using a three-roller mill and characterized the mechanical properties of composites. Their results showed that with 0.1 wt% CNTs doping in the epoxy matrix, the ILSS increased by 5.6%, and a slight increase was observed in tensile and flexural strength. A slightly higher dosage of CNTs (such as 0.3 wt%) did not yield performance improvement in the composites anymore. Similarly, Santos et al. [24] also mixed CNTs into the matrix of a CFRP, and they reported that a higher dosage (0.089 wt%) of CNTs does not benefit the improvement of mode-I fracture toughness of CFRPs compared with a lower dosage (0.043 wt%) of CNTs mixed with CFRP.
CNT grafted on carbon fibers is another way to improve the interfacial and interlaminar properties of CFRPs [25][26][27][28][29][30][31][32]. Grafting CNT on a carbon fiber is a chemical process in vapor deposition. This process is usually conducted at elevated temperatures and involves catalyst coating and formation, which tends to degrade the axial properties of the carbon fibers. Numerous groups [33][34][35][36][37] have observed that both the tensile strength and modulus of carbon fiber reduced after the chemical vapor deposition (CVD) process. For example, Zhang et al. [35] reported a 75% reduction in the tensile strength of carbon fiber after the CVD process. In addition, high CNT grafting density hinders the polymer matrix from flowing onto the carbon fiber surfaces, which resulted in weak interface bonding strength [29], and the uniformity of CNT growth on carbon fibers was not always satisfactory.
To overcome the limitations of the traditional approaches in enhancing CFRP properties, recent research has turned to the use of carbon nanotube (CNT) sheets. Liu et al. [38] and Mu et al. [39] have demonstrated the potential of incorporating CNT sheets in polymer fiber composites, such as CFRPs, to improve the mechanical and electrical performance. However, a challenge arises due to the inherent thinness of CNT sheets pulled from CNT forests. When integrated into continuous fiber composites, these sheets impact only a very thin interphase region between adjacent plies, as indicated in references [40] and [41], leading to limited performance enhancement.
Interestingly, CFRPs constructed with thinner ply thickness have shown increased strength and a reduction in transverse cracks compared to those with thicker plies [7,42]. This suggests a promising direction for further improvement: using even thinner plies in CFRPs. Such an approach would allow the CNT sheets to penetrate a larger volume fraction of the composite material. Consequently, this could not only provide further mechanical enhancements but also potentially improve the electrical properties of the composites. This approach represents a significant step toward fully leveraging the benefits of CNT sheets in advanced composite materials like CFRPs.
While carbon fibers are highly electrically conductive, their polymer matrix composites are in general not electrically conductive. As such, in applications where lightning protection and de-icing are needed, additional metal network is needed.
To reduce the weight of the metal network, numerous researchers have made attempts to enhance the electrical conductivity. The electrical conductivity of CFRPs is highly anisotropic and depends on the orientation of carbon fibers. In the carbon fiber direction, CFRPs are electrically conductive because of the inherent conductivity of carbon fibers but they are insulated in both transverse and thickness directions due to the intrinsic electrical insulation of the polymer matrix. To resolve this issue, CNTs, with high electrical conductivity, are often introduced. When CNTs having a good dispersion reach a percolation threshold or higher, the composites exhibit electrical conductivity [43][44][45].
To address the above-mentioned issues, we herein report the insertion of ultrathin well-aligned CNT sheets into the neighboring plies of thin spread tow carbon fiber fabric to make composites. The aerogel CNT sheets pulled from CNT forests were inserted into the interlaminar region between adjacent plies of CFRPs, followed by ethanol sorption and desorption to form CNT xerogel for enhanced mechanical properties. In spread tow carbon fiber fabrics, spread tow tapes are aligned in both warp and weft directions. Such spread tow fiber plies are much thinner than a typical carbon fiber tow fabric so that more CNT sheets can be inserted between adjacent plies to provide additional enhancement. Moreover, spread tow fabric does not usually contain crimps, which reduce the risk of the formation of possible pin holes and resin shrinkage area [46,47]. The reduced pin holes and shrinkage will further lead to a strong adhesion with the inserted CNT sheets and matrix, leading to increased interlaminar performance. This increased adhesion is also anticipated to increase the fracture toughness when a crack is along the interface between two neighboring plies. Due to the well-aligned structure in CNT sheets, electrical conductivity can be achieved even when the CNT loading is below the percolation threshold reported in the literature, which will be examined in this work.
We made CFRP composites by inserting ultrathin well-aligned CNT sheets at the interface between neighboring plies of the spread tow carbon fiber fabrics. We measured the compressive strength, IFSS, and fracture toughness of the CNT sheet-inserted composites. The mechanisms for the enhancement of fracture toughness were examined by fractographic analysis. The electrical conductivity of CNT sheet-inserted composites was also measured. This paper is organized as follows: the materials used and characterization method are described in Section 2, the results are presented and discussed in Section 3, and conclusions are drawn in Section 4.
In this section, CNT sheets are drawn from the CNT forest grown on a silicon wafer and are subsequently inserted into the interlaminar region of CFRPs. We describe the preparation and insertion of CNT sheets into the CFRPs to prepare composite specimens. We also describe the characterization methods for mechanical properties and electrical conductivity.
The TeXtreme 12k UTS 50 spread tow carbon fiber fabrics were used to make composites. The spread tow fabric has 80 grams per square meter (gsm) in the warp direction and 40 gsm in the weft direction, and the fabric was constructed with a 20-mm width tape in a woven pattern. The matrix material was Epon 862 epoxy resin from Westlake Epoxy, and the hardener was Epikure curing agent W. Ultrahigh purified multi-walled CNT sheets were drawn from 300-μm tall, vertical-aligned CNT forest grown on a 6-inch silicon substrate provided by Lintec NSTC. The sheet has an aerial density of approximately 0.025 gsm, and the CNTs have an outer diameter of approximately 10 nm [48].
Multi-walled CNT sheets were drawn from the CNT forest (Figure 1a), which were then coupled with spread tow carbon fiber fabrics along the warp direction of the fabrics. Both the spread tow carbon fiber fabrics and CNT sheets were bonded together by sorption and desorption of ethanol, which consolidated the as-drawn CNT aerogel sheets to CNT xerogel sheets.
Subsequently, CNT-spread tow carbon fiber fabrics were dried at ambient temperature overnight, and the dried CNT-spread tow carbon fiber fabric is shown in Figure 1b. Thus, the CNT sheets were maintained in the well-aligned structure during fabrication. The surface morphology of CNT-spread tow carbon fiber fabrics was observed under a scanning electron microscope (SEM) as shown in Figure 1c and 1d. The SEM image in Figure 1c showed that CNT sheets were successfully wrapped around carbon fiber filaments. The SEM image in Figure 1d showed the as-received carbon fiber without CNT. Only approximately 60% of spread tow carbon fiber fabrics were covered with CNT sheets (Figure 1b) so that the baseline spread tow CFRPs were made on the remaining uncovered 40% under identical processing conditions with the CNT-spread tow CFRPs. The dried CNTspread tow carbon fiber fabrics were then stacked together (Figure 1e) to fabricate testing coupons (Figure 1f) via vacuumassisted resin infusion molding (VARIM). The curing temperature was set to 75°C, and the curing time was set to 20 hours following a two-hour post-cure at 120°C. Twenty-three plies of spread tow carbon fiber-CNT fabrics were laid up through the stacking sequence in Figure 1e to make a 3.2-mm thick plate for the combined loading compression (CLC) test and the short beam shear (SBS) test (for testing interlaminate shear strength). A total of 32 spread tow carbon fiber fabric layers were stacked together in the same sequence to make CFRP plates for the mode-II fracture toughness. Both the 16th and 17th plies were coupled with a monolayer of ultrathin CNT sheet; these two CNT-CF plies were consolidated and dried before stacking together. A 13-μm thick fluorinated ethylene propylene (FEP) film was inserted with care between the 16th ply and 17th ply as a delamination initiator. The fabricated testing coupons were then cut into testing specimens for testing as shown in Figure 1g-i.
The compressive strength along the fiber warp direction (80 gsm) was determined by the CLC test following ASTM D6641 [49], and the dimensions of each specimen were 140 × 13 × 3.2 mm. The compression test was conducted under the displacement-controlled condition at 1.3 mm/min. Four specimens were tested for each composite. The fractured surface of a failed specimen was observed under a Leica DMi8 inverted optical microscope to investigate the failure mode of the specimen.
The ILSS of the CNT-spread tow carbon fiber composite was determined using the SBS test following ASTM D2344 [50]. Short-beam specimens were cut along the warp direction, and the nominal dimensions of each specimen are 30 × 10 × 3.2 mm. The actual dimensions of each specimen were measured and used in the calculation of ILSS. The span-to-depth ratio was set to 6:1 instead of 4:1 to prevent local crushing/buckling to maintain a large zone of uniform shear stress [14]. The central loading was provided by a 10-mm pin to prevent any indentation into the surface of the beam specimen. The diameter of the two support pins was 3 mm. The displacement rate applied by the loading nose in the center of the beam specimen was 1 mm/min in all the SBS tests. Four specimens were tested for each composite.
Mode-II fracture toughness test was conducted following ASTM D7905 [51]. The end-notched fracture (ENF) specimens were cut by waterjet out of both the baseline plate and 0.025 wt% CNT sheet-inserted plate, respectively. The specimens have a nominal dimension of 210 × 20 × 4.46 mm, with a delamination initiator of 70 mm. The side of the specimen facing the camera was sprayed with a thin layer of speckles for visual determination of the crack. The specimens were marked with vertical lines at 20, 30, and 40 mm from the tip of the insert before the mode-II fracture test. The schematic diagram of the ENF specimen is shown in Figure 2a, and the test configuration is shown in Figure 2b. An Instron 5900 series testing apparatus (Instron, Norwood, MA, USA) was used as the testing platform. The load was applied to the specimen by a 3-point bending fixture, and the loading pin and the support pins have a diameter of 10 mm. The loading pin was placed 20, 30, and 40 mm from the tip of the insert, respectively, while the span was 100 mm. A deflectometer was used to measure the actual displacement at the center of the specimen, and a camera was used to record the crack initiation and propagation. The specimen compliance was calibrated using the method described in ASTM D7905, and the displacement rate was 0.5 mm/min. A total of three specimens were tested for each composite. The mode-II interlaminar fracture toughness was calculated via the compliance calibration method (CCM) using the equation: 22 max 0 IIc 3 2 mP a G B = , where IIc G is the mode-II interlaminar fracture toughness, m is the compliance calibration coefficient, max P is the maximum force in the load-deflection curve, 0
a is the initial crack length, and B is the specimen width.
For examination of the morphology of the specimen after mode-II delamination, a Zeiss SIGMA 500 VP scanning electron microscope (Zeiss, Oberkochen, Germany) was used to conduct the fractography analysis on failed samples for both composites. The samples were cut from the delamination growth region of the ENF specimens with the same distance from the end for both the baseline composite specimen and 0.025 wt% CNT-spread tow CFRP specimen.
The electrical conductivity was measured using a Keithley 2400 source meter (Tektronix, Beaverton, OR, USA). The resistance of the samples was measured along the fiber warp direction by applying voltages between the probes, and the electrical conductivity of the CNT-spread tow carbon fiber composite along the fiber warp direction is calculated by
, where S is the electrical conductivity of the sample, L is the length between two probes, R is the resistance, and A is the cross-sectional area of the sample. Three specimens for each composite were used for measuring the electrical conductivity, and ten measurements were made from each specimen.
The carbon fiber volume fractions, compressive strength along the fiber warp direction, the ILSS, and mode-II fracture toughness of the CFRPs with and without inserted CNT sheets were investigated. Results are reported and discussed in this section.
Optical micrographs as shown in Figure 3 show that the carbon fiber composites with and without insertion of CNT sheets have almost identical carbon fiber volume fractions at approximately 67%, and no difference is seen between the baseline composites (without insertion of CNT sheets) and CNT-spread tow composites, which indicated that the introduction of CNT sheets does not change the structure of CFRPs.
The compressive strength of spread tow carbon fiber composites with different CNT loadings is plotted in Figure 4. The compressive strength along the warp direction of spread tow carbon fiber composite improved with the insertion of CNT sheets reinforcement in the interlaminar region; a 0.025 wt% CNT loading in the composite yields an increase in the compressive strength by 14.7%, and the composite with 0.05 wt% CNT loading shows a 6.6% increase in compressive strength. The compressive strength of composites is governed by the shear strength in the interphase region with a thickness of 100 nm or less [40,41]. The introduction of two layers of CNT sheets increased the asperity and mechanical interlocking of this region, thus directly enhancing the mechanical behavior. The relatively straighter carbon fibers in the spread tow carbon fiber fabrics compared with the regular tow carbon fiber fabrics reduced the possibility of forming wrinkles and folds, resulting in higher compressive strength when compared with the traditional woven carbon fiber composites. The compressive strength of baseline spread tow CFRPs is 551.6 MPa. For the 0.025 wt% CNTspread tow CFRPs prepared herein, the compressive strength is 632.5 MPa, representing 14.7% enhancement over the baseline composites.
It is noted that the compressive strength of the regular tow woven CFRPs is in the range of 400-500 MPa [52,53]. However, multiwalled CNT sheets were joined by van der Waals force between neighboring CNTs and neighboring CNT sheets and, as such, exhibited low normal and shear strength [54,55]; with the increase in the number of CNT layers, the shear load transfer decreases [56]. Therefore, the increase in the number of CNT sheets from two to four leads to a lower compressive strength compared with the case of the insertion of the two CNT sheets into the composite.
The failure mode of the CNT-spread tow carbon fiber composites after the CLC test was observed under an optical microscope. The sample after the CLC test was first embedded in epoxy with care and then polished for optical microscopy, and the micrographs are shown in Figure 4b. The two-dimensional woven structure of the CNT-spread tow carbon fiber composite leads to a combined delamination and shear failure mode. The adjacent horizontal carbon fiber tows shift down and separate from the vertical carbon fiber tows upon longitudinal compression, leading to delamination between plies as shown in the arrowpointed area of the failed specimen in Figure 4b, while in the fiber longitudinal direction, the kink-band was observed in the fiber tows, which is induced by microcracking under the applied axial load. However, no difference was observed between the CNT sheet-inserted carbon fiber composite and the baseline composite.
The load-deflection curves of both the baseline and CNT-spread tow carbon fiber composites under the SBS testing are plotted in Figure 5a. Laminate mismatches were observed at the edge of the SBS specimens for both composites as shown in the inset image of Figure 5a. A nearly linear increase of the curves was observed up to 0.3 mm deflection, then delamination occurred after the curves reached the maximum load. The spread tow composite with CNT sheets inserted between adjacent plies also exhibits a higher stiffness as depicted in the load-deflection curves in Figure 5a. This illustrates that the nanocomposite formed by the insertion of CNT sheets in the interlaminar region of spread tow carbon fiber composites provided improved bending stiffness and critical load at the onset of instability to the in-plane shear load during the SBS test. The ILSS of composites with different CNT loading ratios are plotted as a bar chart in Figure 5b. Interleaving CNT sheets with spread tow carbon fiber/epoxy composite shows an increase in the ILSS, while a higher concentration of CNTs does not provide additional improvement of the ILSS anymore. The CNT-spread tow carbon fiber composite with 0.025 wt% CNT has the highest ILSS (48.4 MPa), which is 33.0% higher than that of the baseline composite.
The 0.05 wt% CNT-spread tow carbon fiber composite exhibits an increase of ILSS by 22.0% compared with the baseline composite. The electrical conductivity of the CNT-spread tow carbon fiber composite was measured using the two-probe method along the warp direction of the spread tow carbon fiber fabrics. The results are plotted as a bar chart in Figure 6. The electrical conductivity improved by introducing CNT sheets to the interlaminar region of spread tow carbon fiber composites. Adding more CNT sheets between plies yields higher electrical conductivity due to the wellaligned structure and high electrical conductivity of CNT sheets, and straighter carbon fibers in spread tow carbon fiber fabric.
Figure 6 • Electrical conductivity of the CNT-spread tow carbon fiber composites with different CNT loadings. The electrical conductivities were measured along the warp direction of spread tow carbon fiber composites, and 0.05 wt% CNT-spread tow carbon fiber composite exhibits the highest electrical conductivity.
Compared with the baseline spread tow carbon fiber composite, the electrical conductivity increased by 16.7%, from 996.0 S/m to 1,162.4 S/m for 0.025 wt% CNT-spread tow carbon fiber composite. For the 0.05 wt% CNT-spread tow carbon fiber composite, the electrical conductivity increased by 64.7%, from 996.0 S/m to 1,640.3 S/m. The increased electrical conductivity in CNT-spread tow carbon fiber composite demonstrates their potential for applications in structural health monitoring [57,58], de-icing, and lightning protection in aircraft and wind blades [59,60].
The above results indicated that the insertion of CNT sheets into the spread tow carbon fiber composites improved their mechanical performance, and a preeminent improvement was demonstrated for spread tow carbon fiber composite with 0.025 wt% CNT loading. As such, the mode-II fracture toughness was compared only between the baseline composite and the composites with 0.025 wt% CNT loading in the following section.
The ENF specimens of the baseline and 0.025 wt% CNT CFRPs were conditioned at 60°C for eight hours before the mode-II delamination test, and then all specimens were tested under room temperature at a humidity of 50% ± 10%. The specimen compliance calibration at the loading point was performed following ASTM D7905 [51], and the specimens were loaded and unloaded within their elastic region at crack lengths of 20 mm and 40 mm, respectively. The fracture test was conducted at a crack length of 30 mm. The compliance C at each crack length then was determined by a linear least squares regression from the displacement and force curves, and the compliance calibration coefficients A and m were calculated using the equation:
. The representative compliance versus crack length cube curves are shown in Figure 7; both curves yielded reasonable goodness of fitting, R-square of 0.9996 for baseline CFRP and R-square of 0.9997 for 0.025 wt% CNT CFRP. As illustrated in Figure 7, CFRP with 0.025 wt% CNT sheets insertion shows a higher m value than that of baseline CFRP; meanwhile, the A value is lower. The compliance coefficients of both baseline and 0.025 wt% CNT CFRPs are summarized in Table 1; the numbers in parentheses show the standard deviations for each coefficient. The representative force versus displacement curves of both composites were plotted as shown in Figure 8a; both composites exhibited linear elastic response until the peak forces. There were neither evident crack observations from the video nor apparent force drops before the peak force from the forcedisplacement curves; therefore, the peak force was identified as the corresponding force at the initiation of the crack under mode-II loading, and the values of peak forces were used to calculate mode-II interlaminar fracture toughness. J/m 2 . The interlaminar fracture toughness of CFRP increased by 34.6% due to the increased loading bearing and compliance after the insertion of CNT sheets in the interlaminar region. The increased toughness is attributed to two aspects: (1) CNT increases the Young's modulus of the epoxy matrix due to the formation of CNT/epoxy nanocomposite, which provides a gradual load transfer between plies to reduce stress concentration at the interphase region, and (2) the CNT provides intimate contact between two neighboring plies to increase the adhesion. This finding also agrees with the result obtained by the SBS test, in which case the ILSS was increased by 33% by introducing 0.025 wt% CNT sheets to the interlaminar region.
The fracture mechanism of both baseline and 0.025 wt% CFRPs under mode-II loading was examined with a high-resolution SEM. The failed ENF specimens for both CFRPs were cut in the crack propagated zone for SEM fractography analysis. The representative SEM fractographic images of baseline CFRP were shown in Figure 9, and the representative SEM fractographic images of 0.025 wt% CNT CFRP were shown in Figure 10. The fracture surface of the baseline composite was relatively clean and smooth, and a typical brittle fracture with hackle markings was observed due to the intensive in-plane shear loading. This shear-induced fracture feature also has been reported by others [61][62][63][64]. There was little resin left on the surface of the carbon fiber near the crack initiation zone. This indicated that the baseline composite has poor resistance to the in-plane shear force and resulted in a low mode-II interlaminar fracture toughness.
As for 0.025 wt% CNT CFRP, more resin was left on the surface of the carbon fiber after mode-II loading, and a similar hackle pattern away from the crack propagation direction was observed as shown in Figure 10a and 10b. The resin river pattern was also observed between stitches of carbon fiber filaments as shown in Figure 10c and 10d. The interlaminarly inserted CNT sheets increased the fracture toughness of the interphase region between neighboring plies. Moreover, the well-aligned structure of inserted CNT sheets formed a connected network as shown in Figure 10c and 10d that increases adhesion, and it helps retain resin on the surface of carbon fiber, which also can absorb more energy during crack initiation and propagation. In addition, the high stiffness of CNT sheets gives rise to relatively lower local strains under in-plane shear loads compared with the baseline without CNT, which helped bridge the in-plane shear-induced matrix cracking as observed in Figure 10d. This in turn delayed the propagation of the cracks compared with the baseline CFRP.
In this work, the well-aligned CNT sheets were inserted into the neighboring plies of spread tow carbon fiber fabrics to make composites via VARIM. The CNT sheets were maintained in a well-aligned structure in the composite. The composite formed with thinner spread tow fabrics and inserted well-aligned CNT sheets demonstrated a 14.7% improvement in compressive strength. The increase in compressive strength is attributed to a 33.0% enhancement in the ILSS and a 34.6% improvement in mode-II fracture toughness. Meanwhile, the maintained, wellaligned structure in CNT sheets in the spread tow composite provides significantly improved electrical conductivity by 64.7% with a low CNT loading even below the electrical percolation threshold for other CNT-filled composites. The results indicate that the insertion of well-aligned CNT sheets directly to the interlaminar region of a spread tow carbon fiber composite improved in the mechanical and electrical performance significantly. The optimal CNT loading was found to be 0.025 wt%, and a further increase in the loading of CNT sheets does not provide continuous enhancement of mechanical properties compared with the lower CNT loading. More research is required to understand the basic mechanism behind this effect.
ACADEMIA MATERIALS SCIENCE 2024, 1