Skip to content
BY-NC-ND 3.0 license Open Access Published by De Gruyter September 22, 2017

Effect of ultrasonic-assisted impregnation parameters on the preparation and interfacial properties of MWCNT/glass-fiber reinforced composites

Shaohua Zeng, Mingxia Shen, Pengpeng Duan, Fengling Lu, Shangneng Chen and Yijiao Xue
From the journal e-Polymers

Abstract

In this study, an ultrasonic-assisted impregnation method was employed to deposit carboxyl multiwalled carbon nanotubes (MWCNTs) onto the E-glass fiber fabric (GFf) for the preparation of the MWCNT-GFf reinforcer. The effects of ultrasonic power, duration and temperature on the dispersion of MWCNTs onto GFf were investigated, and the mechanical properties, interlaminar adhesion, and dynamic viscoelasticity of the resulting MWCNT-GFf-reinforced composites (MGCs) were evaluated. The results indicated that an effective dispersion of MWCNTs onto GFf without obvious breakage of the MWCNTs was achieved under an ultrasonic power of 600 W, duration of 6 min, and processing temperature of about 0°C. Compared with the GFf-reinforced composite, the tensile strength, flexural strength and interlaminar shear strength of the MGCs exhibited maximum increments of 38.4%, 34.6% and 47.1%, respectively. Moreover, the storage moduli and glass-transition temperatures of the MGCs were significantly enhanced. The ultrasonic parameters were of key importance for dispersing MWCNTs onto GFf and improving the interfacial properties of the composites.

1 Introduction

Glass-fiber-reinforced polymer (GFRP) composites have been widely used for structural and functional applications due to their high in-plane specific stiffness and strength, low density and moderate cost (1). Nevertheless, the potential application of GFRP composites is often restricted by poor interfacial adhesion between the fiber and matrix. Interfacial adhesion can be improved by surface treatments of the glass fiber (GF), such as enhancing surface chemical activity or increasing specific surface area (2). However, the fiber/matrix adhesion still remains weak even after such surface treatment (3).

In recent years, carbon nanotubes (CNTs) have been extensively studied for improving the interfacial properties of GFRP composites, which can be attributed to their high specific surface area, aspect ratio, stiffness and strength (4). Multiwalled carbon nanotubes (MWCNTs) have been widely used for composite applications due to their relatively low cost and availability on a large scale, even though their physical and chemical properties are often inferior to those of single-walled carbon nanotubes (5). The application of CNTs provides an opportunity for developing new multifunctional composites with optimized mechanical, thermal and electrical properties (6, 7). To efficiently improve fiber/matrix interfacial adhesion, the incorporation of CNTs with conventional GFs to create multiscale reinforcers has attracted great attention (8). In this way, high loadings of CNTs can be achieved while avoiding the filtering effect of woven GFs during the liquid composite molding process using techniques such as resin transfer molding (RTM) or vacuum-assisted resin transfer molding (VARTM) (9, 10). CNTs on the fiber surface can improve the fiber/matrix interfacial adhesion and allow efficient stress transfer from the matrix phase to the reinforcing GF. The use of functionalized CNTs helps in increasing the chemical interaction and wettability of polymer resins. Multiscale reinforcers have been prepared by a variety of techniques, such as chemical vapor deposition (CVD) (11), electrophoretic deposition (EPD) (12), the spray-up process (13), the use of sizing coatings (14) and dip coating (15). In particular, the in situ growth of CNTs on the GF surface using the CVD technique has been proposed. Nevertheless, high temperatures and additional catalysts are typically required during the CVD process (16), which may degrade the mechanical strength of the GF. The sizing coatings containing CNTs may bundle the GFs in the woven fibers, and these GFs remained adhered even after impregnating with resin and in the final composites, which interfered with the fiber/matrix interfacial adhesion (17). The solution-based methods, including EPD, dip coating and the spray-up process, have been extensively used to deposit CNTs onto the GFs. In this way, the chemical and physical modification of CNTs can ensure a well-dispersed state of CNTs in the liquid solution (18). To facilitate the chemical bonding between the modified CNTs and the GFs, in some cases the GFs were re-silanized after deliberately removing the original sizing coatings from the fibers (19, 20), which was however costly and is not practical for engineering composites.

Recent studies have mostly focused on the deposition of CNTs on GF tows (21), chopped GFs (22), or even single fibers (23), while only a few studies have used woven GFs as the substrate for the deposition of CNTs. Furthermore, the dual-scale pore structure in woven GFs is simply downplayed. Generally, the space between the GF filaments (~50 nm) is far lower than that between the fiber tows (~100 μm) (24). The small intrafiber space may easily block the CNTs at the top surface layer of the woven GFs, resulting in the undesirable re-agglomeration of the CNTs (25). Some studies have attempted to improve the penetration of the CNTs into the spaces between the GFs. Li et al. (26) deposited amine-functionalized MWCNTs onto insulating woven GFs using an EPD method. In the presence of an electric field, the charged MWCNTs were dispersed not only on the GF surface but also among the GF filaments. However, it is still difficult to achieve a homogeneous distribution of MWCNTs on the round GFs, due to the electrically insulating characteristics of the GF (27). Ku-Herrera et al. (28) developed a dipping-procedure-assisted ultrasonic dispersion method for depositing oxidized MWCNTs on commercial E-glass fiber tows. The use of ultrasound prevented the formation of MWCNT agglomerates and assisted the migration of the MWCNTs among the GFs. This method was effective for the deposition of a high density of CNTs on the engineered fibers. Despite these advances, additional efforts are needed to achieve well-dispersed CNTs on woven GF fabrics.

The aims of this study were to investigate the optimum ultrasonic parameters to uniformly disperse MWCNTs onto the E-glass fiber fabric (GFf) and inside the GF tows and to evaluate their effects on the interfacial properties of the resulting composites. To achieve these goals, an ultrasonic-assisted impregnation (UAI) method was used to deposit cost-effective industrial-grade carboxyl MWCNTs onto the commercial GFf. Moreover, three ultrasonic parameters (output power, duration and processing temperature) were chosen to optimize the dispersion procedure of MWCNTs. The effects of dispersion quality of MWCNTs on the interfacial properties were evaluated through the tensile and flexural properties, interlaminar adhesion and dynamic viscoelasticity of MWCNT-GFf-reinforced composites (MGCs).

2 Experimental

2.1 Materials

Unidirectional stitched E-glass fiber fabric with an areal density of 1200 g/m2 was supplied by Saertex Wagener GmbH & Co. KG (Saerbeck, Germany). Industrial-grade carboxyl MWCNTs (>95% purity, 20–40 nm external diameter, 30 μm length, 1.43 wt. % -COOH content) were purchased from Chengdu Organic Chemicals Co., Ltd. (Chengdu, China). Bisphenol-A epoxy (LY1564; 1200–1400 mPa·s viscosity) and its amine hardener (A3486) were obtained from Huntsman Advanced Materials Americas, Inc. (Woodlands, TX, USA) γ-Aminopropyltriethoxysilane (APS; ≥98% purity) was purchased from Nanjing Shuguang Chemical Group Co., Ltd. (Nanjing, China), and chemically pure ethanol was from Shanghai Chemical Reagent Factory (Shanghai, China).

2.2 Preparation of MWCNT-GFf reinforcers

The GFf was cut into several rectangular pieces (200 mm×300 mm) as the substrate for the deposition of MWCNTs. The preparation procedure for the MWCNT-GFf reinforcers, using an ultrasonic power of 200 W, a duration of 6 min, and a processing temperature of 25°C as an example, was as follows.

First, APS (2.5 mg) was added dropwise into ethanol (100 ml) with stirring. Carboxyl MWCNTs (50 mg) were dispersed into the APS/ethanol solution for 45 min using an ultrasonic bath at 600 W and 25°C, to obtain the APS-modified MWCNT suspension. Then, the pre-dried GFf substrate was immersed into the above suspension, and the resulting suspension was further agitated in the ultrasonic bath at 200 W for 6 min, maintaining a processing temperature of 25°C. Finally, the substrate was rinsed and dried under vacuum for 8 h at 90°C to afford the MWCNT-GFf reinforcer with APS-modified MWCNTs, which are referred to here as MG-P2D6T25.

To investigate the effect of the ultrasonic power on the interfacial properties of composites prepared under otherwise identical conditions (a duration of 6 min and a processing temperature of 25°C), MWCNT-GFf reinforcers were also prepared at ultrasonic powers of 600 W and 800 W, referred to as MG-P6D6T25 and MG-P8D6T25, respectively.

Additional MWCNT-GFf reinforcers were prepared using the optimum ultrasonic power of 600 W and ultrasound durations of 1 min, 12 min and 20 min, denoted MG-P6D1T25, MG-P6D12T25 and MG-P6D20T25, respectively. After selecting the optimum duration (6 min), MWCNT-GFf reinforcers were also prepared at temperatures of approximately 0°C (ice/salt bath) and 45°C, referred to as MG-P6D6T0 and MG-P6D6T45, respectively.

The preparation parameters of the MWCNT-GFf reinforcers fabricated in this study are summarized in Table 1. The content of MWCNTs on the GFf surface was evaluated by accurately weighing the fabrics before and after depositing the MWCNTs (18). The MWCNT content is expressed as the mass of MWCNTs in grams per square meter of GFf (g/m2) and is also listed in Table 1.

Table 1:

Preparation conditions of MWCNT-GFf reinforcers and their MWCNT contents.

SampleUltrasonic parameterMWCNT content (g/m2)
Power (W)Duration (min)Temperature (°C)
MG-P2D6T252006250.51±0.02
MG-P6D6T256006250.53±0.02
MG-P8D6T258006250.54±0.03
MG-P6D1T256001250.49±0.02
MG-P6D12T2560012250.54±0.03
MG-P6D20T2560020250.54±0.03
MG-P6D6T0600600.55±0.03
MG-P6D6T456006450.54±0.02

2.3 Fabrication of multiscale composites

The vacuum-assisted resin infusion molding (VARIM) technique was utilized to manufacture multiscale composites. The epoxy resin and its amine hardener at a mass ratio of 100:34 were mixed in a 500 ml beaker and stirred thoroughly. The epoxy mixture was vacuum degassed under vacuum at 40°C for 15 min. Two or four plies of MWCNT-GFf reinforcers were stacked in a sealed bag with a plies orientation of [0°/0°], and then the epoxy mixture was infused into this sealed bag to fully impregnate these reinforcers. After the vacuum infusion process, the MWCNT-GFf reinforcers impregnated with the epoxy mixture in the sealed bag was first cured at room temperature for 24 h, and then post-cured at 70°C for about 6 h to enhance the crosslinking degree of the composites.

Using an accurate calculation, the matrix mass content of the multiscale composites prepared by the above process was controlled at 28±0.5%. The multiscale composite prepared by GFf, MG-P2D6T25, MG-P6D6T25, MG-P8D6T25, MG-P6D1T25, MG-P6D12T25, MG-P6D20T25, MG-P6D6T0 and MG-P6D6T45 was hereafter marked as GFC, MGC-P2D6T25, MGC-P6D6T25, MGC-P8D6T25, MGC-P6D1T25, MGC-P6D12T25, MGC-P6D20T25, MGC-P6D6T0 and MGC-P6D6T45, respectively.

2.4 Measurement and characterization

2.4.1 Mechanical properties

All the mechanical tests were conducted on a universal testing machine CMT-5105 (Shenzhen Sans, Inc., Shenzheng, China). The tensile tests referred to the ISO 527:1997 standard, and type two specimen model was adopted with an extension rate of 2.0 mm/min. The tensile specimens were cut along the longitudinal direction of composites, and the outer dimensions of tensile specimen were 2 mm×25 mm×250 mm. The flexural tests referred to the ISO 14125:1998 standard and used the three-point bending model at a crosshead speed of 2.0 mm/min. The outer dimensions of flexural specimen were 2 mm×15 mm×40 mm, and the span was 32 mm. Each effective datum was the average of five specimens.

2.4.2 Interlaminar shear strength

The interlaminar shear strength (ILSS) was determined by short-beam shear tests referring to ISO 14130:1997, performed on the testing machine CMT-5105 with a cross-head speed of 1.0 mm/min. The outer dimensions of ILSS specimen were 4 mm×20 mm×40 mm, and the span was 20 mm. Each effective datum was the average of five specimens.

2.4.3 Scanning electron microscopy (SEM)

The morphologies of MWCNTs onto the fiber surface were investigated using field-emission SEM SU8010 (Hitachi, Hitachinaka, Japan). The fibers were selected from the inner parts of fiber tows in the MWCNT-GFf reinforcer to explore the MWCNTs dispersion. The fracture surfaces from ILSS tests were examined by field-emission SEM S-4800 (Hitachi, Hitachinaka, Japan).

2.4.4 Dynamic mechanical thermal analysis (DMTA)

The viscoelastic behavior of composites were measured by a DMTA Q800 (TA Instruments, New Castle, DE, USA) under three-point bending mode. Rectangular specimen (60 mm×10 mm×2 mm) was conducted under air from room temperature to 200°C at a heating rate of 5°C/min and a constant frequency of 1 Hz.

3 Results and discussion

3.1 Mechanical properties of multiscale composites

3.1.1 Effect of ultrasonic power

The fiber/matrix interfacial adhesion plays an important role in efficiently transferring stress from the matrix to the reinforcing fibers, which ultimately affects the mechanical properties of the resulting composites. The results of the tensile and flexural tests of the MGCs obtained under different ultrasonic powers are shown in Figure 1A and B, respectively.

Figure 1: Mechanical properties of MGCs obtained under various ultrasonic powers.(A) Tensile strength and modulus; (B) flexural strength and modulus.

Figure 1:

Mechanical properties of MGCs obtained under various ultrasonic powers.

(A) Tensile strength and modulus; (B) flexural strength and modulus.

As shown in Figure 1, with increasing ultrasonic power, both the tensile and flexural properties of the MGCs rose to a peak at an ultrasonic power of 600 W and thereafter declined. MGC-P6D6T25 exhibited the maximum values in terms of both the tensile and flexural properties. Furthermore, the tensile and flexural properties of MGC-P8D6T25 were obviously lower than those of MGC-P6D6T25.

The enhanced tensile and flexural properties of the MGCs were attributed to strong fiber/matrix interfacial adhesion caused by the presence of the MWCNTs. The potential for MWCNTs in improving the mechanical properties in such composites mainly depends on their dispersion state on the GFf surface. A modest ultrasonic power could effectively disperse MWCNTs onto the GFf and thus strengthen the mechanical properties of the composites. However, excessive ultrasonic power (≥800 W) may shear the MWCNTs and diminish their reinforcing effect. In addition, such high power inevitably caused the local breakage of GFs, leading to a reduction in the mechanical strength of the composites. Therefore, very low or high ultrasonic powers are not beneficial for producing high-performance composites.

3.1.2 Effect of ultrasound duration

The results of the tensile and flexural tests of the MGCs obtained under different durations of exposure to the ultrasound are shown in Figure 2A and B, respectively. As seen in Figure 2, both the tensile and flexural properties of MGC were initially improved by increasing the ultrasound duration. However, at durations over 6 min, the tensile and flexural properties of MGC both began to decrease. As such, MGC-P6D6T25 still exhibited the maximum values for tensile and flexural properties.

Figure 2: Mechanical properties of MGCs obtained under various ultrasound durations.(A) Tensile strength and modulus; (B) flexural strength and modulus.

Figure 2:

Mechanical properties of MGCs obtained under various ultrasound durations.

(A) Tensile strength and modulus; (B) flexural strength and modulus.

These results indicated that modest ultrasound durations afforded good dispersion of the MWCNTs and even resulted in increasing MWCNT content among the GF filaments. The MWCNTs anchored on the outside of GFs could engage in mechanical interlocking with the matrix and furthermore formed covalent bonds with the epoxy on curing, all of which enhanced the interfacial adhesion of the composites. However, long durations (>20 min) may cause the MWCNTs to re-agglomerate among the GF filaments due to their high aspect ratio and strong van der Waals forces. These MWCNT agglomerates could obstruct the microflow channels to weaken the impregnation of GFs. Moreover, long durations may damage the GF and the sizing coatings on the GF surface, which would adversely affect the mechanical properties of the composites.

3.1.3 Effect of processing temperature

The results of the tensile and flexural tests of MGCs obtained under different temperatures are shown in Figure 3A and B, respectively. Both the tensile and flexural properties of the MGC were significantly improved at low temperature, but began to worsen as the processing temperature was increased beyond 25°C. The maximum values of the tensile and flexural properties were obtained for MGC-P6D6T0. Compared with GFC, the tensile strength and modulus of MGC-P6D6T0 were improved by 38.4% and 44.7%, respectively, and its flexural strength and modulus increased by 34.6% and 42.4%, respectively.

Figure 3: Mechanical properties of GFC and MGCs obtained under various temperatures.(A) Tensile strength and modulus; (B) flexural strength and modulus.

Figure 3:

Mechanical properties of GFC and MGCs obtained under various temperatures.

(A) Tensile strength and modulus; (B) flexural strength and modulus.

These data indicated that low temperature contributed to the dispersion of MWCNTs, but high temperatures easily caused the collision of MWCNTs and led to increased agglomeration of MWCNTs on the GFf surface. These MWCNT agglomerates even act as stress concentrators to induce the premature cracking of the matrix.

3.2 Interlaminar adhesion of multiscale composites

3.2.1 Effect of ultrasonic power

The ILSS is a widely used indicator to evaluate the interlaminar adhesion of fiber composites. The ILSS values of the MGCs obtained under different ultrasonic powers are shown in Figure 4. Upon increasing the ultrasonic power, the ILSS values of the resulting MGCs first increased and then decreased after a peak value. MGC-P6D6T25 exhibited the maximum value of ILSS.

Figure 4: ILSS values of MGCs obtained under various ultrasonic powers.

Figure 4:

ILSS values of MGCs obtained under various ultrasonic powers.

These results indicated that the appropriate ultrasonic power can promote MWCNT dispersion in the interlaminar region of the composites. The uniform MWCNTs aided in restricting the sliding of neighboring plies under interlaminar shear stress, leading to improved interlaminar adhesion of the composites. However, at high ultrasonic power, it is supposed that the breakage of MWCNTs and/or GFs reduced this enhancement effect on the interlaminar regions of the composites.

3.2.2 Effect of ultrasound duration

The ILSS values of MGCs obtained under different ultrasonic times are shown in Figure 5. With increasing ultrasound duration, the ILSS of the MGCs again first increased and then decreased after a peak value. The ILSS values of the MGCs followed the overall order of MGC-P6D6T25>MGC-P6D12T25>MGC-P6D20T25>MGC-P6D1T25.

Figure 5: ILSS values of MGCs obtained under various ultrasound durations.

Figure 5:

ILSS values of MGCs obtained under various ultrasound durations.

The notable improvement in ILSS values was attributed to the uniform dispersion of MWCNTs in the interlaminar region of the composites. The nanoscale MWCNTs acted as a reinforcer to strengthen the interlaminar region of the composite and reduced the interlaminar stress by eliminating the mismatch in neighboring plies of the composite. However, at long ultrasound durations, MWCNT agglomerates became defects in the interlaminar region of composites and subsequently resulted in debonding of the matrix from the fiber as well as delamination failure.

3.2.3 Effect of processing temperature

The ILSS values of the GFC and MGCs obtained under different processing temperatures are shown in Figure 6.

Figure 6: ILSS values of GFC and MGCs obtained under various processing temperatures.

Figure 6:

ILSS values of GFC and MGCs obtained under various processing temperatures.

The ILSS values of the MGCs can be seen to decrease with increasing processing temperature, and the best results were obtained for MGC-P6D6T0. Compared with GFC, the ILSS of MGC-P6D6T0 exhibited an overall increase of 27.2 MPa, with a maximum increment of 47.1%. Moreover, compared with MGC-P6D6T0, the ILSS of MGC-P6D6T25 showed only a small decrease, but that of MGC-P6D6T45 was significantly lower. This indicated that the high temperature substantially increased MWCNT agglomerates on the GFf surface, which weakened the interlaminar adhesion of composites.

3.3 Morphologies of the MWCNT-GFfs and their composites

3.3.1 Surface morphologies of the MWCNT-GFfs

In order to evaluate the distribution of MWCNTs on the MWCNT-GFf reinforcers, the surface morphologies of GFf and MWCNT-GFf reinforcers photos were taken by three-dimensional (3D) digital microscopy (KH-7700, HIROX, Tokyo, Japan), as shown in Figure 7.

Figure 7: Digital micrographs of MWCNTs distribution on fiber fabrics.(A) GFf, (B) MG-P2D6T25, (C) MG-P6D6T25, (D) MG-P8D6T25, (E) MG-P6D1T25, (F) MG-P6D12T25, (G) MG-P6D20T25, (H) MG-P6D6T0 and (I) MG-P6D6T45.

Figure 7:

Digital micrographs of MWCNTs distribution on fiber fabrics.

(A) GFf, (B) MG-P2D6T25, (C) MG-P6D6T25, (D) MG-P8D6T25, (E) MG-P6D1T25, (F) MG-P6D12T25, (G) MG-P6D20T25, (H) MG-P6D6T0 and (I) MG-P6D6T45.

Figure 7A shows the surface of neat GFf. As seen in Figure 7B–I, a relatively homogeneous distribution of MWCNTs was observed on the GFf surface. However, many MWCNT aggregates could be clearly observed among the fiber tows in MG-P2D6T25 and MG-P6D6T45, as indicated by the yellow arrows in Figure 7B and I, respectively. These indicated that the MWCNTs may not be uniformly dispersed under low output power or at high temperature. Moreover, fractured GF filaments were directly observed on the surface of MG-P8D6T25 and MG-P6D20T25, as indicated by the black arrows in Figure 7D and G, respectively. These indicated that high output powers and long durations could damage the GF. It seemed that few MWCNTs permeated into the fiber tows under low ultrasound duration (1 min), as shown in Figure 7E.

In order to further explore the MWCNT dispersion inside the fiber tows, the fibers chosen from the inner parts of fiber tows were investigated by SEM with a magnification of 5000×, as presented in Figure 8.

Figure 8: SEM images of MWCNTs dispersion on the fiber surface.(A) GFf, (B) MG-P2D6T25, (C) MG-P6D6T25, (D) MG-P8D6T25, (E) MG-P6D1T25, (F) MG-P6D12T25, (G) MG-P6D20T25, (H) MG-P6D6T0 and (I) MG-P6D6T45.

Figure 8:

SEM images of MWCNTs dispersion on the fiber surface.

(A) GFf, (B) MG-P2D6T25, (C) MG-P6D6T25, (D) MG-P8D6T25, (E) MG-P6D1T25, (F) MG-P6D12T25, (G) MG-P6D20T25, (H) MG-P6D6T0 and (I) MG-P6D6T45.

As observed in Figure 8A, the GFf possessed a smooth surface covered with original sizing coatings. Enormous MWCNT agglomerates remained on the surface of MG-P2D6T25, MG-P6D20T25 and MG-P6D6T45, as shown in Figure 8B, G and I, respectively. As seen in Figure 8C, the MWCNTs were uniformly dispersed on the MG-P6D6T25 surface. From Figure 8D, the dispersion state of MWCNTs was further enhanced on the MG-P8D6T25 surface, although the aspect ratio of MWCNTs was substantially decreased due to the high ultrasonic power. Compared with MG-P6D6T25 (Figure 8C), a more limited MWCNT network was formed on the MG-P6D1T25 surface (Figure 8E). When the ultrasound duration during fabrication was over 12 min, the MWCNT density increased and some MWCNT entanglement can be seen on the MG-P6D12T25 surface in Figure 8F, but a good dispersion was still obtained. By comparison, the dispersion state of MWCNTs on the MG-P6D6T0 surface in Figure 8H was the best of all the cases studied.

3.3.2 Fracture morphologies of multiscale composites

In order to evaluate the interfacial adhesion between the GF and matrix, the fracture morphologies of GFC and the MGCs obtained under various ultrasonic parameters after performing short-beam shear tests were investigated by SEM at a magnification of 2000×, as shown in Figure 9.

Figure 9: SEM images of fracture morphologies of multiscale composites.(A) GFC, (B) MGC-P2D6T25, (C) MGC-P6D6T25, (D) MGC-P8D6T25, (E) MGC-P6D1T25, (F) MGC-P6D12T25, (G) MGC-P6D20T25, (H) MGC-P6D6T0 and (I) MGC-P6D6T45.

Figure 9:

SEM images of fracture morphologies of multiscale composites.

(A) GFC, (B) MGC-P2D6T25, (C) MGC-P6D6T25, (D) MGC-P8D6T25, (E) MGC-P6D1T25, (F) MGC-P6D12T25, (G) MGC-P6D20T25, (H) MGC-P6D6T0 and (I) MGC-P6D6T45.

As shown in Figure 9A, the pulled-out GFs exhibited a clear surface with no matrix, which indicates the weak interfacial adhesion in the GFC. Meanwhile, the smooth matrix fracture between GF filaments reveals the brittle behavior of the epoxy matrix. In Figure 9B–I, some matrices are covered on the GF surface, and the interfacial debonding between the fiber and matrix does not present, all of which evidenced the enhanced interfacial adhesion in the MGC. Moreover, MWCNTs at the interface can act as microcrack arresters to toughen the composite through crack deflection and bridging mechanisms in the resin-rich region.

Some matrix ruptures were found in MGC-P2D6T25 (Figure 9B), and many matrix ruptures remained adhered to the GFs in the cases of MGC-P6D6T25 and MGC-P8D6T25, as seen in Figure 9C and D, respectively. However, some fractured GFs can be clearly observed in MGC-P8D6T25 (Figure 9D), indicating that GF breakage occurred under high output power. In comparison to MGC-P6D6T25, fewer matrix ruptures were observed in MGC-P6D1T25 (in Figure 9E), and the matrix between the GFs in the MGC-P6D12T25 was not compact enough after shear tests, as shown in Figure 9F. These indicated that both MGC-P6D1T25 and MGC-P6D12T25 possessed weaker interfacial adhesion than MGC-P6D6T25. In MGC-P6D20T25 (Figure 9G), matrix cracking in the vicinity of the GF surface and fractured GFs were clearly observed. These features indicate that long ultrasound durations may cause some damage to the structure of the GFs, which further resulted in the matrix crack propagation under shear stress. By comparison, the fracture surface of MGC-P6D6T0 was much rougher and the matrix exhibited flexible behavior (Figure 9H). This can be explained by the low temperature contributing to the improved dispersion of the MWCNTs, and such uniform MWCNTs could increase energy dissipation in the resulting composites and delay the fracture failure. In the case of MGC-P6D6T45 (Figure 9I), the matrix was not tightly adhered to the GF. This may be related to the formation of MWCNT agglomerates on the GF surface, which served as stress concentrators to induce the matrix cracking.

3.4 Dynamic viscoelasticity of multiscale composites

3.4.1 Effect of ultrasonic power

The dynamic viscoelasticity is an important indicator to evaluate the interfacial properties of polymer composites. The curves of storage modulus (E′) and loss factor (tan δ) as a function of temperature for the GFC and MGCs obtained under different ultrasonic powers are shown in Figure 10. The main results from the E′ and tan δ curves are summarized in Table 2.

Figure 10: Dynamic viscoelasticity of GFC and the MGCs obtained under various ultrasonic powers.(A) E′ and (B) tan δ curves as a function of temperature.

Figure 10:

Dynamic viscoelasticity of GFC and the MGCs obtained under various ultrasonic powers.

(A) E′ and (B) tan δ curves as a function of temperature.

Table 2:

The E′, Tg and tan δ values from the storage modulus and loss factor curves.a

SampleE′ (GPa)Peak 1Peak 2
E′30E′200Tg1 (°C)Tan δ1Tg2 (°C)Tan δ2
GFC19.42.079.20.201102.00.231
MGC-P2D6T2522.83.384.8 (5.6)0.182106.8 (4.8)0.238
MGC-P6D6T2524.54.586.8 (7.6)0.121107.1 (5.1)0.196
MGC-P8D6T2522.42.281.1 (1.9)0.187103.6 (1.6)0.213
MGC-P6D1T2522.02.382.4 (3.2)0.226102.2 (0.2)0.248
MGC-P6D12T2523.24.485.6 (6.4)0.136107.0 (5.0)0.257
MGC-P6D20T2522.42.783.4 (4.2)0.183108.8 (6.8)0.146
MGC-P6D6T024.95.188.7 (9.5)0.113108.7 (6.7)0.188
MGC-P6D6T4523.71.582.7 (3.5)0.204103.4 (1.4)0.302

  1. aValues within parentheses indicate the increment over that of GFC.

As shown in Figure 10A, the MWCNTs have a strong influence on the E′ values of the composites. Compared with GFC, the E′ values of MGC at 30°C (E′30) were improved by 15.5–26.3%, and the E′ values at 200°C (E′200) were also increased by 10.0–125.0%. The highest E′ values were observed for MGC-P6D6T25. The increase of E′ was attributed to the stiffening effect of MWCNTs and the enhanced interfacial adhesion between the fiber and matrix. However, low ultrasonic power may increase MWCNT agglomerates on the GFf surface, resulting in the more facile movement of epoxy chains at elevated temperatures. For MGC-P8D6T25, fabricated under high ultrasonic power, the significant reduction of E′ value was due to the decrease of the MWCNT aspect ratio.

Two tan δ peaks can be observed in Figure 10B, and the tan δ peak values and their corresponding glass-transition temperatures (Tg) are listed in Table 2. The first tan δ peak value (tan δ1) of the composites plays an important role for engineering applications. Compared with GFC, the Tg1 values of the multiscale composites were 1.9–7.6°C higher (Table 2). With the incorporation of the MWCNTs, the tan δ1 values of the multiscale composites showed a tendency to decrease. MGC-P6D6T25 exhibited the maximum increment of Tg1 and the maximum reduction of tan δ1.

These results can be explained by considering that MWCNTs anchored on the GF surface could restrict the movement of epoxy molecules in multiscale composites and lead to an increase in Tg. The decrease in tan δ indicates that the addition of MWCNTs could reduce the mechanical loss. However, the presence of MWCNT agglomerates due to the low ultrasonic power increased the free volume in the composites and thus resulted in higher conversion of energy into heat. These MWCNT agglomerates at the interface may become a defect to deteriorate the interfacial adhesion of the composites. Additionally, the breakage of MWCNTs under high output power reduced their reinforcing efficiency.

3.4.2 Effect of ultrasound duration

In order to explore the effect of ultrasound duration on the dynamic viscoelasticity, the curves of storage modulus (E′) and loss factor (tan δ) as a function of temperature for the GFC and MGCs obtained under different ultrasound durations are shown in Figure 11.

Figure 11: Dynamic viscoelasticity of GFC and the MGCs obtained under various ultrasound durations.(A) E′ and (B) tan δ curves as a function of temperature.

Figure 11:

Dynamic viscoelasticity of GFC and the MGCs obtained under various ultrasound durations.

(A) E′ and (B) tan δ curves as a function of temperature.

As shown in Figure 11A, upon increasing the ultrasound duration used during fabrication, the E′ values of the resulting MGCs were improved at first, and then began to decline as the ultrasound duration exceeded 6 min. The highest E′30 and E′200 were still obtained for MGC-P6D6T25. These results indicated that the ultrasound duration has a strong influence on the E′ of the resulting composites. A modest ultrasound duration led to good dispersion of the MWCNTs and strong fiber/matrix interfacial adhesion. However, shorter or longer durations produced MWCNT agglomerates on the GFf surface and thus decreased the E′.

Figure 11B shows that MGC-P6D6T25 has the maximum Tg and the minimum tan δ. Compared with MGC-P6D6T25, the Tg of MGC-P6D12T25 exhibited a slight decrease, while those of MGC-P6D1T25 and MGC-P6D20T25 were significantly lower (Table 2). The tan δ1 values of MGC-P6D1T25 and MGC-P6D20T25 were also higher than that of MGC-P6D6T25. These data indicated that the free volume at the interface had been changed by MWCNT agglomerates due to the shorter or longer ultrasound durations. Therefore, the use of a modest ultrasound duration played an important role in ensuring good dispersion of the MWCNTs and improving the interfacial adhesion of the multiscale composites.

3.4.3 Effect of processing temperature

For studying the effect of processing temperature on the dynamic viscoelasticity, the curves of storage modulus (E′) and loss factor (tan δ) for the GFC and MGCs obtained under different temperatures are shown in Figure 12.

Figure 12: Dynamic viscoelasticity of GFC and MGCs obtained under various processing temperatures.(A) E′ and (B) tan δ curves as a function of temperature.

Figure 12:

Dynamic viscoelasticity of GFC and MGCs obtained under various processing temperatures.

(A) E′ and (B) tan δ curves as a function of temperature.

As seen in Figure 12A, the E′ values of the MGCs decreased upon increasing the processing temperature, and the E′200 of MGC-P6D6T45 was even lower than that of GFC. MGC-P6D6T0 exhibited the highest E′. Compared with GFC, the E′30 and E′200 of MGC-P6D6T0 were improved by 28.4% and 155.0%, respectively. These results indicated that the low temperature helped in dispersing MWCNTs onto the GFf surface and thus enhanced the interfacial adhesion strength of the multiscale composites.

Figure 12B shows that the Tg values of MGC-P6D6T0 and MGC-P6D6T25 exhibited an obvious increase compared with that of GFC, but the Tg of MGC-P6D6T45 only showed a slight increase. With the addition of MWCNTs, the tan δ values of the multiscale composites were lower than that of GFC, except in the case of MGC-P6D6T45. By comparison, the maximum increment of Tg1 was attained for MGC-P6D6T0 (9.5°C), and its tan δ1 value exhibited the maximum decrease of 43.8%. These data indicated that the high temperature easily re-agglomerated the MWCNTs on the GFf surface and thus changed the free volume at the interface. As a result, the interfacial adhesion in the composites was significantly reduced.

4 Conclusion

In this study, a multiscale MWCNT-GFf reinforcer was prepared by depositing industrial-grade MWCNTs onto the commercial GFf using an UAI method. The effect of three ultrasonic parameters (output power, duration and processing temperature) on the dispersion of MWCNTs onto the GFf and the mechanical properties, interlaminar adhesion, and dynamic viscoelasticity of the resulting MGCs were systematically investigated. Some interesting findings are as follows:

  1. The ultrasonic treatment during the UAI process can improve the dispersion of MWCNTs and enhance their permeation into the intrafiber space. MWCNTs anchored on the GF surface could interlock the GF with the epoxy and efficiently transfer mechanical stress, all of which toughened the epoxy matrix and improved the interfacial adhesion of multiscale composites.

  2. With low output power (≤200 W) or short duration (≤1 min), it is difficult to disperse MWCNTs uniformly and further insert MWCNTs into the GF tows. With long duration (≥12 min) or high processing temperature (≥45°C), the entanglement or re-agglomeration of MWCNTs readily occurred both on the GF surface and among the GF filaments. High output power (≥800 W) led to well-dispersed MWCNTs but also resulted in reduced aspect ratio for the MWCNTs and breakage of the GF filaments. The best results for the MWCNT dispersion were obtained when using low temperature and moderate duration and output power.

  3. The optimal ultrasonic conditions for dispersing MWCNTs were under 600 W for 6 min at approximately 0°C. Compared with GFC, the tensile strength and modulus of MGC-P6D6T0 increased by 38.4% and 44.7%, respectively, and its flexural strength and modulus were improved by 34.6% and 42.4%, respectively. Furthermore, the ILSS of MGC-P6D6T0 increased by 47.1% and the E′30 and E′200 values increased by 28.4% and 155.0%, respectively. In addition, the interfacial properties of MGC-P6D6T25 were slightly lower than those of MGC-P6D6T0.

Acknowledgments

This work was supported by the Fundamental Research Funds for the Central Universities, China (2016B45414).

References

1. Patel JS, Boddu VM, Brenner MW, Kumar A. Effect of fabric structure and polymer matrix on flexural strength, interlaminar shear stress, and energy dissipation of glass fiber-reinforced polymer composites. Text Res J. 2015;86(2):127–37.10.1177/0040517515586165Search in Google Scholar

2. Cui H, Kessler MR. Glass fiber reinforced ROMP-based bio-renewable polymers: enhancement of the interface with silane coupling agents. Compos Sci Technol. 2012;72(11):1264–72.10.1016/j.compscitech.2012.04.013Search in Google Scholar

3. Zhu J, Imam A, Crane R, Lozano K, Khabashesku VN, Barrera EV. Processing a glass fiber reinforced vinyl ester composite with nanotube enhancement of interlaminar shear strength. Compos Sci Technol. 2007;67(7–8):1509–17.10.1016/j.compscitech.2006.07.018Search in Google Scholar

4. Liu N, Wang J, Yang J, Han G, Yan F. Enhancement on interlaminar shear strength and water-lubricated tribological performance of high-strength glass fabric/phenolic laminate by the incorporation of carbon nanotubes. Polym Adv Technol. 2014;25(12):132–42.10.1002/pat.3405Search in Google Scholar

5. Ajayan PM, Tour JM. Materials science-nanotube composites. Nature 2007;447(7148):1066–8.10.1038/4471066aSearch in Google Scholar PubMed

6. Gude MR, Prolongo SG, Ureña A. Toughening effect of carbon nanotubes and carbon nanofibres in epoxy adhesives for joining carbon fibre laminates. Int J Adhes Adhes. 2015;62:139–45.10.1016/j.ijadhadh.2015.07.011Search in Google Scholar

7. Zeng SH, Shen MX, Duan PP, Liu YR, Zheng HK, Xue YJ. Progress in preparations and applications of encapsulated carbon nanotubes. Adhesion 2016;5:39–42.Search in Google Scholar

8. Díez-Pascual AM, Naffakh M, Marco C, Gómez-Fatou MA, Ellis GJ. Multiscale fiber-reinforced thermoplastic composites incorporating carbon nanotubes: a review. Curr Opin Solid State Mater Sci. 2014;18(2):62–80.10.1016/j.cossms.2013.06.003Search in Google Scholar

9. Reia da Costa EF, Skordos AA, Partridge IK, Rezai A. RTM processing and electrical performance of carbon nanotube modified epoxy/fibre composites. Compos Part A Appl S. 2012;43(4):593–602.10.1016/j.compositesa.2011.12.019Search in Google Scholar

10. Fan ZH, Santare MH, Advani SG. Interlaminar shear strength of glass fiber reinforced epoxy composites enhanced with multi-walled carbon nanotubes. Compos Part A Appl S. 2008;39(3):540–54.10.1016/j.compositesa.2007.11.013Search in Google Scholar

11. Rahmanian S, Thean KS, Suraya AR, Shazed MA, Mohd Salleh MA, Yusoff HM. Carbon and glass hierarchical fibers: influence of carbon nanotubes on tensile, flexural and impact properties of short fiber reinforced composites. Mater Des. 2013;43:10–6.10.1016/j.matdes.2012.06.025Search in Google Scholar

12. An Q, Rider AN, Thostenson ET. Hierarchical composite structures prepared by electrophoretic deposition of carbon nanotubes onto glass fibers. ACS Appl Mat Inter. 2013;5(6):2022–32.10.1021/am3028734Search in Google Scholar PubMed

13. Nag-Chowdhury S, Bellegou H, Pillin I, Castro M, Longrais P, Feller JF. Non-intrusive health monitoring of infused composites with embedded carbon quantum piezo-resistive sensors. Compos Sci Technol. 2016;123:286–94.10.1016/j.compscitech.2016.01.004Search in Google Scholar

14. Romanov V, Lomov SV, Verpoest I, Gorbatikh L. Inter-fiber stresses in composites with carbon nanotube grafted and coated fibers. Compos Sci Technol. 2015;114(18):79–86.10.1016/j.compscitech.2015.04.013Search in Google Scholar

15. Su D, Jin J, Zhang L, Jin L, Li C. Electrostatic layer-by-layer assembly of hierarchical structure of multi-walled carbon nanotubes with glass fiber cloth reinforced epoxy composites. J Macromol Sci B. 2014;53(4):673–82.10.1080/00222348.2013.857549Search in Google Scholar

16. Rahaman A, Kar KK. Carbon nanomaterials grown on E-glass fibers and their application in composite. Compos Sci Technol. 2014;101:1–10.10.1016/j.compscitech.2014.06.019Search in Google Scholar

17. Warrier A, Godara A, Rochez O, Mezzo L, Luizi F, Gorbatikh L, Lomov SV, VanVuure AW, Verpoest I. The effect of adding carbon nanotubes to glass/epoxy composites in the fibre sizing and/or the matrix. Compos Part A Appl S. 2010;41(4):532–8.10.1016/j.compositesa.2010.01.001Search in Google Scholar

18. Zeng SH, Shen MX, Duan PP, Xue YJ, Wang ZY. Effect of silane hydrolysis on the interfacial adhesion of carbon nanotubes/glass fiber fabrics reinforced multiscale composites. Text Res J. 2016:1–13. Available from: http://journals.sagepub.com/doi/abs/10.1177/0040517516679154.10.1177/0040517516679154Search in Google Scholar

19. Tzounis L, Kirsten M, Simon F, Mäder E, Stamm M. The interphase microstructure and electrical properties of glass fibers covalently and non-covalently bonded with multiwall carbon nanotubes. Carbon 2014;73(7):310–24.10.1016/j.carbon.2014.02.069Search in Google Scholar

20. Eskizeybek V, Avci A, Gülce A. The Mode I interlaminar fracture toughness of chemically carbon nanotube grafted glass fabric/epoxy multi-scale composite structures. Compos Part A Appl S. 2014;63(18):94–102.10.1016/j.compositesa.2014.04.013Search in Google Scholar

21. Storck S, Malecki H, Shah T, Zupan M. Improvements in interlaminar strength: a carbon nanotube approach. Compos Part B Eng. 2011;42(6):1508–16.10.1016/j.compositesb.2011.04.039Search in Google Scholar

22. Jin J, Zhang L, Chen W, Li C. Synthesis of glass fiber-multiwall carbon nanotube hybrid structures for high-performance conductive composites. Polym Compos. 2013;34(8):1313–20.10.1002/pc.22544Search in Google Scholar

23. Zhang J, Zhuang R, Liu J, Scheffler C, Mäder E, Heinrich G, Gao S. A single glass fiber with ultrathin layer of carbon nanotube networks beneficial to in-situ monitoring of polymer properties in composite interphases. Soft Mater. 2014;12(1):115–20.10.1080/1539445X.2014.945128Search in Google Scholar

24. Fan ZH, Hsiao KT, Advani SG. Experimental investigation of dispersion during flow of multi-walled carbon nanotube/polymer suspension in fibrous porous media. Carbon 2004;42(4):871–6.10.1016/j.carbon.2004.01.067Search in Google Scholar

25. Gnidakouong JRN, Roh HD, Kim JH, Park YB. In situ assessment of carbon nanotube flow and filtration monitoring through glass fabric using electrical resistance measurement. Compos Part A Appl S. 2016;90:137–46.10.1016/j.compositesa.2016.07.005Search in Google Scholar

26. Li J, Wu Z, Huang C, Li L. Multiscale carbon nanotube-woven glass fiber reinforced cyanate ester/epoxy composites for enhanced mechanical and thermal properties. Compos Sci Technol. 2014;104:81–8.10.1016/j.compscitech.2014.09.007Search in Google Scholar

27. Zhang J, Zhuang R, Liu J, Mäder E, Heinrich G, Gao S. Functional interphases with multi-walled carbon nanotubes in glass fibre/epoxy composites. Carbon 2010;48(8):2273–81.10.1016/j.carbon.2010.03.001Search in Google Scholar

28. Ku-Herrera JJ, Avilés F, Nistal A, Cauich-Rodríguez JV, Rubio F, Rubio J, Bartolo-Pérezc P. Interactions between the glass fiber coating and oxidized carbon nanotubes. Appl Surf Sci. 2015;330:383–92.10.1016/j.apsusc.2015.01.025Search in Google Scholar

Received: 2017-6-8
Accepted: 2017-8-22
Published Online: 2017-9-22
Published in Print: 2018-1-26

©2018 Walter de Gruyter GmbH, Berlin/Boston

This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.