- Email: [email protected]
Flame synthesis of carbon nanotubes onto carbon fiber woven fabric and improvement of interlaminar toughness of composite laminates
Composites Science and Technology 101 (2014) 159–166
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Flame synthesis of carbon nanotubes onto carbon fiber woven fabric and improvement of interlaminar toughness of composite laminates Xusheng Du a,⇑, Hong-Yuan Liu a, Feng Xu a,b, Ying Zeng a, Yiu-Wing Mai a,⇑ a b
Center for Advanced Materials Technology (CAMT), School of Aerospace Mechanical & Mechatronic Engineering J07, The University of Sydney, NSW 2006, Australia School of Aeronautics, Northwestern Polytechnical University, Xi’an, Shaanxi, China
a r t i c l e
i n f o
Article history: Received 20 May 2014 Received in revised form 6 July 2014 Accepted 8 July 2014 Available online 16 July 2014 Keywords: A. Polymer–matrix composites (PMCs) B. Delamination B. Fracture toughness
a b s t r a c t A simple flame synthesis method was utilized for grafting functional carbon nanotubes (CNTs) onto carbon fiber fabrics. Functional organic groups found on CNTs were formed after the flame growth process. Results from electrochemical tests also showed that the accessible surface area was improved by more than 50 times after the carbon fiber fabrics were grafted with CNTs for 3 min. Hence, mode I and mode II interlaminar fracture toughness of these composite laminates, wherein carbon fiber fabrics were grafted with CNTs, increased by 67% and 60%, respectively. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The incorporation of nanofillers, such as carbon nanotubes (CNTs) or carbon nanofibers (CNFs) into the matrix of composites has been developed as an efficient method for improving the mechanical and multifunctional properties of carbon fiber reinforced polymer composites (CFRPs) [1,2]. Two typical routes have been used to incorporate CNTs or CNFs into CFRPs. One utilizes the CNTs or CNFs as reinforced fillers in polymer matrices of CFRPs to improve their mechanical properties. The challenge of this route is that uniform dispersion of CNTs or CNFs in the polymer is hard to achieve, especially at high concentrations, due to the dramatically increased viscosity of the resin. Highly viscous resins with agglomerated CNTs or CNFs are very difficult to process and always lead to poor performance of the polymer nanocomposites [3]. Another route is to attach CNTs or CNFs directly onto CFs to build up a hierarchical reinforcement. Several techniques were successfully developed [1,2]. CNTs could be applied on the fiber surface by spreading CNT powder [4] or CNT-solvent paste [5], and transferring CNT arrays [6]. Grafting of CNTs onto CFs could also be achieved by chemical reaction between the pre-modified functional groups on the surfaces of both CNTs and CFs [7,8]. The electrophoresis technique could be used [9–11], where CNTs were uniformly deposited on the surface of carbon fiber fabric from the CNT dispersion in an applied electric field. ⇑ Corresponding authors. Fax: +61 2 9351 3760. E-mail addresses: [email protected] (X. Du), yiu-wing.mai@sydney. edu.au (Y.-W. Mai). http://dx.doi.org/10.1016/j.compscitech.2014.07.011 0266-3538/Ó 2014 Elsevier Ltd. All rights reserved.
The concept of in situ growing CNTs onto the surfaces of CFs has also been introduced to increase the interfacial shear strength (IFSS) of CFs in the matrix [12]. Chemical vapor deposition (CVD) was the most utilized method for growing CNTs on carbon fibers [12–18]. The growth of CNTs on the surface of CFs leads to the formation of two interfaces: one between CF and CNT and another between CNTs and matrix. Outstanding adhesion between CNTs deposited by CVD and CFs has been obtained [16–18] and a dramatic increase in IFSS was achieved after CNTs were grafted onto CFs [17,18]. In a recent study [19], we showed strong adhesion of in situ flame synthesized CNTs onto CF substrate using an atomic force microscope (AFM) to measure the peel force for a single CNT from a CF [19]. This result indicates a potentially effective method to increase the IFSS and delamination toughness of CFRPs by flame synthesis of CNTs onto carbon fiber fabrics. A problem with CNTs fabricated by normal CVD techniques is their poor affinity with many polymer matrices due to their inert chemical properties. Hence, further functionalization is required to introduce some functional groups on the CNT surface [20]. However, this is not an issue with flame synthesis as oxygen-functional groups are formed on the CNT surface during the growth process [21,22]. Thus, glass fiber/vinyl ester composite laminates with flame synthesized CNTs in the matrix have improved mechanical properties compared to those with CNTs formed by CVD owing to the abundant functional groups present on the flame synthesized CNTs [21]. Hence, the interface between CNTs and matrix in CFRPs can also be promoted by flame synthesized CNTs onto carbon fabrics.
160
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
Fig. 1. (a) Optical image of carbon fiber fabric (CF); CNTs on CF grown for (b) 3 min, and (c) 1 min. (d) SEM image of CNT/CF grown for 3 min.
Based on the abovementioned improved interface properties of CNT/CF and CNT/matrix owing to the flame synthesis technique, we expect the mechanical properties of such CFRPs will be substantially enhanced. Herein, we show that this simple method allows the CNTs to be readily grafted onto the CFs, thus increasing the delamination toughness of woven CFRPs. Compared to CVD, which is an energy intensive batch process requiring costly reagents and equipment, the flame synthesis of CNTs on CFs decreases significantly the growth time and its simple openenvironment deposition process facilitates easy industry scale-up. 2. Experimental work 2.1. Materials Carbon nanotubes were in situ deposited onto the plain woven CF fabric (Inter-Turbine Advanced Logistics Pty Ltd) according to our recent developed flame synthesis method [19]. NiCl2 (0.2 mol/L) was used as a catalyst precursor and ethanol flame the carbon source. Briefly, the plain woven carbon fabric applied with the nickel chloride catalyst precursor was mounted on a metal frame and inserted into the core of the flame at 500– 700 °C for 1 and 3 min to grow the CNTs. Plain woven fabrics with or without CNTs were utilized as the main reinforcement in the CFRPs. Araldite-F (diglycidyl ether of bisphenol A, Huntsman) and piperidine (Sigma–Aldrich) in weight ratio of 100:5 were used as the matrix. Laminates having 16 plies of woven fabrics were prepared by the hand lay-up method. The whole process was maintained at 80–90 °C to ensure the low viscosity of epoxy to fully impregnate the fiber mats. A 25 lm thick polyimide film was inserted in the mid-plane of the laminates to act as the initial crack. The laminates were then wrapped with bleeders and release film within a vacuum bag, first vacuumed in a chamber for 0.5 h followed by curing in a hot-press at 120 °C for 16 h under a pressure of 200 kPa. The fiber volume fraction in the final composite laminates was 58%.
2.2. Characterization Scanning electron microscopic (SEM) images were taken on a Zeiss ULTRA Plus and optical images on a Leica microscope. Transmission electron microscope (TEM) and HRTEM images were recorded on the Philips CM12 (120 kV) and JEOL 2200FS (200 kV), respectively. Fourier transformed infrared (FT-IR) spectra were obtained by a FT-IR spectrometer (Bruker) and differential scanning calorimetry (DSC) datas taken by a TA modulated DSC 2920 in nitrogen. Electrochemical characterization was performed on a CHI1202 Electrochemical Analyzer (CH Instruments). A threeelectrode electro- chemical cell was used for the measurements, where the counter electrode was a Pt foil and the reference electrode was a saturated calomel electrode (SCE). Plain woven CF fabric with or without flame synthesized CNTs was directly used as the working electrode. Mode I and mode II interlaminar fracture toughness values were measured using double-cantilever-beams (DCB) and threepoint end notched flexure (ENF) tests at room temperature on an Instron 5567 machine, following ASTM D5528 for mode I interlaminar fracture toughness tests [23] and the ESIS protocol [24] for mode II interlaminar fracture toughness tests, respectively. For samples modified with CNTs, only the 8th and 9th plies were modified by the flame synthesized CNTs, since the fracture behavior of interest was along the laminate mid-plane in the delamination tests. At least three samples were tested for each reinforced composites system. The mid-plane deflection of the ENF specimens was controlled by a crosshead rate of 1 mm/min and the initial crack length was 25 mm. For Mode I DCB delamination tests, the crack mouth opening displacement rate was 1 mm/min. To ensure the same testing conditions for both neat carbon fabrics and those modified with CNTs, DCB specimens with synthesized CNTs 20 mm ahead of the pre-crack were also prepared (as shown in Fig. 4b later). The delamination toughness results of these specially designed DCBs were obtained and compared with those of normal specimens without the 20 mm bare CF area.
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
Fig. 2. TEM image of CNT/CF (0.2 mol/L NiCl2 and 3 min flame growth).
3. Results and discussion 3.1. Flame growth of CNTs onto carbon fabrics Carbon fiber can be a good reinforcement because it is strong, lightweight, and can withstand the temperatures used in growing the CNTs. Before the synthesis of CNTs, the stability of the bare fabrics in the ethanol flame was examined at identical conditions as used for the growth of CNTs. No obvious morphology changes of
161
the fabrics were observed after the flame treatment. This could be due to the reducing atmosphere in the ethanol flame, protecting the carbon fabrics from combustion. After the flame growth of CNTs, the color of the fabrics changed from grey to black and their surfaces became darker and rougher (Fig. 1(b) and (c)) than the bare fabrics (Fig. 1a). The CNTs uniformly grown on the surface of the CFs can be found in the SEM image (Fig. 1d). In the HRTEM image (Fig. 2), a forest of CNTs is displayed and most have a diameter of 20 nm. No solid core fiber is observed though thicker CNTs with diameters 50 nm can be found occasionally. However, owing to the complex morphology and entangled structure of CNTs around CFs, it is difficult to separate them and measure their exact length directly. But, by analysis of the CNTs scraped from the CF fabrics and dispersed in ethanol, the lengths of most CNTs grown for 3 min are observed to be roughly 1–2 lm, which are much larger than those CNTs grown for 1 min (<500 nm). By measuring the weights before and after the flame synthesis, the weight fractions of CNTs on the CNT/CF fabrics synthesized for 1 and 3 min, are 11% and 18%, respectively. The CNTs obtained here are thinner than those CNTs or CNFs synthesized with similar ethanol flame methods [22,25]. This may be caused by the in situ formation of finer catalyst particles in our flame synthesis. Liao et al. [21] and Liu et al. [22] demonstrated that CNTs grown on nickel plate by a similar flame method has abundant organic oxygen-containing functional groups on their surface. As shown in Fig. 3a, the FTIR characteristic peaks at 1720 and 1580 cm 1 are assigned to the C@O and C@C stretching vibration mode, respectively. The adsorption band at 1230 cm 1 arises from the hydroxyl group and the broad band above 3000 cm 1 may also
Fig. 3. (a) FTIR spectrum of CNTs synthesized by in situ flame method; (b) DSC curves of CNTs synthesized by in-situ flame method and by CVD; and (c) CV curves of CNT/CF and CF in 0.5 M H2SO4 at a scanning rate of 100 mV/s.
162
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
to be an efficient way to improve their electrochemical performance [25]. It is expected that the dramatically improved surface area after flame growth of CNTs onto carbon fabrics and the presence of functional organic groups on CNTs will promote stress transfer between CFs and matrix and enhance the fracture toughness of the composite. 3.2. Mode I interlaminar fracture toughness of composites
Fig. 4. (a) Typical R–Da curves for mode I interlaminar fracture toughness of composite laminates. (b) Geometry of DCB sample; and (c) mode I interlaminar toughness of laminates, GIC versus CNTs growth time.
come from the carboxyl or hydroxyl group. The DSC (Fig. 3b) results confirm the existence of organic functional groups on the flame synthesized CNTs. Compared to the nearly plain curve of CNTs (diameter 10–20 nm, Sigma) fabricated by CVD process, a strong exothermal peak appears around 210 °C in the DSC curve of flame synthesized CNTs. This is similar to the thermal behavior of graphene oxide in the same temperature range [26] which is attributed to the decomposition of organic oxygen-containing groups, such as carboxyl or hydroxyl groups [22]. Since both CNTs and CFs are electrically conductive, accessible surface area of multi-scale carbon materials by liquids can be assessed by the electrochemical method. As shown in Fig. 3c, the current density of CNT/CF fabrics increases by more than 50 times that of plain CF fabrics, indicating huge increase of the surface area of the former. Indeed, flame growth of CNFs onto carbon fabrics has been shown
The mode I fracture toughness GIC, denoted by (N), as a function of crack growth Da for CF grafted with CNTs grown for 3 min is shown in Fig. 4a. Here, the first crack growth was unstable and we could not obtain an accurate GIC. Hence, only GIC values for subsequent Da from 7 mm to 40 mm are plotted which are nearly constant at 0.8 kJ/m2, displaying no R-curve characteristics.1 To illustrate the direct effect of CNTs on interlaminar toughness of CF laminate, specimens with CNTs deposited 20 mm ahead of the initial crack front were prepared as described in Section 2.2 and tested (see Fig. 4b). Thus, crack growth will traverse regions of bare CF and CNT/CF, reflecting their corresponding resistance to delamination in a single sample. GIC results are given by ( ) in Fig. 4a, showing the Rcurve behavior rising from the bare CF value (0.48 kJ/m2) to the CNT/CF plateau value (0.8 kJ/m2) over a crack growth Da of 15– 20 mm. This confirms that the CNTs grafted on CFs for 3 min lead to a significant 67% increase in mode I toughness. Recently, finite element models were used to simulate the pullout of CNTs grafted on CF and study the bridging effect of the CNTs during the hybrid fiber pullout [27]. Both the length and thickness of the CNTs were found to affect fiber bridging. Previous study using CVD deposited CNTs onto alumina fabrics showed that increasing the CNT length increased mode I interlaminar toughness [28]. In our work, for CNTs grafted for 1 min on CF fabric, the fiber length is < 500 nm and there is no effect on GIC with Da (see data ( ) in Fig. 4(a) and (c)). Clearly, the higher the weight fraction of CNTs on CFs, that is, longer synthesis time and higher NiCl2 catalyst concentration, will yield higher interlaminar fracture toughness against delamination. SEM images of fracture surfaces indicate that there exists improved interfacial adhesion between the fiber and matrix after the flame growth of CNTs onto carbon fiber fabrics. These are clearly shown in Fig. 5a for a composite lamina with bare CFs (dominated by fiber debonding with little epoxy resin left on fiber surfaces); and in Fig. 5b for a lamina with CNTs grafted on CFs for 3 min (controlled by both cohesive failure of the CNT/epoxy between adjacent CNT/CFs (typified by blue dash arrows) and on carbon fiber surfaces (see red arrows)). At high magnification, Fig. 5c shows resin-free debonded CFs and brittle fracture of epoxy adjacent the CF bundle in CFRP, confirming the weak interface bonding between bare CFs and epoxy resin. By contrast, at the same magnification, Fig. 5d exhibits two special features on the fracture surface of CFRPs with flame synthesized CNTs for 3 min on the CF fabrics: (a) delamination along the CNT/CF interface peeling off or breaking the CNTs (see white dots indicated by red arrows) which would bridge the interface crack; and (b) transition from well-bonded regions of CNT/CF (see yellow arrows) to cohesive failure of CNT/epoxy matrix with dominant pullout of CNTs (see green arrows). Hence, the composite laminate with synthesized CNTs on CFs has higher mode I delamination toughness. It should be noted that the functional groups on CNTs and their large aspect ratios will promote strong interface bonding between CNTs and epoxy, even higher than that between CNTs and CFs, so that pullout of CNTs is discouraged during delamination growth. The 1 Three tests were performed. Two tests showed fast first crack growth and no Rcurve. One test displayed a full R-curve similar to that shown in Fig. 4a with GIC rising from 0.40 to 0.83 kJ/m2 over a crack growth Da of 20 mm.
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
163
Fig. 5. SEM images of mode I fracture surfaces of laminates reinforced with (a, c) neat CF, (b, d) CNT/CF with 3 min CNT growth, and (e, f) CNT/CF with 1 min CNT growth.
CNT pullout observed in Fig. 5d is due to their large diameter (50 nm). However, our flame synthesized CNTs are mostly 20 nm diameter, their pullout is minimal and fracture is the main failure mode [29]. Although the delamination toughness can in theory be further increased with large diameter CNTs grafted on CF fabric by increasing the catalyst size in the flame growth process [19], the mechanical properties, in particular the tensile strength, of larger diameter CNTs may be decreased compared to those of smaller diameter CNTs. Therefore, further work is required to optimize the morphology, aspect ratio, and areal density of CNTs for toughness enhancement against delamination in CRFPs. Before leaving this section, we note, from Fig. 4c, the laminates with CNTs grafted on CF fabric for 1 min have almost identical mode I toughness GIC as neat CF. As explained, this is primarily because the CNTs are too short (<500 nm) with no pull-out toughness contribution and their weight fraction too small to effectively increase the epoxy matrix toughness or the adhesion between CNTs and CFs, as displayed in Fig. 5(e) and (f) (wherein blue arrows indicate mainly epoxy fracture between CFs, and yellow arrows show CNTs inadequately bonded to CFs). These SEM images are also distinctly different to the corresponding images shown in Fig. 5(b) and (d) for flame synthesized CNTs on CFs for 3 min which impart much higher mode I toughness to the composite laminate.
3.3. Mode II interlaminar fracture toughness of composites The in situ flame growth of CNTs on CF fabric also displays a great effect on the mode II delamination toughness GIIC. Fig. 6a shows the ENF sample and the GIIC results obtained are given in Fig. 6b. Here, GIIC is increased from 1.40 kJ/m2 for the bare carbon fiber fabric to 1.68 kJ/m2 and 2.22 kJ/m2 for those composite laminates with CNTs grafted on CFs for 1 min and 3 min, respectively. In plain woven carbon fiber fabric laminates, there are two directions of the CF bundles. One is along and the other normal to the crack growth direction. In mode II interlaminar shear fracture, the matrix resin between CF fibers within the CF bundles parallel to crack growth (typified by blue dash arrows in Fig. 7b) and inside the cavities (see red arrow in Fig. 7a) between 0°/90° bundles shows characteristic shear hackles. But these hackles cannot be found in areas with the CF bundles normal to crack growth. Instead, broken CFs are observed caused by relative sliding of the top and bottom halves of the ENF sample (typified by yellow dash arrows in Fig. 7(a) and (b). The debonded fibers also appear completely stripped of epoxy resin in the lamina, indicating significant interfacial failure has occurred. Consider the CF fabrics grafted with CNTs for 3 min, shear hackles of CNTs/epoxy between neighboring CFs can still be observed along the crack growth direction (see Fig. 7c); but only with a higher intensity (see blue dash arrows in
164
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
2.5
(a)
(b)
GIIC (KJ/m2 )
2.0 1.5 1.0 0.5 0.0 0
1
2
3
Time for CNT growth on CF (min) Fig. 6. (a) Geometry of ENF sample; and (b) mode II interlaminar fracture toughness of composite laminates, GIIC versus CNTs growth time.
a
b
c
d
e
f
Fig. 7. (a, b) SEM images of mode II fracture surfaces of laminates reinforced with (a, b) neat CF, (c, d) CNT/CF with 3 min CNT growth, and (e, f) CNT/CF with 1 min CNT growth. White arrows indicate crack growth direction.
Fig. 7d), which is a direct result of the CNTs grafted onto CFs, toughening the surrounding epoxy. Higher intensity of hackles means higher fracture surface area, hence more energy dissipation and increased toughness. Resins located within the cavities of the 0°/90° CF bundles also display shear hackles (see red arrow in Fig. 7c). Again, broken CF fibers can be seen on CF/CNT bundles
normal to crack growth (see yellow dash arrows in Fig. 7(c) and (d). All these failure mechanisms give higher mode II delamination toughness. Finally, unlike the mode I delamination results showing no difference of GIC in Fig. 4c, the CNT/CF composite laminates with 1 min growth of CNTs give 20% increase in mode II delamination
165
X. Du et al. / Composites Science and Technology 101 (2014) 159–166 Table 1 Improvements in interlaminar fracture toughness of CNTs-modified CFRPs. Test method
Fracture toughness improvementa
Ref.
Interlayered by CNF bucky paper Growth of CNTs on CF by CVD Growth of CNTs on CF by flame synthesis method
ENF ENF DCB DCB ENF
32% (mode I) 140% (mode II) 23–27% (mode II) 26% (mode I) 20% (mode I) 100–200% (mode II) 5% (mode I) 0% (mode II) 43% (mode I) (for IM7 carbon fabrics) 36% (mode I) (for AS4 carbon fabrics) 200% (mode II) (for AS4 carbon fabrics) 104% (mode II) 46% (mode I) 67% (mode I) 60% (mode II) 350–1350% (mode I) 110–230% (mode II) 1624% (mode I) 815% (mode II)
[30]
Interlayered by (0.5 wt.%) CNT/epoxy film Interlayered by (1 wt.%) CNT/epoxy film Interlayered by CVD grown CNT arrays
DCB ENF ENF DCB DCB ENF DCB ENF DCB
CNTs dispersion techniques 2
Interlayered by spraying CNT dispersion (1.32 g/m ) Interlayered by spraying functionalized CNT dispersion (0.5 wt.% to CF) Interlayered by CNF powder (20 g/m2) Interlayer by CNF paste (20 g/m2)
Stitching (Kevlar or carbon threads and various density) Z-pinning (2% areal density) CF/BMI pins, diameter (0.28 mm) a
DCB ENF DCB ENF
[31] [4] [5] [32] [33] [6]
[2] [13] This work [34] [35] [38]
% Increase is relative to the same CFRPs without CNTs. Fracture toughness, GIC and GIIC, refer to the ‘‘plateau’’ values in the crack growth resistance curves.
toughness GIIC compared to neat CF. This different toughening behavior lies in the different failure mechanism in mode II which is essentially controlled by the formation of shear hackles between CFs and inside the cavities of the 0°/90° CF bundles. Fig. 7(e) and (f) shows such characteristic hackles in the CNT/CF laminate (red2 and dashed blue arrows) are denser in comparison with Fig. 7(a) and (b) with no CNTs on CF fabric, resulting in higher mode II delamination toughness GIIC. Obviously, for CNT/CF laminates with 3 min CNT growth and hence longer CNTs (1–2 lm) and larger weight fraction dispersed in the epoxy matrix, Fig. 7(c) and (d), there are many more shear hackles formed than those with 1 min growth of CNTs during mode II fracture, thus leading to even higher GIIC.
toughening. Indeed, our much earlier studies on CFRPs with micro-sized stitches [34,35] show that up to 1350% increase in GIC and 230% in GIIC can be achieved depending on stitch density. Our previous investigations on through-thickness Z-pins also display their prominent delamination toughening effects [36,37]. Over 1600% increase in GIC and more than 800% increase are obtained using 2% (areal density) 0.28 mm diameter CF/BMI-pins in CFRPs [38]. In fact, with CNTs grafted on CF fabrics or dispersed in interlayers having lengths only in the range of microns and much smaller than the lengths of Z-pins and stitches in millimeters, delamination toughening in modes I and II is less effective as shown in Table 1.
3.4. Comparison of delamination toughening effect due to hierarchical CNT/CF fibers and other methods
4. Conclusion
To put the mode I and mode II delamination toughness results obtained in this work in perspective of similar studies by other researchers, we have compiled relevant toughness data from the literature in Table 1 for easy comparison. Many of these methods include dispersing CNTs or CNFs (carbon nano-fibers) directly as interlayers within the composite laminates [2,4–6,30–33], and CVD deposition of CNTs on plain carbon fiber fabrics [13]. Since all materials have different CFRPs and the interlayer or carbon fabric contains different content of CNTs or CNFs, it is more sensible to compare the percent increase of the plateau toughness relative to that of neat CFRPs as the baseline. Hence, from Table 1, it is clear that the increase in mode I toughness, GIC (67%) by using 3 min flame grown CNTs on CF fabric is larger than most reported values [4–6,13,30–32]. But the increase in mode II toughness, GIIC, in our work (60%) is less effective compared some other methods [2,5,6,30]. It is observed that the CNTs synthesized in this work have diameters 20 nm and lengths 1–2 lm (3 min) and <500 nm (1 min) under an ethanol flame. These CNT forests, Fig. 2, are not stiff enough to stand proud but collapsed on the surface of CFs under hot-press molding. Hence, the CNT forests do not behave like nano-sized pins or stitches (as shown in [6,28]), limiting their effectiveness in mode I and mode II delamination 2 For interpretation of color in Figs. 5 and 7, the reader is referred to the web version of this article.
Flame synthesis is shown to be an efficient technique for in situ growth of CNTs onto CF fabric to establish a hierarchical reinforcement in multi-scale composites to improve the interlaminar toughness. The size and density of the CNTs can be controlled by varying the NiCl2 catalyst molarity and flame growth time, which, in turn, determine the increase of modes I and II toughness against delamination growth. For example, with 0.2 M catalyst solution and 3 min growth time, GIC and GIIC were increased by 67% and 60%, respectively. Under mode I loading, the samples with bare carbon fiber fabric were dominated by CF/epoxy interfacial failure; those samples with CNTs synthesized on CF fabric were largely associated with cohesive CNT/epoxy matrix failure between CNT/CFs and on CF surfaces. Under mode II loading, matrix shear hackles was the predominant toughening mechanisms of neat and CNTmodified CFRP laminates; only the latter has a higher intensity of shear hackles and thus higher fracture toughness. By examining a variety of delamination toughening methods, it appears that nano-size reinforcements using CNTs or CNFs may be less effective compared to the micro-sized reinforcements with through-thickness stitches and Z-pins. Acknowledgments We wish to thank the Australian Research Council for supporting this work. H.Y.L is Future Fellow and X.D. Research Associate, respectively, in the Center for Advanced Materials Technology (CAMT) at the University of Sydney.
166
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
References [1] Qian H, Greenhalgh ES, Shaffer MSP, Bismarck A. Carbon nanotube-based hierarchical composites: a review. J Mater Chem 2010;20(23):4751–62. [2] Khan SU, Kim JK. Improved interlaminar shear properties of multiscale carbon fiber composites with bucky paper interleaves made from carbon nanofibers. Carbon 2012;50:5265–77. [3] Sadeghian R, Gangiredy S, Minaie B, Hsiao KT. Manufacturing carbon nanofibers toughened polyester/glass fiber composites using vacuum assisted resin transfer molding for enhancing the mode-I delamination resistance. Composite A 2006;37:1787–96. [4] Li Y, Hori N, Arai M, Hu N, Liu Y, Fukunaga H. Improvement of interlaminar mechanical properties of CFRP laminates using VGCF. Composite A 2009;40:2004–12. [5] Arai M, Yukihiro N, Sugimoto K-I, Endo M. Mode I and mode II interlaminar fracture toughness of CFRP laminates toughened by carbon nanofiber interlayer. Compos Sci Technol 2008;68:516–25. [6] Garcia EJ, Wardle BL, Hart AJ. Joining prepreg composite interfaces with aligned carbon nanotubes. Composite Part A. 2008;39:1065–70. [7] Li Y, Li Y, Ding Y, Peng Q, Wang C, Wang R, et al. Tuning the interfacial property of hierarchical composites by changing the grafting density of carbon nanotube using 1,3-propodiamine. Compos Sci Technol 2013;85:36–42. [8] Peng Q, He X, Li Y, Wang C, Wang R, Hu PA, et al. Chemically and uniformly grafting carbon nanotubes onto carbon fibers by poly(amidoamine) for enhancing interfacial strength in carbon fiber composites. J Mater Chem 2012;22:5928–31. [9] Bekyarova E, Thostenson ET, Yu A, Kim H, Gao J, Tang J. Multiscale carbon nanotube-carbon fiber reinforcement for advance epoxy composites. Langmuir 2007;23:3940–74. [10] Guo J, Lu C, An F. Effect of electrophoretically deposited carbon nanotubes on the interface of carbon fiber reinforced epoxy composite. J Mater Sci 2012;47:2831–6. [11] Schaefer JD, Rodriguez AJ, Guzman ME, Lim C-S, Minaie B. Effects of electrophoretically deposited carbon nanofibers on the interface of single carbon fibers embedded in epoxy matrix. Carbon 2011;49:2750–9. [12] Thostenson ET, Li WZ, Wang DZ, Ren ZF, Chou TW. Carbon nanotube/carbon fibre hybrid multiscale composites. J Appl Phys 2002;91:6034–7. [13] Kepple KL, Sanborn GP, Lacasse PA, Gruenberg KM, Ready WJ. Improved fracture toughness of carbon fiber composite functionalized with multi walled carbon nanotubes. Carbon 2008;46:2026–33. [14] Sager RJ, Klein PJ, Lagoudas DC, Zhang Q, Liu J, Dai L, et al. Effect of carbon nanotubes on the interfacial shear strength of T650 carbon fiber in an epoxy matrix. Compos Sci Technol 2009;69:898–904. [15] Zhang FH, Wang RG, He XD, Wang C, Ren LN. Interfacial shrearing strength and reinforcing mechanisms of an epoxy composite reinforced using a carbon nanotube/carbon fiber hybrid. J Mater Sci 2009;44:3574–7. [16] De Riccardis MF, Carbone D, Makris Th Dikonimos, Giorgi R, Lisi N, Salernitano E. Anchorage of carbon nanotubes grown on carbon fibres. Carbon 2006;44:671–4. [17] Lv P, Feng YY, Zhang P, Chen HM, Zhao NQ, Feng W. Increasing the interfacial strength in carbon fiber/epoxy composites by controlling the orientation and length of carbon nanotubes grown on the fibers. Carbon 2011;49(14):4665–73.
[18] Qian H, Bismarck A, Greenhalgh ES, Kalinka G, Shaffer MSP. Hierarchical composites reinforced with carbon nanotube grafted fibers: The potential assessed at the single fiber level. Chem Mater 2008;20(5):1862–9. [19] Du XS, Liu HY, Zhou CF, Moody S, Mai Y-W. On the flame synthesis of carbon nanotubes grafted onto carbon fibers and the bonding force between them. Carbon 2012;50(6):2347–50. [20] Xie XL, Mai Y-W, Zhou XP. Dispersion and alignment of carbon nanotubes in polymer matrix: A review. Mater Sci Eng R 2005;49:89–112. [21] Liao L, Wang X, Fang P, Liew KM, Pan C. Interface enhancement of glass fiber reinforced vinyl ester composites with flame-synthesized carbon nanotubes and its enhancing mechanism. ACS Appl Mater Interfaces 2011;3:534–8. [22] Liu H, Wang X, Fang P, Wang S, Qi X, Pan C, et al. Functionalization of multiwalled carbon nanotubes grafted with self-generated functional groups and their polyamide 6 composites. Carbon 2010;48:721–9. [23] Zeng Y, Liu HY, Mai Y-W, Du XS. Improving interlaminar fracture toughness of carbon fibre/epoxy laminates by incorporation of nanoparticles. Compos Part B 2012;43:90–4. [24] European Structural Integrity Society (ESIS). A protocol for interlaminar fracture testing of composites. Delft: European Group on Fracture, Polymer and Composites Task Group; 1989. [25] Khosravi M, Amini MK. Flame synthesis of carbon nanofibers on carbon paper: physicochemical characterization and application as catalyst support for methanol oxidation. Carbon 2010;48:3131–8. [26] Du X, Yu ZZ, Dasari A, Ma J, Mo MS, Meng YZ, et al. New method to prepare graphite nanocomposites. Chem Mater 2008;20:2066–8. [27] Jia Y, Chen ZR, Yan WY. A numerical study on carbon nanotube-hybridized carbon fibre pull out. Compos Sci Technol 2014;91:38–44. [28] Wicks SS, de Villoria RG, Wardle BL. Interlaminar and intralaminar reinforcement of composite laminates with aligned carbon nanotubes. Compos Sci Technol 2010;70:20–8. [29] Tang L, Zhang H, Wu X, Zhang Z. A novel failure analysis of multi-walled carbon nanotubes in epoxy matrix. Polymer 2011;52:2070–4. [30] Joshi SC, Dikshit V. Enhancing interlaminar fracture characteristics of woven CFRP prepreg composites through CNT. J Compos Mater 2011;46(6):665–75. [31] Davis DC, Whelan BD. An experimental study of interlaminar shear fracture toughness of a nanotube reinforced composite. Composite Part B 2011;42:105–16. [32] Sager RJ, Klein PJ, Davis DC, Lagoudas DC, Warren GL, Sue HJ. Interlaminar fracture toughness of woven fabric composite laminates with carbon nanotube/epoxy interleaf films. J Appl Polym Sci 2011;121:2394–405. [33] White KL, Sue HJ. Delamination toughness of fiber-reinforced composites containing a carbon nanotube/polyamide-12 epoxy thin film interlayer. Polymer 2012;53:37–42. [34] Dransfield KA, Jain LK, Mai Y-W. On the effects of stitching in CFRPs – I. Mode I delamination toughness. Compos Sci Technol 1998;58:815–27. [35] Jain LK, Dransfield KA, Mai Y-W. On the effects of stitching in CFRPs – II. Mode II delamination toughness. Compos Sci Technol 1998;58:829–37. [36] Yan WY, Liu HY, Mai Y-W. Numerical study on the mode I delamination toughness of z-pinned laminates. Compos Sci Technol 2003;63:1481–93. [37] Yan WY, Liu HY, Mai Y-W. Mode II delamination toughness of z-pinned laminates. Compos Sci Technol 2004;64:1937–45. [38] Cartié DDR, Partridge IK. Delamination behaviour of z-pinned laminates. In: Proceedings of the 12th international conference on composite materials, ICCM12, Paris; 5–9 July 1999.
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Flame synthesis of carbon nanotubes onto carbon fiber woven fabric and improvement of interlaminar toughness of composite laminates Xusheng Du a,⇑, Hong-Yuan Liu a, Feng Xu a,b, Ying Zeng a, Yiu-Wing Mai a,⇑ a b
Center for Advanced Materials Technology (CAMT), School of Aerospace Mechanical & Mechatronic Engineering J07, The University of Sydney, NSW 2006, Australia School of Aeronautics, Northwestern Polytechnical University, Xi’an, Shaanxi, China
a r t i c l e
i n f o
Article history: Received 20 May 2014 Received in revised form 6 July 2014 Accepted 8 July 2014 Available online 16 July 2014 Keywords: A. Polymer–matrix composites (PMCs) B. Delamination B. Fracture toughness
a b s t r a c t A simple flame synthesis method was utilized for grafting functional carbon nanotubes (CNTs) onto carbon fiber fabrics. Functional organic groups found on CNTs were formed after the flame growth process. Results from electrochemical tests also showed that the accessible surface area was improved by more than 50 times after the carbon fiber fabrics were grafted with CNTs for 3 min. Hence, mode I and mode II interlaminar fracture toughness of these composite laminates, wherein carbon fiber fabrics were grafted with CNTs, increased by 67% and 60%, respectively. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The incorporation of nanofillers, such as carbon nanotubes (CNTs) or carbon nanofibers (CNFs) into the matrix of composites has been developed as an efficient method for improving the mechanical and multifunctional properties of carbon fiber reinforced polymer composites (CFRPs) [1,2]. Two typical routes have been used to incorporate CNTs or CNFs into CFRPs. One utilizes the CNTs or CNFs as reinforced fillers in polymer matrices of CFRPs to improve their mechanical properties. The challenge of this route is that uniform dispersion of CNTs or CNFs in the polymer is hard to achieve, especially at high concentrations, due to the dramatically increased viscosity of the resin. Highly viscous resins with agglomerated CNTs or CNFs are very difficult to process and always lead to poor performance of the polymer nanocomposites [3]. Another route is to attach CNTs or CNFs directly onto CFs to build up a hierarchical reinforcement. Several techniques were successfully developed [1,2]. CNTs could be applied on the fiber surface by spreading CNT powder [4] or CNT-solvent paste [5], and transferring CNT arrays [6]. Grafting of CNTs onto CFs could also be achieved by chemical reaction between the pre-modified functional groups on the surfaces of both CNTs and CFs [7,8]. The electrophoresis technique could be used [9–11], where CNTs were uniformly deposited on the surface of carbon fiber fabric from the CNT dispersion in an applied electric field. ⇑ Corresponding authors. Fax: +61 2 9351 3760. E-mail addresses: [email protected] (X. Du), yiu-wing.mai@sydney. edu.au (Y.-W. Mai). http://dx.doi.org/10.1016/j.compscitech.2014.07.011 0266-3538/Ó 2014 Elsevier Ltd. All rights reserved.
The concept of in situ growing CNTs onto the surfaces of CFs has also been introduced to increase the interfacial shear strength (IFSS) of CFs in the matrix [12]. Chemical vapor deposition (CVD) was the most utilized method for growing CNTs on carbon fibers [12–18]. The growth of CNTs on the surface of CFs leads to the formation of two interfaces: one between CF and CNT and another between CNTs and matrix. Outstanding adhesion between CNTs deposited by CVD and CFs has been obtained [16–18] and a dramatic increase in IFSS was achieved after CNTs were grafted onto CFs [17,18]. In a recent study [19], we showed strong adhesion of in situ flame synthesized CNTs onto CF substrate using an atomic force microscope (AFM) to measure the peel force for a single CNT from a CF [19]. This result indicates a potentially effective method to increase the IFSS and delamination toughness of CFRPs by flame synthesis of CNTs onto carbon fiber fabrics. A problem with CNTs fabricated by normal CVD techniques is their poor affinity with many polymer matrices due to their inert chemical properties. Hence, further functionalization is required to introduce some functional groups on the CNT surface [20]. However, this is not an issue with flame synthesis as oxygen-functional groups are formed on the CNT surface during the growth process [21,22]. Thus, glass fiber/vinyl ester composite laminates with flame synthesized CNTs in the matrix have improved mechanical properties compared to those with CNTs formed by CVD owing to the abundant functional groups present on the flame synthesized CNTs [21]. Hence, the interface between CNTs and matrix in CFRPs can also be promoted by flame synthesized CNTs onto carbon fabrics.
160
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
Fig. 1. (a) Optical image of carbon fiber fabric (CF); CNTs on CF grown for (b) 3 min, and (c) 1 min. (d) SEM image of CNT/CF grown for 3 min.
Based on the abovementioned improved interface properties of CNT/CF and CNT/matrix owing to the flame synthesis technique, we expect the mechanical properties of such CFRPs will be substantially enhanced. Herein, we show that this simple method allows the CNTs to be readily grafted onto the CFs, thus increasing the delamination toughness of woven CFRPs. Compared to CVD, which is an energy intensive batch process requiring costly reagents and equipment, the flame synthesis of CNTs on CFs decreases significantly the growth time and its simple openenvironment deposition process facilitates easy industry scale-up. 2. Experimental work 2.1. Materials Carbon nanotubes were in situ deposited onto the plain woven CF fabric (Inter-Turbine Advanced Logistics Pty Ltd) according to our recent developed flame synthesis method [19]. NiCl2 (0.2 mol/L) was used as a catalyst precursor and ethanol flame the carbon source. Briefly, the plain woven carbon fabric applied with the nickel chloride catalyst precursor was mounted on a metal frame and inserted into the core of the flame at 500– 700 °C for 1 and 3 min to grow the CNTs. Plain woven fabrics with or without CNTs were utilized as the main reinforcement in the CFRPs. Araldite-F (diglycidyl ether of bisphenol A, Huntsman) and piperidine (Sigma–Aldrich) in weight ratio of 100:5 were used as the matrix. Laminates having 16 plies of woven fabrics were prepared by the hand lay-up method. The whole process was maintained at 80–90 °C to ensure the low viscosity of epoxy to fully impregnate the fiber mats. A 25 lm thick polyimide film was inserted in the mid-plane of the laminates to act as the initial crack. The laminates were then wrapped with bleeders and release film within a vacuum bag, first vacuumed in a chamber for 0.5 h followed by curing in a hot-press at 120 °C for 16 h under a pressure of 200 kPa. The fiber volume fraction in the final composite laminates was 58%.
2.2. Characterization Scanning electron microscopic (SEM) images were taken on a Zeiss ULTRA Plus and optical images on a Leica microscope. Transmission electron microscope (TEM) and HRTEM images were recorded on the Philips CM12 (120 kV) and JEOL 2200FS (200 kV), respectively. Fourier transformed infrared (FT-IR) spectra were obtained by a FT-IR spectrometer (Bruker) and differential scanning calorimetry (DSC) datas taken by a TA modulated DSC 2920 in nitrogen. Electrochemical characterization was performed on a CHI1202 Electrochemical Analyzer (CH Instruments). A threeelectrode electro- chemical cell was used for the measurements, where the counter electrode was a Pt foil and the reference electrode was a saturated calomel electrode (SCE). Plain woven CF fabric with or without flame synthesized CNTs was directly used as the working electrode. Mode I and mode II interlaminar fracture toughness values were measured using double-cantilever-beams (DCB) and threepoint end notched flexure (ENF) tests at room temperature on an Instron 5567 machine, following ASTM D5528 for mode I interlaminar fracture toughness tests [23] and the ESIS protocol [24] for mode II interlaminar fracture toughness tests, respectively. For samples modified with CNTs, only the 8th and 9th plies were modified by the flame synthesized CNTs, since the fracture behavior of interest was along the laminate mid-plane in the delamination tests. At least three samples were tested for each reinforced composites system. The mid-plane deflection of the ENF specimens was controlled by a crosshead rate of 1 mm/min and the initial crack length was 25 mm. For Mode I DCB delamination tests, the crack mouth opening displacement rate was 1 mm/min. To ensure the same testing conditions for both neat carbon fabrics and those modified with CNTs, DCB specimens with synthesized CNTs 20 mm ahead of the pre-crack were also prepared (as shown in Fig. 4b later). The delamination toughness results of these specially designed DCBs were obtained and compared with those of normal specimens without the 20 mm bare CF area.
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
Fig. 2. TEM image of CNT/CF (0.2 mol/L NiCl2 and 3 min flame growth).
3. Results and discussion 3.1. Flame growth of CNTs onto carbon fabrics Carbon fiber can be a good reinforcement because it is strong, lightweight, and can withstand the temperatures used in growing the CNTs. Before the synthesis of CNTs, the stability of the bare fabrics in the ethanol flame was examined at identical conditions as used for the growth of CNTs. No obvious morphology changes of
161
the fabrics were observed after the flame treatment. This could be due to the reducing atmosphere in the ethanol flame, protecting the carbon fabrics from combustion. After the flame growth of CNTs, the color of the fabrics changed from grey to black and their surfaces became darker and rougher (Fig. 1(b) and (c)) than the bare fabrics (Fig. 1a). The CNTs uniformly grown on the surface of the CFs can be found in the SEM image (Fig. 1d). In the HRTEM image (Fig. 2), a forest of CNTs is displayed and most have a diameter of 20 nm. No solid core fiber is observed though thicker CNTs with diameters 50 nm can be found occasionally. However, owing to the complex morphology and entangled structure of CNTs around CFs, it is difficult to separate them and measure their exact length directly. But, by analysis of the CNTs scraped from the CF fabrics and dispersed in ethanol, the lengths of most CNTs grown for 3 min are observed to be roughly 1–2 lm, which are much larger than those CNTs grown for 1 min (<500 nm). By measuring the weights before and after the flame synthesis, the weight fractions of CNTs on the CNT/CF fabrics synthesized for 1 and 3 min, are 11% and 18%, respectively. The CNTs obtained here are thinner than those CNTs or CNFs synthesized with similar ethanol flame methods [22,25]. This may be caused by the in situ formation of finer catalyst particles in our flame synthesis. Liao et al. [21] and Liu et al. [22] demonstrated that CNTs grown on nickel plate by a similar flame method has abundant organic oxygen-containing functional groups on their surface. As shown in Fig. 3a, the FTIR characteristic peaks at 1720 and 1580 cm 1 are assigned to the C@O and C@C stretching vibration mode, respectively. The adsorption band at 1230 cm 1 arises from the hydroxyl group and the broad band above 3000 cm 1 may also
Fig. 3. (a) FTIR spectrum of CNTs synthesized by in situ flame method; (b) DSC curves of CNTs synthesized by in-situ flame method and by CVD; and (c) CV curves of CNT/CF and CF in 0.5 M H2SO4 at a scanning rate of 100 mV/s.
162
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
to be an efficient way to improve their electrochemical performance [25]. It is expected that the dramatically improved surface area after flame growth of CNTs onto carbon fabrics and the presence of functional organic groups on CNTs will promote stress transfer between CFs and matrix and enhance the fracture toughness of the composite. 3.2. Mode I interlaminar fracture toughness of composites
Fig. 4. (a) Typical R–Da curves for mode I interlaminar fracture toughness of composite laminates. (b) Geometry of DCB sample; and (c) mode I interlaminar toughness of laminates, GIC versus CNTs growth time.
come from the carboxyl or hydroxyl group. The DSC (Fig. 3b) results confirm the existence of organic functional groups on the flame synthesized CNTs. Compared to the nearly plain curve of CNTs (diameter 10–20 nm, Sigma) fabricated by CVD process, a strong exothermal peak appears around 210 °C in the DSC curve of flame synthesized CNTs. This is similar to the thermal behavior of graphene oxide in the same temperature range [26] which is attributed to the decomposition of organic oxygen-containing groups, such as carboxyl or hydroxyl groups [22]. Since both CNTs and CFs are electrically conductive, accessible surface area of multi-scale carbon materials by liquids can be assessed by the electrochemical method. As shown in Fig. 3c, the current density of CNT/CF fabrics increases by more than 50 times that of plain CF fabrics, indicating huge increase of the surface area of the former. Indeed, flame growth of CNFs onto carbon fabrics has been shown
The mode I fracture toughness GIC, denoted by (N), as a function of crack growth Da for CF grafted with CNTs grown for 3 min is shown in Fig. 4a. Here, the first crack growth was unstable and we could not obtain an accurate GIC. Hence, only GIC values for subsequent Da from 7 mm to 40 mm are plotted which are nearly constant at 0.8 kJ/m2, displaying no R-curve characteristics.1 To illustrate the direct effect of CNTs on interlaminar toughness of CF laminate, specimens with CNTs deposited 20 mm ahead of the initial crack front were prepared as described in Section 2.2 and tested (see Fig. 4b). Thus, crack growth will traverse regions of bare CF and CNT/CF, reflecting their corresponding resistance to delamination in a single sample. GIC results are given by ( ) in Fig. 4a, showing the Rcurve behavior rising from the bare CF value (0.48 kJ/m2) to the CNT/CF plateau value (0.8 kJ/m2) over a crack growth Da of 15– 20 mm. This confirms that the CNTs grafted on CFs for 3 min lead to a significant 67% increase in mode I toughness. Recently, finite element models were used to simulate the pullout of CNTs grafted on CF and study the bridging effect of the CNTs during the hybrid fiber pullout [27]. Both the length and thickness of the CNTs were found to affect fiber bridging. Previous study using CVD deposited CNTs onto alumina fabrics showed that increasing the CNT length increased mode I interlaminar toughness [28]. In our work, for CNTs grafted for 1 min on CF fabric, the fiber length is < 500 nm and there is no effect on GIC with Da (see data ( ) in Fig. 4(a) and (c)). Clearly, the higher the weight fraction of CNTs on CFs, that is, longer synthesis time and higher NiCl2 catalyst concentration, will yield higher interlaminar fracture toughness against delamination. SEM images of fracture surfaces indicate that there exists improved interfacial adhesion between the fiber and matrix after the flame growth of CNTs onto carbon fiber fabrics. These are clearly shown in Fig. 5a for a composite lamina with bare CFs (dominated by fiber debonding with little epoxy resin left on fiber surfaces); and in Fig. 5b for a lamina with CNTs grafted on CFs for 3 min (controlled by both cohesive failure of the CNT/epoxy between adjacent CNT/CFs (typified by blue dash arrows) and on carbon fiber surfaces (see red arrows)). At high magnification, Fig. 5c shows resin-free debonded CFs and brittle fracture of epoxy adjacent the CF bundle in CFRP, confirming the weak interface bonding between bare CFs and epoxy resin. By contrast, at the same magnification, Fig. 5d exhibits two special features on the fracture surface of CFRPs with flame synthesized CNTs for 3 min on the CF fabrics: (a) delamination along the CNT/CF interface peeling off or breaking the CNTs (see white dots indicated by red arrows) which would bridge the interface crack; and (b) transition from well-bonded regions of CNT/CF (see yellow arrows) to cohesive failure of CNT/epoxy matrix with dominant pullout of CNTs (see green arrows). Hence, the composite laminate with synthesized CNTs on CFs has higher mode I delamination toughness. It should be noted that the functional groups on CNTs and their large aspect ratios will promote strong interface bonding between CNTs and epoxy, even higher than that between CNTs and CFs, so that pullout of CNTs is discouraged during delamination growth. The 1 Three tests were performed. Two tests showed fast first crack growth and no Rcurve. One test displayed a full R-curve similar to that shown in Fig. 4a with GIC rising from 0.40 to 0.83 kJ/m2 over a crack growth Da of 20 mm.
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
163
Fig. 5. SEM images of mode I fracture surfaces of laminates reinforced with (a, c) neat CF, (b, d) CNT/CF with 3 min CNT growth, and (e, f) CNT/CF with 1 min CNT growth.
CNT pullout observed in Fig. 5d is due to their large diameter (50 nm). However, our flame synthesized CNTs are mostly 20 nm diameter, their pullout is minimal and fracture is the main failure mode [29]. Although the delamination toughness can in theory be further increased with large diameter CNTs grafted on CF fabric by increasing the catalyst size in the flame growth process [19], the mechanical properties, in particular the tensile strength, of larger diameter CNTs may be decreased compared to those of smaller diameter CNTs. Therefore, further work is required to optimize the morphology, aspect ratio, and areal density of CNTs for toughness enhancement against delamination in CRFPs. Before leaving this section, we note, from Fig. 4c, the laminates with CNTs grafted on CF fabric for 1 min have almost identical mode I toughness GIC as neat CF. As explained, this is primarily because the CNTs are too short (<500 nm) with no pull-out toughness contribution and their weight fraction too small to effectively increase the epoxy matrix toughness or the adhesion between CNTs and CFs, as displayed in Fig. 5(e) and (f) (wherein blue arrows indicate mainly epoxy fracture between CFs, and yellow arrows show CNTs inadequately bonded to CFs). These SEM images are also distinctly different to the corresponding images shown in Fig. 5(b) and (d) for flame synthesized CNTs on CFs for 3 min which impart much higher mode I toughness to the composite laminate.
3.3. Mode II interlaminar fracture toughness of composites The in situ flame growth of CNTs on CF fabric also displays a great effect on the mode II delamination toughness GIIC. Fig. 6a shows the ENF sample and the GIIC results obtained are given in Fig. 6b. Here, GIIC is increased from 1.40 kJ/m2 for the bare carbon fiber fabric to 1.68 kJ/m2 and 2.22 kJ/m2 for those composite laminates with CNTs grafted on CFs for 1 min and 3 min, respectively. In plain woven carbon fiber fabric laminates, there are two directions of the CF bundles. One is along and the other normal to the crack growth direction. In mode II interlaminar shear fracture, the matrix resin between CF fibers within the CF bundles parallel to crack growth (typified by blue dash arrows in Fig. 7b) and inside the cavities (see red arrow in Fig. 7a) between 0°/90° bundles shows characteristic shear hackles. But these hackles cannot be found in areas with the CF bundles normal to crack growth. Instead, broken CFs are observed caused by relative sliding of the top and bottom halves of the ENF sample (typified by yellow dash arrows in Fig. 7(a) and (b). The debonded fibers also appear completely stripped of epoxy resin in the lamina, indicating significant interfacial failure has occurred. Consider the CF fabrics grafted with CNTs for 3 min, shear hackles of CNTs/epoxy between neighboring CFs can still be observed along the crack growth direction (see Fig. 7c); but only with a higher intensity (see blue dash arrows in
164
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
2.5
(a)
(b)
GIIC (KJ/m2 )
2.0 1.5 1.0 0.5 0.0 0
1
2
3
Time for CNT growth on CF (min) Fig. 6. (a) Geometry of ENF sample; and (b) mode II interlaminar fracture toughness of composite laminates, GIIC versus CNTs growth time.
a
b
c
d
e
f
Fig. 7. (a, b) SEM images of mode II fracture surfaces of laminates reinforced with (a, b) neat CF, (c, d) CNT/CF with 3 min CNT growth, and (e, f) CNT/CF with 1 min CNT growth. White arrows indicate crack growth direction.
Fig. 7d), which is a direct result of the CNTs grafted onto CFs, toughening the surrounding epoxy. Higher intensity of hackles means higher fracture surface area, hence more energy dissipation and increased toughness. Resins located within the cavities of the 0°/90° CF bundles also display shear hackles (see red arrow in Fig. 7c). Again, broken CF fibers can be seen on CF/CNT bundles
normal to crack growth (see yellow dash arrows in Fig. 7(c) and (d). All these failure mechanisms give higher mode II delamination toughness. Finally, unlike the mode I delamination results showing no difference of GIC in Fig. 4c, the CNT/CF composite laminates with 1 min growth of CNTs give 20% increase in mode II delamination
165
X. Du et al. / Composites Science and Technology 101 (2014) 159–166 Table 1 Improvements in interlaminar fracture toughness of CNTs-modified CFRPs. Test method
Fracture toughness improvementa
Ref.
Interlayered by CNF bucky paper Growth of CNTs on CF by CVD Growth of CNTs on CF by flame synthesis method
ENF ENF DCB DCB ENF
32% (mode I) 140% (mode II) 23–27% (mode II) 26% (mode I) 20% (mode I) 100–200% (mode II) 5% (mode I) 0% (mode II) 43% (mode I) (for IM7 carbon fabrics) 36% (mode I) (for AS4 carbon fabrics) 200% (mode II) (for AS4 carbon fabrics) 104% (mode II) 46% (mode I) 67% (mode I) 60% (mode II) 350–1350% (mode I) 110–230% (mode II) 1624% (mode I) 815% (mode II)
[30]
Interlayered by (0.5 wt.%) CNT/epoxy film Interlayered by (1 wt.%) CNT/epoxy film Interlayered by CVD grown CNT arrays
DCB ENF ENF DCB DCB ENF DCB ENF DCB
CNTs dispersion techniques 2
Interlayered by spraying CNT dispersion (1.32 g/m ) Interlayered by spraying functionalized CNT dispersion (0.5 wt.% to CF) Interlayered by CNF powder (20 g/m2) Interlayer by CNF paste (20 g/m2)
Stitching (Kevlar or carbon threads and various density) Z-pinning (2% areal density) CF/BMI pins, diameter (0.28 mm) a
DCB ENF DCB ENF
[31] [4] [5] [32] [33] [6]
[2] [13] This work [34] [35] [38]
% Increase is relative to the same CFRPs without CNTs. Fracture toughness, GIC and GIIC, refer to the ‘‘plateau’’ values in the crack growth resistance curves.
toughness GIIC compared to neat CF. This different toughening behavior lies in the different failure mechanism in mode II which is essentially controlled by the formation of shear hackles between CFs and inside the cavities of the 0°/90° CF bundles. Fig. 7(e) and (f) shows such characteristic hackles in the CNT/CF laminate (red2 and dashed blue arrows) are denser in comparison with Fig. 7(a) and (b) with no CNTs on CF fabric, resulting in higher mode II delamination toughness GIIC. Obviously, for CNT/CF laminates with 3 min CNT growth and hence longer CNTs (1–2 lm) and larger weight fraction dispersed in the epoxy matrix, Fig. 7(c) and (d), there are many more shear hackles formed than those with 1 min growth of CNTs during mode II fracture, thus leading to even higher GIIC.
toughening. Indeed, our much earlier studies on CFRPs with micro-sized stitches [34,35] show that up to 1350% increase in GIC and 230% in GIIC can be achieved depending on stitch density. Our previous investigations on through-thickness Z-pins also display their prominent delamination toughening effects [36,37]. Over 1600% increase in GIC and more than 800% increase are obtained using 2% (areal density) 0.28 mm diameter CF/BMI-pins in CFRPs [38]. In fact, with CNTs grafted on CF fabrics or dispersed in interlayers having lengths only in the range of microns and much smaller than the lengths of Z-pins and stitches in millimeters, delamination toughening in modes I and II is less effective as shown in Table 1.
3.4. Comparison of delamination toughening effect due to hierarchical CNT/CF fibers and other methods
4. Conclusion
To put the mode I and mode II delamination toughness results obtained in this work in perspective of similar studies by other researchers, we have compiled relevant toughness data from the literature in Table 1 for easy comparison. Many of these methods include dispersing CNTs or CNFs (carbon nano-fibers) directly as interlayers within the composite laminates [2,4–6,30–33], and CVD deposition of CNTs on plain carbon fiber fabrics [13]. Since all materials have different CFRPs and the interlayer or carbon fabric contains different content of CNTs or CNFs, it is more sensible to compare the percent increase of the plateau toughness relative to that of neat CFRPs as the baseline. Hence, from Table 1, it is clear that the increase in mode I toughness, GIC (67%) by using 3 min flame grown CNTs on CF fabric is larger than most reported values [4–6,13,30–32]. But the increase in mode II toughness, GIIC, in our work (60%) is less effective compared some other methods [2,5,6,30]. It is observed that the CNTs synthesized in this work have diameters 20 nm and lengths 1–2 lm (3 min) and <500 nm (1 min) under an ethanol flame. These CNT forests, Fig. 2, are not stiff enough to stand proud but collapsed on the surface of CFs under hot-press molding. Hence, the CNT forests do not behave like nano-sized pins or stitches (as shown in [6,28]), limiting their effectiveness in mode I and mode II delamination 2 For interpretation of color in Figs. 5 and 7, the reader is referred to the web version of this article.
Flame synthesis is shown to be an efficient technique for in situ growth of CNTs onto CF fabric to establish a hierarchical reinforcement in multi-scale composites to improve the interlaminar toughness. The size and density of the CNTs can be controlled by varying the NiCl2 catalyst molarity and flame growth time, which, in turn, determine the increase of modes I and II toughness against delamination growth. For example, with 0.2 M catalyst solution and 3 min growth time, GIC and GIIC were increased by 67% and 60%, respectively. Under mode I loading, the samples with bare carbon fiber fabric were dominated by CF/epoxy interfacial failure; those samples with CNTs synthesized on CF fabric were largely associated with cohesive CNT/epoxy matrix failure between CNT/CFs and on CF surfaces. Under mode II loading, matrix shear hackles was the predominant toughening mechanisms of neat and CNTmodified CFRP laminates; only the latter has a higher intensity of shear hackles and thus higher fracture toughness. By examining a variety of delamination toughening methods, it appears that nano-size reinforcements using CNTs or CNFs may be less effective compared to the micro-sized reinforcements with through-thickness stitches and Z-pins. Acknowledgments We wish to thank the Australian Research Council for supporting this work. H.Y.L is Future Fellow and X.D. Research Associate, respectively, in the Center for Advanced Materials Technology (CAMT) at the University of Sydney.
166
X. Du et al. / Composites Science and Technology 101 (2014) 159–166
References [1] Qian H, Greenhalgh ES, Shaffer MSP, Bismarck A. Carbon nanotube-based hierarchical composites: a review. J Mater Chem 2010;20(23):4751–62. [2] Khan SU, Kim JK. Improved interlaminar shear properties of multiscale carbon fiber composites with bucky paper interleaves made from carbon nanofibers. Carbon 2012;50:5265–77. [3] Sadeghian R, Gangiredy S, Minaie B, Hsiao KT. Manufacturing carbon nanofibers toughened polyester/glass fiber composites using vacuum assisted resin transfer molding for enhancing the mode-I delamination resistance. Composite A 2006;37:1787–96. [4] Li Y, Hori N, Arai M, Hu N, Liu Y, Fukunaga H. Improvement of interlaminar mechanical properties of CFRP laminates using VGCF. Composite A 2009;40:2004–12. [5] Arai M, Yukihiro N, Sugimoto K-I, Endo M. Mode I and mode II interlaminar fracture toughness of CFRP laminates toughened by carbon nanofiber interlayer. Compos Sci Technol 2008;68:516–25. [6] Garcia EJ, Wardle BL, Hart AJ. Joining prepreg composite interfaces with aligned carbon nanotubes. Composite Part A. 2008;39:1065–70. [7] Li Y, Li Y, Ding Y, Peng Q, Wang C, Wang R, et al. Tuning the interfacial property of hierarchical composites by changing the grafting density of carbon nanotube using 1,3-propodiamine. Compos Sci Technol 2013;85:36–42. [8] Peng Q, He X, Li Y, Wang C, Wang R, Hu PA, et al. Chemically and uniformly grafting carbon nanotubes onto carbon fibers by poly(amidoamine) for enhancing interfacial strength in carbon fiber composites. J Mater Chem 2012;22:5928–31. [9] Bekyarova E, Thostenson ET, Yu A, Kim H, Gao J, Tang J. Multiscale carbon nanotube-carbon fiber reinforcement for advance epoxy composites. Langmuir 2007;23:3940–74. [10] Guo J, Lu C, An F. Effect of electrophoretically deposited carbon nanotubes on the interface of carbon fiber reinforced epoxy composite. J Mater Sci 2012;47:2831–6. [11] Schaefer JD, Rodriguez AJ, Guzman ME, Lim C-S, Minaie B. Effects of electrophoretically deposited carbon nanofibers on the interface of single carbon fibers embedded in epoxy matrix. Carbon 2011;49:2750–9. [12] Thostenson ET, Li WZ, Wang DZ, Ren ZF, Chou TW. Carbon nanotube/carbon fibre hybrid multiscale composites. J Appl Phys 2002;91:6034–7. [13] Kepple KL, Sanborn GP, Lacasse PA, Gruenberg KM, Ready WJ. Improved fracture toughness of carbon fiber composite functionalized with multi walled carbon nanotubes. Carbon 2008;46:2026–33. [14] Sager RJ, Klein PJ, Lagoudas DC, Zhang Q, Liu J, Dai L, et al. Effect of carbon nanotubes on the interfacial shear strength of T650 carbon fiber in an epoxy matrix. Compos Sci Technol 2009;69:898–904. [15] Zhang FH, Wang RG, He XD, Wang C, Ren LN. Interfacial shrearing strength and reinforcing mechanisms of an epoxy composite reinforced using a carbon nanotube/carbon fiber hybrid. J Mater Sci 2009;44:3574–7. [16] De Riccardis MF, Carbone D, Makris Th Dikonimos, Giorgi R, Lisi N, Salernitano E. Anchorage of carbon nanotubes grown on carbon fibres. Carbon 2006;44:671–4. [17] Lv P, Feng YY, Zhang P, Chen HM, Zhao NQ, Feng W. Increasing the interfacial strength in carbon fiber/epoxy composites by controlling the orientation and length of carbon nanotubes grown on the fibers. Carbon 2011;49(14):4665–73.
[18] Qian H, Bismarck A, Greenhalgh ES, Kalinka G, Shaffer MSP. Hierarchical composites reinforced with carbon nanotube grafted fibers: The potential assessed at the single fiber level. Chem Mater 2008;20(5):1862–9. [19] Du XS, Liu HY, Zhou CF, Moody S, Mai Y-W. On the flame synthesis of carbon nanotubes grafted onto carbon fibers and the bonding force between them. Carbon 2012;50(6):2347–50. [20] Xie XL, Mai Y-W, Zhou XP. Dispersion and alignment of carbon nanotubes in polymer matrix: A review. Mater Sci Eng R 2005;49:89–112. [21] Liao L, Wang X, Fang P, Liew KM, Pan C. Interface enhancement of glass fiber reinforced vinyl ester composites with flame-synthesized carbon nanotubes and its enhancing mechanism. ACS Appl Mater Interfaces 2011;3:534–8. [22] Liu H, Wang X, Fang P, Wang S, Qi X, Pan C, et al. Functionalization of multiwalled carbon nanotubes grafted with self-generated functional groups and their polyamide 6 composites. Carbon 2010;48:721–9. [23] Zeng Y, Liu HY, Mai Y-W, Du XS. Improving interlaminar fracture toughness of carbon fibre/epoxy laminates by incorporation of nanoparticles. Compos Part B 2012;43:90–4. [24] European Structural Integrity Society (ESIS). A protocol for interlaminar fracture testing of composites. Delft: European Group on Fracture, Polymer and Composites Task Group; 1989. [25] Khosravi M, Amini MK. Flame synthesis of carbon nanofibers on carbon paper: physicochemical characterization and application as catalyst support for methanol oxidation. Carbon 2010;48:3131–8. [26] Du X, Yu ZZ, Dasari A, Ma J, Mo MS, Meng YZ, et al. New method to prepare graphite nanocomposites. Chem Mater 2008;20:2066–8. [27] Jia Y, Chen ZR, Yan WY. A numerical study on carbon nanotube-hybridized carbon fibre pull out. Compos Sci Technol 2014;91:38–44. [28] Wicks SS, de Villoria RG, Wardle BL. Interlaminar and intralaminar reinforcement of composite laminates with aligned carbon nanotubes. Compos Sci Technol 2010;70:20–8. [29] Tang L, Zhang H, Wu X, Zhang Z. A novel failure analysis of multi-walled carbon nanotubes in epoxy matrix. Polymer 2011;52:2070–4. [30] Joshi SC, Dikshit V. Enhancing interlaminar fracture characteristics of woven CFRP prepreg composites through CNT. J Compos Mater 2011;46(6):665–75. [31] Davis DC, Whelan BD. An experimental study of interlaminar shear fracture toughness of a nanotube reinforced composite. Composite Part B 2011;42:105–16. [32] Sager RJ, Klein PJ, Davis DC, Lagoudas DC, Warren GL, Sue HJ. Interlaminar fracture toughness of woven fabric composite laminates with carbon nanotube/epoxy interleaf films. J Appl Polym Sci 2011;121:2394–405. [33] White KL, Sue HJ. Delamination toughness of fiber-reinforced composites containing a carbon nanotube/polyamide-12 epoxy thin film interlayer. Polymer 2012;53:37–42. [34] Dransfield KA, Jain LK, Mai Y-W. On the effects of stitching in CFRPs – I. Mode I delamination toughness. Compos Sci Technol 1998;58:815–27. [35] Jain LK, Dransfield KA, Mai Y-W. On the effects of stitching in CFRPs – II. Mode II delamination toughness. Compos Sci Technol 1998;58:829–37. [36] Yan WY, Liu HY, Mai Y-W. Numerical study on the mode I delamination toughness of z-pinned laminates. Compos Sci Technol 2003;63:1481–93. [37] Yan WY, Liu HY, Mai Y-W. Mode II delamination toughness of z-pinned laminates. Compos Sci Technol 2004;64:1937–45. [38] Cartié DDR, Partridge IK. Delamination behaviour of z-pinned laminates. In: Proceedings of the 12th international conference on composite materials, ICCM12, Paris; 5–9 July 1999.