Composites Part A 112 (2018) 161–167
Contents lists available at ScienceDirect
Composites Part A journal homepage: www.elsevier.com/locate/compositesa
Simultaneously enhancing the IFSS and monitoring the interfacial stress state of GF/epoxy composites via building in the MWCNT interface sensor
T
⁎
Bin Yanga, Fu-Zhen Xuana, , Hongshuai Leib, Zhenqing Wangc, Yanxun Xianga, Kang Yanga, Xiaojun Tangd, Wenyan Liangc a
School of Mechanical and Power Engineering, East China University of Science and Technology, Shanghai, China Beijing Key Laboratory of Lightweight Multi-functional Composite Materials and Structures, Beijing Institute of Technology, Beijing, China c College of Aerospace and Civil Engineering, Harbin Engineering University, Harbin, China d Beijing Spacecrafts, China Academy of Space Technology, Beijing, China b
A R T I C LE I N FO
A B S T R A C T
Keywords: A. Glass fibers B. Debonding D. Non-destructive testing
This paper presents an effective technology that could simultaneously enhance the interfacial shear strength (IFSS) and monitor the interfacial stress state between glass fiber and epoxy vinyl ester resin (GF/epoxy). Muitiwalled carbon nanotube (MWCNT) was added to aqueous surfactant solution and dispersed by ultrasonic. Subsequently, MWCNT was deposited on GF surface by physical vapor deposition. The results show that the sensing performance of the developed sensor was dependence on MWCNT solution concentration, interface length, ultrasonic dispersion duration, and immersion cycles. Fiber-bundle pull-out tests show that IFSS of GF/ epoxy was enhanced by incorporating MWCNT into the interphase. By measuring resistance change of the MWCNT sensor, the interfacial evolution behaviour was monitored during the pull-out test. The results indicate that the presented technology can be successfully used for in-situ sensing the accumulated interfacial damage and simultaneously enhancing the IFSS of GF/epoxy composites.
1. Introduction Besides the effective interface to guarantee the efficient stress transfer from matrix to load-bearing fibers, it is necessary to establish damage sensing functionality to track the nature and extent of interfacial damage in composites [1]. This is because properties of composite materials are largely determined by fiber/matrix interfacial bonding state, and fiber/matrix debonding is considered as the very early damage during the failure process of composite structures [2–4]. Good interfacial bonding is essential to ensure efficient load transfer from matrix to reinforcements, which helps to reduce stress concentration and improves the overall mechanical properties [5]. The interfacial characterizations of composites are generally influenced by the chemical and physical structure, thickness and morphology, adhesion strength and residual stress of the interface [6]. According to the interfacial characterizations, different physico-chemical and/or frictional strategies have been coined to enhance the load transfer capability [7–9]. Among these methods, fiber surface modification for increasing the wettability and interfacial adhesion with polymeric matrix is believed to be an effective approach [10–13]. For example, by fiber surface treatment with graphene oxide and polyhedral oligomeric
⁎
silsesquioxane, the area and wettability of fiber surface are significantly enhanced, and this further leads to an increase in interfacial shear strength (IFSS) [13]. However, attribute to the complicated interfacial structure, researchers also found that the modified interphases may failure immediately with little advanced warning as soon as the interfacial shear stress increases up to the IFSS [8,14,15]. This phenomenon is mainly due to the accumulation of microcracks in the modified interphase, and it further decreases the overall durability of a composite [15,16]. Therefore, it is required to develop methods for damage detection of the modified interphase [17,18]. In-situ health monitoring can enable prediction of remaining life, thus prevent the catastrophic failures of composites at early stage [17,19]. Principle of in-situ electrical resistance measurement is that the conductive networks can break up with the appearance of damage, which will lead to the macroscopic changes of the electrical resistance [20,21]. Compared with other non-destructive evaluation (NDE) methods, such as guided waves monitoring [22] and fiber bragg grating technology [23], resistance based in-situ monitoring dose not involve the attachment of external sensors or additional fibers input in the structures. By the resistance change of high sensitivity MWCNT sensor, it has been established that adding MWCNT to fiber reinforced
Corresponding author at: School of Mechanical and Power Engineering East China University of Science and Technology, No 130, Meilong Road, Shanghai 200237, China. E-mail address:
[email protected] (F.-Z. Xuan).
https://doi.org/10.1016/j.compositesa.2018.06.006 Received 15 March 2018; Received in revised form 16 May 2018; Accepted 5 June 2018 Available online 05 June 2018 1359-835X/ © 2018 Elsevier Ltd. All rights reserved.
Composites Part A 112 (2018) 161–167
B. Yang et al.
Nomenclature
t ΔR/ R0 Fmax d σd0
interfacial shear strength (MPa) active content (%) resistance of the glass fiber yarn (Ω) volume resistance of the MWCNT (Ω) Yarn length (mm) cross-sectional area of the MWCNT-coating in the GF yarn (mm2) MWCNT solution concentration (g/ml)
τIFSS AC Ryarn ρMWCNT l Acoat N
σdp σd∗
Table 3 Specification of epoxy vinyl ester resin.
Table 1 Technical data of the used –COOH functionalized multiwalled carbon nanotubes. Category
Parameters
Purity - COOH content Outer diameter Inner diameter Length Special surface area Appearance Tap density True density Electric conductivity Making method
> 98% 0.49 wt% > 50 nm 5–15 nm < 10 μm > 60 m2/g Black 0.18 g/cm3 2.1 g/cm3 > 100 s/cm CVD
ultrasonic dispersion duration of MWCNT solution (min) relative resistance maximum load on the load-displacement curve (N) diameter of fiber bundles (mm) initial debonding stress in the fiber bundle pull-out test (MPa) partial debonding stress in the fiber bundle pull-out test (MPa) complete debonding stress in the fiber bundle pull-out test (MPa)
Category
Parameters
Product categories Appearance Viscosity Gelation time Tensile strength Elongation at break Heat deflection temperature
Bisphenol A epoxy vinyl ester resin Light yellow 400–450 MPa s, 25 °C 8–16 min, 25 °C 85 MPa 5% 120 °C
14
Fitting of N=1.25 mg/ml Fitting of N=2.50 mg/ml Fitting of N=3.75 mg/ml Fitting of N=5.00 mg/ml Fitting of N=6.25 mg/ml Fitting of N=7.50 mg/ml N=1.25 mg/ml N=2.50 mg/ml N=3.75 mg/ml N=5.00 mg/ml N=6.25 mg/ml N=7.50 mg/ml
12
Resistance/k
10 Table 2 Specification of sodium dodecyl sulfate. Category
Parameters
Structural formula Active content Non-sulfonated Moisture Whiteness PH value Heavy metals content Inorganic salt content
CH3(CH2)11OSO3Na > 88% < 1.5% < 3% > 78% 7.5–9.5 < 15 mg/kg (NaCl + NaSO4) < 9.0%
4
0
y=0.132x+0.08
y=0.073x+0.01 y=0.044x-0.11 y=0.039x-0.03
0
5
10
15
20
25
30
35
Fig. 2. Dependence of resistance on the MWCNT solution concentrations. Linear lines show least square fit through the origin of the resistance data and the corresponding equations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
(c)
MWCNT coated glass fibers
MWCNT coated glass fibers
Epoxy cylinder
21 mm
y=0.299x+0.25
Test distance/mm
(b)
5 mm
Pristine fibers
6
2
composites is a promising way to detect the formation of microscale damage [1,24,25]. Since MWCNT can provide self-sensing capabilities for damage detection [26,27], it has the potential to provide life extension control or damage mitigation [28]. Although MWCNT sensor
(a)
8
y=0.356x+0.72
R Epoxy cylinder
Fixed surface Tensile load
4 mm
Silver paint 162
Fig. 1. Scheme of fiber-bundle pull-out test and simultaneously recording of resistance change: (a) pristine GF yarn, and (b) the MWCNT-coated GF yarn embedded in an epoxy matrix, and (c) the designed fixture in the test. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Composites Part A 112 (2018) 161–167
B. Yang et al.
(a) Pristine fiber
(b) Control sample
Smooth glass fiber surface
Smooth glass fiber surface
(c) 1.25 mg/ml
(d) 2.50 mg/ml
MWCNT on fiber surface
MWCNT on fiber surface
(e) 3.75 mg/ml
(f) 5.00 mg/ml
Rich MWCNT region
MWCNT on fiber surface
(g) 6.25 mg/ml
(h) 7.50 mg/ml
Fiber surface was fully coated by MWCNT
Rich MWCNT region
Fig. 3. SEM micrographs of MWCNT-coated GF yarns: (a) and (b) are the smooth fiber surfaces, (c–e) are the as-coated GF surfaces with flocculent MWCNTs, and (f–h) show the GF surfaces with fully coverage MWCNT. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
method. Influences of MWCNT solution concentration, interface length, ultrasonic dispersion duration, and immersion cycle on the electrical behavior of the interphase sensor were investigated. Effect of MWCNT on the IFSS of GF/epoxy was evaluated by fiber-bundle pull-out tests. Sensor resistance change was tracked during the pull-out tests. Relationship between the resistance change and the accumulated interfacial damage was discussed in detail.
may sense the interphase failure based on the principles that debonding will lead to its electrical resistance change, researches on this aspect are currently limited. In the present paper, electric conduction functional interphase was fabricated by incorporating MWCNT in the interface between glass fiber and epoxy vinyl ester resin (GF/epoxy). MWCNT in aqueous surfactant solution was firstly dispersed by ultrasonic dispersion technology, and then it was deposited on GF surface by physical vapor deposition (PVD)
163
Composites Part A 112 (2018) 161–167
B. Yang et al.
(a) 0-cycle
(b) 1-cycle
(c) 2-cycle
GF Residual resin
Smooth GF surface
(d)3-cycle 3-cycle (d) Residual resin
Rough fiber surface Fitting of t=30 min Fitting of t=60 min Fitting of t=90 min Fitting of t=120 min t=30 min t=60 min t=90 min t=120 min
Resistance/kΩ
4 3
y=0.131x+0.242
14
21
Test distance/mm
28
35
(b) N=6.25 mg/ml, 5-cycle’s immersion in MWCNT solution. Fig. 4. Dependence of resistance of MWCNT-coated GF yarn on the immersion cycle, yarn length and concentrations. Linear lines show least square fit through the origin of the resistance data and the corresponding equations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
40
IFSS/MPa
30
20
10
0
0
1
2
Rough fiber surface
20 m
MWCNT was purchased from Time Nano Technologies, Chengdu Organic Chemicals Co. Ltd., China. The detailed technical data of the nanotubes were given in Table 1. It was added to aqueous surfactant solution and treated with an ultrasonic processor at constant output power of 300 W. The aqueous surfactant solution was prepared by dissolving sodium dodecyl sulfate (SDS, General Reagent Inc., China) in deionized water. Table 2 listed the specification of the used SDS. In the MWCNT solution, mass ratio of MWCNT to SDS was 100:30 with 40 ml deionized water. The polymer matrix was bisphenol A epoxy vinyl ester resin that can be cured at room temperature with hardening and accelerating agent. The resin was the product from the chemical reaction between bisphenol A epoxy resin and methacrylic acid (structural formula: H2C = C(CH3)COOH). Specification of the resin was given in Table 3, and it was referred to as epoxy in the following paper. The hardening agent was Methyl Ethyl Ketone Peroxide (MEKP), and the accelerating agent was N, N-dimethylaniline (structural formula: C6H5N (CH3)2). The resin was mixed with hardener and accelerator at mass ratio 1:2%:0.5%. The reinforced material was unidirectional glass fiberbundle with average single fiber diameter of 20 μm. It was extracted from the fabric cloth with surface mass density of 800 g/m2. The elongation at break of the fiber is 1.5%, while it was 5% of the resin, thus the resin possess more than 3× the elongation of the fiber in the following prepared pull-out specimens. GF yarn was immersed in the prepared MWCNT solution for different cycles. In each cycle, the GF yarn was removed from the MWCNT solution after immersed for 5 s, and then it was heated at 105 °C for 5 min to evaporate the residual water. Thus, the MWCNT was deposited on GF yarn surface. Control samples were prepared by immersing GF yarn in the solution without MWCNT under the same treatment regime. To evaluate the MWCNT-coated GF surface pattern and the interfacial adhesion quality, scanning electron microscope (SEM, EVO MA 15, Zeiss) observations were conducted. Resistance of the MWCNT sensor was tested by a Keithley 2700 programmable electrometer. Depending on the measured yarn length, the GF yarn was contacted with electrically conductive silver paint at the corresponding distances.
y=0.073x+0.02
7
Residual resin
2. Experimental
y=0.073x+0.01
0
(f) 5-cycle
Fig. 6. Microscopic appearance of GF yarns showing the increasing roughness with immersion cycles. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
y=0.094x+0.286
2 1
(e) 4-cycle
Rough fiber surface
Rough fiber surface
(a) MWCNT dispersion duration t=120 min, and N=6.25 mg/ml.
5
Rough fiber surface
3
Immersion cycles
4
5
Fig. 5. Comparison between IFSS of GF/epoxy filled by MWCNT after various immersion cycles.
164
Composites Part A 112 (2018) 161–167
B. Yang et al.
Complete debonding Shear load Resistance change
than epoxy cylinder diameter. Therefore, the system allowed for friction free sliding of GF yarn while blocking the motion of epoxy host when the device was actuated, thus loading the fiber–matrix interface in shear. Resistance change of the interphase sensor during pull-out test was recorded simultaneously.
* d
2.0
1.5
120 0
Initial debonding
d
Partial debonding
p
1.0
d
80
3. Processing condition selection
R/R0
Shear Load/N
160
As discussed, the sensing ability of MWCNT sensor depends deeply on the perfection of the formed conducting network in the interphase. Because MWCNT must be adequate enough to maintaining a suitable morphology and aspect ratio (length/diameter) in the electrically conductive network, the MWCNT solution concentration, N, plays a significant role in the formation of the conducting network. Fig. 2 shows dependence of the GF yarn resistance on N and yarn length l. For all the concentration cases, resistance shows a linear line with the increasing of l. This is expected since the resistance of the yarn is given by:
0.5 40 0.0
0.2
0.4
0.6
0.8
1.0
0.0 1.2
Displacement/mm (a) 3-cycle’s immersion 200
Ryarn = ρMWCNT ·
2.0
Complete debonding
*
1.5 120
Frictional resistance
1.0
80 0.5 40 0.0
0.2
0.4
0.6
0.8
1.0
0.0 1.2
Displacement/mm (b) 4-cycle’s immersion
Shear Load/N
160
2.5
3-cycle's immersion 4-cycle's immersion 3-cycle's immersion 4-cycle's immersion
2.0 1.5
120 Shear load
1.0
80
R/R0
200
Relative resistance 0.5
40 0.0 0 0.0
0.2
0.4
0.6
0.8
1.0
l Acoat
(1)
Because volume resistance of MWCNT, ρMWCNT , as well as the crosssectional area, Acoat, are constant for a given N, Ryarn linear depends on l. Moreover, as discussed in [24], once a thin coating layer on fiber surface was formed by the homogeneous coating material, Acoat has an increase tendency with increasing of N. The linear depends of Ryarn on l also indicates the uniform distribution of MWCNT along GF length direction. This is necessary in charge transfer due to that GF with homogeneous MWCNT coating layers is crucial as only interconnected areas can form continuous conductive paths in the interphase [24,29]. Meanwhile, for an identical l, Ryarn decreases with increasing of the used MWCNT solution concentrations. The fitting lines for N = 6.25 and 7.50 g/ml in Fig. 2 are nearly overlapped together (the slopes of a straight line are 0.39 and 0.44, respectively), which indicates the saturated MWCNT electric conduction network. Fig. 3 illustrates the SEM micrographs of GF yarn after immersed in MWCNT solution with various concentrations. By comparing the pristine and control sample in Fig. 3a and b, GF treated by aqueous surfactant solution without MWCNT has little influence on the microcosmic appearance of fiber surface. With increasing of solution concentrations, MWCNT thickness on GF yarn keeps increasing until the fiber is fully coated when N = 6.25 and 7.5 mg/ml. This phenomenon illustrates that N = 6.25 mg/ml is in the percolation threshold of the system. Hence, N = 6.25 mg/ml was selected to manufacture the MWCNT based interphase sensor. Depending on the dimension of pull-out specimens in Fig. 1, the measured yarn length was selected as 21 mm. The percolation threshold of MWCNT sensor also depends on the GF yarn immersion cycle and MWCNT ultrasonic dispersion duration t. Intrinsically, these two parameters affect the cross-sectional area, Acoat, in Eq. (1). Fig. 4 shows the dependence of sensor resistance on the two parameters. In Fig. 4a, a significant drop in resistivity is observed after the percolation threshold at around 2-cycle’s immersion, and further addition of immersion cycles has only a minor effect on the resistivity. Fig. 4b shows the relationship between the sensor resistance and t. According to Fig. 4b, the electrical conductivity of the MWCNT sensor increases with the increasing of t. Moreover, resistance of the interphase sensor at t = 90 and 120 min in the figure reaches the same value (with the fitting slope of 0.073 for both cases). This phenomenon demonstrates that resistance of the MWCNT sensor reaches to a constant value when t > 90 min. Therefore, t = 120 min with 5-cycle’s immersion were adopted to perform the fiber-bundle pull-out test and the interphase monitoring experiments.
d
R/R0
Shear Load/N
160
2.5
Shear load Resistance change
1.2
Displacement/mm (c) comparison between the results after 3 and 4-cycle’s immersion Fig. 7. Load-displacement curve and simultaneous recording of the resistance change for a MWCNT-coated GF yarn embedded in epoxy matrix. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Resin casting process was employed to manufacture the fiberbundle pull-out specimens. Fig. 1 shows the scheme of fiber-bundle pull-out specimen with embedded GF yarn and the pull-out test installation. Fiber-bundle pull-out tests were carried out on a Zwick/Roell Z010 tensile testing device with assistance of a designed fixture under displacement control (speed: 2 mm/min). GF yarn with the epoxy host was placed on the fixture in fiber-bundle pull-out tests. One end of GF yarn was fixed to the load cell after across the channel in the designed fixture, as shown in Fig. 1c. The diameter of the channel was 4 mm, and it was sufficiently larger than the fiber-bundle diameter and narrower
4. IFSS enhancement and interfacial damage detection Quite a large number of test methods have been developed in order 165
Composites Part A 112 (2018) 161–167
B. Yang et al.
Contact resistance
Cracks
Tunnel resistance
Fig. 8. Schematic diagram illustrating the changed electric resistance transmission mode by cracks in interphase sensor.
during the pull-out, the microcracks in the interphase initiates and accumulates, and initial debonding happens when the microcrack density nearly approach saturation. With the cracks further propagates, complete interfacial debonding finally happens. Because formation and propagation of these microcracks in interphase sever the electrical paths, the contact resistance is substituted by tunnel resistance between different MWCNTs in the sensor, as shown in Fig. 8.
to determine the interfacial interaction parameters [5]. According to the loaded method, Zhandarov et al. [30] classified these methods into two major categories: Fiber loaded and matrix loaded. Analysis of fracture surfaces of some laminated composites shows that it is fiber bundle rather than single fiber that is often pulled out during fracture process [21,31]. Thus, compared with others, fiber-bundle pull-out test has a wide range of applications since it is more practical to the real failure modes [5,23,32]. From fiber-bundle pull-out results in Fig. 5, it can be concluded that the MWCNT-coated method is effective with respect to enhancing the IFSS between GF yarn and epoxy. In details, IFSS of pristine and control sample is 30.23 MPa. The IFSS increases to 31.34, 32.28, 35.11, 37.95, and 34.95 MPa after immersed in the MWCNT solution for 1 to 5 cycles, respectively. By comparison, IFSS of the specimens after 4-cycle’s immersion increased by 26% relative to the pristine interphase. In our previous work [4], IFSS of the composites was enhanced by 15.5% by incorporating nano-SiO2 particles in interfacial region. Compared with that, application of MWCNT in the interphase shows increased fiber-matrix stress transfer efficiency via the combination of mechanical interlocking and physical adhesion. GF yarn after pulled out were observed in Fig. 6. As can be seen in the figure, the pristine GF yarn is very smooth, while residual epoxy on MWCNTcoated GF yarn increases with the increasing of immersion cycles. Based on the built-in interphase sensor, the interfacial characterizations during pull-out test were monitored. Fig. 7a and b illustrate the shear load-displacement curve and simultaneous recording of the resistance change. For comparison, Fig. 7c plots the curves together to show the difference. Zhou et al. [33] divided the debonding process into 4 stages, including: initial debonding, partial debonding, maximum debonding and complete debonding (or initial frictional pull-out). The stress expression of the corresponding debonding stage is given in Fig. 7a. However, only compete debonding stage is observed in Fig. 7b. This is mainly due to that the interfacial characterizations were affected by the MWCNT, and discussions on this aspect can be fond in our previous work [4]. The electrical resistance change versus load-displacement curve of the interphase in pull-out tests shows that the relative resistance (ΔR/ R0) increases continuously with shear load. In Fig. 7a, upon loading, the resistance shows the continuously increasing tendency with stress until the initial debonding occurs at σd0 . ΔR/R0 shows a sudden increase at the transition point at different interfacial debonding stages. This phenomenon is caused by flaws and cracks induced from tensile load breaking apart the electrically percolating network. Sharply increase in ΔR/R0 at the complete debonding stage and fluctuant resistance in the frictional stage are also found in Fig. 7b. Different damage process in Fig. 7c implies different stress levels and damage modes [15]. Initially
5. Conclusions In summary, MWCNT was utilized to enhance the IFSS and in-situ monitoring the interfacial stress state of GF/epoxy composites. Processing condition of the interphase sensor has been systematically discussed. IFSS of the specimens after 4-cycle’s immersion reaches the maximum value among all the considered cases. IFSS increased by 26% relative to the pristine interphase after immersion in the MWCNT solution with N = 6.25 mg/ml. The solution was ultrasonic dispersed for 120 min. The resistance of MWCNT sensor shows a linear dependence on test length, while it reaches a percolation threshold after 5-cycle’s immersion. Fiber pull-out test in combination with simultaneous recording of the resistance demonstrates the application potential of the developed interphase sensor for detecting interfacial debonding process in GF/epoxy composites. Acknowledgement This work was supported by National Natural Science Foundation of China (No. 11702097) and the Fundamental Research Funds for the Central Universities (No. 222201714015). References [1] Gao L, Thostenson ET, Zhang Z, Chou TW. Sensing of damage mechanisms in fiberreinforced composites under cyclic loading using carbon nanotubes. Adv Funct Mater 2009;19:123–30. [2] Guermazi N, Tarjem AB, Ksouri I, Ayedi HF. On the durability of FRP composites for aircraft structures in hygrothermal conditioning. Compos Part B-Eng 2016;85:294–304. [3] Mahmood H, Tripathi M, Pugno N, Pegoretti A. Enhancement of interfacial adhesion in glass fiber/epoxy composites by electrophoretic deposition of graphene oxide on glass fibers. Compos Sci Technol 2016;126:149–57. [4] Yang B, Zhang J, Zhou L, Wang Z, Liang W. Effect of fiber surface modification on the lifetime of glass fiber reinforced polymerized cyclic butylene terephthalate composites in hygrothermal conditions. Mater Design 2015;85:14–23. [5] Zhou J, Li Y, Li N, Hao X, Liu C. Interfacial shear strength of microwave processed carbon fiber/epoxy composites characterized by an improved fiber-bundle pull-out test. Compos Sci Technol 2016;133:173–83. [6] Liu L, Jia C, He J, Zhao F, Fan D, Xing L, et al. Interfacial characterization, control and modification of carbon fiber reinforced polymer composites. Compos Sci
166
Composites Part A 112 (2018) 161–167
B. Yang et al.
graphene coating. Compos Sci Technol 2016;129:38–45. [21] Hasanzadeh M, Mottaghitalab V, Babaei H, Rezaei M. The influence of carbon nanotubes on quasi-static puncture resistance and yarn pull-out behavior of shearthickening fluids (STFs) impregnated woven fabrics. Compos Part A-Appl S 2016;88:263–71. [22] Yang B, Xuan F, Chen S, Zhou S, Gao Y, Xiao B. Damage localization and identification in WGF/epoxy composite laminates by using Lamb waves: Experiment and simulation. Compos Struct 2017;165:138–47. [23] Hao W, Yao X, Ma Y, Yuan Y. Experimental study on interaction between matrix crack and fiber bundles using optical caustic method. Eng Fract Mech 2015;134:354–67. [24] Rausch J, Mader E. Health monitoring in continuous glass fibre reinforced thermoplastics: Manufacturing and application of interphase sensors based on carbon nanotubes. Compos Sci Technol 2010;70:1589–96. [25] Rausch J, Mader E. Health monitoring in continuous glass fibre reinforced thermoplastics: Tailored sensitivity and cyclic loading of CNT-based interphase sensors. Compos Sci Technol 2010;70:2023–30. [26] Thostenson ET, Chou TW. Real-time in situ sensing of damage evolution in advanced fiber composites using carbon nanotube networks. Nanotechnology 2008;19:215713. [27] Yu B, Jiang Z, Tang X, Yue C, Yang J. Enhanced interphase between epoxy matrix and carbon fiber with carbon nanotube-modified silane coating. Compos Sci Technol 2014;99:131–40. [28] Zhang J, Zhuang R, Liu J, Scheffler C, Mader E, Heinrich G, et al. 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;12:115–20. [29] Naghashpour A, Hoa SV. A technique for real-time detecting, locating, and quantifying damage in large polymer composite structures made of carbon fibers and carbon nanotube networks. Struct Health Monit 2014;14:35–45. [30] Zhandarov S, Mader E. Determining the interfacial toughness from force-displacement curves in the pull-out and microbond tests using the alternative method. Int J Adhes Adhes 2016;65:11–8. [31] Sharma K, Kaushalyayan KS, Shukla M. Pull-out simulations of interfacial properties of amine functionalized multi-walled carbon nanotube epoxy composites. Comp Mater Sci 2015;99:232–41. [32] Qi G, Zhang B, Du S, Yu Y. Estimation of aramid fiber/epoxy interfacial properties by fiber bundle tests and multiscale modeling considering the fiber skin/core structure. Compos Struct 2017;167:1–10. [33] Zhou LM, Mai YW, Baillie C. Interfacial debonding and fibre pull-out stresses. J Mater Sci 1994;29:5541–50.
Technol 2015;121:56–72. [7] Kocsis JK, Mahmood H, Pegoretti A. Recent advances in fiber/matrix interphase engineering for polymer composites. Prog Mater Sci 2015;73:1–43. [8] Hong S, Jolandan MM, Naraghi M. Controlling the wettability and adhesion of carbon fibers with polymer interfaces via grafted nanofibers. Compos Sci Technol 2015;117:130–8. [9] Mader E, Mai K, Pisanova E. Interphase characterization in polymer composites monitoring of interphasial behavior in dependence on the mode of loading. Compos Interfaces 2012;7:133–47. [10] Sharma M, Gao S, Mader E, Sharma H, Wei LY, Bijwe J. Carbon fiber surfaces and composite interphases. Compos Sci Technol 2014;102:35–50. [11] Cech V, Prikryl R, Balkova R, Vanek J, Grycova A. The influence of surface modifications of glass on glass fiber/polyester interphase properties. J Adhes Sci Technol 2003;17:1299–320. [12] Downey MA, Drzal LT. Toughening of carbon fiber-reinforced epoxy polymer composites utilizing fiber surface treatment and sizing. Compos Part A-Appl S 2016;90:687–98. [13] Zhang RL, Gao B, Du WT, Zhang J, Cui HZ, Liu L, et al. Enhanced mechanical properties of multiscale carbon fiber/epoxy composites by fiber surface treatment with graphene oxide/polyhedral oligomeric silsesquioxane. Compos Part A-Appl S 2016;84:455–63. [14] Cai M, Takagi H, Nakagaito AN, Li Y, Waterhouse GIN. Effect of alkali treatment on interfacial bonding in abaca fiber-reinforced composites. Compos Part A-Appl S 2016;90:589–97. [15] Chi PIG, Uicab OR, Barrera CM, Calderon JU, Escamilla GC, Pedram MY, et al. Influence of aramid fiber treatment and carbon nanotubes on the interfacial strength of polypropylene hierarchical composites. Compos Part B-Eng 2017;122:16–22. [16] Dziendzikowski M, Dragan K, Katunin A. Localizing impact damage of composite structures with modified RAPID algorithm and non-circular PZT arrays. Arch Civ Mech Eng 2017;17:178–87. [17] Li W, He D, Dang Z, Bai J. In situ damage sensing in the glass fabric reinforced epoxy composites containing CNT-Al2O3 hybrids. Compos Sci Technol 2014;99:8–14. [18] Nanda N, Kapuria S. Spectral finite element for wave propagation analysis of laminated composite curved beams using classical and first order shear deformation theories. Compos Struct 2015;132:310–20. [19] Wen J, Xia Z, Choy F. Damage detection of carbon fiber reinforced polymer composites via electrical resistance measurement. Compos Part B-Eng 2011;42:77–86. [20] Hao B, Ma Q, Yang S, Mader E, Ma PC. Comparative study on monitoring structural damage in fiber-reinforced polymers using glass fibers with carbon nanotubes and
167
