epoxy composites using amino functionalized MWCNTs

Composite Structures 94 (2012) 2397–2406

Contents lists available at SciVerse ScienceDirect

Composite Structures journal homepage: www.elsevier.com/locate/compstruct

Improvements in mechanical and thermo-mechanical properties of e-glass/epoxy composites using amino functionalized MWCNTs M.M. Rahman a, S. Zainuddin b,⇑, M.V. Hosur b, J.E. Malone a, M.B.A. Salam a, Ashok Kumar c, S. Jeelani b a

Department of Mechanical Engineering, Tuskegee University, Tuskegee, AL 36088, United States Department of Material Science and Engineering, Tuskegee University, Tuskegee, AL 36088, United States c Construction Engineering Research Laboratory, US Army Engineer Research and Development Center, Champaign, IL 61821-9005, United States b

a r t i c l e

i n f o

Article history: Available online 28 March 2012 Keywords: Functionalized MWCNTs Sonication Calendaring Mechanical and thermo-mechanical properties

a b s t r a c t The prime objective of this work is to optimize the mechanical and thermo-mechanical properties of e-glass/epoxy composites by utilizing amino-functionalized multi-walled carbon nanotubes (MWCNTs–NH2) through a combination of dispersion method. At first, 0.1–0.4 wt.% of MWCNT–NH2 was integrated into SC-15 epoxy suspension using a combination of ultra-sonication and calendaring techniques. E-glass/epoxy nanocomposites were than fabricated at elevated temperature with the modified resin using hand layup and compression hot press. 3-Point flexural and dynamic mechanical analysis (DMA) results demonstrated a linearly increasing trend in properties from 0 to 0.3 wt.% loading. Micrographs of MWCNTs incorporated epoxy and e-glass/epoxy samples revealed uniform dispersion of MWCNTs in epoxy, good interfacial adhesion between CNTs and polymer, and improved interfacial bonding between fiber/matrix at 0.3 wt.% loading. An improved dispersion and hence an improved crosslink interaction between MWCNT–NH2 and epoxy lead to the stronger shift of the mechanical and thermo-mechanical properties of the composites. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Glass fiber reinforced polymer (GFRP) composites are increasingly used in many structural applications replacing metallic materials due to their low cost, high strength, high chemical resistance and excellent insulating properties. In most of these GFRP applications, fiber plays a major role in tensile load carrying capacity of a composite structure while compressive, bending, inter-laminar shear properties depend on the selection of matrix [1]. In a fiber/ matrix composite, matrix being the weakest component is first to fail upon such loading. Therefore, enhancement of matrix properties is desired to enhance the overall performance of fiber reinforced polymer (FRP) composites under such loading. In past two decades, researchers have successfully tailored the matrix properties by incorporating inorganic nanoparticles into epoxy polymer and its fiber-reinforced composites [2–5]. Among the nanoparticles, carbon nanotubes (CNTs) have emerged as potential candidates for modification of matrix because of its exceptional strength and stiffness, high flexibility, diameter dependent specific surface area and high aspect ratio. According to Reynaud et al. [6], an interface of 1 nm thick represents roughly 0.3% of the total volume of the polymers in microparticle filled composites, ⇑ Corresponding author. Tel.: +1 334 724 4222; fax: +1 334 724 4224. E-mail address: [email protected] (S. Zainuddin). 0263-8223/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compstruct.2012.03.014

whereas it can reach 30% of the total volume in nanocomposites. The high specific surface area of CNTs provide desirable interface for stress transfer but can introduce strong attractive forces in between CNTs causing excessive agglomeration and produce unwanted stress concentrations which may act as a precursors for failure. However, multi-walled carbon nanotubes (MWCNTs) have an approximate specific surface area of 200 m2/g or less which is lower than SWNTs due to their much larger diameter and multiple graphene walls and thus exhibit better dispersibility. One of the most important parameter in fabricating CNTs reinforced composites is the dispersion of CNTs itself because of their strong tendency to re-agglomerate. Kim et al. have reported that degree of CNTs dispersion into epoxy strongly affected the matrix-dominated mechanical properties [7]. Various methods to disperse nanotubes in polymer resins, such as stirring, sonication and high shear mixing have been reported in literatures [8–10]. Some researchers found that chemical functionalization of the CNT surface improved interfacial interaction between CNTs and matrix, and the dispersion of CNTs into the matrix [11–13]. Previously, it has been shown that amino functionalization in CNTs enhances the dispersibility in the epoxy matrix [14] and multiwalled carbon nanotubes (MWCNTs) exhibits better dispersion in polymer than the single walled or double walled CNTs [1]. In addition, interfacial adhesion between the CNTs and polymer is also a critical issue. In order to have sufficient stress transfer

2398

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

from the matrix to the CNTs and to efficiently use the potential of CNTs as structural reinforcement, a strong interfacial adhesion between the CNTs and polymer is desired. The interfacial adhesion between CNTs and matrix was reported to improve by functionalizing the CNTs. Tailored amino, carboxyl or glycidyl groups enable covalent bonding between CNTs and epoxy resulting in improved interfacial bonding. The positive effects of functionalized CNTs on the mechanical properties are reported by various researchers [9,15,16]. Kim et al. compared the effect of silane and acid treated CNTs in carbon/epoxy composites and found 15% and 10% enhancement in flexural strength and modulus [17]. Moniruzzaman et al. found an improvement of 10–15% in flexural properties with 0.05 wt.% single-walled nanotubes (SWNTs) [18]. While numerous studies are conducted in the last two decades on the CNT reinforced composites, effective dispersion of CNTs in polymer is still a challenge and the experimental mechanical properties are still far below than the theoretical value. In addition, studies on glass fiber reinforced polymer composites incorporated with amino functionalized MWCNTs using a combination of sonication and calendaring techniques is not yet been reported to the best of our knowledge. In this work, MWCNT–NH2 reinforced e-glass/epoxy nanocomposites were processed by combining sonication and calendaring dispersion techniques, utilizing functionalized MWCNTs, combining fabrication processes and curing the composites at higher temperature. Rheological, 3-point bend and dynamic mechanical analysis tests were performed to investigate the effect of functionalized MWCNTs on resin viscosity, flexural and thermo-mechanical properties of e-glass/epoxy composites. Effect of combined sonication and calendaring techniques on MWCNTs dispersion in epoxy suspension were investigated using transmission electron microscope (TEM). In addition, effect of interaction of MWCNTs with epoxide groups and fiber/matrix bonding were investigated using scanning electron microscope (SEM). All results were compared with the control (reference) epoxy composites results containing no MWCNTs. 2. Experimental 2.1. Materials SC-15 epoxy resin used in this study was supplied by Applied Poleramic Inc. It is a two part cyclo-aliphatic amine type epoxy resin (Part A: diglycidylether of bisphenol A, aliphatic diglycidyl and Part B: hardener). Multi-walled carbon nanotubes functionalized using amino groups (–NH2) was purchased from Nanocyl Inc. These nanotubes were of diameter 10 nm, average length of several microns and carbon purity >95%. Fig. 1 shows the scanning electron micrographs of as received carbon nanotubes at different magnifications. High specific surface area and cotton-like entanglements caused the formation of agglomerates as reported by Reynaud et al. [6]. As reinforcement in composites, e-glass woven fabric with a density of 2.58 g/cm3 and a single fiber diameter of 14– 16 lm was procured from Fiber Glast Development Corporation. E-glass fibers were sized with 0.5 wt.% epoxy silanes to increase the compatibility and to have better adhesion between fibers and epoxy matrix.

Fig. 1. SEM pictures of as-received MWCNTs at magnification of 500 (A) and 10000 (B).

pressure and temperature in the system, so the mixture was cooled down to room temperature in a refrigerator cooler maintained at 5 °C. To further improve the dispersion of MWCNTs, the sonicated mixture was then passed through three rollers as shown in Fig. 2. In this three roll process, roller 1 and 3 rotates in the same direction whereas the roller 2 placed in between rotates in the opposite direction thereby inducing high shearing in the mixture. A varying gap setting between the rolls and multiple passes of 20 lm (1st pass), 10 lm (2nd pass) and 5 lm (3rd pass) was used to induce high shear force in the mixture. The induction of high shear forces further de-agglomerates and improves the dispersion of CNTs in resin. Roller speed of the three rolls was maintained at a ratio of 1:3:9 with a maximum speed of 200 rpm was maintained in all the three passes.

2.2. Manufacturing process

2.2.2. Mixing of Part A and Part B The hardener, Part B was added as per stoichometric ratio (Part A: Part B = 10:3) to the modified mixture and mixed with a high speed mechanical stirrer for 10 min at 800 rpm. The mixture was then placed in a desiccator for 30 min to remove the volatile impurities and entrapped air bubbles that were generated due to intense mechanical mixing.

2.2.1. Dispersion of MWCNT–NH2 into epoxy resin At first, MWCNT–NH2 was mixed manually with epoxy resin Part-A as per calculated weight ratio. The mixture was then sonicated at room temperature for 1 h at 35% amplitude and 40 s on/ 20 s off cycle pulse mode. Sonication process induces elevated

2.2.3. Manufacturing of fiber reinforced nanocomposites E-glass/epoxy nanocomposites were fabricated by using a combination of hand lay-up and compression hot press techniques. Eglass woven fabric layers were properly stacked into seven plies and the orientation of fiber within the fabric was kept constant.

2399

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

Sonicator probe Cooling bath Mixtures Collection

3-roll shear mixing process for further MWCNT Dispersion of MWCNTs in epoxy Part A using

Mechanical mixing of modified mixtures and epoxy Thickness 3.25 mm

250 mm

250

Consolidation of laminates in hot press

Final e-glass/epoxy laminates

Fig. 2. Schematic of MWCNTs dispersion and e-glass/epoxy composite fabrication.

Epoxy resin modified with MWCNTs was spread uniformly on each fabric layer and the laminate was consolidated by applying 133 kN force in a hot press. The temperature of hot press was maintained at 60 °C for 4 h for curing. Elevated temperature was selected to reduce the curing period and thus production time as per recommendation of resin manufacturer. After completion of curing, the temperature of the press was gradually reduced at a rate of 2 °C/ min to avoid any unwanted shrinkage in the laminate. Finally, the cured panel was taken out and post cured in an oven for 5 h at 100 °C. The final thickness of the panel was measured to be 3.25 mm. Fig. 2 shows the schematic of matrix modification and e-glass/epoxy composite fabrication.

of 30 mm  30 mm  2 mm were cut from the panels and weighed precisely. The samples were then submerged into a bath of 80% concentrated nitric acid for about 5 h maintained at 75 °C. After 5 h, once the matrix was digested completely, the fibers were taken out and washed several times by using acetone and distilled water. The fibers were then dried at 100 °C and weighed. The corresponding fiber volume, matrix volume and void volume fractions were calculated using the equations:

Fiber volume fraction;

mf ¼

W=F  100 w=C

Matrix volume fraction;mm ¼ 2.3. Material characterization 2.3.1. Rheological properties Rheology measurements were performed to observe the effect of increasing weight percent loading of MWCNTs on viscosity and the changes in viscosity as a measure of shear rate at constant temperature. Measurements were performed with AR 2000 Rheometer in ETC control mode using parallel plate geometry at 1000 lm gap settings. Flow sweep was used to vary the shear rate from 0.1 rad/s to 100 rad/s by keeping the temperature constant. 2.3.2. Calculation of fiber volume fraction and void content Matrix digestion test was conducted according to ASTM D 317199(2004) to determine the fiber volume fraction and void content of the fabricated composite panels [19]. Samples with a dimension

ðw  WÞ=M  100 V

Void volume fraction; mv ¼ 100  ðmm þ mf Þ

ð1Þ

ð2Þ ð3Þ

where W is the weight of fiber in the composite, w is the weight of the initial composite specimen, F is the fiber density, M is the matrix density, V is the volume of composite and C is the composite density. 2.3.3. Mechanical and thermo-mechanical characterization Three point bend flexure test was conducted according to ASTM D790-02 [20]. Samples were precisely cut to 52 mm  3.25 mm  12.25 mm dimensions to maintain span-length ratio of 16:1 and a minimum of seven specimens were tested. The test was performed using a Zwick–Roell Z 2.5 testing machine in displacement control mode with a crosshead speed of 1.2 mm/min. Load–deflection data

2400

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

for each sample was collected. Flexural modulus was calculated from the slope of the tangent to initial straight line portion of stress–strain plot. It was determined by the following equation:

Slope of tangent  ðspan lengthÞ2 4  width  ðthicknessÞ2

ð4Þ

The maximum stress at failure on the tension side of a flexural specimen was considered as the flexural strength of the material. Thus, using the homogeneous beam theory, the flexural strength in a 3-point flexural test was determined using Eq. (5).

Flexure strength ¼

3  peak load  span length 2  width  ðthicknessÞ2

Control (Reference) 0.1 wt. % 0.2 wt. % 0.3 wt. % 0.4 wt. %

Viscosity (Pa. s)

Flexure modulus ¼

100

10

1

ð5Þ

Similarly, flexure strain to failure was calculated from the following equation:

Flexure strain ¼

1

ðspan lengthÞ2 ð6Þ

Dynamic mechanical thermal analysis (DMTA) was performed according to ASTM D4065-01 [21] to study the viscoelastic behavior of composite samples. It is also an effective way to investigate the dispersion state of MWCNTs. In this test, the width of the samples was 12 mm and span length to thickness ratio was 10. Tests were conducted in dual cantilever beam mode with a frequency of 1 Hz and amplitude of 15 lm. The temperature was ramped from 30 to 200 °C at a rate of 5 °C/min. A minimum of five specimens of each type were tested. Storage modulus, loss modulus and glass transition temperature of samples were determined from the tests to evaluate the viscoelastic and damping properties of control and MWCNTs incorporated e-glass/epoxy composites, respectively. 2.3.4. Micrographic analysis Dispersion state of MWCNTs in epoxy resin was investigated by transmission electron microscopy (TEM) using a Zeiss EM10 Transmission Electron Microscope operated at 60 kV. The analysis of fracture surfaces was carried out using a JEOL JSM-6400 scanning electron microscope (SEM) at 5 kV accelerating voltage. Specimen surfaces were coated with a thin gold film to increase their conductance for SEM observation. 3. Results and discussion 3.1. Rheological properties Viscosity calculations were performed immediately after the degasification of mixture (Part A & B + MWCNTs). Fig. 3 shows viscosity as a function of shear rate for control and 0.1–0.3 wt.% MWCNTs–NH2 loaded epoxy resin. Shear viscosity of a polymer is principally divided into Newtonian and shear thinning regions, respectively. With increasing shear rate, a shear thinning behavior was observed and a declining trend in shear viscosity was noticed. However, at high shear rate, a Newtonian behavior was observed in all the samples independent of shear rate. Viscosity of 0.1 wt.% MWCNTs reinforced epoxy sample remains at comparable range with control resin. However, an increasing trend in viscosity was observed with increasing weight percent of MWCNTs content. Viscosity of 0.4 wt.% resin samples was found to increase more than 100% in comparison to control resin samples. This implies that strong particle–particle interaction of CNTs is one major factor that leads to an increase in shear viscosity with increasing CNTs content that may result in poor dispersion.

10

100

Shear rate (1/s)

6  maximum deflection at center  thickness

Fig. 3. Viscosity vs. shear rate response of control and NH2–MWCNTs incorporated epoxy resin.

3.2. Fiber volume fraction Fiber volume fraction and void content in control and 0.1– 0.4 wt.% MWCNTs e-glass/epoxy composites calculated by matrix digestion test is summarized in Table 1. The average fiber volume fraction of the laminates was 58–60% with void content of 4–6%, respectively. It can be seen that with increase of MWCNTs wt.% loading, the percentage of void content was also increased and can be attributed majorly due to increase of resin viscosity. With increasing CNTs wt.% loading, the resin viscosity increased in comparison to control resin shown in Fig. 3. The increase in resin viscosity may hinder the removal of entrapped bubbles and volatile impurities from the systems during processing. 3.3. Flexural properties Typical stress–strain behavior obtained from the 3-point bend tests is shown in Fig. 4. Summary of the flexural test results and the variation in properties as a function of MWCNTs content are shown in Figs. 5a and b, respectively. The positive effect of MWCNT incorporation is clearly evident from Figs. 5a and b. Flexural strength and modulus were improved linearly with a maximum enhancement of 38% and 22% in 0.3 wt.% samples in comparison to control (reference) samples as shown in Fig. 5c. Likewise, flexural strain to failure was also found maximum at 0.3 wt.% loading i.e. 27% higher in comparison to control samples. However a decrease in these properties was found at 0.4 wt.% MWCNTs loading. Fig. 6a and b shows the transmission electron micrographs (TEM) of 0.3–0.4 wt.% MWCNTs reinforced epoxy samples. An improved dispersion of MWCNTs in matrix evident from Fig. 6a provides more sites facilitating higher chances for polymer and MWCNTs interaction. In addition, the amino functional groups present on the outer ring of MWCNTs and epoxide group of epoxy resin causes the interfacial reaction resulting in strong covalent bond evident from scanning electron micrograph (SEM) shown in Fig. 7. Generally after mixing epoxy Part A and MWCNT–NH2, the interfacial reaction takes place between amine functional groups of CNTs and epoxide groups of DGEBA resin which is consisted of ring opening reactions followed by a cross-linking reaction as shown in Fig. 8 [14]. This crosslink reaction creates interlocking structure in the resin blend through the covalent bond which facilitates impediment of the mobility of polymer chains in the system.

2401

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406 Table 1 Matrix digestion test results of control and NH2–MWCNT incorporated e-glass/epoxy composites. Control

0.1

0.2

0.3

0.4

Fiber volume fraction (mf ) Void fraction (md )

59.54 ± 1.3 3.94

58.09 ± 1.6 4.13

58.35 ± 1.3 4.67

58.58 ± 2.7 5.48

58.19 ± 2.2 6.81

Flexural stress (MPa)

600

Increase w.r.t control or reference (%)

MWCNT weight percent content (wt.%) Results of matrix digestion test

Control 0.1 wt.% 0.2 wt.% 0.3 wt.% 0.4 wt.%

500 400 300 200 100 0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

Flexural strain (%)

40

0.1 wt.% 0.2 wt.% 0.3 wt.% 0.4 wt.%

35 30 25 20 15 10 5 0

Flexural strength

Flexural modulus

Flexural strain to failure

Fig. 5c. Improvement of flexural properties of e-glass/epoxy composites using NH2–MWCNTs.

Fig. 4. Flexural stress–strain response of control and NH2–MWCNTs incorporated e-glass/epoxy composites.

600

32

550

Strength

450

28 Modulus

400 26

350 300

Flexural Modulus (GPa)

Flexural Strength (MPa)

30 500

24 250 22

200 0.0

0.1

0.2

0.3

0.4

Weight Percentages (%) Fig. 5a. Effect of NH2–MWCNTs content on flexural strength and modulus of eglass/epoxy composites.

Flexure strain (%)

2.6

2.4

2.2

2.0

Fig. 6. TEM micrographs of (a) 0.3 wt.% and (b) 0.4 wt.% loading of NH2–MWCNTs in epoxy resin.

1.8 0.0

0.1

0.2

0.3

0.4

CNT content (wt.%) Fig. 5b. Effect of NH2–MWCNTs content on flexural strain of e-glass/epoxy composites.

The noticeable increase in flexural strength and modulus from 0 to 0.3 wt.% loading thus can be attributed to the better dispersion of MWCNTs and better interfacial interaction between the amino functionalized MWCNTs and epoxy matrix. This improved

2402

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

Fig. 7. SEM micrograph showing the interfacial adhesion between MWCNTs and epoxy resin.

interfacial interaction may have facilitated in stress transfer during loading and thus resulting increase in flexural properties. Similar reasons for improvement in epoxy/matrix incorporated with MWCNTs were also reported previously [22,23]. However, a decrease in flexural properties at 0.4 wt.% loading can be attributed to strong attractive forces between MWCNTs leading to excessive agglomeration. Due to poor dispersion evident from Fig. 6b, MWCNTs remains in small bundles or agglomerates

form in the resin blends. Dı´ez-Pascual et al. reported that shear slippage of individual nanotubes within these bundles may occur [23]. Due to this factor, a reduction in load transfer capability between CNTs and matrix systems may be occurred and hence the properties may have decreased at 0.4 wt.% loading. So the flexural properties were improved in small loadings as long as it is dispersed in the matrix properly and as well in case of higher loading systems, uniform dispersion of CNTs into the matrix is very difficult. Moreover incorporation of higher weight percent loading increased the viscosity of matrix which may have impeded the dispersion [3]. Song et al. also performed rheological study and stated that poorly dispersed CNTs within epoxy have a higher viscosity than that of uniformly dispersed suspensions [24]. In addition, high viscosity of the resin resulting from poor dispersion may cause poor wetting of the glass fiber during laminate fabrication and hence poor adhesion between glass fiber and matrix [25]. This may be another factor of getting a decrease in strength, strain and modulus s at 0.4 wt.% loading of MWCNTs. A gradual increase in strain to failure was also observed in eglass/epoxy composites from 0 to 0.3 wt.% loading. It was reported that polymer matrix added with amino-functionalized MWCNTs can increase the fracture energy as crack propagation can be resisted by bringing up the crack faces of nanotubes [22]. Zhou et al. reported that the crack propagation changes direction as it crosses CNTs during failure process due to bridge effect which prevents crack opening [5]. As a result, crack initiation and propagation becomes difficult in the laminates reinforced by CNTs than

(a) Reaction of epoxide group (DGEBA) with primary amine of functionalized MWCNT H MWCNT

N

O

+

H 2C

MWCNT

N H

C H

C H2

CH OH

H

Epoxy molecule # 1 (b) Reaction of epoxide group (DGEBA) with primary amine of functionalized MWCNT

MWCNT

N H

C H2

OH H2C

O

CH

+

OH

H2C

C H

MWCNT

C H

N H2C

Epoxy molecule # 2

H C OH

(c) Cross-linking reaction between epoxy and MWCNT

OH

OH H2C MWCNT

C H

N H2C

H C OH

O

O

+

H 2C

H2C

C H

MWCNT

C H2

CH

C H

N H2C

Epoxy molecule # 3 & 4

H C O

C H2

H C OH

Fig. 8. Schematic representation of interfacial reaction between DGEBA (Part A) and MWCNT–NH2.

2403

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406 Table 2 Flexural strain energy of NH2–MWCNT incorporated e-glass/epoxy composites. Specimen Control

0.1 (wt.%)

0.2 (wt.%)

0.3 (wt.%)

0.4 (wt.%)

1.03 –

1.49 +44.66

1.65 +60.19

1.97 +91.26

1.40 +35.92

that of without reinforced CNTs. It may be one of the reasons due to which the flexural strain to failure in the composites was increased. Carbon nanotubes have a high aspect ratio and shows highly flexible elastic behavior during loading [26]. Because of the strong interfacial bonding with this reinforcement, the nanophased resin systems exhibited higher absorption of energy evident from Table 2. Average flexural strain energy calculated from the area of load–deflection curves showed an increase in strain energy with increase in MWCNTs from 0 to 0.3 wt.% loading. The increase in strain energy thus can be attributed to increase in strain to failure in MWCNTs incorporated e-glass/epoxy composites.

1400 1200

Loss Modulus, Mpa

Flexural strain energy (J) % gain/loss

Control 0.1% CNT 0.2% CNT 0.3% CNT 0.4% CNT

Storage Modulus, Mpa

8000 6000 4000 2000 40

60

80

100

120

140

600 400

20

14000

20

800

0

160

180

200

Temperature,°C Fig. 9. Storage modulus vs. temperature response of e-glass/epoxy composites.

40

60

80

100

120

140

160

180

200

Temperature,°C Fig. 10. Loss modulus vs. temperature response of e-glass/epoxy composites.

0.20

Control 0.1% CNT 0.2% CNT 0.3% CNT 0.4% CNT

0.15

Tan Delta

Dynamic mechanical thermal analysis (DMTA) was performed to observe e-glass/epoxy samples stiffness behavior as a function of temperature and analyze the effect of NH2–MWCNTs on thermo-mechanical performance. Figs. 9–11 show the variation in dynamic measurements of storage modulus, loss modulus and loss tangent as a function of temperature obtained from DMTA. A stronger influence of amino-functionalized CNTs was observed in the dynamic mechanical properties evident from Fig. 12a and b showing the mean storage modulus, loss modulus and glass transition temperature plotted as a function of CNTs content. The improvement as obtained in the DMTA test is shown in Fig. 12c. Storage modulus of laminates in the glassy regions and vicinity of the glass transition temperature was increased upon incorporation of MWCNTs from 0 to 0.3 wt.% loading. However, the addition of nanotubes to the composite laminates was observed to have a slight influence in rubbery region. This behavior can be attributed to the improved dispersion and interaction between the MWCNTs and epoxy due to the formation of covalent bonds between them. The interfacial interaction reduces the mobility of epoxy polymer chain around the nanotubes which leads to stronger shift of elastic properties in final composites below the rubbery region. Above the glass transition temperature i.e. in rubbery region, a slight increase of storage modulus was observed. This can be attributed to the relatively higher molecular motion and higher amplitude of this motion in the rubbery region. Montazeri et al. reported that in rubbery

10000

1000

200

3.4. Viscoelastic/thermo-mechanical properties

12000

Control 0.1% CNT 0.2% CNT 0.3% CNT 0.4% CNT

0.10

0.05

0.00 20

40

60

80

100

120

140

160

180

200

Temperature, deg C Fig. 11. Tan delta vs. temperature response of e-glass/epoxy composites.

state the molecular motion and its amplitude remains high and the macromolecule is not practically in contact with particles for which no shear force act between them [27]. This increase in storage modulus was found from 0 to 0.3 wt.% loading of CNTs. However, a decrease in storage modulus was observed when the nanotube loading was increased to 0.4 wt.% loading. It might be due to the incapability of de-agglomeration which facilitates the molecular motion and movement of chain. Loss modulus of MWCNTs reinforced e-glass/epoxy samples was also found to increase in comparison to control samples. Loss modulus of composites indicates the energy used to deform the material that is dissipated into heat and can be used as a measurement of viscous component or unrecoverable oscillation energy dissipated per cycle. The good dispersed nanotubes must dissipate energy due to resistance against viscoelastic deformation of the surrounding matrix [28]. Vlasveld et al. reported that covalent bond between amino-functionalized CNTs and epoxy improve the efficiency of load transfer from matrix to fillers resulting in an increase in loss modulus due to more energy dissipation in composites [29]. So the decrease of loss modulus at 0.4 wt.% nanotube contents can be ascribed to increasing inclination of agglomeration

2404

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

14000

a

1300

12000

1200

11000

1100

10000

1000

9000

900

8000

800

7000

700

6000

Loss Modulus (MPa)

Storage modulus (MPa)

13000

1400

600 0.0

0.1

0.2

0.3

0.4

Weight percentages (%)

Glass Transition Temperature (deg C)

114 112

b

111.62 ºC 110.55 ºC

110

108.55 ºC

108 106

106.08 ºC 104 102

101.66 ºC

100 0.0

0.1

0.2

0.3

0.4

Weight Percentages (%)

Increase w.r.t control or reference (%)

Fig. 12a and b. Effect of NH2–MWCNTs content on viscoelastic properties of e-glass/epoxy composites.

55 50 45 40 35 30 25 20 15 10 5 0

0.1 wt.% 0.2 wt.% 0.3 wt.% 0.4 wt.%

Storage modulus

Loss modulus

Glass transition temp.

Fig. 12c. Improvement of thermo-mechanical properties of e-glass/epoxy composites using NH2–MWCNTs.

tan delta vs. temperature curve as shown in Figs. 11 and 12b. This gain in thermo-stability can again be interpreted as a reduction in mobility of epoxy matrix around the nanotubes by interfacial interactions. Because this interaction induces different cross-linking regions in epoxy matrix which will eventually reduce the polymer chain motion, a strong shift of glass transition temperature in nanocomposites was observed. In contrast, agglomeration of MWCNTs was observed in case of 0.4 wt.%. During impregnation with fiber, these MWCNTs agglomerates at the fiber interfaces could alter the flow behavior of the matrix which results in nanometer scale porosities [30]. As a result, free volume may be created due to these porosities between polymer molecules which facilitates in polymer chain motion. So these may be a reason for drop in glass transition temperature of e-glass/epoxy samples containing 0.4 wt.% MWCNTs. 3.5. Fracture surface analysis

resulting less energy dissipating in the system under viscoelastic deformation. The addition of MWCNTs in epoxy resulted in shift of glass transition temperature of e-glass/epoxy samples. The glass transition temperature was shifted from 101 °C for control samples to 111 °C for 0.3 wt.% loading MWCNTs obtained from the peak of

Fracture surfaces of neat and nanophased glass fiber reinforced composites were investigated by SEM to explore the reason for enhancement in mechanical and viscoelastic properties of composites. Fig. 13a and b shows the SEM micrographs of fractured single fiber of control and 0.3 wt.% samples at higher magnification after

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

2405

Fig. 13. SEM micrographs of fractured single fiber of (a) control and (b) 0.3 wt.% MWCNTs e-glass/epoxy composites.

Fig. 14. SEM micrographs of fractured surface of (a) control and (b) 0.3 wt.% MWCNTs e-glass/epoxy composites.

3-point flexural test. As evident from these micrographs, the fiber in 0.3 wt.% sample contains considerable amount of epoxy resin residue and rougher fracture surface in comparison to control samples. Fig. 14a and b shows the SEM micrographs of fracture surfaces of control and 0.3 wt.% e-glass/epoxy samples taken at a magnification of x300. As evident from Fig. 14a, control samples exhibited poor fiber–matrix bonding as fiber breakage with distinct pullout scene or interface debonding of fiber–fiber tows. In contrast, no sharp fiber pullout or interface debonding was observed in 0.3 wt.% samples. The epoxy silane present on the surface of e-glass fibers may have also reacted with amino groups of MWCNTs in addition to its reaction with epoxy matrix functional groups shown in Fig. 8. This improved chemical reactivity leading to a strong covalent bonding may have resisted the crack propagation and increased the fiber/matrix bonding in 0.1–0.3 wt.% MWCNTs samples. Thus presence of matrix residue and rougher fracture surface is an indication of good adhesion between fiber and matrix contributing to the improvement in properties of 0.1–0.3 wt.% composites. This increase in matrix residue and rougher surface is an indication of good adhesion between fiber and matrix which may have contributed to the improvement in e-glass/epoxy composite flexural and viscoelastic properties.

4. Conclusion In this study, functionalized MWCNTs were incorporated in eglass/epoxy composites to enhance the flexural and viscoelastic properties. Based on the experimental and micrographic results, the following conclusions are reached: (1) MWCNTs incorporation at low concentrations in e-glass/ epoxy composites increased flexural property and thermal stability of composites. In terms of mechanical and thermo-mechanical properties, 0.3 wt.% MWCNTs loading was found optimum. (2) Flexural test results of 0.3 wt.% MWCNTs incorporated eglass/epoxy samples showed a maximum improvement in strength, modulus and strain to failure by 37%, 21% and 21%, respectively in comparison to control samples. (3) Dynamic mechanical thermal analysis (DMTA) results of 0.3 wt.% samples showed a maximum improvement of 41% in storage modulus, 52% in loss modulus and an increase of 10 °C in glass transition temperature in comparison to control samples. (4) Improvement at optimum loading of 0.3 wt.% explores a good interfacial interaction and effective load transfer between CNTs and epoxy system due to better dispersion.

2406

M.M. Rahman et al. / Composite Structures 94 (2012) 2397–2406

Morphological studies have revealed that amino-functionalized MWCNTS promotes good adhesion between glass fiber and epoxy matrix by modifying the matrix adhesive properties and hence, the properties of composite increases. (5) Overall this work showed that amino-functionalized MWCNTs can significantly improve the mechanical and thermo-mechanical/viscoelastic properties of e-glass/epoxy composites. However, the MWCNTs dispersion methods used in this study was able to disperse only from 0 to 0.3 wt.% loading.

Acknowledgements The authors would like to acknowledge US Engineer Research Development Corporation – Construction Engineering Research Laboratory (ERDC-CERL) and NASA-EPSCoR for funding this work. References [1] Mallick PK. Fiber reinforced composites. New York: CRC press; 2007. [2] Farhana P, Zhou Y, Rangari V, Jeelani S. Testing and evaluation on the thermal and mechanical properties of carbon nano fiber reinforced SC-15 epoxy. Mater Sci Eng A 2005;405:246–53. [3] Gojny FH, Wichmann MHG, Fiedler B, Bauhofer W, Schulte K. Influence of nano modification on the mechanical and electrical properties of conventional fibre-reinforced composites. Compos Pt A: Appl Sci Manuf 2005;36(11): 1525–35. [4] Mahfuz H, Adnan A, Rangari VK, Jeelani S, Jang BZ. Carbon nanoparticles/ whiskers reinforced composites and their tensile response. Compos Part A: Appl Sci Manuf 2004;35:519–27. [5] Zhou Y, Farhana P, Lewis L, Jeelani S. Experimental study on the thermal and mechanical properties of multiwalled carbon nanotube-reinforced epoxy. Mater Sci Eng A 2007;452:657–64. [6] Reynaud E, Gauthier C, Perez J. Nanophases Polym. Rev Metall 1999;96: 169–76. [7] Kim M, Park Y-B, Okoli O, Zhang C. Processing, characterization, and modeling of carbon nanotubes reinforced multiscale composites. Compos Sci Technol 2009;69:335–42. [8] Hassan M, Zainuddin S, Parker MR, Tariq A, Rangari VK, Jeelani S. Reinforcement of SC-15 epoxy with CNT/CNF under high magnetic field: an investigation of mechanical and thermal response. J Mater Sci 2009;44:1113–20. [9] Liao YH, Olivier MT, Liang ZY, Zhang C, Wang B. Investigation of the dispersion process of SWNTs/SC-15 epoxy resin nanocomposites. Mater Sci Eng A 2004;385:175–81. [10] Gojny FH, Nastalczyk J, Roslaniec Z, Schulte K. Surface modified multi-walled carbon nanotubes in CNT/epoxy-composites. Chem Phys Lett 2003;370:820–4.

[11] Yaping Z, Aibo Z, Quinghua C, Jiaoxia Z, Rongchang N. Functionalized effect on carbon nanotube/epoxy nano-composites. Mater Sci Eng A 2006;435– 436:145–9. [12] Xie X, Mai YW, Zhou XP. Dispersion and alignment of carbon nanotubes in polymer matrix: a review. Mater Sci Eng 2005;49:89–112. [13] Sahoo NG, Rana S, Cho JW, Li L, Chan SH. Polymer nanocomposites based on functionalized carbon nanotubes. Prog Polym Sci 2010;35:837–67. [14] Ma PC, Mo SY, Tang BZ, Kim JK. Dispersion, interfacial interaction and reagglomeration of functionalized carbon nanotubes in epoxy composites. Carbon 2010;48:1824–34. [15] Allaoui A, Bai S, Cheng HM, Bai JB. Mechanical and electrical properties of a MWNT/epoxy composite. Compos Sci Technol 2002;62:1993–8. [16] Bai J, Allaoui A. Effect of the length and aggregate size of MWNTs on the mechanical and electrical properties of the nanocomposites. Compos Part A 2003;34:689–94. [17] Kim MT, Rhee KY, Lee JH, Hui D, Lau A. Property enhancement of a carbon fiber/epoxy composite by using carbon nanotubes. Compos Part B 2011;42:1257–61. [18] Moniruzzaman M, Du F, Romero N, Winey K. Increased flexural modulus and strength in SWNT/epoxy composites by a new fabrication method. Polymer 2006;47:293–8. [19] Annual book of ASTM standards, D 3171-99. Standard test methods for constituent content of composite materials; 2004. [20] Annual book of ASTM standards, D 790-02. Standard test methods for flexural properties of unreinforced and reinforced plastics and electrical insulating materials; 2002. [21] Annual book of ASTM standards, D 4065-01. Standard practice for determining and reporting dynamic mechanical properties of plastics; 2002. [22] Shen J, Huang W, Wu L, Hu Y, Ye M. Thermo-physical properties of epoxy nanocomposites reinforced with amino-functionalized multi-walled carbon nanotubes. Composites Part A 2007;38:1331–6. ´ nguez J, Johnston A, [23] Dı´ez-Pascual A, Ashrafi B, Naffakh M, Gonza´ lez-Domı Simard B, Martı´nez M, Go´mez-Fatou M. Influence of carbon nanotubes on the thermal, electrical and mechanical properties of poly(ether ether ketone)/glass fiber laminates. Carbon 2011;49:2817–33. [24] Song YS, Yuon JR. Influence of dispersion states of carbon nanotubes on physical properties of epoxy nanocomposites. Carbon 2005;43:1378–85. [25] Hull D, Clyne TW. An introduction to composite materials. New York: Cambridge University Press; 1996. [26] Liu C, Zhang J, He J, Hu G. Gelation in carbon nanotube/polymer composites. Polymer 2003;44:7529–32. [27] Montazeri A, Khavandi A, Javadpour J, Tcharkhtchi A. Viscoelastic properties of multi-walled carbon nanotube/epoxy composites using two different curing cycles. Mater Des 2010;31:3383–8. [28] Wichmann MHG, Sumfleth J, Gojny FH, Quaresimin M, Fiedler B, Schulte K. Glass-fibre-reinforced composites with enhanced mechanical and electrical properties-benefits and limitations of a nanoparticle modified epoxy. Eng Fract Mech 2006;73:2346–59. [29] Vlasveld DPN, Daud W, Bersee HEN, Picken SJ. Continuous fibre composites with a nanocomposites matrix: improvement of flexural and compressive strength at elevated temperatures. Composites Part A 2007;38:730–8. [30] Fan Z, Hsiao K-T, Advani SG. Experimental investigation of dispersion during flow of multi-walled carbon nanotubepolymer suspension in fibrous porous media. Carbon 2004;42(4):871–6.