Materials Science and Nanotechnology

Research Article - Materials Science and Nanotechnology (2017) Volume 1, Issue 2

The effect of 0.5 wt% additions of carbon nanotubes & ceramic nanoparticles on tensile properties of epoxy-matrix composites: a comparative study.

Murat Tiryakioğlu1*, Derya Dispinar2, Muhammet Uludağ3, Şakir Yazman4 and Lokman Gemi5

1School of Engineering University of North Florida Jacksonville, FL 32224, USA

2Engineering Faculty İstanbul University İstanbul, Turkey

3Department of Metallurgical and Materials Engineering Bursa Technical University Bursa, Turkey

4Ilgın Technical Science College Selcuk University Konya, Turkey

5Engineering Faculty Necmettin Erbakan University Konya, Turkey

Corresponding Author:
Murat Tiryakioğlu
School of Engineering
University of North Florida, Jacksonville, FL 32224, USA
Tel: -(904) 620-1393
E-mail: [email protected]

Accepted date: September 26, 2017

Citation: Tiryakioğlu M, Dispinar D, Uludağ M, et al. The effect of 0.5 wt% additions of carbon nanotubes & ceramic nanoparticles on tensile properties of epoxy-matrix composites: a comparative study. Mater Sci Nanotechnol. 2017;1(1):15-22.

Visit for more related articles at Materials Science and Nanotechnology


Carbon nanotube (CNT) and ceramic nanoparticles have different effects on the mechanical properties in epoxy nanocomposites. There are several mechanisms that explain how the differences for mechanical properties in these two different reinforcement are. In this study, multi-walled CNT and three different ceramic nanoparticles (Fe2O3, SiO2 and Al2O3) reinforcements on epoxy were examined. The weight percent of these additions were kept to be 0.5 wt% in all the tests. Tensile tests were carried out in order to characterize the mechanical properties. Statistical analysis was carried out to check the reliability of the test results. Fracture surfaces were examined with scanning electron microscope (SEM). Statistical analyses showed that (i) all additions enhanced tensile strength with the largest increase obtained with SiO2 and Fe2O3 nanoparticles, (ii) elongation was not affected by nanoparticle additions, and (iii) highest reproducibility was obtained in Al2O3 reinforced epoxy.


Nano-structures, Particle-reinforcement, Mechanical properties, Statistical properties/methods.


Nanoparticles are added to polymers as reinforcement particles in to improve mechanical and physical properties. They typically have specific surface areas up to 1000 m2/g [1]. Consequently, load is transferred from the matrix to nanoparticles very effectively, resulting in higher tensile strengths (ST) [2]. Carbon nanotubes (CNT), which have low density, high strength, high thermal and electrical conductivity and high length/diameter ratio, are among the most effective reinforcement particles. The properties of CNTs depend on their diameter, length as well as their directional and surface conditions [3]. The homogenous dispersion of CNTs in matrix plays an important role, particularly for reduced stress concentration [4] under loading.

Recently, as alternatives to CNTs, different nanoparticles such as Al2O3, SiO2 and Fe2O3 have also been studied as reinforcement in the matrix [5-7]. These nanoparticles have almost the same properties as CNTs and have been found to improve tensile strength. However, reports of increase in tensile strength are usually at the expense of ductility. As a result, the fracture toughness of the composite can be also expected to be lower, limiting the use of the composite in certain applications. Ideally, for nanoparticles to give the best reinforcement effect, i.e., simultaneous increase in strength and ductility, the mixing process should produce interparticle spacing that is as homogeneous as possible. If particles are clustered at the end of mixing, their effective size is usually much higher than their individual sizes. Consequently, particle clusters serve as local stress concentrations, finally premature fracture. As a result, strength gained by the amount of force that the reinforcement particles can carry, is accompanied by reduced ductility. Hence nanoparticle reinforcements can potentially enhance strength and ductility simultaneously, provided that the mixing process produces a homogeneous distribution of nanoparticles.

Studies on the effectiveness of the nanoparticle additions as reinforcement particles focused, to a great extent, on the maximization of one mechanical property at the expense of others. Moreover, there is not a study in the literature in which the effectiveness of different nanoparticles was compared at the same addition level. The present study is motivated by this gap in the literature. Tensile properties of neat epoxy and four different reinforcements (CNT, Al2O3, SiO2 and Fe2O3) have been examined at the same addition level. Moreover, the scatter in the results has been characterized statistically to determine which reinforcement addition provides the most consistent change in tensile strength and ductility.


Carbon nanotubes have been among reinforcement particles investigated in the literature. Wang et al. [8] added various CNT concentrations to polylactic acid and found that tensile strength and ductility could be increased simultaneously with CNT additions, regardless of the amount of CNT addition. Despite these promising results, the dispersion of CNTs within the polymer remains to be a challenge. Due to the van der Waals forces and their high surface energy, carbon nanotubes aggregate, leading to stress concentrations which significantly reduce their reinforcing efficiency [9]. Multi-wall CNTs (MWCNT), however, can be functionalized which leads to larger increases in tensile properties, as reported by Li et al. [9]. Starkova et al. [10] studied the effect of small amounts of MWCNT additions to epoxy, up to 1 wt%. Although ST obtained at 1 wt% MWCNT was slightly higher than at 0.5 wt%, ductility was much higher at 0.5 wt%. In a similar study, Ulus et al. [11] observed that the substantial increase in ST at 0.3 wt% MWCNT addition to epoxy was also accompanied by a slight increase in ductility. It is known that while the MWCNT adding displays bridging effect, nanoparticles, such as Al2O3, SiO2 and Fe2O3, display pinning effect.

Alumina (Al2O3) modification was found to improve mechanical properties [5,12]. Hussain et al. [5] produced laminate composite reinforced carbon fiber. They examined epoxy composite reinforced with Al2O3 in nano and micro sizes and reported that mechanical properties were increased. Also, flexural strength, interlaminar shear strength and toughness were increased by Al2O3 reinforcement. Omrani et al. [13] reinforced DGBE-A type epoxy resin composite with nano alumina particles. They found that the presence of very small amounts of Al2O3 played an important role on mechanical properties. Zhao et al. [12] examined mechanisms of increased ductility and toughness in the composite reinforced with nanoparticles. They attributed the increase in mechanical properties with nano-alumina additions which decreased microcracking in tension, plastic void growth, particle pull-out; thus, toughness was increased. Lim et al. [14] prepared epoxy resin nanocomposites including nanoparticles that have platelets and bar shapes. Transmission electron microscopy showed that dispersion of nanoparticles which form platelets in the matrix revealed better properties. Stress modulus, stress strength and fracture toughness were increased. This shows that shape, size and distribution of reinforcements play an important role in determining the mechanical properties.

Similar to alumina, the effect of nano-sized silica (SiO2) particle additions to epoxy have been investigated in several studies. Dmitriev et al. [15] added up to 6 wt.% nano SiO2 as a reinforcement to pure epoxy and determined that it was possible to increase tensile strength (ST) up to 85 MPa, but only by sacrificing significantly from elongation (eF). Similarly, Jumahat et al. [16] investigated the effect of up to 25 wt% nano-SiO2 additions to epoxy on mechanical properties. ST increased with nano-SiO2 particle additions up to 25 wt%. However, ductility was also found to increase slightly along with strength. Yao [7] studied macro and microscopic properties of SiO2-modified epoxy nanocomposites. They determined that an addition of 3 wt% SiO2 to epoxy was sufficient to enhance fracture toughness and deformation resistance of the composite. Similarly, Zhang et al. [17] added SiO2 particles into cycloaliphatic epoxy resins to improve fracture toughness and found that at the optimum value of 3 wt% SiO2 addition. Li [18] investigated the effect modification of epoxy resin matrix with 30-40 nm diameter SiO2 particles at various levels of addition. Results showed that ST increased significantly with modification up to 4 wt% SiO2.

Another nanoparticle addition that has been studied in the literature is Fe2O3. Sun et al. [6] prepared epoxy composites reinforced with Fe2O3 and modified PVP (polivinilpirolidon) and APTES ((3-Aminopropyl)triethoxysilane) with Fe2O3 nanoparticles. They found that strength increased significantly with Fe2O3 addition but at the expense of ductility. They obtained a maximum ST value of 89.1 MPa corresponding to 4 wt% Fe2O3 addition and representing an increase of 50% in ST. Similarly, Pour and Ghaemy [19] investigated the effect of Fe2O3 on mechanical properties of nanocomposites and found that while ST increased with reinforcement content, this was again at the expense of ductility. Maximum ST was found to be 59 MPa, a 20% increase, at a reinforcement content of 8 wt% Fe2O3. In contrast, Zabihi et al. [20] obtained a maximum strength of 66 MPa, a 47% increase, at 10 wt% Fe2O3.

It is hypothesized in this study that the pinning effect, displayed by nanoparticles, has a more pronounced effect on the mechanical properties than the bridging effect, which takes place after CNT additions. This hypothesis is tested in this study by adding the same amount of Al2O3, SiO2, Fe2O3 and MWCNT to epoxy by using a novel mixing technique and measuring the effect of these four reinforcements on tensile strength and elongation (ductility). The results are characterized via scanning electron microscopy (SEM) and statistical analysis.

Materials and Method

The epoxy resin “MGS-L285” was provided from Momentive company (NY, USA). This resin is a lamination resin and it has a viscosity of 600-900 MPa.s, with two phases (diglycidyl ether of bisphenol A 80-90% and aliphatic glycidyl ether 10-20%). Curing was carried out by MGS-H285 (cycloaliphatic amine 70-90% and polyoxyalkylene alkyl amine 10-30%) that was also provided from Momentive Company. MWCNT, Al2O3, SiO2 and Fe2O3 nanoparticles were produced by MKNano Company. The addition of these particles were selected to be 0.5 wt% to determine whether elongation and tensile strength could be enhanced simultaneously at lower addition levels than reported in the literature. The surface area of nanoparticles was calculated by BET (Brunauer–Emmett– Teller) method (Table 1).

Nanoparticles Content (wt.%) Surface area (m2/g) Sizes
Diameter Length
MWCNT 0.50 96.77 5-50 nm 10-30 m
Al2O3 33.38 20-50 nm sphere
SiO2 143.60 15-30 nm sphere
Fe2O3 38.18 20-50 nm sphere

Table 1: The properties of different nanoparticles reinforcement.

For homogeneous distribution of the nanoparticles, solution stirring method was used. The schematic diagram of the process is given in Figure 1. The samples were weighed and then nanoparticles were mixed with acetone (100 g/ml) and stirred for 5 min at 15 min intervals with a probe homogenizer. Epoxy resin was added and stirred for 20 minutes in a cold ice container. The solution was held in a vacuum chamber (-0.75 bar) at 65°C for 24 h. In this way, bubbling of samples was eliminated. The ratio of epoxy and curing agent was 100 to 40 which was mechanically stirred for 10 min. The samples were then held under -0.75 bar vacuum at room temperature for 10 minutes. Epoxy nanocomposites were cured for 24 h at room temperature after poured in the mold and subjected to post curing for 15 h at 80°C. Finally, epoxy nanoparticle mixture was poured into a mold to produced tensile test sample according to ASTM standard (D638-14). For each parameter, five tensile test specimens and epoxy/nanocomposite specimens that produced according to ASTM standards were subjected to tensile testing as per ASTM D638-14, using an Instron 8801 Universal Testing Machine. The fracture surfaces of the neat epoxy and epoxy/ nanocomposites were investigated using scanning electron microscopy (SEM) with a ZEISS EVO/LS10 microscope. All samples surface was coated with gold before SEM examination.

Figure 1: Schematic representation of the sample preparation.

Results and Discussion

Tensile results

The tensile strength and elongation (eF) data are presented in Figure 2. Note that there is almost complete overlap between the tensile strength data of only SiO2 and Fe2O3 specimens, whereas there is no or very minimal overlap between any other datasets. For elongation, there is significant overlap between all datasets despite different levels of scatter. The average and standard deviation of each dataset are given in Table 2. Note that all reinforcement types increase tensile strength with the largest increase achieved by Fe2O3 nanoparticles. The SiO2 nanoparticle addition yielded almost as high an increase in ST whereas the smallest increase occurred after MWCNT additions. These results support the hypothesis tested in this study that the pinning effect displayed by the nanoparticles are more pronounced than the bridging effect of carbon nanotubes. It is significant that all three nanoparticle types yielded higher tensile strengths than MWCNT reinforcements.

Figure 2: The dot plot for tensile strength and elongation for all specimens.

  ST (MPa) eF (%)
  Average S.D. Average S.D.
No Addition 66.35 2.62 4.74 0.77
MWCNT 73.27 1.11 5.30 0.27
Al2O3 78.51 1.52 5.90 0.40
SiO2 83.53 3.81 5.89 0.80
Fe2O3 85.92 2.86 6.31 1.74

Table 2: Mean tensile strength and elongation for different epoxy nanocomposites as well as % changes with respect to neat epoxy.

It is noteworthy that the increase in ST with 0.5 wt% additions is significantly higher in this study than in some studies reported in the literature. For instance, Starkova et al. [10] reported that average tensile strength increased from 63.8 to 65.4 MPa, a 2.5% increase, after the addition of 0.5 wt% MWCNT, which is remarkably lower than the 10.4% increase (66.35 to 73.27, Table 2) in the current study. Similar results were reported for other reinforcement nanoparticles as well. Lim et al. [14] reported no change in ST when 1 wt% Al2O3 was added to neat epoxy whereas Dmitriev et al. [15] found that ST increased only 5 MPa (6.25%) increase with a 6 wt% Al2O3 addition. In the current study, an increase of 12.2 MPa was achieved with 0.5 wt% Al2O3, which corresponds to an improvement of 18.3%. Similarly, the 25.9% increase in ST with SiO2 achieved in the present study is superior to the 12.1% increase reported by Jumahat et al. [16] with a 5 wt% silica addition. Fe2O3 nanoparticle additions also provided superior properties to those reported in the literature; 29.5% increase in average tensile strength is significantly higher than the 6.7% increase reported by Zabihi et al. [20] with 1 wt% Fe2O3 addition.

As stated above, the levels of increase in tensile strength with four nanoparticle additions obtained in this study are superior to those reported in the literature for similar, often higher particle contents. Hence there is strong evidence that the mixing technique has a significant effect on the improvement in tensile strength. The technique described above clearly provided excellent results, which can be speculated to be due to a higher level of homogeneity of the dispersion of particles with the epoxy matrix. Lim et al. [14] had shown how the distribution of particles affected the mechanical properties.

Therefore, in this work, it was found that the establishment of homogeneity of particle distribution had significant contribution the enhancement of tensile properties.

It is also noteworthy that the highest tensile strength levels were obtained with Fe2O3 and SiO2 nanoparticle additions. As listed in Table 1, these two types of nanoparticles have significantly different surface areas and sizes. Moreover, it is affirmed that Fe2O3 particles have better bonding with the matrix, compared to SiO2 and Al2O3. The surface area and size of Fe2O3 and Al2O3 are similar, as listed in Table 1. Therefore, the difference in the extent of strengthening can be attributed to bonding with the matrix. The similarity in tensile strength between Fe2O3 and SiO2 despite different surface areas and sizes is noteworthy. It can be speculated that larger surface area and slightly smaller size of SiO2 nanoparticles compensated for the better bonding with the matrix that Fe2O3 provided.

As shown in Figure 2, there is significant overlap in elongation data between all datasets. Attention is paid only to maximum points, it is clear that the highest elongation was obtained in Fe2O3. However, the minimum elongation values in all datasets are similar. Hence Fe2O3 dataset displays the highest scatter. Whether the scatter and mean of elongation data as well as tensile strength are different between the five datasets will be addressed later in the paper.


The morphologies of fracture surfaces after the tensile tests of neat epoxy (no addition) and epoxy/nanocomposite (MWCNT, Al2O3, SiO2 and Fe2O3) are provided in Figure 3. In Figure 3a, the mirror zone area is quite large; hackle zone takes place in a small place and deformation trails (fracture trails and crack trails) move in the same direction. Hence neat epoxy demonstrates a brittle fracture behavior. For the nanoparticle reinforced samples (Figures 3b-3e), however, the mirror zone is considerably smaller, hackle zone is larger and deformation trails are random and irregularlyoriented. Random orientation of cracks can be explained by transportation of load from weak epoxy matrix to nanoparticles that have much higher strength. With reinforcements, there is an increase in the number and density of fracture branches of the epoxy/ nanocomposites. There is substantial amount of nanoparticles in the extension direction of cracks which can absorb the rupture energy and prevent the propagation of cracks. This shows that nanoparticles in the epoxy resin dispersed quite well and they have a strong interface with epoxy resin, which provides consistently higher levels of strengthening for all nanoparticle additions, than those reported in the literature.

Figure 3: The SEM images of fracture surface of epoxy/nanocomposites (Crack propagation direction is indicated by red color arrows).

The results of the present study coincide with the results of Sun et al. [6] and Zabihi et al. [20]. There is evidence that Al2O3, SiO2 and Fe2O3 nanoparticles cause branching by blocking the crack tip and deflections in the crack directions, which create important toughness mechanisms. While the cracks initiate under load in the matrix, as they propagate, they are blocked by nanoparticles [17,21]. Consequently, blocking by nanoparticles causes atrophies and plastic deformation at the cracks’ tips [17,22,23]. Subsequently, nanoparticles cause deviations in the crack paths, leading to higher strength. It should also be noted that crack initiation depends on epoxy matrix-nanoparticle interface; cracks tend to move through weak epoxy matrixnanoparticle interface and avoid strong interfaces. When the interface is strong, cracks propagate only by creating a secondary crack in the matrix (Figure 3).

Weibull analysis

Based on the “weakest link” theory [24], mechanical properties that involve fracture can be characterized by the Weibull distribution [25-27], the cumulative probability function of which is expressed as:


where, P is the probability of failure at a given stress (or fatigue life) at or lower, σT is the threshold value below which no failure is expected, σ0 is the scale parameter and m is the shape parameter, alternatively known as the Weibull modulus. Note that when σT=0, Equation 1 reduces to the 2-parameter Weibull distribution. The probability density function, f, for the Weibull distribution is expressed as;


The tensile strength and elongation data were statistically analyzed by estimating the Weibull parameters through the linear regression method prescribed by Tiryakioğlu [28] and Hudak [29] that provides unbiased estimates. The probability estimator (plotting position) for unbiased estimates of Weibull parameters is;


where i is the rank of the data point in ascending order, n is sample size and a and b are constants dependent on sample size. For n=5, a and b are 0.173 and 0.500, respectively [28,29]. The estimated parameters for five datasets each for tensile strength and elongation are presented in Table 3, which also lists the coefficient of determination, R2, for each fit. Tiryakioğlu et al. [30] showed that R2 can be used as a test statistic for a goodness-of-fit hypothesis test. The critical R0.052, above which the hypothesis that data come from a Weibull distribution cannot be rejected is found by:


For n=5, R0.052=0.806. Note in Table 3 that all R2 values of nanoparticle reinforced composites exceed 0.806. Only neat epoxy (no addition) R2 values are below 0.806, both for tensile strength and elongation. However, those R2 values are still very close to R0.052 and consequently, distributions for tensile strength and elongation of neat epoxy will be treated as Weibull for practicality and consistency although there is evidence for a positive threshold (3-parameter Weibull) [31].

The probability density functions of the Weibull distributions for tensile strength and elongation for five conditions are plotted in Figure 4, by using Equation 2 and parameter estimates given in Table 3. Note in Figure 4a that the overlap between distributions is minimal, except for Fe2O3 and SiO2, as discussed previously. In Figure 4b, however, there is significant overlap between the elongation distributions. As expected, the distribution for the Fe2O3 reinforcements is the widest whereas the one for MWCNT is the narrowest.

Figure 4: Weibull distributions of (a) tensile strength, and (b) elongation.

ST (MPa) eF (%)
σo (MPa) m R2 σo m R2
No addition 65.36 21.59 0.776 5.54 6.30 0.796
MWCNT 72.84 61.85 0.939 5.19 18.18 0.956
Al2O3 78.85 49.47 0.963 5.84 13.39 0.938
SiO2 85.64 20.67 0.886 6.33 7.13 0.954
Fe2O3 87.22 28.90 0.940 7.06 3.13 0.847

Table 3: Estimated Weibull parameters along with coefficient of determination for tensile strength and elongation data.

A review of Weibull moduli in Table 3 shows that for both tensile strength and elongation, the highest m values were obtained with MWCNT reinforcements. Hence, MWCNT gave the most reliable and reproducible results. The second highest m values were obtained with Al2O3 additions, while obtaining higher tensile strength than MWCNT. Therefore, it is the authors’ opinion that Al2O3 is optimum reinforcement among those investigated in this study.

Statistical comparison of tensile data

To determine whether there is statistically significant difference between mean and standard deviation of all datasets, hypothesis tests were conducted. Because neither tensile strength nor elongation follow normal distribution, nonparametric hypothesis tests, Wilcoxon-Mann-Whitney [32,33] and Levene [34] tests for mean and standard deviation, respectively, were used at a Type I error (α) level of 0.05. The p-values of Wilcoxon-Mann-Whitney tests among the means in tensile strength of all five datasets are presented in Table 4. Note that only Fe2O3-SiO2 comparison exceeds 0.05. Therefore hypotheses that the two mean tensile strengths are equal are rejected, with only the single exception as noted. The p-values for the Levene test for standard deviation are given in Table 5. Because all p-values are above 0.05, none of the hypotheses that standard deviations are equal can be rejected.

  MWCNT Al2O3 SiO2 Fe2O3
No addition 0.0122 0.0122 0.0122 0.0122
MWCNT   0.0122 0.0122 0.0122
Al2O3     0.0122 0.0122
SiO2       0.4034

Table 4: The p-values for the Wilcoxon-Mann-Whitney test for hypotheses that the averages of two tensile strength distributions being compared are equal.
All hypotheses other than that for SiO2 and Fe2O3 can be rejected (p ≤ 0.05), as indicated in bold

  MWCNT Al2O3 SiO2 Fe2O3
No addition 0.4247 0.5707 0.4940 0.8603
MWCNT   0.6258 0.1460 0.1910
Al2O3     0.1969 0.3169
SiO2       0.5340

Table 5: The p-values for the Levene test for hypotheses that the standard deviations of two tensile strength distributions are equal. None of the hypotheses can be rejected.

The results of the hypotheses tests for elongation are presented in Table 6 for mean and Table 7 for standard deviation. All values are above 0.05. Therefore there is not enough evidence to suggest that reinforcement nanoparticles affect either mean or variation in elongation.

  MWCNT Al2O3 SiO2 Fe2O3
No addition 1.0000 0.2963 0.1437 0.6761
MWCNT     0.0947 0.6761
Al2O3     0.5309 0.6761
SiO2       0.6761

Table 6: The p-values for the Wilcoxon-Mann-Whitney test for hypotheses that the averages of two elongation distributions being compared are equal. None of the hypotheses can be rejected.

  MWCNT Al2O3 SiO2 Fe2O3
No addition 0.3457 0.5300 0.9348 0.1678
MWCNT   0.3709 0.2068 0.0800
Al2O3     0.3721 0.0983
SiO2       0.1690

Table 7: The p-values for the Levene test for hypotheses that the standard deviations of two elongation distributions are equal. None of the hypotheses can be rejected.


In all reinforcement trials (MWCNT, Al2O3, SiO2 and Fe2O3), it was found that UTS and elongation were increased at the same time. This indicates that toughness of epoxy resin can be increased by nanoparticle reinforcements.

The increase in tensile strength was achieved, in ascending order, with additions of MWCNT, Al2O3, SiO2 and Fe2O3, respectively. Although the highest Weibull moduli, i.e., the most reliable and reproducible results, were obtained with MWCNT reinforcements, it is the authors’ opinion that Al2O3 is the optimum reinforcement type, based on high Weibull moduli and tensile strength higher than MWCNT specimens.

The crack formation depends on epoxy matrix-nanoparticle interaction. Hence, if this interaction is weak, the cracks move throughout epoxy matrix-nanoparticle interface; otherwise if this interaction is strong, the cracks propagate by creating a secondary crack in the matrix. The interfacial bonding between nanoparticles and the matrix is an important factor improving the mechanical properties of nanocomposites. It is concluded that nanoparticles used in this study were dispersed in matrix perfectly with near-to-perfection wettability. Therefore, uniform stress was distributed along the matrix.

The volumetric distribution of reinforced material is imperative. As the volume gets higher, the strength and strain becomes higher. In addition, surface area, size, shape and density of nanoparticles have an effect on mechanical properties.