Manufacturing and characterization of CNT based polymer composites

2.1. Materials

The fiber chosen were woven roving glass mat for fabricating specimens. Polyester, epoxy, phenolic and vinyl ester are mostly used as thermoset resins, the matrix materials used is LY556 (Bisphenol-A-Diglycidyl-Ether) epoxy resin and hardener is HY951 (Tri-ethylene-Tetramine). Multi walled carbon nanotubes are selected as the filler material for the high tensile modulus and strength, low aspect ratio, low density, high flexibility, excellent thermal and electrical properties. These are allotropes of carbon, made out of at least two concentric round and hollow shells of graphene sheets with diameter in the nano meter scale.

2.2. Preparation of composite specimen

Solution technique is used for fabrication of the CNT composites. The calculated amount of epoxy (LY 556) resin is taken into a glass beaker and four different volume percentages such as 1 %, 3 %, 5 % and 7 % of CNT are added with epoxy resin. Six layered of E glass fibers were used to fabricate. The fibers were directed at [0°/90°], [0°/45°] and [0°/135°] orientation. A magnetic stirrer is used for homogeneous dispersion of multi walled carbon nanotubes in epoxy resin for proposed quantity of volume percentage, stirrer at different RPM for 30 minutes at room temperature. After well dispersion of CNT, HY951 hardener is carefully mixed with the epoxy/CNT solution in the ratio of 1:10 for curing of the composite. A rectangular mould of size (300×300×3) mm³ is used for making the CNT composite. For fabricating composites the mould is coated with one layer film of liquid polyvinyl alcohol as an easy releasing agent. A fifteen number of specimens were prepared with CNT based composites by hand lay-up technique. The fabricated specimens are shown in Fig. 1-3. Entrapped air and volatiles are removed by rolling the composite laminate with a smooth steel roller, which also ensured that the resin is uniformly distributed over the glass fiber surface. The process is repeated to get the composite laminate thickness of 3 mm plate. Finally, the composite has been allowed to fully cure at room temperature for 24 hours.

Fig. 1Specimen containing 0 % MWNT’s for [0°/90°] orientation

Fig. 2Specimen containing 1 % MWNT’s for [0°/90°] orientation

Fig. 3Specimen containing 7 % MWNT’s for [0°/90°] orientation

2.3. Thermogravimetric analysis

Thermal degradation properties of the nanocomposites were measured by thermogravimetric analysis (TGA) from room temperature to 850 °C with a heating rate of 10 °C/min under $N_2$ atmosphere. The measurements were conducted using 6-10 mg samples for fifteen composites. Weight retention/temperature curves were recorded.

2.4. Statistical procedure

A non linear regression analysis was done to build up a relationship for thermal degradation, weight loss and thermal decomposition with respect to the experimental data of specimens. In order to calculate error and to compare between the experimental data and the regression based data for each individual specimen graphs are plotted. Examination was performed utilizing Mini Tab software.

3.1. Morphological properties

Scanning electron microscopy is utilized to investigate the dispersion of carbon nanotubes and surface morphology of the composite. The voids and pores may prompt to stress concentration and leads to premature failure of the composite during loading. As the cracks got started, the matrix materials were lost, generally the form of resin particles. The SEM evidence in Fig. 4 and 5 supports this explanation.

Fig. 4SEM image showing matrix cracking and voids effect

Therefore, investigations of the surface topography on composite provide information on the interfacial adhesion that exists between the fiber and matrix material when reinforcing agent is incorporated. Homogeneous distribution of reinforcing agent MWNT’s prompts to improvement of strength and toughness due to strong interaction between MWNT’s and matrix materials. When reinforcing agent percentage increase beyond 5 %, the uniform dispersion will not take place and some aggregates appear, which prompts to lessen the properties of the composites. Agglomeration prompts to weaking of bonding of the fiber and the matrix material. Hence by loading on the composite, stresses drives the composite to failure in Fig. 6. Fig. 7 demonstrates the crack created in specimen, when placed in tensile experiment.

Fig. 5SEM image showing fracture running along fiber-matrix material and flexural tested specimen

Fig. 6SEM image showing agglomeration for 7 % MWNT’s for [0°/90°] specimen

Fig. 7SEM image of the fracture surface of the tensile tested specimen for [0°/90°] specimen

3.2. Thermal analysis of CNT based composites

There are three primary mass loss processes. In the first stage, the slight mass loss below 100 °C could be attributed to the evaporation of absorbed water and 100°-300 °C the weight loss due to the removal of oxygen containing groups on the surface. In second stage, the temperatures from 300°-500 °C can be assigned to the oxidation process. In the last stage, the temperature from 500°-800 °C weight loss due to decomposition of carbon nanotubes as shown in Fig. 11, 12, 13 for [0°/90°], [0°/45°] and [0°/135°] oriented composites respectively.

Fig. 11TGA curves of different CNT percentages of [0°/90°] oriented nanocomposites

Fig. 12TGA curves of different CNT percentages of [0°/45°] oriented nanocomposites

Fig. 13TGA curves of different CNT percentages of [0°/135°] oriented nanocomposites

In the present study for [0°/90°], [0°/45°] and [0°/135°] oriented composites, the carbon nanotubes based composite at 5 % MWCNT is thermally stable up to 337.6 °C, 360 °C, 350 °C respectively and it was the highest among remaining percentages. The carbon nanotubes based composite at 5 % exhibits better outcome in increasing the thermal stability of specimens. The results indicate that adding carbon nanotubes will raise the temperature of maximum thermal degradation. The reason for this higher thermal stability of composites may be due to increase in the degree of cross-linkage as well as in the crystallinity of matrix material.

The dependence of the decomposition temperature on MWCNTs content. As CNT content is increasing, the thermal decomposition temperature is also increased and later decrease for higher loading of CNT percentage, probably due to the ease of compact char formation for the composites during thermal degradation. The material will not starts decompose until that particular temperature. The composite with 5 % MWCNT is comparatively most thermal stable material among all and does not decompose until 566 °C, 552 °C, 554.5 °C for different orientations. On the other hand, high concentration MWCNTs would definitely prevent its fine dispersion and more aggregation or bundles could often be formed because of Vander Waals force among the MWCNTs, thus deterioration the thermal stability of the composites.

The weight loss is less at 5 % CNT concentration for [0°/90°], [0°/45°] and [0°/135°] oriented composites. As CNT concentration is increased the weight loss is reduced up to 5 % CNT concentration, beyond that it increased as shown in Figs. 14, 15 and 16.

Fig. 14Weight loss of different CNT percentages of [0°/90°] oriented nanocomposites

Fig. 15Weight loss of different CNT percentages of [0°/45°] oriented nanocomposites

Fig. 16Weight loss of different CNT percentages of [0°/135°] oriented nanocomposites

3.3. Development of regression model

In statistics, regression analysis is the technique for the modeling and finding analysis of numerical data consisting of values of dependent variables (response variables) and one or more independent variables(predictors). Several prediction techniques have been used to model the thermal properties of composite materials based on different parameters. Regression model is most helpful technique in determining the thermal properties of fiber-reinforced polymer composite. A positive effect is achieved, so we are considering thermal properties data for statistical prediction using regression model. The experimental data was used to generate a mathematical model using regression analysis. By using that equation, an error can be found by placing the experimental values in that equation. By comparing the experimental values and regression values an error can be found for thermal behaviour of the composites

Regression equation is formulated to get the relation between response/dependent variables i.e., thermal properties ($Y$) and the input parameters i.e., volume percentage ($X$).

3.3.1. Thermogaravometric analysis

This function fits a polynomial regression model to the powers of a single predictor by the method. Interpolation and calculation under the curve are also given. The thermal properties are analysed by using thermogaravometric analysis have been formulated in polynomial equation by regression models.

For [0°/90°] oriented specimen:

1

$y_{deg}=-0.392x^{2}4.081x+326.0,$

2

$y_{decom}=-1.434x^2+11.75x+540.9.$

For [0°/45°] oriented specimen:

3

$y_{deg}=-1.647x^2+16.81x+312.7,$

4

$y_{decom}=-1.546x^2+12.53x+527.2.$

For [0°/135°] oriented specimen:

5

$y_{deg}=-2.307x^2+20.87x+298.8,$

6

$y_{decom}=-0.7170x^2+7.963x+528.3.$

Fig. 23Regression with experimental values for thermal degradation temperature of [0°/90°] oriented specimen

Fig. 24Regression with experimental values for thermal degradation temperature of [0°/45°] oriented specimen

Fig. 25Regression with experimental values for thermal degradation temperature of [0°/135°] oriented specimen

Fig. 26Regression with experimental values for thermal decomposition temperature of [0°/90°] oriented specimen

Fig. 27Regression with experimental values for thermal decomposition temperature of [0°/45°] oriented specimen

Fig. 28Regression with experimental values for thermal decomposition temperature of [0°/135°] oriented specimen

The square residual values ($R^2$) of tensile and flexural strength for proposed percentage of multi walled carbon nanotubes were found to vary between 0.97 to 0.98 respectively in the regression model. As $R^2$ of 1.0 represent that the regression curve perfectly fits the data. $R^2$ is a statistic that give information about the fit of curve for collected experimental data. In regression, the $R^2$ coefficient of determination is the measure of how well the regression line approximates the real data points. The Figs. 23 to 28 shows the comparison between the experimental and predicted strength values.