Resin and an amine-based hardener were sourced from Huntsman. The TiO₂ nanoparticles, with a diameter of approximately 40 nm, and multi-walled carbon nanotubes, with a diameter of around 38 nm, were synthesized using the sol-gel and chemical vapor deposition (CVD) methods, respectively.

To enhance the surface compatibility of MWCNTs with epoxy matrix, functionalization with 3-Aminopropyltriethoxysilane (APTES) was performed. The procedure consisted of several steps as outlined below:

The functionalization process began with 0.5 grams of CVD-synthesized MWCNTs. The MWCNTs were refluxed in an 80 ml mixture of H₂SO₄and HNO₃ in a 3:1 ratio for 4 hours at 70˚C. After refluxing, the MWCNTs modified with -COOH groups were collected by filtration and washed repeatedly with distilled water until the pH reached 7. The modified MWCNTs underwent two washes with acetone to eliminate any trapped water inside the bundles of MWCNTs. Finally, the functionalized MWCNTs were placed in a vacuum oven at 110˚C overnight to ensure complete drying (Figure 1).

The MWCNTs with -COOH groups were placed in a 100 ml round bottom flask along with 40 ml of thionyl chloride (SOCl₂). The mixture was refluxed for 30 hours under an inert atmosphere of nitrogen gas and filtered. The solid residue, which consisted of chlorinated MWCNTs, was washed several times with anhydrous tetrahydrofuran (THF) to remove any residual reactants and dried at 50˚C overnight, resulting in the modified MWCNTs, which were designated as MWCNT/ COCl (Figure 1).

In the final step of the functionalization process, approximately 0.5 grams of MWCNT/COCl was suspended in dry tetrahydrofuran (THF) along with an excess of 3-Aminopropyltriethoxysilane (APTES). The mixture was refluxed for approximately 90 hours, allowing APTES groups to be introduced onto the surface of MWCNTs. After silylation, the functionalized MWCNTs (designated as MWCNT/ APTES) were washed with ethanol to remove excess amine and then with deionized water to ensure thorough cleaning. The MWCNT/APTES were dried under vacuum before use. The all steps of this procedure are visually illustrated in Figure 1.

Figure 1. Pictorial view of chemical reaction during functionalization of MWCNTS with APTES.

A blend of Multi-Walled Carbon Nanotubes (MWCNT/TiO₂) weighing 0.20 grams and Titanium Dioxide (TiO₂) nanoparticles weighing 0.10 grams was subjected to sonication at 50% amplitude with a 5-second on and 5-second off pulse in 100 ml of acetone. This sonication process was carried out for 4 hours at room temperature. The resulting MWCNT/TiO₂ blend was dried for a period of 10 hour sat 100˚C in a vacuum oven (Figure 2).

Figure 2. TEM images of functionalize MWCNTs (a) and nano particles decorated MWCNTs (a1) and (a2), and fesem image nano particles decorated MWCNTs (b1) and (b2).

The MWCNTs (1.5 wt.%) was thoroughly mixed with epoxy resin and 15% acetone using a glass rod. Ultrasonic waves at 50% amplitude were applied to this mixture, which had a volume of 50 ml, for a duration of 45 minutes. External cooling was employed during this process to prevent an increase in temperature and then vaporized the acetone solvent at 60˚C for a duration of 12 hours.

Subsequently, 10 wt.% of amine-basedhardener was uniformly mixed into the mixture of epoxy resin and MWCNTs, and finally cast into a silicone rubber mold to prepare sample testing specimens. The mold with the epoxy resin mixture was placed in a hot air oven at 70˚C for 10 hours to allow the epoxy resin to cure. The same process was followed for the reinforcement of MWCNT/APTES in epoxy matrix (Figure 3).

Figure 3. A schematic illustration of the preparation of nanocomposite.

The dispersion of nano reinforcements in the matrix and the identification of toughening mechanisms on the tensile fracture surface were investigated using a Field Emission Scanning Electron Microscope (FESEM, Zeiss) (acceleration voltage -15 kV). The surfaces of the samples were coated with gold to enhance imaging quality.

Transmission Electron Microscopy (TEM) was utilized to study the morphology of MWCNT/APTES and MWCNT/TiO₂. A small drop of the epoxy resin containing the fillers (MWCNT/APTES or MWCNT/TiO₂), after the sonication process, was used for the analysis. A copper grid coated with carbon, typically with a mesh size of 200, was chosen as the substrate for mounting the sample. The TEM analysis was conducted using the FEI Technai G2-20-S-Twin microscope.

The tensile testing of nanocomposite samples adhered to the ASTM D-638(V) standard. We used dumbbell-shaped tensile specimens and examined them with a Hounsfield Universal Testing Machine (model H25KS). The testing was carried out at a crosshead speed of 1 mm per minute, conducted under normal ambient conditions. To ascertain the tensile strength and elastic modulus, we analyzed stress-strain curves. Each composition underwent testing with a minimum of five replicate specimens, and we reported the average values along with their respective standard deviations.

Cast specimens of neat epoxy and its nanocomposite were prepared to precise dimensions of 9.2 × 7.5 × 2.5 mm³ through fine emery paper polishing. This was done in accordance with the requirements of single cantilever bending mode dynamic testing. Perkin-Elmer DMA 8000 was used for dynamic mechanical analysis from 35 to 180˚C at 1 Hz as per ASTM D4065.

The molecular structural characteristics of the MWCNT/APTES and MWCNT/ TiO₂ were assessed using Fourier transform infrared (FTIR) spectroscopy. To begin, MWCNT/APTES as well as MWCNT/TiO₂ were blended with high-purity potassium bromide (KBr, Aldrich, 99.9%) and compacted using a 10-ton load in a die to create a film suitable for infrared analysis. The infrared spectra were then recorded under normal ambient conditions using an FTIR spectrometer, specifically the Thermo Nicolet Nexus 1600. Figure 4 displays the FTIR spectra of APTESfunctionalized MWCNTs and TiO₂ nanoparticles decorated MWCNTs.

In FTIR spectra of MWCNT/APTES (Figure 4(a)), the broad band at 3411 cm⁻¹ corresponds to the stretching vibration of –OH groups. The peak at 1640 cm⁻¹ is associated with the presence of carbon-carbon double bond (C=C) network in MWCNTs and also with C=O of amide (broad peak) groups attached to MWCNTs. The absorption at wavenumber 1330 cm⁻¹ and 1073 cm⁻¹ are attributed to the -Si-O- and -N-C- amide groups, respectively. The peaks at 2922 cm⁻¹ and 2855 cm⁻¹ are assigned to the –C-H stretching vibration of ethylene, which is produced at the defect sites of MWCNTs after acid treatment. These peaks indicate acid functionalization and confirm the presence of APTES on the surface of MWCNTs. This FTIR analysis provides valuable information about the functionalization of MWCNTs with APTES and the resulting chemical changes in the material, which is important for understanding their properties and compatibility in the epoxy nanocomposites.

Figure 4. (a) FTIR spectra MWCNT/APTES; (b) FTIR spectra of MWCNT/TiO₂.

In the FTIR spectra of TiO₂ nano particles decorated MWCNTs, the broad peak observed at 3425 cm⁻¹ can be ascribed to the presence of -OH groups in both MWCNTs and TiO₂. The peaks corresponding to the -Ti-O- network in the Rutile phase of TiO2 were observed at 1113 cm⁻¹ and 515 cm⁻¹. here it is noted that, this FTIR spectra reveals the presence of all the characteristic peaks associated with MWCNTs. However, a significant shift in the stretching frequency of the -C=C- framework of MWCNTs is evident, moving from 1630 cm⁻¹ to 1571 cm⁻¹ (as depicted in Figure 4). This shift in frequency is attributed to the bonding of TiO₂ nanoparticles onto the surface of MWCNTs [36]. This bonding results in a form of back bonding, where Ti metal interacts with the anti-bonding molecular orbital of the -C=C- groups in CNTs. This shift in the FTIR spectrum signifies the interaction between CNTs and TiO2, shedding light on the chemical alterations and bonding that occur within the MWCNT/TiO₂ nanocomposite.

Tensile testing was conducted to generate stress-strain curves for both the epoxy and its nanocomposites. These nanocomposites included 1.5 wt.% loading of various MWCNT types (MWCNT/APTES and MWCNT/TiO₂). The outcomes of this testing are depicted in Figure 5. Analysis of these curves allowed for the calculation of the materials’ tensile strength and elastic modulus. The tensile characteristics of the epoxy as well as nanocomposites, such as tensile strength and elastic modulus, are provided in Table 1. This data provides valuable insights into the mechanical behavior and performance of the epoxy nanocomposites, especially in comparison to the pure epoxy. The introduction of 1.5 wt.% of MWCNT/TiO₂ into the epoxy matrix results in a significant enhancement of both tensile strength and elastic modulus. Specifically, these properties increase by approximately 20% and 18%, respectively, in comparison to the pure epoxy.

The enhancement in mechanical properties of any nanocomposite can be attributed to the uniform distribution of nanofiller within the polymer matrix. This even dispersion, coupled with effective interfacial interactions, plays a pivotal role in elevating the material’s performance. The substantial improvement observed in tensile strength and elastic modulus of MWCNT/TiO₂ epoxy nanocomposite is primarily a result of the nearly individual dispersion of MWCNT/TiO₂ throughout the epoxy matrix (Figure 6). This dispersion leads to an increased surface area available for interaction with the epoxy, consequently restraining the mobility of epoxy chains and facilitating the transmission of stress from the epoxy matrix to MWCNT/TiO₂.

The inclusion of 1.5 wt.% of MWCNT/APTES in the epoxy likewise leads to enhancements in both tensile strength and elastic modulus, although the increases are relatively lower, approximately around 11% and 8%, respectively, in comparison to the pure epoxy. It’s worth noting that there is a lack of uniformity in the improvement, as evident from the relatively significant data deviations.

The variation in performance may be due to the presence of clusters of MWCNT/APTES in the epoxy matrix, caused by MWCNT-MWCNT interactions. These clusters act as defects in the MWCNT-epoxy nanocomposite and result in poor interaction between the MWCNTs and the epoxy matrix, leading to a degradation of mechanical performance and increased deviation. These observations highlight the importance of achieving a homogeneous dispersion of MWCNTs in the epoxy matrix for optimal mechanical performance. The presence of clusters or agglomerates of MWCNTs can negatively impact the properties of the nanocomposite, underscoring the need for effective dispersion strategies to harness the full potential of MWCNTs in enhancing material performance.

Figure 5. Stress-strain curves of epoxy, MWCNT/APTES epoxy nanocomposite and MWCNT/TiO₂ epoxy nanocomposites.

Table 1. Tensile strength, elastic modulus, storage and glass transition temperature (Tg) of epoxy, MWCNT/APTES epoxynanocomposite and MWCNT/TiO₂ epoxynanocomposite (1.5 wt.%).

Figure 6 presents FESEM images of tensile fracture samples for both the pure epoxy and its nanocomposites. In Figure 6, the direction of crack propagation is indicated by single-headed red arrows. In both the pure epoxy and its nanocomposites, we observed three distinct fracture zones on the tensile fracture surface throughout the entire process of fracture, from the initial stage to the completion of fracture. These zones are referred to as the mirror zone, the mist zone, and the hackle zone on the tensile fracture surfaces. Within the mirror zone, cracks initiate and progress at a relatively slow pace before rapidly accelerating, leading to exceptionally smooth fracture surfaces. The immediate vicinity of the mirror zone, characterized by slight roughness and a thin region, is termed the mist zone. Surrounding the mist zone is the hackle zone, which exhibits a rough and thicker region (Figure 6).

In Figure 6(a2), we observe the mirror zone of the tensile fracture surface at a higher level of magnification. The fracture surface of the pure epoxy is distinctly visible, featuring exceptionally smooth and river-like patterns. In this region, cracks appear to propagate freely and randomly, indicating the inherent vulnerability of the pure epoxy to crack initiation and propagation. This behavior is characteristic of a typical brittle fracture.

Nevertheless, when 0.1 wt.% of MWCNT/TiO₂ is introduced, the fracture surface undergoes a transformation, becoming rough with the presence of numerous winding and deep cracks within the mirror zone (as depicted in Figure 6(b1) and Figure 6(b2)). Typically, increased roughness signifies the dissipation of more fracture energy. These observations suggest the creation of a new fracture surface due to the deflection of crack fronts.

The FESEM image (Figure 6(c3)) of fracture surface of MWCNT/TiO₂ epoxy nanocomposite at higher magnification in mirror zone shows uniformly dispersion of MWCNTs in epoxy without formation of clusters of MWCNTs. The well dispersion of MWCNTs in the epoxy matrix offers large number of obstacles to the crack propagation and generates more number of crack deflection paths in epoxy matrix and leads to enhancement in the performance of MWCNT/TiO₂ epoxy nanocomposite. The FESEM images of tensile fracture surface of MWCNT/APTES epoxy nanocomposite in mirror zone at higher magnification also confirmed that the dispersion of MWCNTs become worsen and some clusters of MWCNTs are formed in epoxy (Figure 6(c2)) at loading of MWCNTs (0.1 wt.%). The presence of cluster of MWCNTs in epoxy matrix generally acts as defect and diminishes the mechanical and physical properties of polymer nanocomposite. So, it is clearly observed that crack deflection is mainly occurred in case of MWCNT/TiO₂ epoxy nanocomposite compared to MWCNT/APTES epoxy nanocomposite, due to better MWCNTs dispersion in earlier one. The FESEM image of MWCNT/TiO₂ epoxy nanocomposite at high magnification in mirror zone (Figure 6(b2)) shows a range of toughening mechanisms, like crack may propagate through the polymer matrix above or below the poles of the MWCNTs resulting in the more energy consumption due to crack deflection [37]. Crack bridging mechanism, which occurs when a crack propagates and found a MWCNTs right in front of its path resulting the bridging of the crack [38].

Figure 6. FESEM images of (a1) neat epoxy, (a2) neat epoxy (High magnification of mirror zone), (b1) MWCNT/TiO₂ epoxy, (b2) MWCNT/TiO₂ epoxy (High magnification of mirror zone), (c1) MWCNT/APTES epoxy, and (c2) MWCNT/APTES epoxy (High magnification of mirror zone).

The effect of dispersion and loading of different types MWCNTs in epoxy for Storage modulus (an indicator for elastic properties) has been shown in Figure 7. The figure shows that the storage modulus is strongly influenced by the dispersion and type of MWCNTs. The storage modulus of MWCNT/TiO₂ epoxy nanocomposite increases with 1.5 wt.% loading of MWCNT/TiO₂ at below the glass transition temperature (Tg). The reinforcement of MWCNT/TiO₂ leads to 19% enhancement in storage modulus respectively at temperature 35˚C (Table 1) in comparison with neat epoxy. This improvement in storage modulus can be attributed due to cluster free nano level homogeneous dispersion of MWCNT/TiO₂ in entire epoxy matrix (Figure 7), which makes available more sites for interaction between MWCNT/TiO₂ and epoxy matrix. It is also an indicator of improvement in stiffness of MWCNT/TiO₂ epoxy nanocomposite in the range of glassy regions. The uniform distribution of MWCNT/TiO₂ in three-dimension epoxy network may diminish the mobility of epoxy chain network and strongly affect the elastic response below glass transition temperature. But with same loading of MWCNT/ APTES, the storage modulus comparatively not so much improved due to incapability of MWCNT/APTES de-agglomeration (Figure 7). As a result, movement of epoxy chain increase, that resulted in reduction of storage modulus. As temperature rises from glass transition region to rubbery region, the storage modulus slightly decreases and then suddenly declined. As the temperature reaches above 110˚C, the storage modulus is not significantly effect by loading of any types of MWCNTs (Figure 7). In this region the reinforcement of MWCNTs have not enough influence for performance of nanocomposites [39][40]. This may arise due to relatively higher movement of epoxy chain in rubbery region.

Figure 7. Storage modulus of neat epoxy, MWCNT/APTES epoxynanocomposite and MWCNT/TiO₂ epoxynanocomposite (1.5 wt.%).

Figure 8. Tan δ vs. temperature of neat epoxy, MWCNT/APTES epoxynanocomposite and MWCNT/TiO₂ epoxynanocomposite (1.5 wt.%).

The ratio of loss modulus to storage modulus is defined as loss factor or tan δ, which indicate the damping behavior of nanocomposite. The temperature corresponding to peak of tan δ curve was treated as glass transition temperature (T_g) of nanocomposite. The effect of the addition of different types of MWCNTs on glass transition temperature of epoxy nanocomposite materials has been shown in Figure 8 and in Table 1. Enhancement in T_g is most prominent for the MWCNT/ TiO₂ epoxy where T_g become 84˚C in comparison of neat epoxy with T_g 75˚C (Table 1) The superior dispersion of MWCNT/TiO₂ in entire epoxy creates hindered relaxation mobility in epoxy segments at interface region and resulting in rise of T_g. However, in the case of MWCNT/APTES there in no significant change in glass transition temperature that indicates lower crosslinking density and higher chain mobility. Further, Figure 8 shows a gradual decrease in tan δ peak height with the addition of MWCNTs. This indicates the raise in stiffness of epoxy due to good interfacial interaction between well-dispersed MWCNTs and epoxy matrix. The reduction in tan δ peak height is most prominent for the MWCNT/ TiO₂ epoxy nanocomposite.