Composition-dependent mechanical properties of epoxy composites embedded with helical multiwalled carbon nanotubes

1. Introduction

Composite materials have long been used in engineering because they make it possible to combine different materials in a single system and obtain properties that are not achievable with one material alone. By selecting an appropriate matrix and reinforcement, composites can be designed to show improved strength, thermal resistance, or durability, depending on the application. Among polymer-based composites, epoxy systems are commonly used in aerospace, automotive, construction, and electrical fields. This is mainly due to their good adhesion to different surfaces, resistance to chemical attack, and relatively stable mechanical behavior during service. Previous studies have shown that carbon fiber–epoxy composites containing helical multiwalled carbon nanotubes (HMWCNTs) exhibit improved resistance to impact loading (Abdullah et al., 2025). The improvement has been linked to the ability of the nanotubes to absorb and redistribute energy during deformation. However, it has been demonstrated that increasing the nanotubes above a specific threshold may lead to performance degradation due to agglomeration. In this context, it is of great importance to determine the optimal nanotube content and require careful experimental investigation.

It is well-known that mechanical and thermal properties of epoxy composites are strongly dependent by concentration, type, and dispersion quality of reinforcement (Sarmin et al., 2023; Bello et al., 2020; Ismail et al., 2023; Jamal-Omidi et al., 2019). Relatively high elasticity, low density, and thermal stability make carbon nanotubes as effective reinforcements (Lau et al., 2002; Harris et al., 2004). Polymer matrix incorporated with carbon nanotubes may promote energy dissipation and hinder crack propagation.

A similar tendency has been found in other nanocomposite systems. Studies on polymer-ferrite and polymer-metal oxide composites, particularly hydrogel-based materials, have revealed that rigid nanoscale fillers may drastically affect stiffness and deformation behavior. Although these materials are primarily designed for soft or functional uses, the basic concepts apply to structural composites. In general, reinforcing efficacy is determined by filler amount, dispersion, and interfacial bonding, which collectively regulate load transmission and resistance to deformation in the composite. (Dave et al. 2025, Bamaniya et al. 2025, Dave et al. 2025, Sirach et al. 2025, Mehta et al. 2025, Dave et al. 2025, Sirach et al. 2025, Dave et al. 2025). In this context, the present work investigates a conceptually related yet mechanically distinct system consisting of epoxy reinforced with HMWCNTs, where the unique nanotube morphology and controlled weight fraction are employed to tailor tensile strength, impact resistance, and hardness in a rigid thermosetting matrix.

Nevertheless, as previously mentioned, a key question remains: what is the optimal fraction of HMWCNT to be included in the composite? To address this, researchers have turned to theoretical, computational, and experimental approaches grounded in solid- state physics. Experimental studies on the requirement of composite materials for specific applications form the foundation for corresponding computational and theoretical investigations. The widespread use of machine learning algorithms in material science also relies heavily on experimental data to train these models (Lec et al., 2021, Liu et al., 2022). While it is impractical and uneconomical to experimentally test every possible composition, a set of necessary and sufficient experiments must still be performed to understand the mechanism of deformation, failure, and other key behaviors. The observations obtained from this work can later be useful for developing broader interpretations using theoretical models, numerical simulations, or data-based methods such as machine learning. For this reason, the study was not designed to answer only one specific performance-related question. Rather, it was intended to examine how the mechanical behavior of HMWCNT–epoxy composites changes with composition over a limited but controlled range, so that clear trends could be identified and compared. Some earlier studies have reported that a carbon nanotube content close to 0.5 wt% often provides a reasonable compromise between mechanical improvement and practical issues such as cost, durability, and ease of processing (Mohamed et al., 2017). Based on these findings, 0.5 wt% HMWCNT was chosen as a reference composition in this work. Additional compositions slightly below and above this value were also investigated. This approach was adopted because there is currently no clearly established optimal weight fraction for HMWCNTs, either in general or specifically for epoxy-based composite systems.

The present study focuses on evaluating the mechanical behavior of epoxy composites containing different amounts of HMWCNTs. Tensile strength, impact resistance, and hardness were selected as the main properties of interest. Before fabrication, a preliminary material selection step was carried out to choose the epoxy resin, hardener, HMWCNTs, and mold materials. Since the aim of the work was to study the effect of composition rather than large-scale production, only a limited number of samples were prepared. The HMWCNTs were chosen instead of straight multiwalled or single-walled carbon nanotubes based on previous reports comparing different nanotube geometries (Mingqi et al., 1994; Abdullah et al., 2025). The helical shape is known to promote mechanical interlocking with the epoxy matrix and can enhance interfacial bonding. However, this same geometry increases the tendency for nanotube entanglement, which makes uniform dispersion more difficult. To address this issue, the fabrication process included mechanical stirring and ultrasonic treatment to improve nanotube distribution within the epoxy.

Owing to good balance of strength, workability, and cost, epoxy was selected as the matrix material (Allaoui et al., 2002). As previously reported in the literature, a preparation process based on mixing epoxy resin with a hardener was carried out in this work and applied consistently to all samples (Ali et al., 2021; Uthale et al., 2023; Nayak et al., 2021). A clean epoxy sample devoid of HMWCNTs was prepared and analyzed under the same conditions as a reference point.

2. Materials and Methods

HMWCNTs were obtained from XFNano in powder form. According to the supplier, the material contained roughly 60% HMWCNTs, with individual nanotubes having diameters between 100 and 200 nm and lengths in the range of 1–10 µm. The reported purity was approximately 90%, and the specific surface area was about 30 m²/g. An epoxy system based on Epotec YD 128 resin was used as the matrix material. The resin was mixed with the curing agent HP100205 in a 2:1 ratio. At room temperature (25°C), the resin exhibited a viscosity between 11,000 and 14,000 cPs and a density of 1.16 g/mL, while its flash point was specified as 252°C.

The mold material was chosen on the basis of practical considerations rather than large-scale production requirements. In this regard, silicone rubber, aluminum, steel, and wood were evaluated with respect to cost, durability, and ease of use (Table 1). Silicone rubber was selected owing to simple mold preparation without specialized equipment. Acrylic sheets were cut to match the shapes of tensile, impact, and hardness test specimens. These acrylic templates are referred as generators. The specimens were prepared following ASTM standards: ASTM D638 Type III for tensile testing, ASTM D6110 for impact testing, and ASTM D2240 for hardness measurements. The generators were arranged inside mold casings. Liquid silicone rubber was poured over them. The silicone molds were removed after curing and used for casting composite samples (Fig. 1).

Close

Fig. 1. Acrylic generators placed inside the mold casing for the preparation of HMWCNT/epoxy composite sample molds used in (a) tensile, (b) impact, and (c) hardness tests. Silicone rubber molds prepared for the synthesis of HMWCNT/epoxy composite samples used in (d) tensile, (e) impact, and (f) hardness tests.

Export to PPT

Composite preparation was done using a fixed epoxy mass of 256.289 g. HMWCNTs were added at weight fractions of 0, 0.5, 1.25, 2, and 2.75%, corresponding to samples A through E. The composite synthesis process is illustrated in Fig. 2. To improve initial dispersion, the required amount of HMWCNT was first mixed with ethanol and stirred magnetically at 600 rpm for 30 minutes. Epoxy resin was then introduced into the mixture, followed by manual stirring and subsequent magnetic stirring at 800 rpm and 80°C until the ethanol had fully evaporated. Ultrasonic agitation at 20 kHz and 60°C was applied for 30 minutes to further improve dispersion. Before casting, the mixture was degassed under vacuum to remove trapped air. The hardener was then added, and the mixture was poured into the silicone molds and allowed to cure at room temperature for 48 hours.

Close

Fig. 2. Synthesis procedure of the HMWCNT/epoxy composite, illustrated for the case of casting a tensile test sample.

Export to PPT

Mechanical testing was performed under identical conditions for all compositions. Tensile tests were conducted using a universal testing machine at a crosshead speed of 5 mm/min in accordance with ASTM D638 Type III, with six specimens tested for each composition. Fig. 3(a) shows a few representative tensile specimens made from neat epoxy and HMWCNT/epoxy composites. Fig. 3(b) illustrates the tensile testing of these specimens using the universal testing machine. The dimensions of the tensile specimen, prepared according to the ASTM D638 Type III standard, are provided in Fig. 3(c). Impact resistance was measured using a Zwick Roell impact testing machine following ASTM D6110. A few representative samples prepared for the impact test are shown in Figs. 3(d-f), where Fig. 3(d) shows neat epoxy specimens, Fig. 3(e) shows HMWCNT-reinforced specimens, and Fig. 3(f) provides the detailed dimensions according to ASTM D6110. Hardness measurements were carried out at room temperature using a Shore D durometer in accordance with ASTM D2240. Figs. 3(g-i) presents representative samples prepared for the hardness test. Fig. 3(g) shows specimens made from neat epoxy without HMWCNT, Fig. 3(h) shows specimens with HMWCNT reinforcements, and Fig. 3(i) gives the dimensions of the hardness specimen used in the tests. In addition, fracture surfaces of selected tensile specimens were examined using a Quanta FEG 250 scanning electron microscope to assess dispersion quality, interfacial bonding, and fracture characteristics.

Close

Fig. 3. (a) Tensile specimens made from neat epoxy without HMWCNT, (b) tensile specimens containing HMWCNT reinforcements, and (c) dimensional drawing of the tensile specimen prepared according to ASTM D638 Type III (all dimensions in mm). Impact test specimens and geometry: (d) neat epoxy samples, (e) HMWCNT-reinforced epoxy samples, and (f) specimen dimensions following ASTM D6110 (all dimensions in mm). (g) Hardness test specimens made from neat epoxy without HMWCNT, (h) HMWCNT-reinforced hardness specimens, and (i) dimensions of the hardness test specimen prepared according to ASTM D2240 (all dimensions in mm).

Export to PPT

3. Results and Discussions

Tensile tests were conducted on neat epoxy samples (0 wt% HMWCNT) using four specimens labelled (A.1, A.2, A.3, and A.4). All samples exhibited brittle failure by the end of the test, as shown in Fig. 4(a). The stress-strain behavior was initially linear up to a stress range of approximately 1–1.35 MPa. Following this region, a softening behavior was observed, where the strain increased at a faster rate for a relatively smaller increase in stress. Beyond a strain value of approximately 1.45%, the material showed signs of regaining strength, indicating a mild ‘work hardening’ effect. After this point, the response returned to a linear trend before eventually reaching failure. While the ultimate tensile strength values varied slightly among the four samples, the overall stress-strain behavior was consistent, confirming the reproducibility of the results.

Close

Fig. 4. (a) Stress-strain response of the neat epoxy sample (HMWCNT/epoxy composite with 0 wt% HMWCNT), (b) Stress-strain response of 0.5 wt% HMWCNT/epoxy composite with different epoxy resin-hardener ratios: H.1 (original ratio), H.2 (10% increased hardener), and H.3 (15% increased hardener), (c) Stress-strain response of 0.5 wt% HMWCNT/epoxy composite after allowing sufficient time for ethanol evaporation, showing highly reproducible behavior across five samples (B.1 to B.5). Stress–strain response of HMWCNT/epoxy composites with increasing nanotube content: (d) 1.25 wt% HMWCNT, (e) 2 wt% HMWCNT, and (f) 2.75 wt% HMWCNT. Each curve represents an individual sample tested under identical conditions.

Export to PPT

However, when 0.5 wt% HMWCNT was added to the epoxy, the resulting sample didn’t cure properly. The analysis of the possible causes for the lack of curing pointed to two potential issues: either the epoxy to hardener ratio was not optimal, or the ethanol used during synthesis had not fully evaporated. As a first corrective step, the hardener content was adjusted, and the stress- strain responses for different epoxy-hardener ratios are presented in Fig. 4(b). Specifically, sample H.1 corresponds to the original ratio without any adjustment, H.2 includes a 10% increase in hardener, and H.3 includes a 15% increase. The results clearly show that increasing the hardener content improved the curing of the composite and led to a more brittle failure mode, indicating a more complete crosslinking of the matrix.

Another set of experiments was also conducted to investigate the second possible reason for incomplete curing: insufficient evaporation of ethanol during the composite synthesis. In this set, special care was taken to allow sufficient time for ethanol to fully evaporate before proceeding with the rest of the fabrication steps. Five samples of the 0.5 wt% HMWCNT/epoxy composite were prepared under these modified conditions. The results of these experiments are presented in Fig. 4(c). Notably, the stress–strain responses were highly reproducible across all five samples, confirming the reliability of the modified synthesis process.

Fig. 4(c) shows that the stress-strain response of the 0.5 wt% HMWCNT/epoxy composite follows a similar trend to that of the neat epoxy (0 wt% HMWCNT), but the values at critical points differ. Specifically, the tensile strength of the 0.5 wt% HMWCNT samples ranged between 16 and 19 MPa. These findings are consistent with prior reports that show mis-matched hardener/resin ratios degrade tensile properties when crosslink density is not optimal (Sritharan et al., 2022) and that residual solvent can significantly hinder curing, lower glass transition temperature, and suppress strength (Wazalwar et al., 2021). Post-curing and proper curing agent proportion have been shown to restore mechanical behavior and reproducibility in similar epoxy systems (Bourchak et al., 2013).

Following the same synthesis conditions —particularly allowing sufficient time for ethanol evaporation— additional tensile tests were conducted on composites with higher HMWCNT content: 1.25 wt%, 2 wt%, and 2.75 wt%. Figs. 4(d-f) illustrates tensile stress vs. strain behavior of the epoxy composites with different HMWCNT ratio. Tensile strength gradually rises with nanotube concentration, indicating that the nanotubes are better load-bearing and stress-transfer agents. This tendency is present in almost all compositions. At a maximum loading of 2.75 weight percent, however, the stress-strain response deviates from this pattern, exhibiting distinct nonlinearity with brief softening periods and rapid resistance loss before breakage.

Fig. 5 illustrates the impact strength measurements, highlighting the importance of HMWCNT composition. Even a small amount of nanotubes can greatly enhance impact resistance compared to pure epoxy. The impact strength reaches its peak at 1.25 weight percent HMWCNT but then drops off quickly at 2 weight percent. When we increase the content to 2.75 weight percent, the impact strength does rise again, but it doesn’t surpass the previous peak. This peak at 1.25 weight percent likely represents the best balance between nanotube surface area and effective stress transfer at the nanotube-matrix interface. The drop in strength afterward could be due to partial agglomeration or saturation of the epoxy matrix, but we’ll need more research to confirm that.

Close

Fig. 5. Impact strength of the HMWCNT/epoxy composite as a function of HMWCNT HMWCNT weight fraction.

Export to PPT

Fig. 6 shows how adding nanotubes affects Shore D hardness. As the HMWCNT concentration increases, so does the hardness, indicating a greater resistance to localized deformation. This effect might stem from the stiff structure of the nanotubes, which boosts resistance to indentation and limits the mobility of polymer chains, as noted in earlier studies (Allaoui et al., 2002). Figs. 7(a-d) present SEM analyses of the tensile fracture surfaces, offering deeper insights into the mechanical behavior. At 0.5 weight percent HMWCNT, the fracture surface displays only a few isolated striations, suggesting brittle failure. In contrast, the composite with 1.25 weight percent shows a denser pattern of distinct striations, indicating improved toughness and more effective crack arrest. At 2 weight percent, the striations begin to combine and thicken, indicating additional plastic deformation before rupture. The nonlinear tensile response seen during testing corresponds to the smoother fracture surface and poorly defined striations of the 2.75 weight percent sample. Together, these characteristics demonstrate that an increase in nanotube content regularly causes a change in fracture behavior.

Close

Fig. 6. Shore D hardness of the HMWCNT/epoxy composite as a function of HMWCNT weight fraction.

Export to PPT

Close

Fig. 7. Scanning electron microscopic (SEM) images of the tensile fractured surfaces of HMWCNT/epoxy composites with HMWCNT weight fractions of (a) 0.5%, (b) 1.25%, (c) 2%, and (d) 2.75%.

Export to PPT

Fracture surface morphology and related mechanical trends were used to calculate nanotube dispersion. Nanofillers were integrated extremely consistently, as evidenced by the repeatability of stress-strain curves (Figs. 4c-f) and predictable fluctuations in striation density (Figs. 7a-d). Improved mechanical performance and finer fracture features at lower HMWCNT concentrations point to effective stress transfer and generally uniform dispersion on the microscale. At lower HMWCNT concentrations, enhanced mechanical performance and finer fracture features point to effective stress transfer and relatively homogeneous dispersion at the microscale. At the highest loading (2.75 wt%), smoother fracture regions and localized changes were observed in fracture morphology. This suggests the onset of filler–filler interaction and entanglement, similar to the agglomeration effects reported in CNT/epoxy literature at higher wt%. (Allaoui et al. 2002).

The observed enhancement in mechanical performance with increasing HMWCNT content can be understood in terms of nanoscale reinforcement and matrix–filler interactions, which are well known to tailor structure–property relationships in composite systems. Similar reinforcement mechanisms have been reported in polymer–metal oxide and polymer–ferrite nanocomposites, particularly in cross-linked polymer networks, where rigid nanofillers restrict polymer chain mobility and improve load-bearing capability through effective stress transfer (Dave et al. 2024, Bamaniya et al. 2025, Dave et al. 2024, Sirach et al. 2025). Although such systems are commonly investigated in the context of hydrogels for adsorption or rheological applications, their fundamental reinforcement principles remain applicable to rigid thermosetting matrices. In the present epoxy–HMWCNT composites, the helical morphology of the nanotubes likely promotes mechanical interlocking and enhanced interfacial contact with the epoxy matrix, thereby improving resistance to crack initiation and propagation. This trend can be seen in the gradual rise in tensile strength and hardness as the nanotube content increases, along with noticeable changes in the fracture surfaces observed under SEM. At higher HMWCNT loadings, the stress–strain response becomes less linear, and the fracture morphology shifts, indicating the increasing influence of nanotube–nanotube interactions and localized stress redistribution. Similar behavior has been reported in other polymer nanocomposites when filler content exceeds an optimal range. Taken together, these observations suggest that the mechanical performance of HMWCNT-reinforced epoxy is controlled by a balance between effective reinforcement, uniform dispersion, and strong interfacial bonding, in agreement with trends widely reported for polymer-based nanocomposite systems.

4. Conclusions

In this work, a planned experimental strategy was adopted to develop epoxy-based composites reinforced with HMWCNTs. Before selecting the mold that best suited the available facilities, batch size, and experimental goals, several options were taken into consideration. In order to produce specimens that matched standard mechanical testing criteria, a straightforward yet reliable mold production method was created. This produced consistent sample geometry and a good surface polish. Tensile strength significantly increased as nanotube concentration increased, according to mechanical testing. Shore D hardness and impact strength generally followed the same pattern, with slight drops seen at certain intermediate compositions. The striations and crack channels found by SEM analysis of the fracture surface supported the mechanical behavior that was seen.

Overall, the results provide a solid foundation for future predictive modeling studies and demonstrate the potential of epoxy–CNT composites for structural, protective, and wear-resistant applications.