1 Introduction

Understanding the mechanical behavior of composite materials for energy absorbing applications such as vehicle crashworthiness, impact resistance devices, and personal protective equipment has gained significant attention from researchers [1]. Particularly, thermosetting polymers have become a prominent research area due to their unique properties, such as stability under high temperatures, high mechanical strength, viscoelastic properties, and resistant to solvent. However, thermosets have a network of covalently bonded macromolecular structure, with a high crosslink density and little dissipated energy. i.e., plastic deformation in front of the crack tip is localized, which leads to a catastrophic brittle failure [1,2,3,4]. Hence, more efforts have been placed to improve the fracture toughness and alter the failure mechanism in the fracture process zone of thermoset polymers [4,5,6,7,8,9,10].

Research has been focused on nanofillers and the resulting nanocomposite polymers to further improve the fracture toughness of nanocomposite materials as compared with that of the commercially available materials (epoxy resin). It has been established that the incorporation of a second micro-phase of a dispersed rubber (core-shell rubber (CSR) particles) can be considered as a promising solution [4, 11,12,13,14]. These particles comprised of a soft rubbery core within a harder shell and are typically formed by emulsion polymerization, and then dispersed in the epoxy resin. These materials have shown to increase the toughness of both bulk polymers and fiber composites [4, 15,16,17]. However, in general, the literature reports that the fracture toughness of rubber-toughened epoxy polymers is susceptible to temperature variation. On the other hand, to improve the mechanical properties, it is suggested that low or high aspect ratio fillers, such carbon nanotubes (CNT), HNTs, nanoclays, and graphene platelets can be incorporated into the epoxy [18,19,20,21,22]. These fillers possess higher rigidity than the base polymer [18], give high specific surface area [19], and provide sufficient filler-matrix bonding to improve strength and to allow a controlled stress transfer from the matrix to the fillers [23, 24]. HNTs in particular are also known for its energy dissipation and toughening mechanisms such as crack pinning and crack deflection [25].

Fillers, with preferably small dimensions, can reduce local stress concentrations and generate high toughness and impact resistance [20]. The reduction of the filler dimensions in brittle polymer composites, such as epoxy-based composite, is a favorable direction to enhance the toughness since it will focus on improving the microstructural perfection of composites by minimizing the size of potential defects (e.g. inclusions, agglomerates, etc.).

Naturally occurring nanomaterials such as HNTs, kaolinite, and montmorillonite, are used in the fabrication of materials. Halloysite was also successfully employed as filler for macromolecules (such as polymers, proteins, resins) in order to obtain nanocomposite materials for cosmetics [26], packaging [27], and restoration [28] applications. They have commonly been functionalized with organosilane groups to increase the thermal and mechanical characteristics of the composite materials [29,30,31]. The organosilane agents and the hydroxyl groups of nanomaterials can be covalently bonded thus allowing to fine-tune the surface of the clays for enhanced reactivity and prevent leaching [32, 33].

Particularly, halloysite nanotubes are commercially available, inexpensive, and have a tubular structure with siloxy groups on the external surface, and aluminol groups on the inner lumen [34, 35]. These external siloxy groups allow for the facile attachment of organosilanes. HNT with its molecular formula (Al2Si2O5(OH)4·2H2O) has an inner diameter of 1–30 nm and an external diameter of 30–70 nm [36,37,38,39,40]. Its characteristic dimensional structure is made by tetrahedral and octahedral building blocks; it forms a very specific nanoscale lumen with better length to diameter (l/d) ratio and can be well utilized in the preparation of nanocomposites. These distinct features indicate the potential of HNTs as a filler for this study. HNTs are also a logical alternative to CNTs because they are cheaper, naturally found, and eco-friendlier. Once functionalized, many material characteristics can be targeted and be significantly improved [41, 42].

There is a gap in the literature that needs to be filled when it comes to studying the fracture toughness of the material. The epoxy resin which is generally used is limited in advanced applications such as adhesives and fiber-reinforced materials due to its intrinsic brittleness. To improve the toughness of the material many epoxy resins are modified via the addition of different fillers which can be covalently bonded to organic materials, inorganic materials, core-shell rubber (CRS) nanoparticles etc. [12, 20, 43, 44].

This study focused on improving the fracture toughness of the existing SC-15 epoxy resin, a two-part toughened epoxy, by incorporating functionalized HNTs as fillers. The pristine HNTs were functionalized with epoxy and thiol functional groups (Scheme 1). The terminal epoxy and thiol functional groups were explored because of their easy reactivity with other organic molecules or ligands. This characteristic makes them preferred coupling agents [45], while generally retaining the beneficial base properties already present in the nanotubes [46].

Scheme 1
scheme 1

General Scheme showing the pure HNTs reaction with organosilane (GLYMO and MEPES) groups to synthesized modified HNTs, which are incorporated as fillers with SC-15 epoxy in Compact Tension (CT) Testing nanocomposite specimen

Full size image

The effect of HNTs on the chemical composition and thermal properties of the resulting nanocomposites were evaluated by physicochemical characterization. The compact tension (CT) testing was conducted to investigate the influence of pure and functionalized HNTs on the fracture toughness of SC-15 epoxy resin in terms of KIC and GIC. Morphology analysis was also performed used a Field Emission Scanning Electron Microscopy (FESEM) to evaluate the effect of nanomodification on the failure mechanisms of nanocomposites.

2 Materials and methods

2.1 Materials

Halloysite nanotube (HNT), (with 30–70 nm outer diameter; 1–3 μm in length and surface area of 64 m2/g) were purchased from Sigma Aldrich. Glacial acetic acid, HPLC grade toluene, and methanol were purchased by Fisher Scientific. The following organosilane, (3-Glycidyloxypropyl) trimethoxysilane (GLYMO) was bought from Tokyo Chemical Industry Co., Ltd. and (3-mercaptopropyl) triethoxysilane (MEPES) (94%), organosilanes were purchased from Alfa Aesar and were both used as coupling agents. All reactants/reagents were used as purchased with no further purification. The resin used was a two-part toughened epoxy, namely SC-15 was obtained from Applied Poleramic.

2.2 Functionalization of HNT with organosilane group

HNT was pre-dried in a vacuum oven overnight. In a general reaction, 2.0 g of HNT was suspended and stirred in 20 mL toluene. The desired organosilane (GLYMO or MEPES; 3 M equivalent) was added dropwise. The solution was sonicated for 10–15 min with occasional stirring. The mixture was further sonicated for 10 more minutes with intermittent stirring. This slurry was then refluxed for 24 h at 120 °C, the final resultant powder was vacuum filtered and rinsed thoroughly with mixed solvents such as methanol and toluene. The product was then dried overnight in a vacuum oven, and fully characterized via TGA, FTIR, PXRD, SEM, EDS Then, the functionalized nanofillers were incorporated in SC-15 epoxy resin to manufacture the CT specimens, as shown in Scheme 1.

2.3 Manufacturing of nanocomposite

Part A of SC-15 epoxy was mixed with the desired concentration (0%, 0.5%, 1%, and 2%) of functionalized halloysite and the resulting material was mixed using a magnetic stirrer, followed by ultrasonication for around 30 min. Intermittent sonicating energy (10 s energy, 5 s pause) was applied to control the rise in temperature of the compound. The resulting mixture was cooled at room temperature for 30 min, followed by the addition of part B of the SC-15 epoxy. The solution was degassed for 20 min and then poured into pre-prepared silicone molds to manufacture the CT specimens. Curing was completed in a convection oven at 60 °C for 2 h and post-cured at 94 °C for 4 h. The specimens were removed from the mold and milled and polished to the desired thickness, as specified in Fig. 1, as per ASTM D5045 standard. A sharp pre-crack was introduced using a razor blade and care was taken to obtain the same pre-crack lengths for all specimens.

Fig. 1
figure 1

Schematic diagram of the CT specimen (All units are in mm)

Full size image

2.4 Fracture toughness test

To study the effect of inclusion of the halloysite nanotube based nanofillers on the fracture toughness of the SC-15 epoxy, five composites were prepared for each test configuration as per ASTM standard, D 5045, as shown in Fig. 1. The CT testing was performed using Mark 10-M5-300, as shown in Fig. 2 at a speed of 0.02 mm per minute. The deformation was monitored and analyzed in video recording using a digital camera to provide further information about crack initiation and propagation. The software program used by the Mark 10- M5-300, MESUR, recorded the force and the crack opening displacement until the specimen broke. The collected data was then used to calculate the critical stress intensity factor (KIC) and the critical strain energy release rate (GIC) using the equations described below.

Fig. 2
figure 2

Experimental setup for compact tension test

Full size image

For the CT specimen, the critical stress intensity factor (KIC) was calculated using the following formula, as per ASTM D5045:

$${K}_{IC}=\frac{P}{B{W}^{1/2}}.f\left(\frac{a}{w}\right)$$
(1)

where

$$f\left(\frac{a}{w}\right)=\frac{\left(2+\frac{a}{w}\right)\left\{0.886+4.64\left(\frac{a}{w}\right)-13.32{\left(\frac{a}{w}\right)}^2+14.72{\left(\frac{a}{w}\right)}^3-5.6{\left(\frac{a}{w}\right)}^4\right\}}{{\left(1-\frac{a}{w}\right)}^{3/2}}$$
(2)

P is the maximum load in kN, B is the specimen thickness in cm, W is the specimen width in cm, and a is the crack length as determined in cm. For the compact tension specimen, the value of the strain energy release rate GIC in units of kJ/m2 from the corrected energy, U, can be calculated as follows:

$${G}_{IC}=\frac{U}{BW\varphi}$$
(3)

where the energy-calibration factor, φ, shall be computed from

$$\varphi =\frac{\left(1.9118+19.118\left(\frac{a}{w}\right)-2.5122{\left(\frac{a}{w}\right)}^2-23.226{\left(\frac{a}{w}\right)}^3+20.54{\left(\frac{a}{w}\right)}^4\right)\left(1-\left(\frac{a}{w}\right)\right)}{\begin{array}{c}\left(19.118-5.0244\left(\frac{a}{w}\right)-69.678{\left(\frac{a}{w}\right)}^2+82.16{\left(\frac{a}{w}\right)}^3\right)\ \left(1-\left(\frac{a}{w}\right)\right)\\ {}+2\left(1.9118+19.118\left(\frac{a}{w}\right)-2.5122{\left(\frac{a}{w}\right)}^2-23.226{\left(\frac{a}{w}\right)}^3+20.54{\left(\frac{a}{w}\right)}^4\right)\end{array}}$$
(4)

The energy, U, is derived from the integration of the load versus displacement curve up to the same load point as used for KIC.

2.5 Characterization of the modified HNT fillers and the nanocomposite

The pristine and the functionalized HNTs were analyzed by thermogravimetric analysis (TGA, TA Instruments, TGA Q50) under a nitrogen atmosphere from room temperature to 700 °C by 10 °C/min to determine the amount of the siloxy loading and the thermal stability of the HNT. The nitrogen flows for both the balance and sample is 20.0 ml/min. The Fourier-Transform Infrared Radiation Spectrometer (FTIR; Thermo Scientific™ Nicolet™ iS™10) was used to detect the prominent functional groups and molecular structures of the clay (~10 mg) before and after the functionalization of the organosilane groups. The FTIR was carried out at room temperature with 32 scans. The Powder X-ray Diffraction (PXRD; X-Ray Powder Diffractor (Rigaku, XtaLAB Mini) study dealt with the intercalation and the basal spacing of the clay and was performed with a Cu Kα radiation (λ = 0.1541 nm) at 45 kV and 40 mA. Scanning was carried out from 0° to 70°, at a rate of 0.2°/s. The d-spacing of the layered particle was then calculated from Bragg’s equation, nλ = 2d sin θ, where λ is the wavelength of the X-ray, d is the interlayer distance, and θ is the angle of incident X-ray radiation.

The morphological characteristics of nanofillers and nanocomposites were carried out by JEOL JSM-7600F Field Emission Scanning Electron Microscope (FESEM) equipped with Energy Dispersive X-Ray Spectrometer (EDS) and was operated at 30.0 kV with a working distance of ~10 mm. Samples for SEM were pre-coated with palladium/gold sputter coating to improve the conductivity before the test, images were then taken to analyze the nanoclay morphology of each sample and were compared with the resulting as-received (AR)-HNT composite materials. For the nanocomposites, each sample was cut in ultra-thin sections and placed on the carbon-coated copper grids. Representative specimens were analyzed with SEM and EDS to determine the elemental composition and morphology of the fracture surfaces of the materials, respectively.

3 Results and discussion

3.1 Thermogravimetric analysis (TGA)

The primary analysis carried out on each synthesized compound was through TGA and was to predominantly determine the percent loading of the organosilane onto the surface of the clay and to analyze the thermal integrity of each compound. Figure 3 shows the thermogravimetric analysis results of HNT (a), HNT modified with epoxy (b), and thiol (c). In general, the mass loss profiles show very similar and small mass differences below 200 °C (<2.8%), which was allocated to the residual solvents, moisture, and gases absorbed on the clay. The HNT at <200 °C shows the 2.8% loss of HNT, which is higher when compared to the grafted HNTs (epoxy—1.37% and thiol—2.07%); this is due to the utilization of the residual water during the silylation process itself [32, 47, 48]. The mass loss between 200 °C and 400 °C is attributed to the decomposition of the organosilane from the surface of the clay, and the gradual loss of the interlayer water. The mass loss beyond 400 °C is by the dehydroxylation of the aluminol groups of the clay and the remaining covalently bonded organosilane groups [25, 49, 50]. Results showed that HNT loses 1.863% of its weight within a range of 200–400 °C, but the functionalized HNT lost 10.15% for epoxy and 3.25% for thiol in the same temperature range. The major loss is seen in the temperature range from 400 to 700 °C (Table 1). The formation of an additional peak in the range 200 °C to 400 °C for epoxy and thiol group also substantiate for the loss of the organosilane group (Fig. 3b, c).

Fig. 3
figure 3

Thermogravimetric analysis results: (a) HNT, (b) HNT-epoxy, and (c) HNT-thiol

Full size image
Table 1 Mass % loss of pure and modified clays, including the amount of silane reacted (mmol of silane per gram) with HNT surface with epoxy and thiol modifiers
Full size table

The data collected from TGA analysis is summarized in Table 1, which shows mass % losses at different temperature ranges, and calculated mmol siloxy group per gram of clay. Results indicated that the lowest and highest mass % losses within the highest temperature range, 200–700 °C, are exhibited by pure HNT and HNT functionalized by epoxy, respectively.

The quantitative analysis (molar amount) of the grafted organosilane loading was determined using the adapted Eq. (5) [51, 52],

$$.X=\frac{A-B}{100-\left(A-B\right)}\times \frac{1}{MW}$$
(5)

where: X is the molar amount of grafted organosilane per gram of clay; A is the mass % loss of a modified HNT from 200 to 700 °C, B is the mass % loss of a pure HNT clay from 200 to 700 °C; and MW is the molar mass of the hydrolyzed organosilane. (A–B) is the difference in mass loss of the functionalized HNT and the pure HNT. For this calculation, the mass % loss from 200 to 700 °C is taken into consideration since all the organosilane is completely decomposed in this range. The molar mass of hydrolyzed GLYMO is 194.34 g/mol and for MEPES 154.42 g/mol. The mass loss for the functionalized HNT is larger than the mass loss in the unmodified clay because of the decomposition of the organosilane group. The results show the molar amount of silane per gram of HNT is higher in epoxy than that of thiol.

3.2 Field emission scanning electron microscopy (FESEM)

FESEM observations were performed to study the morphology of pristine and modified HNTs (Fig. 4). The SEM images show the tubular shape of the pure HNT was retained ever after the functionalization of the material [53]. The average characteristic size of HNT (external diameter of 30–100 nm, an inner diameter of 6–30 nm, and 60–480 nm in length) [54,55,56] was preserved in both functionalized materials [45, 57]. The epoxy modified HNT was found in the form of aggregates with some globules on the surface of the tubes. The thiol modified HNT also retained its characteristic tubular feature. For pure HNT tubes, the outer diameter was found to be 46.06 nm (Fig. 4a). On the other hand, the HNT-epoxy (Fig. 4b) and HNT-thiol (Fig. 4c) exhibited approximately 95.4 nm and 78.6 nm in outer diameters, respectively, showcasing the loading of the organosilane group on the tubes. The inner diameters of pure HNT, HNT-epoxy, and HNT-thiol were 26.09 nm, 13.6 nm, and 18.85 nm respectively. The increase in the outer diameter and the decrease in the inner lumen both support the grafting of the organosilane groups on the clay [45, 57].

Fig. 4
figure 4

SEM images: (a) pristine HNT; (b) HNT-epoxy; and (c) HNT-thiol with X90000 magnification

Full size image

3.3 Energy dispersive X-ray analysis (EDS)

The SEM-EDS elemental mapping of surface-modified HNT with epoxy and thiol is shown in Fig. 5a, b respectively. The major constituents of HNT are oxygen, aluminum, and silicon. The HNT-epoxy shows the presence of carbon along with the oxygen, silicon, and aluminum [57]. The HNT-thiol illustrates the existence of sulfur confirming the loading of thiol group through covalently bonded organosilane [45, 58]. Small peaks of calcium, chloride, palladium, and gold are present because of the calcium salts (calcium carbonate or chloride) from the clay, and pre-coating with palladium/gold on the material [59]. The sulfur % though TGA analysis and EDS experiments are found to be consistent (~0.1%).

Fig. 5
figure 5

EDS spectra of (a) HNT-epoxy; and (b) HNT-thiol

Full size image

3.4 Fourier-transform infrared spectroscopy

Further characterization of clays was carried out through FTIR to understand the molecular structure. As shown in Fig. 6, in each spectrum, the distinct 3690 cm−1 and 3620 cm−1 bands are very typical for the HNTs from two Al–OH inner surface stretching bands [57]. Intense peaks can be seen at 911 cm−1 contributing towards single Al–OH bending frequencies and 1090 cm−1 for the stretching mode of apical Si–O bond. The strongest band of 1029 cm−1 and 1080 cm−1 is resulted by the stretching vibrations of in-plane Si–O–Si. The 534 cm−1 band is a characteristic peak for the deformation vibration of Al–O–Si group for HNTs [60].

Fig. 6
figure 6

FTIR images of pure (a; blue) and modified clay; epoxy (b; red) and thiol (c; gray). The expanded portion from 2600 cm−1 to 3000 cm−1 is shown in the inset

Full size image

Compared with the pristine HNTs, the epoxy modified HNTs showed new weak bands at 2935 cm−1 and 2670 cm−1, as shown in section A–A of Fig. 6. Similarly, HNT-thiol showed peaks at 2885 cm−1, 2960 cm−1 and 2985 cm−1 all corresponding to the asymmetric and symmetric stretching vibration of aliphatic –CH2– groups.

The presence of some interlayer water is suggested by the observation of the –OH stretching and bending vibrations at 3553 cm−1 and 1640 cm−1 for both thiol and epoxy groups, respectively. The epoxy group is generally seen at 1050–950 cm−1 but due to the huge overlap from the 1029 cm−1 peaks, the single C–O weak band is not visible. The band due to the S-H stretching vibration (2600–2540 cm−1) has usually low intensity [32]. The results confirm the loading of the organosilane groups as well as the retention of the molecular structure of HNTs.

3.5 Powder X-ray diffractometer (PXRD)

The PXRD experiments were carried out with a Cu tube operation, at 45 kV and 40 mA. Scanning was done from 0° to 70°, at a rate of 0.2°/s [25]. The data was collected for all the three clays and used to calculate and compare the primary interspacing distances for each compound. The first order basal reflection (001) for the pure HNT is seen at 7.43 Å, which is very close to the previously reported data (Table 2) [25, 61, 62]. Epoxy, and thiol modified HNTs produced spacing of 7.26 Å and 7.29 Å [63], respectively. For the pure HNT, the 2θ positions of 19.98° and 24.92° confirm the dehydrated state of HNT with (020) and (002) basal reflection, shown in Fig. 7. The (020) peak position trend is what is anticipated from each species, due to the addition of the organosilane decreasing the spacing of each molecule. The basal reflection peak at (001) is reported for functionalized HNTs suggesting that no intercalation of organosilane occurred in the lumen of HNT [45]; it is also such a small sizing difference that it can be attributed to such a low amount of loading for each modified clay that the spacing and orientation of the nanotubes is not adjusted greatly [64]. The PXRD data provides evidence that the original structure of the HNT was maintained even after surface grafting with epoxy and thiol groups.

Table 2 The parameters of the peaks identified in the pure and modified HNT materials
Full size table
Fig. 7
figure 7

XRD patterns of pristine HNT (black), HNT with epoxy (red), and HNT with thiol (blue) groups

Full size image

3.6 Fracture toughness test

Figure 8a shows the experimental setup of the compact tension (CT) testing and Fig. 8b–d shows the load-displacement curves obtained from CT test at room temperature for each material configuration: pure SC-15, HNT, HNT-epoxy, and HNT-thiol nano-modified composites. The data showcase the inclusion of nano-fillers improved the load-carrying capacity and crack opening displacement in the mode-I test. Except at 2% HNT (Fig. 8b), 1% HNT-epoxy (Fig. 8c), and 2% HNT-thiol (Fig. 8d) modified SC-15 epoxy, all specimens showed unstable crack propagation under a continuous loading condition followed by a drastic load drop to zero after reaching the peak load. The load drops implied the existence of a linear elastic and brittle behavior in the materials. As shown in the figure, the improved loading carrying capacity observed in the 2% HNT and 2% HNT-thiol modified composites can be due to the presence of better interfacial bonding between the nanofillers and the host SC-15 epoxy. Inclusion of a higher concentration of halloysite nanofillers within the matrix helped to arrest/delay multiple cracks caused by plastic deformation in the specimen from being propagated rapidly, and as a result, allowed the continuing load rise and fall prior catastrophic failures, as can also be seen in Fig. 9. In the case of HNT-epoxy (Fig. 8c), the maximum load-displacement occurred at 1% weight content, unlike the HNT and HNT-thiol nanocomposites that showed the highest value at 2%. However, the value was reduced by about 17% and 29% as compared with HNT and HNT-thiol nanocomposites, respectively. The reason for having a reduced load-carrying capacity enhancement at relatively medium concertation could be attributed to the brittle nature of coated epoxy applied on the halloysite surface that reduced the plastic deformation of the nanocomposite and load-carrying capacity: the crack opening displacement of 1% HNT-epoxy (Fig. 8c) significantly reduced as compared to the other nanocomposites.

Fig. 8
figure 8

Comparison of load-displacement curves of pure SC-15 resin and synthesized nanocomposites at different concentrations: (a) CT experimental setup, (b) HNT, (c) HNT-epoxy, (d) HNT-thiol

Full size image
Fig. 9
figure 9

Comparison of crack openings of pure SC-15 resin and nanocomposites prior to total fracture

Full size image

Figure 10a, b show the comparison of the critical fracture toughness (KIC) and strain energy release rates (GIC) of pure SC-15 and nanocomposites, respectively. Generally, the addition of pure and functionalized HNT nanofillers in SC-15 changed the KIC and GIC. The inclusion of 2% of HNT improved KIC by 71%. Similarly, functionalized 2% HNT-thiol exhibited the highest KIC enhancement, by 111%, compared to the remaining nanocomposites. The 1% weight concentration of HNT-epoxy nanocomposite showed better KIC performance, 67% enhancement, as compared with the 0.5% and 2% HNT-epoxy specimens.

Fig. 10
figure 10

Effect of various functionalized halloysite fillers on: (a) critical fracture toughness (KIC); (b) on critical strain energy release rate (GIC)

Full size image

Similarly, the GIC is improved as the nanofiller contents increased. As observed in Fig. 10b, the 2% HNT, 1% HNT-epoxy, and 2% HNT-thiol modified composites exhibited the highest GIC within their respective categories by 276%, 182%, and 413%. These improved fracture toughness and strain energy released rates were achieved due to the presence of well-dispersed halloysite nanofillers and multiple-energy absorbing mechanisms in the SC-15 epoxy. The presence of halloysite nanofillers caused the initiation of plastic shear of deformation bands, pseudo ductile deformation, blunted the crack tip, and consequently slowed down the crack growth [65]. However, for the higher content of functionalized HNT-epoxy to SC-15, there was a significant reduction in the resistance to material fracture. This can be explained by the brittle nature of the coated epoxy interface presented between the outer surface of the halloysite nanotubes, the SC-15 matrix, and clusters in the matrix that minimized the plastic deformation zone. The brittle nature of the coated epoxy on the halloysite nanofillers introduced sudden abrupt crack propagation with a reduced plastic deformation that detrimentally affected the fracture toughness and the strain energy release rate. As the TGA analysis revealed, the presence of a relatively higher amount of epoxy on the surface of HNT nanotubes could undermine the effectiveness of HNT nanotubes to enhance the fracture toughness of the resin. Generally, it is believed that the synergy between the commercially available toughened SC-15 and HNT nanofiller is the most promising and one of the few best results available in the open literature, this would allow us to open up a new alternative avenue for improving the fracture toughness for brittle thermoset materials.

3.7 Morphology of fracture surfaces

As reported by the manufacturer, SC-15 is a commercially available toughened epoxy used for several Army composite applications [66]. It is a two-phase toughened epoxy cured with a cycloaliphatic amine. At high magnification (X3000), Figs. 11, 12, 13, 14, 15 show the extent and phase-separated domains in pristine and nano modified SC-15 epoxies. The introduction of toughening particles, mostly elastomers, facilitated the two-distinct phase morphology that can be detected by scanning electron microscopy as cavitation. As also indicated by Wang et al. [66], the toughening phase appeared to be around ∅0.4–1 μm in size, as shown in Fig. 11 (Pure SC15 (X3000)).

Fig. 11
figure 11

SEM observation of pristine SC-15 at increasing magnification from X300–X3000

Full size image
Fig. 12
figure 12

SEM observation of HNT modified SC-15 at increasing filler loading and magnification from X300–X3000. (a) 0.5% HNT modified SC-15 resin. (b) 1% HNT modified SC-15 resin. (c) 2% HNT modified SC-15 resin

Full size image
Fig. 13
figure 13

SEM observation of functionalized HNT-epoxy modified SC-15 at increasing filler loading and magnification from X300–X3000. (a) 0.5% HNT-epoxy modified SC-15 resin. (b) 1% HNT-epoxy modified SC-15. (c) 2% HNT-epoxy modified SC-15

Full size image
Fig. 14
figure 14

SEM observation of functionalized HNT-thiol modified SC-15 at increasing filler loading and magnification from X300–X3000. (a) 0.5% HNT-thiol modified SC-15. (b) 1% HNT-thiol modified SC-15. (c) 2% HNT-thiol modified SC-15

Full size image
Fig. 15
figure 15

The stress whitening in HNT-thiol nanocomposite resulted from multiple crack opening, and voids

Full size image

The morphology and failure mechanisms of mode-I fracture surfaces of nanocomposites filled by pure and functionalized halloysites were examined by SEM. Previous studies [8, 13, 62, 67,68,69,70] showed that enhancement in both fracture energy and fracture toughness exhibited by nano-modified composites are strongly related to a large amount of energy dissipated by the presence of various plastic deformation and damage mechanisms taking place at the micro- and nanoscale level.

Figure 11 shows the SEM images of the fractured surfaces of pure SC-15 matrix tested under CT tests at increasing magnification (X300, X600, X1000, and X3000). As noted in the figure, the fractured surfaces were relatively smooth with the presence of some crazing lines at 45° against the crack propagation, after the crack initiated from its origin. The presence of smooth surfaces indicated that crack propagation occurred drastically, which implied the low pseudo-ductility nature of the material compared to the inclusion of the pure and functionalized HNT modified nanocomposites.

Figures 12, 13, 14, 15 show SEM images of that pure and functionalized HNTs modified SC-15 resins at different loading % of HNT fillers. The presence of HNT fillers in the matrix evidently changed the toughness mechanics from smooth and low pseudo-ductile (relative term) to several energy absorption mechanisms such as fibrils, crack bridging, crazing, microvoids, and hackles. For example, in HNT modified nanocomposite, the higher pseudo-ductile strain (plastic deformation) was observed, as shown in Fig. 12. One of the toughening mechanisms in the configuration of this material was the crack deflection mechanism [71]. In this mechanism, the nanofillers caused the deviation of crack front to a slightly different plane, which created river lines, shear yielding, hackles, and feathers (Fig. 12a–c). The river liners were generated on the new cleavage plane and at slightly different planes. As the crack propagated, these levels combined to form a crack advancing on fewer parallel planes [72].

The size of hackles and feathers tend to increase as the filler content of HNT increased to 2%, which implied high pseudo-ductile strain that improved the fracture toughness and strain release energy rate of the nanocomposite. (see Fig. 10a, b). Besides, as seen in Fig. 12b, c, the presence of the hackle region indicated the presence of the shear yielding and crack-front branching mechanisms in the HNT nanocomposite [68]. The shear yielding mechanics could also occur by the stress concertation around the periphery of the cluster (agglomerated) halloysites [13]. When compared with other types of functionalized HNT nanocomposites, the HNT nanocomposite exhibited cluster debonding. This result implied that the functionalization of HNT nanofillers improved the interfacial bond between the outer surface of HNTs and the matrix.

Among pure and functionalized HNT nanocomposites, the lowest KIC and GIC rates were observed in the HNT-epoxy nanocomposite (see Fig. 10a, b). These findings can be evidently correlated with the SEM results shown in Fig. 13. As shown in the figure, the presence of coated epoxy on the surface of HNT exhibited a predominance of the crack deflection mechanism, with a lower amount of energy absorbed through the zone shielding mechanism, crack pinning mechanism (Fig. 13b), and low plastic deformation. For example, Fig. 13 shows the zone shielding mechanism represented by the formation of micro-voids around the multiple crack tips. Similarly, the crack pining mechanism was observed in the matrix due to the presence of clustered (agglomerated) HNT-epoxy nanofillers. Figure 13c shows that as the concentration of functionalized HNT-epoxy increased to 2%, the brittle nature of epoxy situated between the halloysite tube and the toughened SC-15 matrix detrimentally affected the energy absorbing mechanism; formation and size of crack tails or the amount of plastic deformation significantly reduced, supporting the reduction of the fracture toughness value, when compared with the 1% HNT-epoxy nanocomposite, and is also seen in Fig. 10.

The SEM observation of functionalized HNT-thiol revealed the formation of multiple fibrils and deep tail fracture mechanisms within the matrix, as shown in Fig. 14. The high concentration of functionalized HNT-thiol toughening agent increased the occurrence of high plastic deformation, which resulted from crack bridging with the formation of polymer fibrils known to be associated with multiple localized deformations. The highest plastic deformation of the HNT-thiol modified SC-15 was also confirmed by a visual observation shown in Fig. 9. Generally, the fibrils formation mechanism could be associated with: (i) void generation by debonding between the functionalized HNT-thiol and SC-15 matrix, induces formation craze; and (ii) after the crazing initiated, the voids increase in size and elongate along the direction of the maximum principal tensile stress to form thin fibrils [68]. It is also evident in Fig. 14 that excessive stress-whitening occurred in the functionalized HNT-thiol nanocomposite, indicating the onset failure of the material.

The stress-whitening could also be correlated with a recoverable dilatational deformation just as the deformation was involved in fibril formation of the crazes [73]. As proven in this work (Fig. 15), Merz et al. [74] also postulated that stress whitening was caused by scattered light from the microcracks and that void formation, which resulted from the opening of the cracks, permitted large strain. This helps with the energy absorption mechanism of the material.

4 Conclusions

In this work, the effects of the inclusion of pure and functionalized halloysite nanotubes on the critical stress intensity factor (KIC) and the critical strain energy release rate (GIC) of the SC-15 epoxy were explored. The successful modification of HNTs, with an organosilane containing targeted functional groups (epoxy and thiol) was synthesized. These groups were selected to provide improved interfacial bonding between the HNTs and the SC-15 epoxy resin. The results of this study are summarized as follows:

  • The addition of pure and functionalized HNT significantly improved the load-carrying capacity, the KIC, and the GIC of the SC-15 epoxy. Almost all specimens showed unstable crack propagation under a continuous loading condition, followed by a drastic load drop to zero after reaching the maximum. This implied the presence of linear elasticity and brittle behavior in the material except for 2% HNT, 1% HNT-epoxy, and 2% HNT-thiol modified SC-15 epoxy.

  • The inclusion of 2% of HNT-thiol exhibited the highest KIC improvement, 111%, followed by the 2% pure HNT, 71%, and 1% HNT-epoxy, 67%. Similarly, inclusions of the 2% HNT-thiol, 2% HNT-epoxy, and 1% HNT-epoxy enhanced the GIC, by 276%, 182%, and 413%, respectively.

  • SEM observations revealed that the HNT fillers in the matrix changed the toughness mechanics from smooth and low pseudo-ductile to several energy absorption mechanisms such as fibrils, crack bridging, crazing, microvoids, hackles, and stress whitening. The presence of these energy absorption mechanics resulted from the inclusion of HNTs supported the improvement of the fracture toughness and fracture energy.

Overall, the combined effect between the commercially available toughened SC-15 and HNT nanofillers are the most promising results which can give an alternative pathway for improving the fracture toughness of brittle thermoset materials that could be used for energy-absorbing devices and components in several applications such as vehicles, aeronautic, naval, personal protective equipment, and general construction.