Role of chemical funcionalization of carbon nanoparticles in epoxy matrices

Abstract

The effects of the silanization of multi-walled carbon nanotubes and graphene nanoplatelets with 3-APTES on thermal, mechanical and electrical properties of epoxy nanocomposites were investigated. Nanocomposites containing pristine, oxidized and silanized nanoparticles of multi-walled carbon nanotubes or graphene nanoplatelets at two different concentrations (0.15 and 0.50 vol.%) were prepared by in situ polymerization without using solvents. The functionalized nanoparticles were characterized by Fourier-transform infrared, X-ray photoelectron spectroscopy, Raman spectroscopy and transmission electron microscope techniques. The oxidation and the silanization on the surface of both nanoparticles were confirmed by Fourier-transform infrared, X-ray photoelectron spectroscopy, Raman and transmission electron microscope analysis. The thermal properties of all studied materials were analyzed by differential scanning calorimetry and the mechanical properties by nanoindentation. The addition of both nanoparticles (pristine and functionalized) into the matrix did not show significant variations on thermal properties, but decreased values for glass transition temperature (Tg) compared to the neat resin. Higher values for modulus of elasticity and hardness of epoxy/nanocomposites were obtained when silanized multi-walled carbon nanotubes or oxidized graphene nanoplatelets were added into the matrix. Adding 0.15 vol.% of silanized multi-walled carbon nanotubes the modulus of elasticity increased in approximately 60%, whereas 0.50 vol.% this increase was greater than 90% compared to the neat resin. While adding 0.15 vol.% of oxidized graphene nanoplatelets, the modulus of elasticity increased approximately 83%, whereas 0.50 vol.% this increase was greater than 88% compared to the neat resin. The formation of percolating networks has been achieved only by pristine multi-walled carbon nanotubes addition at a concentration of 0.50 vol.% and by silanized graphene nanoplatelets at a concentration of 0.15 vol.%. However, for both carbon-based nanoparticles conductivities on the order of 10⁻⁷ S/m for frequencies near 100 Hz were observed.

Keywords

Silanization epoxy matrix nanocomposites multi-walled carbon nanotubes graphene nanoplatelets

Introduction

During the early 90s, nanocomposites started to be seen as a new class of materials, skirting the performance of classic composites by accessing new properties and exploiting the synergy between constituent materials. Researchers at Toyota, in 1985 enhanced both thermal and mechanical properties of Nylon-6 by dispersing nanoclay in small concentrations and used it in some parts of their cars.¹ On this basis, polymer nanocomposites have a new perspective of expanding its applications to use reinforcements at nanoscale. The nanoparticles have a high surface area to volume ratio, and hence, greater efficiency would be expected in enhancing the final properties of nanocomposites. Thus, lower concentration of them is used in nanocomposites compared to the traditional composites.

Among the most important thermosetting resins in high-performance industries, such as automotive and aerospace industries, are the epoxy resins, due to their properties. The incorporation of some carbon-based nanoparticles such as multi-walled carbon nanotubes (MWCNT) and graphene nanoplatelets within the epoxy resin would enhance various properties of the resin because of their remarkable properties. These nanoparticles have attracted the attention of researchers seeking to improve the final properties of epoxy/nanocomposites.^2–4 However, after the incorporation of these nanofillers into the polymeric matrix, the dispersion state and interfacial interaction nanofiller/matrix are critical issues in the manufacture of polymer nanocomposites.^3,5

The interfacial interaction between the matrix/nanofillers can be improved with non-covalent or covalent functionalization. Thus, the covalent functionalization can insert functional groups on the surface of nanoparticles. It has the advantage of increasing load transfer efficiency through the covalent bonds between polymer matrix and nanofillers. The chemical surface modification of carbon-based nanofillers is generally carried out by preliminary oxidation processes which can induce the formation of carboxylic and hydroxyl groups on surface of nanoparticles making it possible to connect various organic groups.⁶ However, the acid treatment can insert defects on surface of nanoparticles, often reducing the aspect ratio of MWCNT, which in turn results in possible losses in several properties in the final nanocomposite.⁷

The silanization of carbon-based nanoparticles is an approach to compensate the reduction in the aspect ratio resulting from oxidative treatments. Silane coupling agents most commonly used are organosilanes. The organosilanes have been studied in covalent functionalization to improve the interfacial adhesion of the nanoparticles in epoxy resins as well as the final properties of the nanocomposites. The organosilane 3-aminopropyltriethoxysilane (3-APTES) is an important aminosilane that has wide applications in phenolic and epoxy resins, but there are only a few reports in the literature on the use of graphene nanoplatelets^8,9 and carbon nanotubes^3,7,10 in epoxy matrix. Lavorgna et al.⁷ reported the synergistic effect of silanization with 3-APTES-MWCNT enriched with silica nanoparticles in the thermal properties of epoxy matrix nanocomposites containing 1 wt.% of each nanoparticle. Results showed that incorporation of nanoparticles increased the glass transition temperature (Tg) by approximately 20℃, demonstrating that the functionalization has improved the dispersion state of nanoparticles in the polymer matrix. Wang et al.⁹ evaluated the effect of silanization with 3-APTES in graphene nanoplatelets on thermal properties. The addition of 1 wt.% silanized graphene nanoplatelets showed a greater increase than 7% in thermal stability compared to the pure resin and only oxidized graphene nanoplatelets. Li et al.⁸ evaluated the effect of adding 0.1, 0.2, 0.5 wt.% with two different organosilanes, GLYMO and 3-APTES in oxidized graphene nanoplatelets (GO) in LY -5052 epoxy matrix. Adding 0.2 wt.% of 3-APTES-GO increased 32% for modulus of elasticity (3.3 GPa) and 16% for tensile strength (81.2 MPa) compared to the neat resin.

Among these several approaches widely adopted for the designing of nanocomposites enhanced with silanized MWCNT or graphene nanoplatelets, one of the most interesting in this study is the comparison of nanoparticles with different aspect ratio on final properties of epoxy/nanocomposites and seeking the most suitable functionalization/nanoparticle as enhancing of epoxy/nanocomposites. On this basis, the effect of chemical functionalization (oxidation and silanization) with organosilane 3-APTES in MWCNT and graphene nanoplatelets on thermal, mechanical and electrical properties of epoxy/nanocomposites was examined.

Materials and methods

Materials

Multi-walled carbon nanotubes (MWCNT) used in this work were purchased from Chengdu Organic Chemicals Co. Ltd (TNIM4), had a diameter range of 10–30 nm and purity above 90% with surface area about 140 m²/g. Graphene nanoplatelets from Strem Chemicals Company with an average thickness of the 6–8 nm and wide up to 25 µm with surface area more than 300 m²/g and purity above 99%. Sulfuric acid 98% (Sigma-Aldrich, USA) and 65% nitric acid (Biotec) were used to oxidize the surface of both nanoparticles. The 3-APTES organosilane which was used to functionalize the nanoparticles with purity of 99% were purchased from Sigma-Aldrich, USA. Ethanol with 99.5% of purity (Biotec) was used as the reaction medium for the silanization process. The epoxy resin used was Epoxy Phenol Novolac (LY-5052) containing reactive diluent (butanediol diglycidyl ether) with polyamine-based hardener (HY-5052) both from Huntsman.

Chemical functionalization of nanoparticles

Chemical oxidation was carried out by dispersing 600 mg of both pristine nanoparticles (MWCNT or graphene nanoplatelets) in 60 mL of (3:1 v/v) mixture of concentrated H₂SO₄/HNO₃ with magnetic stirring at room temperature for 18 h. After the acid treatment, the solution was washed with excess of deionized water and filtered until the pH value reached 7. The oxidized nanoparticles (MWCNT_o)/(Graf.o) were dried for 24 h in a vacuum oven at 100℃.

Silanization was achieved by first dispersing 500 mg of both oxidized nanoparticles in 125 mL of ethanol via ultrasonication bath (UNIQUE, USC-2800) for 30 min at room temperature. Then the reaction was conducted with organosilane 3-APTES while stirring at 65–70℃ for 4 h. The silanized nanoparticles (MWCNT_s)/(Graf_s) were filtered and washed with water followed by acetone. The resulting silanized nanoparticles were dried for 20 h in a vacuum oven at 80℃.

Preparation of epoxy/nanocomposites

Epoxy/nanocomposites were produced with pristine, oxidized and silanized MWCNT or graphene nanoplatelets in two concentrations (0.15 and 0.5 vol%) by in situ polymerization without using solvents. The nanoparticles were dispersed in the resin by means of a tip sonicator (Sonics VCX 750 at 180 W) totaling 400 J/g. After sonication, air release agents BYK-A500 and BYK-A560 of 0.3 and 0.5 wt.%, respectively, were added for degassing in vacuum for 30 min. After that, the hardener was added in a ratio of 38:100 by weight. Finally, the mixture was casting in silicone molds and cured at 20℃ for 24 h followed by post-curing at 100℃ for 4 h. The nomenclature used in this work for the nanoparticles and epoxy-based nanocomposites is showed in Table 1.

Table 1.

Nomenclature used for nanoparticles and their epoxy/nanocomposites.

Nanoparticle name	Nanoparticle ID	Epoxy/nanocomposite ID
Nanoparticle name	Nanoparticle ID	0.15 vol.%	0.50 vol.%
Pristine MWCNT	MWCNT_p	MWCNT_p_epoxy
Oxidized MWCNT	MWCNT_o	MWCNT_o_epoxy
Silanized MWCNT	MWCNT_s	MWCNT_s_epoxy
Pristine graphene nanoplatelets	Graf._p	Graf._p_epoxy
Oxidized graphene nanoplatelets	Graf._o	Graf._o_epoxy
Silanized graphene nanoplatelets	Graf._s	Graf._s_epoxy

Characterization

Fourier-transform infrared (FT-IR) was recorded for solid nanoparticles in powder form. The equipment used was a Perkin-Elmer Spectrum One B spectrophotometer in transmission mode, 16 scans were performed in the spectral range 4000–550 cm⁻¹ with resolution of 4 cm⁻¹ using KBr pellets. X-ray photoelectron spectroscopy (XPS) was employed to analyze the surface functionalities and elemental compositions of nanoparticles before and after functionalization using VG ESCA 3000 equipment. The base pressure was operated at 2 × 10⁻¹⁰ mbar. The spectra were collected using Mg Kα radiation with an energy resolution about 0.8 eV and the binding energy of 284.5 eV for the C_1S was taken as reference. Raman spectroscopy was performed for both pristine, oxidized and silanized nanoparticles with a Jobin-Yvon equipment with a helium–neon laser using an excitation wavelength of 633 nm and 0.5 mW in the range of 50–3600 cm⁻¹.

Transmission electron microscopy (TEM, JEOL 2100) was used to characterize the surface morphology of the nanoparticles before and after functionalization and the dispersion states of the nanoparticles into the polymeric matrix, whose accelerating voltage was 200 kV. The nanocomposites with nominal thickness about 80 nm were obtained at room temperature on a RCM Power Tome X ultramicrotome using a diamond knife. Thermal characterizations were performed by differential scanning calorimetry (DSC) in NETZSCH DSC-200 F3 Maia equipment from room temperature to 150℃ with heating ratio of 10℃/min under nitrogen flow. Nanoindentation experiments were carried out for measuring the modulus of elasticity and hardness using an MTS Nanoindenter XP with pyramidal Berkovich indenter. All measurements were an average of nine indentations containing 10 cycles of loading/unloading with hold times of 30 s at maximum load. Loading and unloading times were 5 s and the following loads were applied: 0.78; 1.56; 3.13; 6.25; 12.5; 25.0; 50.0; 100.0; 200.0 and 400.0 mN. Impedance spectroscopy was carried out for measuring the electrical properties of the nanocomposites. The parallel plate capacitance model was used for obtaining the electrical conductivity (σ) and the dielectric constant (ɛ_r) of the nanocomposites, according to equations (1) and (2).

σ = \frac{d}{A} \frac{\cos θ}{| Za |}

(1)

ɛ_{r} = \frac{d}{A} \frac{sen θ}{ω ɛ_{0} | Za |}

(2)

where, Z_a is the measured impedance and A, d are sample dimensions, ω is the angular frequency and є₀ is electric permittivity of vacuum. The impedance of the all samples was measured in the interval between 100 Hz and 10 MHz in an impedance analyzer (Agilent Technologies Model 4294A) at room temperature.

Results and discussion

Characterization of nanoparticles

FT-IR spectra for MWCNT and graphene nanoplatelets samples for each functionalization including pristine samples are reported in Figures 1 and 2, respectively. The band 3750 cm⁻¹ is associated to free hydroxyl,¹¹ whereas in 3440 cm⁻¹ and about 1400 cm⁻¹, bands are attributed to the stretching frequencies (–OH) in (O═C-OH) and (C–OH). The infrared absorptions at 2940, 2850 and 825 cm⁻¹ are associated with symmetric and asymmetric vibration modes of groups (C–H) of the both nanoparticles. The bands in the range of 1635–1570 cm⁻¹ are assigned to (–C = C) conjugated stretching.^7,11 The band at 1200 cm⁻¹ is assigned to vibration (–C–O) in alcohols.¹¹ The presence of these bands including in MWCNT pristine sample in Figure 1(a) and graphene nanoplatelets pristine in Figure 2(a) is resulted from either atmospheric moisture of the raw material or oxidation during purification.^12,13

The FT-IR of (MWCNT_o) and (Graf._o) is shown in Figures 1(b) and Figure 2(b), respectively. After the acid treatment, two new bands were detected, the first at 1730 and the second at 1640 cm⁻¹ corresponding to the (C = O) and (C–O) stretching vibration of the carboxylic groups (COOH) carboxyl groups, respectively, which are attached on the surface of the nanoparticles after oxidation.^11–13 An absorptions modes of vibration (−OH) in 3440 and 1385 cm⁻¹ is due to (OH) bending deformation in (COOH) derived from the oxidative treatment on the surface of both nanoparticles. These observations indicate that the surfaces of the MWCNT and graphene nanoplatelets were oxidized.^11–13 Figures 1(c) and 2(c) represent the spectrum of the reaction product of (MWCNT_o) and (Graf._o) with 3-APTES, respectively. Three new absorption bands in the infrared region were detected at 1050, 1010 and 803 cm⁻¹ are assigned to asymmetric stretching vibration modes of (–Si–OR) and deformation vibration (Si–O–C) and (–Si–OH), respectively. This spectral evidence indicated the possible APTES functionalization on the surface of both oxidized nanoparticles through the reaction of acidic groups with amine groups.^7,11–13

XPS results for survey spectra of MWCNT and graphene nanoplatelets after oxidation and silanization with 3-APTES are reported in Figure 3. In particular, Tables 1 and 2 listed the atomic percentages of the elements which constitute the surface of the samples with the relative binding energy values (BE) for oxidized and silanized MWCNT and graphene nanoplatelets, respectively. According to Figure 3, both survey spectra for oxidized and silanized nanoparticles have two main elements (C and O) and four main elements (C, O, N and Si) present on the surface of the analyzed samples, respectively. However, silanized graphene nanoplatelets detected the presence of sulfur (S) in small concentrations (0.63 at.%) at 169.6 eV. It may be associated to neutralization stage of oxidation (–SO₄⁻² sulfuric acid).

Table 2.

Summary of XPS results with atomic percentages and binding energy for MWCNT.

Samples	C1s	O1s	N1s	Si2p
MWCNT_o
B_E (eV)	284.2	533.1	–	–
at.%	85.75	14.25	–	–
MWCNT_s
B_E (eV)	285	533	401	103
at.%	78.957	16.441	1.958	2.643

Figure 3 also shows the principal band (C1s) for both nanoparticles at 284–285 eV region, characteristic of double bonds between carbons (sp² – C = C) in graphitic structures.^7,9,14 Besides, the main band located at 533 eV (O1s) for all survey spectra (Figure 3) is attributed to the oxygen atoms doubly bonded to carbon atoms (−C = O, carbonyls) to quinones, aldehydes and ketones.^7,15 Additionally, Figure 3(b) and (d) illustrates the main bands N1s and Si2p (402 and 103 eV) for MWCNT_s and silanized graphene nanoplatelets, respectively. The main band at 402 eV is attributed to the presence of nitrogen atoms on the surface of the nanoparticles (amide bonds), whereas the main band at 103 eV is attributed with the presence of silicon atoms through the bonds (–Si–O–C). Thus, the presence of O1s, N1s and Si2p elements demonstrate that both functionalizations (oxidation and silanization) occurred successfully on the surface of the studied nanoparticles.^7,13,16,17 These observations are also in agreement with results obtained by FT-IR, which also showed the presence of the vibrational bands of the carbonyl (C = O) and (–Si–O–C) for both nanoparticles.

It can be seen that Tables 2 and 3 have higher concentrations of the elements on the surface of graphene nanoplatelets compared to MWCNT. This is related to greater purity and/or higher surface area of this material (i.e. 99% and 300 m²/g for graphene nanoplatelets against 90% and 140 m²/g for MWCNT). Regarding the oxygen content, an increase of (2.20 at.%) was observed before and after silanization for MWCNT, while the graphene nanoplatelets for the same element were higher (10 at.%). Besides, the nitrogen content was also higher for graphene nanoplatelets due to amide linkages and silicon (Si–O–C) (3.41 and 5094 at.%) against (1.96 and 2.64 at.%) for MWCNT.

Table 3.

Summary of XPS results with atomic percentages and binding energy for graphene.

Samples	C1s	O1s	N1s	S2p	Si2p
Graf._o
B_E (eV)	284.2	532.5	–	–	–
at.%	86.318	13.682	–	–	–
Graf._s
B_E (eV)	285.4	532.4	402.3	169.6	103.7
at.%	67.389	23.124	3.41	0.634	5.094

The Raman spectra for both nanoparticles before and after the functionalization are shown in Figure 4. The presence of three characteristic bands is clearly observed for both pristine, oxidized and silanized nanoparticles.¹⁸ The band D in range of 1330 cm⁻¹, which is associated with disordered graphitic structures due to breathing modes of sp³ atoms of the ring; G band in 1585 cm⁻¹ associated with ordered structures, assigned to stretch links of all the pairs of atoms with sp² hybridization. The G′ band at approximately 2700 cm⁻¹ may indicate the number of graphene nanoplatelets stacked (Figure 4(b)). An alternative to determine the number of layers for graphene nanoplatelets may be performed by the ratio between the intensities of bands G′(2D) and G (I_2D/I_G). The ratio I_2D/I_G ∼ 2–3 indicates the presence of a single sheet of graphene, while 1 < I_2D/I_G < 2 two graphene nanoplatelets and values smaller than unity indicates the presence of multilayers.¹⁹ The results for the ratios (I_D/I_G) for pristine, oxidized and silanized MWCNT as well as the ratios (I_D/I_G) and (I_2D/I_G) for pristine, oxidized and silanized graphene nanoplatelets are showed in Tables 4 and 5, respectively.

Table 4.

Intensity ratio of the D and G bands (I_D/I_G) for MWCNT_p, MWCNT_o and MWCNT_s.

Samples	Band D (a.u)	Band G (a.u)	Band G′(a.u)	I_D/I_G
MWCNT_p	17.410	15.200	9.375	1.15
MWCNT_o	15.800	11.350	5.140	1.40
MWCNT_s	15.820	12.840	7.043	1.23

Table 5.

Intensity ratio of the D and G bands (I_D/I_G) and (I_2D/I_G) for pristine, oxidized and silanized graphene nanoplatelets.

Samples	Band D (a.u)	Band G (a.u)	Band G′(a.u)	I_D/I_G	I_2D/IG
Graf._p	3.860	8.189	3.150	0.47	0.38
Graf._o	4.245	8.525	3.173	0.50	0.37
Graf._s	2.772	6.552	3.342	0.42	0.51

The ratio I_D/I_G (Table 4) slightly increased after acid treatment. The increase is an indication that there was an increase in the number of defects sites created on the surface of MWCNT_p. These defects are due to the insertion of (–COOH) groups after the strong oxidative treatment. Besides, after silanization with 3-APTES, there was a decrease in the ratio I_D/I_G compared with oxidized carbon nanotubes. This behavior was also reported by Velasco-Santos et al.⁴ arguing that after the process of silanization, there is no further significant change in the crystal structure of these nanoparticles. Whereas for graphene nanoplatelets, there was a slightly increase in the ratio (I_D/I_G) for oxidized graphene nanoplatelets in relation to pristine graphene nanoplatelets, due to the insertion of (–COOH) groups after the strong oxidative treatment (Table 5). After the silanization process, there was observed a decrease in the ratio (I_D/I_G) compared with oxidized graphene nanoplatelets. The phenomenon was similar to the one observed for MWCNT_s. Finally, in Table 5, the ratio (I_2D/I_G) for all samples analyzed is smaller than unity. On this basis, and according to the results reported by Nguyen et al.,¹⁹ all samples of this study seem to have multilayered of graphene stacked. These observations found in Raman scattering analysis are consistent with the images obtained by transmission electron microscopy (TEM) for pristine, oxidized and silanized MWCNT and graphene nanoplatelets showed in Figure 5.

Characterization of epoxy/nanocomposites

The values for glass transition temperature (Tg) of the nanocomposites produced were obtained by DSC during the second heating. The results of this analysis for nanocomposites reinforced with pristine, oxidized and silanized MWCNT or graphene nanoplatelets are shown in Table 6 (measurement precision of ± 2℃).

Table 6.

Glass transition temperature (Tg) values of neat epoxy, epoxy/MWCNT and epoxy/graphene nanoplatelets nanocomposites.

Glass transition temperature (Tg)/℃
Volume fraction (vol%)	0.0	0.15	0.50
Neat epoxy	130	–	–
MWCNT_p_epoxy	–	114	124
MWCNT_o_epoxy	–	130	127
MWCNT_s_epoxy	–	126	125
Graf._p_epoxy	–	119	120
Graf._o_epoxy	–	113	130
Graf._s_epoxy	–	124	130

The addition of pristine and functionalized nanoparticles possibly increased the mobility of the polymer chains of the matrix resulting in decreasing of glass transition temperature (Tg) of all the samples compared to neat resin. As it can be seen in Table 6 even after adding pristine carbon nanotubes (MWCNT_p) or pristine graphene nanoplatelets in both the concentrations, a decrease was observed in the Tg values in relation to neat resin. When oxidized/silanized MWCNT or graphene nanoplatelets were added into the resin no changes in the Tg values for volume fraction of 0.15 and 0.50 vol.%, respectively, was observed for both functionalizations compared to neat resin. But a slightly decrease was observed in Tg values for the volume fraction of 0.15 and 0.5 vol.% when oxidized/silanized graphene nanoplatelets or MWCNT, respectively, were added into the matrix compared to the neat resin. The functionalization on the surface of both nanoparticles did not restrict the movement of the polymer chains compared to pristine nanoparticles. The decrease of Tg values of the epoxy/nanocomposites may be associated with network disruption by decreasing both cross-link density (increasing free volume and decreasing glass transition temperature) and/or the interphase region.^8,20 The addition of carbon-based nanoparticles possibly influence the stoichiometric ratio resin/hardener reaction, vicinal nanoparticles, acting as steric limitations near interfaces or changing the formation of the crosslinked network.²⁰ Thus, either a thin layer of soft interphase may have been formed or there is a possibility of increasing the punctual plastic deformation of the produced nanocomposites.⁸

The results for mechanical properties (modulus of elasticity and hardness) obtained by nanoindentation for both pristine and functionalized nanoparticles are shown in Table 7. The effect of the increment of nanoparticle content into the polymeric matrix is shown in Figure 6. For modulus of elasticity of the MWCNT, it is clear that by increasing the volume fraction of MWCNT, the average value of this property increased independent of the functionalization of its surface in relation to neat resin. When MWCNT_p and MWCNT_o were added into the polymeric matrix, there was a slight increase in the modulus of elasticity values for both volume fractions compared to the pure resin; however, after adding silanized MWCNT, the largest increase for this property has been noticed. The addition of 0.15% v/v of silanized MWCNT increased approximately the modulus of elasticity by 60%, whereas for 0.50% v/v of silanized MWCNT, this increase was greater than 90% compared to neat resin (Table 7). For hardness, it can be observed that increasing volume fraction of MWCNT, the numerical value of hardness is incremented as described above for modulus of elasticity for all nanocomposites containing MWCNT as can be seen in Figure 6.

Table 7.

Summary of mechanical properties of neat epoxy, MWCNT/epoxy and graphene nanoplatelets/epoxy nanocomposites.

	Young’s Modulus (GPa)			Hardness (GPa)
Volume fraction (vol%)	0.0	0.15	0.50	0.0	0.15	0.50
Neat epoxy	2.20 ± 0.24	–	–	0.27 ± 0.01	–	–
MWCNT_p_epoxy	–	3.12 ± 0.17	3.31 ± 0.22	–	0.25 ± 0.02	0.26 ± 0.03
MWCNT_o_epoxy	–	2.87 ± 0.25	2.94 ± 0.18	–	0.26 ± 0.01	0.29 ± 0.01
MWCNT_s_epoxy	–	3.50 ± 0.10	4.19 ± 0.23	–	0.30 ± 0.01	0.39 ± 0.04
Graf._p_epoxy	–	1.98 ± 0.33	0.59 ± 0.34	–	0.21 ± 0.03	0.01 ± 0.01
Graf._o_epoxy	–	4.03 ± 0.06	4.14 ± 0.07	–	0.39 ± 0.01	0.35 ± 0.02
Graf._s_epoxy	–	3.70 ± 0.30	3.70 ± 0.10	–	0.33 ± 0.04	0.32 ± 0.01

Instead of increasing the elastic modulus presented before by increasing the volumetric fraction of MWCNT in the polymer matrix, this phenomenon was not observed for nanocomposites reinforced with pristine graphene nanoplatelets (Figure 6). In Table 7, it can be seen a decrease of 9 to 72% for this property to concentrations of 0.15 and 0.50 vol.%, respectively, for the pristine samples analyzed, although it is important to note that higher values were achieved for both functionalized nanoparticles (Graf._o and Graf._s) increasing the filler content compared to the pure resin. The greatest increases were observed with the addition of oxidized graphene nanoplatelets as shown in Figure 6. After the addition of 0.15 vol.% of these nanofillers, the modulus of elasticity increased approximately 83%, while at 0.50 vol.% this increase was greater than 88% compared to the pure resin (Table 7). The hardness values of graphene nanoplatelets can be seen from Figure 6. Thus, with increasing the volume fraction of these nanoparticles, the numerical value was reduced for pristine and oxidized samples. Different phenomenon as previously described for Young's Modulus of MWCNT nanocomposites. After the addition of silanized graphene nanoplatelets, there was no change of the numerical value of hardness with increasing volume fraction of the nanofiller in the polymeric matrix as can be seen in Figure 6 and Table 7.

One of the many reasons to add rigid nanoparticles into epoxy matrix is to improve the mechanical properties. For example, the addition of nanoparticles with high modulus of elasticity (MWCNT and graphene nanoplatelets) can increase modulus and strength of the nanocomposite.²¹ According to Ajayan et al.,²¹ there are three mechanisms that load is transferred from polymer matrices to the filler with addition of nanoparticles. According to the authors, the first mechanism of reinforcing effect is due to the transfer of shear stresses in nanoparticle/matrix interface. However, the load transfer is dependent on the interfacial shear stress existing between the matrix and nanoparticles. A high interfacial shear stress will transfer the applied load to nanoparticles over a short distance, and a low interfacial shear stress will require a long distance. Thus, the higher aspect ratio of nanoparticles, nanocomposite experiences greater modulus due to increased transfer of stress along the dimensions of nanoparticles. The second possible mechanism that load is transferred from polymer matrix to the filler is the formation of a three-dimensional is the formation of a three-dimensional nanostructured network (interphase), which acts by decreasing the mobility of the polymer chains, causes change in glass transition and the modulus of nanocomposite. Finally, the third mechanism that loads transferred from matrix to nanoparticles (based on the second mechanism), is called localized yielding near the particles, may also decrease the modulus at very low and seemingly macroscopic elastic strains.²¹

In an attempt to understand these reinforcement mechanisms of nanocomposites with modified carbon nanoparticles, Li et al.⁸ assumed that there are two factors that compete with each other at the interface matrix/nanoparticle. The first factor is that the formation of the covalent bond can increase crosslinking and facilitate the transfer of stresses of the terminal amino (–NH₂) of the nanoparticles with the matrix components (resin/hardener), which consequently increase the mechanical properties (modulus of elasticity and strength) of the nanocomposite. While the second factor is the formation of a thin flexible interphase layer (soft interphase) that aids the mobility of nanoparticles increasing the ductility of the nanocomposite. Thus, the final properties of the nanocomposite reinforced with silanized (3-APTES) nanoparticles depend on the competition of these two factors.

According to some authors,^21,22 the increase in mechanical properties occurs enhancing the interaction between the nanoparticle with polymer matrix. Besides, according to Ajayan et al.,²¹ when there is a homogeneous distribution of the nanofillers in the polymer matrix, they can improve the modulus and strength of the nanocomposites, because the small sizes of them do not create large stress concentrations. On this basis, the greatest increases in modulus of elasticity and hardness with increasing volume fraction of 0.15 to 0.50 vol.% were observed when MWCNT_s, Graf._o and Graf._s were added . This result corroborates with the images obtained by TEM shown in Figures 7(e) and(f) and 8(c) to(f). However, when MWCNT_p was added, there was a small improvement on the mechanical property probably because of the small agglomerates (see black circles in Figure 7(a) to (d)). These agglomerates can act as stress concentrators, initiating the propagation of internal cracks in these materials and load transfer is impaired. For MWCNT_s or Graf._s, the formation of the covalent bond may have been occurred from the terminal amino (−NH2) of the nanoparticles with the matrix components (resin/hardener) even with the reduction of aspect ratio as can be seen in Figure 7(e) and (f). This can increase crosslinking and facilitate the stress transfer by interfacial shear stress, which consequently improve the mechanical properties of the nanocomposite.⁸ Besides, for oxidized graphene nanoplatelets, improvement in ductility was observed by decreasing the values for the hardness of approximately 10% (Table 7). This behavior can be related with a formation of a thin layer of soft interphase that aided the mobility of nanoparticles, thus higher values for punctual plastic deformation of the nanocomposites were observed with increasing volume fraction.⁸ When pristine graphene nanoplatelets were added, there was a decrease in the mechanical properties (modulus of elasticity and hardness) increasing the volume fraction of 0.15% and 0.50 vol.%. This was probably due to the presence of agglomerates which reduced interfacial area for tension transfer. In Figure 8(a) and (b), it is possible to see these agglomerates (black arrows). Besides, a soft or flexible interphase may have been formed between the nanoparticle-polymer chains due weak interfacial interactions.⁸

The impedance spectroscopy technique was used to evaluate the effect of dispersing pristine, oxidized and silanized multi-walled nanotubes (MWCNT) and graphene nanoplatelets in the electrical conductivity in the epoxy polymer matrix. Figure 9(a) to (d) shows the values of conductivity and dielectric constant versus frequency for nanocomposites reinforced with MWCNT or graphene nanoplatelets, respectively. The values obtained for nanocomposites were compared with those of the neat resin. According to Figure 9(a), all the samples showed an increase in conductivity as the frequency incremented except for those reinforced with pristine multi-walled carbon nanotubes 0.50 vol.%. However, for the nanocomposite with MWCNT 0.15 vol.%, there was only a linear increase of the initial conductivity compared to neat resin with increasing frequency that is related only to the process of interfacial polarization/hopping conduction. These phenomena of interfacial polarization (charged particles that move from one point to another within the dielectric at frequencies up to 10³ Hz) and hopping conduction produced by electrically charged particles jumps between localized states.²³ Thus, this volume fraction was not sufficient to create a percolating network and increase the conductivity of the polymer matrix. The percolating network at low frequencies (10⁻² Hz) was observed only for nanocomposites reinforced with pristine MWCNT 0.50 vol.%. It is assumed that this volume fraction is located above the percolation threshold. The electrical conductivity was observed in the order of 10⁻⁷ S/m. Sene et al.²⁴ also found an increase in electrical conductivity at low frequencies (percolation network) with pristine MWCNT 0.25 wt.% (∼0.15 vol.%) reaching values for electrical conductivity in the order of 10⁻⁷ S/m. This electrical conduction happens by the displacement of the resonant π electrons of benzene rings through the concentric tubes. Since the frequencies were above 10² Hz, the increase in conductivity is related to the interfacial polarization/hopping process. This higher concentration of conductive nanoparticles in insulating polymer matrix increases the number of available paths (probabilities) for electrons creating a percolation network.²³

Additionally, Figure 9(b) illustrates the values for the dielectric constant for all samples analyzed. According to Giacometti and Carvalho,²³ the dielectric constant as a function of frequency may indicate the efficiency of interfacial polarization process. Thus, larger values for relative constant indicate a higher dielectric dispersion effect. This behavior can be correlated with values increase for electrical conductivity as a function of frequency observed for the samples in Figure 9(a), which showed relative values of approximately 12 for neat resin up to 300 for the sample with pristine multi-walled carbon nanotubes 0.50 vol.%.

According to Figure 9(c), all samples containing graphene nanoplatelets showed an increase in conductivity as the frequency is incremented (process interfacial polarization/hopping) except for nanocomposites reinforced with silanized graphene nanoplatelets 0.15 vol.%. Additionally, percolation networks were observed only for the samples reinforced with 0.15 vol.% of silanized graphene nanoplatelets . Different from multi-walled carbon nanotubes, the conduction for graphene nanoplatelets at low frequencies (10² Hz) up to 10⁴Hz showed a constant behavior (resistor behavior) with electric current passage. However, for frequencies higher than 10⁴ Hz, the conductivity increases as the function of frequency as observed for MWCNT by interfacial polarization/hopping conduction.²³ Figure 9(d) shows the values for the dielectric constant for graphene nanoplatelets. Thus, larger values for dielectric constant indicate a higher dielectric dispersion as mentioned before for MWCNT nanocomposites. According to Figure 9(d) relative values for dielectric constant of approximately 11 for neat resin to 19.5 for the sample containing 0.15 vol. % of silanized graphene nanoplatelets was observed. It was lower compared to those ones observed for MWCNT nanocomposites.

Figure 10(a) to (d) illustrates the effect of electrical conductivity for two constant frequencies (40 Hz and 10³) as the concentration of MWCNT or graphene nanoplatelets incremented, respectively. In Figure 10(a) and (b), a slight increase can be observed in electrical conductivity as the volume fraction of nanoparticles increased. Higher increases were achieved for nanoparticles without functionalization, except for the nanocomposite reinforced with silanized MWCNT, which showed a slight decrease in function of the volume fraction for both frequencies as illustrated in Figure 10(a) and (b). Probably because the reduction of aspect ratio of multi-walled carbon nanotubes after the oxidation process and the homogeneous distribution in the polymer matrix observed in TEM images (Figure 7(e) and (f)). The functionalizations may have influenced the formation of a percolation network requiring higher concentrations to achieve the percolation threshold. According to Sreeprasad and Berry,²⁵ the functionalization on the surface of allotropic nanoparticles (i.e. carbon as nanotubes and graphene nanoplatelets) changes the hybridization sp² on the tetrahedral structure with sp³ hybridization disrupting the π electrons of the MWCNT/graphene nanoplatelets impairing the electrical conductivity.

In Figure 10(c) and (d) for pristine graphene nanoplatelets, a decrease was observed in the value of electrical conductivity for both frequencies. This is due to the greater surface area, the nanoplatelets tend to form aggregates in non-homogeneous dispersion state on polymer matrix independent of the concentration impairing the electrical conductivity as shown in TEM images (Figure 8(a) and (b)). According to Potts et al.,²⁶ a slight aggregation of the nanoparticles (homogeneous dispersion state) on the polymer matrix may increase the electrical conductivity of the nanocomposites. A slight increase in the value of the conductivity was observed with increasing volume fraction for oxidized and silanized graphene nanoplatelets reaching values close to the neat resin, as illustrated in Figure 10(c) and (d).

After the chemical functionalizations on the surface of graphene nanoplatelets (oxidation and silanization), percolation networks were only achieved with the lowest concentration of silanized nanoparticles. In higher concentrations of silanized nanoplatelets, the nanoparticles can be covered by the polymer chains due to strong intermolecular interaction with the polymer chains. Besides, a homogeneous dispersion state of nanoparticles in the matrix as shown in TEM images (Figure 8(e) and (f)) was observed. Thus, preventing the nanoparticles to create paths for percolation and resulting in the decay of the electrical conductivity.²⁷ Additionally, the oxygen content for oxidized nanoplatelets also influenced negatively the percolation threshold by breaking the hybridization sp² and impairing the intrinsic electrical conductivity of the nanoplatelets.^25–27

Conclusions

This study evaluated the effect of chemical functionalizations (oxidation and silanization with 3-APTES) on the surface of the MWCNT and graphene nanoplatelets on thermal, mechanical and electrical properties of epoxy/nanocomposites. The oxidation and silanization on the surface of both nanoparticles were confirmed by FT-IR, XPS, Raman and TEM analysis. Epoxy nanocomposites were prepared using these functionalized nanoparticles with two different concentrations. For thermal properties, the oxidation and silanization on the surface of both nanoparticles did not restrict the movement of the polymer chains of the epoxy/nanocomposites compared to neat resin. However, the highest value for glass transition temperature (Tg) was only achieved when 0.15 vol.% of oxidized MWCNT were added in the epoxy matrix. Besides, for graphene nanoplatelets, the highest value for Tg of the epoxy/nanocomposites was only achieved when 0.50 vol% of these nanoparticles were added into the matrix.

For mechanical properties, higher values of tensile modulus and hardness were achieved with the addition of silanized MWCNT, oxidized and silanized graphene nanoplatelets probably because of the homogeneous distribution of the nanofillers in the polymer matrix. This result corroborates with the images obtained by TEM. Additionally, from these results, we can think that stress transfer (interfacial shear stress) is the main mechanism, since Tg was not changed indicating the restriction of the mobility at the interface.

Based on impedance spectroscopy and images obtained by TEM, some distinct behaviors were observed for the two nanoparticles. The percolation networks were achieved with 0.50 vol.% of pristine MWCNT and 0.15 vol.% of silanized graphene nanoplatelets exhibiting the highest electrical conductivity values for low frequencies in the order of 10⁻⁷ S/m. Higher conductivity values of nanocomposites were achieved with pristine MWCNT due to the higher aspect ratios than to chemically functionalized broken tubes. As regards the graphene nanoplatelets, silanized nanoplatelets with higher concentration were covered by the polymer chains due to more homogeneous dispersion state in the matrix resulting in the decay of the electrical conductivity. The oxygen content of oxidized nanoplatelets also influenced negatively the percolation threshold impairing the intrinsic electrical conductivity of the nanoplatelets. Finally, the choice between chemical functionalizations and the type of nanoparticles depends on the final properties and the polymer matrix of the nanocomposite to be produced.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors acknowledge the support received from the State University of Santa Catarina – UDESC as well as the financial support from CAPES and the scholarship to the graduate student Roger Hoél Bello.

References

Kotsilkova

Pissis

Silvestre

et al.

Thermoset nanocomposites for engineering applications, 1st ed. United Kingdom: Smithers Rapra, 2007, pp. 1–1.

Avilès

JVF

Cauich-Rodríguez

Rodríguez-González

JA MP

. Oxidation and silanization of MWCNTs for MWCNT/vinyl ester composites. Polym Lett 2011; 5: 766–776.

Kathi

Rhee

K-Y

Lee

. Effect of chemical functionalization of multi-walled carbon nanotubes with 3-aminopropyltriethoxysilane on mechanical and morphologicalproperties of epoxy nanocomposites. Compos Part A 2009; 40: 800–809.

Velasco-Santos

Martíinez-Hernández

Lozada-Cassou

et al.

Chemical functionalization of carbon nanotubes through an organosilane. Nanotechnology 2002; 13: 495–498.

Tang

et al.

Dispersion, interfacial interaction and re-agglomeration of functionalized carbon nanotubes in epoxy composites. Carbon 2010; 48: 1824–1834.

Tasis

Tagmatarchis

Bianco

et al.

Chemistry of carbon nanotubes. Chem Rev 2006; 106: 1105–1136.

Lavorgna

Romeo

Martone

et al.

Silanization and silica enrichment of multi-walled carbon nanotubes: synergistic effects on the thermal-mechanical properties of epoxy nanocomposites. Eur Polym J 2013; 49: 428–438.

Wang

Young

et al.

Control of the functionality of graphene oxide for its application in epoxy nanocomposites. Polymer 2013; 54: 6437–6446.

Wang

Xing

Zhang

et al.

Covalent functionalization of graphene with organosilane and its use as a reinforcement in epoxy composites. Compos Sci Technol 2012; 72: 737–743.

10.

Lee

J-H

Rhee

Park

. Silane modification of carbon nanotubes and its effects on the material properties of carbon/CNT/epoxy three-phase composites. Compos Part A 2011; 42: 478–483.

11.

Prodana

Ionita

Bojin

et al.

Different methods to functionalization of multi-walled carbon nanotubes for hybrid nanoarchitectures. J Sust Energ 2012; 3: 62–66.

12.

Kathi

Rhee

K-Y

. Surface modification of multi-walled carbon nanotubes using 3-aminopropyltriethoxysilane. J Mater Sci 2008; 43: 33–37.

13.

Kim

Tang

. Functionalization of carbon nanotubes using a silane coupling agent. Carbon 2006; 44: 3232–3238.

14.

W-S

Deng

B-J

et al.

Preparation and characterization of long-chain alkyl silane-functionalized graphene film. J Mater Sci 2012; 48: 156–161.

15.

Stobinski

Lesiak

Kövér

et al.

Multiwall carbon nanotubes purification and oxidation by nitric acid studied by the FTIR and electron spectroscopy methods. J Alloys Comp 2010; 501: 77–84.

16.

Yuen

S-M

C-C M

Chiang

C-L

et al.

Morphology and properties of aminosilane grafted MWCNT/polyimide nanocomposites. J Nanomater 2008; 2008: 1–15.

17.

Yang

Shan

et al.

Covalent functionalization of chemically converted graphene sheets via silane and its reinforcement. Carbon Lett 2009; 19: 4632–4638.

18.

Ferrari

Meyer

Scardaci

et al.

Raman spectrum of graphene and graphene layers. Nat Nanotechnol 2006; 97: 1–6.

19.

Nguyen

et al.

Synthesis of multi-layer graphene films on copper tape by atmospheric pressure chemical vapor deposition method. Adv Nat Sci: Nanosci Nanotechnol 2013; 4: 1–5.

20.

Putz

Palmeri

Cohn

et al.

Effect of cross-link density on interphase creation in polymer nanocomposites. Macromolecules 2008; 41: 6752–6756.

21.

Ajayan

Schadler

Braun

. Nanocomposite science and technology, 1st ed. Weinheim: Wiley, 2003, pp. 77–77.

22.

Kosmidou

TH V

. Structural, mechanical and electrical characterization of epoxy-amine/carbon black nanonocomposites. Express Polym Lett 2008; 2: 364–372.

23.

Giacometti

Carvalho

AJF

Electrical conduction. In: Canevarolo

SVJ

(ed). Techniques for characterizing polymers, 1st ed. Artliber: São Paulo, 2004.

24.

Sene

Ramos

Becker

et al.

Electrical conductivity behavior of epoxy matrix nanocomposites with simultaneous dispersion of carbon nanotubes and clays. Polym Compos 2014; 37: 1–9.

25.

Sreeprasad

Berry

. How do the electrical properties of graphene change with its functionalization? Small 2013; 9: 341–350.

26.

Potts

Dreyer

Bielawski

et al.

Graphene-based polymer nanocomposites. Polymer 2011; 52: 5–25.

27.

Kanham

Ponnamma

AL-Madeed

Electrical properties of graphene polymer nanocomposite. In: Sadasivuni

Ponnamma

Kim

et al. (eds). Graphene-based polymer nanocomposites in electronics, Switzerland: Springer, 2015, pp. 25–47.