Epoxy Nanocomposites Reinforced with Functionalized Carbon Nanotubes

In this article, amino functionalized multiwalled carbon nanotubes (MWCNTs) were prepared by chemical modification of the surface of a MWCNTs using γ-aminopropyltriethoxysilane (APTES) and dispersed into the epoxy composition. The modifying agent (APTES) was directly deposited on the MWCNTs surfaces. For the functionalization of MWCNTs, was used not the APTES concentrate, as it had been described in previous works, but its freshly prepared aqueous solution. Properties of APTES-treated MWCNTs were characterized by transmission electron microscope (TEM), Raman spectroscopy and FT-IR. The results showed that the functionalization and chemical compatibility of APTES-treated MWCNTs with epoxy composition provides their best dispersion in the epoxy composition, had important influence on curing behavior, structure and physicochemical properties of the epoxy composites plasticized with trichloroethyl phosphate. The results showed that the functionalization and chemical compatibility of APTES-treated MWCNTs with epoxy composition provides increased of physicomechanical properties of epoxy composites: bending stress increases by 194% and bending modulus increases by 137%, the tensile strength increases by 108% and the tensile elastic modulus increases by 52%, impact strength increases by 300%, in comparison with plasticized epoxy composite that does not contain MWCNTs.

Currently, nanocomposites are often used in various industries, due to their unique physicomechanical characteristics, which are due to the fact that nanodispersed particles are not stress concentrators in polymer composites [4,[12][13][14][15].
Taking into account excellent physical and mechanical properties, as well as high aspect ratio of carbon nanotubes (CNTs), they are believed to be a promising reinforcement phase for polymer The qualitative characteristics of ED-20 are presented in Table 1. Polyethylene polyamine (PEPA) (TS 6-02-594-85) manufactured by CHIMEX Limited (Russia) was used as a hardener of an epoxy compositions. Its chemical formula: H2 N(-CH2CH2NH-)nH2; n = 1-4. The qualitative characteristics of PEPA are presented in Table 2.  3) 1020 Amine number (mg KOH/g) 1250 Nitrogen content (% by weight) 30.0 To plasticize epoxy composites and making them flame-resistant, we used tri-2-chloroethyl ether of orthophosphoric acid (TCEP) with purity of 95.99%, which is an orthophosphoric acid and ethylene chlorohydrin ester, manufactured by PJSC "Khimprom" (Novocheboksarsk, Russia).
The choice of TCEP is due to the presence of combustion inhibitors-phosphorus and chlorine, which reduce the flammability of the epoxy composite [27,28]. The qualitative characteristics of TCEP are presented in Table 3. Table 3. Properties of TCEP.
The choice of TCEP is due to the presence of combustion inhibitors-phosphorus and chlorine, which reduce the flammability of the epoxy composite [27,28]. The qualitative characteristics of TCEP are presented in Table 3. Multi-walled carboxylated carbon nanotubes (MWCNTs) manufactured by "Nanotechcenter, Ltd." (Tambov, Russia) were used as a nanostructural additive. The qualitative characteristics of MWCNTs are presented in Table 4. The surface of MWCNTs was functionalized with a γ-aminopropyltriethoxysilane (APTES) coupling agent, manufactured by the Penta-91 company (Moscow, Russia). For this purpose, 0.5 g of MWCNTs was dispersed in 100 mL of the H 2 O-APTES (95-5) solution with an ultrasonic homogenizer for 10 min. The suspension was kept under reflux at 80 • C under constant low-speed stirring at 100 rpm for 12 h. The pH of the mixture was adjusted to 5. We selected an acidic media (CH 3 COOH) in order to increase the level of silanol formation and to decrease self-condensation reactions between the hydrolyzed silanol groups. To remove the remaining silane compound around the MWCNTs particles the resulting suspension was centrifuged and washed twice with H 2 O. Then, the product was dried in the laboratory oven [29].
Previously, we determined experimentally the ratio of epoxy oligomer, plasticizer and hardener: 100 parts by mass of ED-20, 40 parts by mass of TCEP and 15 parts by mass of PEPA [27,28]. MWCNTs were added as a modifying agent (0.01-0.50 parts by mass) into a plasticized epoxy composition. Ultrasonic treatment of the composition was used to increase the uniformity of distribution and hinder the aggregation of MWCNTs particles, as well as the activation of its surface and binder. The parameters of the ultrasound exposure were: frequency −22 ± 2 kHz, power −400 W, duration −60 min [30]. Before curing the mixture was degassed under vacuum at 25 ± 5 • C for 30 min. The preparation process of MWCNTs/epoxy nanocomposites was shown in Figure 2. Bending stress and flexural modulus was determined according to ISO 178: 2010, strength and modulus of tensile elasticity was determined according to ISO 527-2: 2012, impact strength was determined according to ISO 179-1: 2010, surface morphology of the samples was studied using a Tescan VEGA 3 SBH scanning electron microscope (Brno, Czech Republic), FT-IR spectroscopy was carried out using the Shimadzu IRTracer-100 (Tokyo, Japan), Raman spectra were measured by DXR Raman Microscope (Thermo scientific, Waltham, Massachusetts, USA) with laser excitation at 532 nm, the MWCNTs structure was studied by transmitting electron microscope JEOL JEM-1400 (Tokyo, Japan). The dispersion of MWCNTs was tested by dynamic light scattering method and Nicomp 380 ZLS (Particle Sizing Systems, Inc., Santa Barbara, California, USA) was applied. Three parallel studies were done, the coefficient of variation was 7%. Nicomp 380 ZLS is a fully automatic particle size and zeta potential analyzer. Measurement of the zeta potential,or the potential of the double layer, allows you to determine the strength of the interaction between the particles, which is a characteristic of the stability of the system and the ability of particles to agglomerate. Electrophoresis is used to determine the zeta potential in the Nicomp 380 ZLS analyzer. A small amount of sample is placed in a measuring cell with palladium electrodes that create an electromagnetic field. The specific surface area of the samples was determined by means of the Quantachrome Nova 2200 specific surface area and porosity analyzer (Boynton Beach, FL, USA) using the initial portion of the isotherm of physical nitrogen sorption (99.999%). The studied samples were placed in a cell preliminarily calibrated with respect to internal volume at a temperature of liquid nitrogen (about 78 K), which was degassed in vacuum to a constant mass at a given temperature (150 °C) for 3 h. After that, the cell was installed in the device and the isotherm of nitrogen adsorption was measured in the pressure range 0.03-0.3 P/Po. The specific surface of the test sample was calculated by the BET method with the Quantachrome Nova 2200 software [29].

Results and Discussion
As a modifying agent (0.01-0.50 parts by mass) pristine MWCNTs were added into plasticized epoxy composition (100ED-20 + 40TCEP + 15PEPA, parts by mass). Previous studies have shown that the addition of TCEP into epoxy composition provides not only an increase in mechanical characteristics (bending stress increases by 32% and impact strength is 2.6 times higher), but also reduces the flammability of the epoxy composite (oxygen index (OI) increases from 19% to 27% by volume) [27,28,30]. Bending stress and flexural modulus was determined according to ISO 178: 2010, strength and modulus of tensile elasticity was determined according to ISO 527-2: 2012, impact strength was determined according to ISO 179-1: 2010, surface morphology of the samples was studied using a Tescan VEGA 3 SBH scanning electron microscope (Brno, Czech Republic), FT-IR spectroscopy was carried out using the Shimadzu IRTracer-100 (Tokyo, Japan), Raman spectra were measured by DXR Raman Microscope (Thermo scientific, Waltham, MA, USA) with laser excitation at 532 nm, the MWCNTs structure was studied by transmitting electron microscope JEOL JEM-1400 (Tokyo, Japan).
The dispersion of MWCNTs was tested by dynamic light scattering method and Nicomp 380 ZLS (Particle Sizing Systems, Inc., Santa Barbara, CA, USA) was applied. Three parallel studies were done, the coefficient of variation was 7%. Nicomp 380 ZLS is a fully automatic particle size and zeta potential analyzer. Measurement of the zeta potential, or the potential of the double layer, allows you to determine the strength of the interaction between the particles, which is a characteristic of the stability of the system and the ability of particles to agglomerate. Electrophoresis is used to determine the zeta potential in the Nicomp 380 ZLS analyzer. A small amount of sample is placed in a measuring cell with palladium electrodes that create an electromagnetic field. The specific surface area of the samples was determined by means of the Quantachrome Nova 2200 specific surface area and porosity analyzer (Boynton Beach, FL, USA) using the initial portion of the isotherm of physical nitrogen sorption (99.999%). The studied samples were placed in a cell preliminarily calibrated with respect to internal volume at a temperature of liquid nitrogen (about 78 K), which was degassed in vacuum to a constant mass at a given temperature (150 • C) for 3 h. After that, the cell was installed in the device and the isotherm of nitrogen adsorption was measured in the pressure range 0.03-0.3 P/Po. The specific surface of the test sample was calculated by the BET method with the Quantachrome Nova 2200 software [29].

Results and Discussion
As a modifying agent (0.01-0.50 parts by mass) pristine MWCNTs were added into plasticized epoxy composition (100ED-20 + 40TCEP + 15PEPA, parts by mass). Previous studies have shown that the addition of TCEP into epoxy composition provides not only an increase in mechanical characteristics (bending stress increases by 32% and impact strength is 2.6 times higher), but also Polymers 2020, 12, 1816 6 of 14 reduces the flammability of the epoxy composite (oxygen index (OI) increases from 19% to 27% by volume) [27,28,30].
The conducted research has shown that the most rational content of pristine MWCNTs as a structuring additive is 0.1 parts by mass, which provides maximum values of physical and mechanical properties, Table 5, at the same time the bending failure stress increases by 135% and the bending elastic modulus increases by 90%, the tensile strength increases by 88% and the tensile elastic modulus increases by 43%, impact resilience increases by 225%. To ensure the chemical interaction of the filler with the polymer matrix and to increase physicomechanical characteristics of epoxy composites based on them, the MWCNTs surface was treated with the APTES sizing additive.
The effect of the APTES treatment on the structure and degree of MWCNTs defectiveness was studied using Raman spectroscopy and transmission electron microscopy (TEM). Analysis of Raman spectra was shown that the R (the ratio of peak D and peak G) value of (1) is significantly smaller than (2), which infers a more regular structure of carbon atoms of (1), suggesting that the modification increased the number of defects on MWCNTs, Figure 3. The conducted research has shown that the most rational content of pristine MWCNTs as a structuring additive is 0.1 parts by mass, which provides maximum values of physical and mechanical properties, Table 5, at the same time the bending failure stress increases by 135% and the bending elastic modulus increases by 90%, the tensile strength increases by 88% and the tensile elastic modulus increases by 43%, impact resilience increases by 225%. To ensure the chemical interaction of the filler with the polymer matrix and to increase physicomechanical characteristics of epoxy composites based on them, the MWCNTs surface was treated with the APTES sizing additive.
The effect of the APTES treatment on the structure and degree of MWCNTs defectiveness was studied using Raman spectroscopy and transmission electron microscopy (TEM). Analysis of Raman spectra was shown that the R (the ratio of peak D and peak G) value of (1) is significantly smaller than (2), which infers a more regular structure of carbon atoms of (1), suggesting that the modification increased the number of defects on MWCNTs, Figure 3. According to TEM, a decrease in aggregation and a degree of particle confusion is observed, which is confirmed by an increase in the distance between individual nanotubes due to the APTES treatment, Figure 4, as well as an increase in the specific surface of MWCNTs from 105 to 182 m 2 /g. According to TEM, a decrease in aggregation and a degree of particle confusion is observed, which is confirmed by an increase in the distance between individual nanotubes due to the APTES treatment, Figure 4, as well as an increase in the specific surface of MWCNTs from 105 to 182 m 2 /g. To study the dispersion of MWCNTs in an epoxy binder, a dynamic light scattering technique was used to estimate the effective particle size of the dispersed phase that would be the true size if these particles were spherical. Since MWCNTs themselves and the agglomerates formed by them are not spherical, the measurement results allow us to evaluate only the relative changes in the dispersed composition of the analyzed suspensions. As the analysis of the diagram of the suspension particles distribution by effective sizes shows, Figure 5, the suspension with pristine MWCNTs is polydisperse, Figure 5a. The average effective size of the smallest agglomerates is 418 nm. The second peak shows an effective particle size of about 7 μm, large agglomerates slightly prevailing over smaller ones in terms of quantity. For a suspension with APTES-treated MWCNTs, the diagram of the suspension particles distribution by effective sizes is monodisperse and the average effective size of the agglomerates is in the region of 900 nm, with no larger agglomerates being observed. Thus, the APTES treatment of MWCNTs surface provides their best dispersion in the epoxy composition.
The formation of strong bonds between MWCNTs and APTES was proved by IR Fourier spectroscopy, Figure 6. As can be seen from the spectrum of the sample after modification, vibration peaks corresponding to APTES appear; the vibration intensity of hydroxyl groups (3400-3200 cm −1 ), which are involved in the formation of the APTES layer, decreases. One can also see a broad To study the dispersion of MWCNTs in an epoxy binder, a dynamic light scattering technique was used to estimate the effective particle size of the dispersed phase that would be the true size if these particles were spherical. Since MWCNTs themselves and the agglomerates formed by them are not spherical, the measurement results allow us to evaluate only the relative changes in the dispersed composition of the analyzed suspensions. As the analysis of the diagram of the suspension particles distribution by effective sizes shows, Figure 5, the suspension with pristine MWCNTs is polydisperse, Figure 5a. To study the dispersion of MWCNTs in an epoxy binder, a dynamic light scattering technique was used to estimate the effective particle size of the dispersed phase that would be the true size if these particles were spherical. Since MWCNTs themselves and the agglomerates formed by them are not spherical, the measurement results allow us to evaluate only the relative changes in the dispersed composition of the analyzed suspensions. As the analysis of the diagram of the suspension particles distribution by effective sizes shows, Figure 5, the suspension with pristine MWCNTs is polydisperse, Figure 5a. The average effective size of the smallest agglomerates is 418 nm. The second peak shows an effective particle size of about 7 μm, large agglomerates slightly prevailing over smaller ones in terms of quantity. For a suspension with APTES-treated MWCNTs, the diagram of the suspension particles distribution by effective sizes is monodisperse and the average effective size of the agglomerates is in the region of 900 nm, with no larger agglomerates being observed. Thus, the APTES treatment of MWCNTs surface provides their best dispersion in the epoxy composition.
The formation of strong bonds between MWCNTs and APTES was proved by IR Fourier spectroscopy, Figure 6. As can be seen from the spectrum of the sample after modification, vibration peaks corresponding to APTES appear; the vibration intensity of hydroxyl groups (3400-3200 cm −1 ), which are involved in the formation of the APTES layer, decreases. One can also see a broad The average effective size of the smallest agglomerates is 418 nm. The second peak shows an effective particle size of about 7 µm, large agglomerates slightly prevailing over smaller ones in terms of quantity. For a suspension with APTES-treated MWCNTs, the diagram of the suspension particles distribution by effective sizes is monodisperse and the average effective size of the agglomerates is in the region of 900 nm, with no larger agglomerates being observed. Thus, the APTES treatment of MWCNTs surface provides their best dispersion in the epoxy composition. The formation of strong bonds between MWCNTs and APTES was proved by IR Fourier spectroscopy, Figure 6. As can be seen from the spectrum of the sample after modification, vibration peaks corresponding to APTES appear; the vibration intensity of hydroxyl groups (3400-3200 cm −1 ), which are involved in the formation of the APTES layer, decreases. One can also see a broad maximum of about 1000 cm −1 , which indicates the presence of Si-O bonds in the sample. In addition, peaks appears at 1040 cm −1 and 630 cm −1 , confirming the formation of the Si-O-C bond.  The chemical interaction of the APTES functional groups and the epoxy oligomer was proved in [29].
Based on this, it can be assumed that the reactions shown in Figure 7 take place in the system. The results showed that the functionalization and chemical compatibility of APTES-treated MWCNTs with epoxy composition provides increased of physicomechanical properties of epoxy composites: bending stress increases by 194% and bending modulus increases by 137%, the tensile strength increases by 108% and the tensile elastic modulus increases by 52%, impact strength increases by 300%, in comparison with plasticized epoxy composite that does not contain MWCNTs, Figure 8. Comparison of the obtained strengthening effects with the literature data (an increase in the physicomechanical characteristics of polymer composites by 20%-40%) [4,17,26] proves the effectiveness of the MWCNTs modification method proposed in this work. The chemical interaction of the APTES functional groups and the epoxy oligomer was proved in [29].
Based on this, it can be assumed that the reactions shown in Figure 7 take place in the system.  The chemical interaction of the APTES functional groups and the epoxy oligomer was proved in [29].
Based on this, it can be assumed that the reactions shown in Figure 7 take place in the system. The results showed that the functionalization and chemical compatibility of APTES-treated MWCNTs with epoxy composition provides increased of physicomechanical properties of epoxy composites: bending stress increases by 194% and bending modulus increases by 137%, the tensile strength increases by 108% and the tensile elastic modulus increases by 52%, impact strength increases by 300%, in comparison with plasticized epoxy composite that does not contain MWCNTs, Figure 8. Comparison of the obtained strengthening effects with the literature data (an increase in the physicomechanical characteristics of polymer composites by 20%-40%) [4,17,26] proves the effectiveness of the MWCNTs modification method proposed in this work. The results showed that the functionalization and chemical compatibility of APTES-treated MWCNTs with epoxy composition provides increased of physicomechanical properties of epoxy composites: bending stress increases by 194% and bending modulus increases by 137%, the tensile strength increases by 108% and the tensile elastic modulus increases by 52%, impact strength Polymers 2020, 12, 1816 9 of 14 increases by 300%, in comparison with plasticized epoxy composite that does not contain MWCNTs, Figure 8. Comparison of the obtained strengthening effects with the literature data (an increase in the physicomechanical characteristics of polymer composites by 20-40%) [4,17,26] proves the effectiveness of the MWCNTs modification method proposed in this work.  Figure 9a presents fractography of the destruction of epoxy composite samples without MWCNTs, which is characterized by a rather smooth fracture surface, showing a low ability to crack resistance [31]. When MWCNTs are added into the epoxy composition they affect the morphology of the matrix-there appear layered structures, Figure 9b. Along with the brittle fracture where numerous scales are formed, there are local areas in the epoxy composite, indicating the flow of material in the process of its destruction. Moreover, fibrous structures can be clearly observed in some places of plastic destruction, which are formed as a result of the intensive drawing out of the polymer matrix, Figure 9c. Considering MWCNTs as a solid phase hardener [32] we can explain the increase in plasticity of the epoxy composite. In this case, a smaller number (compared to the volume of the composition) of crosslinks is formed in the border area of MWCNTs and the epoxy composition, and, therefore, this area will have greater mobility [33].  Figure 9a presents fractography of the destruction of epoxy composite samples without MWCNTs, which is characterized by a rather smooth fracture surface, showing a low ability to crack resistance [31]. When MWCNTs are added into the epoxy composition they affect the morphology of the matrix-there appear layered structures, Figure 9b. Along with the brittle fracture where numerous scales are formed, there are local areas in the epoxy composite, indicating the flow of material in the process of its destruction. Moreover, fibrous structures can be clearly observed in some places of plastic destruction, which are formed as a result of the intensive drawing out of the polymer matrix, Figure 9c. Considering MWCNTs as a solid phase hardener [32] we can explain the increase in plasticity of the epoxy composite. In this case, a smaller number (compared to the volume of the composition) of crosslinks is formed in the border area of MWCNTs and the epoxy composition, and, therefore, this area will have greater mobility [33].
In Figure 9d,e the destruction surface of an epoxy composite modified with amino-functionalized MWCNTs can be seen. A very large fraction of plastic destruction appears in the sample, the borders of the scales become blurred, and their size increases. An increase in the proportion of plastic deformation of the composite material confirms the hypothesis that APTES-treated MWCNTs can act as a solid-state hardener. This may be on account of the functionalization of the MWCNTs surface, which results in the formation of a dense cross-linking network around the nanotube with oligomer epoxy groups.
The study of the curing kinetics of epoxy compounds, Figure 10, has shown that the addition of both pristine MWCNTs and APTES-treated MWCNTs affects the structure formation processes during curing of the epoxy composite.
numerous scales are formed, there are local areas in the epoxy composite, indicating the flow of material in the process of its destruction. Moreover, fibrous structures can be clearly observed in some places of plastic destruction, which are formed as a result of the intensive drawing out of the polymer matrix, Figure 9c. Considering MWCNTs as a solid phase hardener [32] we can explain the increase in plasticity of the epoxy composite. In this case, a smaller number (compared to the volume of the composition) of crosslinks is formed in the border area of MWCNTs and the epoxy composition, and, therefore, this area will have greater mobility [33]. In Figure 9d,e the destruction surface of an epoxy composite modified with amino-functionalized MWCNTs can be seen. A very large fraction of plastic destruction appears in the sample, the borders of the scales become blurred, and their size increases. An increase in the proportion of plastic deformation of the composite material confirms the hypothesis that APTES-treated MWCNTs can act as a solid-state hardener. This may be on account of the functionalization of the MWCNTs surface, which results in the formation of a dense cross-linking network around the nanotube with oligomer epoxy groups.
The study of the curing kinetics of epoxy compounds, Figure 10, has shown that the addition of both pristine MWCNTs and APTES-treated MWCNTs affects the structure formation processes during curing of the epoxy composite. The addition of APTES-treated MWCNTs into the epoxy composition accelerates the curing process, which can be seen in the reduction of the duration of the gelation process from 65 to 52 min， ,and the duration of the curing process from 78 to 61 min, with an increase in the maximum curing temperature from 132 to 172 °C. Table 6, in comparison with a composition containing pristine MWCNTs, which further confirms the participation of APTES functional groups in the curing process. The addition of APTES-treated MWCNTs into the epoxy composition accelerates the curing process, which can be seen in the reduction of the duration of the gelation process from 65 to 52 min, and the duration of the curing process from 78 to 61 min, with an increase in the maximum curing temperature from 132 to 172 • C. Table 6, in comparison with a composition containing pristine MWCNTs, which further confirms the participation of APTES functional groups in the curing process. τ gel is the duration of gelation process, τ cur is the duration of curing, T max is the maximum temperature of the self-heating of the sample during curing.

Conclusions
The conducted research has proven that it is possible to regulate directly operational properties of epoxy composites using small MWCNTs additives that allow to develop epoxy composites with high performance properties. The effectiveness of the surface modification of MWCNTs with γ-aminopropyltriethoxysilane and the formation of strong chemical bonds at the polymer matrix/filler interface have been proved, which ensured an increase in the physicomechanical characteristics of epoxy composites: bending stress increases by 194% and bending modulus increases by 137%, the tensile strength increases by 108% and the tensile elastic modulus increases by 52%, impact strength increases by 300%, in comparison with plasticized epoxy composite that does not contain MWCNTs.
The effect on the structure and structure formation processes during curing of epoxy compositions containing both pristine MWCNTs and APTES-treated MWCNTs has been established. The addition of APTES-treated MWCNTs into the epoxy composition accelerates the curing process, which can be seen in the reduction of the duration of the gelation process from 65 to 52 min, and the duration of the curing process from 78 to 61 min, with an increase in the maximum curing temperature from 132 to 172 • C in comparison with a composition containing pristine MWCNTs, which further confirms the participation of APTES functional groups in the curing process.
As a result of the studies, it has been proved that the addition of 0.05 to 0.5 parts by weight of both pristine MWCNTs and APTES-treated MWCNTs into the epoxy composite increases the thermal diffusivity by 20-42% and the thermal conductivity by 73-87%. Funding: The reported study was funded by RFBR according to the research project No. 18-29-19048.

Conflicts of Interest:
The authors declare no conflict of financial and non-financial interest requiring disclosure in this article.