Enhancing Mechanical Properties of Graft-Type Nanocomposites Using Organically Modified SiO2 and Polypropylene Containing Reactive Methoxy Groups

In situ grafting of a reactive matrix and nanofillers is a promising strategy to fabricate graft-type polypropylene (PP)-based nanocomposites, where the grafting efficiency is affected by the initial dispersion of nanofillers in the matrix. In this work, influences of surface organic modification of nanofillers were investigated on properties of PP/SiO2 nanocomposites using poly(propylene-co-octenyltrimethoxysilane) as a reactive matrix. The surface modification of SiO2, especially with longer alkyl chains, led to improved dispersion of nanoparticles, thus promoting the grafting reaction and mechanical properties. The combination of in situ grafting and surface modification of nanofillers provided several benefits, most notably in balancing the strength and the toughness, which could not be achieved by the grafting alone.


Introduction
Polymer nanocomposites, consisting of at least one phase having dimensions smaller than 100 nm [1], have attracted great interest owing to their potential applications in different industrial sectors. Significant increases in the particle number density and interfacial area due to the small size enable nanocomposites to achieve desired properties at a much smaller loading of fillers compared to conventional microcomposites [2]. Meanwhile, strong interparticle attraction in comparison to weak interfacial interaction with polymer matrices constitutes one of the main impediments in the development of high-performance nanocomposites.

Synthesis of PP-OTMS
PP-OTMS was synthesized by copolymerization of propylene and OTMS using a MgCl 2 -supported Ziegler-Natta catalyst [39]. The polymerization was performed in a 1 L stainless steel reactor in a semi-batch mode. To the reactor blanked by N 2 , 500 mL of heptane as a solvent, 15 mmol of TEA as a cocatalyst, and 10 mmol of OTMS as a comonomer were introduced. The solvent was saturated with 0.5 MPa of propylene at 50 • C for 30 min. Followed by the introduction of 16 mmol of H 2 , catalyst powder (50 mg) was injected to initiate the polymerization. The polymerization was continued at 50 • C and 0.5 MPa for 60 min, followed by depressurization. The solvent was removed by decantation, and the resultant polymer powder was washed repetitively with ethanol and acetone under N 2 . Finally, the powder was purified by reprecipitation (xylene to acetone) and dried in vacuum at room temperature. PP homopolymer (HomoPP) was synthesized in the absence of OTMS under the same conditions, except 5.0 mmol of TEA used as a co-catalyst.

Surface Modification of SiO 2
Surface modification of SiO 2 was performed according to a previously reported method [45]. SiO 2 (0.3 g) was added in 30 mL of toluene in a round-bottom flask, and sonicated for 30 min. A specified amount of a silane coupling agent (0.01, 0.1, or 1 mmol) was added under N 2 . The mixture was stirred at room temperature for 1 h. The obtained product was collected by centrifugation, repetitively washed with anhydrous ethanol, and dried in a vacuum oven at 80 • C for 24 h. The modified SiO 2 samples are denoted as Cx-y-SiO 2 , where x and y represent the alkyl chain length and the amount (mmol) of the silane coupling agents (per 0.3 g of SiO 2 ), respectively. For example, C6-0.1-SiO 2 stands for SiO 2 modified with 0.1 mmol of trichlorohexylsilane.

Preparation of PP-OTMS/SiO 2 Nanocomposites
PP-OTMS/SiO 2 nanocomposites were prepared by the following procedure: First, PP-OTMS (3.7 g) was pre-impregnated with 1.0 wt% of AO-50 in 100 mL of acetone. After evaporation of the solvent under N 2 flow, the polymer was dried at room temperature under vacuum for 12 h. The dried polymer was melt-mixed with 5.0 wt% of unmodified SiO 2 or modified SiO 2 using Micro Compounder MC5 (Xplore) at 185 • C and 100 rpm for 15 min under N 2 atmosphere. The extrudate was hot-pressed into a 200 µm-thick film at 230 • C and 20 MPa for 5 min, followed by quenching at 100 • C for 5 min and then cooling at 0 • C for 3 min. For comparison, a nanocomposite (HomoPP/SiO 2 ) was also prepared using HomoPP and unmodified SiO 2 according to the same procedure.

Characterizations
Fourier transform infrared (FTIR) spectra were recorded on a Perkin Elmer Spectrum 100 spectrometer (PerkinElmer, Inc., Waltham, MA, USA) in the transmission mode between 4000 cm −1 and 450 cm −1 with a resolution of 4 cm −1 . The scans of each FTIR experiment are 64. A sample was mixed with dried KBr and pressed into a disc for the measurement. Water contact angle (WCA) measurements were carried out on SiO 2 samples using a contact angle goniometer (SImageAUTO 100, Excimer. Inc., Kanagawa, Japan). At room temperature, 10.0 µL of deionized water was dropped on a sample surface pressed on a glass slide. The WCA was determined by the tangent method. The weight loss of unmodified and modified SiO 2 was measured by thermal gravimetric analysis (TGA, Thermo plus evo, Rigaku, Tokyo, Japan). A sample was heated from 25 • C to 600 • C at the heating rate of 10 • C min −1 under dry air flow. The weight loss below 200 • C corresponds to the evaporation of physisorped water and surface hydroxyl groups, while the weight loss in a range of 200-600 • C corresponds to thermal decomposition of organic groups for modified SiO 2 .
The methoxy (OMe) content in PP-OTMS before and after melt mixing was analyzed by 1 H NMR (Bruker 400 MHz) operated at 120 • C with 1000 scans. Ca. 60 mg of a sample was dissolved in 0.2 mL of 1,1,2,2-tetrachloroethane-d 2 (an internal lock and a reference) and 0.5 mL of 1,2,4-trichlorobenzene containing 0.006 wt% of 2,6-di-tert-butyl-4-methylphenol (anti-oxidant). The OMe content was calculated using Equation (1), where H a and H c are the peak areas for the methyl protons of the methoxy group and the methine protons of the polymer backbone, respectively. A typical NMR spectrum of PP-OTMS and peak assignment are provided in Figure S1. The dispersion of SiO 2 nanoparticles was observed on a transmission electron microscope (TEM, Hitachi H-7650, Hitachi High-Tech Corporation, Tokyo, Japan) operated at an acceleration voltage of 100 kV. A 100-nm-thick specimen was cut from a nanocomposite film using an ultramicrotome instrument (Leica ULTRACUTS FCS, Leica Microsystems GmbH, Wetzlar, Germany) equipped with a diamond knife. The dispersion of SiO 2 in the matrix was quantitatively evaluated based on a dispersion parameter (D) defined in Equation (2), where µ and σ are the average size and its standard deviation of SiO 2 domains (SiO 2 particles or their aggregates) [46]. The analysis of TEM images was performed using ImageJ software, and covered at least 200 domains of SiO 2 , which corresponded to 3 TEM images taken at different regions and were sufficient to obtain stable D values. Differential scanning calorimetry (DSC) measurements were performed on Mettler Toledo DSC 822 under N 2 atmosphere. Ca. 8 mg of a sample was added in an aluminum pan, and heated to 230 • C at the heating rate of 20 • C min −1 . The melting temperature (T m ) and the crystallinity (X c ) were determined from the melting endotherm. After holding 230 • C for 10 min, the sample was cooled down to 25 • C at the cooling rate of 20 • C min −1 for acquiring the crystallization temperature (T c ), or to 144 • C at the rate of 50 • C min −1 for isothermal crystallization. In the latter case, the crystallization rate was calculated as the inverse of the half time of the crystallization (denoted as t 1/2 −1 ). Tensile properties were measured using a tensile tester (Abecks Inc., Dat-100, Tokyo, Japan) at a crosshead speed of 1 mm min −1 at room temperature. Dumbbell-shaped specimens were die-cut from a 200-µm-thick film. The tensile properties were reported as an average from at least four measurements per sample.

Results and Discussion
Successful organic modification of SiO 2 was confirmed by FTIR. As shown in Figure 1a, unmodified SiO 2 exhibited broad peaks at around 3500 cm −1 and 1617 cm −1 , which are characteristics of surface hydroxyl groups and physisorbed water [47]. The most intense peak around 1100 cm −1 and the sharp peak at 804 cm −1 are respectively attributed to Si-OH/Si-O-Si stretching and Si-O-Si bending [48,49]. After surface modification, new peaks appeared at 2962 cm −1 , 2921 cm −1 , 2847 cm −1 , and 1466 cm −1 . These correspond to asymmetric stretching of −CH 3 , asymmetric stretching of −CH 2 , symmetric stretching of −CH 2 , and C−H bending [50]. It is also noted that the peak ratio between the asymmetric stretching of −CH 2 and −CH 3 changed in line with the CH 2 /CH 3 ratio of the alkyl chain [47,48]. The morphology of the neat and surface modified SiO 2 was conducted. It was found that the shape and size of the SiO 2 before and after modification do not have changed ( Figure S2). TGA was implemented to determine the amount of silanes grafted on SiO2 surfaces. For unmodified SiO2, the weight loss gradually occurred upon heating to 600 °C, due to the vaporization of physisorbed water and the loss of surface hydroxyl groups [51]. For modified SiO2, a sharp weight loss was observed at around 200 °C, as shown in Figure 1b. According to literature, the weight loss of silane-modified SiO2 starts at around 200 °C via dissociative combustion of the alkyl chain, leaving the siloxy groups on the SiO2 surface [52]. The following equation was used to estimate the amount of grafted silanes from the weight loss in the range of 200−600 °C, where m% is the molar percentage of silanes with respect to SiO2. Msilane and MSiO 2 are the molecular weights of silane coupling agents and SiO2, respectively. ∆W is the weight loss of modified SiO2 between 200−600 °C minus that of unmodified SiO2. ra/s is the weight ratio of the alkyl chain in the correspondent silane coupling agent [52]. For example, ∆W is calculated as 3.27 wt% for C6-0.1-SiO2. The ra/s and Msilane values for trichlorohexylsilane are 0.39 and 219.61 g mol −1 , respectively. According to Equation (3), the molar percentage of the grafted silane is calculated as 1.57 mol%. Considering the surface area of SiO2, the silane grafting density can be further calculated as 1.43 nm −2 . Table 1 summarizes the silane molar percentage and the grafting density when silane coupling agents having different alkyl chain lengths were employed. It can be seen that at the fixed mole of silane coupling agents (0.1 mmol), the grafting amount decreased with the increase in the alkyl chain length. It is considered that the adsorption and reaction of a molecule of a bulky silane coupling agent inhibits the approach of a new molecule to the surrounding silanols. By increasing the amount of the silane coupling agent (C16), the grafting density increased TGA was implemented to determine the amount of silanes grafted on SiO 2 surfaces. For unmodified SiO 2 , the weight loss gradually occurred upon heating to 600 • C, due to the vaporization of physisorbed water and the loss of surface hydroxyl groups [51]. For modified SiO 2 , a sharp weight loss was observed at around 200 • C, as shown in Figure 1b. According to literature, the weight loss of silane-modified SiO 2 starts at around 200 • C via dissociative combustion of the alkyl chain, leaving the siloxy groups on the SiO 2 surface [52]. The following equation was used to estimate the amount of grafted silanes from the weight loss in the range of 200-600 • C, where m% is the molar percentage of silanes with respect to SiO 2 . M silane and M SiO 2 are the molecular weights of silane coupling agents and SiO 2 , respectively. ∆W is the weight loss of modified SiO 2 between 200-600 • C minus that of unmodified SiO 2 . r a/s is the weight ratio of the alkyl chain in the correspondent silane coupling agent [52]. For example, ∆W is calculated as 3.27 wt% for C6-0.1-SiO 2 . The r a/s and M silane values for trichlorohexylsilane are 0.39 and 219.61 g mol −1 , respectively. According to Equation (3), the molar percentage of the grafted silane is calculated as 1.57 mol%. Considering the surface area of SiO 2 , the silane grafting density can be further calculated as 1.43 nm −2 . Table 1 summarizes the silane molar percentage and the grafting density when silane coupling agents having different alkyl chain lengths were employed. It can be seen that at the fixed mole of silane coupling agents (0.1 mmol), the grafting amount decreased with the increase in the alkyl chain length. It is considered that the adsorption and reaction of a molecule of a bulky silane coupling agent inhibits the approach of a new molecule to the surrounding silanols. By increasing the amount of the silane coupling agent (C16), the grafting density increased and reached 0.94 nm −2 at 1 mmol of the silane addition. It must be noted that the grafting density was sufficiently lower than the original OH density (ca. 4-5 nm −2 ) for all cases, i.e., the OH sites remained available for grafting to PP-OTMS.  Figure 1c depicts the results of WCA measurements for SiO 2 samples. Without organic modification, the SiO 2 surface was highly hydrophilic (WCA = 0 • ) due to the presence of surface hydroxyl groups. The modification with silane coupling agents significantly increased WCAs to 126-142 • , indicating that the original hydrophilic surface changed to hydrophobic one due to the presence of aliphatic chains. The WCAs were found to be larger when the chain length increased from C6 to C12 regardless of a lower grafting density. These results pointed out that it is easier for a longer alkyl chain to shield the surface from water by forming a hydrophobic network [53].
A series of nanocomposites were prepared by melt compounding 5.0 wt% of unmodified or modified SiO 2 . Figure 2 shows TEM images of the resultant nanocomposites and the dispersion parameter (D) calculated based on Equation (2). Generally, a higher D value indicates better dispersion [46]. As known from literature, SiO 2 nanoparticles have poor compatibility with the PP matrix and easily form agglomerates due to strong particleparticle interactions (Figure 2a). The incorporation of a small amount of OTMS in PP chains helped to improve the dispersion as evidenced by the decrease in the size of SiO 2 domains (Figure 2b). In our previous paper, the improved dispersion was ascribed to the presence of polar functional groups as well as the in situ grafting of OTMS to SiO 2 surfaces to strengthen the interfacial interaction [39]. When modified SiO 2 nanoparticles were employed, the dispersion of SiO 2 in the PP-OTMS matrix further improved (Figure 2c-e). In Figure 2f, the D value increases along the alkyl chain length. These results are explained by the decrease in the cohesive attraction among SiO 2 nanoparticles and improved compatibility with the matrix due to organic modification, and/or the prevention of re-agglomeration due to more efficient grafting of PP-OTMS (described below).
During the melt mixing process, the OMe groups of PP-OTMS can react with surface silanol groups of SiO 2 to form Si-O-Si bonds, by which PP-OTMS grafts to the SiO 2 nanoparticles [39]. To confirm the occurrence of in situ grafting, the OMe content before and after melt mixing was analyzed by 1 H NMR ( Figure S3). As shown in Figure 3, even in the absence of SiO 2 , a part of OMe groups was consumed during melt mixing. This suggests that the OMe groups belonging to different chains can react with each other, most plausibly via hydrolysis/condensation with the aid of residual water. In the presence of SiO 2 , the decrease in the OMe content became more pronounced, which indicated the occurrence of in situ grafting. The consumption of the OMe group was greater for modified SiO 2 as compared to unmodified one, even though surface silanol groups were partially consumed by the organic modification. This result evidenced the importance of initial dispersion on the efficiency of the in situ grafting. The consumption of the OMe groups increased along with the alkyl chain length, in agreement with the improved dispersion parameter in Figure 2f. It is plausible that the improved dispersion of modified SiO 2 provided more contact interfaces to promote the grafting reaction. During the melt mixing process, the OMe groups of PP-OTMS can react with surf silanol groups of SiO2 to form Si-O-Si bonds, by which PP-OTMS grafts to the SiO2 na particles [39]. To confirm the occurrence of in situ grafting, the OMe content before after melt mixing was analyzed by 1 H NMR ( Figure S3). As shown in Figure 3, even in absence of SiO2, a part of OMe groups was consumed during melt mixing. This sugg that the OMe groups belonging to different chains can react with each other, most pla bly via hydrolysis/condensation with the aid of residual water. In the presence of SiO2, decrease in the OMe content became more pronounced, which indicated the occurre of in situ grafting. The consumption of the OMe group was greater for modified SiO compared to unmodified one, even though surface silanol groups were partially c sumed by the organic modification. This result evidenced the importance of initial disp sion on the efficiency of the in situ grafting. The consumption of the OMe groups increa along with the alkyl chain length, in agreement with the improved dispersion param in Figure 2f. It is plausible that the improved dispersion of modified SiO2 provided m contact interfaces to promote the grafting reaction.  During the melt mixing process, the OMe groups of PP-OTMS can react with surface silanol groups of SiO2 to form Si-O-Si bonds, by which PP-OTMS grafts to the SiO2 nanoparticles [39]. To confirm the occurrence of in situ grafting, the OMe content before and after melt mixing was analyzed by 1 H NMR ( Figure S3). As shown in Figure 3, even in the absence of SiO2, a part of OMe groups was consumed during melt mixing. This suggests that the OMe groups belonging to different chains can react with each other, most plausibly via hydrolysis/condensation with the aid of residual water. In the presence of SiO2, the decrease in the OMe content became more pronounced, which indicated the occurrence of in situ grafting. The consumption of the OMe group was greater for modified SiO2 as compared to unmodified one, even though surface silanol groups were partially consumed by the organic modification. This result evidenced the importance of initial dispersion on the efficiency of the in situ grafting. The consumption of the OMe groups increased along with the alkyl chain length, in agreement with the improved dispersion parameter in Figure 2f. It is plausible that the improved dispersion of modified SiO2 provided more contact interfaces to promote the grafting reaction.  The influences of SiO 2 and its surface modification on the melting and crystallization behaviors of nanocomposites were investigated by DSC. The acquired parameters, such as the melting temperature (T m ), crystallization temperature (T c ), and crystallinity (X c ) of the nanocomposites, are summarized in Table 2. The DSC profiles during heating and cooling are respectively provided in Figures S4 and S5. In Table 2, T m and X c of PP-OTMS were slightly higher than those of HomoPP due to slightly higher isotacity of PP-OTMS (98 mol% mmmm for PP-OTMS vs. 95 mol% for HomoPP). On the other hand, T c and t 1/2 −1 were obviously higher for PP-OTMS. This is due to the nucleation ability of PP-OTMS, which forms a long-chain branched (LCB) structure by interchain reaction during melt mixing and shows a nucleation ability similar to LCBPP [39]. The introduction of SiO 2 to HomoPP did not cause any significant change in the thermal properties and the crystallinity of the resultant nanocomposite. Contrary, the introduction of SiO 2 to the PP-OTMS matrix increased T c by 3 • C and doubled the crystallization rate. The significant crystallization acceleration has been reported for the other grafted-type nanocomposites, where grafted PP chains with lower mobility facilitate nucleation [54,55]. The crystallization acceleration was also observed with modified SiO 2 , and became more pronounced for a longer alkyl chain. These results are in perfect agreement with the TEM and NMR results that the surface modification improved the dispersion of SiO 2 nanoparticles and made the grafting reaction more efficient, in particular for a longer alkyl chain. Table 2. Melting and crystallization behaviors of nanocomposites analyzed by DSC.

Sample
T HomoPP 161  Tensile properties of nanocomposites were acquired using a uniaxial tensile tester. The representative stress-strain curves of the nanocomposites are provided in Figure S6. From Table 3 and Figure S7, the yield strength and elongation at break of PP-OTMS were higher than HomoPP as a consequence of LCBPP formation [56]. The addition of SiO 2 to HomoPP did not cause the improvement in the mechanical properties, but rather deteriorated the elongation at break due to the nanoparticle agglomeration and poorly connected interfaces [38,39]. On the other hand, the introduction of SiO 2 to PP-OTMS improved both of the Young's modulus and yield strength as compared to PP-OTMS. Such the improvement was associated with the improved dispersion and the successful grafting of PP-OTMS onto SiO 2 surfaces. The grafting not only improves the dispersion of SiO 2 in PP, but also improves the interfacial bonding (hence the yield strength). Grafted chains cocrystallize with the matrix, and SiO 2 acts as a physical crosslinker, which disappears above the melting point, so nanocomposites are melt processible [38]. This is the largest advantage of graft-type nanocomposites with identical matrix and grafted chains. In contrast, the physical crosslink structure with SiO 2 as the core is much less flexible to deformation, which leads to a large decrease in elongation at break. This is a common disadvantage of the graft-type nanocomposites [38][39][40], and the significant decrease in the elongation at break and the toughness was also observed in PP-OTMS/SiO 2 . By exploiting modified SiO 2 (Cx-0.1-SiO 2 ), the elongation at break was partially recovered in addition to further improvement in the yield strength. The degree of the recovery/improvement was found to be greater for a longer alkyl chain (Table 3). In general, the yield strength is largely determined by the dispersion and interfacial interactions between the polymer matrix and fillers. Hence, the higher yield strength for a longer alkyl chain reasonably arose from better SiO 2 dispersion as well as from a larger extent of in situ grafting. Meanwhile, the elongation at break depends on the incidental presence of agglomerates or defects and the ductility of the original materials. In this light, a long alkyl chain not only provided homogeneous dispersion, but also behaved as a plasticizer to enhance the ductility of the materials, leading to the recovery of the elongation at break [26,28]. With the improvement in the yield strength and the recovery of the elongation at break, the toughness of the resultant nanocomposites was greatly enhanced along with the alkyl chain length ( Table 3). The influence of the silane grafting density on mechanical properties of nanocomposites was also studied for C16 (Table 3). It was found that an excessive addition of the silane coupling agent rather deteriorated the mechanical properties. This was likely associated with the formation of a soft organic layer, which hampered the stress transfer to hard nanoparticles.

Conclusions
In this study, organically modified SiO 2 was exploited for the fabrication of grafttype polymer nanocomposites using polypropylene having reactive functional groups (PP-OTMS) as a matrix. It was shown that the surface modification of SiO 2 with silane coupling agents helped the dispersion of SiO 2 to promote the in situ grafting during melt compounding of the reactive matrix with nanofillers. The influences of the alkyl chain length of silane coupling agents and the grafting amount on the properties of resultant nanocomposites were also studied. It was found that the dispersion improved with the increase in the alkyl chain length due to the improved hydrophobicity of SiO 2 surfaces. This offered a more opportunity for the reactive groups at the PP side chain to react with the remaining surface silanol groups during melt mixing and thus to in situ graft onto nanofiller surfaces. This grafting reaction was confirmed by the significant reduction in the amount of the reactive functional groups and the observed crystallization acceleration in the resultant nanocomposites. The advantage of the combination of surface modification and in situ grafting was evidenced for the mechanical properties, in which the efficient grafting strengthened the interfacial interaction between the matrix and SiO 2 to improve the reinforcement. Furthermore, the modification with plasticizing alkyl chains also helped to recover the toughness deterioration caused by the grafting itself. This provided the opportunity to balance the reinforcement and the toughness of the materials, which is difficult to achieve by in situ grafting alone.