Tóm tắt Luận án Effect of surface fillers modification on phase interaction and mechanical properties of thermoplastic materials

tion frequency of 62.5 kHz, a contact time of 1.75 ms and a recycle delay of 2 s. To record the T1-filtered, domain-selective spectra of the highly mobile components, the single-pulse 13C MAS NMR spectrawere measured with a repetition delay of 2 s. The VT experiments were performed in the temperature range from 295 to 355 K. The frictional heating of the sample was compensated, and the sample temperature was calibrated using the 207Pb chemical shift in Pb(NO3)2. 2.3.5. Factor analysis (FA) FA is a method that uses the Singular Value Decomposition (SVD) algorithm to extract specific information from experimental data. This technique allows visualization and distinguishing subtle differences between the prepared PP/CaCO3 microcomposites in 2D or 3D maps. Chapter 3: RESULS AND DISCUSION 3.1. Composite based on polypropylene and glass beads 3.1.1. Morphology observation The introduction of inorganic filler into a polymer matrix results in a heterogeneous system. Adhesion among different materials is created by physical or chemical bonds between the adhesive and the substrate, and this depends on the selection of coupling agent. Figure 3.1 presents structure morphology taken of the impact fracture cross section of PP/glass beadcomposites with 20% of filler content. a) b) c) Fig. 3.1. Degree of interfacial adhesion between glass bead and PP matrix. As revealed by SEM in the cases of non-treated and NO adhesion, there was a poor interfacial with the strong debonding of particles. While in case of GOOD adhesion, a strong bonding achieved between glass bead particles and PP matrix, coated spheres adhere to the matrix. 3.1.2. Tensile properties The effects of glass bead with different surface properties on the mechanical properties of composites are showed in Fig. 3.2. It can be sheen that the tensile moduli in all cases of composites 6 increase with increasing filler loading. Generally, the addition of rigid particulate fillers increases stiffness, which is measured through Young's modulus. This is due to the fact that fillers often exhibit higher stiffness compared matrix polymer. Besides, Young's modulus is measured at the very beginning of a tensile test, where deformation is insufficient to cause particle-matrix debonding. However, obtained results indicated slightly change in value among cases of no-treated, NO adhesion and GOOD adhesion. This comes from the difference in adhesion between particles and polymer matrix with zero adhesion in case of NO adhesion and a strong adhesion in case of GOOD adhesion. The increasing in Young's modulus of glass bead-filled composites indicates an increase in the rigidity of PP related to the restriction of the mobility in PP matrix due to the presence of fillers. This mechanical restraint resulted from the enhanced surface interaction between two phases in composites. The similar trend also found in other studies that enhancement of the interfacial adhesion between the matrix and glass beads is helpful in improving the stiffness of filled PP composites. Fig. 3.2. Tensile properties of PP/Glass beads composites For particulate filled thermoplastic composites, it is generally believed that the interfacial adhesion between the filler particles and matrix is an important factor affecting strength and toughness of composites. On the other hand, yield stress gives information on filler-matrix interactions and consequently it is one of the preferred methods of composite testing. In thecase of a poor interaction (non- treated and NO adhesion) between the matrixand the filler, the interfacial layer cannot transferstresswas reflected by the lower yield stress value in comparison with a strong interaction in case of GOOD adhesion (Fig. 3.2). Therefore, one can assume that the strengthof a particulate-filled composite is determined by theeffective available area of load borne by the matrix asa result of the absence of the filler. The higher yield stress values in case of GOOD adhesion in comparison with cases of non-treated and NO adhesion at the corresponding concentration reflect the interaction between glass bead particles and polypropylene matrix. On the other hand, it can be seen that in all cases, yield stress decreases with increasing filler loading.This is due to the fact that theconcentration of the inclusions is the main factoraffecting the yield strength of a filled polymer besidesthe interfacial adhesion between the fillers and matrix. The presence of glass beads has a weakening effect on the composite due to debonding. Poor adhesion and debonding reduce the volume fraction which can carry the applied load.Tensile strain at break is a parameter characterizing the extensibility of materialsand it is usually inversely proportional to tensile strength which means that increasing the tensile strength of filled material usually contributes to a decrease in the strain at break. The strain at break value for composite of GOOD adhesion had a lowest strain at break value which corresponds with highest tensile strength value. 7 3.1.3. Impact properties The J-integral was used to characterize the energy absorption of polymer materials at the initial stage of crack and the resistance for crack initiation. Figure 3.3 showed Jld values of polypropylene and glass bead composites. Fig. 3.3. Impact strength of PP/glass beads composites The fracture behavior of polymers is strongly affected by the addition of rigid particles. The incorporation of them into the polypropylene matrix leads to differences in the overall process of crack propagation and fracture. The process starts with the plastic deformation of the matrix ahead of the initial crack. The adsorption of polymer molecules on the filler surface through chemical bonds leads to the rigidity in structure of polymer chains. This leads to earlier crack initiation and propagation with dramatically decreasing of J-integral value in case of GOOD adhesion (Figure 3.3). Plastic deformation of the matrix polymer is the main energy absorbing process in impact and this increases when the interaction between particles and polymer matrix is lowered in case of non-treated and NO adhesion. 3.1.4. ss-NMR In general, narrowing of 1H lineshapes indicates the increase in global segmental dynamics as 1H- 1H dipolar couplings are motionally averaged out. As demonstrated in Figure 3.4 in our particular case, the most mobile polymers segments were found in parent PP. Slightly hindered segmental dynamics in the composites Non-treated and NO-adhesion is reflected by the broadening of the corresponding signals. The most rigid polymer segments were found in the composite GOOD-adhesion that is reflected by the broadest 1H MAS NMR line 8 Fig. 3.4. 1H MAS NMR spectra of the parent PP and composites recorded at 350 K and MAS 6 kHz. Similar trends were found in the 1H MAS NMR spectra measured with Hahn-echo (Figure 3.5) that allowed separation of relatively flexible fractions of polymer chains of amorphous phase. Hình 3.5. 1H MAS Hahn-echo NMR spectra of the parent PP and composites recorded at 350 K and MAS 6 kHz In contrast to solid-state 1H NMR spectra it is generally accepted that narrowing of 13C CP/MAS NMR signals indicates the increase in conformational ordering, and in extreme cases of perfectly evolved crystal phases very narrow signals are expected. In polypropylene systems presence of polymorphic form alpha is reflected by the doublet of CH signal at ca. 44 ppm. The doublet is schematically indicated by dashed lines in Figure 3.6, and evolution of the narrow symmetric doublet requires long term annealing. If considerable amount of amorphous phase is presented than the expected symmetrical doublet disappears. PP Non-treated and NO adhesion GOOD-adhesion PP and NO adhesion Non-treated GOOD-adhesion 9 Rather an asymmetric broad line is expected. Such a signal was detected for the parent PP (Figure 3.6). In contrast, the prepared composites exhibit symmetrization of the doublet and increase in spectral resolution. This finding can be explained by the conformational ordering occurred particularly in the composite NO-adhesion (additional crystallization, reorganization of the amorphous phase, formation of protocrystalline phases etc.). Fig. 3.6. 13C CP/ MAS NMR spectra of the parent PP and composites recorded at 300 K and MAS 6 kHz. 3.3. Composite based on polypropylene and calcium carbonate 3.3.1. Morphology observation The introduction of inorganic filler into a polymer matrix results in a heterogeneous system. Adhesion between different materials is created by physical or chemical bonds between the adhesive and the substrate, and this depends on the selection of coupling agent. The morphology of the compounds up to concentration of 40 wt% treated and untreated CaCO3 are shown in Figure 3.7 and Figure 3.8. The CaCO3 particles are generally supplied as agglomerates, however, it can be seen that aggregates are broken up to the primary particles during the extrusion process. In the case of CaCO3 with average particle size ~ 1.7 µm, the interaction between untreated filler particles with the polymer matrix is stronger compared with treated filler particles (Figure 1a). The strong debonding of particles can be observed in cases of treated particles (Figure 3.7-b and 3.7-c). Figure 3.7. SEM images of morphology of PP filled with a) untreated CaCO3, b) oleic acid treated CaCO3 and c) stearic acid treated CaCO3, average particle size ~ 1.7 µm, filler content 40 wt% a ) b ) c ) PP Non-treated NO-adhesion GOOD-adhesion 10 Similar trend was found in case of CaCO3 with higher size (12 µm, Figure 2) of filler particles, filler-matrix interaction in case of treated filler is lower compared with untreated filler. Figure 3.8. SEM images of morphology of PP filled with a) untreated CaCO3, b) oleic acid treated CaCO3 and c) stearic acid treated CaCO3, average particle size ~12 µm, filler content 40% wt. 3.3.2. Tensile properties Tensile mechanical properties of composites of PP matrix with calcium carbonate particles were investigated (Figure 3.9). Generally, the addition of rigid particulate fillers increases stiffness, which is measured through Young modulus. As can be seen in the figure, all the samples investigated in the present study show monotonic increasing of Young modulus values. This is due to the fact that fillers often exhibit higher stiffness compared with polymer matrix. Furthermore, Young modulus is measured at the very beginning of a tensile test, where deformation is insufficient to cause particle-matrix debonding. Experimental results also indicated that, Young modulus of composites have none effect by particle size and surface modification. Strength and toughness are very important for polymer composites used as structural materials. For particulate filled thermoplastic composites, it is generally believed that the interfacial adhesion between the filler particles and matrix is an important factor affecting strength and toughness of composites. As revealed by SEM, the lower adhesion in case of treated filler led to a decrease in yield stress of composites compared with untreated filler for both particle size, especially at filler content of 40 wt%. In addition to the decreasing yield stress value as the filler content increases, the smaller of the particle size leading to the higher of value of tensile yield stress. On the other hand, surface treatment reduced particle- particle interaction, therefore, reduced the aggregation of filler particles, resulting a better dispersion of particle filler in polymer matrix, increasing interface leading to increasing yield stress. The extent of plastic deformation characterized by strain at break is then very sensitive to composition and morphology of the composites. As shown in Figure 3.9, various types of calcium carbonate fillers have distinctly different effects in strain at break values. a ) b ) c ) 11 Figure 3.9. Effects of particle size, surface modification and filler loading on the tensile mechanical behaviour of injection-moulded specimens. 3.3.3. Độ bền va đập J-integral was originally defined by Rice as a contour integral independent on the path, which express the energy per unit area necessary to create new fracture surfaces in a loaded body containing a crack. Figure 4 show J-integral values of PP/CaCO3 composites with different types of particle size, filler loading and surface properties. 12 Figure 3.10. Fracture behaviour of PP/CaCO3 composites. The addition of calcium carbonate led to a change in the fracture mode from brittle fracture for virgin PP to ductile fracture for the filled composites, which was attributed to the changes of stress fields in the PP matrix around the filler particles. For the unfilled PP, when the notched composite specimen receives an impact load, the crazes will be rapidly developed cracks, and then will propagate toward the whole cross section. The incorporation of rigid particles into the polypropylene matrix leads to differences in the overall process of crack propagation and fracture. The process starts with the plastic deformation of the matrix ahead of the initial crack. Assuming a poor bonding between filler and matrix in cases of treated calcium carbonate, the filler particles detach easily from the matrix by creating voids as observed by SEM. These voids be formed will absorb impact deformation energy. Otherwise, the particle will block the propagation of the crack developed from the crazes to increase the fracture resistance. Jilken et al. [8] also observed that the high impact strength, at high filler content, could be obtained for low aspect ratio filler, such as dolomite and calcium carbonate. And this only obtains for sufficiently fine fraction and if the filler particles are good dispersion. That is why J-integral values for composites of small particle size higher than those of large particle size. The results in Figure 3.10 also showed that fracture resistance was increased up to 20 wt% of calcium carbonate. 3.2.4. ss-NMR Two types of VT ss-NMR experiments were performed in a temperature range from 295 to 355 K. The 13C CP/MAS NMR spectra preferentially detect rigid molecular segments. When the 13C CP/MAS NMR spectra were acquired at 295 K, three signals corresponding to the CH2 (44 ppm), CH (26 ppm) and CH3 (21.3 ppm) groups were easily distinguished. 13 Figure 3.11. 13C CP/MAS NMR spectra at 295 K and 355 K of the pure PP and modified PP. Figure 3.12. 13C MAS NMR spectra at 295 K and 355 K of pure PP and modified PP. At elevated temperatures (far above Tg), the segmental motion in the amorphous phase increased, as indicated by the separate signals corresponding to the free r.c.amorphous phase that appeared in the high-frequency region (Fig. 3.11). The recorded spectra clearly identified the disordered α2 polymorphic form of isotactic polypropylene. However, the corresponding 13C MAS NMR spectra measured with a short repetition delay (1-2 s) reflected a preference for the highly mobile fractions. Consequently, the signals of the CH and CH2 groups were suppressed, and at high temperatures, the signals for the polymer segments in the amorphous r.c. conformation were relatively enhanced (Figure 3.12). Unfortunately, although the prepared PP/CaCO3 systems considerably differed in their mechanical properties, a visual inspection of the recorded 13C CP/MAS NMR and 13C MAS NMR spectra did not reveal any systematic spectral changes or features that reflected the plastic deformation. In a subsequent step, we determined the amount of the highly mobile r.c. amorphous fraction in the prepared composite materials using the 13C MAS NMR spectra measured at 355 K. As shown in Fig. 3.13, the amount of the highly mobile amorphous r.c. fraction systematically decreased as the critical value of the J-integral increased, which represents a measure of the composite. The neat, unfilled PP surprisingly had a dependence in the middle of the other dependences, and regardless of the amount of filler particles the composite systems that were modified by larger particles (12 mm) nearly exclusively exhibited a lower J-integral critical value in comparison with the systems that were modified by smaller particles (1.7 mm). 14 Figure 3.13. Graph of the toughness expressed as the J-integral (JId) critical value and the contents of the highly mobile amorphous fractions in the PP systems calculated from the integral intensities of the CH signal in the 13C MAS NMR spectra. The maximum integral intensity value was calibrated to the maximum content (100%) of the highly mobile fraction. The maximum J-integral value was normalized to 100%. To find other consistent trends and characteristic spectral features related to the changes in the mechanical properties of the samples, we used FA to extract the key information from the relatively large experimental data sets that were obtained. The recorded 13C CP/MAS NMR and/or 13C MAS NMR spectra were separately analyzed in sets representing each temperature. Subsequently, the FA generated subspectra, Sj, and the corresponding singular values, wj, for each set of NMR spectra. Because only the singular values w1 - w4 were significantly high (Fig. 3.14, left panel), the spectral variation in the analyzed dataset was completely described by the corresponding subspectra, S1eS4 Sau đó, FA tạo ra các phổ thế, Sj và các giá trị bất thường tương ứng, wj cho mỗi phổ NMR. Do chỉ các giá trị đơn w1 – w4 ở mức cao đáng kể (hình 3.26), biến thể phổ trong tập dữ liệu được phân tích đã được mô tả hoàn toàn bởi phổ thế tương ứng, S1 – S4. Fig. 3.14. Singular value calculated from the set of 13C MAS NMR spectra at 355 K (on the left) and the relevant subspectra S1, S2 and S4 used for the FA (on the right). The subspectra reflecting the changes in the amorphous (am.) phase and crystalline (cryst.) phase in the CH3 region The first-rank subspectrum, S1, corresponds to the average 13C MAS NMR spectrum for the superposition of the experimental data recorded for all the prepared samples at a given temperature, and the second-order subspectrum, S2, shows the most significant spectral differences found in the analyzed : Jid [N/m] is calibrated to [%] : Highly-mobile fraction of CH % 15 data set. In the subspectrum S2, all the components are equally highlighted, including the highly mobile amorphous and crystalline fractions. Specifically, for the CH3 signal, the amorphous phase was characterized by a positive signal at approximately 22 ppm, and the crystalline phasewas characterized by a negative signal at approximately 21 ppm. In addition, we determined from the analysis of the calculated subspectra that the most significant differences between the samples were in the CH3 regions of the subspectra S1, S2, and S4 (Fig. 3.26). The observed differences were subsequently quantified using the corresponding normalized coefficients, Vi1, Vi2 and Vi4. These parameters and their correlation plots were used to find the relationship between the spectroscopically specific features and the changes in the physicochemical and mechanical properties of the analyzed composites. Consequently, by applying this strategy, especially for the 13C MAS NMR spectra obtained at 355 K with increased segmental motion, we saw clear clustering of the Vij parameters into two groups that were separated by a vertical line, which is visualized in the 3D correlation plot of the relevant parameters, Vi1, Vi2 and Vi4 (Fig. 15 and Fig. 3.16). These groups correspond to composite systems that differ in their: (i) highly mobile amorphous fraction content; (ii) filler particle size; (iii) toughening effects; and (iv) PP/filler aspect ratio. The correlation plot in Fig. 5 shows the two groups of prepared microcomposites that differ in the J-integral critical value. The group on the left represents systems with an average JId value of 3.5 N mm-1 (3.2-3.8 N mm-1), and the cluster on the right represents systems with an average JId value of 4.8 N mm-1. The correlation plots for the other characteristics are shown in Fig. 3.17. Fig. 3.15. 3D correlation plots for the Vi1, Vi2 and Vi4 coefficients evaluated from the 13C MAS NMR data using FA for the full spectra Fig. 3.16. 3D correlation plots for the Vi1, Vi2 and Vi4 coefficients evaluated from the 13C MAS NMR data using FA for the single CH3 signals at 355 K. 16 Fig. 3.17. 3D correlation plots of the Vi1, Vi2 and Vi4 coefficients evaluated from the 13C MAS NMR data at 355 K using FA. The FA results also give information on the addition of fatty acids into the PP/CaCO3 microcomposites. In this case, at high temperatures (330-355 K), clustering of the systems based on the surface treatment of the CaCO3 particles was not observed. However, at lower temperatures (325 K), clustering based on the surface treatment, i.e., with OA or SA and without an acid treatment, was observed (Fig. 3.18). This clustering was likely revealed by the lower temperatures because the amorphous phase segmental motion in the systems is not completely free, and the full differences in the chain mobility and the trans-crystalline layer are not yet well resolved. Therefore, the samples are clustered based on the subtle differences in the system homogeneity, which is influenced by the non- treated or treated regimes of the two fatty acids. 17 Fig. 3.18. 3D correlation plot of the Vi1, Vi2 and Vi4 coefficients evaluated from the 13C MAS NMR data at 325 K using FA. The plot shows the PP samples based on the fatty acid treatment and the corresponding 13C MAS NMR spectra of the grouped systems and the pure PP sample. When searching for the molecular origin of the toughness differences in the prepared PP/CaCO3 microcomposites, we determined that the systems with a higher toughness (higher JId) mostly contained filler particles with a smaller diameter (1.7 mm), smaller amounts of these particles (80/20), and smaller amounts of the highly mobile r.c. amorphous fraction based on the 13C MAS NMR spectra at 355 K. The key role of the amorphous phase in understanding the toughness of the PP/CaCO3 composites was further evidenced by the FA, which highlighted the prominent spectral differences between the prepared composite systems. In particular, the subspectrum S4 unambiguously identified the signal at 22 ppm as corresponding to the methyl groups of the amorphous, free polymer chains in the r.c. conformation. The crystallinity of the prepared systems was determined by wide angle X-ray scattering experiments to be approximately 50%, and the crystallinity did not show a systematic trend with respect to the observed changes in the composite toughness. Therefore, the observed decrease in the free r.c. amorphous fraction amount must be compensated by an increase in the amount of amorphous PP chains with a helical conformation in the confined domains. The PP chains that adopt a helical conformation can be attributed to the transcrystalline phase described in the literature. In general, the formation of this trans-crystalline phase is induced: i) by the adsorption of polymer chains onto the active sites of the filler particles; ii) by a suitable orientation of the backbone bonds in the polymer toward the solid surface; and iii) by the hindered chain mobility, which affects the crystallization kinetics and induces the formation of imperfect crystallites. These imperfect crystallites are thought to undergo a larger plastic deformation in the system than the perfect crystallites, and this deformation increases the composite toughness. Our experimental findings indicate that the key mechanism that is responsible for the enhanced toughness in the PP/CaCO3 composites is related to the PP chain mobility. The mobility is induced by the presence of filler particles, and these particles do not allow perfect crystallites to form and promote the formation of partially ordered amorphous and trans-crystalline domains near the filler surfaces to dissipate the deformation energy. Smaller filler particles with larger specific surface areas offer more active sites for polymer adhesion onto the solid surface, which consequently results in a larger decrease in the polymer chain motion, the creation of imperfect crystallites and increased toughness. However, systematic trends in the 13C NMR parameters were observed only at high temperatures. Although