2.2. Preparation of filler

The rattan pole waste was first cleaned and then cut into small sizes using a knife. It was then dried in the oven at 100°C for 15 min to eliminate impurity and excessive moisture. Furthermore, milling and shaking were performed on the rattan pole waste, followed by sieving to an average particle size of 75 m (sieve 200 mesh). In the current study, a hammer mill with the rate of 25,000 rpm was used to produce particles of the nanometer range. Hammer milling is a mechanical operation used to reduce the particle size with the combined actions of collision, shaking, and milling. In this process, three time variations of 15, 30, and 45 min were explored and applied to obtain particle sizes <100 nm.

2.3. Preparation of nanocomposites

The designations for studied materials associated with their filler compositions are listed in Table 1. For the composite, polypropylene as the matrix was reinforced by rattan nanoparticles, which act as the filler, in the company of the coupling agent maleic anhydride (MA). Due to hydrophilic and hydrophobic properties of the nanofiller (rattan) and matrix (polypropylene), respectively, MA was employed to improve their interfacial interaction. This coupling agent had been proven useful in enhancing the physical-mechanical properties of composites (^{Dogu & Kaynak, 2016}; ^{Essabir, Bensalah, Rodrigue, Bouhfid, & Qaiss, 2016}; ^{Sombatsompop, Yotinwattanakumtorn, & Thongpin, 2005}; ^{Thwe & Liao, 2003}). Note that a good bonding can be formed by the reduction in the particle size as well, in the current case by the use of the nanometer size range, due to a greater surface area. The smallest size range of rattan nanoparticles obtained from the hammer mill process was selected as filler for the composite. The composition of the composite was 5% filler, 3% MA, and 92% matrix. As mentioned in the introduction, we focused on the incorporation of a small content of natural filler into polypropylene, in the current case 5%, due to its better mechanical properties as reported by the previous works (^{Ismail et al., 2002}; ^{Kim et al., 2008}; ^{Nikmatin et al., 2015}; ^{Tajvidi & Ebrahimi, 2003}). For comparison purpose, synthetic composites with the composition of 5% glass fiber were also fabricated (PP/FG5). All samples were fabricated with the twin extrusion molding machine at a temperature of 190°C, with a mixing speed of 45 rpm for 3 h of mixing time, before cooled to a temperature of 40°C.

2.4. Particle size analyzer

Examination of the particle size of rattan powder was done using the particle size analyzer (PSA), VASCO. VASCO is a dynamic light scattering measurement-based and fully agile nanoparticle size analyzer. For the examination, the rattan powder was put into the sample holder. It was performed for three sample types produced by 15, 30, and 45 min milling processes.

2.5. X-ray diffraction analysis

The X-ray diffraction diagrams were recorded by using the EMMA X-ray diffractometer, with operation capacities of 35 kV and 28 mA. The X-ray diffractometer was filtered with a radiation of 1.5A˚ in the 2θ range of 10°-60° with the steps of 0.02° at a scanning rate of 2° min^-1.

2.6. Morphological study

Examinations of the morphology and the composition of rattan fibers were carried out by means of Scanning Electron Microscopy (SEM) using JEOL JED model 2300 at a voltage of 20 kV. SEM micrographs were taken at 500 magnification. Sputter coated with gold was used on the fracture surfaces to prevent electrostatic charging and low image resolution during examinations.

2.7. Thermal property inspection

Thermal properties were analyzed by using the differential scanning calorimeter (model JADE, Perkin Elmer, USA). All the sample weights used in this study were in the range of ±7.9 mg. The samples were assessed with a thermal scan with temperatures from 26°C to 444°C at a heating rate of 10°Cmin^-1, under a nitrogen supply rate of 20ccmin^-1. In the process, heat flow versus temperature change was monitored.

2.8. Tensile test

Tensile tests were carried out at the room temperature of 23°C with a relative humidity of 61%. Dog bone-shaped specimens (see Fig. 2(a)) with overall dimensions of 16.3 × 1.9 × 0.4 cm were prepared. In the present study, a Shimadzu testing machine (EHF-EB20-40L) with a 50 kN load cell was employed. Tensile tests were performed by a crosshead pulling at a rate of 1mmmin^-1 until the failure of the specimens. Three samples were tested for each composite type. The tests were performed according to the American Society for Testing and Materials (ASTM) standard D638.

Fig. 2 Sample dimensions of composite for (a) tensile test and (b) bending and hardness tests. 

2.9. Bending test

Bending tests were also carried out at the room temperature of 23°C and a relative humidity of 61%. Three samples with dimensions of 12.6 × 1.2 × 0.4 cm as shown in Figure 2(b) were tested for each composite type according to the ASTM D790 standard. A Shimadzu testing machine (AG-20KNG) was used to measure the bending properties of the composites. All tests were carried out at a crosshead speed of 2.8mmmin^-1 until the failure of the specimen was reached.

2.10. Hardness test

Hardness tests were carried out using the Shimadzu testing machine (G21 series). The tests were performed using the Rockwell R method. Three samples with dimensions of 12.6 cm × 1.2 cm × 0.4 cm as shown in Figure 2(b) were examined for each composite. The samples were indented by using a steel sphere with a load of 100 kgf. The average of recorded data was reported as the Rockwell R hardness. The test procedure was conducted according to the ASTM E18 standard.

3.1. Sizing

Figure 3(a)-(c) show the distribution of rattan particle sizes obtained from particle size analyses. As shown in Figure 3(a), the distribution of rattan particle sizes is concentrated within the range of 51-489 nm, resulting from the 15 min milling process. There is a significant decrease in the rattan particle size from the initial 75 m to the nanoparticle size range. Hence, the milling period of 15 min can be considered as a good time span to reduce the rattan particle size. As shown in Figure 3(b) for the milling period of 30 min, there is a further substantial decrease in the rattan particle size when compared to that shown in Figure 3(a).

Fig. 3 Size distribution of rattan particles after (a) 15 min milling period, (b) 30 min milling period, (c) zoom in from (b), and (d) 45 min milling period. 

In general, the increase in the milling period provides more force for the breaking event. Consequently, more energy produced from the milling is transferred into the particles. As a result, more cavitation and breakage are induced by the energy for further reduction in the particle sizes. Since all materials have a maximum elasticity, or rather strength under certain deformation, the particles can rupture if the applied force exceeds this ultimate value. Following this logic, the particle sizes become smaller after a longer milling period.

Figure 3(d) shows the distribution of rattan particle size for the milling period of 45 min. Interestingly, the distribution exhibits an increase in particle sizes compared to those of 15 and 30 min. As shown in Figure 3(d), the distribution of rattan particle size spreads within the range of 257 to 6762 nm. From this observation, it is resolved that there exists a certain maximum period for the milling process for an optimal reduction of rattan particle size. ^{Nikmatin et al. (2015)} argued that increasing the milling period causes a rise in temperature in the holder, due to the accumulation of heat in the machine. After a certain duration, the temperature begins to play the role in assisting the recrystallization event. As a result, there exists a surge in particle sizes. Several studies had concluded that in the milling process, the milling period is a key factor affecting the resulted particle size (^{Chaisan, Yimnirun, & Ananta, 2009}; ^{Khamman, Chaisan, Yimnirun, & Ananta, 2007}; ^{Muniandy et al., 2012}). The variation in the rattan nanoparticle size distribution found in the present study is in agreement with the ^{Chaisan et al. (2009)} and ^{Khamman et al. (2007)} findings. This trend of size gain has been further confirmed by ^{Nikmatin et al. (2015)} who also found that smaller nanoparticle sizes were not obtained after a certain threshold of milling period was exceeded. The reverse pattern in particle size distribution such as that observed here is due to a phenomenon defined as agglomeration (^{Inkyo et al., 2006}). Therefore, we note the attractiveness to explore the optimal milling time for desired nanoparticle sizes in the future work.

3.2. X-ray diffraction (XRD) analysis

The existence of glass fiber and rattan nanoparticle fillers can change the internal crystal structures and therefore the crystallinity of a composite. Figure 4 shows XRD patterns of the neat PP versus PP with various fillers as the composites. We can see that the major crystalline peak of each profile occurs at around 2θ = 17.5°. It can be seen also that PP/R5 has the highest peak compared to others.

Fig. 4 X-ray diffractograms of neat PP and PP with various fillers. 

A lattice system presenting the microstructural makeup of a particular material can be described according to the relative lengths of the cell edges (a, b, and c) and the angles between them (α, β, and γ). We find that the neat PP is classified as orthorhombic symmetry with dimensions of a = 10.4, b = 4.64, c = 6.92, and α = β = γ = 90. In addition, it has a degree of crystallinity of 35%. PP/FG5 has a degree of crystallinity of 42%. PP/R5 is classified as monoclinic symmetry with dimensions of a = 7.87, b = 10.31, c = 10.13, α = γ = 90, and β = 122. It has a degree of crystallinity of 64%. Thus, it is clear that PP/R5 has the highest degree of crystallinity compared to PP and PP/FG5. ^{Talbott, Springer, and Berglund (1987)} stated that the mechanical properties of semicrystalline thermoplastics depend strongly on the crystalline structure of the polymer. This is the first sign indicating that the rattan nanoparticles demonstrate a constructive attribute useful to be employed as fillers to improve their composite mechanical properties compared to the glass fiber.

3.3. Morphological property

Figure 5(a) and (b) show the surface morphology of PP/FG5 and PP/R5, respectively. In general, it is clear that there is a significant difference in the surface roughness of the composites. As expected, PP/R5 exhibits a better surface morphology. Voids indicated by black dots in the SEM are lesser compared to composite filled by glass fiber. The surface morphology of a composite is a key factor affecting the physical, thermal, and mechanical properties of the composites. A good contact or bonding between filler and matrix can be affected due to the morphological makeup of the mixture. Its effects on the thermo-mechanical behaviors of the composite are remained to be examined and are the subjects of the next sections. As previously mentioned, this study used rattan nanoparticles with sizes ranging from 16 to 51 nm for the composite filler (see Fig. 3(c)). However, the surface morphology depicted in Figure 5(b) reveals there are particle sizes of around 20 m in a small part of the cured composite. This can be attributed to the agglomeration of nanoparticles that typically occurs in the blending process. ^{Mittal (2013)} stated that nanoparticles commonly agglomerate during and after processing, causing non-uniform dispersion within the polymer matrix.

Fig. 5 Scanning electron microscopy of (a) PP/FG5 and (b) PP/R5. 

3.4. Thermal property

Figure 6(a)-(c) show the differential scanning calorimetry (DSC) curves of PP and PP with various fillers. From these figures, all DSC graphs show that the composites have two melting endothermic peaks. In Figure 6(c), at the first melting endothermic peak, PP/R5 has the onset temperature (T_onset) at 121.44°C and the end temperature (T_end) at 131.39°C. In addition, at the second melting endothermic peak, PP has T_onset at 150.10°C and T_end at 168.24°C.

Fig. 6 Differential scanning calorimetry curves of (a) neat PP, (b) PP/FG5, and (c) PP/R5. 

We notice that, at the first and the second peaks, PP/R5 has the highest T_onset and T_end compared to PP-based composite with glass fiber. The possible reason for this observation is the existence of cellulose in rattan that acts as an efficient nucleating agent for the crystallization of PP and this consequently increases the crystallization rate of PP (^{Amash & Zugenmaier, 1998}; ^{Mi, Chen, & Guo, 1997}). These results are in agreement with those obtained by ^{Qiu, Zhang, Endo, and Hirotsu (2003)}. Their study reported that T_onset of polypropylene composite of high cellulose crystalline increases with the cellulose content from 0 to 60%. Glass fibers are known as the noncrystallizable component since they inhibit the crystal growth of PP.

3.5. Tensile test properties

Figure 7 illustrates the tensile stress-strain relationship of PP and its composites with various fillers. All curves behave somewhat similarly with an early linear portion, defining the slope that reduces after a certain strain level, indicating the material stiffness that softens when the failure load is approached. Extracting all major tensile properties from these plots, they are then offered in terms of average, upper, and lower bounds for tensile modulus, strength, and maximum elongation in Figures 8-10, respectively. It can be seen in Figure 8 that PP/FG5 has the highest average tensile modulus, which is followed by PP/R5 and PP.

Fig. 7 Tensile stress-strain curves for PP and PP with various fillers. 

Fig. 8 Tensile modulus of neat PP and PP with various fillers. 

Fig. 9 Tensile strength of neat PP and PP with various fillers. 

Fig. 10 Maximum elongation of neat PP and PP with various fillers. 

Although the tensile modulus of PP/R5 is found to be slightly lower than PP/FG5, it is still comparable to that of PP/FG5 and higher than PP. This suggests that, with the same composition of fillers, the tensile modulus of PP/R5 is similar to PP/FG5 and greater than PP. These results imply that the rattan nanoparticle fillers can be used as an alternative filler to improve the tensile modulus and to replace the glass fibers in the composite.

Figure 9 summarizes the tensile strength of composites with various fillers including that of neat PP. It is noticed again that PP/FG5 has the highest average tensile strength, and is followed by PP/R5 and PP sequentially. Although it is seen that, with the same filler content, glass fiber fillers produce a slightly higher strength than that of rattan nanoparticle, they are in general in the same order. Also, PP/R5 shows more consistent tensile strength compared with PP/FG5, its variation of which is somewhat larger. This observation offers motivation for our future work to improve the tensile strength with other fillers composition or chemical treatments, focusing chiefly on improving the bonding condition between fillers and matrix.

Figure 10 shows the maximum elongation of composites with various fillers and that of neat PP. The present study finds that PP/FG5 has the highest average elongation, and is followed by PP/R5 and PP. It is interesting to see that 5% glass fiber filler exhibits a greater maximum elongation compared to neat PP. Also, PP/R5 demonstrates almost identical maximum elongation versus that of PP/FG5. ^{Ismail et al. (2002)} reported that the elongation decreases with the rise in the natural filler loading in the natural rubber composites. It is noteworthy to realize that in the present study, nanometer size range was employed for our studied fillers, while ^{Ismail et al. (2002)} used micrometer sizes. These results imply that composite fillers of nanometer size are able to improve the elongation properties, especially when a small amount of nanoparticle is used.

3.6. Bending test properties

Figure 11 shows the force-deflection relationship of PP and its composites with various fillers. In addition, Figure 12 shows their bending modulus of elasticity (MOE). Similar linear initial slope and softening pattern can be seen for the bending curves as compared to those plotted in Figure 7 for tensile loading. PP/FG5 presents the highest average MOE, and is followed by PP/R5 and PP. PP/R5 exhibits an increase in MOE compared to PP due to its filler contribution. This result is in agreement with ^{Tajvidi and Ebrahimi (2003)}, who reported that the modulus of elasticity increases with the purified -cellulose, wastepaper fibers, and wood flour filler loadings of 15, 25, and 35%, respectively, when they were used to reinforce polypropylene.

Fig. 11 Bending force-deflection curves for PP and PP with various fillers. 

Fig. 12 MOE of neat PP and PP with various fillers. 

Figure 13 shows the modulus of rupture (MOR) of composites with various fillers and the neat PP. Again, PP/FG5 demonstrates the highest average MOR, and is followed by PP/R5 and PP. Incorporating rattan nanoparticle filler into polypropylene is found to increase the MOR better than that of PP alone. Our general observation is aligned with ^{Hassan and Nada (2001)}, who found that MOR decreases along with the increment of filler content of old newsprint (ONP) fiber as reinforcing component in the polyester composite. In addition, ^{Mirmehdi, Zeinaly, and Dabbagh (2014)} reported that MOR decreases with the increase in the palm wood flour filler contents of 45%, 60%, and 75% when incorporated into polyethylene. The present study used fillers of the nanometer size, while ^{Mirmehdi et al. (2014)} considered the particle sizes of 25-40 mesh. Therefore, the present study provides an alternative filler size to improve the MOR. Although MOR of PP/R5 is lower than that of PP/R5, it can still be considered as a good alternative due to its nature as a freely available material. Therefore, it offers a better economic advantage over the glass fiber filler.

Fig. 13 MOR of neat PP and PP with various fillers. 

Figure 14 shows the maximum strains of PP and its composites with various fillers. We can see that PP/R5 has the highest average maximum strain compared to PP and PP/FG5. This result suggests that PP/R5 offers a better ductility versus all considered materials and hence provides the best alternative if such property is of importance in applications. This is generally useful for applications that desire ductility or rather energy absorption as the required behavior of structure, for examples soundproofing and impact resisting materials, to name a few.

Fig. 14 Maximum strain of neat PP and PP with various fillers. 

3.7. Hardness properties

Table 2 lists the Rockwell hardness (HRR) obtained from the hardness test. We find that PP has the lowest average HRR amongst all materials studied. It is noteworthy to see that PP/R5 has the highest average HRR. This relates directly to the good surface morphology of the composites as shown in Figure 5(b). The existence of a good bonding between the rattan filler and matrix of the composite offers better mechanical properties in the current case the hardness. Since natural fillers (rattan, palm, and wood) and polymers (polypropylene and polyethylene) are incompatible for good bonding because of their hydrophilic and hydrophobic behaviors, respectively, several treatments have been implemented to improve the interfacial adhesion between these materials by the incorporation of a coupling agent such as maleic anhydride. This is in addition to the use of smaller size particles as a filler, which offers a greater surface area. Various works have reported the obvious enhancement in the biocomposites properties due to the use of the coupling agent, in terms of mechanical (^{Dogu & Kaynak, 2016}; ^{Sombatsompop et al., 2005}; ^{Thwe & Liao, 2003}) and physical behaviors (^{Essabir et al., 2016}). In general, the surface morphology and certain mechanical properties, for instance, the average hardness and maximum strain were enhanced, indicating an improved interaction (compatible) between the matrix and filler.

Another possible reason for the highest average HRR of PP/R5 is that it has the highest crystallinity. Several studies have reported that the crystallinity affects a great deal the composite mechanical properties (^{Mi et al., 1997}; ^{Talbott et al., 1987}), for instance, ^{Talbott et al. (1987)} observed that the tensile strength and tensile modulus of PEEK 150 polymer were increased with the rise in crystallinity.