*3.1. Fiber Characterization*

Well-known in the literature, piassava fibers, as seen in Figure 2, have oval-shaped cross-sections and a relatively high main diameter, as illustrated in Figure 2b. The surface morphology of the piassava fiber shows a rough surface and characteristic structures of spiny nodules made of pure silica (SiO2) that occur in some spherical holes on the fiber surface [43,44]. These SiO<sup>2</sup> nodules are bonded on the fiber surface by weak hydrogen bridges. During the industrial processing steps, some of these nodules are lost, leaving a hollow cavity that provides good anchoring points for the polymeric resin [45]. This is an important feature since the hydrophilic characteristics of the piassava fiber, like those of many other NLFs, tend to form a week interface with hydrophobic polymer resins [25–36], thus creating defects that can accelerate the biocomposite degradation. All these structures were observed on the as-received fiber and are shown in Figure 2a. These structural morphologies were expected since there was no chemical treatment of the piassava fiber at any step of the broom production processes.

at any step of the broom production processes.

at any step of the broom production processes.

*Sustainability* **2022**, *14*, x FOR PEER REVIEW 6 of 13

**Figure 2.** SEM photomicrography of the longitudinal surface of the piassava fiber (**a**) where the spiny silica spheres (indicated as "A") and some hollow holes (indicated as "B") and a cross-section of the piassava fiber (**b**) can be seen. **Figure 2.** SEM photomicrography of the longitudinal surface of the piassava fiber (**a**) where the spiny silica spheres (indicated as "A") and some hollow holes (indicated as "B") and a cross-section of the piassava fiber (**b**) can be seen. **Figure 2.** SEM photomicrography of the longitudinal surface of the piassava fiber (**a**) where the spiny silica spheres (indicated as "A") and some hollow holes (indicated as "B") and a cross-section of the piassava fiber (**b**) can be seen.

hollow cavity that provides good anchoring points for the polymeric resin [45]. This is an important feature since the hydrophilic characteristics of the piassava fiber, like those of many other NLFs, tend to form a week interface with hydrophobic polymer resins [25– 36], thus creating defects that can accelerate the biocomposite degradation. All these structures were observed on the as-received fiber and are shown in Figure 2a. These structural morphologies were expected since there was no chemical treatment of the piassava fiber

hollow cavity that provides good anchoring points for the polymeric resin [45]. This is an important feature since the hydrophilic characteristics of the piassava fiber, like those of many other NLFs, tend to form a week interface with hydrophobic polymer resins [25– 36], thus creating defects that can accelerate the biocomposite degradation. All these structures were observed on the as-received fiber and are shown in Figure 2a. These structural morphologies were expected since there was no chemical treatment of the piassava fiber

Regarding the mechanical behavior of the piassava fiber, the obtained tensile strength results of 87.9 ± 28.1 MPa and elastic modulus of 3.5 ± 1.2 GPa were similar to those reported in the literature [46], which corroborated the idea that, as expected, these fibers did not undergo significant changes during the broom manufacture process. Figure 3, obtained from SEM observations on the fracture surface, shows the characteristic rupture behavior of the piassava fibers. In this figure, the fiber cellulose microfibrils are ruptured in different regions, characterizing a non-uniform failure. Furthermore, Figure 3 displays a detailed image of the piassava fiber presenting columnar hollow lumen structures that justify its relatively low apparent density [39] and tensile strength with a great statistical dispersion. Regarding the mechanical behavior of the piassava fiber, the obtained tensile strength results of 87.9 ± 28.1 MPa and elastic modulus of 3.5 ± 1.2 GPa were similar to those reported in the literature [46], which corroborated the idea that, as expected, these fibers did not undergo significant changes during the broom manufacture process. Figure 3, obtained from SEM observations on the fracture surface, shows the characteristic rupture behavior of the piassava fibers. In this figure, the fiber cellulose microfibrils are ruptured in different regions, characterizing a non-uniform failure. Furthermore, Figure 3 displays a detailed image of the piassava fiber presenting columnar hollow lumen structures that justify its relatively low apparent density [39] and tensile strength with a great statistical dispersion. Regarding the mechanical behavior of the piassava fiber, the obtained tensile strength results of 87.9 ± 28.1 MPa and elastic modulus of 3.5 ± 1.2 GPa were similar to those reported in the literature [46], which corroborated the idea that, as expected, these fibers did not undergo significant changes during the broom manufacture process. Figure 3, obtained from SEM observations on the fracture surface, shows the characteristic rupture behavior of the piassava fibers. In this figure, the fiber cellulose microfibrils are ruptured in different regions, characterizing a non-uniform failure. Furthermore, Figure 3 displays a detailed image of the piassava fiber presenting columnar hollow lumen structures that justify its relatively low apparent density [39] and tensile strength with a great statistical dispersion.

**Figure 3.** Surface of the fractures of piassava fibers at lower (**a**,**c**) and greater (**b**,**d**) magnifications. **Figure 3. Figure 3.** Surface of the fractures of piassava fibers at lower ( Surface of the fractures of piassava fibers at lower (**aa**,,**cc**) and greater ( ) and greater ( **b b**,,**d d** ) magnifications. ) magnifications.

Regarding the FTIR analysis of the piassava powder in Figure 4, the typical cellulose functional groups will appear around 3400 cm−1—in this case, at 3317 cm−1—and are related to OH groups that are linked in the main cellulose chain by a hydrogen bridge [47]. According to the literature [47,48], for CH aliphatic stretches of the methyl and methylene groups, the vibrations will be around 2936 and 2920 cm−<sup>1</sup> , respectively, almost what was found for the piassava powder sample at the bands of 2930 and 2920 cm−<sup>1</sup> . Furthermore, a band at 2116 cm−<sup>1</sup> referred to a Si-H connection attributed to the spiny nodules on the fiber surface, already shown in Figure 2. This provided evidence that these structures

were present even after the broom manufacture processing. The bands at 1424, 1512 and 1608 cm−<sup>1</sup> were attributed to the vibration and elongations in the O-C-O of the aromatic ring present in the lignin. Finally, it was possible to identify the three distinct units that formed lignin molecules: guaiacyl (1278 cm−<sup>1</sup> ), syringyl (1332 cm−<sup>1</sup> ) and *p*-hydroxyphenyl (875 cm−<sup>1</sup> ). present even after the broom manufacture processing. The bands at 1424, 1512 and 1608 cm−1 were attributed to the vibration and elongations in the O-C-O of the aromatic ring present in the lignin. Finally, it was possible to identify the three distinct units that formed lignin molecules: guaiacyl (1278 cm−1), syringyl (1332 cm−1) and *p*-hydroxyphenyl (875 cm−1). present even after the broom manufacture processing. The bands at 1424, 1512 and 1608 cm−1 were attributed to the vibration and elongations in the O-C-O of the aromatic ring present in the lignin. Finally, it was possible to identify the three distinct units that formed lignin molecules: guaiacyl (1278 cm−1), syringyl (1332 cm−1) and *p*-hydroxyphenyl (875 cm−1).

Regarding the FTIR analysis of the piassava powder in Figure 4, the typical cellulose functional groups will appear around 3400 cm−1—in this case, at 3317 cm−1—and are related to OH groups that are linked in the main cellulose chain by a hydrogen bridge [47]. According to the literature [47,48], for CH aliphatic stretches of the methyl and methylene groups, the vibrations will be around 2936 and 2920 cm−1, respectively, almost what was found for the piassava powder sample at the bands of 2930 and 2920 cm−1. Furthermore, a band at 2116 cm−1 referred to a Si-H connection attributed to the spiny nodules on the fiber surface, already shown in Figure 2. This provided evidence that these structures were

Regarding the FTIR analysis of the piassava powder in Figure 4, the typical cellulose functional groups will appear around 3400 cm−1—in this case, at 3317 cm−1—and are related to OH groups that are linked in the main cellulose chain by a hydrogen bridge [47]. According to the literature [47,48], for CH aliphatic stretches of the methyl and methylene groups, the vibrations will be around 2936 and 2920 cm−1, respectively, almost what was found for the piassava powder sample at the bands of 2930 and 2920 cm−1. Furthermore, a band at 2116 cm−1 referred to a Si-H connection attributed to the spiny nodules on the fiber surface, already shown in Figure 2. This provided evidence that these structures were

*Sustainability* **2022**, *14*, x FOR PEER REVIEW 7 of 13

*Sustainability* **2022**, *14*, x FOR PEER REVIEW 7 of 13

**Figure 4.** FTIR spectra of the piassava powder. **Figure 4.** FTIR spectra of the piassava powder. **Figure 4.** FTIR spectra of the piassava powder.

#### *3.2. FTIR Analysis of the COPU Resin and Its Biocomposites 3.2. FTIR Analysis of the COPU Resin and Its Biocomposites 3.2. FTIR Analysis of the COPU Resin and Its Biocomposites*

Figure 5 shows the FTIR spectra of plain COPU along with its piassava powder-incorporated biocomposites. As can be seen, the spectra obtained for the COPU biocomposites presented almost the same bands and were very similar to the COPU resin. Figure 5 shows the FTIR spectra of plain COPU along with its piassava powderincorporated biocomposites. As can be seen, the spectra obtained for the COPU biocomposites presented almost the same bands and were very similar to the COPU resin. Figure 5 shows the FTIR spectra of plain COPU along with its piassava powder-incorporated biocomposites. As can be seen, the spectra obtained for the COPU biocomposites presented almost the same bands and were very similar to the COPU resin.

**Figure 5.** FTIR spectra for COPU and its composites reinforced by 10% (COPU + 10%P), 20% (COPU + 20%P) and 30% (COPU + 30%P) volume fraction of piassava powder. **Figure 5.** FTIR spectra for COPU and its composites reinforced by 10% (COPU + 10%P), 20% (COPU + 20%P) and 30% (COPU + 30%P) volume fraction of piassava powder. **Figure 5.** FTIR spectra for COPU and its composites reinforced by 10% (COPU + 10%P), 20% (COPU + 20%P) and 30% (COPU + 30%P) volume fraction of piassava powder.

It should be noticed in Figure 5 that the characteristic groups of the COPU resin were found, as reported by [49,50]. The bands at 3320, 2929 and 2850 cm−<sup>1</sup> corresponded, respectively, to amine (NH), methyl (CH3) and methylene (CH2). The 1728 cm−<sup>1</sup> band corresponded to the stretching of the pre-polymer carbonyl and, finally, the 1226 and 1053 bands corresponded to ether groups. The 2273 cm−<sup>1</sup> band was attributed to groups of free isocyanates (NCO) in the COPU resin. Its presence indicated the excess of the NCO group in the polymeric structure. Moreover, as the amount of piassava powder increased

in volume, the 3317 cm−<sup>1</sup> band (OH cellulose group) on the composites had its intensity diminished along with the 2273 cm−<sup>1</sup> band (free isocyanates in the COPU resin). Indeed, at 30% volume fraction of piassava powder, the 3317 cm−<sup>1</sup> band almost disappeared and the 2273 cm−<sup>1</sup> band had its intensity drastically lowered. ume, the 3317 cm−1 band (OH cellulose group) on the composites had its intensity diminished along with the 2273 cm−1 band (free isocyanates in the COPU resin). Indeed, at 30% volume fraction of piassava powder, the 3317 cm−1 band almost disappeared and the 2273 cm−1 band had its intensity drastically lowered.

It should be noticed in Figure 5 that the characteristic groups of the COPU resin were found, as reported by [49,50]. The bands at 3320, 2929 and 2850 cm−1 corresponded, respectively, to amine (NH), methyl (CH3) and methylene (CH2). The 1728 cm−1 band corresponded to the stretching of the pre-polymer carbonyl and, finally, the 1226 and 1053 bands corresponded to ether groups. The 2273 cm−1 band was attributed to groups of free isocyanates (NCO) in the COPU resin. Its presence indicated the excess of the NCO group in the polymeric structure. Moreover, as the amount of piassava powder increased in vol-

*Sustainability* **2022**, *14*, x FOR PEER REVIEW 8 of 13

The interaction between the cellulose and the free isocyanates was expected [50,51]. The FTIR results in Figures 4 and 5 suggested that, not only was the interaction occurring, but the level of interaction was also greater with higher amounts of piassava powder on the biocomposites, compared to lower volume fractions. The interaction between the cellulose and the free isocyanates was expected [50,51]. The FTIR results in Figures 4 and 5 suggested that, not only was the interaction occurring, but the level of interaction was also greater with higher amounts of piassava powder on the biocomposites, compared to lower volume fractions.

#### *3.3. Flexural Strength 3.3. Flexural Strength*

Figure 6 shows the flexural strength and flexural elastic modulus obtained for the COPU resin (0%), and the 10, 20 and 30 vol% piassava powder-reinforced biocomposites along with a statistical analysis of the data. It was noticed that, in all cases, none of the biocomposite materials was able to meet the standard requirements presented in Table 1. However, it was evident that the use of piassava powder significantly enhanced both the flexural strength and elastic modulus of the COPU matrix. Indeed, the 30 vol% reinforced composite presented a resistance more than four times and stiffness more than seven times greater than the plain COPU. Moreover, the results showed almost irrelevant data dispersion, which is not a common behavior for composites reinforced by NLFs. Figure 6 shows the flexural strength and flexural elastic modulus obtained for the COPU resin (0%), and the 10, 20 and 30 vol% piassava powder-reinforced biocomposites along with a statistical analysis of the data. It was noticed that, in all cases, none of the biocomposite materials was able to meet the standard requirements presented in Table 1. However, it was evident that the use of piassava powder significantly enhanced both the flexural strength and elastic modulus of the COPU matrix. Indeed, the 30 vol% reinforced composite presented a resistance more than four times and stiffness more than seven times greater than the plain COPU. Moreover, the results showed almost irrelevant data dispersion, which is not a common behavior for composites reinforced by NLFs.

**Figure 6.** Flexural strength and elastic modulus for COPU resin and its composites along with linear fit of the results. **Figure 6.** Flexural strength and elastic modulus for COPU resin and its composites along with linear fit of the results.

In fact, the COPU resin tended to present elastomeric characteristics, and the addition of reinforcement was enough, at least partially, to change this behavior [52]. Moreover, both good approximated linear fit (*r2* = 0.87871) and exponential fit (*r2* = 0.99679) indicated that the flexural strength and the stiffness, respectively, of the biocomposites could be further increased with higher amounts of piassava powder, enough to meet the standard criteria. In addition, the discussion of the FTIR results for the biocomposites indicated that not all the free NCO groups were completely bonded with the piassava cellulose OH, which indicated that more reinforcement could be added to the matrix without loss of strength of the interface between reinforcement and matrix. In contrast to what was observed in this work, piassava powder was reported to present a decrease of around 30% flexural resistance when incorporated with up to 30 vol% in other polymer matrixes [38], In fact, the COPU resin tended to present elastomeric characteristics, and the addition of reinforcement was enough, at least partially, to change this behavior [52]. Moreover, both good approximated linear fit (*r <sup>2</sup>* = 0.87871) and exponential fit (*r <sup>2</sup>* = 0.99679) indicated that the flexural strength and the stiffness, respectively, of the biocomposites could be further increased with higher amounts of piassava powder, enough to meet the standard criteria. In addition, the discussion of the FTIR results for the biocomposites indicated that not all the free NCO groups were completely bonded with the piassava cellulose OH, which indicated that more reinforcement could be added to the matrix without loss of strength of the interface between reinforcement and matrix. In contrast to what was observed in this work, piassava powder was reported to present a decrease of around 30% flexural resistance when incorporated with up to 30 vol% in other polymer matrixes [38], reaching around 30–35 MPa. This also suggested that the flexural strength of the composites can be enhanced, since the curves presented in Figure 6 showed a tendency to growth.

Figure 7 shows the fracture surface of the biocomposites. In this figure, we noticed the presence of bubbles on the matrix that were associated with existing retained gases during the polymer curing. On one hand, the use of the open mold tended to mimic the application of this composite as an HPCF. On the other hand, this also facilitated the appearance of defects such as the bubbles in Figure 7, which further explained the lower flexural strength than that required by the standard [10].

flexural strength than that required by the standard [10].

**Figure 7.** Surface of fracture for a 30% volume fraction reinforced by piassava powder COPU resin composites. **Figure 7.** Surface of fracture for a 30% volume fraction reinforced by piassava powder COPU resin composites.

reaching around 30–35 MPa. This also suggested that the flexural strength of the composites can be enhanced, since the curves presented in Figure 6 showed a tendency to

Figure 7 shows the fracture surface of the biocomposites. In this figure, we noticed the presence of bubbles on the matrix that were associated with existing retained gases during the polymer curing. On one hand, the use of the open mold tended to mimic the application of this composite as an HPCF. On the other hand, this also facilitated the appearance of defects such as the bubbles in Figure 7, which further explained the lower

Regarding the crack propagation, also evidenced in Figure 7, it occurred on the matrix, as expected, by the change of direction when in contact with the reinforcement. The fiber powder marks presented on the matrix indicated a relatively stronger interface between reinforcement and matrix, which is not common on this kind of material, as already discussed. Regarding the crack propagation, also evidenced in Figure 7, it occurred on the matrix, as expected, by the change of direction when in contact with the reinforcement. The fiber powder marks presented on the matrix indicated a relatively stronger interface between reinforcement and matrix, which is not common on this kind of material, as already discussed.

#### *3.4. Impact Resistance 3.4. Impact Resistance*

growth.

As seen in Table 2, the addition of piassava powder to the COPU matrix tended to decrease the Izod impact resistance of the biocomposites. However, one may notice that, owing to greater data dispersion, almost all the conditions studied were statistically equal. Actually, with ANOVA tests, *p*-values of 0.0321 and 0.210 were obtained for notch resistance and impact resistance, respectively, which indicated that there was a statistical difference in at least one of the analyzed conditions for notch resistance and none for the impact resistance. In fact, the honest significant difference (HSD) of 14.59 was obtained on the Tukey's test for the notch resistance. This meant that only the 30 vol% reinforced COPU resin presented a significant HSD compared to the plain resin. As seen in Table 2, the addition of piassava powder to the COPU matrix tended to decrease the Izod impact resistance of the biocomposites. However, one may notice that, owing to greater data dispersion, almost all the conditions studied were statistically equal. Actually, with ANOVA tests, *p*-values of 0.0321 and 0.210 were obtained for notch resistance and impact resistance, respectively, which indicated that there was a statistical difference in at least one of the analyzed conditions for notch resistance and none for the impact resistance. In fact, the honest significant difference (HSD) of 14.59 was obtained on the Tukey's test for the notch resistance. This meant that only the 30 vol% reinforced COPU resin presented a significant HSD compared to the plain resin.

 **Table 2.** COPU resin (0%) and its biocomposites' notch resistance and impact resistance obtained by Izod tests.

