3.1. Flow Properties
The melt flow rates (MFRs) of the prepared samples are shown in
Table 2. PLA shows a higher MFR, implying lower viscosity and ease of processability. The addition of DCP and Joncryl individually decreased the MFR of PLA by 32.3 and 50.2%, respectively. A decrease in the MFR of PLA/DCP could be attributed to chain–chain coupling, which results from the interaction of radicals generated by DCP on PLA chains, as shown in
Supplementary Figure S1. On the other hand, Joncryl as a chain extender could have increased the molecular weight and viscosity of PLA due to the reaction between the epoxide groups of Joncryl and the carboxylic groups of PLA (see
Supplementary Figure S2) [
25]. The FTIR spectroscopy shows the reactions between the chain extenders, and will be discussed in the FTIR section. The simultaneous addition of both Joncryl and DCP did not significantly affect the MFR of PLA/DCP/J compared to the PLA/J system. Upon the addition of BA at 2 and 3 wt.% to PLA/DCP/J, the MFR decreased slightly, by 12.7 and 8.5%, respectively, suggesting the immobilization of PLA chains by the nanoparticles. However, with further increase in BA particles, the MFR started to increase, probably due to the separation and weakening of PLA chains by the rigid BA particles.
DCP and Joncryl have a significant influence on the structural properties of polymers. In particular, the molecular weight (M
W) and distribution (MWD) can be affected due to chain scission or chain–chain coupling in the presence of DCP. In addition, chain extenders such as Joncryl can also modify the structural properties of a polymer. Rheology is a powerful tool to elucidate the changes in the molecular structures of polymers in melt states.
Figure 2 shows the plots of G′ and G″ against angular frequency from 0.1 to 100 rad/s. In these plots, the crossover point between G′ and G″ provides information about the changes in the M
W and MWD. The horizontal shift of the crossover frequency (G
ω) to lower values is related to an increase in M
W. In contrast, the vertical shift of the crossover modulus (G
m) indicates an increased broadening of the MWD. It is worth mentioning that the crossover point is determined within the tested range (0.1 to 100 rad/s). For neat PLA (
Figure 2a), the crossover point could not be determined within the tested range. However, looking at the distance between G′ and G″ at 100 rad/s, it can be noticed that the G′ and G″ curves are closer, suggesting lower angular frequency for the PLA/DCP (
Figure 2b) system compared to neat PLA.
Moreover, a dramatic decrease in G
ω was noticed in the PLA/J system (
Figure 2c), indicating an increase in the M
W of PLA. A horizontal shift to higher values could be noticed in the PLA/DCP/J system (
Figure 2d); this suggests that there was a reduction in the M
W of PLA, which could be attributed to chain scission when both DCP and Joncryl were added. The incorporation of BA did not significantly change the M
W of the PLA/DCP/J system (
Figure 2e–i), as can be seen from the distance between G′ and G″ at 100 rad/s.
Figure 3a shows the complex viscosity of the prepared PLA/DCP/J-based composites against angular frequency. As evident from
Figure 3, the flow behavior of PLA, PLA/DCP, PLA/J, and PLA/DCP/J followed a similar trend to that noted from the MFR analysis. However, PLA/J showed the highest viscosity due to an increase in the M
W when Joncryl was introduced to PLA. This observation is attributable to the branches formed by the Joncryl and the introduction of long-chain branching (LCB) in the PLA structure, conveying pseudosolid-like behavior. The viscosity of the composites was solely dependent on the distribution of BA particles in the PLA matrix. The BA3 system showed better distribution, and exhibited a higher viscosity than the other composites (
Figure 3b). With the increase in filler concentration, the viscosity decreased, possibly due to the poor distribution and agglomeration of BA particles, forming the weak points in the matrix. Further increase in BA concentrations resulted in an increase in viscosity, due to the reinforcing effect of the particles.
3.2. Chain Extension
FTIR spectroscopy was used to compare the processed PLA/BA chain extender and crosslinking systems, in order to identify the reaction that may have transpired between the PLA/BA and chain extender/crosslinking.
Figure 4a shows the FTIR spectra of neat PLA and composites. In the case of neat PLA, the peaks at 1754, 1455, and 1369 cm
−1 are due to C=O stretching, C–H deformation, and C–O–H bands, respectively. Meanwhile, 1186 and 1080 cm
−1 are assigned to –C–O stretching, 868 cm
−1 is attributed to –C–C stretching, and 755 cm
−1 is attributed to C–H bending. The H–O–H peak at 1630 cm
−1 was not obtainable from the processed PLA, due to the existence of thermal chain scission at the C–O bond [
26].
Supplementary Figure S3 shows the vibrations at 2852 and 3000 cm
−1 assigned to the O–H stretching, and 2922 cm
−1 due to the axial C–H stretching bond [
27]. The FTIR spectra of neat BA (
Supplementary Figure S4a) reveal that the vibrations at 3309 and 3095 cm
−1 relate to the O–H stretching of BA [
28]. The vibration at 1397 cm
−1 is attributed to the amorphous surface structure that exists in crystalline BA [
29]. In the case of neat DCP, as shown in
Supplementary Figure S4b, vibrations are observed at 1727 cm
−1 due to the C=O stretching, at 910 cm
−1 due to C=C stretching, and at 762 cm
−1 and 698 cm
−1 due to C–H bending.
Supplementary Figure S4c shows that in the FTIR spectra for neat Joncryl, vibrations are observed at 907 and 843 cm
−1, attributed to the symmetric and asymmetric ring deformation of cyclic epoxide [
30,
31].
Figure 4e shows that there is no significant change to the PLA matrix upon addition of BA, due to the low concentration of BA embedded in the PLA matrix. Further, upon addition of DCP to the PLA/BA composite, as shown in
Figure 4f, DCP undergoes homolytic cleavage when heat is applied, breaking down into free radicals and assisting in the removal of H’s from the PLA chains in order to create free radicals on the backbone structure of PLA, as shown in
Supplementary Figure S1. This phenomenon is attributed to the propagation of the radical reaction to form a crosslinked/branched structure of PLA [
13]. Therefore, PLA produces free radicals on the tertiary C atoms, which become stabilized in reactive extrusion [
4]. In the case of Joncryl, the 907 and 843 cm
−1 peaks disappear, owing to the interaction of the epoxy groups with carboxyl groups on the PLA, suggesting that the reaction occurrs between the Joncryl epoxy and the PLA terminal functional group [
32]. Surprisingly, there is a synergistic effect of peak decreases for Joncryl at 2922 cm
−1. These results relate to the decrease observed in the MFR results and the peak decrease shown in
Supplementary Figure S2. Upon addition of all the components, the peak at 2922 cm
−1 increases, suggesting that the initiator has created macroradicals after the chain extender has truly extended the PLA chain for the BA attachment. In all of the composites, vibrations at 1752, 1455, 1186, 1080, and 868 cm
−1 remain unchanged. Inata and Matsumura [
26] reported that the epoxides might react with carboxyl and hydroxyl end groups of polyesters, and the electrophilic group with the carboxyl end groups. It can be concluded that the chain extension/crosslinked/branched structures in the polymer composites play an important role in improving the properties of the reactive composites in a controlled way.
3.3. BA Distribution
To measure the distribution of BA particles in the PLA matrix, samples containing different BA concentrations were cryosectioned and viewed under TEM.
Figure 5a illustrates that the particles of BA formed more agglomerates in the PLA matrix, and that the particles were not well distributed.
Figure 5b shows that the addition of the initiator in the PLA matrix decreased the agglomerations of BA particles and distributed the particles better than in PLA/BA. This is attributed to the fact that the viscosity of the PLA matrix was increased by chain extension and/or branching, which assisted in breaking the BA agglomerates. On the other hand, better distribution of BA particles was observed in the presence of Joncryl, due to more chain branching that was created in the PLA structure. Furthermore,
Figure 5d shows that the addition of all components at once produced a fair distribution of BA particles and strong intercomponent bonding. This observation correlates with the FTIR spectroscopy results in the next section, which show the interfacial bonding between all components.
Furthermore, it is evident from
Figure 5e that upon the addition of 3 wt.% BA to the PLA matrix, the finest distribution was observed, showing an optimal distribution amongst all composites, owing to better intercomponent bonding amongst the neat PLA and BA particles. Das et al. [
2] reported similar results in PLA/BA, revealing that the 3 wt.% BA loading in the PLA matrix was the optimal distribution. This led to improved mechanical properties, which were affected by the distribution of the BA particles in the PLA composites. On the other hand, the increase in BA concentration (i.e., 4, 5, 6, 10, and 20 wt.%) produced poorer distribution and more agglomerations of the BA particles in the PLA matrix.
3.4. Non-Isothermal Crystallization of the Modified PLA Systems
DSC was used to study the effects of BA, DCP, Joncryl, and the resultant structures on the crystallization and melting temperature of the PLA matrix. The DSC data are summarized in
Table 3. The degree of crystallinity (
χm) during cold crystallization, during heating (
χcc), and total crystallinity (
χc) were calculated using the following equations [
2,
23]:
or
where Δ
Hm is the melting enthalpy, Δ
Hcc is the enthalpy of cold crystallization,
∅PLA is the weight fraction of PLA, and Δ
H°m is the enthalpy of fusion of 100% PLA, taken as 93.7 J/g [
33].
Figure 6 shows the DSC thermograms from the second heating. PLA shows diverse transitions; the first transition is related to the PLA
Tg at (60 °C), the second transition is allied with the
Tcc at (110.24 °C), and the last transition is linked with the
Tm at (169.11 °C). In addition,
Figure 6a also shows various melting temperature peaks at 164.13 and 169.11 °C. Moreover, there was no
Tc detected for PLA during the cooling cycle, because PLA crystallizes very slowly [
34]. Upon addition of BA, and chain extension/branching by DCP/Joncryl, the
Tm of the samples moved to the low side of the graph compared to neat PLA. This observation suggests that BA acted as a weak nucleating agent in the samples; meanwhile, 3 wt.% showed a profound shift to the lower side of the
Tm peak (165.5 °C), suggesting a strong nucleating effect. Similar nucleating effects were reported by Malwela et al. [
10] and Das et al. [
2]. On the other hand, the
Tcc temperatures were also affected by the addition of BA, due to the nucleating effect. Upon addition of BA, the
Tc value of the composites decreased, moving towards the lower crystallization temperatures, indicating enhanced nucleation. Malwela et al. [
10] reported a similar nucleating effect.
Additionally, BA limited the mobility of the PLA macromolecules, restricting their chain arrangement. When the molecular structure of PLA was altered by DCP and Joncryl, the
Tcc was reduced, and the crystallinity increased. This phenomenon is related to PLA degradation [
35,
36,
37]. However, when the BA content was increased, the crystallinity in 3 wt.% decreased due to the well-dispersed BA particles causing a physical barrier in the PLA matrix. The
Tg of the samples remained unchanged regardless of incorporating BA or alterations to the molecular structure of PLA. This observation is related to the DMA results that will be discussed later. Overall, the
Tm and
Tc of the composites decreased compared to the neat PLA, confirming that BA, DCP, and Joncryl are good nucleating agents. Small loading of the nucleating agent assisted in forming the polymer crystals; meanwhile, high loading of the nucleating agent restricted the ordered arrangement of the molecular chain, leading to low crystallinity. Moreover, during heating, more crystals were formed; as a result, the PLA crystallinity was improved.
The XRD patterns of neat PLA, BA powder, and composites are shown in
Figure 7. The diffraction patterns on the as-received BA powder were observed at 2θ = 13.98°, 28.12°, 38.36°, 49.46°, 55.11°, and 64.60°, attributed to the (20, 120, 031, 200, 002, and 151 crystallographic planes, respectively. In the case of neat PLA, the broad amorphous peak at 16.50° was observed and ascribed to the 200/100 crystallographic plane of PLA crystal, consistent with the features of PLA [
38,
39,
40,
41]. Upon the addition of various BA concentrations to PLA, the features of BA at the peak of interest (2¦È = 13.98°) were also recognized, signifying the presence of the filler in the composites. On the other hand, the XRD patterns of all of the composites show an intensive peak around 16.21°, and have slightly moved to a higher angle, suggesting that the crystal size of the composites has decreased due to the interaction and distribution of BA in the PLA matrix [
2]. Chain branching contributes to the higher crystallinity of PLA.
Similarly, this tends to reduce the crystal sizes. Therefore, it was important to calculate the crystallite size of the samples, using the Scherrer equation shown below (Equation (3)) as a mathematical expression of the relationship between full width at half maximum FWHM and the crystallite size. The results of the crystalline size for all samples are listed in
Table 4.
where
FWHM is the full width at half maximum attained from the instrument, λ is the wavelength of the X-ray that was used for the diffraction,
L is the crystalline size,
θ is the peak position (2θ/2) in radians obtained from the instrument, and
K is a shape factor constant with a value of 0.9 [
42]. Neat BA shows a crystalline size of 39.59, and PLA 12.8, with a
Tm of 169.11 °C from DSC curves. When a polymer is heated at the minimal
Tm and the equilibrium
Tm (
Tm0), the remaining well-ordered structures in the melt will significantly influence the crystallinity [
43]. Upon alteration of the PLA structure, the crystal size decreased, and the addition of various BA loading further increased the crystal size with 5 wt.% as the threshold. The small crystalline size is due to the crystal growth of the polymer, which is attained by the extra addition of folded polymer chain segments, meaning that the sample has a lower
Tm value [
44]. Farid et al. [
45] reported a similar observation. Further, the crystal size was not dependent on the BA concentration. In conclusion, we observed that the BA and the resultant branched structure acted as good nucleating agents, and this observation was consistent with the DSC analysis.
3.6. Thermomechanical Properties
The effects of BA distribution on the thermomechanical properties of PLA were examined.
Figure 8a shows the storage modulus (E′) of the samples as a function of temperature. The E′ of the samples is discussed at two different phases: glassy phase, below the
Tg, where the polymer chains are highly restricted; and transition phase, at the
Tg of PLA (60 °C). The glassy phase shows that neat PLA has a very low E′ compared to the composites. The reinforcing effect of BA in increasing the Eʹ of PLA was noted. However, the addition of DCP further increased the E′ of PLA due to the enhanced distribution of BA and the possibility of branched chains and/or crosslinking, which contributed to the rigidity of the PLA matrix. On the other hand, when Joncryl was added to PLA/BA, it further increased the E′ higher than DCP in PLA/BA, because of the chain extender used to extend the PLA chains.
Furthermore, the concurrent addition of DCP and Joncryl to PLA/BA further increased the E′. The presence of both DCP and Joncryl induces chemical bonding between PLA and BA, as shown in
Figure 4; hence, better distribution of BA, as shown in
Figure 5. This results in strong interfacial bonding between PLA and BA; hence, the E′ increases when both DCP and Joncryl are added. The increase in E′ can also be attributed to the increase in crystallinity, as shown in
Figure 3. With an increase in temperature, the E′ decreased, as expected. Region 2 illustrates the
Tg of all of the samples, as listed in
Table 6 and shown in
Figure 7; the
Tg of all of the samples remained almost the same, indicating no effect of BA distribution on the
Tg. The
Tg of all of the samples examined from the tan delta curve (
Figure 7) clearly shows no effect on the
Tg, although it was expected that the
Tg would move to higher temperatures due to the chain restriction in the presence of BA particles. Overall, it is evident that the storage modulus was dependent on the distribution of BA.
3.7. Tensile
Figure 9 displays the tensile modulus (E′) and elongation at break (ε
ba) of the neat PLA, PLA/DCP, PLA/J, PLA/DCP/J/, and PLA/DCP/J/BA composites at various concentrations of BA. The neat rigid PLA shows a high E′ of 2040 MPa. Expectedly, PLA exhibits a low ε
ba of 4.8%. Upon the structural modification of PLA using DCP and Joncryl individually, the ε
ba did not change significantly. However, the E′ of PLA/DCP was higher than that of PLA and PLA/J. The entanglement of crosslinked structures could have caused this increase.
On the other hand, the structural modification of PLA using both DCP and Joncryl concurrently did not change the ε
ba, while the E′ slightly increased with respect to neat PLA. Upon the addition of 2–4 wt.% BA to the matrix, the E′ slightly decreased, steadily increasing as the BA concentration increased. At the same time, the ε
ba increased in low concentrations (2 and 3 wt.%), with a decrease at 4 and 5 wt.% loading. We believe that the good distribution of the filler and improved matrix interaction enhanced the stress transfer of the materials [
2]. In this case, the better distribution observed in 3 wt.% loading, as shown in
Figure 5, did not lead to the highest enhanced E′ and ε
ba. Overall, the structural modification of PLA and the incorporation of BA in all of the systems did not significantly affect the E′ and ε
ba of PLA.