3.1. Optimization of the drying stage
Curves of lost moisture percentage of PET in IR oven vs. drying time for different operation power are depicted in Fig. 2. It is clearly observed that the moisture loss for 1500 W and 1000 W is similar, and it increases for a power of 1900 W, which can be associated with a higher internal temperature in the polymer for the same residence time. In fact, the temperature achieved by PET surface inside the oven varied from 110°C at 100W, to 148°C at 1500 W and 160°C at 1900 W.
Likewise, stabilization in moisture loss was achieved for all cases at approximately 40 minutes, so this time was chosen as the optimal one to dry PET.
Dried samples were then compression molded, and their crystallinity evaluated, as it has been reported to be sensitive to degradation reactions in PET, such as hydrolysis [15]. Results are shown in Fig. 3a. A slight increase in crystallinity is observed with decreasing operation power. This increase in crystallinity has been observed when degradation occurred due to the scission of previously entangled chains in the amorphous regions, then having sufficient mobility to crystallize. Moreover, FTIR analysis of same samples was conducted, and results are shown in Fig. 3b. FTIR analysis can monitor changes in the bond density of polymers, which may be related to degradation processes or any other event that may modify the chemical characteristics of the material [16]. Enlargements of two zones of the FTIR spectra are also shown in Fig. 3b: the high wavelength zone, in which the peak at 2960 cm− 1 corresponds to the asymmetric stretching of the C-H bond of the ethylene-glycol segment of the PET molecule; and the mid-wavelength zone, in which the peak at 1720 cm− 1 corresponds to the vibration and stretching of the C = O group of the ester bond. It is clearly observed that as the drying power decreases, the intensity of the peak at 2960 cm− 1 increases and the intensity of the peak at 1720 cm− 1 decreases. This has been reported in literature as indicators of degradation due to hydrolysis reactions in PET [17, 18].
Our results then indicate that optimal operation parameters of the IR oven for PET are 40 minutes and 1900 W. Lower operation powers (which translate into lower polymer temperatures) or times are not enough to extract water from the polymer to prevent hydrolysis degradation reactions during subsequent processing.
3.2. Comparison of global performance
Table 2 presents a summary of the obtained results, both for virgin and recycled PET dried for 40 min in the IR oven and for 3,5 h in the conventional oven. It is easily seen that, beyond expected differences due to recycling process, there is no substantial difference in the values of evaluated properties. Differences are within error intervals when properties of each material dried by IR or conventional oven are compared. The only difference is found between fracture toughness of vPET-C and vPET-IR. Moreover, a noticeable difference was observed in the fracture during the propagation stage of vPET-C and vPET-IR, as shown in Fig. 4. Typical load-displacement curves along with pictures of the vPET samples during the tests are displayed. It is observed that although both materials have a similar initiation behavior, differences in the propagation mode between vPET-C and vPET-IR are very noticeable. vPET-C presents a post yielding behavior, characterized by a stable crack propagation with an evident crack tip rounding and a large plastic deformation of the remnant ligament. On the other hand, vPET-IR presents a semi-ductile behavior, characterized by crack tip opening, accompanied by slight plastic deformation and a final unstable propagation with a catastrophic failure.
Table 2
properties of vPET and rPET dried in conventional and IR ovens.
| Sample | Cristallinity % | MFI (g/10min) | Flexural modulus (GPa) | Fracture Toughness (N/mm) |
| vPET-C | 27 ± 2 | 31 ± 2 | 7 ± 2 | 93 ± 24 |
| vPET-IR | 25 ± 3 | 30 ± 6 | 7 ± 1 | 113 ± 20 |
| rPET-C | 28 ± 4 | 25 ± 11 | 7 ± 1 | 82 ± 15 |
| rPET-IR | 29 ± 3 | 27 ± 10 | 7 ± 1 | 83 ± 10 |
Typical SEM micrographs of fracture surfaces are shown in Fig. 5. The vPET-C sample exhibits large plastic deformation, with tapered edges and highly elongated material. One ligament is marked with a red oval, in which striations can be observed at 45 degrees as a result of plastic deformation. In fact, the ligament was longer and it was cut for observation purposes. In the vPET-IR sample, on the other hand, a restricted plastic deformation is observed, with voids on the fracture surface due to the abrupt breakage of material, and much smaller ligaments (also indicated by a red oval). Both rPET samples exhibit even less plastic deformation, with almost no elongation of the material prior to its fracture. A lower plastic deformation of rPET-IR is also observed, compared to rPET-E, although the difference is minimal when compared to the difference observed between vPET-IR and vPET-E. It is then clear that both drying in the IR oven (from left to right) and recycling (from top to bottom) reduce the plastic deformation that PET can develop.
It is important to notice here that fracture toughness is indicative of the material resistance to fracture in the presence of a sharp crack, which is not completely avoidable in the processing of a polymeric material. Moreover, it has been found that fracture tests also provide interesting information to detect and understand degradation mechanisms, even in early stages of aging [19]. Our results are in great concordance with this affirmation, since it is the only test that detected some effect of IR radiation on vPET.
It is then evident that IR drying may generate an inhibition of the plastic deformation mechanisms of PET. This observation could imply that some molecular or chemical events have occurred during IR drying [16]. To study these events in a deeper way FTIR tests were carried out. For the analysis, the vPET-C sample was considered as a reference – because conventional drying did not generate chemical alterations in it – and the peaks associated with conformational changes of the chain were excluded. Figure 6 shows the obtained FTIR spectra, and peaks corresponding to ester bond [20] are depicted in Table 3.
It is clearly observed in Fig. 6 an increase in the typical peaks of the PET ester in vPET-IR sample, when compared with vPET-C sample. If hydrolysis reactions occurred in the samples that were subjected to IR radiation, the intensity of these three peaks should decrease [18]. Therefore, vPET-IR has experienced esterification reactions, in which the chains continue to react and link together, releasing water as a product, instead of hydrolysis reactions (see Fig. 7) [21]. These esterification reactions lead to the formation of longer chains, i.e., a polymer with a higher molecular weight. This result is consistent with the observation of the inhibition of plastic deformation mechanisms experienced by vPET-IR. There is evidence that the configuration of the amorphous region (inter and intra spherulites) and the molecular weight (which determines both the size and configuration of amorphous regions) are important parameters to represent the mechanisms of plastic deformation of semicrystalline polymers [22]. Particularly, tie molecules (referring to the amorphous fraction between lamellas inside spherulites) play a central role in the plastic deformation, since they are responsible for transmitting stress under deformation. Thus, tie molecules support the external force required for the lamellas’ plastic deformation mechanisms to happen. An increase of tie molecules fraction (which depends directly on the length of tie molecules) reduces the support span of the lamella, leading to a lower deflection: hence, a higher stress will be required for plastic deformation mechanisms to be activated, stiffening the material [22].
Table 3
Characteristic peaks of PET.
| Wavelength (cm-1) | Assigned group |
| 1721 | C = O stretch |
| 1245 | C-C-O stretch |
| 1100 | O-C-C stretch |
There are no indications of this type of secondary reactions in the recycled material, probably since it was dried in the form of scrap with a much smaller thickness than the virgin material pellets. It is possible then, that heat concentration is produced in vPET pellets and not in rPET flakes, since heat is dissipated faster in flakes, avoiding conditions where esterification reactions occur. To prove this hypothesis, a thin film of vPET was obtained after conventional drying. This film was stored at ambient humidity for a week and then chopped to obtain a thin scrap. Finally, this scrap was dried in the IR oven and processed as previously described. The FTIR curve corresponding to this material is denoted as flat vPET-IR in Fig. 6, in which no evidence of hydrolysis or esterification reactions are observed. This result allowed us to corroborate our hypothesis: the thickness of the material to be dried influences the occurrence of esterification reactions.
Based on our findings, esterification reactions in thick PET seem to be facilitated by two distinct mechanisms. Firstly, IR drying significantly accelerates water extraction, reducing available water for hydrolysis by 80% compared to conventional methods. Secondly, thick PET pellets likely experience concentrated heat, resulting in inner temperatures exceeding 170°C. These factors collectively promote more favorable conditions for esterification reactions in thick PET materials.