Starch powder (60 g/100 g of system) was dispersed in a mixture of glycerol (20 g/100 g of system) and water (20 g/100 g of system). The suspension was manually mixed until complete absorption of the liquid phase into the starch. The mixture was processed by extrusion technique using a twin screw extruder ThermoScientific^® process 11, using the following temperature profile: 100–110–120–130–130–130–120–100°C, and a screw speed of 25 rpm. Then, the filaments were pelletized in a cryogenic mill IKA^® A11 BASIC, and subsequently used to prepare the blends.

Before being processed, PLA pellets were dried in a Cole Parmer vacuum oven for 24 h in order to prevent PLA hydrolysis during processing. Then, extruded samples of neat and plasticized PLA were obtained. PLA pellets were mixed with one of the plasticizer manually and then introduced in the same twin screw extruder previously described, using the following temperature profile: 195–198–195–198–198–198–198–200°C and a screw speed of 20 rpm. The proportions of plasticizers (TB or TEC) used was 12.5 wt.% of the final mixture weight. Neat PLA pellets were also processed following the same procedure than for the plasticized samples.

Additionally, PLA/TPS blends were extruded following the last temperature profile and screw speed described previously. Neat PLA (or plasticized PLA) was manually mixed with TPS pellets prior to be introduced in the extruder hopper, maintaining the proportion of 30 wt.% TPS of the final mixture weight.

Films of all the extruded samples were obtained by compression molding in a hydraulic press at 195°C following the procedure: they were kept between the two hot plates at atmospheric pressure for 5 min until melting and then under a pressure of 70 kgf cm⁻² for another 5 min. Finally, the films were fast-cooled to room temperature.

The processing conditions, temperature and screw speed, were selected considering the thermal properties of the polymers and in order to obtain a homogeneous extruded product. These processing conditions contribute to partial degradation of PLA leading to chain scission, plasticizer volatilization, and initial crystallization (Velghe et al., 2023). This effect was quantified by measuring the molecular weight of the PLA pellets using GPC, and the results are M¯n = 159,684 and M¯w = 196,087, which yielded values that were clearly higher than those measured after processing (shown in Section 3.1.1).

Accelerated weathering tests were carried out in the test chamber QUV Accelerated Weathering Tester Q-Lab equipped with 4 UVA-340 lamps (wavelength from 365 to 295 nm), irradiance of 0.77 W/m² at up to 60°C temperature and 100% relative humidity. The weathering conditions were determined in accordance with the ASTM D G154 Operating Fluorescent Light Apparatus for UV AcExposure of Nonmetallic Materials, cycle 1. The samples were exposed to 8 h of UV irradiance at 60°C followed by 4 h of condensation without irradiation at 50°C for 996 h (41.5 days). Samples were removed from the weathering chamber for characterization at 24, 48, 120, 168, 228, 360, 528, 864, and 996 h.

Outdoor exposure of the neat and plasticized PLA, TPS, and PLA/TPS samples was carried out in the city of Mar de Cobo, Buenos Aires, Argentina (Latitude 37°46′00′ S, Longitude 57°26′00′ W). The samples were mounted on a support and directly exposed to the natural environmental conditions between the months of April to December 2024 and January 2025. Summary of the relevant values of the weathering conditions, obtained from National Meteorological sources corresponding to the location and time period of the outdoor test, is showed in Table 1 to provide the relevant context where the degradation occurred. Samples were collected for characterization at 15, 50, 162, 239, and 324 days.

The molecular weight of PLA in the samples before and after exposed in weathering conditions were evaluated with a Shimadzu system with a series of Shim-Pack columns and equipped with a refractive index detector. The eluent used was tetrahydrofuran (THF) at a flow rate of 1 mL/min at 25°C. Sample solutions were dissolved in THF at 3 mg/mL at 40°C and filtered through a 0.5 μm polytetrafluoroethylene membrane prior to injection. Calibration was performed using polystyrene standards.

The average number of chain scissions per macromolecule was determined by Equation 1 following the methodology described by Sadi et al. (2010):

where M¯n0 and M¯nt are the number-average molecular weight of unweathered and weathered samples for a given time t, respectively.

Thermal characterization of the films was performed with a Pyris 1 (Perkin Elmer) under nitrogen atmosphere, with a heating rate of 10°C/min from −50 to 220°C. The following thermal properties of the samples were calculated from the DSC scans: the glass transition temperature (T_g), the crystallization temperature (T_c) and the melting temperature (T_m). The T_g was measured at the half-height of the heat capacity increase from the DSC curve. The degree of crystallinity (X_c) was calculated by means of Equation 2:

where the ϕ is the mass fraction of the PLA in the blend, ΔH_m is the melting enthalpy, ΔH_c is the enthalpy of crystallization and ΔHm0 is the melting enthalpy of PLA 100% crystalline (93 J/g; Avérous, 2008).

FTIR spectra of the films were obtained using a FTIR Nicolet 6,700 spectrometer (Thermo Scientific Instrument) in ATR mode in the 4,000–400 cm⁻¹ range. The measurements were averaged over 64 scans. The C–H stretching band at 2,988 cm⁻¹ were chosen as internal standard (Kaplan et al., 2019).

The crystalline structure of the samples was evaluated on an PANalytical X'PERT PRO diffractometer (Netherlands) with CuKα radiation (λ = 1.5406 Å). The X-ray tube was operated at 40 kV and 40 mA. The diffractograms were acquired at a scanning speed of 2°/min in the range from 5 to 60°.

The overall crystallinity degree (X_c) was calculated by means of Equation 3, that relates amorphous and crystalline areas assessed from the spectra (Seoane et al., 2017):

The crystallite size was calculated using the Scherrer Equation (Giełdowska et al., 2022):

where L(hkl) is the average sizes of the crystalline areas orthogonal to lattice planes (hkl), K is the Scherrer constant (K = 0.89), λ = 0.15406 is the X-ray wavelength, B is the half-peak width of the XRD peak, and θ is the Bragg angle for planes (hkl). Equation 4 was applied to the main diffraction peaks at 2θ values of 16.5 and 19°, corresponding to the (110)/(200) and (203) crystallographic planes, respectively.

Thermal degradation analyses of the films were carried out under nitrogen atmosphere (50 mL/min flow rate) at a heating rate of 10°C/min in the 25–700°C range, using a TGA-50 SHIMADZU thermogravimetric analyzer (Japan).

An INSTRON Emic 23–50 universal testing machine operated at a speed of 1 mm/min was used to measure the uniaxial tensile properties of the films according to ASTM D 638–03. The tested samples were cut from the films using a scalpel, measuring 25 mm in length and 5 mm in width. The Young's modulus, tensile strength and strain at break were determined from the assays as the average value of five samples for each system. The standard deviation was calculated as a measure of dispersion.

The morphology of the samples were examined by means of field emission scanning electron microscopy (FESEM) in a ZEISS Crossbeam 350. The samples were previously subjected to brittle fracture by immersion in liquid nitrogen, placed on a support and then covered with a thin layer of gold.

The light transmission of the films was determined in by quantifying the absorption of light at wavelengths between 300 and 1,000 nm using an UV–visible spectrophotometer Agilent 8453.

A CIELab color space by means of a Lovibond Colorimeter RT500 (Amesbury, United Kingdom) was used to analyze the color properties of the films. It was calibrated against white standard tile and the evaluated area was 8 mm in diameter. Color coordinates, L^* (lightness), a^* (green–red), and b^* (blue–yellow) and percent opacity (%Op) were measured. The average of three measurements at random positions over the film surfaces were reported. Total color difference (ΔE) was evaluated with respect to the white control according to Equation 5:

where ΔL*=L*-L0*,Δa*=a0*, and Δb*=b*-b0*; being =92.28,a0*= 1.38 and b0*= -0.86.

Accelerated weathering is expected to cause polymer chain scission due to photolysis and hydrolysis reactions, leading to a reduction in molecular weight. To corroborate this, the variation in PLA molecular weight with exposure time in the chamber was determined by means of GPC tests. Table 2 shows the values of M¯n, the weight-average molecular weight (M¯w) and Polydispersity Index (PDI) measured for the films of neat and plasticized PLA. As predicted, the molecular weight progressively decreased with increasing exposure time for all the samples (Zhang et al., 2021). However, the value dropped dramatically at 528 h for the plasticized samples, while for the pure PLA films, this occurred at longer times. This may be because the presence of the plasticizer reduces the interaction between the polymer chains, facilitating the entry of water and subsequent hydrolysis and chain rupture. Chain scission is also reflected in the increase in the PDI with increasing exposure time, indicating the formation of low molecular weight chains and a greater heterogeneity in polymer chain sizes (Rebelo et al., 2022). This was further confirmed by assessing the number of chain scissions by means of Equation 1, which is directly related to the extent of PLA chain degradation after weathering (Lv et al., 2018). As shown in Table 2, this number increased with exposure time for all the samples, with a more pronounced effect observed in the plasticized samples. The decrease in M¯w and M¯n, along with the increase in PDI and number of chain scissions, indicates significant polymer degradation, which likely contributes to the deterioration of the mechanical and physical properties of the weathered materials, as discussed below.

The samples containing TPS degraded quickly under the chamber conditions, disintegrating within 24 h of exposure, which made further characterization impossible. This suggests that starch degrades more readily than PLA under high humidity and temperature conditions due to its hydrophilic character (Lv et al., 2018). Moreover, the molecular weight of PLA is lower when blended with TPS (Table 2), likely as a result of hydrolytic degradation triggered by the presence of hydrophilic components.

DSC tests were carried out on the films at different exposure times in the chamber and the curves obtained for the heating scan are shown in Figure 2. The disappearance of the crystallization peak at short exposure times suggests a rapid increase in the crystallinity of the samples. The percentage of crystallinity (X_c) of the different samples was calculated using Equation 2 and the values are reported in Table 3. It was observed that the plasticized samples exhibited a faster increase in crystallinity compared to neat PLA. The plasticized samples achieved a remarkable 150% increase in X_c within just 48 h of exposure, whereas the neat PLA films required 120 h to reach a comparable X_c value of 130%. Furthermore, the crystallinity of the plasticized films remained higher than that of neat PLA at 996 h, the final point of the test, regardless of the type of plasticizer used. This occurs because the chamber's humidity promotes the hydrolysis of PLA chains, facilitating their breakage and movement, which in turn enhances crystallization. This process is further accelerated in the presence of plasticizers, as they increase the free volume, allowing for greater mobility and rearrangement of the polymer chains (Iglesias Montes et al., 2019). The crystallization temperature (T_c) of neat PLA gradually shifted to lower values with increasing weathering time, likely due to the reduction in molecular weight, as shorter polymer chains tend to crystallize more easily at lower temperatures. In parallel, a slight decrease in the melting temperature (T_m) was also observed at longer exposure times, which can be attributed to the formation of less perfect crystals under the prevailing humidity and temperature conditions.

The plasticizing effect of TB and TEC was also reflected in the diminution of the glass transition temperature (T_g) of PLA, which has already been corroborated in previous studies (Malbos et al., 2024). Initially, the T_g of the plasticized films was lower than that of pure PLA (55.45°C), with values of 38.1°C for PLA-TB and 35.1°C for PLA-TEC. The difference in T_g values measured for samples plasticized with TB and TEC can be attributed to the structural differences between the two plasticizers and their interactions with PLA. This is consistent with previous studies, which have shown that the thermal properties of PLA are influenced by the type of plasticizer used, depending on its molecular structure and compatibility with the polymer, as was also related below with the tensile test results (Maiza et al., 2016; Ruiz et al., 2021). Then, under accelerated weathering conditions, the T_g of neat PLA remains almost unchanged throughout the test (Figure 2), a finding consistent with the results reported by Zhou and Xanthos (2008) who studied the hydrolytic degradation of PLLA subjected to accelerated hydrolytic degradation over a temperature range of 50–70°C. However, the T_g of PLA in the plasticized samples increased over the course of the test (Figure 2), which can be attributed to the loss of plasticizer under the applied conditions, as further confirmed by the TGA results discussed below. Additionally, the glass transition became progressively less pronounced in the DSC analyses as weathering progressed, likely due to a reduction in the amorphous phase content of PLA associated with the increase in crystallinity (X_c).

Figure 2 also displays the thermal test results for the samples containing TPS up to 24 h of exposure because these samples degraded completely after this time.

It is evident that TPS acts as a nucleating agent when added to PLA. This observation aligns with findings reported by other authors, who analyzed the crystallization kinetics of PLA and concluded that the presence of TPS enhances both the crystallization rate and the degree of crystallinity in the samples (Cai et al., 2011; Li and Huneault, 2008; Li et al., 2013). They found that the interfacial area within the PLA/TPS blend can influence the nucleation and crystallization process of the material. The blends showed a behavior comparable to that of neat and plasticized PLA films after 24 h of exposure, marked by an increase in both glass transition temperature (T_g) and PLA crystallinity. However, the presence of TPS also promotes PLA crystallization under these conditions, as the increase in crystallinity observed in the blends after 24 h is comparable to that reached by the plasticized samples at 48 h and by neat PLA at 120 h.

Samples of neat and plasticized PLA were analyzed using thermogravimetric analysis, providing valuable insights into the thermal degradation behavior of the materials. Figure 3 shows the curves of percentage of mass loss and its derivative with respect to temperature for all the films before to be exposed to accelerated weathering in the chamber. Moreover, the thermal degradation profiles of plasticizers TB and TEC were also reported in Figure 3. PLA was observed to degrade in a single-step process, with a maximum degradation temperature of 353°C, and complete mass loss at the end of the test. The decomposition of PLA has been found to occur via a non-radical mechanism, specifically through a back-biting ester interchange reaction, which produces acetaldehyde, carbon dioxide, carbon monoxide, cyclic oligomers, and lactide (Chen et al., 2012; Kervran et al., 2022; Mofokeng and Luyt, 2015). On the other hand, both TB and TEC degrade at lower temperature than the PLA samples, with a maximum degradation temperature around 200°C, exhibiting similar thermal degradation behavior due to their comparable chemical structures. It is possible to observe in Figure 3 both the peaks corresponding to the degradation of PLA and TB or TEC in the plasticized samples. The peak at lower temperature appeared in the region corresponding to the decomposition of the plasticizers, while the second one at higher temperature was attributed to the degradation of PLA. Then, the addition of plasticizers did not significantly affect the maximum degradation temperature of PLA, which remained higher than the temperature chosen for the processing conditions of the samples. For thermoplastic processability, an essential characteristic is the material's thermal stability at the processing temperature. This temperature must be significantly higher than the melting temperature (T_m) and lower than the degradation temperature (T_d) to prevent polymer degradation during processing.

Figure 4 shows the curves of percentage of mass loss for the plasticized PLA samples exposed to different accelerated aging times in the range between 100 and 300°C. A progressive loss of plasticizer was observed as the material undergoes degradation. In the case of the PLA-TEC sample, at 996 h of exposure, practically no loss of mass was detected due to the plasticizer, and this drop was minimal for PLA-TB at the same exposure time. These results suggest that nearly all of the plasticizer was eliminated by the end of the exposure period. Additionally, they align with the progressive increase in T_g over time for these samples, previously observed.

From the DTG curves of the neat and plasticized PLA films (Figures 5A–C), it was observed that the temperature corresponding to the maximum degradation rate of PLA initially increased slightly after 24 h of exposure, then gradually decreased, and later increased again at longer exposure times. This behavior can be correlated with the initial rise in crystallinity after 24 h, which was subsequently counteracted by degradation effects and, ultimately, by the loss of plasticizer as the exposure time progressed (Figure 4).

On the other hand, Figure 6 shows that neat TPS thermally degraded in two steps. It was reported that the first step, with a DTG peak at ~130°C, corresponds to water evaporation, while the second step is associated with the thermal decomposition of starch (Ojogbo et al., 2018). As expected, the percentage of the weight loss in the first step is directly related to the moisture content of the starch. Liu et al. (2008) found that the initial water content of the sample had no impact on the onset of decomposition temperature of starch, as these are two distinct processes. The second step involves the breakdown of chemical bonds in the starch chain, as well as oxidation and dextrinization (Winkler et al., 2014). Thermal reactions for starch began around 300°C, involving condensation between the hydroxyl groups of starch chains, leading to the formation of ether segments and the release of water molecules and other small species. The structure of starch disintegrates at 400°C, and beyond this point, forms a highly cross-linked system. Carbonization reactions occur in the system at temperatures above 500°C, leading to the formation of relatively large conjugated aromatic structures at around 600°C. With further heating, these structures transform into amorphous carbon (Ojogbo et al., 2018; Zhang et al., 2002). The TPS film, after being exposed in the chamber under accelerated aging conditions for 24 h, showed an increase in the height of the first peak at a lower temperature, due to the loss of the water absorbed by the sample during that period (Figure 5D).

For the PLA/TPS samples prior to exposure in the chamber, the DTG curve in Figure 5D revealed a single intermediate peak located between the characteristic peaks of each individual polymer. However, after 24 h of exposure, the peaks corresponding to each polymer became distinguishable due to the increase in the degradation temperature of PLA, as illustrated in Figure 5A. The DTG curves of the plasticized PLA/TPS samples (also shown in Figure 5D) exhibited a similar evolution, indicating that plasticization did not significantly alter the phase separation behavior upon weathering.

This is consistent with the analogous shift of the PLA peak observed after the addition of both plasticizers, as previously explained. These results suggest that PLA and TPS are immiscible, forming two separate phases.

From the diffractograms obtained by XRD (Figure 7), an increase in the crystallinity of the samples was observed with the exposure time in the chamber. It was clearly evidenced by the progressive increase in the intensity of the peaks in the 2θ = 16.5 and 19° corresponding to the lattice planes (110/200) and (203) of the PLA crystals (Hsieh et al., 2020) and others of lower intensity in 2θ = 24.3, 28.6, and 32.2°. In the case of neat PLA, these peaks began to be more noticeable after 48 h of exposure, while for plasticized samples and blends with TPS, they could already be clearly identified after 24 h. To better visualize the changes in the diffractograms over the exposure period, the spectra of both neat and plasticized PLA samples at the initial and final test times are plotted in Figure 8. The percentage of overall crystallinity of the different samples was calculated from the XRD spectra using Equation 3, and the results are shown in Table 4. It was observed that the values follow the same trend as the results obtained by DSC (Table 3). The combination of UV irradiation cycles at 60°C, humidity and the decrease in the molecular weight of PLA favor the crystallization of the polymer chains (Kaynak and Sari, 2016). The X_c values calculated from XRD do not exactly match those calculated from DSC, due to the different technique used. In addition, there is a greater inaccuracy in the calculation at low exposure times, where the peaks are less defined in the spectra (Kaavessina et al., 2013).

The crystallite size of PLA after 996 h of exposure in the chamber was estimated using the Scherrer (Equation 4), considering the two main diffraction peaks at 2θ = 16.5 and 19°, which correspond to the (110)/(200) and (203) crystallographic planes, respectively. As shown in Table 5, the crystallites in the plasticized samples were smaller compared to those in the neat PLA films. This observation aligns with the lower molecular weight measured by GPC for the films containing TB and TEC at the end of the weathering test. Moreover, the melting temperature (T_m) also decreased accordingly (Table 5).

The FTIR spectra in the 400–4,000 cm⁻¹ region obtained for neat PLA before being exposed in the chamber showed the characteristic peaks of the polymer in Figure 9. The main characteristic bands of PLA are the following: –C=O ester stretching bands at 1,752 cm⁻¹ (crystalline) and 1,744 cm⁻¹ (amorphous); –C–O–C– stretching in the range 1,265–1,080 cm⁻¹; C–H deformation and stretching at 1,356 and 2,996–2,946 cm⁻¹, respectively; C–CH₃ stretching at 1,044 cm⁻¹; CH₃ asymmetric and symmetric deformation at 1,450 and 1,380 cm⁻¹, respectively; and the –C–C– stretching was observed at 865 cm⁻¹ (Furukawa et al., 2005; Kaplan et al., 2019). No significant differences were found between the spectra of the plasticized films and the neat PLA, indicating that all samples have a similar chemical structure initially.

It was reported that photodegradation of PLA results in random chain scission via the Norrish II type photocleavage reaction, which involves the ester group and ethylidene group next to the ester oxygen, leading to the formation of C=C double bonds and carboxylic acid groups on the formed chains (Kaynak and Sari, 2016). Photolysis and/or photooxidation are found to be the mechanisms that occur during UV irradiation of PLA. During the photolysis takes place especially the C–O bonds scission of the PLA ester backbone structure. While photooxidation of PLA leads to the formation of hydroperoxide derivative and its subsequent degradation into carboxylic acid and diketone end groups (Bocchini et al., 2010; Man et al., 2012; Zhang et al., 2011). Moreover, hydrolysis due to moisture reduces the molecular weight of PLA chains by cleaving the ester bonds and also leads to the formation of carboxylic acid and ketone end groups. This process also causes swelling and plasticization (Deroiné et al., 2014; Shinzawa et al., 2012).

FTIR analysis of the neat and plasticized PLA revealed some spectral variations as weathering time in the chamber progressed, as illustrated in Figure 9. The observed changes are in accordance with the variation in the chemical groups mentioned previously, and they are described below. It was observed an increase of the band at 1,752 cm⁻¹ consistent with the formation of C=O groups due to photo-oxidation and hydrolysis as well as to the increment in the percentage of crystallinity of PLA during the test. The increase in the band at 865 cm⁻¹, attributed to C-C stretching, provides further evidence of the increase in crystallinity of the samples with the exposure time (Kaplan et al., 2019). Additionally, the band at 1,264 cm⁻¹ decreased, which can be attributed to the cleavage of the C-O bond in the ester backbone due to photolysis. Finally, a broad band appeared in the 3,500–3,000 cm⁻¹ region of some spectra, which could be attributed to moisture retained in the films after their removal from the chamber.