PLA’s isothermal crystallization behavior was investigated using a regular DSC. Figure 1 shows the DSC curves of the isothermal crystallization of the PLA samples at different temperatures. At a low temperature of 80 °C, no obvious crystallization peak was found, indicating that the crystallization rate of PLA was very slow, and 10 h was not enough to induce its crystallization. As shown in the following figure (Figure 2), the T_g and T_m of the PLA is about 58 °C and 148 °C, respectively. Therefore, a temperature of 80 °C was too low to induce PLA crystallization within a short time due to low chain mobility. At 100 °C, however, PLA could crystallize, as observed by the small exothermal peak from 245 to 549 min. With further increasing the temperature to 120 °C, the crystallization rate became faster with crystallization occurring from 155 min, and the ΔH_c is bigger. When the temperature is increased to 140 °C, PLA crystallization becomes difficult, because the high chain mobility interferes with the regular arrangement of the polymer chain.

Based on these results, PLA crystallization was difficult. It was found that a longer time and narrower temperatures’ window might be needed to induce PLA crystallization. In addition, compared to 100 and 140 °C, it seems that 120 °C was a proper temperature to induce PLA crystallization under atmospheric pressure.

Amorphous PLA films were treated under compressed CO₂, and the different processing parameters of temperature, pressure, and time were used during the treatment process. DSC was used to investigate the cold-crystallization and melting behavior of the PLA samples in this study.

During the gas releasing process, PLA samples could foam under some conditions due to the strong plasticization effects of CO₂, even though a very thin CO₂ film and a low depressurization rate was applied in these experiments. It is known that a biaxial extensional flow is usually formed in cell walls during cell growth, and this process facilitates strain-induced crystallization. Therefore, the final crystallinity obtained in foamed PLA samples tends to include both the contributions of CO₂-induced crystallization and strain-induced crystallization. In this study, we tried to show their individual effects on the crystallization process.

Figure 9 shows the foaming window of PLA samples as a function of treatment temperature and time. The shadowed region in the figure shows where PLA could foam during the gas release process. Below the shadowed region, the PLA samples could not foam; while above the shadowed region, PLA tended to melt in a high-pressure vessel. It is seen that PLA could not foam at the low temperature of 25 °C. With an increase in temperature to 60 °C, however, foaming began to occur in PLA after being saturation within 20 min, and then could not foam again at the extended time. At a higher temperature, the foaming window of PLA tended to broaden gradually. For example, at 80 °C and 100 °C, foaming occurred after the PLA sample was saturated with compressed CO₂ for 40 min and 1 h, respectively. At 120 °C, PLA could foam even after being saturated for 12 h. With temperature increased to 140 °C, however, PLA foamed only within 10 min, and then tended to melt completely when the time was more than 20 min. Under those conditions, no foams were obtained.

The curved line shown in Figure 9 represents the time required for crystallization equilibrium that was estimated by using the effects of time on crystallinity. The crystallization equilibrium time of PLA at 6.89 MPa exhibited an obvious temperature dependency. At 25 °C, 40 min is needed to reach crystallization equilibrium. With increased temperature, the required time linearly decreases to 35 min at 60 °C, 25 min at 80 °C, and 20 min at 100 °C. By further increasing the temperature, however, the required time increases to 25 min at 120 °C and about 40 min at 140 min, respectively.

As mentioned above, both CO₂-induced crystallization and strain-induced crystallization contributed to crystallinity development in foamed PLA samples. For unfoamed conditions such as at 25 °C, however, it is believed that only CO₂-induced crystallization contributed to the crystallinity of PLA because of the absence of foaming. Therefore, the data at 25 °C supplied a useful reference to describe the possible effect of polymeric foaming on the final crystallinity of PLA. In unfoamed conditions, a reduction of about 5 min in the time needed for crystallization equilibrium was observed by increasing the temperature of 35 °C (from 25 °C to 60 °C), resulting from the increased polymer chain mobility at high temperature. Once the foaming occurred, however, a reduction of about 15 min was realized by increasing the temperature of 20 °C (from 60 °C to 80 °C). These results demonstrated that the foaming also presented a positive effect in reducing the time requirement for crystallization equilibrium of PLA.

Figure 10 shows the SEM micrographs of PLA foams obtained after saturated at 6.89 MPa and 100 °C for different times. At the saturation time of 1 min, only a very few big cells are shown in the foam, which might be resulted from an insufficient CO₂ exposure time; thus, CO₂ solubility equilibrium was not reached. By increasing the time to 3 min, the cell size tended to decrease in the foam center, but a large amount of small cells were observed near the foam skin. At extended times of 5 min and 10 min, the cell size distributions become more uniform due to good gas dispersion in the PLA matrix. At 20 min, however, the obvious heterogeneous cell distribution is observed in the foam center, and lots of very small cells are present in those regions (shown in Figure 10 in the regions pointed by arrows). Similar cell morphology has been observed in PP [36] and polyester amide [37] foams with the foaming temperature that were slightly lower than their melting points, which was attributed to the occurrence of cell nucleation at/around crystal regions, i.e., crystal phase induced heterogeneous cell nucleation. Based on cell morphology obtained after saturation of 10 min, as well as high crystallinity at 20 min, it is suggested that higher crystallinity might have been obtained at 20 min, before foaming. At the extended times of 30 min and 60 min, several spherical particles with a diameter of about 30–40 μm, which seems to be coated by polymer matrix, are observed in the junction of cells. Similar spherical particles have also been observed in the CO₂ induced-crystallization’s polymeric foaming systems like PC [38], where spherical particles were thought of as spherulite crystal. Because no extra solid particles were added in the PLA foaming process, we believed that the same phenomenon occurred in the PLA foaming system. Furthermore, we believed that these spherulites should be formed during the CO₂ saturation process. We made this claim based on two reasons: the first is that the cell wall stretching process during cell growth tended to make the cell wall and junction area thinner, while the particle size was much thicker than that of cell wall (30–40 v.s., 5 μm); the second is that the stretching process that induced spherulite crystal formation was difficult. We also note that the thickness of foamed samples tended to decrease significantly from 30 to 60 min, and then to 120 min, where no foaming occurred.

Foaming-induced crystallization seems to be associated with foam expansion, and a high volume expansion ratio tended to induce high crystallinity in the PLA sample. Mihaela et al. [18] investigated the effect of biaxial stretching on the crystallinity of PLA. They found that a high stretch ratio of 9 induced high crystallinity of about 20% in the same PLA as was used in this study, while a low stretch ratio of 1–4 exhibited a small effect on the PLA’s crystallinity. These results suggest that the stretching process during cell growth might exhibit less effect in low expansion ratio foam compared with high expansion ratio foam. Figure 11 shows the expansion ratio of PLA foams as a function of treatment time. It is seen that the expansion ratio of PLA foams significantly increases within 10 min; that is, 3.21 at 1 min, 7.27 at 3 min, 11.43 at 5 min, and 15.72 at 10 min, and then quickly decreased to 6.07 at 20 min, 3.43 at 30 min, 2.12 at 40 min, 1.56 at 60 min, and to 1.01 at 120 s., respectively. The obvious foam expansion at 5 min and 10 min might contribute highly to crystallinity. This could be due to a rapid increase in the PLA foam’s crystallinity with the addition of the original crystallites formed during the CO₂ saturation process. At the extended time of 30–60 min, PLA foams exhibited a low expansion ratio, because the high crystallinity that was formed during the CO₂ saturation process, as confirmed by SEM micrographs, increased the stiffness of the polymer matrix and inhibited cell growth. Consequently, foam expansion might only slightly affect PLA’s final crystallinity. This result was consistent with the DSC measurement, where the equilibrium crystallization was reached at about 20 min, and no increase in crystallinity was observed by further extending the time up to 12 h. Therefore, the crystallinity formed by CO₂ exposure determined the foaming behavior of PLA, and thus affected the contribution of strain-induced crystallization to the final crystallinity of PLA foams. To be specific, in the case of high CO₂-induced crystallization, strain-induced crystallization had a small effect on the final crystallinity of PLA foams; while in low CO₂-induced crystallization, strain-induced crystallization played an obvious role. Therefore, the PLA sample crystallized very quickly, and the short time required for equilibrium crystallization could be achieved after the sample was treated by the compressed CO₂.

A commercial linear PLA (Ingeo^™, 2002D) in pellet form was provided by Natureworks LLC, USA. According to the supplier, its d-isomer content is 4.3%, melt flow rate is 5–7 g/10 min, and its density is 1.24 g/cm³. CO₂ with a purity of 99.5% (Linde gas) was used as the physical blowing agent in all experiments. The PLA pellets were dried in the oven for 4 h at 80 °C before use. Then, polymer specimens with a thickness of about 0.3 mm were prepared by compression molding at 200 °C. The samples were cut into disc-shaped sheets for CO₂ treatment.

A high-pressure apparatus was used for compressed CO₂ treatment. The 0.3 mm-thick PLA film was placed in a high-pressure vessel preheated to the experimental temperature. The vessel was flushed with low-pressure CO₂ for about 1 min and then pressurized to the desired value. Once the CO₂ pressure reached the desired value, the timing began. At the end of the experiment, the vessel was released at a low depressurization rate, and the sample was removed quickly from the vessel.

DSC measurements were carried out to determine the T_g, cold crystallization temperature (T_c) and melting peak (T_m) by using a Q2000 (TA Instruments) that was calibrated with indium. All measurements were made at a heating rate of 10 °C/min over a temperature range of 20–200 °C in a dry nitrogen environment. The enthalpy of crystallization on heating ΔH_c and melting enthalpy ΔH_m were measured. The crystallinity of PLA CO₂-treated samples was calculated by [(ΔH_m ΔH_c)/ΔH_f] × 100%, where ΔH_f is the theoretical heat of fusion of 100% crystalline PLA with a value of 93 J/g [39]. Isothermal crystallization from the amorphous state was used to measure the crystallization rate of PLA. The samples were first heated to 200 °C at a rate of 10 °C/min to eliminate any thermal history, and then quickly cooled to the present crystallization temperature to initiate isothermal crystallization. In the isothermal crystallization study, the samples were held at the crystallization temperature for 10 h. For CO₂-treated PLA samples, the first heat scanning curve was used to calculate crystallinity.

WAXD measurements were taken on a Siemens D5000 diffractometer equipped with a Kevex solid-state detector that uses a Cu Kα radiation source with a wavelength of 1.54Å. The measurements were performed at 50 kV and 35 Ma. The data were recorded in reflection mode at a range of 2θ = 5–35° using a STEP scan mode. The step size was 0.02 degrees and the counting time was 2.0 s per step.

After being saturated by compressed CO₂, PLA samples tended to foam under some conditions during the gas release process. Cell morphologies of the foamed samples were observed with a JEOL JMS 6060 scanning electron microscope (SEM). Prior to the SEM observation, the samples were freeze-fractured in liquid nitrogen and sputter-coated with platinum. The volume expansion ratio of the polymer foams can be calculated according to Equation (1): (1) φ ρ ρ fwhere ρ and ρ_f are the mass densities of the samples before and after foaming, respectively, which were measured by the water displacement method in accordance with ASTM D792.