Physico chemical and phase separation characterization of high molecular PLLA blended with low molecular PCL obtained from solvent cast processes

Poly(L-lactide) (PLLA, intrinsic viscosity in chloroform: 3.8 dl g⁻¹, M_w = 320 kDa) was purchased from Evonik Industries AG (Darmstadt, Germany). Poly-ε-caprolactone (PCL, intrinsic viscosity in chloroform: 0.8 dl g⁻¹, M_w = 40 kDa) was supplied by LACTEL Absorbable Polymers (Birmingham, AL, USA). Chloroform was purchased from Baker (Fisher Scientific GmbH, Austria, Vienna). All chemicals were used as received without further purification.

For cast film production, polymer solutions (1 g Polymer in 25 ml chloroform) of PLLA, PCL and their blends in different ratios (90/10, 50/50, 10/90 w/w) were poured into glass petri dishes (Ø = 9 cm). After evaporation of the solvent under ambient conditions in a fume hood, films were washed for 2 days in methanol (LiChrosolv, Merck KGaA) with daily exchange, two days in distilled water (MilliQ Ultrapure, R = 18.2 MΩ.cm at 25 °C) with daily exchange, and dried for 7 days in a vacuum oven (Memmert GmbH + Co.KG) at 40 °C and 40 mbar. With this preparation method film samples with a thickness of ∼100 μm were achieved. For annealing experiments, samples of PLLA, PCL and their blends were placed in a chamber furnace and kept at different temperatures (40 °C, 80 °C and 200 °C) for 60 min at ambient atmosphere (KLS 05/13 Thermconcept). An annealing temperature of 40 °C was selected that corresponds to a temperature lower than the melting temperature (T_m) of PCL and PLLA. 80 °C were used due to the temperature lower than T_m of PLLA but higher than T_m of PCL. Furthermore 200 °C was selected, since this temperature exceeded the melting temperature for both polymers. Subsequently, all samples were cooled down with 1 K min⁻¹ in the furnace, see figure 2. The PLLA/PCL blend samples were kept in the laboratory refrigerator at 8 °C until the investigations were completed.

Zoom In Zoom Out Reset image size

SEM

AFM

Raman imaging

Uniaxial tensile testing was performed with a Zwicki ZN 2.5 (Zwick, Ulm, Germany). Tests were conducted with a 50 N load cell and a crosshead speed of 25 mm min⁻¹. During the procedure samples were kept at room temperature. Dumbbell-shaped tensile test specimens with effective dimensions of 12 mm × 2 mm × 0.1 mm were prepared with a blanking tool according to DIN EN ISO 527-2 1BB standards [17]. An image with the detailed dumbbell-shaped size is given in the supplementary data available online at stacks.iop.org/MRX/7/095302/mmedia. The tensile force as a function of elongation was measured. Based on these data the elastic modulus (E) was calculated in the linear elastic region. Furthermore, the tensile strength (σ_max) and elongation at break (ε_B) were determined. All results were averaged over n = 6 samples.

The thermal behavior of the polymer films was investigated in a DSC 1 system (Mettler Toledo, Schwerzenbach, Switzerland) using the conventional calibration methods with highly pure standards. The specimens were heated at a rate of 10 K min⁻¹ operating under nitrogen at atmospheric pressure. The samples were heated from −90 °C up to 200 °C. The sample weights were in the range of 3–6 mg. Samples were analyzed with respect to glass transition (T_g), melting temperature (T_m), and degree of crystallinity (χ). The T_g were determined by the midpoint of the inflection of the measured heating curve. The heats of fusion ΔH and crystallization χ_C were quantitatively evaluated by means of equation (1)

$$\begin{eqnarray}&&{\chi }_{c}=100 \% \cdot \left[\displaystyle \frac{{\rm{\Delta }}{H}_{m}}{{\rm{\Delta }}{H}_{m0}}\right]\cdot \displaystyle \frac{1}{W}\end{eqnarray} \tag{ 1 }$$

where ΔH_m0 is the enthalpy value of a pure crystalline material, ΔH_m is the enthalpy corresponding to the fusion process and W the amount of each homopolymer in the blend. The reference value taken for ΔH_m0 of PLLA was χ₁₀₀ = 93.7 J g⁻¹ [18] and for PCL χ₁₀₀ = 135.44 J g⁻¹ [19]. All results were averaged over n = 6 samples.

Figure 3 shows SEM documentation of the polymer morphology with increasing PCL content after 40 °C, 80 °C and 200 °C annealing temperature. Pure PLLA appears to have one uniform phase with spherulite crystals of about 20 μm at an annealing temperature of T = 40 °C (PLLA_T40) and T = 80 °C (PLLA_T80). In comparison to this, the radius of such spherulites of PLLA annealed at T = 200 °C (PLLA_T200) increased dramatically to about 100 μm. These results are comparable with the literature, which show that PLLA spherulites increase with increasing temperature [7, 20, 21]. The smaller crystal size of PLLA film obtained from solution compared to PLLA film from melt may be explained by the surrounding solvent and evaporation processes during thermal annealing. PLLA/PCL 90/10 blends show significant increase in crystal size with increasing annealing temperature. However, two separated phases could not be clearly observed.

Zoom In Zoom Out Reset image size

Interestingly, the surface morphology of PLLA/PCL 50/50 differs with different annealing conditions, see figure 3. From an initial smooth surface at T = 40 °C (PLLA/PCL 50/50_T40), the surface at T = 80 °C (PLLA/PCL 50/50_T80) changes to a perforated structure, which develops into two phases at T = 200 °C (PLLA/PCL 50/50_T200), a smooth and a spherulite-seeded structure. It is known from literature that a mixing gap of PLLA/PCL occurs from 80 °C and higher. When reaching mixing gap, PLLA spherulites are formed and small PCL crystals are deposited on the surface of the spherulite, thereby enhancing PLLA crystal growth [9]. This can also be found considering the solubility parameters of the two polymers which show a difference of approximately 0.9. Coleman et al describe a relatively weak favorable intermolecular interaction, if the difference between the solubility parameters of these polymer pairs is Δδ = 1.0 (cal cm⁻³)^0.5, respectively [16]. They conclude in their work, that for high molecular polymers Δδ should be much smaller than 1 to ensure that the components are miscible. The fact that the value of PLLA and PCL are δ_PLLA = 10.1 (cal cm⁻³)^0.5 and δ_PCL = 9.2 (cal cm⁻³)^0.5 and therefore Δδ = 0.9 suggests that the blend has only weak intermolecular interactions.

PLLA/PCL 10/90 shows large spherulites of about 300 μm at an annealing temperature of T = 40 °C (PLLA/PCL 10/90_T40) and T = 80 °C (PLLA/PCL 10/90_T80). The crystal shape of PLLA/PCL 10/90 at T = 200 °C (PLLA/PCL 10/90_T200) changes from circular to oval and the diameter decreased dramatically to about 50 μm.

Pure PCL exhibits remarkably smaller spherulites compared to PLLA/PCL 10/90 with diameters of about 200 μm at annealing temperatures of T = 40 °C (PCL_T40) and T = 80 °C (PCL_T80). The radii of spherulites of PCL at T = 200 °C (PCL_T200) increased dramatically to about 300 μm and the crystal height seems to regrade to a smooth surface.

Figure 4 shows a typical error, used for topography interpretation and height image of PLLA_T200 (image size 100 μm × 100 μm), which are used for the roughness evaluation R_t, see table 1. PLLA that was annealed at 200 °C forms spherulites with different sizes and heights.

Zoom In Zoom Out Reset image size

Figure 5 shows error signal images of the tested polymer samples with a picture size of 100 μm × 100 μm. The top row represents the polymers and their blends at an annealing temperature of T = 40 °C. In the middle, the samples are shown at T = 80 °C and below at T = 200 °C. Pure PLLA, annealed at 40 °C, crystallizes into small conical spherulites between 10 and 30 μm in size. The size of the crystals increases drastically at T = 200 °C. Based on the height profiles, an increase from 2.68 ± 0.14 μm to 4.2 ± 0.8 μm is determined. PLLA/PCL 90/10, annealed at 40 °C, exhibits a smoother substructure, which is pervaded with approximately 5 μm pores and suggestively of small PCL inclusions. Chloroform diffuses rapidly through the PCL, leading to higher chloroform concentrations in PCL areas, which leads to pore formation during evaporation. The pore size increases at an annealing temperature of 200 °C. Dell'Erba et al discovered that low interfacial tensions are obtained in binary PLLA/PCL blends due to the similar chemical nature of the blends components, which allows interpolymer polar interactions across phase boundaries [22]. This can be also observed at PLLA/PCL 50/50. In PLLA/PCL 50/50, annealed at 40 °C, lamellar structures in circular arrangements of about 60 μm in size are embedded in a smooth surface. In addition, approximately pores with a diameter of 5 μm can be observed. With PLLA/PCL 50/50_T80, the surface becomes very porous exhibiting pore diameters up to 50 μm. The height difference increases drastically to R_t = 16.2 ± 0.9 μm at T = 80 °C. Similar SEM images and pore sizes have been observed after solvent etching of PLLA/PCL blends [23]. This may be due to the melting of PCL in contrast to PLLA at the thermal processing step at 80 °C due to its lower melting temperature. Hence, PCL is liquefied, similar to the etching process where PCL is solved. As a result of this liquefying process a highly porous structure mainly built by PLLA would be expected. Such pores are not observed at T = 40 °C, where PCL remains solid and at T = 200 °C, where both polymers are melted, which results in a smooth surface without pores. For the PLLA/PCL 10/90_T80 also a smooth surface is expected, due to the low amount of PLLA, witch rather forms particles than a connected structure. At T = 200 °C height difference drops down to 7.9 ± 0.5 μm. Clearly, two types of structures are visible: First, a smooth surface combined with a lamellar structure which corresponds to PCL crystals was observed in accordance to [22]. However, PLLA/PCL 10/90 exhibits a rough surface that can be interpreted as partial segments of spherulites. In part, the crystal boundaries are still visible. The height values (seen in table 1) show the same results of approx. 5.7 μm at all annealing temperatures. A similar picture is visible with pure PCL. When applying an annealing temperature of 200 °C, the structure changed. Plate-shaped structures were formed on a generally rough base structure. The AFM results obtained are consistent with the SEM images.

Zoom In Zoom Out Reset image size

Confocal Raman imaging was used in order to detect the phase separation in the different polymeric blends. In figure 6 the Raman images received by mapping the PLLA/PCL samples and their blends are shown. The corresponding Raman spectra are given in the supplementary data. In particular, the yellow area is related to the intensity of PLLA signal from the band centered at 878 rel. cm⁻¹. The red area is related to the intensity of the PCL signal, centered in the band at 1727 rel. cm⁻¹. The dark regions of the Raman images correspond to regions out of focus plane. From this analysis it becomes clear that phase separated morphology exists for all the three different blends, and it is possible to detect both, the PLLA-rich phase and the PCL-rich phase. For PLLA/PCL 90/10 it can be supposed that PLLA forms a continuous phase with dispersed spheres of PCL with different sizes (red regions of figure 6). Most interestingly, for PLLA/PCL 50/50 _T40 and PLLA/PCL 50/50_T80, two well differentiated areas of each homopolymer are detected, indicating the existence of large domains of each homopolymer. It is known from the literature that if polymer components are being phase separated, the interface width (w) can be related to the Flory-Huggins parameters (χ) by a general relation in the form w ∼ (χ−χcrit)^−0.5 [14]. That means for our PLLA/PCL system a narrow interfacial width indicates relatively low interface interactions, which is consistent with the reported results of incompatibility between PCL and PLA [24]. López et al noted that in phase-separated PLA/PCL blends sharp interfaces are existing in the microstructure, whereas Broz et al found poor adhesion at the phase boundary interface and confirmed immiscibility and phase separation through NMR [6, 9]. It was hypothesized that to improve the mechanical properties of the blend, the samples should be annealed in the single-phase region of the LCST phase-diagram to enhance interfacial adhesion. It was concluded that interfacial adhesion may occur when the majority phase is PCL [25]. This could be also seen at T = 200 °C, but it is a more dominant PLLA phase in that case. Finally, PLLA/PCL 10/90_T200 presents the PCL continuous region with small irregular PLLA domains of different sizes. At T = 40 °C, red regions indicating PCL phases are overlayed with yellow phases of PLLA, hence the areas appear orange. Thus, Raman maps are well related with the morphology previously received from the SEM and AFM images.

Zoom In Zoom Out Reset image size

The mechanical properties of the PLLA/PCL blends are plotted in figure 7. In this graphic, pure PLLA shows a stiff and brittle behavior, but after annealing to a temperature of 40 °C PLLA shows an elongation at break of 60 ± 18%. This is probably due to the chain alignment during solvent evaporation [26]. Annealing above melting point (200 °C) reduces the elongation at break of pure PLLA down to a value of approximal 3%, which is known from literature [27].

Zoom In Zoom Out Reset image size

PCL shows, in contrast to the literature, a brittle behavior. This can be associated with its high crystallinity, as described later in this work. Due to the washing and annealing processes, PCL reached such a high degree of crystallization and as a result the mechanical properties varied. Comparing the elongation values of PCL at 80 °C and 200 °C [6], it is evident that the crystallization values are significantly lower and consequently the elongation at break increases. Annealing of pure PCL above the melting temperature led to the degradation of the samples, which is why no mechanical properties could be determined. Addition of 10% PCL to PLLA and annealing to 40 °C results in a substantial increase in elongation at break, decreasing again with further addition of PCL (figure 7). From the literature, it is known from blending PLLA with PEG, that there is probably still an adhesion between PLLA and PEG based on the crystallization kinetics [27]. After annealing at 80 °C the 90/10 PLLA/PCL blend reveals the same effect, which indicates that the temperature has no significant effect on the PLLA/PCL boundary interfaces. Table 1 shows all mechanical properties of PLLA/PCL obtained by solvent casting, followed by an annealing process. The corresponding stress strain curves are given in the supplementary data. With further addition of PCL tensile strength (table 1) and elastic modulus decreases, which is comparable with the reported results [6, 9]. In summary, blending can substantially change the mechanical properties by influencing factors such as crystallization kinetics. To address this issue, thermal properties of the blends were investigated.

As typical examples, figure 8 shows the DSC traces of PLLA, PCL, and their blends in different ratios (90/10, 50/50, 10/90 w/w) of the cast film samples recorded by the first heating scan. The DSC spectrum of pure PLLA film exhibits a glass transition temperature at approximal 67 °C, see table 2 for all annealing temperatures and are consistent compared to literature values [26]. No exothermic peak, but an endothermic melting peak at 179 °C for an annealing temperature of 40 °C and 80 °C is observed. At 200 °C the melting peak of pure PLLA decreases to 173.4 ± 1.4 °C. With increasing annealing temperature two PLLA melting peaks are formed, see figure 8(c). The DSC spectrum of pure PCL film exhibits a glass transition temperature at −64.66 ± 0.03 °C for an annealing temperature of 40 °C (figure 8(a)) and increases with increasing annealing temperature up to −61,6 ± 1,9 °C, see table 2. Pure PCL and all blends show a different melting behavior at 200 °C compared to annealing temperatures of 40 °C and 80 °C. For an annealing temperature of 200 °C, two melting peaks are formed for PCL (figure 8(c)). The appearance of two PCL and two PLLA melting peaks might be an indication of formation of mixed crystalline structures. There are some explanations for the PCL and PLLA double melting peaks phenomenon in the literature. Finotti et al state that the lower peak temperature can be associated with the partial migration of PLLA molecules to the PCL phase. The low molecular weight allows the diffusion of PLLA from the PLLA/PCL interface to the interior of the PCL rich-phase, decreasing the size and degree of perfection of the crystalline lamellae, hence lowering the T_M of the PCL. They also described the second melting peak at 56 °C as typical of pure PCL [28]. Regarding PLLA, Zhou et al stated the possibility to crystallize into two different crystal structures, the α-form, melting at higher temperature and the β-form, melting at lower temperature [29]. They also pointed out that the presence of the two endothermic peaks can also be explained by the growth of small crystals, which at low heating rates melt and recrystallize before melting again. Shan et al suggested the higher temperature peak may be a result of the melting of the thicker lamellae, possibly generated during the production process. In contrast, the lower temperature peak is attributed to the crystallization of the sample during the annealing process [30].

Zoom In Zoom Out Reset image size

Table 2 shows the peak temperatures T_m and the T_g of PLLA and PCL as determined from the DSC curves as well as the percentage PLLA and PCL crystallinity, χ_PLLA (%) and χ_PCL (%) values as determined according to equation (1). Heat of fusion (ΔH) of PLLA, PCL and their blends are given in the supplementary data. T_g is not discussed here due to insufficient heating rate (10 K min⁻¹) and some overlay characteristics.

Thus, it can be concluded that blending PLLA with PCL changes the inherent crystallization behaviors of both polymers in combination. χ_PLLA at T₄₀ is about 39.3 ± 1.7% and increases with increasing annealing temperature up to 62.8 ± 0.9%. The crystallinity of PLLA_T80 in the blend decreases at PLLA/PCL 90/10 and increases with higher PCL content. This can be explained in a way that PCL separates from the PLLA crystal due to the PLLA/PCL mixing gap, which is known from the literature [31]. This approaches the original crystallinity of PLLA 100. At T₂₀₀ χ_PLLA in the 10/90 blend decreases from 62 ± 4% to 43 ± 8%. Other studies report an increase in the crystallinity of PLLA in the presence of PCL obtained from melt processes, probably due to the increase in nucleation rate [22, 32, 33]. However, this effect could not be observed in this work which may be explained by the fact that a solvent casting process was chosen in this paper. This has a different effect on the crystallization mechanism of PLLA and underlines the impact of the manufacturing process of such blend materials. The trend of increasing crystallinity with increasing PCL content in the blend can be seen in the case of all applied annealing temperatures, in a way that the high temperature thermal treatment at 200 °C approximates melt cast conditions.

Annealing at 200 °C is comparable to the thermally induced equilibration from melting process, which is widely applied for the investigation of crystalline polymers. Notably, it can be seen in figure 9 that the degree of crystallinity of PCL gradually decreases with increasing PLLA content, while the crystallinity of PLLA increases slowly with increasing PCL content (see table 2). Liu et al showed that a higher crystallization rate coefficient (CRC) of the blend in comparison to pure PLLA causes faster crystallization [34]. Therefore, χ_PLLA increases with increasing PCL content. Clearly, high PCL content substantially decreases the crystallinity of PLLA/PCL blends. This means that the large nucleus formation induced by PCL molecules inhibit the mobility of PLLA chains. As a result, chain rearrangement is limited, so low χ_PLLA are noted. Our assumption is that similar processes on the surface of the PLLA nuclei lead to a decrease in the crystallinity of the PCL when raising the PLLA concentration. Due to the increase in the annealing temperature, the crystal sizes of both polymers may decrease, resulting in a decrease in the melting temperature.

Zoom In Zoom Out Reset image size

In general, solvent cast processes followed by low thermal annealing lead to substantially different properties of the yielded films in comparison to melt cast process data from the literature [22, 35, 36]. In addition to thermal equilibration, solvent evaporation plays an important role and can lead to different crystal growth rates compared to the cooling rate.