Miscibility, Morphology and Crystallization Behavior of Poly(Butylene Succinate-co-Butylene Adipate)/Poly(Vinyl Phenol)/Poly(l-Lactic Acid) Blends

Amorphous poly(vinyl phenol) (PVPh) is introduced into poly(butylene succinate-co-butylene adipate)/poly(l-lactic acid) (PBSA/PLLA) blends via solution casting. Fourier transform infrared spectroscopy (FTIR) analysis verifies that intermolecular hydrogen bonding formed in PBSA/PVPh/PLLA blends. The miscibility between PBSA and PLLA is improved with PVPh incorporation as evidenced by approaching Tgs of the two components. When PVPh content reaches up to 50 wt %, the blend sample exhibits only one Tg, meaning complete miscibility between PBSA and PLLA. The improved miscibility of PBSA/PLLA blends is further confirmed by scanning electron microscope (SEM). Typical “see-island” phase separation structure for PBSA/PLLA blend transforms into homogenous phase structure for blend samples with 5 wt % PVPh and above. Non-isothermal crystallization analysis shows that the crystallization temperature and crystallization enthalpy of PBSA decrease with PVPh addition, and those of PLLA also show a decreasing trend. Isothermal crystallization rate of PBSA in blend samples distinctly decreases with PVPh incorporation, whereas that of PLLA in blend samples increases slightly with PVPh addition. Wide angle X-ray diffraction (WAXD) analysis indicated that PLLA in blend samples remained partly crystallized, while PBSA turned into amorphous state with increasing PVPh contents.


Introduction
The indiscriminate use of fossil-based plastics had produced severe environmental problems, which made it urgent to find biodegradable substitutes. Poly(L-lactic acid) (PLLA) as a biodegradable aliphatic polyester derived from biomass resources has attracted intensive attention and already been applied in many fields such as medicine and packaging. To further expand its application range, however, much more efforts are still required to overcome some shortcomings including low crystallization rate, low toughness and poor heat resistance [1,2].
Many works aimed at improving the performances of PLLA have been reported. Among these, blending is a favorable method due to low cost and simple operation. Some works involve blending nucleating agents into PLLA to accelerate its crystallization rate because the degree of crystallinity, crystal size and morphology significantly influence the microstructure of materials and cause changes in mechanical and thermal properties [3][4][5][6]. While some other works are focused on blending PLLA with flexible polymers, flexible biodegradable polyesters, such as poly(ε-caprolactone)

Scanning Electron Microscope (SEM)
All samples were etched in THF at room temperature for 12 h to dissolve PVPh for higher contrast. The films were then dried and fractured. The fracture surfaces were used for observation with a Hitachi 5-3400 scanning electron microscope (Hitachi, Tokyo, Japan) after sputter coated with gold.

FTIR Analysis
It had already been reported that the hydroxyl groups of PVPh had the capability to form intermolecular hydrogen bonding with carbonyl groups of PBSA and PLLA in binary PVPh/PLLA and PVPh/PBSA blends, respectively [14][15][16][17]. Combined with the chemical structure of these polymers, it would be reasonable to anticipate that the carbonyl groups of PBSA and PLLA will form intermolecular hydrogen bonding interaction with hydroxyl groups of PVPh. To verify the intermolecular hydrogen bonding in PBSA/PVLh/PLLA blends, the FTIR spectra of samples in the carbonyl stretching vibration region were analyzed using a peak fitting program (Peakfit, V4.12, Seasolve Software Inc., San Jose, CA, USA). This program detects hidden peaks by second derivative minima. It adjusts the peak location, line shape, peak width and height based on a least-squares parameter-adjustment criterion using the Gaussian amplitude function to obtain the best fit [19,20].
The curve fitting result for PVPh-10 in the carbonyl stretching vibration region is shown in Figure 1 as an example. It can be seen that the fitted spectrum is highly coincident with the experimentally obtained one, proving that this peak fitting program is applicable. The two strong absorption bands at 1760 and 1718 cm −1 are attributed to carbonyl stretching vibration in PLLA and PBSA, respectively. The band at 1760 cm −1 is resolved into three hidden peaks at 1777, 1760 and 1745 cm −1 , which are ascribed to amorphous, crystalline and hydrogen bonded carbonyl vibration absorption in PLLA chains [21]. The band at 1718 cm −1 is also resolved into three hidden peaks at 1732, 1716 and 1695 cm −1 , which are ascribed to amorphous, crystalline and hydrogen bonded carbonyl vibration absorption in PBSA chains [20]. The area of each hidden peaks were also derived to estimate the fraction of amorphous, crystalline and hydrogen bonded carbonyl groups in PLLA and PBSA chains. That is, f i = A i /∑A, where f i is the fraction of component i, A i is the area of resolved hidden peak corresponding to component i, and ∑A is the total peak area. The estimated fractions of each component in blend samples are listed in Table 1. It is interesting to find that the fraction of hydrogen bonded carbonyl groups in PLLA fluctuates in the small range of 12.0%-13.2%, while that in PBSA increases from 2.6%-4.0% with increasing PVPh content.

Miscibility of PBSA/PVPh/PLLA Blends
The miscibility of the blending components in an amorphous state can be assessed based on glass transition behavior of blends, provided that the difference in Tgs of the two components is more than 20 °C [22]. A fully miscible polymer blend exhibits only one Tg, whereas completely immiscible blends usually have two independent Tgs close to those of the individual component. If the Tgs of components shift to each other after blending, the two components are partially miscible. Figure 2 shows the DSC heating curves for melt-quenched samples. As can be seen from the DSC curve for PVPh-0, the Tgs for PBSA and PLLA in blends were around −37.7 °C at 51.6 °C, respectively, which are close to those of neat components [19,20,23]. This suggests that PBSA and PLLA are completely immiscible in the absence of PVPh, which is in accordance with that reported in literature [10]. PVPh is an amorphous polymer with high glass transition temperature of about 183 °C as shown in Figure  2b. The Tgs for PBSA increased with PVPh addition, while the Tgs for PLLA remained almost constant. The approaching Tgs indicate that the miscibility of PBSA and PLLA were improved. When 50 wt % of PVPh was incorporated into a PBSA/PLLA blend, only one single Tg around 45.0 °C could be observed, implying that PBSA and PLLA are miscible in an amorphous state. The Tg1 for PBSA and Tg2 for PLLA are listed in Table 2.

Miscibility of PBSA/PVPh/PLLA Blends
The miscibility of the blending components in an amorphous state can be assessed based on glass transition behavior of blends, provided that the difference in T g s of the two components is more than 20 • C [22]. A fully miscible polymer blend exhibits only one T g , whereas completely immiscible blends usually have two independent T g s close to those of the individual component. If the T g s of components shift to each other after blending, the two components are partially miscible. Figure 2 shows the DSC heating curves for melt-quenched samples. As can be seen from the DSC curve for PVPh-0, the T g s for PBSA and PLLA in blends were around −37.7 • C at 51.6 • C, respectively, which are close to those of neat components [19,20,23]. This suggests that PBSA and PLLA are completely immiscible in the absence of PVPh, which is in accordance with that reported in literature [10]. PVPh is an amorphous polymer with high glass transition temperature of about 183 • C as shown in Figure 2b. The T g s for PBSA increased with PVPh addition, while the T g s for PLLA remained almost constant. The approaching T g s indicate that the miscibility of PBSA and PLLA were improved. When 50 wt % of PVPh was incorporated into a PBSA/PLLA blend, only one single T g around 45.0 • C could be observed, implying that PBSA and PLLA are miscible in an amorphous state. The T g1 for PBSA and T g2 for PLLA are listed in Table 2.
that the introduction of PVPh retarded crystallization of PBSA. The crystallization rate of PBSA is too fast in PVPh-0, so it crystallized to a relatively high crystallinity degree during the rapid cooling process. Remaining crystallizable segments completed crystallization following the heating process at a rate of 10 °C/min, which formed a weak exothermic peak on the DSC curve. For PBSA in PVPh-10, the crystallization was somewhat impeded due to intermolecular hydrogen bonding formation and did not crystallize or crystallized very little after rapid cooling, resulting in a highly amorphous state. Cold crystallization occurred during following heating process and formed a stronger exothermic peak in a higher temperature range than those of PBSA in PVPh-0. With further PVPh addition, the crystallization of PBSA in blend samples was largely hindered and did not crystallize in rapid cooling and subsequent heating process.  SEM micrographs can provide further information about miscibility and microstructure of blend samples. Figure 3 shows the fracture surface morphologies for PBSA/PVPh/PLLA blends. A typical "sea-island" phase separation structure was observed in SEM photographs for blend without PVPh incorporation, which means poor miscibility between PBSA and PLLA. With 5 wt % PVPh incorporation, however, the morphologies of PBSA/PVPh/PLLA blends moved into the homogenous phase. Such an effect became more evident when PVPh content further increased to 10 wt % and above, for example, 40 wt % as presented in Figure 3. These results again verify that the miscibility between PBSA and PLLA was improved. In addition, more ductile fractures were observed for samples with higher PVPh content compared with those without PVPh incorporation, which suggests that the addition of PVPh may result in better toughness or ductility.   It is also interesting to note that a weak exothermic peak around −6.0 • C and a much stronger one at 18.0 • C were present on DSC curves for PVPh-0 and PVPh-10, respectively, which could be attributed to cold crystallization of PBSA component in blends. However, no discernible peaks were observed at corresponding locations for samples with higher PVPh content. These can be explained that the introduction of PVPh retarded crystallization of PBSA. The crystallization rate of PBSA is too fast in PVPh-0, so it crystallized to a relatively high crystallinity degree during the rapid cooling process. Remaining crystallizable segments completed crystallization following the heating process at a rate of 10 • C/min, which formed a weak exothermic peak on the DSC curve. For PBSA in PVPh-10, the crystallization was somewhat impeded due to intermolecular hydrogen bonding formation and did not crystallize or crystallized very little after rapid cooling, resulting in a highly amorphous state. Cold crystallization occurred during following heating process and formed a stronger exothermic peak in a higher temperature range than those of PBSA in PVPh-0. With further PVPh addition, the crystallization of PBSA in blend samples was largely hindered and did not crystallize in rapid cooling and subsequent heating process.
SEM micrographs can provide further information about miscibility and microstructure of blend samples. Figure 3 shows the fracture surface morphologies for PBSA/PVPh/PLLA blends. A typical "sea-island" phase separation structure was observed in SEM photographs for blend without PVPh incorporation, which means poor miscibility between PBSA and PLLA. With 5 wt % PVPh incorporation, however, the morphologies of PBSA/PVPh/PLLA blends moved into the homogenous phase. Such an effect became more evident when PVPh content further increased to 10 wt % and above, for example, 40 wt % as presented in Figure 3. These results again verify that the miscibility between PBSA and PLLA was improved. In addition, more ductile fractures were observed for samples with higher PVPh content compared with those without PVPh incorporation, which suggests that the addition of PVPh may result in better toughness or ductility.

Non-Isothermal Crystallization and Melting Behavior of PBSA/PVPh/PLLA Blends
Non-isothermal crystallization DSC curves for PBSA/PVPh/PLLA blends are presented in Figure  4. The non-isothermal crystallization parameters are listed in Table 2. As can be seen from Figure 4, the crystallization peaks of PBSA (Tc1) and PLLA (Tc2) are around 50.9 °C with a ΔHc1 of 21.7 J/g and 99.9 °C with a ΔHc2 of 16.8 J/g for PVPh-0, respectively, which are close to those of neat components [20,23]. The crystallization peak for PBSA shifts to 40.1 °C with a ΔHc1 of 16.6 J/g when 10 wt % PVPh was added into PBSA/PLLA blends. With further PVPh incorporation, however, the crystallization peaks for PBSA disappear on cooling DSC curves, implying that the crystallization of PBSA was completely impeded due to the formation of intermolecular hydrogen bonding and the dilution effect [14][15][16]. Tc2 almost remains constant in blends with 20 wt % and less PVPh incorporation as shown in Figure 4. When further increasing PVPh content, Tc2 decreases within a narrow range. The crystallization enthalpy of PLLA decreases with PVPh addition as presented in Table 2.

Non-Isothermal Crystallization and Melting Behavior of PBSA/PVPh/PLLA Blends
Non-isothermal crystallization DSC curves for PBSA/PVPh/PLLA blends are presented in Figure 4. The non-isothermal crystallization parameters are listed in Table 2. As can be seen from Figure 4, the crystallization peaks of PBSA (T c1 ) and PLLA (T c2 ) are around 50.9 • C with a ∆H c1 of 21.7 J/g and 99.9 • C with a ∆H c2 of 16.8 J/g for PVPh-0, respectively, which are close to those of neat components [20,23]. The crystallization peak for PBSA shifts to 40.1 • C with a ∆H c1 of 16.6 J/g when 10 wt % PVPh was added into PBSA/PLLA blends. With further PVPh incorporation, however, the crystallization peaks for PBSA disappear on cooling DSC curves, implying that the crystallization of PBSA was completely impeded due to the formation of intermolecular hydrogen bonding and the dilution effect [14][15][16]. T c2 almost remains constant in blends with 20 wt % and less PVPh incorporation as shown in Figure 4. When further increasing PVPh content, T c2 decreases within a narrow range. The crystallization enthalpy of PLLA decreases with PVPh addition as presented in Table 2.

Non-Isothermal Crystallization and Melting Behavior of PBSA/PVPh/PLLA Blends
Non-isothermal crystallization DSC curves for PBSA/PVPh/PLLA blends are presented in Figure  4. The non-isothermal crystallization parameters are listed in Table 2. As can be seen from Figure 4, the crystallization peaks of PBSA (Tc1) and PLLA (Tc2) are around 50.9 °C with a ΔHc1 of 21.7 J/g and 99.9 °C with a ΔHc2 of 16.8 J/g for PVPh-0, respectively, which are close to those of neat components [20,23]. The crystallization peak for PBSA shifts to 40.1 °C with a ΔHc1 of 16.6 J/g when 10 wt % PVPh was added into PBSA/PLLA blends. With further PVPh incorporation, however, the crystallization peaks for PBSA disappear on cooling DSC curves, implying that the crystallization of PBSA was completely impeded due to the formation of intermolecular hydrogen bonding and the dilution effect [14][15][16]. Tc2 almost remains constant in blends with 20 wt % and less PVPh incorporation as shown in Figure 4. When further increasing PVPh content, Tc2 decreases within a narrow range. The crystallization enthalpy of PLLA decreases with PVPh addition as presented in Table 2.

AvramiAnalyses for Isothermal Crystallization of PBSA/PVPh/PLLA Blends
PVPh-0, PVPh-5 and PVPh-10 were selected for isothermal crystallization analysis because samples with higher PVPh content crystallized too slowly or did not crystallize as a result of hydrogen bonding formation. The Avrami equation was employed to analyze the isothermal crystallization of blend samples with 0, 5 and 10 wt % of PVPh in the following form [24]: where X t is the relative crystallinity degree at time t, k is the crystallization rate constant and n is the Avrami exponent dependent on crystallization mechanism and crystal growth dimensions. Using the double logarithmic form of Equation (1), one can obtain k and n values from slopes and intercepts of plots of ln[−ln(1 − X t )] versus lnt, respectively. The Avrami plots for isothermal crystallization of PBSA in blend samples at designated temperatures are presented in Figure 5. As can be seen, the plots of ln[−ln(1 − X t )] against lnt generally show good linear correlation, which means that the Avrami theory was applicable in the isothermal crystallization of PBSA in PBSA/PVPh/PLLA blends. However, it is also noticeable that deviations from Avrami plots occur in the later stages of crystallization, which is ascribed to spherulite impingement [20,25]. The Avrami parameters were obtained using the linear portion of plots of ln[−ln(1 − X t )] versus lnt and are listed in Table 3. Avrami exponent n values listed in Table 3 were decimals between 2.3-2.7, indicating that crystalline branching and/or two stage crystal growth might present in the crystallization process, and the PBSA spherulites assumed a two-dimensional to three-dimensional growth with a combination of homogeneous and heterogeneous nucleation [20].

AvramiAnalyses for Isothermal Crystallization of PBSA/PVPh/PLLA Blends
PVPh-0, PVPh-5 and PVPh-10 were selected for isothermal crystallization analysis because samples with higher PVPh content crystallized too slowly or did not crystallize as a result of hydrogen bonding formation. The Avrami equation was employed to analyze the isothermal crystallization of blend samples with 0, 5 and 10 wt % of PVPh in the following form [24]: where Xt is the relative crystallinity degree at time t, k is the crystallization rate constant and n is the Avrami exponent dependent on crystallization mechanism and crystal growth dimensions. Using the double logarithmic form of Equation (1), one can obtain k and n values from slopes and intercepts of plots of ln[−ln(1 − Xt)] versus lnt, respectively. The Avrami plots for isothermal crystallization of PBSA in blend samples at designated temperatures are presented in Figure 5. As can be seen, the plots of ln[−ln(1 − Xt)] against lnt generally show good linear correlation, which means that the Avrami theory was applicable in the isothermal crystallization of PBSA in PBSA/PVPh/PLLA blends. However, it is also noticeable that deviations from Avrami plots occur in the later stages of crystallization, which is ascribed to spherulite impingement [20,25]. The Avrami parameters were obtained using the linear portion of plots of ln[−ln(1 − Xt)] versus lnt and are listed in Table 3. Avrami exponent n values listed in Table 3 were decimals between 2.3-2.7, indicating that crystalline branching and/or two stage crystal growth might present in the crystallization process, and the PBSA spherulites assumed a two-dimensional to three-dimensional growth with a combination of homogeneous and heterogeneous nucleation [20].
ln t ln t PVPh-10  The crystallization half time (t 1/2 ) was derived from the following equation to investigate effects of PVPh addition on the crystallization rate: The crystallization half time of PBSA (t 1/2 ) isothermally crystallized at designated temperatures (T c s) presented in Table 3 increases notably with weight fraction of PVPh and T c s, which means that the addition of PVPh decreases the crystallization rate of PBSA in PBSA/PVPh/PLLA blends. Avrami analyses for isothermal crystallization of PLLA in blend samples were performed using similar methods. The Avrami plots for PLLA isothermal crystallization in blend samples are shown in Figure 6. Deviations in the later stages of crystallization are also interpreted as spherulite impingement [20,25]. Avrami parameters for PLLA isothermal crystallization in blend samples are presented in Table 4. The values of Avrami exponent n for PLLA isothermal crystallization are between 2.6-2.8 and very close to 3, implying that the PLLA spherulites assumed three-dimensional growth with heterogeneous nucleation [19]. t 1/2 for PLLA isothermally crystallized at indicated T c s listed in Table 4 decreases slightly with PVPh addition, which means that the crystallization rate of PLLA was also increased. impingement [20,25]. Avrami parameters for PLLA isothermal crystallization in blend samples are presented in Table 4. The values of Avrami exponent n for PLLA isothermal crystallization are between 2.6-2.8 and very close to 3, implying that the PLLA spherulites assumed three-dimensional growth with heterogeneous nucleation [19]. t1/2 for PLLA isothermally crystallized at indicated Tcs listed in Table 4 decreases slightly with PVPh addition, which means that the crystallization rate of PLLA was also increased.

WAXD Analysis
WAXD patterns for PBSA/PVPh/PLLA blend samples are shown in Figure 7. The diffraction peaks at 2θ = 16.5°, 18.8°, 28.8° are attributed to (110), (111), (300) crystallographic plane of PLLA crystal in α′ form, respectively [26,27]. However, those at 2θ = 19.4°, 21.9°, 22.5° are ascribed to (020), (021), (110) crystallographic plane of PBSA, respectively [23]. The diffraction peaks for PLLA remained relatively sharp for samples with all investigated PVPh content, suggesting that PLLA was partly crystallized. However, the diffraction peaks for PBSA vanished gradually with increasing PVPh content, indicating that PBSA turned into a fully amorphous state when introducing high content of PVPh into PBSA/PLLA blends. These results are in accordance with preceding nonisothermal crystallization analysis. This is a very interesting phenomenon because it seems that PVPh in blends "chooses" to interact more with PBSA than with PLLA. However, investigations about possible reasons are beyond the scope of this work.

WAXD Analysis
WAXD patterns for PBSA/PVPh/PLLA blend samples are shown in Figure 7. The diffraction peaks at 2θ = 16.5 • , 18.8 • , 28.8 • are attributed to (110), (111), (300) crystallographic plane of PLLA crystal in α form, respectively [26,27]. However, those at 2θ = 19.4 • , 21.9 • , 22.5 • are ascribed to (020), (021), (110) crystallographic plane of PBSA, respectively [23]. The diffraction peaks for PLLA remained relatively sharp for samples with all investigated PVPh content, suggesting that PLLA was partly crystallized. However, the diffraction peaks for PBSA vanished gradually with increasing PVPh content, indicating that PBSA turned into a fully amorphous state when introducing high content of PVPh into PBSA/PLLA blends. These results are in accordance with preceding non-isothermal crystallization analysis. This is a very interesting phenomenon because it seems that PVPh in blends "chooses" to interact more with PBSA than with PLLA. However, investigations about possible reasons are beyond the scope of this work.

Conclusions
PVPh was introduced into PBSA/PLLA blends via solution casting to improve the miscibility between biodegradable PBSA and PLLA. It was proved by FTIR analysis that PVPh had formed intermolecular hydrogen bonding with PBSA and PLLA, which acted as crosslinking agents between PBSA and PLLA. The miscibility between PBSA and PLLA were improved with PVPh introduction

Conclusions
PVPh was introduced into PBSA/PLLA blends via solution casting to improve the miscibility between biodegradable PBSA and PLLA. It was proved by FTIR analysis that PVPh had formed intermolecular hydrogen bonding with PBSA and PLLA, which acted as crosslinking agents between PBSA and PLLA. The miscibility between PBSA and PLLA were improved with PVPh introduction evidenced by approaching T g s of the two components in DSC analyses. When PVPh content reached up to 50 wt %, the blend sample exhibited only one T g , meaning complete miscibility between PBSA and PLLA. The improved miscibility of PBSA/PLLA blends was further confirmed by SEM. Typical "sea-island" phase separation structure for PBSA/PLLA blends transformed into one phase structure for PBSA/PVPh/PLLA blend samples. In addition, the melt crystallization peak of PBSA in blend samples shifted to lower temperatures with PVPh addition and vanished when adding 20 wt % PVPh and above. PLLA also shows a decreasing trend. Avrami analyses suggested that crystallization of PBSA in blend samples assumed two-to three-dimensional crystal growth with a combination of homogeneous and heterogeneous nucleation, while that of PLLA assumed three-dimensional crystal growth with homogeneous nucleation. The crystallization rate of PBSA was distinctly decreased by PVPh incorporation. On the contrary, the crystallization rate of PLLA increased slightly with PVPh addition. WAXD analysis indicated that PLLA in blend samples remained partly crystallized, while PBSA turned into an amorphous state with increasing PVPh content.