J. Phys. Chem. B 2005, 109, 19175-19183
19175
Miscibility and Hydrogen-Bonding Interactions in Biodegradable Polymer Blends of Poly(3-hydroxybutyrate) and a Partially Hydrolyzed Poly(vinyl alcohol) He Huang,*,†,‡ Yun Hu,† Jianming Zhang,† Harumi Sato,† Hongtao Zhang,§ Isao Noda,| and Yukihiro Ozaki*,† Department of Chemistry, School of Science and Technology and Research Center for EnVironmental Friendly Polymers, Kwansei-Gakuin UniVersity, Gakuen, Sanda 669-1337, Japan, College of Materials Science & Engineering, Donghua UniVersity, Shanghai 200051, China, School of Chemistry and Materials Science, Hubei UniVersity, Wuhan 430062, China, and The Procter & Gamble Company, 8611 Beckett Road, West Chester, Ohio 45069 ReceiVed: June 15, 2005; In Final Form: August 23, 2005
Miscibility and hydrogen-bonding interactions, as well as the morphological properties, of biodegradable polymer blends of poly(3-hydroxybutyrate) (PHB) and a 80% hydrolyzed poly(vinyl alcohol) (PVA80) were studied using Fourier transform infrared spectroscopy (FTIR) and differential scanning calorimetry (DSC). It was found that PHB is miscible with PVA80 in the amorphous phase over the whole composition range. PVA80 or PHB assumes the amorphous state when its content in the blend is lower than 30 or 20 wt %, respectively. Due to the heavy overlapping of CdO stretching bands from both PVA80 and PHB and the nonmeasurable peak shift in the OH stretching band region, hydrogen-bonding interactions between the OH group of PVA80 and the CdO group of PHB were not detectable at room temperature, but were observed at a higher temperature of 180 °C. This is because hydrogen-bonding interactions are promoted above the melting points of these two crystalline polymers, by increasing the mixing entropy and reducing the ∆χ effect. Blending PHB with PVA80 does not have a significant effect on the OH groups of PVA80 that are hydrogen bonded with each other. Instead, the CdO groups of PHB dispossess some of the OH groups that are hydrogen bonded to the CdO groups of PVA80, which gives rise to the miscibility between PVA80 and PHB in the amorphous phase.
Introduction Polyhydroxyalkanoates (PHAs) are biodegradable high molecular weight aliphatic copolyesters that can be produced in microorganisms. Poly(3-hydroxybutyrate) (PHB), which has similar thermal and mechanical properties to isotactic polypropylene, is the most abundant polyester found in bacteria and also the most extensively studied one in the family of PHA.1-7 Nevertheless, PHB is still far from finding a broad range of applications, even with its intriguing potential as an environment-friendly polymeric material. This is because PHB is rigid and brittle due to its excessively high crystallinity. In addition, PHB is difficult to handle with the conventional melt processing, due to its high melting temperature, which is close to the thermal decomposition temperature. Furthermore, the relatively high price, compared to conventional synthetic polymers, might be one of the major current restrictions to its practical applications. Therefore, different approaches have been explored to improve the performance of PHB polymeric material, including copolymerization and blending. Examples of PHB copolymers include those containing 3-hydroxyvalerate,8-14 4-hydroxybutyrate,15-18 5-hydroxyvalerate,19 and hydroxypropionate20,21 units. However, these copolymers are still in their developmental stage and are usually much more expensive than conventional * Corresponding author. E-mail:
[email protected] (Y.O.) or
[email protected] (H.H.). † Kwansei-Gakuin University. ‡ Donghua University. § Hubei University. | The Procter & Gamble Co.
plastics. Compared to the copolymerization method, blending may be a much easier and faster way to achieve the desired properties. More importantly, through blending, other less expensive polymers could be incorporated with PHB. Many blends containing PHB have been studied, such as blends with poly(ethylene oxide),22-30 cellulose derivatives,31-35 poly(vinyl phenol),34-39 and poly(vinyl acetate).40-45 Poly(vinyl alcohol) (PVA) is a water-soluble polymer, which is also known to be biodegradable.46,47 Compared to PHB, PVA has excellent mechanical properties. It has been widely used in the preparation of plastic articles and as additives in the paper, wood, tannery, paint, textile, and agro industries.48 Therefore, PVA is a good candidate for a biodegradable blend partner for PHB, when the blends are designed for biodegradable materials. Nevertheless, PHB/PVA blends did not receive much attention until the beginning of the 1990s.49-54 Inoue’s group used infrared analysis51 on blends of PHB and stereoregular PVAs. It was found51 that in blends with highly isotactic PVA (i-PVA), PHB crystallizes in the same way as in pure PHB. The degree of crystallinity of PHB decreases via blending with atactic PVA (a-PVA), when a-PVA was the predominant component of the blend (g80% a-PVA). In PHB blends with highly syndiotactic PVA (s-PVA), a significant suppression of PHB crystallization was observed in a wider range of blend composition (g70% s-PVA) than in PHB/a-PVA blends. These results were consistent with their previous results from 13C nuclear magnetic resonance (NMR) and differential scanning calorimetry (DSC),49,50 suggesting the influence of tacticity on compatibility55 of polymer blends.
10.1021/jp0532162 CCC: $30.25 © 2005 American Chemical Society Published on Web 09/28/2005
19176 J. Phys. Chem. B, Vol. 109, No. 41, 2005 Hydrogen-bonding interaction was believed to be the driving force for the miscibility55 of PHB and PVA in the amorphous phase, and it was investigated by high-resolution solid-state 13C NMR spectroscopy.49,50 It was found that, in the NMR spectra, the carbonyl carbon resonance from PHB showed a downfield shift for the composition range between 80 and 95 wt % a-PVA or s-PVA, which is indicative of a hydrogen-bonding interaction in the amorphous phase. No downfield shift was observed for the blends containing less than 75 wt % PVA. However, whether any hydrogen-bonding interaction formed in these blends containing less than 75 wt % PVA was unclear, due to the possible screening of the amorphous phase by the peaks of the crystalline phase for the highly crystalline sample PHB. Besides, the resonances of the PVA component are partially overlapped with those of PHB, so that the methine carbon resonance of PVA, the expected hydrogen-bonding site in the blends, cannot be observed. There is also an obvious drawback in the DSC method; the DSC melting peak of PVA is severely overlapped with the thermal decomposition peak of PHB. This problem makes it difficult to observe the thermal behavior and morphological properties of PVA in the blends, such as its crystallinity. In addition, no glass transition temperatures (Tg) of the blends were reported in the above studies,49,50 where it was difficult to observe precise Tg of the blends with relatively high crystallinity. Therefore, miscibility in the amorphous phase of the blend was not judged from the Tg data but analyzed by density measurement and NMR spectroscopy.49,50 It was concluded from density measurement49 that PHB/PVA blends are partially miscible in the amorphous phase. NMR results,49,50 measurements of T1 (1H spin-lattice relaxation time in the laboratory frame) and T1F (1H spin-lattice relaxation time in the rotating frame), indicated that the increase of PVA content in the blend enhances the miscibility. FTIR spectroscopy was employed to detect hydrogen-bonding interactions between PHB and PVA51 because of two reasons. One is that there are no overlapping bands in the CdO stretching band region of PHB and the OH stretching band region of PVA. The other is that IR spectroscopy is sensitive to conformational changes, especially for semicrystalline polymers. It was found51 that, in the blends of PHB/a-PVA and PHB/s-PVA, an increasing contribution from the amorphous band at 1744 cm-1 was clearly observed for the spectra of blends containing less than 20% PHB. This observation, however, does not serve as direct evidence of hydrogen-bonding formation. It only indicates that PHB remains in the amorphous state in these blends. In the OH stretching band region of PVA, all band shapes of the blends are almost the same as those of the pure a-PVA and s-PVA. No shift to a lower wavenumber of the OH stretching band indicates that not many hydrogen bonds are formed between PHB and PVA. In other words, no direct and strong evidence for the hydrogen-bonding interaction between PHB and PVA was found using FTIR.51 It is noticed that all PVAs used in the above studies49-51 were completely hydrolyzed. A straightforward but also interesting question then can be asked: what will happen to the miscibility between PHB and PVA if a partially hydrolyzed PVA is used? In addition, can we detect hydrogen-bonding interaction directly by FTIR in these blends? Furthermore, fully hydrolyzed PVA has high melting point (around 250 °C), and the DSC peak of this PVA melting peak is severely overlapped with the thermal decomposition peak of PHB. This problem makes it not very practical to observe the details of thermal behavior and morphological properties of PVA in the blends. However, a
Huang et al. CHART 1: Chemical Structures of PVA80 and PHB
partially hydrolyzed PVA, with a saponification degree of 7882% (thereafter called PVA80 for the sake of convenience), has a much lower melting point (ca. 180 °C, see Experimental Section), which is close but slightly higher than that of PHB. Therefore, it becomes feasible now to observe the thermal behavior and morphological properties of PVA80 in the blends. In the present study, we investigate the miscibility and hydrogenbonding interaction in the amorphous phase of PHB/PVA80 blends, as well as the morphological properties of PVA80 and PHB in the blends mainly using FTIR spectroscopy. DSC result will also be provided whenever necessary. The detailed crystallization and melting behaviors of the blends will be reported later.56 Experimental Section Materials and Sample Preparations. Bacterially synthesized PHB (Tm ∼ 172 °C, DSC result) with Mn ) 2.9 × 105 and Mw ) 6.5 × 105 was kindly provided by the Procter & Gamble Co., Cincinnati, OH. PHB was purified by first dissolving the sample into hot chloroform and then precipitating in methanol. A PHB thin film was cast from the chloroform solution at room temperature. Partially hydrolyzed PVA (Tm ∼ 180 °C, DSC result) sample with a saponification degree of 78-82% (PVA80) and degree of polymerization of 2000, was purchased from Wako Co., Japan. Thin films of PVA80 and all their blends with PHB were prepared by casting from solutions of 1,1,1,3,3,3-hexafluoro2-propanol (HFIP), which was also purchased from Wako Co. All the thin films were left in air for 1 day and dried in a vacuum oven at 60 °C for 1 day and then 80 °C for another day. The samples made were kept in a desiccator at room temperature until use. Fourier Transform Infrared Measurement. Infrared spectra were obtained on a Thermo Nicolet NEXUS Fourier transform infrared (FTIR) spectrometer using a minimum of 128 coadded scans at a resolution of 2 cm-1. The film was sufficiently thin to be within the absorbance range where the Beer-Lambert law is obeyed. The temperature of the sample cell was controlled by a thermoelectric device (CN4400, OMEGA) with an accuracy of (0.2 °C. DSC Measurement. DSC measurement was performed on a Perkin-Elmer apparatus over a temperature range from -30 to 200 °C. The heating and cooling rates were 10 °C/min. Results and Discussion IR Spectra of PVA80 and PHB in the CdO Stretching Band Region. PVA80 is a partially hydrolyzed PVA. It can be regarded as a copolymer of vinyl alcohol and vinyl acetate, as depicted in Chart 1, where the chemical structure of PHB is also shown. From the chemical structure, one might presume that hydrogenbonding interaction between the carbonyl groups and the hydroxyl groups will be a significant feature in this polymer. Figure 1 shows an IR spectrum of PVA80 at room temperature
Polymer Blends of PHB and PVA80
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19177
Figure 1. (A) An IR spectrum in the CdO stretching band region of PVA80 at room temperature. (B) The second derivative of the spectrum shown in part A.
Figure 2. (A) An IR spectrum in the CdO stretching region of PHB measured at room temperature. (B) The second derivative of the spectrum shown in part A.
in the region from 1800 to 1650 cm-1, where bands due to the CdO stretching vibrations are observed. Also shown in Figure 1 is its second-derivative spectrum. Two bands at 1739 and 1715 cm-1 are clearly resolved, corresponding to the well-known “free” and hydrogen-bonded CdO bands in many hydrogenbonded polymer blends containing carbonyl groups.57 The CdO stretching band region of PVA80 is simple, consisting only of the “free” and hydrogen bonded CdO bands. In the CdO stretching band region of PHB, however, it is not so straightforward to determine exactly how many bands there are, even via the analysis of the second-derivative spectrum (Figure 2). Nevertheless, it is generally accepted that there are at least three major bands around 1748, 1735, and 1723 cm-1. The peak assignments have been discussed in detail, with some discrepancies remaining on the assignment of the 1735 cm-1 band.7,51,58 Nevertheless, the first point we wish to make is that it would be safe to assign the 1748 and 1723 cm-1 bands, respectively, to amorphous and crystalline CdO bands. Obviously, the CdO bands from PHB are heavily overlapping with those from PVA80 in the IR spectra of PHB/PVA blends (see Figure 3). Composition-Dependent IR Spectra of PVA80/PHB Blends. Figure 3 shows the scaled FTIR spectra in the region from 1800 to 1660 cm-1 of PHB, PVA80, and their blends with compositions of PVA80/PHB ) 1/9 to 9/1, indicating the heavy overlapping of CdO stretching bands between PHB and PVA80. Nevertheless, it is interesting to notice that, in the PVA80/PHB blends with compositions of PVA80/PHB ) 1/9 to 7/3, the crystalline band of PHB centered around 1723 cm-1 dominates, with gradual increase of the amorphous CdO band around 1740 and the hydrogen-bonded CdO band around 1715 cm-1. An abrupt increase in the intensity of amorphous CdO band occurs
Figure 3. Scale-expanded FTIR spectra in the CdO stretching region of PHB, PVA80, and their blends with different compositions measured at room temperature: 1, PVA80; 2, PVA80/PHB ) 9/1; 3, PVA80/ PHB ) 8/2; 4, PVA80/PHB ) 7/3; 5, PVA80/PHB ) 6/4; 6, PVA80/ PHB ) 5/5; 7, PVA80/PHB ) 4/6; 8, PVA80/PHB ) 3/7; 9, PVA80/ PHB ) 2/8; 10, PVA80/PHB ) 1/9; 11, PHB.
when the composition of the PVA80/PHB blend is changed to PVA80/PHB ) 8/2 and 9/1, and the dominant band changes to the amorphous CdO band at 1740 cm-1, with the band profile similar, but not identical, to the CdO stretching band of pure PVA80. The above observation is very important at least for two reasons. First, it suggests that PHB is no longer able to crystallize at room temperature when the weight percentage of PVA80 in the blend is at or greater than 80 wt %. This conclusion can be first verified by the disappearance of a minor band centered around 1687 cm-1 when the weight percentage of PVA80 is
19178 J. Phys. Chem. B, Vol. 109, No. 41, 2005
Figure 4. Scale-expanded FTIR spectra in the region from 1695 to 1675 cm-1 of PHB, PVA80, and their blends with different compositions measured at room temperature: 1, PVA80; 2, PVA80/PHB ) 9/1; 3, PVA80/PHB ) 8/2; 4, PVA80/PHB ) 7/3; 5, PVA80/PHB ) 6/4; 6, PVA80/PHB ) 5/5; 7, PVA80/PHB ) 4/6; 8, PVA80/PHB ) 3/7; 9, PVA80/PHB ) 2/8; 10, PVA80/PHB ) 1/9; 11, PHB.
Figure 5. FTIR spectra in the region from 1700-1675 cm-1 of PHB measured at 30, 100, 150, 170, 180, and 190 °C.
Figure 6. FTIR spectra in the region from 1000-880 cm-1 of PHB measured at 30, 100, 150, 170, 180, and 190 °C.
80% in the blend, as shown in Figure 4. The 1687 cm-1 band, undoubtedly, is a crystalline band, which can be demonstrated by the melting behavior of PHB (Figure 5). It can be seen from Figure 5 that the 1687 cm-1 band disappears completely when the temperature is around 180 °C, which is a little bit higher than the Tm of PHB.7 There is also a number of C-C stretching bands59 in the spectrum of PHB, located between 1000 and 900 cm-1. Their crystalline features can be easily demonstrated by a temperature study, as shown in Figure 6. It can be found that the bands centered around 895, 910, 928, and 937 cm-1 disappear when the temperature is above the melting temperature of PHB. Above
Huang et al.
Figure 7. FTIR spectra in the region from 1000 to 880 cm-1 of PHB, PVA80, and their blends with different compositions measured at room temperature.
Figure 8. Scale-expanded FTIR spectra in the region from 3050 to 2750 cm-1 of PHB, PVA80, and their blends with different compositions measured at room temperature: 1, PHB; 2, PHB/PVA80 ) 9/1; 3, PHB/PVA80 ) 8/2; 4, PHB/PVA80 ) 7/3; 5, PHB/PVA80 ) 6/4; 6, PHB/PVA80 ) 5/5; 7, PHB/PVA80 ) 4/6; 8, PHB/PVA80 ) 3/7; 9, PHB/PVA80 ) 2/8; 10, PHB/PVA80 ) 1/9; 11, PVA80.
the Tm of PHB, the sharp crystalline band centered around 978 cm-1 collapses and shifts toward a lower frequency, which emerges into a broad peak with the crystalline band centered around 953 cm-1. Therefore, the crystalline C-C stretching bands between 1000 and 900 cm-1 provide another useful region to monitor the morphological properties of PHB in the blends. Figure 7 shows the FTIR spectra in the region from 1000 to 880 cm-1 of PHB, PVA80, and their blends with different compositions measured at room temperature. Band profiles in Figure 7 are strikingly similar to those in Figure 6, except the two with 80 and 90 wt % PVA80; they clearly contain the contribution from PVA80 around 945 cm-1. Thus, the amorphous state of PHB in the PVA80/PHB ) 8/2 and 9/1 blends is further confirmed by the disappearance of these crystalline C-C stretching bands in the region from 1000 to 880 cm-1. The diminishing crystalline bands of PHB at 3007, 2995, and 2975 cm-1, which can be assigned to the CH3 asymmetric stretching bands, as well as the feature at 2875 cm-1 for CH3 symmetric stretching band in the C-H stretching region7 (Figure 8) all together provide additional evidence to support the above conclusion that PHB assumes the amorphous state in the blends, when the compositions of the blends are PVA80/PHB ) 8/2 and 9/1. It should be noted that, with the disappearance of the CH3 asymmetric stretching bands at 3007, 2995 and 2975 cm-1, the band around 2983 cm-1 becomes significant. This band has been assigned to CH3 asymmetric stretching modes from the
Polymer Blends of PHB and PVA80
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19179
Figure 9. IR spectra of PVA80 in the region from 1160 to 950 cm-1 measured at 30, 90, 110, 130, 150, 160, 170, 180, 190, 200, and 210 °C.
amorphous state.7 This further confirms that PHB assumes the amorphous state in the blends of PVA80/PHB ) 8/2 and 9/1. The second implication of the above important observation is that phase inversion of the polymer matrix may happen when the amount of PVA80 reaches 80% in the blends. While studying the water transport, structures, and mechanical behavior of biodegradable PHB/PVA blends, where PVA is also fully hydrolyzed, Olkhov60 et al. found that the tensile strength and elongation at break of PHB/PVA blends decrease dramatically when the concentration of PHB reaches 30 and 20%. This observation suggests that possible phase inversion of the polymer matrix takes place in the concentration range of about 30% of PHB. Similarly, in the blends studied here, when the amount of PVA80 in the blends is lower than 80%, the mechanical behavior, tensile strength and elongation, of the blends would be analogous to that of the PHB matrix. On the other hand, at low PHB concentrations (e20%), the mechanical behavior, tensile strength and elongation, of the blends would be determined by PVA80. This finding is important in terms of controlling the mechanical properties of the blends. We have found that PHB remains in the amorphous state when the blend composition is PVA80/PHB ) 8/2 and 9/1. A relevant question arises: what is the state of PVA80 in the blends? To answer this question, it is necessary to follow the changes of crystalline bands from PVA80. There is only one crystalline band in PVA,48 located around 1145 cm-1. Although there are some discrepancies in the opinions on the possible origin of this band,48,61 e.g., whether it is a C-O stretching, C-C stretching, or H-bonded OH in the crystalline phase, the crystalline feature of this band can be easily verified by its temperature sensitivity, as shown in Figure 9. It can be seen that the intensity of the 1145-cm-1 band starts to decrease at 130 °C and disappears around 170 °C. This observation is consistent with the DSC result of PVA80 shown in Figure 10, where the starting temperature of the melting peak of PVA80 is located around 130 °C, and the melting peak is found around 180 °C (first heating). Although the intensity of the 1145 cm-1 band decreases with the increase of PHB content in the blend, it is not decisive enough to tell the morphological property of PVA80 in the blends. This is because this band is partially overlapping with the bands from PHB (Figure 11). Therefore, additional DSC measurement is necessary to follow the morphological property of PVA80 in the blends. It can be easily found from Figure 12 that the melting peak from PVA80 is undetectable when the composition of PVA80 in the blend is kept below 30%. In other words, PVA80 will assume the amorphous state when its content is below 30% in
Figure 10. DSC results of PVA80 for the first and second heating. Heating/cooling rate, 20 °C/min.
Figure 11. Scale-expanded FTIR spectra in the region from 1695 to 1675 cm-1 of PHB, PVA80, and their blends with different compositions measured at room temperature: 1, PVA80; 2, PVA80/PHB ) 9/1; 3, PVA80/PHB ) 8/2; 4, PVA80/PHB ) 7/3; 5, PVA80/PHB ) 6/4; 6, PVA80/PHB ) 5/5; 7, PVA80/PHB ) 4/6; 8, PVA80/PHB ) 3/7; 9, PVA80/PHB ) 2/8; 10, PVA80/PHB ) 1/9; 11, PHB.
Figure 12. DSC thermograms of PHB, PVA80, and their blends with different compositions: 1, PVA80; 2, PVA80/PHB ) 9/1; 3, PVA80/ PHB ) 8/2; 4, PVA80/PHB ) 7/3; 5, PVA80/PHB ) 6/4; 6, PVA80/ PHB ) 5/5; 7, PVA80/PHB ) 4/6; 8, PVA80/PHB ) 3/7; 9, PVA80/ PHB ) 2/8; 10, PVA80/PHB ) 1/9; 11, PHB. (first heating; heating rate, 10 °C/min).
the blend. The DSC result in Figure 12 also demonstrates that the melting peak from PHB disappears when the composition of PHB in the blend is kept below 20%, which is in good agreement with the FTIR results discussed earlier.
19180 J. Phys. Chem. B, Vol. 109, No. 41, 2005
Figure 13. DSC thermograms of PHB, PVA80, and their blends with different compositions: 1, PVA80; 2, PVA80/PHB ) 9/1; 3, PVA80/ PHB ) 8/2; 4, PVA80/PHB ) 7/3; 5, PVA80/PHB ) 6/4; 6, PVA80/ PHB ) 5/5; 7, PVA80/PHB ) 4/6; 8, PVA80/PHB ) 3/7; 9, PVA80/ PHB ) 2/8; 10, PVA80/PHB ) 1/9; 11, PHB. (first heating; heating rate, 10 °C/min).
In summary, composition-dependent IR spectra and DSC thermograms of PVA80/PHB blends indicate that a significant phase change occurs at PVA80/PHB ) 8/2. When the amount of PVA80 reaches about 80%, PHB assumes the amorphous state and is homogeneously distributed in PVA80 in the amorphous phase, and the mechanical properties of the blends, e.g., tensile strength and elongation, would become analogous to that of pure PVA80. When the PVA80 concentration is less than 30%, PVA80 will assume the amorphous state in the blends. Miscibility of PHB/PVA80 Blends in the Amorphous Phase. To investigate the miscibility of two polymers by FTIR, the well-known Painter-Coleman model57,62 is often employed. However, the miscibility of blends is a property usually associated with the amorphous phase, such that the PainterColeman model57,62 only works for amorphous polymer blends. In other words, it is inappropriate to use the Painter-Coleman model to predict miscibility of semicrystalline polymer blends, such as PHB/poly(vinyl phenol) (PVPh)38 or the PHB/PVA80 blend studied here. Nevertheless, from above compositiondependent IR spectra and DSC thermograms of PVA80/PHB blends, we found that when the amount of PVA80 reaches 80%, PHB takes an amorphous state and is homogeneously distributed in PVA80 in the amorphous phase; when the PVA80 concentration is less than 30%, PVA80 will assume an amorphous state in the blends. These data naturally suggest that PHB and PVA80 are miscible in the amorphous phase of the blends. To obtain direct evidence for the miscibility of PHB/PVA80 blends in the amorphous phase, a conventional DSC method is employed. Figure 13 shows the DSC result from the first heating run of cast films of PVA80, PHB, and PHB/PVA80 blends from the solvent HFIP. The Tg of pure PHB (∼0 °C) is not easily measurable under the experimental condition, as also reported in the literature.49,50 The difficulty probably arises from the high crystallinity of PHB, which prevents the observation of the amorphous phase. Only one Tg was found for all the blend compositions. Observed Tg of the blends shifts more or less to a lower temperature than that of pure PVA80 (∼50 °C), but this shift is not significant at all. Nevertheless, the broadening of the Tg peaks in the blends is indicative of a certain level of miscibility in the amorphous phase of the blends.63,64 The miscibility of PHB/PVA80 blends in the amorphous phase can be more directly observed from the DSC result of PHB/PVA80 blends on the second heating run (Figure 14). It
Huang et al.
Figure 14. DSC thermograms of PHB, PVA80, and their blends with different compositions: 1, PVA80; 2, PVA80/PHB ) 9/1; 3, PVA80/ PHB ) 8/2; 4, PVA80/PHB ) 7/3; 5, PVA80/PHB ) 6/4; 6, PVA80/ PHB ) 5/5; 7, PVA80/PHB ) 4/6; 8, PVA80/PHB ) 3/7; 9, PVA80/ PHB ) 2/8; 10, PVA80/PHB ) 1/9; 11, PHB. (second heating run; heating rate, 10 °C/min).
is now very clear that there is only one Tg for the PHB/PVA80 blends in the whole composition range, and the Tg is located between that of pure PVA80 and PHB, which shifts to a lower temperature with the increase of the PHB component. It is noticed that, when the composition of PHB in the blends reaches 70%, no measurable glass transition can be observed, again probably due to the high crystallinity of PHB. It is also interesting to note that the Tg of PVA80 and the blends in the second heating run are larger than those in the first run. As demonstrated in Figure 10 for PVA80, hydrogen-bonding interaction is promoted after increasing the temperature of PVA80, and the blends up to as high as 200 °C. This is an important finding in terms of monitoring the hydrogen-bonding interactions in the blends. More on this later. The depression of Tm of the PHB component in the blends of PHB/PVA80 also demonstrates the miscibility in the amorphous region65 (Figure 12). Although some more interesting issues can be addressed regarding the crystallization and melting behaviors of PHB/PVA80 blends, detailed discussion is beyond the scope of this report. It will be the topic of another publication.56 Hydrogen-Bonging Interactions Between PVA80 and PHB. Hydrogen-bonding interaction between the OH group of PVA and the CdO group of PHB was believed to be the major driving force for the miscibility of PHB and a fully hydrolyzed PVA in the amorphous phase and was partially verified by highresolution solid-state 13C NMR spectroscopy.49,50 It was found that50 there is no detectable hydrogen-bonding interaction between the carbonyl oxygen of PHB and the hydroxyl hydrogen of PVA in the crystalline phase. In the NMR spectra of the amorphous phase, the carbonyl carbon resonance from PHB showed a downfield shift, which is indicative of hydrogenbonding interaction. However, there are some obvious drawbacks in using NMR. The most significant one is that the resonances of the PVA component are partially overlapped with those of PHB, so the methine carbon resonance of PVA, the expected hydrogen-bonding site in the blends, cannot be directly observed. Besides, results from 13C-PST (pulse saturation transfer) MAS (magic angle spinning) NMR spectrum might be misleading. It was found that50 the downfield shift was observed only for the composition range between 80 and 95 wt % a-PVA or s-PVA, suggesting that no hydrogen bonding was formed in the blends containing less than 75 wt % PVA. This result is clearly open to question. A 13C-PST MAS NMR
Polymer Blends of PHB and PVA80
Figure 15. FTIR spectra of PVA80 (1) and PVA80/PHB ) 8/2 (2), the difference spectrum (3) obtained by subtraction the spectrum of PVA80 from that of PVA80/PHB ) 8/2, and the spectrum of PHB (4). All spectra were measured at 180 °C.
spectrum emphasizes the resonances from the amorphous phase but also contains the contributions from 13C NMR included in the crystalline phase. In other words, the information on the amorphous phase might be obscured by the peaks of the crystalline phase for highly crystalline samples, like PHB. Thus, it is probable that the downfield shift of the amorphous peak is hidden by the crystalline peak in the blends containing the highly crystalline PHB.50 Therefore, FTIR spectroscopy had to be employed to detect hydrogen-bonding interactions between PHB and PVA.51 However, as stated in the Introduction, no direct and strong evidence for the hydrogen-bonding interaction between PHB and PVA was found in that study.51 In the particular blends studied here, the situation is more complicated in that the CdO bands from PHB and PVA80 are heavily overlapping at room temperature, as shown in Figure 3. As mentioned earlier, with the increase of PVA80 in the blends, the relative intensity of the hydrogen-bonded CdO band around 1715 cm-1 increases gradually. One might think that this is indicative of the hydrogen-bond formation between the carbonyl group of PHB and hydroxyl group of PVA. But this may not be true, because the relative intensity of the hydrogenbonded CdO band around 1715 cm-1 in the blend is lower than that in PVA80, and it will eventually increase with increasing PVA80 in the blend, even if there is no hydrogenbonding interaction between PHB and PVA80. It seems that it is impossible to detect hydrogen-bond formation between the carbonyl group of PHB and hydroxyl group of PVA by simply looking at the CdO stretching region at room temperature. To overcome this difficulty, analyzing spectra collected at a higher temperature, say 180 °C, which is right above the melting point of PHB or PVA80, might be helpful. This is because, at 180 °C, both PHB and PVA80 are in the molten state, and the crystalline component of PHB disappears. This will not only promote hydrogen-bonding interaction between PVA80 and PHB, as found above, but also transfer all contributions from crystalline CdO bands of PHB into amorphous bands. The latter effect makes it much more practical to detect a hydrogenbonding interaction between the carbonyl groups of PHB and hydroxyl groups of PVA, if there is any, by monitoring the Cd O groups of the PHB component in the blend at 180 °C. This advantage can be readily demonstrated by a difference spectrum obtained by the subtraction of the PVA80 spectrum from the spectrum of a blend, say PVA80/PHB ) 8/2, at 180 °C (Figure 15), based on a reference peak from PVA80 only, the CH2 bending mode around 1430 cm-1.48 Comparing with the spectrum of PHB, a weak shoulder around 1715 cm-1 in the difference spectrum (curve 3) was found and can obviously be
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19181
Figure 16. FTIR spectra of PVA80 (solid line) and PVA80/PHB ) 6/4 (dash line) in the region from 3700 to 3000 cm-1 measured at 180 °C.
ascribed to the hydrogen-bonded CdO band in PHB at 180 °C. Therefore, the observation demonstrates unambiguously the existence of hydrogen-bonding interaction between PVA80 and PHB at 180 °C, though the percentage of the hydrogen-bonded CdO band is not high at this temperature. Because of the limited level of hydrogen bonds formed between PHB and PVA80, it is not surprising to see that there would be no major observable peak shift in the OH stretching band region from 3700 to 3100 cm-1 in the PVA80/PHB blends (not shown) at room temperature, as also observed in the blends of PHB with a fully hydrolyzed PVA.51 Nevertheless, it should be possible to see some hydrogen-bonding interaction at a higher temperature, say 180 °C, because hydrogen bonding is promoted above the melt temperature, as observed in the CdO stretching region (Figure 15). Figure 16 shows the FTIR spectra of PVA80 and blend PVA80/PHB ) 6/4 obtained at 180 °C. It can be found that the OH peak shifts to a lower wavenumber in the blend of PVA80/PHB ) 6/4, forming another evidence for the hydrogen-bonding interaction between PVA80 and PHB. Once again, the extent of shift is not so big, confirming that there are not many hydrogen bonds formed between PVA80 and PHB. It was noticed that the intensity of the hydrogen-bonded Cd O band in the blend is much less than that in pure PVA80 (Figure 3). This result might be counterintuitive, because mixing PHB with PVA80 was expected to increase the intensity of hydrogen-bonded CdO band, due to the anticipated hydrogenbonding interaction between the OH groups of PVA80 and the CdO groups of PHB, assuming only the hydroxyl groups that do not participate in the intramolecular hydrogen-bonding interaction can form the intermolecular hydrogen bond.50 The above assumption turns out to be wrong. In the OH stretching region (3700-3000 cm-1), the line shapes are almost the same (not shown) as those observed in the blends of PHB with fully hydrolyzed PVA,51 indicating no significant effect of mixing PHB with PVA80 on the OH groups. It was assumed51 that intermolecular hydrogen bonds between PVA hydroxyl groups are to be formed prior to the formation of those between PHB carbonyls and PVA hydroxyls. In other words, hydrogen bonds were preferentially formed among PVA hydroxyl groups, instead of those between PHB carbonyls and PVA hydroxyls. The hydrogen bond formed between two hydroxyl groups is stronger than that formed between a hydroxyl group and a carbonyl group, as in the blends of poly(vinyl phenol) and poly(methyl methacrylate).66,67 Besides, both PVA80 and PHB keep their semicrystalline states in the blends, except for the blends with high PVA80 or high PHB contents, which further limits the formation of hydrogen bond between a hydroxyl group of PVA80 and a carbonyl group of PHB, due
19182 J. Phys. Chem. B, Vol. 109, No. 41, 2005 to the effect of unfavorable entropy of mixing. In addition, under the experimental conditions, where the solvent HFIP was used to cast films at room temperature, the ∆χ factor might play a role.68,69 In this case, minor phase separation could possibly occur, which also gives rise to the formation of fewer hydrogen bonds between PVA80 and PHB. On the basis of these factors, it would be not so difficult to understand the counterintuitive phenomenon mentioned above. It is likely that, when PHB is blended with PVA80, some of the OH groups which were hydrogen bonded to CdO groups of PVA80 are now turning to form hydrogen bonds with the CdO groups of PHB, which in turn gives rise to the observed miscibility of PVA80 and PHB in the amorphous phase. In other words, mixing PHB with PVA80 does not affect much the OH groups of PVA80 that are already hydrogen bonded among themselves prior to the mixing. Conclusions PHB is miscible with PVA80 in the amorphous state over the whole composition range of the blends. When the amount of PHB is at or lower than 20 wt % in the blends, PHB assumes the amorphous state and the IR spectrum in the CdO stretching band region of a blend is analogous to that of pure PVA80. Likewise, when the PVA80 concentration is at or lower than 30 wt %, PVA80 assumes the amorphous state in the blends. Blending PHB with PVA80 does not have measurable influence on the hydroxyl groups of PVA80 that are hydrogen bonded together at room temperature. Instead, PHB dispossesses some hydroxyl groups of PVA80 that are hydrogen bonded to its carbonyl groups, which in turn leads to the miscibility of PHB and PVA80 in the amorphous phase. The hydrogen-bonding interactions between the OH groups of PVA80 and the CdO groups of PHB were not detectable at room temperature, due to very few hydrogen bonds formed, but was observed at a higher temperature of 180 °C, where hydrogen-bonding interaction is promoted by increasing the mixing entropy and reducing the ∆χ effect. Acknowledgment. H.H. thanks JSPS (Japan Society for the Promotion of Science) for providing a postdoctoral fellowship and financial support. This work was partially supported by the “Open Research Center” project for private universities and a matching fund subsidy from MEXT (Ministry of Education, Culture, Sports, Science and Technology), 2001-2005. This work was also supported by Kwansei-Gakuin University “Special Research” project, 2004-2008. References and Notes (1) Dawes, E. A. Microbial Energetics; Blackie: Glasgow, U.K., 1986. (2) Holmes, P. A. In DeVelopments in Crystalline Polymers. II. Bassett, D. C., Ed.; Elsevier: Landon, 1988; p 1. (3) Doi, Y. Microbial Polyester; VCH: New York, 1990. (4) An, Y.; Dong, L.; Mo, Z.; Liu, T.; Feng, Z. J. Polym. Sci. Polym. Phys. 1998, 36, 1305. (5) Sudesh, K.; Abe, H.; Doi, Y. Prog. Polym. Sci. 2000, 25, 1503. (6) Ha, C.; Cho, W. Prog. Polym. Sci. 2002, 27, 759. (7) Sato, H.; Murakami, R.; Padersmshoke, A.; Hirose, F.; Senda, K.; Noda, I.; Ozaki, Y. Macromolecules 2004, 37, 7203. (8) Holmes, P. A.; Wright, L. F.; Collins, S. H. Eur. Pat. Appl. 0052459, 1981. (9) Bloembergen, S.; Holden, D. A.; Hamer, G. K.; Bluhm, T. L.; Marchessault, R. H. Macromolecules 1986, 19, 2865. (10) Bluhm, T. L.; Hamer, G. K.; Marchessault, R. H.; Fyfe, C. A.; Veregin, R. P. Macromolecules 1986, 19, 2871. (11) Doi, Y.; Kunnioka, M.; Nakamura, Y.; Soga, K. Macromolecules 1987, 20, 2988. (12) Doi, Y.; Tamaki, A.; Kunnioka, M.; Soga, K. Appl. Microbiol. Biotechnol. 1988, 28, 330.
Huang et al. (13) Orts, W. J.; Marchessault, R. H.; Bluhm, T. L. Macromolecules 1991, 24, 6435. (14) Pearce, R. P.; Marchessault, R. H. Macromolecules 1994, 27, 3869. (15) Doi, Y.; Segawa, M.; Kunika, M. Int. J. Boil. Mcromol. 1990, 12, 106. (16) Nakamura, K.; Yoshie, N.; Sakurai, M.; Inoue, Y. Polymer 1994, 35, 193. (17) Kamiya, N.; Yamamoto, Y.; Inoue, Y.; Chujo, R.; Doi, Y. Macromolecules 1989, 22, 1676. (18) Kamiya, N.; Yamamoto, Y.; Inoue, Y.; Chujo, R.; Doi, Y. Macromolecules 1989, 22, 3800. (19) Kunnioka, M.; Nakamura, Y.; Doi, Y. Polym. Commun. 1988, 29, 174. (20) Doi, Y.; Tamaki, A.; Kunnioka, M.; Soga, K. Makromol. Chem. Rapid Commun. 1987, 8, 631. (21) Nakamura, S.; Kunnioka, M.; Doi, Y. Makromol. Rep. 1991, A28, 15. (22) Avella, M.; Martuscelli, E. Polymer 1988, 29, 1731. (23) Avella, M.; Greco, P.; Martuscelli, E. Polymer 1991, 32, 1964. (24) Avella, M.; Martuscelli, E.; Raimo, M. Polymer 1993, 34, 3234. (25) Li, R. Q.; An, Y. X.; Zhuang, Y. G.; Dong, L. S.; Teng, F. G.; Feng, Z. L. Acta Polym. Sin. 1999, (4), 498. (26) Li, R. Q.; An, Y. X.; Zhuang, Y. G.; Dong, L. S.; Teng, F. G.; Feng, Z. L. Acta Polym. Sin. 2001, (2), 143. (27) Park, S. H.; Lim, S. T.; Shin, T. K.; Choi, H. J.; Jhon, M. S. Polymer 2001, 42, 5737. (28) Chee, M. J. K.; Ismail, J.; Kummerlowe, C.; Kammer, H. W. Polymer 2002, 43, 1235. (29) Chen Y.; Chen, Q. Acta Polym. Sin. 2003, (1), 115. (30) Yang, H.; Li, Z. S.; Qian, H. J.; Yang, Y. B.; Zhang, X. B.; Sun, C. C. Polymer 2004, 45, 453. (31) Scandola, M.; Ceccorulli, G.; Pizzoli, M. Macromolecules 1992, 25, 6441. (32) Ceccorulli, G.; Pizzoli, M.; Scandola, M. Macromolecules 1993, 26, 6722. (33) Zhang, L. L.; Deng, X. M.; Huang, Z. T. Polymer 1997, 38, 5379. (34) Zhang, L. L.; Deng, X. M.; Zhao, S. J.; Huang, Z. T. Polymer 1997, 38, 6001. (35) Wang, T. Z.; Cheng, G. X.; Ma, S. H.; Cai, Z. J.; Zhang, L. G. J. Appl. Polym. Sci. 2003, 89, 2116. (36) Iriondo, P.; Iruin, J. J.; Fernandez-Berridi, M. J. Polymer 1995, 36, 3235. (37) Xing, P.; Dong, L.; Feng, H.; An, Y.; Zhuang, Y.; Feng, Z. Chem. J. Chin. UniV. 1996, 11, 1813. (38) Iriondo, P.; Iruin, J. J.; Fernandez-Berridi, M. J. Macromolecules 1996, 29, 5605. (39) Xing, P.; Dong, L.; An, Y.; Feng, Z.; Avella, M.; Martuscelli, E. Macromolecules 1997, 30, 2726. (40) Greco, P.; Martuscelli, E. Polymer 1989, 30, 1475. (41) Kumagai, Y.; Doi, Y. Polym. Degrad. Stab. 1992, 36, 241. (42) An, Y.; Dong, L.; Xing, P.; Mo, Z.; Zhuang, Y.; Mo, Z.; Feng, Z. Eur. Polym. J. 1997, 33, 1449. (43) An, Y.; Dong, L.; Li, L.; Mo, Z.; Feng, Z. Eur. Polym. J. 1999, 35, 365. (44) Chiu, H. J.; Chen, H. L.; Lin, T. L.; Lin, J. S. Macromolecules 1999, 32, 4969. (45) An, Y.; Li, L.; Dong, L.; Mo, Z.; Feng, Z. J. Polym. Sci. Polym. Phys. 1999, 37, 443. (46) Suzuki, T.; Ichihara, Y.; Yamada, M.; Tonomura, K. Agric. Biol. Chem. 1973, 37, 747. (47) Chiellini, E.; Corti, A.; D’Antone, S.; Solaro, R. Prog. Polym. Sci. 2003, 28, 963. (48) Finch, C. A., Ed. PolyVinyl alcoholsdeVelopment, 2nd ed.; Chichester: John Wiley: 1992. (49) Azuma, Y.; Yoshie, N.; Sakurai, M.; Inoue, Y.; Chujio, R. Polymer 1992, 33, 4763. (50) Yoshie, N.; Azuma, Y.; Sakuri, M.; Inoue, Y. J. Appl. Polym. Sci. 1995, 56, 17. (51) Ikejima, T.; Yoshie, N.; Inoue, Y. Macromol. Chem. Phys. 1996, 197, 869. (52) Ikejima, T.; Cao, A.; Yoshie, N.; Azuma, Y.; Inoue, Y. Polym. Degrad. Stab. 1998, 62, 463. (53) Ikejima, T.; Inoue, Y. Macromol. Chem. Phys. 2000, 201, 1598. (54) Zhao, L.; Tsuchiya, K.; Inoue, Y. Macromol. Biosci. 2004, 4, 699. (55) The use of the terms compatibility and miscibility is very arbitrary in the literature. It might be necessary to distinguish these two terms. Miscibility is a property in the amorphous phase.57 Crystalline/crystalline polymer blends are normally immiscible. Nevertheless, they might be miscible in the amorphous region, as in the case of the PHB/PVA80 blend studied here, and give rise to the compatibility of the blends. Those blends are often referred to as “compatible” polymer blends.
Polymer Blends of PHB and PVA80 (56) Huang, H.; Ozaki, Y. In preparation. (57) Coleman, M. M.; Graf, J.; Painter, P. C. Specific Interactions and the Miscibility of Polymer Blends; Technomic: Lancaster, PA, 1991. (58) Padersmshoke, A.; Sato, H.; Katsumoto, Y.; Ekgasit, S.; Noda, I.; Ozaki, Y. Vib. Spectrosc, 2004, 36, 241. (59) Lin-Vien, D.; Colthup, N. B.; Fateley, W. G.; Grasselli, J. G. The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules; Academic Press: New York, 1991. (60) Olkhov, A. A.; Vlasov, S. V.; Iordanskii, A. L.; Zaikov, G. E.; Lobo, V. M. M. J. Appl. Polym. Sci. 2003, 90, 1471.
J. Phys. Chem. B, Vol. 109, No. 41, 2005 19183 (61) Fujii, K. J. Polym. Sci. 1972, D1, 431. (62) Coleman, M. M.; Painter, P. C. Prog. Polym. Sci. 1995, 20, 1. (63) Roovers, J.; Toporowski, P. M. Macromolecules 1992, 25, 1096. (64) Roovers, J.; Toporowski, P. M. Macromolecules 1992, 25, 3454. (65) Painter, P. C.; Shenoy, S. L.; Bhagwager, D. E.; Fishburn, J.; Coleman, M. M. Macromolecules 1991, 24, 5623. (66) Huang, H.; Ni, C.; Huang, S. Polym. Prepr. 2003, 44 (2), 735. (67) Huang, H.; Ni, C.; Huang, S. Polym. Prepr. 2003, 44 (2), 737. (68) Zeman, L.; Patterson, D. Macromolecules 1972, 5, 513. (69) Paterson, D. Polym. Eng. Sci. 1982, 22, 64.