L Spectra Data
2.4 WAXD Measurements
WAXD patterns of the as-prepared blend films were measured with a Rigaku
RINT2100 X-ray diffiractometer equipped with a scintillation detector (Rigaku PT30). The diffraction patterns were recorded at room temperature from 2e= 1 1 .5 to 350 at a scanning rate-33-ofO.2 O/min. The X-ray beam of Cu-Kct radiation was generated at 40 kV and 50 mA and
passed through a Ni filter (wavelength: O.154 nm).3. Results and discussion 3.1 DSC Studies
DSC thermograms of the blends with various compositions measured in the first
cooling run and the second heating run are shown in part (a) and (b) in Figure 2, respectively.Hereafter the blend composition wi11 be denoted by weight fraction of CAB, wcAB, too. The pure PHB sample was rapidly crystallized in the first cooling process, while the crystallization
in the blends was suppressed as typically shown for PHB/CAB = 9011O and 80120 (wcAB= O.1 and O.2) (see Figure 2a). The cold crystallization temperature (T..) of the blends, which is
clearly observed as an endothermic peak, in the second heating process was slightly shifted to a higher temperature with increasing wcAB, as obviously seen for the blends with O.2S wcAB S O.7 in Figure 2(b). On the contrary, the endotherm for the cold crystallization was not clearly discernible for pure PHB. This may be because the first cooling process involves enough crystallization, which in turn suppresses tn' e cold crystallization in the second heating run. On the other hand, the blends did not undergo enough crystallization in the first cooling process so that they underwent the remarkable cold crystallization in the second heating process.
The DSC thermogram of pure PHB in the second heating run shows double
melting-endothermic peaks. The melting temperature (T.) of the blends shifts to a lower temperature with wcAB (Table 1). The melting endotherm cannot be observed for the blends with wcAB ) O.8, so that the blends are amorphous, which will be due to formation of physically crosslinked network in the blends via inter between PHB and CAB as well as intramolecular hydrogen bondings within CAB, as will be discussed later in section 3.5. The enthalpy ofmelting also decreases with,wcAB (see Table 1).On the basis ofthe report by Gunaratne et al.25, the double melting thermogram can be explained by the effect ofthermal history, i.e., the crystallization condition ofPHB. In the neat PHB, the first endothermic peak is due to the melting of the crystals crystallized in the first cooling process, while the second peak is due to the recrystallization and melting of the crystals. The melting temperature of the blend decreases with wcAB, partly (i) because of an increased entropy of melting upon mixing PHB and CAB and partly (ii) be,cause of a greater suppression of mobility of PHB component for the cold crystallization with increasing wcAB, that creates less perfect crystallites with enhanced distortions of the intramolecular hydrogen bondings within PHB crystals, as will be detailed later. We think the latter effect (ii) dominates the former effect (i) on the melting point depression with wcAB, because the entropy ofmelting of the blends ass'ociated with mixing PHB and CAB chains is expected to be small for the
-35-blends with high molecular weights.
Although these blends have been reported to be entirely miscible,i9-22 the relationship of Tg vs wcAB is far from that expected to the Fox law.26 The Tg's of blends observed slightly shift to a lower temperature than pure PHB with wcAB up to wcAB - O.6, followed by a slight increase of the Tg with further increase of wcAB from O.6 to O.7. wnen wcAB > O.7, this low Tg is obscured and the Tg shifts steeply to a higher temperature with wcAB at wcAB > O.8. For the blend with wcAB = O.9, the thermogram around Tg (shown by the arrow) was highlighted in the inset to Figure 2(b). The characteristic parameters for the thermal properties ofthese blends are summarized in Table 1. Even though all the blends apparently exhibited,single Tg, The Tg vs wcAB behavior may infer existence of the two Tg's due to the local heterogeneities ofthe blend composition mediated by a difference in the "self-concentration" of the stiff (CAB) and flexible components (PHB) in the blends.27' 28 Probably the higher Tg for the blends with wcAB S O.7, if it existed, is obscured by the cold crystallization ofPHB.
Figure 2(c) and its inset present a typical DSC thermogram taken in the first heating process of the as-prepared blend sample with wcAB = O.5 at a slow heating rate of2 OC/min.
The experimental conditions chosen here are almost the same as those chosen for the FT-IR studies to be discussed in the sections 3.2 and 3.5 for a fair comparison of the thermal analyses of DSC and FTIR. It is noted that the cold crystallization of the as-prepared blend sample at the slow heating rate is not so obvious as that found for the same blend sample in the second heating run shown in Figure 2(b). We interpret the disparity described above is due to the fact that the crystallization occurred sufficiently during the preparation process of the as-prepared sample, hence suppressing the cold crystallization in the as-prepared sample in the first heating process. As will be shown later in conjunction with Figures 8 and 9, the FTIR clearly indicated the cold crystallization in the temperature range between 80 and
-- 1300C. Based on the FTIR result, we attempted to draw an expected DSC thermogram by dotted line in the case where the cold crystallization did not occur. The downward deviation
of the real thermogram from that shown by the dotted line is very roughly assigned to be an endotherm due to the cold crystallization and partial melting of the crystals formed. Figure 2(c) also exhibits the double melting thermogram. The first endothermic peak may be due to the melting of the crystals formed in the preparation process of the as-prepared sample and those formed in the cold crystallization process in the first heating process, while the second endothermic peak may be due to the recrystallization ofmolten crystals in the first endotherm and melting process of the recrystallized crystals.
3.2 Temperature-dependentIRspectralvariationsofpureCAB
To explore molecular interactions, specifically hydrogen bondings, of the blends, temperature-dependent IR spectra ofneat CAB were first investigated.
Figure 3(a) presents IR spectra in the OH stretching region of3700-3300 cm'i ofpure CAB measured in the heating process of the as-prepared sample from 30 to 190 OC. It is
important to note the fact that the OH stretching bands can clarify or focus on the
intermolecular and intramolecular interactions ofOH groups within CAB, no matter how small is a number fraction ofOH groups per single CAB chain [34.01(34.0+442+165) == O.053 molO/o]compared with other groups, acetyl and butyry1 groups. The selected second derivative spectra at 30 and 190 OC are given in part (b). One can easily recognize one broad band at 3485 cm"i and two weak shoulders at 3587 and 3549 cm-i. It is noted that the intensity of the 3485 cm-i band decreases significantly, as indicated by the downward arrow, and shifts toward a higher
wavenumber of -3489 cm-i with temperature. The intensity of the 3549 cm-i band also
decreases, as indicated also by the downward arrow, and shifts towatd a higher wavenumber.On the other hand, the intensity at 3587 cm'i certainly increases as shown by the' upward arrow.
The present results together with the previous ones on cellulose and CAB22'29 lead us to assign
the band at 3485 cm-i to the OH stretching mode of inter-chain hydrogen bondings
(C=OeeeH-O) (defined as inter CAB-CAB),22 while the band at 3549 cm-i may be ascribed to intramolecular hydrogen bondings of OHeee-O- (ether) and O-HeeeO==C in CAB (defined asintra CAB-CAB). This is because the absorbance of these bands (3549 and 3485 cm-i)
decrease with temperature, as typically observed for bands associated with hydrogen bondings, and because of the relative peak positions of the two bands relative to that at 3587 cm"i.29 We should note here that CAB contains much more C=O groups than cellulose and its derivatives contain. The third peak at 3587 cm-i arises from free O-H groups, because it appears at the highest wavenumber and increases with temperature as shown by the upward arrow in Figure 3(a).29-3i The results in Figure 3 reveal that most of the OH groups of CAB are involved in the hydrogen bondings, although the fractions of OH groups themselves are small among all the substituent groups.The gradual intensity decreases in the bands at 3485 and 3549 cm-i together with the
higher wavenumber shift with temperature suggest that both inter CAB-CAB and intra CAB-CAB become weak and some of them undergo dissociations with temperature. The
intensity increase in the free OH stretching band is not so 1arge, qualitatively indicating that a rather small fraction of the hydrogen bondings are dissociated into the free OH groups,
-37-although they become weak significantly with temperature.
Figure 4(a) shows IR spectra in the C=O stretching band region ofthe as-prepared pure CAB sample measured in the heating process from 30 to 190 OC. The second derivative spectra presented in Figure 4(b) reveal that at least two kinds of C=O stretching bands appear at 1759 and -1743 cm-i at 30 OC. Firstly, we should note the fact that (1) most ofthe C=O groups of CAB exist as free C=O, because 94.7 molO/o of the substitutes are acetyl and butyryl groups, and only 5.3 molO/o are OH groups; only one out of 18 C==O groups can encounter OH groups, because the number ratio ofC=O groups to OH groups is - 18 to 1. From this view point, it is
not easy to identify the C==O stretching band due to inter CAB-CAB and/or intra CAB-CAB -C =O•••H-O- in the spectra shown in Figure 4. Neyertheless, we should note also the fact that
(2) there are certainly the C=OeeeH-O hydrogen bonds, as already evidenced by the OH
stretching band region shown in Figure 3. From these facts (1) and (2) described above, we can conclude that there are two stretching bands due to the free C=O groups which exist in slightly different environments.Detailed band assignments for the neat CAB are summarized in Table 2, together with those ofPHB and PHB/CAB which will be discussed later.