Higher-order structure formation of a poly(3-hydroxybutyrate) film during solvent evaporation

Abstract

Solvent evaporation crystallization (SEC) of poly(3-hydroxybutyrate) (PHB) in a PHB/chloroform solution was investigated by time-resolved attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy and grazing incidence wide angle X-ray diffraction (GI-WAXD). The ATR-FTIR investigation reveals that the PHB/chloroform solution was in a homogeneous state at first, and with the evaporation of chloroform, the separated PHB from the chloroform solvent was in the mixture of intermediate and amorphous states, but no crystal structure formed due to the presence of chloroform. Subsequently, further evaporation induced a transition from intermediate to crystal structure and the formation of C=O⋯H–C intramolecular interactions within the latter. As the crystal structure developed, the intra-molecular interaction changed from weak to strong due to the reduced intra-molecular distance within the lamella structure. The results of the GI-WAXD investigation suggest the presence of two kinds of intermediate structures with different order (less ordered and highly ordered). During SEC, the intermediate structures formed first, subsequently transforming into a crystal structure.

---

1. Introduction

Crystallization is the most important step during the crystal structure formation of semicrystalline polymers.^1–5 For polymer materials, a crystalline state is a highly stable state with low free energy. Different crystallization pathways can yield different crystal structures, which is reflected in the polymorphic behavior of polymers. For example, poly(L-lactic acid) (PLA) has four different crystalline modifications (α, α′, β, γ); the α- and α′-forms are obtained by crystallization from amorphous states at temperatures higher and lower than 120 °C,^6,7 respectively, while the β- and γ-forms are prepared using a high draw ratio at high temperature and epitaxial crystallization,^8–11 respectively. The formation of final crystal structure involves the transformation of the molecular chain package from an amorphous state to a crystalline state. During this process, the molecular structure passes many states since long molecular chains are highly entangled,¹² and every macromolecular chain has to pass selections on different length and time scales.^13,14 Therefore, investigations of structural changes during crystallization are very important for understanding the growth mechanism of polymer polymorphisms.

Being different from crystallization from pure polymers system, such as cold and melt crystallization, solvent evaporation crystallization (SEC) is in general more complex.^15–21 For a polymer solution, the solvent molecules usually exhibit some interactions with polymer chains,²² and phase separation occurs during solvent evaporation when the concentration reaches a saturation point. A SEC polymer film is usually formed in a very short time, so that in situ synchrotron radiation X-ray has usually been used to investigate this process.^{16,19,20,23–25}

Heinzer et al.²³ studied the spacing change of hexagonally packed cylinders in a poly(styrene-b-butadiene) copolymer film during solvent drying by in situ grazing incidence small-angle X-ray scattering (GI-SAXS) measurements. The authors found that the above spacing increases on solvent evaporation, with the segregation of blocks increasing first. When the solution concentration reaches a critical point, the spacing decreases due to the loss of solvent. Using same method, Ogawa et al.²⁴ studied the structural development of symmetric poly(styrene-b-2-vinylpyridine) block copolymers thin films during spin-coating. The authors suggested that the microstructure of poly(styrene-b-2-vinylpyridine) is first created from randomly distributed spherical micelles forming a body-centered cubic (BCC) lattice, with further evaporation inducing a transition from spheres in the BCC lattice to cylindrical structures.

Compared to X-ray analysis, it is well known that Fourier-transform infrared (FTIR) spectroscopy is more suitable for investigating the conformational and local molecular environments changes of polymers.^26–28 Therefore, time-resolved FTIR spectroscopy has been extensively used to investigate the crystallization of semicrystalline polymers.^7,29–31 However, to ensure better signal-to-noise ratio (SNR), IR spectra are usually obtained using 64 or 128 scans at 2–4 cm⁻¹ resolution. Thus, it takes about 1–2 min to obtain one spectrum. Therefore, FTIR investigations of solvent evaporation crystallization are difficult. Due to this limitation, to the best of our knowledge, no reports on FTIR characterization of SEC exist.

During the last two decades, biodegradable polymers have attracted considerable interest. Among them, poly(3-hydroxybutyrate) (PHB) has been extensively studied,^32–37 since it has mechanical properties similar to those of conventional synthetic polymers.^38–41 The α crystal modification of PHB is the most common form, and its helical chain is packed as an orthorhombic unit cell, P2₁2₁2₁-D⁴₂, with dimensions of a = 5.76 Å, b = 13.20 Å, and c = 5.96 Å.^42–44 Based on IR and X-ray crystallographic studies, Sato et al. suggested the existence of weak hydrogen bonding between the carbonyl and methyl groups (C=O⋯H–C) in the unit cell. The distance between them is 2.62 Å,⁴⁴ which is shorter than the sum of van der Waals radii of O and H atoms, 2.72 Å. The FTIR investigation of Sato et al.⁴⁵ showed that the asymmetric CH₃ stretching band appears at an abnormally high frequency (3009 cm⁻¹), providing more evidence for the C=O⋯H–C hydrogen bonding. Very recently, Wang and Tashiro⁴⁴ studied the crystal structure and intermolecular interactions of the α-form of PHB in detail, confirming the existence of C=O⋯H–C hydrogen bonding.

PHB is also a good model for investigating polymer crystallization behavior due to its exceptional purity and low nucleation density.^46,47 Therefore, it was chosen as a candidate to investigate SEC in this work. Until now, studies on the crystallization of PHB have been mainly concerned with melt crystallization (around 110 °C).^34,37,48 However, no related studies on the cold crystallization of PHB at lower temperature (e.g., room temperature) exist, probably the T_g (glass transition temperature) of PHB is relatively low (around 5 °C). Moreover, the cold crystallization of PHB at room temperature is very fast, and thus, it is rather difficult to carry out the cold crystallization of PHB. Therefore, investigating the SEC of PHB is very meaningful. In this research, SEC studies were carried out at room temperature (25 °C), even though this temperature is higher than the T_g of PHB. The crystallization from an amorphous state can still be traced very clearly due to the presence of solvent. It should be noticed that by SEC, PHB can form only one kind of crystal structure (α-form) from amorphous state, changing SEC condition will not change the crystallization process and the final crystal structure.

In the present study, we focus on such two points: (1) the structure change of PHB from molecular level during SEC; (2) the multi molecular interaction change between PHB and chloroform and within PHB crystal structure during SEC. Attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy was used to investigate the SEC of a PHB/chloroform solution. To overcome the conflict between high IR signal-to-noise ratio and high evaporation rate of chloroform, a glass tube was used to hold the PHB/chloroform solution (Fig. 1). This can slow down the evaporation rate due to the reduction of exposed area and increase the chloroform concentration in vapor on the solution surface. Therefore, the PHB/chloroform solution in the glass tube provided enough time for FTIR measurements during SEC. ATR-FTIR spectroscopy can be used to investigate the SEC of the PHB/chloroform solution in detail, especially in terms of the formation of intermediate structures and hydrogen bonding. Moreover, time-resolved in situ grazing incidence wide angle X-ray diffraction (GI-WAXD) was also used to support the results of ATR-FTIR.

Fig. 1 The picture (a) and the schema (b) of the ATR accessory during measurement.

2. Experimental

2.1 Materials and sample preparation

Bacterially synthesized PHB with a number-average molecular weight of M_n = 42 000 g mol⁻¹ was purchased from Sigma-Aldrich Chemical Co., Ltd., and was used without further purification. A PHB/chloroform solution was prepared by dissolving PHB powder in hot chloroform at 80 °C to produce a homogeneous solution with a concentration of 5 wt%.

2.2 Time-resolved in situ ATR-FTIR measurements

Time-resolved in situ ATR-FTIR measurement of the PHB/chloroform solution were carried out during SEC using a Thermo Nicolet Magna 6700 Fourier-transform FTIR spectrometer with a liquid nitrogen-cooled mercury-cadmium-telluride (MCT) detector. The IR spectra were collected using a PIKE MIRacle (WI, USA) single reflection ATR cell with a 45° ZnSe ATR crystal, which was connected to a dry air supply for purging and aligned for the measurements. Fig. 1 shows a picture (Fig. 1(a)) and a scheme (Fig. 1(b)) of the ATR accessories used. A glass tube was used to hold the PHB/chloroform solution in the ATR cell. Each IR spectrum was obtained at room temperature (25 °C) by co-adding 32 scans at a spectral resolution of 2 cm⁻¹. There was no interval time between the collection of two adjacent spectra. The 0 min point was defined as the time when the addition of the PHB/chloroform solution into the glass tube was finished. The amount of the added solution was about 200 μL, and it took 160 min to complete the SEC of PHB in this experiment. It should be noted that the time required for completing SEC does not have a real physical meaning, since it depends on the amount of solution. Actually, SEC is polymer weight fraction dependent, but it is difficult to simultaneously measure (by gravimetrically method) that change during the spectral measurement, necessitating the use of time instead of polymer weight fraction change.

2.3 Time-resolved in situ GI-WAXD measurements

Time-resolved in situ GI-WAXD measurements were performed using an X-ray diffractometer (Nanoviewer, Rigaku Co., Japan). The system consisted of a rotating anode X-ray generator (Cu K_α, 40 kV, 30 mA) and a specifically designed confocal X-ray mirror with three slit optic collimators and a two-dimensional (2D) detector (Pilatus 100K, Dectris, Switzerland). A Si (100) wafer was used as a substrate. The SEC of PHB in chloroform was investigated using GI-WAXD by placing a 50 μL drop of the above solution onto the substrate. Each WAXD pattern was acquired using a 5 s exposure, with no interval time between subsequent data collections. The angle of incidence was fixed to be 0.19°, which corresponds to 1.14 times of the critical angle for total reflection (θ_c), indicating that the incident X-way fully illuminate the surface region of the sample as well as the deeply-buried interfacial region. The camera distance was chosen to be 131 mm, allowing the (020) and (110) reflections to be measured simultaneously with sufficient angular resolution.

3. Results and discussion

3.1 Structure evolution of PHB during SEC studied by ATR-FTIR spectroscopy

Fig. 2 shows the IR spectra and their second derivatives in the C–H stretching region (a), C=O stretching region (b), and C–Cl stretching region (c) of pure PHB and chloroform. The corresponding assignments are summarized in Table 1. Chloroform shows only two bands, at 3020 and 742 cm⁻¹ (Fig. 2(a) and (c)). Attention should be paid to the PHB bands at 1722 and 3009 cm⁻¹, which are characteristic of the crystalline state (Fig. 2(a) and (b)). The frequency of these bands indicates the existence of C=O⋯H–C hydrogen bonding in crystalline PHB.^34,44,45 Bands at 1738 and 1748 cm⁻¹ are due to the C=O stretching modes of different PHB conformations in the main chain^34,45 or to differently ordered⁴⁹ amorphous parts of PHB. The bands at 2974, 2934, and 2874 cm⁻¹ are assigned to CH₃, CH₂ asymmetric, and CH symmetric stretching modes in the crystalline region, respectively,³⁶ while the band at 2998 cm⁻¹ exists both in the crystalline and amorphous regions (Fig. 2(a)).³⁴ During the crystallization of PHB, these IR bands show sequential changes, which can be used to track the molecular structure change of PHB during SEC.

Fig. 2 ATR-FTIR spectra (top panels) and their second derivatives spectra (bottom panels) in the C–H stretching region (a), C=O stretching region (b) and C–Cl stretching region (c) of pure PHB and chloroform.

Table 1 Assignment for major IR absorbance bands of PHB and chloroform

IR frequencies (cm^−1)	Assignments
3020	CH3 asymmetric stretching mode in chloroform
3009	CH3 asymmetric stretching mode in PHB (C)
2998	CH3 asymmetric stretching mode in PHB
2974	CH3 asymmetric stretching mode in PHB (C)
2934	CH2 antisymmetric stretching mode in PHB (C)
2874	CH symmetric stretching mode in PHB (C)
1748	C=O stretching mode in PHB (A)
1738	C=O stretching mode in PHB (A)
1731	C=O stretching mode in PHB (I)
1722	C=O stretching mode in PHB (C)
1686	C=O stretching mode in PHB (C)
742	C–Cl stretching mode in chloroform

---

Fig. 3(a) and (b) show time-dependent variations of IR spectra in the 3050–2800 and 1770–1680 cm⁻¹ regions during the SEC of PHB, respectively. The yellow and blue regions in Fig. 3(a) stand for the C=O stretching bands of PHB in the amorphous and crystalline regions, respectively. Fig. 4 shows changes in the amorphous (a) and crystalline (b) C=O stretching bands of the PHB/chloroform solution during SEC. SEC prior to the appearance of the crystal C=O band (1722 cm⁻¹) is discussed first. Based on Fig. 4(a), the changes in the amorphous PHB C=O stretching band (denoted C=O (amorphous)) can be divided into two regions, depending on the absorbance and wavenumber changes: time range Δt₁, from 0 to 117 min, and time range Δt₂, from 117 to 146 min. In the first time range, the wavenumber and absorbance of C=O (amorphous) do not change much. Chloroform is good solvent for PHB, and it is reasonable to consider that intermolecular interactions exists between PHB molecular chains and chloroform molecules (denoted as inter).⁵⁰ Jacquel et al.⁴⁹ studied the solubility of polyhydroxyalkanoates (PHAs) in different solvents. They found that among tetrachloromethane, chloroform, dichloromethane, and 1,2-dichloroethane, chloroform and dichloromethane exhibited high solubilizing properties for PHAs, while tetrachloromethane and 1,2-dichloroethane exhibited lower solubilizing properties. The carbon atom in solvent molecules with high PHA solubilizing properties should carry at least one chlorine atom and one hydrogen atom. These high solubilizing properties can be explained by a polar interaction between the chloride atom and the carbonyl group carbon, together with the fact that the electron-deficient hydrogen atom of the halogenated compound is linked to the carbonyl group of the polymer [see Fig. 4 in ref. 50].

Fig. 3 ATR-FTIR spectra in the C=O stretching region (a) and C–H stretching region (b) from 0 to 160 min during SECB.

Fig. 4 ATR-FTIR absorbance and wavenumber change of amorphous (a) and crystal (b) C=O stretching bands of PHB/chloroform solution with time during SECB.

Fig. 5(a) and (b) shows IR spectra of the PHB/chloroform solution in the C=O (amorphous) stretching region and the chloroform CH₃ asymmetric stretching band region during Δt₁, respectively. The absorbance of C=O (amorphous) increases with time due to chloroform evaporation, but the wavenumber remains constant at around 1739 cm⁻¹ (Fig. 4(a) and 5(a)). Similar behavior is also observed for the CH₃ asymmetric stretching band region of chloroform. These results imply that even though the PHB concentration increases due to solvent evaporation, the inter changes little. In the other words, the PHB/chloroform solution is still homogenous, and no phase separation takes place.

Fig. 5 ATR-FTIR spectra of the C=O (amorphous) (a) and C–H stretching band in chloroform (b) during Δt₁.

Time range Δt₂. This time range shows a spectral variation clearly different from that of Δt₁ in terms of both absorbance and wavenumber. It can be seen from Fig. 4(a) that the wavenumber shows a red shift from 1739 to 1735 cm⁻¹, while the absorbance keeps increasing. These spectral changes can also be seen in the raw spectra (Fig. 6(a)). It is interesting to note that the changes of absorbance and wavenumber for Δt₂ are not linear, and that there is a sudden increase in absorbance starting from 127 min. After 132 min, the absorbance curve enters another linear change region. Similarly to the absorbance change, the wavenumber also shows a discrete red shift between 127 and 132 min. Therefore, Δt₂ can be divided into three ranges: 117–127, 127–132, and 132–146 min. The wavenumber is related to the local chemical environment. For the PHB/chloroform solution during SEC, changes of the latter may be very complex due to the occurring phase separation. Therefore, to investigate the structural changes of PHB in the 117–146 min interval, second derivatives of IR spectra are used, as shown in Fig. 6(a). Two sudden change points, 127 and 132 min, are indicated by red and blue broken lines, respectively (Fig. 6(a)).

Fig. 6 ATR-FTIR spectra (top panels) and their second derivatives spectra (bottom panels) in the C=O stretching band region of PHB during 117–146 min (a) and 146–148 min (b).

〈1〉 117–127 min. The second derivative spectrum at 117 min in Fig. 6(a) features two bands at around 1741 and 1727 cm⁻¹. The former band is due to the C=O (amorphous) of PHB, while the new band at 1727 cm⁻¹ is present in neither pure PHB nor pure chloroform spectrum. As discussed above for Δt₁, the solution is still homogeneous, with no phase separation and nucleation taking place. Thus, no crystal structure is formed. Therefore, this new peak of the PHB/chloroform system may reflect inter C=O. In the 117–127 min interval, the 1741 cm⁻¹ band becomes stronger and is shifted to lower frequency, while the intensity of the 1727 cm⁻¹ band slightly increases. The absorbance increase in this range is due to the increase of PHB concentration (Fig. 5(a)). However, the reason behind the hardly detectable concentration change in the Δt₁ range may be due to the sensitivity of the FTIR instrument, since even though chloroform undergoes extensive evaporation, the solution is still dilute, and the amount of PHB on the surface of the ATR crystal does not change much. Thus, the absorbance of PHB changes little.
〈2〉 127–132 min. In this time range, the absorbance shows an abnormal increase (Fig. 4(a)). If the solution is still homogeneous, and only the concentration changes, the rate of absorbance increase should be equal to that in the 117–127 min range. Therefore, it is very likely that other processes happen in this time domain. It can be seen from Fig. 6(a) that in the 127–132 min range a band at around 1731 cm⁻¹ gradually becomes more distinct, while the one at 1727 cm⁻¹ becomes less distinct. In the spectrum recorded at 132 min, the 1731 cm⁻¹ band shows up more clearly. The weaker 1727 cm⁻¹ band indicates that the amount of inter C=O decreases. In other words, the interaction between chloroform and PHB starts to gradually disappear. This means that the amount of chloroform is not sufficient to completely dissolve PHB. Therefore, PHB starts to separate out from chloroform (phase separation). It is reasonable to speculate that the IR absorbance of PHB separated out from chloroform is stronger than that of PHB in solution. Therefore, the absorbance in Fig. 4(a) shows a sudden increase from 127 min onwards.

The most interesting finding in this time domain is that a new band appears at 1731 cm⁻¹. Since this band shows up after phase separation occurs, it is not due to the interaction between PHB and chloroform. In our previous research, we also observed the 1731 cm⁻¹ band during the isothermal crystallization of PHB from melt.^34,48 This band was assigned to an intermediate structure between the amorphous and crystalline states of PHB following the concept proposed by Strobl,⁵¹ since the intensity of the 1731 cm⁻¹ band tends to decrease as the intensity of the crystalline band around 1722 cm⁻¹ increases. The results in Fig. 6(b) show that the intensity of the 1731 cm⁻¹ band decreases, while that of the crystalline band at 1722 cm⁻¹ increases, which is similar to the results of previous studies.^34,48 Therefore, we assign the 1731 cm⁻¹ band to the intermediate structure. However, contrary to the previous results, the 1731 cm⁻¹ band (e.g., in the spectrum at 132 min) does not appear simultaneously with the 1722 cm⁻¹ band.

It should be noted that the so-called crystalline band of PHB at 1722 cm⁻¹ is the crystalline C=O stretching band featuring intramolecular interactions (denoted intra) within PHB, and this band starts to appear just after secondary crystallization. During the introduction period, only the 1732 cm⁻¹ band (due to the C=O stretching mode without intra) shows complex change.⁴⁸ Therefore, studying the intermediate structure of PHB is very important for understanding the crystallization process. However, the isolated intermediate structure is difficult to form during melt crystallization of bulk PHB, since the crystal with intra has a lower energy than the intermediate state, and the transformation of the latter into the former is very fast and spontaneous. However, in the present study, the intermediate state is relativity stable during SEC. This may be due to the presence of chloroform solvent, which weakly interacts with PHB to loosen the crystal structure of the latter. In other words, the distance between CH₃ and C=O groups in the molecular chain of PHB may become longer than the van der Waals separation between the O and H atoms (2.72 Å), thus excluding the formation of intra. Therefore, only the amorphous and intermediate structures appear before intra.

〈3〉 132–146 min. In this time domain, the increase in absorbance becomes slower than in the 127–132 min interval. Since the latter is a transition region between the 117–127 min and 132–146 min intervals, complex changes (e.g., phase separation) start to happen between 127–132 min. Therefore, the absorbance increase is faster between 127–132 min than between 117–127 min and 132–146 min. A similar trend is also observed for the wavenumber changes. Fig. 6(a) reveals that the 1731 cm⁻¹ band becomes stronger in the 132–146 min interval, and that the 1739 cm⁻¹ band shows no frequency shift compared to the 117–127 min and 127–132 min intervals. However, the intensity of the 1727 cm⁻¹ band continually decreases. This information indicates that the phase separation continues with the evaporation of chloroform in the 132–146 min region, but the rate of phase separation is lower than that in the 127–132 min interval, since most of the PHB has already separated out between 127–132 min. However, chloroform can still affect the structure of PHB, and the crystal structure with intra does not appear.

The 1722 cm⁻¹ band starts to appear from 146.5 min, as shown in Fig. 6(b). The absorbance and wavenumber changes between 147.5–160 min are shown in Fig. 4(b) (since the absorbances at 146.5 and 147 min are difficult to read from the raw IR spectra, the absorbance in Fig. 4(b) is plotted starting from 147.5 min). The crystallization process is almost complete after 153 min, as shown by the black broken dotted line in Fig. 4(b). Fig. 4(b) shows that the wavenumber is shifted to lower frequency (from 1722 to 1720 cm⁻¹), accompanied by an increase of absorbance intensity between 146.5–153 min. This change can also be seen from the second derivative spectra in this time domain, as shown in Fig. 7. The low frequency shift of the 1722 cm⁻¹ band is due to the intra becoming stronger during crystallization. This behavior is easy to understand, since the formation of intra involves the creation of C=O⋯H–C hydrogen bonds in PHB, where the distance between the crystal lattice planes is becoming shorter. With closer packing of the crystal lattice planes, the distance between the C=O and CH₃ groups becomes shorter, so that intra becomes stronger during 146.5–153 min.

Fig. 7 Second derivative spectra in C=O stretching band region of PHB during 146.5–153 min.

The IR and corresponding second-derivative spectra of the 3040–2950 cm⁻¹ CH stretching region during SEC in the 146.5–153 min interval are shown in Fig. 8. The three bands at 3008, 2997, and 2976 cm⁻¹ are assigned to the CH₃ asymmetric stretching modes of PHB,³⁴ with their intensities increasing concomitantly with that of the 1722 cm⁻¹ band (Fig. 7). The 3008 cm⁻¹ band is due to the weak C=O⋯H–C hydrogen bonds⁴⁴ and shows a shift to higher frequency from 3008 to 3010 cm⁻¹, presenting additional evidence of the intra becoming stronger during SEC.

Fig. 8 ATR-FTIR spectra (top panels) and their second derivatives spectra (bottom panels) in the C–H stretching region of PHB/chloroform solution during SECB from 146.5 to 153 min.

3.2 Structure evolution of PHB during SEC studied by time-resolved in situ GI-WAXD

Time-resolved in situ GI-WAXD was used in the present study to support the results of ATR-FTIR. Since GI-WAXD measurements cannot employ the glass tube used in the latter measurements for lowering the evaporate rate of chloroform, a PHB/chloroform solution was directly dropped on a Si substrate, letting the solvent evaporate at room temperature. Subsequently, GI-WAXD measurements were carried out.

Time-resolved evolution of two-dimensional GI-WAXD patterns during SEC of the PHB/chloroform solution is shown in Fig. 9. No diffraction pattern of PHB is observed before 50 s, and the diffraction profile of crystalline PHB appears after 55 s (not always clear in the 2D pattern in Fig. 9, see Fig. 10(a)). The diffraction arcs of the (020) and (110) lattice planes of PHB⁴⁴ become more distinct with the evaporation of chloroform starting from 75 s (as shown by the white arrows in Fig. 9(c)). Since each diffraction arc is continuous, there is no preferred orientation within the film. Time-resolved GI-WAXD profiles of the (020) reflection during SEC between 0–300 s integrated from 2D diffraction patterns are shown in Fig. 10(a). It can be seen from the latter figure that the diffraction profiles are very broad and unsymmetrical, implying the presence of more than one diffraction peak.

Fig. 9 Time-resolved two-dimensional GI-WAXD patterns of PHB/chloroform solution during SECB at 0 s (a), 50 s (b), 75 s (c), 150 s (d), 200 s (e) and 300 s (f).

Fig. 10 Time-resolved GI-WAXD profiles of PHB/chloroform solution during SECB from 0–300 s (a). Enlarged GI-WAXD pattern of PHB/chloroform solution during SECB at 300 s (b). Decomposition of the diffraction profiles at 75 s (c) and 100 s (d) (grey solid line) into less ordered intermediate structure (black solid line); highly-ordered intermediate structure (purple solid line); crystal (blue solid line); ghost diffraction (red solid line), the reconstructed diffraction profile (dotted line) was obtained by summing the diffraction of the elemental peaks.

To quantitatively investigate the SEC of PHB, diffraction profiles of the (020) reflection were deconvoluted assuming a Gaussian shape of the underlying peaks. Fig. 10(b) and (c) depict two typical curve fitting results of the diffraction profiles at 75 and 100 s, respectively. Four peaks are present at 2θ ≈ 12.96°, 13.21°, 13.68° and 14.21°. The diffraction arc at 2θ ≈ 14.21°, indicated by the black arrow in Fig. 10(a), is clearly separated from the (020) diffraction arc, as can be seen in Fig. 10(d). Therefore the above ghost peak does not arise from the (020) lattice plane and necessitates future research. It is very interesting that three peaks overlap in the (020) diffraction region, with their relative intensity also changing during SEC. From the relative intensity change and diffraction angle, it is reasonable to assign the peak at 2θ ≈ 13.68° to the crystalline PHB. Meanwhile, the intensities of the other two diffraction peaks at 2θ ≈ 12.96° and 13.21° decrease with time. Their lattice plane distance is slightly larger than the one of the crystal structure, so it is reasonable to assign them to intermediate structures. Very recently, Khasanah et al.⁵⁰ found that ultra-thin PHB films contain stable intermediate structures, and the GI-WAXD diffraction profile of the (020) lattice plane is overlapped by two peaks, assigned to the intermediate and highly-ordered structures.

The two kinds of intermediate structure observed in this study may also be due to their different order (the reason why only one intermediate structure was reported by Khasanah et al., but two are reported in this study may be due to the different confinement in ultrathin and thin films); the lower angle one arises from the less ordered intermediate structure (inter, L), whereas the higher angle one is due to the highly-ordered intermediate structure (inter, H). One may wonder about the controversy between the GI-WAXD and ATR-FTIR observations in this study: the IR spectra indicate the presence of only one intermediate state (1731 cm⁻¹ band), whereas the GI-WAXD results show two kinds of intermediate states. The reason for this controversy may be ascribed to the different principles of IR and WAXD, which are sensitive to changes of the dipole moment and the crystal lattice, respectively. Therefore, IR spectra usually reflect different conformations of polymers, whereas WAXD is usually used to investigate ordered structures having three-dimensional periodicity, such as crystals. Therefore, one possibility is that the two intermediate structures possess similar conformations. To find the true reason, it is crucial to conduct simultaneous GI-WAXD and ATR-FTIR measurements.

Time-resolved fractions of less ordered intermediate structure (X_inter,L(t)), highly-ordered intermediate structure (X_inter,H(t)), and crystalline structure (X_Crys(t)) during SEC between 75–100 s are shown in Fig. 11. Even though the diffraction peaks of crystalline PHB appeared after 55 s, their intensity was very low, and the noise level was high. Therefore, it was impossible to successfully carry out curve-fitting of the profiles in the 55–70 s domain. Thus, the plot starts at 75 s in Fig. 11. The fraction of each part was calculated based on the curve-fitting result using the following equation (using X_inter,L(t) as an example):

where A_inter,L is the diffraction peak area of inter, L. From Fig. 11 (at 75 s), X_Crys is calculated to be around 30%, whereas X_inter,L and X_inter,H are around 25 and 45%, respectively. With the progress of SEC, X_Crys increases to about 41% and X_inter,L and X_inter,H decrease to about 8% and 21%, respectively, at 80 s. Starting from 85 s, X_Crys, X_inter,L, and X_inter,H are stable around 70, 8, and 22%, respectively. It should be noted that GI-WAXD could not detect scattering due to the amorphous region in this study, so the transition from amorphous to intermediate and then to crystalline structure cannot be reflected. These changes very clearly indicate the existence of an intermediate structure in the PHB/chloroform solution system, which can transform into crystalline structure during SEC. Moreover, based on the behavior observed from 75 to 85 s, it is reasonable to speculate that the pure intermediate structure is formed before the onset of crystalline structure formation. This also agrees with the ATR-FTIR results of the present study.

Fig. 11 Percentage fraction of the less ordered intermediate structure (black solid symbols line), highly-ordered intermediate structure (purple solid symbols line) and crystal (blue solid symbols line) as a function of time.

4. Conclusion

Solvent evaporation crystallization of a PHB/chloroform solution was investigated at room temperature by time-resolved ATR-FTIR and GI-WAXD. A glass tube was used to hold the solution, slowing down the evaporation rate of chloroform during ATR-FTIR measurements. The SEC in the ATR-FTIR measurement can be divided into the following five steps: 〈1〉 only the PHB concentration increases with time, and the solution is still homogeneous; 〈2〉 the concentration is beyond the saturation point of the PHB/chloroform solution. Phase separation starts to occur and PHB begins to separate out from the chloroform solvent. The separation of PHB produces a mixture of intermediate and amorphous structures; 〈3〉 PHB continues to separate out from solution as chloroform evaporates, but the rate of phase separation is lower than between step 〈2〉; 〈4〉 intra starts to appear in the crystal structure of PHB, becoming stronger as crystallization progresses; 〈5〉 the absorbance and wavenumber of the intra C=O stretching band show almost no change, indicating that the SEC is complete.

Time-resolved GI-WAXD measurements were performed by casting the PHB/chloroform solution on a Si substrate. The diffraction arcs of PHB during SEC were very broad. The curve fitting procedure revealed that the diffraction peak of the (020) lattice plane contained three overlapping peaks, attributable to the less ordered intermediate, highly-ordered intermediate, and crystalline structures. During SEC, the intermediate structure was formed first, subsequently transforming into the crystalline structure.

Conflict of interest

The authors declare no competing financial interest.

Acknowledgements

This study was supported in part by the Japan Society for the Promotion of Science (Grant-in-Aid for Scientific Research (C) 24560033). We would like to thank Dr Shigesaburo Ogawa (Kwansei Gakuin University, Japan) for help with GI-WAXD measurements. We would also like to thank Prof. Shigeaki Morita (Osaka Electro-Communication University, Japan) for the kind discussion.

References

1. L. Mandelkern, Crystallization of polymers, McGraw-Hill, New York, 1964.
2. B. Wunderlich, Macromolecular Physics, Academic Press, New York, 1980, vol. 3.
3. A. Sharples, Introduction to Polymer Crystallization, Edward Arnold Ltd., London, 1966.
4. K. Armitstead and G. Goldbeck-Wood, Polymer crystallization theories, Adv. Polym. Sci., 1992, 100, 219–312.
5. G. Reiter and G. R. Strobl, Progress in Understanding Polymer Crystallization, Springer-Verlag, Berlin, 2007.
6. J. Zhang, Y. Duan, H. Sato, H. Tsuji, I. Noda, S. Yan and Y. Ozaki, Crystal Modifications and Thermal Behavior of Poly(l-lactic acid) Revealed by Infrared Spectroscopy, Macromolecules, 2005, 38, 8012–8021.
7. J. Zhang, H. Tsuji, I. Noda and Y. Ozaki, Weak Intermolecular Interactions during the Melt Crystallization of Poly(l-lactide) Investigated by Two-Dimensional Infrared Correlation Spectroscopy, J. Phys. Chem. B, 2004, 108, 11514–11520.
8. W. Hoogsteen, A. R. Postema, A. J. Pennings, G. Ten Brinke and P. Zugenmaier, Crystal structure, conformation and morphology of solution-spun poly(L-lactide) fibers, Macromolecules, 1990, 23, 634–642.
9. B. Eling, S. Gogolewski and A. J. Pennings, Biodegradable materials of poly(l-lactic acid): 1. Melt-spun and solution-spun fibres, Polymer, 1982, 23, 1587–1593.
10. J. Puiggali, Y. Ikada, H. Tsuji, L. Cartier, T. Okihara and B. Lotz, The frustrated structure of poly(L-lactide), Polymer, 2000, 41, 8921–8930.
11. L. Cartier, T. Okihara, Y. Ikada, H. Tsuji, J. Puiggali and B. Lotz, Epitaxial crystallization and crystalline polymorphism of polylactides, Polymer, 2000, 41, 8909–8919.
12. R. S. Porter and J. F. Johnson, The entanglement concept in polymer systems, Chem. Rev., 1966, 66, 1–27.
13. B. Lotz, What can polymer crystal structure tell about polymer crystallization processes?, Eur. Phys. J. E: Soft Matter Biol. Phys., 2000, 3, 185–194.
14. S. Cheng, C. Y. Li and L. Zhu, Commentary on polymer crystallization: selection rules in different length scales of a nucleation process, Eur. Phys. J. E: Soft Matter Biol. Phys., 2000, 3, 195–197.
15. M. Lee, J. K. Park, H. Lee, O. Lane, R. B. Moore, J. E. McGrath and D. G. Baird, Effects of block length and solution-casting conditions on the final morphology and properties of disulfonated poly(arylene ether sulfone) multiblock copolymer films for proton exchange membranes, Polymer, 2009, 50, 6129–6138.
16. M. J. Heinzer, S. Han, J. A. Pople, D. G. Baird and S. M. Martin, In Situ Measurement of Block Copolymer Ordering Kinetics during the Drying of Solution-Cast Films Using Small-Angle X-ray Scattering, Macromolecules, 2012, 45, 3471–3479.
17. Y. Gong, H. Huang, Z. Hu, Y. Chen, D. Chen, Z. Wang and T. He, Inverted to Normal Phase Transition in Solution-Cast Polystyrene–Poly(methyl methacrylate) Block Copolymer Thin Films, Macromolecules, 2006, 39, 3369–3376.
18. Y. Yuan, J. Shu, P. Liu, Y. Zhang, Y. Duan and J. Zhang, Study on π–π Interaction in H- and J-Aggregates of Poly(3-hexylthiophene) Nanowires by Multiple Techniques, J. Phys. Chem. B, 2015, 119, 8446–8456.
19. W. Zhao, F. Liu, X. Wei, D. Chen, G. M. Grason and T. P. Russell, Formation of H* Phase in Chiral Block Copolymers: Morphology Evolution As Revealed by Time-Resolved X-ray Scattering, Macromolecules, 2013, 46, 474–483.
20. D. T. W. Toolan, A. Isakova, R. Hodgkinson, N. Reeves-McLaren, O. S. Hammond, K. J. Edler, W. H. Briscoe, T. Arnold, T. Gough, P. D. Topham and J. R. Howse, Insights into the Influence of Solvent Polarity on the Crystallization of Poly(ethylene oxide) Spin-Coated Thin Films via in Situ Grazing Incidence Wide-Angle X-ray Scattering, Macromolecules, 2016, 49, 4579–4586.
21. H. Huang, Z. Hu, Y. Chen, F. Zhang, Y. Gong, T. He and C. Wu, Effects of casting solvents on the formation of inverted phase in block copolymer thin films, Macromolecules, 2004, 37, 6523–6530.
22. M. Rubinstein and R. Colby, Polymers physics, Oxford, UK, 2003.
23. M. J. Heinzer, S. Han, J. A. Pople, D. G. Baird and S. M. Martin, In Situ Tracking of Microstructure Spacing and Ordered Domain Compression during the Drying of Solution-Cast Block Copolymer Films Using Small-Angle X-ray Scattering, Macromolecules, 2012, 45, 3480–3486.
24. H. Ogawa, M. Takenaka, T. Miyazaki, A. Fujiwara, B. Lee, K. Shimokita, E. Nishibori and M. Takata, Direct Observation on Spin-Coating Process of PS-b-P2VP Thin Films, Macromolecules, 2016, 49, 3471–3477.
25. K. Shimokita, T. Miyazaki, H. Ogawa and K. Yamamoto, Development of a simultaneous measurement system for SAXS–WAXD and the thickness of coating films during film formation by solvent evaporation, J. Appl. Crystallogr., 2014, 47, 476–481.
26. J. M. Chalmers, Spectra–Structure Correlations: Polymer Spectra, Handbook of Vibrational Spectroscopy, 2002.
27. S. Krimm, Infrared Spectra of High Polymers, Springer, Berlin, 1960, pp. 51–172.
28. D. I. Bower and W. F. Maddams, The vibrational spectroscopy of polymers, Cambridge University Press, New York, 1992.
29. J. Zhang, Y. Duan, D. Shen, S. Yan, I. Noda and Y. Ozaki, Structure Changes during the Induction Period of Cold Crystallization of Isotactic Polystyrene Investigated by Infrared and Two-Dimensional Infrared Correlation Spectroscopy, Macromolecules, 2004, 37, 3292–3298.
30. C. Yan, Y. Zhang, Y. Hu, Y. Ozaki, D. Shen, Z. Gan, S. Yan and I. Takahashi, Melt crystallization and crystal transition of poly(butylene adipate) revealed by infrared spectroscopy, J. Phys. Chem. B, 2008, 112, 3311–3314.
31. K. Tashiro, Y. Ueno, A. Yoshioka and M. Kobayashi, Molecular mechanism of solvent-induced crystallization of syndiotactic polystyrene glass. 1. Time-resolved measurements of infrared/Raman spectra and X-ray diffraction, Macromolecules, 2001, 34, 310–315.
32. P. J. Barham, A. Keller, E. L. Otun and P. A. Holmes, Crystallization and morphology of a bacterial thermoplastic: poly-3-hydroxybutyrate, J. Mater. Sci., 1984, 19, 2781–2794.
33. Y. Furuhashi, Y. Imamura, Y. Jikihara and H. Yamane, Higher order structures and mechanical properties of bacterial homo poly(3-hydroxybutyrate) fibers prepared by cold-drawing and annealing processes, Polymer, 2004, 45, 5703–5712.
34. J. Zhang, H. Sato, I. Noda and Y. Ozaki, Conformation Rearrangement and Molecular Dynamics of Poly(3-hydroxybutyrate) during the Melt-Crystallization Process Investigated by Infrared and Two-Dimensional Infrared Correlation Spectroscopy, Macromolecules, 2005, 38, 4274–4281.
35. R. Murakami, H. Sato, J. Dybal, T. Iwata and Y. Ozaki, Formation and stability of β-structure in biodegradable ultra-high-molecular-weight poly(3-hydroxybutyrate) by infrared, Raman, and quantum chemical calculation studies, Polymer, 2007, 48, 2672–2680.
36. H. Sato, Y. Ando, J. Dybal, T. Iwata, I. Noda and Y. Ozaki, Crystal Structures, Thermal Behaviors, and C–H⋯O=C Hydrogen Bondings of Poly(3-hydroxyvalerate) and Poly(3-hydroxybutyrate) Studied by Infrared Spectroscopy and X-ray Diffraction, Macromolecules, 2008, 41, 4305–4312.
37. L. Guo, N. Spegazzini, H. Sato, T. Hashimoto, H. Masunaga, S. Sasaki, M. Takata and Y. Ozaki, Multistep Crystallization Process Involving Sequential Formations of Density Fluctuations, “Intermediate Structures”, and Lamellar Crystallites: Poly(3-hydroxybutyrate) As Investigated by Time-Resolved Synchrotron SAXS and WAXD, Macromolecules, 2012, 45, 313–328.
38. Y. Doi, S. Kitamura and H. Abe, Microbial synthesis and characterization of poly(3-hydroxybutyrate-co-3-hydroxyhexanoate), Macromolecules, 1995, 28, 4822–4828.
39. E. Chiellini and R. Solaro, Recent Advances in Biodegradable Polymers and Plastics, Wiley-VCH, Weinheim, 2003.
40. M. Vert, Aliphatic polyesters: great degradable polymers that cannot do everything, Biomacromolecules, 2005, 6, 538–546.
41. C. Bastioli, Handbook of Biodegradable Polymers, Rapra Technology Limited, Shawbury, U.K., 2005.
42. J. Cobntbekt and R. H. Mabchessault, Physical properties of poly-β-hydroxybutyrate: IV. Conformational analysis and crystalline structure, J. Mol. Biol., 1972, 71, 735–756.
43. M. Yokouchi, Y. Chatani, H. Tadokoro, K. Teranishi and H. Tani, Structural studies of polyesters: 5. Molecular and crystal structures of optically active and racemic poly(β-hydroxybutyrate), Polymer, 1973, 14, 267–272.
44. H. Wang and K. Tashiro, Reinvestigation of Crystal Structure and Intermolecular Interactions of Biodegradable Poly(3-Hydroxybutyrate) α-Form and the Prediction of Its Mechanical Property, Macromolecules, 2016, 49, 581–594.
45. H. Sato, R. Murakami, A. Padermshoke, F. Hirose, K. Senda, I. Noda and Y. Ozaki, Infrared Spectroscopy Studies of CH⋯O Hydrogen Bondings and Thermal Behavior of Biodegradable Poly(hydroxyalkanoate), Macromolecules, 2004, 37, 7203–7213.
46. P. J. Barham, A. Keller, E. L. Otun and P. A. Holmes, Crystallization and morphology of a bacterial thermoplastic: poly-3-hydroxybutyrate, J. Mater. Sci., 1984, 19, 2781–2794.
47. P. J. Barham, Nucleation behaviour of poly-3-hydroxy-butyrate, J. Mater. Sci., 1984, 19, 3826–3834.
48. N. Suttiwijitpukdee, H. Sato, J. Zhang and T. Hashimoto, Effects of Intermolecular Hydrogen Bondings on Isothermal Crystallization Behavior of Polymer Blends of Cellulose Acetate Butyrate and Poly(3-hydroxybutyrate), Macromolecules, 2011, 44, 3467–3477.
49. N. Jacquel, C. W. Lo, H. S. Wu, Y. H. Wei and S. S. Wang, Solubility of polyhydroxyalkanoates by experiment and thermodynamic correlations, AIChE J., 2007, 53, 2704–2714.
50. Khasanah, K. R. Reddy, S. Ogawa, H. Sato, I. Takahashi and Y. Ozaki, Evolution of Intermediate and Highly Ordered Crystalline States under Spatial Confinement in Poly(3-hydroxybutyrate) Ultrathin Films, Macromolecules, 2016, 49, 4202–4210.
51. G. R. Strobl, The physics of polymers, Springer, Berlin, 1997.

---