In situ POM and FTIR investigation on quiescent phase transition of polybutene-1 from form II to form I

2.1 Materials

The isotactic PB-1 sample employed in this study was purchased from Lyondell Basell Industries, which trade name is PB0110M. The melt flow rate is 0.4 g/10 minutes (190°C/2.16 kg), the weight-average molecular weight M_w is 711 kg/mol, while the dispersity is 3.5.

2.2 Polarized optical microscopy

Crystalline morphologies during isothermal crystallization and heating after forms II-I phase transformation were observed with an Olympus POM equipped with a charge coupled device and a heating device. During in situ POM measurements, a PB-1 film with a thickness of about 50 μm was sandwiched between two SiO₂ windows and then fixed to the heating device. During isothermal crystallization, a POM was taken automatically every 5 minutes with self-contained software.

2.3 Fourier transform infrared spectroscopy

Structural changes during isothermal crystallization and phase transition were monitored with a PerkinElmer Spectrum Two FTIR spectrometer equipped with a heating device. During FTIR measurements, the sample was sandwiched between two ZnSe windows, which were then fixed on the heating devices. The IR spectrum in the range of 400 to 4000 cm⁻¹ was collected at the resolution of 4 cm⁻¹. Every 5 minutes, an IR spectrum was collected automatically. The band at 816 cm⁻¹ was employed to monitor the change of form I during phase transition.

3.1 Various morphologies formed at high temperature

To observe form I nucleation clearly, large form II crystals with sizes of around 100 μm were formed first at a higher crystallization temperature after heating at 180°C for 10 minutes to remove thermal history. A temperature of 103°C was chosen because it was suitable for POM measurements. Various morphologies of spherulites were observed during isothermal crystallization, for example, hexagonal-shape, round-shape, and octagonal-shape spherulites observed in Figure 2A-C. Square-shape hedrites can be also observed, as seen in Figure 2D. The morphological changes during crystallization are unique and interesting. For round-shape spherulite and square-shape hedrites, their shape remains unchanged but their sizes increase constantly. However, for hexagonal-shape spherulite, a columnar crystal is formed first (see Figure 2A). After reaching a length, it starts to branch. Meanwhile, four small fibrillar crystals form on the lateral surface. Obviously, these small crystals form under the help of the columnar crystal. The columnar crystal can be regarded as mother lamellae, while four new crystals as daughter lamellae. Furthermore, if daughter lamellae on the same side of the mother lamellae had enough distance, a small fibrillar crystal forms in between, forming so-called octagonal-shape crystal (see Figure 2C). Among these crystals, hexagonal-shape and round-shape spherulite were dominant.

Actually, these crystals have been reported by Fu52, 53 and Acierno.54 The formation of the hedrites depends not only on the molecular weight but also on the crystallization temperature. The lower the molecular weight and the higher the crystallization temperature, the higher the probability for the formation of the hedrites. Conversely, the probability of the formation of spherulites will be high. An important difference between hedrites and spherulite is that the hedrites have no intercrystalline links but the spherulite has. An important evidence is that the radius of gyration R _g is smaller than or equal to the addition of the lamellar thickness d _c and the half of amorphous thickness in the hedrites, while in the spherulite.24, 52, 53 Imaging easily, if R _g > 2d _c + d _a, a polymer chain can exist in two adjacent lamellar crystals simultaneously.24 A part of the polymer chain will exist in the amorphous layer between these two lamellar crystals, forming tie chains. Entangled loops with other chains will also exist. With the increases of d _c and d _a, tie chains and entangled loops will decrease. When , a chain cannot cross two lamellar crystals again, leading to the disappearance of tie chains. It almost cannot form entangled loops. When , the entangled loops will also disappear.24 Qiao et al determined d _c and d _a of PB-1 samples with different molecular weights at various crystallization temperatures using small-angle X-ray scattering (SAXS) and estimated R _g.24 It was found that for small molecular weight of PB-1, can be achieved at a lower temperature; however, for the high molecular weight of PB-1 sample PB0110M, this condition can be satisfied only above 100°C. PB0110M is exactly the sample employed in this study. This could be the reason that a few hedrites can be observed at 103°C.

Another difference of the hedrites from the spherulite is the orientation. With AFM, Fu et al found also that polymer chains in the hedrites were perpendicular to the surface of glass slide, while the polymer chains in spherulites were parallel to the glass surface.52 This point was confirmed recently by Tashiro et al using two-dimensional polarized IR imaging, SAXS and wide-angle X-ray scattering.27, 55

The morphological changes during heating were also recorded. All form II crystals disappeared at 120°C. In hexagonal-shape and octagonal-shape spherulites, the mother lamellae seemingly have a slightly higher thermal stability than the daughter lamellae. The daughter lamellae disappeared at 118°C, while the mother lamellae disappeared at 120°C (see Figure S1). The melting point of 120°C is close to the value reported for form II crystal.24 This implies that only form II was formed, since if form I was formed also some crystals would be found at a higher temperature. To further confirm the formation of the form II crystal, in situ FTIR measurement was carried out during isothermal crystallization at 103°C. Figure 3 shows IR spectra collected every 5 minutes during isothermal crystallization. It took 11 hours for complete crystallization. Three new bands appeared gradually, locating at 905, 1014, and 1296 cm⁻¹, respectively. As is well known,7, 10, 27, 56, 57 the bands at 905 and 923 cm⁻¹ are the characteristic bands of form II and form I, respectively. Only the 905 cm⁻¹ band appeared during isothermal crystallization, meaning that only form II formed. In the following, nucleation of form I within form II crystals will be investigated with form II square-shape hedrites and hexagonal-shape spherulite as representatives.

The formation of form II first could be attributed to fast dynamics of ethyl groups of PB-1. Compared to polyethylene and iPP, PB-1 has a long ethyl group. The ethyl groups have higher mobility in melt, as revealed by solid-state ¹³C NMR spectroscopy.44-46 It is not hard to image that high mobility of ethyl groups would prevent molecular stems from approaching too close. If two molecular stems stayed close, the ethyl group movement will be restricted significantly, leading to a large decrease of conformational entropy. As is well known, polymer crystallization depends on two factors, melt enthalpy, and conformational entropy. Polymer crystal has a lower enthalpy than melt. During polymer crystallization, the enthalpy decreases, leading to the decrease of Gibbs free energy. Therefore, it always favors polymer crystallization. Conformational entropy decreases also during crystallization due to formation of helices and ordered arrangement of the helices in the crystal. Nevertheless, it leads to the increase of Gibbs free energy, which is unfavorable for polymer crystallization. The ethyl groups have also fast dynamics in form II,45, 46 which will not lead to large decrease of conformational entropy. While ethyl groups in form I are completely fixed up to the melting point, which inevitably results in large loss in conformational entropy. Therefore, PB-1 would like to crystallize into form II first, such that conformational entropy does not need to decrease so much.

3.2 The role of annealing temperatures

In recent years, it was found that low-temperature annealing benefited form I nucleation.25 Wang et al called these low-temperature formed nuclei as latent nuclei, because conventional techniques, such as DSC, WAXD, POM, and FTIR, hardly detected them.49 They only can be inferred indirectly from subsequent nucleation effect. In order to observe these so-called latent nuclei, we increased the preparation temperature of form II crystals. In their experiment, the crystallization temperature was 85°C.49 At the lower temperature, it hardly formed large form II crystallites because of higher nucleation density. Besides, the annealing time was extended from 3 minutes to 1 hour. Although after such a long annealing time they may be not nuclei again, it can infer where the nuclei would like to form. This is what we are interested really. During experiments, we prepared first a large form II spherulite at the isothermal crystallization temperature of 103°C. After identifying a suitable spherulite or hedrites in POM, the temperature was decreased to an annealing temperature and kept for 1 hour. Then, the temperature was increased until the spherulite melted completely.

Figure 4 shows POMs taking during increasing temperature after annealing at various temperatures for 1 hour. For comparison, the spherulites before cooling were also given. Let us see first the case with an annealing temperature of 25°C. The first micrograph in Figure 4A is the spherulite obtained after isothermal crystallization at 103°C. It is a hexagonal-shape spherulite, consisting of a mother crystal and four daughter crystals. The second micrograph was taken after cooling to 25°C. Small crystals surrounding the large spherulite were formed during cooling. After annealing at 25°C for 1 hour, the temperature was increased again. The third micrograph was taken at 112°C. Small crystals surrounding the spherulite disappeared since their melting point was around 110°C. The lower melting point probably is due to smaller lamellar thickness. As is well known, a lower temperature is benefited to the formation of thin folded-chain lamellae. Further increasing temperature, the large spherulite started to melt. Daughter crystals disappeared gradually; nevertheless, mother crystal existed still at 130°C. At such a higher temperature, form II crystal cannot exist. The crystal was most likely form I crystal. From these micrographs, it can be inferred easily that the form I is preferable to nucleate in the mother lamellar crystals, or said, thick radial lamellar crystals.

If observing more carefully, we have another finding. Compared to form II spherulite in the first graph, the form I spherulite at 120°C and 130°C seems to be composed of numerous fibrillar crystals, similar as the finding in the study of Jiang.38 These fibrillar crystals can be seen more clearly at higher annealing temperatures (40°C, 50°C, and 55°C). The formation of fibrillar crystals should be resulted from thermal stress induced by difference in thermal expansion coefficients. The lower the annealing temperature, the bigger the thermal stress and the more the form I fibrillar crystals. In previous FTIRI studies,27, 32 it was found that form I crystals were easier to form at the edge of big spherulite, which was attributed to larger internal stress induced by big spherulite and small crystals around it. Actually, this could be due to unique crystallization behavior of PB-1. As seen in Figure 2, with the growth of the mother crystal, the lateral size increases constantly due to branching.58 During sequent annealing, more form I nuclei will be formed at the edge of the spherulite. Therefore, they are more easily found by FTIRI. There did not exist a lag in phase transition between the center and the edge of spherulite. In all of our experiments, form I crystals were observed simultaneously in the center and at the edge of the spherulite.

With the increase of the annealing temperature, the number of fibrillar crystals decreases constantly. At the annealing temperature of 25°C, numerous fibrillar crystals can be observed. However, after increasing the annealing temperature to 55°C, only a few fibrillar crystals can be observed. At 60°C or higher, all crystals melt at 120°C. This means that form I hardly nucleates at such high annealing temperatures. This finding is consistent with previous report by Wang, where form I nuclei were found only below 60°C.49 This indicates that form I nucleation is temperature dependence.

In Figure 4, all spherulites investigated are hexagonal-shape spherulites, because they are seen commonly at 103°C. We also observed nucleation in other types of crystals. Bright fibrillar crystals can be seen also in other types of spherulites but hardly seen in the hedrites. To show this point, we chose purposely a square-shape hedrites and a round-shape spherulite, which stayed close in a PB-1 sample. Although they were annealed simultaneously at 25°C for 1 hour, fibrillar crystals only can be seen clearly in the round-shape spherulite (see Figure 5). In the hedrites, fibrillar crystals were very weak. Compared to spherulite, the most important difference of the hedrites from spherulite is that no intercrystalline links exist.24, 52-54 This implies that intercrystalline links play an important role in promoting form I nucleation.

With above POM investigations, some conclusions can be obtained: (1) form I would like to nucleate first in thick radial lamellar crystals, leading to the formation of fibrillar crystals and (2) form I nucleation rate depends strongly on the annealing temperature. Only below 60°C, fibrillar crystals can be formed in a short time. (3) Intercrystalline links play an important role in promoting form I nucleation.

Despite having above findings, we have many things unclear. First, when do the fibrillar crystals form, on cooling, during annealing or on heating? Second, is the formation of fibrillar crystals immediate or time-consuming? Third, Can form I nucleate also in thin form II lamellar crystals also? In the following sections, we will answer these questions one by one.

To answer the first question, POM obviously cannot undertake such task, since it is hard to identify form I before melting of form II. To observe form I nucleation, in situ FTIR measurement was employed. Wang et al employed FTIR also to investigate form I nucleation.49 No characteristic peak of form I appeared during cooling. What was observed only was a constant increase of the band at 905 cm⁻¹, meaning that cooling only led to the formation of more kinetically favored form II crystals. Similar finding was found in our study (not shown here). Therefore, we directly show IR spectra collected during the annealing process.

Figure 6 shows IR spectra collected every 5 minutes during annealing at 25°C. During annealing, the bands at the 816, 850, 923, 1026, 1060, 1100, 1208, and 1329 cm⁻¹ increase constantly. In contrast, the bands at 905, 1014, 1112, 1150, 1214, and 1296 cm⁻¹ decrease continuously. As mentioned above, the band at 923 cm⁻¹ is the characteristic band of form I crystal, while the band at 905 cm⁻¹ is the characteristic band of form II crystal. The increase of the band at 923 cm⁻¹ in the price of the band at 905 cm⁻¹ means that form II transforms ceaselessly to form I during annealing process.

IR bands of PB-1 have been assigned successfully.56, 57, 59 For example, both bands at 905 and 923 cm⁻¹ were assigned to the coupling of the CH₂ rocking mode in the main chain r(CH₂) _m with the CH₃ rocking mode in the ethyl group r_a′(CH₃). The band at 816 cm⁻¹ was assigned to coupling vibration of the CH₂ rocking mode in the main chain r(CH₂) _m and the CC stretching mode in the ethyl group ν_s(CC) _s; the band at 850 cm⁻¹ was from the coupling of the CH₃ rocking mode in the ethyl group r_a′(CH₃) and the CC stretching mode in the main chain ν_s(CC) _m, while the band at 1208 cm⁻¹ was due to the coupling of the CH₂ twisting t(CH₂) _m and the CH wagging modes in the main chain ω(CH). To differentiate CH₂ and CC groups in the main chain and the ethyl group, the subscripts m and s were added, which represent the groups in the main chain and the ethyl group, respectively. Vibrations of the CH₃ and CH groups were not noted with the subscripts because they only exist in the ethyl group and main chain, respectively.

From above assignments, it can be found that all of characteristic bands are from couplings of the vibrations in the main chain and the ethyl group. Their intensities depend on spatial relationship of the main chain and the ethyl group or said helical conformation. When the helical conformation changes from 11₃ to 3₁, characteristic bands will change also, for example, transition from the 905 cm⁻¹ band to 923 cm⁻¹ band, as shown in Figure 7.

IR spectra collected at t = 0 and 1 hour are given in Figure 6C. Characteristic bands of form I at 816, 850, and 923 cm⁻¹ are absent at t = 0 hour; however, they can be seen clearly at t = 1 hour. This indicates that form I indeed can nucleate at the annealing temperature of 25°C. In the study of Wang group, the annealing time was only 3 minutes.49 In such a short time, it hardly formed enough form I nuclei to be detected by FTIR or WAXD.

The change of form I crystal during annealing at 25°C was monitored with the band at 816 cm⁻¹. The band at 923 cm⁻¹ was not employed because it overlapped seriously with the band at 905 cm⁻¹. The band at 816 cm⁻¹ is also sensitive to form I nucleation, which can be also regarded as the characteristic band of from I.56 From the beginning, the peak area at 816 cm⁻¹ increased (see Figure 5D). No induction time was observed, indicating that the phase transition is spontaneous.

The phase transition at the annealing temperature of 60°C was also monitored with FTIR. Unlike at 25°C, phase transition at 60°C was very slow. After 24 hours, only a trace of form I was formed (see Figure 8A). The slow phase transition only led to a tiny form I crystal after 1 hour, which probably was the reason that form I was not observed in POM measurements. We monitored also the change of form I with the peak at 816 cm⁻¹. Similar as that at 25°C, the peak area at 816 cm⁻¹ increased from the beginning (see Figure 8B). No induction time was observed. This indicates again that the phase transition is spontaneous. And form I nucleation depends strongly on the annealing temperature.

Above FTIR results actually implies also that the formation of form I fibrillar crystals is time-consuming. To see the evolution process of form I fibrillar crystals, we returned to POM. Figure 9 shows crystalline morphologies taking during increasing temperature after annealing at 25°C for 15 minutes or 12 hours. When a large spherulite formed at 103°C was annealed for 15 minutes, only a few fibrillar crystals were observed at 128°C. However, when the annealing time was extended to 12 hours, thousands of fibrillar crystals were found at the temperature. This means that the formation of fibrillar crystals was not immediate.

3.3 The role of crystallization temperatures

In Figures 4 and 5, form I nucleates only in the large spherulite. No form I fibrillar crystals can be seen around the spherulite. Around the spherulite, only thin form II lamellar crystals formed since these crystals formed during cooling. Will these thin lamellar crystal accommodate form I nuclei? This could be a more important question, since most of PB-1 pipes form at a lower temperature in real processing. To clarify such question, form I nucleation in a lower temperature crystallized form II sample was investigated also with POM. The temperature of 90°C was selected as the crystallization temperature. At such a temperature, the nucleation density was too high to identify single form II spherulite, and crystallization completed within 20 minutes. Figure 10 shows micrographs at 128°C after annealing at various temperatures for 1 hour. Different from large spherulite formed at 103°C, only below the annealing temperature of 40°C, fibrillar crystals formed. And only below 0°C, a great number of clear fibrillar crystals were observed. This means that the lower the crystallization temperature, the harder the form I nucleation. Or said, the thinner the lamellar crystal, the harder the form I nucleation, which is also seen in Figure 9.

The change of form I during the phase transition of the 90°C crystallized sample at 25°C was also monitored with FTIR. Figure 11 shows the change of peak area at 816 cm⁻¹ during phase transition. For comparison, the change of peak area at 816 cm⁻¹ in the 103°C crystallized sample was also plotted. The transition in the lower temperature crystallized sample is much slower than that in the higher temperature crystallized sample, as reported by Azzurri et al.40 The half time of transition for the 103°C crystallized sample is 4.75 hours, while 6.32 hours for the 90°C crystallized sample. The higher transition rate in the higher temperature crystallized sample can be attributed to the higher nucleation density, deducing from Figures 4 and 10.

3.4 The role of growth temperatures

As mentioned above, except the crystallization temperature and the annealing temperature, there exists also a temperature affecting the phase transition, that is, the growth temperature. Actually, Men et al had investigated the transformation rate dependence on the growth temperature.25 In their study, standard form II samples were annealed first at various temperatures for same time and then subjected to increasing temperature. From melting peaks in DSC, the transformation rates at various annealing temperatures were inferred. Such indirect method is rigorous but time-consuming. To determine a transformation rate at one temperature, it needed at least 4 hours. Here we employed a fast and direct method. After isothermal crystallized sample was annealed 25°C for a time, the temperature was increased by 5°C every 5 minutes. Before every increasing temperature, an IR spectrum was collected. The increased peak area at 816 cm⁻¹ within 5 minutes was regarded as the increased form I at the temperature. Specifically, we collected an IR spectrum at 25°C. After 5 minutes, another spectrum was collected at the temperature. Immediately after the collection, the temperature was increased immediately to 30°C. At t = 10 minutes, an IR spectrum was collected again and then the temperature was increased to 35°C. The times of every increasing temperature and collection of FTIR are short, normally just a few seconds, which can be omitted. The difference between the first and the second spectra at 816 cm⁻¹ is increased form I at 25°C, while the difference between the second and the third spectra is increased form I at 30°C. In such way, it can obtain transformation rates between 25 and 80°C in an hour.

Figure 12A shows IR spectra collected during increasing temperature after annealing at 25°C for 3 hours for a form II sample formed at 103°C. The difference between adjacent spectra increases first and then decreases. The difference between 35°C and 40°C is the maximum, meaning the maximum transformation rate at 40°C, as found by Men and coworkers.25 To see the transformation rate dependence on the temperature more clearly, the difference in peak area between adjacent spectra was plotted in Figure 12B. Indeed, the transformation rate increases first and then decreases. At 40°C, it obtains the greatest transformation rate. Above 75°C, it becomes zero. The temperature of 75°C should be the highest growth temperature for form I. The inhibition of the phase transition at higher temperature could be attributed to fast dynamics of ethyl groups. With NMR, Miyoshi et al found that at 100°C, side chains in form II vibrated in extremely high frequency.46 Imaging easily, fast vibration will prevent one stem from approaching another stem to form I nucleus or crystal.

Transformation rates in the 90°C crystallized sample were also determined. Similar as the finding in the 103°C crystallized sample, the transformation rate increases first and then decreases. At 40°C, it reaches the greatest transformation rate, and at 75°C, the transformation rate becomes zero (also see Figure 12B).

One probably questions the rigorousness of the method. Even at a constant temperature, crystallization rate changes. Polymer crystallizes first slowly, then rapidly and finally slowly again.60 The first slow crystallization is due to nucleation, while the second slow crystallization is due to secondary crystallization. In this study, the transformation rate was determined after annealing for 3 hours. As seen in Figure 6D, the transformation rate has entered the linear region after 3 hours in both samples. It would not enter the later stage because the measurement time was only 1 hour. Therefore, it does not need to worry about the rigorousness of the method.

After identifying the best temperatures for nucleation and growth of form I, it could obtain faster phase transition. As seen in Figure 10, higher nucleation density can be seen at 0°C, while from Figure 12B, the maximum growth rate locates at 40°C. If annealing first at 0°C and then at 40°C, it could obtain faster phase transition. Figure 13A shows IR spectra collected during phase transition at 40°C after annealing at 0°C for 3 hours. Form II can phase transit to form I rapidly. Within 3 hours, the phase transition has finished. It is much faster than that at 25°C without annealing. The half time of the phase transition is reduced from 6.25 to 0.68 hour (see Figure 13B). It is only one-ninth of previous one if not counting the annealing time.