Synchronous architecture of ring-banded and non-ring-banded morphology within one spherulite based on in situ ring-opening polymerization of cyclic butylene terephthalate oligomers

Abstract

Ring-banded morphology of polymerized cyclic butylene terephthalate (pCBT) accompanied by the synchronous evolution of non-ring-banded patterns was investigated for the first time through the crystallization of pCBT prepared by ring-opening polymerization (ROP) of cyclic butylene terephthalate oligomers (CBT). The result of gel permeation chromatography (GPC) shows that pCBT has a broader molecular weight distribution than commercial poly(butylene terephthalate) (PBT). The polar optical microscope (POM), atomic force microscopy (AFM) and scanning electron microscope (SEM) demonstrate that the ring-banded patterns originate from the lamellae twisting. The radial growth rates, in situ evolution and melting observation from the POM prove that the crystals' melting point in ring-bands is lower than that in non-ring-banded parts. And the crystals in ring-bands are deduced to be constructed by the low molecular weight pCBT fractions. Finally, the possible formation mechanism is discussed.

---

Introduction

Polymer ring-banded spherulites have been investigated for more than half a century.¹ There are several competing models to interpret the bands' formation mechanism including lamellar crystal twisting (Keith and Padden^2,3), screw dislocation queues (Schultz⁴ and Bassett⁵), compositional fields (Schultz⁶) and stress fields (Toda^7,8) etc. The most popular explanation is periodic twisting of ribbon-like crystalline lamellae along the radial growth direction of the spherulites. And the unbalanced surface stresses as the mechanical origin of lamellae twisting has been intensively reviewed by Lotz and Cheng.⁹ In addition, the structural discontinuities caused by rhythmic crystal growth have been proved to be the other kind of mode suit for ring-banded spherulites in films.^10–16 Nevertheless, there are special ring-banded morphologies. Liu et al.¹⁷ disclosed a morphology composed of ring-banded and non-ring-banded regions within one poly(butylene adipate) spherulite and demonstrated that this structure originated from the crystal forms changing between alpha and beta. Recently, Li et al.¹⁵ reported the nested ring-banded structure of poly(ethylene adipate) induced by rhythmic growth with lamellae twisting.

Molecular weight (MW), as an important parameter, has significant effect on polymer crystallization behaviors and morphology.^18–23 For example, it is difficult to obtain the ring-banded morphology of poly(L-lactide) (PLLA) though the conventional crystalline schedule of melting followed by quenching to crystallization temperature (T_C). However, Xu et al.²⁴ reported the PLLA ring-banded spherulite through annealing followed by slow cooling to T_C for isothermal crystallization, and proved that the part decrease of MW resulting from thermal degradation maybe a reason for the appearance of ring-bands. Since MW is a typical factor for polymer crystallization, it is possible that the spherulite composed of ring-banded and non-ring-banded regions can be obtained by using polymer with large MW distributions.

To substantiate this idea, cyclic butylene terephthalate (CBT) oligomers was chosen as an example. CBT is composed of two to seven cyclic components. Ring-opening polymerization (ROP) of CBT has received considerable attention because its melt viscosity is only about ca. 30 mPa s at 190 °C and its athermal ROP can quickly occur without evolution by-products.^25–29 The polymerized CBT, (named as pCBT to distinguish from conventional polycondensed poly(butylene terephthalate) (PBT)) is structurally equal to PBT, a well-known engineering thermoplastic. Because the ROP of CBT and crystallization of pCBT could occur simultaneously around 190 °C, and the crystalline solidification of pCBT during ROP might result in the large distributions of MW.^25–29 Attractively, at the same T_C, ring-banded morphology accompanied by the synchronous evolution of non-ring-banded parts is observed for the first time.

Results and discussion

This kind of pCBT ring-banded spherulite was firstly studied by polarized optical microscope (POM). As labeled with arrows in Fig. 1, non-ring-banded (arrow 1) and ring-banded regions (arrow 2) co-constitute an integral pCBT spherulite, in which the non-ring-banded regions show axialite morphology and the other parts are symmetrically filled by ring-bands. The ring-bands' period is about 0.5–1.5 μm. The alternative change between blue and yellow color (rectangular region in Fig. 1) indicates that the ring-bands may be originated from the periodic change of lamellae orientation.²⁴

Fig. 1 POM image of ring-banded and non-ring banded morphology within one pCBT spherulite, inserted image indicates the enlarged view corresponding to the red rectangle. Scale bar: 20 μm.

In order to clearly observe the morphology, the obtained sample is immersed in trichloromethane (CHCl₃) to dissolve the residual CBT crystals which seriously impact the AFM scan (Fig. ESI1 in ESI†). Accordingly, the ring-banded patterns are presented in Fig. 2a–c. The flat-on lamellae can be distinguished (Fig. 2d) in valley bands, and bands in ridge are probably consisted of edge-on lamellae bundles. Also the ridge (bright) and valley (grayish black) ring-bands can be observed by SEM (Fig. 3b and d). Moreover, Fig. 3e shows a twisted lamella originated from the branching of its parent lamella (arrow in Fig. 3e).

Fig. 2 AFM images. (a) Height image, inserted image corresponding to POM result; (b) enlarged view of rectangular region in (a), inserted image represents height profile of the line in (b); (c) enlarged view of rectangular region in (b); (d) phase image corresponding to rectangle in (c).

Fig. 3 (a) Overall SEM view; (b–e) enlarged view, respectively, ellipse (e) indicates a twisted lamellar crystal.

To elucidate the evolution mechanism of this kind of ring-banded spherulite, in situ POM investigation is performed and the final image is presented in Fig. 4a. Intriguingly, four kinds of morphologies appear simultaneously. They are half versus half morphology of ring-banded and non-banded parts (labeled with number 1); coexisting morphology but ring-bands are dominant (number 2); normal ring-banded spherulite (number 3) and conventional spherulite enclosed by ring-bands (number 4). The whole evolution scenarios can be found in Fig. ESI2 in ESI.†

Fig. 4 (a) POM images of four spherulite's morphologies obtained at 170 °C for 30 min; (b) in situ scenarios of spherulite growth labeled with 1 in (a); (c) in situ melting process from 180 °C of spherulite labeled with 1 in (a). Note that the contrast ratio of some images has been adjusted for better observation. Scale bars for b and c: 10 μm.

When we focus on the number 1 spherulite in Fig. 4a, ring-banded and non-ring-banded regions synchronously grow up at initial stage (Fig. 4b and ESI2 in ESI†) and they have almost same radial growth rates, G, (within (t₀ + 2 min)). The detailed plots of radial growth rates are displayed in Fig. 5, which shows that ring-bands' growth rates, however, become faster than that in non-ring-banded regions with the lapse of time. They all experience moderating process before (t₀ + 10) min. (Non-linear radial growth rates had been intensively studied by Keith and Padden.^18,19) Hereafter, both of them reach a constant rate and non-ring-banded regions stop growing at a certain time. Therefore, as long as there are enough material supply and sufficient time, spherulite will be enclosed by ring-bands (spherulite 1, 2 and 4 in Fig. 4a). Cheng et al.²¹ has concluded that the crystal growth rates decrease with increasing MW in the study of poly(ethylene oxide) molecular segregation. This suggests that, in our case, the MW in ring-banded regions maybe lower than that in non-ring-banded regions.

Fig. 5 Radial growth rates, G, of number 1 spherulite with ring-banded and non-ring-banded morphology in Fig. 4a.

Therefore, the MW of samples was measured by gel permeation chromatography (GPC) (Fig. 6). The number average molecular weight (M_n) are 28.6 and 33.6 kg mol⁻¹ obtained from ROP of CBT at 190 and 230 °C, named as pCBT190 and pCBT230, respectively. Although commercial PBT has the same crystal form, chain structure^25–29 and similar M_n (35.9 kg mol⁻¹) to pCBT, the similar morphology cannot be achieved. It is worth mentioning that there is a certain amount of un-reacted CBT in the sample from the results of GPC (Fig. 6).^25,27 In addition, the cyclic oligomers, about 5 (wt)%, in polyester or polycarbonate has been pronounced and investigated for a long time.³⁰ In Fig. 6, there are almost the same two characteristic peaks for the three samples, located by the rectangular dashed line, which can be ascribed to the cyclic species. However, an excess broad peak marked by arrows in Fig. 6a and b for pCBT190 and pCBT230, which should be the low MW pCBT and lead to the wider polydispersity index (PDI = 2.59 and 2.90, respectively) than PBT4500 (PDI = 2.08) (shown in Table 1). This agrees well to the Parton's GPC results.²⁷ Moreover, all of the samples' MW are beyond the critical entanglement M_e of PBT, 50 kg mol⁻¹.²⁷ It is necessary to know this, since for most ring-bands' formation mechanism models, entanglement is required. Thus, the difference is PDI. The PDI of pCBT obtained from ROP of CBT at 190–230 °C in hot stage are higher than commercial PBT. And there is a shoulder peak in GPC curves of pCBT (Fig. 6b), the corresponding M_n of these shoulder peaks for ROP at 190 and 230 °C is 3.71 and 4.01 kg mol⁻¹, respectively, which can be ascribed to the low MW pCBT. Therefore, the broad PDI maybe the key factor for the formation of this kind spherulite morphology.

Fig. 6 GPC results (a) and enlarged view (b).

Table 1 Summary of MW data

Sample	Mn (kg mol^−1)	MW (kg mol^−1)	PDI
PBT4500	35.9	74.7	2.08
pCBT190	28.6	73.9	2.59
pCBT230	33.6	97.5	2.90

---

But the difficulty to advance the understanding of its evolution is how to accurately know the difference of MW in such intimate two regions. And the experimental separation of crystals in two regions within one spherulite seems to be unrealistic. For simplification, according to Hu's method,³¹ we postulate that pCBT in number 1 spherulite is composed of long-chain (L-pCBT) and short-chain (S-pCBT) fractions. Thus, combined with the residual CBT,²⁸ the system can be considered as a ternary blends (L-pCBT/S-pCBT/CBT). Thereupon, the molecular segregation^18–22,31 may be a scientific possibility to explain the issue mentioned-above. Namely, low MW fractions are rejected by the crystalline phase of high MW fractions, in which, low MW fractions are unable to crystallize under the given conditions and will crystallize at longer time or lower T_C. In our case, as a result, L-pCBT crystallize firstly during quenching to a given T_C, and the rejected S-pCBT will crystallize at later time.

Moreover, it is well known that the melting point (T_m) of polymer crystals not only depends on the lamellae thickness, but the polymerization degree or MW. On the one hand, Balijepalli and Schultz²³ proved that the thin lamellae were mainly constructed by the low MW fractions, on the contrary, crystals are thick. Unfortunately, to the best of our efforts, the lamellae thickness in both regions are not be successfully measured by AFM because the details of sample surface are always masked by residual CBT.²⁸ Alternatively, the software of Digital Micrograph was used to measure the lamellar crystal thickness according to the method in the literature.³² Within one sample, as displayed in Fig. 7 and 8, at least five solo edge-on lamellar crystal (like the illustration in left image of Fig. 9) was selected. The results of lamellae thickness between the non-ring-banded spherulite and the ring-banded spherulite (Fig. 8 and 9) shows that the lamellae thickness are 11.9 ± 0.2 nm vs. 12.1 ± 0.4 nm for non-ring-banded and ring-banded spherulite, respectively. In addition, the possible error from platinum sprayed on the sample surface can be effectively avoided due to the lamellae for thickness detecting comes from the same sample and the same measure method, illustrated in Fig. 9.

Fig. 7 POM image of ring-banded and non-ring banded morphology within one pCBT spherulite. Scale bar: 20 μm.

Fig. 8 SEM images obtained from the same sample corresponding to Fig. 7 and etched by methylamine vapor. (a–c) normal pCBT spherulite; (a′–c′) ring-banded and non-ring-banded morphology within one pCBT spherulite.

Fig. 9 Illustration for the thickness measurement for edge-on lamellae corresponding to Fig. 8c.

Thus, the MW should be a decisive factor for T_m. On the other hand, assisted by the Flory equation³³ (eqn ESI1 in ESI†), low MW results in more chain-ends which act as impurity^18,19 and frustrate the crystals' T_m. In this study, the in situ melting process (Fig. 4c and ESI3 in ESI†) demonstrates that the T_m in ring-banded parts is lower than that in non-ring-banded parts. The T_m depression in ring-banded parts, therefore, may originate from the crystals that are constructed by the S-pCBT chains. Based on the former discussions, briefly, in this ternary system, the crystals in ring-banded regions should be constructed mainly by the S-pCBT. This is also strengthened by the POM result in Fig. ESI4.† All of these discussions are mutually supported by the in situ melting process (Fig. 4c), radial growth rates (Fig. 5) and GPC results (Fig. 6).

Herein, the remaining issue is why the crystals of ring-banded regions adopt twisting fashion? Firstly, under the same T_C, L-pCBT fractions crystallize preferentially and build spherulite's backbone, accordingly, the schematic view is displayed in Fig. 10a–d on the basis of in situ AFM studies of spherulite evolution from Li and collaborators.^34,35 During this time, S-pCBT fractions and the residual CBT are rejected by the crystalline phase of L-pCBT (Fig. 10e). When the S-pCBT begin to crystallize, however, only the residual CBT with ultra-low melt viscosity are rejected, and S-pCBT (purple color in Fig. 10f) will easily diffuse in the CBT melt (green balls in Fig. 10f). As displayed in Fig. 3e and 10f, the twisted lamellae are derived from the branching of its parent lamella. Here, Toda et al.^7,8 has proved that the stress from pressure gradient between crystals and melts result in the instability of crystal growth fronts, which can drive lamellar crystal to branch accompanied with the propagation of dislocation. Thus, the branched lamellae adopt re-orientation (edge-on or flat-on) and finally produce the ring-banded morphology.^7,8,36

Fig. 10 (a–d) Schematic evolution of number 1 spherulite in Fig. 4a; (e) non-ring-banded part; (f) ring-banded part and the lamellae twisting model.

Secondly, because the lamellae in ring-bands are mainly constructed by S-pCBT and there are more chain-ends than that of L-pCBT, and these chain-ends will be excluded by crystals growth and deposited on its surface.^3,9 Therefore, the unbalanced surface stress of lamellae in ring-banded regions will be larger than that of non-ring-banded regions. Oppositely, L-pCBT crystals mainly reject the S-pCBT and CBT. The melt viscosity of L-pCBT crystal growth front is larger than that of S-pCBT which only reject CBT based on Fig. 10e and f. Namely, lamellae in ring-bands may have larger unbalanced surface stress than those in non-ring-banded regions, but they only need to conquer the lower resistance force from low viscosity CBT melt because the melt viscosity of pCBT is larger with 10⁴ magnitude than CBT.³⁷ In a word, the re-orientation difficulty of a branched baby lamella in ring-banded parts is smaller than that in non-ring-banded regions.

Thirdly, it has been proved that the formation of ring-banded spherulites for some polymers can be achieved by blending with other compatible polymer due to the dilution effect.^36,38,39 Fig. 11 shows that there is no ring-banded morphology in the sample after washing off the residual CBT. Namely, when the sample is quenched from 250 °C to T_C (160–185 °C), above the main melting peaks of CBT (120–160 °C) (Fig. ESI5 in ESI†), the molten CBT, as a self-compatible “diluent”, may promote the S-pCBT to form ring-banded patterns. The predicted content of CBT should locate from 5 to 10% (wt) for the formation of ring-banded structure although it is difficult to accurately record the content of CBT. In addition, the appearance of ring-bands often need suitable T_C window. In our study, pCBT shows distinct ring-bands at relatively low T_C (160–185 °C), on the contrary, no well-defined ring-bands can be found at high T_C (>185 °C). Furthermore, PBT has no ring-banded spherulites' report,⁴⁰ which is also strengthened by our POM results (Fig. ESI6 in ESI†).

Fig. 11 In situ POM images of melting process (10 °C min⁻¹) followed by quenching (liquid nitrogen) to T_C for second crystallization, sample obtained from hot stage was treated in CHCl₃ for 10 min to wash off the residual CBT. (a) 185 °C; (b) 195 °C; (c) 205 °C; (d) 250 °C; (e) 185 °C; (f) 170 °C. Arrows represent a reference point for accurate recording the location. Scale bars: 20 μm.

Experimental

Samples were prepared by solution casting on cleaned glass slide with 5% (m/v) CHCl₃ solution of CBT (CBT100, Cyclics Corporation) and 0.45 wt% catalyst (butyltin chloride dihydroxide, Fascat 4101, Sigma-Aldrich). After the solvent evaporation, the obtained film was dried at 105 °C for 4 h in vacuum oven. Morphology observation was achieved by POM (Olympus BX 51) equipped with a CCD camera and hot stage (Instec H601). The ROP and crystallization are performed according to the following steps: (1) heating up to 190–230 °C at 100 °C min⁻¹ and holding for 10–30 min; (2) heating up to 250 °C at 10 °C min⁻¹ and holding for 5 min; (3) quenching (liquid nitrogen) to TC (160–185 °C) to isothermal crystallization. Obtained sample was directly used for scanning probe microscope (SPM, Veeco Dimension 3100V) observations by tapping mode. SPM was performed with a SPA-300HV AFM equipped with a SPI 3800N controller. Sample surface sprayed by platinum was photographed by SEM (Hitachi S4800) at an accelerating voltage of 15 kV. GPC (PL-GPC50, detector worked at 254 nm) was used to determine the MW with 10 μL injection volume, 0.1 mg mL⁻¹ concentration from phenol-based solvent and the polystyrene standards.

Conclusions

In summary, ring-banded morphology of pCBT accompanied by the synchronous evolution of non-ring-banded patterns was investigated, and the ROP of CBT can result in diverse morphologies of pCBT spherulite. The results shows that pCBT has broader molecular weight distribution and ring-banded patterns originate from the lamellae twisting. The radial growth rates, in situ evolution and melting observation from POM prove that the crystals' melting point in ring-bands is lower than that in non-ring-banded parts. And the crystals in ring-bands are deduced to be constructed by the low molecular weight pCBT fractions. There are multifactor synergies that contribute to the synchronous evolution of ring-banded and non-ring-banded parts within one spherulite, including molecular segregation, lamellae twisting from unbalance surface stress and diluent effect. This widens the view and understanding of PBT's crystalline morphology. Especially, it provides an excellent exemplification, that crystallization morphology of polymer is determined by the crystal structure itself, as well as by the complicated growth process. And the systemic study is ongoing.

Acknowledgements

This effort was supported by the National Natural Science Foundation of China (21364004), Gansu Province University Fundamental Research Funds and Doctor Research Fund of Lanzhou University of Technology, P. R. China. And we are grateful to Prof. B. Lotz, (Université de Strasbourg, France), Prof. A. Toda, (Hiroshima University, Japan) and Prof. J. Xu (Tsinghua University, P. R. China) for their helpful discussions during 11th International Symposium on Polymer Physics.

Notes and references

1. A. Keller, J. Polym. Sci., 1955, 17, 291.
2. H. D. Keith and F. J. Padden, Polymer, 1984, 25, 28.
3. H. D. Keith and F. J. Padden, Macromolecules, 1996, 29, 7776.
4. J. M. Schultz and D. R. Kinloch, Polymer, 1969, 10, 271.
5. D. C. Bassett and A. M. Hodge, Polymer, 1978, 19, 469.
6. J. M. Schultz, Polymer, 2003, 44, 433.
7. A. Toda, M. Okamur, K. Taguchi, M. Hikosaka and H. Kajioka, Macromolecules, 2008, 41, 2484.
8. A. Toda, K. Taguchi and H. Taguchi, Macromolecules, 2008, 41, 7505.
9. B. Lotz and S. Z. D. Cheng, Polymer, 2005, 46, 577.
10. T. Kyu, H. W. Chiu, A. J. Guenthner, Y. Okabe, H. Saito and T. Inoue, Phys. Rev. Lett., 1999, 83, 2749.
11. Y. X. Duan, Y. Jiang, S. D. Jiang, L. Li, S. K. Yan and J. M. Schultz, Macromolecules, 2004, 37, 9283.
12. J. Chen and D. C. Yang, Macromol. Rapid Commun., 2004, 25, 1425.
13. Z. B. Wang, Z. J. Hu, Y. Z. Chen, Y. M. Gong, H. Y. Huang and T. B. He, Macromolecules, 2007, 40, 4381.
14. Z. B. Wang, G. C. Alfonso, Z. J. Hu, J. D. Zhang and T. B. He, Macromolecules, 2008, 41, 7584.
15. Y. G. Li, H. Y. Huang, T. B. He and Z. B. Wang, ACS Macro Lett., 2012, 1, 154.
16. Y. G. Li, H. Y. Huang, T. B. He and Z. B. Wang, ACS Macro Lett., 2012, 1, 718.
17. J. Liu, H. M. Ye, J. Xu and B. H. Guo, Polymer, 2011, 52, 4619.
18. H. D. Keith and F. J. Padden, J. Appl. Phys., 1964, 35, 1270.
19. H. D. Keith and F. J. Padden, J. Appl. Phys., 1964, 35, 1286.
20. A. Mehta and B. Wunderlich, Colloid Polym. Sci., 1975, 253, 193.
21. S. Z. D. Cheng and B. Wunderlich, J. Polym. Sci., Part B: Polym. Phys., 1986, 24, 595.
22. P. Kohn, S. Huettner, U. Steiner and M. Sommer, ACS Macro Lett., 2012, 1, 1170.
23. S. Balijepalli and J. M. Schultz, Macromolecules, 1996, 29, 6601.
24. J. Xu, B. H. Guo, J. J. Zhou, L. Li, J. Wu and M. Kowalczuk, Polymer, 2005, 46, 9176.
25. D. J. Brunelle, J. E. Bradt, J. Serth-Guzzo, T. Takekoshi, T. L. Evans, E. J. Pearce and P. R. Wilson, Macromolecules, 1998, 31, 4782.
26. A. R. Tripathy, A. Elmoumni, H. H. Winter and W. J. MacKnight, Macromolecules, 2005, 38, 709.
27. H. Parton, J. Baets, P. Lipnik, B. Goderis, J. Devaux and I. Verpoest, Polymer, 2005, 46, 9871.
28. J. Q. Zhang, Z. B. Wang, B. J. Wang, Q. T. Gou, J. W. Zhang, J. Zhou, Y. Li, P. Chen and Q. Gu, Polym. Chem., 2013, 4, 1648.
29. J. Q. Zhang, Z. F. Li, Z. Zhang, H. X. Feng, Z. B. Wang, Y. Li, P. Chen and Q. Gu, Chin. J. Polym. Sci., 2015, 33, 1104.
30. D. J. Brunelle, Cyclic polymers, Kluwer Academic Publishers, 2nd edn, 2000, ch. 6, pp. 185–228.
31. W. B. Hu, Macromolecules, 2005, 38, 8712.
32. Y. M. Shi, W. Zhou, A. Y. Lu, W. J. Fang, Y. H. Lee, A. L. Hsu, S. M. Kim, K. K. Kim, H. Y. Yang, L. J. Li, J. C. Idrobo and J. Kong, Nano Lett., 2012, 12, 2784.
33. B. Wunderlich, Crystal Melting, Macromolecular Physics, Academic Press, New York, 1976, vol. 3, p. 104.
34. L. Li, C. M. Chan, J. X. Li, K. M. Ng, K. L. Yeung and L. T. Weng, Macromolecules, 1999, 32, 8240.
35. L. Li, C. M. Chan, K. L. Yeung, J. X. Li, K. M. Ng and Y. G. Lei, Macromolecules, 2001, 34, 316.
36. E. M. Woo and P. L. Wu, Colloid Polym. Sci., 2006, 284, 357.
37. Z. A. Ishak, K. G. Gatos and J. Karger-Kocsis, Polym. Eng. Sci., 2006, 46, 743.
38. T. C. Wang, H. H. Li, F. Wang, J. M. Schultz and S. K. Yan, Polym. Chem., 2011, 2, 1688.
39. L. X. Zeng, M. J. Liu, S. C. Chen, J. B. Zeng and Y. Z. Wang, Polym. Chem., 2012, 3, 2537.
40. Y. F. Chen, E. M. Woo and S. H. Li, Langmuir, 2008, 24, 11880.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20471f

---