Novel polymorphism behavior of poly(butylene adipate) in its nanocomposites with carbon nanofibers

Abstract

Nanocomposites formed between poly(1,4-butylene adipate) (PBA) and carbon nanofibers (CNF) were prepared and characterized. The results showed that CNF had significant α-form crystal nucleating ability on PBA, and the α-form crystals could appear at rather low crystallization temperatures in PBA/CNF composites, for example 0 °C. Furthermore, a novel plateau content phenomenon of α-form crystals was discovered, and the value of the plateau content almost increased linearly with increasing CNF loading content. Microscopic morphology data suggested that PBA chains could wrap (or coat) around the nanofillers after the solution mixing process to form an interaction region. And in situ FTIR spectra proved the existence of multiple-weak CH–π interactions between CH₂ groups on PBA chains and sp² structures on the surface of CNF in composites, which helped to induce some precursory structures in the interaction region containing a chain conformation close to that in α-form crystals. Thus, the wrapping (or coating) behavior and precursory structures are responsible to the formation of α-form crystals at low temperature and the appearance of a novel plateau content phenomenon of α-form crystals. The change of PBA polymorphism behavior in nanocomposites provides a pathway for evaluating the interaction between polymer chains and nanofillers more intuitively.

---

Introduction

One-dimensional carbonous nanofillers have been widely used as noticeable nucleating agents for various semi-crystalline polymers, such as carbon nanofibers (CNF)¹ and carbon nanotubes (CNT).² They effectively enhance the crystallization kinetics of different polymers, induce certain crystal forms, and manipulate spherulitic morphology, such as, the formation of β-form crystals is favored in syndiotactic polystyrene (sPP) after the addition of CNT,³ both multi-wall carbon nanotubes (MWCNT)⁴ and CNF⁵ can induce the formation of β-phase poly(vinylidene fluoride) (PVDF). The mechanism for the formation of β-form PVDF crystals induced by MWCNT was speculated to be that PVDF chains with zigzag conformation could be more easily formed due to π-electrons interaction with the CNT surface.⁴ So the presence of higher CNT content in the PVDF composites is beneficial for promoting PVDF β-phase formation during stretching. The conformation preference of polymer chains absorbed on nanofiller surface would reduce the potential barrier of nucleation for a certain crystal form before the crystallization. The important function of CNT and CNF driving the formation of preferred chain conformation in polymer matrix is associated with their special features, including the long slenderness ratio and the sp² sidewall structures.

Polymorphism phenomenon provides a suitable pathway to study the influence of nucleating agent on polymer. Therefore, poly(1,4-butylene adipate) (PBA), an aliphatic polyester with polymorphism, was often chosen as study subject.^6–8 The crystal form of PBA is very sensitive to the crystallization temperature. The α-form crystal was characterized as a monoclinic unit cell, adopting skew CH₂–CH₂–O–C(=O) conformers and a P2₁/n space group; while β-form crystal was orthorhombic unit with planar zigzag conformation.^9–11 It is considered that α-form crystal is thermodynamically stable, while the β-form crystal is kinetically favored.¹² The pure α- and β-form PBA crystals form in the temperature ranges above 32 °C and below 28 °C, respectively. In the temperatures between 28 °C and 32 °C, both crystal forms appear.^13,14 The polymorphism phenomenon decides the physico-chemical properties of PBA, including spherulitic morphology, thermal and enzymatic stability.^12,15,16 Therefore many works had been carried out on controlling the PBA crystal structure by epitaxial growth,¹⁷ copolymerization,^18,19 blending,^20,21 inclusion complex method,^22,23etc. Many researches showed that adding nucleating agent would promote the formation of α-form crystals, while the mechanism for this phenomenon was suggested differently in different PBA composites. Uracil, talc and carbon-functionalized MWCNT were reported to improve the heterogeneous nucleation resulting in the PBA crystallization in the α-form crystal formation temperature range during the nonisothermal crystallization process.^24–26 The multi methyl-benzylidene sorbitol (TM6) was suggested to reduce the surface free energy of PBA nucleus, so that the formation of thermodynamically stable α-form crystals was favorable than kinetically preferential growth of β-form crystals.²⁷ Tang et al. speculated that PBA molecular chains were anchored at the surface of hexagonal boron nitride nanosheets (BNNSs), which helped to induce the formation of α-form crystals in response to “memory effect”.²⁸ Another possible explanation they offered was the nuclei density increases tremendously due to the heterogeneous nucleation property of BNNSs resulting in β-form crystals which have slower primary nucleation rate compared with α-nuclei had no time to (secondary) nucleate on the periphery of α-form spherulites. The above mentioned possible mechanisms are interlinked at some level, so they might not work alone in particular PBA composites, but work cooperatively.

In this work, the PBA/CNF nanocomposites were prepared by solution blending. The crystallization behavior of nanocomposites was detected using various methods including differential scanning calorimetry (DSC), wide angle X-ray diffraction (WAXD), and the Fourier transform infrared spectroscopy (FTIR). The effect of CNF nanofillers on PBA crystallization was reflected by the polymorphism property; and in situ FTIR was used to explore the interaction between the CNF and PBA chains. In that way, a plausible mechanism for the effect of one-dimensional carbonous nanofillers on the crystallization behavior of PBA was proposed.

Experimental

Materials and sample preparation

PBA was synthesized from 1,4-butanediol and adipic acid by a two-step reaction of esterification and polycondensation in the molten state, and details of the synthesis process were reported previously.²⁹ The number average molecular weight (M_n) and polydisperse index (PDI), determined by gel permeation chromatography (GPC, Viscotek, M302 TDA), are 5.7 × 10⁴ g mol⁻¹ and 1.82, respectively. CNF, 25–50 nm in width and 10–30 μm in length, was purchased from Beijing Daojin Technology Limited Company. MWCNT, ∼20 nm in width and 10–20 μm in length, was purchased from Beijing Cnano Company. PBA/CNF nanocomposites were prepared as follows: CNF was first suspended in dimethyl formamide (DMF) by ultrasonication (200 W) for 30 min to obtain uniform dispersion, and then a certain amount of PBA was added to the dispersion. Another 15 min of ultrasonication was performed to ensure an adequate dispersion. The mixture was air dried at 50 °C for 2 days and then vacuum dried at 70 °C for 48 hours to remove the residual DMF. The mass fractions of CNF in as-prepared nanocomposites ranged from 0.5% to 5%, and the samples were abbreviated as CNF-n, where “n” is the weight percentage of the nanofillers. In order to obtain isothermally crystallized specimens, the specimen sandwiched between two thin glass slides was placed on a hot stage at 110 °C for 5 min to eliminate thermal history. After that, the specimen was quickly dipped into an oil bath preset at required temperature for 30 min, longer enough for completion of the crystallization of PBA.

Characterizations

Differential scanning calorimetry (DSC) measurement was performed on a NETZSCH 204F1 instrument under argon atmosphere equipped with an intercooler as the cooling system. The heating and cooling rate was set as 10 °C min⁻¹. Wide angle X-ray diffraction (WAXD) analysis was carried out on a Bruker AXS D8 instrument using a graphite-filtered Cu Kα radiation target (λ_Cu = 0.154 nm). Data were collected in the 2θ interval from 16° to 28° with a scan rate of 2° min⁻¹. The transmission electron microscopy (TEM) images were obtained on a FEI Tecnai G² F20 instrument. The scanning electron microscope (SEM) images were observed on Hitachi SU8010 scanning electronic microscopy with an accelerating voltage of 30 kV. In situ Fourier transformation infrared (FTIR) spectra were collected using a Bruker Hyperion FTIR spectrometer with a resolution of 4 cm⁻¹. Two KBr plates were used as IR glass to clamp neat PBA and CNF-5 films. Each specimen was kept at 80 °C for 10 min to erase any thermal history. Then the melt was cooled to 50 °C at a rate of 2 °C min⁻¹ with simultaneous data collection.

Results and discussion

Melting and crystallization behavior

Fig. 1 shows the melt-crystallization DSC curves of PBA/CNF nanocomposites. The crystallization temperature (T_c) increases with increasing nanofillers content. The T_c value of PBA in nanocomposites was 12.9 °C higher than that of neat PBA when nonisothermally crystallized at a cooling rate of 10 °C min⁻¹ by adding 5 wt% CNF. Based on the temperature requirement for the formation of crystal form, it was clear that addition of CNF benefited the formation of α-form PBA. During the subsequent heating process, the nanocomposites showed single or double melting peaks, which could be assigned to the melting decalescence of the original and the recrystallized α-form crystals, while the neat PBA displayed a characteristic triple melting behavior. Particularly, the appearance of the highest melting peak in neat PBA (indicated by arrow) was due to the melt of the newly-formed α-form crystals from original β-form crystals through a solid–solid crystal transformation process.¹² The CNF-5 specimen showed only one melting peak instead of double melting peaks in other nanocomposites, indicating the melt-recrystallization behavior of the original α-form crystals did not occur during the heating process. The rather stability of nonisothermal crystallization α-form crystals in CNF-5 specimen should due to the rather high T_c, above 40 °C. By comparing the T_c values of PBA/CNF nanocomposites with other PBA composites containing the same mass ratio of different kinds of nucleating agents reported in literature,^24–27 it is clearly that CNF displays rather strong nucleating ability on PBA matrix.

Fig. 1 The (a) melt-crystallization and (b) subsequent heating DSC thermograms of PBA/CNF nanocomposites at a rate of 10 °C min⁻¹.

Polymorphic crystalline structure

X-ray diffraction was employed to distinguish the PBA crystal form in nanocomposites after completely isothermal crystallization at various temperatures. Fig. 2a and b show the WAXD results of CNF-5 and CNF-2 specimens. Surprisingly, PBA in nanocomposites continued to form a considerable amount of α-form crystals even the isothermal crystallization temperature was lower than 28 °C, which is the upper critical temperature for forming pure β-form crystals in neat PBA. The characteristic diffraction peaks of α-form crystals were still significant even at rather low temperature in nanocomposites, while the neat PBA could only form α-form crystals above 28 °C. In order to quantify the specific contents of the two crystal forms, Gaussian functions were employed to separate the diffraction peaks of PBA mixture crystals and fit them with an amorphous peak and several crystalline peaks so as to calculate the content of α-form crystals.^8,14 The relative fraction of α-form crystals was calculated by the following equation:

Fig. 2 X-ray diffractograms of (a) CNF-5 and (b) CNF-2 specimens isothermally crystallized at various temperatures, X-ray diffractogram of β-form of neat PBA was inserted for comparison in (b). (c) Variation of α-form crystals content in PBA/CNF nanocomposites after complete crystallization at various temperatures.

Fig. 2c shows the dependence of α-form crystal content in nanocomposites on the crystallization temperatures for different specimens. Below 32 °C, where is the beginning for β-form crystals appearance in neat PBA, the α-form crystals content gradually reduced with decreasing crystallization temperature for all specimens. The less CNF content in nanocomposites, the faster decrease of α-form crystal content. However, the α-form crystal content was still rather high even at very low crystallization temperature (e.g., 0 °C) in nanocomposites, which illustrated that CNF had great promotion on the formation ability of α-form crystals. Recently, Tang et al. had investigated the isothermal crystallization behavior of PBA/BNNSs nanocomposites.²⁸ They found that BNNSs could enlarge the temperature range for the formation of PBA mixture crystals toward lower temperature, and the α-form crystals in PBA nanocomposites containing 1 wt% BNNSs disappeared at isothermal crystallization temperature as low as 11 °C. In this study, the α-form crystal content in CNF-1 and CNF-5 specimens could remain at 9% and 44% respectively when crystallized at 8 °C. Although the nucleating abilities of CNF is similar to that of BNNSs during the nonisothermal crystallization process, the ability of enhancing the formation of α-form crystals at relative low temperature is much stronger in PBA/CNF. Zhao et al. had investigated the nonisothermal crystallization behavior of PBA/carboxyl-functionalized MWCNT nanocomposites, addition of 0.5 wt% carboxyl-functionalized MWCNT could lead to the formation of α-form PBA even at a cooling rate as high as 25 °C min⁻¹.²⁶ The reason for that phenomenon was because PBA in the nanocomposites could still crystallize in a relatively higher temperature range at that cooling rate. However, in this study the molten specimens were quickly dipped into the silicon oil bath preset at the required crystallization temperatures, which could be corresponding to a very high cooling rate process, and the possibility of crystals formation during the cooling process should be negligible. Furthermore, only β-form crystals formed when the CNF-5 melt was dipped directly into oil bath preset at −15 °C (Fig. 2a), hence formation of α-from crystals could be prevented during the cooling process. Therefore, it could be deduced that (1) the α-form crystals appeared at low isothermal crystallization temperature in PBA/CNF nanocomposites should not originate from the cooling process; (2) the mechanism for CNF enhancing formation of PBA α-form crystals might be different from BNNSs (a type of two-dimensional nanofiller).

Another novel phenomenon shown in Fig. 2c is that the amount of α-form crystals decreased with decreasing crystallization temperature until 12 °C for PBA/CNF nanocomposites; then the α-form crystal content seemed to reach a plateau value. This phenomenon has not been reported before. The plateau content of α-form crystals increased progressively with the increasing nanofillers content. While the α-form crystal content continued to decrease as the crystallization temperature further decreased, and eventually disappeared at −15 °C. The advent of the plateau phenomenon very likely means that a certain fraction of PBA chains are confined to form the α-form crystals at a certain crystallization temperature range. When the crystallization temperature locates in the special temperature range, only the confined faction of PBA retains to crystallize into α-form crystals, and other free fraction chains crystallize into β-form crystals as in neat PBA. The more CNF was incorporated, the larger amount of confined PBA chains, leading to the higher content of α-form crystals obtained under low temperature. Hence, the fraction of PBA chains surrounding CNF is likely to interact with CNF, which causes such special feature. Fig. 3 plots the relationship between the plateau values and CNF loading percentages for the different specimens. Interesting to notice that the plateau value is approximately linear proportion to the nanofillers content in PBA/CNF nanocomposites. The result confirmed that the content of α-form crystals formed at such low temperature was indeed associated with the surface area of CNF, and the dispersion situation of CNF in PBA matrix should be almost the same for all PBA nanocomposites. In Fig. 4, the cryo-fractured section surface of CNF-5 specimen reveals that the CNF was well separated and dispersed in PBA matrix. Generally, the pristine CNF tends to aggregate together into bundles under the mutual van der Waals interaction. To avoid the aggregation, a common method is to adsorb molecules onto the nanofibers surface by covalent or non-covalent interaction. Here, the one-dimensional CNF was well dispersed in the DMF solvent by ultrasonication first, then the polymer was added and potentially wrapped (or coated) around the one-dimensional nanofiller during the solution mixing process which led to the good dispersion of CNF in nanocomposites. Many literature had presented the wrapping (or oating) behavior of polymer chains on nanotubes,^30–34 and the interaction between PBA and CNF will be further studied below.

Fig. 3 The relationship between the plateau values and CNF percentages for PBA/CNF nanocomposites.

Fig. 4 SEM image of the cryo-fractured section of CNF-5 specimen.

Interaction between PBA and CNF

Fig. 5a and b clearly show that the surface of CNF was encapsuled by a layer of PBA chains, which illustrated that PBA chains were attracted to the CNF surface and surround them. Flexible polymer chains tend to wrap along the one-dimensional nanotubes and have been observed in many literature.^32,33,35 In some cases, the polymer chains contain conjugated structure (e.g., benzene rings) or other special functional groups, which can bind CNT with π–π interaction or other type of strong interaction. However, some polymers, such as polyethylene and nylon 66 would form “nano hybrid shish-kebab” (NHSK) around the CNT by crystallize in solution, although there are not special functional groups that could strongly interact with CNT.³⁶ The formation mechanism of NHSK was demonstrated as “soft epitaxy”, in which strict lattice matching between the CNT surface and the polymer crystal was not required. Geometry (radius in nanometer scale) of CNT and the high flexibility of polymer chains in solution should provide strong driving force for polymer chains aligning on CNT surface before crystallization. That meant polymer chains could adsorb on the CNT surface even there was not strong interaction force between them. Since PBA chains are rather flexible and CNF shares similar surface structure and long slenderness ratio with pristine CNT, further the nanocomposites preparing procedure included an ultrasonication process in solution. Ultrasonication in solution is an effect method to prepared polymer wrapped CNT.³⁷ Therefore, it is reasonable to assume that some PBA chains wrapped (or coated) around the CNF during the solution mixing process, the well dispersion of CNF in PBA matrix also supports this view. Fig. 5c is the SEM image of original CNF, and Fig. 5d is the morphology of CNF-5 specimen isothermal crystallization at 35 °C following by etching in 36% w/v hydrochloric acid for 10 min. The PBA was removed during the etching process,¹⁵ while the residual CNF showed rougher surface than the original CNF and the diameter ranged in 40–60 nm. The thicker diameter and rougher surface of CNF in nanocomposites after etching than original CNF is probably caused by the strong absorption behavior of PBA on the surface, which was in accordance with the TEM image, Fig. 5b.

Fig. 5 TEM images of (a) CNF-5, (b) enlarge picture in different areas, and SEM images of (c) original CNF and (d) CNF-5 isothermal crystallization at 35 °C for 1 h, followed by 36% w/v hydrochloric acid etching for 10 min.

The wrapping (or coating) behavior of PBA chains on nanofillers can be used to explain the nucleating behavior, while the reason why CNF exclusively significantly enhance α-form crystals content in the PBA nanocomposites at low crystallization temperature still needs further research. For PBA, α-form crystal is thermodynamic advantage but kinetic disadvantage compared with β-form crystal, the appearance of rather high proportion of α-form crystals when the nanocomposites was crystallized at low temperature could be possibly due to the situation that some PBA chains were preordered into special conformation which benefited crystallization into α-form crystal rather than β-form crystal in the molten state of nanocomposites. In that way, formation of α-form crystal kept one step ahead of β-form crystal in the initial stage of crystallization process which counteracted its kinetic disadvantage in the region with preordering chain conformation. In order to demonstrate whether there exist preordering chain conformation and what kind of interaction between CNF and PBA chains, in situ FTIR was employed to investigate the differences between neat PBA and CNF-5 specimen during the cooling process from 80 to 55 °C. If any special preordering conformation exists, it would show up much clearer before crystallization in the FTIR spectra. The temperature range was chosen to ensure that the samples were maintained in their molten state. Fig. 6a shows the evolution of FTIR spectra of neat PBA in the wavenumber range of 3100–2700 cm⁻¹, the positions and intensities of all absorption peaks were basically unchanged during the temperature variation process. Two main absorption peaks detected in the wavenumber range are 2956 and 2873 cm⁻¹, which are both assigned to the C–H stretching modes. While as shown in Fig. 6b, beside the 2956 and 2873 cm⁻¹ peaks, there are two more splitting peaks toward lower wavenumber direction appeared at 2930 and 2857 cm⁻¹, respectively. The most plausible appearance of those new peaks could be attributed to the occurrence of interaction between CNF and PBA chains. The red-shift behavior of C–H stretching vibration modes indicated special confinement effect arose for some C–H bonds on PBA chains in molten state, analogous but more common confinement-induced red-shift cases can be found in carbonyl groups on polyester chains after melt-crystallization³⁸ or forming hydrogen-bonds with other substance.³⁹ However, the absorption peaks of carbonyl groups in neat PBA and CNF-5 specimen were the same and kept constant during the cooling process. Thus, some PBA chains should be absorbed to the CNF surface by forming CH–π interaction in nanocomposites. The CH–π interaction has been detected in many types of polymer composites, which is a kind of weak hydrogen-bond-like attractive force.^35,40,41 Although the strength of CH–π interaction is only one-tenth of the hydrogen bond, their cooperative multiple interaction significantly influence many physical and chemical phenomena.⁴² For example, a fine dispersion of carbon black within a matrix rubber is due to the present of many CH–π interactions. Therefore, it was reasoned that PBA chains containing many CH linkages would sufficiently interact with CNF under the preparation condition of nanocomposites. PBA plays as CH donor and sp² sidewall of CNF plays as conjugated structure donor. CH–π interaction is the key for PBA wrapping (or coating) the surface of CNF, also this interaction should be the incentive for the possible preordering chain conformation.

Fig. 6 Time-resolution FTIR spectra in the range of 3100–2700 cm⁻¹ for (a) neat PBA and (b) CNF-5 specimen during the cooling process from 80 to 55 °C at a rate of 2 °C min⁻¹.

Fig. 7a and b show the FTIR spectra of CH₂ wagging and bending vibration modes ranged in 1500–1380 cm⁻¹ for neat PBA and CNF-5 specimen during the cooling process from 80 to 55 °C, respectively. The absorption peaks at 1392, 1421 and 1460 cm⁻¹ are assigned to the CH₂ wagging mode, CH₂ bending mode and CH₂ asymmetric wagging mode in amorphous phase, respectively. While there was no significant difference between these two sets of spectra, the change of CNF-5 specimen exhibited somewhat apparently varying dependence on temperature compared with neat PBA. In order to make the variation more explicit, difference spectra of neat PBA and CNF-5 specimen were calculated through subtracting all spectra by the initial spectrum collected at 80 °C and presented in Fig. 7c and d, respectively. Two obviously peaks in difference spectra located at 1419 and 1396 cm⁻¹ appeared and their intensities increased during the cooling process, the evolution of FTIR spectra for molten amorphous specimens during cooling might indicate the formation of some precursory structures. Several examples for precursory ordering process prior to crystallization had been proved in various polymer melts.^43–45 Compared to neat PBA, the difference FTIR spectra of CNF-5 specimen displayed three distinct characteristics: (1) displaying faster evolution rate; (2) showing larger increased degree at peak 1419 cm⁻¹ (while neat PBA showed larger increased degree at peak 1396 cm⁻¹); (3) exhibiting a new peak at 1471 cm⁻¹ in the difference spectra. The peak at 1419 cm⁻¹ have been assigned to the CH₂ stretching vibration mode in PBA α-form crystals, and the peak at 1471 cm⁻¹ is close to PBA crystalline FTIR peak.⁴⁶ The significant increase of 1471 cm⁻¹ with decreasing the temperature indicated the obvious formation of precursory structures. These experiments had been repeated for several times, and the above mentioned phenomena were always observed. Therefore, it is reasoned that CNF in nanocomposites formed apparent CH–π interaction with PBA chains which helped to promote the precursory ordering process. Particularly, the precursory structure in CNF-5 molten state was likely to adopt dominant fraction of chains conformation similar as it in α-form crystal.

Fig. 7 Time-resolution FTIR spectra in the range of 1500–1380 cm⁻¹ for (a) neat PBA and (b) CNF-5 specimen during the cooling process from 80 to 55 °C at a rate of 2 °C min⁻¹. Difference spectra of (c) neat PBA and (d) CNF-5 specimen were obtained by subtracting the initial spectra collected at 80 °C. The arrow “↑” indicates the temperature cooling direction.

In order to verify whether the novel plateau content phenomenon of α-form crystal can also exist in other PBA/carbonous nanofillers composites, PBA/MWCNT nanocomposites with different MWCNT contents were prepared through the same procedure. Fig. 8a shows the α-form crystals contents in different PBA/MWCNT specimens after isothermal crystallization at various temperatures. Obviously, the α-form crystal content plateau phenomenon indeed existed again, and the start temperature of plateau was around 16 °C, which was slightly higher than PBA/CNF composites. Fig. 8b plots the relationship between the plateau values and MWCNT contents of PBA/MWCNT nanocomposites. The plateau value is also approximate in linear proportion to the MWCNT content but with somewhat larger slope value. The curvature radius of MWCNT (∼20 nm) was smaller than that of CNF (25–50 nm), so the specific surface area of MWCNT was larger than that of CNF due to their similar length, probably resulting in a higher amount of interacted PBA chains in PBA/MWCNT than PBA/CNF with the same weight percentage loading of nanofillers, and a somewhat larger fitting slope value in PBA/MWCNT. The result indicates that the appearance of PBA α-form crystal content plateau in PBA/CNF nanocomposites is not a single event. It is likely to appear in various PBA/one-dimensional carbonous nanofillers composites, the α-form crystal content plateau and the linear fitting slope value could probably be used as quantitative criteria to evaluate the effect of nanofillers on the polymorphism of PBA.

Fig. 8 (a) The α-form crystals content in PBA/MWCNT nanocomposites after complete crystallization at various temperatures. “MWCNT-n” means the content of MWCNT content in composites is n%. (b) The relationship between the plateau values and MWCNT percentages for PBA/MWCNT nanocomposites.

Then we can deduce the mechanism for formation of α-form crystal at low crystallization temperature and the existence of plateau content phenomenon in PBA/CNF nanocomposites as follows. When being introduced into PBA matrix, the CNF with large aspect ratio provides a large amount of sites on the intrinsic sp² sidewall for forming nonspecific CH–π interaction with PBA chains during the solution mixing and further processes, which induces the restrained PBA chains to wrap (or coat) on the surface of nanofillers with the assist of entropically driven advantage for their flexible character. In this way, PBA chains within as-prepared nanocomposites in molten state could be divided into two distinct parts, restrained chains surrounding nanofillers and free chains. The former part is forced to pre-establish some certain conformation under the effect of CH–π interaction and geometry of nanofillers, plausibly some gauche structures that are close to the conformers in α-form crystal as revealing in FTIR spectra. The similar conformation on some positions of the PBA chains between restrained chains part and α-form crystal lowers down the energy barrier for melt crystallization, resulting in the significant α-form crystal nucleating effect of CNF on PBA matrix. Also, the kinetic advantage for formation of β-form crystal at low temperature range is counteracted and inhibited in the restrained region. As to the free chains part, the crystallization behavior is similar to neat PBA. Therefore, a schematic diagram in Fig. 9 is used to illustrate the effect of nanofillers on the melt-crystallization behavior of PBA matrix at different crystallization temperature ranges. The restrained PBA chains surrounding nanofillers form special regions, where the chains tend to crystallize into α-form crystals due to the pre-established conformation during cooling or isothermal crystallization. When the isothermal crystallization is carried out higher or equal to 32 °C, both restrained and free chains crystallize into α-form crystals; the restrained chains crystallize earlier and play as nucleating sites for adjacent free chains. When the crystallization temperature is between 32 °C and plateau temperature, β-form crystals appear due to the crystallization of free chains. However, PBA chains in restrained regions still crystallize into α-form crystals, then serve as α nucleating sites for adjacent free chains, so the α-form crystals contents in nanocomposites are higher than neat PBA and generally increase with nanofillers content. With decreasing crystallization temperature, the advantage for the formation of β-form crystals in free chains regions becomes more remarkable and the spread out of α-form crystals from the restrained region is gradually inhibited. Once the crystallization temperature is low enough to reach the plateau temperature, such as 12 °C for PBA/CNF and 16 °C for PBA/MWCNT, the α-form crystals can only form within the restrained regions, then α-form crystals content attains a limited value and displays a plateau value in the special crystallization temperatures range; and the plateau value obviously only depends on the nanofillers content. When the crystallization temperature is further decreased, the advantage for formation of α-form crystals in restrained regions could not compete with the formation of β-form crystals at such large supercooling degree, thus the α-form crystals content further decreases.

Fig. 9 Illustration for the effect of CNF on the PBA melt-crystallization behavior at various crystallization temperature ranges. “d”,“d₁”,“d₂” and “d₃” represent the diameter of the α-form crystal region around the CNF in various temperature ranges, respectively.

Conclusions

In this study, CNF showed strong α-form crystal nucleating ability on PBA matrix and induced a novel polymorphism behavior of PBA. The formation of α-form crystal in PBA/CNF nanocomposites could reach as low as 0 °C, and a novel α-form crystal content plateau phenomenon was discovered. Similar phenomenon was also found in PBA/MWCNT nanocomposites. Morphology observation and in situ FTIR study confirmed the existence of restrained regions surrounding CNF through the occurrence of specific CH–π interaction between PBA chains and CNF surface, the restriction promoted formation of precursor structure containing conformation close to α-form crystal. Therefore, CNF exclusively significantly enhanced the α-form crystals content in the PBA nanocomposites, even at low crystallization temperature; and induced the appearance of novel content plateau phenomenon of α-form crystals. The unique change of polymorphism behavior of PBA in its nanocomposites with CNF and MWCNT provides a more intuitive method to evaluate the interaction between polymer chains and nanofillers.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 21304108) and the Science Foundation of China University of Petroleum – Beijing (No. YJRC-2013–14, 2462013BJRC001).

Notes and references

1. M. H. Al-Saleh and U. Sundararaj, Composites, Part A, 2011, 42, 2126.
2. M. T. Byrne and Y. K. Gun'ko, Adv. Mater., 2010, 22, 1672.
3. C. L. Huang and C. Wang, Eur. Polym. J., 2011, 47, 2087.
4. K. Ke, P. Pötschke, D. Jehnichen, D. Fischer and B. Voit, Polymer, 2014, 55, 611.
5. L. L. Sun, B. Li, Y. Zhao and W. H. Zhong, Polymer, 2010, 51, 3230.
6. Z. Liang, J. Yang, L. Hua, P. Pan, J. Huang, J. Zhang, H. Abe and Y. Inoue, J. Appl. Polym. Sci., 2014, 131, 39600.
7. Y. A. Chen, Y. T. Hang and T. M. Wu, J. Appl. Polym. Sci., 2015, 132, 42526.
8. J. Yang, Y. Chen, S. Qin, J. Liu, C. Bi, R. Liang, T. Dong and X. Feng, Ind. Eng. Chem. Res., 2015, 54, 8048.
9. E. Pouget, A. Almontassir, M. T. Casas and J. Puiggalí, Macromolecules, 2003, 36, 698.
10. T. Iwata, S. Kobayashi, K. Tabata, N. Yonezawa and Y. Doi, Macromol. Biosci., 2004, 4, 296.
11. K. Noguchi, H. Kondo, Y. Ichikawa, K. Okuyama and J. Washiyama, Polymer, 2005, 46, 10823.
12. Z. Gan, K. Kuwabara, H. Abe, T. Iwata and Y. Doi, Biomacromolecules, 2004, 5, 371.
13. E. M. Woo and M. C. Wu, J. Polym. Sci., Part B: Polym. Phys., 2005, 43, 1662.
14. L. Zhao, X. Wang, L. Li and Z. Gan, Polymer, 2007, 48, 6152.
15. J. Liu, H. M. Ye, J. Xu and B. H. Guo, Polymer, 2011, 52, 4619.
16. Z. Gan, K. Kuwabara, H. Abe, T. Iwata and Y. Doi, Polym. Degrad. Stab., 2005, 87, 191.
17. Y. Sun, H. Li, Y. Huang, E. Chen, L. Zhao, Z. Gan and S. Yan, Macromolecules, 2005, 38, 2739.
18. L. Zhao and Z. Gan, Polym. Degrad. Stab., 2006, 91, 2429.
19. Z. Liang, P. Pan, B. Zhu and Y. Inoue, Macromolecules, 2010, 43, 6429.
20. J. Yang, P. Pan, L. Hua, Y. Xie, T. Dong, B. Zhu, Y. Inoue and X. Feng, Polymer, 2011, 52, 3460.
21. J. Yang, P. Pan, L. Hua, X. Feng, J. Yue, Y. Ge and Y. Inoue, J. Chem. Phys. B, 2012, 116, 1265.
22. T. Dong, W. Kai and Y. Inoue, Macromolecules, 2007, 40, 8285.
23. A. Prasannan, T. L. Bich-Tram, D. Y. Hsu, P. D. Hong and G. R. Pan, CrystEngComm, 2013, 25, 5119.
24. W. Kai, B. Zhu, Y. He and Y. Inoue, J. Polym. Sci., Part B: Polym. Phys., 2005, 43, 2340.
25. M. Weng, Y. He and Z. Qiu, Ind. Eng. Chem. Res., 2012, 51, 13862.
26. Y. Zhao and Z. Qiu, CrystEngComm, 2011, 13, 7129.
27. N. Jiang, L. Zhao and Z. Gan, Polym. Degrad. Stab., 2010, 95, 1045.
28. Y. R. Tang, J. Xu and B. H. Guo, Ind. Eng. Chem. Res., 2015, 54, 1832.
29. H. M. Ye, R. D. Wang, J. Liu, J. Xu and B. H. Guo, Macromolecules, 2012, 45, 5667.
30. Z. Qi, H. Ye, J. Xu, J. Chen and B. Guo, Colloids Surf., A, 2013, 421, 109.
31. C. Caddeo, C. Melis, L. Colombo and A. Mattoni, J. Phys. Chem. C, 2010, 114, 21109.
32. A. Liu, I. Honma, M. Ichihara and H. Zhou, Nanotechnology, 2006, 17, 2845.
33. A. A. B. Davijani and S. Kumar, Polymer, 2015, 70, 278.
34. Y. K. Kang, O. S. Lee, P. Deria, S. H. Kim, T. H. Park, D. A. Bonnell, J. G. Saven and M. J. Therien, Nano Lett., 2009, 9, 1414.
35. D. Baskaran, J. W. Mays and M. S. Bratcher, Chem. Mater., 2005, 17, 3389.
36. L. Li, C. Y. Li and C. Ni, J. Am. Chem. Soc., 2006, 128, 1692.
37. A. E. Hsiao, S. Y. Tsai, M. W. Hsu and S. J. Chang, Nanoscale Res. Lett., 2012, 7, 240.
38. O. P. Valmikanathan, O. Ostroverkhova, I. S. Mulla, K. Vijayamohanan and S. V. Atre, Polymer, 2008, 49, 3413.
39. X. Sun, F. Pi, J. Zhang, I. Takahashi, F. Wang, S. K. Yan and Y. Ozaki, J. Phys. Chem. B, 2011, 115, 1950.
40. Z. Tang, D. Sun, D. Yang, B. Guo, L. Zhang and D. Jia, Compos. Sci. Technol., 2013, 75, 15.
41. X. Hu, H. An, Z. M. Li, Y. Geng, L. Li and C. Yang, Macromolecules, 2009, 42, 3215.
42. M. Nishio, M. Hirota and Y. Umezawa, CH/pi interaction, evidence, nature and consequences, Wiley-VCH, New York, 1998.
43. K. Kaji, K. Nishida, T. Kanaya, G. Matsuba, T. Konishi and M. Imai, Adv. Polym. Sci., 2005, 191, 187.
44. C. H. Su, U. Jeng, S. H. Chen, S. J. Lin, Y. T. Ou, W. T. Chuang and A. C. Su, Macromolecules, 2008, 41, 7630.
45. J. F. Mano, Y. Wang, J. C. Viana, Z. Denchev and M. J. Oliveira, Macromol. Mater. Eng., 2004, 289, 910.
46. J. Yang, Z. Li, P. Pan, B. Zhu, T. Dong and Y. Inoue, J. Polym. Sci., Part B: Polym. Phys., 2009, 47, 1997.

---