Unusual hierarchical distribution of β-crystals and improved mechanical properties of injection-molded bars of isotactic polypropylene

Abstract

In this study, the microstructures, hierarchical distribution of β-phase crystalline morphology and mechanical properties of neat isotactic polypropylene (iPP) prepared by conventional injection molding (CIM) and gas-assisted injection molding (GAIM) were intensively examined. The obtained samples were characterized via two-dimensional small-angle X-ray scattering (2D SAXS), polarizing light microscopy (PLM), scanning electron microscopy (SEM), differential scanning calorimetry (DSC), wide-angle X-ray diffraction (XRD) and tensile testing. It was found that the strong shear effect introduced during the gas penetration of the GAIM process greatly influences the morphology and the formation of β-crystals in the pure PP specimens. Shish-kebab and β-cylindrite morphologies were observed simultaneously in the sub-skin layer of the GAIM part for the first time, where the flow pattern was complex and the shear strength was believed to be the maximum. Furthermore, an iPP specimen with β-crystals existing in the entire cross section of the molded bar were obtained for the GAIM sample without adding β-nucleating agents or other components. However, the pure iPP sample with a low content of β-crystals were only found in the skin layer of the CIM part. In addition, due the coexistence of shish-kebab and β-cylindrite structures, the mechanical properties of the GAIM specimen were significantly improved when compared with the CIM sample. Based on these experimental observations, a schematic illustration was proposed to interpret the mechanism of the formation of the unusual hierarchical distribution of β-crystals during the GAIM process.

---

Introduction

Semi-crystalline polymers are known to crystallize in two main stages; the formation of nuclei and the subsequent growth of crystals. It is well-known that the polymer melt, under the shear field, exhibits an increased rate of crystallization or a different morphology when compared with those of a quiescent melt.¹ Flow-induced crystallization (FIC) of semi-crystalline polymers, which is very common in industrial processing, has attracted increasing attention.^2–5 On one hand, FIC attracts considerable attention for its intrinsic theoretical interest because it is a typical example of a non-equilibrium, externally controlled phase transition process. On the other hand, FIC is also practically significant because the processing of semi-crystalline polymers constitutes the largest proportion of commercially useful polymeric materials. During most industrial processing operations (e.g., extrusion, molding, film blowing, fiber spinning, etc.), the molten polymer is exposed to complex and intense flow fields, and the kinetics, semi-crystalline morphology and material properties are considerably influenced by the external flow field.^6–8

Gas-assisted injection molding (GAIM), an innovative technology for producing plastic parts, has received wide attention in modern industry during the past decade due to its flexibility in designing and manufacturing of plastic parts.⁹ As presented in Fig. 1a, a pre-determined amount of polymer melt is injected to partially fill the mold cavity. Then, a high-pressure gas is subsequently introduced through the runner or nozzle after a short gas delay time (Fig. 1b). Finally, the gas pressure remains unchanged until all the polymer material solidifies after the completion of the melt filling stage (Fig. 1c). It should be noted that during the CIM process, the polymer melt is confined by the mold wall, which is a rigid and cold medium. However, during the gas penetration stage of the GAIM process, the polymer melt is confined not only by the mold wall but also by the compressed gas, and is subjected to severe instantaneous flow fields during the gas penetration stage, which is much more complicated than the CIM process. Therefore, many fascinating supermolecular structures (e.g. transcrystallinity, oriented crystals, spherulites with ring bands, shish kebab, etc.) were formed in the GAIM parts of high-density polyethylene, polycarbonate/polyethylene blends, glass fiber reinforced polyamide-6, etc.^10–15

Fig. 1 Schematic diagram for the GAIM process: (a) partial melt filling; (b) gas-assisted filling; and (c) gas-assisted packing.

Isotactic polypropylene (iPP) is one of the most commonly used polymeric materials due to its versatility, relatively good mechanical properties, easy processability and recyclability, and a favorable price-to-performance ratio.¹⁶ The supermolecular structure and final properties of iPP products are essentially influenced by crystallization conditions, in particular, iPP polymorphism. There are three basic crystalline forms of iPP (α, β, γ).¹⁷ It has been reported that among all the crystalline forms of iPP, β-crystals demonstrate different performance characteristics such as improved elongation at break and impact strength. β-Crystals can be obtained using special techniques such as the temperature gradient method,¹⁸ adding specific nucleating agents^19–21 or flow-induced crystallization.^22–24

In the injection molding process, polymers crystallize from a melt, which has been exposed to complex flow and temperature variations, and the CIM part is characterized by an intrinsic heterogeneous microstructure, featuring a gradual and hierarchical variation of morphology evolving throughout the part. Much research concerning the microstructure of injection-molded PP bars have been carried out in recent years with considerable attention being paid to the formation of different crystal forms, such as the α- and β-phases, and their relative content in different zones along the cross section.^25–27 However, for pure PP, the controversy regarding the relative content of the β-phase in different zones has been an ongoing debate for a long time.^28–32 Furthermore, only a few studies have been conducted on the microstructure of pure iPP molded by GAIM, and studies focusing on the formation of different crystal forms of the GAIM parts have been reported yet.

In this research, the hierarchical distribution of the β-phase crystalline morphology and the mechanical properties of neat iPP specimens molded by GAIM were investigated. For comparison, the crystalline morphology and crystal forms of its CIM counterpart were also examined. To observe the microstructure and crystalline forms of iPP, the samples molded by different processing methods were characterized using synchrotron two-dimensional small-angle X-ray scattering (2D SAXS), polarizing light microscopy (PLM) and scanning electron microscopy (SEM). The crystalline forms and their distributions were characterized via differential scanning calorimetry (DSC) and wide-angle X-ray diffraction (WAXD). Moreover, the mechanical properties of the molded bars were characterized by their tensile test. From a practical viewpoint, attempting to explore the microstructures and crystalline morphology of iPP during the GAIM process, which have been widely used in practical production, is of great significance.

Experimental

Materials

The iPP used was commercially available from the ExxonMobil Chemical Company Houston, Texas, USA with a melt flow index of 0.9 g/10 min (230 °C, 2.16 kg). The weight-average molecular weight (M_w) was 690 kg mol⁻¹.

Sample preparation

A PS40E5ASI injection molding machine and a gas pressure system (Model: MPC-01) supplied by ZhongTuo (Beijing) Co, P.R. China were utilized for the molding process. To provide an even route for gas penetration, a mold with a circular cross-section cavity was employed. It should be noted that a sub-cavity connected to the main cavity by a narrow flow channel was designed as shown in Fig. 1, which was different from our previously used mold.³³ With this new design, an even and thorough penetration of the high-pressure gas was guaranteed. The processing parameters for the GAIM process are listed in Table 1. For comparison, the conventional injection molding (CIM) process was also carried out using the same processing parameters but without the gas penetration and gas-assisted packing process. It should be noted that the melt injection volume for CIM is as large as 100 vol.% and the main body of CIM part exhibits a solid structure, while the GAIM part has a hollowed structure.

Table 1 Processing variables used in the GAIM experiment

Processing variables	Values
Gas pressure (MPa)	9.5
Gas delay time (s)	1.0
Gas packing time (s)	15
Short shot size (vol.%)	70
Melt temperature (°C)	210
Mold temperature (°C)	15
Injection pressure (MPa)	80

---

Synchrotron two-dimensional small-angle X-ray scattering (2D SAXS) experiments

The synchrotron 2D SAXS experiments were conducted on the BL16B1 beamline in the Shanghai Synchrotron Radiation Facility (SSRF), Shanghai, China. The wavelength used was 0.124 nm, and the rectangle-shaped beam had dimensions of 0.2 × 2 mm². The backgrounds of all the 2D SAXS patterns were extracted. Each 2D SAXS pattern were obtained in different zones along the thickness direction of the molded bars, and were collected within 40 s at a sample-to-detector distance of 1950 mm. At this distance the effective scattering vector q (q = 4π sin θ/λ, where 2θ is the scattering angle and λ the wavelength) range is 0.17–2.08 nm⁻¹. The samples were placed with their orientation (flow direction) perpendicular to the projection beams, as shown in Fig. 2. Fit2D software from the European Synchrotron Radiation Facility was used to analyze the SAXS data, and the 2D SAXS patterns were integrated azimuthally to obtain the one-dimensional (1D) scattering profile as a function of azimuthal angle. For 1D SAXS data, the long period, L = 2π/q, was obtained.^34,35

Fig. 2 Schematic of the sample preparation procedure for SAXS: FD, the flow direction; TD, the transverse direction; ND, the direction normal to the MD-TD plane.

Polarizing light microscopy

The molded bars were sectioned using a KD-3558 rotary microtome provided by Zhejiang Jinhua Kedi Strumntal Equipment Co., Ltd, China. Thin sections of about 10 microns in thickness were cut from the midpoint of the bars parallel to the flow direction. These microtomed specimens were then sandwiched between a glass slide and a coverslip, and then observed using a Leica DMIP polarizing light microscopy (PLM) equipped with a Canon PowerShot 550 digital camera (DC).

Scanning electron microscopy

The selective permanganic etching approach proposed by Olley and Bassett³⁶ was employed to prepare the specimens for morphological observations under a scanning electron microscope (SEM, Model: JSM-5900LV, JOEL Co., Ltd., Japan). Before the etching procedure, the specimen was cut in the middle of the bar along the melt flow direction; these segments were then immersed in liquid nitrogen for about 30 min before they were cryogenically fractured. After etching, the surfaces of the specimens were covered with a thin layer of gold, and the crystalline morphology in different zones of these samples was observed with an accelerating voltage of 20 kV.

Differential scanning calorimetry

Thermal analysis of the samples, which were cut at various distances from the surface, was conducted using differential scanning calorimetry (DSC, TA-Q20). All the measurements were conducted under a nitrogen atmosphere. A sample of about 5 mg was used and the heating rate was set at 10 °C min⁻¹ from 25 to 160 °C. The percentage of β-crystals of iPP (ϕ_β) was determined by the relative crystallinity of the α- and β-crystals according to eqn (1),where X_α and X_β are the crystallinity of the α- and β-crystals, respectively, based on the measured heat of fusion.

Because of the coexistence of α- and β-crystals in the molded bars, the crystallinity of each of the crystal forms could be calculated separately according to the following equation,

where ΔH_i is the calibrated specific fusion heat of either the α- or β-crystals and ΔH⁰_i is the standard fusion heat of either pure α- or β-crystals of iPP. The values of ΔH⁰_α for α-crystals and ΔH⁰_β for β-crystals were 177.0 and 168.5 J g⁻¹, respectively.³⁷

The DSC curves of the samples exhibited both an α-fusion peak and a β-fusion peak. The specific fusion heats for the α- and β-crystals were approximate according to the following calibration method. The total fusion heat, ΔH, can be obtained by integrating from 100 °C to 180 °C on a DSC thermogram. A vertical line was drawn through the minimum between the α- and β-fusion peaks and the total fusion heat was divided into the β-component, ΔH^*_β, and the α-component, ΔH^*_α. The less perfect α-crystals were melted before the maximum point during heating, and thus they contributed to ΔH^*_β, and consequently, the true value of β-fusion heat, ΔH_β, could be approximated by multiplying ΔH^*_β with a calibration factor A,³⁸

ΔH_α = ΔH − ΔH_β (4)

ΔH_β = A × ΔH^*_β (5)

In eqn (3), h₁ and h₂ are the heights from the base line to the β-fusion peak and the minimum point, respectively.

Wide-angle X-ray diffraction

To investigate the hierarchical crystalline structures and distribution of the β-crystals of the molded bars, wide-angle X-ray diffraction (WAXD) measurements were carried out using a DX-1000 X-ray diffractometer (Dandong Fanyuan Company, China) with curved graphite crystal filtered Cu K radiation source (λ = 0.154 nm) at 50 kV and 30 mA at 25 °C. The samples for WAXD characterization were cut from the middle section of the injection molded samples.

The relative content of the β-crystals (K_β) was determined from the WAXD profiles according to the following relation,

K_β = H_β(300)/[H_α(110) + H_α(040) + H_α(130) + H_β(300)] (6)

where H_α(110), H_α(040), H_α(130) are the intensity of the (110), (040) and (130) reflections of the α-crystals in PP, appearing at 2θ around 14°, 16.8° and 18.4°, respectively; H_β(300) is the intensity of the (300) reflection of the β-crystals at 2θ around 16.0°.³⁹

Tensile testing

For the hollow structures of the GAIM samples, the standard ASTM sample for tensile test cannot be applied directly. Specimens with a modified shape, as shown in Fig. 3, were prepared for the tensile test. The tensile tests were conducted on an Instron universal testing machine (Model 5567) with a crosshead speed of 50 mm min⁻¹ at room temperature (∼25 °C). The specimen broke around the center during the test, and the average values were obtained from over five specimens for each measuring condition.

Fig. 3 Shape and dimensions of the GAIM molded bar.

Results and discussion

Shear-induced orientation

Fig. 4 shows the selected 2D SAXS patterns at different zones of two samples. The flow direction was vertical, and can be considered the same as the fiber axis when assuming that the deformed scatterers have a fiber symmetry.⁴⁰ It is widely accepted that the appearance of an equatorial streak in the 2D SAXS patterns can be attributed to the shish-like crystalline structure, which are oriented parallel to the flow direction, while the appearance of meridional maxima can be attributed to kebab-like lamellar stacks, which are oriented perpendicularly to the flow direction.⁴¹ As shown in Fig. 4a, the 2D SAXS pattern of the skin layer of GAIM sample clearly exhibited the appearance of meridional maxima from the oriented scatterers, indicating that the packages or stacks of crystal lamellae (kebab) have already formed in the skin layer. In addition, a relatively weak equatorial streak existed in the skin layer, which verified the existence of a shish structure parallel to the flow direction. This weak signal was mainly attributed to the small size or low volume fraction of shish, which was consistent with our previously reported results of low shear field in the skin layer.^10–12 As depicted in Fig. 4b, an equatorial streak in the 2D SAXS pattern was clearly observed at the sub-skin layer of the GAIM sample in addition to the meridional maxima, indicating that a typical shish-kebab morphology was formed in the sub-skin layer of the GAIM specimen.² However, the preferentially aligned lamellar structure existed only in the skin and the sub-skin zones of the GAIM sample. As illustrated in Fig. 4c, the gas-channel zone of the GAIM sample was characterized by an isotropic scattering circle, indicating the absence of any preferred orientation or a very low degree of orientation within this region. Under the mutual interaction of the shear gradient introduced during the high pressure gas penetration process and the temperature gradient across the thickness direction, hierarchical oriented structures existed from the skin to the gas-channel layer of the GAIM part. For comparison, the 2D SAXS patterns of the CIM sample at different zones were also investigated, and are illustrated in Fig. 4. It can be observed that the dominant patterns of the CIM sample (Fig. 4d and e) from the skin to the core zone were mainly characterized by isotropic rings due to the random orientation of the chain.

Fig. 4 2D SAXS patterns at different zones of the GAIM part: (a) skin layer; (b) sub-skin layer; (c) gas channel layer and CIM part, (d) skin zone; (e) core zone. Melt flow direction is parallel to the meridian direction.

The variables of the crystalline structures can be quantified in the azimuthal scans of the SAXS patterns at the different zones of the two samples (Fig. 5). As shown in Fig. 5a, the skin layer of the GAIM sample was characterized by clear meridional scattering and a weak equatorial scattering. For the sub-skin layer of the GAIM sample, Fig. 5b clearly displays both the strong meridional and clear equatorial scattering, indicating that a typical shish-kebab morphology was formed. However, the gas channel zone only exhibited a weak and broad meridional scattering without any signal of equatorial scattering, implying the formation of a less oriented arrangement of iPP lamellae. Therefore, the skin and sub-skin layers of the GAIM specimen contained iPP lamellae (periodicities) oriented along and perpendicular to the fiber axis. In contrast, for the CIM sample (Fig. 5b), apart from no trace of equatorial scattering, the meridional scattering was very broad in the skin and core zones, which validated the absence of a shish-kebab or oriented structure in the CIM specimen.

Fig. 5 Azimuthal distribution of the scattered intensity in the 2D SAXS images of skin, sub-skin and gas channel layers of the (G1, G2, G3) GAIM samples and skin and core layers of the (C1, C2) CIM samples.

According to the ref. 42 and 43, the lamellar crystal thickness and interlamellar amorphous layer thickness can be obtained from the one-dimensional correlation function of the SAXS patterns but because of the difficulty and complexity in calculating the one-dimensional correlation function, we compared the contributions of the isotropic and oriented scattering, which may be more important for oriented samples used in this study. To quantify the contributions of the isotropic and oriented scattering, the total scattered intensity of the 2D SAXS patterns were separated into two contributions; isotropic and oriented. The oriented component was calculated by subtracting the azimuthally independent component from the total SAXS intensity. The 1-dimensional scattering profiles were then extracted along the meridian line as well as equatorial line for the anisotropic scattering patterns, and averaged 1D scattering profiles were obtained for the isotropic 2D SAXS patterns.^44,45 The isotropic part of the 2D SAXS patterns was sectioned in the range of 0–180°, and the Bragg's long spacing (L) attributed to the isotropic PP was determined. Two sections of the oriented patterns were prepared: (i) along the meridian (−45° to +45°) giving rise to L^Eq and (ii) along the equator (45° to 135°) for L^Mer. The 1D SAXS intensity profiles of circularly integrated 2D SAXS patterns at various zones of the two samples are shown in Fig. 6. The long spacing (L), reflecting the thickness of the lamellar and amorphous regions between the two lamellae, were obtained and are listed in Table 2. For the skin layer of the GAIM sample, it can be observed that L^Eq (16.8 nm) was bigger than L^Mer (15.9 nm). The variables of the periodicities for the sub-skin layer followed the same tendency, and the corresponding periodicities of the sub-skin layer were larger than that of the skin layer, indicating a larger thickness of the iPP lamellae and more regularly aligned lamellae in the sub-skin layer. For the sub-skin zone of the GAIM sample, the velocity gradient between the interface of the high-pressure gas and molten polymer introduced a strong shear field during the gas penetration process. In general, the iPP lamellae along the equator have bigger long spacings when compared to those of the iPP crystallized along the fiber direction. As shown in Table 2, one can observe that the iPP lamellae of the oriented zones (skin and sub-skin layers) of the GAIM specimen have bigger periodicities than the non-oriented zones (gas channel). Furthermore, it can be seen that the GAIM iPP specimen exhibited higher long periods when compared to the pure CIM sample, indicating a larger thickness of iPP lamellae or more perfect crystalline structure. This originated from the high melt shearing induced crystallization during the gas penetration of the GAIM process. In addition, the long spacing of the core zone of the CIM sample was larger than the skin layer, which was caused by the low cooling rates in the core zone of the CIM sample.

Fig. 6 1D SAXS intensity profiles of skin, sub-skin and gas channel layers of the (G1, G2, G3) GAIM samples and skin and core layers of the (C1, C2) CIM samples. Gx^Eq (x = 1, 2) denotes the 1D SAXS intensity profiles along the equator, and Gx^Mer (x = 1, 2) denotes the 1D SAXS intensity profiles along the meridian.

Table 2 Long spacing values at different zones of the GAIM and CIM samples

Samples	GAIM					CIM
Zones	Skin		Sub-skin		Gas channel	Skin	Core
	L^Eq	L^Mer	L^Eq	L^Mer	L	L	L
Long spacing (nm)	16.8	15.9	17.7	16.2	15.7	12.9	14.4

---

Crystal morphologies of iPP

The crystal morphologies in the different zones of pure iPP molded via the GAIM process was investigated by PLM, and are shown in Fig. 7. It should be emphasized that PLM provides useful information regarding the entire structure of a sample in spite of its relatively low resolution. As depicted in Fig. 7, the molded bar fabricated by GAIM displayed a typical skin-core morphology, and when observed in the cross section, the boundary between the non-spherulitic and spherulitic zone appeared quite sharp. However, the corresponding areas observed parallel to the flow direction were found to contain a layer between the skin and gas channel zone, which was not readily apparent in the cross sections. This sub-skin layer, or what will be called the shear zone, was spherulitic but the morphology differed from that observed in the gas channel zone, in which the spherulites (or sheaves) in the shear zone were row nucleated while those in the gas channel zone were randomly nucleated. The appearance of three zones in a longitudinal thin section is shown in Fig. 7a. As shown in Fig. 7b, the exterior layer (skin and shear layer) consisted of regions of small, bright spherulites exhibiting negative birefringence. The gas channel zone was formed by a spherulitic domain structure as shown in Fig. 7c. The spherulites or sheaves (oriented lamellar) in the skin and shear layer were tightly bunched together along thin, negatively birefringent rows. The spherulites that were nucleated from each row were nearly identical in size and shape; their size increased with increasing distance from the mold surface. The majority of the spherulites in the exterior were the negatively birefringent type III variety, which is characteristic of the β-crystal form of polypropylene.⁴⁶ The gas channel zone contained larger but randomly sized spherulites. With the exception of the occasional type III spherulites, the spherulites in the gas channel zone were positively birefringent type I variety, which is characteristic of the stable, monoclinic polymorph.⁴⁶ For comparison, the crystal morphologies in the CIM part were also investigated to explore the role of secondary gas penetration on the formation of different crystal forms. As shown in Fig. 7d, the trend of morphological evolution at corresponding position along the residual thickness was similar for both the GAIM and CIM specimens. However, the thickness of the exterior layer of the GAIM sample was about 620 μm, which was much larger than that of the CIM part (about 300 μm). The elucidation of the morphological fine structure of each layer is of interest because it might cast some light on the mode of formation of these layers, and particularly the role played by secondary gas penetration in their formation. Because PLM can not give information on the fine structures of the molded bars, in the subsequent section, scanning electron micrographs of etched samples will be presented, which clearly identify some major fine crystalline features.

Fig. 7 PLM micrographs of the crystal morphologies in different zones of the GAIM sample: (a) the whole picture; (b) near the skin zone; (c) near the gas channel zone; and CIM sample (d).

Hierarchy in crystal structures

To obtain the fine architecture of the molded bars, the morphological features of the etched iPP at different zones along the flow direction were studied by SEM. Macroscopically, the dominating feature of the GAIM sample was the shear-induced morphologies with the gas channel layer in the center, which is a highly-oriented layer surrounding the gas channel zone and the skin layer in the cross sectional areas of the specimen. Under a flow field, the polymer chains can be stretched and oriented into rows along the flow direction in the molten state. These prearranged rows of molecule chains can be transformed into crystalline clusters and act as crystallization nuclei for the subsequent crystallization process. This characteristic linear type of self-nucleus is termed a row nucleus.^47,48 Row nuclei can induce an epitaxial growth of crystallites in folded chains, leading to the characteristic supermolecular formations of cylindrical symmetry, i.e., cylindrites.⁴⁸ As shown in Fig. 8a, it is found that the skin layer of the GAIM sample was characterized by β-cylindrites aligned in rows along the flow direction. The cylindritic structure is one kind of several kinds of supermolecular formations, which always corresponds to another crystalline morphology called “transcrystalline”. The difference between them is that the former is developed via homogeneous or self-nucleation, while the latter is formed by heterogeneous nucleation.⁴⁸

Fig. 8 SEM microphotographs of the crystalline morphologies of the GAIM part: (a) skin layer; (b) sub-skin layer; (c) gas channel layer; (d) is the magnification of (b). The melt flow direction is parallel to the meridian direction.

Most interestingly, when compared to the skin layer, one can observe shish-kebab-like fibrillar structure in the sub-skin layer (see Fig. 8b) as well as the β-cylindrites. The shish-kebab structure, which was first found in stirred polyethylene solution by Pennings and co-workers,⁴⁹ is usually a predominant morphological feature when a polymer crystallizes under shear flow. The so-called “shish-kebab” structure, consisting of a long central fiber core (shish) and lamellar crystalline structure (kebab) periodically attached to the shish, substantially influences the final physical and mechanical properties of the compound.⁵⁰ In this sub-skin zone or shear layer, as the gas penetrates, the velocity gradient between the gas front and melt front induced a strong shear field, in which the shear rate might exceed the critical value for shish-kebab formation, inducing the extension and crystallization of molecular chain segments along the shear direction and developing into a stable structure. The stable structure, i.e. shish, as primary nuclei, further induced the folded-chains to overgrow along the flow direction, and finally the shish-kebab morphology could be formed. However, the shish kebab structures were not observed in the skin zone probably due to the weak shear field in this region.¹⁰ As mentioned above, one of the most notable differences in the crystalline morphologies between the skin and sub-skin layers of the GAIM sample was the existence of shish-kebab structures in the sub-skin layer. As shown in the magnified micrographs in Fig. 8c, the shish-kebab structure can be clearly seen in the sub-skin zone of the GAIM sample. It is noted that it is hard to find shish-kebab structures in the skin layer of the GAIM sample. However, it does not necessarily imply that there are no shish-kebab structures in the skin layer because the formed shish-kebab may be too small or too diluted to be detected. To the best of our knowledge, the coexistence of β-cylindrites and shish-kebab morphology is observed for the first time in pure iPP parts molded by GAIM. The presence of β-cylindrites in the injection-molded parts of neat iPP is also rarely reported. In the gas channel layer (Fig. 8d), common α-spherulites of polypropylene were obtained, and the lamellae primarily exhibited a random arrangement with much less oriented structures. In addition, it should be mentioned that there were some black grains in the SEM microphotographs, which may be caused by the etching procedures or raw materials; therefore, further research is required. For the CIM part, the dominant morphological feature from skin to core zone was the α-spherulites of polypropylene, which was similar to the gas channel zone of the GAIM part (see Fig. 9).

Fig. 9 SEM microphotographs of the CIM part: (a) skin layer; (b) core layer. The melt flow direction is parallel to the meridian direction.

Thermal properties

The thermal behaviors in different zones of the iPP sample molded by GAIM are shown in Fig. 10, and the corresponding melting parameters are listed in Table 3. As shown in Fig. 10a, the skin and sub-skin layer of the GAIM sample exhibited three endothermic melting peaks, whereas the gas channel layer of PP exhibited only a single endothermic peak at around 166 °C. The first two peaks at around 146–152 °C of the skin and shear layers of the GAIM sample were caused by the melting of the original β₁-crystals and subsequent recrystallization for producing a more stable structure (β₂) during scanning, while the last peak was associated with the melting of the original and recrystallized α-crystals.⁵¹ However, such double melting phenomenon of the β-crystals was not observed in the CIM sample, and only a single melting peak of α-crystals was observed in Fig. 10b for both the skin and core layer of the CIM sample.

Fig. 10 DSC heating curves of the different zones of GAIM (a) and CIM parts (b).

Table 3 Melting parameters of the iPP at different layers

	Layer	ΔHα J g^−1	Xα %	ΔHβ J g^−1	Xβ %	ϕβ %	Kβ
GAIM	Skin	66.01	37.29	8.15	4.84	11.48	0.28
	Sub-skin	64.28	36.32	12.72	7.55	17.20	0.33
	gas channel	78.23	44.20	—	—	—	0.13
CIM	Skin	69.97	39.53	—	—	—	0.06
	Core	73.90	41.75	—	—	—	0

---

As listed in Table 3, the total fusion heat (α- and β-crystals) of the GAIM sample slightly increased from the skin layer to the sub-skin layer and then to the gas channel zone. This was because of the cooling rate decreasing along the residual thickness, and thus the gas channel layer exhibited the slowest cooling rate and therefore the longest time for crystallization. Moreover, the flow field also contributed to the enhancement of crystallinity.⁵² In addition, the total fusion heat at each zone of the GAIM sample was larger than that of the CIM part. These differences were mainly due to the strong shear field introduced during the secondary shear process of the GAIM sample. These results were firmly confirmed by the abovementioned morphological results, which clearly showed that the GAIM sample had a more oriented lamellar and shish-kebab-like structure. There are two reasons,¹ both thermodynamic and kinetic, for the enhancement of crystallization by chain extension. On one hand, the flow-induced orientation increases the melting temperature of the material, and hence the crystallization temperature, i.e. the entropy of the melt is lowered and the free energy is increased. On the other hand, the extended chain is closer to its final configuration state in the crystal, and therefore has less kinetic barrier to overcome than that of a chain in the random state to form a crystal structure. Therefore, the GAIM part has larger values of fusion heat because of a stronger shear field and a higher cooling rate.

The percentage of β-phase (ϕ_β) was determined according to eqn (1) and is listed in Table 3. Based on the DSC results, it can be seen that only the skin and sub-skin layer of the GAIM sample had β-crystals, and the ϕ_β of the skin and sub-skin layer were 11.48% and 17.20%, respectively. However, it does not necessarily imply that there were no β-crystals in the gas channel layer of the GAIM or CIM samples because the β → α phase transition can occur in PP during thermal treatment and depends on the structure and content of the β-crystals in molded bars induced by the flow field.¹⁶ Due to the β → α phase transition that occurs in PP during thermal treatment, wide-angle X-ray diffraction (WAXD) measurements were carried out in the following analysis.

Distribution of crystalline morphologies

Fig. 11 depicts the WAXD patterns of the molded bars of iPP. The reflections at 2θ angles of 14.0°, 16.8° and 18.6° correspond to the (110), (040) and (130) lattice planes of the monoclinic α-crystals of iPP, respectively, and the reflection at 2θ of 16° corresponds to the (300) plane of the hexagonal β-crystals of iPP.³⁹

Fig. 11 The WAXD patterns of the different layers of the iPP molded by (a): GAIM and (b) CIM.

As depicted in Fig. 11a, it was surprising that β-crystals were detected in the entire cross section of the GAIM sample, which was different from the abovementioned DSC results. Thus, by using the GAIM method, we can obtain a pure iPP sample with β-crystals existing in the entire molded bars without adding β-nucleating agents or other secondary components. In previous studies, pure iPP samples with β-crystals were found only in the skin layer or core layer of the molded bars depending on the different processing methods and the molecular peculiars of polypropylene.^28–31 β-Nucleating agents or other secondary component were required to be added to obtain a part with β-crystals in the entire cross section.^19–21 To the best of our knowledge, this peculiar finding is observed for the first time in a pristine iPP part molded by GAIM, which was correlated with a strong shear field introduced by the secondary shear and high cooling rate of the GAIM process. It has been widely accepted that β-crystals can be formed in a sheared iPP melt,^53–55 in which melt-shearing yields α-row-nuclei, and the surface of these α-row-nuclei may induce an α → β transition in the crystalline modification of iPP, leading to β-modification. Most of these studies were mainly performed under isothermal crystallization conditions under low shear, low strain and single flow field conditions.^52,55 However, polymer materials usually experience more complex flow fields, consisting of shear, elongational, or mixed, thermal fields, and the interactions between melts in practical processing. Therefore, our results may provide some tips for the practical production of pure iPP specimens. For the CIM sample, the β-crystals of iPP only existed in the skin layer, which was in accordance with other studies on injection molding parts.³⁰

The relative content of the β-phase (K_β) of the molded bars were calculated according to eqn (6), and the results were summarized in Table 3. For the GAIM sample, the K_β value in the sub-skin layer was the largest and the gas-channel zone possessed the lowest K_β, which was consistent with the hierarchical structures observed by SEM. However, there was a low value of K_β in the skin layer of the CIM sample, which indicated a much lower shear field in the CIM process. Compared with CIM, the K_β value of the GAIM sample was substantially increased, which indicated that the stronger shear field induced during the gas penetration of GAIM process promoted the formation of the β-crystals of iPP. As shown in Table 3, it can be observed differences existed in the β-crystals of iPP measured by DSC and WAXD methods. The obvious differences between the two values of K_β and ϕ_β in Table 3 may be due to the occurrence of the β → α phase transition during the heating process of DSC, and thus the β-crystals of iPP measured by DSC was smaller than the real values. Because the WAXD measurement was performed at ambient temperature and did not destroy the microstructure of the molded bars, it was reasonable to assume that the results of WAXD were more close to the real values.

Mechanical properties

The selective stress–strain curve of the GAIM sample is shown in Fig. 12, including the CIM sample for comparison. There was obvious necking for both the GAIM and CIM samples with ductile failure. The obtained values of yield stress and strain are listed in Table 4. Yield stress, which is a measure of the strength at the point of deviation from Hooke's law (the yield point), is important for polymeric materials. As shown in Table 4, a mechanical enhancement from 32.2 MPa for the CIM sample to 43.0 MPa for the GAIM sample for yield stress was observed, indicating a better strength for the GAIM sample. The difference in the values of yield stress were mainly caused by the microstructure and crystalline structures among the iPP articles, and the yield stress was directly proportional to the content of the different crystal forms and the area of the oriented zones.^56,57 It is known that the strength of the α-crystals is higher than the β-crystals of iPP. Because the content of α-crystals of the GAIM was lower than the CIM sample, the higher strength of GAIM sample can be attributed to the larger area of the oriented zone, which was consistent with the morphological results. The yield strain can be considered the ratio of material deformability at a low strain. It is known that the values are increasing with the ductility of the material. Regarding the higher ductility of the β-crystals when compared to the α-crystals, a higher value of the yield strain of the GAIM sample was obtained. As is known, the strain at break represents cold drawability, the area under the stress–strain curve, in fact, manifests the work spent on the failure, i.e., specimen toughness. More importantly, there was a sharp enhancement in the strain at break of the GAIM sample, which was five times larger than that of the CIM sample. Therefore, the GAIM specimen was characterized by a large enhancement of yield stress and strain at break when compared with the CIM sample, indicating that a high strength and high toughness molded bar was obtained using the GAIM technique. This was probably due to the larger values of highly oriented zones and higher content of β-crystals in the GAIM part.

Fig. 12 Strain–stress curves for the GAIM and CIM parts.

Table 4 Tensile properties for the GAIM and CIM samples

Sample	Yield stress (MPa)	Yield strain (%)	Strain at break (%)
GAIM	43.0 ± 2.1	16 ± 2	460 ± 45
CIM	32.2 ± 1.4	12 ± 2	95 ± 12

---

Formation mechanisms of hierarchical structures in GAIM

During the GAIM process, iPP crystallized from a melt, which was exposed to complex flow and temperature variations, and the final specimen was characterized by an intrinsic heterogeneous microstructure, featuring a gradual and hierarchical variation of morphology that evolved throughout the part. The effect of the shear field on the hierarchical structures of the crystal is expected to depend on the extent of flow-induced orientation and the rate at which this orientation relaxes (τ·γ′).⁵⁸ The shear field facilitates the formation of β-crystals^31–33 and according to the model of “row nucleated structures” proposed by Janeschitz-Kriegl and co-workers,⁵⁹ the morphological development of iPP under different shear strength can be depicted in Fig. 13. As shown in Fig. 13a, under no shear or low shear (τ·γ′ ≪ 1), there were no shearing prenuclei and only some random point nuclei existed, and finally α-spherulites were obtained. At medium shear (τ·γ′ < 1), shear flow promoted local chain alignment and generated some precursors with an intermediate degree of order. These precursors, comprising locally oriented, stretched chain segments and chains, may be the “prenuclei” of amorphous character proposed by Varga and Karger-Kocsis,⁶⁰ the quasi-ordered clusters whose size and orientation distribution strongly depend on the intensity of the flow field,⁶¹ or the oriented, metastable and non-crystalline phase.⁶² Point nuclei (under shear flow) came from shear prenuclei, which were aligned parallel to the flow direction, and can facilitate the formation of β-spherulites, as shown in Fig. 13b. At strong shear (τ·γ′ > 1), when the applied shear stress exceeded the critical shear stress, threadlike nuclei formed. The high aspect ratio of the threads causes lamellar growth for them to be laterally constrained, resulting in the formation of a highly oriented, row-nucleated morphology (β-cylindrites). At very high strong shear (τ·γ′ ≫ 1), numerous threadlike nuclei were formed; the lamellar growth from them was highly laterally constrained and then shish-kebab can be obtained. The cylindritic and shish-kebab structures are developed via homogeneous or self nucleation, and the main difference between them is the density of row nuclei or threadlike nuclei.

Fig. 13 Schematic illustration of the morphological development of iPP under different shear strengths.

To understand the different crystalline structure formation during the GAIM process, the flow behaviors in the separated filling stages (such as short shot and gas penetration process) should be made clear. The molecular chains were exposed to the shear field during both the short shot and gas penetration process. It is proven by our previous studies that the short shot process mainly affected the superstructures near the mold surface and the gas penetration process played a dominating role on the crystalline structures away from the mold surface.^13–15 A schematic illustration of the shear rate profile in GAIM is shown in Fig. 14. As indicated in Fig. 14, the shear rate showed a sharp enhancement from the skin layer and reached its maximum at the sub-skin layer. Then, it began to fall to a minimum at the gas-channel layer. Based on these analyses and the morphological results, the diagram of hierarchical structures of the GAIM part is depicted in Fig. 14. The shear strength was considered to be critical for the formation of β-spherulites, β-cylindrites and shish-kebab structures in the pure iPP specimen. When τ·γ′ was sufficiently high, shish-kebab-like structures were formed. As the value of τ·γ′ decreased, β-cylindrites and β-spherulites appeared. When there was no shear or low shear, β-crystals disappeared and α-spherulites dominated the corresponding area.

Fig. 14 Schematic drawing of the shear rate profile during the GAIM process and a diagram of the hierarchical structures of the GAIM part. For brevity, the dimensionless distance was adopted and the number “0” and “1” denote the mold surface and gas channel, respectively.

Conclusion

The orientation behavior, hierarchical distribution of β-crystals and mechanical properties of neat iPP molded by GAIM were intensively investigated in this work. Spatial variation of the superstructure with changed orientation and crystalline morphology along the thickness direction of the molded bar was developed because of the interaction between the temperature and shear field. It was observed for the first time that both the shish-kebab and β-cylindrite morphology coexisted at the sub-skin layer of the GAIM part, where the shear strength was the greatest. More importantly, a neat iPP specimen with β-crystals existing in the entire cross section of the molded bar was obtained for GAIM sample, while for CIM sample, a low value of β-crystals was only observed in the skin layer. In addition, when compared with the CIM sample, the yield stress of the GAIM sample was significantly increased, accompanying with a sharp enhancement (five times) of the strain at break. These results firmly authenticated that a high strength and high toughness sample was obtained for the GAIM specimen due to a larger value of the highly oriented zone and higher content of β-crystals. Finally, a schematic illustration was proposed to interpret the formation mechanism of the unusual hierarchical distribution of β-crystals in GAIM parts.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Grant no. 21174092, 20874066 and 51121001). We acknowledge the assistance of Mr Chao-liang Zhang from the Huaxi College of Stomatology, Sichuan University for SEM experiments.

References

1. A. Kelle and H. W. H. Kolnaar, in Processing of polymers, ed. Meijer HEH, VCH New York, 1997, vol. 18.
2. R. H. Somania, L. Yang, L. Zhu and B. S. Hsiao, Polymer, 2005, 46, 8587.
3. X. F. Wei, R. Y. Bao, L. Gu, Y. Wang, K. Ke, W. Yang, B. H. Xie, M. B. Yang, T. Zhou and A. M. Zhang, RSC. Adv., 2014, 4, 2733.
4. H. Fukushima, Y. Ogino, G. Matsuba, K. Nishida and T. Kanaya, Polymer, 2005, 46, 1878.
5. K. Zhang, Z. Y. Liu, B. Yang, W. Yang, Y. Lu, L. Wang, N. Sun and M. B. Yang, Polymer, 2011, 52, 3871.
6. L. Xu, Y. F. Huang, J. Z. Xu, X. Ji and Z. M. Li, RSC. Adv., 2014, 4, 1512.
7. G. Kalay and M. J. Bevis, J. Polym. Sci., Part B: Polym. Phys., 1997, 35, 241.
8. G. Kalay, P. Zhong, P. Allan and M. J. Bevis, Polymer, 1996, 37, 2077.
9. J. Avery, Gas-assist injection molding: principles and application, Hanser publishers, Munich, 2001.
10. G. Q. Zheng, L. Huang, W. Yang, B. Yang, M. B. Yang, C. T. Liu and C. Y. Shen, Polymer, 2007, 48, 5486.
11. L. Huang, W. Yang, B. Yang, M. B. Yang and H. N. An, Polymer, 2008, 49, 4051.
12. L. Wang, B. Yang, W. Yang, N. Sun, B. Yin, J. M. Feng and M. B. Yang, Colloid Polym. Sci., 2011, 289, 1661.
13. L. Wang, B. Yang, N. Sun, K. Zhang, J. M. Feng and M. B. Yang, Colloid Polym. Sci., 2012, 290, 1133.
14. G. Q. Zheng, W. Yang, L. Huang, Z. M. Li, M. B. Yang, B. Yin, Q. Li, C. T. Liu and C. Y. Shen, J. Mater. Sci., 2007, 42, 7275.
15. G. Q. Zheng, W. Yang, C. T. Liu, C. Y. Shen and M. B. Yang, Mater. Lett., 2007, 61, 3436.
16. C. Maier and T. Calafut, Polypropylene-the definitive user's guide and databook, Plastics Design LibraryNorwich, 1998.
17. B. Lotz, J. C. Wittmann and A. J. Lovinger, Polymer, 1996, 37, 4979.
18. A. J. Lovinger, J. O. Chua and C. C. Gryte, J. Polym. Sci., Polym. Phys. Ed., 1977, 15, 641.
19. K. Cho, D. N. Saheb, H. Yang, B. Kang, J. Kim and S. Lee, Polymer, 2003, 44, 4053.
20. H. B. Chen, J. Karger-Kocsis, J. S. Wu and J. Varga, Polymer, 2002, 43, 6505.
21. H. W. Bai, Y. Wang, Z. J. Zhang, L. Han, Y. L. Li, L. Liu, Z. W. Zhou and Y. F. Men, Macromolecules, 2009, 42, 6647.
22. R. H. Somani, B. S. Hsiao, A. Nogales, S. Srinivas, A. H. Tsou, I. Sics, F. J. Balta-Calleja and T. A. Ezquerra, Macromolecules, 2000, 33, 9385.
23. R. H. Somani, B. S. Hsiao, A. Nogales, H. Fruitwala, S. Srinivas and A. H. Tsou, Macromolecules, 2001, 34, 5902.
24. P. W. Zhu, J. Tung, A. Phillips and G. Edward, Macromolecules, 2006, 39, 1821.
25. Q. X. Zhang, J. B. Song, S. Y. Wang and Z. S. Mo, Polymer, 2005, 46, 11820.
26. P. W. Zhu and G. Edward, Macromol. Mater. Eng., 2003, 288, 301.
27. H. W. Bai, Y. Wang, B. Song, Y. L. Li and L. Liu, Polym. Eng. Sci., 2008, 48, 1532.
28. M. R. Kantz, H. D. Newman and F. H. Stigale, J. Appl. Polym. Sci., 1972, 16, 1249.
29. D. R. Fitchmun and Z. Mencik, J. Polym. Sci., Polym. Phys. Ed., 1973, 11, 951.
30. H. Dragaun, H. Hubeny and H. Muschik, J. Polym. Sci., Polym. Phys. Ed., 1977, 15, 1779.
31. W. Wenig and F. Herzog, J. Appl. Polym. Sci., 1993, 50, 2163.
32. Č. Roman, O. Martin, P. Petr, P. Martina, S. Karel and L. Anežka, Eur. Polym. J., 2005, 41, 1838.
33. L. Wang, J. H. Wang, Y. Wang, B. Yang, Q. P. Zhang, M. B. Yang and J. M. Feng, Colloid Polym. Sci., 2013, 291, 1503.
34. B. Lee, T. J. Shin, S. W. Lee, J. Yoon, J. Kim and M. Ree, Macromolecules, 2004, 37, 4174.
35. B. Lee, T. J. Shin, S. W. Lee, J. Yoon, J. Kim, H. S. Youn, K. B. Lee and M. Ree, Polymer, 2003, 44, 2509.
36. R. H. Olley and D. C. Bassett, Polymer, 1982, 23, 1707.
37. J. X. Li, W. L. Cheung and D. Jia, Polymer, 1999, 40, 1219.
38. J. X. Li and W. L. Cheung, Polymer, 1998, 39, 6935.
39. A. Turner-Jones, J. M. Aizlewood and D. Beckett, Makromol. Chem., 1964, 75, 134.
40. R. H. Somani, L. Yang, B. S. Hsiao, T. Sun, N. V. Pogodina and A. Lustiger, Macromolecules, 2005, 38, 1244.
41. R. H. Somani, L. Yang, B. S. Hsiao, P. K. Agarwal, H. A. Fruitwala and A. H. Tsou, Macromolecules, 2002, 35, 9096.
42. K. Heo, J. Yoon, K. S. Jin, S. Jin, H. Sato, Y. Ozaki, M. M. Satkowski, I. Noda and M. Ree, J. Phys. Chem. B, 2008, 112, 4571.
43. Y. Y. Kim, K. Heo, K.-W. Kim, J. Kim, T. J. Shin, J. R. Kim, I. S. Yoon and M. Ree, Macrmol. Res., 2014, 22, 194.
44. B. Ahn, T. Hirai, S. Jin, Y. Rho, K.-W. Kim, M.-A. Kakimoto, P. Gopalan, T. Hayakawa and M. Ree, Macromolecules, 2010, 43, 10568.
45. T. J. Shin, B. Lee, B. S. Seong, Y. S. Han, C.-H. Lee, H. H. Song, R. S. Stein and M. Ree, Polymer, 2010, 51, 5799.
46. H. D. Keith, F. J. Padden, J. N. M. Walter and H. W. Wyckoff, J. Appl. Phys., 1959, 30, 1485.
47. A. Kellar and J. Manchin, J. Makromol. Sci. B, 1967, 1, 41.
48. J. Varga, J. Mater. Sci., 1992, 27, 2557.
49. A. J. Pennings and A. M. Kiel, Kolloid Z. Z. Polym., 1965, 205, 160.
50. F. Mai, K. Wang, M. J. Yao, H. Deng, F. Chen and Q. Fu, J. Phys. Chem. B, 2010, 114, 10693.
51. G. Shi, Z. Zhang, Y. Cao and J. Hong, J. Makromol. Chem., 1993, 194, 269.
52. J. K. Keum, F. Zuo and B. S. Hsiao, Macromolecules, 2008, 41, 4766.
53. J. Varga and J. Karger-Kocsis, J. Polym. Sci., Part B: Polym. Phys., 1995, 34, 657.
54. J. Moitzi and P. Skalicky, Polymer, 1993, 34, 3168.
55. R. H. Somani, B. S. Hsiao, A. Nogales, H. Fruitwala, S. Srinivas and A. H. Tsou, Macromolecules, 2001, 34, 5902.
56. R. Su, J. X. Su, K. Wang, C. Y. Yang, Q. Zhang and Q. Fu, Eur. Polym. J., 2009, 45, 747.
57. J. Varga, J. Macromol. Sci. Phys., 2002, 41, 1121.
58. A. Elmoumni, H. H. Winter, A. Waddon and H. Fruitwala, Macromolecules, 2003, 36, 6453.
59. S. Liedauer, G. Eder and H. Janeschitz-Krigel, Int. Polym. Process., 1995, 3, 243.
60. J. Varga and J. Karger-Kocsis, J. Polym. Sci., Part B: Polym. Phys., 1996, 34, 657.
61. B. J. Monasse, J. Mater. Sci., 1992, 27, 6047.
62. R. Androsch, M. L. Di Lorenzo, C. Schick and B. Wunderlich, Polymer, 2010, 51, 4639.

---