Influence of crystal orientation on stretching induced void formation in poly(4-methyl-1-pentene) investigated by in-situ small-angle and wide-angle X-ray scattering

2.1 Materials and sample preparation

P4MP1 used in this study was purchased from Mitsui Chemicals with a trademark of MX004, the melt flow rate is 25 g/10 min (260°C, 5 kg).³² The as-received P4MP1 pellets were injection molded by Demag 80 (Germany) to get plates with preferred orientation along the machine direction. The barrel temperature was set at 270, 275, 255, and 100°C from the nozzle to the hopper. The mold temperature was 90°C. The injection pressure and holding pressure were 1100 and 600 bar. The injection molding speed was 80 cm³/s. The width, length, and thickness of the specimen were 60, 80, and 0.3 mm. To enlarge the melt flow time and enhance the orientation, a spew groove was added at the end of the cavity. After injection molding, the plates were suffered a heating treatment shown in Figure 1A to adjust the orientation state inside. The heating temperature (T_h) in the thermal protocol was selected according to the melting behavior of the injection-molded plate (Figure 1B). The duration of the treatment at T_h was 10 min. For simplification, the sample was labeled as P4MP1-T_h. P4MP1-ref refers to the sample without any heating treatment after injection molding.

2.2 Differential scanning calorimetry (DSC) measurement

The melting behavior of the sample was characterized by TA instrument Q2000 (USA) under nitrogen atmosphere. Five milligrams sample was heated from room temperature to 280°C with a heating rate of 10°C/min and then cooled to 80°C with a cooling rate of 10°C/min. The heat absorption during heating was recorded.

2.3 Scanning electronic microscopy (SEM) measurement

The morphology of the sample before and after stretching was examined by JEOL JSM-6380 (Japan) operated at an accelerating voltage of 10 kV. Before the observation, the sample was etched by a mixture of KMnO₄-H₂SO₄-HNO₃ and then sputter-coated with a layer of gold.

2.4 Atomic force microscopy (AFM) measurement

The morphology of the sample was also observed by Bruker MultiMode 8 (Germany) by tapping mode. Before the measurement, the sample was cut by Cryo-ultra-microtome (Leica EM FC7) at −80°C.

2.5 Uniaxial stretching and in-situ synchrotron X-ray scattering experiments

Uniaxial stretching was performed on the Linkam TST 350 tensile stage (UK). Before stretching, all specimens were kept at 30°C for 2 min to reach the thermal equilibrium state. The reason for choosing 30°C instead of the room temperature is to assure that the environment temperature during stretching is always the same. A step-stretching method was adopted with a step of 80 um, the stretching speed was 20 μm/s. To keep the sampling position where X-ray passes unchanged, a waist-shape geometry was adopted for the tensile specimen. The details of the specimen's geometry can be found in our previous reports.^{29, 33, 34} Additionally, optical images were taken every step by an auto-shoot high-resolution camera. The local strain where X-ray past was calculated by the following equation

()

with and are the initial and instant widths of samples during stretching which were obtained by gauging optical photo images.

In-situ two-dimensional (2D)-SAXS and 2D-WAXS measurements were performed on Bruker Nanostar (Germany) with Vantec 2000 area detector and Bruker D8 ADVANCE (Germany) with Vantec 500 area detector, respectively. The resolution of the detectors is 2048 × 2048 pixels, and the pixel size is 68 × 68 μm². The sample to detector distance was 1087.5 and 148.8 mm, respectively. The wavelength of the X-ray was 0.154 nm and the energy of the X-ray is ca. 8 keV. To make sure that the scattering patterns have enough resolution, the collection time was set at 30 s. The diameter of the X-ray beam is 0.5 mm. The SAXS and WAXS patterns were further processed by self-written Python scripts.

3.1 The microstructure and morphology of P4MP1 suffered different heating treatment

Figure 2 presents the melting behavior of the sample treated at different T_h. On the curve of P4MP1-ref, the melting peak was located at 234.2°C. Treating at 150°C influenced slightly the melting behavior. As T_h was elevated to 223 and 228°C, a broad shoulder peak appeared at 228°C due to the melting of the imperfect crystals formed during the heat treatment.²⁵ When T_h was in the range of 234–238°C, the shoulder peak disappeared and the main peak shifted obviously to a higher temperature as a result of crystal perfection. When T_h was higher than 240°C, the temperature of the main peak started to decrease and finally stayed at 234°C. In addition, a small shoulder peak can be recognized as T_h was higher than 250°C. The appearance of this shoulder peak should be ascribed to the melting of metastable crystals formed during the cooling process in the heating treatment.

To check the orientation of the sample before stretching, 2D-WAXS patterns are provided in Figure 3. On the pattern one can see a series of diffractions from inner to outer region, which are originated from different crystal planes. It is well known that P4MP1 is a polymer with polymorphic behavior: Form I owns a tetragonal unit cell (a = 18.88 and c = 13.8 Å),^{35, 36} Form II exhibits a monoclinic unit (a = 18.50, b = 10.43, c = 7.22 Å and γ = 113°),³⁷ Form III presents a tetragonal cell (a = 19.16 and c = 7.12 Å),³⁸ and Form IV shows a hexagonal structure (a = 22.17 and c = 6.69 Å).³⁸ Among all crystal forms, Form I is the most stable one and is also the only one that can be crystallized directly from the melt. In Form I, the chains are characterized by 7₂ helical conformations. As verified by the results in Figure 2, no characteristic peak of Form II (133°C) or Form III (74°C)³⁹ can be found, so it is inferred that only Form I exists in the sample. These diffractions were assigned to (001), (200), (220), (3 1), (2 2), (3 1), (4 1), and (420) crystal planes of Form I considering their two-theta value. For the sample without any treatment, the intensity distribution of the diffraction indicates an anisotropic structure in the sample.⁴⁰ As T_h was higher than 236°C, the intensity distribution of the diffraction was changed obviously from a combination of arc and circle to full circle with increasing T_h, indicating a transition of orientation from oriented to randomly distributed one.

The intensity distribution of the (200) crystal plane is plotted in Figure 4A to provide a semi-quantitate evaluation of the crystal orientation. In the temperature range of 150–228°C, the height of the peak located at 90° was enlarged slightly, suggesting that the orientation was enhanced a little bit; in the temperature of 228–240°C, the height of the peak was decreased and the width of the peak was enlarged, indicating that the orientation started to decrease. Furthermore, the orientation degree of the c-axis in the crystal was calculated by the Hermann's orientation factor f_H given below

()

with Z refers to the flow direction (namely the machine direction) during injection molding and represents the mean-square cosine (average overall crystallites) of the angle between the c-axis and the Z direction. f_H = 1 means that the c-axis is perfectly oriented along the flow direction, f_H = −0.5 suggests the c-axis is perpendicular to the flow direction, and f_H = 0 means a random orientation. Since that the intensity of (001) crystal plane is quite weak and disappeared as T_h is higher than 240°C, so is given by

()

()

with is the intensity distribution of (200) crystal plane as a function of the azimuthal angle ( ). f_H is summarized in Figure 4B. In addition, the crystallinity of the sample calculated from the 1D-WAXS curve through a standard peak fitting procedure is provided in Figure 4C. 1D-WAXS curves were obtained by circular integration of the 2D-WAXS patterns. The peak fitting procedure is shown in Figure S1.

It can be found that f_H was 0.49 and the crystallinity was 0.38 in P4MP1-ref, and they were enlarged to 0.52 and 0.40, respectively, when T_h was 150°C. As shown in Figure 2, no additional melting peak arose, so the increase of f_H and the crystallinity was probably caused by the lamellae perfection and molecular chains rearrangement in the amorphous phase.^{25, 41, 42} As T_h was higher than 223°C and lower than 234°C, part of the previously existed crystals was melted. The residual part served as nuclei to induce the formation of the oriented crystal during the cooling process in the heat treatment, as a result of the “memory effect”.^{43, 44} So f_H was further increased to 0.58. Meanwhile, the crystallinity was enlarged drastically to 0.47. “Memory effect” refers to the phenomenon that the polymer melt containing clusters of molecular segments with a more ordered conformation than the equilibrated random coil will crystallize faster due to a depression in free energy change for nucleation.^{43, 44} When the semicrystalline polymer is heated to a temperature close to its nominal melting temperature or well below the equilibrium melting point, the “memory effect” will be triggered.^{45, 46} As T_h was higher than 236°C, f_H began to decrease and reached 0 when T_h was 250°C due to that the majority part of the previously existed crystals was destroyed and lost the orientation nature, although the residual part can still serve as nuclei to increase the crystallinity of the sample (see Figure 4C). Lastly, as T_h was higher than 250°C, the crystallinity of the sample was decreased to 0.41, which was only slightly larger than P4MP1-ref and P4MP1-150.

Normally, due to the existence of the shear rate and the temperature gradient, a “skin-core” structure will be formed in the injection-molded sample^{47, 48}: in the skin region a frozen layer and a highly oriented layer can be recognized, and in the core region a homogenous region with spherulite can be observed. The morphology of P4MP1-ref and P4MP1-270 are shown in Figure 5. The horizontal dark line in Figure 5A is the amorphous phase which was removed during etching. Interestingly, no “skin-core” structure can be found in P4MP1-ref, which is greatly different from the morphology of normal injection molded semi-crystalline polymer. The suppression of the “skin-core” structure herein is of great possibility caused by the high shear rate, high mold temperature, and fast crystallization rate of P4MP1. As the melt was injection molded into the cavity, since that the thickness of the cavity is very small, so dense nuclei were induced by the shear flow at the edge of the cavity. Then the lamellae grew towards the center in a very fast way and occupied the full cavity. Therefore, the formation of the core region was suppressed. As P4MP1-ref was treated at 270°C, the previously formed oriented lamellae melt totally as proved by the DSC and WAXS results. So, spherulite was formed during the cooling process.

3.2 The stretching induced void formation

Figure 6 shows the selected optical images and the stress–strain curves of P4MP1 stretched at 30°C. The mechanical behaviors can be divided into three groups, which are hard-elastic deformation where no obvious necking can be found (T_h ≤ 236°C), ductile deformation where the sample is broken after necking (T_h = 238 and 240°C), and cold drawing (T_h = 250 and 270°C) where a stable neck is developed. The results in Figure 6 prove that the deformation behavior of P4MP1 correlates intimately with its microstructure and morphology. The hard-elastic deformation belongs to the sample with high orientation when the stretching direction is parallel with the c-axis of the crystal.^{31, 49} And the separation of lamellae by the molecular network extension within the amorphous phase contributed mainly to the hard-elastic deformation.⁵⁰ For those samples which exhibited hard-elastic behavior, no load drop can be found after the linear deformation region. Due to lamellae perfection and molecular rearrangement in the amorphous phase, the stress at the same Hencky strain was enlarged obviously and the Hencky strain at failure was decreased.^{42, 51} The sample with moderate orientation was broken shortly after necking. In these samples, both lamellae separation and lamellae/chain slip existed during necking.^{52, 53} In addition, these samples were fractured earlier than the fully homogeneous sample.

The optical images showed that the stress-whitening behavior of the sample was altered greatly as T_h was changed, implying a variation of the cavitation behavior. With increasing T_h, the transparency of the sample was gradually weakened at fracture, indicating an intensification of void formation.²⁸ Moreover, for the specimen with hard elastic deformation, a straight fracture line can be observed on the optical image. For the specimen with ductile deformation and cold-drawing, the fracture line is less straight. In addition, a large elongation can be found for the specimen with cold drawing deformation.

SAXS is powerful in detecting the structure with tens to hundreds of nanometers. Representative 2D-SAXS patterns during stretching are provided in Figure 7. Since that the density contrast between the crystalline phase and the amorphous phase is extraordinarily small in P4MP1,⁵⁴ no lamellae scattering can be found before stretching, although the lamellar structure has been observed in Figure 5. Whereas upon stretching the lamellae scattering (indicated by the green arrows in Figure 7) showed up at the Hencky strain of ca. 0.05, presenting solid evidence that in hard-elastic P4MP1 uniaxial stretching will lead to a decrease of the density in the amorphous phase. The decrease of the density in the amorphous phase is of great possibility caused by the expansion of the molecular NETWORK CONFINED WITHIN ADJACENT LAMELLAE.

To evaluate the long period (L_p), a cut was made on the 2D-SAXS pattern in the direction of the lamellar scattering, and the 1D-SAXS curve obtained from the cut as shown in Figure 8. The cut was made along the direction of lamellae scattering and the angle range of the cut (φ) is 30° as shown in Figure 7. Then L_p is calculated by , is the peak position on the 1D-SAXS curve. L_p at the Hencky strain of ca. 0.05 was increased from 23.8 nm (without heating treatment) to 39.9 nm (T_h = 228°C). As T_h was higher than 234°C, no lamellae scattering can be recognized from the pattern, which can be understood from the following aspects: (1) the lamellae scattering was redistributed in the whole azimuthal range due to orientation reduction, so the intensity in the equator direction was decreased; (2) L_p was so large that the maximum scattering moved to a position which was overlapped with the beam-stop.

In the 2D-SAXS pattern of P4MP1-ref, as the Hencky strain reached 0.137 a strong streak scattering showed up after the appearance of the lamellae scattering, as a result of void formation. The streak is along the horizontal direction, suggesting that the longitude axis of the void is perpendicular to the stretching direction. At a larger Hencky strain, the scattering in the horizontal direction is weakened, but a weak scattering rises in the vertical direction. This suggests that a new group of voids with their longitude direction parallel with the stretching direction were formed. The appearance of the new group of voids can be caused by the direction transition of initially formed voids,^{55, 56} or the formation of new voids directly. Different from P4MP1-ref, the sample with T_h of 150°C was broken at a Hencky strain of 0.135, so no voids direction transverse was observed. As T_h was increased to 234°C, the voids scattering became much stronger implying that more voids were formed. In the sample with T_h of 240 and 270°C, the horizontal void scattering can be found at a smaller Hencky strain, and the scattering intensity was quite strong. In addition, due to the occurrence of yielding and cold drawing, in the sample with T_h of 270°C the void scattering was transformed to a strong vertical scattering at the final stage of stretching.

Scattering invariant (Q) is related to the volume fraction ( ) of voids and the electron density contrast between voids and the matrix by

()

with and are the volume fraction of the matrix and voids, and are the electron density of the matrix and voids. Equation (5) can be rewritten as

()

So Q is calculated to quantitively assess the void fraction evolution during stretching by the following equation

()

The evolution of Q as a function of the Hencky strain is plotted in Figure 9. Heating treatment at a temperature higher than 223°C could greatly enlarge the Q value during stretching, suggesting that the heating treatment could promote the formation of voids. And at the same Hencky strain, Q was largest as T_h was 238°C. In the sample with T_h of 250 and 270°C where cold-drawing happened, one can see a slight decrease on the plot after Q reached the maximum value. The slight decrease is probably caused by the coalescence of the void.³³

The size of the void can be obtained using the Ruland streak method⁵⁷ given below

()

where B_obs is the integral breadth of the peak on the azimuthal scan of the streak (see Figure 10A) considering that the orientation distribution of voids can be described well by a Lorentzian model. L_void is the length of the void, and is the misorientation of the void. The results in Figure 10B show that in the sample with T_h of 270°C, L_void increased gradually from 190 to 220 nm as the Hencky strain was increased from 0.4 to 0.8. Meanwhile, the misorientation of the void is decreased suggesting an enhanced orientation of the void.

As well known, the SAXS measurements provide information of the entity in the reciprocal space. To check the morphology of voids in real space SEM and AFM were performed on P4MP1-270 after stretching (Figure 11). The stretching direction is horizontal. It can be found that voids are distributed homogeneously in the whole scope of the image. In the AFM image given in Figure 11B, more details about the voids can be observed. Generally, two types of voids can be distinguished: which are voids containing fibrils (type I) inside and voids containing nothing (type II). The voids containing fibrils is formed by the coalescence of small voids, and the fibrils in the voids are caused by the stretching of the boundary between these voids. The voids containing nothing is a result of the growth of nascent small voids.

The size of the voids in real space is analyzed statistically and the results are given in Figure 12. The statistical values of the voids length and voids width are 382 and 81 nm, respectively. The size of the voids presented here is larger than that calculated from SAXS results. The discrepancy between them may be caused by the different detecting resolutions and detecting areas of these two methods. During SAXS measurement, the beam size was set at 500 × 500 μm² which covers a region much larger than the scope of SEM. Additionally, the etching procedure used in SEM may also contribute to the enlargement of the void size.

3.3 The deformation of the crystalline phase

Representative 2D-WAXS patterns during stretching are provided in Figure 13. The diffractions on the pattern are assigned to (001), (200), (220), (3 1), (2 2), (3 1), (4 1), and (420) crystal planes of Form I from inner to outer region. In P4MP1-ref, before stretching (001) diffraction is focused on the equator due to crystal orientation. Upon stretching, (001) diffraction disappeared at the Hencky strain of 0.15 as a result of the orientation reduction. The orientation reduction can be also found from the evolution of the azimuthal intensity distribution of (200) diffraction. The broader (200) diffraction at the final stage of stretching indicating that the orientation was reduced after stretching. Similar to that of P4MP1-ref, the orientation of P4MP1-150 and P4MP1-234 also exhibits a decreasing trend upon stretching. However, in P4MP1-240 the orientation was varied very slightly. In P4MP1-270, the crystal orientation was enhanced by stretching where the (200) diffraction was changed from a circle to an arc.

The quantitative analysis of the c-axis orientation during stretching is summarized in Figure 14A. To be more specific, in P4MP1-ref f_H decreased slightly from 0.49 to 0.44, as a result of lamellae fragmentation and melting. For the sample with T_h higher than 150°C and lower than 236°C, the evolution trend of f_H is similar with that of P4MP1-ref. In P4MP1-238 f_H was decreased from 0.25 to 0.21. Whereas in P4MP1-270, f_H grew gradually from 0 to 0.32 after stretching. The increase of f_H in P4MP1-270 was caused by the combination effect of lamellae rotation and stretching induced crystallization.⁵⁸

In addition to the orientation evolution, the crystal strain on (200) crystal plane (ε₍₂₀₀₎) was also calculated by

()

()

with is the initial d-spacing of (200) crystal plane, is the d-spacing during stretching, is the wavelength, and is the Braggs angle of (200) on the 1D-WAXS curve. The results presented in Figure 14B showed that (200) crystal plane suffered compression stress upon stretching due to the contraction that occurred in the transverse direction of the c-axis. Then, the compression was saturated very fast at the Hencky strain of 0.06, suggesting that the affine deformation happened only in a very small Hencky strain range. As the Hencky strain exceeded 0.06, the macroscopic stretching of the sample was completed mainly by the extension of the molecular network in the amorphous phase, the crystal slip, lamellae slip, and melting-recrystallization. In P4MP1-250 and P4MP1-270, the fibrillation process that occurred during cold-drawing gave rise to the increase of f_H.

The crystal size ( ) along the normal direction of (200) crystal plane was evaluated by Scherrer equation

()

where is the half-width of (200) diffraction on the 1D-WAXS curve. As one can see in Figure 14C, although the orientation was greatly changed by the heating treatment, the evolution of was influenced negligibly. in all specimens was reduced upon stretching. In P4MP1-270, was reduced gradually from 32 to 15 nm when the Hencky strain reached 0.6 as a result of chain slip and lamellae fragmentation. Afterward, was kept constantly at 15 nm until fracture. The crystallinity is related to the fraction of the crystalline phase. Therefore, the evolution of the crystallinity in Figure 14D provides a broad view of the deformation of the crystalline phase. In general, the crystallinity was reduced as the Hencky strain was increased. In P4MP1-ref, the crystallinity was decreased from 0.38 to 0.32 in the whole stretching process. The variation of the crystallinity is much larger than the detection resolution of WAXS (0.005–0.01),⁵⁹ suggesting that the stretching led to a fragmentation and melting of lamellae. In the sample showing hard elastic behavior, the evolution trend was similar with that of P4MP1-ref although the crystallinity was larger. In the sample exhibiting ductile deformation and cold drawing, the crystallinity was decreased much faster. For instance, at the Hencky strain of 0.31 the crystallinity was decreased to 0.33 in P4MP1-ref and 0.22 in P4MP1-270, respectively.

3.4 Final discussion

A summary is made here necessary before the final discussion. By injection molding, P4MP1 plates with oriented morphology were obtained. In addition, the traditional “skin-core” structure was suppressed in the injection-molded plates. Subsequently, the microstructure and morphology of the plates were adjusted by the designed heating treatment. The mechanical behavior of the final plates was categorized into three groups which are hard elastic deformation, ductile deformation, and cold drawing deformation. The discussion below will be focused on the hard elastic deformation and cold drawing, since that the ductile deformation is more like a combination of them.

For the specimens exhibiting hard elastic deformation, the volume fraction of the void was increased as T_h was elevated. Before the initiation of the void, a density decrease in the amorphous phase was observed. In addition, the longitude direction of voids was perpendicular to the stretching direction when voids were initiated. Before the specimen was fractured, only a few voids transferred its longitude direction to the stretching direction.

For the specimens showing cold drawing deformation, the volume fraction of the void was in the same order as that of the highest one of the specimens showing hard elastic deformation. Once the voids were formed their longitude direction was perpendicular to the stretching direction. After yielding, a new group of voids appeared with their longitude direction aligned in the stretching direction. Upon further stretching, the length of the voids was increased and the orientation of the voids was enhanced along the stretching direction. The average length and width of the voids obtained from the SEM image were 382 and 81 nm.