Unique impact behavior and toughening mechanism of the polypropylene and poly(ethylene-co-octene) alternating multilayered blends with superior toughness

Abstract

In this study, a novel approach is proposed to significantly toughen polypropylene (PP) by fabricating PP and poly(ethylene-co-octene) (POE) into alternating multilayered blends instead of conventional blends. POM, SEM, polarized-FTIR, DSC and XRD were performed to investigate and characterize the microstructure of the alternating multilayered and conventional blends. The crack-initiation term, impact fracture surface and bulk morphologies beneath the fracture surface are inspected in order to understand the differences in the impact behaviors of the alternating multilayered blends and the conventional blends. The results show that the unique multilayered structure has a great advantage in toughening PP. The notable improvement of the toughness of the alternating multilayered blends is ascribed to the synergetic effects of the interfaces' delaminations, craze deflection, larger subcritical damage zone (stress whitening zone) and the combination of the voids and deformation during the fracture process. Moreover, the alternating multilayered blends exhibit high toughness with a low POE content; thus, this work also offers a new method to toughen the materials without an obvious sacrifice of their strength.

---

1. Introduction

Polypropylene (PP), one of the most widely used commodity plastics, has been the subject of intensive studies with the objective to improve its mechanical properties, especial impact toughness. One of the most effective methods is blending PP with elastomers.¹ Unfortunately, in such a manner, PP blends with high toughness usually are accompanied with obvious drops in strength, and the cost also increases because of the high price of the elastomers. Therefore, how to balance the strength and the toughness of PP/elastomers blends is highly desirable for many industrial applications. In general, the addition of rigid particles into polymers usually leads to dramatic drops in fracture toughness, compared to their corresponding pristine polymers. Only a few studies have reported the enhancement of both toughness and strength.^2–5 However, the extent of toughness enhancement, if at all, is not always significant. Thus, several studies^6–9 have been conducted on the polymer/elastomer/rigid filler ternary system. It has been determined that a good dispersion of the rigid filler or elastomer plays a key role for obtaining the best combination of mechanical properties. In order to obtain a high impact toughness in the polymer/elastomer/rigid filler ternary system, compatibilizers and surfactants are often used, which indicates tedious fabrication processes and the incorporation of costly steps. In addition, more or less drops in the flow-ability of the polymer melts always occur, resulting in processing difficulty. Therefore, it is highly desirable for achieving a high toughness of the PP/elastomer composites without obvious sacrifice of the strength.

On the other hand, the design of the composites inspired by the hierarchical micro-structure, which is found in nature's load-bearing materials, is an efficient route to obtain materials with a unique combination of strength and toughness.^10,11 Bio-composites such as wood, spider silk, bone, tendon and nacre are characterized by ordered structures combining large fractions of hard and reinforcing segments with a small amounts of soft, energy-absorbing and lubricating biopolymer. Inspired by such biological architectures, several advanced technologies, such as in electrical fields,¹² tape-casting,¹³ layer-by-layer,¹⁴ and gel casting combined with hot-pressing,^15,16 have been developed for preparing novel composites with superior mechanical properties, despite these approaches are often limited to fabricating thin films and/or require multiple processing steps.¹⁷ However, in past decades, multilayered micro-co-extrusion¹⁸ was developed as a novel technology in polymer melt processing to tune the material properties, such as barrier,¹⁹ optical,²⁰ electrical,²¹ and especially mechanical properties.^22–32

Baer et al.^22–27 investigated the crazing behaviors of co-extruded multilayered sheets of polycarbonate (PC) and styrene–acrylonitrile copolymer (SAN) alternating layers. They observed a shift in the deformation mode from craze opening to shear yielding as the individual layer was thin enough, which dramatically toughened the multilayered PC/SAN composites. Similarly, the brittle-to-ductile transition of the PS/SEPS multilayered films was also determined when the thickness of the SEPS layer was thin enough.^28,29 Except the crazing behaviors under quasi-static uniaxial tension, PC/SAN and PC/PMMA multilayered sheets were also examined in dynamic ballistic tests.^27,30 With increasing the layer number, the multilayered sheets fractured with more profuse cracking, delaminations, and emergence of a circular impression, indicating that more impact energy was absorbed, and finally the projectile did not penetrate the specimens. Shen et al.³¹ simulated the mechanical properties of the PP/POE multilayered blends by equivalent box modeling and found that the yield strength of multilayered blends was higher than that of the conventional blends, which can be explained well by the high phase continuity of the multilayered blends. Herein, it can be concluded that combining two polymers with alternating multilayered architecture indeed endows the materials with excellent mechanical properties. However, one may note that all these excellent properties were determined from sheets or films, as the thickness of multilayered co-extrusion samples is usually lower than 1.5 mm. Because the polymers are viscoelastic, their toughness generally exhibits geometry and strain-rate dependence. Moreover, most materials are used as load-bearing bulk materials with different shapes and sizes. Consequently, it is necessary and significant to evaluate the toughness of multilayered materials in different test standards. In our previous work,³² we fabricated multilayered materials by alternating the PP and the PP/POE blend, which indicates an alternating distribution of the POE particles in the PP matrix, and then the multilayered sheets were hot-pressed into impact bars for the Izod test. Most interestingly, compared with the random distribution of POE particles, the unique alternating distribution of POE particles endowed the materials with a great enhancement of toughness at −40 °C. Damage is prone to occurring at the interfaces, and a large number of interfaces is a great advantage in accumulating micro-cracks, delaminations and crack deflection, all of which will obviously toughen the materials.^33–35 Therefore, we speculate that there should be a more striking enhancement of toughness for the PP/POE alternating multilayered blends where the POE phase is distributed with a laminar morphology.

In this work, we attempt to imitate the micro-structures of nacre partly by introducing the alternating multilayered structure into the PP toughening system, in which higher fractions of “stiff” PP layers and lower fractions of “soft” POE layers alternate through multilayered micro-co-extrusion technology. The fracture behaviors of the alternating multilayered and conventional blends are investigated. The unique alternating multilayered micro-structure with PP layers and POE layers, as well as large amounts of weak PP/POE interfaces will obviously toughen the multilayered blends without an obvious drop in strength.

2. Experimental

2.1 Materials

Polypropylene is of grade 1300 with a MFI of 2.0 g per 10 min at 230 °C, 2.16 kg, supplied by Mao Ming Petro-chemistry Co. The POE is Engage 8100 manufactured by Dow Chemical, with an octane-co-monomer molar content of 9.8% and a MFI of 1.0 g per 10 min, at 190 °C, 2.16 kg. Their densities are 0.9029 and 0.8785 g cm⁻³ (measured by MatsuHaku, GH-120M), respectively.

2.2 Sample preparation

PP and POE were dried in an oven at 85 °C and 50 °C for 12 h prior to the processing, respectively. Each stratified sample was co-extruded as a sheet of about 1.3 mm thick and 30 mm wide through the multilayered micro-co-extrusion system designed by our lab, the schematic of which is illustrated in Fig. 1. PP melt and POE melt were simultaneously extruded from two different extruders and combined as 2-layer melt in the co-extrusion block, and then the 2-layer melt flowed through a series of layer multiplying elements (LMEs). In a LME, the melt was sliced into two left and right sections by a divider, and then vertically recombined. An assembly of n LMEs could produce a multilayered blend with 2⁽ⁿ⁺¹⁾ layers. In this work, 8-, 32- and 128-layer samples were extruded with 2, 4, and 6 LMEs, respectively. The multilayered sheets with different POE volume content were produced by controlling the PP/POE feeding ratio. In the processing, the extruder temperature was 160–180–190–200–200 °C for PP section and 90–180–190–200–200 °C for POE section, respectively. In comparison, conventional PP/POE blends were prepared by mixing equal mass fractions of 8-, 32- and 128-layer multilayered samples with the same PP/POE feeding ratio. The mixing condition was 30 rpm min⁻¹ for 8 min at 200 °C. Neat PP samples were also prepared under the same shearing force history using one of the extruders in the multilayered co-extrusion system with 0, 2, 4, and 6 LMEs. The average thickness of the PP or POE layer was calculated by averaging the individual PP or POE layer thickness of the co-extrusion sheets, which was obtained by measuring the POM micrograph of the multilayered samples with Image-Pro-Plus software. Table 1 shows the sample code, total sheet thickness, average PP or POE layer thickness, and total POE volume content in the multilayered and conventional blends, which could be calculated from the following formula:where, V_POE is the volume fraction of POE; ρ_B is the density of the blends; ρ_PP and ρ_POE represent the densities of neat PP and POE, respectively. The densities of ρ_B, ρ_PP and ρ_POE were measured through a high precision density tester (MatsuHaku, GH-120M). The blends are coded according to the abbreviation for alternating multilayered blends “A”, the number of the layer and the volume content of POE. For example, A2-128 represents the 128 layers multilayered blend with 16.21 vol% POE. As for the conventional blends and neat PP, the abbreviation is “C”.

Fig. 1 Sketch of multilayered co-extrusion technology: (A and B) single screw extruder; (C) connector; (D) layer multiplying element (LME); (E) die.

Table 1 Sample code, total sheet thickness, average thickness of PP and POE layers, and total POE content of alternating multilayered and conventional blends

Sample code		Total sheet thickness (μm)	Average PP layers thickness (μm)	Average POE layers thickness (μm)	POE content (vol%)
Neat PP	C0, C0-8, C0-32, C0-128	1280	—	—	0
Conventional blends	C1	—	—	—	6.79
	C2	—	—	—	16.57
Alternating multilayered blends	A1-8	1340	311.2	23.1	6.80
	A1-32	1330	77.3	5.5	6.91
	A1-128	1250	18.2	1.3	6.71
	A2-8	1390	289.4	57.3	16.53
	A2-32	1360	70.4	14.8	16.98
	A2-128	1225	16.6	3.4	16.21

---

2.3 Molecular orientation, crystalline characterization

To eliminate the effects of molecular orientation and crystalline differences on the testing and the remove shearing history, before mechanical test, except some comparison samples, all the samples were put into a compress machine with set condition, which is under 10 MPa for 10 min at 200 °C (for imitating the hot-press condition) (more information can be obtained from the ESI†).

2.4 Mechanical properties tests

2.4.1 Notched Izod impact test. The notched Izod impact strength of the specimens was measured with a XJU-22 Izod machine according to GB/1943–2007. In order to clearly understand the impact behaviors, two types of layer arrangements of 80 × 10 × 4 mm thick impact bars were prepared. The alternating multilayered sheets were first cut into 80 × 15 × 1.3 mm thick sheets and followed by hot-pressing into 80 × 15 × 4 mm thick, 80 × 15 × 10 mm thick, respectively, the hot-pressed condition is under 10 MPa for 10 min at 200 °C. As shown in Fig. 2, the 80 × 15 × 4 mm thick bars were polished into 80 × 10 × 4 mm thick for the impact parallel to the layer plane (Fig. 2(B)); the 80 × 15 × 10 mm thick bars were polished into 80 × 10 × 4 mm thick for the impact vertical to the layer plane (Fig. 2(A)). Then, the suffixes the “P” and “V” were added to the obtained samples, which indicate that the impact is parallel and vertical to the layer plane, respectively. As for the neat PP and conventional blends, 80 × 10 × 4 mm thick bars were directly hot-pressed for the impact test (Fig. 2(C)). The depth of the notch is 2 mm. Testing was carried out at ambient temperature (23 °C). Each impact test included at least 5 parallel experiments, and the results were averaged.

Fig. 2 Schematic representation of impact test through different directions of alternating multilayered and conventional blends and the position for the SEM. (A) Alternating multilayered blends impact vertical to the layer plane; (B) alternating multilayered blends impact parallel to the layer plane; and (C) conventional blends and neat PP. ((1) Impact direction; (2) guide chute; (3) fixed clamp). (a)–(c) Illustration for SEM position.

2.4.2 No-notched impact tensile test. As shown in Fig. 3, the testing geometry was cut from the centre of the heat treated samples. The impact of the tensile of the samples was carried out by an instrument XJ-50 (JJ-TEST, China) impact pendulum machine, according to the GB/T13525. The impact velocity was 3.8 m s⁻¹. The impact energy was 15 J. Testing was carried out at ambient temperature (23 °C). Each impact test included at least 8 parallel experiments, and the results were averaged.

Fig. 3 Schematic representation of impact tensile test.

2.4.3 Tensile test. The standard dog-bone tensile sheets were cut from the centre of the heat treated samples. Tensile tests were carried out by the extrusion direction at 10 mm min⁻¹. At least five parallel samples were tested in each test.

2.5 Polarized optical microscopy (POM) observation

To examine the structure morphologies of multilayered with different layers, a rotary microtome (YD-2508B) was used to cut a 20 μm slice from the multilayered sheets along the transverse direction. The sample slice was placed between two glass slices, and then inspected on a polarized optical microscope (POM, Leica, DM2500P).

In order to understand the fracture and toughening mechanism of multilayered and conventional blends, the crack initiation and propagation stages were observed through a part-impact test, which was performed with the XJU-22 impact test machine. The pendulum was raised at an angle of 30° from the vertically fixed specimen, and then released to hit the specimen with appropriately constant impact energy of about 0.3 J. The specimen was not broken into two halves as expected, and the propagating crack or craze stopped in the interior of the specimen. The initiation and propagation patterns of crack or craze were collected by transmission optical microscopy (TOM), with the 20 μm sample slices cut from part-impact specimens along the crack propagation direction, but parallel to the impact direction. The images collected by TOM were all recorded with a Pixelink camera (PL-A662).

2.6 Scanning electron microscopy (SEM)

The impact fracture behaviors of the samples were also investigated by SEM (JSM-5900LV, Japan). The impact-fracture surfaces were obtained from the notched Izod impact testing. To observe the sub-damage zone (stress whitening zone) underneath the impact-fracture surfaces, the impacted samples were cryogenically polished along the extrusion direction, but perpendicular to the layer plane with a tungsten knife at −110 °C using the Leica RM2265 microtome (as shown in Fig. 2(a)–(c)). The impact tensile samples were also polished in the same manner. Before SEM characterization, all the surfaces were sputter-coated with a gold layer.

3. Results and discussion

3.1 Phase morphology

The most significant difference between the multilayered and conventional blends with the same POE content can be understood reasonably as follows: through the multilayered co-extrusion technology, the alternating multilayered microstructure of the PP layers and the POE layers was successfully fabricated, which can be well observed through the POM images presented in Fig. 4. The macro-anisotropy in the multilayered samples is very obvious and the continuity of each layer is very good. The darker layers in the images belong to the amorphous POE layers, whereas the brighter layers belong to the crystalline PP layers. Moreover, all of the multilayered samples have a clear laminar morphology, in which the PP and POE layers align alternately vertical to the interfaces and continuously parallel to the extrusion direction. It should be noted that the laminar morphology of the A1-128 does not appears very clear. It is attributed that the thickness of POE layers in A1-128 is too low to be distinguished through the POM, although its continuously laminar morphology can be clearly demonstrated through the SEM of its impact fracture surface (Fig. 7–9). As for the conventional blends (C1 and C2), the addition of POE appears to decrease the sizes of the PP spherulites. Given that the unique micro-structure of the multilayered samples can significantly influence the PP/POE toughening systems, the investigations from macroscopic perspective and microscopic perspective will be conducted in detail later.

Fig. 4 Polarized optical micrographs of the conventional and multilayered blends morphologies.

3.2 Crystalline structure and molecular orientation

The multilayered PP/POE blends with different layer numbers can be fabricated with different LMEs. More LMEs indicates considerably stronger shearing force that the multilayered blends will suffer from, resulting in distinct differences of crystalline structure and molecular orientation, which will have evident effects on the toughness of the PP/POE system.^36,37 As the major goal in this work is to investigate whether the alternating multilayered micro-structure can toughen the PP/POE blends; therefore, it is necessary to eliminate the shear and thermal history of the samples. Just as shown in the ESI,† the thermal treatment condition, which was under 10 MPa for 10 min at 200 °C, indeed removed the shear and thermal history.

3.3 Mechanical testing

As indicated in Fig. 5(a), the notched Izod impact strength of the PP/POE multilayered blends increases with increasing their layer number. Moreover, the Izod values, which are measured vertical to the layer plane direction, are higher compared to the ones that are measured parallel to the layer plane direction. There is no obvious variation of Izod values for the neat PP (C0) (about 31.1 kJ m⁻²) with increasing its layer number. In detail, the impact strengths of alternating multilayered blends A1-128-V (91.1 kJ m⁻²), A1-32-V (76.9 kJ m⁻²), and A1-8-V (69.6 kJ m⁻²) are 2, 1.7 and 1.5 times as high as that of their corresponding conventional blend C1 (46.5 kJ m⁻²), respectively, although the impact strengths of A1-128-P (69.6 kJ m⁻²), A1-32-P (51.5 kJ m⁻²), A1-8-P (44.1 kJ m⁻²) are only 1.5, 1.1 and 0.95 times as high as that of C1 (46.5 kJ m⁻²), respectively. All these results indicate that it does not matter whether parallel or vertical to the layer plane, the multilayered blends with low POE content (6.79 vol%) and high layer number exhibit higher toughness than their conventional blend. Whereas for the high POE content (16.57 vol%) multilayered system, although the impact strengths slightly increases by increasing their layer number, except A2-128-V, the Izod values of the others are slightly lower compared to that of their corresponding conventional blend. As for the no-notched impact tensile strength, the impact tensile strengths of multilayered blends obviously enhance by increasing their layer number (Fig. 5(b)). For the low POE content system, the impact tensile strengths of alternating multilayered blends A1-128 (781.3 kJ m⁻²), A1-32 (442.0 kJ m⁻²), and A1-8 (381.8 kJ m⁻²) are about 1.8, 1 and 0.86 times as high as that of their conventional blend C1 (443.8 kJ m⁻²), respectively. For the high POE content system, the impact tensile strengths of alternating multilayered blends A2-128 (855.6 kJ m⁻²), A2-32 (666.2 kJ m⁻²), and A2-8 (411.7 kJ m⁻²) are only about 1.2, 0.94 and 0.58 times as high as that of their conventional blend C2 (712.0 kJ m⁻²), respectively. The neat PP with different layers presents a stable value of about 520 kJ m⁻². Interestingly, compared with C1, A1-8, A1-32 and A2-8, neat PP with different layers exhibits a high impact tensile value. The multilayered blends present higher toughness compared to their corresponding conventional blends, and the Izod values of the multilayered blends with high POE content are lower than those with low POE content. Therefore, the corresponding toughening mechanisms should be discussed in detail.

Fig. 5 Mechanical properties of the alternating multilayered, conventional blends and neat PP. (a) notched Izod impact strength through different directions; (b) no-notched impact tensile strength; (c) tensile yield strength.

On the other hand, the differences of the tensile yield strength of neat PP and the multilayered blends with different layers are little (Fig. 5(c)). The tensile yield strengths of the multilayered blends, in particular the high POE content system, are higher than those of their corresponding conventional blends, which is consistent with our previous work.³² According to the equivalent box model, the yield strength of a dispersed, co-continuous structure would be lower than those with multilayered structures.^38,39

3.4 Fracture and toughening mechanisms

The images of Izod samples taken after the notched impact test often obtain some macroscopic information to understand the toughening mechanisms. As shown in the Fig. 6(a), the neat PP (C0) and conventional blends (C1 and C2) exhibit a hinged breakage and a semi-circular stress whitening zone. As for the multilayered blends, when measured parallel to the layer plane, the semi-circular stress whitening zone is also observed, while only the A2-8-P and all low POE content systems (A1) exhibit obviously hinged breakage. When the samples are measured vertical to the layer plane, except a large rhombus-shaped stress whitening zone, no obvious breakage can be found. In order to analyze the differences of the stress whitening zone between conventional and the multilayered blends, some of the images shown in the Fig. 6(b) are treated by the Image-Pro-Plus software with the Invert Contrast Mode, in which the black zone corresponds to the stress whitening zone in the Fig. 6(a). Stress whitening, the tendency of polymer materials to display a white appearance under imposed stress, is attributed to the scattering entities or localized stress concentration sites scatter light. For PP and PP/elastomer blends, stress whitening is caused by the crazes and micro-voids during the deformation process.^40,41

Fig. 6 Images of Izod samples taken after the notched impact test ((a) normal images; (b) typical images treated by the Image-Pro-Plus software with the Invert Contrast Mode).

The total energy of fracture can be partitioned into two components; (i) one is consumed to create new fracture surface, and (ii) the other is dissipated in the stress whitening zone (bulk damage) for the deformation of matrix or dispersed phase. Therefore, most of the fracture energy must be consumed through the bulk damage for the fracture without obvious breakage. Consequently, only the impact fracture surface morphologies of the samples with obvious breakage are inspected by SEM. Fig. 7–9 show the representative impact fracture surface morphologies of C0, C1, A1-8-P and A1-128-P. As shown in the Fig. 7 and 8, the micrographs clearly illustrate three distinct types of fracture morphologies: (a) a relatively smooth zone “A” in the vicinity of the origin or primary crack initiation sites; (b) the coarse zone “B” with several strip-like protrusions and pleats arranging vertical to the impact direction, which represent the plastic deformation zone, and (c) unbroken part zone “C” shown by the rectangular-shaped dashed lines. Most interestingly, the fracture morphologies of the multilayered blends, particularly for the A1-128-P, are characterized by amounts of delaminations of the PP/POE interfaces. It should be noted that the relatively smooth region in the left of the unbroken part zone “C” of the A1-128-P was cut through with a blade. Generally, good toughness materials cannot break completely during the impact test. Even with the C0, there is still some small unbroken zone. Moreover, though the POE content is the same, the area of unbroken zone increased by the sequence of A1-8-P, C0, C1 and A1-128-P, which is roughly consistent with their Izod values. In addition, as shown in the high magnification micrograph Fig. 8, there is almost no difference between C0 and A1-8-P (taken only from PP layer) in either the A zone, showing surface characters without any obvious plastic deformation and micro-voids, or the B zone, showing the local deformation of the matrix characters with large amounts of protrusions and pleats arranging perpendicular to the impact direction. In comparison with the C0 and A1-8-P, the visualized surface morphologies of the conventional blend C1 are a little different. First, the A zone becomes considerably coarser, and small scale local matrix deformation is found; second, one can notice that more visible protrusions and pleats are formed in B zone, which means considerably greater local deformation of the matrix, while the voids may still not be observed. In addition, the fracture surface of the PP layers in A1-128-P totally reveals different morphologies. The surface, particularly in A zone, is very smooth, and unexpectedly, there is no obvious local deformation in B zone. More interestingly, although the interfaces exist between PP and POE phases for all blends, the delaminations of interfaces are only observed in multilayered blends. It is been proven that the delaminations of interfaces can dramatically enhance the toughness of materials.^5,42,43 Consequently, it is highly desirable to inspect the delaminations of the interfaces. As shown in Fig. 9, for the A1-8-P, interfaces delaminations are found in both the A and the B zone, whereas in the A zone, there is nearly no deformation for the POE layers, while in the B zone, the POE layers are deformed into asperities and slightly crumple. In contrast, for the A1-128-P, besides the delaminations between PP and POE layers, the POE layers are torn into fibrils in B zone and into pellets in A zone. The fibrils or pellets are named “ligaments” and these ligaments bridge across the adjacent PP layers. As mentioned in the introduction section, one of the important explanations for the superior toughness of nacre is the energy-dissipating of the 5 vol% fraction of the organic phase (just similar to the POE phase in PP/POE multilayered blends) by the formation of organic ligaments between platelets.⁴⁴ In addition, the ligaments are also found in the PLA/PBSA system with a compatibilizer, which finally dramatically toughen the PLA/PBS system.⁴³ Therefore, the observation of amounts of ligaments in A1-128-P can partially account for its higher toughness. Moreover, the crumple deformation of PP layers in A1-128-P can also consume energy, which may be another contributor to its high toughness.

Fig. 7 SEM images of the impact fracture surface of neat PP (C0), conventional blend (C1) and the multilayered blends (A1-8-P and A2-128-P) at low magnification. (A and B) Indicate the crack initiation zone and crack propagation zone, respectively. Zone C marked by rectangular-shaped dashed line represents the unbroken part of the samples during the impact fracture process.

Fig. 8 SEM images of the impact fracture surface of neat PP (C0), conventional blend (C1) and the multilayered blends (A1-8-P and A2-128-P) at high magnification. The images were obtained from the different zones shown in Fig. 7. For the A1-8-P, the images were taken from the PP layers. The scale bars represent 20 μm.

Fig. 9 SEM images of the interfaces delaminations from the impact fracture surface of the multilayered blends (A1-8-P and A2-128-P) at high magnification. The images were obtained from the different zones shown in Fig. 7. The scale bars represent 50 μm.

In order to get more in-depth evidence for understanding the toughening mechanisms, the cross sections beneath the impact fracture surface of the samples are also observed with SEM. As shown in Fig. 10, the locations, where the micrographs were obtained, are the zone underneath the impact fracture surface, 50 μm, 500 μm and 5000 μm away and labeled 1, 2 and 3, respectively (as shown by Fig. 2(a)–(c)). For the C0, slight shear yielding and few voids are visualized at least 50 μm beneath the fracture surface (C0-1), at the distant 500 μm away, the shear yielding is inconspicuous while the voids still can be found (C0-2). Moving a farther distance leads to the region absolutely unaffected by the fracture surface (C0-3). For the conventional blend C1, moderate intensive shear yielding and voids are observed at least 500 μm (C1-1, C1-2). Both the shear yielding and voids are not apparent in C1-3. As for the high POE content C2, both the intensive shear shielding and extensive elongated voids with bigger sizes are visualized in C2-1. Compared with C2-1, slightly intensive shear shielding and less elongated voids are found in C2-2. Moving farther from the fracture surface also leads to the region affected slightly (C2-3). It's worthy of noting that the small and regular dark spheres in C2-3 are POE particles instead of voids. Generally, under impact stress, the POE particles act as stress concentrator and cavitation sites. After cavitation, the triaxial stress disappears and the matrix behaves as if it were under plane-stress conditions, where more shear yielding readily occurs. Additionally, the voids created by the POE particles further act as the stress concentrator.⁴⁵ Therefore, the addition of POE toughens PP by accelerating the formation of voids and shear yielding, which can be shown well by the Fig. 10 (C0-1–C1-3). In particular, when the POE content is higher than a critical value, the distance between two neighboring POE particles is smaller than the critical matrix ligament thickness. In such a case, the overlap of the adjacent stress fields will first initiate the local shear yielding in the PP matrix, and subsequently result in the deformation of the PP matrix, just shown as the Fig. 14 (C2-1, C2-2). The cooperative motion of PP matrix and POE phase consumes large amounts of energy,^46,47 which finally dramatically toughens the PP.

Fig. 10 SEM micrographs of the cross sections underneath the impact fracture surface of the neat PP (C0), conventional blends (C1 and C2) and multilayered blends (A1-128-V, A1-128-P, and A2-128-V, A2-128-P). The scale bars represent 20 μm. The location, from where the micrographs were taken, is the zone under the impact fractured surface, 50 μm, 500 μm and 5000 μm away and labeled 1, 2 and 3, respectively.

As for the multilayered blends A1 and A2, extraordinary and versatile bulk morphologies are found. In the case that the impact vertical to the layer plane, in the zone 50 μm away beneath the fracture surface, a relatively smooth surface with only a few scattered voids occurs (A1-128-V-1), which is caused by the adiabatic heating process generated during the fracture,^5,48 while in the area farther away from the fracture surface (500 μm away for A1-128-V-2), the shear yielding becomes obvious; moreover, both the sizes and density of the voids dramatically increases. For the farthest area 5000 μm away (A1-128-V-3), the surface with moderate shear shielding and voids is still visible. In comparison, an overall deformation of the PP layers accompanied by the interfaces delaminations can be observed well in A2-128-V-1. Sea-island-like surface caused by the local deformation of the PP layers and strip-shaped smooth surface are observed in A2-128-V-2. A considerably thicker strip-shaped smooth surface compared to that in A2-128-V-1, this strip-shaped smooth surface may not only include the POE layers, but also the partial PP layers that were close to the POE layers. Neither the shear yielding nor the voids are obvious in A2-128-V-3. As the deformation is an adiabatic process during the impact test, considerate stress can be released, which finally results in a relaxation zone with fewer voids and deformation in the micrographs (Fig. 11). Moreover, the occurrence of the relaxation zone is accompanied by an increase in fracture energy, which may result from its crack blunting effects.^48,49 Interestingly, the relaxation zone is only present in A1-128-V and A1-128-P, which partially accounts for the high impact strength of the A1-128-V. On the other hand, the formation of voids can obviously toughen the materials.^50–52 Additionally, the formation of the voids will consume more energy in homogeneous materials rather than heterogeneous materials as the homogeneous materials usually have higher surface energy. Several voids are observed in the homogeneous PP layers and spread over at least 5000 μm away under the fracture surface (except for the relaxation zone), which is an important contributor for the superior toughness of A1-128-V. Moreover, the large shear yielding zone may be an another explanation for its high toughness. As for the A2-128-V, although the overall deformation and delaminations are observed near the fracture surface, the severity of the deformation sharply decreases with the distance. Additionally, the voids are inconspicuous. In other words, it can be deduced that the lower toughness of A2-128-V compared to A1-128-V results from its smaller subcritical damage region, the absence of a relaxation zone and the inability to form plenty of voids.

Fig. 11 SEM micrographs of the cross sections underneath the impact fracture surfaces of the A1-128-V and the A1-128-P, the fracture surface is located at the left. The dashed zone presents the relaxation zone, which formed during the impact process.

Versatile bulk morphologies are also found on the condition where the impact direction is parallel to the layer plane. One can observe a peculiar stalactite lava-like morphology with intensive voids in A1-128-P-1 (Fig. 10). Interestingly, this morphology appears to be in partial relaxation or melting state, which is caused by the adiabatic process during the impact tests (Fig. 11). Obvious shear yielding and voids are observed in A1-128-P-2. As for the farthest region A1-128-P-3, the morphology appears unaffected by the fracture surface. On the other hand, for the A2-128-P, strong shear flow of the PP and POE accompanied by large elongated voids can be readily revealed in A2-128-P-1. Compared with the A2-128-P-1, slightly shear flow is only observed in PP layers and the elongation extent of voids decreases while their sizes become bigger in A2-128-P-2. As for the A2-128-P-3, the morphology also appears unaffected by the fracture surface. Just as the above description, the higher unbroken part (as shown by the Fig. 6) and the absence of the relaxation zone in A2-128-P result in its lower Izod value than that of the A1-128-P.

In order to further ascertain the special impact behaviors of the multilayered blends, POM micrographs from the subcritical damage zone of the part-impact samples are shown in Fig. 12. For comparison, POM micrographs of the conventional blend and neat PP are also obtained. For the multilayered blends impacted vertical to the layer plane (A1-128-V), a mass of craze deflects along the PP/POE interfaces and only a handful of craze can propagate through the soft POE layers. As for the A1-8-V, only a certain amount of multiple craze is restricted in only one of the PP layers with a standard rectangular-shaped zone. Interestingly, these crazes appear to be initiated from the POE layer, which is located at the crack tip and arrested by the next adjacent POE layer. On the other hand, for the multilayered blends impacted parallel to the layer plane (A1-8-P and A1-128-P), massive craze is initiated around the crack tip, and then propagates along the impact direction. Finally, a fan-shaped craze zone is formed. Unlike the A1-128-V, craze deflection is observed not only near the crack tip, but also at the root of the crack in A1-128-P. However, one may question the rationality of the occurrence of craze deflection in the POM micrograph of A1-128-P because the slice for POM is taken parallel to the layer plane in A1-128-P. In fact, it is difficult to only take the slice from one of the layers as the thin layer in A1-128-P. Because of the thicker PP layers in the A1-8-P, the effects of the POE layers can be negligible for those craze initiation sites are far enough, which results in similar craze patterns of the C0 and the A1-8-P. Compared with the C0, large multiple crazes (much larger dark zone) are observed for the conventional blend (C1) indicating that the incorporation of POE can toughen the materials by accelerating the generation of massive craze. As craze is prone to propagating along the weakness of materials,⁵³ most of the fracture energy must be dissipated through the craze deflecting along to the weak PP/POE interfaces in A1-128-V. Besides, the existence of the soft POE layers can blunt the craze through their deformation. In simple words, the effects of the soft POE layers on the crazing behaviors of the multilayered blends are to provide deflection interfaces and blunt the craze, such that craze has more difficulty in propagating through the layers. The A1-8-V still presents high toughness despite its low layer number because the craze in A1-8-V can be perfectly arrested by the adjacent thick and soft POE layer; in addition, more craze initiation sites are formed along the interfaces, which can prevent the transformation of damage from craze to crack, and finally avoid catastrophic fracture of the multilayered materials. When the impact is parallel to the layer plane, the craze pattern of the A1-8-P is similar to that of the C0 and this can account for its poor impact strength well. Compared with the A1-8-P, craze deflection is also found in A1-128-P. Most importantly, a plenty of interfaces delaminations are found at its fracture surface (Fig. 7). All these factors act in toughening the multilayer materials and finally endow the A1-128-P with higher toughness than that of the A1-8-P and C1.

Fig. 12 Craze initiation patterns of the samples after the Izod notched part-impact test (C0: neat PP; C1: conventional blend; A1-8-P and A1-128-P: multilayered blends impacted parallel to the layer plane; A1-8-V and A1-128-V: multilayered blends impacted vertical to the layer plane).

We have to admit some of the samples did not break completely during the Izod impact test. Thus, the Izod values of those not completely broken samples cannot fully reflect their toughness. In order to comprehensively reflect the toughness of all samples, the impact tensile test was also carried out. Just as shown in Fig. 13, the fracture surface morphologies and the bulk morphologies underneath the fracture surface were also inspected by SEM. The fracture surface of the A1-128 is smooth (Fig. 13(A)). More information can be revealed through the close-up of the rectangular dashed zones in A1-128 (Fig. 13(a1) and (a2)). The surface of the a1 is rather smooth while slight delaminations and ductile tearing align parallel to the layer plane in the surface of a2. For the A1-8, smooth fracture surface is also observed (Fig. 13(B)). Although obvious delaminations are observed in the right edge of the fracture surface, most of the interfaces still remain perfect just as shown in the close-up b2. As for its PP layers, slight deformation arranges horizontally and ductile tearing arranges vertically downward (b1). Similar but more severe deformation and ductile tearing are observed in c1. In addition, smooth surface with few voids is illustrated in c2. Compared with the blends, the most distinct difference of the C0 (Fig. 13(D)) is the “necking” during the fracture process, which can mainly account for its high impact tensile value. As for its fracture surface, a coarser surface with little voids is found in d1, while smooth surface without the voids is formed in d2. Although the samples exhibit complicated fracture surfaces, it is clear that the information from the fracture surface cannot fully account for their distinct impact tensile performances. Herein, it is necessary to reveal the bulk morphologies underneath the fracture surfaces (Fig. 13(A1)–(D3)). For the A1-128, obvious delaminations and warping PP layers with few voids are observed at least 100 μm beneath the fracture surface (Fig. 13(A1)). Severe local deformation and voids are observed at 700 μm beneath the fracture surface (Fig. 13(A2)). Moving a further distance, only plenty of voids with bigger sizes are found (Fig. 13(A3)). In comparison, in A1-8, delaminations are absent and the voids are obvious at least 100 μm beneath the fracture surface (Fig. 13(B1)). Moreover, away from the fracture surface, the severity of the plastic deformation decreases, massive plastic deformation can be observed at least 100 μm beneath the surface, whereas the local plastic deformation can be observed at the distance of 700 μm and 2000 μm away underneath the fracture surface (Fig. 13(B1)–(B3)). On the other hand, for the C1, slight local deformation and voids can be seen at least 700 μm underneath the surface (Fig. 13(C1) and (C2)), moving a farther distance results in an absolutely unaffected bulk morphology by the fracture surface (Fig. 13(C3)). As for the C0, at the distance 700 μm away from the fracture surface, an apparent boundary lies between the deformation zone and the smooth zone (Fig. 13(D2)). The morphology before the boundary belongs to the necking zone with obvious deformation (Fig. 13(D1)). Beyond the necking zone, the morphology appears not to be affected during the impact test (Fig. 13(D3)). By combining the analysis of fracture surface morphologies and the bulk morphologies, it can be deduced that the higher impact tensile strength of the A1-128 results from its versatile bulk morphologies, which can be characterized with delaminations, severe local deformation, plenty of voids and the formation of ductile tearing in the fracture surface. Compared with the A1-8, the C1 is characterized with severer deformation and apparent ductile tearing at the fracture surface but a small bulk damage zone beneath the fracture surface. Therefore, the A1-8 and C1 present proximate impact tensile value. As for the C0, the “necking” phenomenon of C0 leads to its higher impact tensile value. As the impact tensile strength of the high POE content system, similar morphology characteristics are also observed (not shown in here), which can also interpret the difference of their impact on tensile performance.

Fig. 13 SEM images of the no-notched impact tensile fracture surface morphologies and the bulk morphologies beneath the fracture surface of the multilayered blends (A1-8 and A1-128), conventional blend (C1) and the neat PP (C0) at different magnifications. For the low and high magnifications, the scale bars represent 2 mm and 50 μm, respectively. The subscripts 1, 2 and 3 of the A1–D3 represent 100 μm, 700 μm and 2000 μm away from the fracture surface, respectively.

3.5 Structure and property relations

Based on the above discussion, we herein attempt to summarize the influences of the soft and stiff alternating multilayered structure on the mechanical properties of the PP/POE blends, in particular the Izod impact strength. Schematic illustration is proposed for the craze patterns of the typical multilayered blends (Fig. 14(a)–(c)) and conventional blend C1 (Fig. 14(d)). For the A1-128-V, most of the fracture energy is consumed through craze deflecting along the PP/POE interfaces, besides, the existence of the soft POE layers can blunt the craze through its deformation as the craze propagation direction vertical to the layer plane (Fig. 14(a)), which facilitates the enhancement of impact toughness. In the case of fewer layers multilayered blends A1-8-V, although the craze deflection is invisible, more craze initiation sites are found along the PP/POE interfaces and the initiated craze is perfectly arrested by the next adjacent POE layer (Fig. 14(b)), which also enabled the enhancement of its impact toughness. As for the A1-128-P and C1, there is no obvious difference in the amounts of the craze except the shape of craze zone. Therefore, the corresponding toughening mechanism will be revealed by the combination analysis of the fracture surface morphologies and bulk morphologies as shown in the schematic Fig. 15. The neat PP (C0) and the conventional blend C1 exhibit the same sort of morphologies, indicating the same fracture mechanism in them. In detail, several remarkable protrusions, pleats and the unbroken part are found in the fracture surface of the C1 (Fig. 15(A) and (B)), at the same time, considerably larger sub-damage zones with extensive shear deformation and voids are also formed in C1 (Fig. 15(a) and (b)). Such morphology characteristics result in the ultimate toughness enhancement of C1 compared to that of C0. Compared with the C1, the most distinct fracture surface morphologies of A1-128-P are the PP/POE interfaces delaminations (Fig. 15(C)). Moreover, near the fracture surface, a relaxation zone with smooth bulk morphology is formed (Fig. 15(c)). Just as the discussion in Section 3.4, the occurrence of the relaxation zone and delaminations led to the high toughness of the A1-128-P. Compared with the A1-128-P, although the A2-128-P presents severer shear deformation, considerably bigger sized voids and nearly same large sub-damage zone (Fig. 15(d)), the relaxation zone is absent. In addition, due to little breakage of the A2-128-P (Fig. 6) during impact testing, the new facture surface created by the fracture process is little and the interfaces delaminations, if they exist, should be also little. All these result in its lower Izod value than that of A1-128-P despite its high POE content. On the other hand, the relative smooth surface morphology caused by the relaxation zone near to the fracture surface, the large sub-damage zone (at least 5000 μm away from the impact fracture surface) and plenty of large sized voids in a moderate distance are developed during the impact testing, which ultimately enabled the superior impact strength of A1-128-V (Fig. 15(e)). As for A2-128-V, its sub-damage is small and the relaxation zone is absent, and Moreover, the delaminations and obvious plastic deformation is concentrated close to the fracture surface (Fig. 15(f)), which facilitates the A2-128-V with a poorer impact strength compared to A1-128-V despite its high POE content.

Fig. 14 Schematic for the craze patterns of the typical multilayered and conventional blends during the part-impact test. (a) A1-128-V; (b) A1-8-V; (c) A1-128-P; (d) C1.

Fig. 15 Schematic illustrations for the fracture surface morphologies and the bulk morphologies beneath the fracture surface. (A–C) Present the facture surface of C0, C1 and A1-128-P, respectively. (a–f) Present the bulk morphologies beneath the fracture surface 50 μm, 500 μm and 5000 μm away. ((A), (a) C0; (B), (b) C1; (C), (c) A1-128-P; (d) A2-128-P; (e) A1-128-V and (f) A2-128-V).

Although the soft and stiff alternating multilayered structure has been proven to be efficient to toughen the PP blends in this work, the toughness of the multilayered blends is enhanced 2 times as high as that of their corresponding conventional blends with proper POE content and impact direction, which is not as outstanding as nacre. Therefore, several challenges still need to be overcome. First, we have to admit that the synergic factors that act over multiple scales to toughen nacre are only partially observed in PP/POE multilayered blends. Additionally, it is really a great challenge to copy the multiscale fine features of nacre, such as platelet waviness, mineral bridges and nano-asperities through micro-co-extrusion. Moreover, the toughness of polymers changes widely and for those possessing high toughness polymers, the potential improvement of the toughness cannot be as outstanding as the inorganic materials such as calcium carbonate in nacre or ceramics in layer ceramics. Therefore, we wonder if a relative “brittle” material is chosen, or if the toughness were measured in a lower temperature, whether more outstanding improvement of the toughness can be gained using multilayered microstructure. This work is now being undertaken in our group.

4. Conclusion

In this work, two different POE content (6.79 and 16.57 vol%) multilayered blends were successfully fabricated and the shear history of the blends was successfully removed through heat treatment. Compared with the conventional blending samples, no matter that the impact is vertical or parallel to the layer plane, the alternating multilayered blends with low POE content and high layer number present high Izod impact values. Simultaneously, the impact tensile strength of the multilayered blends with high layer number also present high values. This is attributed to the unique fracture mechanisms during the fracture process. In detail, the craze deflects along the PP/POE interfaces, the large sub-damage zone and the occurrence of the relaxation zone can mainly account for their high toughness. For the high POE content multilayered blends, the inefficiency to form voids, as well as the absence of the relaxation zone and the small sub-damage zone led to their low Izod value. Pertaining to the abnormal phenomenon that the multilayered blends with low POE content exhibit high toughness, this work provides us not only a deep insight for understanding the toughening mechanism, but also a new route to toughen PP or other polymer materials without obvious sacrifice of their strength.

Acknowledgements

Financial supports of the National Natural Science Foundation of China (51273132, 51227802 and 51121001) and Program for New Century Excellent Talents in University (NCET-13-0392) are gratefully acknowledged.

References

1. W. V. Vander, R. Nijhof and R. J. Gaymans, Polymer, 1999, 40, 6031–6044.
2. C. M. Chan, J. S. Wu, J. X. Li and Y. K. Cheung, Polymer, 2002, 43, 2981–2992.
3. D. Shah, P. Maiti, E. Gunn, D. F. Schmidt and D. D. Jiang, Adv. Mater., 2004, 16, 1173–1177.
4. W. C. J. Zuiderduin, C. Westzaan, J. Huetink and R. J. Gaymans, Polymer, 2003, 44, 261–275.
5. Y. Lin, H. Chen, C. M. Chan and J. Wu, Macromolecules, 2008, 41, 9204–9213.
6. L. P. Li, B. Yin and M. B. Yang, Polymer, 2012, 53, 3043–3051.
7. O. Masami and S. Yoshihiro, Polymer, 1993, 23, 4868–4873.
8. X. L. Gao, C. Qu, Q. Zhang and Q. Fu, Macromol. Mater. Eng., 2004, 1, 41–48.
9. L. Hyuksoo, D. F. Paula and R. R. William, Polymer, 2005, 46, 11673–11689.
10. D. A. Stone and L. T. J. Korley, Macromolecules, 2010, 43, 9217–9226.
11. M. A. Meyers, P. Y. Chen, A. Y. M. Lin and Y. Seki, Prog. Mater. Sci., 2008, 53, 1–206.
12. T. H. Lin, W. H. Huang, I. K. Jun and P. Jiang, J. Colloid Interface Sci., 2010, 344, 272–278.
13. R. Libanori, F. H. L. Munch, D. M. Montenegro and A. R. Studart, Compos. Sci. Technol., 2012, 72, 435–445.
14. L. J. Bonderer, A. R. Studart and L. J. Gauckler, Science, 2008, 319, 1069–1073.
15. L. J. Bonderer, K. Feldman and L. J. Gauckler, Compos. Sci. Technol., 2010, 70, 1966–1972.
16. L. J. Bonderer, K. Feldman and L. J. Gauckler, Compos. Sci. Technol., 2010, 70, 1958–1965.
17. I. Corni, T. J. Harvey, J. A. Wharton, K. R. Stokes, F. C. Walsh and R. J. K. Wood, Bioinspiration Biomimetics, 2012, 7, 031001.
18. C. D. Mueller, S. Nazarenko, T. Ebeling, T. L. Schuman, A. Hiltner and E. Baer, Polym. Eng. Sci., 1997, 37, 355–362.
19. M. Gupta, Y. J. Lin, D. Taneisha, A. S. David and E. Baer, Macromolecules, 2010, 43, 4230–4239.
20. R. Tangirala, E. Baer, A. Hiltner and C. Weder, Adv. Funct. Mater., 2004, 14, 595–604.
21. S. X. Xi, M. Wang, J. Li and S. Y. Guo, Polymer, 2008, 49, 4861–4870.
22. E. Baer, A. Hiltner and H. Keith, Science, 1987, 235, 1015–1022.
23. E. Shin, A. Hiltner and E. Baer, J. Appl. Polym. Sci., 1993, 47, 269–288.
24. D. Haderski, K. Sung, A. Hiltner and E. Baer, J. Appl. Polym. Sci., 1994, 52, 121–133.
25. K. Sung, D. haderski, A. Hiltner and E. Baer, J. Appl. Polym. Sci., 1994, 52, 135–146.
26. K. Sung, D. haderski, A. Hiltner and E. Baer, J. Appl. Polym. Sci., 1994, 52, 147–162.
27. J. Kerns, A. Hsieh, A. Hiltner and E. Baer, J. Appl. Polym. Sci., 2000, 77, 1545–1557.
28. T. M. Burt, J. Keum, A. Hiltner and E. Baer, ACS Appl. Mater. Interfaces, 2011, 3, 4804–4811.
29. T. M. Burt, A. M. Jordan and T. J. K. Korley, ACS Appl. Mater. Interfaces, 2012, 4, 5155–5161.
30. J. Kerns, A. Hsieh, A. Hiltner and E. Baer, Macromol. Symp., 1999, 147, 15–25.
31. J. B. Shen, M. Wang, J. Li and S. Y. Guo, Eur. Polym. J., 2009, 45, 3269–3281.
32. J. F. Wang, C. L. Wang, X. L. Zhang and H. Wu, RSC Adv., 2014, 4, 20297–20307.
33. J. D. Currey, Proc. R. Soc. London, Ser. B, 1977, 196, 443–463.
34. A. P. Jackson, J. F. V. Vincent and R. M. Turner, Proc. R. Soc. London, Ser. B, 1988, 234, 415–440.
35. R. Z. Wang, H. B. Wen, F. Z. Cui and H. B. Zhang, J. Mater. Sci., 1995, 30, 2299–2304.
36. C. Z. Gen, J. J. Su, S. J. Han, K. Wang and Q. Fu, Polymer, 2013, 54, 3392–3401.
37. J. B. Bao, T. Liu, L. Zhao and G. H. Hu, Ind. Eng. Chem. Res., 2011, 50, 9632–9641.
38. L. E. Nielsen and E. Lawrence, Mechanical properties of polymers, Reinhold Publishing Corporation, London, 1962.
39. J. Kolarik, Polymer, 1996, 37, 887–891.
40. Y. Liu, C. H. L. Kennard, R. W. Truss and N. J. Calos, Polymer, 1997, 38, 2797–2805.
41. A. Dasari and R. D. KMisra, Acta Mater., 2004, 52, 1683–1697.
42. H. Tan, Y. Huang, C. Liu, G. Ravichandran, H. M. Inglis and P. H. Geubelle, Int. J. Solids Struct., 2007, 44, 1809–1822.
43. O. Vincent, S. R. Suprakas and S. Rotimi, ACS Appl. Mater. Interfaces, 2012, 4, 6690–6701.
44. G. Mayer and M. Sarikaya, Exp. Mech., 2002, 42, 395–403.
45. W. G. Perking, Polym. Eng. Sci., 1999, 39, 2445–2460.
46. J. Z. Liang and R. K. Y. Li, J. Appl. Polym. Sci., 2000, 77, 409–417.
47. G. Yang, L. Han, H. Ding, H. Wu, T. Huang and X. Li, Mater. Sci. Eng., A, 2011, 528(3), 1382–1390.
48. A. Wal and R. J. Gaymans, Polymer, 1999, 40, 6067–6075.
49. K. Dijkstra and R. J. Gaymans, J. Mater. Sci., 1994, 29, 3231–3238.
50. A. Pawlak and A. Galeski, Macromolecules, 2008, 41, 2839–2851.
51. H. W. Bai, H. Xiu, J. Gao, H. Deng, Q. Zhang, M. B. Yang and Q. Fu, ACS Appl. Mater. Interfaces, 2012, 4, 897–905.
52. A. Dasari, Q. X. Zhang, Z. Z. Yu and Y. W. Mai, Macromolecules, 2010, 43, 5734–5739.
53. J. Sun and B. Bhushan, RSC Adv., 2012, 2, 7617–7632.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra09302j

---