Cellular Morphology Optimization of Polypropylene/CaCO 3 Composite Films Designed for Piezoelectric Applications

2.1 Materials and Film Development Conditions and Properties

The PP used was supplied by Lyondell Basell, Canada under the trademark Pro-fax 6323. This homopolymer has a density of 900 kg/m³, a tensile strength and elongation at yield of 34 MPa and 10%, respectively, and a melt flow index of 12 g/10 min (230°C/2.16 kg). Its intrinsic viscosity at room temperature is 1.99 dl/g and its weight-average molecular weight is 2.36×10⁵ [32]. Since piezoelectric films sustain several cyclic stresses, PP was chosen because it has good fatigue resistance. It is also low cost, easy to process and has well-known processing conditions. The three CaCO₃ used [Snowhite 3 (SW3), Omyacarb 6 (OM6) and Snowhite 12 (SW12)] were supplied by Omya Canada Inc. Their respective average particle diameters were 3, 6 and 12 μm.

Initial sheets of around 1.0 mm in thickness were obtained using a co-rotating Leistritz ZSE18HP-40D twin-screw extruder with a temperature profile of 210-210-215-215-215-215-210-205-185°C (from the main hoper to the die) and a screw rotational speed maintained constant at 275 rpm. A flat die of 150 mm was adjusted to an opening of 1 mm and heated at 185°C. The three rolls (200 mm width and 140 mm diameter) calendaring system was set at 68°C using an oil bath cooling circuit. The rolls were positioned at 0.9 mm apart and set at a pulling rotational speed of 1.15 rpm. The die to rolls distance was approximately 4 cm. The PP feeding rate was fixed at 3.33 kg/h, while that of CaCO₃ was changed to get various concentrations.

Square samples of 8.5 cm x 8.5 cm were cut from the extruded sheets for the post-extrusion stretching step. Before their biaxial stretching, all samples were annealed at 140°C (10 minutes) under a force of approximately 4500 N to make their thickness more uniform and increase their crystallinity. For each blend composition, three samples were stretched using a Bruckner Karo IV laboratory-stretching machine (stretched area of 7 cm x 7 cm). The samples were rapidly pre-heated at 160°C for 60 s then biaxially stretched using a pantograph clipping system as shown in Figure 2. The stretching parameters were: r_s = 4 (biaxial) and stretching rate = 1 m/min [11].

OPEN IN VIEWER

Figure 2

Brückner stretcher showing the stretched PP/CaCO₃ sample after its extraction from the hot air oven

Following the stretching step, all samples were submitted to gas (nitrogen) diffusion expansion (GDE) using a progressive method in a homemade device enabling a pressure treatment up to 6 MPa at temperatures from 25°C to 150°C [10]. Since a high pressure is applied to the sample under high temperature, this makes it easier for nitrogen to diffuse inside the PP matrix then to penetrate inside the cells until pressure equilibrium. A subsequent sudden release of the external pressure leads to cell inflation due to the high pressure already built inside the cells. This step is followed by a rapid cooling to solidify the sample and maintain its cellular structure. A second method, the two-step inflation (long- time GDE), was also used for comparison [18]. Both methods sequences are detailed in Figure 3 .

OPEN IN VIEWER

Figure 3

GDE sequences for: (a) long-time GDE, and (b) progressive GDE. (solid lines: applied pressure; dashed lines: applied temperature)

Film thickness was measured using a Mitutoyo Digimatic micrometer with a resolution of 1 urn. For each extruded sample, 21 measurements were randomly taken before stretching and 63 measurements were taken after stretching. Moreover, 7 measurements were also taken at the centre of the stretched sample, where the thickness is the most uniform.

The crystallinity was measured using a Perkin-Elmer DSC 7 and Pyris Software. Samples of 5 to 22 mg were tested under the following thermal sequence: i) a first heating from 50 to 190°C at 10°C/min, ii) cooling from 190 to 50°C at 10°C/min, and iii) a second heating from 50 to 190°C at 10°C/min. All measurements were performed under a nitrogen atmosphere (20 mL/min). Since the crystallinity of the samples after each processing step was needed, only the area under the first peak (120 to 170°C) was used to compute the melting enthalpy. The value of 207.1 J/g was used as the 100% crystalline PP melting enthalpy [35].

Finally, the cellular morphology was characterized using scanning electron microscopy (SEM) and ImageJ software to perform the analysis of the manually identified cellular structure. The microscope used was a JEOL JSM-840A operated at 15 kV and pictures were taken at three magnifications: 150, 300 and 1000X. To get a good statistical analysis, at least three different pictures per film were taken at 300X, which was enough to measure cell dimensions greater than 2 urn (see Figure 4). This analysis was performed in two directions: parallel (MD) and perpendicular (transverse, TD) to the extrusion direction.

OPEN IN VIEWER

Figure 4

Example of SEM image (300X) of a film cross-section showing how cell morphology is characterized using the ImageJ software. The image identifies ellipsoidal cells from which the length (a), the height (b) and the aspect ratio (a/b) were computed

3.1 Developed PP/CaCO₃ Films and their Corresponding Thickness and Composition

To measure the exact CaCO₃ concentration inside the extruded PP films, different PP/SW3, PP/OM6 and PP/SW12 films were characterized by thermogravimetric analysis (TGA) from 50 to 950°C at 10°C/min under a nitrogen atmosphere (TA instruments Q5000IR). For all films, the CaCO₃ mass was assumed to be equal to that of the residues after complete polymer decomposition at around 475°C. The corresponding results are shown in Table 2 together with the average unstretched film thickness after annealing and the melting temperature obtained from the corresponding DSC curves.

As mentioned above in Section 2.1, during sample levelling (before stretching), all samples were subjected to annealing at 140°C for a short time then cooled down slowly to increase their crystallinity which is important for piezoelectric films since the interface between the amorphous and crystalline zones constitute an obstacle to electric charges drift retaining these charges for longer time. Figure 5 shows the crystallinity variation of PP/SW12 and PP/OM6 films as a function of filler concentration for annealed films before and after biaxial stretching. For both unstretched and stretched films, it is seen that PP crystallinity does not depend on the type of filler used (SW12 or OM6), but only on filler concentration. For the range of concentrations studied, the crystallinity increases with increasing filler content due to the fact that solid filler particles act as crystal nucleation sites from which crystal growth takes place. This heterogeneous nucleation effect leads to film crystallinity increase with increasing filler concentration. The crystalline areas are denser around the particles than further inside the PP matrix since the corresponding crystallites grow from the filler particles toward the PP matrix. Figure 5 also shows that for the same SW12 or OM6 filler concentration, film crystallinity is higher after film biaxial stretching which is due to the alignments of PP chains during stretching, i.e. orientation-induced crystallinity. This increase is more pronounced at higher filler concentrations. Since the crystalline areas are denser around the solid CaCO₃ particles, the PP/CaCO₃ interface it then subjected to higher stress during the biaxial-stretching, which explains the fact that PP/SW12 films showed higher cellular structure due to the higher interfacial decohesion generated by these high stresses.

OPEN IN VIEWER

Figure 5

Crystallinity of PP/SW12 (closed symbols) and PP/OM6 (open symbols) films as a function of filler concentration before and after biaxial stretching

The size distribution of CaCO₃ particles is an important parameter to generate a filler/matrix decohesion over a wide range of particle size. Figure 6 shows the particle size distribution of the three CaCO₃ particles (SW3, OM6, and SW12) used based on their equivalent spherical diameter. As shown, more than 70% of OM6 and SW3 particles are smaller than 3 μm in diameter. The presence of these small particles in the PP film will not serve as crack initiators during film biaxial stretching. Figure 7 shows the cross-section SEM pictures (taken along MD) of cellular films developed with the above three fillers. It is obvious from those micrographs that not all these fillers led to a cellular structure after biaxial stretching and GDE. None of the films obtained with the smaller size SW3 filler (3 μm size) showed a cellular structure, as shown in Figure 7(a) . Due to the fact that they are very small, these particles just move along with the surrounding PP matrix during stretching without initiating cracks. Figure 7(b) shows that films developed with OM6 (6 μm size) had only few cells at a filler concentration of 26 wt.%. On the other hand, Figure 7(c) shows that films obtained with SW12 (12 μm size) at a filler concentration of 44 wt.% had the highest cellular (layered) structure in the form of elongated cells around the particles. Also, due to the high cell inflation during the GDE step, the film is thicker. This is due to the delamination occurring at the PP/CaCO₃ interface in the MD and TD directions only, but not perpendicular (normal) to the film surface since planar biaxial stretching was applied in these directions. Due to their large particle size distribution, only OM6 and SW12 have a certain amount of particles with a diameter over 10 μm, which corresponds respectively to a fraction of 5% and 8% of the whole OM6 and SW12 fillers inside the sample. The SW12 filler is then the most adequate to selectively initiate and propagate cracks during stretching to generate a cellular film under the processing conditions used. It also shows larger cells that are known to decrease the Young's modulus.

OPEN IN VIEWER

Figure 6

Particle size distribution of the CaCO₃ particles (SW3, OM6 and SW12) based on their equivalent spherical diameter using the ImageJ software

OPEN IN VIEWER

Figure 7

SEM micrographs (300X magnification in the perpendicular direction, MD) of PP/CaCO₃ cellular films developed using the progressive GDE method: a) 18 wt.% of SW3; b) 26 wt.% of OM6; and c) 45 wt.% of SW12

Analyzing the film thickness reduction ratio h_r after the biaxial stretching step gives an important information about the interactions between the CaCO₃ particles and the PP matrix. First, it should be mentioned that the standard deviation of the thickness reduction ratio over the whole sample surface qualifies film thickness uniformity. Figure 8 shows the average h_r as a function of CaCO₃ filler concentration for both SW12 and OM6 fillers. At all concentrations, both OM6 and SW12 show an approximately linear decrease of h_r with increasing filler concentration and the SW12 filler leads to lower h_r values. Also, it can be seen that smaller particles (OM6) yield better overall film uniformity (lower standard deviations). For all concentrations studied, SW12 led to lower film thickness reduction ratio. The expected h_r obtained from Equation (1) for r_s = 4 is given by the horizontal dashed line. Since Equation (1) assumes a constant volume of the samples during stretching, samples having h_r below that line (PP/SW12 samples) show a total volume increase, thus a density reduction due to void formation. On the other hand, for samples having no cells, which are those with h_r above the horizontal dashed line, it can be concluded that the PP matrix was not subjected to enough stress to create interfacial delamination (voids). However, under stretching, the PP molecular chains were aligned, resulting in an initial compaction of the structure and leading to higher h_r.

OPEN IN VIEWER

Figure 8

Average film thickness reduction ratio as a function of CaCO₃ concentration. Open symbols: SW12 and closed symbols: OM6

Since the literature reports several GDE techniques used to inflate the cells, we first compared the cellular morphology of selected PP/SW12 films obtained by the progressive GDE technique with the so-called long-time GDE technique [10, 18]. Both techniques are presented in Figure 3. Table 3 reports on the cell wall thickness t (μm) and average cell aspect ratio (a/b) in the machine (MD) direction, for PP/SW12 films at 11.0 and 44.0 wt.% of SW12 after gas diffusion expansion using both GDE techniques. Taking into account the standard deviation values, the average t values are not significantly affected by the GDE technique used to inflate the cells. For both long-time and progressive GDE techniques, the average t fluctuated around 20 and 6 μm for 11.0 and 44.0 wt.% of SW12, respectively. On the other hand, for the highest SW12 concentration (44 wt.%) where the cells have thinner walls, the average aspect a/b ratio also did not show any significant difference between both inflation techniques (a/b was around 8). However, this was not the case for the lowest SW12 concentration (11.0 wt.%) where the cell walls were more than three times thicker (around 20 μm). Although the average t was not significantly affected by the GDE method used, the average a/b was more than doubled (increased from around 15 for long-time GDE to around 33 for progressive GDE). So to get an aspect ratio within the specification of Table 1 (4 < a/b < 10), the progressive GDE method which is 5 times shorter than the long-time GDE, as shown in Figure 3 , is more suitable for high SW12 concentrations where the cell walls are thinner (t ≈ 7 μm).

The progressive GDE technique was then selected to develop PP/SW12 cellular films at four various SW12 concentrations (11.0, 23.0, 38.0 and 44.0 wt.%). These films were characterized for their cellular morphology in terms of cell aspect ratio (a/b), cell height (b), and cell wall thickness (t). It is important to mention that the aspect ratio is a key factor in reducing the Young's modulus of the cellular films. In Figure 9(a) , the average aspect ratio in both MD and TD is reported as a function of SW12 concentration. For both directions, a/b decreases with increasing filler concentration. This is due to the fact that cell wall thickness decreases with increasing filler concentration and consequently become easier to stretch under internal pressure, leading to higher cell height (b) values. Also, the figure shows that, for all the SW12 concentrations, the a/b ratio is higher in MD than in TD, which is due to PP chains orientation along MD. Taking into account the standard deviation, it can be observed from Figures 9(a) and (b) that the cell size distribution is not uniform. However, as seen in Figure 9(a) , a more uniform ellipsoidal cell shape was obtained by increasing SW12 concentration (lower standard deviation on the aspect ratio). A filler concentration higher than 23.0 wt.% is then considered as adequate to generate cells with an average aspect ratio within the specification of Table 1 (4 ≤ a/b ≤ 10) in both MD and TD, which is needed for good piezoelectric cellular films. Cell height b is also an important parameter since the micro-discharge barrier takes place only at a minimum cell height, depending on the gas pressure within the cells [34]. As seen from Figure 9(b) , very good cell height was achieved with respect to the minimum required value of around 5 μm. Nevertheless, apart from the fact that the cells obtained are above 5.2 μm, the samples with at higher concentration have thicker cells. This is attributed to the fact that for these filler concentrations, the cell walls are thinner and, as seen in Figure 9(a) , lead to better GDE effectiveness, and consequently a lower Young's modulus (flexible walls for better inflation and charge displacement). Figures 9(c) shows the cell wall thickness as a function of SW12 concentration. For a concentration of 23 wt.% and higher, only a small variation of the average t was observed (from 9 to 6 μm). It is seen that after GDE, cell walls thickness remained slightly higher than the target values presented in Table 1 (3 μm < t < 7 μm).

OPEN IN VIEWER

Figure 9

Average a/b aspect ratio (a); average cell height b (b), and average cell wall thickness t (c), as a function of SW12 concentration, in both MD and TD after progressive GDE