Biaxial orientation of poly(vinyl chloride) compounds: interaction between drawing, structure, and properties.
Oriented polylvinyl chloride) (PVC) products (for example bottles, films, corrugated roofing sheet, pipes) are becoming increasingly important. The earliest detailed studies of the relationship between stretching conditions and properties for biaxially oriented PVC were those reported by Brady (1) and de Vries and Bonnebat (2). Some earlier work in our own laboratories (3) used a small stretching rig to investigate the effect of biaxial drawing on the tensile properties and subsequent recovery of both flexible and rigid PVC. This work was later extended to explore the effect of a wide range of drawing and annealing conditions on various properties (4). A larger scale biaxial stretcher, the "BASE" was developed for this purpose (5). Both uniaxial and biaxial stretching were found to produce increased order in the plane of the film, as demonstrated by X-ray diffraction (3, 6, 7). In a recent paper, Karacan et al. (8) have characterized uniaxially and biaxially drawn PVC films using infra red spectroscopy, refractive index and density measurements, and have also measured their isochronal extensional creep moduli and dynamic tensile moduli. Orientation functions for the drawn samples were calculated, and it was found that there was no significant difference between samples stretched simultaneously and sequentially to the same ultimate draw ratios. In another study, enhancements in the tensile strength and impact resistance of biaxially oriented PVC sheet were also found to be largely independent of stretching mode (9).
The biaxial stretching equipment used in our laboratories has recently been modified (9) for simultaneous and sequential stretching to equal or unequal draw ratios. A set of equal and unequal rigid PVC samples produced by simultaneous stretching has been characterized using X-ray diffraction, birefringence, and density measurements, and results compared with their tensile properties to develop further understanding of structure/property relationships in these systems.
Commercially produced rigid PVC sheet (1.00 mm thick, and based on a formulation containing K61 PVC) was used throughout this work. This sheet is post treated after processing to remove any orientation and frozen-in stresses.
Samples were stretched using the BASE equipment described previously (5, 9). Three stretching modes were used as follows: uniaxial restrained (i.e. with the transverse direction held constant); simultaneous equal; simultaneous unequal, giving major and minor draws in two directions at right angles. Previous work has shown that the simultaneous equal mode produces planar orientation, with identical properties in all directions in the plane of the sheet. Stretching was carried out at 90 [degrees] C and at a rate of 50 mm/min in the major draw direction. The same rate was applied in the minor draw direction to achieve equal biaxial stretching. For unequal biaxial stretching the rate in the minor draw direction was reduced from 50 mm/min to obtain the required relative draw ratios in the sample. [Changes in rate of this order of magnitude have been shown to have no significant effect on the properties of oriented PVC (4}.] Naturally, the rate in the minor draw direction was zero for uniaxial stretching. The sample was then annealed for 5 min at the same temperature and subsequently cooled to room temperature whilst still in the stretched state. Figure 1 shows samples prepared by the three different stretching routes. To eliminate any effects of thermal treatment, an unstretched sample was annealed in the BASE temperature cabinet for 10 rain at 90 [degrees] C.
Sample densities were measured at 23 [degrees] C using a Davenport density gradient column filled with an aqueous calcium nitrate solution. Accuracy is within [+ or -]0.1 kg/[m.sup.3]. Each measurement was made by averaging results from four specimens.
An initial attempt to measure the out-of-plane birefringence of stretched PVC sheets using conventional methods failed because samples cut with a glass knife using a microtome suffered considerable shrinkage during the cutting process. Attempts to avoid this problem by cooling with liquid carbon dioxide during cutting, and extending annealing time to reduce internal stresses in the stretched sheets were unsuccessful. A wedge method, which had been used successfully for transparent polyethylene terephthalate and polycarbonate moldings (10) was therefore investigated. Figure 2 shows the relationship between birefringence, principal refractive indices and stretching direction. For each measurement, an isosceles right-angled triangular sample was cut from the stretched sheet as shown in Fig. 2, so that the two equal sides of the wedge were parallel to the major draw (MA) and minor draw (MI) directions of the stretched sample, respectively. The edges of the sample were cut carefully using a microtome, with a freshly prepared glass knife. The angle of the wedge was checked using a polarizing microscope to ensure that it was precisely 45 [degrees] . The experimental arrangement for the measurement is shown in Fig. 3. The sample wedge was mounted on an optically isotropic glass block with a thin layer of benzyl benzoate so that the upper edge of the sample wedge was approximately parallel to the lower surface of the glass block. Monochromatic green light ([Lambda] = 546 nm) was passed through the mounted sample as shown, and the whole assembly, arranged so that the wedge was at 45 [degrees] to the crossed polars, was photographed. Figure 4 shows typical images for a series of restrained uniaxially stretched samples. The birefringence an was calculated using the following formula:
[Delta]n=[Lambda]/[Delta]l tan [Theta] (1)
where [Lambda] is the wavelength of the green light, [Theta] the wedge angle and [Delta]l the distance between adjacent bands of the wedge image. To minimize error [Delta]l was the average distance between all the bands visible on the photograph. It was intended to measure the birefringences ([n.sub.z] - [n.sub.x]) and ([n.sub.y] - [n.sub.x]) for all samples. However, for the uniaxially drawn samples the birefringence in the MI (transverse) direction was too low for accurate measurement by the method used. For the equal biaxial samples the measurements in the two directions were, as anticipated, not significantly different; i.e. ([n.sub.z] - [n.sub.y]) is low. The maximum value measured was 0.07 x [10.sup.-3].
Wide Angle X-Ray Diffraction
Wide angle X-ray diffraction traces were obtained in the reflection mode using a Hiltonbrooks X-ray generator fitted with a vertical goniometer and operated at 40 kV, 30 mA. Measurements were made in an air atmosphere over the range 10-51 [degrees] 2[Theta] with Ni filtered Cu [K.sub.[Alpha]] radiation. No correction was made for air scattering. Typical results for restrained uniaxial and equal biaxial samples are shown in Fig. 5. (Traces are displaced along the vertical axis for the sake of clarity.) Commercial PVC typically has a crystallinity of [approximately]10%. The crystalline PVC has an orthorhombic unit cell. Intensities at [approximately]17 [degrees] 2[Theta] and 18.6 [degrees] 2[Theta] correspond to (200) and (110) planes resp. and [approximately]24.5 [degrees] 2[Theta], (210), (201) and (111) planes all contribute to the measured intensity. Peak intensities at these three angles were measured for oriented samples and ratioed to corresponding intensities from a trace for an unstretched sample.
Dumbbell-shaped specimens with a waisted section 25 mm long by 4 mm wide were cut from the oriented sheets parallel to both draw directions. These directions will subsequently be referred to as the major (MA) and minor (MI) draw directions [ILLUSTRATION FOR FIGURE 1 OMITTED]. (For the uniaxially drawn samples the minor draw direction is effectively the transverse direction.) The samples were tested at room temperature at a crosshead speed of 20 mm/min using a Lloyd testing machine. The failure strength was calculated using the initial cross-section area. Five specimens were used for each test, and average values calculated.
For convenience, all results were plotted against planar strain, defined as final area of the sample divided by its initial area. This method of presentation, used previously by Brady (1) enables all data to be compared on a single graph. Density results are shown in Fig. 6. Draw ratios in the two directions are shown in the Figure; the first corresponds to the minor draw ratio, and the second to the major draw ratio. (Densities are ratioed to that of an unstretched sample measured at the same time to enable changes to be followed more easily). All stretching conditions produce an increase in density, which increases with orientation, with the maximum effect being observed for the uniaxial samples, and the least effect for equal biaxial ones. As might be anticipated, unequal biaxial samples produce intermediate behavior.
Birefringence results are shown in Table 1, and the full set of values for ([n.sub.z] - [n.sub.x]) is plotted in Fig. 7. Calculated values for ([n.sub.z] - [n.sub.y]) for the unequal biaxial samples show expected trends with draw ratio, providing some confidence in the measurement method used. The three sets of samples are compared in Fig. 7. As expected, the undrawn sample has zero birefringence, and birefringence is increased by drawing. Again, the highest values are obtained for uniaxial samples. For uniaxially drawn samples these values would be expected to be similar to the birefringence ([n.sub.z] - [n.sub.y]) in the plane of the sheet, for example the data reported by Hibi et al. (11). However, restraint in the transverse direction actually causes some stretching in this direction, so results obtained for ([n.sub.z] - [n.sub.x]) in the present work are lower. Equal biaxial stretching further reduces ([n.sub.z] - [n.sub.x]). When unequal biaxial stretching is used, significant values are obtained for ([n.sub.z] - [n.sub.x]) and ([n.sub.y] - [n.sub.x]). Averages for these values, which may be used to compare planar orientation, are shown in Table 1. The direction of measurement is shown by the underlined draw ratios in Fig. 7. As could be anticipated, pairs of data lie above and below the "equal biaxial" line, and values in the major draw direction are higher.
X-ray results are shown in Fig. 8. Drawing has no significant effect on the (110) peak, while there appears to be a small decrease in peak intensity in the 24.5 [degrees] 20 region as planar strain increases. The most significant effect is the increase in intensity observed for the (200) peak. This tends to mirror the density results, with the largest increase being observed for uniaxial samples, the smallest for equal biaxial, and unequal biaxial samples lying in an intermediate position.
Failure strengths are plotted in Fig. 9. As with birefringence, pairs of data are obtained for unequal biaxial samples, since tensile strengths are measured in two directions at right angles. Variations in tensile strength with planar strain are similar to the birefringence variations. These results are also similar to those for yield stress, as reported by Brady (1).
The aim of this work is to relate structure and tensile properties of drawn samples. A general comparison of data suggests clear similarities in observed effects. For directional properties, i.e., birefringence and failure strength, the effect of mode of orientation is as follows:
[TABULAR DATA FOR TABLE 1 OMITTED]
uniaxial [greater than] major draw biaxial [greater than] equal biaxial [greater than] minor draw biaxial
For nondirectional properties, i.e., density and X-ray diffraction intensity,
uniaxial [greater than] unequal biaxial [greater than] equal biaxial
(It should be noted that orientation in the uniaxial samples is not purely in the stretch direction, since the transverse direction is held fixed, providing some strain in that direction.)
These results suggest that the application of a second force inhibits orientation in the major draw direction; as the draw ratio in the second direction is increased, orientation in the first is inhibited further.
The X-ray diffraction data provide some specific information about crystallite orientation. Figure 8c shows that the amount of (200) planes parallel to the surface of the film increases when the sample is oriented, which corresponds to the alignment of PVC chains in ordered regions in the plane of the film. Changes observed in Fig. 5 suggest that the structure produced as orientation increases is mesomorphous in nature (12), i.e., associated with the alignment and packing together of chains rather than the formation of three dimensional structure. The observed shift of the peak maximum to a lower 20 value (i.e. higher d spacing) is also consistent with a mesomorphous structure. The lack of increased intensity shown in Figs. 8a and 8b also supports the view that there are no significant changes in three dimensional order. The observed intensity increases could be due to alignment of existing structure in the plane of the film, or development of new structure. The small decrease in intensity in the shown 24.5 [degrees] 20 region suggests that some realignment may have occurred. It is noticeable that this decrease is greatest for the uniaxial samples, and least for the equal biaxial samples, suggesting that this is a real effect.
Density measurement was used to provide further evidence to distinguish between the two possibilities. It is noticeable that density values seemed to reach a limiting value as draw ratio increased [ILLUSTRATION FOR FIGURE 6 OMITTED], while [I.sub.(110)]/[I.sub.o(110)] continued to increase [ILLUSTRATION FOR FIGURE 8c OMITTED]. This difference in behavior suggests more than one effect. The maximum observed increase in density corresponds to about 2 kg/[m.sup.3], which is similar to values reported by Karacan et al. (8). This value is also of a similar order to the changes observed in unoriented samples annealed at temperatures in the region of 100 [degrees] C (13, 14), with the resulting formation of "secondary" crystallinity. Unoriented commercial PVC contains about 10% crystallinity. Assuming a crystalline density for PVC of 1530 kg/[m.sup.3] (15), and an amorphous density of 1373 kg/[m.sup.3] (16), a change in density of 2 kg/[m.sup.3] would represent a change in crystallinity of [approximately]1.3%. However, X-ray diffraction results suggest that increased order produced by stretching is mesomorphous in nature, rather than three dimensional crystallinity. Such regions would have a lower density than crystalline regions. It is known that the separation between chains (corresponding to the 'a' direction of the unit cell) is increased to 0.54 nm in the mesomorphous structure (12). Assuming packing in the other directions to be similar to that in the crystalline structure would suggest conversion of [approximately]2% of amorphous material into the mesomorphous form. There is no suggestion of a reduction in crystallinity with draw ratio, as tentatively proposed by Karacan et al. (8).
Birefringence measurements provide information about total orientation (crystalline, mesomorphous and amorphous). Since the samples were stretched above their glass transition temperature, and crystallinity is relatively low, amorphous orientation will provide the major contribution to the birefringence values. As observed by Karacan et al. (8), birefringence results for the uniaxial samples are concave to the draw ratio axis, although the actual birefringence results are higher in the work of Karacan et al. (8) than in this work. (Our sheets were prepared by calendering, followed by subsequent relaxation in a press, and had an initial birefringence of zero in the plane of the sheet, unlike the commercial sheets used in the former work.)
Using the birefringence data shown in Table 1, it is possible to assess the general character and magnitude of the molecular orientation. A fully uniaxially oriented PVC gives a birefringence of 12.5 x [10.sup.-3] (8). All the measured values are low compared with this maximum figure, indicating low orientation. Orientation is predominantly in the plane of the sheet, more so in the samples with the highest planar strain. The results therefore show that significant property enhancements are achieved for PVC samples with quite low planar orientation.
One of the objectives of this work was to relate structure and properties. Failure strength is plotted against birefringence in Fig. 10. It is seen that there is a reasonable correlation between the two parameters. Deviations are not systematic, so appear to arise from experimental error rather than other factors such as crystallite orientation.
For the range of sample types studied, uniaxial stretching of samples restrained in the transverse direction produces the highest overall and crystallite orientation, density and tensile strength. Lowest values of all parameters are produced by equal biaxial orientation. Nondirectional properties for unequal biaxial samples lie between these two extremes; directional properties are higher than equal biaxial when measured in the major direction, and lower when measured in the minor draw direction.
Orientation causes some reorientation of crystallites into the plane of the stretched sheet. At lower draw ratios new mesomorphous structures ([approximately]2% of the PVC) are also produced, causing an increase in density. These effects are most significant for uniaxially oriented PVC, and least significant for equal biaxially oriented samples.
Total orientation seems to have the greatest influence on tensile failure strength, and there is a reasonable correlation between birefringence and failure strength over the range of draw ratios investigated.
1. T. E. Brady, Polym. Eng. Sci., 16, 638 (1976).
2. A. J. de Vries and C. Bonnebat, Polym. Eng. Sci., 16, 93 (1976).
3. M. Gilbert and Z. Liu, Plast. Rubb. Proc. & Appl., 9, 67 (1988).
4. M. Gilbert, D. J. Hitt, and M. J. B. Harte, Plast. Rubb. & Comp. Proc. & Appl., 22, 177 (1994).
5. D. J. Hitt and M. Gilbert, Polym. Test., 13, 219 (1994).
6. J. C. Vyvoda, M. Gilbert, and D. A. Hemsley, Polymer, 21, 72 (1980).
7. Z. Liu and M. Gilbert, Polymer, 28, 72 (1987).
8. I. Karacan, D. I. Bower, and I. M. Ward, Polymer, 35, 3411 (1994).
9. D. J. Hitt, Conf. Proc. PVC 96, Brighton. (April 1996).
10. D.A. Hemsley and M. A. Robinson, Polym. Test., 11,373 (1992).
11. S. Hibi, M. Maeda, H. Kubota, and T. Miura, Polymer, 18, 137 (1977).
12. M. Mammi and V. Nardi, Nature, 199, 247 (1963).
13. J. A. Juijn, J. H. Gisolf, and W. A. de Jong, Kolloid.-Z. Z. Polym., 235, 1157 (1969).
14. A. Gray and M. Gilbert, Polymer, 17, 44 (1976).
15. C. E. Wilkes and V. L. Folt, Macromolecules, 6, 235 (1973).
16. Y. V. Glazkosvskii, A. N. Zav'yalov, N.M. Bakardzhiyev, and I. I. Novak, Polym. Sci. USSR, 12, 3061 (1970).
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||First Symposium on Oriented Polymers|
|Author:||Gilbert, M.; Liu, Z.; Hitt, D.J.|
|Publication:||Polymer Engineering and Science|
|Date:||Nov 1, 1997|
|Previous Article:||Evolution of hot strain induced crystalline texture of poly(ethylene terephthalate) films.|
|Next Article:||Multi-cavity pressure control in the filling and packing stages of the injection molding process.|