SEARCH

SEARCH BY CITATION

Keywords:

  • mechanical properties;
  • on-line X-ray scattering;
  • polypropylene;
  • SAXS;
  • strain-induced crystallization;
  • structural characterization;
  • WAXS

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

The mechanical behavior of polymer materials is strongly dependent on polymer structure and morphology of the material. The latter is determined mainly by processing and thermal history. Temperature-dependent on-line X-ray scattering during deformation enables the investigation of deformation processes, fatigue and failure of polymers. As an example, investigations on polypropylene are presented. By on-line X-ray scattering with synchrotron radiation, a time resolution in the order of seconds and a spatial resolution in the order of microns can be achieved. The characterization of the crystalline and amorphous phases as well as the study of cavitation processes were performed by simultaneous SAXS and WAXS. The results of scattering experiments are complemented by DSC measurements and SEM investigations. © 2010 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 48: 1574–1586, 2010


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

The stress-strain curves of semicrystalline materials generally show three characteristic regions. Although the initial region before the yield point apparently behaves elastically, stress relaxation due to rearrangements in the amorphous phase can be observed. The mobility in the amorphous phase depends on the difference between the ambient temperature and the temperature characteristic of the glass transition, which is the dominant relaxation process in the temperature range under investigation. After the yield point, typically local necking occurs with high local strain, whereas the overall strain remains moderate. In the case of dog-bone specimens then the neck propagates over the whole specimen during constant load. In the true stress-strain diagram necking is a fast local deformation from the yield strain to the strain of the fully yielded specimen. In the third step during further elongation strain hardening occurs, until finally the specimen fails.

Already in his first model of structural changes during deformation, Peterlin1 discussed spherulites consisting of lamellae separated by amorphous regions. Because of tie molecules stress is transferred between the lamellae inducing deformation and finally disintegration and rearrangement in the form of fibrillae. Later numerous authors modified and enhanced Peterlin's approach to describe their experiments. The discussion gained new impetus after the first online structure investigation during deformation by Chen et al.2

Breese and Beaucage3, 4 gave an overview of older models. Finally, they describe mechanical behavior with a model refining the models of Weeks and Porter5 and Gibson et al.6 in the sense that the effects of the orientation process on the modulus of the nonfibrous gel component are incorporated into the model by allowing a transition to fibers.

To describe the deformation behavior of semicrystalline polymers Strobl and coworkers7, 8 separated different mechanisms of stress transfer with respect to amorphous and crystalline units. They distinguished between four phases: Onset of local block sliding (1), collective motion slightly below the yield point (2), disassociation of crystalline blocks and transformation into fibrils (3), and the start of disentangling (4).

In most models, the common process of cavitation in polymers during deformation is not yet incorporated. Many systems show a characteristic whitening during plastic deformation due to the formation of voids or cavities with a typical length scale growing up to the wavelength of visible light, that is, some hundreds of nanometers (see Pawlak and Galeski9 and Men et al.10). It is obvious that these large scale structures are not accessible with a conventional small angle X-ray scattering setup. But at least the beginning of this process should be well-detectable by X-ray scattering due to the strong difference in electron density between the polymer and the voids in the relevant angular range accessible by USAXS (see Lode et al.,11 Gehrke et al.,12 and Roth et al.13).

In previous studies, Davies et al.14 demonstrated a continuous generation of the voids in the plastic phase starting at the yield point, which led to a conspicuous increase in the scattering power. Furthermore, a change in the diffuse scattering profile indicates a monotone change in the size and in the shape of the voids.

Stribeck et al.15 described nanostructure evolution in polypropylene during online mechanical testing with simultaneous SAXS and refined details of the interaction between the different phases including cavitation.

Processing conditions strongly affect the mechanical properties caused by the interaction of the different phases within the samples. For instance in an injection molded plate, the properties can be extremely dependent on the position and direction of a specimen, changing from relatively brittle to highly stretchable.

As an example the properties of isotactic polypropylene samples, produced by compression and injection molding, respectively, were investigated under stepwise loading. Simultaneously, the structural changes were characterized by Synchrotron-SAXS and discussed together with the mechanical properties. It was found that there is a highly preoriented structure within the injection molded specimen, which strongly influences the further temperature-dependent deformation behavior. The influence of temperature and preorientation on the deformation behavior is described extending the present models of plastic deformation.

EXPERIMENTAL

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

General Procedure

To investigate structural modification during deformation X-ray scattering experiments were performed on samples mounted in a miniaturized tensile rig placed in the synchrotron X-ray beam. The synchrotron radiation is necessary to have sufficient intensity to get high time resolution. A general description of the equipment was given by Davies et al.14

The actually used experimental arrangement and the specimen geometry are illustrated in Figure 1.

thumbnail image

Figure 1. Sketch of the experimental arrangement for SAXS and WAXS during deformation (left) and waisted specimen (mini-dumbbell) used for simultaneous structure and mechanical investigation (right), the dimensions of the specimen can be scaled.

Download figure to PowerPoint

To investigate local strain-dependent properties, small waisted specimens were used to concentrate the stress in the center of the specimen. By using a relatively large radius of curvature (12 mm) compared with the specimen width (3 mm), the stress state is in good approximation maintained constant in the middle of the specimen. The strain was determined optically by observing the deformation of a grid pattern applied on the specimen surface. The grid pattern was applied using a self-made flexible ink and a mesh size of 0.35 mm. The middle of the specimen, where the beam crosses, was left blank. Alternatively, also a classical image correlation analysis for strain estimation can be used. A comparison of stress-strain curves of the waisted specimens with standardized dog-bone specimens show a good consistency: The curve progression from the beginning to the yield point as well as during strain hardening, that is, where the parallel region of the dog-bone specimen is deformed in a uniform way, are identical. For the range of neck formation and propagation in the dog-bone specimen, an estimation of true stress and strain via global strain measurement are not possible. Only a local strain measurement solves the problem. The local measurement of the waisted specimens provides true stress and strain data in a good approximation.

In this context, the specimens are assumed as incompressible. This allows calculating the true stress σt from the measured force F, initial cross section A, and tensile strain εt as σt = F × (1+εt)/A.

To characterize the structure at certain strains mostly step-loading experiments were performed, stretching the sample to a certain strain and then recording the patterns at this constant strain. This is also the reason for the presented slightly fluttering stress-strain curves. In some cases also continuous stretching was performed.

To keep the beam in a fixed position relative to the gauge length of the waisted sample throughout the measurement, both grips were moved simultaneously in opposite direction.

To get simultaneous SAXS and WAXS patterns, the WAXS could be monitored only in a limited range. By using a horizontal tensile direction, the vertically arranged WAXS detector monitors mainly the equatorial scattering of the sample.

For the validity check, the sample was rotated around the tensile direction and the patterns were compared. To detect oriented crystallites, the sample can be rotated around an axis perpendicular to the beam.

To perform temperature-dependent tensile and scattering experiments, a small heating device locally blows preheated air at the sample.

Besides the global scanning of samples, the experimental setup also permits spatially resolved pattern recording, for example, around a crack tip, if it is used in a microfocus beamline. The described arrangement was successfully used investigating semicrystalline polymers during deformation (Schneider et al.16) and fracture (Schneider17, 18).

WAXS and SAXS measurements were performed at the synchrotron beamline BW4 at HASYLAB in Hamburg, Germany.13 The wavelength of the X-ray beam was 1.3808 Å, the beam diameter was about 400 μm. The SAXS images were collected by a two-dimensional MarCCD-detector (2048 × 2048 pixels of 79.1 × 79.1 μm2). The sample-to-detector distance was set to 4080 mm. The WAXS images were collected by a two-dimensional PILATUS 100K-detector (487 × 195 pixels of 172 × 172 μm2). By a special procedure, the position of the tilted WAXS-detector was determined to have a tilt angle of 22.7°. The distance between the sample and the point of normal incidence was 247 mm.

Exposure times were chosen in the range of 5–60 s per pattern. The frame rate, determined by exposure and data storage, was 15–70 s per pattern.

For the discussion, all SAXS and WAXS 2d-patterns are shown with vertical tensile direction. For specimens with fiber symmetry, this means that the fiber axis and so the scattering vector s3 also are vertical, the scattering vector s12 always are horizontal.

For quantitative comparison of crystallinity, some DSC measurements were performed on drawn and undrawn samples using a DSC Q 1000 from TA-instruments. The specimens were measured between −80 °C and 230 °C with heating and cooling rates of 20 K/min. Before each scan, the sample was equilibrated at constant temperature for 300 s.

As qualitative check of the discussed structures, some SEM images of the drawn samples were taken on a Zeiss Gemini Ultra plus SEM.

Material

The investigations were done on isotactic polypropylene (iPP), grade HD 120 MO from Borealis. Both injection and compression molded plates were used. Melt temperature in injection molding was 245 °C, mold temperature 40 °C. The mold has a film gate with a width of 0.5 mm, the size of the plate is 80 mm × 80 mm. The compression molded plates were produced by heating injection molded plates to 190 °C for 5 min in a vacuum press. Then the plates were cooled down to room temperature at 15 K/min. All plates have a thickness of about 1 mm. Mini-dumbbell specimens were produced by CNC milling. From the injection molded plate, specimens were taken near the inlet in injection direction and perpendicular to the injection direction.

Evaluation of Scattering Data

SAXS Data Evaluation

The evaluation of SAXS images is strongly related to the approach for materials with fiber symmetry developed by Stribeck.19 Data processing was realized with the software package pv-wave from Visual Numerics.

The images were normalized with respect to the incident flux and blind areas were masked. Considering the sample absorption, the instrument background was subtracted. By translation and rotation, the images were aligned in that way that the tensile direction is vertical and the beam position in the middle of the patterns. Finally, the patterns were averaged with respect to the four quadrants of the detector. The blind area of the beam stop is interpolated assuming a Guinier-type behavior of intensity in this range.

Generally, SAXS realizes a projection of the specimen structure into the reciprocal space, where it produces a slice of the Fourier transform (FT) of the structure whose amplitude is measured. A single back-transformation would deliver a projection of the autocorrelation function of the structure.

Complete information about the specimen would be available only by tomographic methods with a stepwise rotation of the sample (see e.g., Schroer et al.20) or using inherent symmetry properties of the sample. Under the assumption of fiber symmetry of the stretched specimen around the tensile axis, from the slices through the squared FT-structure, the three-dimensional squared FT-structure in reciprocal space can be reconstructed and hence also the projection of the squared FT-structure in reciprocal space. The Fourier back-transformation of the latter delivers slices through the autocorrelation function of the initial structure. Stribeck pointed out that the chord distribution function (CDF) as Laplace transform of the autocorrelation function can be computed from the scattering intensity I(s) simply by multiplying I(s) by the factor L(s) = −4π2(smath image + smath image) before the Fourier back-transformation. Here s is the scattering vector with the component s12 in transversal direction and s3 in fiber direction.

The interpretation of the CDF is straightforward,21 because it has been defined by the Laplacian of Vonk's multidimensional correlation function.22 It presents autocorrelations of surfaces from the scattering entities in that way that positive values characterize distances between surfaces of opposite direction, negative values distances between surfaces of the same direction, see Figure 2.

thumbnail image

Figure 2. Information of CDF's: In any particular structure, positive values of the CDF represent correlations of interfaces with opposite direction (light arrows); negative values represent correlation of interfaces with the same direction (dark arrows).

Download figure to PowerPoint

Hence, positive peaks in the vicinity of the origin characterize size distributions of the primary domains. In the case of semicrystalline materials, this can be crystallites as well as amorphous regions in-between. If cavitation occurs, it will be superimposed by the size of the cavities. Also a relatively small amount of cavities will be visible because the difference in electron density is here much higher than between crystalline and amorphous phase within the polymer.

Following negative peaks characterize distances between adjacent repetition units (long periods).

Positive peaks at greater distances describe size and orientation of domains (from the beginning of the first domain to the end of the second one). In-well oriented systems also peaks of higher order can be observed.

For detailed discussion, the CDF's can be presented as contour plots or as density plots in the plane.

WAXS Data Handling

According to the experimental arrangement, the WAXS data unfortunately are restricted to only a certain range of angles. The position of the WAXS detector was estimated using certain reference reflexes of iPP. The procedure developed for this purpose will be published separately.

In a first step, the individual images were normalized with respect to the incident flux and blind areas were masked. The instrument background was subtracted considering sample absorption. The images were always transformed to reference coordinate systems with respect to the scattering angles (scattering angle 2θ and azimuthal angle φ or scattering vectors s12 and s3).

For further discussion, the scattering intensity was projected over a selected range of φ on the 2θ-axis or discussed over a selected range of 2θ on φ-axis.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

Preliminary Mechanical Tests

Because of the waisted sample geometry, the whole deformation of the samples is different compared with standardized dog-bone specimens. We no longer observe the formation of a neck which moves at constant load across the parallel region of the specimen. Once the neck has formed the sample immediately continues stretching within the necked region due to the increasing cross section of the sample adjacent to this region. The comparison of technical and true stress-strain curves for the three specimens under consideration is shown in Figure 3. The strain is measured optically. True stress is calculated assuming constant volume as σt = F × (1+εt)/A, naturally it is always higher than the technical stress.

thumbnail image

Figure 3. Stress-strain curves of waisted specimens from compression and injection molded iPP specimen at room temperature. The technical stress is always the curve with the lower stress values.

Download figure to PowerPoint

The injection molded specimens in injection direction as well as the compression molded and quickly cooled specimens are highly stretchable. By contrast, the specimens transversal to the injection direction fail very soon. Here in some cases, even crazing could be observed before failure. This points out that there will be a strong structural anisotropy due to the processing history (shear stress as well as cooling rate).

Check of Fiber Symmetry

The evaluation of SAXS patterns uses the constraint of fiber symmetry. To check the validity of this constraint several samples were investigated under three different angles of incident beam. By rotating the sample, the X-ray beam penetrated the sample from the flat side, the narrow side, and 45° in-between. Comparison of the normalized patterns showed no significant differences, which would be expected, for example, in the case of any affine deformation in one of these directions. Small differences arise due to a skin-core-structure of the samples. With increasing strain, the differences become smaller.

Crystallite Identification

According to Bragg's law, the positions of WAXS reflexes refer to the distance between crystalline planes within the crystallites. In the compression molded as well as in the injection molded plates, a couple of crystalline reflexes could be identified. The reflexes within the relevant angular region are summarized in Table 1.

Table 1. Crystalline Reflexes of iPPa
  • a

    Crystalline reflexes were used for calibration of the position of the WAXS-detector as well as for further discussion of changes during deformation.

inline image

Elastic Crystallite Deformation

Loading of the samples below the yield point causes a certain shift of the crystalline peaks. According to Bragg's law, the peak positions refer to spacing between lattice planes. Their shift inversely reflects the strain vertical to the corresponding plane. In this context, the shift of the observed (hk l)-reflexes in equatorial direction (perpendicular to tensile direction) display the transversal crystalline strain of the crystallites aligned in tensile direction.

Cyclic loading and unloading gives an estimation of the elastic crystalline strain also during plastic deformation. Figure 4 shows transversal crystalline strain of the (1 1 0)-reflex versus the overall tensile strain for three consecutive loading cycles. The first strain amplitude was 0.12, the remaining strain after unloading 0.04. The following cycles were conducted within the overall elastic region. The shift of other (hk l)-reflexes yield the same result.

thumbnail image

Figure 4. Transversal crystalline strain versus optically measured tensile strain of semicrystalline iPP specimen estimated via the shift of the (1 1 0)-reflex at room temperature.

Download figure to PowerPoint

In a range of up to about 8% strain, the crystallites deform transversally by about −0.8%. The mismatch between crystalline and global strain indicates the different stiffness of the crystallites compared with the whole specimen due to the relatively soft amorphous regions. During further loading, the strain of the crystallites remains constant. This means that further deformation is realized on the microscale only by the amorphous phase and by additional shear dislocation of the crystallites.

Crystallite Reorientation During Drawing

The stress-strain diagrams captured during the scattering experiments are shown in Figure 5. The optical measured strain did not exhibit the stability as in the preliminary mechanical tests mainly due to the stepwise loading, but the general trend is representative. As expected, the stress-level of the stress-strain decreases with increasing temperature. Otherwise, it is generally lower than in the preliminary tests due to the generally lower strain rate.

thumbnail image

Figure 5. Temperature-dependent stress-strain diagrams of the compression molded iPP specimens under investigation.

Download figure to PowerPoint

During deformation, the WAXS patterns change in a quite characteristic way, see Figure 6. Initially the intensity of the crystalline peaks is relatively constant over the azimuthal angle pointing to a homogeneous distribution of crystallite directions. At deformations below the yield point, slight changes in the intensity of the peaks are reversible.

thumbnail image

Figure 6. Equatorial scattering intensity of undeformed iPP at room temperature and during strain hardening at elevated temperatures. The curves always show a projection of the angular range of ±2° around the equatorial direction. Right: WAXS-patterns, azimuthal angle versus scattering angle 2θ, at three characteristic points: Undeformed as well as immediately before failure at room temperature and 150 °C. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Above the yield point, the deformation behavior is strongly temperature-dependent. Generally in the investigated equatorial direction only the (hk 0)-peaks remain and the scattering intensity concentrates azimuthally in equatorial direction. This is generally an indication of orientation of crystallites with the c-axis (chain direction) in tensile direction. Although the peaks sharpen with increasing temperature, which is an indication of growing crystallites, they decrease strongly and broaden at temperatures in the range immediately above room temperature. This indicates a gradual disruption of crystallites into smaller fragments. Finally, at room temperature, the distinct peaks disappear but give a broad halo in the angular range of the previous peaks. This indicates a rough orientation of the chains or very small crystallite fragments in tensile direction, probably in the form of fibrils. However, the thermal mobility of the chains seems to be all in all insufficient to establish new crystallites.

Transmission and Total/Maximal Scattering Intensity

For the flat samples under investigation, the transmission was 0.83 in the unstretched state. During stretching and simultaneous thinning, the transmission generally increased to 0.96.

SAXS During Stretching

From the set of successive scattering patterns during the deformation for the following discussion, only patterns and the corresponding CDF's were chosen where characteristic changes could be seen.

The respective results of SAXS from stretching at room temperature are shown in Figure 7. The regular circular form of the CDF from the beginning is nearly unchanged to about 6% strain also during unloading. It represents the randomly distributed crystallites, which were deformed elastically, as discussed in the section about WAXS above. The diameter of the inner ring in the CDF indicates a lamellar thickness of 9.8 nm. The isotropic ring in the negative direction represents the long period of 17.7 nm. The following second positive and negative reflexes indicate a certain correlation with the third neighbors. Missing higher-order reflexes indicate that there are surprisingly no remarkable correlations above the third neighbors. During deformation below the yield point (loading, unloading, relaxation, and repeated loading), this general situation is unchanged.

thumbnail image

Figure 7. Deformation of iPP at room temperature, from left to right: SAXS patterns, positive and negative CDF's (always log-scale, pseudo color) as well as a surface plot of the CDF's (linear scale and for two strains also log scale) at different strains ε = 0.0, 0.125, 0.37, and 5.3 (from top). Each square of the pattern covers a range −0.12 nm−1 < s1, s3 < 0.12 nm−1, each square of the 2d-CDF covers a range of −100 nm < r12, r3 < 100 nm, fiber direction always vertical. The stretching (fiber) direction is indicated in the surface plots by arrows.

Download figure to PowerPoint

Beyond the yield point, the situation changes drastically. First of all it is noticeable that the total scattering intensity as well as the extrapolated intensity I0 increase dramatically suggesting the activity of new strong scatterers—presumably cavities, which exhibit a strong scattering contrast to iPP due to their extremely low electron density. They start as small cracks transversal to the tensile direction and finally deform to long drawn cavities between fibrils. According to their successive growth and hence their broadly distributed dimensions, they hardly show distinct scattering signals, but overlap with the signals arising from the lamellae or their fragments. The growing cavities are also responsible for the whitening of the cold drawn samples, see also Pawlak and Galeski.9 In contrast to the scanning experiments with a microbeam reported by Roth et al.,23 which allows to resolve certain individual voids, here over a large ensemble of cavities is averaged. Within the CDF's, these cavities are reflected by the broad initial decay in fiber direction, at higher strains perpendicular to fiber direction, see Figure 7, CDF's in log scale on the right.

Yet at a strain of 12.5%, the beginning of the yielding region, the structure becomes anisotropic. Although thickness of the lamellae remains constant, they align perpendicular to the tensile direction. The correlation to the next neighbors in transversal direction is lost and concentrated within a 4-point-pattern. This indicates occasionally an internal shear deformation. Under the external stress, the lamellae are sheared and finally break down into certain blocks as described by Strobl and coworkers.7

A new transversal correlation length is established at about 37% strain. By shearing, the lamellar blocks break down to a size which later represents the dimensions of the fibrils. According to the relatively low internal mobility, the aligned chains are not able to form new crystallites. Their mean dimension in transversal direction is with 8.4 nm clearly below the lamellar thickness. This correlates with the WAXS results, which also indicate the final disappearance of the lamellar transversal order during cold drawing. Ultimately, at high strains, this transversal correlation dominates the whole CDF.

The general changes within the oriented chains—within crystallites as well as noncrystallized stretched chains—are shown in the sketch in Figure 8, deformation at room temperature is shown in steps (a–c), (f), and g).

thumbnail image

Figure 8. Sketch of the transformations of the oriented chains during deformation, in-between are additional amorphous chains: (a) Lamellae with some tie-molecules, (b) elastic shear-deformation of lamellae under small load and reorientation with respect to the load, (c) fracture of lamellae into smaller blocks due to local stress concentration caused by tie molecules, (d) stretched aligned, but not recrystallised chains between the bocks during stretching at higher temperatures, (e) some of the fibrillar arranged molecules crystallize, final stage in the case of hot stretched PP, (f) further dissolution of the blocks creating more extended chains at room temperature, and (g) finally, there are several strands of extended chains, not crystallized, with some amorphous regions in between, final stage in the case of cold stretched PP. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

With increasing temperature, the deformation behavior totally changes. The patterns and the CDF's of the stretching of iPP at 130 °C are shown in Figure 9. Here the averaged and the extrapolated maximum intensity I0 does not change dramatically above the yield point. Instead soon the lamellae align in tensile direction increasing their correlation in this direction, indicated by higher order of reflexes in the CDF's. The transversal displacement of the reflexes suggests that the aligned lamellae are shifted against each other. The situation does not change generally also at high strains. This interpretation is also supported by the WAXS results described earlier. The deformation is shown schematically in Figure 8(a–e).

thumbnail image

Figure 9. Deformation of iPP at at 130 °C, from left to right: SAXS patterns, positive and negative CDF's (always log-scale, pseudo color) as well as a surface plot of the CDF's (linear scale) at different strains ε = 0.0, 0.08, and 1.3 (from top). Each square of the pattern covers a range −0.12 nm−1 < s1, s3 < 0.12 nm−1, each square of the 2d-CDF covers a range of −100 nm < r12, r3 < 100 nm, fiber direction always vertical. The stretching (fiber) direction is indicated in the surface plot by arrows. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

With increasing temperature, also the whitening is remarkably reduced indicating that due to the higher mobility of the amorphous phase nearly cavitation is negligible as compared with the extent observed at room temperature.

SAXS of Injection Molded Specimen

In comparison with the compression molded specimens also injection molded specimens were investigated. Samples were cut longitudinally and transversally to the injection direction near the inlet of the plates of 1 mm thickness. The SAXS patterns and the CDF's during deformation are shown in Figures 10 and 11.

thumbnail image

Figure 10. Deformation of injection molded iPP in injection direction at room temperature, from left to right: SAXS patterns, positive and negative CDF's (always log-scale, pseudo color) as well as a surface plot of the CDF's (linear scale) at different strains ε = 0.0, 0.1, 0.81, and 2.81 (from top). Each square of the pattern covers a range −0.12 nm−1 < s1, s3 < 0.12 nm−1, each square of the 2d-CDF covers a range of −100 nm < r12, r3 < 100 nm, fiber direction always vertical. The injection and stretching (fiber) directions are indicated in the surface plot by arrows. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

thumbnail image

Figure 11. Deformation of injection molded iPP transversal to injection direction at room temperature, from left to right: SAXS patterns, positive and negative CDF's (always log-scale, pseudo color) as well as a surface plot of the CDF's (linear scale) at different strains ε = 0 and 0.59 (from top). Each square of the pattern covers a range −0.12 nm−1 < s1, s3 < 0.12 nm−1, each square of the 2d-CDF covers a range of −100 nm < r12, r3 < 100 nm, fiber direction always vertical. The stretching (fiber) direction is indicated in the surface plot by arrows. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

It is apparent that the CDF's of the initial sample are comparable with the compression molded one stretched at 130 °C. The only difference is that there are nearly no small blocks, but larger lamellae aligned vertically to the tensile direction. Here a strong correlation in tensile (that means injection) direction is apparent. Only the characteristic length of the lamellae with 7.24 nm and a long period of 14.1 nm are slightly smaller than in the compression molded specimens. The later evolution of the patterns of the sample in injection direction correlates with the patterns of the cold drawn compression molded sample described earlier, however, the lamellar thickness slightly increases to 14.7 nm. The final fibril width is 5.5 nm.

The sample transversal to injection direction changes directly from the transversal correlation caused by the processing induced orientation to the general fibrillar correlation pattern typical for the stretched compression molded sample, see Figure 7, bottom row. Here the characteristic dimension of the fibrils is about 8.1 nm.

The WAXS patterns of the injection molded specimens show rings representing a random lamellae distribution and some intensity maxima along the rings representing a preferential orientation of some of the lamellae.

Over the whole deformation range stress relaxation is observed as soon as the tensile rig was stopped. However, because this can be ascribed to the amorphous phase, SAXS results remain mostly unaffected.

DSC Investigation of Unstretched and Stretched Specimens

The heating peaks of the first heating curves of the stretched samples are very different, see Figure 12. It is possible to split each curve into two individual peaks whose positions shift with the stretching temperature to higher values. Furthermore, although the high-temperature peak is nearly constant, the low-temperature peak increases with stretching temperature.

thumbnail image

Figure 12. DSC-measurements: First heating curves of compression molded iPP-specimens stretched at different temperatures. For clarification the curves are vertically shifted.

Download figure to PowerPoint

The two peaks indicate that due to stretching at different temperatures, certain different crystallites evolve, lamellar and fibrillar, with different extension. This is in accordance with the scattering results. The lower melting peak should be that of the fibrils. The frozen stresses in the samples stretched at lower temperatures may shift the peaks additionally to lower temperatures.

These curves are very sensitive to the individual piece of the stretched sample. This is also the reason for the outlier of the curve of the 130 °C-stretched sample.

The second heating curves of all stretched specimens are almost identical, but they differ from the second heating curve of the unstretched sample, see Figure 13. Melting of the latter is finished almost 5 °C later. This means that during heating to 230 °C and annealing there for 5 min, the crystallites indeed melt, but the global alignment of the chains remained. Therefore, again a fibrillated crystalline morphology will be obtained by spontaneous crystallization during cooling.

thumbnail image

Figure 13. DSC-measurements: Second heating curves of stretched and unstretched compression molded iPP-specimens.

Download figure to PowerPoint

When comparing the second heating of the samples with the first one shows that the stretched samples have up to 15% higher melting enthalpy, respectively, crystallinity, increasing with the stretching temperature.

SEM of Stretched Specimens

SEM images were taken to verify the described general deformation behavior. Figures 14–17 show the compression molded unstretched reference specimens and specimens stretched at room temperature, 90 °C and 150 °C. The specimens were cut in tensile direction, and vapor deposited with platinum. Thus, stretching direction and s3-axis are always vertical in the figures, the s1-axis horizontal.

thumbnail image

Figure 14. Scanning electron micrograph of the unstretched reference iPP sample.

Download figure to PowerPoint

thumbnail image

Figure 15. Scanning electron micrograph of the iPP sample stretched at room temperature. Stretching direction vertical.

Download figure to PowerPoint

thumbnail image

Figure 16. Scanning electron micrograph of the iPP sample stretched at 90 °C. Stretching direction vertical.

Download figure to PowerPoint

thumbnail image

Figure 17. Scanning electron micrograph of the iPP sample stretched at 150 °C. Stretching direction vertical.

Download figure to PowerPoint

Figures 14–17 show the voids in the stretched samples, the formation of which was discussed earlier. Their size decreases with increasing stretching temperature. In the sample stretched at 150 °C, they are not found. The lamellae and fibrillae are not visible at the present magnification, only a slight texture in stretching direction can be seen. A similar trend was reported recently by Pawlak and Galeski,9 who found voids only at stretching temperatures below 60 °C, at higher temperatures only deformed shperulites and fibrils.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

The experiments revealed the continuously changing mechanisms of elastic and plastic deformation and energy dissipation (the transformations of amorphous and crystalline phases and void formation). Because of the high number of parallel processes, the use of structure characterization by X-ray scattering techniques was realized in wide as well as small angle range and complemented by DSC and by electron microscopy. This combination of methods is likely to provide a well-founded basis for material development and optimization. DSC is an essential extension to characterize crystallization behavior.

It was found that there is a strong temperature-dependent interaction of deformations within the amorphous phase and reorientations as well as transformations of crystalline units. They are strongly determined by the molecular structure and the processing dependent initial morphology of the samples. According to the temperature-dependent mobility within the amorphous phase, the stress transfer to the crystalline phase and so the changes within this phase are also strongly temperature-dependent.

The mechanisms found and discussed here in the case of isotactic polypropylene seem to be very universal. It might also apply to other polymers. This will be checked in the future by additional investigations.

The detailed study of the multiple deformation processes is an essential base for the development of semicrystalline materials with well-defined properties.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

This article is dedicated to M. Stamm in the occasion of his 60th birthday. He had encouraged the author and his group to establish the presented experimental and evaluation environment and to perform the experiments. Furthermore, the author is grateful to HASYLAB for beamtime within the project II-20060086, A. Timmann and St. Roth for local support, and N. Stribeck (University of Hamburg) for the support in the course of data evaluation. Finally, the author thanks W. Jenschke for the software support for the tensile rig, A. Schöne for support in data evaluation, R. Sommerschuh and D. Krause for sample preparation and preliminary tests, S. Frenzel for the DSC measurements, R. Boldt for SEM images, and K. Brüning for revising the manuscript.

REFERENCES AND NOTES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES