The volume of polyolefins produced dwarfs that of all other classes of polymers, due to the low cost of the monomers and the range of mechanical properties that can be achieved.1, 2 Recent catalyst discoveries have enabled the synthesis of multiblock copolymers with high melting point semicrystalline syndiotactic polypropylene (sPP) blocks and rubbery, low glass transition temperature (Tg) ethylene-r-propylene (EPR) blocks.3, 4 These polymers have remarkable mechanical properties, as neat thermoplastic elastomers, as gels in mineral oil, and as “dried gels” prepared from gels from which the mineral oil has been extracted. They can be used as thermoplastic elastomers and can potentially compete in cost and performance with hydrogenated styrenic block copolymers and thermoplastic urethane elastomers.4, 5
The mechanical properties of these semicrystalline block copolymers can be dramatically improved by step cyclic mechanical deformation. The semicrystalline polyolefins sPP have crystalline regions with high melting temperature (Tm) [Tm(sPP) ∼186 °C with [rrrr] = 0.946] higher than polyethylene [Tm(PE) ∼105–130 °C7] but they also have amorphous regions exhibiting a relatively high Tg [Tg ∼0 °C8]. For this reason alone, they are unsuitable as elastomeric materials.9 The random copolymerization of propylene with low Tg monomers such as 1-hexene lowers the Tg but also decreases the crystallinity and, therefore, decreases Tm which is not favorable for high temperature applications.
The development of new metal centered, single site polymerization catalysts have led to the previously unimagined capability to produce multiblock copolymers with high Tm stereoregular polypropylene blocks and low Tg EPR blocks. This enables the formation of block copolymers that have both high use temperatures and excellent elastomeric properties for weight percentage of sPP blocks less than 30%. The properties include a low initial Young's modulus, a high elastic recovery, and a good extension ratio to break.3–5, 10, 11 One major difference between chemically crosslinked elastomers and those elastomers based on the semicrystalline block copolymers is the fact that once a strand in the crosslinked network has been pulled tight and subsequently broken, it no longer can contribute to either the elasticity or the fracture resistance of the network. Because the “crosslinks” in the semicrystalline block copolymer elastomers are crystals that can deform plastically, this plastic deformation could protect the network from total failure and promote a more equal load sharing between network strands. This plastic deformation may be expected to break up the connections between lamellar crystals and transform them into rod-like fibrils. The incremental plastic deformation of the crystals has dramatic effects on the true stress versus true strain curves.
In this article, we investigate the mechanical properties of “dried gels” prepared from triblock copolymers with sPP crystalline blocks. Small angle X-ray scattering and wide angle X-ray scattering, SAXS and WAXS, respectively, are used to determine the morphological and structural changes of the crystals which occur as the material is subjected to step cyclic tensile processing. This structural information allows a more complete understanding of the changes in mechanical properties that occur as a result of the cycling.
The syndiotactic polypropylene-b-ethylene-co-propylene-b-syndiotatic polypropylene triblock copolymer (sPP-EPR-sPP) (Fig. 1) has been synthesized from ethylene and propylene by using stereoselective, living alkene polymerization catalysts. More precisely, a titanium catalyst was used to synthesize the triblock by a sequential polymerization procedure.4 The fluorinated bis(phenoxyimine)-based titanium catalyst system gave highly syndiotactic polypropylenes (sPP, ([rrrr] = 0.96) with very narrow molecular weight distributions (Mw/Mn = 1.1).3, 12
The triblock copolymer studied and called sPP24 contains 24 wt % of hard blocks and has a molecular weight of 298 kg/mol. The sPP has a melting temperature of 134 °C. The ethylene fractions in the EPR block is 0.66 (Table 1).
Table 1. Characterization of the Semicrystalline Triblock Copolymers
Smith et al.13, 14 reported that high molecular weight polyethylene homopolymer can be processed in dilute solution using a diluent such as decalin to produce a monofilament fiber. Subsequent mechanical extension of the gel fiber after removal of the diluents produced highly crystalline fibers with ultra-high strength and high modulus. We used a similar strategy with gels of the semicrystalline polyolefin block copolymer sPP-EPR-sPP prepared in mineral oil (Fisher Chemical, Heavy (USP/FCC) 8042-47-5-paraffin, Fisher Scientific, Fair Lawn, NJ, 07410).
The starting gel comprised of the triblock copolymer with mineral oil as a diluent and was prepared as follows: A mixture of powder of block copolymer in mineral oil was prepared (∼6 wt % sPP-EPR-sPP powder in mineral oil). The mixture was heated in a vacuum oven to 200 °C for about 2 h until a homogeneous blend was obtained, then cooled to room temperature. After this treatment the mixture was a clear and transparent gel. The mineral oil was then extracted by stirring the gel in hexanes for 2 h. Thermogravimetric analyses proved that such a treatment was sufficient to remove all the mineral oil from the gel (Fig. 2).
Many entanglements between EPR chains are removed from the original gel by dilution and will not be able to reform when the diluent is removed (Fig. 3). This loss of entanglements tend to lower the initial modulus compared with the neat polymer. The initial modulus of the dried gel (3.6 MPa) is about three times lower than the initial modulus of the neat polymer (11 MPa). An additional advantage of the dried gels is that, compared with the gels, they have a higher crystallinity, nearly the same as the neat polymers (Fig. 3). From the heat fusion ratio ΔH/ΔHsPP and considering the heat fusion of 100% crystalline sPP to be 207 J/g,15 crystallinities of 12% and 9% for the dried gel and the neat polymer, respectively, have been measured. This enables a more complete characterization of the evolution of the crystalline morphology and microstructure.
Time-resolved SAXS and WAXS experiments in conjunction with uniaxial tensile measurements were carried out using a modified Instron tensile apparatus16 at the X27C beamline in the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory (BNL).
Step Cycle Tensile Testing
Rectangular strips were cut in the polymer gel film. After the removal of the mineral oil, the tensile specimens were about 0.8 mm thick and about 6.5 mm wide.
For the step cycle tensile experiments, we used a load frame that allows the sample to be stretched and unloaded symmetrically so that the same area of the sample is in the X-ray beam during each cycle. The mechanical tests were performed at room temperature. The stretching rate was kept at 0.083 mm s−1 corresponding to an initial strain rate of about 0.006 s−1 with an original clamp-to-clamp length of 15 mm. Step cycle tests were performed so that they combined a stepwise stretching of the tensile specimen with unloading–reloading cycles. The sample was extended step-by-step up to different strains. Once the sample reached the appropriate strain, the crosshead direction was reversed and the sample strain was decreased at the same crosshead velocity until zero stress was achieved. The sample was then extended again at the same constant crosshead speed until it reached the next targeted step-strain. The step cycle test was performed (ε = 0.1, 0.5, 1, 2, 3, 4…, where ε is the nominal tensile strain) until the sample fractured.
X-Ray Scattering Experiments
For the X-ray scattering experiments, the wavelength used was 0.1371 nm. Two different two-dimensional X-ray detectors were used for data collection, a MAR-CCD (MAR USA) for SAXS and FUJI imaging plates for WAXS. A BAS-2500 scanner was used to extract the digital image from the imaging plate. The sample-to-detector distance was 114.7 mm for WAXS and 1937.7 mm for SAXS. An exposure time of 40 s was chosen. This time must be long enough to have a high enough intensity and short enough to capture the evolution of the structure as the sample is deformed. A schematic of the setup is shown in Figure 4.
During the step cycle tensile test, SAXS patterns are recorded continuously, whereas the WAXS patterns are only collected at the maximum strain of each cycle and at the zero load after the unloading as shown in Figure 5.
RESULTS AND DISCUSSION
The mechanical data were analyzed using the true stress (σT = σNL/L0, where σN is the nominal stress, L the actual length of the tensile specimen, and L0 the initial clamp to clamp distance) and the true (or Hencky) strain εH = ln (L/L0). We defined the maximal true strain of each cycle as total true strain or maximal true strain εH. The calculation of the true stress and true strain is facilitated because no necking occurred during the stretching of the sample.
Figure 6 shows the true stress versus true strain curve obtained during the step cycle tensile test performed on the dried gel prepared from the 6 wt % gel of sPP24. The shape of this curve is commonly obtained in the case of highly semicrystalline polymers17–22 and results from the plastic deformation that gives rise to progressively larger residual strain at zero stress. The initial slope of each ascending curve corresponds to the Young's modulus of the specimen and decreases slightly at low strain before stabilizing at higher strains. In contrast, the maximum tangent modulus at higher strains tends to increase as the material is further processed. Similar results are obtained whether the sample is exposed to the X-ray beam or not, indicating that radiation damage is not important. Also the fact that the failure did not occur where the beam hits the sample suggests that the radiation does not greatly influence the mechanical behavior.
The maximum true strain in each cycle can be written as the sum of an elastic component εH,e and a plastic component εH,p. εH,e is the strain recovered after unloading determined from the length of the sample after the stress has been decreased to zero; εH,p is the remaining “plastic” part of the strain.
The mechanical results can be analyzed using the methodology suggested by Strobl and coworkers, for the study of random copolymers of polyolefins23 and of syndiotactic polypropylenes,22 in which the elastic part εH,e and the plastic part εH,p of the true strain are plotted versus the total true strain.18, 22–27 They demonstrated that the mechanical properties can actually be understood as the breakup of a network of lamellar crystals at small strain (collective onset of inter- and intra-lamellar slip processes) followed by the destruction of these lamellar crystals and the formation of crystalline fibrils at higher strain.
Figure 7 shows the evolutions of εH,e and εH,p of the dried gel of sPP24. At low true strain, the elastic true strain increases much faster than the plastic part, until the trend is reversed at higher strain; the curves representing εH,e and εH,p are made of two quasi linear portions. We note that there is some plastic true strain at strains less than point C. From the model of Strobl and coworkers, this plastic strain arises from the breaking down of connections between lamellae as well as from breakup of lamellae into blocks. For most of the elastomers studied, the same trend is observed, namely, a trend similar to that of Strobl and coworkers.18, 22–27 Thus, we decided to denote the critical strain at which the slopes of the curves change, “point C.” This point C was originally defined by Strobl and coworkers as the critical strain where the elastic component reaches a plateau25 and it is believed to correspond to the strain at which the fibril formation sets in.25 From our mechanical and X-ray results, we believe that this conversion is mostly complete by a true strain of about 0.8 so that further strain requires plastic deformation of the fibrils themselves. They are then aligned and plastically deformed in the stretching direction giving rise to the strain hardening.
To address the evolution of the microstructure to the mechanical behavior observed, we conducted some X-ray scattering experiments. The overall crystallinity of the dried gels is quite low because the amorphous EPR block represents about 70% of the volume. Thus, to follow the changes of both the crystal structure and the crystal morphology throughout each step cycle, we used a synchrotron X-radiation source because even in the most favorable cases, exposures of several hours are needed with a conventional X-ray source.
Selected SAXS and WAXS images collected during the step cycle tensile test performed on the dried gel made of 6 wt % of sPP24 in mineral oil are shown in Figures 8 and 9, respectively.
During the first cycles, the SAXS image (Fig. 8) changed from a circular to an oval shape pattern with its long axis perpendicular to the stretching direction. This indicates that the lamellar crystals orient along the tensile direction. Then, a streak develops in the equatorial direction (normal to the tensile direction) and starts to be visible after the third cycle (εH,max = 0.7). As the maximal deformation increases, this diffraction streak becomes thinner and longer. After unloading, the diffraction streak tends to disappear, but not completely and an oval shape with short and broad streaks at the equator is still observed. The copolymer does not return to its initial microstructure, which is consistent with the remaining plastic strain observed (Fig. 6).
The development of the long horizontal intensity streak is believed to be due to the conversion of lamellar crystals into crystalline fibrils that orient under tension along the tensile axis. We used the method proposed by Ruland28 to analyze the length and width of the streak as a function of q to obtain information regarding the average length lf of the fibrils and the fibril misorientation θ from the fiber axis (Fig. 10). We calculated the angular spread Bobs from the azimuthal profiles for values of q in a range between 0.3 and 0.55 nm−1. Bobs is the full width at half maximum of the azimuthal profiles of the equatorial streak fitted with a Lorentzian function. The average fibril length lf and the average misorientation θ have been obtained from the slope and the y-intercept of the Bobs = f(1/q) and qBobs = f(q) curves (see eqs 1 and 2).
The values of the average length lf and of the average misorientation θ reported in Figure 12 are the average values between the two values obtained using both equations (eqs 1 and 2).
The average diameter D has been calculated from the length of the streak qD as 2π/qD, where qD is determined to be the point where the streak intensity is just above the background (Fig. 10).
Figure 11 shows the evolution of the average diameter of the fibrils as a function of the plastic true strain (i.e., plastic true strain after the maximal strain at which the SAXS pattern has been collected; and in the remainder of this article the plastic true strain always corresponds to the plastic true strain after the cycle during which the SAXS pattern has been recorded). For plastic true strains lower than 0.2 (corresponding to the maximal true strain of 0.7), the length, the misorientation, and the diameter were not calculated because the diffraction streak was not observed. The critical plastic true strain at which the streak starts to develop is close to the point C of the Strobl analysis (Fig. 7). These results tend to confirm the Strobl hypothesis that the onset of the fibril formation coincides with a change in slopes of the elastic εH,e and plastic εH,p true strains versus total true strain εH.
For plastic true strain higher than 0.2 and until the failure of the sample, the fibril diameter slightly decreases from 12 nm to 8.6 nm.
The evolution of the average length and of the misorientation of the fibrils as a function of the plastic true strain of the dried gel of sPP24 is shown in Figure 12. The average length increases from 65 nm at the plastic true strain of 0.2 to the value of 250 nm at the plastic true strain of 1.4. Over the same true strain range, the misorientation plateaus at a value of about 2°.
If we assume that the fibrils deform plastically, at constant volume, the average diameter should decrease so that D = DC/exp(εH,p,C/2) and the average length should increase so that l = lCexp(εH,p,C), where εH,p,C is the plastic true strain after point C in Figure 7 and DC and lC are the fibril diameter and length at point C. The predicted exponential evolution of the diameter and the length with the plastic true strain correspond to the dashed lines in Figures 11 and 12, respectively. From the fits, we obtain a value of 14 nm for DC and of 57 nm for lC. The evolution of the average length is quite well-described by the fit, whereas the evolution of average diameter is in reasonable, but not perfect, agreement with the data.
As described earlier, we investigated the evolution of the crystalline fibril morphology using the SAXS results. In the following section, the WAXS results are analyzed to obtain the evolution of the crystalline structure.
In the literature, four crystalline forms of sPP have been identified so far. Using the most recent nomenclature,9, 29, 30 they are called Form I, Form II, Form III, and Form IV. The most stable form is Form I, whereas Form III and Form IV are the less stable. Forms I and II are characterized by chains in helical conformation (T2G2)n,31 whereas forms III and IV present chains in trans-planar and (T6G2T2G2)n32, 33 conformations, respectively.
Form I is characterized by an orthorhombic unit cell having antichiral helical chain packing along both the a- and b-axes. The unit cell dimensions are a = 1.45 nm, b = 1.12 nm, and c = 0.74 nm and α = β = γ = 90°. Lotz and coworkers34, 35 found that sPP crystals grown from the melt have a lath-like rectangular shape with long and short sides parallel to the b- and a-axes of the unit cell, respectively. From electron diffraction the primary growth front has been identified as the b-axis.34 Large transverse crazes normal to the b-axis with fibrils oriented along the b-axis and bridging the two edges of the craze have been observed by transmission electron microscopy (TEM).34 Also, from TEM and atomic force microscopy (AFM) images a sectorization exists in sPP lamellar single crystals with different thickness in the (100) and (010) sectors.35
Form III is characterized by chains in trans-planar conformation packed in an orthorhombic cell with axes a = 0.522 nm, b = 1.117 nm, and c = 0.506 nm. This form is obtained by stretching highly stereoregular sPP samples.9 Vittoria and coworkers36, 37 showed that drawing sPP samples initially in the helical Form I leads first to an orientation of the crystals and then to the transformation of the helical form into the crystalline trans-planar Form III. We also see evidence for this transformation in our dried gel samples from Fourier transform infrared spectroscopy (FTIR) (not shown). On the undeformed sample, both helical and trans-planar bands are present with a smaller fraction of chains in the trans-planar conformation than in the helical conformation. As the sample is stretched, the helical bands decrease, whereas the trans-planar bands increase; more chains in trans-planar conformation are formed during the tensile deformation process. On the sample drawn to a true strain of 2.5 both helical and trans-planar conformations are observed in nearly equal proportions.
The scattering intensity I(ε) at a given strain ε depends on the thickness of the sample tε as I(ε) = I0σtε exp(−μtε), with I0 the incident intensity and σ the vertical section of the sample, perpendicular to the beam. The WAXS patterns have been corrected taking into account the fact that the thickness of the sample decreases as the deformation increases. The selected WAXS patterns shown in Figure 9 have been obtained after the correction.
From the intensity profiles (Fig. 13) and based on the diffraction analysis on single crystals of sPP conducted by Lotz and Lovinger31, 38–41 and on the extensive study of the sPP crystalline structure by Auriemma and coworkers,9, 29, 30, 42, 43 the crystalline form I of the sPP crystals has been identified but the presence of Form II cannot be ruled out.
At low strains, the WAXS patterns (Fig. 9) consist of circular rings. As the sample is deformed, these rings are first transformed into arcs and then into more localized spots. The first strong arc and the third weak arc localized at the equator correspond to the (200)hI (hI standing for helical form I) plane (q = 8.7 nm−1) and to the (400)hI plane (q = 17.5 nm−1), respectively. These equatorial arcs suggest that there is an orientation of the unit cell along the tensile axis. From Figure 9, the sample drawn at a true strain of 0.4 is already well-oriented. Another strong and broad spot is observed at the equator. From the equatorial scan of the intensity profile (the equatorial scan is not shown here), this peak extends from q values of about 12.5 nm−1 to about 15 nm−1 with its maximum between 13.6 nm−1 and 14 nm−1. We have seen that there is a progressive transition from the helical to the trans-planar conformation of the sPP chains as the sample is deformed. Thus, this spot could correspond to the (110)tIII reflection (tIII standing for trans-planar form III) presumably from the drawn fibrils that give rise to the SAXS streak.
In the intensity profile (Fig. 13), an additional peak at q values around 11.2–11.3 nm−1 is observed. This peak corresponds to the arcs which develop at the meridian. We believed that these arcs on the meridian correspond to the (020)hI reflection. If correct, this means that the long axis of the original lath-like crystals (b-axis of the hI cell) become aligned with the tensile direction, a highly unusual orientation indeed. We speculate that this comes about as follows due to the fact that the lath-like lamellae are embedded in a very soft elastomeric matrix. Initially, all orientations of the long axis of the laths are present in the sample. Those crystals whose long axes lie parallel to the tensile axis break up into fibrils first and these fibrils have to continue to plastically deform as the sample is deformed further. However, as the strain increases further, those lath-like crystals whose long axes make larger angles to the tensile axis rotate so that their b-axis lies more nearly parallel to the tensile axis. Eventually, the stress on these off-axis oriented crystals will increase so that they form fibrils but some portion of these (hI) crystals with their b-axes parallel to the tensile axis survive even at a true strain of 2.5 and these crystals give rise to the oriented hI pattern observed.
The effects of the morphological changes in the dried gel are also apparent in a plot of the initial modulus Ei and the maximum tangent modulus Emax as a function of the plastic true strain at the end of the previous cycle shown in Figure 14. The moduli are computed from the slopes of the true stress versus total true strain curves. As the plastic true strain increases from zero to a value between 0.24 and 0.49, corresponding roughly to the plastic true strain of point C (Fig. 7), the initial modulus decreases from about 3.6 MPa to 1.3 MPa. In subsequent cycles, Ei tends to increase slightly to reach a value of 4 MPa during the last loading before the failure of the sample. The initial drop of Ei can be attributed to a breakdown of the initial network of lamellar crystals (Fig. 15).
The maximal tangent modulus was calculated as soon as a change of the curvature of the curve was observed at high strains; this occurs for a value of the plastic true strain similar to that of the point C. Subsequently, the maximum tangent modulus increases exponentially with plastic true strain from 1.14 MPa to 16.3 MPa. The SAXS results suggest that the fibrils, once formed, can be further plastically deformed, thus increasing their aspect ratio. In analogy to the elasticity of short fiber composite, the increase of Emax may result from the reinforcement of the elastomeric network after the formation and orientation of the fibrils whose aspect ratio increases with increasing strain (Fig. 15). Clearly, the reorientation of the lath-like crystals also contributes to Emax.
In this study, we have demonstrated that combined SAXS/WAXS experiments is a powerful tool for achieving a better understanding of the stress versus strain curves obtained during step cyclic tensile tests via the investigation of the evolution of the morphology and the microstructure of the sPP crystals of semicrystalline polyolefin block copolymers. From the SAXS results, we have shown that as the deformation increases lamellar crystals are transformed into thin and long crystalline fibrils oriented in the stretching direction and that the true strain at which the onset of the fibril formation occurs corresponds to the true strain of the point C of the Strobl analysis. The WAXS results showed that there is also an orientation of the crystalline unit cell of the lath-like crystals as the sample is stretched. The methodology outlined here will be applied in the future to better understand the elastomeric properties of other semicrystalline thermoplastic elastomers, such as block copolymers with iPP blocks as the hard blocks.
The authors acknowledge for the support from the Mitsubishi Chemical Center for Advanced Materials at UCSB. They also thank the critical comments of Christian Burger and Yimin Mao of Stony Brook that improved the manuscript. We made use of the UCSB MRL Central Facilities funded through the NSF MRSEC Program (DMR 0520415) and the Cornell Center for Materials Research Shared Experimental Facilities funded through the NSF MRSEC Program (DMR 0520404). Use of the National Synchrotron Light Source, Brookhaven National Laboratory, was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-98CH10886.