Reactive blending of thermoplastic polyurethane in situ with poly(vinyl chloride).
Blends of PVC with other polymers, used to produce permanently plasticized or impact modified compounds in the absence of monomeric liquid plasticizers, have significant commercial significance but are often difficult to mix on conventional PVC processing equipment and command significant premiums. Of these blend systems, PVC/TPU blends are described in the literature, but are not well known commercially. Remarkable abrasion resistance, permanence, low temperature flexibility, and impact resistance are claimed for such blends, which are said to be good candidates for cable jacketing, tubing, hose, and shoe sole applications [1-4].
Although it is possible to melt blend PVC and softer grades of TPU, the high processing temperatures (approximately [greater than or equal to]190[degrees]C) and melt viscosities required to homogenize such blends often result in partial PVC degradation, even in the presence of thermal stabilizers . Moreover, many TPUs are not completely miscible with PVC, making it difficult to produce desirable blend morphologies with conventional melt blending techniques. While solution blending eliminates many of the problems associated with melt blending, it is not a commercially viable mixing technique due to solvent removal. Consequently, to overcome the processing and mixing difficulties associated with these blend systems, it would be advantageous to develop an alternative blending technique.
In the synthesis of semi-interpenetrating polymer networks  and dynamic vulcanizates , polymer blending is often achieved by mixing an already polymerized high polymer with miscible monomers or oligomers of a second polymer that are subsequently polymerized. Such a reactive blending technique offers an attractive, alternative method for blending PVC and TPUs. More specifically, preblending PVC with TPU monomers with which it is miscible and the subsequent polymerization of high molecular weight TPU in situ with PVC under a set of processing conditions benign to PVC overcomes the previously mentioned melt blending problems. Several other advantages that such a reactive blending technique might offer include reduced costs and a means of controlling final blend morphology.
PVC/TPU miscibility is dependent on TPU chemical structure and thus formulation. According to the literature, poly(butylene adipate) (PBA) polyesters are completely miscible with PVC [8-10]. However, TPUs with PBA based polyol soft segments (SS) and 1,4-butanediol/4,4'-diphenylmethane diisocyanate (BDO/MDI) based hard segments (HS) are reported to be partially miscible with PVC [11-13] if their HS content is low. A PBA based TPU with a low MDI/BDO HS content (approximately 23%) is used throughout this study. When reactive blended with PVC, this TPU formulation allows for miscible PVC/TPU reactant mixtures prior to in situ TPU polymerization and partially miscible PVC/TPU blends afterward.
The primary objective of this study is to develop a novel reactive blending process for producing PVC/TPU blends. Furthermore, the PVC/TPU reactive blending process can be broken down into two fundamental stages: the compounding and/or plasticization of PVC with PBA and BDO (stage 1), and, upon adding MDI, the subsequent polymerization of TPU in situ with PVC to produce a PVC/TPU blend (stage 2). Hence, the second objective of this study is to characterize the morphology of PVC/PBA/BDO blends at the completion of stage 1 of the PVC/TPU reactive blending process. Using optimal processing conditions, the third objective of this study is to characterize the morphology of PVC/TPU reactive blends at the completion of stage 2 of the process. Finally, the fourth objective of this study is to characterize the tensile properties of PVC/TPU reactive blends.
The poly(vinyl chloride) (PVC) homopolymer used in this study was obtained from the Oxy Vinyls company, formerly Geon. According to the supplier's specifications, this injection molding grade, suspension polymerized PVC resin had a number average and weight average molecular weight of approximately 30,000 g/mol and 63,000 g/mol, respectively. Except where noted, PVC was always stabilized with 1 wt% of the tin mercaptide, dibutyltin bis(2-ethylhexyl mercaptoacetate) (T31), obtained from Elf Atochem. In the PVC industry, thermal stabilizers are usually added to PVC with a process called dry blending. In this process, a high intensity, nonfluxing mixer such as a Henschel mixer is used to disperse the stabilizer with PVC under extremely high shear. On a smaller scale, domestic liquid blenders used in the kitchen, also known as a Waring type blenders, can be used for the same purpose. In this study, such a device was used to dry blend T31 with PVC prior to compounding.
It was found that in addition to functioning as a thermal stabilizer for PVC, T31 functioned as a very powerful catalyst for TPU polymerization in stage 2 of the PVC/TPU reactive blending process. This is not surprising since T31, being a tin mercaptide, is quite electrophilic and very similar in structure to the commonly used tin carboxylate TPU catalyst, dibutyltin dilaurate. Above a critical T31 concentration of approximately 0.05 wt%, linear, high-molecular weight TPU was polymerized in situ with PVC as evidenced by PVC/TPU reactive blend solubility in tetrahydrofuran and gel permeation chromatograph measurements. However, T31 concentrations of less than approximately 0.05 wt% resulted in PVC/TPU reactive blends that were only partially soluble in tetrahydrofuran. Soxlet solvent extraction experiments coupled with DSC and Raman spectroscopy revealed that the insoluble gel fraction of these blends was TPU with a larger HS content than the original formulation. It is speculated that at low T31 concentrations, low rates of TPU polymerization favor HS phase separation from the PVC/TPU reactant mixture before complete conversion is achieved. This in turn results in low molecular weight TPU with poor SS/HS phase connectivity and large HS crystalline domains that are insoluble in tetrahydrofuran (THF). The same phenomenon has been observed in uncatalyzed polyether based TPUs [14-16].
The soft segment of the TPU used in this study was an oligomer diol of PBA supplied by Bayer Material Science LLC. According to the certificate of analysis from Bayer, this diol had a number average molecular weight of approximately 2000 g/mol. The hard segment of the TPU was derived from flaked MDI and BDO supplied by Bayer and ARCO chemical company, respectively. Synthesized with equimolar quantities of PBA and BDO, the TPU contained 76.84% PBA, 3.536% BDO, and 19.62% MDI by mass (based on PBA with an equivalent number average molecular weight of 979.1 g/mole) when the stoichiometric ratio of hydroxyl to isocyanate functionality was maintained at unity. However, MDI was always used in 2% excess of stoichiometry in an effort to compensate for trace amounts of residual water in the reactants.
In preparation for TPU synthesis, PBA was melted and dried under a vacuum at 100[degrees]C for a minimum of 4 hours while BDO was dried over type 3A molecular sieves at room temperature for at least 2 weeks prior to synthesis. MDI was used as received but was stored under a vacuum at 0[degrees]C until required for synthesis. In all studies in which TPU was polymerized in situ, PVC was dried under a vacuum at 60[degrees]C for 12 hours to remove trace quantities of water.
Processing Equipment and Procedures
All studies of the PVC/TPU reactive blending process were performed on a C. W. Brabender batch type internal mixer operated at 50 rpm. Equipped with cam style rotors, this mixer had an internal mixing volume of approximately 82 [cm.sup.3]. Linked to a data acquisition software system, it was capable of measuring both temperature and torque as a function of mixing time. The bowl of the mixer was blanketed with dry nitrogen gas at all times. The internal mixer was always operated with a fill factor of 85% by volume and a mixing temperature of 120[degrees]C. This fill factor/temperature combination was used because it yielded homogenous PVC/PBA blends at all compositions in a reasonable amount of time. In addition, a mixing temperature of 120[degrees]C minimizes PVC dehydrochlorination and is a favorable temperature for high molecular weight TPU polymerization. In this study, a PVC/PBA blend was considered homogeneous when the suspension polymerized grains of PVC were destroyed.
In stage 1 of the PVC/TPU reactive blending process, three different PVC/PBA/BDO blend compositions were compounded in preparation for MDI addition and subsequent TPU polymerization in situ with PVC. Initially, stabilized PVC was added to the preheated internal mixer and allowed to equilibrate for 5 minutes. Next, at the start of the mixing sequence, PBA was added in proportions required to yield PVC/PBA blend compositions of 25/50, 50/50, and 75/25 wt%. After 9 minutes of mixing time, all PVC/PBA blend compositions were homogenized as measured by an increase in mixing torque. Next, stoichiometric quantities of BDO (depending on composition) were added to yield PVC/PBA/BDO blend compositions of approximately 24/73/3, 49/49/2, and 74/25/1 wt%. Finally, after 15 minutes of total mixing time, BDO was adequately incorporated into all three PVC/PBA/BDO blend compositions and stage 1 of the PVC/TPU reactive blending procedure was considered complete. If required, samples were removed from the internal mixer and characterized within 2 days.
[FIGURE 1 OMITTED]
In stage 2 of the PVC/TPU reactive blending process, stoichiometric quantities of MDI were added to the PVC/PBA/BDO blend compositions produced in stage 1 of the process. MDI was added and allowed to incorporate until a maximum in mixing torque was observed (approximately 3 minutes as shown in Fig. 1), indicating that TPU polymerization in situ with PVC was essentially complete. The resulting PVC/TPU reactive blends, with compositions of 20/80, 44/56, or 70/30 wt%, were immediately removed from the internal mixer and characterized within 2 days.
For purposes of comparison, neat TPU of the same formulation present in PVC/TPU reactive blends was also synthesized in this study. In this procedure, dewatered PBA (heated to 100[degrees]C), BDO (at room temperature), and MDI (at room temperature) were gravimetrically metered into a 500-ml polypropylene beaker and vigorously hand mixed for 15 seconds. The resulting TPU reactant mixture was immediately cast on aluminum trays coated with a mold release agent and fully polymerized at 110[degrees]C for 16 hours in a normal convection oven.
Characterization Equipment and Procedures
A JEOL model JEM 100 CX II transmission electron microscope (TEM) was used to characterize the morphology of all blends investigated. Ultrathin ([less than or equal to]100 nm) sections were microtomed from samples with the aid of liquid nitrogen. These sections were placed on 400 mesh copper grids and analyzed with at various magnifications. Due to the electron dense nature of PVC, samples were not stained.
A Thermal Advantage 2920 modulated differential scanning calorimeter (DSC) was used to characterize the glass transition temperature ([T.sub.g]) and/or crystalline melting behavior in all blends investigated. Before analysis, samples were carefully weighed to 10 [+ or -] 2 mg and sealed in aluminum hermetic pans and lids. After placing encapsulated samples in the DSC, all temperature scans were performed with a heating rate of 20[degrees]C/min. PVC/PBA/BDO blends were scanned over a temperature range of -100 to 120[degrees]C, while PVC/TPU reactive blends were scanned over a temperature range of -100 to 200[degrees]C. Upon completing one temperature scan, all blend samples were immediately quenched to -100[degrees]C in liquid nitrogen and a second temperature scan was performed.
A Kaiser Optical Systems Series 5000 Holoprobe Raman spectrometer was used to characterize any hydrogen bonding behavior existent in all blends investigated. Equipped with a thermoelectrically cooled charge coupled device (CCD) detector, the system was capable of collecting spectra over a Raman shift spectral range of approximately 300 to 3300 [cm.sup.-1]. A 100 mW, 785 nm GaAlAs diode laser was used as the excitation radiation source. Using a 180[degrees] backscattering Raman measurement geometry, samples approximately 1.0 mm in thickness were sandwiched between fused quartz cover slips and placed in a hot stage. Upon aligning the aperture of the hot stage, and thus the sample, with the focused Raman laser beam, all spectra were acquired at room temperature. To obtain an acceptable signal to noise ratio, all Raman spectra were the result of co-adding 100 individual scans. Each scan had a resolution of approximately 5 [cm.sup.-1] and was acquired over 30 seconds of exposure time. All PVC/PBA/BDO blend samples were quenched to room temperature from 120[degrees]C immediately prior to characterization.
In preparation for analysis, all Raman Spectra were processed with several chemometric spectral manipulation techniques using Grams/386 software from Galactic Industries Corp. In order to remove Raleigh/fluorescence induced background scattering, a best-fit, fourth order, polynomial baseline was subtracted from all spectra. Because Raman spectroscopy is a single beam method and because the number of scattering sites can never be known in the analysis of solids, all Raman spectra were normalized with respect to an internal standard. For PVC/TPU reactive blend samples, peak height of the 1616 [cm.sup.-1] band was used. This band was the result of aromatic ring breathing/stretching vibrational modes present in the phenylene groups of MDI. Unfortunately, a suitable internal standard was not available for normalizing PVC/PBA/BDO blend samples. Therefore, spectra for all compositions were normalized in such a way that peak area of the 1730 [cm.sup.-1] band remained constant. This band originated from carbonyl stretching vibrations present in the ester groups of PBA.
A Waters 150-C gel permeation chromatograph (GPC), in tandem with refractive index, light scattering, and solution viscosity equipment, was used to measure the molecular weight and molecular weight distribution of all PVC/TPU reactive blend systems that were completely soluble in tetrahydrofuran. The GPC component utilized five Ultrastyragel columns with tetrahydrofuran as the carrier solvent.
A Flexsys Tensometer 2000 with a 1-kN load cell and extensometer was used to make uniaxial stress-strain measurements on all blends investigated. In preparing tensile specimens, all blend samples were compression molded into 1 mm thick sheets at 190[degrees]C for 2 minutes. Then, in accordance with ASTM D638, type V dumbbell-shaped tensile specimens were cut out of each sheet. All measurements were made at room temperature with a crosshead speed of 50 mm/min and a gauge length of 10.287 mm. In determining the elastic modulus, yield stress, 100% strain secant modulus, stress at failure, and strain at failure for a particular sample, measurements from five specimens were averaged.
RESULTS AND DISCUSSION
Stage 1 of the PVC/TPU Reactive Blending Process
In stage 1 of the PVC/TPU reactive blending process, PVC is compounded and plasticized with PBA and BDO in preparation for MDI addition and subsequent TPU polymerization in situ with PVC. Figure 1 shows characteristic mixing torque versus mixing time curves from stage 1 of the PVC/TPU reactive blending process. For the purposes of this study, compounding PVC/PBA/BDO blends in a batch type internal mixer would ideally result in one-phase blend morphologies prior to MDI addition. This expectation is actually quite realistic when one considers that amorphous PVC/PBA blends are reported to be miscible at all compositions and experimentally accessible temperatures [8-10]. However, suspension polymerized PVC is unique in that it exhibits a particulate structure hierarchy that extends well below length scales of 1 [micro]m . Because of this particulate morphology, compounding PVC with PBA and BDO at relatively low mixing temperatures (below approximately 160[degrees]C) may not result in one-phase blend morphologies.
The state of PVC/PBA/BDO blend morphology at the completion of stage 1 of the PVC/TPU reactive blending process was characterized with TEM. Figure 2 shows TEM micrographs of all three PVC/PBA/BDO blend compositions removed from the internal mixer after 15 minutes of mixing time. The darker regions in the micrographs represent phases rich with PVC. A result of PBA and BDO plasticization, and the stresses imposed by the mixing process, the PVC/PBA/BDO = 74/25/1 wt% blend composition is characterized by the almost complete destruction of PVC particulate microstructure. Vestiges of PVC microstructure may still exist as evidenced by nodular structures on the order of 0.1 [micro]m in diameter, but this structure could be an artifact left over from microtoming. Relative to the PVC/PBA/BDO = 74/25/1 wt% blend composition, the PVC/PBA/BDO = 49/49/2 wt% blend composition is much more heterogeneous. This blend composition is characterized by broken up remnants of PVC primary particles less than 1 [micro]m in diameter. Analogously, primary particles are also observed in the PVC/PBA/BDO = 24/73/3 wt% blend composition. However, the primary particles appear to be relatively intact and dispersed evenly throughout a continuous phase rich in PBA and BDO. Plasticized and swollen with PBA and BDO, these primary particles are approximately 1-2 [micro]m in diameter.
Figure 3 shows DSC temperature scan traces of different PVC/PBA/BDO blend compositions. In an effort to minimize PBA crystallization, only second-run DSC temperature scan traces are shown. These traces correspond to samples quenched in liquid nitrogen from a temperature of approximately 120[degrees]C. For reference, DSC temperature scan traces of neat PVC and PBA are also shown in the Fig. 3. PVC exhibits a [T.sub.g] of 83[degrees]C while PBA has a [T.sub.g] of -66[degrees]C. Both the PVC/PBA/BDO = 74/25/1 and 49/49/2 wt% blend compositions display one, broad [T.sub.g], intermediate between those of pure PVC and PBA. While the PVC/PBA/BDO = 24/73/3 wt% blend may also exhibit a single [T.sub.g], the results are inconclusive since rapid quenching failed to yield a completely amorphous blend as evidenced by the presence of a broad PBA melting endotherm in the temperature regime of interest. While not proof of miscibility in terms of a thermodynamically stable single phase, the appearance of a single, composition dependent [T.sub.g] suggests extensive phase mixing and that the scale of any phase separation and dispersion approaches the scale of cooperative segmental motion responsible for the glass transition. The unusually broad [T.sub.g]s observed in Fig. 3 are indicative of miscible blends with local composition fluctuations in excess of normal density and temperature fluctuations, or blends with a broad distribution of finely dispersed phases with varying composition . In other words, broad [T.sub.g]s of this nature usually indicate some degree of heterogeneity. This assessment is in agreement with the TEM results discussed earlier.
[FIGURE 2 OMITTED]
Figure 4 shows room temperature acquired Raman spectra of different PVC/PBA/BDO blend compositions. In an effort to minimize PBA crystallization, all spectra were acquired from samples quenched to room temperature from a temperature of 120[degrees]C immediately prior to characterization. Specifically, the carbonyl stretching vibrations in the ester groups of PBA were analyzed for changes in Raman shift and/or intensity that would indicate the presence of intermolecular interactions with PVC. The PVC/PBA/BDO = 74/25/1 and 49/49/2 wt% blend compositions exhibit broad carbonyl bands relative to that observed for the PVC/PBA/BDO = 24/73/3 wt% blend composition and pure PBA. Consistent with the DSC results discussed earlier, an increase in conformational freedom and the resulting band broadening indicates that PBA is amorphous in blend compositions with high and intermediate concentrations of PVC. In addition, similar to Fourier Transform Infrared spectroscopy (FTIR) results obtained from analogous PVC/PCL blends , a progressive shift to lower wavelength and broadening of the carbonyl band associated with PBA is observed as PVC concentration increases. This behavior reveals the presence of specific intermolecular interactions (e.g., hydrogen bonding) between PVC and the carbonyl groups of PBA, which could play a significant role in the apparent miscibility of these blend systems when PBA is amorphous. The results obtained from Fig. 4 indicate that at a processing temperature of 120[degrees]C, PVC/PBA/BDO blends are intimately mixed at the molecular level.
[FIGURE 3 OMITTED]
The results indicate that PVC/PBA/BDO blends are miscible at the processing temperature of 120[degrees]C. Why then, do some of the PVC/PBA/BDO blend compositions exhibit heterogeneous blend morphologies at the conclusion of stage 1 of the PVC/TPU reactive blending process? At the processing temperature of 120[degrees]C used in this study, amorphous regions of PVC can mix with PBA via self-diffusion of the latter because they are well above their [T.sub.g] (approximately 83[degrees]C) and are miscible with PBA. When PBA diffuses into PVC and functions as a plasticizer, amorphous PVC molecules that were rigidly held together by intermolecular bonding are more easily disentangled and free to move when exposed to stress. The same explanation can also hold true for BDO. However, according to the supplier, the suspension polymerized PVC homopolymer used in this study contained a small amount of crystallinity due to small syndiotactic sequences in the main chain. Therefore, recognizing that the crystalline regions of PVC have the wide melting range of 120-210[degrees]C  and act as physical cross-links [21, 22], PVC/PBA/BDO blend phase mixing is hindered. These crystalline regions of PVC, also known as microcrystallites, limit the rate and extent to which PVC microstructure destruction can take place. The degree of PVC microstructure breakup and thus homogeneity achieved in these blends is also related to the composition. For the PVC/PBA/BDO = 49/49/2 wt% blend composition and especially the 24/73/3 wt% blend compositions, PBA and BDO exist as the continuous phase. With a relatively low melt viscosity, this phase limits the amount of mechanical stress that can be applied to PVC in the mixing process. Hence, one-phase PVC/PBA/BDO blend morphologies are not realized under this set of processing conditions.
[FIGURE 4 OMITTED]
Stage 2 of the PVC/TPU Reactive Blending Process
In stage 2 of the PVC/TPU reactive blending process, the addition of MDI to PVC/PBA/BDO blends from stage 1 of the process and subsequent TPU polymerization in situ with PVC result in PVC/TPU reactive blends. Figure 1 shows characteristic mixing torque versus mixing time curves from stage 2 of the PVC/TPU reactive blending process. From the rapid rise in measure mixing torque for all PVC/TPU reactive blend compositions, it is apparent that TPU polymerization in situ with PVC occurs quite rapidly and appears to be finished 3 minutes after MDI addition, or 18 minutes of total mixing time. This is consistent with the measured values of molecular weight for these blends shown in Table 1. The values given represent the average of both PVC and TPU for each blend composition. Due to changes in system miscibility upon chain extending PBA with MDI/BDO HSs and thus polymerizing a relatively high molecular weight TPU in situ with PVC, reaction induced phase separation may occur in stage 2 of the PVC/TPU reactive blending process. In support of this, TPUs with PBA SSs and MDI/BDO HSs are reported to exhibit partial miscibility and multiphase behavior with PVC [11-13]. Consequently, both TPU polymerization in situ with PVC and possibly reaction induced PVC/TPU blend phase separation are recognized as occurring in stage 2 of the PVC/TPU reactive blending process.
In an effort to characterize the state of PVC/TPU reactive blend morphology at the completion of stage 2 of the PVC/TPU reactive blending process, Fig. 5 shows TEM micrographs of all three PVC/TPU reactive blend compositions removed from the internal mixer after 18 minutes of total mixing time. The darker regions in the micrographs presumably represent regions rich in PVC but could also represent regions rich with TPU HSs. Under the former assumption, morphology of the PVC/TPU = 70/30 wt% reactive blend composition consists of TPU rich phases approximately 100 nm in diameter dispersed in a PVC-rich continuous phase. Conversely, the morphology of the PVC/TPU = 44/56 wt% reactive blend composition is characterized by PVC rich phases approximately 1 [micro]m in diameter dispersed throughout a TPU-rich continuous phase. Therefore, at some blend composition between that of the PVC/TPU = 70/30 and 44/56 wt% blend compositions, phase inversion has occurred. The morphology of the PVC/TPU = 20/80 wt% reactive blend composition is also characterized by PVC rich phases approximately 1 [micro]m in diameter dispersed throughout a TPU rich continuous phase. However, in this composition the correlation length between dispersed phases is larger than in the PVC/TPU = 44/56 wt% reactive blend composition. In all compositions, a diffuse interface exists between the dispersed phase and the continuous phase, suggesting significant phase mixing. In addition, all compositions appear to display a broad distribution of phase sizes and compositions.
Figure 5 indicates that PVC/TPU reactive blends exhibit heterogeneous, multiphase blend morphologies with significant phase mixing. This blend morphology is not a PVC microstructure remnant from stage 1 of the PVC/TPU reactive blending process. Blend viscosities in stage 2 of the PVC/TPU reactive blending process are much higher than those observed in stage 1 of the PVC/TPU reactive blending process (see Fig. 1). Therefore, the correspondingly high shear and normal stresses destroy the majority of PVC microstructure remaining from stage 1 of the PVC/TPU reactive blending process. The heterogeneous morphology observed for PVC/TPU reactive blends in Fig. 5 is a product of reaction-induced PVC/TPU blend phase separation. In support of this, TEM was used to characterize the PVC/TPU = 44/56 wt% reactive blend sample shown in Fig. 5b both 1 hour and 1 day (as shown) after preparation. Although not shown here, the former exhibited similar yet finer blend morphology than the latter.
Figures 6 and 7 show first and second runs of DSC scans, respectively, of all three PVC/TPU reactive blend compositions. Except for room temperature aging/annealing effects, blend samples in Fig. 6 came directly from stage 2 of the PVC/TPU reactive blending process with no further thermal treatment. For reference, DSC temperature scan traces of neat PVC and TPU are also shown in Fig. 6. PVC exhibits a [T.sub.g] of 83[degrees]C while TPU has a [T.sub.g] of -34[degrees]C. It appears that two broad [T.sub.g]s intermediate between those of neat PVC and TPU exist for every PVC/TPU reactive blend composition in Fig. 6. Initially, it was speculated that the high temperature thermal transition for each composition was actually an endotherm corresponding to the melting of a crystalline SS polymorph in the TPU component of the blends . However, based on the TEM evidence displayed in Fig. 5 and the relatively large change in specific heat flow observed for the single blend [T.sub.g]s observed in Fig. 7, it was concluded that this thermal transition was indeed a [T.sub.g]. Two blend [T.sub.g]s intermediate between those of neat PVC and TPU is indicative of blend morphology characterized by two major phases and is evidence for a partially miscible blend system in which the two components spontaneously mix yet do not form a single, thermodynamically stable phase. This assessment is in agreement with the TEM results discussed earlier. Interestingly, some degree of HS phase separation and domain ordering is present in the high TPU content blends.
[FIGURE 5 OMITTED]
The second-run DSC temperature scan traces in Fig. 7 represent PVC/TPU reactive blend samples that were quenched in liquid nitrogen from a temperature of approximately 200[degrees]C before characterization. The idea here was to minimize TPU SS crystallization and investigate the effect of high temperature thermal histories on PVC/TPU reactive blend phase morphology. For reference, DSC temperature scan traces of neat PVC and TPU are also shown in Fig. 7. PVC exhibits a [T.sub.g] of 83[degrees]C while TPU has a [T.sub.g] of -28[degrees]C. As shown in Fig. 7, every PVC/TPU reactive blend composition displays one broad [T.sub.g] intermediate between those of neat PVC and TPU. Apparently, the degree of phase separation and dispersion in this blend system has decreased. In confirmation of this assessment, blend samples in Fig. 6 were turbid before characterization while blend samples in Fig. 7 were optically transparent. Relative to the heterogeneous, multiphase morphology noted for PVC/TPU reactive blends at room temperature, high temperature annealing results in finer, if not one-phase, PVC/TPU reactive blend morphologies. Nevertheless, the [T.sub.g]s in Fig. 7 are still rather broad and may suggest that some blend heterogeneity still exists. It is speculated that the dependence of PVC/TPU reactive blend morphology on temperature is the result of upper critical solution temperature phase behavior. However, a change in blend morphology may also be related to TPU degradation and the corresponding increase in blend miscibility. Several publications [24-26] have shown that urethane bond thermal dissociation becomes significant at 190-200[degrees]C for analogous TPU formulations.
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
Figure 8 shows room temperature acquired Raman spectra of all three PVC/TPU reactive blend compositions. Except for room temperature aging/annealing effects, blend samples came directly from stage 2 of the PVC/TPU reactive blending process with no further thermal treatment. Specifically, the carbonyl stretching vibrations in the urethane and SS ester linkages of TPU were analyzed for changes in Raman shift and/or intensity that would indicate the presence of intermolecular interactions with PVC. Two overlapping carbonyl bands with Raman shifts of approximately 1700 and 1730 [cm.sup.-1] are clearly present in all blend compositions. The former band actually represents two species of stretching vibration corresponding to urethane and SS ester carbonyls H-bonded with donor N-H groups of adjacent urethane linkages. In the same way, the latter band represents two species of stretching vibration corresponding to urethane and SS ester carbonyls free of intermolecular interactions [11, 27, 28]. Of course, this simple description of carbonyl stretching vibrations in TPU assumes any carbonyl band splitting resulting from specific intermolecular interactions with PVC is negligible. Figure 8 shows that the relative intensity of the 1700 [cm.sup.-1] composite band to the 1730 [cm.sup.-1] composite band decreases with increasing PVC composition. At the expense of H-bonded carbonyls, the concentration of free carbonyls increases with increasing PVC composition. This suggests the presence of specific intermolecular interactions between PVC and the carbonyl groups of TPU, and that phase mixing in PVC/TPU reactive blends occurs at the molecular level. However, the relative intensity of the 1700 [cm.sup.-1] composite band to the 1730 [cm.sup.-1] composite band for the PVC/TPU = 20/80 wt% reactive blend composition is nearly the same as that for neat TPU. Hence, this blend composition may be more phase separated than the others or simply has less PVC to significantly interrupt TPU carbonyl H-bonding. A combination of both scenarios is probably the case here.
[FIGURE 8 OMITTED]
The results in Figs. 5-8 indicate that PVC/TPU reactive blends taken directly from stage 2 of the PVC/TPU reactive blending process are partially miscible at room temperature. Not a result of PVC microstructure remnant from stage 1 of the PVC/TPU reactive blending process, they are characterized by heterogeneous, multiphase morphologies, with all phases containing a certain amount of both polymers. As in neat TPU, some degree of HS phase separation (from the blends) and domain ordering is present in high TPU content blends. Similar results have been reported in the literature for analogous solution blended systems [11-13]. Therefore, chain extending PBA with MDI/BDO HSs in miscible blends of PVC, PBA, and BDO, and thus polymerizing a relatively high molecular weight TPU in situ with PVC, results in reaction induced PVC/TPU blend phase separation. Such phase separation is caused by an increase in TPU molecular weight, and thus a decrease in the combinatorial entropy of mixing, coupled with variations in the enthalpy of mixing with increasing conversion. To state the latter effect in other terms, MDI/BDO based TPU HSs are speculated to be immiscible with PVC. This is consistent with other studies pointing out that with increasing TPU HS content, a TPU's miscibility with PVC decreases [29-32]. It is also consistent with the significant difference in solubility parameters reported for these two materials (e.g., [delta] = 11.5-13.2 and 9.4-10.8 (cal/[cm.sup.3])[.sup.0.5] for MDI/BDO HSs [33, 34] and PVC , respectively). However, as evidenced by significant phase mixing, complete phase separation does not occur in this blend system.
Acquired at room temperature, typical engineering tensile stress-strain curves from different PVC/TPU reactive blend compositions are shown in Fig. 9. Their average engineering tensile properties are listed in Table 2. Neat PVC, with a [T.sub.g] well above room temperature, shows characteristic yielding and necking behavior before it fails at relatively small strains. The PVC/TPU = 70/30 wt% reactive blend composition is similar to neat PVC in that it has a well defined yield point prior to which elastic deformation occurs. This was expected since its [T.sub.g] is well above room temperature as evidenced by DSC. In addition to necking instability beyond the yield point, this composition exhibits strain-hardening behavior that does not exist for neat PVC. Both the PVC/TPU = 44/56 and 20/80 wt% reactive blend compositions are similar to neat TPU in that they exhibit monotonic, elastomeric, stress-strain behavior prior to failure without forming a neck during elongation. This was expected since both compositions' major blend [T.sub.g] is below room temperature as evidenced by DSC. Such behavior is a product of coherent blend phase morphology in which high melting TPU HS domains act as physical cross-links for the low [T.sub.g], deformable, PVC/TPU SS matrix, giving high elasticity and strength to the resulting blends. In general, all three PVC/TPU reactive blend compositions exhibit stress-strain behavior intermediate between that of neat PVC and TPU. In particular, elastic modulus and 100% strain secant modulus decrease with increasing TPU content while stress and strain at failure increase. Moreover, unlike PVC plasticized with conventional low molecular weight plasticizers such as dioctyl phthalate , modulus and strain at failure do not decrease and increase, respectively, at the expense of decreasing stress at failure.
[FIGURE 9 OMITTED]
The results of TEM, DSC, and Raman spectroscopy indicated that PVC/PBA/BDO blends produced at the conclusion of stage 1 of the PVC/TPU reactive blending process were characterized by heterogeneous morphologies. While it appears that PBA and BDO were completely miscible with PVC at a processing temperature of 120[degrees]C, remnants of the particulate structure hierarchy in suspension polymerized PVC remained in the blends. However, the degree of heterogeneity increased in blends rich with PBA and BDO. In these blends, PBA and BDO predominantly exist as the continuous phase in the early stages of the mixing process. With a relatively low melt viscosity, this phase limits the amount of mechanical stress that can be applied to the PVC rich dispersed phase. This coupled with microcrystallites in PVC limits the rate and extent to which PVC microstructure destruction can take place under this set of processing conditions.
The results of TEM, DSC, and Raman spectroscopy showed that PVC/TPU reactive blends produced at the completion of stage 2 of the PVC/TPU reactive blending process were characterized by heterogeneous, multiphase morphologies at room temperature. Such morphologies displayed a broad distribution of phase sizes and compositions, with all phases containing a certain amount of both polymers. As in neat TPU, some degree of HS phase separation (from the blends) and domain ordering was present in high TPU content blends. Not the result of PVC microstructure remnant from stage 1 of the PVC/TPU reactive blending process, the above morphological picture for the PVC/TPU reactive blend system is evidence for partially miscibility where the two components spontaneously mix yet do not form a single, thermodynamically stable phase at room temperature. Therefore, chain extending PBA with MDI/BDO HSs in miscible blends of PVC, PBA, and BDO, and thus polymerizing a high molecular weight TPU in situ with PVC, results in reaction induced PVC/TPU blend phase separation.
Depending on composition, PVC/TPU reactive blends exhibited excellent tensile properties intermediate between that of neat PVC and TPU. This is consistent with the high values of molecular weight measured for these blends and with their coherent blend phase morphologies. In conclusion, the primary objective of developing a novel reactive blending process for the production of viable PVC/TPU blends has been achieved.
TABLE 1. Mean molecular weight of different PVC/TPU reactive blend compositions. PVC/TPU blend [M.sub.n] [M.sub.w] [M.sub.w]/ composition (wt%) (g/mol) (g/mol) [M.sub.n] 100/0 (neat PVC) 3.0 * [10.sup.4] 6.3 * [10.sup.4] 2.1 70/30 4.2 * [10.sup.4] 7.3 * [10.sup.4] 1.7 44/56 8.0 * [10.sup.4] 1.7 * [10.sup.5] 2.1 20/80 8.8 * [10.sup.4] 2.3 * [10.sup.5] 2.6 0/100 (neat TPU) 1.1 * [10.sup.5] 2.6 * [10.sup.5] 2.4 TABLE 2. Mean engineering tensile properties of different PVC/TPU reactive blend compositions. PVC/TPU blend Elastic Yield composition modulus stress (wt%) (MPa) (MPa) 100/0 (neat PVC) 3200 [+ or -] 640 (a) 62 [+ or -] 3.8 70/30 2100 [+ or -] 800 42 [+ or -] 3.4 44/56 N/A N/A 20/80 N/A N/A 0/100 (neat TPU) N/A N/A PVC/TPU blend 100% strain Stress at Strain at composition secant modulus failure failure (wt%) (MPa) (MPa) (%) 100/0 (neat PVC) N/A 42 [+ or -] 7.2 51 [+ or -] 36 70/30 N/A 48 [+ or -] 7.8 320 [+ or -] 76 44/56 9.0 [+ or -] 0.36 57 [+ or -] 11 470 [+ or -] 46 20/80 4.6 [+ or -] 0.58 59 [+ or -] 12 480 [+ or -] 72 0/100 (neat TPU) 3.3 [+ or -] 0.30 64 [+ or -] 9.2 570 [+ or -] 34 (a) 2 X SD.
1. A. Reischl, W. Gobel, and K. Schmidt, U.S. Patent No. 3,444,266 (1969).
2. R.P. Carter, Jr., U.S. Patent No. 3,487,126 (1969).
3. W. Keberle and W. Gobel, U.S. Patent No. 3,637,553 (1972).
4. H.W. Bonk, A.A. Sardanopoli, H. Ulrich, and A.A.R. Sayigh, J. Elastoplast., 3, 157 (1971).
5. R.S. Brookman, Handbook of PVC Formulating, E.J. Wickson, ed., John Wiley & Sons, New York (1993).
6. L.H. Sperling, Interpenetrating Polymer Networks, Advances in Chemistry Series No. 229, American Chemical Society, Washington D.C. (1994).
7. J. Karger-Kocsis, in Polymer Blends and Alloys, G.O. Shonaike and G.P. Simon, eds., Marcel Dekker, New York (1999).
8. J.J. Ziska, J.W. Barlow, and D.R. Paul, Polymer, 22, 919 (1981).
9. R.E. Prud'homme, Polym. Eng. Sci., 22, 90 (1982).
10. E.M. Woo, J.W. Barlow, and D.R. Paul, Polymer, 26, 763 (1985).
11. F. Xiao, D. Shen, X. Zhang, S. Hu, and M. Xu, Polymer, 28, 2335 (1987).
12. Y.Q. Zhu, Y.J. Huang, Z.G. Chi, and H.J. Shen, Eur. Polym. J., 30(12), 1493 (1994).
13. Y.Q. Zhu, Y.J. Huang, and Z.G. Chi, J. Appl. Polym. Sci., 56, 1371 (1995).
14. R.E. Camargo, C.W. Macosko, M.V. Tirrell, and S.T. Wellinghoff, Polym. Eng. Sci., 22, 719 (1982).
15. R.E. Camargo, C.W. Macosko, M.V. Tirrell, and S.T. Wellinghoff, Polymer, 26, 1145 (1985).
16. T.A. Speckhard and S.L. Cooper, Rubb. Chem. Technol., 59, 405 (1986).
17. P.H. Geil, J. Macromol. Sci. Phys., B14, 171 (1977).
18. L.A. Utracki, Polymer Alloys and Blends, Hanser, New York (1990).
19. M.M. Coleman and J. Zarian, J. Polym. Sci. Polym. Phys. Ed., 17, 837 (1979).
20. D.E. Witenhafer, J. Macromol. Sci. Phys., B4, 915 (1970).
21. H. Munstedt, J. Macromol. Sci. Phys., B14, 195 (1977).
22. J.W. Summer, J. Vinyl Technol., 3, 107 (1981).
23. S.R. Parnell, "The Reactive Blending of Thermoplastic Polyurethane In Situ With Polyvinyl Chloride," Ph.D. Dissertation, University of Akron, Ohio (2002).
24. W.P. Yang, C.W. Macosko, and S.T. Wellinghoff, Polymer, 27, 1235 (1986).
25. D. Joel and A. Hauser, Die Angewandte Makromolekulare Chemie, 217, 191 (1994).
26. T. Hentschel and H. Munstedt, Polymer, 42, 3195 (2001).
27. M.M. Coleman, K.H. Lee, D.J. Skrovanek, and P.C. Painter, Macromolecules, 19, 2149 (1986).
28. R.E. Camargo, C.W. Macosko, M. Tirrell, and S.T. Wellinghoff, Polym. Commun., 24, 314 (1983).
29. K.R. Gifford, D.R. Moore, and R.G. Pearson, Plast. Rubb. Mater. Appl., 5, 161 (1980).
30. V.V. Shilov, A.N. Bliznyuk, and Y.S. Lipatov, J. Mater. Sci., 22, 1563 (1987).
31. T.O. Ahn, K.T. Han, H.M. Jeong, and S.W. Lee, Polym. Int., 29, 115 (1992).
32. Y. Kim, W. Cho, and C. Ha, J. Appl. Polym. Sci., 71, 415 (1999).
33. A.J. Ryan, J.L. Stanford, and R.H. Still, Polym. Commun., 29, 196 (1988).
34. Y. Camberlin and J.P. Pascault, J. Polym. Sci. Polym. Phys. Ed., 22, 1835 (1984).
35. D.W. Van Krevelen, Properties of Polymers, 2nd ed., Elsevier, New York (1976).
36. P. Ghersa, Mod. Plast., 36(2), 135 (1958).
Shane Parnell, Kyonsuku Min
Department of Polymer Engineering, University of Akron, Akron, Ohio 44325
Correspondence to: K. Min; e-mail: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Parnell, Shane; Min, Kyonsuku|
|Publication:||Polymer Engineering and Science|
|Date:||Jun 1, 2005|
|Previous Article:||Experimental and numerical studies of injection molding with microfeatures.|
|Next Article:||Morphological influence on mechanical characterization of ethylene-vinyl acetate copolymer-clay nanocomposites.|