Surface Modification Strategies for Multicomponent Polymer Systems

Apr 2, 1997 - A specific functionalized copolymer of poly(isoprene-4-vinylpyridine) (PIP-P4VP) is demonstrated to be an effective interfacial modifier...
0 downloads 0 Views 297KB Size
Ind. Eng. Chem. Res. 1997, 36, 1211-1217

1211

Surface Modification Strategies for Multicomponent Polymer Systems. 2. Study of the Interfacial Tension and Morphology of Linear Low-Density Polyethylene/Poly(vinyl chloride) Blends Hua Liang and Basil D. Favis* Centre de Recherche Applique´ e sur les Polyme` res (CRASP), Department of Chemical Engineering, Ecole Polytechnique de Montreal, P.O. Box 6079, Station Centre Ville, Montreal, Quebec H3C 3A7, Canada

The breaking thread method was used to measure the interfacial tension between linear lowdensity polyethylene and poly(vinyl chloride) (PVC). The thermal sensitivity of PVC renders this type of study extremely difficult. By choosing a low molecular weight of the PVC that showed Newtonian plateau behavior at low shear rates, using an appropriate temperature, and substantially reducing the thread thickness to accelerate the experiment, the breakup is constrained to occur before severe thermal degradation. In fact, the onset of thermal degradation can be observed as a deviation of the distortion amplitude from the predicated linearity of the Tomotika theory. A specific functionalized copolymer of poly(isoprene-4-vinylpyridine) (PIPP4VP) is demonstrated to be an effective interfacial modifier for this system. The effectiveness of this modifier is shown by its ability to significantly reduce the particle size as well as the interfacial tension. Introduction The interfacial tension between two components in an immiscible polymer blend plays an important role in determining the morphology, and therefore the final mechanical properties of the blend or alloy. Among the several experimental methods applicable to polymer systems (Wu, 1982; Luciani, 1996), the breaking thread method is relatively fast and suitable for systems with high viscosities and relatively poor thermal stability. Another dynamic method, called imbedded fibre retraction (IFR) has been developed by Carriere et al. (1989). Some methods, like the pendant drop or spinning drop, require several hours to attain an equilibrium state. The breaking thread method does not require density data and specialized apparatus but does need the zero shear rate viscosities of the two melt phases at the processing temperature. It is limited to systems that have no yield stress at low shear rates. Until now, it has been used in some immiscible and compatibilized polymer blend systems (Chappelear, 1964; Chapleau et al., 1995; Elemans et al., 1990; Elmendorp, 1986; Lepers and Favis, 1995; Luciani et al., 1996; Mekhilef et al., 1997; Van Gisbergen and Meijer, 1990; Watkins and Hobbs, 1993). To our knowledge, however, no data concerning the interfacial tension between linear low-density polyethylene (LLDPE) and poly(vinyl chloride) (PVC) melts measured has been reported in the literature. The particular difficulty in measuring the interfacial tension between PVC and other polymer melts mainly lies in the high viscosity and the poor thermal stability of PVC at ordinary processing temperatures. Studying a system that is sensitive to thermal degradation such as PVC made this study an order of magnitude more difficult than typical breaking thread studies on other polymer systems. The breaking thread method is based on Tomotika theory (1935) and was first applied to polymer blends by Chappelear (1964). It involves the breakup of a polymer thread embedded in another polymer matrix. The sinusoidal distortions of the thread will occur when * To whom correspondence should be addressed. S0888-5885(96)00486-1 CCC: $14.00

the perturbation of the wavelength is greater than the circumference of the thread. The amplitude of the distortion, R, is related exponentially to time, t, as

R ) R0eqt

(1)

where R0 is the initial amplitude (m) and q is the growth rate that can be obtained from the slope of the plot of the thread diameter evolution over time. Thus, from the following equation, the interfacial tension σ can be calculated

q)

σΩ(λ,p) ηmD0

(2)

where σ is the interfacial tension (N/m), ηm the zero shear rate viscosity of the matrix phase (Pa‚s), p the viscosity ratio of the thread phase to the matrix phase (ηt/ηm), D0 the initial diameter of the thread, and Ω(λ,p) a tabulated function of viscosity ratio and wavelength. It has been convincingly demonstrated that the addition of a block copolymer to a blend system decreases the interfacial tension between two phases and changes the morphology of the blend (Chapleau et al., 1995; Lepers and Favis, 1995; Mekhilef et al., 1997; Wagner, 1993). The blend of PVC with PE is a typical immiscible system with very poor adhesion at the interface. Suitable block copolymers, e.g., ethylene-propylene-diene copolymers (EPDM), styrene-butadiene-styrene copolymer (SBS), styrene-ethylene (SE), or chlorinated polyethylene (CPE), can improve the compatibility of the blends (Paul et al., 1973; Chen and White, 1993; Elemans et al., 1990; Wagner and Wolf, 1993), resulting in a decrease of the dispersed phase size and an increase of mechanical properties. The diblock copolymers need to move to the interface in order to be efficient in compatiblizing blends by lowering the interfacial tension and stabilizing the morphology. It has been reported, for other systems, that the interfacial tension decreases from one-half to one-fifth of the original value when a block copolymer is added to a polymer blend (Elemans © 1997 American Chemical Society

1212 Ind. Eng. Chem. Res., Vol. 36, No. 4, 1997

et al., 1990; Lepers and Favis, 1995; Mekhilef and Favis, 1996; Sundararaj and Macosko, 1994; Wagner and Wolf, 1993). In this work, the interfacial tension between LLDPE and the thermally sensitive PVC is measured in the melt by the breaking thread method through careful control of the experimental conditions. The effect of addition of a specific functionalized copolymer, poly(isoprene-4-vinylpyridine), on the dispersed phase particle size and the interfacial tension of the blend is also investigated. Experimental Section Materials. The LLDPE/PE blend selected for this study is typical of problematic recycle combinations. The LLDPE used (grade HS-7028 Natural 7; Mn, 33 000 g/mol; Mw, 115 000 g/mol) was supplied by Union Carbide Chemicals and Plastics Company Inc.; its melt flow index is 1.0 g/10 min according to the manufacturer, and its melting point is 125.5 °C as measured on the DSC. The PVC (grade SE-950ED) was supplied by Synergistic Chemicals Inc. and was based on a commercially available sample of PVC with a K value ) 51 (Mw, 54 000 g/mol); it was stabilized by seven parts liquid methyltin heat stabilizer (Advastab TM-281 SP). The interfacial modifier was synthesized by Yu and Eisenberg at McGill University; it is a diblock copolymer of polyisoprene (PIP)-poly(4-vinylpyridine) (P4VP); the degree of polymerization of this diblock copolymer determined by GPC and NMR methods is 256-494 (PIP-P4VP). Blending. The original PVC powder mixed with the heat stabilizer was compounded at 175 °C on a Brabender Plasticorder with a chamber volume of about 25 mL. LLDPE/PVC blends in the proportion of 75/25 (by weight, unless stated otherwise) were prepared on the Brabender at 200 °C. The PE was first added into the chamber at a set temperature; the PVC was introduced into the chamber when the PE was molten; then the mixture was blended for another 5 min with a roller blade rotation speed of 50 rpm at the desired temperatures. In the compatibilization study, the interfacial modifier in different concentrations based on the weight of the PVC or the PVC dispersed phase in the blend was added to the PVC melt immediately after fusion. Blends were immediately cooled in cold water and then prepared for morphology observation. Rheology. Rheological characterization of the raw materials and the blends was conducted on a Bohlin constant stress rheometer (CSM) in the oscillation mode at desired temperatures. A parallel plate configuration was used with a gap of about 1.3 mm. Strain sweeps were performed to define the region of linear viscoelasticity. The Cross generalized power law equation (Cross, 1970) was used to extrapolate zero shear rate viscosities of raw materials and the blends of PVC with the copolymer. The accuracy of the LLDPE viscosity is within (3% and that for the PVC is (6%. It should be pointed out that the zero shear rate of the LLDPE, i.e., the matrix, rather than that of the PVC, is crucial in determining the interfacial tension as shown in eq 2. In addition, for the viscosity ratio, p, in the range of this study (0 < p < 1), the tabulated function Ω(λ,p) varies slowly with p. Preparation of PVC Thread and LLDPE Film. PVC threads were produced by drawing out the molten PVC pellets on a hot plate by carefully controlling the heating temperature. The PVC threads prepared in this

way had diameters ranging from 3-20 mm and are much smaller than those prepared by a capillary rheometer, melt index apparatus, or a single screw extruder. The uniformity of the threads was expressed by the diameter standard deviation that was obtained by measuring 8-10 diameters. In a similar way, threads containing different concentrations of the diblock copolymer were made from the PVC blends. An LLDPE film with a thickness of 0.2-0.4 mm was obtained by compressing LLDPE pellets at 200 °C for 5 min. Interfacial Tension Measurement. The interfacial tension measurement was conducted by preparing a sandwich of a PVC thread about 10 mm long between two LLDPE films (10 mm × 10 mm). The sandwich was placed between a glass slide and a cover slip. The sample was first heated up to 150 °C and kept isothermally for 5 min. The process was followed by heating the sample to the desired temperature on a hot stage (Mettler FP82 HT) at a rate of 20 °C/min. While the sample is being heated up to the desired temperature, care is taken so that no pressure is exerted on the sample to avoid undesired deformation of the thread. The temperature was controlled by the Mettler FP90 control processor to (0.8 °C. The temperature calibration was made by using normal calibration substances, such as benzoic acid (melting point 122.4 °C) and caffeine (melting point 236.2 °C). A Nikon transmission light microscope connected to a CCD-IRIS/RGB video camera was used to observe and record the distortion amplitude of the thread with time at regular intervals. A magnification of 400 times was used so that 5-6 complete wavelengths could be observed depending on the diameter of the threads. To make the measurement, a computer equipped with an image analysis system was used to capture the digitized images from the microscope. The measurement of the distortion amplitude evolution of the thread over time and the wavelength was obtained by the Visilog 4.1.3 image analysis software package modified in-house for the breaking thread experiment. In addition, a Nikon camera was used to record the breaking thread process photographically. Morphology Analysis. Freeze-fractured surfaces were prepared cryogenically by immersing the Brabender blended samples in liquid nitrogen for 10-15 min prior to the fracturing. The fracture surfaces were coated with a gold-palladium alloy prior to the examination on a JEOL scanning electron microscope (JSMT300). The semiautomatic image analyzer was used for measuring the size of the PVC phase in the blends from plane-faced samples. The number average diameter dn and the volume average diameter dv were calculated on the basis of 400 diameter measurements made on SEM photos taken from different areas of the same blend sample. A correction factor developed by Saltikov (1958) was applied to the diameters determined from SEM micrographs. This step was carried out to account for polydispersity and for the fact that the observed plane does not necessarily cut through the particles at their equators. The uncertainty on the average diameter measurements by this method is better than (10%. Results and Discussion 1. The Measurement of the Interfacial Tension of LLDPE/PVC. At the beginning of the study, when a PVC grade of higher molecular weight (K ) 66, Mw ) 140 000 g/mol) was used, no plateau on the viscosity vs shear rate curves was observed in the low shear rate

Ind. Eng. Chem. Res., Vol. 36, No. 4, 1997 1213 Table 1. Zero Shear Viscosities and Viscosity Ratios of Raw Materials and Blends at 200 °C samples

η0, Pa‚s

p ) ηt/ηm



LLDPE PVC PVC + 0.1%PIP-P4VP PVC + 0.5%PIP-P4VP PVC + 1%PIP-P4VP PVC + 2%PIP-P4VP PVC + 3%PIP-P4VP PVC + 5%PIP-P4VP

8140 1940 2077 2299 2997 3843 3960 5523

0.24

0.15

0.26 0.28 0.37 0.47 0.49 0.68

0.14 0.13 0.12 0.10 0.10 0.09

region. This may indicate that a yield stress exists for this PVC and that no thread breakup would occur if the breaking thread method were used for measuring the interfacial tension. A PVC of lower molecular weight (Mw) 54 000 g/mol) was then selected and used throughout this study. When the measuring temperature was above 190 °C, low shear rate plateaus on the viscosity/ shear rate curves were observed. Table 1 gives the zero shear rate viscosities of the selected raw materials and blends of PVC with the copolymer. The zero shear rate viscosities were obtained by extrapolating the rheological data at the very low shear rate region to zero shear rate through the application of the Cross model. The value of the tabulated function of the specific viscosity ratio is also given in Table 1. The breakup time from a thread to droplets can be estimated from Tomotika’s theory (Tomotika, 1935). It is predicted that when the arbitrary wavelength λ > πD0, the distortion amplitude R must grow exponentially with time; the thread breakup occurs when R ) R ) 0.81R0. R is the average thread radius. If the interfacial tension is known, one can derive from eq 1 the breakup time for a thread with diameter D0

tb ) 1/q ln(Rb/R0)

(3)

where Rb is the amplitude at breakup and R0 the original amplitude that can be estimated from the expression derived by Kuhn (1953)

R0 )

(

21kT 8π3/2σ

)

1/2

(4)

where k is the Boltzmann constant and T is the absolute temperature. The zero shear rate viscosities of LLDPE and PVC at 200 °C are found to be 8.14 × 103 and 1.94 × 102 Pa‚s, respectively. This viscosity ratio (pPVC/PE ) 0.24) gives a tabulated function of Ω ) 0.15. Using an arbitrary interfacial tension value of 5.0 × 10-3 N/m as an example, threads with 0.1, 1, 5, 10, and 50 µm diameters would take 5, 70, 425, 920, and 5500 s, respectively, to break into droplets according to the Tomotika theory. The viscosity of PVC is very sensitive to temperature. Hence, when the temperature is increased to 210 °C, the corresponding breakup times would be remarkably decreased to 1.5, 25, 150, 340, and 2000 s, respectively. The above calculation indicates that the increase of the measuring temperature can significantly speed up the breakup process. In order to conduct the measurement of the interfacial tension of the LLDPE/PVC blend system in a relatively short time, it is necessary therefore to use both an appropriate measuring temperature and threads with small diameters. The thermal stability of PVC, which decreases with temperature, may affect the quality of the experiment. Therefore, it is critical to optimize the measuring temperature, the experiment time, and the quality of

Figure 1. Semilog plot of the time dependence of the relative deformation of a PVC thread in an LLDPE matrix at different temperatures.

the experiment. When the measuring temperature was below 190 °C, no obvious thread distortions were observed, but the thermal degradation of PVC was evident from the gradually darkening color of the threads with time. The effect of further temperature increases on the relative distortion (2R/D0) of some threads of similar diameters is shown in Figure 1. It should be noted that in Figure 1, t ) 0 represents the time at which the recording of the evolution on the video camera starts. It is usually after the time when the hot stage reaches the measuring temperature. The dotted lines indicate the experimental values, and the drawn straight lines are the best fits as determined by linear regression of the relative distortion with time at an early stage of the process. It can be seen that at 190 °C in the early stage of the experiment, the dependence of log(2R/D0) on time is virtually linear. However, with an increase of time, the PVC thermal degradation, in which cross-linking is widely believed to be involved, occurs. The degradation and cross-linking are the main factors that lead the dependence to deviate from the straight line. Further increases in the time do not result in any further distortion development of threads. This phenomenon is called “thread stiffening” for the purposes of this study. At 190 °C, a PVC thread with a diameter slightly less than 10 mm did not break into droplets within 15 min and at that time was starting to show signs of thermal degradation as evidenced by the discoloring of the thread. The time duration in which R vs time shows a linear relationship decreased with increasing temperature. When the temperature was increased to 200 °C and above, the PVC threads broke into droplets, and some deviation from the best-fit straight line was found in the latter stage of the process. Deviation from linearity of R vs time was generally found to correspond to the onset of discoloration in the samples. Apparently another parameter, the initial diameter of the thread, has a significant effect on the relative distortion of the thread and determines if the thread breaks up into droplets at a certain temperature in an acceptable period of time for PVC thermal stability. In

1214 Ind. Eng. Chem. Res., Vol. 36, No. 4, 1997

Figure 2. Semilog plot of the time dependence of the relative deformation of a PVC thread in an LLDPE matrix at 210 °C.

Figure 2, the relative distortion amplitude of two threads with different diameters plotted against time at 210 °C is presented. The thread with a smaller diameter (D0 ) 4.75 µm) broke up in less than 3 min during which a very good linear relationship between relative distortion amplitude and time was observed. By contrast, the thread with a larger diameter (D0 ) 12.24 µm) did not break up before the thread became “stiff”. Nevertheless, the relative distortion amplitude and time showed a linear relationship within about 280 s before the deviation and thread stiffening. Since the breakup time of the PVC thread is very much dependent on the temperature and initial diameter of the thread, it is evident that the experimental error can be possibly minimized only when the measurement is done in the linear relationship range of time and relative distortion. Considering the restriction arising from the PVC thermal degradation, it is desirable to decrease the initial thread diameter as much as possible to accelerate the disintegration process and avoid the experimental error brought about by degradation of the PVC. By following the above considerations, the growth rate q of various threads is obtained from the slope of the linear relationship range on a plot of the relative amplitude distortion against time at 200 °C. Only those breakup processes where the distortion’s wavelength numbers equal to Xm ( 10% were taken into account. From the growth rate, the initial diameter of the thread, and the zero shear rate viscosity of LLDPE as well as the viscosity ratio, the interfacial tension of LLDPE and PVC was calculated according to eq 2 and is listed in Table 2. The value of the interfacial tension is 3.4 × 10-3 ( 0.5 N/m, which is the average of several samples. One of the typical thread distortion processes was photographically recorded and is shown in Figure 3. In Table 2, the experimental wavenumbers (Xexp) are found generally to be close to the theoretical one (Xm). The interfacial tension was also measured at 210 °C. The viscosity ratio, at zero shear rate, at this temperature was 0.066. Table 3 gives the calculated interfacial tension results at 210 °C, which is 3.1 ( 0.8 × 10-3 N/m. The interfacial tension of the PE/PVC blend can be compared to that predicted from the empirical harmonic-

Figure 3. Sinusoidal distortions of a PVC thread (diameter 12.10 µm) embedded in a LLDPE matrix at 200 °C. The photos were taken at time intervals 0, 110, 270, 390, and 630 s from top to bottom. Table 2. Interfacial Tension Measurement Results for PVC Threads in LLDPE Matrix at 200 °C D0 (SD), µm



Xm

Xexp

q, s-1 × 103

7.53 ((0.21) 8.22 ((0.15) 10.39 ((0.22) 10.71 ((0.14) 5.56 ((0.20) 6.47 ((0.18) 8.03 ((0.25) 10.25 ((0.25) 11.25 ((0.21)

0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15

0.58 0.58 0.58 0.58 0.58 0.58 0.58 0.58 0.58

0.54 0.61 0.64 0.58 0.58 0.62 0.59 0.56 0.64

4.96 6.15 6.78 6.21 10.25 9.10 9.05 6.11 4.54

σ, N/m × 103

3.4 ( 0.5

Table 3. Interfacial Tension Measurement Results for PVC Threads in LLDPE Matrix at 210 °C D0 (SD), µm



Xm

Xexp

q, s-1 × 103

10.22 ((0.24) 11.10 ((0.19) 9.54 ((0.25) 12.12 ((0.19) 10.23 ((0.23) 8.85 ((0.24)

0.27 0.27 0.27 0.27 0.27 0.27

0.57 0.57 0.57 0.57 0.57 0.57

0.55 0.60 0.63 0.57 0.58 0.61

4.96 6.15 6.78 6.21 10.25 9.10

σ, N/m × 103 3.1 ( 0.8

mean method (Wu, 1982)

σ12 ) γ1 + γ2 -

4γ1dγ2d

4γ1pγ2p

γ1d + γ2

d

γ1p + γ2p

(5)

where γ1 and γ2 are the surface tensions of two phases, γ1d and γ2d the dispersion components of liquid surface tension of two phases, and γ1p and γ2p the polar components of liquid surface tension of two phases. For LLDPE, a nonpolar polymer, it is reasonable to assume its polarity to be zero, i.e., XPEp ) 0. On the basis of the data given by Wu (1982), the calculated interfacial tension between PE and PVC at 20 °C is approximately 5.0 × 10-3 N/m. Increasing the temperature from 20 to 200 °C will tend to lower this value. Unfortunately, the data for PE and PVC required for the harmonic-mean calculation at higher temperature is not available.

Ind. Eng. Chem. Res., Vol. 36, No. 4, 1997 1215

Figure 4. Chemical structure of copolymer (PIP-P4VP)

Figure 5. Effect of copolymer (PIP-P4VP, 256-494) on the interfacial tension between LLDPE and PVC.

2. The Effect of the Interfacial Modifier. A potential compatibilizer, i.e., a diblock copolymer of poly(isoprene-4-vinylpyridine) was used. The molecular structure of the copolymer is illustrated in Figure 4. The objective of using this copolymer is to modify the interfacial properties and enhance the adhesion between LLDPE and PVC phases in the blends. It is expected that the diblock copolymer locates at the LLDPE/PVC interface due to the similar polarity of PE to the isoprene segment in the copolymer and the basic-acid interaction between PVC and 4-vinylpyridine in the copolymer (Schreiber, 1995; Bosse et al., 1994). The breaking thread method was used to measure the effect of the interfacial modifier on the interfacial tension between LLDPE and PVC blend at 200 °C. The copolymer was introduced into the PVC phase, i.e., the thread, by melt blending, and its effect on the interfacial tension of the blend is presented in Figure 5. It should be noted that generally adding a copolymer to an immiscible polymer blend lowers the interfacial tension, as is the case in this study. The interfacial tension-driven breakup process of the PVC thread was retarded in the presence of diblock copolymer at the interface. When the copolymer content was higher than 1%, distortion occurred, but no ultimate thread breakup was observed during the experiment. In addition, due to the addition of the copolymer, the zero shear rate viscosity of the PVC was increased as shown in Table 1. Some caution should be exercised with respect to the 5% copolymer/PVC system. For the 5% copolymer/PVC rheological analysis, it was necessary to carry out the extrapolation to zero shear rate without a clear plateau behavior being observed since it was difficult to obtain

data in the low shear rate range and because of the thermal degradation. However, the plot of relative distortion of the thread vs time presents a similar shape to those of PVC without copolymer, i.e., the deviation from the straight line at longer time. The measurement of the growth rate was carried out in the linear relationship range of the plot. As can be seen from Figure 5, the interfacial tension decreases significantly as the copolymer content increases. Beyond 3% copolymer concentration, further increasing the copolymer content results in no further change. The interfacial interaction can be estimated from the acid-base pair interaction and the dispersion forces of the individual component. In this blend system, PVC acts as an acid and the P4VP segment in the copolymer acts as a base. The potential for strong adsorption exists between the short segment of the high-energy component P4VP in the copolymer and the PVC (Bosse et al., 1994; Schreiber, 1995). Therefore, the original interface between PVC and LLDPE can be partly replaced by a new interface where the P4VP component interacts strongly with PVC while the PIP component is toward the LLDPE phase. Due to dispersion forces and the higher chemical affinity of the PIP block in the copolymer with LLDPE than PVC, the LLDPE should be more compatible with PIP than with PVC, which is probably one of the reasons for the drop of the interfacial tension when the copolymer is added. With the increase of copolymer content, more modifier molecules will locate at the interface of PVC and PE until the copolymer saturates the interface. 3. Melt Mixing and Morphology. The morphologies of the uncompatibilized and compatibilized LLDPE/ PVC blend with 4% interfacial modifier concentration, based on the weight of the PVC phase, are compared in Figure 6a,b. The PVC dispersed phase size decreases with the interfacial modifier, and in both cases the PVC dispersed phase and LLDPE matrix are clearly seen as two separate phases, illustrating their immiscibility. Using semiautomatic image analysis, the PVC domain size change with increased interfacial modifier content was measured. The statistical results of the PVC domain average diameter are shown in Figure 7. The effect of the copolymer on dV is slightly more pronounced than that on dn. It can also be noticed that the difference between dV and dn decreases with the increase of copolymer content, implying that polydispersity of the dispersed phase can be reduced by addition of the diblock copolymer. The morphology data in Figure 7 known as an emulsification curve indicates saturation of the interface by the copolymer at about 8% copolymer content (based on the dispersed phase). Emulsification curves and their relation to the interface have been studied extensively in this laboratory (Cigana and Favis, 1996; Favis, 1994; Matos and Favis, 1995). The log-normal particle size distribution is shown in Figure 8, and the slope gives an idea of the blend polydispersity. The steeper the slope, the narrower the size distribution. It can be observed that the distribution becomes narrower when the copolymer is added. It can be seen that the diameters measured for the blends containing the copolymer follow a log-normal distribution substantially, while the blend without copolymer deviates slightly from the log-normal distribution. When 8% copolymer was added to the blend

1216 Ind. Eng. Chem. Res., Vol. 36, No. 4, 1997

Figure 7. Dependence of the dispersed phase size in LLDPE/ PVC blends (75/25) on the copolymer concentration (at 200 °C). The copolymer content is calculated on the basis of the dispersed phase.

Figure 8. Minor phase size distribution curve (log-normal scale) for the LLDPE/PVC (75/25) blends containing 0% A, 0.4% B, 2% C, 8% D, and 20% E copolymer at 200 °C. The modifier concentration was based on the weight of the PVC dispersed phase.

Figure 6. Scanning electronic micrographs for the LLDPE/PVC blend: (a) LLDPE/PVC(75/25) blend compounded at 200 °C; (b) LLDPE/PVC(75/25) blend compatibilized with 4% diblock copolymer (based on the dispersed phase) compounded at 200 °C; (c) LLDPE/PVC(75/25) blend compounded at 175 °C.

during compounding, the reduction in the dispersity of the PVC particles can be as much as 23%. Upon further increasing the copolymer content, the PVC domain size distribution becomes only slightly narrower.

A similar morphological measurement was made on the blend compounded at 175 °C. The number average diameter of the PVC dispersed phase, dn, is 3.25 µm; the volume average diameter, dV, is 5.63 µm. A photomicrograph is shown in Figure 6c, where PVC phases are seen to be much larger than those in Figure 6a. It is interesting to note that comparing the change of dn and dV caused by addition of the copolymer and the change by increasing temperature, the reduction of dV and dn due to the temperature increasing from 175 to 200 °C is much more pronounced. From this comparison, one can expect that temperature has a significant effect on the viscosity ratio of LLDPE and PVC since PVC is a heat sensitive polymer whose viscosity is dramatically affected by temperature.

Ind. Eng. Chem. Res., Vol. 36, No. 4, 1997 1217

Conclusion The interfacial tension between the incompatible LLDPE and PVC melts was measured by the breaking thread method and was evaluated at 3.4 mN/m at 200 °C. The influence of the thermal degradation of PVC is substantially overcome by producing threads having initial diameters as small as possible so that the measurement of the thread distortion amplitude was conducted before the occurrence of the thermal degradation of PVC at a relatively high temperature. The degradation of PVC results in a deviation of the distortion amplitude from the predicted linearity of the Tomotika theory. The use of the specific functionalized copolymer of poly(isoprene-4-vinylpyridine) as a potential interfacial modifier was shown by its ability to reduce the interfacial tension between PVC and LLDPE and the PVC dispersed phase size. When 3% copolymer was added, the interfacial tension was reduced from 3.44 to 1.23 mN/m. When the same amount of copolymer was added to the blend, the average diameter of the dispersed PVC phase was decreased from 1.82 to 0.89 µm. With increasing copolymer content the dispersed phase size distribution becomes less polydisperse. Acknowledgment The authors are grateful to the Natural Science and Engineering Research Council of Canada (NSERC) for sponsoring this research. The authors also wish to express their gratitude to Prof. A. Eisenberg and Dr. Y. S. Yu for supplying the copolymer. Literature Cited Bosse, F.; Eisenberg, Adi; Xu, Ruijian; Schreiber, H. P. IGC Study of Surface Properties of Adsorbed Styrene/Methacrylic Acid Diblock Polymers. J. Appl. Polym. Sci. 1994, 51, 521-527. Carriere, C. J.; Cohen, A.; Arends, C. B. Estimation of Interfacial Tension Using Shape Evolution of Short Fibres. J. Rheol. 1989, 33, 681-689. Chapleau, N.; Favis, B. D.; Carreau, P. J. Evaluation of the Interfacial Tension Using the Breaking Thread Method: Effect of Temperature and of Interfacial Modifier. Proceedings of the American Chemical Society Division of Polymeric Materials: Science and Engineering, Chicago, 1995; Vol. 73, pp 18-19. Chappelear, D. C. Interfacial Tension Between Molten Polymers. Polym. Prepr. 1964, 5, 313-371. Chen, C. C.; White, J. L. Compatibilizing Agents in Polymer Blends: Interfacial Tension, Phase Morphology, and Mechanical Properties. Polym. Eng. Sci. 1993, 33, 923-930. Cigana, P.; Favis, B. D.; Je´roˆme, R. Diblock Copolymers as Emulsifying Agents in Polymer Blends: Influence of Molecular Weight, Architecture, and Chemical Composition. J. Polym. Sci., Polym. Phys. Ed. 1996, 34, 1289-1700. Cross, M. M. Kinetic Interpretation of non-Newtonian Flow. J. Colloid Interface Sci. 1970, 33, 30-35 .

Elemans, P. H. M.; Janssen, J. M. H.; Meijer, H. E. H. The Measurement of Interfacial Tension in Polymer/Polymer systems: The Breaking Thread Method. J. Rheol. 1990, 34, 13111325. Elmendorp, J. J. A Study on Polymer Blending Microrheology. Polym. Eng. Sci. 1986, 26, 418-426. Favis, B. D. Phase Size/Interface Relationships in Polymer Blends: the Emulsification Curve. Polymer 1994, 35, 15521555. Kuhn, W. Spontane Aufteilung Von Flu¨ssigkeitszylindern in kleine Kugeln. Kolloid-Z. 1953, 132, 84-99. Lepers, J.-C.; Favis, B. D. Morphology/Interface Relationships in Polyester/Polyolefin Blends. Tech. Pap.-Soc. Plast. Eng. 1995, 41, 1588-1592. Luciani, A.; Champagne, M. F.; Utracki, L. A. Interfacial Tension in Polymer Blends, Part 2: Measurements. Polym. Networks Blends, 1996, 6, 51-62. Matos, M.; Favis, B. D.; Lomellini, P. Interfacial Modification of Polymer Blends - the Emulsification Curve: 1. Influence of Molecular Weight and Chemical Composition of the Interfacial Modifier. Polymer 1995, 36, 3899-3907. Mekhilef, N.; Favis, B. D.; Carreau, P. J. Morphological Stability, Interfacial Tension and Dual Phase Continuity in PolystyrenePolyethylene Blends. J. Polym. Sci., Polym. Phys. Ed. 1997, 35, 293-308. Paul, D. R.; Locke, C. E.; Vinson, C. E. Chlorinated Polyethylene Modification of Blends Derived from Waste Plastics. Part I: Mechanical Behaviour. Polym. Eng. Sci. 1973, 13, 202-208. Saltikov, S. A. Stereometric Metallography, 2nd ed.; Metallurgizdat: Moscow, 1958. Schreiber, H. P. Inverse Gas Chromatography: A Versatile Tool for Polymer Surface characterization. Tech. Pap.-Soc. Plast. Eng. 1995, 41, 2446-2448. Sundararaj Uttandaraman; Macosko, C. W. Drop Breakup and Coalescence in Polymer Blends: The Effects of Concentration and Compatibilization. Macromolecules 1994, 28, 2647-2657. Tomotika, S. On the Instability of a Cylindrical Thread of a Viscous Liquid Surrounded by Another Viscous Fluid. Proc. R. Soc. London 1935, A150, 322-337. Van Gisbergen, J. G. M.; Meijer, H. E. H. Influence of Electron Beam Irradiation on the Microrheology of Incompatible Polymer Blends: Thread Break-up and Coalescence. J. Rheol. 1990, 35, 63-87. Wagner, M.; Wolf, B. A. Effect of Block Copolymer on the Interfacial Tension between Two ‘Immiscible’ Homopolymers. Polymer 1993, 34, 1460-1464. Watkins, V. H.; Hobbs, S. Y. Determination of Interfacial Tensions between BPA Polycarbonate and Styrene-acrylonitrile Copolymers from Capillary Thread Instability Measurements. Polymer 1993, 34, 3955-3959. Wu, S. Polymer Interface and Adhesion; Marcel Dekker, Inc.: New York, 1982.

Received for review August 8, 1996 Revised manuscript received November 20, 1996 Accepted November 21, 1996X IE960486L

X Abstract published in Advance ACS Abstracts, February 15, 1997.