Miscibility and phase behavior of IR/BR and BR/BR blends.

Blends of two or more polymers are often used to combine the desired physical properties of each polymer to obtain an improved product. The polymers, when blended, may be thermodynamically miscible, exhibiting a single phase morphology, or the polymers may be immiscible and exhibit a two phase morphology. The miscibility of the blend can significantly affect physical properties. Understanding miscibility and how to control it is a key to obtaining optimum blend properties.

In this article, the miscibility of cis-1,4-polyisoprene/polybutadiene and polybutadiene/polybutadiene blends as a function of 1,2-polybutadiene content was investigated. The phase diagrams of several blends which exhibited phase separation behavior were obtained by differential scanning calorimetry (DSC) and small angle laser light scattering (SALS). Several blends were characterized by solid state nuclear magnetic resonance relaxation measurements (NMR). In addition, pressure dilatometry measurements were used to calculate the equation of state parameters for the polybutadienes.

Experimental

The physical characteristics of the polymers used are shown in table 1. The microstructure was determined by NMR spectrometry in solution using a probe set for 1H analysis. Polymer Tgs were determined by differential scanning calorimetry. The Tg reported represents the onset temperature of the glass transition. Molecular weight was measured using gel permeation chromatography. Nomenclature for the polybutadienes indicate the % 1,2, content (i.e. PBD-8 contains 8% 1,2 units) and for the polyisoprenes indicates the molecular weight (i.e. PI-125 has a MN of 125,000).

Polymers were synthesized anionically in hexane in a batch reactor using various modifiers to control microstructure. Consequently, all the polymers have a very narrow molecular weight distribution, with the exception of PBD-35 and PBD-46, which were synthesized in a continuous reactor system. Blends were prepared by dissolving the polymers at 5% by weight in toluene. The solution was cast into a film and dried at room temperature for one day, then in a vacuum oven at 60[degrees]C for 1-2 days. Samples for small angle laser light scattering were cast directly from solution onto microscope slides, placed into a covered Petri dish and slowly dried at room temperature and then in a vacuum oven at 60[degrees]C.

Analytical techniques

The Tg of the blends was measured using a differential scanning calorimeter (DSC). The DSC was calibrated using an indium standard. Samples of 10-15 mg were quenched with liquid nitrogen to -120[degrees]C then heated at 10[degrees]C/minute to 40[degrees]C. The reported Tg is the temperature measured at the onset of the glass transition. This is the temperature at which tangents drawn to the baseline and to the steepest part of the transition intersect.

Annealing experiments were used to determine the phase diagrams of the blends. Samples were held in the DSC at the desired annealing temperature for various times (5-30 minutes) and then rapidly quenched in liquid nitrogen to -120[degrees]C. The sample was then reheated at 10[degrees]C/minute to 40[degrees]C and the blend Tg measured. Comparison of the blend Tg before and after annealing indicated if any phase separation occurred at the annealing temperature.

Small angle laser light scattering was also used to determine the blend phase diagram. A 4 mW He-Neon laser was passed through the sample held in a hot stage to control temperature. The light scattered by the sample was viewed on a ground screen with a Nikkor 60 mm Macro lens attached to a MI453A EG&G photodiode and detector system. The scattered intensity as a function of scattering angle from 2[degrees] to 30[degrees] was recorded. The cloud point was measured by heating the sample from 50[degrees]C to 150[degrees]C at 2[degrees]C/minute and recording the scattered light intensity as a function of temperature. The detector made a measurement every minute. The cloud point was defined as the temperature at which the scattering intensity began to significantly increase. Pressure-volume-temperature behavior of the polymers was measured using pressure dilatometry. The experimental apparatus used is described elsewhere (ref. 1). The volume of each polymer was measured as a function of pressure (0-200 MPa) and temperature (25[degrees]C-250[degrees]C). Data obtained from P-V-T measurements were used to calculate the equation of state parameters for each polymer. These equation of state parameters were used in the Sanchez-Lacombe equation of state to predict blend miscibility.

Results

Differential scanning calorimetry A blend with a single Tg is defined as miscible and one with two Tgs is defined as being immiscible. Miscible blends exhibited no strong scattering peaks in SALS measurements.

DSC results for blends of PI-125 with various polybutadienes are shown in figure 1. Results indicate the 1,4-polyisoprene is miscible at room temperature with polybutadienes of >24% 1,2 content. Blends containing very high 1,2, contents had extremely broad Tgs much closer to the polyisoprene Tg. These blends did not have a Tg intermediate between the two component polymer Tgs.

A typical result from a DSC annealing experiment is shown in figure 2. The DSC scan is shown for the PI-174/PBD-35 blend at various annealing temperatures. As the annealing temperature increased, the glass transition broadened and began to separate into the individual polymer Tgs, Because the polymer Tgs are very close (within 14[degrees]C) completely resolved Tgs were not observed. This suggests that the blend exhibits lower critical solution temperature (LCST) behavior. Similar results have been reported in the literature (refs. 2 and 3). Annealing experiments indicated that a PI-I25/PBD-24 blend also has a LCST at comparable temperatures.

The width of the Tg was measured by subtracting the temperature at the end of the transition (Tend) for the onset temperature (Tonset) (figure 2). This method has been used in the literature (ref. 2). The cloud point was defined as the temperature at which the width of the Tg as calculated by this method increased by more than 10% over the initial width. This phase diagram for the blend of PI-174/PBD-35 determined by this technique is shown in figure 3. Each point is the average of three or more annealing experiments at each temperature. In all cases, the rate of phase separation was very fast. Blends held at 100[degrees]C exhibited complete phase separation as evidenced by two distinct glass transitions within five minutes. No significant changes occurred with longer annealing times. After annealing at room temperature, phase separated blends eventually returned to the single phase state. This indicates that the phase separation is reversible.

The critical point can be calculated from the Flory-Huggins equation:

[MATHEMATICAL EXPRESSION OMITTED]

where Ni is the degree of polymerization of polymer A or B and [Pi.sub.CA] is the critical volume fraction of component A. For the PI-174/PBD-35 system, [Pi.sub.C,PDB = .51. This indicates that the phase diagram should be fairly symmetric about the 50/50 composition, as was observed experimentally.

Small angle laser light scattering

The cloud point of the PI-174/PBD-35 blend was also determined by laser light scattering. Samples were heated at 2[degrees]C/minute from 50[degrees]C up to 180[degrees]C and the scattered light intensity measured. The cloud point was defined as the temperature at which the intensity of scattering dramatically increased. The cloud point curve defines the binodal curve for the system. The phase diagram for this system as determined by SALS is shown in figure 4. SALS confirms the existence of an LCST for this blend system and also for the PI-125/PB D-24 blend sy stem.

The exact cloud point temperatures as determined by SALS are different than those determined by DSC. Each technique is sensitive to different domain sizes, therefore the temperature at which the phase separation can be observed is different for each technique. Also, the DSC phase diagram was determined by static annealing techniques and represents a system at equilibrium, as compared to SALS which was a dynamic technique. The cloud point temperature measured by SALS depends on the heating rate. In general, faster heating rates give higher cloud point temperatures. Measuring the cloud point temperature as a function of heating rate and extrapolating to zero heating rate would reduce the measured cloud point temperature, bringing it close to that measured by DSC.

Dilatometry results

High pressure dilatometer data were obtained for five polybutadiene microstructures, ranging from 8% to 87% 1,2 content. A representative data set is given in figure 5 for PBD-8. Each data point was obtained in isothermal data collection mode. That is, at each temperature, the hydrostatic pressure was varied from 10 to 200 MPa in increments of 10 MPa. Data at 0 MPa for each temperature was extrapolated from the data using a 2nd degree polynomial. The data were fit to two equations of state. In table 2 each polybutadiene microstructure is fit to the Tait equation. For a molecular interpretation, the same data are fit to the Sanchez-Lacombe polymer equation of state (ref. 5) and are given in table 3.

Discussion

Both the DSC and SALS results indicate that 1,4-polyisoprene is miscible at room temperature with polybutadienes of greater than 24% 1,2 content for the molecular weights tested. Blends of polyisoprene with polybutadienes of intermediate vinyl contents (24-35%, depending on polymer molecular weight) exhibit LCST behavior. Isoprene and butadiene monomers are very similar and there are no specific interactions such as hydrogen bonding between these polymers to explain the miscibility. The miscibility is due to the copolymer effect (ref. 3). The random copolymer effect accounts for the repulsion between different segments (i.e. 1,2 and 1,4 units) within a random copolymer. The greater the repulsion, the more miscible the copolymer is with any other polymer. The large repulsion between the 1,4 and 1,2 polybutadiene units forces these segments to be miscible with the polyisoprene. As the temperature increases, this repulsive effects decreases and blends exhibit lower critical solution temperature behavior.

The miscible blends, especially those with very high 1,2 contents, have extremely broad Tgs. Polymer relaxation times as measured by nuclear magnetic resonance experiments on these blends indicate that even though the blend is miscible, each polymer maintains its own separate relaxation times. The polymer chains, because they are miscible, are in close contact, but the movement of one polymer chain does not restrict the movement of the second chain. This is especially true since there are no specific interactions between two polymers such as hydrogen bonding which could act to restrict their motion. However, the polyisoprene does act to plasticize the high vinyl polybutadiene, and Shifts its response to different frequencies. Therefore, since the DSC is sensitive only to certain frequencies of motion, it cannot detect the changed motion of the high viny polybutadiene and subsequently gives very broad Tgs for these blends.

Similar behavior has been noted in the literature (ref. 4). This work with 1,4-polyisoprene/1,2-polybutadiene blends indicated that despite their miscibility, carbons of the respective polymers exhibited distinct glass transitions.

The DSC results indicate that blends of polybutadienes with only modest differences in 1,2 content are miscible. Large differences in 1,2 levels cause the blends to become immiscible and exhibit two Tgs.

The polybutadiene dilatometry data appear to follow a group additive concept. In figure 6 the Sanchez-Lacombe characteristic pressure for each microstructure is plotted as a function of % 1,2 content. Except for the apparent discrepancy in the value at 50% 1,2 content, P* varies in a linear fashion. P* is equivalent to the cohesive energy density (CED) of the polymeric liquid at absolute zero. P* is a measure of the volatility per monomer unit, since it is equal to the heat of vaporization per unit volume. This means that P* is essentially the binding energy of the polymer melt.

The characteristic pressure has an analogous molecular interpretation. P* = [Epsilon/v*, where [Epsilon]* is the attractive mean-field potential for a molecular segment (in KJ/mol) and v* is the close-packed (excluded) volume of a polymer segment (ref. 6). In table 4, the Sanchez-Lacombe constants are recast in terms of their equivalent molecular parameters. The striking feature of these data is that we see that the mean-field potential varies only marginally, on the order of about .10 KJ/mol. On the other hand, the excluded volume of each molecular segment increases about 20% as a function of vinyl content. As a measure of the size of each segment, a methylene group, -[CH.sub.2]-, of polyethylene, has a size of about 15cc/mole. The molecular segments for cis and trans polybutadiene have a molecular volume of roughly 9-10 cc/mole, which is a little less than a CH group. The change in volume as a function of vinyl content is therefore an indicator that the short pendant side chain is causing a disruption in the packing of the polymer chain.

Some further consequences of the packing effects can be described. Since the primary contribution to the variation in P* is due to packing effects, we have for all practical purposes a "bland" interaction between cis, trans and vinyl units. This means that thermal expansivity will drive the phase separation process. Blends with similar vinyl contents should be miscible, as is shown experimentally. However, when we have even modest differences in microstructure of the same monomer, a lower critical solution temperature (LCST) should be found.

Conclusions

1,4-polyisoprene is miscible at room temperature with polybutadienes of greater than about 24% 1,2 content for the molecular weights tested. Blends of polyisoprene with polybutadienes of intermediate vinyl content (24-35%, depending on polymer molecular weight) exhibit lower critical solution temperature behavior. Both differential scanning calorimetery and small angle laser light scattering techniques are useful in measuring the cloud point of these blends. The phase separation of the blends is extremely rapid, occurring in less than five minutes at 100[degrees]C.

Miscible blends have extremely broad Tgs very close to the polyisoprene Tg. NMR relaxation measurements indicate that this is because the polymers maintain their own individual motions in the blend, with the polyisoprene acting as a plasticizer for the high vinyl polybutadiene. Furthermore, it appears that thermal expansivity effects dictate the phase separation process in polybutadienes of varying microstructure. With similar microstructures, polybutadienes are miscible with one another. Once a modest difference in vinyl content occurs, an LCST is formed. The origin of this process has been determined to be the difference in the packing of vinyl groups versus cis or trans units. The equation of state properties of the melts follows a group additivity relationship based upon the microstructure content.

[TABULAR DATA OMITTED]

References

1. P. Zoller, Gnomix Inc., Gnomix PVT Apparatus Information, Boulder, CO, 1990.

2. S. Kawahara and S. Akiyama, Polym. J., 22, (5), 361 (1990).

3. S. Sakurai, H. Jinnai, H. Hasegawa, T. Hashimoto and C. Han, Macromolecules 24, 4839 (1991).

4. J.B. Miller, K.J. McGrath, C..M. Roland, CA. Trask and A.N. Garroway, Macromolecules 23, 4543 (1990).

5. I.C Sanchez and R.H. Lacombe, J. Polvm. Sci., Polvm. Lett. Ed. 15, 71 (1977).

6. I.C. Sanchez and R.H. Lacombe, Macromolecules 11, 1145 (1978).