The PPO melt was processed at 280 °C. As discussed above, it is observed that the minimum operating temperature required for efficient mixing and extrusion of the blends decreases with increasing epoxy or MCDEA content, as expected. For example, for 80PPO:20DGEBA340, the extruder temperature used is 240 °C while the temperature is reduced 180 °C for 50PPO:50DGEBA. All blends are transparent at the processing temperature, but the blended samples of the epoxies with less than 60 wt% PPO are opaque at room temperature suggesting a two-phase structure caused by partial immiscibility of PPO- and DGEBA-rich phases or by partial crystallization of PPO from the melt. PPO crystallization occurs in solvents and is observed in DSC studies of some of the PPO:DGEBA samples as discussed below. For the PPO:MCDEA blends, the minimum operating melt blending temperature is reduced as the content of PPO is reduced as noted above. All PPO:MCDEA blends are transparent at room temperature, suggesting high miscibility, which is consistent with the close correspondence of their solubility parameters (Table 1).
The isothermal rheological properties of blends of PPO with either DGEBA381 or DGEBA340 were investigated in order to examine the improvement in processability of PPO on addition of epoxy resin at high PPO levels, as well as the impairment in flow behaviour of the epoxy caused by PPO at low PPO levels. Figure 1 shows the steady shear viscosities of PPO:DGEBA340 blends at 200 °C at various shear rates. Very similar behaviour was also found for PPO:DGEBA381 (not shown). The blends possess typical polymer melt characteristics with plateau viscosities at low shear rates (Newtonian region) and a reduction of viscosities at high shear rates (pseudoplastic or shear rate thinning behaviour). For blends with 50 wt% PPO or more, the transition from Newtonian to pseudoplastic region occurs at a lower shear rate with increasing PPO content, as is normally found in other polymer solutions.[27, 28] For blends with 40 wt% PPO or less, the viscosity is close to Newtonian over the entire range of shear rates studied.
Figure 2 shows that the logarithm of viscosity in the Newtonian region drops rapidly as the concentration of DGEBA340 increases in the melt blend; this type of behaviour is usually interpreted as being due to an increase of free volume with the presence of solvent.[29, 30] The logarithm of the viscosity decreases in an approximately linear fashion as the PPO content is increased, as commonly observed in polymer blends,[19, 31, 32] but deviation occurs at lower levels of PPO. Thus at high concentrations of PPO, the viscosity of blends decreases approximately tenfold per 10 wt% increase in epoxy content and this behaviour gives a quantitative basis for the previously noted improved processability of PPO after blending with epoxy. However at low PPO concentrations, the change is approximately twofold per 10 wt% increase in PPO which is more desirable for its use in toughening the crosslinked epoxy.
Figure 2. Steady shear viscosity of PPO:DGEBA340 and PPO:MCDEA at 0.1 s−1 and 200 °C and complex viscosity of PPO:DGEBA340, PPO:DGEBA381 and PPO:MCDEA at 0.1 rad s−1 and 200 °C as a function of PPO content.
Download figure to PowerPoint
Figure 3 shows the frequency dependence of the complex viscosity of PPO:DGEBA340 at 200 °C in the range 0.1 to 100 rad s−1. As found for the steady shear data, the viscosity progressively decreases with increasing epoxy content suggesting a significant improvement of processability. Nearly identical behaviour is observed for PPO:DGEBA381, but as demonstrated in Figure 2, the PPO:DGEBA381 blends have slightly higher complex viscosities than the PPO:DGEBA340 blends. This may be due to the higher molecular weight of the epoxy in the former system, or may be caused by the formation of hydrogen bonds in the presence of hydroxyl groups in DGEBA381. Figure 3 also shows that the dynamic viscosity becomes frequency dependent at higher frequencies, particularly for higher levels of PPO in the blend, and similar behaviour has been reported by Venderbosch et al.
A comparison of the steady shear viscosity (Fig. 1) and complex viscosity (Fig. 3) shows very similar behaviour, indicating that the system complies with the empirical Cox–Merz relationship:
where η is the steady shear viscosity in Pa.s, η* is the complex viscosity in Pa.s, ω is the frequency in rad s−1 and is the shear rate in s−1. The steady shear viscosity at 0.1 s−1 and the complex viscosity at 0.1 rad s−1 are very similar as shown in Figure 2, but at shear rates greater than 10 s−1, the steady shear viscosity tends to be lower than the dynamic viscosity.
The variations of storage modulus (G′) and loss modulus (G″) and tan δ with angular frequency of the PPO:DGEBA340 blends at 200 °C are plotted in Figures 4 and 5. Similar behaviour was also observed for the PPO:DGEBA381 blends (data omitted for brevity). Most of the blends show power-law dependencies on frequency over the majority of the frequency sweep. Blends with 50 wt% PPO and more show a higher power-law dependence of G′ on frequency than G″, and tan δ decreases with increasing frequency suggesting liquid-like behaviour.[34, 35] However, tan δ is virtually independent of frequency for blends with 20–40 wt% PPO, which is characteristic of a gel-like state. This behaviour may be caused by small amounts of crystallization of PPO in the DGEBA solution or by an entanglement structure of the PPO chains. Physical gels such as hydrogen-bonded proteins, polysaccharides, block copolymer solutions and crystallizing polymer solutions[36-39] have also been found to obey the power-law dependencies of G′ and G″ on frequency which is similar to those for chemically crosslinked incipient gels.
Figure 4. Storage modulus (filled symbols) and loss modulus (open symbols) versus frequency for PPO:DGEBA340 at 200 °C.
Download figure to PowerPoint
The dynamic and steady-state rheology of 57PPO:43MCDEA was also studied at 200 °C—other blends with higher PPO levels were too viscous at the chosen processing temperature (200 °C) for the fully formulated blends to be measured. Figure 6 illustrates the dependence of moduli (G′ and G″) and tan δ on frequency. Tan δ decreases with increasing frequency and the scaling exponents for G′ and G″ are close to 2 and 1, respectively, which are typically found for simple liquids and polymer melts at high temperature. The steady shear viscosity (not shown) of 57PPO:43MCDEA has a zero shear rate viscosity of 500 Pa.s and starts showing shear-rate thinning above 1 to 10 s−1 and this is comparable to the complex viscosity at 1 rad s−1 of 466 Pa.s. As observed for the PPO:DGEBA data, the shear rate and frequency dependence of the steady shear viscosity and dynamic viscosity are similar indicating that the system complies approximately with the Cox–Merz relationship.
Dynamic mechanical properties
DMTA was performed on the PPO:DGEBA340 blends to determine their Tg. Blends with less than 50 wt% PPO could not be studied because the samples were very fragile and so attempts to carry out DMTA of these samples were unsuccessful. Figure 7 shows the storage modulus (E′) and loss factor (tan δ) of uncured PPO with different contents of DGEBA340 fron which it can be seen that the glass transition region shifts to lower temperature with increasing levels of DGEBA due to the plasticization effect which reduces Tg by approximately 40 °C per 10 wt% DGEBA340. Almost identical behaviour (not shown here) was observed for PPO:DGEBA381 blends. The PPO:DGEBA blends show similarly shaped tan δ curves—a main symmetric transition between 70 and 120 °C followed by a rise in tan δ at high temperatures which is commonly seen in uncrosslinked polymers with few entanglements. Similarly the elastic modulus passes through a sigmoidal transition before entering a brief plateau (entanglement) region and then the flow regime. However, the tan δ curves also exhibit a small shoulder below the main glass transition peak which increases in strength as the concentration of DGEBA is increased, suggesting the presence of an uncured epoxy-rich phase. Concomitantly, the modulus curves have small inflections between 20 and 150 °C indicating the glass transition of a second phase. These observations are consistent with the observed cloudiness at room temperature of the blends with more than 40 wt% DGEBA. It should be noted that optical transparency of polymer blends does not prove miscibility because the differences in refractive indices or the size of the domains may be too small to significantly scatter visible radiation. Further evidence of phase separation is apparent in the dependence of the peak temperature in tan δ on composition (Fig. 8)—Tg of the blends continuously decreases with increasing DGEBA340 content from 0 to 50 wt% DGEBA340. However, the blend with 60 wt% DGEBA has almost the same Tg as the 50PPO:50DGEBA340 blend (Fig. 8). This is further demonstrated when the Tg vales are compared with those predicted from the Fox equation—for 60 to 80 wt% PPO, the experimental values are slightly lower than those calculated but a large deviation can be observed for the 50 wt% DGEBA381 blend. This behaviour is consistent with large-scale phase separation for this blend at room temperature.
Figure 8. Tg calculated (Fox equation) and observed (DMTA) of PPO:DGEBA340 and PPO:DGEBA381 blends with various concentrations of PPO. Tg of DGEBA340 and DGEBA381 was measured using DSC—note that Tg determined using DMTA at 1 Hz is usually about 20 °C above the value determined using DSC.
Download figure to PowerPoint
PPO:DGEBA-based epoxy blends have been reported to exhibit upper critical solution temperature phase behaviour with cloud point temperatures ranging from 60 to 190 °C (and thus immiscibility below these temperatures) for 10–30 wt% PPO, depending on the structure or molecular weight of the DGEBA epoxy resin[3, 31] and on the molecular weight of the PPO.[10, 21] Venderbosch et al.[9, 10] observed that PPO:DGEBA blends with high concentrations of PPO (≥70 wt% PPO) are visually homogeneous at room temperature and have Tg values that appear to obey the Fox equation. In addition, using dynamic rheology during cooling the blends from the molten miscible state, Meijer et al. reported that blends with lower concentrations of PPO (10–60 wt% PPO) exhibit one constant Tg irrespective to PPO content. This behaviour can be explained by a thermoreversible phenomenon[8-10, 15] in which phase separation into polymer-rich and plasticizer-rich phases of the equilibrium composition is hindered by the vitrification of the polymer-rich phase, thus producing a single and similar Tg for differing polymer concentrations, known as the Berghmans point.[44, 45] Thus for blends with low PPO concentrations, the observed Tg values are equal to the temperatures at the intersection of the cloud point curve (for up to 60 wt % PPO) and the ideal Tg–composition line of a homogeneous system.[9, 10, 43]
For the PPO:DGEBA381 blends, the main Tg values are approximately 20 °C higher than those for blends with DGEBA340. This difference can only be partly ascribed to the higher Tg of DGEBA381 (−14 °C) compared to DGEBA340 (−17 °C) which would result in higher Tg of the PPO:DGEBA381 blends. However, the difference is more likely to be associated with a lower solubility of DGEBA381 in PPO, in agreement with the miscibility results of Merfeld et al. causing less of a plasticizing effect in terms of reduction of Tg of the blends. As found with PPO:DGEBA340, the Tg–composition plot (Fig. 8) shows that Tg of the main relaxation is approximately independent of composition at less than 60 wt% PPO. These results suggest that the Berghmans points for PPO:DGEBA340 and PPO:DGEBA381 are approximately 90 and 112 °C at 50–60 wt% PPO, respectively. In agreement, Ishi and Ryan observed a Berghmans point (using DSC) of 100 °C at 63 wt% PPO in a blend with an oligomer similar to DGEBA381.
Pure PPO crystals have a melting temperature of approximately 260 °C and a heat of fusion of approximately 40 J g−1. To determine whether PPO is able to crystallize from the PPO:epoxy melt, dynamic DSC studies (Fig. 9) were conducted on some of the samples with low and high concentrations of PPO. For the blend with 20 wt% PPO, one endothermic melting transition is observed at 34 °C. This transition could be due to either the melting of DGEBA crystals (DGEBA has a melting point near 42 °C) or impure PPO crystals. The former assignment is unlikely because the presence of PPO in DGEBA should decrease the melting point of DGEBA whereas melting point depression of PPO in the blends could cause this low melting point. The heat of fusion measured for the peak at 34 °C is 2.72 J g−1 (based on the total mass of the blend) which equates to 33.7% crystallinity of PPO. After cooling and reheating, this transition is replaced by a broad maximum at 157 °C with an endotherm of 7.47 J g−1 which might be due to a devitrification process. For the blend with 50 wt% PPO, an endotherm is observed at 70 °C in the first heating stage, which is most likely due to the fusion of PPO crystals because the endotherm is too high for DGEBA. The heat of fusion for this peak (13.3 J g−1) equates to 66.3% crystallinity for PPO. On cooling and reheating, this transition is replaced by two DSC glass transitions at 22 and 130 °C with heat capacity step (ΔCp) of 0.23 and 0.16 J g−1 °C−1, respectively. At first sight this result appears to contradict the DMTA data which give a DMTA Tg of 90 °C for the PPO-rich phase and another Tg due to the DGEBA-rich phase near 25 °C. Aside from the small difference in Tg determined using DSC or DMTA, the reason for the differences in these data is probably associated with the differing times allowed for phase separation to occur in the two specimens. In contrast, the 80PPO:20DGEBA340 blend shows only one DSC glass transition during heating and reheating at temperatures near 120 °C (ΔCp of 0.21 J g−1 °C−1). Since this system is believed to be a miscible blend, this value is in accord with the DMTA Tg of ca 155 °C, since Tg values measured using DMTA are invariably higher. Thus it appears that PPO can crystallize slowly from solutions in DGEBA when the blend composition is of an intermediate value, such as 50 wt% PPO, but not for high PPO contents, as with the 80 wt% PPO blend. The values measured here are also lower than the temperature (200 °C) used in the dynamic rheology experiments (Fig. 5) which indicate pseudo-gelation behaviour for blends with 20–40 wt% PPO, and so crystallization of PPO is unlikely to be the cause.
DMTA was also performed on the PPO:MCDEA blends. They only exhibit one glass transition (Fig. 10) which is consistent with the observation that all samples are transparent at room temperature. The variation Tg of the blend with composition is also shown in Figure 10 with Tg being reduced by approximately 38 °C per 10 wt% MCDEA. These results are in agreement with the qualitative information discussed above which indicates that addition of MCDEA to PPO makes it more readily processable.
Figure 10. E′ and tan δ of various uncured PPO:MCDEA blends at 1 Hz as a function of temperature and plot of Tg versus PPO content. Tg of pure MCDEA is estimated from its melting point based on the two-thirds rule.
Download figure to PowerPoint