PHASE-SEPARATION BEHAVIOR IN BLENDS OF POLY(BENZIMIDAZOLE) AND POLY(ETHER IMIDE)

A 50/50 blend of a poly(ether imide) (PEI) and a poly(benzimidazole) (PBI) has been studied by small-angle X-ray scattering and solid-state nuclear magnetic resonance. This blend, cast from a dimethylacetamide solution, is shown to be intimately blended on a scale of 2.5 nm. Phase separation is detectable upon annealing for 1 h at temperatures of 310-degrees-C and above even though the DSC-determined glass transition temperature for this blend composition is 344-degrees-C. Proton spin-diffusion concepts, along with proton multiple-pulse methods and magic angle spinning, are used to estimate the minimum domain dimensions (MDD's) in those annealed blends which display phase separation. The process of phase separation is interesting in this blend since kinetics and thermodynamics are expected to influence phase separation. The disparate T(g)'s (220-degrees-C for PEI and 420-degrees-C for PBI) suggest that the vitrification of the PBI-rich phase will arrest the growth and composition of phases in a phase-separation process. In keeping with these expectations, relatively small MDD's are observed; they range from 5 to 23 nm for annealing temperatures from 310 to 400-degrees-C. Stoichiometric information regarding the composition of the PBI-rich phase has also been extracted from the spin-diffusion data. This composition moves closer toward a pure PBI phase as the annealing temperature is raised. Surprisingly, the composition of the PBI-rich phase has a DSC-determined T(g) of about 400-degrees-C after annealing for 1 h at only 340-degrees-C. Thus one must be cautious about how one relates DSC-determined T(g)-versus-composition data to long-term blend stability. The SAXS scattering from the phase-separated blend samples is very weak. SAXS curves showed no peaks m the small-angle region; therefore, the data were analyzed using a random-size, two-phase model from which a correlation length, xi, was obtained. The xi values were 3.1-4.6 times smaller than the MDD's when analysis was carried out on the same samples. This disparity, it is argued, implies that the dominant electron density fluctuations occur within the phases rather than originating from the difference m chemical composition between the phases. It is conjectured that such fluctuations could arise from stress buildup during cooling, owing to differential contraction in a bicontinuous-phase morphology. Finally, some simple solubility experiments were conducted on annealed samples. Evidence for cross-linking was seen in both the homopolymers and the blend. However, the suggestion was strong that a PEI/PBI cross-link was favored over a homopolymer cross-link. It remains an open question whether cross-linking or kinetics is responsible for arresting the growth of domains.