Spinodal nanostructures in polymer blends and the Cahn-Hilliard lengthscale prediction

Spinodal decomposition of partially miscible polymer blends has the potential to generate well-defined polymeric nanostructured materials, with precise control of lengthscale and connectivity, and vast applications ranging from membranes to photovoltaics. We examine experimentally the validity of the classical Cahn-Hilliard theory prediction for the initial spinodal lengthscale q*=sqrt(-G''/4k), where G’’ is the second derivative of the free energy of mixing with respect to composition, and k is the prefactor of the square gradient term, accounting for additional free energy arising from concentration gradients. We contrast an unprecedented series of Λ≡2π/q* estimates, self-consistent with the theory, and independent -G’’(T) and k measurements, and overall find the prediction to be remarkably accurate for all blends and conditions examined. No breakdown of this prediction is observed at the smallest Λ recorded experimentally, which are however still several times larger than Rg. Finally, we outline design considerations and limitations for generating polymeric materials via spinodal decomposition, bound by thermodynamics of available polymer systems, coarsening kinetics governed by rheology, as well as by engineering constraints.