Simulations of sieving characteristics of macromolecules in porous membranes at high concentrations

Abstract

The effect of concentration on the diffusive hindrance coefficient D eff/ D ∞ and the convective reflection coefficient σ of random-coil polymers in porous membranes were evaluated. Membrane sieving of nonideal flexible polymers at higher concentrations was modeled in molecular simulations. The partition coefficient Φ of flexible chains in equilibrium with channel- or slit-like pores was computed in good and pseudo-ideal (theta) solvents as a function of coil-to-pore size ratio λ = 2 R g/ d. An increase in concentration brings about enhancement of the partitioning and the change of the partitioning curve Φ versus λ from a convex to a concave shape. Enhanced concentration brings about a considerable permeation of polymers into pores smaller than coil size ( λ > 1) and thus a shift of the cut-off separation limit of membranes. The steric retention data from simulations were combined with the hindrance functions from the hydrodynamic theory to evaluate the coefficients of hindered transport at finite concentrations. For a given coil-to pore size ratio λ, the computations predict an increase of hindered diffusion D eff/ D ∞ with concentration ϕ, in harmony with measurements of dextran and polyethyleneoxide. On the other hand, the computed reflection coefficient σ, a measure of hindered convection in ultrafiltration, decreases with ϕ for a given λ, in qualitative accord with ultrafiltration data reported for polystyrene. The reduction of the coefficient σ by concentration is more intense for random-coil solutes than for hard-sphere solutes.