A model for osmotic flow in porous membranes is developed from classical transport and thermodynamic relations. Mathematical expressions for the reflection coefficient as a function of solute dimension and shape, and more generally pore/bulk distribution coefficient, are derived for long cylindrical pores of circular cross section. For a rigid, spherical macromolecule the osmotic reflection coefficient equals (1 - Phi)(2), where Phi is the solute distribution coefficient; this result differs significantly from expressions found in the literature. The effect of weak solute adsorption to (or repulsion from) the pore wall can also be accounted for in the derivation. The driving force for osmotic flow arises from solute-pore wall interactions which cause radial variations in concentration and concomitant gradients in pressure normal to the wall. Implications of this three-dimensionality of osmotic phenomena are discussed with particular reference to the adequacy of one-dimensional treatments in relating reflection coefficient to membrane and solute properties.

No abstract

Diffusional mass transfer across membranes with uniform but low pore densities was studied experimentally as a function of stirring rate and pore area fraction. The results were analyzed in terms of a stagnant film boundary-layer model specially formulated for a heterogeneous membrane containing discrete pores. The overall membrane diffusional resistance is linearly related to the inverse of the pore area fraction of the membranes for a constant stirring rate. An equivalent boundary-layer thickness can be defined which is independent of membrane properties but a unique function of stirrer speed. These experimental boundary-layer thicknesses are greater by a factor of 3 than those predicted by a published Sherwood number correlation determined for homogeneous surfaces, but the stirring rate dependence is in excellent agreement with this same correlation.

Aqueous sodium alginate solutions were subjected to various heat sterilization treatments. Sodium alginate powder was also treated by both gamma-irradiation and ethylene oxide sterilization. The effects of these treatments on the viscosities of sodium alginate solutions and both the diameter and strength of the beads formed in 0.1 M CaCl2 solutions were determined quantitatively. The viscosity of sodium alginate solutions and the gel strength of the calcium alginate beads decreased with increasing sterilization temperature while the bead diameters were found to increase. All these effects can be attributable to a reduction in the degree of polymerization of the alginate molecules as a result of the heat treatments. Ethylene oxide and gamma-irradiation treatments caused similar effects. Standard conditions for sterilization are necessary for comparative studies with alginate beads.

No abstract