Modeling virus filtration based on a multilayer membrane morphology and pore size distribution

Abstract

Obtaining adequate viral clearance in the manufacture of biopharmaceuticals is a major challenge. While virus filtration is routinely used for this purpose, there have been few attempts to model virus filter performance. Here, virus filter performance is modeled in consideration of a multilayered membrane structure with a membrane pore size distribution that is assumed to be the same log-normal distribution in each membrane layer. Comparison of experimental data both for constant pressure and constant flux filtration to the model calculations showed good agreement. This method relies on three fitted parameters that can easily be determined from experimental data: particle size that blocks the membrane pores, particle density and the maximum pore size that can capture a particle. In addition, high virus removability was confirmed by the model calculations. By applying the calculation method introduced here, it becomes possible to quantitatively reproduce and analyze the filtration characteristics of virus filters. Based on analysis using the numerical calculation method presented in this study, we demonstrated the equivalence of constant pressure filtration and constant flux filtration. This method also shows potential for predicting filtration behavior based on minimal experimental data.