ABSTRACT
Two computational approaches, namely Brownian dynamics and network modeling, are presented for predicting effective diffusion coefficients of probes of different sizes in three chromatographic adsorbents, the structural properties of which were determined previously using electron tomography. Three-dimensional reconstructions of the adsorbents provide detailed, explicit characteristics of the pore network, so that no assumptions have to be made regarding pore properties such as connectivity, pore radius and pore length. The diffusivity predictions obtained from the two modeling approaches were compared to experimental diffusivities measured for dextran and protein probes. Both computational methods captured the same qualitative results, while their predictive capabilities varied among adsorbents.
