Pore network model for permeability characterization of three-dimensionally-printed porous materials for passive microfluidics.

Abstract

Stand-alone and portable lab-on-chips (LOC) can be obtained by exploiting capillary flow in porous media. Polymethylmethacrylate (PMMA) platforms obtained through powder-based three-dimensional (3D) printing are appropriate for capillarity-driven LOCs. However, fluid flow in such platforms needs to be characterized well. For this purpose, a 3D pore network (PN) was extracted from high-resolution μCT images of printed PMMA through a watershed algorithm and a PN model was developed with the final goal of characterizing material permeability. The effect of all parameters involved in the PN extraction and modeling was investigated. The study focused in particular on the effect of the number of seeds for the watershed segmentation, pore sphericity, and pore-to-pore channel shape that was modeled as a bicylindrical or biconical object. The results proved that all PN extraction and modeling parameters influenced the permeability, which was found to be lower the higher the number of seeds and when using sphericity and biconical channels. Eventually, the Calinski-Harabasz index value was used to identify the optimal number of watershed seeds.