Filtration by porous media: a microfluidics approach

The transport of colloids in porous media is governed by deposition on solid surfaces and pore-scale flow variability. Classical approaches, like colloid filtration theory (CFT), do not capture behaviours observed experimentally, such as non-exponential steady state deposition profiles and heavy tailed BreakThrough Curves (BTC). In the framework of CFT, a key assumption is that the colloid attachment rate $k$ is constant and empirically estimated via a posteriori macroscopic data fitting. We design a novel experimental set-up based on time-lapse microscopy and continuous injection of fluorescent monodisperse colloids into a folded microfluidics device (1mt total length) designed with a controlled level of 2D spatial disorder. This set-up allows us to i) measure both BTC and deposition profile over several orders of magnitude and ii) to perform particle tracking and Lagrangian analysis of single colloid's trajectories. Based on this analysis, we propose a stochastic model that takes into account pore scale heterogeneities in terms of correlation length, velocity and attachment rate distribution, that captures the anomalous behaviour shown by the experimental data.