Chemotactic Bacteria are Retained Near NAPL Ganglia in Saturated Porous Media: Microscopy, Microfluidics and Modeling over Multiple Scales

Abstract:

In situ bioremediation in natural aquifers can be accelerated by chemotaxis due to its efficiency in transporting bacteria toward contaminants that serve as chemoattractants and carbon sources. However, the transport and distribution of bacteria in porous media under flow is challenging to predict, especially for chemotactic bacteria that respond to nonaqueous phase liquids (NAPLs). In this work, we directly imaged Pseudomonas putida G7 (PpG7, chemotactic to naphthalene) and nonchemotactic Pseudomonas putida G7 Y1 (PpG7 Y1) near NAPL ganglia (chemoattractant naphthalene dissolved in 2,2,4,4,6,8,8-heptamethylnonane) trapped within the pore network of a microfluidic device. Then, we simulated bacterial transport at the pore scale using a convection-dispersion equation with the addition of chemotactic velocity to the convective term and first-order sorption-desorption kinetics for retention around NAPL ganglia. Previous simulations at the core scale in granular media showed that the heterogeneous hotspots of chemotactic PpG7 around NAPL ganglia yielded the overall greater loss of biomass compared to nonchemotactic PpG7 Y1 in breakthrough curves. Our experimental observations at the pore scale confirmed the simulation results by revealing greater accumulation of PpG7 over PpG7 Y1 near ganglia of naphthalene sources. Our simulation results showed that greater retention of PpG7 was due to its chemotactic response to naphthalene gradients and sorption to NAPL ganglia; interestingly, no PpG7 accumulation was observed in regions with steep naphthalene gradients where higher shear rates (beyond 0.13 s^-1) interfered with chemotaxis. Direct experimental observations revealed that accumulated bacteria were actively bouncing back and forth on NAPL surfaces, instead of adsorbing irreversibly as we originally expected. Nevertheless, our model can sufficiently characterize the transport of chemotactic bacteria by including the chemotactic velocity and adjusting sorption kinetic parameters at NAPL surfaces. Our modeling predictions in combination with laboratory experiments at varying scales can be a useful tool to predict the effects of chemotaxis in in situ bioremediation.