Bacterial biofilms are sessile communities that develop on surfaces and represent the dominant form of life for bacteria [1][4] [Figure1.A]. Biofilms play a key role in a variety of engineering problems found in porous media sciences, ranging from soils bioremediation to bioreactors for the production of proteins of interest or for filtering wastewater [2]. On the other hand, uncontrolled growth can induce clogging and loss of efficiency in technical systems, biofouling in industrial and medical devices, or even resistant contaminations [3][6] [Figure1.B]. Developing approaches to regulate the development of the biomass in porous structures could pave the way towards a new field in control engineering, aiming at controlling the permeability of porous media colonized by biofilm [5].

In particular, it has been observed recently that a competition between growth and detachment leads to large amplitude fluctuations of the pressure drop [6]. For example, in a bioreactor system for the filtration of wastewater or the production of a protein of interest, such fluctuations could induce important changes of yield [2][3]. In this work, we ask the question of whether we can avoid such fluctuations and more generally control the permeability of porous structures colonized by biofilms. We hypothesized that a closed loop control could be implemented to reach a target permeability in a demonstration microfluidic device.

We will first present a novel experimental technology that allows us to measure the dynamics of the pressure drop across a porous structure, while working in a microfluidic system with accurate control of environmental conditions. The core of the system is a 3D printed microbioreactor containing a porous structure, where biofilms develop. The porous structure is printed by stereolithography and composed of 300µm wide/700µm long channels, with a connectivity of 3. This system is adapted for control because we can: precisely define and change the structure and the material of the porous media; change nutrient types, concentrations and flow rate through programmable gear pumps; impose temperature to favour or not growth with a thermostat cell, while measuring the effect of those different parameters through a pressure sensor and oxygen sensor [Figure2]. The system can also be imaged via X-ray microtomography with a newly developed contrast agent based upon gold nanoparticles.

We will show a simple example of how pressure fluctuations can be controlled for a biofilm of Pseudomonas aeruginosa and discuss the implications in terms of the definition of a cost function. We will further present our global strategy for control, which relies on feedback control of four environmental conditions [Figure2.A]: nutrient concentrations, temperature, flow and the presence of bacterial predators-in particular Bdellovibiro bacteriovorous [8].