Real time monitoring of photoautotrophic microalgae growth in photobioreactor, based on gas analysis

International audienceMicroalgae growth in a photobioreactor (PBR) is a complex process influenced by multiple parameters, such as photosynthetic light capture and attenuation, nutrient uptake, PBR hydrodynamics and gas – liquid (G – L) mass transfer. Like for any bioprocess, culture monitoring is a key aspect, as it allows controlling and optimizing operating parameters for improved production. In practice, real-time monitoring is mainly limited to pH and temperature. Online measurement of biomass concentration using optical density sensors needs improvements for providing accurate measurements. One alternative is to use on-line measurements such as pH, CO2 and O2 concentrations in- and off- gas stream. Software sensors based of first principle in combination with items as CO2 and O2 concentrations have been successfully implemented in fermentation since many years and in most recently in recombinant processes [1]. Despite their interest, the development of such tool in the context of microalgae application is underrepresented in the literature. Indeed, in phototrophs, CO2 and O2 are quantitatively the most important substrate and product, respectively. They allow quantifying on-line growth related parameters as net O2 production rate, CO2 uptake rate. Based on these quantities, biomass and nutrients concentrations could be estimated on-line. This study focusses on the development of such an estimation tool in the context of photoautotrophic growth cultivation of microalgae in closed PBR, in non-limiting mineral conditions. The proposed method uses a minimalistic model relying on macroscopic mass balances in liquid and gas phases and G - L mass transfer laws, without any assumption about the kinetic rate. Biomass concentration and growth rate have been estimated from net O2 production, assuming constant conversion yield. Combining in-situ O2 and CO2 measurements, the estimation of key parameters as dissolved CO2 and / or transfer coefficients is discussed. Although the methodology proposed here is rather general, online computation of net O2 and CO2 rates could be more or less complex, depending on gas vector through the PBR, G - L mass properties, and hydrodynamic conditions. A simulation study, based on a previous validated model [2] has been used to discuss limitations depending on the made assumptions. To demonstrate the methodology, the proposed tool has been experimentally implemented in a laboratory PBR throughout the photoautotrophic growth of a C. reinhardtii culture, its efficiency being proved under several operating conditions. Online estimation of biomass concentration has been in agreement with dry weight off-line measurements, even in batch operation, on a wide range of incident illumination conditions and gas flow rate through the PBR. In continuous conditions, online estimates have been successfully coupled with a feedback linearizing controller for biomass concentration. Since key process variables are available online, the proposed tool enables further control strategies. This is part of our ongoing research