Continuous butanol fermentation with submerged ceramic membranes

Introduction. Building block chemicals, such as butanol, can be produced via fermentation and biobased feedstocks as a more sustainable alternative compared to the current petrochemical methods. However, the butanol fermentation is subjected to low product titres (<20 g/L) due to product inhibition of the bacteria [1]. To overcome this continuous fermentation can be applied, as it offers considerably higher volumetric productivities compared to batch fermentation. However, it suffers from lower yields and concentrations due to the continuous dilution effects due to the feed stream. An integrated membrane-fermentation process would retain the microorganism in the fermenter, resulting in higher biomass concentrations and thereby enhancing both the yield and the volumetric productivity. Most research combining membrane and fermentation technology uses an external membrane unit [2]; but this leads to problems with sterility, increased process volume and cell viability due to shear stress in the circulation pump. In contrast, less research has been performed using submerged membranes. The key problem with membrane bioreactors for continuous fermentation is a stable membrane operation with small membrane area to fermenter volume ratio due to internal space limitations. Especially with regards to minimising membrane fouling, allowing operation for 150+ hours with no chemical cleaning agents.Experimental/methodology. Continuous acetone butanol ethanol fermentation was performed using Clostridium beijerinckii in a 5 L fermenter. A silica carbide membrane (0.1 μm, 0.01 m2, Cembrane) was submerged in the fermenter. The permeate was constantly logged, providingonline flux measurements. The membrane was periodically backflushed with deionised water, with the frequency of thebackflushing varied to find the optimal operating conditions, aiming to maintain the flux at 20 kg/m2·h.Results and discussion. While a continuous fermentation process without a membrane achieved a cell density of 0.6 g/L and butanol productivity of 0.10 g/L·h, incorporating a membrane into the fermenter tripled the cell density to 1.8 g/L. However, this high density could only be maintained for 50 hours before a decrease in membrane flux resulted in cell loss. This flux decline caused more broth toleave through the waste line rather than as the desired product in the permeate, therefore proportionally increasing cell loss. Maintaining a flux of 20 kg/m2·h was crucial for fermentation performance. Backflushing the membrane was implemented, initially every 2 minutes for 10 seconds. This maintained the flux at 19 kg/m2·h, but significantly diluted the broth, keeping the cell density at 0.7 g/L. By optimizing backwash duration and frequency (as shown in Figure 1), a steady flux of 16-18 kg/m2·h was achieved when the backflushing occurred for 10 seconds. Reducing the backflush frequency to every 20 minutes minimized dilution by 85%. This significantly improved the fermentation: peak cell density reached 3 g/L, with an average of 2 g/L (three times higher than the control without a membrane), and butanol productivity increased by 24% to 0.13 g/L·h for a stable 150 hours of operation.Conclusions and outlook. This work has demonstrated the successful operation of a submerged membrane bioreactor under sterile operation for 150 hours. With further optimisation of the fermentation and membrane operation it will be possible to increase the productivity gains further, improving the viability of sustainable manufacturing routes.References [1] Jones, D.T.; Woods, D.R. Microbiol Rev 1986, 4, 44-524.[2] Carstensen, F. et al. J. Membr Sci 2012, 394-395, 1-36.