Process development of propionic acid production by fermentation
The fermentation kinetics of Propionibacterium thoenii strain P20 were studied with calcium alginate bead-immobilized cells without mass transfer limitations. The fermentation was found to be significantly affected by the end-product, propionic acid, in the medium. High levels of the acid not only inhibited cell growth, but also changed cell metabolism to produce more byproducts such as acetic, succinic and lactic acids. As a result, substrate consumption rate and propionic acid productivity and yield also decreased;Coupling liquid-liquid extraction of the acid with the fermentation, that is, extractive fermentation, alleviated the inhibitory effects. Solvent selection, based on partitioning and toxicity, was carried out among three solvent systems: Alamine°ler 304-1 in 2-octanol, 1-dodecanol, and Witcohol°ler 85 NF (oleyl alcohol). For each system, the acid loading per amine molecule decreased with increasing amine concentration; hence, an optimal amine concentration existed. The solvent toxicity to five strains of propionibacteria was determined. Strain P20 was the most solvent-resistant strain and Alamine 304-1 in Witcohol 85 NF was the only nontoxic solvent. Addition of vegetable oils to scavenge the toxic diluents was also able to reduce toxicity;Extractive fermentation was carried out with 40% (w/v) bead load of immobilized P20 cells and a hollow-fiber membrane extractor. By maintaining the propionic acid concentration at about 13 g/L, the yields of propionic and acetic acids were more than doubled compared to those of the nonextractive process. Higher acid productivities, an 80% reduction of base consumption for pH control, and solvent selectivity of propionic over acetic acid were also achieved;Vacuum distillation produced purified acids as the overheads and recyclable solvent as the bottoms. No solvent degradation occurred under the distillation conditions;Based on the experimental findings, economic evaluations of the extractive fermentation process on a 50,000-ton annual production scale were conducted. The process was modified to use whole-cell extraction with a mixer-settler extractor. This process suggested that economic feasibility could only be realized with favorable assumptions such as long-term bead usage, byproduct revenue, and inexpensive substrate.