François-Nicolas de Glutz
The objective of this thesis was to evaluate the feasibility for the production of fuel bioethanol from whey permeate. The identification of the limiting factors of the process revealed that the stabilization of the substrate was critical to the use of whey as a fermentation substrate and that the conversion yields and ethanol tolerance limited the process. A comparison of the possible scenarios for the production of bioethanol from whey was first performed. It was demonstrated that fresh whey should be concentrated at the production site before being transported to a centralized treatment plant, since transporting dilute material or producing ethanol at the dairy would result in too high operating costs. An economic comparison of a treatment plant, which (i) "directly" ferments lactose into ethanol using Kluyveromyces fragilis and (ii) "indirectly" ferments hydrolyzed whey using Saccharomyces cerevisiae, resulted in similar production costs of 1.35 CHF/LEtOH and 1.32 CHF/LEtOH respectively. Whey stabilization was investigated by testing the addition of chemical compounds to whey. These were evaluated according to their ability to prevent the growth of lactic acid bacteria which were identified as being mainly responsible for whey instability. Formic acid (50 mM) or hydrogen peroxide (100 mM) were shown to extend the stability from 2-3 days, at 4°C to 21 days at ambient temperature. However, such concentrations of preservative also inhibit the growth of yeasts, therefore they must be removed prior to fermentation. Of the compounds tested, formic acid was preferred due to the high level of bacterial growth inhibition at pH 4, and its non-toxicity for yeast at neutral pH. This would enable whey to be stored at room temperature for a three-week period without negatively influencing the subsequent yeast fermentation. Global productivity of the initial fermentation cultures was considerably reduced as a result of a long non-productive lag phase. In order to improve the understanding of the principle factors which, not only influence the duration of the lag phase, but also the biomass produced during a pre-culture period of 24h and the maximum growth rate in fermentation cultures, six pre-culture parameters were tested alone or in combination on two ethanol-producing yeasts, K. marxianus CBS 5795 and CBS 397, using whey permeate as substrate through the application of design of experiment procedures. The key parameters identified through this strategy were: influence of temperature, type of sugar, culture mode, initial biomass concentration and initial sugar concentration. Optimum ethanol productivity was achieved by cultivating the pre-culture anaerobically on medium containing lactose, which resulted in an improvement of the productivity by 10-11% compared to an aerobic pre-culture with glucose. The principal organism studied for ethanol fermentation from whey permeate was K. fragilis due to its ability to directly ferment lactose. However, such direct fermentation yeasts generally suffer from low conversion yields and poor tolerance to ethanol (2-3% v/v). An alternative is to utilize indirect fermentation yeasts, such as S. cerevisiae, which show considerably better ethanol fermentation performance but has the disadvantage that an expensive enzymatic hydrolysis step is required prior to fermentation. In this study both types of process have been characterized involving eight ethanol producing yeasts. The culture conditions were optimized for each strain using a design of experiment methodology. Highest conversion yield and alcohol tolerance were achieved with S. cerevisiae Ethanol Red (YP/S= 0.662 C-mol/C-mol, cEtOHmax= 148 g/L), of the indirect fermentation yeasts, and with K. marxianus CBS 5795 (YP/S= 0.660 C-mol/C-mol, cEtOHmax= 79 g/L) of the direct fermentation yeasts studied. Introducing the data obtained from cultures with these yeasts to former economic evaluations of both scenarios, showed that direct fermentation should be preferred for fermenting whey permeate to ethanol. A maximum volumetric productivity of 6.24 g/(L·h) at 37°C and pH 4 was achieved with K. marxianus CBS 5795. Fermentation of non-sterile whey permeate with a consortium (CEKI) of K. marxianus (S1), I. orientalis (S2) and E. faecalis (S3), isolated from spontaneous cultures, was then studied. Maximal ethanol yield of 0.65 C-mol/C-mol, as highest ethanol concentration of 55 g/L, was achieved with CEKI, at 37°C and pH 4, compared to isolated cultures of these organisms. The results also suggest that E. faecalis exhibits a protective effect against lactic acid bacteria. Specific productivity of CEKI was 0.21 gEtOH/(gbiomass·h). Finally, a novel in-situ product recovery method, based on capsular perstraction with an organic solvent, was developed for ethanol extraction in batch fermentation systems using CEKI, which utilizes microcapsules. The production of capsules of 2 mm diameter that contained a hydrophobic core of laurinaldehyde and an alginate-based membrane enabled (i) to reduce the toxicity of the solvent for the growing cultures, and (ii) make the separation of the organic phase easier for a subsequent ethanol recovery. For the produced capsules mass transfer was determined by the solvent layer (0.27·10-5 cm2/s), which resulted in a maximum specific ethanol recovery of 3.17 gEtOH/(gsolvent·s).
EPFL2009
