Lactate production from lactose-rich wastewater: A comparative study on reactor configurations to maximize conversion rates and efficiencies.

About 90 % of global lactate production is derived from bacterial fermentation of sugars via pure homofermentative cultures in batch mode. Acid whey, which is a lactose-rich wastewater from the yogurt industry, can be used as an alternative substrate for commercial lactate production. Operating reactor microbiomes reduces the lactate production costs by circumventing sterilization, while continuous operation with biomass retention achieves higher productivity at shorter production times. To find the best reactor configuration with biomass retention for lactate production from acid whey, we operated three different reactor configurations: (1) an upflow anaerobic sludge blanket (UASB) reactor; (2) an anaerobic filter reactor (AFR); and (3) an anaerobic continuously stirred tank reactor (CSTR) with a hollow-fiber membrane module. We operated at different hydraulic retention times (HRTs) to find the optimum production parameters at a temperature of 50 °C and a pH of 5.0. We did not use an inoculum but enriched the endogenous D-lactate-producing Lactobacillus spp. that later dominated the reactor microbiomes (> 90 % relative abundance). Undissociated lactic acid concentrations of more than 60 mmol C L inhibited the microbiomes. We alleviated the inhibition effect by shortening the HRT to 0.6 days and using diluted acid-whey substrate (1.67-fold dilution) to achieve almost complete conversion of the acid-whey sugars to lactate. At the 0.6-day HRT, the AFR and CSTR performed better than the UASB reactor due to their better cell retention abilities. During the period between Day 365-384, we experienced an error in the pH control of the CSTR system during which the pH value dropped to 4.3. After this pH-error period, the lactose and galactose-into-lactate (LG-into-LA) conversion efficiency for the CSTR considerably improved and surpassed the AFR. We achieved the highest lactate conversion rate of 1256 ± 46.3 mmol C L d (1.57 ± 0.06 g L h) at a LG-into-LA conversion efficiency of 82.2 ± 3.4 % (in mmol C), with a yield of 0.85 ± 0.02 mmol C mmol C (product per consumed substrate) for the CSTR.