A 5 L UASB reactor with 4.5 L working volume (R1) was made of stainless steel and cylindrical in shape (Fig. 12). The reactor was constantly stirred at 50 rpm. The pH of mixed liquid in R1 was controlled automatically with 6 M NaOH. The temperature in R1 was maintained at 37°C by inserting the reactor in the thermostatic chamber. For start-up, the reactor was filled with undiluted UF whey permeate and was operated anaerobically at a batch mode. When hydrogen production reached its peak value, the bioreactor feeding mode was turned to a continuous one at a HRT of 24 h (OLR of 10 g COD L-1 d-1) or at a HRT of 12 h (OLR of 20 g COD L-1 d-1). The R1 performance (biogas production and composition in H2 and CH4, COD, Total Volatile Fatty Acids — TVFA concentration, pH) was monitored twice a week throughout the experimental period (84 days).

A 5 L UASB reactor with 4.5 L working volume (R2) was made of stainless steel and cylindrical in shape (Fig. 12). The reactor was constantly stirred at 50 rpm. The pH of mixed liquid in R2 was controlled automatically at pH 7.2 (±0.05) with 6 M NaOH. The temperature in R2 was maintained at 37°C by inserting the reactor in the thermostatic chamber. The R2 was fed with the effluent from R1, which was collected in a 3 L container used as a storage tank which was constantly stirred at 50 rpm (Fig. 12). The temperature in the tank was maintained at 37°C by placing it in the thermostatic chamber. Overflow effluent flowed out in the top part of the storage tank and was collected in a separate container. The R2 was operated at an HRT of 3 d. The R2 performance (biogas production and composition in CH4, COD, TVFA concentration, pH) was monitored twice a week throughout the experimental period, from day 51 to 84. Before R2 was fed with R1 effluent, the diluted UF whey permeate had been used as a feedstock to reach the OLR of 2 g COD L-1 d-1 and HRT of 3 d.