3. Results and discussion
3.1. Start-up of the AnMBR
The start-up period of AnMBR lasted for 37 d, and an overview of the system performance is presented in Fig. 2. As shown in Fig. 2A, the COD removal rate increased significantly with an increase in influent COD concentration during the first 10 d. At the end of start-up (from day 33 to day 37), COD removal averaged 92.9% corresponding to an average permeate COD concentration of 82.8 mg/L. By contrast, the averaged soluble COD concentration in bioreactor was as high 138 mg/L. The difference in bioreactor and permeate soluble COD concentration averaged 55 ± 10 mg/L, which corresponded to 9 ± 5% of the total COD removal. This finding suggested that removal of soluble COD across the membrane resulted in a high-quality effluent in AnMBR. The mechanism of soluble COD removal across the membrane may be related to microbial activity, size or charge exclusion, and/or adsorption [29]. Moreover, the production of biogas as a source of energy is an important factor in characterizing AnMBR. As shown in Fig. 2B, the methane yield increased from 60 mL/d (day 2) to 2453 mL/d (day 28) with increasing OLR from 0.83 kg COD/m3 to 2.32 kg COD/m3. During the stable phase of OLR (on average 2.36 ± 0.09 kg COD/m3, from 28 d to 37 d), the methane yield averaged 2179 mL/d, corresponding to 263 mL/g COD removed. Nevertheless, this value represents only 75.1% of the maximal theoretical value of 350 mL/g COD removed at 35 °C [30].