Denitrification of drinking water by the association of an electrodialysis process and a membrane bioreactor: feasibility and application

A hybrid process combining electrodialysis and a membrane bioreactor was investigated to treat ground waters with excessive nitrate concentrations. Electrodialysis (ED) allowed the nitrate separation producing, on one hand, partially demineralized waters whose ion concentrations were in agreement with the norm, and on the other hand, brines that were treated by a membrane bioreactor. Experiments performed at laboratory scale with synthetic solutions and pre-industrial scale to treat a ground water contaminated by nitrates showed high efficiency of the hybrid process. The nitrate concentration of the treated water remained below the acceptable value (50 mg/l−1) and even below the recommended value (25 mg/l−1) for drinking water. Moreover, the ED treatment induced a softening of the treated water. The biological denitrification allowed the almost total removal of nitrates (99%) with kinetics close to 0.3 kgN–NO3/kg MVS/d and a limited sludge production of 0.5 gV.S.S/gN–NO3.

− Preferably [NO3 ]>500 mg/l Optimum for 7.3<pH<8.4 No influence Denitrification rates mgN-NO3-/mgVSS/h 0.003 < K0 < 0.037 Mean value: 0.012 Literature values (ethanol as carbon source): Martin et al. [7] 0.027 Richard and Leprince [8] 0.003–0.015 Henze et al. [9] 0.01–0.02 each experiment [5]. We can observe that the denitrification is almost complete 10 h after the injection. The linear evolution of nitrate concentration during these 10 h led us to suppose that the kinetics were of zero order with respect to nitrate concentrations. The values calculated from the slope of the linear regression of this linear evolution allows calculation of the specific − denitrification rate K0 (mgN-NO3 /mg VSS/h). Table 2 reports the operating conditions, results and comments about the efficiency of denitrification: • The biokinetics quantified with various initial nitrate concentrations (ranging from 300 to − 1600 mg NO3 l−1, with a constant C/N ratio of 1.3) showed that the specific denitrification rates were better with high initial concentrations and that no inhibition occurred. • The optimum pH range (7.3–8.4) corresponds to the optimum values noticed in the literature with ethanol as the carbon source [6,7]. • In our operating conditions (conductivity ranging from 3.3 to 19.3 mS cm−1), no signifi- cant influence of the salinity was observed on the biokinetics. In these batch conditions, where the biomass concentration ranged from 500 to 2500 mg VSS/l according to the different experiments, the kinetics were of zero order and the specific nitrate removal rates were close to 0.3 mg − N-NO3 /mg VSS/d, which is in good agreement with the literature values [7–9]. The subsequent experiments on the continuous mode were performed with synthetic solutions whose compositions were very close to those of ED concentrates. Three reconstituted concentrates were used: concentrate I, corresponding to the concentrate obtained by treatment of the natural water 2 (Table 1); concentrates II and III, corresponding to concentrates obtained in ED plants for nitrate removal. They constituted the MBR influent. Hydraulic and solids residence times in the MBR were equal to 5 h and 7 d, respectively. Table 3 gives the operating conditions and the performance of the denitrification process.