STUDIES ON DENITRIFICATION-DEFLUORIDATION OF WATER USING SBR SYSTEM

Abstract

Nitrate and fluoride concentrations in groundwaters exceed desirable levels at various places throughout the world. According to the WHO guidelines and recommendations, the allowable concentrations ofnitrate and fluoride in drinking water are 10 mg/L (asN) and 1.5 mg/L respectively (WHO, 1984). Excessive intake of nitrate and fluoride can cause methemoglobinemia and fluorosis respectively. The Problem is more acute in rural and small urban communities particularly in third world countries. In India, 27845 habitations consisting approximately 25 million people are supposed to have water supply contaminated with fluoride (Paramasivam and Nanoti, 1997; Iyengar and Venkobachar, 1997). Another example is the Rift Valley region in Ethiopia (Ashley and Burley, 1994). Simultaneous occurrence of both nitrate and fluoride in high concentrations in groundwater samples has been noticed at several places. Three such examples are Rajasthan State and Agra region in India, and the Rift Valley region in Ethiopia where nitrate-N and fluoride concentrations as high as (i) 251 mg/L and 3.2 mg/L, (ii) 178 mg/L and 21 mg/L, and (ii) 224 mg/L and 26 mg/L respectively have been reported (Gupta, 1992; Pal, 1983; Ashley and Burley, 1994). Methods for fluoride and nitrate removal include ion exchange, reverse osmosis, and electrodialysis. Chemical precipitation and adsorption, and chemical reduction and biological denitrification are also used for fluoride and nitrate removal respectively. Of these methods, chemical precipitation by alum and lime and adsorption by activated alumina or bone char are extensively used for fluoride removal (Iyengar and Venkobachar, 1997), whereas ion exchange and biological denitrification are widely used for nitrate removal (Hoek and Klapwijk, 1987; Gayle et al., 1989). In India, defluoridation technique based on the principle offluoride precipitation using alum and lime, which could be operated in batch or continuous mode of operation, was developed and referred to as "Nalgonda" technique. This technique is simple with significant cost savings compared to other methods (Bulusu, et al., 1979). It is utilized for defluoridation of water with fluoride concentrations only up to 10 mg/L. Waters containing high fluoride concentrations (> 10 mg/L) require high dose of alum which results in an increase of sulfate and aluminum concentrations in the treated water to unacceptable levels (Bulusu, 1984; Gupta et al., 1999). Alum and powdered activated carbon (PAC) both have been found to remove fluoride from water separately. However, alum and PAC have not been used together. Therefore, in the present investigation their combination as alum-PAC slurry was tried for defluoridation particularly to reduce alum dose requirement at high initial fluoride concentrations. Although ion exchange and biological denitrification are widely used in nitrate removal from water, the treatment cost of the latter is relatively lower (Kapoor and Viraraghavan, 1997). Fluidized bed, fixed bed, and upflow sludge-blanket reactors have been tried for drinking water denitrification (Richard et al., 1980; Gayle et al., 1989; Green et al, 1994). Sequencing batch reactors (SBRs) have not been tried in drinking water denitrification, although there are many advantages and they are being used in biological wastewater treatment particularly for small communities (Wun-Jern, 1989). Some of the advantages of these reactors include simplicity in operation, cost effectiveness, and suitability to match with the operation of fill-and-draw (sequencing batch) type defluoridation plant working based on the Nalgonda technique. Hence the in possibilities of using an SBR for denitrification of water have been explored in the present investigation. Although several treatment methods are known for separate nitrate and fluoride removal, there is a need to explore the possibility of a treatment option for water containing relatively high fluorides and nitrates simultaneously. Thus, the objectives of this research work were to explore the defluoridation potential of alum-PAC slurry, to evaluate a dentification SBR unit, and to develop an integrated nitrate and fluoride treatment process using two SBRs for community water supply schemes. Studies were conducted in three phases using jar test apparatus (phase 1) and bench-scale reactors (phase 2 and 3). In the study on defluoridation by alum-PAC slurry (phase 1), optimum pH and operation procedure were assessed, and alum and PAC doses were optimized for initial fluoride concentrations of6,10,15, and 20 mg/L. Role ofeach component (i.e. alum, PAC, and lime) on fluoride removal and kinetics of the process were examined. In addition, effects of some water quality parameters i.e. phosphate, silica, alkalinity, sulfate, nitrate, chloride and dissolved organic matter (ethanol) on the treatment were investigated. In the denitrfication part of the treatment process (phase 2), using ethanol as external carbon source, the COD/N03"-N ratio was optimized, the efficiency of the SBR for initial nitrate concentrations in the range of 40-250 mg/L (as N) was investigated, and kinetics ofdenitrification was studied. Furthermore, effects oflength ofidle time and fluoride concentrations on denitrification were also investigated. Lastly in phase 3, treatment integration sequence options were assessed and the overall efficiency of the selected method was evaluated for various combinations of fluoride IV (6-20 mg/L) and nitrate (40-250 mg/L as N) by analysing eleven water quality parameters which were considered to be affected by the treatment process. The optimum pH for defluoridation by alum-PAC slurry was found to be in the range of 5.8 to 6.5. Operation procedure having two-step feeding (at the beginning of rapid and slow mixing) and two-phases of mixing with rapid mixing at 100 rpm for 40 minutes and slow mixing at 30 rpm for 20 minutes was found most appropriate to reduce initial fluoride to the acceptable level. Alum-PAC slurry having 500, 600, and 800 mg/L ofalum along with 100 mg/L ofPAC were found optimum to reduce initial fluoride concentrations of 6, 10, and 15 mg/L respectively to the acceptable level (1.5 mg/L). Use of alum-PAC in place of alum and lime reduced the requirement of alum dose by 40 and 43% at fluoride concentrations of 10 and 15 mg/L respectively. However, it had no significant advantage for lower fluoride concentrations (< 6mg/L). Fluoride concentration of 20 mg/L also could not be reduced to 1.5 mg/L with alum doses up to 900 mg/L along with 100 mg/L ofPAC. Combination ofalum with PAC doses greater than 100 mg/L did not provide improvement in fluoride removal. It was found that in this process fluoride is removed by alum during rapid and slow mixing, whereas direct removal by PAC took place only during rapid mixing. In this treatment process lime was used only for adjustment of pH and alkalinity. It did not provide fluoride removal by itself. The kinetics of the treatment process revealed that competitive fluoride removal by alum and PAC took place, which resulted in minor desorptions of fluoride from PAC mainly during the transition to slow mixing. The mechanism of fluoride removal by alum-PAC slurry is possibly adsorption and/or complexation on the aluminum hydroxide floe available in the bulk water and floe available on the surface of the PAC, and by adsorption directly on the surface of the PAC. Gradual diffusion of fluoride through the floe surrounding the particle to the PAC surface particularly at higher initial fluoride concentrations (higher concentration gradient) is also possible. Sulfate, nitrate, chloride, and organic matter (ethanol) did not show adverse effect on the process, whereas phosphate and silica did in different magnitudes. Sufficient alkalinity (0.5 to 0.63 mg/L as CaC03 for each mg/L of alum dose) is a prerequisite for the treatment, whereas excess of it would exert demand on extra dose of alum. In the denitrification treatment process, ethanol at COD/N03~-N ratio of 2.00 was found to be sufficient to reduce the initial nitrate to the acceptable level. Length of idle time (1-14 hours) and fluoride concentrations (6-20 mg/L) had no significant effect on the treatment process. Almost at all initial nitrate concentrations significant nitrate removal (85.7- 91.5 %) took place in the first hour of reaction. Nearly in all cases peak nitrite accumulations were noticed within the first six minutes ofreaction. It decreased rapidly thereafter. Considering the quality of the treated water in terms ofboth nitrate and nitrite, the SBR was found to be efficient in denitrification of nitrate in the concentration range of40 to 250 mg/L at anoxic reaction times (ARTs) of3, 5, and 7 hours for initial nitrate concentrations of 40-160, 200, and 250 mg/L (as N) respectively. The Monod kinetic parameters i.e. maximum specific denitrfication rate (kmax), half saturation coefficient (Ks), yield coefficient (Y), endogenous decay coefficient (k,,), and maximum specific growth rate (um) for denitrifcation were estimated and found to be 0.31 d"1, 0.46 mg N03"-N/L, 1.54 mg VSS/mg NOf-N, 0.009 d"1, and 0.48 d"1 respectively. Denitrification-defluoridation treatment sequence was found to be a better choice than defluoridation followed by denitrification. This treatment sequence was vi found to be promising only for treatment of water with fluoride and nitrate concentrations up to 15 mg/L and 80 mg/L (as N) respectively. Even though the quality ofthe treated water was within the acceptable limit, slight pH adjustment, filtration, and disinfection will also be additionally required to make sure the safety and potability of the water. At higher nitrate concentrations (> 120 mg/L as N), production of excessive alkalinity during preceding denitrification inhibited reduction offluoride to 1.5 mg/L when alum doses were restricted to 900 mg/L. Use of alum-PAC slurry for defluoridation of denitrified water reduced the alum dose by 40 and 43% (at initial nitrate concentration of40 mg/L as N), and 20 and 39% (at initial nitrate concentration of 80 mg/L as N) for initial fluoride concentrations of 10 and 15 mg/L respectively, compared to defluoridation by alum and lime alone. Some ofthe other advantages of this treatment sequence are (i) the alkalinity produced by the denitrification process is used up during defluoridation in raising the pH avoiding the need oflime for the same purpose, and (ii) residual amounts of turbidity, COD, sulfide, nitrogen, and denitrifying microorganisms in the denitrified effluent are removed by alum and PAC along with fluoride at the defluoridation stage. vn