Streamlined Ammonia Removal from Wastewater Using Biological Deammonification Process

Streamlined Ammonia Removal from Wastewater Using Biological Deammonification Process Abstract Matias Vanotti 1 *, José Martinez 2, Albert Magrí 3, Ariel Szögi 1, Takao Fujii 4 1 USDA-ARS, 2611 W. Lucas St., Florence, S.C., 29501, USA 2 IRSTEA, 17, avenue de Cucille, Rennes, 35044, France 3 GIRO Technological Centre, Mollet del Vallès, Barcelona, Spain 4 Sojo University, Department of Applied Life Sciences, Kumamoto, Japan *Corresponding author. E-mail: matias.vanotti@ars.usda.gov In this work we evaluated biological deammonification process to more economically remove ammonia from livestock wastewater. The process combines partial nitritation (PN) and anammox. The anammox is a biologically mediated reaction that oxidizes ammonia (NH 4 + ) and releases di-nitrogen gas (N 2 ) under anaerobic conditions using nitrite (NO 2 - ) as the electron acceptor. We tested two deammonification configurations: 1) a two-stage process (PN in one tank and anammox in another), and 2) a one-stage deammonification process (PN and anammox in a single tank). In both cases, a high performance nitrifying sludge, HPNS (NRRL B-50298) was used for PN and Brocadia caroliniensis (NRRL B-50286) was used for the anammox reaction. The two-tank process was carried out using sequencing batch reactors (SBR), one aerated and the other anaerobic, at 32 o C water temperature. The single-tank process was carried out in an aerated vessel operated under continuous flow. It contained biofilm plastic carriers at 30% v/v packing ratio that were fluidized by the aeration. The process water temperature was 22 o C and DO generally below 0.5 mg/l. In the PN-SBR reactor, the NO 2 - production rate obtained at was 0.91 g NO 2 - -N/L-reactor/d. The PN effluent contained 1.38 g NO 2 - -N/g NH 4 + -N, which was within 5% of the target oxidation of ammonium to nitrite of 57%, which corresponds with the reaction ratio of anammox. In the second SBR, the anammox process was effectively applied to the PN effluent, attaining a nitrogen conversion rate of 0.36 g N/L-reactor/day (14.3 mg N/g VSS/ h) and anammox stoichiometric ratios of 1.00:1.32:0.15. For the single-tank deammonification, total N removal rates of 0.75 to 1.0 g N/L-reactor/day were obtained for both synthetic and swine wastewater. The stoichiometry of the reaction obtained in the single-tank process was consistent with deammonification process theory combining partial nitritation and anammox: NH 4 + + 0.88 O 2 0.44 N 2 + 0.11 NO 3 - + 1.43 H 2 O + 1.14 H + Compared to conventional nitrification/denitrification, a partial nitritation/anammox pathway (deammonification process) can reduce 58% of the oxygen requirement to biologically eliminate the ammonia from wastewater. Thus, it offers the potential to reduce the cost of treatment of ammonia in livestock wastewaters and other effluents containing high ammonia nitrogen. Key words: ammonia abatement, wastewater treatment, anammox, swine manure. 1. Introduction 1.1. Biological Nitrogen Removal Processes The combined bioprocess of nitrification-denitrification (NDN) has been successfully used for the removal of ammonia nitrogen from animal wastewater through the conversion of ammonium (NH + 4 ) into dinitrogen gas (N 2 ) via nitrate (NO - 3 ) as intermediate (Béline et al., 2004; Vanotti et al. 2009). Removing ammonia from wastewater has many benefits. In a fullscale demonstration of a second-generation environmentally superior technology (EST) for

treatment of swine manure in the USA that used NDN, Vanotti et al. (2009) found that animal health and productivity were enhanced hog sales increased 32,900 kg/cycle- as a result of recycling mostly ammonia-free treated wastewater into the barns. However, the operational cost of NDN was an important consideration in livestock effluents. In the same study, NDN consumed 87% of the total electrical power used by the system, which included solids separation, NDN, and phosphorus removal/disinfection unit processes; furthermore, 80% of NDN requirement was to power the air blower for the nitrification process. They concluded that significant savings in power requirements by the EST system in the future will come from changes in the N treatment, such as the incorporation of anaerobic ammonium oxidation (anammox) that requires about half the aeration required by NDN (Vanotti et al., 2009). Biological N removal processes Traditional Nitrogen Removal Processes (nitrification / denitrification) Deammonification process (partial nitritation / anammox) Q NH 4 + NO 2 - nitrification denitrification MeOH denitrification 4Q nitrification Anammox Uses external organic carbon addition for DN Uses endogenous carbon for DN N 2 No carbon needed FIGURE 1: Biological processes to remove nitrogen from wastewater Deammonification is a completely autotrophic nitrogen removal approach that combines partial nitritation (PN) and anammox (Fig. 1). Thus, it can be a promising approach for the biological removal of ammonia in effluents that are low in BOD and high in ammonia concentration such as effluents from the anaerobic digestion process. Compared to traditional nitrification/denitrification, the new process has several advantages: 1. it reduces 58% of the oxygen requirements; 2. it eliminates carbon needs for denitrification; 3. it removes nitrogen at higher rates. Therefore, deammonification could be key technology for development of more economical and energy efficient ammonia treatment systems in the future. Our objectives were to test the deammonification concept for swine wastewater using two configurations: a two-stage process (PN in one tank and anammox in another), and a one-stage deammonification process (PN and anammox in a single tank). 2. Bacterial cultures 2.1. Nitrifying bacteria For partial nitritation we used a high performance nitrifying sludge (HPNS) developed for treatment of high ammonium concentration and low temperature wastewater (Vanotti et al., 2011a). The HPNS is deposited under the provisions of the Budapest Treaty of the United Nations in the Agricultural Research Service Culture Collection in Peoria, IL, on June 26, 2009, with deposit accession number NRRL B-50298. The HPNS is a composition of

bacteria comprised of 35 strains or populations of isolated bacteria; 26 of the bacteria are affiliated with Proteobacteria, 7 with Bacteroidetes, and 2 with Actinobacteria. The HPNS was derived from swine lagoon nitrifying sludge after prolonged cultivation in a suspended biomass reactor at low temperature (10ºC) at the USDA-ARS laboratory in Florence, SC. The HPNS was maintained in an aeration tank with fine bubble aeration at 10ºC water temperature using the fill-and-draw cultivation method and an inorganic salts medium (Vanotti et al., 2011a). The HPNS had the following characteristics (Vanotti et al., 2011a): (i) a specific nitrification activity of 51.0 mg N per g total suspended solids (TSS) per hour [62.1 mg N per g volatile suspended solids (VSS) per hour] obtained at 30ºC using inorganic salts medium with 300 mg NH 4 + -N/L, biomass concentration 2.01 g VSS/L and process DO 5.0 ±0.6 mg/l; and (ii) a sludge volume index (SVI) of 62 ml/g TSS. 2.1. Anammox bacteria The anammox bacteria used was Candidatus Brocadia caroliniensis deposited under the provisions of the Budapest Treaty in the Agricultural Research Service Culture Collection (NRRL) at Peoria, IL with accession number NRRL B-50286 (Vanotti et al., 2011b). It was maintained at the USDA-ARS laboratory (Florence, SC) in a 10-L jacketed up-flow continuous reactor (120 cm) packed with a biomass carrier to enhance retention of microorganisms (parent reactor). At the time of sludge harvesting, the parent reactor was being fed with synthetic wastewater containing 153 mg NH 4 + -N/L and 153 mg NO 2 - -N/L and operated with a flow rate of 60 L d-1, a N-loading rate (NLR) of 1735 mg N/L/d, and a water temperature of 30ºC. Under these conditions, the N-conversion efficiency (NCE) obtained was 94% and the total N-removal efficiency (NRE) was 85% (Vanotti et al., 2011b). The anammox sludge extracted from the parent reactor had a granular structure and a reddish color. The initial specific conversion rate (measured after harvesting using synthetic wastewater fed into the parent reactor) was 16.7 ±4.4 mg N/g VSS/h (13.3 ±3.9 mg N/g TSS/ h) (Magrí et al., 2012a). Start-up of the A-SBR was done using 700 ml of the anammox sludge containing 7.9 g VSS (9.8 g TSS). 3. Two-tank process 3.1. Partial nitritation reactor Partial nitritation (PN) on swine wastewater was investigated in a Sequencing Batch Reactor as described in detail by Magrí et al. (2012b). Partial nitritation was carried out in a 3.5-L jacketed glass reactor (32ºC) operated under batch mode (Fig. 2, right). Air was supplied through aquarium air pumps and porous stones, which provided fine bubbles. A complete cycle lasted 8 h and consisted of 15 reaction phases of 30 min each and one settling and withdrawing phase. According to the anammox reaction ratio, 1.32 g NO - 2 -N are consumed per g NH + 4 -N removed (Strous et al., 1998). Thus, optimal performance of PN for its coupling with the anammox process requires 57% oxidation of NH + 4 -N to NO - 2 -N: NH + 4 + 0.86 O 2 0.57 NO - 2 + 0.43 NH + 4 + 0.58 H 2 O + 1.14 H + Characteristics of the swine wastewater used (collected after liquid-solid separation) were low content of biodegradable organic matter and high alkalinity-to-ammonium ratio (Magri et al., 2012b). Target oxidation of NH4 + -N to NO - 2 -N) was 57%, which corresponds with the reaction ratio of the anammox as represented in the equation above. This target was successfully achieved just by controlling the inflow rate and corresponding nitrogen loading rate (NLR). An average NLR of 1.47 g NH4 + -N/L/d was applied to the PN-SBR during a period of 70 days. The nitrite production rate obtained was 0.91 g NO - 2 -N/L/d. No nitrate was produced. The PN effluent contained 1.38 g NO - 2 -N/g NH4 + -N, which was within 5% of the target ratio (Magri et al., 2012b).

FIGURE 2: Deammonification treatment using two-stage process: partial nitritation (right) and anammox (left) SBR reactors. 3.2. Anammox reactor The anammox reactor (A-SBR) set-up to perform the anammox process was similar to the PN-SBR reactor, but in a sealed reactor to provide anaerobic conditions (Fig. 2, left). Mixing inside the reactor was achieved through a mechanical stirrer operated at 50 rpm. A 10-L Tedlar bag filled with N 2 was connected with tubing to the top of the reactor and worked as a gas regulation tank to prevent air from entering the reactor during the withdrawal step. Process temperature was also 32 o C. It was operated under 2-h cycles. Continuous feeding and mixing were supplied during the first 90 min, the feeding was stopped and at 110 min in the cycle the mixing was stopped, and the effluent was withdrawn during the last 5 min of the cycle. The anammox process was effectively applied to the PN effluent, attaining a nitrogen conversion rate of 0.36 g N/L/d (Magri et al., 2012b). The specific N conversion rate with PN swine wastewater was estimated to be 14.3 N/g VSS/h. Van Hulle et al. (2010) summarized up to 14 different studies regarding the anammox process which provided an average value for specific N-removal of 17.0 mg N/g VSS/h. Thus, the anammox sludge NRRL B-50286 contained in the A-SBR presented a relatively good activity. Moreover, the specific NCR obtained with PN effluent (14.3 mg N/g VSS/h) was consistent with the specific NCR of the same sludge obtained with inorganic synthetic wastewater (16.7 N/g VSS/h) (Magri et al., 2012a). Stoichiometry ratios obtained (NH 4 + -N consumed, NO 2 - -N consumed, NO 3 - -N produced, and N 2 -N produced) were 1.00:1.32:0.15:1.08, which are very similar to those of 1.00:1.32:0.26:1.02 proposed by Strous et al. (1998) or those of 1.00:1.30:0.18:1.06 measured by Vanotti et al. (2011b) for Brocadia caroliniensis NRRL B-50286 used in this study. Since the specific activity and reaction ratios of the anammox was maintained with PN effluent, we concluded that the overall deammonification process using two tanks was suitable for the treatment of swine wastewater incorporating anammox(magri et al., 2012b). 4. Single-tank process The single reactor was a 5-L aerated vessel operated under continuous flow (Fig. 3). It contained biofilm plastic carriers (1200 m 2 /m 3 ) at 30% v/v packing ratio that were fluidized by the aeration. The reactor was started by adding anammox bacteria Brocadia caroliniensis

(NRRL B-50286) and the biofilm plastic carriers into the reactor. During the first week, the reactor was operated as anammox reactor and fed continuously a solution containing equal amounts of ammonia and nitrite (140 mg/l) (as described by Vanotti et al., 2011) without aeration being applied. Then the nitrifying sludge HPNS (NRRL B-50298) was added at once into to the reactor, which already contained the active anammox bacteria. At that point, continuous aeration was started. During the first five weeks of operation of the single-tank reactor, the influent contained 140 mg/l NH 4 -N and 20 mg/l NO 2 -N. Thereafter, the NO 2 -N was eliminated from the influent leaving only ammonia as the N source. We tested different influent wastewaters (synthetic and swine wastewater) containing increasing ammonia (190 to 440 mg N/L and 750 to 2000 mg/l carbonate alkalinity.the process water temperature was 22 o C. The aeration rate applied varied from 300 to 800 ml/min as the activity increased. DO was generally below 0.5 mg/l. Under these conditions, N removal rates of 0.75 to 1.0 kg N/m 3 /day were obtained for both synthetic and swine wastewater using a single tank. FIGURE 3: Deammonification treatment using single-tank to perform partial nitritation and anammox in a fluidized continuous flow aerated reactor. A. Reactor set-up, B. Addition of HPNS to reactor, C. Reactor detail with biofilm carriers inside. N Concentration (mg/l) 100 90 80 70 60 50 40 30 20 10 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time (hours) Nitrite Nitrate Ammonia Alkalinity FIGURE 4: Experimental data used to estimate single-tank stoichiometry (Table 1). 1000 900 800 700 600 500 400 300 200 100 0 Alkalinity (mg/l) Batch tests were conducted to determine the stoichiometry of the single-tank deammonification process (Fig. 4). The stoichiometry was derived from the concentration profiles of ammonia, nitrite, nitrate, and carbonate alkalinity obtained in the batch (Fig. 4).

The results are summarized in the single-stage equation shown in Table 1 and compared with the theory after combining equiations 1 and 2. The results obtained indicate that the stoichiometry of the reaction in the single-tank process was consistent with the theory of the deammonification process combining partial nitritation and anammox. Compared to separate reactors, a single-tank approach for deammonification treatment was much easier to manage. There was not need to use controls to ensure a balanced partial nitritation, nor to keep anaerobic conditions for anammox. The two bacteria groups (NRRL B-50286 and NRRL B-50298) were able to associate effectively in a single tank providing a streamlined ammonia removal process. Thus, the single tank configuration offers the potential to further reduce the cost of treatment of ammonia in livestock wastewaters containing high ammonia nitrogen. TABLE 1: Comparison of the theoretical stoichiometry of deammonification with experimental reaction data (Figure 4) Deammonification stoichiometry theory: (1) Partial nitritation: 2 NH 4 + + 1.5 O 2 NH 4 + + NO 2 - + H 2 O + 2 H + (2) Anammox: NH 4 + + 1.32 NO 2-1.02 N 2 + 0.26 NO 3 - + 2 H 2 O Single-stage theory (1+2): NH 4 + + 0.85 O 2 0.44 N 2 + 0.11 NO 3 - + 1.43 H 2 O + 1.14 H + Single-stage stoichiometry obtained in this research: NH 4 + + 0.88 O 2 0.44 N 2 + 0.11 NO 3 - + 1.43 H 2 O + 1.14 H + Reference list Béline, F., M.L. Daumer, and F. Guiziou. 2004. Biological aerobic treatment of pig slurry in France: nutrients removal efficiency and separation performances. Trans. ASAE 47:857-864. Magrí, A., M.B. Vanotti, and A.A. Szögi. 2012a. Anammox sludge immobilized in polyvinyl alcohol (PVA) cryogel carriers. Bioresour. Technol. 114:231-240. Magrí, A., M.B. Vanotti, A.A. Szögi, and K.B. Cantrell. 2012b. Partial nitritation of swine wastewater in view of its coupling with the anammox process. Journal of Environmental Quality doi:10.2135/jeq2012.0092. Strous, M., J.J. Heijnen, J.G. Kuenen, and M.S.M. Jetten. 1998. The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms. Appl. Microbiol. Biotechnol. 50:589-596. Van Hulle, S.W.H., H.J.P. Vandeweyer, B.D. Meesschaert, P.A. Vanrolleghem, P. Dejans, and A. Dumoulin. 2010. Engineering aspects and practical application of autotrophic nitrogen removal from nitrogen rich streams. Chem. Eng. J. 162:1-20. Vanotti, M.B., A.A. Szogi, and T.F. Ducey. 2011a. High performance nitrifying sludge for high ammonium concentration and low temperature wastewater treatment. US Patent Publication No. US2011/0000851 A1. US Patent and Trademark Office, Washington, DC. Vanotti, M.B., A.A. Szogi, P.D. Millner, and J.H. Loughrin. 2009. Development of a secondgeneration environmentally superior technology for treatment of swine manure in the USA. Bioresour. Technol. 100:5406-5416. Vanotti, M.B., A.A. Szogi, and M.J. Rothrock. 2011b. Novel anammox bacterium isolate. US Patent Application No. 13/013,874. US Patent and Trademark Office, Washington, DC.