CHEMICAL ENGINEERING TRANSACTIONS VOL. 57, 2017 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš, Laura Piazza, Serafim Bakalis Copyright © 2017, AIDIC Servizi S.r.l. ISBN 978-88-95608- 48-8; ISSN 2283-9216 Microbial Culture Dynamics and Performance of an Anammox Sequencing Batch Reactor System Phumza. V. Tikilili*a, Evans. M. N. Chirwab aSoshanguve East, 172 Tshega street, Pretoria bUniversity of Pretoria, Department of Chemical Engineering, Water Utilization division, University of Pretoria, South campus. Pretoria, South Africa. phumza2@yahoo.com In this study, a 5L sequencing batch reactor was successfully operated for the treatment of simulated wastewater using anammox process. The reactor was inoculated with a pre enriched anammox biomass obtained from a wastewater treatment plant. The reactor was operated in two stages differentiated by different nitrogen loading rate (NLR) i.e. 5 g.N•L -1 •d -1 and 6.3 gN•L-1•d-1. The suspension culture was highly active. The highest removal efficiency for total nitrogen achieved by the reactor was 93% at the highest nitrogen loading rate of 6.3 gN•L -1 •d -1. Furthermore, a modified Stover Kincannon model was used to evaluate the performance of the reactor. A maximum substrate removal rate of 34 gN.L-1.d-1 was predicted. The modified Stover- Kincannon was more suitable for the description of nitrogen removal in the reactor, with the regression coefficient of R2 =0.9739. 1. Introduction Biological treatment processes are recommended for the treatment of wastewater effluents before discharging into receiving water bodies due to lower operation cost and energy saving . Nitrogen removal is one of the most crucial wastewater treatments required due to its contribution to eutrophication of the water bodies (Daims et al., 2006). Various innovative biological nitrogen removal processes such as Single reactor High activity Ammonia Removal Over Nitrite (SHARON), completely autotrophic nitrogen removal over nitrite (Canon) process, De-ammonification and oxygen-limited autotrophic nitrification-denitrification (Oland) process, have been developed (Verstraete and Philips., 1998). However, anaerobic ammonium oxidation (anammox) process is the latest technique for improved nitrogenous compounds removal from wastewater. This process is carried out by autotrophic bacteria of the type Planctomycetes called anammox bacteria which combine ammonium and nitrite to produce nitrogen gas and a small amount of nitrate in anoxic conditions. The anammox process is considered economical and low energy alternative to the conventional biological nitrogen removal, which is generally accomplished through successive aerobic autotrophic nitrification and anoxic heterotrophic denitrification (Cho et al., 2010). The application of anammox process has been developed for wastewater treatment using different reactors. The types of commonly used microbial cultures for the anammox process are granule-based and attached growth cultures. However, a gap still exists on the use of suspension cultures for anammox process. Therefore the aim of this study was to establish the anammox process with a sequencing batch reactor (SBR) inoculated with suspension microbial culture. Emphasis was set on performance of the reactor and substrate removal kinetics. 2. Kinetic model The Stover Kincannon model is a mostly used mathematical model for determining the substrate removal rate as a function of substrate loading rate. The model was initially used for the attached growth biomass performance in a rotating biological contactor (Stover and Kincannon., 1982) using the following equation: DOI: 10.3303/CET1757055 Please cite this article as: Tikilili P., Chirwa E.M.N., 2017, Microbial culture dynamics and performance of an anammox sequencing batch reactor system, Chemical Engineering Transactions, 57, 325-330 DOI: 10.3303/CET1757055 325   A QS K A QS U A SSQ dt dS i B i ei     max (1) Where dS/dt is the substrate removal rate (mg.L-1.d-1), Q is the flow rate (L/d), Si is the influent substrate concentration, Se is the effluent substrate concentration (mg/L), A is the total disc surface area on which biomass concentration is immobilized (m2). KB represents the saturation value constant (g/d.m 2) whereas Umax is the maximum substrate removal rate constant (g.L-1.d-1). The original model was later modified and used to predict the bioreactor performance (Yu et al., 1998). In this approach, the suspended biomass concentration was compared with the attached biomass. When the surface area (A) is replaced by reactor volume (V), the Stover Kincannon model is modified as follows:         V QS K V QS U dt dS i B i max (2) For this equation, units of KB change to g.L -1.d-1 Equation 2 can be linearized as follows   max 1 . 1 max UQS V U K SSQ V dtdS i B ei    (3) If 1/(dS⁄dt) is deduced as V/(Q(Si-Se)), the inverse of the removal rate and is plotted against V/QSi, the inverse of the loading rate, a straight line plot is obtained. From this plot, the slope gives KB/Umax and the intercept of the straight line gives 1/Umax . The substrate balance for the reactor can be expressed as follows: ei QSV dt dS SQ  .. (4) Replacing equation 2 in the above equation gives e i B i i QSV V QS K V QS U SQ          .. max (5) This equation can be solved for effluent substrate concentration by introducing the values of Umax and KB values using the following equation         V S QK SU SS i B i ie . max (6) 326 3. Materials and methods 3.1 Reactor description The SBR with a working volume of 5L was used. The pH was maintained between 7.5 and 8 without specific control. The medium was homogenized by a magnetic stirrer. A set of two peristaltic pumps was used to introduce the feeding solution and to discharge the effluent. Timers controlled the actuations of the pumps and valves and regulated the different periods of the operational cycle. The reactor was flushed continuously Argon to maintain anaerobic conditions. All the tubing was norprene tubes, to prevent the diffusion of oxygen inside the system 3.2 Operational conditions The SBR was operated in 12 h cycles distributed as follows during operation: 600 min of feeding and mixing, 45 min of settling, 15 min of effluent withdrawal. The feeding supplied to the reactor was prepared using the mineral salt medium. The reactor was operated in 2 different stages depending on the nitrogen loading rate. The Hydraulic Retention Time (HRT) was maintained at 0.4 days. 3.3 Culture media and inoculum Synthetic medium was supplemented with ammonium and nitrite (at required concentrations) in the form of (NH4)2SO4 and NaNO2, respectively. The composition of the synthetic medium was (per litre deionized water) 1.25 g of KHCO3, 0.05 g of NaH2PO4, 0.2 g of MgSO4·7H2O, 0.3 g of CaCl2·2H2O, 0.006 g of FeSO4, 0.006 g of EDTA and 1.25 ml trace elements solution ((Jetten et al. , 2005). The trace elements solution contained (per litre deionized water) 0.4 g of ZnSO4·7H2O, 0.04 g of CuSO4.5H2O, 0.1 g of KI, 0.2 g of FeCl3.6H2O, 0.4 g of MnSO4·H2O, 0.2 g of Na2MoO4.H2O. 0.4 of ZnSO4.7H2O, 1 g of NaCl, 0.1 g of CoSO4, 0.1g of CaCl2, 0.01g of AlK(SO4)2.12H2O and 0.05 g of H3BO3. The reactor was inoculated with a suspension culture of anammox biomass. 3.4 Analysis Ammonium, nitrite and nitrate were analysed calorimetrically according to the following methods: •Nitrate analysis – add 10µl saturated sulfumic acid and 40µl reactor effluent together. To the mixture add a total of 0.2 ml reagent containing 5% salicylic acid in 98% sulphuric acid and 2ml 4M NaOH (4ºC). This solution is analysed in a spectrophotometer at 420nm after a 30 minutes reaction. •Ammonium analysis – add 760 µl of a solution containing 0.54% ortho-pthalaldehyde, 0.05% β- mercaptoethanol and 10% ethanol in 400mM potassium phosphate buffer (pH 7.3) to a 40 µl reactor effluent sample. This solution is analysed in a spectrophotometer at 420nm after a 30 minutes reaction. •Nitrite analysis – add 950 µl of a reagent containing 1% sulfanilic acid and 0.05% N- naphthylethylenediamine in 1 M H3PO4 to 50µl of reactor effluent. This is followed by a spectrophotometric analysis at 540nm after 5 minutes reaction. 4. Results and Discussion 4.1 Performance of the reactor The SBR for an anammox process was started-up with nitrogen loading rate (NLR) of 5 g.N•L-1•d-1. The reactor was operated for about 120 days. The reactor was operated in two stages. In the second stage, the initial NLR was increased from 5 g.N•L -1 •d -1 to 6.3 gN•L-1•d-1 by increasing the concentrations of nitrogen compounds. However, only NH4 + concentrations were increased. Nitrite concentrations remained the same as in the first to avoid the inhibitory effect of nitrite in the system. Figure 1 illustrates the performance of the reactor. Over the period of 120 days the nitrogen removal efficiency was about 93% (Table 1) with nitrite almost completely consumed (>97%) (Figure 1). When NLR was increased to 6.3 gN•L -1 •d -1 concentrations of NH4 +were slightly increased in the effluent and later they started to decrease gradually. 327 Time (days) 0 20 40 60 80 100 120 140 N itr og en c on ce nt ra tio n (m g/ L) 0 20 40 60 80 100 EffluentNitrate Inffluent nitrate Effluent Ammonium Influent Ammonium Effluent Nitrite Influent Nitrite Figure 1: Performance of anammox suspension in SBR Influent: NH4 + ∆, NO2 - □, NO3 - ○, Effluent NH4 + ▼, NO2 - ■, NO3 - ● In an anammox process nitrogen gas has to be produced. In order to reach the conclusion that anammox was taking place in the reactor, gas production was also monitored. The production of the gas in the reactor was confirmed by a gradual increase in readings of the gas metre that was used. In addition, calculations of nitrogen balance were made and it was found that 1.22 moles of nitrite were consumed and 0.2 moles of nitrate were produced per mole of ammonium consumed. Though these values are not exactly the same as those of the anammox process proposed by Strous et al., (1998) they are too close and are comparable to the theoretical stoichiometry of anammox. Therefore it can be concluded that anammox process definitely took place in the reactor and was able to remove up to 93% of total nitrogen in the system. Table 1: Nitrogen removal rate and efficiency of the anammox reactor Time (days) Substrate removal rate (gN/L.d) Substrate removal efficiency % NH4-N NO2- N Total nitrogen NH4-N NO2- N Total nitrogen 1 0.5085 0 0.5085 12.2054 0 10.1875 14 0.576 0.498 1.074 13.8234 60.3432 21.5145 19 0.6668 0.7561 1.4229 16.0042 91.6192 28.5056 22 0.447 0.7963 1.2433 10.7281 96.4841 24.9061 26 1.4274 0.8078 2.2352 34.2595 97.8743 44.777 30 2.0898 0.8253 2.9151 50.1583 100 58.3986 35 4.239 0.3637 4.6028 77.4742 44.0715 73.0963 40 5.1125 0.3674 5.4799 93.4378 44.5217 87.0266 100 4.6992 0.8017 5.5009 85.8845 97.138 87.3595 120 5.0568 0.8054 5.8622 92.42 97.5881 93.0974 4.2 Substrate removal kinetics The nitrogen removal kinetics of the anammox reactor was determined using modified Stover-Kincannon model. Figure 2 shows the plot of V/[Q(Si – Se)], the reciprocal of substrate removal rate against V/(QSi), the reciprocal of substrate loading rate. Saturation value constant KB and maximum substrate removal rate Umax were calculated from the slope and intercept of the line plotted in Figure 2 and determined to be 35.8 and 328 34gN.L-1.d-1 respectively. This implied that the anammox reactor had a maximum total nitrogen removal rate of 34 g N•L -1 •d -1. These values were higher than those obtained in other studies (Gong et al. , 2008, Jin and Zheng. , 2009, Ni et al. , 2010). The maximum total nitrogen removal rate from the experimental data was 5.8622 g.N.L-1d-1. This value was much less than that of predicted values and only accounted for 17% of Umax indicating the nitrogen removal full capacity of the reactor has not been reached yet. The plot also gave correlation coefficient of R2=0.7061 thus supporting the appropriateness of the modified Stover-Kincannon model. The effluent nitrogen concentration was calculated using equation 6 after introducing the KB and Umax values in equation 3. The calculated values were compared with experimental data in order to validate the model. Figure 3 illustrates the comparison between predicted and experimental effluent concentrations. The linear relationship was obtained from the comparison. The linear relationship represented a good agreement between experimental and predicted total nitrogen effluent concentrations and gave a high correlation coefficient (R2=0.9739). This indicated that Stover-Kincannon model was suitable for nitrogen removal kinetics in sequencing batch reactor. Figure 2: Nitrogen removal model plot– modified Stover-Kincannon model Figure 3: Validation of the modified Stover-Kincannon model 5. Conclusions This study presents the achievability of successful start-up of the anammox process in a laboratory-scale sequencing batch reactor. The start-up duration of an anammox reactor with pre-enriched non-granular sludge was significantly reduced compared to the ones reported in literature. The study showed that anammox non- granular culture had a great activity. A maximum substrate removal rate of 34 gN.L-1d-1 was predicted. The 329 model simulation matched the experimental data very well, proving the Stover-Kincannon model to be appropriate for nitrogen removal kinetics for the anammox process. Acknowledgments The authors gratefully acknowledge South African Nuclear Human Asset & Research Programme (SANHARP) and Water Research Commission (WRC) for providing funding for this study. Reference Cho, S., Takahashi, Y., Fujii, N., Yamada, Y., Satoh, H. & Okabe, S. 2010, Nitrogen removal performance and microbial community analysis of an anaerobic up-flow granular bed anammox reactor, Chemosphere, 78, 1129-1135. Daims, H., Taylor, M.W. & Wagner, M. 2006, Wastewater treatment: a model system for microbial ecology, Trends in biotechnology, 24, 483-489. Gong, Z., Liu, S., Yang, F., Bao, H. & Furukawa, K. 2008, Characterization of functional microbial community in a membrane-aerated biofilm reactor operated for completely autotrophic nitrogen removal, Bioresource technology, 99, 2749-2756. Jetten, M., Schmid, M., van de Pas‐Schoonen, K., Sinninghe Damsté, J. & Strous, M. 2005, Anammox organisms: enrichment, cultivation, and environmental analysis, Methods in enzymology, 397, 34-57. Jin, R. & Zheng, P. 2009, Kinetics of nitrogen removal in high rate anammox upflow filter, Journal of hazardous materials, 170, 652-656. Ni, S., Lee, P. & Sung, S. 2010, The kinetics of nitrogen removal and biogas production in an anammox non- woven membrane reactor, Bioresource technology, 101, 5767-5773. Stover, E.L. & Kincannon, D.F. 1982, Rotating biological contactor scale-up and design, Proceedings of the 1st International Conference on Fixed Film Biological Processes, 1-21 Strous, M., Heijnen, J., Kuenen, J. & Jetten, M. 1998, The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms, Applied Microbiology and Biotechnology, 50, 589-596. Verstraete, W. & Philips, S. 1998, Nitrification-denitrification processes and technologies in new contexts, Environmental pollution, 102, 717-726. Yu, H., Wilson, F. & Tay, J. 1998, Kinetic analysis of an anaerobic filter treating soybean wastewater, Water research, 32, 3341-3352. 330