EFFECTS OF RESIDUAL COD ON MICROBIAL GROWTH KINETICS IN A NITRIFYING UCBR

Three reactor set-ups were investigated, namely nitrification UCBR prior to coupling, coupled UCBR-packed bed system for nitrogen removal, and de-coupled recovering UCBR. Batch studies were conducted on the biofilm particles taken from the UCBR. A mathematical model was developed to estimate the growth kinetics of the biofilm autotrophic and heterotrophic bacteria. Heterotrophs was noted to accumulate on the nitrifying biofilm due to the effect of residual COD from the packed bed column during the coupled reactor phase. The growth of heterotrophs had affected the substrate removal rates, biofilm morphology, and growth kinetics of nitrifying bacteria. For example, μm of ammonium oxidizing bacteria decreased from 0.55 to 0.19 d-1, while Ks increased from 1.42 to 3.34 mg N/L. This study demonstrated that μm and Ks of the nitrify bacteria changed with the type of substrate. Hence, growth kinetics would not be the same for nitrifying bacteria that is exposed to residual carbon in a coupled UCBR-packed bed system and a single UCBR that is fed on organic-carbon free substrate.

- 91 - EFFECTS OF RESIDUAL COD ON MICROBIAL GROWTH KINETICS IN A NITRIFYING UCBR L.F. Song, S.L. Ong, J.Y. Hu, K.B. Chia, L.Y. Lee and W.J. Ng Wastewater Biotreatment Group, Department of Civil Engineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260 ABSTRACT Three reactor set-ups were investigated, namely nitrification UCBR prior to coupling, coupled UCBR-packed bed system for nitrogen removal, and de-coupled recovering UCBR. Batch studies were conducted on the biofilm particles taken from the UCBR. A mathematical model was developed to estimate the growth kinetics of the biofilm autotrophic and heterotrophic bacteria. Heterotrophs was noted to accumulate on the nitrifying biofilm due to the effect of residual COD from the packed bed column during the coupled reactor phase. The growth of heterotrophs had affected the substrate removal rates, biofilm morphology, and growth kinetics of nitrifying bacteria. For example, µ m of ammonium oxidizing bacteria decreased from 0.55 to 0.19 d -1 , while K s increased from 1.42 to 3.34 mg N/L. This study demonstrated that µ m and K s of the nitrify bacteria changed with the type of substrate. Hence, growth kinetics would not be the same for nitrifying bacteria that is exposed to residual carbon in a coupled UCBR-packed bed system and a single UCBR that is fed on organic-carbon free substrate. KEYWORDS Biofilm; half saturation constant; heterotroph; maximum specific growth rate; nitrifying bacteria; UCBR INTRODUCTION The Ultra Compact Biofilm Reactor (UCBR) has been shown to be efficient for nitrification (Yu, 1998). The design of a biofilm system is usually governed by the intrinsic process parameters such as growth and decay rates of microorganisms. The reported growth kinetics by various researchers have been different for different bioreactor configurations such as fluidized bed, chemostat and activated sludge (Stevens et al., 1989; Hanaki et al., 1990; Sheintuch et al., 1995). This is because µ m and K s are related to operational conditions, such as sludge age, bacteria genus, degree of turbulence and mixing in the bioreactor. This study aimed to determine the values of growth kinetics, µ m and K s , of a nitrifying UCBR and the effects different system set-ups, through single UCBR, reactors coupled and de-coupled reactors, have on these growth parameters. MATERIALS AND METHODS This study was carried out in three phases. 1. Phase 1 investigated and optimized UCBR nitrification performance at a NH 4 + - N loading rate of 2.8 kg NH 4 + - N/m 3 .d. 2. Phase 2 studied biofilm nitrification activity in the nitrification UCBR column coupled with a denitrifying packed bed column under the influence of residual COD remaining in the effluent of denitrifying column. The NH 4 + -N and COD loading rates for the coupled reactors system were 2.8 kg NH 4 + - N/m 3 .d and 10.2 kg COD/m 3 .d, respectively. - 92 - 3. Phase 3 studied biofilm recovery in a single UCBR after de-coupling the reactors. The NH 4 + - N loading was decreased to 0.5 kg NH 4 + -N/m 3 .d and then gradually increased to 2.8 kg NH 4 + -N/m 3 .d over a period of 42 days. Experimental set-up UCBR The UCBR had two concentric draught tubes which divided the reactor column into two zones, the riser and downcomer. A three-phase separator was mounted at the top of the reactor. Compressed air was introduced at 2.0 cm/s into the riser column via a metallic sparger located at the bottom of the UCBR. The air supplied provided mixing and dissolved oxygen to the UCBR. Packed Bed Column. The packed bed column was a long cylindrical column packed with SIRAN  rings (Schotts  Pte Ltd, Germany) with a dimension of 25mm x 25mm (diameter x height). Backwashing to prevent clogging was performed once every two days using an industrial grade N 2 gas. Coupled UCBR- Packed Bed System. During the coupled reactor phase, a recycle ratio of 4.0 was used. This recirculation ratio was chosen to avoid short circuit in the flow of the wastewater. Recirculation between the two columns was achieved using a peristaltic pump to connect the flow from the bottom of the packed bed column to the bottom of the UCBR column. The geometry of the UCBR and packed bed column are given in Table 1. Table 1. Geometry of UCBR and Packed Bed column Parameters UCBR Packed bed Total reactor volume (L) 4.35 3.30 Reactor height (m) 0.8 0.7 Riser diameter (mm) 50 - Downcomer diameter (mm) 80 - Packed bed diameter (mm) - 80 Batch Test. Batch tests were conducted to obtain the specific COD and NH 4 + -N removal rates. Each batch test was conducted in a 1 L beaker. Compressed air was supplied through a diffuser and the reactor was adequately mixed using a floating magnetic stirrer. Dissolved oxygen was maintained at above 4.0 mg/L throughout the experiment. pH of the reactor was maintained at 7.5 + 0.3 with manual addition of acid (HCl) and base (NaHCO 3 ). Approximately 15 ml of biofilm particles was taken from the UCBR to inoculate 1 L of synthetic feed for each batch test. The synthetic feed consisted of 382.0 mg/L NH 4 Cl, 294.0 mg/L CH 3 COONa, 600.0 mg/L NaHCO 3 , 22.5 mg/L K 2 HPO 4 and 0.2 ml/L trace elements. Each liter of trace elements solution contains 10 g CaCl 2 ⋅H 2 O, 8 g FeCl 3 ⋅6H 2 O, 5 g MgSO 4 ⋅7H 2 O, 2 g CoCl 2 ⋅6H 2 O, 2 g Thiamine-HCl, 1 g NaSiO 3 ⋅9H 2 O, 550 mg Al 2 (SO 4 ) 3 ⋅16H 2 O, 50 mg MnCl 2 ⋅2H 2 O, 1 mg (NH 4 ) 6 Mo 7 O 24 ⋅6H 2 O, 1 mg CuSO 4 ⋅5H 2 O, 1 mg ZnSO 4 ⋅7H 2 O and 1 mg H 3 BO 4 . Samples were collected at 20 to 30 minutes intervals. All the samples were filtered through 0.45 µm pore size filter papers. The filtrate was tested for NH 4 + -N, NO 2 - -N, NO 3 - -N and COD in accordance with Standard Methods for Water and Wastewater (APHA, 1995). DO and pH were recorded at each sample collection interval. MLSS concentration, particle size and biofilm density were determined at the end of each experiment. The biofilm particle size was determined using an Image Analyzing System (Image- Pro Plus version 3.0 for Windows from Media Cybernatics, U.S.A). The batch tests were terminated when ammonium has been completely converted to nitrate. - 93 - Modeling with AQUASIM. AQUASIM 2.0 (Reichert, 1998) was used to estimate the kinetic parameters for autotrophs and heterotrophs as shown in Table 2. Their values would be based on data obtained from the batch tests. Table 2. Kinetic coefficients obtained with AQUASIM Autotrophs Maximum specific growth rate (d -1 ) Mass Fraction Half saturation constant (mg/l) Ammonium Oxidizers µ m,NH4 K S,NH4 Nitrite Oxidizers µ m,NO2 n A (X A = n A X) K S,NO2 Heterotrophs µ m,H n H (X H = n H X) K S,H A simple mathematic model based on Monod’s kinetics for substrate ultilization within the batch reactor was derived and used to estimate values of the parameters shown in Table 3. The mass fraction parameters, n A and n H have been introduced to differentiate the mixed-population biofilm into functions of ammonium oxidizing and COD oxidizing fractions. Convergence tolerance was set to 0.0005 and the secant minimization algorithm was selected as the optimization routine for parameter estimation. Table 3. Process matrix for parameter estimation Stoichiometric coefficient Bacteria Type Substrate utilization rate S COD S NH4-N S NO2-N S NO3-N Heterotrophs         +         +         = DOK DO SK S Y Xn dt dS HOCODHS COD H HHm COD ,, , µ -1 Nitrosomonas         +         +         = −− −− DOK DO SK S Y Xn dt dS AONNHNNHS NNH A ANHm NNH ,44, 4 4, 4 µ -1 +1 Nitrobacter         +         +         = −− −− DOK DO SK S Y Xn dt dS AONNONNOS NNO A ANOm NNO ,22, 2 2, 2 µ -1 +1 RESULTS & DISCUSSION In Phase 1, nitrification efficiency in the UCBR was virtually 100%. During Phase 2, 80% nitrogen removal efficiency was achieved initially. Over a coupled reactors period of 16 days, the nitrogen removal efficiency decreased to 55%. The corresponding conversion rate of NH 4 + -N to NO 3 - -N deteriorated from 100% to 55% over the same period. The deterioration in the specific ammonium oxidation rate was due to the rapid growth of heterotrophs over the slower growing nitrifiers. The decrease in nitrification activities could be due to 3 reasons: 1. Competition between heterotrophs and nitrifiers for common substrate such as dissolved oxygen and ammonium. 2. Increase in diffusional resistance of substrate into the biofilm due to the growth of the outer heterotrophic layer over the slower growing nitrifers. During Phase 2, the diameter of biofilm particles increased from around 600 to 850 µm. At the same time, density of particle decreased from 153 to 97 g/L. Heterotrophic bacteria growth had contributed to a less dense biofilm during Phase 2. According to Tijhuis et al. (1995), biofilm density of the heterotrophs was about 7 times lower than the biofilm formed by the predominantly nitrifying bacteria. - 94 - 3. Reduction in gas-liquid mass transfer due to decreased turbulence in the UCBR caused by a higher solids holdup. This reduction in turbulence could have reduced the oxygen transfer necessary for nitrification. As shown in Fig. 1, the specific activities of heterotrophs increased from 0.29 g COD/g VSS.d to 0.68 gCOD/gVSS.d while the ammonia oxidation specific activity decreased from 0.33 g NH 4 + -N/g VSS.d to 0.04 gNH 4 + -N/gVSS.d during transition from Phase 1 to Phase 2. Tijhuis et al. (1995) reported that biofilm specific ammonium oxidation activity was approximately 1.4 g NH 4 + -N /g VSS.d in a Biofilm Airlift Suspension Reactor (BAS). In comparison, the value obtained in this study was rather low. However, it must be noted that the value reported in this study was an average specific rate for a batch system. The maximum value of specific ammonium oxidation rate obtained in the batch test during Phase 1 was approximately 0.97 g NH 4 + -N/g VSS.d. This value was closer to the value reported by Tijhuis et al. (1995). The different operating conditions between the two studies could have led to the difference. The NH 4 + -N loading rate used in Tijhuis et al. (1995) study was 5 kg N/m 3 .d which is higher than the 2.8 kg N/m 3 .d used in this study. As specific substrate utilization rate will increase with substrate loading, one would therefore expect that a higher NH 4 + -N loading rate would lead a higher specific ammonium oxidation rate. Fig. 1. Specific activity vs time The microbial growth kinetics was obtained in this study by fitting the experimental batch test results to the mathematical model. A comparison between the nitrifying and heterotrophs growth kinetics obtained in this study with those reported in literature is summarized in Table 4. Table 4. Kinetic coefficients for nitrifiers Nitrosomonas Nitrobacter Heterotrophs Type of reactor µ m,NH4 (1/d) K s,NH4 (mgN/l) µ m,NO2 (1/d) K s,NO2 (mgN/l) µ m,H (1/d) K s,H (mgCOD/l) UCBR 1 Phase 1 0.55 + 0.08 1.42 + 0.09 0.03 + 0.01 3.68 + 0.32 0.95 + 0.04 4.45 + 0.17 Phase 2 0.19 + 0.12 3.34 + 2.05 0.06 + 0.04 4.14 + 0.29 1.26 + 0.13 5.18 + 0.30 Phase 3 0.58 + 0.25 1.57 + 0.22 0.05 + 0.08 3.58 + 0.13 1.04 + 0.38 4.40 + 0.32 Chemostat 2 0.66 1.00 0.58 0.35 - - Biofilm tube reactor 3 0.66 0.50 0.14 0.50 5.50 10.00 Activated sludge process 4 0.70 0.60 1.00 1.40 5.00 40.00 Biofilm Air-lift Suspension reactor 5 - - - - 9.60 1.00 1 This study 2 Hanaki et al., (1990) 3 Harald and Dietmar (1997) 4 Metcalf and Eddy (1991) 5 Tijhuis et al. (1995) 0.00 0.20 0.40 0.60 0.80 1.00 0 20406080 Day of opera tion g/g.d Specific NH4-N rate Specific COD rate Phase 1 Phase 2 Phase 3 - 95 - A sensitivity analysis of the model parameters using the absolute-relative sensitivity function showed that maximum specific growth rate, µ m , and mass fraction, n, had high sensitivity, while half saturation concentration, K s , had a low sensitivity. The purpose of a sensitivity analysis is to check if the model parameters can be adequately determined with the aid of the available data and to estimate the uncertainty of the parameter estimated. The µ m obtained in this study was lower for both heterotrophs and nitrifiers as compared to the activated sludge process (Metcalf & Eddy, 1991). This could be due to a lower specific surface loading rate in the UCBR. The K S on the other hand was higher, which could be due to a higher diffusional resistance within the immobilized biofilms as compared to the suspended flocs in an activated sludge process. As noted from Fig. 2 and Fig. 3, the increase in µ m,H due to the influx of COD into the nitrification UCBR was rapid (from 0.99 to 1.36 d -1 in 3 days), whereas the reduction in µ m,NH4 was relatively gentle (from 0.55 to 0.10 d -1 in 21 days). On the other hand, the change in µ m,NO2 due to the coupled reactors arrangement was less compared to µ m,NH4 . It was also noted that in the recovery phase, µ m,H reached its original values faster than µ m,NH4 . As shown in Fig. 4 and Fig. 5, K S for both heterotrophs and autotrophs increased due to the coupled reactors arrangement. K s for heterotrophs increased from 4.8 to 5.3 mgCOD/L over a period of 13 days, while K S for ammonium oxidizers and nitrite oxidizers increased from 1.42 to 3.34 mgN/L and from 3.68 to 4.41 mgN/L, respectively. This could be due to the growth of the outer heterotrophic layer leading to a higher diffusional resistance of substrate to the microorganisms in the inner layer of the biofilm. 0 0.5 1 1.5 2 0 20406080 Days of Operation Maximum specific growth rate (1/d) 0 0.2 0.4 0.6 0.8 0 20406080 Days of Operation Maximum specific growth rate (1/d) Nitrosomonas Nitrobacter Fig 3. Maximum specific growth of nitrifiers. 0.00 1.00 2.00 3.00 4.00 5.00 0 10203040506070 Days of operation Half saturation constant (mgN/l) nitrosomonas nitrobacter 0 2 4 6 0 20406080 Days of Operation Half saturation constant (mgCOD/l) Fig 4. Half saturation concentration Ks of nitrifiers . Fig 5. Half saturation concentration Ks of heterotrophs . Mass fraction of the microorganism also indicated that percentage of heterotrophs increased, while nitrifying bacteria population decreased during coupled reactor period (Fig. 6). The mass fraction, n H , for heterotrophic bacteria indicated the presence of heterotrophic bacteria in the biofilm even in the absence of COD in Phases 1 and 3. This observation is similar to the findings of Rittmann et.al. (1994). After the reactors have been de- coupled in Phase 3, the fraction of nitrifiers was restored to its Phase 1 values within a few days. However performance of the reactor and specific removal rates were still below the values obtained in Phase 1. During Phase 1, the average particle diameter was around 600 µm, while experimental results showed that the diameter of particles in Phase 3 was smaller, approximately 400µm. This suggested that sloughing of outer biofilm layers Phase 1 Phase 2 Phase 3 Fig 2. Maximum specific growth of heterotrophs. Phase 1 Phase 2 Phase 3 - 96 - may have taken place after de-coupling. This could have resulted in nitrifying bacteria in inner layers being sheared off during the sloughing process and washed out from the reactor. 0 0.2 0.4 0.6 0.8 1 5 1215192226334047576269 Days of operation Fraction by Mass Nitrifiers Heterotrophs CONCLUSIONS This study showed that the growth kinetics of nitrifiers, µ m and K s , changed in accordance to the prevailing type of substrate and other operational parameters, such as system configuration. Due to the effect of residual COD, µ m,NH4 decreased from 0.55 to 0.19 d -1 over a period of 21 days, whereas K S,NH4 increased from 1.42 to 3.34 mgN/L over a period of 13 days. Heterotrophic activity had also increased due to residual COD. µ m,H increased from 0.99 to 1.36 d -1 in 3 days, while K S,H increased slightly from 4.8 to 5.3 mgCOD/L . Experimental results showed that the kinetic parameters, µ m and K S , for UCBR nitrification would differ between a coupled UCBR- packed bed system and a single UCBR. Hence appropriate choice of kinetic values would have to be made when applying kinetic model for reactor design. REFERENCES 1. Hanaki K., Wantawin C., and Ohgaki S. (1990) Effects of the activity of heterotrophs on nitrification in suspended growth reactor. Wat. Res., 24 (3), 289-296. 2. Harald H. and Dietmar C.H. (1997) Growth and decay in an auto/heterotrophic biofilm. Wat. Res., 31(9), 2243-2252. 3. Metcalf and Eddy Inc., (1991). Wastewater engineering: treatment, disposal, reuse. 3 rd edn, McGraw- Hill Inc., USA. 4. Reichert P. (1998). Aquasim 2.0 User Manual . Swiss Federal Institute for Environmental, Science and Technology (EAWAG), CH-8600, Diibendorf, Switzerland. 5. Rittmann B. E., Regan J.M. and Stahl D. A. (1994). Nitrification as a source of soluble organic substrate in biological treatment. Wat. Sci. Tech., 30 (6),1-8. 6. Sheintuch M., Tartakovsky B., Narkis N. and Rebhun M. (1995). Substrate inhibition and multiple states in continuous nitrification process. Wat. Res., 29(3), 953-963. 7. Standard Methods for the Examination of Water and Wastewater (1995). 19th ed., American Public Health Association/ American Water Works Association/ Water Environment Federation, Washington DC, USA. 8. Stevens D.K., Berthouex P.M. and Chapman T.W. (1989) Dynamic model of nitrification in fluidized bed. J. Envir. Eng., 32 (1), 71-80. 9. Tijhuis L., Huisman J.L., Hekkelman H.D., van Loosdrecht M.C.M. and Heijnen J.J. (1995). Formation of Nitrifying Biofilm on small suspended particles in airlift reactors. Biotechnol. Bioeng., 47, 585-595. 10. Yu J. (1998). Study of ultra-compact biofilm reactors for nitrogen removal. MEng thesis, Department of Civil Engineering, National University of Singapore. Fig 6. Fraction of auto/heterotrophs by mass P1 P2 P3 . 4.45 + 0 .17 Phase 2 0 .19 + 0 .12 3.34 + 2.05 0.06 + 0.04 4 .14 + 0.29 1. 26 + 0 .13 5 .18 + 0.30 Phase 3 0.58 + 0.25 1. 57 + 0.22 0.05 + 0.08 3.58 + 0 .13 1. 04 +. 9.60 1. 00 1 This study 2 Hanaki et al., (19 90) 3 Harald and Dietmar (19 97) 4 Metcalf and Eddy (19 91) 5 Tijhuis et al. (19 95) 0.00 0.20 0.40 0.60 0.80 1. 00