Nitrogen removal from wastewater in an anoxic aerobic biofilm reactor

165 IWA Publishing 2012 Journal of Water Reuse and Desalination 02.3 2012 Nitrogen removal from wastewater in an anoxic aerobic biofilm reactor M. F. Hamoda and R. A. Bin-Fahad ABSTRACT A pilot plant, using a four-compartment reactor packed with Biolace media, was operated in the anoxic/aerobic submerged fixed-film (A/ASFF) and the aerobic (ASFF) modes at loadings 0.03 to 0.3 g BOD. g 1 BVS. d 1, 0.01 to 0.11 g NH 3.g 1 BVS d 1, HRTs 0.7 to 8 h, C/N of 6, and 28 ± 2 W C. The system proved to be very effective in treating municipal wastewater, achieving removals up to 98% for biological oxygen demand (BOD), 75% for chemical oxygen demand (COD) and 97% for ammonia. Performance was not adversely affected by a 10-fold increase in loading rate. Both modes of operation showed high specific nitrification rates up to 96 mg N. g 1 BVS d 1, but the A/ASFF was more stable and efficient at higher loadings. Its anoxic stage removed more than 90 and 60% for BOD and COD, respectively. The A/ASFF reactor also achieved denitrification, which eliminated 3.35 mg BOD (or 6.6 mg COD) versus 1 mg denitrified NO 3 -N, that resulted in higher organic removals. Denitrification rate increased linearly with the TON (total oxidised nitrogen) loading applied, and specific substrate removal reached up to 114 mg TON. g 1 BVS. d 1. Key words fixed-film processes, nitrification denitrification, substrate removal rates, wastewater treatment M. F. Hamoda (corresponding author) Department of Environmental Technology Management, CFW, Kuwait University, PO Box 5969, Safat 13060, Kuwait E-mail: mfhamoda@gmail.com R. A. Bin-Fahad Ministry of Environment and Water Resources, Government of UAE, Abu Dhabi, UAE INTRODUCTION In line with developments in the water reuse sector, the concerned authorities in arid and semi-arid countries have implemented policies calling for construction of new wastewater treatment plants and upgrading existing plants to cope with increasing flows and produce treated effluents suitable for reuse in irrigation. Currently, treated wastewater effluents in the Arabian Gulf countries are reused extensively in landscape and greenery irrigation and, to a lesser extent, in agricultural lands (Hamoda 2004). Biological suspended growth systems, such as the activated sludge process, are commonly used for the secondary treatment of municipal wastewaters (Metcalf & Eddy 2003). However, such systems generally suffer from poor nitrification and excessive sludge production. On the other hand, attached growth (fixed-film) systems are gaining much attention for their merits (Ong et al. 2004; Von Sperling 2007; Love et al. 2010). Attached growth processes are biological systems in which the microorganisms responsible for biodegradation and stabilization of organic matter are attached or fixed to a solid inert media forming a biofilm. Such processes have the advantages of long solids retention, better nitrification, good sludge settleability, low sludge production and stable operation (Lessel 1994; Metcalf & Eddy 2003; Farabegoli et al. 2004; Ryu et al. 2008; Love et al. 2010). A number of innovative processes have been developed, such as the aerated submerged fixed-film (ASFF) process (Hamoda 1989). This process employs a four compartment-in-series reactor equipped with an array of submerged media (fixed ceramic plates) for biomass attachment that is maintained under continuous diffused aeration. Modification of the ASFF process to operate in the anoxic aerobic (A/ASFF) mode could have some advantages based on studies on other biological systems (Hao & Huang 1996). Moreover, combining both attached growth and suspended growth of microorganisms doi: 10.2166/wrd.2012.015

166 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 has become a viable option to upgrade the activated sludge process (Odegaard & Rusten 1990; Su & Ouyang 1996; Hamoda & Al-Sharekh 2000). Performance of such hybrid systems requires further investigation. This study was conducted in order to investigate the effect of hydraulic loading on nitrogen removal in the ASSF hybrid system. A new approach was to alternate the anoxic aerobic mode in an A/ASFF bioreactor in order to promote nitrogen recovery from wastewater for potential water reuse. MATERIALS AND METHODS Description of the pilot plant and experimental set-up The aerated submerged fixed-film (ASFF) bioreactor was used for conducting the pilot-scale experiments. This reactor is made of 6-mm thick plexiglass sheets and is divided into four equal-sized compartments connected in series. Each reactor has a total liquid volume of 115 l. A pilot plant was installed at Al-Awir WWTP in Dubai (Figure 1). The experimental programme involved in-parallel testing of ASFF and A/ASFF reactors, each packed with the Biolace support medium. Biolace is a structured medium of cross-linked textile fibres which is fixed vertically and stretched in a high-grade stainless steel cage. Five sheets of the Biolace (each 390 mm long and 230 mm wide), spaced at 25 mm, were fixed in each cage, occupying approximately 44% of the compartment s volume. The biolace is manufactured by UTS, Germany. It has a specific surface area of 17.5 m 2 /m 3. Each reactor was operated continuously at a preset feed flow rate. Different flow rates were tested in each reactor over a total period of 9 months to obtain HRTs (hydraulic retention times) in the range of 0.7 to 8 hours. Aeration was provided in the second to fourth compartments in the A/ASFF and in all compartments in the ASFF reactors through medium-to-fine, tubularmembrane, air diffusers placed underneath the media and operated at a constant pressure of approximately 200 250 kpa. Operating parameters The operating parameters applied in the ASFF and the A/ASFF bioreactors are summarized in Table 1. The reactors were operated at ambient temperature of 28 ± 2 W C. The mass loadings were based on filtered (soluble) biological oxygen demand (BOD) and chemical oxygen demand (COD) concentrations. There are some differences in the operation of the two bioreactors used in this study in terms of estimating the hydraulic and mass loadings. The ASFF bioreactor is classified as an attached growth process with minimal suspended growth. In contrast, the A/ASFF Figure 1 Schematic diagram of the A/ASFF experimental set-up.

167 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 Table 1 Operating parameters applied to the pilot plant Parameter 8 6 4 2 (a) The ASFF bioreactor Flow rate (m 3 d 1 ) 0.346 0.460 0.691 1.382 BVS (g) 291.55 339.02 410.83 488.03 Hyd. load (m 3 kg 1 BVS d 1 ) 1.185 1.356 1.682 2.833 BOD load (g BOD g 1 BVS d 1 ) 0.154 0.182 0.257 0.382 COD load (g COD g 1 BVS d 1 ) 0.330 0.354 0.463 0.787 NH 3 N load (g NH 3 -N g 1 BVS d 1 ) 0.049 0.056 0.076 0.115 Parameter 8 6 4 2 1.5 1 0.7 (b) The A/ASFF bioreactor Flow rate (m 3 d 1 ) 0.173 0.229 0.389 0.821 0.864 0.950 1.382 NRS (m 3 d 1 ) 0.173 0.229 0.389 0.821 0.864 1.037 1.382 NML (m 3 d 1 ) N/A N/A N/A N/A N/A 0.691 1.210 Total inflow (m 3 d 1 ) 0.346 0.458 0.691 1.382 1.814 2.678 3.974 Recycle ratio 1.0 1.0 0.8 0.7 1.1 1.8 1.9 BVS (g) 694.69 574.46 544.33 580.23 464.02 617.60 601.35 Hyd. load (m 3 kg 1 BVS d 1) 0.497 0.797 1.270 2.382 3.910 4.337 6.609 BOD load (g BOD g 1 BVS d 1 ) 0.033 0.056 0.109 0.218 0.290 0.222 0.290 COD load (g COD g 1 BVS d 1 ) 0.076 0.090 0.186 0.379 0.525 0.369 0.526 NH 3 N load (g NH 3 -N.g 1 BVS d 1 ) 0.009 0.015 0.034 0.064 0.084 0.065 0.099 HRT: hydraulic retention time; NRS: nitrified return sludge; NML: nitrified mixed liquor; BVS: biomass volatile solids. bioreactor involves a hybrid-growth biological process where both attached and suspended growth patterns exist because of sludge recirculation. Therefore, estimation of loading rates based on available surface area for fixed-film growth could not be applied to the A/ASFF bioreactor. It will generally undermine the clear contribution of the suspended growth biomass which constituted about 30% of the total biomass. Loading rates were calculated per total biomass present in the system; that is, per total attached and suspended volatile biomass solids. Mixed liquor volatile suspended solids (MLVSS) and attached volatile solids (AVS) were added up to express the biomass volatile solids (BVS). Hence, hydraulic loading was estimated as m 3 per kg BVS per day and mass loading as g BOD (or g COD) per g BVS per day. Internal recirculation was achieved in the A/ASFF bioreactor by: (1) recycling the thickened nitrified sludge (NRS) into the first compartment (anoxic zone) to support anoxic conditions; and (2) circulating the effluent nitrified mixed liquor (NML) into the first compartment to further support the anoxic process, especially at the shorter hydraulic retention times (HRTs) of 1 h or less. Therefore, such internal loading was accounted for in the overall hydraulic loading applied to the system. Analytical methods Samples were collected daily from each reactor and analysed on the same day of collection. The samples were filtered using Whatman Qualitative Filters size 4. The following parameters were determined on the filtrate: BOD 5, COD, ammonia (NH 3 -N), nitrites (NO 2 -N), nitrates (NO 3 -N), and total oxidized nitrogen (TON). Unfiltered samples

168 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 were used for other measurements such as suspended solids (SS) and volatile suspended solids (VSS). The ph, dissolved oxygen (DO) concentration and temperature were measured on all samples collected. Compartmental attached biofilm mass was determined at the end of each experimental run. Representative compartmental samples of each medium were collected and oven-dried at 105 W C. The volatile (organic) fraction of the attached and suspended solids was determined by further burning the samples at 550 W C. All laboratory analyses were performed according to Standard Methods (1998). RESULTS AND DISCUSSION Major performance parameters such as organics removal, nitrogen transformation and nitrification denitrification rates were used to compare the performance of the ASFF and A/ASFF bioreactors operating at constant, but different hydraulic loading rates. Removal of organics Tables 2 and 3 present percentage BOD and COD removal, respectively, along the compartments in both the A/ASFF and the ASFF bioreactors at different hydraulic retention times. The BOD and COD removal profiles, generated at all HRTs in each of the bioreactors, indicated that the majority of organic removal occurred in the first compartment, in the range of 85 92% and 60 67% for BOD and COD, respectively. No appreciable removal occurred in the remaining compartments as the substrate removal followed a first-order kinetic model (IWA 2006). The overall BOD and COD removal efficiencies obtained in the ASFF and A/ASFF bioreactors at different HRTs are shown in Figure 2. It is evident that the A/ASFF bioreactor performance was slightly better than that of the ASFF bioreactor. The first compartment of the A/ASFF bioreactor had DO concentrations between 0.1 and 0.2 mg l 1. This is due to the long retention time in the settling tank, thus Table 2 Compartmental percentage BOD removal in the A/ASFF and ASFF bioreactors A/ASFF bioreactor % BOD removal ASFF bioreactor % BOD removal Reactor s compartment 8 6 4 2 1.5 1 0.7 8 6 4 2 1st 85.0 88.2 92.5 91.7 90.8 86.9 89.0 77.4 68.2 69.9 71.9 2nd 11.4 9.30 5.03 6.01 6.84 10.3 5.4 15.0 21.5 17.6 7.9 3rd 1.82 0.78 1.29 1.11 0.73 0.66 1.7 5.0 4.0 5.5 11.6 4th 0.55 0.47 0.31 0.1 0.4 0.74 0.4 0.4 0.4 3.0 1.1 Overall 97.5 97.0 98.5 98.0 96.5 95.5 96.4 96.8 93.2 95.5 92.0 Table 3 Compartmental percentage COD removal in the A/ASFF and ASFF bioreactors A/ASFF bioreactor % COD removal ASFF bioreactor % COD removal Reactor s compartment 8 6 4 2 1.5 1 0.7 8 6 4 2 1st 66.5 60.9 70.1 66.8 61.1 65.8 62.0 46.2 37.8 43.3 35.6 2nd 8.95 8.38 1.79 4.75 5.73 8.25 7.1 11.4 14.9 9.5 8.8 3rd 2.34 2.03 0.96 4.70 2.03 1.47 2.0 3.4 9.2 12.7 16.3 4th 1.87 0.33 3.49 1.08 3.09 1.68 1.7 2.3 5.1 1.0 2.0 Overall 78.5 72.1 77.5 78.2 72.5 78.0 74.5 67.1 67.5 67.3 63.0

169 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 Figure 2 Overall BOD and COD removal in the A/ASFF and ASFF bioreactors. return sludge was recycled with low DO. The NML, on the other hand, was recycled with a high DO concentration of approximately 4 5mgl 1, but no drawback on the anoxic process was observed at the short HRTs applied of 1 and 0.7 h. Meanwhile, low DO concentration in the first compartment was a good sign of attaining anoxic conditions. This was the great advantage of implementing the anoxic process where the carbonaceous substrate (BOD and COD) can be utilized as a carbon source for the denitrifying bacteria. This is achieved without any outside energy source (oxygen) and without encountering any operational difficulties such as attached growth bridging (excessive biomass growth). In the second, third and fourth aerated compartments, the DO concentrations increased rapidly to 6 mg l 1. In general, high DO concentrations prevailed at all HRTs regardless of loading. Usually, as was the case with the ASFF bioreactor, longer HRTs resulted in longer contact periods between substrate and microorganisms compared with shorter HRTs, thus allowing better removal efficiencies and exhibition of higher DO concentrations. It is evident that, apart from the advantage of implementing the anoxic stage in improving the carbonaceous substrate removal, having high mixed liquor solids concentration resulting from recirculation of return sludge and additional recycling of mixed liquor solids, had greatly helped to support high removal rates where the process could be classified as hybrid growth (i.e. attached and suspended growth). Analysis of the specific oxygen consumption rate for the aerobic compartments has shown that suspended growth was superior in the removal efficiency especially at high loadings. This in turn had a positive impact in maintaining thinner biofilms that allow for better substrate and oxygen penetration. Nitrogen transformation The biological process of nitrogen transformation is currently the most economical and widely employed means of removing nitrogen from wastewater (Metcalf & Eddy 2003). Nitrogen is primarily removed from wastewater by two biological processes: nitrification (oxidation of ammonium to nitrate nitrogen by aerobic autotrophic bacteria) followed by denitrification (reduction of nitrate to nitrogen gas by facultative heterotrophic bacteria). At normal ph values, ammonia in wastewater is actually in the form of the ammonium cation (NH þ 4 ). Bacteria from the genera Nitrosomonas convert ammonia to nitrite nitrogen (NO 2 ), then bacteria from the genus Nitrobacter further convert nitrite to nitrate (NO 3 ), i.e nitrification is a two-step process. On the other hand, denitrification is classified into heterotrophic and autotrophic denitrification according to electron donor. Heterotrophic denitrification requires an organic carbon source as food for the bacteria to live. Facultative bacteria can get their oxygen by taking dissolved oxygen out of the water and by taking it off the nitrate molecules. Denitrification occurs under anoxic conditions when oxygen levels are depleted (<0.5 mg/l) and nitrates become the primary oxygen source for microorganisms.

170 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 Recent research has demonstrated discrepancies in the ability of suspended growth treatment systems compared with attached growth systems (Metcalf & Eddy 2003). Attached growth treatment systems have shown improved rates and maintenance of ammonia removal over traditional, suspended growth treatment systems (Delatolla et al. 2010). In general nitrogen removal rates are most influenced by temperature, hydraulic retention time (i.e. hydraulic loading rates) and carbon:nitrogen (C:N) ratio (i.e. substrate loading rates) among a number of other factors (Metcalf & Eddy 2003). An understanding of the impact of C:N ratio on nitrogen removal from wastewater is imperative for optimizing the biofilm reactor (Biplop et al. 2011). In this study, fixed organic (BOD or COD) loading was applied to simulate the growth of autotrophs since the autotrophs decrease with the incremental increase of influent COD (Ni et al. 2008). The compartmental percentage ammonia removal and the total oxidized nitrogen (TON) production of the A/ASFF bioreactor were determined as illustrated in Table 4. Generally, the observed compartmental efficiency suggests the presence of active nitrifying microorganisms mainly in the second and third compartments at all loading rates. This allowed the fourth compartment of the A/ASFF bioreactor to contribute to process stabilization and to act as a buffering zone to make up for any fluctuations in performance. Therefore, the A/ASFF bioreactor would offer stable nitrification performance and quickly adapt to fluctuations in the organic and hydraulic loading rates applied to the system. In contrast, nitrification in the ASFF bioreactor was mainly accomplished in the third and fourth compartments. The A/ASFF bioreactor effluent was characterized by <1.0 mg l 1 NH 3 -N, <12.2 mg l 1 NO 3 -N, <16 mg l 1 TON-N, <5mgl 1 BOD and <10 mg l 1 SS which can easily meet the requirements for nonpotable agricultural or industrial reuse applications (Hamoda 2004). An effluent NO 3 -N concentration of <3mgl 1 exists in wastewaters that are fully nitrified and denitrified. An effluent that is fully nitrified but has not been denitrified will generally contain an NO 3 -N concentration of approximately 20 mg l 1 (Metcalf & Eddy 2003). Meanwhile, the ph of the influent during experimentation varied in the range of: 6.9 7.2 in the influent, 7.2 7.5 in the anoxic compartment and 7.3 7.7 in the aerobic compartments. The alkalinity (as CaCO 3 ) averaged 290 ± 35 mg l 1 for the influent, 185 ± 40 mg l 1 for the anoxic compartment and 130 ± 45 mg l 1 for the aerobic compartments since the nitrification reaction consumes alkalinity. Achievement of such good effluent characteristics emphasizes the favourable response of the A/ASFF bioreactor to a wide range of loading rates and the sustainability of performance without loss of nitrogen removal capacity compared with other biofilm systems (Liu et al. 1996; Chowdhury et al. 2010; Biplop et al. 2011). Removal rates Ammonia removal efficiencies were higher in the A/ASFF bioreactor compared with those observed in the ASFF bioreactor especially at shorter HRTs. This is illustrated in Figure 3 which shows percentage ammonia oxidation of up to 97% at HRTs of 6 8 h. Meanwhile, Figure 4 illustrates the mean concentrations of nitrates obtained in all Table 4 Compartmental mean (steady state) percentage NH 3 -N oxidized and TON produced in the A/ASFF bioreactor 2nd compartment 3rd compartment 4th compartment % NH 3 -N removal % TON produced % NH 3 -N removal % TON produced % NH 3 -N removal % TON produced 8 31.40 52.52 17.68 43.31 2.55 2.85 6 33.02 54.93 13.23 33.80 1.56 4.38 4 19.81 55.60 17.98 33.78 0.64 5.33 2 25.14 34.33 26.12 67.17 0.79 11.10 1.5 19.00 26.20 17.00 57.50 2.60 10.60 1.0 20.00 11.60 19.00 66.00 2.20 4.07 0.7 12.00 18.60 13.00 51.50 4.50 14.10

171 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 Figure 3 Mean percentages of overall ammonia oxidation in the ASFF and A/ASFF bioreactors. compartments of the A/ASFF bioreactor at all HRTs. Presence of nitrates in the aerobic compartments (second, third and fourth) confirms the activity of the nitrifying bacteria in these compartments. In contrast, only traces of nitrates were observed in the anoxic first compartment where denitrification takes place. Moreover, this study was conducted at a constant C:N (COD:ammonia) ratio of about 6 in both the ASFF and the A/ASFF systems and showed a maximum ammonia removal efficiency of 97%. Other studies reported in the literature (Ahmed et al. 2007; Biplop et al. 2011) showed that the percentage ammonia removal ranged from about 69 to 91% at C:N ratios of about 20 to 2, respectively. It is clear that the ammonia removal efficiency increases at lower C:N ratios at similar ph values and operating temperatures. Meanwhile, the nitrification rates obtained in this study are comparable to those reported in the literature as displayed in Table 5. One of the benefits gained in adopting the anoxic step in the A/ASFF bioreactor is evident in securing steady ammonia removal at higher hydraulic loading rates than applied in the ASFF bioreactor which gives an edge to the A/ASFF bioreactor as a biological nitrogen removal (BNR) process (Tay et al. 2003; Chowdhury et al. 2010; Biplop et al. 2011). Also, high removals of both the carbonaceous and nitrogenous matter were achieved by the A/ASFF bioreactor under the anoxic conditions as compared with the aerobic systems. Moreover, aeration requirements in the A/ASFF bioreactor are much lower. It can be noticed from Table 6 that a 10- fold increase in the organic loading rate from 0.133 to 1.360 g BOD.g 1 BVS d 1 resulted in a 10-fold increase in the denitrification rate from 0.44 to 4.74 g TON kg 1 BVS h 1. Such rates are higher than those reported by Galvez et al. (2003) for a submerged fixed-film reactor but are comparable to those reported by other researchers for either autotrophic or heterotrophic denitrification as presented in Table 7. The specific substrate (TON) utilization rate increased linearly with the increase in the TON loading rate applied in the first compartment of the A/ASFF bioreactor, in the range of loadings studied. The correlation coefficient obtained was high (r 2 ¼ 0.9923) as shown in Figure 5. This indicates that the system was not under TON limitation. Similarly, the carbonaceous (BOD) substrate utilization rate was not adversely affected despite the increase in mass or hydraulic loading rate applied (i.e. decrease in HRT) and followed a first-order kinetic model (IWA 2006). The BOD consumed per TON removed (g BOD/g TON) was determined. Figure 6 shows that this ratio was increased at shorter HRTs (i.e. higher loading rates) and was associated with higher denitrification rates. The denitrification process was able to eliminate about 3.35 mg BOD (or 6.6 g Figure 4 Mean (steady state) nitrate concentrations in the A/ASFF bioreactor.

172 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 Table 5 Nitrification rates in some aerobic treatment systems System Operating conditions Maximum removal rate Reference Packed bed (BioStyr) Domestic wastewater 0.26 kg N m 3 d 1 at 8 W C Delatolla et al. (2010) COD/N ¼ 4.5 7.5 0.18 kg N m 3 d 1 at 4 W C HRT ¼ 1.0 1.2 h Moving bed bioreactor Domestic wastewater 0.21 kg N m 3 d 1 at 8 W C Delatolla et al. (2010) (MBBR) COD/N ¼ 4.5 7.5 0.14 kg N m 3 d 1 at 4 W C HRT ¼ 1.8 2.0 h Modified Bardenpho process Upflow biological aerated filter (BAF) Aerated submerged fixed film Synthetic wastewater 0.127 kg Nm 3 d 1 at room temperature Park et al. (2010) COD/N ¼ 2 HRT ¼ 8h Synthetic wastewater 0.167 kg N m 3 d 1 at 30 ± 2 W C Biplop et al. (2011) COD/N ¼ 1 Loading 0.184 kg NH 3 -N m 3 d 1 Domestic wastewater 0.096 g N g 1 BVS d 1 at 28 ± 2 W C This study COD/N ¼ 6 0.192 kg N m 3 d 1 at 28 ± 2 W C HRT ¼ 0.7 8h Loading 0.009 0.099 g NH 3 -N g 1 BVS d 1 Table 6 Effect of organic loading rates on denitrification rates in the first compartment of the A/ASFF bioreactor Organic loading rate (OLR) Denitrification rate HRT (h) g BOD g 1 BVS d 1 gcodg 1 BVS d 1 (g TON g 1 BVS d 1 ) (g TON kg 1 BVS h 1 ) 8 0.133 0.306 0.011 0.44 6 0.368 0.647 0.036 1.50 4 0.388 0.664 0.015 0.63 2 1.000 1.737 0.027 1.12 1.5 1.245 2.250 0.045 1.87 1 0.837 1.395 0.037 1.52 0.7 1.360 2.471 0.114 4.74 COD) versus 1 mg denitrified NO 3 -N. This contributed to higher organic removal in the A/ASFF system. The denitrification rate in the A/ASFF system reached up to 114 mg TON g 1 BVS d 1. From the results obtained above, the observed nitrification/denitrification patterns of behaviour are very coherent. There are many operational advantages to be obtained from the application of fixed-film processes such as the A/ASFF process compared with suspended growth processes for nitrification and denitrification because of the retention of high biomass and long mean cell residence time in the A/ASFF process. Reduction of more than 25% in sludge production under anoxic conditions and low aeration rates in the A/ASFF process compared with aerobic conditions in the ASFF process results in reduced operational cost. CONCLUSIONS Based on the experimental results obtained, the principal findings of this study are the following: 1. The ASFF and A/ASFF bioreactors achieved high carbonaceous and nitrogenous substrate removal efficiency of up to 98% and 75% for BOD and COD, respectively, and up to 97% for ammonia. Process performance was not adversely affected by a 10-fold increase in loading rates in the range of 0.5 to 6.6 m 3 kg 1 BVS d 1 for hydraulic loadings and in the range of 0.03 to 0.3 g BOD g 1 BVS d 1 for organic loadings. 2. For nitrification at the shorter HRTs (0.7 h), the A/ASFF bioreactor was superior to the ASFF bioreactor in terms of ammonia removal efficiencies. The denitrification rate was also high in the first compartment of the A/ ASFF bioreactor at the shorter HRTs. A 10-fold increase in the specific denitrification rate from 11 to 114 mg TON g 1 BVS d 1 was associated with a 10-fold increase in specific organic loading rate from 0.133 to 1.360 g BOD g 1 BVS d 1.

173 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 Table 7 Denitrification rates in some anoxic/anaerobic treatment systems System Operating conditions Maximum removal rate Reference Laboratory-scale anoxic vials Synthetic wastewater Autotrophic denitrification using NO 2 as 158.4 mg NO 3 -N g 1 VSS d 1 at 30 W C electron acceptor Batch anaerobic reactors Synthetic wastewater 169.2 mg NO 3 -N g 1 VSS d 1 at Autotrophic denitrification using reduced sulfur 30 ± 1 W C compounds Modified Bardenpho anoxic process Synthetic wastewater Autotrophic denitrification using spent caustic 74 g TN m 3 d 1 at room temperature Anoxic denitrification process Methanol as carbon source 320 mg NO 3 -N g 1 VSS d 1 at Heterotrophic denitrification 25 W C Anoxic denitrification process Domestic wastewater as carbon source 110 mg NO 3 -N g 1 VSS d 1 at Heterotrophic denitrification 27 W C Anoxic submerged fixed film Domestic wastewater as carbon source 114 mg TON g 1 BVS d 1 at 28 ± Heterotrophic denitrification 2 W C Fajardo et al. (2010) Souza et al. (2010) Park et al. (2010) Metcalf & Eddy (2003) Metcalf & Eddy (2003) This study 3. Success of the A/ASFF bioreactor was mainly due to the efficient performance of the anoxic stage incorporated in the first compartment. Removal efficiencies of more than 90% and 60% for BOD and COD, respectively, were achieved in this stage. For treatment of wastewater in hot climates, utilization of an anoxic process provides efficient carbonaceous/nitrogenous substrate removal without additional aeration requirements. REFERENCES Figure 5 Effect of TON loading rate applied on TON substrate utilization rate. Figure 6 Effect of hydraulic retention time on the ratio of BOD consumed to TON removed in the first compartment of the A/ASFF bioreactor. Ahmed, M., Idris, A. & Adam, A. 2007 Combined anaerobicaerobic system for treatment of textile wastewater. J. Eng. Sci. Technol. 2 (1), 55 69. Biplop, P., Fatihah, S., Shahrom, Z. & Ahmed, E. 2011 Nitrogen removal efficiency in an upflow partially packed biological aerated filter (BAF) without backwashing process. J. Wat. Reuse and Desal. 1 (1), 27 35. Chowdhury, N., Zhu, J. & Nakhla, G. 2010 Effect of dynamic loading on biological nutrient removal in a pilot-scale liquidsolid circulating fluidized bed bioreactor. J. Environ. Eng., ASCE 136, 906 913. Delatolla, R., Tufenkji, N., Comeau, Y., Gaadbois, A., Lamarre, D. & Berk, D. 2010 Investigation of laboratory-scale and pilotscale attached growth ammonia removal kinetics at cold temperature and low influent carbon. Wat. Qual. Res. J. Canada 45 (4), 427 436. Fajardo, C., Mosquera-Corral, A., Campos, J. L. & Mendez, R. 2010 Evaluation of NO 2 as electron acceptor for autotrophic denitrification. In: Proceedings IWA World Water Congress,

174 M. F. Hamoda & R. A. Bin-Fahad Nitrogen removal in a biofilm reactor Journal of Water Reuse and Desalination 02.3 2012 19 24 September 2010, Montreal, Canada. IWA Publishing, London. Farabegoli, G., Chiavola, A., Rolle, E. & Stracquadanio, S. 2004 Experimental study on nitrification in a submerged aerated biofilter. Wat. Sci. Tech. 49 (11 12), 107 113. Galvez, J. M., Gomez, M. A., Hontoria, E. & Gonzalez-Lopez, J. 2003 Influence of hydraulic loading and air flow rate on urban wastewater nitrogen removal with a submerged fixed film reactor. J. Hazar. Mater. B 101, 219 229. Hamoda, M. F. 1989 Kinetic analysis of aerated submerged fixedfilm (ASFF) bioreactor. Wat. Res. 23, 1147 1154. Hamoda, M. F. 2004 Water strategies and potential of water reuse in the southern Mediterranean countries. Desalination 165,31 41. Hamoda, M. F. & Al-Sharekh, H. 2000 Performance of a combined biofilm-suspended growth system for wastewater treatment. Wat. Sci. Tech. 41, 167 175. Hao, O. & Huang, J. 1996 Alternating aerobic anoxic process for nitrogen removal: Process evaluation. Water Res. 68, 83 93. IWA Task Group on Biofilm Modeling 2006 Mathematical Modeling of Biofilms. IWA Publishing, London. Lessel, T. 1994 Upgrading and nitrification by submerged biofilm reactors: Experience from a large scale plant. Wat. Sci. Tech. 29, 167 174. Liu, J., Groenestijn, J. & Doddema, H. 1996 Removal of nitrogen and phosphate using a new biofilm-activated sludge system. Wat. Sci. Tech. 34, 315 322. Love, N. G., Daigger, G. T., Leslie Grady Jr, C. P. & Filipe, C. D. M. 2010 Biological Wastewater Treatment. 3rd edition. CRC Press, Boca Raton, FL. Metcalf & Eddy, Inc. 2003 Wastewater Engineering: Treatment, Disposal, Reuse. 4th edition. McGraw-Hill Book Co., New York. Ni, B. J., Yu, H. Q. & Sun, Y. J. 2008 Modelling simultaneous autotrophic and heterotrophic growth in aerobic granules. Wat. Res. 42 (6 7), 1583 1594. Odegaard, H. & Rusten, B. 1990 Upgrading of small municipal wastewater treatment plants with heavy dairy loading by introduction of aerated submerged biological filters. Wat. Sci. Tech. 22, 191 198. Ong, S. L., Liu, Y., Lee, I. Y., Hu, J. Y. & Ng, W. 2004 A novel high capacity biofilm reactor system for treatment of domestic sewage. Wat. Air Soil Pollut. 157 (1 4), 245 256. Park, S., Byun, I., Park, T. & Lee, T. 2010 Effect of influent organics and nitrogen concentration on autotrophic denitrification using spent caustic. In: Proceedings IWA World Water Congress, 19 24 September 2010, Montreal, Canada, IWA Publishing, London. Ryu, H. D., Kim, D., Lim, H. E. & Lee, S. L. 2008 Nitrogen removal from low carbon-nitrogen wastewater in four-stage biological aerated filter system. Process Biochem. 43, 729 735. Souza, T. S. O., Moracs, B. S., Vich, D. V., Garcia, M. L., Varesche, M. B. A. & Foresti, E. 2010 Autotrophic denitrification using sulfide as electron donor: kinetic and microbial characterization. In: Proceedings IWA World Water Congress, 19 24 September, Montreal, Canada, IWA Publishing, London. Standard Methods for the Examination of Water and Wastewater 1998 20th edition. American Public Health Association/ American Water Works Association/Water Environment Federation, Washington, DC. Su, J. L. & Ouyang, C. F. 1996 Nutrient removal using a combined process with activated sludge and fixed biofilm. Wat. Sci. Tech. 34, 477 486. Tay, J. H., Chui, P. C. & Li, H. 2003 Influence of COD:N:P ratio on nitrogen and phosphorus removal in fixed-bed filter. J. Environ. Eng. ASCE 129, 285 290. Von Sperling, M. 2007 Activated Sludge and Aerobic Biofilm Reactors. IWA Publishing, London. First received 11 January 2012; accepted in revised form 2 May 2012