Kinetic Model for Biological Nitrogen Removal Using Shortcut

Jun 11, 2010 - Harbin Institute of Technology. , ‡. Northeast Forestry University. , §. Beijing University of Technology. , ∥. Stanford Universit...
7 downloads 0 Views 2MB Size
Environ. Sci. Technol. 2010, 44, 5015–5021

Kinetic Model for Biological Nitrogen Removal Using Shortcut Nitrification-Denitrification Process in Sequencing Batch Reactor D A W E N G A O , * ,†,‡ Y O N G Z H E N P E N G , § A N D WEI-MIN WU| State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, P.R. China, School of Forestry, Northeast Forestry University, Harbin 150040, P.R. China, School of Energy and Environmental Engineering, Beijing University of Technology, Beijing 100022, P.R. China, and Department of Civil and Environmental Engineering, Stanford University, Stanford, California 94305-4020

Received May 26, 2009. Revised manuscript received May 23, 2010. Accepted May 28, 2010.

A kinetic model for shortcut nitrification-denitrification process with sequencing batch reactor (SBR) was developed. To test this model, the kinetic parameters of the model including maximum specific rates and half-maximum rate concentrations for shortcut nitrification and denitrification were estimated from the results obtained from a laboratory-scale SBR fed with a soybean curd processing wastewater (400-800 mg COD L-1, 50-65 mg NH4+-N L-1) at 26 °C. In the nitrification step, two DO levels (0.5 and 3.5 mg L-1) were tested and the predicated nitrification rates under different NH4+-N concentrations using this model fit well with correlation coefficient R ) 0.9902. In the denitrification step, the process of nitrite removal was close to a zero-order reaction if the concentration of electron donor was not so low (COD > 100 mg L-1), and concentrations of nitrite and organic matter (as COD) had limited effect on denitrification rate. The model can be used to predict nitrogen removal performance with different influent NH4+-N and COD concentrations and under various DO concentrations.

1. Introduction Because of environmental pollution problems, including eutrophication, deterioration of water quality, and potential hazards to human health, the removal of various forms of nitrogen (organic nitrogen, ammonium, nitrite, and nitrate) from water and wastewater has become mandatory in recent years (1). A lot of researchers have investigated biological nitrogen removal process, including mechanisms, microbial ecology, and optimization of parameters, as well as kinetic modeling (2-4). Shortcut biological nitrogen removal (SBNR) process has recently received more research attention because of its potential cost-effectiveness (5-11). In this process, ammonium is oxidized to nitrite ammonium* Corresponding author phone: 86-451-82192657; fax: 86-45186282104; e-mail: [email protected]. † Harbin Institute of Technology. ‡ Northeast Forestry University. § Beijing University of Technology. | Stanford University. 10.1021/es100514x

 2010 American Chemical Society

Published on Web 06/11/2010

oxidizing bacteria (AOB) but not further to nitrate by nitriteoxidizing bacteria (NOB) under aerobic condition, and then nitrite is reduced to nitrogen gas by denitrifying bacteria (DNB) under anoxic condition with electron donor. Researchers have studied SBNR with various wastewaters, parameters influencing nitrification/denitrification, and reactor combination, etc. (12, 13). Although two-step nitrification has been modeled previously (14, 15), the model for shortcut nitrification-denitrification kinetics has not been reported. Sequencing batch reactor (SBR) has been widely used in wastewater treatment not only for organic removal but also for nutrient removal including nitrogen and phosphorus (2, 16, 17). SBR system is especially effective for biological nitrogen removal. By controlling the duration of operational cycles, dissolved oxygen concentration, and pH, microbial community dominated with AOB and DNB or/ and anaerobic ammonium-oxidizing bacteria can be selected for SBNR with low or no NOB activity (18-20). In our previous study, SBR has been successfully tested for SBNR process to remove both COD and nitrogen up to 90% and 95% (18, 19). During SBNR process, ammonium is oxidized to nitrite by AOB with less nitrate produced, i.e., low NOB activity (21). Unlike traditional nitrification, shortcut nitrification process does not require further oxidization of nitrite to nitrate. This means that up to 25% of the oxygen consumed for nitrification and 40% of the carbon source for denitrification could be saved using the SBNR process to achieve low COD/N ratio effluents (22). Mathematical models have been developed for conventional nitrification and denitrification for years (23-25). Recently, a model was developed for the membrane-areated biofilm reactor with a community consisting of hereothrophic bacteria, AOB, and NOB (4). These models are useful for specific processes but do not focus on the shortcut nitrification in SBR system. The biological kinetics of denitrification have been studied for many years and several models with varying degrees of complexity have been developed to facilitate the design and operation of biological nitrogen removal plants (26-29). However, these models are unable to predict the concentration of nitrite produced during denitrification since the models are based on denitrification as a one-step process, i.e., nitrate reduction to nitrogen gas (1). Shortcut nitrification-denitrification process does not have the reaction of nitrite to nitrate by control microbial community via operation and is different from the traditional nitrogen removal process from the aspects of denitrification and demand for carbon source. These differences require development of different model and kinetic parameters for the shortcut denitrification. In this study, we developed a kinetic model of SBNR for SBR and experimental approaches to determine the major kinetic parameters including the maximum specific rates and half-maximum rate concentrations for shortcut nitrification and denitrification from experimental results. This model can be used to predict operational performance of shortcut nitrification-denitrification processes in SBR with different NH4+-N and various DO concentrations.

2. Materials and Methods A laboratory-scale SBR was used in this study. The upper part of the reactor was cylindrical, and the bottom was a cone with a height of 70 cm and a diameter of 30 cm. The total effective volume of the reactor was 38 L. The temperature was maintained at 26 °C ( 0.5 in the reactor. Each cycle of the SBR system consisted of filling (0.25 h), aeration (6.5-8.0 h), anoxic mixing (2.0 h), settling (2.0 h), and decanting stages VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5015

(1.0 h). The DO concentration in the SBR was controlled from 0.5 or 3.5 mg L-1, as test required. The wastewater used was from a soybean curd production plant with chemical oxygen demand (COD) about 400-800 mg L-1. Ammonium chloride (NH4Cl) was supplemented to the wastewater to maintain a concentration of ammonium nitrogen (NH4+-N) about 50-65 mg L-1 in the influent of SBR. NaHCO3 was added to the wastewater to get alkalinity of 280-300 mg L-1 (as CaCO3) and pH of 6.5. The seed sludge was obtained from the Harbin City Municipal Wastewater Treatment Plant, Harbin, China. During the operation period, DO, redox potential (ORP), and pH were continuously monitored and recorded. Biomass, or mixed liquor suspended solids (MLSS), in the reactor was maintained at 3.5-4.0 g SS L-1. Wastewater samples were analyzed for COD, NH4+-N, NO2--N, NO3--N, MLSS, and alkalinity according to the Standard Methods (30).

3. Model Development, Results, and Discussion The SBNR model was developed based on two distinct sequential reactions, i.e., partial nitrification of ammonium to nitrite and denitrification of nitrite to nitrogen gas. 3.1. Partial Nitrification Model. Important assumptions of this model are the following: (1) the quantity and quality of influent are constant; (2) the influent filling is instantaneous, which means that substrate is not degraded by microorganisms during influent filling stage; (3) the biomass is in a completely mixed state; (4) during shortcut nitrification process, AOB are not in endogenous respiration phase, i.e., the endogenous-decay coefficient b ) 0; (5) the metabolic characteristics of AOB are stable; and (6) the shortcut nitrification kinetics is on the basis of Monod equation (31). In the shortcut nitrification reaction, ammonium is the substrate. The mass balance for ammonium is as follows: the change of ammonium includes ammonium production from degradation of nitrogen-containing organic matter, ammonium assimilated for microbial cell synthesis, and ammonium converted to nitrite. The mass balance for the change of total ammonium is written as

As shown in eq 3, the molar mass of NH4+-N used for AOB cell synthesis is less than 2% of total NH4+-N during shortcut nitrification process. To simplify the model, the assimilation of NH4+-N is ignored as (dSNH /dt)assimilation ≈ 0 Substituting eqs 2 and 4 into eq 1, we obtain: dSNH /dt ) (dSNH /dt)nitration

(dSNH /dt)ammonified ≈ 0

(2)

During the degradation of organic matter, as mentioned above, the nitrogen source required by heterotrophic bacteria is mainly organic nitrogen but not NH4+-N. Therefore, the NH4+-N used by the heterotrophic bacteria for assimilation is also negligible. The autotrophic bacteria also synthesize cells during the oxidation of ammonium in nitrification process. The reaction equation (32) is AOB

NH+ 4 + 1.382O2 + 1.982HCO3 98 0.982NO2 +

1.036H2O + 1.891H2CO3 + 0.018C5H7O2N 5016

9

(3)

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

(5)

Conventional nitrification is a two-step process (2). In the first step, NH4+ is oxidized to NO2- by AOB as follows (2): + NH+ 4 + 1.5O2 f NO2- + 2H + H2O f

∆G0′ ) -277.68 kJ/mol

(6)

The second stage is the oxidation of NO2- to NO3- by NOB: NO2- + 0.5O2 f NO3- ∆G0′ ) -74.14 kJ/mol

(7)

The growth rate of AOB is faster than that of NOB (2, 4) but in traditional nitrification process, the first step is the rate-limiting process. However, in a shortcut nitrificationdenitrification process, especially using SBR reactor, nitrite can be produced as major nitrification product with limited or no nitrate produced (18, 19). In our experiments described later, by controlling the aeration duration, the concentration of NO3--N was below 1.0 mg L-1. Therefore, the step of NO3--N production can be ignored in the model. We obtain the following equation: (dSNO /dt)nitration ) -(dSNH /dt)nitration

(8)

Based on the Monod equation, the relationship between the specific growth rate of AOB and concentrations of NH4+-N and DO is µAOB ) µA

dSNH /dt ) (dSNH /dt)ammonified + (dSNH /dt)assimilated + (dSNH /dt)nitrated (1) Heterotrophic bacteria preferentially use organic nitrogen for anabolism and catabolism. In most wastewater treatment systems, depending on the composition ratio of carbon, nitrogen, and phosphorus in the wastewater, organic nitrogen is composed of a limited amount of total nitrogen source because most organic nitrogen matter associated with protein in sewage is removed by primary sedimentation and high nitrogen-containing wastewater such as anaerobic digester liquid and landfill leachate contains mainly ammonium. Therefore, to consider treating this type of wastewater, that the mass of NH4+-N produced from degradation of organic matter can be ignored, i.e.,

(4)

So SNH · KA,NH + SNH KA,O + So

(9)

The growth of microorganisms is related to substrate degradation, the relationship between microbial growth and NH4+-N consumption can be expressed as dSNH (consumption of NH4+-N) and dXA (increment of AOB biomass) in the following equation: YA ) -

r dXA /XA dXA µAOB /dt r ) - dS ) - ) -υ )NH dSNH υ qammonia /dt /XA (10)

where r ) (dXA)/(dt), v ) (dSNH)/(dt), qammonia ) (v)/(XA). From eq 10 µAOB ) -YA · qammonia ) -YA ·

dSNH υ ) -YA · XA XAdt

(11)

Substitute eq 11 to 9, then we obtain: dSNH /dt ) -

SNH So 1 ·µ · · ·X YA A KA,NH + SNH KA,O + So A (12)

Assume that the percentage of AOB in biomass of the SBR is constant during the nitrification process during a short time period, so XA

/X ) M

(13)

where M is constant. Therefore, a shortcut nitrification kinetic equation can be obtained after eq 13 is substituted into 12, which is dSNH /dt ) -

SNH SO 1 ·µ · · ·M·X YA A KA,NH + SNH KA,O + SO (14)

In eq 14, the concentration of NH4+-N varies with time during shortcut nitrification process in the SBR, where SNH, SO, and X are the function of time, which is SNH ) SNH(t), So ) So(t), and X ) X(t), and YA, µA, KA,NH, KA,O, and M are constant coefficients under certain operational conditions. These kinetic constants can be determined using experimental data from SBR operation. 3.2. Determination of Shortcut Nitrification Kinetic Parameters. When DO concentration (SO) is controlled at a constant level, the term (SO)/(KA,O + SO) in eq 14 can be regarded as a constant. Because of slow growth of AOB, the amount of AOB in biomass can be assumed to be invariant during the nitrification cycle in SBR. We can use the average concentration of biomass for X in eq 12. Since the amount of AOB does not change over time in the experiment with constant DO, the eq 14 can be simplified as dSNH /dt ) -A

SNH KA,NH + SNH

(15)

in which A)

SO 1 ·µ ·M·X· YA A KA,O + SO

(16)

The reciprocal style of eq 15 is 1 KA,NH + SNH dt )- · dSNH A SNH

(17)

The integral of eq 17 is as follows: KA,NH + SNH

∫ - A1 · S K 1 - ∫ (1 + A S

t)

NH

)

) -

A,NH NH

)

dSNH

dSNH

(18)

1 (S + KA,NHln SNH) + C A NH

where A and C are constant, and t stands for the time of shortcut nitrification. To make the equation reflect the change of NH4+-N with time accurately, in other words, if we want the predicted value to be close to the measured value, the difference between them should be minimal. Therefore, using the principle of least-squares method we can obtain A and C. To obtain these two kinetic parameters, we used the following approaches to design SBR tests in this study: (1) First, a high concentration of DO was maintained in the reactor. Under this condition, SO . KA,O, therefore, in eq 16, SO/(KA,O + SO) ≈ 1 and A ≈ µA/YA · M · X. Because of the value of M · X is assumed to be constant, so µA/YA can be obtained. During the experiment, a constant DO concentration was maintained to avoid the influence of DO fluctuation on nitrification rate. (2) After the test for µA/YA, we resumed DO concentration to a low constant level. Thus, KA,O can be obtained from experimental data with the low DO concentration. 3.3. Shortcut Nitrification Kinetic Parameters from SBR Experiments. During the first shortcut nitrification experiment, DO concentration in the SBR was maintained constantly at 3.5 mg L-1 prior to the test. Then samples were taken to monitor the concentrations of NH4+-N, NO2--N,

and NO3--N over operational cycles. Using the principle of least-squares method and the data obtained (Figure S1 in Supporting Information), the solution was A ) 0.317, C ) 139.55, and KA,NH ) 1.35 mg L-1. The value of KA,NH is close to the empirical value provided by ASM3 (1.0 mg L-1, 20 °C) (33). In the first experiment, DO concentration was maintained at the constant of about 3.5 mg L-1, which was much greater than KA,O, so eq 16 can be expressed as A ≈ (1)/ (YA) · µA · M · X. Since the concentration of AOB biomass was assumed to be unchanged, when value M ≈ 5%, the biomass concentration (X) was measured as 5.106 g SS L-1, and the observed yield coefficient YA ≈ YA,obs ) 0.26 mg SS mg-1 NH4+N. Therefore, it was calculated as µA ) 0.0193 h-1. Compared to the maximum growth rate of nitrate autotrophic bacteria reported by ASM3 (µA ) 0.042 h-1) (33), the value obtained in this study was much lower. This is because the rate of ASM3 is based on the traditional nitrification process with nitrate as an end product, and the rate is the overall rate of both AOB and NOB; but the rate obtained in this study was from shortcut nitrification process with nitrite as major product (>95% of total N) and basically represented the growth of AOB. The result also confirmed that the biomass produced in shortcut nitrification process is less than that of conventional nitrification process. Less sludge produced is an advantage of the shortcut nitrification over traditional nitrification. After µA was obtained, we did the second experiment. The DO concentration was decreased and then maintained constantly at 0.5 mg L-1. The system was stabilized after 2-3 cycles. Subsequently, the concentrations of NH4+-N, NO2-N, and NO3--N were monitored. Similarly, using the data of the change in NH4+-N concentrations vs reaction time (Figure S2 in Supporting Information), the values of A, µA, and KA,O under this DO level were estimated as A ) 0.1962, C ) 343.0173, and KA,O ) 0.307 mg DO L-1. The calculated value of KA,O is smaller than that from ASM3 (KA,O ) 0.5 mg DO L-1, 20 °C), but greater than the half-maximum rate concentration of heterotrophic bacteria (KO ) 0.2 mg DO L-1, 20 °C) (33). The relatively lower KA,O value than that of ASM3 is likely due to the fact that the biomass dominated with AOB in the SBR. The DO concentration had greater influence on NOB than on AOB during a nitrification process. Because most NOB has much higher KA,O than AOB (2, 4) the operation with relatively low DO concentration would be favorable for the selection of AOB. Substitute µA, KA,NH, and KA,O into eq 14, and the final shortcut nitrification dynamic equation is: dSNH /dt ) -0.074 ·

SNH SO · ·M·X 1.35 + SNH 0.31 + SO (19)

Using eq 19, the nitrification rates under different NH4+-N concentrations in the SBR were calculated. The estimated values had good correlation with experimental results (with the correlation coefficient R ) 0.9902) as is shown in Figure 1. This suggests that, using the kinetic parameters obtained from experimental data, the kinetic model eq 19 can be used to simulate the shortcut nitrification process in SBR. Using eq 19, we simulated the shortcut nitrificantion rate vs NH4+-N concentration (Figure 2A) and the change in NH4+-N concentration vs reaction time (Figure 2B) at DO concentrations of 0.5, 1.0, 1.5, 2.0, and 3.5 mg L-1. Higher rate and better NH4+-N removal pattern were observed at relatively higher DO concentrations. However, the simulated data show that the NH4+-N removal extent was almost identical for DO of 1.5 and 3.5 mg L-1 after reaction for 250 min (Figure 2). This suggests that relatively lower DO such as 1.5-2.0 mg L-1 may be favorable for the operation to save aeration energy. Operation at low DO also provides negative selection pressure for the growth of NOB in SBR. VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5017

With methanol (CH3OH) as electron donor and carbon source for cell synthesis, the reaction of shortcut denitrification of nitrite can be described as (2) DNB

NO2 + 0.67CH3OH + 0.53H2CO3 98 0.48N2v + 1.23H2O + HCO3 + 0.04C5H7O2N

FIGURE 1. Correlation between calculated values and test values of shortcut nitrification rate. 3.4. Shortcut Denitrification Model. Assumptions for this model are made as follows: (1) The filling of organic matter as electron donor is instantaneous; (2) the biomass is in a completely mixed state; (3) in denitrification process, denitrifying bacteria (DNB) are not in endogenous respiration period, i.e., the endogenous-decay coefficient b ) 0; (4) the metabolic characteristics of DNB are stable; and (5) the denitrification kinetics is based on the Monod equation (31). In a SBR system, there was no influent and effluent NO2--N at the end of the denitrification process. Therefore, the NO2--N balance is described as follows: dSNO /dt ) (dSNO /dt)assumilated + (dSNO /dt)denitrated (20) Denitrifying bacteria (DNB) are heterotrophic bacteria that use nitrate or nitrite as electron acceptor under anoxic condition (2). They use organic compounds as energy and electron donor source, and generate alkalinity during denitrification. In the case of nitrite, the reaction is NO2- + 4H+ + 3e- f 0.5N2 + 2H2O 0′

∆G ) 277.68 kJ/mol

During SBNR operation, nitrite is the major nitrification product (18, 19). This is also confirmed in this study, i.e., the NO3--N concentration was below 1.0 mg L-1 during the nitrification process. Therefore, only reduction of nitrite to nitrogen gas was considered in the model. As shown in eq 22, when 1 mol nitrite is reduced, only 0.04 mol of nitrogen is used for the synthesis of DNB cells with methanol as electron donor. The amount of NO2--N used for cell synthesis would be also limited when other electron donor sources were used. To simplify the model, the assimilation of denitrification process is ignored, that is (dSNO /dt)assimilated ) 0

(23)

Substitute eq 23 into 20 and we obtain dSNO /dt ) (dSNO /dt)denitrated

(24)

Using Monod equation, the relationship between NO2--N and organic electron donor can be described as dSNO /dt ) -

SNO 1 S ·µ · · ·N·X YN H KNO + SNO KS + S

(25)

where N ) (XH)/(X) is the fraction of DNB in total biomass (as MLSS). As the amount of NO2--N consumed for cell synthesis is ignored, the removal of NO2--N can be considered simply due to denitrification of nitrite (eq 21). The specific denitrification rate is expressed as -

(21)

(22)

SNO S 1 dSNO · · ) qN · XH dt KNO + SNO KS + S

(26)

FIGURE 2. (A) Simulated shortcut nitrification rate vs ammonium concentration. (B) Changes in ammonium concentration during nitrification process in a SBR at five different DO concentrations. 5018

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

FIGURE 3. (A) Simulated shortcut denitrification rate vs NO2--N concentration. (B) Changes in NO2--N concentration during denitrification process in a SBR with different initial COD concentrations (0-50 mg L-1) vs elapsed reaction time.

3.5. Shortcut Denitrification Kinetic Parameters from SBR Experiment. Equation 26 contains two variables, i.e., NO2--N concentration (SNO) and organic matter concentration (S as mg COD L-1), and cannot be directly calculated with definite integral. To obtain these parameters, we used the following experimental approaches: (1) During the denitrification process, first we applied relatively high NO2--N concentrations (>15 mg L-1) and organic matter (>100 mg COD L-1) which were much higher than the respective half-maximum rate concentrations. Therefore, the denitrification reaction was performed as zero order reaction and the rate of denitrification was independent of the concentrations of NO2--N and organic matter. In eq 26, when SNO . KNO, S . KS: SNO S · ≈1 KNO + SNO KS + S

(27)

Substituting eq 27 into eq 26 and rearranging yields: -

1 dSNO · ) qN XH dt

-

(29)

S 1 dSNO · ) qN · XH dt KS + S

(30)

Using the data of COD concentration obtained with a high initial NO2--N concentration (35 mg NO2--N L-1) and qN obtained above, KS was estimated by plotting the integrated equation from eq 30 as KS ) 9.98 mg L-1. This value is slightly higher than that reported by ASM3 (33). This is probably mainly because the organic matter used was from beancurd production wastewater rather than sewage. (3) The third step was to obtain the half-maximum rate concentration of nitrite denitrification (KNO). Similarly, we used a concentration of organic matter which was much higher than KS and a relatively low initial nitrite concentration to make the denitrification rate dependent on NO2--N concentration to determine the KNO. In eq 26, when S . KS: S ≈1 KS + S

(28)

In this study, the temperature of the reactor was maintained at 26 ( 0.5 °C. Using the data from five batch denitrification tests, we plotted data of NO2--N concentration vs time into integrated equation from eq 28 and obtained the maximum specific denitrification rate qN ) 0.0275 h-1. The value is within the range reported in the literature for complete denitrification of nitrate (qN ) 0.0083-0.0333 h-1) (34). This suggests that the maximum rate of shortcut denitrification in this study was similar to that of conventional denitrification at high organic matter concentration. (2) The second step was to change the test condition to determine the half-maximum rate concentration of organic matter (KS). At a NO2--N level greater than its half-maximum rate concentration (KNO) with limited organic matter, the shortcut denitrification rate is dependent on the concentrations of organic matter. Therefore, the KS can be obtained as following. In eq 26, when SNO . KNO: SNO ≈1 KNO + SNO

Substitute 29 into 26, the reaction equation is simplified and the rate is influenced by only organic matter (as COD):

(31)

Substitute 31 into 26, and rearranging yields -

SNO 1 dSNO · ) qN · XH dt KNO + SNO

(32)

With the experimental data of NO2--N at high initial COD concentration (400 mg L-1), KNO was calculated to be 0.72 mg L-1 by plotting NO2--N vs qN using eq 32. Finally, substitute all these calculated parameters, i.e., qN, KS, and KNO, into eq 26, and then the shortcut denitrification kinetic equation for soybean wastewater at 26 ( 0.5 °C was obtained: dSNO /dt ) -0.0275 ·

SNO S · ·X 0.72 + SNO 9.98 + S H

(33)

Equation 33 suggests that the shortcut denitrification equation could be close to a zero order reaction for a wastewater with NO2--N > 3.0 mg L-1 and COD > 20-30 mg L-1 because of low values of both KS and KNO. In this study, both NO2--N and organic matter had limited influence on denitrification rate because the wastewater tested had high VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5019

organic concentration and relatively high nitrite level produced from the nitrification. The changes of NO2--N and COD concentrations in SBR were simulated using eq 33. The results confirm that denitrification occur as a zero order reaction in SBR when the concentration of electron donor is within the range of the soybean curd wastewater (Figure 3A and B). The results indicate that the denitrification reaction rate is high when organic mater is not limited (>100 mg L-1) as shown in Figure 3A. The NO2--N was removed rapidly and the NO2--N concentration reached near zero in a short period (Figure 3B). 3.6. Implementation of the Models and Further Study. In this study, we have developed the model to describe shortcut nitrification-denitrification process in SBR. Using the laboratory-scale SBR treating soybean curd wastewater, the kinetic parameters were estimated and the nitrogen removal performance was simulated. For other wastewater, the kinetic parameters can be estimated using similar experimental approaches and the results can be used for performance analysis and serve as reference for process design. Based on the simulation results with the model, relatively low DO (2.0 mg L-1) could be favorable for the shortcut nitrification process and also save energy in operation. In this study, we tested the model at only one temperature condition (26 °C). For application, the impact of temperature on the system performance for both nitrification and denitrification must be evaluated (2). The influence of pH should also be addressed when designing a SBR system for the removal of COD and nitrogen. More operational conditions including different temperature and pH should be tested prior to applying the model for fullscale SBR design. In addition, this model is developed for the most common wastewaters containing limited or no organic nitrogen matter. For a wastewater containing high concentration of proteins and other organic nitrogen sources, ammonification reaction has to be included in the process and anaerobic or anoxic treatment is needed to degrade these compounds prior to nitrogen removal.

Acknowledgments This research was supported by the Foundation for Author of National Excellent Doctoral Dissertation of the P.R. China (FANEDD; 200544), and the Program for New Century Excellent Talents in the University (NCET; NCET-05-0330), Ministry of Education of the P. R. China, and the Scientific Research Foundation for the Returned Overseas Chinese Scholars, Heilongjiang Province (LC07C06), and the Scientific Research Foundation for the Innovative Talents, Harbin City Government (2007RFLXS002).

Appendix A Nomenclature ASM (ASM3) AOB DO NOB DNB SBNR SBR b KA,NH KA,O KO KNO

5020

9

Activated sludge models reported by IWA Ammonium-oxidizing bacteria Dissolved oxygen Nitrite-oxidizing bacteria Denitrifying bacteria Shortcut biological nitrogen removal Sequencing batch reactor endogenous-decay coefficient (h-1) Half-maximum rate concentration for SNH for autotrophs (NH4-N in mg · L-1) Half-maximum rate concentration for DO for autotrophs (mg O2 L-1) Half-maximum rate concentration for DO for heterotrophs (mg O2 L-1) Half-maximum rate concentration for nitrite (mg NO2--N L-1)

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

KS M qammonia qN SNH SNO S SO XA X XH YA YN

Half-maximum rate concentration for organic matter (mg COD L-1) Percent of AOB in total biomass (%) Specific nitrification rate of ammonium (h-1) Maximum specific denitrification rate of nitrite (h-1) Ammonium concentration (mg NH4+-N L-1) Nitrite concentration (mg NO2--N L-1) Organic matter concentration (as mg COD L-1) DO concentration (mg DO L-1) Active AOB biomass concentration (mg SS L-1) Total biomass concentration in SBR (as mg SS L-1) Active DNB biomass concentration (mg SS L-1) Yield coefficient for AOB (mg SS/mg NH4+-N) Yield coefficient for NOB (mg SS/mg NO2--N)

GREEK LETTERS µH µAOB µA

Maximum specific growth rate of DNB (h-1) Specific growth rate of AOB (h-1) Maximum specific growth rate of AOB (h-1)

Supporting Information Available Time courses of nitrification in SBR. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Marazioti, C.; Kornaros, M.; Lyberatos, G. Kinetic modeling of a mixed culture of Pseudomonas denitrificans and Bacillus subtilis under aerobic and anoxic operating conditions. Water Res. 2003, 37, 1239–1251. (2) Rittmann, B. E.; McCarty, P. L. Environmental Biotechnology: Principles and Applications; McGraw-Hill: Columbus, OH, 2001. (3) Peng, Y. Z.; Wang, X.; Wu, W. M.; Li, J.; Fan, J. Optimisation of anaerobic/anoxic/oxic process to improve performance and reduce operating costs. J. Chem. Technol. Biotechnol. 2006, 81, 1391–1397. (4) Matsumoto, S.; Terada, A.; Tsuneda, S. Modeling of membraneaerated biofilm: effect of C/N ratio, biofilm thikness and surface loading of oxygen on fesibility of simultaneous nitrification and denitrification. Biochem. Eng. J. 2007, 37, 98–107. (5) Gorska, J. S.; Cichon, A.; Miksch, K. Nitrogen removal from wastewater with high ammonia nitrogen concentration via shorter nitrification and denitrification. Water Sci. Technol. 1997, 36, 73–78. (6) Eum, Y.; Choi, E. Optimization of nitrogen removal from piggery waste by nitrite nitrification. Water Sci. Technol. 2002, 45, 89– 96. (7) Han, D. W.; Chang, J. S.; Kim, D. J. Nitrifying microbial community analysis of nitrite accumulating biofilm reactor by fluorescence in situ hybridization. Water Sci. Technol. 2003, 47, 97–104. (8) Ruiz, G.; Jeison, D.; Chamy, R. Nitrification with high nitrite accumulation for the treatment of wastewater with high ammonia concentration. Water Res. 2003, 37, 1371–1377. (9) Kim, D. J.; Chang, J. S.; Lee, D. I.; Han, D. W.; Yoo, I. K.; Cha, G. C. Nitrification of high strength ammonia wastewater and nitrite accumulation characteristics. Water Sci. Technol. 2003, 47, 45–51. (10) Chung, J.; Bae, W.; Lee, Y. W.; Rittmann, B. E. Shortcut biological nitrogen removal in hybrid biofilm/suspended growth reactors. Process Biochem. 2007, 42, 320–328. (11) Park, S.; Bae, W.; Rittmann, B. E.; Kim, S.; Chung, J. Operation of suspended-growth shortcut biological nitrogen removal (SSBNR) based on the minimum/maximum substrate concentration. Water Res. 2010, 44, 1419–1428. (12) Hwang, S.; Jang, K.; Jang, H.; Song, J.; Bae, W. Factors affecting nitrous oxide production: a comparison of biological nitrogen removal processes with partial and complete nitrification. Biodegradation 2006, 17, 19–29. (13) Blackburne, R.; Yuan, Z. G.; Keller, J. Demonstration of nitrogen removal via nitrite in a sequencing batch reactor treating domestic wastewater. Water Res. 2008, 42, 2166–2176. (14) Volcke, E. I. P.; Hellinga, C.; Van Den, Broeck, S.; van Loosdrecht, M. C. M.; Vanrolleghem, P. A. Modelling the SHARON process in view of coupling with Anammox. In Proceedings 1st IFAC

(15)

(16)

(17) (18)

(19)

(20)

(21) (22)

International Scientific and Technical Conference on Technology, Automation and Control of Wastewater and Drinking Water Systems (TiASWiK’02), Gdansk-Sobieszewo, Poland, 2002; pp 65-72. Pambrun, V.; Paul, E.; Sperandio, M. Modeling the partial nitrification in sequencing batch reactor for biomass adapted to high ammonia concentrations. Biotechnol. Bioeng. 2006, 95, 120–131. Zeng, W.; Peng, Y. Z.; Wang, S. Y.; Peng, C. Y. Process control of an alternating aerobic-anoxic sequencing batch reactor for nitrogen removal via nitrite. Chem. Eng. Technol. 2008, 31, 582– 587. Alessandro, S.; Stefano, M. L. Nitrogen removal via nitrite in a sequencing batch reactor treating sanitary landfill leachate. Bioresour. Technol. 2009, 100, 609–614. Wang, S. Y.; Gao, D. W.; Peng, Y. Z.; Wang, P.; Yang, Q. Nitrification-denitrification via nitrite for nitrogen removal from high nitrogen soybean wastewater with on-line fuzzy control. Water Sci. Technol. 2004, 49, 121–127. Yang, Q.; Peng, Y. Z.; Liu, X. H.; Zeng, W.; Mino, T.; Satoh, H. Nitrogen removal via nitrite from municipal wastewater at low temperatures using real-time control to optimize nitrifying communities. Environ. Sci. Technol. 2007, 41, 8159–8164. Joss, A.; Salzgeber, D.; Eugster, J.; Kn ¨ ig, R.; Rottermann, K.; Burger, S.; Fabijan, P.; Leumann, S.; Mohn, J.; Siegrist, H. Fullscale nitrogen removal from digester liquid with partial nitritation and anammox in one SBR. Environ. Sci. Technol. 2009, 43, 5301–5306. Tchobanoglous, G.; Burton, F. L.; Stensel, H. D. Wastewater Engineering Treatment and Reuse, 4th ed; McGraw-Hill: Columbus, OH, 2002. Fux, C.; Boehler, M.; Huber, P.; Brunner, I.; Siegrist, H. Biological treatment of ammonium-rich wastewater by partial nitritation and subsequent anaerobic ammonium oxidation (anammox) in a pilot plant. J. Biotechnol. 2002, 99, 295–306.

(23) Henze, M.; Grady, C. P. L.; Gujer, W.; Marais, G.v.R.; Matsuo, T. A general model for single-sludge wastewater treatment systems. Water Res. 1987, 21, 505–515. (24) Bark, P. L.; Dold, P. L. Nutrient removal in activated sludge system. Water Environ. Res. 1998, 68, 969–984. (25) Grady, C. P. L.; Daigger, G. T.; Lim, H. C. Biological wastewater treatment, 2nd ed.; Marcel Dekker Inc.: New York, 1999. (26) Batchelor, B. Simulation of single-sludge nitrogen removal. J. Environ. Eng. 1983, 109, 1–16. (27) Grady, C. P. L. A general model for single-sludge wastewater treatment systems. Paper presented at a Specialized Seminar on Modelling of Biological Wastewater Treatment, IAWPRC; Copenhagen, Denmark, 1985. (28) Brenner, A.; Argaman, Y. Effect of feed composition, aerobic volume fraction and recycle rate on nitrogen removal in the single-sludge system. Water Res. 1990, 24, 1041–1049. (29) Gujer, W.; Henze, M.; Mino, T.; Matsuo, T.; Wentzel, M. C.; Marais, G.v.R. The activated sludge model No. 2: biological phosphorous removal. Water Sci. Technol. 1995, 31 (2), 1–11. (30) APHA. Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association: Washington, DC, 1998. (31) Monod. The growth of bacteria cultures. Annu. Rev. Microbiol. 1949, 3, 371–394. (32) McCarty, P. L. Stoichiometry of biological reaction. Prog. Water Technol. 1975, 71, 157–172. (33) Gujer, W.; Henze, M.; Mino, T.; van Loosdrecht, M. Activated sludge model No. 3. Water Sci. Technol. 1999, 39, 183–193. (34) Tchobanoglous, G.; Burton, F. L. Wastewater Engineering: Treatment, Disposal, Reuse, 3rd ed.; McGraw Hill: Columbus, OH, 1991.

ES100514X

VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5021