Effect of anaerobic reaction time on denitrifying phosphorus removal and N2O production

Bioresource Technology 102 (2011) 5674–5684

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Effect of anaerobic reaction time on denitrifying phosphorus removal and N2O production Yayi Wang ⇑, Junjun Geng, Zhongjia Ren, Weitao He, Meiyan Xing, Min Wu, Shaowei Chen State Key Laboratory of Pollution Control and Resources Reuse, School of Environmental Science and Engineering, Tongji University, Siping Road, Shanghai 200092, PR China

a r t i c l e

i n f o

Article history: Received 2 January 2011 Received in revised form 18 February 2011 Accepted 18 February 2011 Available online 22 February 2011 Keywords: Nitrous oxide (N2O) Denitrifying phosphorus removal Anaerobic reaction time Poly-b-hydroxyalkanoates Microbial activity

a b s t r a c t Nitrous oxide (N2O) is a highly potent greenhouse gas; however, the characteristics of N2O production during denitrification using poly-b-hydroxyalkanoates (PHA) as a carbon source are not well understood. In this study, effects of anaerobic reaction time (AnRT) on PHA formation, denitrifying phosphorus removal and N2O production were investigated using a laboratory-scale anaerobic/anoxic/oxic sequencing batch reactor (An/A/O SBR). The results showed that operation of the An/A/O SBR for 0.78 SRT (47 cycles) after the AnRT was shortened from 90 min to 60 min resulted in anaerobically synthesized PHA improving by 1.8 times. This improvement was accompanied by increased phosphorus removal efficiency and denitrification. Accordingly, the N2O–N production was reduced by 6.7 times. Parallel batch experiments were also conducted with AnRTs of 60, 90 and 120 min. All results indicated that in addition to the amount of anaerobically synthesized PHA, the kinetics of PHA degradation also regulated denitrifying phosphorus removal and N2O production. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Denitrifying phosphorus removal, which can achieve simultaneous denitrification and P removal, is a novel biological nutrient removal technology. Denitrifying phosphorus removal occurs due to the capacity of denitrifying phosphorus accumulating organisms (DPAOs) to use nitrate and/or nitrite as an electron acceptor for P removal instead of oxygen (Kuba et al., 1996a). When compared with conventional enhanced biological phosphorus removal (EBPR) processes, the denitrifying phosphorus removal process can reduce the need for aeration by 30%, as well as reduce sludge production and the demand for carbon sources by 50% (Kuba et al., 1996a). N2O is a significant greenhouse gas with an approximately 300-fold greater warming potential than CO2 (IPCC, 2001). Numerous studies have reported that both nitrification and denitrification could lead to N2O production in wastewater treatment systems (Colliver and Stephenson, 2000; Kampschreur et al., 2008). Recently, Kampschreur et al. (2009) reviewed 28 studies relevant to N2O emissions from full-scale and laboratory-scale wastewater systems. They found that N2O emissions reached 0–14.6% of the nitrogen load in full-scale wastewater systems, whereas N2O emissions varied from 0% to 95% of the nitrogen load in laboratory-scale wastewater systems. Moreover, several studies revealed that N2O rather than N2 was the major denitrification end-product in simultaneous denitrification and phosphorus removal systems (Zeng et al., 2003b; Meyer ⇑ Corresponding author. Tel./fax: +86 21 65984275. E-mail addresses: [email protected], [email protected] (Y. Wang). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.02.080

et al., 2005; Lemaire et al., 2006). If so, this reduces the potential advantages of denitrifying phosphorus removal technology (Zhou et al., 2008a,b). Denitrification for denitrifying phosphorus removal is conducted using poly-b-hydroxyalkanoates (PHA) as a carbon source. Poly-b-hydroxybutyrate (PHB) is the main part of PHA and is the rate-limiting substance in organisms that grow on PHB (Beun et al., 2002). The slower nature of PHB degradation can produce competition for electrons between denitrifying enzymes, resulting in a higher NO reduction rate compared to the N2O reduction rate (Kampschreur et al., 2009). Indeed, Schalk-Otte et al. (2000) reported that N2O accumulated as soon as the stored PHB became the growth substrate due to a shortage of carbon. Moreover, kinetic studies of the PHA degradation indicated that PHB degradation follows first-order kinetics (Beun et al., 2002; Third et al., 2003), and that a decrease in the rate of PHB availability can result in limited reducing power for denitrification, leading to increased N2O production (Third et al., 2003). These results indicate that PHA is associated with denitrification and N2O production. However, most studies of N2O production during denitrifying phosphorus removal have focused on the effects of nitrite or free nitrous acid (FNA) (Zhou et al., 2008a,b), and few have considered the impact of intracellular PHA. Therefore, the mechanism for N2O production in the denitrifying phosphorus removal process may not have been described properly to date, specifically at the intracellular compound level. The anaerobic reaction time (AnRT) is of primary importance for designing EBPR processes. The AnRT guides the process design required to achieve the desired treatment efficiency. Specifically, the

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

AnRT has a significant effect on PHA synthesis. A shorter AnRT may result in insufficient PHA synthesis, which finally leads to low P removal efficiency. Conversely, a longer AnRT may cause microorganisms to suffer a longer famine period, which also negatively affects PHA synthesis and the N and P removal efficiencies. More recently, Freitas et al. (2009) examined the effect of cycle length on biological nutrient removal in a sequencing batch reactor (SBR) and found that a short cycle was more favorable for nutrient removal. However, it is still unclear how and why AnRT or cycle length affects P removal. Particularly, the relationship between AnRT and PHA synthesis, which is important for understanding N&P removal as well as the N2O production mechanisms involved in DPAOs, are not fully understood. Therefore, in this study, both a continuous experiment and batch tests were conducted to investigate the effect of AnRT on anaerobic PHA syntheses during denitrifying phosphorus removal and to determine if these effects were correlated with N2O production and denitrifying phosphorus removal efficiency.

2. Methods 2.1. Process setup and operation A sealed laboratory-scale SBR with a working volume of 7.5 L and overhead space of 1 L was operated at 22 ± 2 °C. The SBR was fed with synthetic wastewater and operated under alternating anaerobic–anoxic–aerobic conditions. At the end of the cycle, 125 mL of sludge was removed to achieve a solids retention time (SRT) of approximately 20 days and a mixed liquor suspended solid (MLSS) level of 3.7–4.7 g/L. The SBR was inoculated with activated sludge from the Quyang Sewage Treatment Plant (Shanghai, China) and a laboratory-scale plug-flow anaerobic–aerobic (A/O) reactor in which polyphosphate-accumulating organisms (PAOs) were successfully enriched (Wang et al., 2010). After the SBR was operated over 60 days, DPAOs were enriched as reflected by the constant P removals and biomass concentrations. Thereafter, the experiments reported below were conducted. The SBR operation consisted of the following four phases: Phase I (cycles 1–126): the SBR was operated with three 8 hcycles per day. Specifically, each cycle consisted of 15 min feeding, 90 min anaerobic reaction, 210 min anoxic reaction, 30 min aerobic reaction followed by 45 min settling, 15 min decanting and 75 min idle time. During the first 15 min feeding period, 5.5 L of synthetic wastewater was pumped into the reactor, while 100 mL KNO3 solution was pulse added into the reactor at the end of the anaerobic period, giving an initial NO 3 -N concentration of 40 N /L. Phase II (cycles 127–182): the operating mode was similar to that of phase I, except that the AnRT for the SBR was reduced to 60 min; accordingly, the idle period was extended to 105 min. The HRT of the SBR for this phase was 10.9 h. Phase III (cycles183–228): the operating conditions were similar to that of phase II, while the initial NO 3 -N concentration was increased to 50–60 mg N/L. Phase IV (cycles 229–399): the operating conditions were similar to that of phase I, except that the initial NO 3 -N concentration was maintained at 50 mg N/L. The liquid and solid-phase samples were taken for chemical analysis with sampling intervals of 10 min during the first 30 min of the anaerobic period and anoxic period and each 30 min afterwards. Gas samples were collected using gas-tight collecting bags and syringes and then analyzed immediately. 2.2. Synthetic wastewater Synthetic wastewater used in this study contained (per liter): 512.53 mg of CH3COONa (400 mg of COD); 32.9 mg of KH2PO4

5675

(7.5 mg of P); 42 mg of K2HPO4 (7.5 mg of P); 57.4 mg of NH4Cl; 85 mg of MgSO47H2O; 10 mg of CaCl2; 110 mg of NaHCO3. Additionally, the trace salt solution (0.3 mL/L) described by Wang et al. (2010) was added. 2.3. Batch experiments Three series batch experiments were conducted to characterize the effect of AnRT on denitrifying phosphorus removal and N2O production. For Experiments 1 and 2, a 2.5 L sealed reactor with a working volume of 2.4 L and overhead space of 0.1 L was used. For Experiment 3, a 0.5 L sealed reactor without overhead space was used. The temperature for all tests was controlled at 22 ± 2 °C. The pH values in Experiments 1 and 2 were maintained at 7.5 ± 0.1 by adding 0.3 M HCl or 0.3 M NaOH; for Experiment 3, there was no pH control. 2.3.1. Effect of AnRT on denitrifying phosphorus removal and N2O production (Experiment 1) The sludge (2 L) for the batch tests was withdrawn from the An/A/O SBR at the end of the decanting phase during phase III. After being washed three times, the sludge was evenly divided into three parts and then transferred into three batch reactors. Next, 1.7 L of synthetic wastewater was rapidly added to each reactor at the beginning of the cycle, which resulted in mixed liquor volatile suspended solids (MLVSS) levels of about 2.6 g/L in each bioreactor. The anoxic reaction time for all reactors was 210 min, while the AnRTs for three reactors were controlled at 60 min (R1), 90 min (R2) or 120 min (R3). The anoxic reaction was initiated by adding KNO3 solution to give an initial NO 3 -N concentration of 55 mg N/L. 2.3.2. Denitrification rates of nitrate and nitrite with PHA as a carbon source (Experiment 2) The sludge for Experiment 2 was withdrawn from the An/A/O SBR at the end of the anaerobic phase during phase IV. The washed sludge was evenly divided into two parts and then transferred into two reactors. Next, 1.7 L of synthetic wastewater was added to the two reactors to give an initial P concentration of 40 mg/L. One reactor was then amended with KNO3 solution to give an initial NO 3 -N concentration of 40 mg/L, while the other received NaNO2 solution to provide an initial NO 2  N concentration of 30 mg/L. The anoxic reaction lasted for 210 min. The MLVSS for these two reactors were both approximately 3.0 g/L. The tests are conducted in triplicate. 2.3.3. Denitrification rate of N2O with PHA as a carbon source (Experiment 3) The sludge mixture (500 mL) for Experiment 3 was withdrawn from the An/A/O SBR at the end of the anaerobic phase in phase IV. An N2O solution was prepared in a 500 mL reactor by aerating 450 mL of synthetic wastewater with 10,000 ppm N2O gas for 15 min, after which the washed sludge (50 mL) was injected rapidly into the reactor. The anoxic reaction was 30 min. The experiment was repeated three times. 2.4. Analytical methods The liquid samples were immediately filtered through Millipore  filter units (0.45 lm pore size) for analysis of COD, NHþ 4 -N, NO3 -N, 3   3 þ NO -N and PO -P. NH -N, NO -N, NO -N, PO -P, MLSS and MLVSS 4 2 4 3 2 4 were measured according to the Chinese State Environmental Protection Agency (SEPA) Standard Methods (Chinese SEPA, 2002). Merck COD reagents were used for the COD test (Merck, Germany).

5676

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

Dissolved oxygen (DO), oxidation–reduction potential (ORP) and pH were measured online using oxygen, ORP and pH meters (oxi 3310 and pH 3310, WTW company, Germany), respectively. FNA-N was calculated according to Ma et al. (2010). Glycogen was determined using the method described by Jenkins et al. (2003). The N2O concentrations in the gas samples were analyzed using a gas chromatograph (GC) (Agilent 7820, USA) equipped with an electron capture detector (ECD). The GC was equipped with a pre-column (Porapak Q 80/100 mesh, 1 m  2 mm) and a main column (Porapak Q 80/100 mesh, 3 m  2 mm). A 1 mL sample loop and a 10-port valve were used to inject the gas samples. The temperatures of the columns and the ECD were 100 °C and 300 °C, respectively. A mixture of 91% Ar + 9% CH4 was used as the carrier gas. The dissolved N2O was measured by GC using a headspace method. The equilibrium temperature and time were 25 °C and 3 h, respectively. N2O was quantified using the standard curves generated from certified standard gases (National Institute of Metrology, PR China). Acetic acid (HAc) was measured using an Agilent 6890N gas chromatograph (GC) equipped with a 30  0.53 mm  1 lm (length  ID  film) DB-WAXetr column and a flame ionization detector (FID) at 220 °C. PHB, poly-hydroxyvalerate (PHV) and poly-3-hydroxy-2-methylvalerate (PH2MV) were measured according to a method described by Oehmen et al. (2005). The total PHA in the samples was calculated as the sum of the measured PHB, PHV and PH2MV.

3. Results and discussion 3.1. Typical cycles of the An/A/O SBR in phases I and II During phases I and II, the mean effluent PO3 4 -P concentrations remained at 3.23 and 2.12 mg/L, respectively. For the effluent NO 2 -N concentration, it averaged at 11 and 3.28 mg/L during phases I and II, respectively. The effluent NO 3 -N concentrations were close to zero during both phases. 3.1.1. Variations of the nitrogen and phosphorus in typical cycles In typical cycles of 77 (phase I), 119 (phase I), 149 (phase II) and 173 (phase II) for the An/A/O SBR, the PO3 4 -P concentrations increased remarkably during the anaerobic period due to the polyphosphate hydrolysis to soluble ortho-phosphorus (SOP) and the release of phosphorus (Figs. 1b and e and 2b and e). However, anaerobic PO3 4 -P uptake occurred in all of these cycles. Specifically, the amounts of P absorbed anaerobically in cycles 77 (phase I), 119 (phase I), 149 (phase II) and 173 (phase II) were approximately 2.48, 18.32, 1.57 and 0.4 mg/L, respectively. Moreover, this absorption was significantly higher in phase I than in phase II (Fobserved = 8.04, Fsignificance = 6.61, P (0.05) = 0.04 < 0.05). Similar anaerobic P uptake was also observed by Kong et al. (2002), but they did not provide relevant reasons. As the SBR was properly sealed and the overhead space was only 1 L, the DO level during the entire testing period was lower than 0.10 mg/L. Further studies

Fig. 1. Variations in nitrogen, phosphorus, N2O, HAc, glycogen and PHA in cycles 77 and 119 for the An/A/O SBR (a, b and c: cycle 77; d, e and f: cycle 119).

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

5677

Fig. 2. Variations in nitrogen, phosphorus, N2O, HAc, glycogen and PHA during cycles 149 and 173 for the An/A/O SBR (a, b and c: cycle 149; d, e and f: cycle 173).

are needed to fully determine the mechanisms of this anaerobic phosphate uptake. In the anoxic phase, the denitrification efficiency of the An/A/O SBR increased after the AnRT was reduced from 90 min to 60 min (Table 1). Notably, the denitrification efficiency in cycle 173 was higher than that in cycle 149, even though the AnRTs for the two cycles were the same. Additionally, anoxic phosphorus uptake was observed in all typical cycles (Figs. 1b and e and 2b and e).

typical cycles (Figs. 1c and f and 2c and f). Thereafter, the N2O production rate decreased gradually as the amount of N2O decreased. Indeed, the N2O-N production showed an obvious and gradual decrease when the AnRT was shortened from 90 min to 60 min (Table 1).

3.1.2. Transformations of PHA and glycogen in the typical cycles As shown in Table 2, after the AnRT was shortened from 90 min to 60 min, the PHA synthesis and consumption as well as the glycogen degradation and synthesis all increased either during the anaerobic period or the anoxic period. Moreover, for phase II, the transformations of PHA and glycogen were enhanced gradually with the operation of the system (Table 2). It should be noted that PHA degradation occurred in the latter part of the anaerobic phase along with glycogen degradation (Figs. 1b and e and 2b and e). Specifically, in cycles 77 (phase I), 119 (phase I), 149 (phase II) and 173 (phase II), over 0.49, 0.15, 0.12 and 0.21 mmol C/g-MLVSS of PHA, respectively, were degraded anaerobically. Moreover, from a long-term operation point of view, this reduction in the amount of PHA for phase I was greater than that of phase II.

3.2.1. Comparation of nitrogen and phosphorus removal in batch Experiment 1 Of R1, R2 and R3, which had AnRTs of 60, 90 and 120 min, respectively, the highest and lowest P release were observed in R2 and R1, respectively (Figs. 3b, 4b, 5b and 6a). Moreover, anaerobic P uptake occurred in R3 during the latter part of the anaerobic phase, which directly led to a reduced net P release when compared with that of R2. This occurred because a fraction of the released P was taken up again by DPAOs. It seems that an extension of AnRT itself cannot efficiently increase the release of phosphorus (Figs. 3b, 4b and 5b). After the batch reactors were fed KNO3 solution, denitrification and PO3 4 -P uptake occurred simultaneously. The highest P removal was obtained in R2, while the lowest was observed in R1 (Fig. 6a). Moreover, the added NO3--N was rapidly reduced in R1, R2 and R3, and this reduction was accompanied by the accumulation and reduction of NO 2 -N (Figs. 3a, 4a and 5a), which resulted in corresponding denitrification efficiencies of 74%, 91% and 83%, respectively (Fig. 6a). These findings indicate that an AnRT of 90 min

3.1.3. N2O production in the typical cycles After the KNO3 was pulse added at the beginning of the anoxic period, a remarkable increase in the N2O production occurred in all

3.2. Effect of AnRT on denitrifying phosphorus removal and N2O production in batch Experiment 1

5678

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

Table 1 Comparison of nitrogen and phosphorus removals as well as N2O production for typical cycles of the An/A/O SBR. Item

P removal efficiency at the end of the anoxic phase (%) a P removal efficiency at the end of the post aerobic phase (%) b Denitrification efficiency (%) c NO 2 -N concentration at the end of the anoxic phase (mg/L) NO 3 -N denitrification rate (mg N/g-MLVSSh) Net NO 2 -N consumption rate (mg N/g-MLVSSh) Anoxic N2O production (mg N/L) Aerobic N2O production (mg N/L) Total N2O production (mg N/L) Ratio of anoxic N2O-N production to denitrified nitrogen (%)

Phase I (AnRT = 90 min)

Phase II (AnRT = 60 min)

Cycle 77

Cycle 119

Cycle 149

Cycle 173

16

39

46

19

4

64

4

42

60 14.94

66 12.16

94 1.76

100 0.04

6.07

13.59

15.34

28.05

0.62

1.05

2.14

3.41

6.09 0.66 6.74 15.20

7.52 0.33 7.85 18.80

6.72 0.16 6.88 16.79

0.94 0.12 1.05 2.34

a P removal efficiency at the end of the anoxic phase (%) = (P0  P1)/P0; P0: PO3 4 -P concentration at the beginning of the anaerobic phase; P1: PO3 4 -P concentration at the end of the anoxic phase; b P removal efficiency at the end of the post aerobic phase (%) = (P0  P2)/P0; P0: P concentration at the beginning of the anaerobic phase; P2: PO3 4 -P concentration at the end of the aerobic stage.     c Denitrification efficiency (%) = (NOx -N1  NOx -N2 Þ=NOx -N1 ; NO x -N1 : NO3 -N þ   NO 2 -N concentration at the beginning of the anoxic phase; NOx -N2 : NO3 -N þ NO -N concentration at the end of the anoxic phase. 2

Table 2 Anaerobic and anoxic transformations of PHA and glycogen in typical cycles of the An/ A/O SBR. Item

Anaerobic PHB synthesis (mmol C/g-MLVSS) Anaerobic PHV synthesis (mmol C/g-MLVSS) Anaerobic PHA synthesis (mmol C/g-MLVSS) PHA amount at the end of the anaerobic phase (mmol C/g-MLVSS) PHA synthesis/HAc uptake (mmol C/mmol C) Anaerobic glycogen degradation (mmol C/g-MLVSS) Anoxic PHB degradation (mmol C/g-MLVSS) Anoxic PHV degradation (mmol C/g-MLVSS) Anoxic PHA degradation (mmol C/g-MLVSS) Anoxic PHA degradation rate a (mmol C/g-MLVSSh) Anoxic glycogen synthesis (mmol C/g-MLVSS) a

Phase I (AnRT = 90 min)

Phase II (AnRT = 60 min)

Cycle 77

Cycle 119

Cycle 149

Cycle 173

1.83

1.85

2.55

3.24

0.29

0.34

0.52

0.66

2.13

2.18

3.07

3.90

2.28

2.32

3.30

5.48

0.71

0.71

0.89

1.29

1.07

1.02

1.55

2.12

1.76

1.73

2.60

2.98

0.29

0.34

0.51

0.71

2.11

2.08

3.12

3.71

1.71

1.83

2.52

3.27

1.12

1.20

1.95

2.19

The mean PHA degradation rate during the initial 60 min of anoxic reaction.

was more favorable for the denitrifying phosphorus removal for these cases. 3.2.2. Effect of AnRT on transformations of HAc, PHA and glycogen in batch Experiment 1 For R2 and R3, the added HAc was completely consumed during the anaerobic period, while a fraction of HAc still remained in R1

(Figs. 3b, 4b and 5b). Consequently, the PHA synthesis in R1, R2 and R3 was 2.10, 2.41 and 2.20 mmol C/g-MLVSS, respectively, and the corresponding glycogen degradation was 1.11, 1.32 and 1.54 mmol C/g-MLVSS, respectively (Fig. 6b). These findings appear to disagree with those of Zheng et al. (2009), who reported that higher PHA synthesis was generally associated with greater glycogen degradation. However, as shown in Fig. 5b, at least 0.21 mmol C/g-MLVSS of PHA was degraded during the latter part of the anaerobic phase in R3, which induced a decline in the net amount of the synthesized PHA. With respect to the anoxic phase, the greatest PHA degradation and glycogen synthesis both occurred in R2, while the lowest was obtained in R1 (Fig. 6b). The characteristics of PHA transformation in the three reactors were consistent with the corresponding denitrification and phosphorus removal efficiencies (Figs. 3–5). 3.2.3. Effect of AnRT on N2O production in batch Experiment 1 The variations in N2O production, NO 2 -N and FNA concentrations displayed similar trends in the three reactors (Figs. 3c, 4c and 5c). For R1, R2 and R3, the ratios of N2O-N production to denitrified nitrogen were 4.84%, 4.52% and 6.43%, respectively (Table 3). These findings suggest that the minimum N2O production could be obtained in the case of an AnRT of 90 min. 3.3. Effect of the AnRT on the anaerobic PHA synthesis by the phosphorus removal organisms 3.3.1. Effect of the short or long AnRT on PHA synthesis Batch Experiment 1 confirmed that HAc could not be completely consumed if the AnRT was insufficient, which led to a reduction in the maximum amount of PHA synthesis that could be produced. Moreover, it should be noted that PHA synthesis still continued in some cases even though the HAc was close to depletion during the latter part of the anaerobic phase (Figs. 1e, 2e and 5b), indicating that PHA synthesis may lag behind HAc uptake. Therefore, it is likely that the apparent HAc was exhausted within the set AnRT, while complete PHA synthesis was not yet accomplished. This is likely to result in incomplete PHA synthesis and poor P removal if the set AnRT is too short. In contrast, as shown in Table 2 and Fig. 6b, a long AnRT also lowered the net anaerobic PHA synthesis in the present study, which resulted in deteriorated phosphorus removal and denitrification. As the SBR was sealed, the lowered anaerobic PHA storage when with a long AnRT might be due primarily to the endogenous respiration for the maintenance of microbial cells. Indeed, Qin et al. (2005) found that anaerobic endogenous respiration of granular sludge occurred when external carbon and nitrate were not available, which subsequently led to a decrease in the PHB pool in microbial granules. In several studies a slight increase or decrease in the amount of PHA for the PAOs sludge was observed during the latter part of the anaerobic phase (Kong et al., 2002; Oehmen et al., 2005). This means that the increase in the PHA amount is not significant within the prolonged AnRT and instead a decrease in the PHA amount may occur. Therefore, an AnRT that is too long is not appropriate in terms of minimizing operating costs and ensuring good phosphorus removals. Generally, during anaerobic endogenous period, poly-P hydrolysis and glycogen degradation produce maintenance energy, and part of glycogen degraded can be converted to PHV which makes the overall amount of PHB not vary significantly (Lopez et al., 2006; Lu et al., 2007). Interestingly, our observations that a decline in PHA amount occurred during the latter part of the anaerobic phase are not in accordance with the work by Lopez et al. (2006) and Lu et al. (2007), who found that an increase in PHA content occurred when activated sludge used for phosphorus removal was subject to long-term anaerobic starvation. This is primarily be-

5679

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

14

anaerobic

anoxic

KNO3

48

a)

40

-

N O 3 -N

12

-

N O 2 -N

10

+

8

24

6

16

4

8

2

(mmol-C/gVSS or mmol-C/L)

0 8 7 6

20

2

10

1 0 5

0 24

c)

dissolved N2 O off-gas N2 O

4 N 2O (mgN/L)

-4

30

3

20

total N2 O -

16

FNA

12

NO 2 -N

3

-

FNA (10 mg N/L)

40

2

8

4 2 0

1 0

NO2-N (mg/L)

6

50 3-

8

60

b)

4

14 10

0 70

G ly V FA P

5

16 12

PH B PH V PH 2M V PH A

PO4 -P (mg/L)

PHA, PHB, PHV, PH2MV, Gly or VFA

-

-

32

+

N H 4 -N

NH4 -N (mg/L)

NO3-N and NO2-N (mg/L)

56

4 0 0

30

60

90

120 150 Time (min)

180

210

240

270

Fig. 3. Variations in nitrogen, phosphorus, N2O, HAc, glycogen and PHA during one cycle in batch R1.

cause the biomass used in their experiments was withdrawn from the reactor at the end of the aerobic phase and initially with high levels of glycogen and poly-P and low level of PHA. However, when the levels of internal poly-P and glycogen were both lower, e.g. during the latter part of the anaerobic phase in the system evaluated here, microorganisms may also use PHA to generate maintenance energy. Indeed, the anaerobic decrease in PHA has been observed in some previous studies (Kong et al., 2002; Oehmen et al., 2005; Lopez et al., 2006). For example, observations of changes in PHA compositions by Lopez et al. (2006) showed that a decrease in the PHA amount occurred after 14 d of anaerobic starvation, although the long-term trend for the PHA amount (from day 0 to day 27.9) had increased. Further studies are required to confirm the findings. 3.3.2. Effect of AnRT on the activity of DPAOs linked to PHA synthesis The PHA synthesis in the An/A/O SBR was much greater in phase II than in phase I, possibly due to a change in the microbial activities in the system. Lopez et al. (2006) found that the activity of PAOs dropped after long-term anaerobic starvation. Similar results were also reported by Pijuan et al. (2009). Therefore, it is likely that endogenous processes caused by a suboptimal AnRT affected the activities of DPAOs in the system evaluated here, which ultimately

led to different capacities of PHA synthesis in DPAOs during phases I and II. Moreover, the lowered PHA storage due to the declined activities of the DPAOs in denitrifying phosphorus removal systems would cause incomplete denitrification, resulting in the accumulation of nitrite. The toxicity due to the accumulated nitrite or high FNA inhibited the activities of DPAOs, which in turn could lower PHA synthesis (Zhou et al., 2010). It must be noted that after the An/A/O SBR was operated at an AnRT of 60 min, the maximum phosphate release rate was increased from 49.9 mg P/g-MLVSSh for phase I to 68.6 mg P/g-MLVSSh for phase II, and the maximum HAc uptake rate was also increased from 7.27 mmol C/g-MLVSSh for phase I to 9.61 mmol C/g-MLVSSh for phase II, indicating that the activity of the DPAOs was gradually increased. This directly resulted in the net PHA synthesis in one standard cycle increasing concurrently with a decrease in the nitrite accumulation (Table 2). Subsequently, the microbial activities were recovered smoothly due to the reduced nitrite inhibition. Additionally, according to the proposed PAOs and glycogen accumulating organisms (GAOs) metabolism (Filipe et al., 2001), GAOs tend to produce more PHV than PAOs when fed with acetate, due to the partial glycogen hydrolysis through the succinate-propionate pathway. In our studied SBR, the PHB fractions during phase

5680

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

14

anaerobic

anoxic

KNO 3

48

a)

-

4

8

2

6 5 4

50

b)

3

N2O (mg N/L)

-4

FNA (10 mg N/L)

60

40 30

2

20

1

10

0 5

0 24

dissolved N 2 O

4

off-gas N 2 O total N 2 O

3

NO 2 -N

c)

-

20 16

-

0

0 70

12

FNA

2

8

1

4

0

NO2-N (mg/L)

2

Gly VFA P

3-

(mmol-C/gVSS or mmol-C/L)

7

12

4

PHB PHV PH2MV PHA

PO4 -P (mg/L)

PHA, PHB, PHV, PH2MV, Gly or VFA

16

+

6

14

6

8

24

16

8

10

+ 4

NH -N

32

0 8

10

12

2

NO -N

-

40

-

NO 3 -N

NH4 -N (mg/L)

NO3-N and NO2-N (mg/L)

56

0 0

30

60

90

120 150 180 210 240 270 300 Tim e (m in)

Fig. 4. Variations in nitrogen, phosphorus, N2O, HAc, glycogen and PHA during one cycle in batch R2.

II and phase I contributed to 83% and 85% of PHA, respectively. Also, the P release/VFA-uptake (P mmol/C mmol) ratios during phases I and II were both around 0.24. All of these results indicate that the percentage of the GAOs in the SBR was remained constant during phases I and II. Therefore, the increase in the PHA synthesis capacity during phase II was due primarily to the increased activity of DPAOs.

utilized for the biomass synthesis in an operating cycle can be calculated according to the reduced amount of NHþ 4 -N (Table 4). Theoretically, the denitrification reactions using PHB and PHV as the carbon source can be expressed as Eqs. (1) and (2), respectively. CH1:5 O0:5 ðPHBÞ þ 0:9NO3 ! 0:45N2 þ 0:1CO2 þ 0:3H2 O þ 0:9HCO3 CH1:6 O0:4 ðPHVÞ þ 0:96NO3 ! 0:48N2 þ 0:04CO2 þ 0:32H2 O þ 0:96HCO3

ð1Þ ð2Þ

3.4. Effect of AnRT on denitrification and phosphorus removal

Accordingly, approximately 1.11 mol C of PHB or 1.04 mol C of PHV is required for a complete reduction of 1 mol NO 3 -N to N2.

3.4.1. Intracellular carbon flow during the anoxic reaction To characterize the contribution of the internal carbon (PHA) and identify the mechanism of N2O production during the anoxic period, a simple intracellular carbon flow for anoxic reaction was proposed (Table 4). According to the metabolic model of the EBPR proposed by Kuba et al. (1996b), carbon supplied by PHA was used to replenish glycogen, biomass synthesis and oxidized to CO2. In the present study, nitrification was negligible during the anoxic  period because the DO was below 0.2 mg/L. NO 3 /NO2 ammonification is considered not to occur under these conditions, and the removal of NHþ 4 -N in the anoxic phase was only used for microbial cell synthesis. The biomass composition can be expressed as CH2.09O0.54N0.20P0.015 (Smolders et al., 1994); therefore, the carbon

3.4.2. Effect of AnRT on denitrification According to the anoxic internal carbon flow in Table 4, the difference between the PHB or PHV amount demanded for the complete denitrification and the PHA amount used for denitrification for the An/A/O SBR was larger in phase I than in phase II. In other words, the An/A/O SBR during phase I lacked more amount of anaerobically synthesized PHA to meet the demands of complete denitrification when compared to phase II, which led to lower denitrification efficiencies for the SBR in phase I (Table 1). On the other hand, when the amount of the anaerobically synthesized PHA increased in the An/A/O SBR, the sequential anoxic degradation rate of PHA increased accordingly (Table 2). Furthermore, lower PHA degradation rates always corresponded with lower rates

5681

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

32

8

24

6

16

4

8

2

6 4 2 0

60 50

5

40

4

b)

30

3 2

20

1

10

0 5

dissolved N 2 O

c)

off-gas N 2 O

4

0 24 20

total N 2 O

16

-

NO2 -N

3

-

8

6

12

FNA

2

8

1

4

0

NO2-N (mg/L)

10

7

0 70

Gly VFA P

3-

12

PHB PHV PH2MV PHA

PO4 -P (mg/L)

(mmol-C/gVSS or mmol-C/L)

0 8

N2 O (mg N/L)

-

2

10

+

NH 4 -N

3

-

12

-

NO 2 -N

+

PHA, PHB, PHV, PH2MV, Gly or VFA

-

NO 3 -N

a)

40

14

-4

14

anoxic

KNO 3

48

16

FNA (10 mg N/L)

anaerobic

NH 4-N (mg/L)

NO -N and NO -N (mg/L)

56

0 0

30

60

90

120

150 180 Time (min)

210

240

270

300

330

Fig. 5. Variations in nitrogen, phosphorus, N2O, HAc, glycogen and PHA during one cycle in batch R3.

 of NO 3 -N denitrification and net NO2 -N consumption, as well as lower denitrification efficiencies (Tables 1 and 2). This was primarily because a low rate of PHB availability results in limitation of reducing power for denitrification, which may lead to competition for electrons between denitrifying enzymes and long-term accumulation of nitrite or FNA (Figs. 1a and d and 2a and d). Accordingly, denitrification efficiency decreased because nitrite or FNA has an inhibitory effect on denitrifiers (Zhou et al., 2007; Ma et al., 2010). Similar results were also obtained in batch Experiment 1 (Tables 3 and 4 and Fig. 6a). All of these results confirm that not only the amount of anaerobically synthesized PHA but the kinetics of PHA degradation has a significant effect on denitrification.

3.4.3. Effect of AnRT on phosphorus removal The long-term performance of the An/A/O SBR showed that more nitrite or FNA accumulated during phase I (AnRT of 90 min) than in phase II (AnRT of 60 min), due to the lower amount of PHA synthesized during phase I. Nitrite and FNA accumulation are known to inhibit phosphorus uptake (Zhou et al., 2007), which may be involved in damage to polyphosphate kinases (PPK) or PHA oxidation (Zhou et al., 2007). As a result, the phosphorus removal efficiency was lower during phase I than during phase II (data not shown). Similar phenomena were also observed in Experiment 1. Moreover, higher nitrite or FNA accumulation was always associated with reduced phosphorus removal efficiency (Figs. 3c, 4c, and 5c). In particular, a phosphorus release occurred in R1 in the latter part of the anoxic phase concurrently with a severe nitrite

accumulation (Fig. 3b), indicating that the energy production by PHA may have been inhibited by nitrite or FNA. 3.5. Identifying the causes of N2O production during the denitrifying phosphorus removal 3.5.1. Effect of the amount of anaerobically synthesized PHA on N2O production As shown in Table 4, the amounts of PHA available for the complete denitrification were not sufficient in all cases, which inevitably led to N2O accumulation. Nevertheless, as the anaerobic PHA synthesis increased, the ratio of N2O-N production to denitrified nitrogen decreased in the anoxic phase (Tables 1–3 and Fig. 6b). Numerous studies have reported that COD/N (m/m) has a significant impact on N2O production during denitrification (Hwang et al., 2006; Yang et al., 2009). In the present study, by expressing PHA as mg COD, the COD/NO x -N in the initial anoxic reaction was 6.24, 5.95, 7.77 and 14.1 in cycles 77 (phase I), 119 (phase I), 149 (phase II) and 173 (phase II), respectively, while they were 4.42, 4.98 and 4.64 in R1, R2 and R3, respectively. Clearly, the COD/NO x -N at the end of the anaerobic phase increased when the anaerobic PHA synthesis increased, which led to a reduction in the ratio of N2O-N production to denitrified nitrogen (Tables 1–3 and Fig. 6b). However, relatively more N2O was produced in our experiment, even though the COD/N in the initial anoxic phase was higher than that observed in some studies in which COD/N impacts on N2O production were examined during denitrification based on

5682

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

80 P removal efficiency denitrification efficiency

90

a)

60

100

80

50

70

40

60

30

50 40

20

removal efficiency (%)

phosphorus or nitrogen (mg/L)

70

110 phosphorus release phosphorus uptake denitrified nitrogen

30

10

20

0

10 1h (R1)

1.5h (R2)

2h (R3)

PHB, PHV, PH2MV and Gly (mmol-C/gVSS)

Anearobic reaction time (h) 4.0 3.5

anaerobic PHB synthesis anoxic PHB degradation anaerobic glycogen degradation anoxic glycogen synthesis

anaerobic PHV synthesis anoxic PHV degradation anaerobic PH2MV synthesis anoxic PH2MV degradation

3.0

b)

2.5 2.0 1.5 1.0 0.5 0.0 1h (R1)

1.5h (R2)

2h (R3)

Anaerobic reaction time (h) Fig. 6. The effect of anaerobic reaction time on N&P removals and intracellular compounds transformation in three batch reactors (a: variations in denitrification, phosphorus release and uptake, as well as denitrification and phosphorus removal efficiencies; b: the transformations of PHA and glycogen).

Table 3 Comparison of denitrification rate, N2O production and the transformation of PHA in R1, R2 and R3. Item

R1*

R2*

R3*

NO 3 -N

13.41 0.68 1.97 4.84

10.58 1.95 2.25 4.52

10.45 1.03 2.93 6.43

2.51

2.91

2.68

0.86 0.69

0.86 0.77

0.79 0.73

denitrification rate (mg N/g-MLVSSh) Net NO 2 -N consumption rate (mg N/g-MLVSSh) Anoxic N2O production (mg N/L) Ratio of anoxic N2O-N production to denitrified nitrogen (%) PHA amount at the end of the anaerobic phase (mmol C/g-MLVSS) PHA synthesis/HAc uptake (mmol C/mmol C) PHA degradation rate a (mmol C/g-MLVSSh)

a *

The mean rate of PHA degradation during the initial 60 min of anoxic reaction. R1: AnRT = 60 min; R2: AnRT = 90 min; R3: AnRT = 120 min.

external carbon (Yang et al., 2009; Hwang et al., 2006). Hwang et al. (2006) demonstrated that a negligible amount of N2O was produced during denitrification when the COD/N was 3. Similarly, Yang et al. (2009) reported that when the COD/NO3--N ratios were 4.7 and 3.8, nitrate was completely denitrified without N2O production, whereas when the COD/NO 3 -N ratios were 1.5 and 0.4, N2O accumulation occurred. Indeed, the denitrification using intracellular PHB produces more N2O compared to that using the external soluble carbon.

It should also be noted that the amount of PHA at the beginning of the anoxic phase in R3 was higher than that in R1, but that the ratio of N2O production to denitrified nitrogen in R3 was greater than that obtained in R1. This may have been due to the fact that unconsumed HAc during the anaerobic period was residual to the anoxic phase in R1, and this was exhausted by ordinary denitrifying bacteria, which therefore improved the denitrification rate and reduced N2O emission. 3.5.2. Effect of the kinetics of PHA degradation and denitrification on N2O production In addition to the amount of PHA, the kinetics of PHA degradation also had a significant effect on denitrification phosphorus removal efficiency as well as N2O production. A slower rate of PHA degradation can lead to competition for electrons between denitrifying enzymes, and even influence the expression of denitrifying enzymes (Kampschreur et al., 2009). Specifically, under certain conditions, e.g. with a limited carbon source or nondegradable carbon sources, nitrate reductase (Nar) may have a competitive advantage at capturing electrons when compared to nitrite reductase (Nir). This imbalance of the two reduction rates easily results in nitrite or FNA accumulation and N2O production (Zeng et al., 2003a,b; Lemaire et al., 2006; Zhou et al., 2008a,b). Moreover, the reduction of N2O is the last step of denitrification, which makes it more difficult for N2O reductase (Nos) to compete for electrons than other enzymes, particularly when the amount of the anaerobically synthesized PHA is low. Therefore, if more PHA is synthesized, more carbon sources are available for denitrification and the PHA degradation rate can be improved. This ultimately leads to a lower ratio of N2O-N production to denitrified nitrogen. Nevertheless, PHB is degraded 6–20 times slower than soluble COD (Third et al., 2003). Consequently, more N2O was produced in the present study by using of internal PHA for denitrification, even though the initial COD/N of the anoxic phase in our experiment was comparable to that of ordinary denitrifying systems (using external carbon). The denitrification rates with the addition of nitrate, nitrite and N2O were evaluated in Experiments 2 and 3 based on PHA consumption. The NO3--N denitrification rate was 13.43 ± 0.17 mg N/g-MLVSSh, which was higher than the NO 2 -N denitrification rate of 9.90 ± 0.15 mg N/g-MLVSSh, while the lowest value of 4.23 ± 1.21 mg N/g-MLVSSh was the N2O denitrification rate. These findings support our assumption that Nos was less competitive in obtaining electrons during denitrification, and also explain the fact that NO 2 -N accumulation and N2O production occurred during the anoxic phase in both the continuous operation and batch tests. 3.5.3. Effect of nitrite or FNA on N2O production As shown in Figs. 1c, 1f, 2c, 2f, 3c, 4c and 5c, significant nitrite or FNA accumulation was always associated with a high ratio of N2O-N production to denitrified nitrogen. Therefore, nitrite or FNA had an important effect on N2O production. In batch Experiment 2, nitrite addition to the DPAOs sludge stimulated the net N2O production rate by 4.7 times when compared with that obtained in the nitrate addition batch (data not shown). This finding is consistent with that of Lemaire et al. (2006), who reported that the addition of nitrite to the sludge led to a net N2O production rate that was five times higher than that obtained in response to nitrate addition. Similarly, Zeng et al. (2003a) demonstrated that once nitrite was close to 5 mg N/L, N2O was produced instead of N2. They later found that Nos was inhibited when the nitrite concentration reached approximately 1 mg N/L (Zeng et al., 2003b). Recently, Zhou et al. (2008a) reported that the N2O reduction activity decreased by 50% at FNA concentrations of 0.0007–0.001 mg HNO2-N/L and was totally inhibited

5683

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684 Table 4 Intracellular carbon flow in the anoxic stage for the An/A/O SBR and batch tests. Item

An/A/O SBR Cycle77 (Phase I)

c

Cycle 149 (Phase II)

Cycle 173 (Phase II)

R1

R2

R3

2.11 1.12 0.53

2.08 1.20 0.36

3.12 1.95 0.55

3.71 2.19 0.63

2.34 1.40 0.48

2.71 1.58 0.28

2.57 1.5 0.37

PHA amount used for denitrification PHB demanded for the complete denitrification a PHV demanded for the complete denitrification a

0.46 1.06 0.97

0.52 1.13 1.03

0.62 1.23 1.13

0.89 1.13 1.04

0.46 1.64 1.51

0.85 1.70 1.56

0.70 1.67 1.54

17.43 18.59

18.46 19.69

20.18 21.52

31.50 33.60

15.39 16.41

27.56 29.39

23.03 24.56

4.48

2.82

3.98

4.94

0.40

0.53

0.51

Anoxic NHþ 4 -N consumption (mg N /L) b

Cycle 119 (Phase I)

Anoxic PHA degradation a Anoxic Glycogen synthesis a PHA amount used for cell growth

b (mg N /L) The amount of denitrified NO 3 -N using the synthesized PHB c The amount of denitrified NO 3 -N using the synthesized PHV (mg N /L)

a

Batch tests

The unit was mmol C/g-MLVSS. Completely denitrified NO 3 —N amount using PHB as a sole carbon source, providing that PHB amount accounted for 100% of the PHA supplied for denitrification. Completely denitrified NO 3 —N amount using PHV as a sole carbon source, providing that PHV content accounted for 100% of the PHA content supplied for denitrification.

when the FNA concentration was greater than 0.004 mg HNO2-N/L. For our studied An/A/O SBR, the maximum FNA concentrations were 0.0004, 0.0003, 0.0005 and 0.0007 mg N/L, respectively, in cycles 77 (phase I), 119 (phase I), 149 (phase II) and 173 (phase II), and that they were 0.0010, 0.0009 and 0.0013 mg N/L, respectively, in R1, R2 and R3. This suggests that the inhibitory effect of FNA on the N2O reduction activity also occurred in our studied systems. In addition, for cycles 77 (phase I), 119 (phase I), 149 (phase II) and 173 (phase II) in the An/A/O SBR, as the nitrite concentrations increased at the end of the anoxic phase, the N2O production increased gradually in the post oxic period (Table 1), possibly via the denitrification pathway. Indeed, Kim et al. (2010) reported that N2O emission was greatly enhanced when nitrite was added to an NHþ 4 -N oxidizing system, and they demonstrated that N2O was produced via the denitrification pathway by AOB. 4. Conclusions AnRT significantly influenced anaerobic PHA synthesis. A reduction in AnRT from 90 min to 60 min in the An/A/O SBR resulted in an improvement in PHA synthesis of about 1.8 times. The PHA storage and the kinetics of PHA degradation are both key variables in controlling the N2O emission for anoxic P removal reactor. Particularly, a lower PHA storage leads to a lower PHA degradation rate, and the combined effects of lower PHA storage and degradation rate stimulate the N2O production during denitrification. Nitrite or FNA can inhibit the activity of denitrifying enzymes and phosphorus uptake, leading to poor denitrification efficiency and a large amount of N2O production. Acknowledgements This study was supported by the National Natural Science Foundation of China (NSFC) (no. 51078283), the Program for New Century Excellent Talent in University (No. NCET-08-0404), and the Shanghai Science and Technology Development Funds (no. 09QA1406100). References Beun, J.J., Dircks, K., Van Loosdrecht, M.C.M., Heijnen, J.J., 2002. Poly-bhydroxybutyrate metabolism in dynamically fed mixed microbial cultures. Water Res. 36 (5), 1167–1180. Chinese SEPA, 2002. Water and Wastewater Monitoring Methods, fourth ed. Chinese Environmental Science Publishing House, Beijing, China. Colliver, B.B., Stephenson, T., 2000. Production of nitrogen oxide and dinitrogen oxide by autotrophic nitrifiers. Biotechnol. Adv. 18, 219–232. Filipe, C.D.M., Daigger, G.T., Grady, C.P.L., 2001. A metabolic model for acetate uptake under anaerobic conditions by glycogen accumulating organisms: stoichiometry, kinetics, and the effect of pH. Biotechnol. Bioeng. 76 (1), 17–31.

Freitas, F., Temudo, M.F., Carvalho, G., Oehmen, A., Reis, M.A.M., 2009. Robustness of sludge enriched with short SBR cycles for biological nutrient removal. Bioresour. Technol. 100, 1969–1976. Hwang, S.J., Jang, K., Jang, H., Song, J., Bae, W., 2006. Factors affecting nitrous oxide production: a comparison of biological nitrogen removal processes with partial and complete nitrification. Biodegradation 17, 19–29. IPCC, 2001. Climate Change 2001: The Scientific Basis. Cambridge University, Cambridge. Jenkins, D., Richard, M.G., Daigger, G.T., 2003. Manual on the Causes and Control of Activated Sludge Bulking, Foaming and other Solids Separation Problems, second ed. Lewis Publishers. Kampschreur, M.J., van der Star, W.R.L., Wielders, H.A., Mulder, J.W., Jetten, M.S.M., van Loosdrecht, M.C.M., 2008. Dynamics of nitric oxide and nitrous oxide emission during full-scale reject water treatment. Water Res. 42, 812–826. Kampschreur, M.J., Temmink, H., Kleerebezem, R., Jetten, M.S.M., van Loosdrecht, M.C.M., 2009. Nitrous oxide emission during wastewater treatment. Water Res. 43, 4093–4103. Kim, S.W., Miyahara, M., Fushinobu, S., Wakagi, T., Shoun, H., 2010. Nitrous oxide emission from nitrifying activated sludge dependent on denitrification by ammonia-oxidizing bacteria. Bioresour. Technol. 101, 3958–3963. Kong, Y.H., Beer, M., Reers, G.N., Seviour, R.J., 2002. Functional analysis of microbial communities in aerobic–anaerobic sequencing batch reactors fed with different phosphorus/carbon (P/C) ratios. Microbiology 148, 2299–2307. Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1996a. Phosphorus and nitrogen removal with minimal cod requirement by integration of denitrifying dephosphatation and nitrification in a two-sludge system. Water Res. 30, 1702–1710. Kuba, T., Murnleitner, E., van Loosdrecht, M.C.M., Heijnen, J.J., 1996b. A metabolic model for biological phosphorus removal by denitrifying organisms. Biotechnol. Bioeng. 52, 685–695. Lemaire, R., Meyer, R., Taske, A., Crocetti, G.R., Keller, J., Yuan, Z.G., 2006. Identifying causes for N2O accumulation in a lab-scale sequencing batch reactor performing simultaneous nitrification, denitrification and phosphorus removal. J. Biotechnol. 122, 62–72. Lopez, C., Pons, M.N., Morgenroth, E., 2006. Endogenous processes during long-term starvation in activated sludge performing enhanced biological phosphorus removal. Water Res. 40, 1519–1530. Lu, H.B., Keller, J., Yuan, Z.G., 2007. Endogenous metabolism of Candidatus Accumulibacter phosphatis under various starvation conditions. Water Res. 41, 4646–4656. Ma, J., Yang, Q., Wang, S.Y., Wang, L., Takigawa, A., Peng, Y.Z., 2010. Effect of free nitrous acid as inhibitors on nitrate reduction by a biological nutrient removal sludge. J. Hazard. Mater. 175, 518–523. Meyer, R.L., Zeng, R.J.X., Giugliano, V., Blackall, L.L., 2005. Challenges for simultaneous nitrification, denitrification, and phosphorus removal in microbial aggregates: mass transfer limitation and nitrous oxide production. FEMS Microbiol. Ecol. 52 (3), 329–338. Oehmen, A., Keller-Lehmann, B., Zeng, R.J., Yuan, Z.G., Keller, J., 2005a. Optimisation of poly-beta-hydroxyalkanoate analysis using gas chromatography for enhanced biological phosphorus removal systems. J. Chromatogr. A 1070, 131–136. Oehmen, A., Teresa Vives, M., Lu, H.B., Yuan, Z.G., Keller, J., 2005b. The effect of pH on the competition between polyphosphate accumulating organisms and glycogen-accumulating organisms. Water Res. 39, 3727–3737. Pijuan, M., Werner, U., Yuan, Z.G., 2009. Effect of long term anaerobic and intermittent anaerobic/aerobic starvation on aerobic granules. Water Res. 43, 3622–3632. Qin, L., Liu, Y., Tay, J.H., 2005. Denitrification on poly-b-hydroxybutyrate in microbial granular sludge sequencing batch reactor. Water Res. 39, 1503– 1510.

5684

Y. Wang et al. / Bioresource Technology 102 (2011) 5674–5684

Schalk-Otte, S., Seviour, R.J., Kuenen, J.G., Jetten, M.S.M., 2000. Nitrous oxide (N2O) production by Alcaligenes faecalis during feast and famine regimes. Water Res. 34, 2080–2088. Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M., Heijnen, J.J., 1994. Stoichiometric model of the aerobic metabolism of the biological phosphorus removal process. Biotechnol. Bioeng. 44, 837–848. Third, K.A., Burnett, N., Cord-Ruwisch, R., 2003. Simultaneous nitrification and denitrification using stored substrate (PHB) as the electron donor in an SBR. Biotechnol. Bioeng. 83, 706–720. Wang, Y.Y., Jiang, F., Zhang, Z.X., Xing, M.Y., Lu, Z.B., Wu, M., Yang, J., Peng, Y.Z., 2010. The long-term effect of carbon source on the competition between polyphosphorus accumulating organisms and glycogen accumulating organism in a continuous plug-flow anaerobic/aerobic (A/O) process. Bioresour. Technol. 101, 98–104. Yang, Q., Liu, X.H., Peng, C.Y., Wang, S.Y., Sun, H.W., Peng, Y.Z., 2009. N2O production during nitrogen removal via nitrite from domestic wastewater: main sources and control method. Environ. Sci. Technol. 43, 9400–9406. Zeng, R.J., Yuan, Z., Keller, J., 2003a. Enrichment of denitrifying glycogenaccumulating organisms in anaerobic/anoxic activated sludge system. Biotechnol. Bioeng. 81, 397–404.

Zeng, R.J., Lemaire, R., Yuan, Z., Keller, J., 2003b. Simultaneous nitrification, denitrification, and phosphorus removal in a labscale sequencing batch reactor. Biotechnol. Bioeng. 84, 170–178. Zheng, X., Tong, J., Li, H.J., Chen, Y.G., 2009. The investigation of effect of organic carbon sources addition in anaerobic–aerobic (low dissolved oxygen) sequencing batch reactor for nutrients removal from wastewaters. Bioresour. Technol. 100, 2515–2520. Zhou, Y., Pijuan, M., Yuan, Z.G., 2007. Free nitrous acid inhibition on anoxic phosphorus uptake and denitrification by poly-phosphate accumulating organisms. Biotechnol. Bioeng. 98 (4), 903–912. Zhou, Y., Pijuan, M., Zeng, R.J., Yuan, Z.G., 2008a. Free nitrous acid inhibition on nitrous oxide reduction by a denitrifying-enhanced biological phosphorus removal sludge. Environ. Sci. Technol. 42, 8260–8265. Zhou, Y., Pijuan, M., Yuan, Z.G., 2008b. Development of a 2-sludge, 3-stage system for nitrogen and phosphorous removal from nutrient-rich wastewater using granular sludge and biofilms. Water Res. 42, 3207–3217. Zhou, Y., Ganda, L., Lim, M., Yuan, Z.G., Kjelleberg, S., Ng, W.J., 2010. Free nitrous acid (FNA) inhibition on denitrifying poly-phosphate accumulating organisms (DPAOs). Appl. Microbiol. Biotechnol. 88, 359–369.