Optimizing sulfur-driven mixotrophic denitrification process: System performance and nitrous oxide emission

Chemical Engineering Science 172 (2017) 414–422

Contents lists available at ScienceDirect

Chemical Engineering Science journal homepage: www.elsevier.com/locate/ces

Optimizing sulfur-driven mixotrophic denitrification process: System performance and nitrous oxide emission Yiwen Liu a, Huu Hao Ngo a,⇑, Wenshan Guo a, Junliang Zhou a, Lai Peng b, Dongbo Wang c, Xueming Chen d, Jing Sun e, Bing-Jie Ni e,⇑ a

Centre for Technology in Water and Wastewater, School of Civil and Environmental Engineering, University of Technology Sydney, Sydney, NSW 2007, Australia Research Group of Sustainable Energy, Air and Water Technology, Department of Bioscience Engineering, University of Antwerp, Antwerp 2020, Belgium Key Laboratory of Environmental Biology and Pollution Control, Ministry of Education; College of Environmental Science and Engineering, Hunan University, Changsha 410082, China d Advanced Water Management Centre, The University of Queensland, St. Lucia, Brisbane, QLD 4072, Australia e State Key Laboratory of Pollution Control and Resources Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, PR China b c

h i g h l i g h t s 0

 Mitigating N2O from S -driven mixotrophic denitrification was studied with a model.  The model is successfully validated using data from two experimental systems.  Increasing SRT and influent COD/N ratio substantially reduces N2O emission.

a r t i c l e

i n f o

Article history: Received 17 April 2017 Received in revised form 27 May 2017 Accepted 1 July 2017 Available online 3 July 2017 Keywords: Nitrous oxide Autotrophic denitrification Heterotrophic denitrification Sulfur Sulfate production Mathematical modeling

a b s t r a c t Nitrate contamination of groundwater has been recognized as a significant environmental problem world widely. Sulfur-driven mixotrophic denitrification has been demonstrated as a promising groundwater treatment process, which though plays an important role in nitrous oxide (N2O) emissions, significantly contributing to the overall carbon footprint of the system. However, the current process optimizations only focus on nitrate removal and excess sulfate control, with the N2O emission being ignored. In this work, an integrated mathematical model was proposed to evaluate the N2O emission as well as the excess sulfate production and carbon source utilization in sulfur-driven mixotrophic denitrification process. In this model, autotrophic and heterotrophic denitrifiers use their corresponding electron donors (sulfur and organic matter, respectively) to reduce nitrate to nitrogen gas, with each modeled as three-step denitrification (NO3 to N2 via NO2 and N2O) driven by sulfur or organic matter to describe all potential N2O accumulation steps. The developed model, employing model parameters previously reported in literature, was successfully validated using N2O and sulfate data from two mixotrophic denitrification systems with different initial conditions. Modeling results revealed substantial N2O accumulation due to the relatively low autotrophic N2O reduction activity as compared to heterotrophic N2O reduction activity, explaining the observation that higher carbon source addition resulted in lower N2O accumulation in sulfur-driven mixotrophic denitrifying system. Based on the validated model, optimizations of the overall system performance were carried out. Application of the model to simulate long-term operations of sulfur-driven mixotrophic denitrification process indicates that longer sludge retention time reduces N2O emission due to better retention of active biomass. High-level total nitrogen removal with significant N2O emission mitigation, appropriate excess sulfate control and maximized COD utilization can be achieved simultaneously through controlling the influent nitrate and COD concentrations. Ó 2017 Published by Elsevier Ltd.

1. Introduction

⇑ Corresponding authors. E-mail (B.-J. Ni).

addresses:

[email protected] (H.H. Ngo), [email protected]

http://dx.doi.org/10.1016/j.ces.2017.07.005 0009-2509/Ó 2017 Published by Elsevier Ltd.

Due to the intensive utilization of nitrogenous fertilizers and inappropriate discharge of wastewaters and solid wastes (Chen et al., 2016; Wakida and Lerner, 2005), nitrate contamination of groundwater has become a significant environmental issue

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

throughout the world (Rivett et al., 2008). It has been reported that 10–25% of groundwater used for drinking water supply in the US contains nitrate concentrations over the maximum allowable contaminant level of 10 mg-N/L (Lee and Rittmann, 2002). Such elevated nitrate concentrations in groundwater can cause serious human health problems (i.e., methaemoglobinaemia and cancer) (Peng et al., 2016a) and ecological disturbances (i.e., eutrophication of water bodies) (Roy and Bickerton, 2010). Traditional nitrate treatment approaches include ion exchange, reverse osmosis, electrodialysis and distillation, which are expensive in terms of operation cost and also not suitable for in-situ applications (Peng et al., 2015). For this reason, biological denitrification (heterotrophic and autotrophic) is considered as an alternative process. Typically, groundwater contaminated with nitrate contains no organic matter, therefore sulfur-driven autotrophic denitrification has attracted more attentions compared to heterotrophic denitrification which requires massive external organic carbon (Chen et al., 2015a; Di Capua et al., 2015; Koenig and Liu, 2001; Liu et al., 2009; Sahinkaya et al., 2014; Sahinkaya et al., 2015; Sierra-Alvarez et al., 2007). The main disadvantage of sulfur-driven autotrophic denitrification process is the inevitable generation of a large amount of sulfate. Theoretically, ca. 33 mg-N/L nitrate in groundwater can be removed via sulfur-driven autotrophic denitrification, without exceeding the US allowable sulfate limit in drinking water (i.e., 250 mg/L) (Oh et al., 2001). Therefore, for treating groundwater with high nitrate concentrations, a more effective strategy to control sulfate formation is to combine autotrophic and heterotrophic denitrification processes (Sahinkaya et al., 2011). A number of studies have been carried out on this promising mixotrophic denitrification process, focusing on the selection of carbon source, effects of C/N ratio, kinetics analysis, microbial community structure, reactor types and other operation parameters (Chen et al., 2014; Chen et al., 2009; Garcia-de-Lomas et al., 2007; Liu et al., 2016a; Mora et al., 2015; Qian et al., 2015; Sahinkaya and Dursun, 2012; Sahinkaya and Kilic, 2014; Xu et al., 2015a; Xu et al., 2015 b; Xu et al., 2013; Xu et al., 2014; Yu et al., 2016). However, recent studies have revealed that nitrous oxide (N2O) can be produced and accumulated as a significant intermediate product (i.e., maximum accumulation concentration of ca. 24% of the nitrogen load) during this mixotrophic denitrification process with both sulfur and organic matter as electron donors (Zhang et al., 2015a), which has raised increasing concerns owing to its potent greenhouse gas effect and its ability to deplete stratospheric ozone (Chen et al., 2012; Ravishankara et al., 2009). For example, it has been showed that 1% increase in N2O emission would induce a 30% increase in the carbon footprint during the wastewater treatment due to the high global warming potential of N2O (Law et al., 2012). Therefore, understanding and reducing N2O production during this mixotrophic denitrification process is of great importance to optimize the application of such system. Mathematical modeling is important toward a full understanding of biological nitrogen removal process, which has been applied to predict and mitigate N2O emission during wastewater treatment processes (Liu et al., 2015b, 2015c; Ni et al., 2011; Peng et al., 2016b). However, little effort has been dedicated to modeling the N2O dynamics during mixotrophic denitrification process for high-concentration nitrate removal despite of considerable N2O production. Under such circumstances, the competition for nitrogen compound between heterotrophic and autotrophic denitrifiers could induce a different scenario on N2O emission, which is not accounted for in previous studies. In addition, the previous process optimizations only focus on nitrate removal and sulfate control in mixotrophic denitrification, with the N2O emission being not considered (Sahinkaya et al., 2011).

415

This study aims to propose a model for the prediction of N2O production during sulfur-driven mixotrophic denitrification and to optimize the process performance taking N2O emission into consideration. The model is validated using experimental data under two different conditions. The validated model is applied to investigate the optimal conditions for achieving high level of nitrate removal with relative low N2O emission and sulfate production during mixotrophic denitrification, which is expected to provide first insights into the improvement of the design and operation of such systems for future applications. 2. Materials and methods 2.1. Model development In mixotrophic denitrification process, autotrophic and heterotrophic denitrifiers use their corresponding electron donors (sulfur and organic matter, respectively) to reduce nitrate to nitrogen gas. The biological reactions in this model were based on Liu et al. (2016c) and Schulthess and Gujer (1996), considering the threestep denitrification (NO3 to N2 via NO2 and N2O) driven by both sulfur and organic matter to describe all potential N2O accumulation steps (Fig. 1). Nitric oxide (NO) is not taken into account in our model since the NO reduction related parameters are beyond the ability of measurement. Indeed, NO reduction is usually prioritized by bacteria to avoid its toxicity and thus ensure no accumulation of NO as intermediate (Liu et al., 2015c). Also, thiosulfate and sulfite are not considered in the biological model since they are the intermediate products of chemical oxidation rather than biological oxidation (Liu et al., 2015a; Xu et al., 2016). The model describes the relationships among seven compounds involved in autotrophic denitrifiers (XSOB) and heterotrophic denitrifiers (XHB), namely NO3 (SNO3), NO2 (SNO2), N2O (SN2O), N2 (SN2), S0 (SS0), SO24 (SSO4) and soluble chemical oxygen demand (SCOD). The units are g-N m 3 for all nitrogenous species, g-S m 3 for sulfurous species and g-COD m 3 for other compounds (Table S1 in SI). Two groups of biological processes (Table S2 and S3 in SI) were considered, namely, sulfur-driven autotrophic denitrification processes (Process 1–3) and heterotrophic denitrification processes (Process 5–7), each modeled as three sequential denitrification processes from NO3 to N2 via NO2 and N2O with individual reaction-specific Monod-type kinetics. In addition, biomass decay of autotrophic denitrifiers (Process 4) and heterotrophic denitrifiers (Process 8) was also included. All model parameter values were obtained from literature, as presented in Table S4 in SI (including definitions, values, units, and sources). 2.2. Model evaluation with experimental data Experimental data previously reported by Zhang et al. (2015a) was used to evaluate the proposed mixotrophic denitrification model. An enriched autotrophic denitrifying culture was employed as the inoculum and further developed in two 2.3-L lab-scale continuous-flow anaerobic fluidized bed membrane bioreactors. Initially, 200 g of sulfur were added to each reactor as electron

Fig. 1. Schematic representation of the proposed N2O model concept in mixotrophic denitrification processes.

416

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

donor for sulfur-based autotrophic denitrification, and additional 50–100 g of sulfur was supplemented every 2-month, to ensure a stable sulfur concentration. The reactor was fed with N2-sparged synthetic wastewater mainly consisting of KNO3, resulting in an influent nitrate concentration from 25 to 80 mg-N/L. From day 31, methanol as external carbon source was added to one reactor (R1) and ethanol to the other reactor (R2), which were supplemented according to ca. 40–200% of theoretical requirements of heterotrophic denitrification. More details of the reactor operation and performance can be found in Zhang et al. (2015a). Batch experiments with an initial 15N–NO3 concentrations of 50 mg-N/L as a sole nitrogen substrate, were conducted with the denitrifying culture and elemental sulfur from bioreactor R1 (in the presence of methanol) and bioreactor R2 (in the presence of ethanol), respectively, in a 0.5-L glass serum flask supplemented with 0.2 L medium free of NH+4. The headspace was flushed with helium to exclude oxygen and background nitrogen. Mixed liquor and gas samples were taken periodically for NO3 , NO2 , N2O, N2 and sulfate analysis, respectively. More detailed batch experimental setup and analysis methods can be found in Zhang et al. (2015a). The batch experimental data sets of NO3 , NO2 , N2O and SO24 from both mixotrophic denitrifying cultures above (Fig. 3 in Zhang et al. (2015a)) were then used to verify the validity and applicability of the model.

ent nitrate concentration of 50 mg N/L and a SRT of 50 d. Scenario 3 simulates the impact of influent nitrate concentrations on process performance and N2O emission at steady state, with the influent nitrate concentration varying from 30 to 80 mg-N/L, given an influent COD concentration of 150 mg/L and a SRT of 50 d. Scenario 4 assesses the combined impact of COD (0–500 mg/L) and nitrate (30–80 mg-N/L) concentrations (i.e., typical ranges in experimental studies (Zhang et al., 2015a, 2015b)) on the steady-state process performance and N2O emission at a SRT of 50 d in order to acquire the optimal operating window for sulfur-driven mixotrophic denitrification process. The selection of SRT as well as the concentrations of influent nitrate and COD was based on the typical reactor operation conditions reported in literature (Liu et al., 2016c; Yang et al., 2016; Zhang et al., 2015a). For each scenario, the influent sulfur is set to be 1000 mg-S/L, as sulfur is usually not a limiting factor in practical application where sulfur concentration reached over 1000 mg-S/L (Zhang et al. 2015a). Therefore, no sulfur limitation would occur during the simulations. In this way, we can thus analyze other parameters affecting the optimizations of sulfur driven mixotrophic denitrification without interference of sulfur availability 3. Results and discussion 3.1. Model evaluation with experimental data

2.3. Optimizing the mixotrophic denitrification After verification of the model, further simulations were performed to evaluate the impacts of operational conditions, i.e., sludge retention time (SRT), COD and nitrate concentrations, on steady-state process efficiency and N2O emission during sulfurdriven mixotrophic denitrification, employing the software AQUASIM 2.1d (Reichert, 1998). The simulated continuous-flow sludge system in this work has a working volume of 2 L and a hydraulic retention time (HRT) of 2 h, mimicking a typical experimental condition (Sierra-Alvarez et al., 2007; Zhang et al., 2015a, 2015b). Four different scenarios are considered to optimize the mixotrophic denitrifying performance in this work, as detailed in Table 1. Scenario 1 tests the effect of SRT on process performance and N2O emission for the mixotrophic denitrifying system at steady state. The operational SRT for simulations is chosen from 25 to 200 d, with the applied influent nitrate and COD concentrations at 50 mg N/L and 150 mg/L, respectively. Scenario 2 examines the effect of influent COD concentrations on process performance and N2O emission at steady state, with the variations of influent COD concentrations from 0 to 250 mg/L (Zhang et al., 2015a), given an influ-

The developed model was firstly used to predict NO3 , NO2 , N2O and SO24 dynamics in two data sets from batch tests of sulfurdriven mixotrophic denitrifying cultures, by employing model parameter values previously reported in literature (see Table S4). The predicted NO3 , NO2 , N2O and SO24 profiles with the established model are shown in Fig. 2, along with the experimental results. At the beginning ca. 7 h of the batch test for sulfur-driven methanol-feeding mixotrophic denitrifying culture with an initial COD concentration of 180 mg/L, both nitrite and N2O gradually accumulated along with the mixotrophic nitrate reduction (Fig. 2a). After nitrate depletion, nitrite was used up within ca. 9 h. Afterwards, N2O was the only electron acceptor for anoxic sulfur and organic oxidation, and consumed up within ca. 13 h. Also, higher sulfate production rates were observed in the first 9 h while lower rates afterwards (Fig. 2b), coincident with the lower sulfur conversion rate by N2O (0.010 h 1) as compared to nitrate (0.020 h 1) and nitrite (0.035 h 1). The model captured these trends reasonably well (R2NO3 = 0.98, R2NO2 = 0.98, R2N2O = 0.94, R2SO4 = 0.91).

Table 1 An overview of the scenarios for the model-based assessments. Scenarios Scenario 1 Effect of SRT on N2O emission and Process efficiency during sulfur-driven mixotrophic denitrification Scenario 2 Effect of influent COD concentrations on N2O emission and process efficiency during sulfur-driven mixotrophic denitrification Scenario 3 Effect of influent nitrate concentrations on N2O emission and process efficiency during sulfur-driven mixotrophic denitrification Scenario 4 Combined effects of influent COD and nitrate on N2O emission and process efficiency during sulfur-driven mixotrophic denitrification

Simulation conditions

Variable conditions

SNO3 = 50 mg-N/L SCOD = 150 mg/L SS0 = 1000 mg-S/L

SRT = 25–200 d

SNO3 = 50 mg-N/L SRT = 50 d SS0 = 1000 mg-S/L

SCOD = 0–250 mg/L

SCOD = 150 mg/L SRT = 50 d SS0 = 1000 mg-S/L

SNO3 = 30–80 mg-N/L

SRT = 50 d SS0 = 1000 mg-S/L

SCOD = 0–500 mg/L SNO3 = 30–80 mg-N/L

417

Concentration (mg-N/L)

60

a

measured N2O modelled N2O

50

measured NO3modelled NO3measured NO2-

40 30 20

modelled NO2-

10 0 0

5

10

15

20

Sulfate concentration (mg/L)

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

1000

b

modelled SO42measured SO42-

800 600 400 200 0 0

25

5

b

measured N2O modelled N2O

50

measured NO3modelled NO3measured NO2-

40 30 20

modelled NO2-

10 0 0

5

10

15

20

25

Sulfate concentration (mg/L)

Concentration (mg-N/L)

60

10

15

20

25

Time (h)

Time (h) 1000

d

modelled SO42measured SO42-

800 600 400 200 0 0

5

10

15

20

25

Time (h)

Time (h) SO24

Fig. 2. Model evaluation with experimental data on (a) NO3 , NO2 , N2O and (b) from batch test of sulfur-driven methanol-feeding mixotrophic denitrifying culture with an initial nitrate and COD concentrations of 50 mg-N/L and 180 mg/L, respectively; and model evaluation with experimental data on (c) NO3 , NO2 , N2O and (d) SO24 from batch test of sulfur-driven ethanol-feeding mixotrophic denitrifying culture with an initial nitrate and COD concentrations of 50 mg-N/L and 209 mg/L, respectively.

The developed model was further tested for their ability to predict experimental data of NO3 , NO2 , N2O and SO24 from the batch test of sulfur-driven ethanol-feeding mixotrophic denitrifying culture with an initial COD concentration of 209 mg/L. The model predictions and the experimental results are shown in Fig. 2c and d. NO3 , NO2 and N2O dynamics was similar to those of the methanol-feeding culture batch test with an initial COD concentration of 180 mg/L (Fig. 2c and a), except for the higher NAspecies consumption rates (i.e., NO3 , NO2 and N2O depletion at ca. 5 h, 6 h and 7 h, respectively). This could be attributed to the higher heterotrophic denitrification activity of ethanol-feeding culture, due to a higher initial organic content. Heterotrophic denitrification rates by NO3 , NO2 and N2O (0.053, 0.056 and 0.134 h 1) are much quicker than those of sulfur-driven autotrophic denitrification rates (0.020, 0.035 and 0.010 h 1), thus resulting in faster NO3 consumption as well as lower NO2 and N2O accumulation. Consistently with the higher heterotrophic denitrification activity in this case, a much lower sulfate accumulation (Fig. 2d) was observed as compared to Fig. 2b. The good agreement between these simulated and measured data supported that the developed model properly captures the relationships among N2O dynamics, nitrogen reduction and sulfur oxidation (R2NO3 = 0.99, R2NO2 = 0.92, R2N2O = 0.95, R2SO4 = 0.91). The long sampling interval (ca. 2 h per sample) during sulfur-based mixotrophic denitrification might decrease the data resolution and lead to slight mismatches between model prediction and experimental data. Overall, the validation results showed that the model predictions well matched the measured data of concentrations in the experiment by employing literature reported values, which supports the validity of the developed model. Model validation results indicated that a higher amount of N2O could accumulate at a lower initial COD concentration during sulfur-driven mixotrophic denitrification process, due to the relatively lower autotrophic N2O reduction rate (0.010 h 1) as compared to the heterotrophic N2O reduction rate (0.134 h 1), as the fact that heterotrophic and autotrophic denitrifying cultures utilize completely different substrates

as electron donors (carbon and sulfide/sulfur, respectively) for growth, thus resulting in distinct denitrifying microbial community with different growth kinetics on nitrogen. This is different from Yang et al. (2016), who reported that the N2O emission during sulfide-driven (with sulfide as electron donor rather than sulfur) autotrophic denitrification process was lower than literature value of heterotrophc denitrification, but did not compare N2O emission with a reference heterotrophic denitrification reactor operated under similar conditions. Therefore, for simultaneous heterotrophic and autotrophic denitrification treating highconcentration nitrate contamination in the groundwater, carbon source addition not only alleviates the excess sulfate production (Sahinkaya et al., 2011; Zhang et al., 2015a), but also the N2O accumulation and emission. 3.2. Impact of SRT on N2O emission and process performance In activated sludge systems, SRT is an important process parameter determining the process performance and N2O emission of both autotrophic and heterotrophic denitrification processes (Liu et al., 2015c). To reveal the detailed role of SRT on sulfur-driven mixotrophic denitrification, simulations (Scenario 1 in Table 1) were conducted (Fig. 3). The influent nitrate and COD were fixed at 50 mg-N/L and 150 mg/L, typical high-concentration nitrate contaminated groundwater requiring external carbon source addition (Sahinkaya et al., 2011). With the increase of SRT from 25 to 80 d, the sulfur-driven mixotrophic denitrification system achieves a significant N2O mitigation and N2O emission factor decreases substantially from 3.54% to 0.78% (Fig. 3a), resulting in a substantial carbon footprint drop (Law et al., 2012). Further increasing SRT from 80 to 200 d leads to a slow decline of N2O emission factor from 0.78% to 0.32% as compared to above. Meanwhile, both nitrate and total nitrogen removal (i.e., both nitrate and nitrite) efficiencies of the system could maintain >98% and >97%, respectively (Fig. S1), when SRT ranges from 30 to 200 d. A lower SRT (i.e., 25 d) could not retain

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

a Sulfate N2O emission

300 250 200 150 100 50

50

75

100 125 SRT (d)

150

HB

80 60 40 20 0

0 200

175

SOB

b

100

350

Active biomass fraction (%)

10 9 8 7 6 5 4 3 2 1 0 25

Effluent sulfate (mg/L)

N2O emission per N load (%)

418

25

50

75

100 125 SRT (d)

150

175

200

Fig. 3. Model simulation results of the impact of SRT on (a) N2O emission and sulfate production, and (b) microbial abundance during sulfur-driven mixotrophic denitrification. The applied influent nitrate and COD concentrations are 50 mg N/L and 150 mg/L, respectively.

thus affect process efficiency and N2O emission. Therefore, simulations (Scenario 2 in Table 1) were conducted with varying influent COD concentrations of 0–250 mg/L (Fig. 4). The influent nitrate concentration and SRT were kept at 50 mg-N/L and 50 d, respectively. As expected, increasing influent COD concentrations from 0 to 200 mg/L leads to a substantial decrease in N2O emission (ca. 6– 0.04%, Fig. 4a) and sulfate production (ca. 307 to 228 mg/L) while the system maintains >99% total nitrogen removal efficiency, due to the proliferation of heterotrophic denitrifiers (XH) (Fig. 4b). The heterotrophic N2O reduction rate (0.134 h 1) is approximately two order of magnitude higher than that of autotrophic N2O reduction rate (0.010 h 1), and also heterotrophic nitrate/nitrite reduction rates (0.053 and 0.056 h 1) are about half order of magnitude higher than those of autotrophic rates (0.020 and 0.035 h 1), thus resulting in decreasing N2O emission and sulfate production. Afterwards, N2O emission keeps relatively constant low level (i.e., below 0.04%) with further increasing the influent COD from 200 to 250 mg/L. It should be noted that our simulation results are in agreement with the previous studies that higher organic concentration causes lower N2O accumulation and sulfate production (Sahinkaya et al., 2011; Zhang et al., 2015a).

3.3. Impact of influent COD concentrations on N2O emission and process performance

3.4. Impact of influent nitrate concentrations on N2O emission and process performance

Regarding a sulfur-driven mixotrophic denitrifying system treating high-concentration nitrate contaminated groundwater, the amount of external carbon source dosed is also an important factor (Zhang et al., 2015a), as it could regulate the individual contributions of autotrophic and heterotrophic denitrification and

As both autotrophic and heterotrophic denitrification processes are also regulated by the nitrate availability, simulations (Scenario 3 in Table 1) were also conducted with varying influent nitrate concentrations of 30–80 mg/L (Fig. 5) to evaluate process efficiency and N2O emission in a sulfur-driven mixotrophic denitrifying

a

350

100

Sulfate N2O emission

300 250 200 150 100 50

0

50

100 150 COD (mg/L)

200

0 250

Active biomass fraction (%)

10 9 8 7 6 5 4 3 2 1 0

Effluent sulfate (mg/L)

N2O emission per N load (%)

biomass well in the system, thus resulting in the lower nitrate and total nitrogen removal (i.e., 92% and 85%, respectively). Consistently, the relative fraction of autotrophic denitrifiers (XSOB) increases with the increase of SRT (Fig. 3b). In addition, the sulfate production only increases from 219 mg/L at a SRT of 25 d to 255 mg/L at a SRT of 200 d, almost within the US EPA allowable sulfate limit (i.e., 250 mg/L) in drinking water (Oh et al., 2001). Overall, the model simulations show that employing a longer SRT could effectively reduce N2O emission factor while achieving a high-level nitrogen removal efficiency. For example, increasing SRT from 25 d to 50 d helps to lower N2O emission factor by ca. 2% (Fig. 3), likely due to the better retention of the active biomass under the applied shortened HRT of 2 h. Further increasing SRT from 50 d to 200 d leads to ca. 1% decrease in N2O emission factor. Together with the fact that an extended SRT over 50 d may not be realistic for activated sludge system due to the massive excessive sludge (both heterotrophic and autotrophic biomass) production (Tchobanoglous et al., 2003), a designed SRT of 50 d for sulfurdriven mixotrophic denitrifying system would be suitable for process operation and N2O emission control purposes. Therefore, a SRT of 50 d was also selected for the subsequent simulations in terms of influent COD and nitrate concentrations.

b

SOB

HB

80 60 40 20 0

0

50

100 150 COD (mg/L)

200

250

Fig. 4. Model simulation results of the impact of influent COD concentrations on (a) N2O emission and sulfate production, and (b) microbial abundance during sulfur-driven mixotrophic denitrification. The applied influent nitrate concentration and SRT of the system are 50 mg N/L and 50 d, respectively.

419

a

500

100

Sulfate N2O emission

400 300 200 100

40

50 60 Nitrate (mg-N/L)

70

0 80

Active biomass fraction (%)

10 9 8 7 6 5 4 3 2 1 0 30

Effluent sulfate (mg/L)

N2O emission per N load (%)

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

b

SOB

HB

80 60 40 20 0

30

40

50 60 Nitrate (mg-N/L)

70

80

Fig. 5. Model simulation results of the impact of influent nitrate concentrations on (a) N2O emission and sulfate production, and (b) microbial abundance during sulfur-driven mixotrophic denitrification. The applied influent COD concentration and SRT of the system are 150 mg/L and 50 d, respectively.

system. The influent COD concentration and SRT were set at 150 mg/L and 50 d, respectively. The trend of N2O emission increase is not obvious with increasing influent nitrate concentrations from 30 to 35 mg-N/L, i.e., from ca. 0.01% to 0.02% (Fig. 5a). Further increasing influent nitrate concentrations from 35 to 65 mg-N/L results in a substantial increase in N2O emission (from ca. 0.02% to 2% Fig. 5a), which afterwards plateaus (ca. 2%) at an influent nitrate concentration between 65 and 80 mg-N/L. Meantime, the system maintains >99% and >97% total nitrogen and COD removal efficiency during the course. In comparison, sulfate production showed a continuous increase from ca. 118 to 436 mg/L with increasing influent nitrate concentrations from 30 to 80 mg-N/L, due to the proliferation of autotrophic denitrifiers (XSOB) at an increasing N/COD ratio (Fig. 5b), coincident with the N2O emission profile. This is also in agreement with Zhang et al. (2015a), who reported that the percentage of the dominant sulfur-oxidizing autotrophic denitrifiers i.e., Thiobacillus and Sulfurimonas, increased by 2–4 times with increasing influent N/COD ratio. 3.5. Optimal operational conditions for sulfur-driven mixotrophic denitrification The above model predictions present a strong dependency of N2O emission and process performance of sulfur-driven mixotrophic denitrifying system on influent COD concentrations and influent nitrate concentrations, therefore, the joint effects (Scenario 4 in Table 1) of influent nitrate (30–80 mg-N/L) and COD (0–500 mg/L) concentrations on the steady-state performance of sulfur-driven mixotrophic denitrifying system are explored in Fig. 6. It should be noted that the total nitrogen removal efficiency is over 99% within all these operational conditions (data not shown). Fig. 6a illustrates the N2O emission from steady-state mixotrophic denitrifying system under the simulation conditions of Scenario 4. Higher N2O emission (i.e., emission factor >3%) from the system is observed when influent COD concentrations are below ca. 100 mg/L (Fig. 6a), due to the proliferation of autotrophic denitrifiers (XSOB) which contribute significantly to N2O accumulation (Li et al., 2017) during denitrification (Fig. 6d). In contrast, the region for achieving lower N2O emission (i.e., emission factor <1%) is located in the white dash line covered trapezoid-shaped area (Fig. 6a) where the corresponding higher influent COD concentrations (i.e., higher than ca. 120–250 mg/L) are applied. Fig. 6b illustrates the sulfate production from steady-state mixotrophic denitrifying system under the simulation conditions of Scenario 4. In the absence of carbon source addition (COD = 0 mg/L), sulfate production from autotrophic denitrification is within

the US EPA allowable sulfate limit (i.e., 250 mg/L) in drinking water (Oh et al., 2001) when influent nitrate concentration is below ca. 40 mg-N/L. Further increase in the influent nitrate concentration leads to the effluent sulfate over 250 mg/L, when carbon source dosing becomes necessary, i.e., 500 mg-COD/L is required at an influent nitrate concentration of 80 mg-N/L. The white dash line covered trapezoid-shaped area indicates the region where sulfate production meets the US EPA allowable sulfate limit. Fig. 6c illustrates COD removal efficiency of steady-state mixotrophic denitrifying system under the simulation conditions of Scenario 4. The region for achieving high COD removal (i.e., >95%) is located in the white dash line covered dark red area (Fig. 6c). Overdose of carbon source in a large amount will not only result in decline in the COD removal efficiency (i.e., <60% in blue zone in Fig. 6c) and effluent COD contamination due to the incomplete COD utilization by heterotrophic denitrifiers (XH), but also extra operational cost. Therefore, the carbon source should be wisely added. Overall, from the perspective of system operation for achieving low N2O emission (i.e., <1%), appropriate sulfate production (i.e., <250 mg/L) and COD removal efficiency (i.e., >95%) in a sulfurdriven mixotrophic denitrifying system, both influent nitrate and COD concentrations should be well controlled in the highlighted optimal ridge-shape area (white dash line covered zone in Fig. 6d), especially in the bottom line of this zone considering avoiding of extra operational cost of carbon source addition, in order to achieve high-level total nitrogen removal, N2O emission mitigation, excess sulfate control and maximized COD utilization.. Also, it should be noted that the slope of the bottom line is higher when nitrate is over ca. 60 mg-N/L due to sulfate control reason; while it is much lower when nitrate is below ca. 60 mg-N/L, thus resulting in lower operational cost on carbon source dosing in this range. Under these optimal conditions in terms of influent nitrate and COD concentrations, autotrophic denitrifiers (XSOB) and heterotrophic denitrifiers (XH) can coexist well and interact properly (Fig. 6d). In addition, an accidental oversupply of carbon source over the bottom line but still in the optimal zone would not lead to system deterioration. Also, as the system might encounter a dynamic influent loading with variant nitrate concentration (high or low), a slight overdose of carbon source would ensure the stability of the system. Therefore, the information of this work would be very useful for appropriate carbon source addition and effective N2O mitigation in sulfur-driven mixotrophic denitrifying system. 3.6. Implications of this work Sulfur-driven mixotrophic denitrification process is a promising technology for high-level nitrate removal from contaminated

420

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

a

c

b

d

Fig. 6. Model simulation results of the dependency of the (a) N2O emission per N load, (b) sulfate production, (c) COD removal efficiency and (d) relative fraction of autotrophic denitrifiers (XSOB) on the simultaneous variations of influent nitrate and COD concentrations. The color scale represents% in (a), (c) and (d), and mg/L in (b). The applied SRT is 50 d.

groundwater (Liu et al., 2009; Sierra-Alvarez et al., 2007; Zhang et al., 2015b). However, recent studies have demonstrated substantial N2O accumulation during this process (Liu et al., 2015d; Zhang et al., 2015a). Also, inappropriate amount of carbon source addition may cause the process failure when effluent sulfate and COD concentrations do not meet discharge standard (Oh et al., 2001). Mathematical modeling has been recognized as a powerful tool to understand complex wastewater treatment systems and to support system operation optimization (Chen et al., 2015 b; Liu et al., 2016 b; Liu et al., 2015e; Wang et al., 2017), thus potentially being valuable for N2O mitigation, sulfate production control and carbon material saving purposes in sulfur-driven mixotrophic denitrification systems. In this work, a mathematical model is developed to describe N2O production and process performance of sulfur-driven mixotrophic denitrification process, through integrating the activated sludge model (ASM) -based heterotrophic denitrifying N2O model (Ni et al., 2011) and sulfur-driven autotrophic denitrifying N2O model (Liu et al., 2016c). The integrated model was validated by two independent data sets (Zhang et al., 2015a), by employing model parameter values previously reported in literature (Table S4). The model was robust in its ability to predict nitrate, nitrite, N2O, and sulfate dynamics under different initial organic substrate conditions, indicating the applicability of the developed model. The findings in this work are useful to design and optimize sulfur-driven mixotrophic denitrification process in terms of N2O mitigation, excess sulfate control and carbon source utilization.

Model simulation results indicate that SRT and influent COD/N ratio play an important role in regulating N2O emission from sulfur-driven mixotrophic denitrification process. We believe the preliminary results obtained from this study will support our understanding on future optimizations of sulfur-driven mixotrophic denitrifying system.

4. Conclusions In this work, a mathematical model is proposed to evaluate N2O emission and excess sulfate production in sulfur-driven mixotrophic denitrification process, through integration of autotrophic and heterotrophic denitrification N2O models. The developed model was successfully validated from experimental data obtained from two mixotrophic denitrification systems with different conditions, by employing model parameters previously reported in literature. Modeling results demonstrated substantial N2O accumulation in sulfur-driven mixotrophic denitrifying system due to the relatively low autotrophic N2O reduction rate. Increasing SRT and influent COD/N ratio would substantially reduce N2O emission. High-level total nitrogen removal with significant N2O mitigation, appropriate excess sulfate control and maximized COD utilization can be achieved simultaneously through controlling influent nitrate and COD concentrations, supporting the effective N2O mitigation and process optimization in sulfur-driven mixotrophic denitrifying system.

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

Acknowledgements This work was partially supported by the Recruitment Program of Global Experts and the Natural Science Foundation of China (No. 51578391). Dr. Yiwen Liu acknowledges the support from the UTS Chancellor’s Postdoctoral Research Fellowship. The authors are grateful to the research collaboration among University of Technology Sydney, University of Antwerp, Hunan University and Tongji University.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ces.2017.07.005.

References Chen, C., Liu, L., Lee, D.-J., Guo, W., Wang, A., Xu, X., Zhou, X., Wu, D., Ren, N., 2014. Integrated simultaneous desulfurization and denitrification (ISDD) process at various COD/sulfate ratios. Biores. Technol. 155, 161–169. Chen, C., Wang, A., Ren, N., Lee, D.-J., Lai, J.-Y., 2009. High-rate denitrifying sulfide removal process in expanded granular sludge bed reactor. Biores. Technol. 100 (7), 2316–2319. Chen, H., Liu, Y., Ni, B.-J., Wang, Q., Wang, D., Zhang, C., Li, X., Zeng, G., 2016. Fullscale evaluation of aerobic/extended-idle regime inducing biological phosphorus removal and its integration with intermittent sand filter to treat domestic sewage discharged from highway rest area. Biochem. Eng. J. 113, 114– 122. Chen, H.-B., Wang, D.-B., Li, X.-M., Yang, Q., Zeng, G.-M., 2015a. Enhancement of post-anoxic denitrification for biological nutrient removal: effect of different carbon sources. Environ. Sci. Pollut. Res. 22 (8), 5887–5894. Chen, X., Guo, J., Xie, G.-J., Liu, Y., Yuan, Z., Ni, B.-J., 2015b. A new approach to simultaneous ammonium and dissolved methane removal from anaerobic digestion liquor: a model-based investigation of feasibility. Water Res. 85, 295– 303. Chen, Y., Wang, D., Zhu, X., Zheng, X., Feng, L., 2012. Long-term effects of copper nanoparticles on wastewater biological nutrient removal and N2O generation in the activated sludge process. Environ. Sci. Technol. 46 (22), 12452–12458. Di Capua, F., Papirio, S., Lens, P.N.L., Esposito, G., 2015. Chemolithotrophic denitrification in biofilm reactors. Chem. Eng. J. 280, 643–657. Garcia-de-Lomas, J., Corzo, A., Portillo, M.C., Gonzalez, J.M., Andrades, J.A., SaizJimenez, C., Garcia-Robledo, E., 2007. Nitrate stimulation of indigenous nitratereducing, sulfide-oxidising bacterial community in wastewater anaerobic biofilms. Water Res. 41 (14), 3121–3131. Koenig, A., Liu, L.H., 2001. Kinetic model of autotrophic denitrification in sulphur packed-bed reactors. Water Res. 35 (8), 1969–1978. Law, Y., Ye, L., Pan, Y., Yuan, Z., 2012. Nitrous oxide emissions from wastewater treatment processes. Philos. Trans. R. Soc. B 367 (1593), 1265–1277. Lee, K.-C., Rittmann, B.E., 2002. Applying a novel autohydrogenotrophic hollowfiber membrane biofilm reactor for denitrification of drinking water. Water Res. 36 (8), 2040–2052. Li, P., Wang, Y., Zuo, J., Wang, R., Zhao, J., Du, Y., 2017. Nitrogen removal and N2O accumulation during hydrogenotrophic denitrification: influence of environmental factors and microbial community characteristics. Environ. Sci. Technol. 51 (2), 870–879. Liu, C., Han, K., Lee, D.-J., Wang, Q., 2016a. Simultaneous biological removal of phenol, sulfide, and nitrate using expanded granular sludge bed reactor. Appl. Microbiol. Biotechnol. 100 (9), 4211–4217. Liu, H., Jiang, W., Wan, D., Qu, J., 2009. Study of a combined heterotrophic and sulfur autotrophic denitrification technology for removal of nitrate in water. J. Hazard. Mater. 169 (1), 23–28. Liu, Y., Ngo, H.H., Guo, W., Peng, L., Pan, Y., Guo, J., Chen, X., Ni, B.-J., 2016b. Autotrophic nitrogen removal in membrane-aerated biofilms: Archaeal ammonia oxidation versus bacterial ammonia oxidation. Chem. Eng. J. 302, 535–544. Liu, Y., Ni, B.-J., Ganigué, R., Werner, U., Sharma, K.R., Yuan, Z., 2015a. Sulfide and methane production in sewer sediments. Water Res. 70, 350–359. Liu, Y., Peng, L., Chen, X., Ni, B.-J., 2015b. Mathematical modeling of nitrous oxide production during denitrifying phosphorus removal process. Environ. Sci. Technol. 49 (14), 8595–8601. Liu, Y., Peng, L., Guo, J., Chen, X., Yuan, Z., Ni, B.-J., 2015c. Evaluating the role of microbial internal storage turnover on nitrous oxide accumulation during denitrification. Sci. Rep. 5, 15138. Liu, Y., Peng, L., Ngo, H.H., Guo, W., Wang, D., Pan, Y., Sun, J., Ni, B.J., 2016c. Evaluation of nitrous oxide emission from sulfide- and sulfur-based autotrophic denitrification processes. Environ. Sci. Technol. http://dx.doi.org/10.1021/acs. est.1026b02202.

421

Liu, Y., Sharma, K.R., Ni, B.-J., Fan, L., Murthy, S., Tyson, G.Q., Yuan, Z., u et al., 2015 d. Effects of nitrate dosing on sulfidogenic and methanogenic activities in sewer sediment. Water Res. 74, 155–165. Liu, Y., Zhang, Y., Ni, B.-J., 2015e. Zero valent iron simultaneously enhances methane production and sulfate reduction in anaerobic granular sludge reactors. Water Res. 75, 292–300. Mora, M., Fernández, M., Gómez, J.M., Cantero, D., Lafuente, J., Gamisans, X., Gabriel, D., 2015. Kinetic and stoichiometric characterization of anoxic sulfide oxidation by SO-NR mixed cultures from anoxic biotrickling filters. Appl. Microbiol. Biotechnol. 99 (1), 77–87. Ni, B.-J., Ruscalleda, M., Pellicer-Nacher, C., Smets, B.F., 2011. Modeling nitrous oxide production during biological nitrogen removal via nitrification and denitrification: extensions to the general ASM models. Environ. Sci. Technol. 45 (18), 7768–7776. Oh, S.E., Yoo, Y.B., Young, J.C., Kim, I.S., 2001. Effect of organics on sulfur-utilizing autotrophic denitrification under mixotrophic conditions. J. Biotechnol. 92 (1), 1–8. Peng, L., Liu, Y., Gao, S.-H., Chen, X., Ni, B.-J., 2016a. Evaluating simultaneous chromate and nitrate reduction during microbial denitrification processes. Water Res. 89, 1–8. Peng, L., Liu, Y., Gao, S.H., Chen, X., Xin, P., Dai, X., Ni, B.J., 2015. Evaluation on the nanoscale zero valent iron based microbial denitrification for nitrate removal from groundwater. Sci. Rep. 5, 12331. Peng, L., Liu, Y., Ni, B.-J., ng et al., 2016 b. Nitrous oxide production in completely autotrophic nitrogen removal biofilm process: a simulation study. Chem. Eng. J. 287, 217–224. Qian, J., Lu, H., Jiang, F., Ekama, G.A., Chen, G.-H., 2015. Beneficial co-treatment of simple wet flue gas desulphurization wastes with freshwater sewage through development of mixed denitrification–SANI process. Chem. Eng. J. 262, 109– 118. Ravishankara, A.R., Daniel, J.S., Portmann, R.W., 2009. Nitrous oxide (N2O): the dominant ozone-depleting substance emitted in the 21st century. Science 326 (5949), 123–125. Reichert, P., 1998. AQUASIM 2.0—user manual. Swiss Federal Institute for Environmental Science and Technology. Dubendorf, Switzerland. Rivett, M.O., Buss, S.R., Morgan, P., Smith, J.W.N., Bemment, C.D., 2008. Nitrate attenuation in groundwater: a review of biogeochemical controlling processes. Water Res. 42 (16), 4215–4232. Roy, J.W., Bickerton, G., 2010. Proactive screening approach for detecting groundwater contaminants along urban streams at the reach-scale. Environ. Sci. Technol. 44 (16), 6088–6094. Sahinkaya, E., Dursun, N., 2012. Sulfur-oxidizing autotrophic and mixotrophic denitrification processes for drinking water treatment: elimination of excess sulfate production and alkalinity requirement. Chemosphere 89 (2), 144– 149. Sahinkaya, E., Dursun, N., Kilic, A., Demirel, S., Uyanik, S., Cinar, O., 2011. Simultaneous heterotrophic and sulfur-oxidizing autotrophic denitrification process for drinking water treatment: control of sulfate production. Water Res. 45 (20), 6661–6667. Sahinkaya, E., Kilic, A., 2014. Heterotrophic and elemental-sulfur-based autotrophic denitrification processes for simultaneous nitrate and Cr (VI) reduction. Water Res. 50, 278–286. Sahinkaya, E., Kilic, A., Duygulu, B., 2014. Pilot and full scale applications of sulfurbased autotrophic denitrification process for nitrate removal from activated sludge process effluent. Water Res. 60, 210–217. Sahinkaya, E., Yurtsever, A., Aktasß, Ö., Ucar, D., Wang, Z., 2015. Sulfur-based autotrophic denitrification of drinking water using a membrane bioreactor. Chem. Eng. J. 268, 180–186. Schulthess, R.V., Gujer, W., 1996. Release of nitrous oxide (N2O) from denitrifying activated sludge: verification and application of a mathematical model. Water Res. 30 (3), 521–530. Sierra-Alvarez, R., Beristain-Cardoso, R., Salazar, M., Gómez, J., Razo-Flores, E., Field, J.A., 2007. Chemolithotrophic denitrification with elemental sulfur for groundwater treatment. Water Res. 41 (6), 1253–1262. Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2003. Wastewater Engineering: Treatment and Reuse. McGraw-Hill, Boston. Wakida, F.T., Lerner, D.N., 2005. Non-agricultural sources of groundwater nitrate: a review and case study. Water Res. 39 (1), 3–16. Wang, D., Liu, Y., Ngo, H.H., Zhang, C., Yang, Q., Peng, L., He, D., Zeng, G., Li, X., Ni, B.J., 2017. Approach of describing dynamic production of volatile fatty acids from sludge alkaline fermentation. Biores. Technol. 238, 343–351. Xu, G., Feng, C., Fang, F., Chen, S., Xu, Y., Wang, X., 2015a. The heterotrophiccombined-with-autotrophic denitrification process: performance and interaction mechanisms. Water Sci. Technol. 71 (8), 1212–1218. Xu, G., Peng, J., Feng, C., Fang, F., Chen, S., Xu, Y., Wang, X., et al., 2015 b. Evaluation of simultaneous autotrophic and heterotrophic denitrification processes and bacterial community structure analysis. Appl. Microbiol. Biotechnol. 99 (15), 6527–6536. Xu, G., Yin, F., Chen, S., Xu, Y., Yu, H.-Q., 2016. Mathematical modeling of autotrophic denitrification (AD) process with sulphide as electron donor. Water Res. 91, 225–234. Xu, X., Chen, C., Lee, D.-J., Wang, A., Guo, W., Zhou, X., Guo, H., Yuan, Y., Ren, N., Chang, J.-S., 2013. Sulfate-reduction, sulfide-oxidation and elemental sulfur bioreduction process: modeling and experimental validation. Biores. Technol. 147, 202–211.

422

Y. Liu et al. / Chemical Engineering Science 172 (2017) 414–422

Xu, X., Chen, C., Wang, A., Guo, W., Zhou, X., Lee, D.-J., Ren, N., Chang, J.-S., 2014. Simultaneous removal of sulfide, nitrate and acetate under denitrifying sulfide removal condition: modeling and experimental validation. J. Hazard. Mater. 264, 16–24. Yang, W., Zhao, Q., Lu, H., Ding, Z., Meng, L., Chen, G.-H., 2016. Sulfide-driven autotrophic denitrification significantly reduces N2O emissions. Water Res. 90, 176–184. Yu, M., Lu, H., Wu, D., Zhao, Q., Meng, F., Wang, Y., Hao, X., Chen, G.-H., 2016. Effects of carbon-to-sulfur (C/S) ratio and nitrate (N) dosage on denitrifying sulfur

cycle-associated enhanced biological phosphorus removal (DS-EBPR). Sci. Rep. 6, 23221. Zhang, L., Zhang, C., Hu, C., Liu, H., Bai, Y., Qu, J., 2015a. Sulfur-based mixotrophic denitrification corresponding to different electron donors and microbial profiling in anoxic fluidized-bed membrane bioreactors. Water Res. 85, 422–431. Zhang, L., Zhang, C., Hu, C., Liu, H., Qu, J., 2015b. Denitrification of groundwater using a sulfur-oxidizing autotrophic denitrifying anaerobic fluidized-bed MBR: performance and bacterial community structure. Appl. Microbiol. Biotechnol. 99 (6), 2815–2827.