Model-based evaluation on simultaneous nitrate and

Chemical Engineering Science 152 (2016) 488–496

Contents lists available at ScienceDirect

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

Model-based evaluation on simultaneous nitrate and arsenite removal in a membrane biofilm reactor Xueming Chen, Bing-Jie Ni n Advanced Water Management Centre, The University of Queensland, St. Lucia, Brisbane, QLD 4072, Australia

H I G H L I G H T S








A novel approach to simultaneous NO3  and As(III) removal in MBfR was proposed. Kinetics of autotrophic As(III) oxidation process was determined. A biofilm model coupling DAMO and autotrophic As(III) oxidation was developed. MBfR performance and microbial structure were assessed under different conditions. The optimal CH4 supply was found inversely proportional to the corresponding HRT.

art ic l e i nf o

a b s t r a c t

Article history: Received 1 April 2016 Received in revised form 29 May 2016 Accepted 20 June 2016 Available online 21 June 2016

Nitrate (NO3  ) and arsenite (As(III)) are two major contaminants in groundwater, which could cause significant risks to human wellbeing and ecological system. In this work, a single-stage membrane biofilm reactor (MBfR) coupling denitrifying anaerobic methane (CH4) oxidation (DAMO) and autotrophic As(III) oxidation processes was proposed for the first time to achieve the in-situ or ex-situ simultaneous removal of NO3  and As(III) from groundwater. CH4 is supplied to the MBfR through gas-permeable membranes while NO3  and As(III) are provided in the bulk liquid. A mathematical model was developed by integrating the well-established biokinetics of DAMO microorganisms with the kinetics of As (III)-oxidizing bacteria (AsOB). The key parameter values of AsOB were specifically estimated using the batch experimental data of an enriched pure AsOB culture in conjunction with thermodynamic state calculations. The maximum specific growth rate of AsOB ( μAsOB ) and the yield coefficient for AsOB (YAsOB ) were determined to be 0.00161 h  1 and 0.016 g COD g  1 As, respectively. The modeling results demonstrated that both influent surface loading (or hydraulic retention time (HRT)) and CH4 surface loading played important roles in controlling the steady-state microbial community structure and thus significantly affected the system performance. The As(III)/NO3  ratio between 0.1 and 2 g As g  1 NO3  -N in the influent would have no significant impact on the overall system performance despite the varying microbial composition in the biofilm. Through properly adjusting the influent surface loading (or HRT) and CH4 surface loading whilst maintaining a sufficient biofilm thickness at a suitable influent As(III)/ NO3  ratio, the maximum removal efficiencies of total nitrogen and As(III) could both reach above 95.0%, accompanied by a high CH4 utilization efficiency of up to 99.0%. & 2016 Elsevier Ltd. All rights reserved.

Keywords: Membrane biofilm reactor Nitrate removal Arsenite removal Denitrifying anaerobic methane oxidation Mathematical modeling

1. Introduction The remediation of contaminated groundwater represents a significant issue which needs to be addressed urgently, considering the potable use of groundwater by over 50% of the global population (Jadhav et al., 2015). Arsenic (As) has been recognized as a crucial pollutant present in groundwater, with arsenite (As

n

Corresponding author. E-mail address: [email protected] (B.-J. Ni).

http://dx.doi.org/10.1016/j.ces.2016.06.049 0009-2509/& 2016 Elsevier Ltd. All rights reserved.

(III)) and arsenate (As(V)) being the predominantly occurring species. Arsenic has been classified as a human carcinogen by the United States Environmental Protection Agency (EPA), and arsenicosis could arise via ingestion of As-containing water and its subsequent accumulation in the body (Jadhav et al., 2015). Though the permissible limit of As in water is recommended at 10 mg L  1 by the World Health Organization (WHO), the actual As contamination level in groundwater varies greatly and could reach up to 300 mg L  1 (Mukherjee et al., 2006), depending on the specific contamination type and source of the site. A number of technologies have been established and applied to treat As-laden

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

groundwater, including the injection of nitrate (NO3  ) (Harvey et al., 2002). However, as a matter of fact, NO3  itself serves as another important pollutant commonly observed in groundwater. Due to the widespread application of nitrogen-containing fertilizers as well as the discharge of industrial and municipal wastewaters, the NO3  concentration in groundwater has been remarkable (Della Rocca et al., 2007), although its level varies in different places. The maximum acceptable contamination level of NO3  has been documented at 10 mg N L  1 by EPA (Yang and Lee, 2005), while it's set/recommended at 11.3 mg N L  1 by European Union countries and the WHO (Zhu and Getting, 2012). Nevertheless, the current NO3  contamination level in the groundwater of the United States and some European countries has been reported to increase to 200 mg L  1 (Chen et al., 2014a). Excessive NO3  not only favours the eutrophication of receiving waters (Fennessy and Cronk, 1997; Koren et al., 2000) but also causes serious health problems, such as methemoglobinemia (Knobeloch et al., 2000). Therefore, the potential in-situ or ex-situ integration of As and NO3  removal would be fairly desirable for groundwater with joint or separate contamination of As and NO3  , respectively. The discovery of denitrifying anaerobic methane (CH4) oxidation (DAMO) microorganisms (Raghoebarsing et al., 2006) and autotrophic As(III)-oxidizing bacteria (AsOB) (Oremland et al., 2002; Rhine et al., 2006) offers a potential solution to the above-mentioned problem by

489

means of implementing biological treatment processes. Through utilizing CH4 as the electron donor, DAMO archaea are capable of reducing NO3  to nitrite (NO2  ) (Haroon et al., 2013) while DAMO bacteria are able to convert NO2  to nitrogen gas (N2) (Ettwig et al., 2010). AsOB can use the energy and reducing power from As(III) oxidation for carbon dioxide (CO2) fixation and cell growth under both aerobic and anaerobic NO3  -reducing conditions (Zhang et al., 2015a). The oxidized As(V) produced is generally regarded as less mobile in the environment and less toxic to organisms than As(III) (Sun et al., 2010). Through proper introduction of CH4 in conjunction with the creation of a favorable reactor configuration, the coculture of DAMO microorganisms and AsOB could be facilitated to likely achieve the simultaneous removal of NO3  and As(III) from groundwater with various contamination conditions. Attached growth (i.e., biofilm) has been considered particularly suitable for retaining sufficient biomass of microorganisms with slow growth kinetics, such as DAMO archaea and DAMO bacteria (Chen et al., 2014b; Winkler et al., 2015). To ensure a high transfer and utilization efficiency of externally supplied CH4, the membrane biofilm reactor (MBfR) has been commonly applied/proposed to couple DAMO with other processes in a single-stage system, with high achievable removal performance (Chen et al., 2014b, 2015; Shi et al., 2013). Therefore, a novel MBfR integrating DAMO and autotrophic As(III) oxidation is proposed hereby for the

(A)

Gas-permeable membrane Biofilm

As(III) NO3-

CH4 Membrane biofilm reactor

Groundwater with As(III) and NO3-

(B)

Gas-permeable membrane

As(III) Groundwater with As(III)

Pump station

Biofilm

NO3CH4 Groundwater with NO3

-

Membrane biofilm reactor

Fig. 1. The conceptual application of the MBfR to achieve (A) the in-situ remediation of groundwater contaminated with NO3  and As(III) simultaneously and (B) the ex-situ treatment of the combined stream of groundwater polluted by NO3  and As(III) separately.

490

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

simultaneous removal of NO3  and As(III) from groundwater. In such an MBfR, CH4 is supplied through gas-permeable membranes while groundwater containing NO3  and As(III) is provided in the bulk liquid. The separation and counter-diffusional design of gas and liquid substrates not only renders a flexible control strategy, but also supplies a redox-stratified environment supportive of microorganisms of distinct physiological properties. From the perspective of application, as proposed in Fig. 1A, such an MBfR could be placed in situ to treat groundwater which is contaminated with NO3  and As(III) simultaneously. Alternatively, if the conditions permit, one stream contaminated with As(III) could be combined with another NO3  -containing stream to get treated in the proposed MBfR ex situ, with the treated water being sent back to groundwater (as depicted in Fig. 1B). However, to date no effort has yet been dedicated to investigating the feasibility of such an MBfR system with the coexistence of DAMO microorganisms and AsOB to achieve the simultaneous NO3  and As(III) removal. Due to the slow growth rates of DAMO microorganisms as well as the uncertainty of their ability to compete/cooperate with AsOB, experimental work aiming at process understanding and optimization could be very time-consuming and effort-demanding. Mathematical modeling provides a powerful and reliable tool to evaluate emerging technologies, such as DAMO-based wastewater treatment systems (Chen et al., 2014b, 2015; Winkler et al., 2015), and was thus adopted to probe into the MBfR coupling DAMO and autotrophic As(III) oxidation. For the construction of a reliable model, appropriate parameter values play a crucial role. In this work, the yield coefficient of AsOB was obtained by thermodynamic state calculations, while the maximum specific growth rate of AsOB by fitting the batch experimental data from an enriched pure culture of AsOB (Zhang et al., 2015a), with other parameters directly taken from reliable literature reported values. The derived/assessed parameters related to AsOB were integrated with previously well-established biokinetics of DAMO microorganisms to form a multi-species biofilm model, which was then applied to evaluate the conceptual feasibility of the simultaneous NO3  and As(III) removal using a single-stage MBfR coupling DAMO and autotrophic As(III) oxidation. The impacts of key process parameters including influent surface loading (or hydraulic retention time (HRT)), CH4 surface loading, and influent As(III)/ NO3  ratio on the microbial community structure and the system performance (including NO3  and As(III) removal efficiencies and CH4 utilization efficiency) of the MBfR were comprehensively investigated through a series of simulations. The findings of this work would benefit the experimental demonstration and process optimization of this new MBfR technology as well as its further investigations.

2. Materials and methods 2.1. Determination of the biomass yield of AsOB by thermodynamic state calculations Due to the unavailability of direct measurements, the biomass yield of AsOB was calculated based on thermodynamic state analysis. Similar to Winkler et al. (2015), the thermodynamic state calculations for the stoichiometric reaction of AsOB were performed according to the method of Kleerebezem and Van Loosdrecht (2010). First, the mass and electron balances and reaction stoichiometry were set up. The biomass of AsOB was assumed to take the form of CH1.8O0.5N0.2 (Kleerebezem and Van Loosdrecht, 2010), with CO2 and ammonium (NH4 þ ) being the carbon and nitrogen source, respectively. Then, a vector-based calculation method was applied to derive the stoichiometry of the metabolic redox reactions. The theoretical yield and the resulting overall

stoichiometry were obtained through combining the anabolic metabolism and the catabolic reaction without considering the energy dissipation. 2.2. Estimation of the maximum specific growth rate of AsOB using experimental data The multiple Monod equation was applied to describe the growth and endogenous respiration processes of AsOB (Eqs. (1) and (2)), with the inhibitory effect of As(III) included. The maximum specific growth rate of AsOB was estimated through minimizing the sum of squares of the deviations between the batch experimental data and model predictions of a pure culture of AsOB (strain SY) enriched from an As-contaminated soil (Zhang et al., 2015a). The enriched pure culture of AsOB was amended with 5 mM NO3  (70 g N m  3) as the electron acceptor and 1 mM As(III) (75 g As m  3) as the electron donor and cultivated under anaerobic conditions in triplicate. The concentrations of nitrate and nitrite were measured using SEAL Analytical segmented continuous-flow Auto Analyzer 3, while As species were extracted with their concentrations determined using high-performance liquid chromatography and inductively coupled plasma mass spectrometry (HPLC-ICP-MS). Cell concentration was assayed on spectrophotometer at 600 nm and converted approximately to cell number (Bahar et al., 2013), volatile suspended solids (Davis et al., 1973), and finally to chemical oxygen demand (COD) (Contreras et al., 2002), which was applied as the initial biomass concentration in g COD m  3 to the parameter estimation. The fact that NO2  was not detected in the medium confirmed the capability of AsOB to completely reduce NO3  to N2 (Zhang et al., 2015a). The As(III) inhibition constant for AsOB was assigned as the calculated EC50 value on the basis of 24 h growth data of strain SY (Zhang et al., 2015a). Since no direct measurements have been conducted on this AsOB strain, the reported values for the NO3  and As(III) affinity constants of similar species (Bahar et al., 2013; Peng et al., 2015) were adopted while the endogenous respiration rate was assumed as 1/20 of the maximum specific growth rate to be estimated (Winkler et al., 2015).

AsOB growth rate = μAsOB

SAs(III )

SNO3

AsOB AsOB SAs(III ) + KAs (III ) SNO3 + KNO3

×

KIAsOB , As(III ) SAs(III ) + KIAsOB , As(III )

AsOB endogenous respiration rate = bAsOB

XAsOB

(1)

SNO3 AsOB SNO3 + KNO 3

XAsOB

(2)

3

where XAsOB is AsOB concentration, g COD m ; SAs(III ) is As(III) concentration, g As m  3; SNO3 is NO3  concentration, g N m  3; μAsOB is maximum specific growth rate of AsOB, h  1; bAsOB is endogenous AsOB respiration rate of AsOB, h  1; KAs (III ) is As(III) affinity constant for AsOB, g As m  3; g As

m3 ;

and

KIAsOB , As(III ) is As(III) inhibition constant for AsOB,

AsOB KNO 3

is NO3  affinity constant for AsOB, g N m  3.

2.3. Development of the biofilm model integrating DAMO and AsOB The obtained stoichiometric and kinetic parameters of AsOB were integrated with the reported well-established parameters of DAMO microorganisms (Chen et al., 2014b) to construct a biochemical reaction model characterizing the metabolisms of DAMO microorganisms and AsOB. Kinetic control of all the reaction rates of microbial growth and endogenous respiration processes was described by the Monod equation considering all substrates

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

involved. The components, kinetics, and stoichiometry of the developed model are summarized in Tables S1 and S2 in the Supporting Information (SI). A multi-species and multi-substrate one-dimensional biofilm model was developed and employed on the software AQUASIM 2.1d (Reichert, 1998) to simulate the bioconversion processes and microbial community structure in the MBfR, the schematic diagram of which is delineated in Fig. 1. The MBfR was modeled as composed of a completely mixed gas compartment (representing the membrane lumen operated as flowthrough) and a biofilm compartment (containing the biofilm and the bulk liquid). The CH4 concentration in the gas compartment was determined by the applied gas flow rate and pressure. The methane flux from the gas to the biofilm matrix compartment through membranes was implemented via a defined diffusive link and modeled based on Henry's law (Terada et al., 2007). The biofilm porosity was kept constant at 0.75 m3 liquid m  3 biofilm (Ni and Yuan, 2013), while the biomass density was 50000 g COD m  3 (Koch et al., 2000). The simulated influent (i.e., single or merged groundwater containing NO3  and As(III)) of the MBfR was supplied to the bulk liquid, with the flow rate varied to regulate the influent surface loading which also corresponds to HRT. Similar to Terada et al. (2007), the diffusion coefficients for soluble components in the liquid phase of the biofilm were set at 0.8-fold of the values in water. Parameters regarding the diffusion coefficients for nitrogen species were selected according to Hao et al. (2002). The diffusion coefficient for CH4 was taken from Cussler (2003), while those for As(III) and As (V) were adopted from Tanaka et al. (2013). More details of the biofilm model and MBfR setup could refer to the SI as well as Chen et al. (2015). 2.4. Design of the simulation scenarios Model simulations were then performed under different process conditions to evaluate the implementation of the MBfR coupling DAMO and autotrophic As(III) oxidation for the simultaneous removal of NO3  and As(III). Five different scenarios are considered in this work, as detailed in Table 1. The first simulation scenario (Scenario 1 of Table 1) analysed the mechanisms behind the system performance through generating depth profiles of substrates (including nitrogen species, As species, and CH4) as well as microbial distribution in the MBfR biofilm. The As(III) and NO3  concentrations in the influent were set at 50 g As m  3 and 50 g N m  3, respectively, giving rise to an influent As(III)/NO3  ratio of 1. The applied influent surface loading (LIN), CH4 surface loading (LCH4), and biofilm thickness (Lf) were 0.004 m day  1 (i.e., an HRT of 1 day), 0.14 g m  2 day  1, and 200 mm, respectively. Scenarios 2–4 in Table 1 explored the effects of LCH4, LIN, and influent As(III)/NO3  ratio, respectively, on the system performance and the microbial community structure of the MBfR biofilm at steady-state. The process parameters for simulation were chosen systematically over wide ranges of LCH4 (0.02–0.35 g m  2 day  1), LIN (0.002– 0.008 m day  1, i.e., HRT of 0.5–2 day), and influent As (III)/NO3  ratio (0.1–2.25 on the basis of elemental mass with the NO3  concentration being fixed at 50 g N m  3). Scenario 5 in Table 1 examined the joint impact of LCH4 and LIN on the system performance and optimized the operation of the MBfR to achieve the high-level simultaneous NO3  and As(III) removal as well as CH4 utilization. For each simulation scenario, the initial concentrations of all soluble components were assumed to be zero in the biofilm and the bulk liquid. An average biofilm thickness was applied in the model without consideration of its variation with locations. All simulations assumed an initial biofilm thickness of 20 mm. Simulations were typically run for up to 1000 days to reach steady-state conditions in terms of biofilm thickness, effluent quality, and

491

Table 1 An overview of the simulation strategies for the reported results. Simulation strategies Scenario 1 Standard simulation of the MBfR

Scenario 2 Effect of LCH4 on the MBfR

Scenario 3 Effect of LIN on the MBfR

Simulation conditions

LIN ¼ 0.004 m day  1 (HRT ¼ 1 day) As(III)/NO3  ¼ 1 LCH4 ¼ 0.14 g m  2 day  1 Lf ¼ 200 mm LIN ¼ 0.004 m day–1 (HRT ¼ 1 day) As(III)/NO3  ¼ 1 Lf ¼ 200 mm

LCH4 ¼0.02– 0.35 g m  2 day  1

As(III)/NO3– ¼ 1

LIN ¼0.002– 0.008 m day  1 (HRT ¼0.5–2 day)

LCH4 ¼ 0.14 g m  2 day  1 Lf ¼ 200 mm Scenario 4 Effect of influent LIN ¼ 0.004 m day  1 As(III)/NO3– ratio on the MBfR (HRT ¼ 1 day) LCH4 ¼ 0.14 g m  2 day  1Lf ¼200 mm Scenario 5 Combined effect of LCH4 and LIN on the MBfR

Variable conditions

As(III)/NO3– ¼ 1 Lf ¼ 200 mm

As(III)/NO3– ¼0.1– 2.25

HRT¼ 0.5–2 day (LIN ¼ 0.002– 0.008 m day  1) LCH4 ¼0.05– 0.35 g m  2 day  1 (360 simulation runs)

biomass compositions in the biofilm. In addition to the removal efficiencies of total nitrogen (TN), i.e., sum of NO3  and NO2  , and As(III), the steady-state CH4 utilization efficiency was also calculated and used as a criterion to assess the system performance and applicability of the MBfR.

3. Results 3.1. Sensitivity analysis of the kinetic model describing AsOB Prior to the parameter estimation, a sensitivity analysis was conducted to locate the most important parameters of the kinetic model of AsOB. The key parameters in the parameter estimation process were determined based on the sensitivities of these parameters to the model output. For a parameter to be reliably identifiable, the measured experimental data should be highly sensitive to the changes of this parameter. The sensitivity analysis involved the maximum specific growth rate of AsOB ( μAsOB ), yield coefficient for AsOB (YAsOB ), As(III) affinity  AsOB affinity constant for AsOB constant for AsOB ( KAs (III ) ), and NO3

AsOB AsOB ( KNO 3 ), while the As(III) inhibition constant for AsOB ( KI , As(III ) ) was set to the experimentally determined value without further evaluation. In this work, the “absolute-relative” sensitivity function was used, with the base values of parameters and initial conditions obtained according to the literature reported values of similar species (Bahar et al., 2013; Peng et al., 2015; Zhang et al., 2015a) and the specific batch experimental settings. The sensitivities of the As(III) and NO3  concentrations to the changes of these parameters are demonstrated in Figure S1 in the SI. The sensitivity analysis revealed that the As(III) and NO3  concentrations were AsOB AsOB highly sensitive to μAsOB and YAsOB , while KAs (III ) and KNO3 showed no significant effect. Based on the findings of sensitivity analysis, the

492

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

AsOB AsOB two non-sensitive parameters, KAs (III ) and KNO3 , were directly set to the literature reported values. In contrast, the two sensitive parameters, μAsOB and YAsOB , could be reliably identified based on experimental data together with the thermodynamic state calculations.

3.2. Parameter determination of the kinetic model describing AsOB Assuming an optimal energy usage of NO3  and As(III) yielded a stoichiometric reaction (Eq. (3)) which characterizes the overall metabolism of AsOB. As indicated in Eq. (3), the production of 1 mol of AsOB biomass in the form of CH1.8O0.5N0.2 requires 28.1 mol of As(III), resulting in a theoretical YAsOB of 0.016 g COD g  1 As. CO2 þ10.4NO3  þ28.1H3AsO3 þ0.2NH4 þ 46H þ þ 4.6H2Oþ 5.2N2 þCH1.8O0.5N0.2 þ28.1HAsO42 

(3)



The batch experimental profiles of NO3 , As(III), and As (V) reported in Zhang et al. (2015a) were then used to estimate the value of μAsOB , with the fits of the model to the batch experimental data shown in Fig. 2. μAsOB was adjusted and obtained at 0.00161 h  1, leading to a corresponding endogenous respiration rate of AsOB ( bAsOB ) of 0.00008 h  1. As delineated in Fig. 2, the model predictions not only matched the consumption profiles of NO3  and As(III), but also captured the accumulation trend of As(V) produced. The good agreement between the model predictions and measured data in this work supported the validity of the estimated parameters as well as the model structure to describe AsOB. Therefore, these derived/evaluated parameters related to AsOB could be reliably applied to subsequent simulations. 3.3. Simultaneous removal of NO3  and As(III) in MBfR

Fig. 3. Modeling results of the MBfR for simultaneous NO3  and As(III) removal from Scenario 1 in Table 1 (depth zero represents the membrane surface at the base of the biofilm): (A) Microbial population distribution; (B) distribution profiles of nitrogen species, As species, and CH4; and (C) species-specific consumption rates of nitrogen species and As(III).

substrate supplied (i.e., CH4) were comparable to those reported in similar H2-based and CH4-based MBfR systems (Chen et al., 2016a; Rittmann, 2007; Tang et al., 2013, 2012) and confirmed the feasibility of the proposed MBfR approach to simultaneous NO3  and As(III) removal.

80 NO3

-

3.4. Profiles of microbial community structure and substrates in the biofilm

60

-3

-3

Concentration (g N m or g As m )

Table S3 in the SI shows the influent and effluent characteristics as well as the system performance of the steady-state MBfR under the operational conditions of Scenario 1 in Table 1. The influent NO3  (50 g N m  3) was significantly removed by the MBfR, with 2.9 g N m  3 remaining in the effluent. No NO2  accumulation was observed in the MBfR, thus leading to a TN removal efficiency of 94.2%. The influent As(III) (50 g As m  3) was mainly converted to As(V) at an effluent concentration of 49.2 g As m  3, corresponding to an As(III) removal efficiency of 98.4%. The accompanying CH4 utilization efficiency was calculated at 97.7%. These high-level removal efficiencies as well as the efficient utilization of gaseous

40

0

As(III)

As(V)

20

0

1

2 3 Time (day)

4

5

Fig. 2. Parameter estimation results based on the batch experimental data of a pure AsOB culture (measured data, symbols; model predictions, solid lines).

The steady-state microbial population distribution and the concentration profiles of nitrogen species, As species, and CH4 as well as the species-specific substrate consumption rates within the MBfR biofilm under the operational conditions of Scenario 1 (Table 1) are shown in Fig. 3. DAMO microorganisms were dominant with the coexistence of AsOB in the biofilm (Fig. 3A). The abundance of DAMO microorganisms was around 98% at the base of the biofilm (including 46% for DAMO archaea and 52% for DAMO bacteria) and decreased gradually to 90% at the surface of the biofilm (including 42% for DAMO archaea and 48% for DAMO bacteria). In contrast, the abundance of AsOB was only 2% at the base of the biofilm and increased to 10% at the surface of the biofilm. The associated substrate profiles within the biofilm are delineated in Fig. 3B. Both NO3  and As(III) concentrations decreased from the surface to the base of the biofilm due to their consumptions by DAMO archaea and AsOB. Nevertheless, a

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

The impacts of LCH4 on the system performance and the microbial community structure in the biofilm of the MBfR coupling DAMO and autotrophic As(III) oxidation (Scenario 2 in Table 1) are shown in Fig. 4A. At LCH4 of less than 0.14 g m  2 day  1, the As(III) removal efficiency stayed above 98.4%. However, the As(III) removal efficiency dropped abruptly with the further increase of LCH4, reaching zero at LCH4 of more than 0.17 g m  2 day  1. A different trend was observed for the TN removal efficiency, which was only 17.3% at LCH4 of 0.02 g m  2 day  1 but gradually increased to and remained at the maximum of 99.5% at LCH4 of over 0.17 g m  2 day  1. The changing microbial community structure under different LCH4 conditions was responsible for the varying removal performance of the MBfR. When LCH4 was below 0.17 g m  2 day  1, DAMO microorganisms coexisted with AsOB in the biofilm. The increase of LCH4 to 0.17 g m  2 day  1 favored the competitiveness and growth of DAMO microorganisms to a higher degree. Therefore, the abundance of DAMO archaea and DAMO bacteria increased from 39% and 44% to 45% and 55%, respectively, while AsOB were gradually outcompeted and disappeared from the biofilm, with the abundance decreasing from 17% to zero. The CH4 supplied was efficiently used by DAMO microorganisms which was reflected by the high-level CH4 utilization efficiency in Fig. 4A. Further increase of LCH4 represented the excessive CH4 supply, which was confirmed by the decreasing CH4 utilization efficiency (reaching 43.8% at LCH4 of 0.35 g m  2 day  1). The relationships between LIN and the system performance as well as the microbial community structure in the biofilm (Scenario 3 in Table 1) are illustrated in Fig. 4B. When LIN was lower than 0.0033 m day  1, DAMO archaea had the competitive advantage in NO3  utilization over AsOB. Thus, DAMO microorganisms dominated the biofilm without the presence of AsOB, resulting in the zero As(III) removal efficiency. Meanwhile, the TN removal efficiency remained above 99.6% while the CH4 utilization efficiency kept increasing with LIN. A higher LIN offered sufficient substrates to AsOB, stimulated the growth, and hence conduced to the appearance of AsOB in the biofilm. At LIN of 0.0040 m day  1, the active

80

0.6

60

0.4

40

0.2

20

0.0 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 -2

-1

CH4 surface loading (g CH4 m day )

(B) 1.0 Active biomass fraction

Efficiency (%)

0.8

100

0.8

80

0.6

60

0.4

40

0.2

20

Efficiency (%)

Active biomass fraction

100

0

0.0 0.002 0.004 0.006 0.008 -1 Influent surface loading (m day )

(C) 1.0

100

0.8

80

0.6

60

0.4

40

0.2

20

0.0 0.0

0.5

1.0 -

1.5

Efficiency (%)

3.5. Effects of key process parameters on the MBfR

(A) 1.0

Active biomass fraction

different trend was observed for CH4, which quickly decreased from the base of the biofilm (i.e., the membrane surface) and dropped below 0.01 g COD m  3 at the biofilm thickness of more than 125 μm, owing to its consumption by DAMO microorganisms. Over the entire biofilm range, there was no obvious change in the concentrations of NO2  and As(V), which were generated by DAMO archaea and AsOB, respectively. The counter-diffusional supply of liquid (NO3  and As(III)) and gas (CH4) substrates resulted in not only heterogeneous microbial distribution but also stratified activity in the biofilm of the MBfR. As shown in Fig. 3C, the nitrogen removal mediated by DAMO microorganisms mainly occurred in the inner layer of the biofilm. The NO3  consumption rate by DAMO archaea was comparable to the NO2  consumption rate by DAMO bacteria, both of which significantly decreased towards the bulk liquid. By comparison, though the activity of AsOB was higher in the outer layer of the biofilm and decreased towards the membrane surface, it basically prevailed across the biofilm (Fig. 3C). Under the given operational conditions of Scenario 1, the cooperation between DAMO microorganisms approximately accounted for around 93% of the TN removed, with the remaining 7% achieved by AsOB. This TN distribution directly led to the dominance of DAMO microorganisms in the biofilm over AsOB, which possessed a relatively low yield (i.e., 0.016 g COD g  1 As in this work). Overall, DAMO microorganisms played the leading role in the TN removal. AsOB were the sole contributor to the As(III) removal whilst further enhancing the nitrogen removal.

493

0 2.5

2.0 -1

As(III)/NO3 ratio (mg As mg N) DAMO bacteria As(III) removal

DAMO archaea TN removal

AsOB CH4 utilization

Fig. 4. Modeling results of the MBfR for simultaneous NO3  and As(III) removal from Scenarios 2–4 in Table 1: (A) Effect of CH4 surface loading; (B) Effect of influent surface loading; and (C) Effect of influent As(III)/NO3  ratio on the TN and As (III) removal efficiencies, CH4 utilization efficiency, and microbial community structure of the biofilm.

biomass fractions of DAMO archaea, DAMO bacteria, and AsOB were 45%, 51%, and 4%, respectively. The corresponding As(III) removal efficiency, TN removal efficiency, and CH4 utilization efficiency were 98.4%, 94.2%, and 97.7%, respectively. As indicated in Fig. 4B, the active biomass fractions of both DAMO archaea and DAMO bacteria slightly decreased with the further increase of LIN, compensated by the increasing active biomass fraction of AsOB. Though high-level As(III) removal (498.0%) and CH4 utilization ( 97.0%) could still be obtained at a higher LIN (40.0040 m day  1), the TN removal efficiency would decrease, reaching 50.4% at LIN of 0.0080 m day  1. The dependence of the system performance and the microbial community structure of the MBfR on the influent As(III)/NO3  ratio of the MBfR (Scenario 4 in Table 1) is shown in Fig. 4C. Within the range studied, there was a positive relationship between the abundance of AsOB and the influent As(III)/NO3  ratio. To be

494

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

specific, the abundance of AsOB increased straight from less than 1% at the influent As(III)/NO3  ratio of 0.1–10% at the influent As (III)/NO3  ratio of 2.25. The abundance of both DAMO archaea and DAMO bacteria decreased accordingly. A higher influent As(III)/ NO3  ratio provided more electron donors (i.e., As(III)) for the denitrification of NO3  , thus ensuring a higher nitrogen removal performance, as evidenced by the increasing TN removal efficiency in Fig. 4C. Regarding the As(III) removal efficiency, it first increased from 88.6% at the influent As(III)/NO3  ratio of 0.1 to 98.0% at the influent As(III)/NO3  ratio of 0.75. Then, it remained above 98.0% until the influent As(III)/NO3  ratio exceeded 1.75. An influent As (III)/NO3  ratio of more than 1.75 restricted the As(III) removal due to the relatively limited availability of NO3  as the electron acceptor for AsOB. As a result, the As(III) removal efficiency decreased to 79.5% at the influent As(III)/NO3  ratio of 2.25. Over the range of the influent As(III)/NO3  ratio studied, the CH4 utilization efficiency maintained above 97.5% because of the consistent dominance of DAMO microorganisms in the biofilm. 3.6. Optimization of the system performance for simultaneous NO3  and As(III) removal Additional simulations based on the operational conditions of Scenario 1 showed that although the TN removal and CH4 utilization of the MBfR were poor under thin biofilm thickness (o100 μm) conditions due to the limited total biomass as well as the impaired abundance of DAMO microorganisms, a biofilm thickness of over 100 μm was sufficient to achieve high-level simultaneous NO3  and As(III) removal and CH4 utilization (Figure S2 in the SI). Likewise, as already demonstrated in Fig. 4C, though the microbial composition varied, no significant change was observed for the overall system performance of the MBfR (including As(III) removal efficiency, TN removal efficiency, and CH4 utilization efficiency) at an influent As(III)/NO3  ratio of between 0.1 and 2. In contrast, Figs. 4A and B reveal that both LCH4 and LIN played important roles in regulating the microbial community structure of the MBfR coupling DAMO and autotrophic As(III) oxidation and thus significantly affected the system performance. Therefore, the joint optimal operational conditions in terms of LCH4 and LIN at an influent As(III)/NO3  ratio of 1 and a sufficient biofilm thickness of 200 μm were assessed to optimize the system performance of the MBfR for simultaneous NO3  and As(III) removal. Considering that LIN is practically equivalent to HRT, HRT instead of LIN was applied to the following analysis. Figs. 5A–C illustrate the As(III) removal efficiency, TN removal efficiency, and CH4 utilization efficiency, respectively, of the steady-state MBfR under the extensive simulation conditions of Scenario 5 in Table 1. The white dot line in Fig. 5C highlights the optimal operational conditions in terms of HRT and LCH4 to achieve high-level ( 495%) simultaneous NO3  and As(III) removal and CH4 utilization of the MBfR. The optimal LCH4 was inversely proportional to the corresponding HRT; the optimal LCH4 decreased with the increasing HRT. Under these optimal operational conditions, DAMO microorganisms dominated in symbiosis with AsOB in the biofilm to remove NO3  and As(III) simultaneously. Below the optimal region indicated in Fig. 5, the increase of HRT or LCH4 with the other one being fixed would increase the TN removal efficiency but have no significant effect on the As(III) removal efficiency and CH4 utilization efficiency, both of which stayed over 95.0% (the dark red regions in Figs. 5A and C). A higher HRT or LCH4 above the optimal region highlighted in Fig. 5 would cause the washout of AsOB from the biofilm, resulting in the abrupt loss of

Fig. 5. Modeling results for (A) As(III) removal efficiency, (B) TN removal efficiency, and (C) CH4 utilization efficiency under different CH4 surface loading and HRT conditions from Scenario 5 in Table 1. The color scales on the right represent efficiency in %. The optimal condition for high-level simultaneous NO3  and As(III) removal as well as efficient CH4 utilization is highlighted in (C) using a white dot line. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

the As(III) removal efficiency (the dark blue region in Fig. 5A). DAMO microorganisms therefore solely dominated the biofilm, leading to a high TN removal efficiency of up to 99.8% (the dark red region in Fig. 5B). Both too long HRT at a certain LCH4 and too high LCH4 at a certain HRT therein represented the excessive CH4 provision for DAMO microorganisms, with a fraction of CH4 leaving the system unused. This CH4 wastage was verified by the decreasing CH4 utilization efficiency in Fig. 5C.

4. Discussion NO3  and As(III) are introduced into groundwater by either natural or anthropogenic activities or both. The migration and transformation of arsenic in groundwater are influenced by hydrodynamic and redox conditions, adsorption-desorption actions, and biochemical reactions (Chen and Chai, 2008). In general, concentration gradient leads to arsenic migration in groundwater. Microorganisms play an important role in the transformation of arsenic; As(V) is the dominant form under oxidizing conditions while As(III) mainly exists in reduced anaerobic environments (Sun et al., 2010). Efforts should be dedicated to achieving the efficient removal of NO3  and As(III) from groundwater where these two toxicants are solely or collectively present, considering their individual as well as synergistic risks to human health as well as ecological system. Physical- and chemical-based methods have been commonly applied to the removal of NO3  and As(III) (Jadhav et al., 2015). Regardless of the high achievable removal performance, these methods are susceptible to interference and are usually associated with high capital and operational costs as well as difficulty in the toxic by-product disposal (Chen et al., 2014a). Bioremediation of As(III)-laden groundwater has been proposed through natural or transgenic plants or engineered microorganisms (Rahman et al., 2014). Compared to the conventional heterotrophic denitrification, hydrogen- and sulfur-based autotrophic denitrification processes elude the requirement of external organic carbon, produce less sludge, and therefore have been extensively investigated and applied to remove NO3  from groundwater (Haugen et al., 2002; Moon et al., 2006; Schnobrich et al., 2007; Soares, 2002; Zhang et al., 2015b). The discovery of methane-dependent denitrification (i.e., DAMO) processes provides another alternative to treating groundwater through using inexpensive and widely available CH4 as the carbon source and electron donor. However, to date, few study has been reported on technologies to achieve the simultaneous removal of NO3  and As(III) from groundwater, not to mention the related process evaluation or optimization. In this work, a new approach to simultaneous NO3  and As(III) removal through integrating DAMO and autotrophic As(III) oxidation processes in a single MBfR was proposed for the first time, with its feasibility proved using mathematical modeling. The MBfR could potentially enable the in-situ remediation of groundwater contaminated with NO3  and As(III) simultaneously as well as the exsitu treatment of the combined stream of groundwater polluted by NO3  and As(III) separately, as proposed in Fig. 1. The extensive simulation results of this work showed that the simultaneous removal of NO3  and As(III) and CH4 utilization highly relied on process parameters such as LIN, LCH4, and influent As(III)/NO3  ratio of the MBfR (Fig. 4). These process parameters directly controlled substrates availability and therefore determined the competition/cooperation among microorganisms (e.g., DAMO and AsOB in this work) in the multi-species biofilm, which was also observed in similar O2-based and CH4-based MBfRs (Chen et al., 2014b, 2015, 2016a, 2016b; Terada et al., 2007). Under the conditions of low LIN or high LCH4, DAMO microorganisms outcompeted AsOB for NO3  (Fig. 4). The cooperative mechanism between DAMO microorganisms and AsOB was

495

only favored at high LIN or low LCH4. A high LIN provided more NO3  for AsOB while a low LCH4 reduced the substrate for DAMO microorganisms and hence limited their growth and contribution. Therefore, the control strategies derived from this work would significantly benefit the demonstration and operation of this novel MBfR system. Though this work mainly targeted groundwater with a moderate As(III)/NO3  ratio of between 0.1 and 2, the proposed MBfR could be easily adjusted to achieve the simultaneous removal of NO3  and As(III) from groundwater with an As(III)/NO3  ratio outside the studied range (i.e., 0.1–2). For example, at a relatively low As(III)/NO3  ratio (e.g., o 0.1 under the simulation conditions in this work), AsOB would lose the competition against DAMO archaea for NO3  , resulting in the failure of the As(III) removal in the MBfR. In this case, a small amount of oxygen could be introduced into the bulk liquid of the MBfR to stimulate the aerobic oxidation of As(III). However, the introduction of oxygen might potentially incur the existence of aerobic CH4 oxidation process, which would increase the operational complexity of the MBfR and affect the effective utilization of CH4 supplied for the denitrification of NO3  by DAMO microorganisms. Comparatively, a high As (III)/NO3  ratio (e.g., 42.0 under the simulation conditions in this work) would strengthen the role of autotrophic As(III) oxidation process. Therefore, LCH4 of the MBfR should be properly reduced to depress the contribution from DAMO microorganisms to the TN removal. Given that the obtained metabolic reaction of AsOB (Eq. (3)) is correct, an As(III)/NO3  ratio of 14.5 on the basis of elemental mass is sufficient to support the simultaneous removal of NO3  and As(III) by AsOB alone. Under this critical influent condition of the MBfR, the simultaneous removal of NO3  and As(III) does not necessitate the contribution from DAMO processes, thus reducing LCH4 to zero. It should be noted that this work only focused on exploring the feasibility of the proposed technology without considering its economic and process comparison to other technologies as well as the disposal of the converted As(V), both of which warrant further assessment. Furthermore, the model parameters applied in this work are specific to the species of microorganisms involved and that their values might change under different conditions. A sensitivity analysis (Figure S3 in the SI) showed that the TN removal efficiency was highly sensitive to the yield coefficients of DAMO bacteria ( YDb) and DAMO archaea ( YDa ), while the maximum specific growth rates of AsAOB ( μAsOB ) and DAMO archaea ( μDa ) were the most decisive parameters for the As(III) removal efficiency under the operational conditions of Scenario 1 in Table 1. Although these four parameters have all been calibrated and/or validated using experimental data, they might still need to be accurately determined in the future specific application of the model. Moreover, further experimental work of the proposed MBfR would help to validate the results as well as the model.

5. Conclusions A novel approach was proposed to remove NO3  and As(III) simultaneously from groundwater through coupling DAMO and autotrophic As(III) oxidation processes in a single-stage MBfR. By fitting the batch experimental data of an enriched pure AsOB culture together with the thermodynamic state calculations, the maximum specific growth rate of AsOB ( μAsOB ) was determined to be 0.00161 h  1 while the yield coefficient for AsOB ( YAsOB ) was calculated at 0.016 g COD g  1 As. A biofilm model developed by integrating the well-established biokinetics of DAMO with the assessed/derived parameters of AsOB was then applied to evaluate the system performance and microbial community structure of the MBfR under different operational conditions. The simulation results demonstrated that the simultaneous removal of NO3  and As

496

X. Chen, B.-J. Ni / Chemical Engineering Science 152 (2016) 488–496

(III) and CH4 utilization highly relied on the competition/cooperation between DAMO and AsOB in the biofilm, which could be controlled by process parameters such as influent and methane surface loadings and influent As(III)/NO3  ratio of the MBfR. The maximum simultaneous removal efficiencies of TN and As(III) could both reach above 95.0% together with a high CH4 utilization efficiency of up to 99.0%, by adjusting the HRT (or influent surface loading) and CH4 surface loading whilst maintaining a sufficient suitable biofilm thickness at a moderate influent As(III)/NO3  ratio.

Acknowledgements This study was supported by the Australian Research Council (ARC) through Project DP130103147. Xueming Chen acknowledges the scholarship support from China Scholarship Council (CSC). Dr. Bing-Jie Ni acknowledges the support of ARC Discovery Early Career Researcher Award (DE130100451) and University of Queensland Foundation Research Excellence Award. The authors declare that they have no competing interests.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ces.2016.06.049.

References Bahar, M.M., Megharaj, M., Naidu, R., 2013. Kinetics of arsenite oxidation by variovorax sp. MM-1 isolated from a soil and identification of arsenite oxidase gene. J. Hazard. Mater. 262, 997–1003. Chen, D., Yang, K., Wang, H.Y., Lv, B., 2014a. Nitrate removal from groundwater by hydrogen-fed autotrophic denitrification in a bio-ceramsite reactor. Water Sci. Technol. 69, 2417–2422. Chen, X., Guo, J., Xie, G.-J., Yuan, Z., Ni, B.-J., 2016a. Achieving complete nitrogen removal by coupling nitritation-anammox and methane-dependent denitrification: a model-based study. Biotechnol. Bioeng. 113, 1035–1045. Chen, X., Liu, Y., Peng, L., Yuan, Z., Ni, B.-J., 2016b. Model-based feasibility assessment of membrane biofilm reactor to achieve simultaneous ammonium, dissolved methane, and sulfide removal from anaerobic digestion liquor. Sci. Rep. 6, 25114. Chen, X., Guo, J., Shi, Y., Hu, S., Yuan, Z., Ni, B.-J., 2014b. Modeling of simultaneous anaerobic methane and ammonium oxidation in a membrane biofilm reactor. Environ. Sci. Technol. 48, 9540–9547. Chen, X., Guo, J., Xie, G.-J., Liu, Y., Yuan, Z., Ni, B.-J., 2015. 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.-n., Chai, L.-y., 2008. Migration and transformation of arsenic in groundwater. Nonferr. Met.-Beijing 60, 109. Contreras, E.M., Bertola, N.C., Giannuzzi, L., Zaritzky, N.E., 2002. A modified method to determine biomass concentration as COD in pure cultures and in activated sludge systems. Water SA 28, 463–467. Cussler, E.L., 2003. Mass transfer and diffusion. In: Meyers, R.A. (Ed.), Encyclopedia of Physical Science and Technology, Third ed. Academic Press, New York, pp. 171–180. Davis, B.D., Dulbecco, R., Eisen, H., Ginsberg, H., 1973. Bacterial Physiology: Microbiology. Harper and Row, Maryland, pp. 96–97. Della Rocca, C., Belgiorno, V., Meric, S., 2007. Overview of in-situ applicable nitrate removal processes. Desalination 204, 46–62. Ettwig, K.F., Butler, M.K., Le Paslier, D., Pelletier, E., Mangenot, S., Kuypers, M.M.M., Schreiber, F., Dutilh, B.E., Zedelius, J., de Beer, D., Gloerich, J., Wessels, H.J.C.T., van Alen, T., Luesken, F., Wu, M.L., van de Pas-Schoonen, K.T., den Camp, H.J.M.O., Janssen-Megens, E.M., Francoijs, K.J., Stunnenberg, H., Weissenbach, J., Jetten, M.S. M., Strous, M., 2010. Nitrite-driven anaerobic methane oxidation by oxygenic bacteria. Nature 464, 543–548. Fennessy, M.S., Cronk, J.K., 1997. The effectiveness and restoration potential of riparian ecotones for the management of nonpoint source pollution, particularly nitrate. Crit. Rev. Environ. Sci. Technol. 27, 285–317. Hao, X., Heijnen, J.J., van Loosdrecht, M.C.M., 2002. Sensitivity analysis of a biofilm model describing a one-stage completely autotrophic nitrogen removal (CANON) process. Biotechnol. Bioeng. 77, 266–277. Haroon, M.F., Hu, S.H., Shi, Y., Imelfort, M., Keller, J., Hugenholtz, P., Yuan, Z.G., Tyson, G.W., 2013. Anaerobic oxidation of methane coupled to nitrate reduction in a novel archaeal lineage. Nature 500, 567–570.

Harvey, C.F., Swartz, C.H., Badruzzaman, A.B.M., Keon-Blute, N., Yu, W., Ali, M.A., Jay, J., Beckie, R., Niedan, V., Brabander, D., Oates, P.M., Ashfaque, K.N., Islam, S., Hemond, H.F., Ahmed, M.F., 2002. Arsenic mobility and groundwater extraction in Bangladesh. Science 298, 1602–1606. Haugen, K.S., Semmens, M.J., Novak, P.J., 2002. A novel in situ technology for the treatment of nitrate contaminated groundwater. Water Res. 36, 3497–3506. Jadhav, S.V., Bringas, E., Yadav, G.D., Rathod, V.K., Ortiz, I., Marathe, K.V., 2015. Arsenic and fluoride contaminated groundwaters: a review of current technologies for contaminants removal. J. Environ. Manag. 162, 306–325. Kleerebezem, R., Van Loosdrecht, M.C.M., 2010. A generalized method for thermodynamic state analysis of environmental systems. Crit. Rev. Environ. Sci. Technol. 40, 1–54. Knobeloch, L., Salna, B., Hogan, A., Postle, J., Anderson, H., 2000. Blue babies and nitrate-contaminated well water. Environ. Health Perspect. 108, 675–678. Koch, G., Egli, K., Van der Meer, J.R., Siegrist, H., 2000. Mathematical modeling of autotrophic denitrification in a nitrifying biofilm of a rotating biological contactor. Water Sci. Technol. 41, 191–198. Koren, D.W., Gould, W.D., Bedard, P., 2000. Biological removal of ammonia and nitrate from simulated mine and mill effluents. Hydrometallurgy 56, 127–144. Moon, H.S., Nam, K., Kim, J.Y., 2006. Initial alkalinity requirement and effect of alkalinity sources in sulfur-based autotrophic denitrification barrier system. J. Environ. Eng. ASCE 132, 971–975. Mukherjee, A., Sengupta, M.K., Hossain, M.A., Ahamed, S., Das, B., Nayak, B., Lodh, D., Rahman, M.M., Chakraborti, D., 2006. Arsenic contamination in groundwater: a global perspective with emphasis on the Asian scenario. J. Health Popul. Nutr. 24, 142–163. Ni, B.J., Yuan, Z.G., 2013. A model-based assessment of nitric oxide and nitrous oxide production in membrane-aerated autotrophic nitrogen removal biofilm systems. J. Membr. Sci. 428, 163–171. Oremland, R.S., Hoeft, S.E., Santini, J.M., Bano, N., Hollibaugh, R.A., Hollibaugh, J.T., 2002. Anaerobic oxidation of arsenite in mono lake water and by a facultative, arsenite-oxidizing chemoautotroph, strain MLHE-1. Appl. Environ. Microbiol. 68, 4795–4802. 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. Raghoebarsing, A.A., Pol, A., van de Pas-Schoonen, K.T., Smolders, A.J.P., Ettwig, K.F., Rijpstra, W.I.C., Schouten, S., Damste, J.S.S., Op den Camp, H.J.M., Jetten, M.S.M., Strous, M., 2006. A microbial consortium couples anaerobic methane oxidation to denitrification. Nature 440, 918–921. Rahman, S., Kim, K.H., Saha, S.K., Swaraz, A.M., Paul, D.K., 2014. Review of remediation techniques for arsenic (As) contamination: a novel approach utilizing bio-organisms. J. Environ. Manag. 134, 175–185. Reichert, P., 1998. AQUASIM 2.0-Computer Program for the Identification and Simulation of Aquatic Systems. EAWAG, Dubendorf, Switzerland. Rhine, E.D., Phelps, C.D., Young, L.Y., 2006. Anaerobic arsenite oxidation by novel denitrifying isolates. Environ. Microbiol. 8, 899–908. Rittmann, B.E., 2007. The membrane biofilm reactor is a versatile platform for water and wastewater treatment. Environ. Eng. Res. 12, 157–175. Schnobrich, M.R., Chaplin, B.P., Semmens, M.J., Novak, P.J., 2007. Stimulating hydrogenotrophic denitrification in simulated groundwater containing high dissolved oxygen and nitrate concentrations. Water Res. 41, 1869–1876. Shi, Y., Hu, S.H., Lou, J.Q., Lu, P.L., Keller, J., Yuan, Z.G., 2013. Nitrogen removal from wastewater by coupling anammox and methane-dependent denitrification in a membrane biofilm reactor. Environ. Sci. Technol. 47, 11577–11583. Soares, M.I.M., 2002. Denitrification of groundwater with elemental sulfur. Water Res. 36, 1392–1395. Sun, W.J., Alvarez, R.S., Field, J.A., 2010. The role of denitrification on arsenite oxidation and arsenic mobility in an anoxic sediment column model with activated alumina. Biotechnol. Bioeng. 107, 786–794. Tanaka, M., Takahashi, Y., Yamaguchi, N., Kim, K.W., Zheng, G.D., Sakamitsu, M., 2013. The difference of diffusion coefficients in water for arsenic compounds at various pH and its dominant factors implied by molecular simulations. Geochim. Cosmochim. Acta 105, 360–371. Tang, Y., Ontiveros-Valencia, A., Feng, L., Zhou, C., Krajmalnik-Brown, R., Rittmann, B.E., 2013. A biofilm model to understand the onset of sulfate reduction in denitrifying membrane biofilm reactors. Biotechnol. Bioeng. 110, 763–772. Tang, Y.N., Zhao, H.P., Marcus, A.K., Krajmalnik-Brown, R., Bruce, E.R., 2012. A steady-state biofilm model for simultaneous reduction of nitrate and perchlorate, part 1: model development and numerical solution. Environ. Sci. Technol. 46, 1598–1607. Terada, A., Lackner, S., Tsuneda, S., Smets, B.F., 2007. Redox-stratification controlled biofilm (ReSCoBi) for completely autotrophic nitrogen removal: the effect of coversus counter-diffusion on reactor performance. Biotechnol. Bioeng. 97, 40–51. Winkler, M.K.H., Ettwig, K.F., Vannecke, T.P.W., Stultiens, K., Bogdan, A., Kartal, B., Volcke, E.I.P., 2015. Modelling simultaneous anaerobic methane and ammonium removal in a granular sludge reactor. Water Res. 73, 323–331. Yang, G.C.C., Lee, H.L., 2005. Chemical reduction of nitrate by nanosized iron: kinetics and pathways. Water Res. 39, 884–894. Zhang, J., Zhou, W.X., Liu, B.B., He, J., Shen, Q.R., Zhao, F.J., 2015a. Anaerobic arsenite oxidation by an autotrophic arsenite-oxidizing bacterium from an arseniccontaminated paddy soil. Environ. Sci. Technol. 49, 5956–5964. Zhang, L.L., Zhang, C., Hu, C.Z., Liu, H.J., Qu, J.H., 2015b. Denitrification of groundwater using a sulfur-oxidizing autotrophic denitrifying anaerobic fluidized-bed MBR: performance and bacterial community structure. Appl. Microbiol. Biotechnol. 99, 2815–2827. Zhu, I., Getting, T., 2012. A review of nitrate reduction using inorganic materials. Environ. Technol. Rev. 1, 46–58.