Simultaneous removal of sulfide, nitrate and acetate under denitrifying sulfide removal condition: Modeling and experimental validation

Journal of Hazardous Materials 264 (2014) 16–24

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Simultaneous removal of sulfide, nitrate and acetate under denitrifying sulfide removal condition: Modeling and experimental validation Xijun Xu a , Chuan Chen a , Aijie Wang a , Wanqian Guo a , Xu Zhou a , Duu-Jong Lee a,b,c,∗ , Nanqi Ren a,∗∗ , Jo-Shu Chang d a

State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, China Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan c Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei 106, Taiwan d Research Center for Energy Technology and Strategy, National Cheng Kung University, Tainan, Taiwan b

h i g h l i g h t s

g r a p h i c a l

• This work developed a mathematical

Model evaluation applied to case study 1: (A-G) S2− , NO3 − -N, NO2 − -N, and Ac− -C profiles under initial sulfide concentrations of 156.2 (A), 539 (B), 964 (C), 1490 (D), 342.7 (E), 718 (F), and 1140.7 (G) mg L−1 . The solid line represents simulated result and scatter represents experimental result.

model for DSR process. • Kinetics of sulfur–nitrogen–carbon and interactions between denitrifiers were studied. • Kinetic parameters of the model were estimated via data fitting. • The model described kinetic behaviors of DSR processes over wide parametric range.

a r t i c l e

i n f o

Article history: Received 23 August 2013 Received in revised form 22 October 2013 Accepted 24 October 2013 Available online 1 November 2013 Keywords: Denitrifying sulfide removal Heterotrophic denitrifier Autotrophic denitrifier Kinetics Mathematical modeling

a b s t r a c t

a b s t r a c t Simultaneous removal of sulfide (S2− ), nitrate (NO3 − ) and acetate (Ac− ) under denitrifying sulfide removal process (DSR) is a novel biological wastewater treatment process. This work developed a mathematical model to describe the kinetic behavior of sulfur–nitrogen–carbon and interactions between autotrophic denitrifiers and heterotrophic denitrifiers. The kinetic parameters of the model were estimated via data fitting considering the effects of initial S2− concentration, S2− /NO3 − -N ratio and Ac− -C/NO3 − -N ratio. Simulation supported that the heterotrophic denitratation step (NO3 − reduction to NO2 − ) was inhibited by S2− compared with the denitritation step (NO2 − reduction to N2 ). Also, the S2− oxidation by autotrophic denitrifiers was shown two times lower in rate with NO2 − as electron acceptor than that with NO3 − as electron acceptor. NO3 − reduction by autotrophic denitrifiers occurs 3–10 times slower when S0 participates as final electron donor compared to the S2− -driven pathway. Model simulation on continuous-flow DSR reactor suggested that the adjustment of hydraulic retention time is an efficient way to make the reactor tolerating high S2− loadings. The proposed model properly described the kinetic behaviors of DSR processes over wide parametric ranges and which can offer engineers with basis to optimize bioreactor operation to improve the treatment capacity. © 2013 Elsevier B.V. All rights reserved.

∗ Corresponding author at: State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, China. Tel.: +886 233663028. ∗∗ Corresponding author. E-mail addresses: [email protected], [email protected] (D.-J. Lee), [email protected] (N. Ren). 0304-3894/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2013.10.056

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24

1. Introduction Sulfide, nitrate and organic compounds are pollutants in receiving waters. Nitrogenous compounds in water bodies afford the risks associated with toxicity and bad odors in waters [1,2]. Sulfide is a toxic, corrosive and odorous compound that is harmful to human health that concentrations as low as 10 mg L−1 [3,4]. Biological process to remove these pollutants from water has received increasing interests [5]. Autotrophic denitrifiers use reduced sulfur compounds (S2− , S0 , S2 O3 2− ) as an energy source [6–16], while the heterotrophic denitrifiers use numerous organic compounds as electron donor [17,18]. Reyes-Avila et al. [1] proposed that the rate of nitrate reduction to nitrite is faster via autotrophic denitrification than via heterotrophic denitrification pathway, while the rate of nitrite reduction to nitrogen gas is faster via heterotrophic pathway than via autotrophic pathway. With sufficiently incubated symbiotic heterotrophs and autotrophs, these authors achieved simultaneous removal of sulfide, nitrogen and carbon under well defined denitrifying conditions at loading rates of 0.294 kg S m−3 d−1 , 0.209 kg N m−3 d−1 , and 0.303 kg C m−3 d−1 . Chen et al. [5] promoted these loadings to 3.0 kg S m−3 d−1 , 1.45 kg N m−3 d−1 , and 2.77 kg Ac m−3 d−1 using a novel sludge cultivation strategy. Show et al. [19] reviewed those studies for simultaneous removal of N, S and C. Kinetic models on heterotrophic denitrifiers alone [2,20–23] or autotrophic denitrifiers alone [24–29] were available. Simultaneous removal of sulfide, nitrate, and acetate in wastewater can be attributed to the mixed culture denitrification [1,2,14]. Limited efforts have been dedicated to model this mixed culture process. Few measurements, with most of them limited one substrate with the rest in excess were available for parameter estimation. Kinetic models on DSR process considering complex interactions between autotrophic and heterotrophic denitrifiers are desired. This study aims at developing a comprehensive kinetic model on autotrophic/heterotrophic denitrification dynamics. The kinetic processes for simultaneous sulfide, nitrate, and carbon (acetate) removal via mixed-culture autotrophic and heterotrophic denitrification were evaluated and validated by comparison with experimental data obtained in this work and in literature.

2. Materials and methods 2.1. Experimental setup and analysis Sludge sample was collected from a bench-scale expanded granular sludge bed (EGSB) which has been operated for more than six months. The EGSB reactor has a working volume of 4 liter, height of 120 cm and internal diameter of 50 mm. The temperature was maintained at 30 ± 1 ◦ C. The reactor was fed with a synthetic wastewater containing as follows (in mg L−1 ): S2− , 200; NO3 − -N, 105; Ac− -C, 113.4; NH4 Cl, 50; K2 HPO4 , 50; NaHCO3 , 1500; MgSO4 , 50; and trace element solution (in mg L−1 ): EDTA, 50; NaOH, 11; CaCl2 ·2H2 O, 7.34; FeCl2 ·4H2 O, 3.58; MnCl2 ·2H2 O, 2.5; ZnCl2 , 1.06; CoCl2 ·6H2 O, 0.5; (NH4 )6 Mo7 O24 ·4H2 O, 0.5; CuCl2 ·2H2 O, 0.14. To assess the effect of initial sulfide concentration on DSR process, seven different sets of experiments (I–VII) were conducted at substrate levels of 156, 539, 962, 1490 mg S2− L−1 for parameter values estimation and of 343, 718, 1140 mg S2− L−1 for model verification with S2− /NO3 − -N = 5/6 and Ac− -C/NO3 − -N = 1.26 to meet the stoichiometric requirement for maintaining high performance of DSR process (Table S1) [19]. The inoculums were sampled from the reactor under the steady state (variation of substrate effluent concentrations less than 10% over three HRT’s), centrifuged at 10,000 rpm for 20 min, and the pellets anaerobically collected were

17

washed twice with distilled water to remove residual substrates eliminating the disturbance of background. The batch tests were carried out in 250 mL serum bottles and inoculated with 200 mL inoculums with 10,000 mg L−1 volatile suspended solids (VSS). The bottles were flushed with argon gas for 5 min to remove oxygen from both the aqueous phase and headspace, and sealed with butyl rubber stoppers and aluminium crimps. Then, the sterile anaerobic stock solution was syringe injected to generate a final substrate concentration as listed above. All serum bottles were regularly sampled anaerobically for the analysis of S2− , NO3 − , NO2 − , Ac− and VSS. The experiments were conducted at 30 ◦ C and pH of 7.5. In order to assess the steady state of the system, successive batch experiments under each tested condition were carried out until the residual substrate concentrations in serum bottle varied less than 10% and then the averaged results of the steady state were recorded and reported herein. The effect of Ac− -C/NO3 − -N molar concentrations ratio (1.26, 2, and 3) on DSR process was determined by conducting additional experiments in serum bottles containing 200 mg L−1 S2− , 105 mg L−1 NO3 − -N, and 113.4, 180 or 270 mg L−1 Ac− -C. All other conditions were similar to those described above. To further validate the proposed model, another experiment was conducted with 200 mg L−1 S2− , 140 mg L−1 NO3 − -N, and 151.2, 240 or 360 mg L−1 Ac− -C (Table S1). Experimental data from Zhou et al. [30] concerning DSR process under different S2− /NO3 − -N molar ratio were used for testing the validity and applicability of the model. Zhou et al. [30] carried out a series of batch experiments with 200 mg L−1 S2− , 240 mg L−1 Ac− , and 194, 387.5, 542.5, 620 or 775 mg L−1 NO3 − giving sulfide to nitrate molar ratios of 5/2.5, 5/5, 5/7, 5/8 and 5/10 (Table S1) and the inoculum was obtained from Chen et al. [16]. The concentrations of sulfate, thiosulfate, nitrate, nitrite and acetate in liquor samples after 0.45 ␮m filtration were measured using ion chromatography (ICS-3000, Dionex, Bannockburn, IL, USA). Aqueous sulfide was determined spectrophotometrically (UV759S, Shanghai, China) with N, N dimethyl-p-phenylene diamine [31]. Concentrations of nitrogenous species (NO, N2 O, N2 ) were determined by gas chromatography (GC-6890, Agilent, Foster City, CA, USA). Both the measurement of VSS and COD were performed according to the standard methods [32]. 2.2. Model development A mathematical model was developed on the kinetics involving NO3 − , NO2 − , S2− and Ac− (Fig. S1) [1]. Although the heterotrophic denitratation (NO3 − to NO2 − ) and denitritation (NO2 − to N2 ) are complicated biological processes [23], the overall heterotrophic denitrification reaction was illustrated as that in Fig. S1. Moreover, since there is sufficient support for S0 as the dominant end-product of S2− oxidation at high S2− /NO3 − (or O2 ) ratio (about 1.6) [11,33–35], the autotrophic denitrifiers mediating SO4 2− production from S0 oxidation with NO3 − or NO2 − reduction was ignored. NO3 − was the preferred electron acceptor for S2− oxidation compared NO2 − [36], thus when NO3 − was presented in the medium, S2− oxidation coupling NO2 − reduction process was ignored (except for the S2− /NO3 − -N = 5:2.5 with insufficient nitrate as electron acceptor which would be discussed latter). The model describes the relationships between three bacterial groups: autotrophic denitrifiers (XA ), heterotrophic denitrifiers (XH ), and residual inert biomass (XI ) and four soluble compounds: NO3 − (concentration referred to as SNO− for the remaining part 3

of this paper), NO2 − (SNO− ), S2− (SS2− ), and Ac− (SAc− ). The units 2

are in mg-S L−1 for all sulfur species, mg-N L−1 for all nitrogen species, mg-COD L−1 for carbon species, and mg-VSS L−1 for each biomass. In the present model, there are two processes catalyzed by autotrophic denitrifiers and heterotrophic denitrification, modeled as two sequential processes with individual, reaction-specific

18

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24

Table 1 Best-fit parameters capturing mixed-culture denitrification obtained in three case studies. Parameter

Definition

Autotrophic denitrification Maximum reaction rate for S2− oxidation, NO3 − reduction to NO2 − , h−1 ␮BA R1 Maximum reaction rate for S2− oxidation, NO2 − reduction to N2 , h−1 ␮BA R2 KI,S2− R1 S2− inhibiting coefficient for NO3 − driven autotrophic denitrification, mg S L−1 NO2 − inhibiting coefficient for NO3 − driven autotrophic denitrification, mg N L−1 KI,NO2 − R1 KI,S2− R2 S2− inhibiting coefficient for NO2 − driven autotrophic denitrification, mg S L−1 NO2 − inhibiting coefficient for NO2 − driven autotrophic denitrification, mg N L−1 KI,NO2 − R2 ␮BA R5 Maximum reaction rate for S0 oxidation to SO4 2− , NO3 − reduction to NO2 − , h−1 Maximum reaction rate for S0 oxidation to SO4 2− , NO2 − reduction to N2 , h−1 ␮BA R6 KS0 R5 S0 affinity for S0 oxidation to SO4 2− , NO3 − reduction to NO2 − , mg S L−1 KS0 R6 S0 affinity for S0 oxidation to SO4 2− , NO2 − reduction to N2 , mg S L−1 KNO3 − R5 NO3 − affinity for S0 oxidation to SO4 2− , NO3 − reduction to NO2 − , mg N L−1 KI,NO2 − R6 NO2 − affinity for S0 oxidation to SO4 2− , NO2 − reduction to N2 , mg N L−1 NO3 − inhibiting coefficient for S0 oxidation to SO4 2− , NO2 − reduction to N2 , mg N L−1 KI,NO3 − R6

0 − + S2− + NO− 3 + 2H → S + NO2 + H2 O

S2−

(R1) −

S0

oxidation to with NO2 as electron acceptor Process 2: catalyzed by autotrophic denitrifiers, consuming S2− and NO2 − , and yielding biomass. 0 + 1.5S2− + NO− 2 + 4H → 1.5S + 0.5N2 + 2H2 O −

(R2)

−

Process 3: NO3 reduction to NO2 by heterotrophic denitrification, consuming NO3 − and Ac− , and yielding biomass. − − − NO− 3 + 0.25CH3 COO → NO2 + 0.5CO2 + 0.25H2 O + 0.25OH (R3)

Process 4: NO2 − reduction to N2 by heterotrophic denitrification, consuming NO2 − and Ac− , and yielding biomass. NO2 − + 0.375CH3 COO− + 0.125H2 O → 0.5N2 + 0.75CO2 + 1.375OH−

(R4)

Process 5: S0 oxidation to SO4 2− with NO3 − reduction to NO2 − when in excess by autotrophic denitrification, consuming S0 and NO3 − , and yielding biomass. S0 + 3NO3 − + H2 O + 3H+ → SO4 2− + 3NO2 − + 5H+ S0

2−

(R5)

−

Process 6: oxidation to SO4 with NO2 reduction to N2 by autotrophic denitrification, consuming S0 and NO2 − , and yielding biomass. S0 + 2NO2 − + 2H+ → SO4 2− + N2 + 2H+

(R6)

The oxidation rate of S2− to S0 with reduction NO3 − to NO2 − was higher than the oxidation of Ac− to reduce NO3 − to NO2 − , and the oxidation rate of S0 to reduce NO2 − to N2 is lower than the oxidation of Ac− for reducing NO2 − to N2 [1]. A substrate (S2− ) inhibition function was introduced to the Monod-expression describing heterotrophic denitrification (Table S2). Additionally, accumulated NO2 − may inhibit autotrophic denitrifiers leading to the process breakdown [15], so the corresponding kinetic rate incorporated the substrate inhibition function of NO2 − . The kinetics and stoichiometry of the interactions and transformations among model

Case 2

Case 3

0.245 0.135 2050 0.698 1.38 0.651

0.219

0.320

0.020 0.083 0.215 175 0.183 0.107 15.9

Heterotrophic denitrification Maximum reaction rate for heterotrophic NO3 − reduction, h−1 ␮BH R3 Maximum reaction rate for heterotrophic NO2 − reduction, h−1 ␮BH R4 KI,S2− R3 S2− inhibiting coefficient for heterotrophic NO3 − reduction, mg S L−1 S2− inhibiting coefficient for heterotrophic NO2 − reduction, mg S L−1 KI,S2− R4

kinetics (Table S2) similar to those described in ADM1 [37]. These processes include: Process 1: S2− oxidation to S0 with NO3 − as electron acceptor catalyzed by autotrophic denitrifiers, consuming S2− and NO3 − and yielding biomass.

Case 1

0.047 0.065 0.181 697

0.057 0.047

0.038 0.077

components were listed in Tables S2 and S3. In addition, mass balances for the autotrophic biomass, heterotrophic biomass and total amount of organics are made below to determine the autotrophic denitrifier concentration XA , and heterotrophic denitrifier concentration XH according to [38]. Autotrophic denitrifier, XA bXA +

1 XA = YA SG X

(R7)

Heterotrophic denitrifier, XH bXH +

1 XH = H XH X

(R8)

where b: cell lysis rate, h−1 ;  X : sludge age, d; YA : yield coefficient of autotrophic denitrifier, g VSS g−1 N; SG : sulfide oxidation rate for satisfying the energy demand by the growth process, mg-S L−1 d−1 ; H : specific growth rate of heterotrophic denitrifier, h−1 . To find appropriate values for new kinetics for DSR process by autotrophic and heterotrophic denitrifiers, the values for their kinetic parameters (see Table 1) were adjusted based on their sensitivity until the predicted dynamics of S2− , NO3 − , NO2 − and Ac− concentrations were matched to those measured. Mass balances for microbial growth, sulfide, nitrate, nitrite and acetate utilizations in a batch system were integrated based on each kinetic expression in Table S2 and these five equations were solved simultaneously to minimize the objective function (Eq. (R9) in Supplementary Information). Modeling and simulation are performed with a nonlinear least-squares algorithm in the software package AQUASIM [53], which offers a free definition of the biokinetic model, flow scheme, process control strategies, graphic support of the simulation, experimental data, and communication with spreadsheet programs. 3. Results Besides those typically used parameters to describe the autotrophic and heterotrophic denitrification [24,27,36,39–41], the model required process parameters to characterize the mixedculture denitrification. For autotrophic denitrification, they are ␮BA R1 , ␮BA R2 , ␮BA R5 , ␮BA R6 , KS2− R1 , KNO − R1 , KS2− R2 , KNO − R2 , KI,S2− R1 , KI,NO KI,NO

3

−

R6

2

−

R1

, KI,S2− R2 , KI,NO

2

−

␮BA

3

, KS0 R5 , KS0 R6 , KNO

3

R5

−

2

, KNO

2

−

R6

,

. Some of these parameters were scarcely reported in liter-

ature (␮BA R1 , ␮BA R2 , KI,S2− R1 , KI,NO R6 ,

R2

KS0

R5

, KS0

R6

, KNO

3

−

R5

2

−

, KNO

R1

2

−

, KI,S2− R2 and KI,NO

R6

, KI,NO

3

−

R6

2

−

R2

, ␮BA R5 ,

for autotrophic

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24 20

25

150

200

150

100

50

50

0

2400

5 20

10 40

15 Time(h)

20

60

30

100

10

15

20

8

1600

2- 2 S -R =0.997 2 NO3-N-R =0.974 2 NO2-N-R =0.931

1200 800

6

4

Ac--C-R2=0.983 2

400 0 0

20

40

0

20

40

60

80

100

0

100

40

45

50

55 1000

800

600

2 Ac -C-R =0.996

400

400

200

200

0

0

5

5

10

15

20

25

S2--R2=0.999 NO-3-N-R2=0.992 2 NO2-N-R =0.886

600

30

(E)

35

500

400

400

20

40

60

80

S2--R2=0.988 2 NO3-N-R =0.947 NO-2-N-R2=0.826 2 Ac -C-R =0.996

1600 1200

100 1600

(C) 1200

800

800 400

400

0

600

2 Ac -C-R =0.997

0

2000

0 0

10 15 20 25 30 35 40 45 50 55 Time(h)

1500

20

40 20

60 Time(h) 40

80 60

2- 2 S -R =0.993 2 NO3-N-R =0.994

1200

NO-2-N-R2=0.861 2 Ac -C-R =0.995

900

0

100 80 1000

(F) 800

600

300

600

400

300

200

200

200

100

0

0

5

10

15

20

Time(h) 80

35

600

Time(h) 60

30

(B)

(D)

2000

2000

25

800

800

10

5

2- 2 S -R =0.997 NO-3-N-R2=0.985 2 NO2-N-R =0.961

0 0

0

25

80

1000

250

100

0 0

Concentration(mg/L)

(A)

2 NO2-N-R =0.865 Ac--C-R2=0.999

200

0

30 300

2- 2 S -R =0.999 2 NO3-N-R =0.994

250 Concentration(mg/L)

15

Concentration(mg/L)

10

Concentration(mg/L)

5

Concentration(mg/L)

0

Concentration(mg/L)

300

19

25

30

0

35

0 0

20

40

60

0

80

Time(h)

10

(G) Concentration(mg/L)

1600

8

1200

6

800 400 0 0

20

40

2- 2 S -R =0.993 2 NO3-N-R =0.981

4

NO-2-N-R2=0.868 2 Ac -C-R =0.989

2

60 Time(h)

80

100

0

Fig. 1. Model evaluation applied to case study 1: (A–G) S2− , NO3 − -N, NO2 − -N, and Ac− -C profiles under initial sulfide concentrations of 156.2 (A), 539 (B), 964 (C), 1490 (D), 342.7 (E), 718 (F), and 1140.7 (G) mg L−1 .The solid line represents simulated result and scatter represents experimental result.

denitrification and ␮BA R3 , ␮BA R4 , KI,S2− R3 and KI,S2− R4 for heterotrophic denitrification). These parameters were estimated by data fitting. The other parameters were available from literature (Table S3). Data fitting was done by minimizing the sum of squares of the deviations between the measured data and the model simulations (available in Supplementary Information) [42]. The model was then used to simulate the process kinetics with the best-fit parameters in the following cases: case 1 for effects of initial S2− concentrations; case 2 for effect of S2− /NO3 − -N ratio; case 3 for effect of Ac− -C/NO3 − -N ratio. The best-fit parameters were listed in Table 1. 3.1. Case1: initial S2− concentrations Here, the model was applied to describe the effect of initial S2− concentrations on mixed culture denitrification with S2− /NO3 − N = 5/6 and Ac− -C/NO3 − -N = 1.26. The data fitting derived model parameters from individual batch experiments (Fig. 1A–D), including ␮BA R1 , ␮BA R2 , KI,S2− R1 , KI,NO − R1 , KI,S2− R2 , KI,NO − R2 for 2

2

autotrophic denitrification and ␮BH R3 , ␮BH R4 , KI,S2− R3 , KI,S2− R4 for heterotrophic denitrification (Table 1). The fittings were quite nice in correlation of the experimental data (Fig. 1A–D). In autotrophic denitrification, all S2− was removed at tested concentrations except for 1485 mg L−1 . The concentration of S2− did not markedly affect its elimination rate in the range being investigated. Rapid conversion of S2− may be explained by the low electron

equivalent required for oxidation of S2− to S0 coupled to the reduction of NO3 − to NO2 − (Eq. (R1)). Conversely, 718–1485 mg L−1 S2− partially inhibited NO3 − and Ac− conversion (Fig. 1). BeristainCardoso et al. [10] also noted significant inhibition of S2− on NO3 − conversion rate: increase in S2− from 80 to 320 mg L−1 decreased the denitrification rate by approximately 21 folds, from 99.82 to 4.34 mg L−1 NO3 − g−1 VSS d−1 . The sharp decrease in the rate of denitrification and acetate oxidation suggested that high levels of S2− were toxic to both autotrophic and heterotrophic denitrifiers [43–45]. Accumulation of NO2 − occurred temporarily during the tests (Fig. 1A–D), whose occurrence was delayed in the presence of high levels of S2− , likely a consequence of delay of NO3 − reduction. Conversely, reduction of NO2 − was not significantly affected by S2− at up to 718 mg L−1 (Fig. 1). However, 1140 mg L−1 S2− would lead to significant accumulation of NO2 − in the medium. High concentration of NO2 − was revealed to inhibit the activities of both autotrophic and heterotrophic denitrifiers [15]. The present model with best-fit parameters correlates well the system performances with different initial sulfide concentrations. As a validation, the so-obtained parameters were used to generate three additional curves for fitting the experiments with 343, 718 and 1140 mg L−1 S2− (Fig. 1E–G), the correlation between model output and experimental data was also satisfactory. Parameter sensitivity of the obtained process parameters was studied by procedures by Siegrist et al. [46] and Zhang et al. [47] (details in Supporting Materials). As initial sulfide concentrations

20

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24 0

2

4

6

8

10

12

14

16

(A)

250

300

0

5

10

15

(B)

S2--R2=0.984 2 NO3-N-R =0.982 2 NO2-N-R =0.730 Ac--C-R2=0.971

250

200

200

2- 2 S -R =0.998 NO-3-N-R2=0.972 NO-2-N-R2=0.851 Ac--C-R2=0.980

150

100

150

100

50

50

0

Concentrations (mg/L)

Concentrations (mg/L)

250 200 150

3

6

9

12

0

0

0

5

10

Time(h) 5

15

Concentrations (mg/L)

250

0

S2--R2=0.997 2 NO3-N-R =0.986 2 NO2-N-R =0.831 Ac--C-R2=0.980

200

150

(C)

200

150

100

50

50

0 10

10

2- 2 S -R =0.993 2 NO3-N-R =0.995 2 NO2-N-R =0.812 2 Ac -C-R =0.988

200

150

(D)

250

200

150

100

100

50

50

15

0

0

5

10

Time (h)

15

Time (h)

0

5

10 15

250

Concentrations (mg/L)

15

0

0

5

5

250

250

100

0

15

Time (h)

10

Concentrations (mg/L)

0

150

50

50

15

200

100

100

0

0

250

S2--R2=0.996 2 NO3-N-R =0.998

200

2 NO2-N-R =0.810 2 Ac -C-R =0.987

150

(E)

250

200

150

100

100

50

50

0

0

0

5

10

15

Time (h) Fig. 2. Model evaluation applied to case study 2: (A–E) effect of different S2− /NO3 − -N ratios (A, 5/2.5; B, 5/5; C, 5/7; D, 5/8; E, 5/10) with S2− and Ac− -C concentrations of 200 and 96 mg L−1 respectively by Zhou et al. [30]. The solid line represents simulated result and scatter represents experimental result.

and nitrite accumulation impacted strongly on the DSR process, this means that DSR process was greatly dependent on those parameters that were used by the “the uptake of sulfide” process (␮BA R1 , KI,NO − R1 ) and the “the uptake of nitrite” process (␮BH R4 , KI,S2− R4 ). In 2

this case, the parameters ␮BA R1 , ␮BH R4 , KI,S2− R4 , and KI,NO

2

−

R1

have

higher sensitivity than the others while the parameters KI,S2− R1 , ␮BH R3 , and KI,S2− R3 are not sensitive to the model [48] (Table S4).

3.2. Case 2: effect of S2− /NO3 − -N ratio The study by Zhou et al. [30] on the effects of S/N ratio on mixed-culture DSR process on the S2− -declining phase were used

for model validation (Fig. 2A–E). Experimental data in Fig. 2A and B were used to estimate the parameter values for substrate utilization and the so-obtained parameters were applied to verify the model with the other data sets (Fig. 2C–E). As S2− oxidation with NO2 − reduction to N2 was a significant pathway as experimental results revealed, parameter values for R2 were used for estimating parameters as employed in case 1. The values of the estimated parameters in this case were also listed in Table 1, which correlated with those obtained from case 1. As Fig. 2 shows, fully S2− oxidation was attained independent of S2− /NO3 − -N ratio. NO3 − consumption rate, on the other hand, was decreased with increasing S2− /NO3 − -N ratio. NO2 − accumulation was observed at all S2− /NO3 − -N ratio tested, with the peak concentration increasing with S2− /NO3 − -N ratio, and then

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24 0

5

10

20 300

(A)

250

Concentration(mg/L)

15

250

S2--R2=0.993 2 NO3-N-R =0.996 NO-2-N-R2=0.820 Ac--C-R2=0.995

200 150

Concentration(mg/L)

300

500

0

1

21 2

3

(B)

S2--R2=0.982 NO-3-N-R2=0.969 2 NO2-N-R =0.846 Ac--C-R2=0.991

400

200

300

400

300

200

200

100

100

100

50

50

0 0

5

10

0

0

15

20

0

0

1

2

Time(h) 700

0

1

2

3

4

(C)

600

700

400

0

2

4

3

Time(h)

6

400 300 200

400 300

100

0

0

1

240

160

2

3

4

2

3

4

(E)

0

0

2

4

1000

Concentration(mg/L)

200

400

600

400

200

100

0

0

1

2

8

0

10

3

4

Time(h)

0.5

1.0

1.5

2.0

2.5

800

2- 2 S -R =0.984 NO-3-N-R2=0.965 NO-2-N-R2=0.744 2 Ac -C-R =0.998

600

400

200

0 0.0

3.0 1000

(F)

800

300

100

0

6

600

500

2- 2 S -R =0.982 2 NO3-N-R =0.981 NO2-N-R2=0.750 Ac--C-R2=0.992

300

160

Time(h)

500 400

240

80

0.0

1

320

80

Time(h) 6000

400

200

100

0

10

2- 2 S -R =0.993 NO-3-N-R2=0.974 2 NO2-N-R =0.733 2 Ac -C-R =0.989

500

2- 2 S -R =0.976 2 NO3-N-R =0.882 NO-2-N-R2=0.832 2 Ac -C-R =0.950

8

600

Concentration(mg/L)

500

4

(D)

320

Concentration(mg/L)

500

150

100

Concentration(mg/L)

4

200

0.5

1.0

1.5

2.0

2.5

0

3.0

Time(h)

Fig. 3. Model evaluation applied to case study 3: (A–F) effect of different Ac− -C/NO3 − N ratios (A, 1.26; B, 2; C, 3) with a S2− /NO3 − -N ratio of 5/6 obtained herein; (I–K) effect of different Ac− -C/NO3 − N ratios (D, 1.26; E, 2; F, 3) with a S2− /NO3 − -N ratio of 5/8.The solid line represents simulated result and scatter represents experimental result.

exhausted, indicating that heterotrophic reduction of NO2 − to N2 was affected by the S2− /NO3 − -N ratio. The S2− /NO3 − -N ratio has a strong impact on the reactions involving S2− [9,49,50]. Also, elemental sulfur is an expected product of partial oxidation of S2− when electron acceptor supply is limited [51,52]. The present model properly described the experimental data with the best-fit parameters from case 1 and case 2.

3.3. Case 3: effect of Ac− -C/NO3 − -N ratio The mixed-culture denitrification was experimentally monitored at different Ac− -C/NO3 − -N ratios with data being used

for parametric fitting. The parameters from case 3 were also listed in Table 1, which correlated with those obtained from case 1.The model output properly described the kinetics in the batch experiments (Fig. 3A–C). The contents of organic carbon impact significantly the activities of heterotrophic denitrifiers and thus deteriorate the balanced growths of autotrophic and heterotrophic denitrifiers [5]. The lag phase of S2− oxidation (by autotrophic denitrification) was increased with increasing Ac− -C/NO3 − -N ratio, and the increase in Ac− consumption indicated that heterotrophic denitrifiers rather than autotrophic denitrifiers dominated at high Ac− -C/NO3 − -N ratios. The resulting parameter values were used to generate three additional sets of curves for the comparison with three other data sets (S/N ratio of 5:8 with Ac− -C/NO3 − -N ratio of 1.26, 2, and 3 in

22

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24 0

Fig. 3D–F). The satisfactory agreement between experimental data and the model output verifies the capability of the present model for describing effects of Ac− -C/NO3 − -N ratio on DSR processes.

Removal efficiency (%)

Simultaneous removal of sulfide, nitrate, and acetate in wastewater can be attributed to the mixed culture denitrification [1,5,14]. Limited efforts have been dedicated to model this mixed culture process. Most available modeling works were on either autotrophic denitrification or heterotrophic denitrification, which were commonly empirical and casespecific, without general applicability. Conversely, the model presented herein is a comprehensive one considering conversion of sulfur, nitrogenous and carbonaceous compounds in the DSR reactor. The model parameters presented in the present study (Table 1) are of practical interest such as the development of models involving sulfur and nitrogenous compounds cycles. Discrepancy in parametric values may be present when different microbial communities were involved; however, since the present inoculums was from environmental samples, the apparent kinetic parameters thus obtained should be “typical” amongst microbial communities existing in environment. The obtained ␮BH R3 and ␮BH R4 are of the same order of magnitude and commensurate with the typical maximum denitrification reaction rate reported [2,39]. Similarly, the maximum growth rates for autotrophic denitrification, ␮BA R1 and ␮BA R2 , are close to those reported in literature (0.11–0.20 h−1 ) [24,26,27]. The calibrated ␮BH R3 and ␮BH R4 were 0.020 and 0.083 h−1 respectively, indicating far lower (3–10 times) NO3 − reduction rates when S2− vs S0 is the final electron donor, consistent with Beristain-Cardoso et al. [9]. Cai et al. [36] reported that weak inhibition effects were noted for nitrate and sulfide as substrates and strong inhibition effects for nitrite and sulfide systems. However, literature works regarding inhibition effects of sulfide on autotrophic denitrification are limited. The sulfide inhibition parameter obtained for denitratation (0.181 mg S2− L−1 ) was approximately three orders of magnitude less than that for denitritation (697 mg S2− L−1 ). This difference reflects the denitratation step (NO3 − reduction to NO2 − ) was inhibited first by S2− , consistent with the observation in case 1.The calibrated S2− inhibition parameter obtained for autotrophic NO3 − driven denitrification (2050 mg S2− L−1 ) and autotrophic NO2 − driven denitrification (1.38 mg S2− L−1 ) were significantly different, reflecting that NO3 − is the preferred electron acceptor than NO2 − for sulfide oxidation. The present model simulated the DSR performances in a continuous-flow system. The mass-balance equations of each substance in Eqs. (R1)–(R4) were with QC0 /V − QC/V; where V is the liquid volume, Q is the flow rate, and C0 and C represent substrate concentrations in influent and inside reactor, respectively. As the value of decay coefficient determined from the batch data was small, decay term was not included. Mass-balance equations were solved simultaneously to determine the transient concentrations of biomass, sulfide, nitrate, nitrite and acetate at each substrate loading rate. The theoretical steady-state concentrations at each substrate loading rate were determined by conducting the calculations in the transient model over a sufficiently long time [2]. In order to predict reactor operating under different substrate loading rates, the rate term in each mass balance and the parameters for model simulation were the same as for case 1. During simulation the loading rates were changed via varying influent substrate concentrations or hydraulic retention time (HRT). For changing HRT (A), the initial conditions for simulation were

22

3

44

5

66

7

88

910111213141516 10 12

14

16

(A)

100

2-

S

100

100 NO-3-N

COD 80

80 80

60

60 60

40

40 40

20

20 20

0 0

2

4

0

2

4

6

8

10

12

Simulation number 6

8

10

12

14

16

14

16

(B)

100

Removal efficiency (%)

4. Discussion

1

0

2-

S

-

NO3-N COD

80

100

80

60

60

40

40

20

20

0 0

2

4

6

8

10

12

14

16

0

Simulation number Fig. 4. Predicted mixed culture denitrification in continuous reactors performing simultaneous removal of sulfide, nitrate and acetate under different substrate loading rates: for changing HRT (A), the initial conditions for simulation were reactor volume = 4 L, influent S2− = 718 mg L−1 , NO3 − –N = 406 mg L−1 , Ac− C = 1150 mg COD L−1 , with HRT range from 1 to 48 h; for changing influent substrates concentrations (B), the initial conditions for simulation were reactor volume = 4 L, HRT = 12 h, influent substrates concentration of 269–8620 mg S2− L−1 , 152–4870 mg NO3 − -N L−1 , and 431–13,800 mg COD L−1 . The set loading rates were detailed in Table S5.

reactor volume = 4 L, influent S2− = 718 mg L−1 , NO3 − −1 − −1 N = 406 mg L , Ac -C = 1149 mg COD L , with HRT 1–48 h. For changing influent substrate concentrations (B), the initial conditions for simulation were reactor volume = 4 L, HRT = 12 h, influent substrate concentration of 269–8616 mg S2− L−1 , 152–4870 mg NO3 − -N L−1 , and 431–13,790 mg COD L−1 . The detailed information could be found in Table S5. The simulation period was 30 days. The simulated results (Fig. 4) indicated that high substrate loading rates and/or high influent S2− would reduce substrate removal efficiencies. The consumption rates of NO3 − -N and Ac− -C were more sensitive to high loading rates, indicating that the activities of heterotrophic denitrifiers were easily affected by the high loading rates. Adjusting HRT was shown as a favorable option for tolerating high loadings to maintain reactor performance. Also, the toxicity of S2− was shown as an important role determining reactor performance. The practical implication of these predictions was that the treatment of sulfur, nitrogen and carbon containing wastewaters by mixed culture denitrification should be operated under conditions ensuring that the effective steady-state sulfide concentration in the reactor is below a critical level of strong inhibition [9]. Table S5 lists the

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24

performance of mixed culture denitrification under substrate loading rates reported in the literature for comparison with our model predictions (Fig. 4). The simulation results were in general accordance with the reported results [15]. Overall, the present kinetic model properly described the kinetic behavior of typical DSR process with best-fit kinetic parameters reported. The effects of initial S2− concentrations, S2− /NO3 − -N and Ac− -C/NO3 − -N ratios were used to validated the derived model. To void the rate constants of autotrophic denitrification (R1 and R2) the model can be simplified to heterotrophic denitrification process; while to set the rate constants for R3 and R4 as zero can the model be simplified to autotrophic denitrification process. This model should be useful for further process development and online monitoring and control for sulfide and/or nitrate involving processes. 5. Conclusions This study modeled the DSR process considering simultaneous removal of sulfide, nitrate, and acetate in wastewater. The kinetic behaviors of sulfur, nitrogen and carbon compounds with growths of autotrophic denitrifiers and heterotrophic denitrifiers, were considered. The model successfully describes the effects of initial S2− concentration, S2− /NO3 − -N ratio and Ac− -C/NO3 − -N ratio on DSR performance. Additionally, the model output suggested possible way of mitigating toxicity of sulfide on DSR consortium. The developed model presents a practical tool for development, optimization and control of sulfur–nitrogen–carbon removal process. Acknowledgements This research was supported by National High-tech R&D Program of China (863 Program, Grant no. 2011AA060904), by the National Natural Science Foundation of China (Grant no. 51176037 and 51308147), Project 51121062 (National Creative Research Groups), by the State Key Laboratory of Urban Water Resource and Environment (2012DX06), by the Academician Workstation Construction in Guangdong Province (2012B090500018). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat. 2013.10.056. References [1] J. Reyes-Avila, E. Razo-Flores, J. Gomez, Simultaneous biological removal of nitrogen, carbon and sulfur by denitrification, Water Res. 38 (2004) 3313–3321. [2] S.J. An, H. Stone, M. Nemati, Biological removal of nitrate by an oil reservoir culture capable of autotrophic and heterotrophic activities: kinetic evaluation and modeling of heterotrophic process, J. Hazard. Mater. 190 (2011) 686–693. [3] J.M. Visser, L.A. Robertson, H.W. van Verseveld, J.G. Kuenen, Sulfur production by obligately chemolithoautotrophic Thiobacillus species, Appl. Environ. Microbiol. 63 (1997) 2300–2305. [4] J. Garcia-de-Lomas, A. Corzo, M.C. Portillo, J.M. Gonzalez, J.A. Anderades, C. Saiz-Jimenez, E. Garcia-Robledo, Nitrate simulation of indigenous nitratereducing, sulfide-oxidizing bacterial community in wastewater anaerobic biofilms, Water Res. 41 (2007) 3121–3131. [5] C. Chen, A.J. Wang, N.Q. Ren, H. Kan, D.J. Lee, Biological breakdown of denitrifying sulfide removal process in high-rate expanded granular sludge bed reactor, Appl. Microbiol. Biotechnol. 81 (2008) 765–770. [6] K. Tang, V. Baskaran, M. Nemati, Bacteria of the sulfur cycle: an overview of microbiology, biokinetics and their role in petroleum and mining industries, Biochem. Eng. J. 44 (2009) 73–94. [7] K. Tang, S. An, M. Nemati, Evaluation of autotrophic and heterotrophic processes in biofilm reactors used for removal of sulfide, nitrate and COD, Bioresour. Technol. 101 (2010) 8109–8118. [8] E. Vaiopoulou, P. Melidis, A. Aivasidis, Sulfide removal in wastewater from petrochemical industries by autotrophic denitrification, Water Res. 39 (2005) 4101–4109.

23

[9] R. Beristain-Cardoso, R. Sierra-Alvarez, P. Rowlette, E.R. Flores, J. Gomez, J.A. Field, Sulfide oxidation under chemolithoautotrophic denitrifying conditions, Biotechnol. Bioeng. 95 (2006) 1148–1157. [10] R. Beristain-Cardoso, A.C. Texier, R. Sierra-Alvarez, J.A. Field, E. Razo-Flores, J. Gomez, Simultaneous sulfide and acetate oxidation under denitrifying conditions using an inverse fluidized bed reactor, J. Chem. Technol. Biotechnol. 83 (2008) 1197–1203. [11] S. Gadekar, M. Nemati, G. Hill, Batch and continuous biooxidation of sulfide by Thiomicrospira sp. CVO: reaction kinetics and stoichiometry, Water Res. 40 (2006) 2436–2446. [12] Q. Mahmood, P. Zheng, J. Cai, D. Wu, B. Hu, J. Li, Anoxic sulfide oxidation using nitrite as electron acceptor, J. Hazard. Mater. 147 (2007) 249–256. [13] R. Sierra-Alvarez, R. Beristain-Cardoso, M. Salazar, J. Gomez, E. Razo-Flores, J.A. Field, Chemolithotrophic denitrification with elemental sulfur for groundwater treatment, Water Res. 41 (2007) 1253–1262. [14] S.J. An, K. Tang, M. Nemati, Simultaneous biodesulfurization and denitrification using an oil reservoir microbial culture: effects of sulfide loading rate and sulfide to nitrate loading ratio, Water Res. 44 (2010) 1531–1541. [15] C. Chen, N.Q. Ren, A.J. Wang, Z.G. Yu, D.J. Lee, Simultaneous biological removal of sulfur, nitrogen and carbon using EGSB reactor, Appl. Microbiol. Biotechnol. 78 (2008) 1057–1063. [16] C. Chen, A.J. Wang, N.Q. Ren, D.J. Lee, J.Y. Lai, High-rate denitrifying sulfide removal process in expanded granular sludge bed reactor, Bioresour. Technol. 100 (2009) 2316–2319. [17] M.R. Betlach, J.M. Tiedje, Kinetic explanation for accumulation of nitrite, nitric oxide, and nitrous oxide during bacteria denitrification, Appl. Environ. Microbiol. 42 (1981) 1074–1084. [18] J.C. Akunna, C. Bizeau, R. Moletta, Denitrification in anaerobic digester: possibilities and influence of wastewater COD/N-NOx ratio, Environ. Technol. 13 (1992) 825–836. [19] K.Y. Show, D.J. Lee, X.L. Pan, Simultaneous biological removal of nitrogen–sulfur–carbon: recent advances and challenges, Biotechnol. Adv. 31 (2013) 409–420. [20] I.A. Vasilliadou, S. Pavlou, D.V. Vayenas, A kinetic study of hydrogenotrophic denitrification, Process Biochem. 41 (2006) 1401–1408. [21] I. Ersever, V. Ravindran, M. Pirbazari, Biological denitrification of reserve osmosis brine concentrations: I. Batch reactor and chemostat studies, Environ. Eng. Sci. 6 (2007) 503–518. [22] E. Ficara, R. Canziani, Monitoring denitrification by pH-stat titration, Biotechnol. Bioeng. 98 (2007) 368–377. [23] O. Soto, E. Aspe, M. Roeckel, Kinetics of cross-inhibited denitrification of a high load wastewater, Enzyme Microb. Technol. 40 (2007) 1627–1634. [24] S.E. Oh, K.S. Kim, H.C. Choi, I.S. Kim, Kinetics and physiological characteristics of autotrophic denitrification by denitrifying sulfur bacteria, Water Sci. Technol. 42 (2000) 59–68. [25] J.L. Campos, S. Carvalho, R. Portela, A. Mosquera-Corral, R. Mendez, Kinetic of denitrification using sulfur compounds: effects of S/N ratio, endogenous and exogenous compounds, Bioresour. Technol. 99 (2008) 1293–1299. [26] G. Claus, H.J. Kutzner, Physiology and kinetics of autotrophic denitrification by Thiobacillus denitrificans, Appl. Microbiol. Biotechnol. 22 (1985) 283–288. [27] H. Zeng, T.C. Zhang, Evaluation of kinetic parameters of a sulfur–limestone autotrophic denitrification biofilm process, Water Res. 39 (2005) 4941–4952. [28] H.S. Moon, K.H. Ahn, S. Lee, K. Nam, J.Y. Kim, Use of autotrophic sulfur-oxidizers to removal nitrate from bank filtrate in a permeable reactive barrier system, Environ. Pollut. 129 (2004) 499–507. [29] A. Koenig, L.H. Liu, Kinetic model of autotrophic denitrification in sulfur packedbed reactors, Water Res. 35 (2001) 1969–1978. [30] X. Zhou, L.H. Liu, C. Chen, N.Q. Ren, A.J. Wang, D.J. Lee, Reduction of produced elementary sulfur in denitrifying sulfide removal process, Appl. Microbiol. Biotechnol. 90 (2011) 1129–1136. [31] A. Grigoryan, A. Lambo, S. Lin, S.L. Cornisn, T.R. Jack, G. Voordouw, Souring remediation by field-wide nitrate injection in an Alberta oil field, J. Can. Petrol. Technol. 48 (2009) 58–61. [32] APHA, Standard Methods for the Examination of Water and Wastewater, 19th ed., American Public Health Association, New York, 1995. [33] A.J.H. Janssen, R. Sleyster, C. van der Kaa, A. Jochemsen, J. Bontsema, G. Lettinga, Biological sulfide oxidation in a fed-batch reactor, Biotechnol. Bioeng. 47 (1995) 327–333. [34] S. Alcantara, A. Velasco, A. Munoz, J. Cid, S. Revah, E. Razo-Flores, Hydrogen sulfide oxidation by a microbial consortium in a recirculation reactor system: sulfur formation under oxygen limitation and removal of phenols, Environ. Sci. Technol. 38 (2004) 918–923. [35] A.J. Wang, D.Z. Du, N.Q. Ren, J.W. van Groenestign, An innovative process of simultaneous desulfurization and denitrification by Thiobacillus denitrificans, J. Environ. Sci. Health A 40 (2005) 1939–1949. [36] J. Cai, P. Zheng, M. Qaisar, Influence of various nitrogenous electron acceptors on the anaerobic sulfide oxidation, Bioresour. Technol. 101 (2010) 2913–2937. [37] D.J. Batston, J. Keller, I. Angelidaki, S.V. Kalyuzhnyi, S.G. Pavlostathis, A. Rozzi, W.T.M. Sanders, H. Siegrist, V.A. Vavilin, Anaerobic Digestion Model No.1 (ADM1), IWA Publishing, London, 2002, IWA Scientific and Technical Report No.1. [38] V.M. Vadivelu, Z.G. Yuan, C. Fux, J. Keller, Stoichiometric and kinetic characterization of Nitrobacter in mixed culture by decoupling the growth and energy generation process, Biotechnol. Bioeng. 94 (2006) 1176–1188.

24

X. Xu et al. / Journal of Hazardous Materials 264 (2014) 16–24

[39] R. von Schulthess, W. Gujer, Release of nitrous oxide (N2 O) from denitrifying activated sludge: verification and application of a mathematical model, Water Res. 30 (1996) 521–530. [40] M. Henze, W. Gujer, T. Matsuo, M.C.M. van Loosdrecht, Activated Sludge Models ASM1, ASM2, ASM2d and ASM3, Scientific and Technical Reports, IWA Publishing, London, 2000. [41] M. de Graaff, J.B.M. Klok, M.F.M. Bijmans, G. Muyzer, A.J.H. Janssen, Application of a 2-step process for the biological treatment of sulfidic spent caustics, Water Res. 46 (2012) 723–730. [42] B.J. Ni, M. Ruscalleda, C. Pellicer-Nacher, B.F. Smets, Modeling nitrous oxide production during biological nitrogen removal via nitrification and denitrification: extensions to the general ASM models, Environ. Sci. Technol. 45 (2011) 7768–7776. [43] R. Knowles, Denitrification, Microbiol. Rev. 46 (1982) 43–70. [44] B. Schonharting, R. Rehner, J.W. Metzger, K. Krauth, M. Rizzi, Release of nitrous oxide (N2 O) from denitrifying activated sludge caused by H2 S-containing wastewater: quantification and application of a new mathematical model, Water Sci. Technol. 38 (1998) 237–246. [45] J. Sorensen, J.M. Tiedje, R.B. Firestone, Inhibition by sulfide of nitric and nitrous-oxide reduction by denitrifying Pseudomonas fluorescens, Appl. Environ. Microbiol. 39 (1980) 105–108. [46] H. Siegrist, D. Vogt, J.L. Garcia-Heras, W. Gujer, Mathematical model for mesoand thermophilic anaerobic sewage sludge digestion, Environ. Sci. Technol. 36 (2002) 1113–1123.

[47] P. Zhang, Y.G. Chen, Q. Zhou, X. Zheng, X.Y. Zhu, Y.X. Zhou, Understanding short-chain fatty acids accumulation enhanced in waste activated sludge alkaline fermentation: kinetics and microbiology, Environ. Sci. Technol. 44 (2010) 9343–9348. [48] M.A.J. Huijbregts, W. Gilijamse, A.M.J. Ragas, L. Reijnders, Evaluating uncertainty in environmental life-cycle assessment. A case study comparing two insulation options for a Dutch one-family dwelling, Environ. Sci. Technol. 37 (2003) 2600–2608. [49] P.J.F. Gommers, W. Bijleveld, J.G. Kuenen, Simultaneous sulfide and acetate oxidation in a denitrifying fluidized bed reactor, I: start-up and reactor performance, Water Res. 22 (1988) 1075–1083. [50] B. Krishnakumar, B.V. Manilal, Bacterial oxidation of sulfide under denitrifying conditions, Biotechnol. Lett. 21 (1999) 437–440. [51] D.P. Kelly, J.K. Shergill, W.P. Lu, A.P. Wood, Oxidative metabolism of inorganic sulfur compounds by bacteria, Antonie Leeuwenhoek 71 (1997) 95–108. [52] C. Buisman, G. Geraats, P. Ijspeert, G. Lettinga, Optimization of sulfur production in a biotechnological sulfide-removing reactor, Biotechnol. Bioeng. 35 (1990) 50–56. [53] P. Reichert, AQUASIM 2.0-User Manual, Computer Program for the Identification and Simulation of Aquatic Systems, Swiss Federal Institute for Environmental Science and Technology (EAWAG), Dubendorf, Switzerland, 1998.