- Email: [email protected]
Simultaneous biological removal of nitrogen, carbon and sulfur by denitrification
ARTICLE IN PRESS
Water Research 38 (2004) 3313–3321
Simultaneous biological removal of nitrogen, carbon and sulfur by denitrification Jesu´s Reyes-Avilaa, El!ıas Razo-Floresa,1, Jorge Gomezb,* b
a Instituto Mexicano del Petro´leo, Programa de Biotecnolog!ıa, Eje Central La´zaro Ca´rdenas 152, C.P. 07730, Mexico Departamento de Biotecnolog!ıa, Universidad Auto´noma Metropolitana-Iztapalapa, Divisio´n CBS, San Rafael Atlixco 186, Col. Vicentina, C.P. 09340, Mexico
Received 7 March 2003; received in revised form 20 April 2004; accepted 29 April 2004
Abstract Refinery wastewaters may contain aromatic compounds and high concentrations of sulfide and ammonium which must be removed before discharging into water bodies. In this work, biological denitrification was used to eliminate carbon, nitrogen and sulfur in an anaerobic continuous stirred tank reactor of 1.3 L and a hydraulic retention time of 2 d. Acetate and nitrate at a C/N ratio of 1.45 were fed at loading rates of 0.29 kg C/m3 d and 0.2 kg N/m3 d, respectively. Under steady-state denitrifying conditions, the carbon and nitrogen removal efficiencies were higher than 90%. Also, under these conditions, sulfide (S2) was fed to the reactor at several sulfide loading rates (0.042–0.294 kg S2/m3 d). The high nitrate removal efficiency of the denitrification process was maintained along the whole process, whereas the carbon removal was 65% even at sulfide loading rates of 0.294 kg S2/m3 d. The sulfide removal increased up to B99% via partial oxidation to insoluble elemental sulfur (S0) that accumulated inside the reactor. These results indicated that denitrification is a feasible process for the simultaneous removal of nitrogen, carbon and sulfur from effluents of the petroleum industry. r 2004 Elsevier Ltd. All rights reserved. Keywords: Denitrification; Anoxic sulfide oxidation; Nitrate reduction; Sulfur production
1. Introduction At present, the contamination by carbon, nitrogen and sulfur compounds in wastewaters and water bodies is a critical problem. Some wastewaters as those from the oil industry represent a tremendous challenge for
*Corresponding author. Tel.: +52-55-5804-6408; fax: +5255-5804-6407. E-mail address: [email protected] (J. Gomez). 1 Present address: Instituto Potosino de Investigacio´n Cient!ıfica y Tecnolo´gica A.C., Departamento de Ingenier!ıa Ambiental y Manejo de Recursos Naturales, Camino a la Presa San Jose´ 2055, Col. Lomas 4a, Seccio´n, C.P. 78216, San Luis Potos!ı, SLP, Mexico.
treatment before discharge because of its chemical complexity. These effluents may contain a high concentration of organic compounds such as phenol and cresols that can be mineralized by biological processes. Likewise, some inorganic compounds like sulfide and ammonia are also frequently found. Nitrogen compounds contribute mainly to eutrophication of water bodies, besides the risks associated with toxicity and bad odors [1,2]. Sulfide is a very toxic compound for many microorganisms, even at concentrations as low as 10 mg/ L; this is due to the fact that sulfide reacts with the iron from cytochromes inhibiting the respiration [3]. Additionally, it is corrosive and possesses a high chemical oxygen demand (COD). Ammonia, as well as sulfide, has a high COD and is toxic for the aquatic fauna even at concentrations of 4 mg/L.
0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2004.04.035
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3314
In order to remove these contaminants, biological treatments are preferred technologies rather than physical–chemical methods, which are expensive and may generate toxic residuals [4]. Ammonium can be biologically oxidized to nitrate under aerobic conditions and subsequently reduced to molecular nitrogen (N2) via denitrification under anoxic conditions. The heterotrophic denitrification process uses many organic compounds as carbon and energy sources [5]. Cervantes et al. [6] reported that for high denitrifying efficiencies the C/N ratio was the main parameter of control to achieve a dissimilative respiratory process. The stoichiometric reaction between acetate and nitrate (C/N ratio of 1.07) is shown in Eq. (1):
biological activity of the biomass will also determine the rate of the reaction. The objective of this work was to evaluate the simultaneous biological removal of nitrogen (nitrate), carbon (acetate) and sulfur (sulfide) under well defined denitrifying conditions. First, a continuous stirred tank reactor cultivating a heterotrophic population was operated. After the reactor reached steady-state denitrification, sulfide was fed to the reactor at several sulfide-loading rates to evaluate the performance of the reactor. Finally, batch experiments were conducted measuring the specific removal rates and the influence of the abiotic reactions.
1:25CH3 COOH þ 2NO 3 -2:5CO2 þ N2 þ 1:5H2 O þ 2OH ;
2. Materials and methods
0
DG0 ¼ 1054:8 kJ=reaction:
2.1. Denitrifying reactor ð1Þ
The heterotrophic denitrification can be a highrate process. Cuervo-Lopez et al. [7] reported that for a denitrifying sludge in the presence of acetate, a C/N 3 ratio of 2 and a nitrate loading rate of 2 kg NO 3 -N/m d, a nitrate removal efficiency of 100% and a denitrifying yield (Y-N2, g N2/g NO 3 -N consumed) of 0.9 were obtained. Bernet et al. [8] and Chen et al. [9] applied 3 nitrate loading rates above 2.1 kg NO 3 -N/m d, but the nitrate removal efficiencies were around 70%. Lithotrophic denitrification using reduced sulfur compounds as an energy source has also been observed. A culture of Thiobacillus denitrificans, at low C/N ratio, used S0, thiosulfate and sulfide as electron donors for nitrate reduction [10–14]. Generally, CO2 is the carbon source and the final products of the autotrophic process are sulfate and N2. Eq. (2) shows the reduction of nitrate using sulfide as the energy source (S/N ratio of 1.43): 2 þ 1:25S2 þ 2NO 3 þ 2H -1:25SO4 þ N2 þ H2 O; 0
DG0 ¼ 972:8 kJ=reaction:
ð2Þ
Scarce evidence exists about the occurrence of this lithoautotrophic process when organic matter is present. Gommers et al. [15] used a fluidized bed reactor to study the effect of sulfide and acetate on denitrification under limited conditions of both substrates. The authors observed that nitrate reduction was partial as nitrite accumulated in the system. Acetate was consumed at high efficiencies and sulfide was completely oxidized to sulfate. Nevertheless, not all the end products of the biological reactions were determined. More recently, similar studies have been conducted, but elemental sulfur was used as electron donor instead sulfide [16,17]. As denitrification is a redox process, the thermodynamic reactions involved (using acetate and sulfide) should influence the overall efficiency of the process. The
An anaerobic continuous stirred tank reactor with a biomass retention device and a working volume of 1.3 L was used as illustrated in Fig. 1. The reactor was instrumented to control: temperature (30 C), agitation rate (250 rpm) and pH (8.370.2). Gas production rate was measured in a calibrated column by liquid displacement. The reactor was inoculated with 0.13 L of methanogenic sludge giving a biomass concentration of 1.75 g volatile suspended solids (VSS)/L. Nitrate and acetate (org-C) were fed separately to the reactor using two media, named as medium 1 (M1) and medium 2 (M2). The chemical composition of M1, was (g/L): CH3COONa 3H2O, 12.4; CaCl2 2H2O, 1.0; Na2MoO4 2H2O, 0.05. The chemical composition of M2 was (g/L): NaNO3, 7.6; KH2PO4, 1.5; MgSO4 7H2O, 1.45; FeCl3 6H2O, 0.01; CuSO4 5H2O, 0.03. The flow rate of each media was 0.32 L/d (total flow rate 0.64 L/d) and the hydraulic retention time (HRT) was of 2 d. During the start up the C/N ratio was 1.75, being adjusted afterwards to 1.45 and maintaining the reactor in operation under this condition for more than 90 d. 2.2. Denitrifying sulfide oxidizing reactor In order to avoid production and precipitation of metallic sulfides, adjustments were made to M1 and M2 as follows: Sodium sulfide (Na2S 9H2O) was added to medium 1 as alternative energy source, whereas sulfate (MgSO4 7H2O) concentration was reduced from 1.45 to 0.5 and 0.48 g/L of MgCl2 6H2O was added for complementing the magnesium. Finally, Na2MoO4 2H2O was added to M2 instead M1. The C/ N ratio in the reactor was fixed to 1.45. Organic carbon (as sodium acetate) and nitrogen loading rates were 303 mg org-C/L d and 209.4 mg NO 3 -N/L d, respectively. The pH in the reactor was controlled at 8.370.2 and the HRT was kept at 2 d. Five sulfide loading rates
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3315
Fig. 1. Scheme of the denitrifying continuous stirred reactor used in the experiments.
were applied to the reactor: 42.2, 83.6, 167.1, 258.3 and 294 mg S2/L d. The last value is equivalent to a S/N stoichiometric ratio of 1.43.
trophic, with acetate and nitrate, (b) lithotrophic, with sulfide and nitrate, and (c) mixed, with acetate, sulfide and nitrate.
2.3. Batch cultures
2.4. Analytical methods
Two types of batch tests (biotic and abiotic) were conducted out in order to measure the substrate consumption. The assays were carried out in 120 mL serum flasks sealed with butyl rubber stoppers. 100 mL of mineral medium were used with the following basal chemical composition (g/L): CaCl2 2H2O, 0.09; Na2MoO4 2H2O, 0.005; MgCl2 6H2O, 0.044; KH2PO4, 0.12; MgSO4 7H2O; 0.02; FeCl3 6H2O; 0.0008; CuSO4 5H2O, 0.002. For the biotic test, 12 mL of denitrifying biomass from the continous stirred tank sulfide oxidizing reactor was added to each flask for a final concentration of 1.45 g SSV/L. The conditions for abiotic tests were similar but no inoculum was added. The assays were conducted in presence and absence of either acetate or sulfide. The pH, temperature and agitation were 8.370.2, 30 C and 95 rpm, respectively. Sulfide, nitrate and bicarbonate were used at concentrations of 104 mg S2/L, 73 mg NO 3 -N/L and 0.065 g HCO -C/L, respectively. In sulfide absence, 3 acetate and nitrate concentrations of 102 mg org-C/L and 73 mg NO 3 -N/L were used. In order to obtain the specific substrate consumption rates of acetate [qAce ], sulfide [qS2 ], nitrate [qNO3 ] and the denitrifying activity [qN2 ], additional batch assays were performed in a 1.3 L reactor under similar culture media as described previously. The experiments were conducted under the following conditions: (a) hetero-
To measure nitrate, nitrite, thiosulfate and sulfate, liquid samples were centrifuged in an Eppendorf centrifuge and the supernatant filtered with a 0.45 mm filter and injected into a capillary electrophoresis ion analyzer (Waters 4000) as described by Gomez et al. [28]. An ammonia-specific electrode (Phoenix Electrode Co.) was used to measure ammonium according to standard methods [18]. N2, N2O, CO2 and CH4 were analyzed by gas chromatography (Varian Star 3400) equipped with a thermal conductivity detector and a Poropak Q column (mesh of 80–100 mm). Helium was used as carrier gas at a flow rate of 16 mL/min. The column, injector and detector temperatures were 35 C, 100 C and 110 C, respectively. Soluble organic carbon was determined with a total organic carbon analyzer (Shimadzu TOC analyzer TOC-5000). Liquid samples were centrifuged at 5000g for 10 min and directly quantified. Elemental sulfur analysis was made by a modification of the method described by Bartlett and Skoog [19]. A sludge sample dried at 80 C for 2 h was mixed with petroleum ether to dissolve sulfur which was analyzed by cianolysis. A calibration curve was prepared using dilutions of a 50 ppm of elemental sulfur dissolved in petroleum ether and measuring absorbance at 465 nm. Total sulfide was iodometrically quantified [18]. Biomass as VSS and total solids were measured by standard methods [18].
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3316
3. Results and discussion 3.1. Denitrifying reactor under steady state (control reactor) The denitrifying reactor was initially operated under heterotrophic conditions at a C/N ratio of 1.75 using acetate as the carbon and energy source. The NO 3 -N and org-C removal in the denitrifying reactor were very high as can be seen in Fig. 2(a and b). In this period there was a significant variation of the Y-N2 due to a gas leakage from the reactor, which was corrected on day 150. After approximately 180 d of operation, the C/N ratio of the culture was adjusted from 1.75 to 1.45. This change in the C/N ratio was to adjust stoichiometric biological culture conditions for the nitrate and acetate reaction considering microbial growth. No effect on the nitrate consumption was seen when the C/N ratio decreased, as the consumption efficiency (calculated as [mg NO 3 -N consumed/mg NO3 -N fed] 100) remained constant and close to 100%. The Y-N2 was 0.93. This respiration process was constant during the last 90 d
350
(a) 300
C/N 1.7
C/N 1.45
Q, mg /L.d
250 200 150 100 50 0
500
(b)
Q, mg /L.d
400
before addition of sulfide. The response of the culture at the C/N ratio of 1.45 is shown in Table 1. The variation coefficient of the N2 production rate was low (75%) throughout the denitrifying process, thus the respiration process was in steady state. Due to the high Y-N2 at the C/N ratio of 1.45, the process was clearly dissimilative, thus no significant production of VSS was observed being controlled in the reactor at 2.470.4 g VSS/L. Accumulation of intermediates from the denitrification, + such as NO 2 and N2O was not detected. NH4 was rarely detected in the effluent and represented less than 3% of the influent nitrogen-loading rate. The org-C consumption coincided with the NO 3 -N depletion and was consumed as much as 94%. Evolution of CO2 from acetate was low because it was mainly solubilized in the medium due to its alkaline pH conditions.
3.2. Denitrifying sulfide oxidizing reactor After 90 d of operation at steady-state denitrification under anoxic heterotrophic conditions, additionally to acetate several sulfide-loading rates were applied to the reactor as shown in Table 2. At any sulfide-loading rate, the nitrate consumption rate did not change, as it was similar to the one observed with acetate as the sole electron donor (control reactor). In all sulfide-loading rates the molecular nitrogen production rate was 17473.9 mg N2/L d, 12% less than the control. As the denitrifying rate consistently had low variation, the reactor achieved a steady-state condition. Although the Y-N2 decreased close to 0.83 at any sulfide loading rate (runs 1–5), the denitrifying yield was continuously high (Fig. 3). Thus, NO 3 -N was always efficiently denitrified to N2. A low fraction of nitrous oxide gas (N2O) accounting for less than 3.3% of the NO 3 -N fed was detected in the biogas produced. Consequently, the addition of sulfide to the reactor did not significantly modify the denitrifying process. In contrast to nitrate consumption, the carbon consumption efficiency decreased. At a maximum
300
Table 1 Operational parameters and treatment efficiency during continuous conditions of the denitrifying reactor under heterotrophic conditions at a C/N ratio of 1.45
200
100
0 0
50
100
150
200
250
300
Operational Biomass (g VSS/L) N-NO 3 load rate (mg/L d) CH3COO-C load rate (mg/L d)
2.4 209.4 303
Efficiency N-NO 3 consumption (%) C-CH3COO consumption (%) Denitrifying yield, Y-N2 (g N2/g N-NO 3)
9970.1 9476 0.9370.05
Time, d
Fig. 2. (a) Nitrogen compound profile: (J) N-NO 3 loading rate, (m) N2 production rate, and ( ) N-NO 3 in the effluent. (b) Carbon compound profile: (K) org-C loading rate, (n) CO2-C in the produced gas, and (&) org-C in the effluent.
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3317
Table 2 Sulfide, nitrate and acetate transformations under different sulfide loading rates applied to the continuous reactor. During all sulfide loading rates around 20% of acetate consumed was used for biomass production. # Feed
Input
Output
0 42.2 83.6 167.1 258.3 294
0 3277 5773 67715 1377 772
0.87 0.94 0.96 1 0.99
Q, N-NO 3 (mg/L d) Input
Output
209.4 209.4 209.4 209.4 209.4 209.4
0 0 0 0 0 0
Y-N2a
0.94 0.83 0.84 0.8 0.8 0.83
Q, org-C (mg/L d) Input
Output
Q, org-C consumed for denitrification (mg/L d)
303 303 303 303 303 303
1274 3778 71713 111723 114711 9378
185.9 182.9 172.6 140.8 137.3
% N-NO 3 for sulfide oxidation
% N-NO 3 for acetate oxidation
1.1 2.7 9.8 24.6 28.1
98.9 97.3 90.2 75.4 71.9
Y-: Yield (mg product/mg substrate consumed)
Yield, Y-N2
a
Y-S a
1.0
100
0.8
80
0.6
60
0.4
40
0.2
20
0.0 0
50
100
150
200
250
300
Consump. Efficiency, %
1 2 3 4 5
Q- S2 (mg/L d)
0 350
Q-S2-, mg /L.d
Fig. 3. Denitrifying yield (J), and consumption efficiencies of sulfide (m) and acetate (&) at different sulfide loading rates (Q-S2) and constant C/N rate.
sulfide-loading rate, the org-C consumption was 69% compared to the 94% of the control reactor. On one hand, at low sulfide loading rates (runs 1, 2 and 3, that is, 42.2, 83.6 and 167.1 mg S2/L d, respectively) the anoxic sulfide oxidation efficiency was low, but at higher sulfide loading rates it increased (Fig. 3). As the S/N ratio reached the stoichiometric value (1.43, according to Eq. (2)) the sulfide oxidation efficiency reached almost 100%. In all cases, the anoxic sulfide oxidation was partial as elemental sulfur was produced instead of sulfate. This result could be due to the simultaneous feeding of acetate. At sulfide loading rates smaller than 167 mg S2/L d, a possible competitive pattern between both reducing sources (acetate and sulfide) could have occurred. However, at higher sulfide loading rates the acetate consumption remained constant, but the sulfide oxidation efficiency increased. Heterotrophic denitrification using acetate as electron donor is well described
[12,20,21]. In contrast, the pathway of anoxic sulfide oxidation under denitrifying lithoautotrophic conditions is not yet well understood. Hence, it is possible to assume that both respiratory processes are different. Nevertheless, the sulfide and acetate oxidation coexisting at the culture conditions established here, possibly indicate the presence of both oxidizing metabolic pathways. There are evidences in the literature of simultaneous oxidation of elemental sulfur or thiosulfate together with organic matter [16,22,23], but sulfide oxidation in the presence of organic matter is rarely observed [24]. These simultaneous respiratory processes might be explained in terms of the microbial diversity present in the consortium, where it could be possible to find groups of microorganisms simultaneously carrying out the biological reduction of nitrate using acetate and sulfide as electron donors. However, further work must be
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3318
conducted in order to clarify which organisms are required for simultaneous heterotrophic and autotrophic denitrification. To get further evidence about the nature of the respiratory process, a stoichiometric analysis of the consumption of both electron donors was carried out. For acetate consumed at each sulfide loading rate, the nitrate reduced was calculated, according to Eq. (1). The results indicated that at the highest sulfide loading rate (run 5, Table 2), the acetate consumption rate for nitrate reduction was 137.3 mg org-C/L d, which reduced 72% of NO 3 -N. The 28% nitrate remaining in the influent (58 mg NO oxidized near 3 -N/L d) 287 mg S2/L d, producing elemental sulfur which accumulated inside the reactor as shown in Fig. 4. The nitrate and sulfide consumed agreed with the stoichiometry of Eq. (3) 0 5S2 þ 2NO 3 þ 6H2 O-5S þ N2 þ 12OH ; 0
DG0 ¼ 1168:4 kJ=reaction:
ð3Þ
In general, during all sulfide loading rates around 20% of acetate consumed was used for biomass production. 3.3. Batch cultures The Gibbs free energy changes (DG 00 ) associated with the oxidation of acetate (–1054.8 kJ/reaction) and sulfide (1168.4 kJ/reaction) in the presence of nitrate are exergonic. The DG 00 changes of biological nitrate reduction are also influenced by the concentration of acetate and sulfide. Likewise, the metabolic activity, like the specific substrate consumption rate, is also influenced by the changes in concentration. In any biological kinetic processes the type of electron donors for nitrate reduction must also be considered, since at similar concentrations the metabolic rate might be different.
25
20
15
15 10 10
Sulfur (%w)
Total Solids, g/L
20
5
5 0
0 0
50
100
150
Q-S
2-
200
250
300
350
(mg/L.d)
Fig. 4. Total solids (’) and elemental sulfur (J) accumulated in the reactor under the gradual increase of the sulfide loading rates (Q-S2). Elemental sulfur is expressed as percentage weight related to the total solids inside the reactor.
Thus, it was important to evaluate the specific biological oxidation rates of acetate and sulfide for nitrate reduction using the biomass produced at steady state in the continuous denitrifying sulfide oxidizing reactor. Batch assays in serum bottles under anoxic heterotrophic conditions with acetate indicated that the denitrifying biomass completely consumed both acetate and nitrate in 18 h, with N2 as the main gas produced. Depletion of both compounds in the abiotic controls in the same period of time was 5% and 15%, respectively (Table 3, column A). Experiments conducted under lithotrophic conditions shown that biological sulfide removal was complete while nitrate consumption efficiency was close to 90%. Under abiotic conditions, 17% of sulfide disappeared and 100% of NO 3 was reduced to NO 2 (Table 3, column B). Thus, as the biological conversions were higher than the abiotic ones, the denitrification process was mainly due to biological activity. Three series of experiments to measure the specific consumption rates of acetate, nitrate and sulfide were carried out in a 1.3 L stirred reactor. The results are shown in Table 4. The C/N and S/N ratios were 1.45 and 1.43, respectively. The heterotrophic specific denitrification rate [qN2 ]h was 0.3 kg N2/kg VSS d. The specific consumption rates for nitrate [qNO3 ]h and acetate ½qAce had the same value (1.9 kg substrate/kg VSS d). The lithotrophic specific denitrification rate [qN2 ]l was 6.9 103 kg N2/Kg VSS d, while for sulfide consumption [qS2 ] it was 5.3 kg S2/kg VSS d and for nitrate [qNO3 ]l was 0.38 g NO3-N/kg VSS d. It can be observed that the sulfide consumption rate was three times higher than for acetate, although the nitrate consumption rate in presence of sulfide was five times lower than under heterotrophic conditions. Ikemoto-Yamamoto et al. [25] found a similar value for the specific consumption rate of nitrate in a denitrifying lithoautotrophic culture, but using thiosulfate as energy source and a culture enriched with sulfur utilizing bacteria. During the batch experiments conducted in the 1.3 L reactor under lithotrophic conditions, it was observed that sulfide oxidation proceeded in two steps: sulfide was first oxidized to thiosulfate and elemental sulfur (S0), then both compounds oxidized further to sulfate in the second step as shown in Fig. 5b. The first step was faster than the second one. This behavior was also observed by Gommers et al. [26] under anoxic conditions and by Buisman et al. [27] under aerobic conditions. Visser et. al. [3] suggested that the slow consumption of sulfur might be due to the saturation of the electron transport chain. Thus, it seems that the sulfate formation from intermediates (thiosulfate and elemental sulfur) is the bottleneck of the lithotrophic denitrification. In the third case using acetate and sulfide mixed as electron donors (Table 4, column C), the specific consumption rates of sulfide, nitrate and the qN2 showed
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3319
Table 3 Effect of the biomass and chemical activity in the denitrification batch cultures using: (A) acetate-nitrate at a C/N ratio of 1.4, (B) sulfide-nitrate at an S/N ratio of 1.43 Time, h
A
B
With biomass
Abiotic control
With biomass
Abiotic control
0 18 66
C-organic (mg/L) 102 0.7 0
102 97 96
S2 (mg/L) 104 0 0
104 86 86
0 18 66
N-NO 3 (mg/L) 73 0 0
73 62 65
N-NO 3 (mg/L) 73 6.8 0
73 0 0
0 18 66
N-NO 2 (mg/L) 0 0 0
0 5 5
N-NO 2 (mg/L) 0 26 6
0 70.2 71
qAce (kg C/kg VSS d) qS2 (kg S2/kg VSS d) qNO3 (kg N-NO 3 /kg VSS d) qN2 (kg N2/kg VSS d)
A
B
C
1.9 — 1.9 0.3
— 5.3 0.38 6.9 103
0.6 8.2 1.1 8.4 102
80
(a) 60 conc. mg/L
Table 4 Specific consumption rates in the denitrification using different electron donors: (A) acetate, 102.2 mg org-C/L; (B) sulfide, 104 mg S2/L; (C) acetate, 102.2 mg org-C/L and sulfide, 104 mg S2/L
20 0
73 mg N-NO 3 /L.
In all cases the electron acceptor was nitrate, qAce ; qS2 and qNO3 : specific consumption rates for organic-C, sulfide and N-NO 3 . qN2 : N2 specific denitrification rate.
160
(b) 120 conc. mg/L
an increase of 1.54, 2.8 and 12 times, respectively, with respect to the lithotrophic conditions. This suggests that in the lithotrophic denitrification the presence of acetate could enhance sulfide and nitrate consumption rates. During the transient accumulation of thiosulfate and elemental sulfur in batch cultures, nitrite was also accumulated and reduced slowly to N2 as shown in Fig. 5a. The specific sulfide consumption rate increased when acetate was present. In contrast, in the continuous culture cultivated in the presence of both sulfide and acetate at C/N and S/N ratios of 1.45 and 1.43, respectively, nitrate was efficiently converted to N2 without nitrite accumulation. Acetate consumption was 69% and sulfide oxidation was partial, as elemental sulfur was the end product. Under batch heterotrophic conditions with acetate and nitrate, the [qN2 ]h was always higher than both lithotrophic and mixed (sulfide and acetate) conditions. The specific rates of the batch cultures can be used to attempt a possible explanation of the partial sulfide oxidation to elemental sulfur in
40
80 40 0 0
20
40
60
80
100
time, h
Fig. 5. Lithotrophic denitrification profiles in batch experi ment: (a) N-NO 3 , (J); N2 production, (m); N-NO2 (’) and (b) S2, (K); S-S2O2 , (x); and S-SO (&). 3 4
continuous culture that is depicted in Fig. 6. The oxidation rate of sulfide to sulfur and to reduce nitrate to nitrite ðr1 Þ is higher than the oxidation rate of acetate to reduce nitrate into nitrite. However, the oxidation rate of sulfur ðr3 Þ in order to reduce nitrite to N2 is slower than the oxidation rate of acetate ðr2 Þ for reducing nitrite to N2. Therefore, sulfur accumulates
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3320
Acknowledgements
NO3-
S2-
This work was financed by IMP projects D.00021 and FIES 98-106-IV. r1 So
NO2-
CH3COO-
References
r3 SO42-
r2
N2
CO2
Fig. 6. Suggested sulfide and acetate oxidation pathway under denitrifying conditions.
when acetate is present. Likewise, these rate differences account for the decreasing consumption efficiency of org-C in continuous culture. Furthermore, the minimal inhibitory effect of sulfide could also be explained by a similar way, namely, the potential toxic effect was eliminated due to the rapid oxidation of sulfide to thiosulfate and sulfur. Moreover, this agreed with the behavior observed in batch culture where a concentration of 104 mg S2/L without acetate present resulted in a low denitrification rate due to the toxic effect of sulfide and the slow oxidation rate of sulfur up to sulfate.
4. Conclusions The results of this work demonstrated that denitrification is a feasible process for the simultaneous removal of nitrogen, carbon and sulfur. Sulfide was eliminated via partial oxidation to elemental sulfur that accumulated inside the reactor without any signs of inhibition on the process. In this way, elemental sulfur can be removed from the reactor closing the sulfur cycle. In continuous denitrifying sulfide oxidizing reactor the removal efficiencies for nitrate, acetate and sulfide were close to 100%, 69% and 100%, respectively. Additionally, it was demonstrated that the denitrification process was biologically mediated, as the chemical transformation reactions in batch experiments were incomplete and proceeded at very low rates. Sulfide drove denitrification at a rate over two orders of magnitude less than acetate, however, the presence of acetate in addition to sulfide increased both sulfide oxidation and denitrification rate by roughly 55% and an order of magnitude, respectively. The results showed in this work suggest this approach could be applied to the treatment of wastewaters from the petroleum industry.
[1] Baker LA. Design considerations and application for wetland treatment of high-nitrate waters. Water Sci Technol 1998;38(1):389–95. [2] Oenema O, Roest CW. Nitrogen and phosphorus losses from agriculture into surface waters: the effects of policies and measures in The Netherlands. Water Sci Technol 1998;37:19–30. [3] Visser JM, Robertson LA, Van Verseveld HW, Kuenen JG. Sulfur production by obligately chemolithoautotrophic Thiobacillus species. Appl Environ Microbiol 1997;63:2300–5. [4] Cadena F, Peters RW. Evaluation of chemical oxidizer for hydrogen sulfide control. J Water Pollut Control Fed 1988;60:1259–63. [5] Betlach MR, Tiedje JM. Kinetic explanation for accumulation of nitrite, nitric oxide, and nitrous oxide during bacterial denitrification. Appl Environ Microbiol 1981;42:1074–84. [6] Cervantes F, Monroy O, Go´mez J. Influence of ammonium on the performance of a denitrifying culture under heterotrophic conditions. Appl Biochem Biotechnol 1999;81:13–21. [7] Cuervo-Lopez F, Mart!ınez F, Gutierrez-Rojas M, Noyola RA, Go´mez J. Effect of nitrogen loading rate and carbon source on denitrification sludge settleability in upflow anaerobic sludge blanket (UASB) reactors. Water Sci Technol 1999;40(8):123–30. [8] Bernet N, Habouzit F, Moletta R. Use of a industrial effluent as carbon source for denitrification of a highstrength wastewater. Appl Microbiol Biotechnol 1996;46:92–7. [9] Chen SD, Chen CY, Shen YC, Chiu CM, Chen HJ. Treatment of high strength nitrate wastewater by biological methods—operational characteristic study. Water Sci Technol 1996;34(1–2):269–76. [10] Driscoll CT, Bisogni JJ. The use of sulfur and sulfide in packed-bed reactors for autotrophic denitrification. J Water Pollut Control Fed 1978;50(3):569–77. [11] Flere MJ, Zhang TC. Sulfur-based autotrophic denitrification pond systems for in-situ remediation of nitratecontaminated surface water. Water Sci Technol 1998;38: 15–22. [12] Sorensen HB, Jorgensen SE. The removal of nitrogen compounds from wastewater. Amsterdam, The Netherlands: Elsevier Science; 1993. [13] Koening A, Liu HL. Autotrophic denitrification of landfill leachate using elemental sulphur. Water Sci Technol 1996;34:469–76. [14] Trouve C, Chazal PM, Gueroux B, Sauvaitre N. Denitrification by new strains of Thiobacillus denitrificans under non-standard physicochemical conditions. Effect of
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
temperature pH and sulphur source. Environ Technol 1998;19:601–10. Gommers PJ, Buleveld W, Zuiderwijk FJ, Kuenen JG. Simultaneous sulfide and acetate oxidation in a denitrifying fluidized bed reactor-I. Water Res 1988;22:1075–83. Kim EW, Bae JH. Alkalinity requirements and the possibility of simultaneous heterotrophic denitrification during sulfur-utilizing autotrophic denitrification. Water Sci Technol 2000;42:233–8. Oh SE, Yoo YB, Young JC, Kim IS. Effect of organics on sulfur-utilizing autotrophic denitrification under mixotrophic conditions. J Biotechnol 2001;92(1):1–8. APHA, Standard methods for the examination of water and wastewater, 17th ed. New York, USA: American Public Health Association; 1989. Bartlett JK, Skoog DA. Colorimetric determination of elemental sulfur in hydrocarbons. Anal Chem 1954; 26(6):1008–11. Fass S, Ganaye V, Urbain V, Manem J, Block JC. Volatile fatty acids as organic carbon sources in denitrification. Environ Technol 1994;15:459–67. Tiedje JM. Ecology of denitrification and dissimilatory nitrate reduction to ammonium. In: Zehnder AJB, editor. Biology of anaerobic microorganisms. New York: Wiley; 1988. p. 179–244. Kim IS, Son JH. Impact of COD/N/S ratio on dentrification by the mixed cultures of sulfate reducing bacteria and sulfur denitrifying bacteria. Water Sci Technol 2000;42:69–76.
3321
[23] Lee DU, Lee IS, Choi YD, Bae HO. Effects of external carbon source and empty bed contact time on simultaneous heterotrophic and sulfur-utilizing autotrophic denitrification. Process Biochem 2001;36(12):1215–24. [24] Krishnakumar B, Manilal VB. Bacterial oxidation of sulphide under denitrifying conditions. Biotechnol Lett 1999;21:437–40. [25] Ikemoto-Yamamoto R, Komori T, Nomura M, Ide Y, Matzukami T. Nitrogen removal of hydroponic culture wastewater by autotrophic denitrification using thiosulfate. Water Sci Technol 2000;42:369–76. [26] Gommers PJ, Buleveld W, Zuiderwijk FJ, Kuenen JG. Simultaneous sulfide and acetate oxidation in a denitrifying fluidized bed reactor-I. Water Res 1988;22:1085–92. [27] Buisman CJN, Geraats BG, Ijspeert P, Lettinga G. Optimization of sulphur production in a biotechnological sulphide-removing reactor. Biotechnol Bioeng 1990;35:50–6. [28] Gomez J, Mendez R, Lema J. The effect of antibiotics on nitrification process: batch assays. Appl Biochem Biotechnol 1996;57:869–76.
Further reading Janssen AJ, Sleyster R, Van der Kaa C, Jochemsen A, Bontsema J, Lettinga, G. Biological sulphide oxidation in a fed-batch reactor. Biotechnol Bioeng 1995;47:327–33.
Water Research 38 (2004) 3313–3321
Simultaneous biological removal of nitrogen, carbon and sulfur by denitrification Jesu´s Reyes-Avilaa, El!ıas Razo-Floresa,1, Jorge Gomezb,* b
a Instituto Mexicano del Petro´leo, Programa de Biotecnolog!ıa, Eje Central La´zaro Ca´rdenas 152, C.P. 07730, Mexico Departamento de Biotecnolog!ıa, Universidad Auto´noma Metropolitana-Iztapalapa, Divisio´n CBS, San Rafael Atlixco 186, Col. Vicentina, C.P. 09340, Mexico
Received 7 March 2003; received in revised form 20 April 2004; accepted 29 April 2004
Abstract Refinery wastewaters may contain aromatic compounds and high concentrations of sulfide and ammonium which must be removed before discharging into water bodies. In this work, biological denitrification was used to eliminate carbon, nitrogen and sulfur in an anaerobic continuous stirred tank reactor of 1.3 L and a hydraulic retention time of 2 d. Acetate and nitrate at a C/N ratio of 1.45 were fed at loading rates of 0.29 kg C/m3 d and 0.2 kg N/m3 d, respectively. Under steady-state denitrifying conditions, the carbon and nitrogen removal efficiencies were higher than 90%. Also, under these conditions, sulfide (S2) was fed to the reactor at several sulfide loading rates (0.042–0.294 kg S2/m3 d). The high nitrate removal efficiency of the denitrification process was maintained along the whole process, whereas the carbon removal was 65% even at sulfide loading rates of 0.294 kg S2/m3 d. The sulfide removal increased up to B99% via partial oxidation to insoluble elemental sulfur (S0) that accumulated inside the reactor. These results indicated that denitrification is a feasible process for the simultaneous removal of nitrogen, carbon and sulfur from effluents of the petroleum industry. r 2004 Elsevier Ltd. All rights reserved. Keywords: Denitrification; Anoxic sulfide oxidation; Nitrate reduction; Sulfur production
1. Introduction At present, the contamination by carbon, nitrogen and sulfur compounds in wastewaters and water bodies is a critical problem. Some wastewaters as those from the oil industry represent a tremendous challenge for
*Corresponding author. Tel.: +52-55-5804-6408; fax: +5255-5804-6407. E-mail address: [email protected] (J. Gomez). 1 Present address: Instituto Potosino de Investigacio´n Cient!ıfica y Tecnolo´gica A.C., Departamento de Ingenier!ıa Ambiental y Manejo de Recursos Naturales, Camino a la Presa San Jose´ 2055, Col. Lomas 4a, Seccio´n, C.P. 78216, San Luis Potos!ı, SLP, Mexico.
treatment before discharge because of its chemical complexity. These effluents may contain a high concentration of organic compounds such as phenol and cresols that can be mineralized by biological processes. Likewise, some inorganic compounds like sulfide and ammonia are also frequently found. Nitrogen compounds contribute mainly to eutrophication of water bodies, besides the risks associated with toxicity and bad odors [1,2]. Sulfide is a very toxic compound for many microorganisms, even at concentrations as low as 10 mg/ L; this is due to the fact that sulfide reacts with the iron from cytochromes inhibiting the respiration [3]. Additionally, it is corrosive and possesses a high chemical oxygen demand (COD). Ammonia, as well as sulfide, has a high COD and is toxic for the aquatic fauna even at concentrations of 4 mg/L.
0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2004.04.035
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3314
In order to remove these contaminants, biological treatments are preferred technologies rather than physical–chemical methods, which are expensive and may generate toxic residuals [4]. Ammonium can be biologically oxidized to nitrate under aerobic conditions and subsequently reduced to molecular nitrogen (N2) via denitrification under anoxic conditions. The heterotrophic denitrification process uses many organic compounds as carbon and energy sources [5]. Cervantes et al. [6] reported that for high denitrifying efficiencies the C/N ratio was the main parameter of control to achieve a dissimilative respiratory process. The stoichiometric reaction between acetate and nitrate (C/N ratio of 1.07) is shown in Eq. (1):
biological activity of the biomass will also determine the rate of the reaction. The objective of this work was to evaluate the simultaneous biological removal of nitrogen (nitrate), carbon (acetate) and sulfur (sulfide) under well defined denitrifying conditions. First, a continuous stirred tank reactor cultivating a heterotrophic population was operated. After the reactor reached steady-state denitrification, sulfide was fed to the reactor at several sulfide-loading rates to evaluate the performance of the reactor. Finally, batch experiments were conducted measuring the specific removal rates and the influence of the abiotic reactions.
1:25CH3 COOH þ 2NO 3 -2:5CO2 þ N2 þ 1:5H2 O þ 2OH ;
2. Materials and methods
0
DG0 ¼ 1054:8 kJ=reaction:
2.1. Denitrifying reactor ð1Þ
The heterotrophic denitrification can be a highrate process. Cuervo-Lopez et al. [7] reported that for a denitrifying sludge in the presence of acetate, a C/N 3 ratio of 2 and a nitrate loading rate of 2 kg NO 3 -N/m d, a nitrate removal efficiency of 100% and a denitrifying yield (Y-N2, g N2/g NO 3 -N consumed) of 0.9 were obtained. Bernet et al. [8] and Chen et al. [9] applied 3 nitrate loading rates above 2.1 kg NO 3 -N/m d, but the nitrate removal efficiencies were around 70%. Lithotrophic denitrification using reduced sulfur compounds as an energy source has also been observed. A culture of Thiobacillus denitrificans, at low C/N ratio, used S0, thiosulfate and sulfide as electron donors for nitrate reduction [10–14]. Generally, CO2 is the carbon source and the final products of the autotrophic process are sulfate and N2. Eq. (2) shows the reduction of nitrate using sulfide as the energy source (S/N ratio of 1.43): 2 þ 1:25S2 þ 2NO 3 þ 2H -1:25SO4 þ N2 þ H2 O; 0
DG0 ¼ 972:8 kJ=reaction:
ð2Þ
Scarce evidence exists about the occurrence of this lithoautotrophic process when organic matter is present. Gommers et al. [15] used a fluidized bed reactor to study the effect of sulfide and acetate on denitrification under limited conditions of both substrates. The authors observed that nitrate reduction was partial as nitrite accumulated in the system. Acetate was consumed at high efficiencies and sulfide was completely oxidized to sulfate. Nevertheless, not all the end products of the biological reactions were determined. More recently, similar studies have been conducted, but elemental sulfur was used as electron donor instead sulfide [16,17]. As denitrification is a redox process, the thermodynamic reactions involved (using acetate and sulfide) should influence the overall efficiency of the process. The
An anaerobic continuous stirred tank reactor with a biomass retention device and a working volume of 1.3 L was used as illustrated in Fig. 1. The reactor was instrumented to control: temperature (30 C), agitation rate (250 rpm) and pH (8.370.2). Gas production rate was measured in a calibrated column by liquid displacement. The reactor was inoculated with 0.13 L of methanogenic sludge giving a biomass concentration of 1.75 g volatile suspended solids (VSS)/L. Nitrate and acetate (org-C) were fed separately to the reactor using two media, named as medium 1 (M1) and medium 2 (M2). The chemical composition of M1, was (g/L): CH3COONa 3H2O, 12.4; CaCl2 2H2O, 1.0; Na2MoO4 2H2O, 0.05. The chemical composition of M2 was (g/L): NaNO3, 7.6; KH2PO4, 1.5; MgSO4 7H2O, 1.45; FeCl3 6H2O, 0.01; CuSO4 5H2O, 0.03. The flow rate of each media was 0.32 L/d (total flow rate 0.64 L/d) and the hydraulic retention time (HRT) was of 2 d. During the start up the C/N ratio was 1.75, being adjusted afterwards to 1.45 and maintaining the reactor in operation under this condition for more than 90 d. 2.2. Denitrifying sulfide oxidizing reactor In order to avoid production and precipitation of metallic sulfides, adjustments were made to M1 and M2 as follows: Sodium sulfide (Na2S 9H2O) was added to medium 1 as alternative energy source, whereas sulfate (MgSO4 7H2O) concentration was reduced from 1.45 to 0.5 and 0.48 g/L of MgCl2 6H2O was added for complementing the magnesium. Finally, Na2MoO4 2H2O was added to M2 instead M1. The C/ N ratio in the reactor was fixed to 1.45. Organic carbon (as sodium acetate) and nitrogen loading rates were 303 mg org-C/L d and 209.4 mg NO 3 -N/L d, respectively. The pH in the reactor was controlled at 8.370.2 and the HRT was kept at 2 d. Five sulfide loading rates
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3315
Fig. 1. Scheme of the denitrifying continuous stirred reactor used in the experiments.
were applied to the reactor: 42.2, 83.6, 167.1, 258.3 and 294 mg S2/L d. The last value is equivalent to a S/N stoichiometric ratio of 1.43.
trophic, with acetate and nitrate, (b) lithotrophic, with sulfide and nitrate, and (c) mixed, with acetate, sulfide and nitrate.
2.3. Batch cultures
2.4. Analytical methods
Two types of batch tests (biotic and abiotic) were conducted out in order to measure the substrate consumption. The assays were carried out in 120 mL serum flasks sealed with butyl rubber stoppers. 100 mL of mineral medium were used with the following basal chemical composition (g/L): CaCl2 2H2O, 0.09; Na2MoO4 2H2O, 0.005; MgCl2 6H2O, 0.044; KH2PO4, 0.12; MgSO4 7H2O; 0.02; FeCl3 6H2O; 0.0008; CuSO4 5H2O, 0.002. For the biotic test, 12 mL of denitrifying biomass from the continous stirred tank sulfide oxidizing reactor was added to each flask for a final concentration of 1.45 g SSV/L. The conditions for abiotic tests were similar but no inoculum was added. The assays were conducted in presence and absence of either acetate or sulfide. The pH, temperature and agitation were 8.370.2, 30 C and 95 rpm, respectively. Sulfide, nitrate and bicarbonate were used at concentrations of 104 mg S2/L, 73 mg NO 3 -N/L and 0.065 g HCO -C/L, respectively. In sulfide absence, 3 acetate and nitrate concentrations of 102 mg org-C/L and 73 mg NO 3 -N/L were used. In order to obtain the specific substrate consumption rates of acetate [qAce ], sulfide [qS2 ], nitrate [qNO3 ] and the denitrifying activity [qN2 ], additional batch assays were performed in a 1.3 L reactor under similar culture media as described previously. The experiments were conducted under the following conditions: (a) hetero-
To measure nitrate, nitrite, thiosulfate and sulfate, liquid samples were centrifuged in an Eppendorf centrifuge and the supernatant filtered with a 0.45 mm filter and injected into a capillary electrophoresis ion analyzer (Waters 4000) as described by Gomez et al. [28]. An ammonia-specific electrode (Phoenix Electrode Co.) was used to measure ammonium according to standard methods [18]. N2, N2O, CO2 and CH4 were analyzed by gas chromatography (Varian Star 3400) equipped with a thermal conductivity detector and a Poropak Q column (mesh of 80–100 mm). Helium was used as carrier gas at a flow rate of 16 mL/min. The column, injector and detector temperatures were 35 C, 100 C and 110 C, respectively. Soluble organic carbon was determined with a total organic carbon analyzer (Shimadzu TOC analyzer TOC-5000). Liquid samples were centrifuged at 5000g for 10 min and directly quantified. Elemental sulfur analysis was made by a modification of the method described by Bartlett and Skoog [19]. A sludge sample dried at 80 C for 2 h was mixed with petroleum ether to dissolve sulfur which was analyzed by cianolysis. A calibration curve was prepared using dilutions of a 50 ppm of elemental sulfur dissolved in petroleum ether and measuring absorbance at 465 nm. Total sulfide was iodometrically quantified [18]. Biomass as VSS and total solids were measured by standard methods [18].
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3316
3. Results and discussion 3.1. Denitrifying reactor under steady state (control reactor) The denitrifying reactor was initially operated under heterotrophic conditions at a C/N ratio of 1.75 using acetate as the carbon and energy source. The NO 3 -N and org-C removal in the denitrifying reactor were very high as can be seen in Fig. 2(a and b). In this period there was a significant variation of the Y-N2 due to a gas leakage from the reactor, which was corrected on day 150. After approximately 180 d of operation, the C/N ratio of the culture was adjusted from 1.75 to 1.45. This change in the C/N ratio was to adjust stoichiometric biological culture conditions for the nitrate and acetate reaction considering microbial growth. No effect on the nitrate consumption was seen when the C/N ratio decreased, as the consumption efficiency (calculated as [mg NO 3 -N consumed/mg NO3 -N fed] 100) remained constant and close to 100%. The Y-N2 was 0.93. This respiration process was constant during the last 90 d
350
(a) 300
C/N 1.7
C/N 1.45
Q, mg /L.d
250 200 150 100 50 0
500
(b)
Q, mg /L.d
400
before addition of sulfide. The response of the culture at the C/N ratio of 1.45 is shown in Table 1. The variation coefficient of the N2 production rate was low (75%) throughout the denitrifying process, thus the respiration process was in steady state. Due to the high Y-N2 at the C/N ratio of 1.45, the process was clearly dissimilative, thus no significant production of VSS was observed being controlled in the reactor at 2.470.4 g VSS/L. Accumulation of intermediates from the denitrification, + such as NO 2 and N2O was not detected. NH4 was rarely detected in the effluent and represented less than 3% of the influent nitrogen-loading rate. The org-C consumption coincided with the NO 3 -N depletion and was consumed as much as 94%. Evolution of CO2 from acetate was low because it was mainly solubilized in the medium due to its alkaline pH conditions.
3.2. Denitrifying sulfide oxidizing reactor After 90 d of operation at steady-state denitrification under anoxic heterotrophic conditions, additionally to acetate several sulfide-loading rates were applied to the reactor as shown in Table 2. At any sulfide-loading rate, the nitrate consumption rate did not change, as it was similar to the one observed with acetate as the sole electron donor (control reactor). In all sulfide-loading rates the molecular nitrogen production rate was 17473.9 mg N2/L d, 12% less than the control. As the denitrifying rate consistently had low variation, the reactor achieved a steady-state condition. Although the Y-N2 decreased close to 0.83 at any sulfide loading rate (runs 1–5), the denitrifying yield was continuously high (Fig. 3). Thus, NO 3 -N was always efficiently denitrified to N2. A low fraction of nitrous oxide gas (N2O) accounting for less than 3.3% of the NO 3 -N fed was detected in the biogas produced. Consequently, the addition of sulfide to the reactor did not significantly modify the denitrifying process. In contrast to nitrate consumption, the carbon consumption efficiency decreased. At a maximum
300
Table 1 Operational parameters and treatment efficiency during continuous conditions of the denitrifying reactor under heterotrophic conditions at a C/N ratio of 1.45
200
100
0 0
50
100
150
200
250
300
Operational Biomass (g VSS/L) N-NO 3 load rate (mg/L d) CH3COO-C load rate (mg/L d)
2.4 209.4 303
Efficiency N-NO 3 consumption (%) C-CH3COO consumption (%) Denitrifying yield, Y-N2 (g N2/g N-NO 3)
9970.1 9476 0.9370.05
Time, d
Fig. 2. (a) Nitrogen compound profile: (J) N-NO 3 loading rate, (m) N2 production rate, and ( ) N-NO 3 in the effluent. (b) Carbon compound profile: (K) org-C loading rate, (n) CO2-C in the produced gas, and (&) org-C in the effluent.
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3317
Table 2 Sulfide, nitrate and acetate transformations under different sulfide loading rates applied to the continuous reactor. During all sulfide loading rates around 20% of acetate consumed was used for biomass production. # Feed
Input
Output
0 42.2 83.6 167.1 258.3 294
0 3277 5773 67715 1377 772
0.87 0.94 0.96 1 0.99
Q, N-NO 3 (mg/L d) Input
Output
209.4 209.4 209.4 209.4 209.4 209.4
0 0 0 0 0 0
Y-N2a
0.94 0.83 0.84 0.8 0.8 0.83
Q, org-C (mg/L d) Input
Output
Q, org-C consumed for denitrification (mg/L d)
303 303 303 303 303 303
1274 3778 71713 111723 114711 9378
185.9 182.9 172.6 140.8 137.3
% N-NO 3 for sulfide oxidation
% N-NO 3 for acetate oxidation
1.1 2.7 9.8 24.6 28.1
98.9 97.3 90.2 75.4 71.9
Y-: Yield (mg product/mg substrate consumed)
Yield, Y-N2
a
Y-S a
1.0
100
0.8
80
0.6
60
0.4
40
0.2
20
0.0 0
50
100
150
200
250
300
Consump. Efficiency, %
1 2 3 4 5
Q- S2 (mg/L d)
0 350
Q-S2-, mg /L.d
Fig. 3. Denitrifying yield (J), and consumption efficiencies of sulfide (m) and acetate (&) at different sulfide loading rates (Q-S2) and constant C/N rate.
sulfide-loading rate, the org-C consumption was 69% compared to the 94% of the control reactor. On one hand, at low sulfide loading rates (runs 1, 2 and 3, that is, 42.2, 83.6 and 167.1 mg S2/L d, respectively) the anoxic sulfide oxidation efficiency was low, but at higher sulfide loading rates it increased (Fig. 3). As the S/N ratio reached the stoichiometric value (1.43, according to Eq. (2)) the sulfide oxidation efficiency reached almost 100%. In all cases, the anoxic sulfide oxidation was partial as elemental sulfur was produced instead of sulfate. This result could be due to the simultaneous feeding of acetate. At sulfide loading rates smaller than 167 mg S2/L d, a possible competitive pattern between both reducing sources (acetate and sulfide) could have occurred. However, at higher sulfide loading rates the acetate consumption remained constant, but the sulfide oxidation efficiency increased. Heterotrophic denitrification using acetate as electron donor is well described
[12,20,21]. In contrast, the pathway of anoxic sulfide oxidation under denitrifying lithoautotrophic conditions is not yet well understood. Hence, it is possible to assume that both respiratory processes are different. Nevertheless, the sulfide and acetate oxidation coexisting at the culture conditions established here, possibly indicate the presence of both oxidizing metabolic pathways. There are evidences in the literature of simultaneous oxidation of elemental sulfur or thiosulfate together with organic matter [16,22,23], but sulfide oxidation in the presence of organic matter is rarely observed [24]. These simultaneous respiratory processes might be explained in terms of the microbial diversity present in the consortium, where it could be possible to find groups of microorganisms simultaneously carrying out the biological reduction of nitrate using acetate and sulfide as electron donors. However, further work must be
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3318
conducted in order to clarify which organisms are required for simultaneous heterotrophic and autotrophic denitrification. To get further evidence about the nature of the respiratory process, a stoichiometric analysis of the consumption of both electron donors was carried out. For acetate consumed at each sulfide loading rate, the nitrate reduced was calculated, according to Eq. (1). The results indicated that at the highest sulfide loading rate (run 5, Table 2), the acetate consumption rate for nitrate reduction was 137.3 mg org-C/L d, which reduced 72% of NO 3 -N. The 28% nitrate remaining in the influent (58 mg NO oxidized near 3 -N/L d) 287 mg S2/L d, producing elemental sulfur which accumulated inside the reactor as shown in Fig. 4. The nitrate and sulfide consumed agreed with the stoichiometry of Eq. (3) 0 5S2 þ 2NO 3 þ 6H2 O-5S þ N2 þ 12OH ; 0
DG0 ¼ 1168:4 kJ=reaction:
ð3Þ
In general, during all sulfide loading rates around 20% of acetate consumed was used for biomass production. 3.3. Batch cultures The Gibbs free energy changes (DG 00 ) associated with the oxidation of acetate (–1054.8 kJ/reaction) and sulfide (1168.4 kJ/reaction) in the presence of nitrate are exergonic. The DG 00 changes of biological nitrate reduction are also influenced by the concentration of acetate and sulfide. Likewise, the metabolic activity, like the specific substrate consumption rate, is also influenced by the changes in concentration. In any biological kinetic processes the type of electron donors for nitrate reduction must also be considered, since at similar concentrations the metabolic rate might be different.
25
20
15
15 10 10
Sulfur (%w)
Total Solids, g/L
20
5
5 0
0 0
50
100
150
Q-S
2-
200
250
300
350
(mg/L.d)
Fig. 4. Total solids (’) and elemental sulfur (J) accumulated in the reactor under the gradual increase of the sulfide loading rates (Q-S2). Elemental sulfur is expressed as percentage weight related to the total solids inside the reactor.
Thus, it was important to evaluate the specific biological oxidation rates of acetate and sulfide for nitrate reduction using the biomass produced at steady state in the continuous denitrifying sulfide oxidizing reactor. Batch assays in serum bottles under anoxic heterotrophic conditions with acetate indicated that the denitrifying biomass completely consumed both acetate and nitrate in 18 h, with N2 as the main gas produced. Depletion of both compounds in the abiotic controls in the same period of time was 5% and 15%, respectively (Table 3, column A). Experiments conducted under lithotrophic conditions shown that biological sulfide removal was complete while nitrate consumption efficiency was close to 90%. Under abiotic conditions, 17% of sulfide disappeared and 100% of NO 3 was reduced to NO 2 (Table 3, column B). Thus, as the biological conversions were higher than the abiotic ones, the denitrification process was mainly due to biological activity. Three series of experiments to measure the specific consumption rates of acetate, nitrate and sulfide were carried out in a 1.3 L stirred reactor. The results are shown in Table 4. The C/N and S/N ratios were 1.45 and 1.43, respectively. The heterotrophic specific denitrification rate [qN2 ]h was 0.3 kg N2/kg VSS d. The specific consumption rates for nitrate [qNO3 ]h and acetate ½qAce had the same value (1.9 kg substrate/kg VSS d). The lithotrophic specific denitrification rate [qN2 ]l was 6.9 103 kg N2/Kg VSS d, while for sulfide consumption [qS2 ] it was 5.3 kg S2/kg VSS d and for nitrate [qNO3 ]l was 0.38 g NO3-N/kg VSS d. It can be observed that the sulfide consumption rate was three times higher than for acetate, although the nitrate consumption rate in presence of sulfide was five times lower than under heterotrophic conditions. Ikemoto-Yamamoto et al. [25] found a similar value for the specific consumption rate of nitrate in a denitrifying lithoautotrophic culture, but using thiosulfate as energy source and a culture enriched with sulfur utilizing bacteria. During the batch experiments conducted in the 1.3 L reactor under lithotrophic conditions, it was observed that sulfide oxidation proceeded in two steps: sulfide was first oxidized to thiosulfate and elemental sulfur (S0), then both compounds oxidized further to sulfate in the second step as shown in Fig. 5b. The first step was faster than the second one. This behavior was also observed by Gommers et al. [26] under anoxic conditions and by Buisman et al. [27] under aerobic conditions. Visser et. al. [3] suggested that the slow consumption of sulfur might be due to the saturation of the electron transport chain. Thus, it seems that the sulfate formation from intermediates (thiosulfate and elemental sulfur) is the bottleneck of the lithotrophic denitrification. In the third case using acetate and sulfide mixed as electron donors (Table 4, column C), the specific consumption rates of sulfide, nitrate and the qN2 showed
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3319
Table 3 Effect of the biomass and chemical activity in the denitrification batch cultures using: (A) acetate-nitrate at a C/N ratio of 1.4, (B) sulfide-nitrate at an S/N ratio of 1.43 Time, h
A
B
With biomass
Abiotic control
With biomass
Abiotic control
0 18 66
C-organic (mg/L) 102 0.7 0
102 97 96
S2 (mg/L) 104 0 0
104 86 86
0 18 66
N-NO 3 (mg/L) 73 0 0
73 62 65
N-NO 3 (mg/L) 73 6.8 0
73 0 0
0 18 66
N-NO 2 (mg/L) 0 0 0
0 5 5
N-NO 2 (mg/L) 0 26 6
0 70.2 71
qAce (kg C/kg VSS d) qS2 (kg S2/kg VSS d) qNO3 (kg N-NO 3 /kg VSS d) qN2 (kg N2/kg VSS d)
A
B
C
1.9 — 1.9 0.3
— 5.3 0.38 6.9 103
0.6 8.2 1.1 8.4 102
80
(a) 60 conc. mg/L
Table 4 Specific consumption rates in the denitrification using different electron donors: (A) acetate, 102.2 mg org-C/L; (B) sulfide, 104 mg S2/L; (C) acetate, 102.2 mg org-C/L and sulfide, 104 mg S2/L
20 0
73 mg N-NO 3 /L.
In all cases the electron acceptor was nitrate, qAce ; qS2 and qNO3 : specific consumption rates for organic-C, sulfide and N-NO 3 . qN2 : N2 specific denitrification rate.
160
(b) 120 conc. mg/L
an increase of 1.54, 2.8 and 12 times, respectively, with respect to the lithotrophic conditions. This suggests that in the lithotrophic denitrification the presence of acetate could enhance sulfide and nitrate consumption rates. During the transient accumulation of thiosulfate and elemental sulfur in batch cultures, nitrite was also accumulated and reduced slowly to N2 as shown in Fig. 5a. The specific sulfide consumption rate increased when acetate was present. In contrast, in the continuous culture cultivated in the presence of both sulfide and acetate at C/N and S/N ratios of 1.45 and 1.43, respectively, nitrate was efficiently converted to N2 without nitrite accumulation. Acetate consumption was 69% and sulfide oxidation was partial, as elemental sulfur was the end product. Under batch heterotrophic conditions with acetate and nitrate, the [qN2 ]h was always higher than both lithotrophic and mixed (sulfide and acetate) conditions. The specific rates of the batch cultures can be used to attempt a possible explanation of the partial sulfide oxidation to elemental sulfur in
40
80 40 0 0
20
40
60
80
100
time, h
Fig. 5. Lithotrophic denitrification profiles in batch experi ment: (a) N-NO 3 , (J); N2 production, (m); N-NO2 (’) and (b) S2, (K); S-S2O2 , (x); and S-SO (&). 3 4
continuous culture that is depicted in Fig. 6. The oxidation rate of sulfide to sulfur and to reduce nitrate to nitrite ðr1 Þ is higher than the oxidation rate of acetate to reduce nitrate into nitrite. However, the oxidation rate of sulfur ðr3 Þ in order to reduce nitrite to N2 is slower than the oxidation rate of acetate ðr2 Þ for reducing nitrite to N2. Therefore, sulfur accumulates
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
3320
Acknowledgements
NO3-
S2-
This work was financed by IMP projects D.00021 and FIES 98-106-IV. r1 So
NO2-
CH3COO-
References
r3 SO42-
r2
N2
CO2
Fig. 6. Suggested sulfide and acetate oxidation pathway under denitrifying conditions.
when acetate is present. Likewise, these rate differences account for the decreasing consumption efficiency of org-C in continuous culture. Furthermore, the minimal inhibitory effect of sulfide could also be explained by a similar way, namely, the potential toxic effect was eliminated due to the rapid oxidation of sulfide to thiosulfate and sulfur. Moreover, this agreed with the behavior observed in batch culture where a concentration of 104 mg S2/L without acetate present resulted in a low denitrification rate due to the toxic effect of sulfide and the slow oxidation rate of sulfur up to sulfate.
4. Conclusions The results of this work demonstrated that denitrification is a feasible process for the simultaneous removal of nitrogen, carbon and sulfur. Sulfide was eliminated via partial oxidation to elemental sulfur that accumulated inside the reactor without any signs of inhibition on the process. In this way, elemental sulfur can be removed from the reactor closing the sulfur cycle. In continuous denitrifying sulfide oxidizing reactor the removal efficiencies for nitrate, acetate and sulfide were close to 100%, 69% and 100%, respectively. Additionally, it was demonstrated that the denitrification process was biologically mediated, as the chemical transformation reactions in batch experiments were incomplete and proceeded at very low rates. Sulfide drove denitrification at a rate over two orders of magnitude less than acetate, however, the presence of acetate in addition to sulfide increased both sulfide oxidation and denitrification rate by roughly 55% and an order of magnitude, respectively. The results showed in this work suggest this approach could be applied to the treatment of wastewaters from the petroleum industry.
[1] Baker LA. Design considerations and application for wetland treatment of high-nitrate waters. Water Sci Technol 1998;38(1):389–95. [2] Oenema O, Roest CW. Nitrogen and phosphorus losses from agriculture into surface waters: the effects of policies and measures in The Netherlands. Water Sci Technol 1998;37:19–30. [3] Visser JM, Robertson LA, Van Verseveld HW, Kuenen JG. Sulfur production by obligately chemolithoautotrophic Thiobacillus species. Appl Environ Microbiol 1997;63:2300–5. [4] Cadena F, Peters RW. Evaluation of chemical oxidizer for hydrogen sulfide control. J Water Pollut Control Fed 1988;60:1259–63. [5] Betlach MR, Tiedje JM. Kinetic explanation for accumulation of nitrite, nitric oxide, and nitrous oxide during bacterial denitrification. Appl Environ Microbiol 1981;42:1074–84. [6] Cervantes F, Monroy O, Go´mez J. Influence of ammonium on the performance of a denitrifying culture under heterotrophic conditions. Appl Biochem Biotechnol 1999;81:13–21. [7] Cuervo-Lopez F, Mart!ınez F, Gutierrez-Rojas M, Noyola RA, Go´mez J. Effect of nitrogen loading rate and carbon source on denitrification sludge settleability in upflow anaerobic sludge blanket (UASB) reactors. Water Sci Technol 1999;40(8):123–30. [8] Bernet N, Habouzit F, Moletta R. Use of a industrial effluent as carbon source for denitrification of a highstrength wastewater. Appl Microbiol Biotechnol 1996;46:92–7. [9] Chen SD, Chen CY, Shen YC, Chiu CM, Chen HJ. Treatment of high strength nitrate wastewater by biological methods—operational characteristic study. Water Sci Technol 1996;34(1–2):269–76. [10] Driscoll CT, Bisogni JJ. The use of sulfur and sulfide in packed-bed reactors for autotrophic denitrification. J Water Pollut Control Fed 1978;50(3):569–77. [11] Flere MJ, Zhang TC. Sulfur-based autotrophic denitrification pond systems for in-situ remediation of nitratecontaminated surface water. Water Sci Technol 1998;38: 15–22. [12] Sorensen HB, Jorgensen SE. The removal of nitrogen compounds from wastewater. Amsterdam, The Netherlands: Elsevier Science; 1993. [13] Koening A, Liu HL. Autotrophic denitrification of landfill leachate using elemental sulphur. Water Sci Technol 1996;34:469–76. [14] Trouve C, Chazal PM, Gueroux B, Sauvaitre N. Denitrification by new strains of Thiobacillus denitrificans under non-standard physicochemical conditions. Effect of
ARTICLE IN PRESS J. Reyes-Avila et al. / Water Research 38 (2004) 3313–3321
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
temperature pH and sulphur source. Environ Technol 1998;19:601–10. Gommers PJ, Buleveld W, Zuiderwijk FJ, Kuenen JG. Simultaneous sulfide and acetate oxidation in a denitrifying fluidized bed reactor-I. Water Res 1988;22:1075–83. Kim EW, Bae JH. Alkalinity requirements and the possibility of simultaneous heterotrophic denitrification during sulfur-utilizing autotrophic denitrification. Water Sci Technol 2000;42:233–8. Oh SE, Yoo YB, Young JC, Kim IS. Effect of organics on sulfur-utilizing autotrophic denitrification under mixotrophic conditions. J Biotechnol 2001;92(1):1–8. APHA, Standard methods for the examination of water and wastewater, 17th ed. New York, USA: American Public Health Association; 1989. Bartlett JK, Skoog DA. Colorimetric determination of elemental sulfur in hydrocarbons. Anal Chem 1954; 26(6):1008–11. Fass S, Ganaye V, Urbain V, Manem J, Block JC. Volatile fatty acids as organic carbon sources in denitrification. Environ Technol 1994;15:459–67. Tiedje JM. Ecology of denitrification and dissimilatory nitrate reduction to ammonium. In: Zehnder AJB, editor. Biology of anaerobic microorganisms. New York: Wiley; 1988. p. 179–244. Kim IS, Son JH. Impact of COD/N/S ratio on dentrification by the mixed cultures of sulfate reducing bacteria and sulfur denitrifying bacteria. Water Sci Technol 2000;42:69–76.
3321
[23] Lee DU, Lee IS, Choi YD, Bae HO. Effects of external carbon source and empty bed contact time on simultaneous heterotrophic and sulfur-utilizing autotrophic denitrification. Process Biochem 2001;36(12):1215–24. [24] Krishnakumar B, Manilal VB. Bacterial oxidation of sulphide under denitrifying conditions. Biotechnol Lett 1999;21:437–40. [25] Ikemoto-Yamamoto R, Komori T, Nomura M, Ide Y, Matzukami T. Nitrogen removal of hydroponic culture wastewater by autotrophic denitrification using thiosulfate. Water Sci Technol 2000;42:369–76. [26] Gommers PJ, Buleveld W, Zuiderwijk FJ, Kuenen JG. Simultaneous sulfide and acetate oxidation in a denitrifying fluidized bed reactor-I. Water Res 1988;22:1085–92. [27] Buisman CJN, Geraats BG, Ijspeert P, Lettinga G. Optimization of sulphur production in a biotechnological sulphide-removing reactor. Biotechnol Bioeng 1990;35:50–6. [28] Gomez J, Mendez R, Lema J. The effect of antibiotics on nitrification process: batch assays. Appl Biochem Biotechnol 1996;57:869–76.
Further reading Janssen AJ, Sleyster R, Van der Kaa C, Jochemsen A, Bontsema J, Lettinga, G. Biological sulphide oxidation in a fed-batch reactor. Biotechnol Bioeng 1995;47:327–33.