Environ. Sci. Technol. 2003, 37, 402-408
Activated Carbon as an Electron Acceptor and Redox Mediator during the Anaerobic Biotransformation of Azo Dyes FRANK P. VAN DER ZEE,* IEMKE A. E. BISSCHOPS, AND GATZE LETTINGA Agrotechnology and Food Sciences, Subdepartment of Environmental Technology, Wageningen University, P.O. Box 8129, 6700 EV, Wageningen, The Netherlands JIM A. FIELD Department of Chemical and Environmental Engineering, University of Arizona, P.O. Box 210011, Tucson, Arizona 85721-0011
Activated carbon (AC) has a long history of applications in environmental technology as an adsorbent of pollutants for the purification of drinking waters and wastewaters. Here we describe novel role of AC as redox mediator in accelerating the reductive transformation of pollutants as well as a terminal electron acceptor in the biological oxidation of an organic substrate. This study explores the use of AC as an immobilized redox mediator for the reduction of a recalcitrant azo dye (hydrolyzed Reactive Red 2) in laboratory-scale anaerobic bioreactors, using volatile fatty acids as electron donor. The incorporation of AC in the sludge bed greatly improved dye removal and formation of aniline, a dye reduction product. These results indicate that AC acts as a redox mediator. In supporting batch experiments, bacteria were shown to oxidize acetate at the expense of reducing AC. Furthermore, AC greatly accelerated the chemical reduction of an azo dye by sulfide. The results taken as a whole clearly suggest that AC accepts electrons from the microbial oxidation of organic acids and transfers the electrons to azo dyes, accelerating their reduction. A possible role of quinone surface groups in the catalysis is discussed.
Introduction Azo dyes represent the largest class of dyes used in textileprocessing and other industries. The release of these compounds into the environment is undesirable, not only because of their color but also because many azo dyes and their breakdown products are toxic and/or mutagenic to life. Azo dyes are generally persistent under aerobic conditions (1). However, under anaerobic conditions, they undergo facile reductive fission, yielding colorless aromatic amines (2), compounds that in turn generally require aerobic conditions for their biodegradation (3, 4). Sequential or integrated anaerobic-aerobic treatment is therefore the most logical strategy for the complete removal of azo dyes from wastewater. * Corresponding author present address: Department of Chemical Engineering, Faculty of Sciences, University of Valladolid, Paseo de Prado de la Magdalena s/n, 47005 Valladolid, Spain; telephone: ++34 983 423166; fax: ++34 983 423013; e-mail:
[email protected]. 402
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003
Though most azo dyes are fortuitously reduced under anaerobic conditions, the rate of the reaction may be low, especially for reactive azo dyes (5). This presents a problem for the application of high-rate anaerobic bioreactors for the treatment of dye-containing wastewater as long hydraulic retention times would be necessary to reach a satisfactory extent of dye reduction. However, this problem can be solved by making use of the property of redox mediating compounds to increase the rate of azo dye reduction by shuttling electrons from microorganisms or chemical electron donors to the electron-accepting azo dye. In different experimental systems, redox mediators have been demonstrated to accelerate azo dye reduction. Enzyme cofactors such as flavin adenine dinucleotide are known as effective redox mediators for azo dye reduction (6-8), and also certain quinones can act as redox mediators: in abiotic systems, quinones accelerated chemical azo dye reduction by sulfide (9) as well as electrochemical azo dye reduction (10). In biological systems, quinones were also shown to accelerate azo dye reduction by anaerobically incubated aerobic biomass (11-13) as well as by anaerobic granular sludge (14). Theoretically, feasible redox mediators for azo dye reduction should have redox potentials between those of the two eventual half reactions, the reduction of an azo dye and the oxidation of a primary electron donor. Although standard electron potentials (E0′) for the reduction of azo dyes to their constituent aromatic amines are not available, an indication can be derived from polagraphic data. Reported half-wave potentials (E1/2) range between -530 and -180 mV (15-17). For biological azo dye reduction, i.e., coupled to the oxidation of organic primary electron donors by anaerobically incubated bacteria, the E0′ value of NAD(P)H, the cofactor with the lowest electron potential (-320 mV), can be taken into account; whereas for the chemical azo dye reduction by sulfide, the E0′ value of HS-/S0 is -270 mV. Rau et al. (13) tested a wide range of quinones in their ability to mediate the reduction of the model azo dye, amaranth. Quinones with E0′ values up to about -50 mV were shown to effectively transfer electrons from microorganisms to the dye. Previous research in our laboratory demonstrated that continuous dosing of anthraquinone-2,6-disulfonate (AQDS) at catalytic concentrations strongly increases the azo dye reduction efficiencies of anaerobic bioreactors operated at hydraulic retention times realistic for wastewater treatment practice (14, 18). Although the effective AQDS dosage levels were low, continuous dosing implies continuous expenses related to procurement of the chemical as well as continuous discharge of this biologically recalcitrant compound. Therefore, it is desirable to immobilize the redox mediator in the bioreactor. For this purpose, activated carbon (AC) was considered since it is known to contain surface quinone (carbonyl) structures (19, 20). By various methods that specifically block quinone functional groups, it has been shown that the redox active groups on glassy carbon used in electrodes are quinones. These groups were shown to be solely responsible for oxidation reduction reactions of Eu3+/Eu2+, V3+/V2+, and Fe3+/Fe2+, with E0′ values ranging from -430 to +771 mV (19). Also S2- (E0′ ) -270 mV) is known to be oxidized by AC, although quinone moieties have not yet been implicated (21-23). Depending on type and pretreatment of AC, the surface quinone concentration may amount up to a few millimoles per gram (20). In this study, we investigated the feasibility of AC as a redox mediator for the anaerobic reduction of azo dyes. Its potential to improve the azo dye reducing activity of 10.1021/es025885o CCC: $25.00
2003 American Chemical Society Published on Web 12/06/2002
FIGURE 1. Structure formulas of the dyes used in this study. anaerobic granular sludge was tested in a number of labscale upflow anaerobic sludge bed (UASB) reactors and in supporting batch experiments.
Experimental Section Materials. The azo dye reactive red 2 (RR2, 45%) was obtained from BASF (Arnhem, The Netherlands) as an industrial preparation, and the azo dye acid orange 7 (AO7, 98%) was obtained from Aldrich (Gillingham, U.K.). For the experiments described in this paper, RR2 was hydrolyzed by heating at alkaline pH according to previously described protocol (24), which has the effect of replacing the chloro with hydroxy groups located on the triazyl ring. The hydrolysis was carried out to simulate the form of RR2 that is discharged with textile industry effluent. Hydrolyzed RR2 (HRR2) is also less toxic than RR2 (14). (See Figure 1 for structures). AC was obtained from Norit b.v. (Amersfoort, The Netherlands). Two types of AC were used: SA-4 (steam activated carbon) and SX-4 (acid washed, steam activated carbon). Both carbon types were used as obtained from the supplier, without pretreatment of any kind. Reactor seed matter was anaerobic granular sludge from a full-scale UASB reactor treating alcohol distillery wastewater (Nedalco, Bergen op Zoom, The Netherlands). The inoculant of the biological AC reduction experiments was either crushed and diluted granular sludge from a lab-scale UASB reactor in which acetate oxidation was coupled to the reduction of AQDS or an AQDS/acetate enrichment culture derived from this sludge, which was predominated by a bacterium with 97% similarity to Geobacter sulfurreducens (25). Adsorption Isotherms. Adsorption isotherms were conducted to estimate the extent of adsorption of the HRR2 to AC (Norit SA-4). Variable quantities of HRR2 were added from a 4 mM stock solution to 117-mL serum flasks containing 50 mg of AC in 50 mL of 0.1 M pH 7.0 phosphate buffer. The flasks were sealed and incubated at 22 °C in a rotary shaker at 100 rpm. After 24 h, samples were centrifuged (10 min at 10 000 rpm), and absorbance of the residual soluble HRR2 was measured spectrophotometrically at λmax (539 nm). Continuous Bioreactors. Lab-scale UASB reactors (liquid volume 0.25 L) were initiated with 35 g of volatile suspended solids (VSS) L-1 of the anaerobic granular sludge and with a neutralized volatile fatty acid (VFA) mixture (1.5 g of COD L-1 at a 1:1:1 COD-based ratio of acetate, propionate, and butyrate) in basal nutrient medium containing (mg L-1) NH4Cl (280), CaCl2 (5.7), KH2PO4 (250), MgSO4‚7H2O (100), H3BO3 (0.05), FeCl2‚4H2O (2), ZnCl2 (0.05), MnCl2‚4H2O (0.5), CuCl2‚2H2O (0.04), (NH4)6Mo7O24‚5H2O (0.05), CoCl2‚6H2O (1), NiCl2‚6H2O (1), and Na2SeO3‚5H2O (0.16). The hydraulic retention times of the reactors were kept constant within the range of 5-5.5 h. Effluent was recycled at a 1:1 influent: effluent flow ratio. After a 15-day start-up phase in the absence of dye, HRR2 was added to the influent at a concentration of 42 mg/L (0.073 mM) of the reactors. The day in which dye was first added to the reactor influents was defined as day 0 of the reactor operation. Three reactors were used. One reactor was started up with 2.5 g of AC (Norit SA-4) mixed
with the granular sludge. The second reactor, which lacked any AC addition, served as the control reactor. The third reactor, originally operated identical to the control reactor, was amended with 0.1 g of AC (Norit SA-4) at day 46. The reactors’ dye and VFA removal efficiencies were monitored regularly. Chemical Dye Reduction. A series of batch experiments was conducted to determine whether AC could accelerate the chemical reduction of an azo dye by sulfide. The defined azo dye, acid orange 7 (AO7) was utilized in these experiments. AO7 reduction by sulfide was monitored in the presence and in the absence of 100 mg L-1 AC. Controls without sulfide were also evaluated to correct for dye adsorption as well as to verify the stability of the dye. The experiments were performed in triplicate. AC (5.0 mg of either Norit SA-4 or Norit SX-4) was added to 117-mL glass serum flasks. A 60 mM NaHCO3 solution was added to a liquid volume of 50 mL. The flasks were sealed with butyl rubber stoppers, and the gas headspace was flushed for 5 min with oxygen-free flush gas containing N2:CO2 (80%:20%). The thus obtained pH was 7.4-7.5. Sulfide was added with a syringe from a 0.1 M Na2S stock solution to obtain an initial total sulfide concentration of either 0.5 or 1.7 mM. All vials were incubated at 25 °C in a rotary shaker at 50 rpm. After 1-day preincubation, the total sulfide concentration was measured, and AO7 was added from a concentrated stock solution to obtain an initial AO7 concentration of 0.14 mM. At selected time intervals samples were withdrawn. Samples for dye measurements were centrifuged (3 min at 10 000 rpm) and analyzed spectrophotometrically for dye adsorbance at λmax (484 nm). Samples for HPLC measurement of the dye reduction product sulfanilic acid (SA) were diluted with zinc acetate (40 mM) to stop the reaction by precipitating sulfide as ZnS. Controls showed that there was no reaction between sulfanilic acid and Zn2+ or ZnS. Furthermore, it was investigated whether AO7 adsorbed to AC could be reduced by sulfide. AC with adsorbed AO7 from sulfide-free control flasks was sampled (five aliquots of 1.5 mL), and each aliquot was centrifuged for 10 min at 10 000 rpm. The liquid phase was decanted, and the carbon pellet was washed (three times) with demineralized water. The washed pellet was mixed with 0.75 mL of 5 g L-1 NaHCO3, and the centrifuge cup was placed in a glass flask. Next, the flask was sealed with a rubber septum, and the gas phase was flushed for 5 min with oxygenfree flush gas (N2:CO2 80%:20%). Sulfide (0.25 mL of a 0.1 M Na2S stock solution) was added with a syringe to each centrifuge cup, and the glass flask was incubated at 25 °C. SA in the liquid phase was sampled after 1 h and after 1 day and was measured by HPLC. Control experiments showed that there was no adsorption of SA to AC. AC as Electron Acceptor for Microorganisms. Batch experiments were conducted to determine whether AC could act as the terminal electron acceptor for the biological oxidation of acetate. Acetate concentrations and the amount of reduced AC was followed in flasks incubated anaerobically with either the diluted crushed granular sludge or the Geobacter sp. enrichment culture. A set of controls excluding either bacteria or AC or acetate was incorporated. The VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
403
experiments were performed with triplicate vials. AC (1.00 or 2.00 g of either Norit SA-4 or Norit SX-4) was added to glass serum vials (V ) 117 mL). A 60 mM NaHCO3 solution in basal nutrient medium was added to a liquid volume of 50 mL, and acetate was added by pipet from a neutralized stock solution. The vials were sealed with butyl rubber stoppers, and the gas headspace was flushed for 5 min with oxygen-free flush gas (N2:CO2 70%:30%). The thus obtained pH was 7.2-7.3. Next, the vials were autoclaved in a pressure cooker (20 min at 120 °C) and allowed to cool. Under sterile conditions, the methanogenesis inhibitor 2-bromoethane sulfonic acid (BES) was now added with a syringe from a filter-sterile concentrated stock solution to a concentration of 30 mM. Finally, the crushed sludge or the enrichment culture was added. At selected time intervals, medium or activated carbon was sampled under sterile or aseptic conditions, respectively. Acetate was measured by gas chromatography, and the reduction equivalents of activated carbon were measured by the ferrozine technique. At the end of the experiment, headspace samples for methane and medium samples for sulfide were taken. Analytical. AO7 and RR2 was measured spectrophotometrically at the dyes’ wavelengths of maximum visible absorbance (484 and 539 nm, respectively). Liquid-phase samples (0.75 mL) were centrifuged (2 min at 10 000 rpm) and diluted to an absorbance of less than 0.8 in a 0.1 M phosphate buffer at pH 7.0. The buffer contained freshly added ascorbic acid (200 mg L-1) to prevent autoxidation of dye degradation products that would interfere in the measurement (reactor experiments only). Without dye, light absorbance of medium and buffer was less than 1% of the absorbance right after dye addition and could therefore be neglected. Sulfide was determined colorimetrically after reaction with N,N-dimethyl-p-phenylenediamine oxalate according to the method described previously (26). Volatile fatty acids (VFA) and methane were determined by gas chromatography as described earlier (14). SA was measured by high-performance liquid chromatography. The chromatograph (Hewlett-Packard 1090) was equipped with a column LiChrocart 250 × 3 mm packed with LiChrospher 100 RP-18 (5 µm) from Merck. Samples were diluted 1:1 with a 40 mM zinc acetate solution to precipitate the sulfide as ZnS. After centrifugation (10 min at 10 000 rpm), 5 µL was injected with an autosampler. The carrier liquid, composed of 98% acetic acid (0.5% at pH 5.9) and 2% methanol with demineralized water (98%:2%), was pumped at a flow rate of 300 µL min-1. SA was detected at its absorbance maximum (248 nm) with a HP diode array detector. Analysis of reduced AC using the ferrozine technique (27) was carried out in an anaerobic chamber under N2/H2 (95%:5%) atmosphere. Samples reacted with iron(III)citrate at low pH, yielding Fe(II). Next, the absorbance of the purple complex of Fe(II)-ferrozine [3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine, monosodium salt; molar extinction coefficient 28 × 103 cm-1 M-1] was measured spectrophotometrically. When no Fe3+ was added, no Fe2+ was formed: the AC itself did not contain any ferrous iron. Volatile suspended solids (VSS) were determined according to standard methods (28).
Results Adsorption isotherms of HRR2 to AC (Norit SA-4) are plotted in Figure 2 together with their fits to the Freundlich and Langmuir equations. The Freundlich equation provided a better fit as compared to the Langmuir equation. The maximum adsorption capacity that could be estimated from the Langmuir fit was 54.9 mg of dye/g of AC, and the affinity of the carbon for the dye was very high. The observed maximum HRR2 sorption capacity was 59.2 ( 3 mg of HRR2/g of Norit SA-4. 404
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003
FIGURE 2. Adsorption isotherm for the adsorption of HRR2 on AC (Norit SA-4). Experimental data (full circles) with standard deviations (error bars); Freundlich fit: Q ) KCn ) 28.2C0.17 (dotted line), where Q is the mg of sorbed dye/g of AC; C is the equilibrium concentration (mg/L); K and n are constants. Langmuir fit: Q ) (Q0C)/(K + C) ) (54.9C)/(0.31 + C) (dashed line), where Q is the mg of sorbed dye/g of AC; C is the equilibrium concentration (mg/L); Q0 is the maximum sorption capacity (mg of sorbed dye/g of AC); and K is the affinity constant (mg/L).
FIGURE 3. Dye removal efficiency (circles) and effluent aniline concentration (triangles) of the reactor with 2.5 g of Norit SA-4 (open markers) and the control reactor without AC (closed markers). Aromatic amine analysis was only conducted between day 3 and day 58. The effect of AC on reductive biotransformation of azo dyes was studied in laboratory-scale UASB reactors. The reactors were initiated with anaerobic granular sludge (8.75 g of VSS) and fed with the VFA substrate (1.5 g L-1 chemical oxygen demand) in basal nutrient medium. The dye, HRR2 (42.3 mg L-1; 0.073 mM), was added to the influent after a 15-day start-up phase in the absence of dye. The moment of dye addition was defined as the start of the experiment (day 0 in Figures 2 and 3). At this point in time, acetate and butyrate were removed with a high efficiency (>95%) that continued for the rest of the reactor operation. Stable propionate removal (>95%) was achieved from day 25 onward. In the first phase of the experiment, two reactors were used. One reactor was started up with 2.5 g of Norit SA-4 mixed with the granular sludge, whereas the other reactor was the control reactor to which no AC was added. The dye removal efficiencies (based on λmax absorbance measurements) of these reactors are depicted in Figure 3. The dye removal efficiency of the control reactor was much lower than that of the AC-amended reactor. In the control reactor, the dye removal during the first 2-4 weeks could be partly attributed to dye adsorption onto the reactor sludge. As the adsorption capacity of the sludge became exhausted, the dye removal efficiency decreased to about 35%. This level, presumably corresponding to biological reduction of HRR2, remained more or less stable for the rest of the experiment (up to day 130). In the AC-amended reactor, the initial dye removal efficiency was almost complete (97%) and gradually decreased, probably due to some wash-out of AC, to
FIGURE 4. Dye removal efficiency (circles) and effluent aniline concentration (triangles) of the reactor with 0.1 g of Norit SA-4 (open markers) and the control reactor without AC (closed markers). Aromatic amine analysis was only conducted between day 3 and day 58. approximately 90% after 40 days and remained at that level for the rest of the experiment. During the entire experiment (130 day), the AC-amended reactor had removed 5.542 g of HRR2, whereas the control reactor had removed 2.242 g of HRR2. Therefore, 3.300 g of HRR2 had been additionally removed in the carbon-amended reactor. With the observed HRR2 sorption capacity of the AC that was used in these experiments (maximum 59.2 ( 3 mg of HRR2/g of Norit SA-4), it can be estimated that maximally 0.1481 g of HRR2 would be removed by AC adsorption. Consequently, an additional dye removal of 3.300 g of HRR2 exceeds the sorption capacity by more than 22 times. To verify the enhanced conversion of the dye, the reaction product (aniline) was monitored at selected time points between day 3 and day 58. From the aniline data presented in Figure 3, it can be seen that the higher dye removal efficiency of the AC-amended reactor corresponds to higher aniline concentrations in the reactor effluent during steadystate operation, confirming an enhanced reduction of the dye because of AC catalysis. In both reactors, the aniline recovery corresponded to a large extent to the expected stoichiometry based on dye removal (0.68 ( 0.15 and 0.79 ( 0.12 mol:mol aniline formed:dye removed for the control reactor and the AC-amended reactors, respectively), indicating that the reduction was the major mechanism of dye removal. In the second part of the experiment, an additional reactor was used in which a much smaller amount of AC was supplied (0.1 g of Norit SA-4). The performance of this reactor was also compared in parallel with the performance of the control reactor described previously. Addition of AC was postponed until day 46 to verify that both the AC-amended and the control reactors performed the same before AC addition. Figure 4 confirms that both reactors had the same pattern of dye removal preceding the addition of AC. Immediately after AC addition, dye removal efficiency increased to 76% from the prior value of 32%. In the following 3 weeks, the dye removal efficiency rapidly declined from 78% to 50%. This decline was possibly due to a rapid exhaustion of the dye adsorbing capacity as well as to the wash out of some of the added AC, thereby lowering concentration of AC as chemical catalyst. Thereafter, the dye removal efficiency declined gradually further to approximately 43% after 112 days of reactor operation with AC. The dye removal efficiency of the AC-amended reactor throughout the entire experiment remained significantly higher than that of the control reactor. During the 158 days of operation, the AC reactor had removed 3.377 g of HRR2, whereas the control reactor had removed 2.665 g of HRR2.The difference of 0.712 g of HRR2 had been
FIGURE 5. Activated carbon catalysis of AO7 reduction by sulfide. Dye concentration (full markers) and SA concentration (open markers): squares, AO7 + sulfide + activated carbon; circles, control without AC; triangles, control without sulfide; diamonds, dye only. The error bars represent standard deviations between triplicate measurements. (A) 100 mg L-1 Norit SA-4; initial totalsulfide concentration 0.51 ( 0.02 mM; (B) 100 mg L-1 Norit SA-4; initial total sulfide concentration 1.73 ( 0.05 mM. additionally removed in the AC-amended reactor. In this reactor, the estimated dye removal due to sorption is very limited: using the observed maximum RR2 sorption capacity of 59.2 mg of HRR2/g of Norit SA-4, AC sorption would only account for the removal of 0.0059 g of HRR2. Consequently, the observed additional dye removal of 0.712 g of HRR2 exceeds the dye sorption capacity of AC by more than 120 times. These data suggest that the major role of AC was to enhance the chemical conversion of dye rather than dye adsorption. The results from these continuous experiments indicate a role of AC in the catalysis of azo dye reduction. To get more evidence for the catalysis, we conducted two supporting batch experiments. The first experiment was meant to assess catalysis of AC in the direct chemical reduction of an azo dye by sulfide. The second experiment was designed to demonstrate that AC could accept electrons from the microbial oxidation of the VFA substrate. A first series of batch experiments was performed to determine whether low amounts of activated carbon could accelerate the simple chemical reduction of azo dyes by sulfide. The conversion of a structurally well-defined azo dye of high purity, Acid Orange 7 (AO7), to sulfanilic acid (SA), one of the dye’s reductive cleavage products, was monitored in time in the presence and absence of AC. Controls without sulfide were incorporated to determine the extent of dye adsorption and to verify the stability of the dye. The results of these experiments, presented in Figure 5, show that AC strongly accelerates the reduction of AO7 and the concomitant production of SA. With 100 mg L-1 Norit SA-4 and an initial sulfide concentration of 0.5 mM, the AO7 absorbance rapidly decreased by 80% within 5 days to its final level; whereas, without AC, only 40% was removed within 2 weeks (Figure 5A). The AO7 absorbance did not completely disappear but remained at 20% of its original value. Although there is a certain threshold level of ∼0.1 mM sulfide below which no dye reduction occurs (data not shown), 0.5 mM VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
405
sulfide should have been sufficient to reduce 0.15 mM AO7. However, since the sulfide concentration was lowered by AC adsorption, these data suggest that the adsorbed sulfide is not available for azo dye reduction, accounting for incomplete reduction AO7. The reduced product, SA, was formed only in treatments with sulfide. No SA was formed, but a dye removal of 22% was observed if AC was used in the absence of sulfide. These results indicate that the removal of AO7 in the presence of sulfide and AC was a chemical reaction, whereas the removal of AO7 in the presence of AC alone was only due to adsorption. The AO7 adsorption capacity was 107.9 mg (0.31 mmol)/g of Norit SA-4. A second experiment was conducted using the same quantity (100 mg L-1) of a different type of AC (Norit SX-4, acid washed Norit SA-4). The results of this experiment (data not shown) were similar to those of the experiment with Norit SA-4, aside from a relatively higher degree of dye adsorption (42%) by the AC in sulfide-free controls, indicating an AO7 adsorption capacity of 206 mg (0.59 mol)/g of Norit SX-4). At the end of this experiment, dye-sorbed carbon from vials with AC in the absence of sulfide was sampled, washed, and centrifuged. Next, a concentrated sulfide solution (25 mM) was added, and samples were taken to determine whether the adsorbed dye would be reduced. After 1 h of anaerobic incubation, 0.02 mmol of SA was released per gram of Norit SX-4 and 1 day of incubation resulted in the release of 0.59 mmol SA/g of Norit SX-4, a molar amount almost equal to the AO7 that was estimated to have been adsorbed. Therefore, the adsorbed dye is reversibly bound to the carbon and can be reduced. In a third experiment, the AC Norit SA-4 was used again but with a higher initial sulfide concentration (1.73 ( 0.05 mM). Figure 5B shows that in this experiment the reduction of AO7 to SA in the treatment with AC and sulfide proceeded more rapidly, mainly within the first day of incubation. In contrast, reduction of the dye by sulfide in the absence of AC took more than 7 days. Also in this experiment, no SA was formed in the assays without sulfide. The molar fraction of removed AO7 that was recovered as SA in the assays with AC and sulfide was 85 ( 1 and 80 ( 5% for the experiments presented in Figure 5, panels A and B, respectively. If AC accelerates the reduction of azo dyes in the bioreactors via redox mediation, the bacteria in the sludge must be able to transfer electrons to AC. Therefore, it should be feasible to demonstrate that AC is a terminal electron acceptor for the oxidation of organic substrates. To test the hypothesis for biological oxidation of acetate, two quinone respiring bacterial consortia were evaluated on the basis of the assumption that quinone surface groups on AC were the redox mediating moieties. Previous research has demonstrated that certain bacteria (e.g., G. metallireducens) can utilize model quinone compounds such as AQDS as terminal electron acceptors to support anaerobic respiration (29). Various anaerobic consortia from sludges and sediments were also shown to utilize AQDS as an electron acceptor (30). The two consortia utilized for the experiment were both cultivated with acetate and AQDS (25). The microbial consortia were incubated with acetate and a large amount of AC (20 or 40 g L-1). To prevent the flow of electrons to methanogens, the cultures were supplemented with 30 mM 2-bromoethanesulfonate (BES). The acetate concentration as well as the concentration of reduced carbon (measured after reaction with Fe3+ as Fe2+ equivalents) was monitored in time. Controls without acetate, without AC, and without bacteria were incorporated. The time course of acetate consumption and the concomitant formation of additional reducing equivalents in AC is shown in Figure 6 for a typical experiment with crushed granular sludge. The results clearly show that the microbial consortium is only able to significantly consume acetate when AC is provided as an electron acceptor and 406
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003
FIGURE 6. Oxidation of acetate and concomitant reduction of AC (20 g L-1 Norit SA-4) by a crushed granular sludge inoculum in basal nutrient medium with 60 mM NaHCO3 buffer under a 80%:20% N2:CO2 atmosphere at pH 7.2 and T ) 30 °C. Acetate concentration (full symbols) and AC-reducing equivalents concentration (open symbols): treatments with acetate, AC and inoculum (circles); treatments with acetate and inoculum (triangles); treatments with acetate and AC (squares); treatments with AC and inoculum (diamonds). The error bars represent standard deviations between triplicate measurements.
TABLE 1. Microbial Acetate Oxidation with AC as the Terminal Electron Acceptor change in concn (mequiv e-/L)d
ratio (mequiv e-:mequiv e-)
AC type (concn, g L-1)a
culturec
acetate consumede
reduced AC formedf
reduced AC formed:acetate consumed
Norit SA-4 (20) O2-Norit SX-4 (20)b Norit SX-4 (20) Norit SX-4 (40) Norit SA-4 (20) Norit SA-4 (20)
CGS CGS CGS CGS EC EC
14.1 ( 0.7 16.1 ( 0.3 10.5 ( 0.5 25.3 ( 0.8 14.4 ( 0.3 ndg
6.3 ( 0.6 6.3 ( 0.3 5.4 ( 0.3 13.1 ( 1.5 7.3 5.2 ( 0.6
0.45 0.39 0.51 0.52 0.50 ndg
a Per liter of liquid. b O -pretreated Norit SX-4 (moist AC was 2 preincubated with 100% O2 for 18 h). c CGS, crushed granular sludge originating from anaerobic bioreactor oxidizing acetate at the expense of AQDS reduction; EC, a Geobacter-dominated acetate-oxidizing AQDSreducing enrichment culture derived from CGS. d All units are in electron milliequivalents per liter (mequiv e-/L), 1 mmol of acetate is equivalent to 8 mequiv e-; 1 mmol of Fe2+ formed from the reaction of reduced AC with Fe3+ is equivalent to 1 mequiv e-. e Overall decrease in acetate concentration due to the presence of AC (corrected for acetate consumption in the presence of bacteria in the absence of AC). f Overall increase in AC reducing equivalents per liter in the presence of acetate and bacteria. g nd, no data (not measured).
that AC is only reduced when acetate is provided as an electron donor. The results from additional experiments utilizing either a different type of AC or the enrichment culture provided similar results, as are summarized in Table 1. Headspace samples analyzed for methane showed that methanogenesis did not occur in any of the assays (data not shown). The removal of acetate in the control assays without bacteria was negligible. The removal of acetate in the controls without AC was less than 20% of the removal of acetate in the assays with AC (
