Electrolysis within anaerobic bioreactors stimulates breakdown of toxic products from azo dye treatment
Azo dyes are the most widely used coloring agents in the textile industry, but are difficult to treat. When textile effluents are discharged into waterways, azo dyes and their degradation products are known to be environmentally toxic. An electrochemical system consisting of a graphite-plate anode and a stainless-steel mesh cathode was placed into a lab-scale anaerobic bioreactor to evaluate the removal of an azo dye (Direct Black 22) from synthetic textile wastewater. At applied potentials of 2.5 and 3.0 V when water electrolysis occurs, no improvement in azo dye removal efficiency was observed compared to the control reactor (an integrated system with electrodes but without an applied potential). However, applying such electric potentials produces oxygen via electrolysis and promoted the aerobic degradation of aromatic amines, which are toxic, intermediate products of anaerobic azo dye degradation. The removal of these amines indicates a decrease in overall toxicity of the effluent from a single-stage anaerobic bioreactor, which warrants further optimization in anaerobic digestion.
KeywordsAnaerobic treatment Electric potential Azo dye Direct Black 22 Aromatic amines Toxicity Electrolysis eAD
Textile wastewater is often treated through physicochemical processes such as coagulation and flocculation. However, large volumes of chemical coagulants are consumed, resulting in the production of substantial volumes of sludge (Solanki et al. 2013). Biological treatment is an alternative with a lower sludge production potential. Azo dyes are the most commonly used coloring agents in the textile industry, and can be removed biologically through two subsequent steps: (1) an anaerobic step in which reducing equivalents reduce the azo bond, producing aromatic amines and CO2; and (2) an aerobic step in which the aromatic amines are mineralized (Pandey et al. 2007; Sarayu and Sandhya 2012). Considering the toxicity of aromatic amines, in addition to azo dyes (Carliell et al. 1995), studies have reported the success of these sequential steps (anaerobic and aerobic) to remove the toxicity related with textile effluents (Ferraz et al. 2011; Silva et al. 2012).
Bioelectrochemical systems (BESs) consist of electrodes with bacteria performing extra-cellular electron transfer either at the anode (using the electrode as an electron acceptor) or cathode (using the electrodes as an electron donor). If the potential difference (~0.5 V) between anode and cathode is spontaneous, electric power can be generated, and the BES is referred to as a microbial fuel cell (MFC). Besides anaerobic oxidation of organic materials at the anode, MFCs have also been investigated to reduce aromatic compounds at the cathode, including nitrobenzene (Mu et al. 2009a), p-nitrophenol (Shen et al. 2012), and azo dyes (Mu et al. 2009b). The MFCs that were studied for azo dye removal consisted of two-chamber reactors separated by an ion exchange membrane (Sun et al. 2013; Fernando et al. 2013), which would increase capital and operating costs for full-scale deployments. Here, we will discuss an electrochemical system with electrodes in an anaerobic bioreactor that is not a BES because the applied potential between the electrodes is too high to sustain extra-cellular electron transfer. Instead, this high potential will split water into oxygen (anode) and hydrogen (cathode) via electrolysis.
Such an electrochemical system with power-source-controlled electrodes, and without membranes, had already been developed in anaerobic environments by at least three approaches. First, Tartakovsky et al. (2011) treated domestic wastewater in an anaerobic bioreactor with electrodes, which they referred (and here we will refer) to as electrolysis-enhanced anaerobic digestion (eAD). By applying 2.8–3.5 V, the electrolysis of water supplied oxygen and hydrogen, which facilitated hydrolysis of the wastewater and improved methane production by 10–25 % by hydrogenotrophic methanogens, respectively. Second, Chun et al. (2013) treated polychlorinated biphenyls (PCBs) in anaerobic sediments with and without an applied potential difference between electrodes. The authors referred to this process as electrical stimulation and found 40–60 % PCB removal only when the potential was applied. Third, Zhu et al. (2014) placed electrodes in upflow anaerobic sludge blanket (UASB) reactors and applied 2.5–5 V to enhance the reductive transformation of p-chloronitrobenzene through the generation of H2 gas.
The application of eAD to stimulate the degradation of azo dyes and removal of aromatic amines has not been previously investigated. We theorized that by utilizing applied potentials of ~2–5 V, hydrogen, which is produced at the cathode, can improve azo bond breakdown under anaerobic conditions; while the generation of small amounts of oxygen at the anode can promote aerobic cleavage of the aromatic amine rings. The use of small amounts of oxygen in anaerobic digesters by, for example, air sparging was previously investigated and methanogenesis was not inhibited (Kato et al. 1993). In-situ oxygen generation with electrodes may be advantageous compared to localized oxygen sparging because of a better control to prevent overdosing of oxygen during full-scale applications. However, research with large-scale digesters must be performed to prove this point.
The aromatic amines that are formed after azo dye degradation range from aniline to complex molecules with conjugated aromatic or heterocyclic structures and multiple substituents (Pinheiro et al. 2004). The production of oxygen through electrolysis would result in a similar complete degradation of, for example, aniline, which was reported by Tan et al. (1999) after addition of oxygen to a methanogenic system. Here, we utilized a lab-scale eAD system to investigate whether in-situ electrolysis can promote the anaerobic degradation of an azo dye and the removal of the corresponding aromatic amines.
Materials and methods
Reactors and inoculum
Parameters for the five operating phases (PI–PV)
Period of operation (day)
Applied potential (V)a
Dye concentration (mM)
Abiotic assays were conducted after the end of the operating period to verify the occurrence of adsorption and electrochemical reduction of the azo dye DB22 (without interference of microbes). Both reactors were cleaned and soaked in a bath of household bleach and tap water (1:5 v/v) for 1 h of disinfection. The reactors were then rebuilt and fed the same substrate as used in PII, to quantify oxygen and hydrogen, which were electrolytically produced with an applied potential of 2.5 V in R1 without biomass. The experiment was considered completed when the effluent dye concentration equaled that of the influent.
The reactors were monitored by measuring flow rate, dye concentration, and biogas production daily; while COD, pH, alkalinity, and volatile fatty acids (VFAs) were measured three times per week. VFAs were analyzed by gas chromatography, with samples filtered at 0.22 μm and diluted with formic acid to 2 % (1:1 v/v). An HP 6890 series gas chromatograph (Hewlett Packard, Anaheim, California) was used at the following operating conditions: capillary column (NUKOL, fused silica, 15 m × 0.53 mm × 0.50 µm film thickness; Supelco Inc., Bellefonte, PA); air, helium and hydrogen flow rate of 380, 30, and 35 mL min−1, respectively; FID detector at 275 °C; inlet at 200 °C; and oven ramp temperature program (initial temperature 70 °C for 2 min; temperature ramp of +12 °C per min to 200 °C; final temperature 200 °C for 2 min). The biogas composition was measured by an SRI 8610C gas chromatograph (Torrance, CA) with the detector and TCD cell set to 100 °C and oven to 40 °C. Alkalinity was measured as reported by Dilallo and Albertson (1961). The azo dye concentration was measured photometrically at the wavelength of maximum absorbance for DB22 (475 nm). Samples were filtered (5 µm) and diluted in a phosphate buffer (9.61 g L−1 NaH2PO4, 4.78 g L−1 Na2HPO4, and 0.2 g L−1 ascorbic acid) to prevent auto-oxidation.
Scanning in the light absorption range of 200–550 nm was performed with influent and effluent reactor filtered samples (0.45 µm) every 15 days to qualitatively evaluate the aromatic amines formation, as suggested by Pinheiro et al. (2004). This method is commonly used for this purpose (Baeta et al. 2012; Amaral et al. 2014) because the aromatic amine absorbs light in the 288–300 nm range, without interference of contaminants that are usually present in textile wastewaters or are byproducts of anaerobic degradation. According to Pinheiro et al. (2004) other compounds that can exhibit considerable influence are residual dyes, but their absorption characteristics are much higher than the aromatic amines [in the visible region (400–700 nm)]. All other parameters were determined according to Standard Methods for the Examination of Water and Wastewater (APHA 2005). Student’s t tests were applied to verify the statistical difference between the performance of R1 and R2.
Results and discussion
From the beginning of PII, about 10 % of hydrogen was detected in the biogas of R1 with an applied potential of 2.5 V due to hydrogen production via water electrolysis (Fig. 3c). When the applied potential was temporarily changed to 5 V (day 39 in PII), enhanced water electrolysis resulted in an immediate increase in gas production from 0.356 to 1.662 L day−1. The headspace gas composition changed from 51 % CH4, 27 % CO2, 11 % H2, 10 % N2, and 0.3 % O2 to 5 % CH4, 22 % CO2, 69 % H2, 1 % N2, and 2 % of O2. On that same day, the dye removal efficiency dropped from 65 to 27 %, which is illustrated by an increased dye concentration (DB22 concentration) in the effluent of R1 only (Fig. 3d). Because anaerobic microbes are responsible for methane formation and the azo dye removal process, it is likely that the excess oxygen (2 % of the biogas) was the main contributing factor for the decrease in the methane production and the dye removal efficiency. Species members of the genera Klebsiella (Franciscon et al. 2009), and Clostridium (Cui et al. 2012) are the most commonly microbes detected in azo dye anaerobic degrading reactors. Most of Clostridium species show growth inhibition in the presence of oxygen in the headspace ranging from 0.5 % (Loesche 1969) to 1 % (Kawasaki et al. 2005); while the oxygen sensitivity of Klebsiella species has also been demonstrated (Kong et al. 1986). Therefore, the potential was returned to 2.5 V to reestablish the performance observed in R1 during PII. No hydrogen was detected in the R2 biogas during the operating period as anticipated without the applied potential (Fig. 3c).
We did not find significant differences in DB22 removal efficiencies between the reactors during the rest of the operating period (p = 0.47). This confirms our assumption that the microbes were unable to directly use the electrons from the cathode to perform azo dye reduction at such high applied potential differences, and, in addition, indicates that the generated hydrogen did not stimulate microbial breakdown of the dye. Wang et al. (2013b) had previously reported a bioelectrochemical activity of microbes at a biocathode by using a single-chamber system with lower potentials. Although these authors had found a 13.3 % increase in dye removal efficiency using the electrons at their biocathodes, their dye removal rate was 3-fold lower than our dye removal rate. Thus, we may not have been able observe such a relatively small stimulating effect of dye removal at the cathode (even if it existed, which is unlikely). In addition, in our work the excess of electrons from starch (0.14 mol e− day−1) in relation to the electrons that are required to break the azo bonds (1.22 × 10−3 mol e− day−1) may have provided a non-ideal environment for stimulating (bio)electrochemical dye removal (additional electrons and H2 are not necessary). Regardless, the chosen applied potentials differences of 1–3 V were likely too high to stimulate bioelectrochemical activity for azo dye breakdown, and microbial breakdown due to additional hydrogen was not observed.
When considering the treatment of the dye DB22, the production of aromatic amines as intermediate products must be accounted for because of their toxic nature. We started to detect aromatic amines on day 24 of the operating period (PII) by spectrophotometer absorbance measurements with a maximum aromatic amines absorption in the UV range of 288–300 nm (Fig. 5b). We found a decreased absorbance for aromatic amines (lower concentrations) in the effluent of R1 when compared with R2 (Fig. 5b). With the increase in the applied potential from 1 to 2.5 V for PII, favorable conditions for water electrolysis were created (Vanýsek 2012), leading to the detection of hydrogen in the biogas (Fig. 3c). The low amount of O2 detected in the R1 biogas during PII (0.82 mL of O2 day−1) compared with those from the abiotic assay (26.75 mL of O2 day−1) indicates that the oxygen was consumed in R1. Some of the oxygen consumption in R1 was likely used because of the aerobic cleavage of aromatic amines.
The measured aromatic amines, which were formed via azo dye reduction, likely consisted of sulfonated aromatic amines because: (1) the degradation of one molecule of azo dye DB22 can release three anilines, one sulfanilic acid, and two naphthalenesulfonic acid as aromatic amines; and (2) sulfonated aromatic amines are not biodegradable under methanogenic conditions (Tan et al. 2005). Sulfanilic acid was the sulfonated aromatic amine that had accumulated in the anaerobic reactor containing BES with membranes, which was operated by Kong et al. (2015). The authors used the azo dye acid orange 7, and mentioned that they did not observe other aromatic amines, similar to aniline due to its low stability and aerobic autoxidation. The aerobic degradation of sulfanilic acid has been demonstrated (Chen et al. 2012). The aerobic desulfonation reactions generally involve hydroxylation catalyzed by either mono or dioxygenases with catechol sulfonates as intermediates (Pandey et al. 2007). In addition, the toxic and carcinogenic potential of the aromatic amine aniline and naphthalenesulfonic acid was reported (Pinheiro et al. 2004). Both can also be removed under aerobic conditions (Wang et al. 2011; Keck et al. 1997).
Considering the behavior of R1 during PII in which stable operation was observed with low energy consumption (1.17 Wh/d to supply 2.5 V), a greater methane yield was measured in R1 (Yapp = 0.15 L of CH4/g of removed COD) compared to R2 (Yapp = 0.12 L of CH4/g of removed COD). This makes the energy balance favorable (1.71 Wh/d from 0.15 L of CH4/g of removed COD, considering 890 kJ/mol heat from methane combustion) to apply this potential difference between electrodes for the removal of DB22 and its corresponding aromatic amines in a single-stage anaerobic reactor (eAD). When compared to a two-stage anaerobic and aerobic reactor configuration for azo dye treatment, the benefits of energy production and increased amine mineralization from this system look promising for further evaluation. Future studies in system optimization will explore the effect of applied potentials, electrode materials, and reactor architecture over longer periods of operating time.
The removal efficiency of azo dye DB22 was not improved with an eAD system compared to a conventional anaerobic digestion system. However, the oxygen generated by electrolysis at the applied potentials of 2.5 and 3.0 V promoted the removal of aromatic amines, which are important toxic compounds in textile effluent. Further increases in applied potential (from 2.5 to 5.0 V) caused a decrease in both the methane production and DB22 removal efficiency due to excessive oxygen generation. Before any practical application should be envisioned, more experiments are needed with other azo dyes, which can be more recalcitrant than DB22, or have auto-oxidizing breakdown products.
The authors thank the Brazilian agency CNPq for a Post-Doctorate Scholarship (Process number 202290/2012-3) granted to S.G. and the National Science Foundation through CAREER Grant No. 0939882 to L.T.A. We also thank anonymous reviewers for helpful comments.
- American Public Health Association, American Water Works Association, and Water Environment Federation (2005) Standard methods for the examination of water and wastewater. American Public Health Association, Washington, DCGoogle Scholar
- Dillalo R, Albertson OE (1961) Volatile acids by direct titration. J Water Pollut Control Fed 3:356–365Google Scholar
- Fernando E, Keshavarz T, Kyazze G (2013) Simultaneous co-metabolic decolourisation of azo dye mixtures and bio-electricity generation under thermophillic (50 °C) and saline conditions by an adapted anaerobic mixed culture in microbial fuel cells. Bioresour Technol 127:1–8. doi:10.1016/j.biortech.2012.09.065 CrossRefPubMedGoogle Scholar
- Franciscon E, Zille A, Fantinatti-Garboggini F, Serrano Silva I, Cavaco-Paulo A, Durrant LR (2009) Microaerophilic-aerobic sequential decolourization/biodegradation of textile azo dyes by a facultative Klebsiella sp. strain VN-31. Process Biochem 44:446–452. doi:10.1016/j.procbio.2008.12.009 CrossRefGoogle Scholar
- Kawasaki S, Watamura Y, Ono M, Watanabe T, Takeda K, Niimura Y (2005) Adaptive responses to oxygen stress in obligatory anaerobes Clostridium acetobutylicum and Clostridium aminovalericum. Appl Environ Microbiol 71(12):8442–8450. doi:10.1128/AEM.71.12.8442-8450.2005 CrossRefPubMedCentralPubMedGoogle Scholar
- Vanýsek P (2012) Electrochemical series. In: Haynes WM (ed) Handbook of chemistry and physics, 93rd edn. Chemical Rubber Company, ClevelandGoogle Scholar