Xian Cao,
Feng Gu,
Hui Wang,
Zhou Fang and
Xian-ning Li*
School of Energy and Environment, Southeast University, Nanjing 210096, China. E-mail: loveveolkswagen@163.com; 504629959@qq.com; lwcq306@163.com; shaka-fz@163.com; lxnseu@163.com; Fax: +86 25 83795618; Tel: +86 137 76650963
First published on 20th March 2017
In this study, biofilm electrode reactors (BERs) were constructed to degrade the azo dye Reactive Brilliant Red (RBR) X-3B. Three different BERs, namely, cathodes with differently structured reactors, cathodes filled with a granular activated carbon (GAC) reactor, and a dimensionally stable anode (DSA) electrode anode reactor, were individually studied to investigate their influence on the removal of both X-3B and chemical oxygen demand (COD). Experimental results showed that it was the internal resistance rather than the total surface area of the cathode that influenced the X-3B removal efficiency in the different cathode reactors. The smaller the internal resistance, the larger the current in the reactor. The larger the current in the whole system, the more electrons that could be utilized for the reduction of X-3B, resulting in higher removal efficiency. The cathode filled with a GAC-BER did not improve the COD removal efficiency, while the X-3B removal efficiency actually declined. Using the DSA electrode as an anode increased the current in the system and improved the removal efficiency. Scanning electron microscopy results showed that the DSA electrode anode had a longer service life than a graphite anode. UV-vis spectral analysis determined that the conjugate system in X-3B was destroyed.
In recent years, bioelectrochemical systems have been explored to reduce refractory substances including nitro aromatics,9 antibiotics,10 heavy metals11,12 and azo dyes13 into biodegradable and less toxic products.4 A moderate electrical current supply on an anode can provide electrons for the cathodic reduction of refractory substances. The energy consumption would be lower and the operational requirements would be much less complex than those for conventional physico-chemical technologies or biological treatment processes,14,15 while the treatment efficiency would be increased at the same time. In recent years, the biofilm electrode reactors (BERs) have been investigated for reduction of nitrate,16 hydrocarbon pollutant,17 chlorinated organic compound,18 perchlorate,19 azo dye20 and antibiotics.21 Previous studies demonstrated that some operational parameters can affect the removal of refractory substances BERs, such as the voltage applied, different reactor materials, and co-substrates. Mu14 et al. reported that the decolorization efficiency of azo dye, AO7, was increased from 70.9% to 98.7% by controlling the cathodic electrode potential in the range of −350 to −550 mV vs. a standard hydrogen electrode (SHE). Kong22 et al. reported that 4-chlorophenol dechlorination could be improved with composite materials rather than carbon-based materials. Sun23 et al. reported that a BER supplied with glucose had higher removal efficiency of Alizarin Yellow R than a BER supplied with acetate. In addition, different electrode materials were used in BERs. Zhou24 et al. reported a 6-times higher maximum biocurrent density was obtained by using a carbon-coated hematite electrode. Zhou25 et al. used a third electrode of graphite in three-dimensional BER which made the 24% higher Fe(III) EDTA reduction efficiency.
However, few studies have reported the removal efficiency could be affected by the use of different cathodes and anodes in BER. In this study, the treatment efficiency of cathodes with different structures, including a cathode filled with granular activated carbon (GAC), and anodes composed of different materials, was thoroughly evaluated in BER. Reactive Brilliant Red (RBR) X-3B was used as a model azo dye because it is often present in dyeing wastewater. The removal of RBR X-3B was investigated under different conditions in BERs. Scanning electron microscopy (SEM) was used to examine the anodes composed of different materials.
Fig. 1 Schematic diagram of the reactor: (1) inlet tank; (2) peristaltic pump; (3) stirrer; (4) the anode; (5) the cathode; (6) the reactor; (7) DC regulated power supply. |
The experiment consisted of three stages. In the first stage, to investigate the impact of the different surface areas and structures of the cathode reactors on X-3B removal efficiency, four types of BER, with two types of different single cathode and two types of different double cathode (inner cathode and outer cathode) were constructed. The structural parameters of these cathodes are listed in Table 1. The cathode was composed of a stainless steel (1 mm thickness) ring and activated carbon fiber (ACF). The ACF (1.0 mm in thickness) was attached to both the inner and the outer surfaces of the stainless steel ring using conductive adhesive (Nanjing Xilite Adhesive Co., Ltd., Nanjing, China). A graphite rod (18 cm in length and 2 cm in diameter) was installed in the center of each type of reactor as an anode, with the cathode surrounding it. A single cathode and anode were connected by titanium wire (1 mm in diameter) to a DC power supply (1 V DC) to create a loop circuit, while the double cathode, with a parallel connection, was connected to the anode and by titanium wire to the DC power supply. 5# BER had the same setup with 1# BER except the cathode and anode were not connected DC power supply. The concentration of X-3B was 100 mg L−1 at the beginning of the experiment, and reached 150 and 200 mg L−1 on days 46 and 60, respectively.
Number | 1# | 2# | 3# | 4# | 5# | ||
---|---|---|---|---|---|---|---|
Inner cathode | Outer cathode | Inner cathode | Outer cathode | ||||
Height (mm) | 90 | 135 | 40 | 40 | 60 | 60 | 90 |
Diameter (mm) | 80 | 80 | 70 | 110 | 70 | 110 | 80 |
Superficial area (m2) | 0.045 | 0.068 | 0.045 | 0.068 | 0.045 |
In the second stage, to investigate the X-3B removal efficiency in a cathode filled with a GAC (3–5 mm in diameter with a specific area of 500–900 m2 g−1) reactor, GAC-BERs were constructed. In these reactors, the cathode was the same as in the 3# BER; GAC was used to fill the space between the inner and outer cathodes and then wrapped with gauze. The double cathode was connected to the graphite rod anode and by titanium wire to a DC power supply (1 V DC). The concentration of X-3B was 100 mg L−1 at the beginning of the experiment, and reached 150 and 200 mg L−1 on days 46 and 60, respectively.
In the third stage, to investigate the influence of the X-3B removal efficiency in different anode material reactors, Ti/IrO2–Ta2O5-coated dimensionally stable anode (DSA; Baoji Ruicheng Co., Ltd., China; 18 cm in length and 2 cm in diameter) BERs were built. The structure of a DSA-BER was the same as that of the 3# BER, except for the anode. The concentration of X-3B was 200 mg L−1.
E = (A1 − A2)/A1 × 100% | (1) |
The influent chemical oxygen demand (COD) concentrations (C1) and effluent COD concentrations (C2) were measured according to standard methods. The COD removal efficiency (C) was calculated as shown in eqn (2)
C = (C1 − C2)/C1 × 100% | (2) |
All samples were filtered through a 0.45 lm syringe filter to remove suspended solids from the liquid media prior to the measurements.1,26
The internal resistance was calculated by the linear region of polarization curve.27
For scanning electron microscopy (SEM) observation, the electrode was immobilized in a solution of glutaraldehyde for 24 h at 4 °C, then processed through an ethanol dehydration series (i.e., 30, 50, 70, 85, 95 and 100%, v/v, ethanol, 0.5 h each treatment), dried with the critical point drying method and evaluated using scanning electron microscopy (HIMADZU SSX-550, Shimadzu, Japan).20
Fig. 2B shows that the average X-3B removal efficiency in the four different reactors was relatively stable when the concentration of X-3B was 200 mg L−1. It can be seen that the average concentrations of X-3B in the four different reactors were 84.90, 73.37, 66.42, and 100.72 mg L−1, respectively. The average X-3B removal efficiencies were 57.55%, 63.31%, 66.79%, and 49.64%, respectively. Liu20 et al. constructed a BER and found that the decolorization efficiency increased from 39.30 to 76.23% when the initial dye concentration decreased from 200 to 25 mg L−1. Zhang28 et al. also constructed a BER which got a 35.8% higher X-3B decolorization efficiency than the control group. Rahmani29 et al. constructed another BER with C.I. Acid Red 18 decolorization efficiency of 73.5% when the optimized current density was 8.6 mA cm−2. It can be concluded that the BER constructed in this study achieved the similar level of decolorization efficiency. Meanwhile, it was found that the average X-3B removal efficiency did not change with the total surface area of the cathode in the four reactors with different cathode structures. It is generally believed that the removal efficiency increases as the total surface area of a cathode increases; however, there was no direct link between the removal efficiency and the total surface area of the cathode in this experiment.
The currents in the four different reactors were measured over time. The average current densities were 3.93, 5.67, 9.16, and 1.54 mA m−2, respectively. As shown in Fig. 2C, there was a positive correlation (R2 = 0.9197) between the average X-3B removal efficiency and the average current density, with the removal efficiency rising as the current in the reactor increased. The internal resistances of the four different reactors were 839.56, 437.33, 381.90, and 1035.63 Ω, respectively, with these values obtained via polarization curves. It can be seen that the removal efficiency decreased with increasing internal resistance. In bioelectrochemical systems, the internal resistance composes by ohm internal resistance, mass transfer internal resistance and activation polarization internal resistance. Ohm resistance mainly comes from the electrode materials; mass transfer resistance mainly comes from transmission of reactants and products to the electrode surface and internal resistance due to activation polarization mainly comes from the activation energy which the electrochemical reactions require in cathode and anode. In our experiment, because the resistance of cathode and anode materials were small, ohm internal resistance was not the major influence factor which led to different internal resistance. Meanwhile, Wang30 et al. reported that higher current always accompanied by lower mass transfer resistance in microbial electrolysis cell. The current (less than 0.1 mA) in our study was low, so the mass transfer resistance was relatively high which had a relatively large impact on internal resistance. Moreover, Elmekawy31 et al. reported the internal resistance due to activation polarization could be change by changing the projected surface area of anodic and cathodic electrodes. The projected cathodic surface areas of the four reactors in our study were different, so were the activation polarization internal resistances. Thus, the different on mass transfer and the polarization resistance between four kinds of cathode construction led to different internal resistance. The removal of X-3B was a process that required electrons to be gained to reduce the action of microorganisms under anaerobic conditions. The increase in current density in the whole system meant that more electrons could be utilized for the reduction of X-3B, and the X-3B removal efficiency then increased. Therefore, the main factor influencing the X-3B removal efficiency in the reactors with different cathode structures was the different internal resistances caused by the different cathode constructions.
Fig. 3B shows the average COD and X-3B removal efficiencies in BER and GAC-BER when the concentration of X-3B was 200 mg L−1. The COD removal efficiencies were 78.60% and 72.37%, and the X-3B removal efficiencies were 64.68% and 46.93%. The cathode filled with GAC-BER did not improve the COD removal efficiency, while the X-3B removal efficiency actually declined. When the cathode was filled with electrode material GAC, it was easy for short circuits in the current to reduce the current efficiency in the reactor, which led to fewer electrons being available for the reduction of X-3B and decreased the removal efficiencies as a consequence. The average maximum current densities in the two reactors were 9.16 and 1.02 mA m−2. It was clear that the current in the BER was higher than in the GAC-BER, and the removal efficiencies were also better.
Fig. 4 (A) The average COD and X-3B removal efficiencies in reactors with anodes composed of different materials; (B) SEM images of two different anodes before and after 120 days. |
The components of the influent were analyzed, with the strong oxidizing substances formed at the anode including ClO− and OH˙. The Cl− in NaCl and NH4Cl were first converted to Cl2 and then to ClO−. The standard oxidation potential for the conversion of Cl− to Cl2 was 1.358 V (vs. SHE), and the standard oxidation potential for the conversion of H2O to OH˙ was 2.8 V (vs. SHE). In this experiment, the DC power supply (1 V). The anode potential was 0.27 V measured by calomel electrode, and was 0.51 V when converted to hydrogen electrode reference (vs. SHE) which was much lower than the oxidation potential for ClO− and OH˙. Therefore, the strong oxidizing substances ClO− and OH˙ would not be formed.
The internal resistances of the two different anode reactors were 318.9 and 142.46 Ω, respectively, with the values obtained via polarization curves. The current in the BER was 0.412 mA, while in the DSA-BER it was 0.715 mA. The lower internal resistance of the DSA-BER resulted in a higher current density for the same DC power supply, which we considered to be the reason that the removal efficiencies were improved.
Fig. 4B shows SEM images of the two different anodes. Fig. 4B(a) and (b) are characterizations of the DSA anode before and after 120 days, and Fig. 4B(c) and (d) are characterizations of the graphite anode. By comparing (a) and (b) with (c) and (d), it was found that the surface of the DSA anode changed little, and was still smooth with good stability. We speculate that this may result in the DSA anode having a longer service life. Marshall et al.34 also reported a similar phenomenon. In contrast, the surface of the graphite anode was loose, broken, and uneven. Owing to long-term soaking, solution had penetrated into the center of the graphite rod, which may have been the cause of the increase in internal resistance.
Fig. 5 The UV-vis spectrum of (A) the influent (X-3B 200 mg L−1); (B) the influent and the effluent from the BER, GAC-BER, and DSA-BER. |
Fig. 5B shows the UV-vis spectrum of the effluent from the BER, GAC-BER, and DSA-BER as well as the influent. As shown in this figure, the conjugate system in the effluent of the three types of reactors was destroyed. There were no new absorbance features in the visible region (380–780 nm). The extent of the removal of the conjugate system in the DSA-BER and BER was better than in the GAC-BER, which was explained by the DSA-BER and BER having higher X-3B removal efficiency. The four absorbance features (198, 235, 285, and 332 nm) all disappeared at the same time. However, there were some new absorbance features between 240 and 265 nm, which implies that the azo dye degraded into lower-molecular-weight products. A continual treatment process is needed to degrade these breakdown substances.
This journal is © The Royal Society of Chemistry 2017 |