The aim of this study was to investigate the different performance of bioelectricity generation and wastewater treatment between constructed wetland (CW) respectively coupled with air-cathode microbial fuel cell (ACMFC) and microbial fuel cell (MFC) under a fed-batch mode. During a 75-day-operation, the voltage of CW-ACMFC and CW-MFC ranged from 0.36 to 0.52 V and from −0.04 to 0.07 V, indicating that the bioenergy output of CW-ACMFC was significantly higher than that of CW-MFC system. In addition, the maximum of power density of CW-ACMFC and CW-MFC was 4.21 and 0.005 mW m−2. Notably, the chemical oxygen demand (COD) and NH3-N removal efficiency of CW-ACMFC was slightly higher than that in CW-MFC, which resulted from a higher voltage accelerating the transport of electron donors and the growth of microorganisms and plants. This study possesses a probability of using ACMFC coupled with CW to enhance the pollutant removal performance in CW system.
The domestic sewage treatment in rural areas of developing countries is an increasing pollution burden due to the lack of adequate sewage treatment facilities. Pollutants, including ammonia nitrogen, organic matter and heavy metals, may affect significantly the self-purification capacities of water bodies. This is the primary reason for the higher rates of water eutrophication in these areas of developing countries. Typically, the cost of electricity account for 5% to 30% of total operating costs in municipal wastewater treatment (Chae & Kang 2013). Hence, the low-cost operation and maintenance, sustainable and energy-efficiency constructed wetland (CW) has been considered as a worldwide technology for domestic sewage and industries treating in small communities. Recently, the bioenergy recovery from organic matter in the process of wastewater treatment has been got focused by integrating microbial fuel cell (MFC) technology into CW.
In anode of MFC, organic matter is degraded by microorganisms especially electrochemically active microorganisms (EAM) to generate electrons (e−) and protons (H+). Protons shuttle through proton exchange membrane (PEM) into the cathode chamber, and electrons under the effect of potential gradient transfer from the anode electrode to the cathode, where they combine with protons and oxygen to produce water (Yadav et al. 2012). However, the application of MFC usually showed a low power density, current density and coulombic efficiency, especially under fed-batch mode (Liu et al. 2013). To ameliorate MFC performance, air-cathode microbial fuel cell (ACMFC) has been frequently observed in organic matter degradation with a higher bioelectricity generation (Cheng et al. 2011; Campo et al. 2014; Kim et al. 2014).
The variation of dissolved oxygen (DO) and oxidation reduction potential (ORP) in CW provide a natural habitat for the growth of microorganisms with different metabolic functions (Ciria et al. 2005; Ojeda et al. 2008). In this regard, CW can be designed, by ameliorating operational parameters, to act as MFC for wastewater treatment and bioelectricity generation simultaneously (Yadav et al. 2012). These considerations make it more excitable that CW-MFC has been developed for treating various types of wastewater under different operational conditions (Liu et al. 2013; Zhao et al. 2013). By integrating MFC into CW (Zhao et al. 2013), the CW were in suit to observe with a peak power density of 12.83 μW m−2 and 71.5% removal performance of chemical oxygen demand (COD). In a CW-MFC developed by Liu et al. (2013), the power density of 12.42 mW m−2 was observed with Ipomoea aquatic. In addition, in an ACMFC designed by Cheng et al. (2011), the maximum power density of 1,070 mW m−2 was obtained. Hence, the application of ACMFC into CW might be useful for the enhancement of bioelectricity generation.
In this study, a comparative study on the bioelectricity generation performance of CW-ACMFC and CW-MFC was operated under a fed-batch mode. To assess the efficiency of two kinds of reactors, the maximum power density, COD and ammonia nitrogen removal performance were measured in each cycle.
MATERIAL AND METHODS
System construction, inoculation and operation
Two polyvinyl chloride tanks (0.6 L × 0.4 W × 0.4 H m) filled with quartz sand (φ 4–6 mm; porosity 37%) with an average thickness of 0.35 m were used to simulate the laboratory-scale CW-MFC. The containers were planted with disease-free Cannas indica of initial stem lengths 25 ± 3 cm. Cannas were transplanted into every wetland under a density of 17 rhizomes per square meter and propagated for 30 days' acclimation.
As schematically shown in Figure 1, an anode electrode (graphite felt, 0.6 L × 0.36 W × 0.003 H m) was placed 0.15 m near to the bottom. The larger surface of the electrode could provide a habitat for the growth of microorganisms that will benefit the process of electron production of EAM. A similar graphite felt (dimensions of 0.6 L × 0.08 W × 0.003 H m) was placed 0.05 m near to the top of the microcosm to form a good aerobic condition. The difference between CW-ACMFC (System 1) and CW-MFC (System 2) was that the cathode electrode of CW-ACMFC was placed at the top of influent level, while cathode chamber of CW-MFC was filled with wastewater during the influent process. The total volume of System 1 and 2 was 84 L with a liquid volume of 28.8 L and 33.6 L, respectively. The anode to cathode electrode was connected through an insulated electric wire (copper) and an external electrical resistance of 2,000 Ω. Each microcosm was inoculated with equal volume (15 L) activated sludge, which was collected from Songjiang sewage treatment plant, Shanghai, China.
The main composition of synthetic wastewater is listed as follows (in mg L−1): glucose·H2O (112.33), NH4Cl (48.59), KH2PO4 (7.69), Na2HPO4·12H2O (17.73), MgCl2·6H2O (17.67), ZnCl2 (7.54), CaCl2 (17.64), CH3COONa (100.20). This study was performed for a total of 75 days (from September 28 to December 12, 2014), which consisted of 30 days' acclimation for the growth of biofilm and wetland plants. During the experiment, both reactors were worked under a fed-batch mode with the hydraulic retention time (HRT) of 24 h. The collection of influents and effluents were manipulated at 16:00 every day.
Water quality analysis
The COD solution of water sample was digested by a HH-6 COD meter and measured by HACH DR900 Spectrophotometer (USA). The concentration of NH3-N was analyzed by HANNA HI 93733 instrument (Italy). Water temperature, pH and DO were calculated by a multi-parameter meter (HQ40d, HACH, USA). ORP was measured by an HI 9135, HANNA equipment.
Bioenergy output determination
The voltage between the edge of external electrical resistance of two reactors were registered every 5 min by an automatic recorder intelligent instrument. To obtain the polarization curve of Systems 1 and 2, the current and voltage were recorded by a digital multimeter (VICTOR 86E, Shenzhen) through the variation of external electrical resistance from 40,000 to 5 Ω every 3 min on Day 65. The current and power density were calculated by the following formula: J = U/SRex, P = UI/S, where the J (mA m−2), U (V), Rex (Ω), S (m2), P (mW m−2) and I (A) is current density, cell voltage, external electrical resistance, cathode surface, mean power density and current, respectively.
DO, COD, and ammonia nitrogen removal by the CW-ACMFC and CW-MFC reactor under similar operational condition was statistically analyzed with the one-way analysis of variance (ANOVA) method in SPSS 22.0 software, and p < 0.05 were considered as the signiﬁcant level.
RESULTS AND DISCUSSION
DO and ORP profile
DO and ORP of the effluent samples are presented in Figure 2(a). In general, ORP values of Systems 1 and 2 were obviously affected by DO level. In System 1, the values of effluent ORP and DO were −170 mV and 1.2 mg L−1, respectively, while effluent ORP and DO were obtained with value of −180 mV and 1.1 mg L−1 in System 2. A previous study in CW found similar results, indicating that DO concentration of the effluent at CW are usually maintained at a low level (García et al. 2010). Notably, the concentration of these parameters showed a relatively obvious fluctuation at the middle and end of the experiment. It was well recognized that the drop of average air temperature (from 19 to 14 °C in Days 55–60, and from 16 to 5 °C in Days 60–75, respectively) and might result in the instability of DO and ORP of both reactors. In addition, average pH of Systems 1 and 2 were maintained at 6.79 and 6.85, respectively.
COD and NH3-N removal performance
As shown in Figure 2(b), the average COD concentration of influents were 198.4 mg L−1, and the average COD concentration of effluent samples collected from Systems 1 and 2 were 59.8 and 67.2 mg L−1, respectively. The average COD removal efficiencies of Systems 1 and 2 were respectively 69.9 and 66.0%. In addition, the maximum efficiency of COD was found to be 79.13% in System 1, indicating that an anode of MFC integrated into the CW was beneficial for the COD removal (Fang et al. 2013). From a previous study, a higher voltage output might stimulate the growth of microorganisms and plants (the synthesis of chlorophyll), which resulted in a better COD removal performance in System 1 (Song et al. 2011). In addition, the present data also revealed that the COD removal was limited to the fluctuation of air temperature, the first drop of air temperature (from 19 to 14 °C, at Days 55–60) was more obvious than that of the second drop (from 16 to 5 °C, at Days 60–75) during the experiments. A decrease of air temperature could reduce the process of oxidation reaction in the anode chamber of the reactor.
Figure 2(c) shows the performance of NH3-N removal efficiency in these reactors. Average NH3-N concentration of influents was 14.9 ± 0.2 mg L−1, and average NH3-N concentration of effluent sampled collected from Systems 1 and 2 were found to be 9.5 ± 0.5 mg L−1 and 10.0 ± 0.3 mg L−1, respectively. This suggested the NH3-N removal efficiency of CW-ACMFC and CW-MFC was 36.2% and 32.8%, respectively. Notably, the ammonia nitrogen removal efficiency in both reactors was not high.
The mechanisms of nitrogen removal in CW were plant adsorption, ammonia-volatilizing and simultaneous nitrification–denitrification (Bigambo & Mayo 2005; Song et al. 2011). A generally accepted cause of low efficiency wastewater treatment was the DO concentration of wetland under a lower level during the biodegradation of pollutants (Horne 2002), which showed a significant influence on the efficiency of nitrification. The DO concentration of these reactors were maintained at a relatively lower concentration. In addition, the presence of electrons production under higher cell voltage of CW-ACMFC might promote removal of NO3-N or NO2-N. However, Song et al. (2011) indicated that a large sum of NH3-N could be transformed into NO3–N in the circumstance of DO exceeding 1.5 mg L−1 in a horizontal subsurface-flow (HSSF) CW for high concentration ammonia nitrogen wastewater treatment. With a DO concentration lower than 1.5 mg L−1, the efficiency of nitrification has been inhibited. Hence, the average NH3-N removal efficiency of CW-ACMFC and CW-MFC was observed with a value below than 35%. However, some studies reported that the integrating micro-electric field (MEF) into CW could stimulate the production of chlorophyll, the permeability of cells membrane and transporting capacity of plants (Song et al. 2011). A higher voltage found in CW-ACMFC system might be beneficial for microorganism and plant growth. Therefore, the average ammonia nitrogen removal efficiency of System 1 was higher than that in System 2.
Bioelectricity generation performance
After stabilization, the cell voltage of Systems 1 and 2 is presented in Figure 3(a) and 3(b) with 45 cycles. As described in the section ‘System construction, inoculation and operation’, two reactors were carried out in a fed-bath mode, the feeding time of wastewater was performed at 16:00 every day, so the starting point of each cycle of cell voltage was plotted at same time. The cell voltage of Systems 1 and 2 were immediately rising at feeding time, which was largely driven by wetland MFC were fed with new substrate. After the feeding time, stable voltage of both reactors was obtained, the voltage of CW-ACMFC varied in the ranges of 0.36 to 0.52 V, while the voltage of CW-MFC changed from −0.04 to 0.07 V. For example, on Day 56, stable voltage of System 1 was approximately constant with a value around 0.5 V, whereas the voltage of System 2 was approximately constant values of 0.06 V after the replacement of substrate. The constant voltage produced by System 1 was 8.3 times higher than that of System 2, indicating a significant difference on the bio-electricity generation between the in two reactors. When the wastewater dropped from wetland MFC, cell voltage of Systems 1 and 2 was also descended sharply to about 0.0 V, resulted from the lack of fuel substrate in the reactors.
The air diffusion could enhance oxygen concentration in the cathode of wetland MFC, where the electrons produced by EAM combine with protons and oxygen to form water: 4e− + 4H+ + O2 = 2H2O. In addition, plants also could promote the oxygen concentration in the cathode through their photosynthesis that had been indicated in the experiment designed by Fang et al. (2013). The average voltage output of CW-MFC with plants was about 15% higher than that of unplanted CW-MFC. Furthermore, the continuous mode and high substrate concentration are beneficial to the amplitude of bioenergy output in CW-MFC (Zhao et al. 2013; Liu et al. 2014). The CW-MFC designed by Zhao et al. (2013) produced its peak voltage of 58 mV with initial average COD concentrations of 1,058 ± 421 mg L−1 under the sequencing batch mode. By integrating CW into MFC, Villaseñor et al. (2013) observed that an average cell voltage of CW-MFC was approximately maintained at 240 mV, at COD concentrations of 250 mg L−1 under a continuous mode. Therefore, a higher efficiency of recycling the bioenergy contained in organic matter was found with CW-ACMFC by changing the configuration of CW during wastewater treatment.
In addition, the cell voltage performance of System 2 was unstable, especially in the night cycle, which resulted from root exudates of plants consuming parts of oxygen around the rhizosphere. According to previous studies, there are numerous factors, such as air temperature, photosynthesis, affecting on macrophytes releasing low-molecular weight organic compounds into the soil or filler according to previous studies (Vancura & Hanzliko 1972; Carvalhais et al. 2011). The release of dissolved organic carbon (DOC) from P. australis was 9.0 ± 0.9 μg g−1 root dry mass h−1 (Zhai et al. 2013). Also, organic carbon released by roots would cut down because of the descendant efficient photosynthesis during the night cycle (Jonasson et al. 1999). However, because of negative value of voltage were obtained for several days, the release of DOC from plants might have had a deep influence on the bioelectricity generation in CW-MFC.
Cell voltage of Systems 1 and 2 had dropped from 0.81 to 0.64 V and 0.74 to 0.35 V when the external resistance was reduced from 40 to 6 kΩ, whereas the cell voltages of Systems 1 and 2 had dropped from 0.50 to 0 V and 0.01 to 0 V when the external resistance was reduced from 2,000 to 5 Ω, respectively (Figure 4). The maximum power density and current density of System 1 was obtained to be 4.21 mW m−2 and 20.30 mA m−2, while the maximum power density and current density of System 2 was obtained to be 0.005 mW m−2 and 1.47 mA m−2. This suggested maximum power outputs of Systems 1 and 2 were 362.93 and 0.43 J m−2 d−1, respectively, and the maximum power density of System 1 was 842 times higher than that of System 2. It was observed that the power density of Systems 1 and 2 increased to a peak value with an external electrical resistance of 1,000 and 4,000 Ω. This indicated that CW-MFC with air-cathode was conducive to the loss of internal resistance in MFC. Similar results were found in a previous study with the air-cathode MFC having 18% less ohmic resistance than the two-chamber reactor (Cheng et al. 2011).
In this study, the performance of bioelectricity generation in CW-ACMFC and CW-MFC was compared under an operational mode of fed-batch. In general, the bioenergy output in CW-ACMFC was significantly higher than that of CW-MFC. The peak power density (4.21 mW m−2) was observed in CW-ACMFC. In addition, the removal efficiency of COD and NO3-N in CW-ACMFC was slightly higher than that of CW-MFC, which might be resulted from the growth and metabolism of microorganisms and plant stimulated by a higher voltage produced in former reactor. In this study, it was demonstrated that the CW-ACMFC system was an advantageous technology for the harvest of bioelectricity from organic matter during wastewater treatment.
This study was supported by the Chinese National Natural Science Foundation (91547209, 51679041 and 41571037).