Graphite fiber fabric (GFF), a material consisting of fibers about 5–10 μm in diameter, was used for electrode material in this study. The carbon nanotube was deposited on the surface of GFF via Electrophoresis Deposition (EPD) method for enhancing its conductivity. The EPD solution was prepared with CNT (1 g/L), Polyethylenimine (PEI) and nickel dichloride in deionized water. The concentration of PEI & Ni were varied with1000 & 500 mg/L, 500 & 500 mg/L, 500 & 250 mg/L, 250 & 250, which were referred to set A, B, C, D, respectively [21]. In order to make the CNT particles positively charged, the pH of EPD solutions were judged in the range of 9.3-10.2 by using 1 mol/L of sodium hydrate [22][23]. In the EPD system, the stainless steel mesh was used as the anode, while the GFF was used as the cathode (deposition electrode) with 5 mm distance. A direct current voltage of 30 V was applied between the anode and cathode using a DC power supplier for 5mins [23][24]. Then, the deposited GFF was dried at 120℃, repeat three or four times for the EPD process. The EPD-prepared GFF was used as the bio-cathode, and the anode was further modified by screen-printing with expanded graphite (EG) and CNT on the surface of EPD-prepared GEE to increase its compatibility

A single-chamber BEAD reactor was fabricated with acrylic tube, with a total volume of 18L and a working volume of 12L (Figure 1). On the cover plate, a motor was settled for mixing sludge in the reactor, a gas sampling hole was be installed for biogas analysis, and an air sealed hole was for feeding the sewage sludge. A water displacement gas collector, fill with acidified saturated aqueous NaCl, was connected to the cover plate with a rub tube for the biogas collection. 6 sets of anode (24 × 6 cm) and cathode (24 × 6 cm) separated by non-woven PP sheet were placed in the BEAD reactor (Figure 1). The reactor was settled in a water tank with a heater and temperature sensor to maintain the reaction temperature at 35 ± 2℃. Seed sludge was collected in an anaerobic digester of S municipal wastewater treatment plant, while the feeding sludge was collected in sludge settling tank of Y wastewater treatment plant in Busan, Korea. The Characteristics of feed sludge and seed sludge are shown in Table 1. For the BEAD reactor, 30% of seed sludge and 70% of feed sludge was injected in the reactor for the start-up, and operated by adding the feed sludge of 600 mL once a day with a HRT of 20 days. The potential difference between the anode and cathode was maintained at 0.3 V by using a power supplier [25].

Figure 1: 
Schematic diagram and operation of a bioelectrochemical anaerobic digester and electrode assembly

The morphology of the surface of the carbon based-electrode was observed with a scanning electron microscopy (SEM, JSM-6700, JEOL) with energy dispersive spectrometer (EDS) equipment. The resistance was measured by 4-point method, the electrochemical feature of GFF was measured by cyclic voltammetry (CV) in 0.1 N sulphuric acid solution with a 0.1 V/s of scan rate. The generated biogas was measured periodically by the biogas collector with graduated cylinder [26]. The composition of biogas were analyzed by a gas chromatograph (GowMac Series 580, Porpak Q 6ft × 1/8 "SS) equipped with a thermal conductivity detector (TCD). The analyzed methane production was calculated to standard temperature and pressure (STP) status by using Equation (1) [2].

where, V_CH4 (at T) is the measured methane production at room temperate, T is the room temperature, W is the water vapor pressure at Temperature T. The concentrations of TS, VS, TCOD, SCOD and Alkalinity were determined according to Standard Methods. The pH were immediately measured after sampling with pH meter (Orion 370). Total VFAs were determined according to a titration method [27], and the compositions of VFAs were analyzed by HPLC with Aminex HPX-87H column (300 × 7.8mm, BIO-RAD) and 0.005M (or 0.008M) H₂SO₄ as the mobile phase. The potential of anode and cathode were measured by multimeter with an Ag/AgCl reference electrode. Current between anode and cathode was monitored by digital multimeter (DMM), and the volumetric current density (A/m³) was calculated by the monitored current over a 24 h period and the working volume. Energy recovery was calculated from the Gibbs free energy of methane (W_CH4 = n _CH4 × G _CH4), where G _CH4 = -890.4 kJ/mol (oxidation by oxygen to water). The energy efficiency relative to electrical input (%) was the ratio of energy content produced (methane) to the substrate energy (-872.9 kJ/mol for acetic acid) added. The overall energy recovery based on both electric and substrate input (%) took into account both electric and substrate energy inputs [1][11].

Figure 2 (a) shows the SEM images of the GFF (set B) treated with EPD method at different magnification times. The SEM images of GFF treated with EPD method show comparably good deposition of CNT. Figure 2 (b) shows the images of the GFF after EPD treatment with different concentration of PEI-Ni, which is clear that a large quantity of CNT was deposited on GFF averagely with a 30V of applied voltage at constant deposition time from set A to set D. Set A-D are varied by the concentrations of PEI and Ni, which indicates that PEI-Ni concentrations didn’t effect much for the deposition quality on the surface of GFF. Qualitative analysis were carried out using an SEM equipped with an EDS shown in Figure 3. The untreated GFF has a carbon content as 93%, the Content of MWCNT in composite coatings was about 50 wt% (70 At%), Nickel was 30 wt% (10 At%), Oxygen was 15 wt% (15 At%) and less iron was measured from sea A to set D. This indicates CNT, nickel, oxygen exist on the surface of the GFF, and the element composition has a small change at different concentrations of PEI-Ni.

Figure 2: 
SEM images of (a) the GFF treated with EPD method at different magnification times and (b) the GFF treated with different concentration of PEI-Ni.

Figure 3: 
EDS data for GFF before and after EPD method with different concentrations of PEI-Ni, (a) Percentage by weight, (b) Percentage by atomic number

The resistance of initial GFF is 0.12 Ω/cm² measured by 4-point method (Figure 4), but the resistance of treated GFF has a significant decrease, especially set B was reduced to 0.0006 Ω/cm², which was almost 20 times less than the untreated GFF. This indicates that CNT and nickel was well deposited on the surface of GFF and the conductivity of GFF was significantly enhanced. The electrochemical feature of GFF was measured by CV in 0.1 N sulphuric acid [28]. Figure 5 shows the treated GFF have a lager current peak than the untreated GFF, which means the redox reaction of treated GFF was be enhanced. Especially, set B has a more lager range of reduction peak, which is much higher than others. This result indicates that the set B, which is the GFF deposited with CNT and nickel via EPD method at a appreciate concentration has a great potential for using as a good bio-electrode in the BEAD process.

Figure 4: 
Electrical resistance for GFF before and after EPD method as a function of PEI-Ni

Figure 5: 
Cyclic voltammetry for GFF before and after EPD method as a function of PEI/Ni

In order to investigate the stability of BEAD reactor operation, the pH and alkalinity of effluent was measured once a day. The pH of influent sludge has small changes between 5.8 ~ 6.2 shown in Figure 6 (a), and the pH of effluent sludge of BEAD was unstable in start-up, which was in the range of 6.8-7.6. However, the effluent pH was turned to a stable state of 7.2 ~ 7.5 after operating for 40 days, which is an applicative pH range for methanogenesis [29]. The alkalinity also showed a stable range of 4,500 ~ 5,200 (mg/L as CaCO₃) after start-up in Figure 6 (b). It indicateds that the stable range of pH was due to the appreciate alkalinity, which has a function of buffer solution. Furthermore, the concentration of VFA was measured less than 250mg/L (as HAc), which means almost all of VFA produced from hydrolysis was converted to methane gas without accumulation.

Figure 6: 
(a) pH and (b) alkalinity changes of influent and effluent sludge in BEAD reactor

The potential of electrode can affect the microbial growth and redox reaction of BEAD, the amount of energy of the electrons is expressed in the anode and cathode potential and it’s determined by thermodynamics of the anode and cathode reaction. To calculate these potentials, the Gibb’s free energy (kJ/mol) of the reaction needs to be determined. The anodic oxidation reaction for acetate is -486mV (mV vs Ag/AgCl) which means oxidation reaction will appear when anode potential is higher than -486mV, and the cathode potential should be less than – 445 mV (mV vs Ag/AgCl) to lead the cathodic reduction reaction for carbon dioxide to methane [1][25]. In the BEAD reactor, anode potential and cathode potential were -200 ± 24mV and -503 ± 18mV, respectively, which indicates that the potential of anode and cathode were perfect theoretical agreement with experiment during operation. This is implies that the BEAD process has a very strong stability for oxidizing organic matters and producing methane [30][31].

The biogas production and composition during the continuous operation of BEAD reactor is shown in Figure 7. Predictably, biogas production and methane content were very low amount up to 40 days, and there was a small amount of hydrogen was produced (Figure 7 (a)). It indicates that the electro-active bacteria was not well established on the bioelectrode at the start-up period. However, after the operation of BEAD reactor changed to stable, the methane production rate was stabilized at 412ml/L.d, which is much higher than a conventional AD process [2]. The content of methane was stabilized at 77.3%, which is higher than 50-60% of the conventional AD (Figure 7 (b)). As well known, methane was produced by methanogenic bacteria from acetic acid in conventional AD. However, in the BEAD, the methane is believed to be produced on the surface of cathode by reducing of carbon dioxide, as well as the direct interspecies electron transfer (DIET) between exoelectrogenic fermentation bacteria and electrotrophic methanogenic archaea in the bulk solution [18]. In this study, it was believed that modified GFF electrode has a good ability to reduce carbon dioxide and activate DIET for methane production, which related to a high methane content in biogas.

Figure 7: 
(a) Biogas production and (b) composition in the BEAD recator

The VS removal efficiency was stabilized at 72.5% (65% for COD) at the stable state, which was significantly higher than 43.5% of conventional AD at the same conditions [2]. The high VS reduction rate means high hydrolysis rate in BEAD reactor. In the conventional AD, the rapid hydrolysis step generally cases unstable digester operation due to the VFA accumulation. In this study, the methane production could be significantly reduced when stopping the applied voltage in the BEAD reactor. It indicates that the electrode reactions was shocked, the anode will stop to oxidize the organic matters, and the cathode won’t utilize the proton and carbon dioxide fast enough to produce methane, which caused an accumulation of VFA in the reactor and a possible reduction in pH. However, after applying voltage to the BEAD reactor, methane production was increased rapidly, which indicates that the BES system could assist AD process to reach an equilibrium state of the fermentation reaction steps [32].

The potential difference was controlled at 0.3 V between anode and cathode. The energy efficiency of BEAD related to the electrical energy input was 296%. By taking into account the both electric and substrate energy inputs, the energy efficiency became 163%. The energy efficiencies obtained in this study are yet comparable with those achieved in other studies (Table 2). This is largely due to the fact that potential of the cathode at a highly reduction potential allowed achieving a very high coulombic efficiency (89%) relative to proton reduction. On the other hand, the current density was calculated with measured current by DMM is 2.265 A/m³ in the present study was higher than those typically reported in the literature. This could be due to the fact that lots electrochemical microbial inoculated to the anode. Main advantage resulting from the 0.3V potential difference of operation is that the anode potential was helping the rate of substrate oxidation.