Performance of air-cathode stacked microbial fuel cells systems for wastewater treatment and electricity production

Nowadays, wastewater represents a source of recyclable water, despite the fact that it contains many contaminants that may affect public health and the environment. Wastewater contains large amounts of renewable energy in the form of chemical bonds. Domestic wastewater could potentially generate up to 2.2 kW/h.m³ (assuming 500 mg/L chemical oxygen demand (COD)) of energy, which corresponds to 29.3 terawatt-hours or 0.10 quadrillion kJ. If this energy is recovered correctly, it can generate electricity from wastewater (Virdis et al. 2011). Currently, microbial fuel cells (MFCs) can be utilized as a decentralized wastewater treatment, having the advantage of producing bio-energy (electricity, methane, and hydrogen) from wastewater (Logan 2012). Several studies have been carried out in single MFC units (smaller scale) to generate electricity. MFCs could produce up to 15 W/m³ using domestic wastewater (Logan 2008). However, there are only few examples of MFCs with multiple anodes and cathodes. The scaling-up of MFCs requires an intensification process, in order to reduce the size of the reactor, generate high power and COD removal, and reduce the ohmic losses, thus minimizing the voltage reversal in the reactors (Oh & Logan 2007; Kim et al. 2012). Multi-electrode MFCs are the best option so far to troubleshoot some of these problems (Jiang et al. 2010; Ahn & Logan 2012). The theoretical voltage of an individual MFC in open circuit voltage (OCV) mode is ∼1.25 V, using glucose as electron donor and oxygen as the electron acceptor, according to the Nerst equation. However, in practice, the voltages produced in a single-chamber MFC are ∼0.2 to 0.5 V (Oh & Logan 2007; Kim et al. 2012). In order to increase the voltages from a single unit cell, MFCs had to be connected in series. Xinmin et al. (2016), An et al. (2015a, 2015b), Yazdi et al. (2015), An et al. (2014), Ieropoulos et al. (2013), Kim et al. (2013), Rahimnejad et al. (2012), Zhuang et al. (2012), Kim et al. (2012), Jiang et al. (2010), Gálvez et al. (2009), Zhuang & Zhou (2009), Ieropoulos et al. (2008), Shimoyama et al. (2008), Aelterman et al. (2006), and Shin et al. (2006) showed a successful use of series stacked MFCs with voltages ranging from 2 V to 23 V, at OCV. However, these studies were limited to 2–6 individual MFC units connected in series, and a few stacked MFCs were operated under continuous flow. Previous studies have demonstrated that during the electricity production, in stacked MFC systems, a decreased power density has been observed (Gurung & Oh 2012). Furthermore, the scalability of MFCs is fundamental for real-world application, not only for increasing the electricity production but also (in terms of the treatment capacity) for removing the contaminants. In order to make a scaled-up MFC that is practical and sustainable, it is necessary to carry out research focused on the terms of installation of multiple anodes and cathodes (stacked MFC configurations), flow distribution, and the effect of the operational variable hydraulic retention time (HRT) on the performance of stacked MFCs. The purpose of this is to increase the capacity of organic utilization and to reduce the cathode limitations and the ionic cross-conduction. The effects of HRT on voltage profiles, organic matter, and nutrient removal efficiencies were not studied in stacked MFCs. Therefore, the optimal selection of HRT in a stacked MFC is an important factor when designing a larger-scale stacked MFC to remove contaminants and to simultaneously generate electricity. The present study has been carried out to evaluate the performance of two new air-cathode stacked MFC configurations (20 and 40 single-chamber MFCs) operated under continuous flow at OCV and closed circuit voltage (CCV), during wastewater treatment and electricity production, at three different HRTs.

Stacked MFC 2 contained 40 units of air-cathode MFC in a shared reactor. Stacked MFC 2 module was formed by four chambers: each chamber of 70 cm height, 10 cm length, and 5.7 cm width (∼4 L volume) containing 10 MFC units without separator (Figure 1(b)). An individual MFC unit was composed of three carbon felts (5 cm length, 5 cm height, and 0.65 cm thickness) and was used as anode (surface area of 0.0075 m²). The three anodes were connected externally by a single copper wire. An MEA (5 cm length, 5 cm width) was exposed to air and was utilized to separate the anodic chamber from the cathode. The distance between anode and cathode was 2 cm and the distance between each MFC unit was 3 cm. The distances between anode and cathode and surface area were selected according to Estrada-Arriaga et al. (2017), which allowed a high power density using a single air-cathode MFC to be obtained. Cheng et al. (2006) showed that a minor distance between the anode and the cathode (0.5–3 cm) in an air-cathode MFC reduces the ohmic losses and the oxygen crossover from the anode to the cathode, thus improving the MFCs' performance.

The anode electrodes were inoculated with a mixture of 50% raw wastewater from residential housing at Jiutepec, Mexico, and 50% anaerobic granular sludge (56,000 mg volatile suspended solids per litre) from the upflow anaerobic sludge blanket reactor at a paper industry wastewater treatment plant. In order to establish a good microbial community on the anode electrodes, the anodes were immersed into the inoculum for 20 days, and then they were placed into the anodic chambers of the two stacked MFC systems. After that, the two systems were operated at HRT of 10 d at OCV until ensuring a continuous voltage in each MFC unit. This acclimation period allowed the microorganism to adapt to raw wastewater. After achieving a steady acclimation, stacked MFC systems were operated at OCV and then the OCV was changed to CCV with an external resistor of 1,000 Ω. In each circuit mode, the stacked MFCs were operated at three different HRTs: 3, 1, and 0.5 d. Throughout all tests, stacked MFC systems were operated at shared anolyte mode under continuous flow. The anolyte is the wastewater in the anodic chamber of the stacked MFC.

OCV is voltage which is not connected to any load in a circuit in the absence of current. The electromotive force is a thermodynamic value that does not take into account internal losses. The CCV is voltage obtained when the circuit is connected to an external resistor (load); one current is generated. When the MFCs are connected in OCV or CCV the electrons flow at different rates, then the metabolism of bacteria change, and then the production of electricity and removals of organic matter are different, increasing or decreasing the MFC performance. Also the circuit mode operation in the MFC allows the voltage drop in the fuel cell to be determined. Furthermore, it is very important to study the stacked MFC system performance in OCV and CCV.

The stacked MFC systems were fed with raw wastewater from residential housing throughout all tests. COD of raw wastewater was 209 ± 41 mg/L, and total nitrogen (TN) and total phosphorus (TP) concentrations were 38 ± 11 mg/L, and 15 ± 3 mg/L, respectively. Other parameters were analyzed in the raw wastewater: total suspended solids (147 ± 20 mg/L), oil and grease (39.5 ± 12 mg/L), heavy metal (As 0.0021 mg/L, Cd 0.030 mg/L, Cu 0.05 mg/L, Cr 0.1 mg/L, Hg 0.0010 mg/L, Ni 0.050 mg/L, Pb 0.10 mg/L and Zn 0.1314 mg/L) and cyanides (0.02 mg/L).

An electronic device was developed to measure the voltages generated in each individual MFC unit and the voltages of MFC units connected in series during the operation. The voltages generated in each individual MFC unit were monitored every 5 h, and the ones generated in series were monitored every 12 h. The measurements of voltages were collected using a data acquisition system based on LabVIEW software, connected to a personal computer.

Current, I (mA), was calculated using ⁠, where V (mV) is the voltage and R_ex (Ω) is the external resistor. Power density, P (mW/m²), and current density, j (mA/m²), were calculated according to and ⁠, respectively, where A (m²) is the surface area of the anode electrode. Power density–current density (P-j) curves in stacked MFCs were obtained by changing the external resistor from 68 Ω to 10,000 Ω (20 min intervals in each resistor) using a resistor portable box/load bank (10 Ω–40 kΩ) developed in-house. The polarization curves were controlled using LabVIEW software, connected to a personal computer. For P-j curves connected in series, power and current density were normalized, based on the total surface area anode electrode (0.0036 m² for stacked MFC 1; 0.0075 m² for stacked MFC 2, multiplied by the number of MFC units in each system). COD, TN, and TP were measured using Standard Methods (APHA/AWWA/WEF 2005).

When the individual MFC units were switched to a connection in series, the maximum voltage during the acclimated period was 707 mV for stacked MFC 1 and 568 mV for stacked MFC 2. When stacked MFC 1 was switched to a different HRT, the voltages in all individual MFC units decreased. For stacked MFC 2, four individual MFC units were affected by the change of HRT. The electricity productions were generated steadily in all individual MFC units and remained positive. The electricity production for stacked MFC 1 and stacked MFC 2 at OCV were 165 ± 51 mV and 535 ± 30 mV, respectively.

Thermodynamically the maximum voltage of one MFC is ∼1.1 V at open circuit (without external resistor) depending on the type of substrate being used, e.g. acetate. When the MFC is connected to an external resistance, the voltage decreases to 0.7 V if the conditions are ideal. The voltage decreases due to the internal resistance that occurs inside the MFC (ohmic losses, concentration losses and transfer mass). Hence, the size of the MFC does not increase the voltage. If the surface area of the cathode with respect to the anode is increased, the voltage will not increase to >1.1 V, but will help to better the MFC performance in terms of drop voltage; hence, the internal resistances decrease inside MFC.

The voltage obtained from stacked MFC 1 connected in series was 580 ± 65 mV, using 20 individual cells. For stacked MFC 2 connected in series, the voltage was 540 ± 35 mV, which was similar to the one generated in each individual MFC unit. According to the law of conservation of energy, when the cell units are connected in series, it is expected that the voltages generated in each cell are the sum of all voltages generated from all cell units (V_total=V_{MFC 1}+V_{MFC 2}+V_MFCn+1). This principle was not observed in both stacked MFC systems, even when both systems were operated at open and closed circuits. This was due to a voltage drop, generated within the systems. This behavior has already been reported for other studies from stacked systems during series connections (Oh & Logan 2007; Zhuang & Zhou 2009; Kim et al. 2012). When the individual MFC units were connected in series, a voltage drop phenomenon or voltage reversal occurred due to different factors such as: substrate concentration gradient between cells, insufficient oxygen at the cathode, insufficient fuel, impedance differences, and a lack of catalyst and higher internal resistance (ohmic loss) (Oh & Logan 2007). The protons generated in the anode travel through an anolyte up to the cation-exchenge membrane and to the cathode, due to electroneutrality. This oxidation-reduction reaction is faster in the anode than in the cathode. For this reason, when the individual MFC units are connected in series and share the same anolyte, or when the individual cell is in a shared reactor, the voltage drop is caused by ionic cross-conduction between units. This phenomenon was observed in the two stacked MFCs. This ionic cross-conduction was generated by the presence of an internal short current flow (parasitic current) that occurred between anodes and cathodes, due to the way the cells were connected and due to the fact that the anode kinetic loss was lower than the cathode kinetic loss, which generated voltage loss in the two systems. Another factor that caused the voltage drop between units was the architecture of the MFCs, mainly of stacked MFC 2. The individual cells that formed stacked MFC 2 were placed in a shared reactor and these cells were not divided by a separator.

When the MFC units are sharing the same reactor and the distance between each one of them is short, the force with which protons travel from anode to cathode through the anolyte is low, compared to the potential difference between the distance from anode to cathode within their respective MFC (cell_n). In order to increase the transportation force of a proton, which travels from the anode of the cell_n to the cathode of the cell_n+1, it would be necessary to move cell_n away from cell_n+1 enough to generate a potential difference between the anode of cell_n and its cathode, without generating voltage drops between cell_n and cell_n+1. However, it would not be a practical approach, since the reactor design would be very large, impacting directly on the operating costs of the system. A practical alternative to decrease the voltage drop and increase the power production in a stacked MFC is to test different external resistors in each individual cell unit when the system is working in series or parallel circuit, until the minor voltage loss of the system is found. In this way, each individual cell will work as a single MFC without the need to change the configuration of the stacked one.

Figures 4 and 5 show P-j curves obtained from non-stacked units and from stacked MFCs connected in series. The polarization curves were obtained after running at least three HRT values under continuous flow. Polarization curves help to understand the performance of MFCs (internal resistances: activation losses, ohmic losses, and concentration losses) and determine the maximum power generation of the MFC. The peaks indicate the maximum power density of the MFC with respect to the current density. Likewise, P-j curves were obtained under continuous flow mode. The power and current densities generated by each individual cell under different HRTs were not the same due to differences in the substrate concentrations, different microbial activities and anode potential generated in each cell and, also, due to the configuration of the stacked MFC. For the two stacked MFC systems, the maximum power densities generated in both systems were obtained with HRT of 3 days. For individual MFC units from stacked MFC 1, the maximum power density was 1,106 ± 1.2 mW/m², which corresponds to the last individual MFC unit (MFC unit 20). The individual cell that showed low power density (peak of power density) was MFC Unit 10 (10 ± 1 mW/m²). For stacked MFC 1 (connected in series), the maximum power density observed was 79 ± 0.65 mW/m² for 14 ± 2 mV, working voltage at external resistor of 680 ohms. The current density was 1.3 ± 0.4 mA/m². The polarization curves of individual cells from stacked MFC 2 showed a maximum power density of 473 ± 2.5 mW/m² for individual cell 10. Individual MFC Unit 13 was the cell that presented a low power density (maximum power 73 ± 0.5 mW/m²). Lower current densities (below 1 mA/m²) were obtained when the individual MFC units were connected in series. The maximum power density obtained from stacked MFC 2 (connected in series) was 4.2 ± 0.6 mW/m². The magnitude of the voltage drop observed in the two stacked systems was higher when the individual MFC units were connected in series. The P-j curves suggest that the ohmic and activation losses were the main action mechanisms during the voltage drop in the two stacked MFC systems connected in series (Oh & Logan 2007; Ieropoulos et al. 2008; Gurung & Oh 2012; An et al. 2015a, 2015b). Current densities were smaller in stacked MFC 2 and hence the anode potentials were lower, which altered the metabolic activities on the anode biofilm and reduced the electro-active bacteria activity as well as the power densities. When all MFCs were connected in series, the power and current density were calculated based on the surface area of the anode electrode. For this reason, the power and current densities were lower with respect to each individual MFC. Also, the lower current densities in the two MFC systems connected in series connection showed that activation losses were the principal mechanism of voltage drop. The two stacked MFC systems at high HRT showed that the power densities were increased due to an increase of microbial activities, both exoelectrogenic and anaerobic, thus obtaining a high power production and COD removal.

The two tested stacked MFC systems connected in series were not effective for power production at OCV and CCV (low power densities), due to higher voltage drop generated by activation and ohmic losses inside the systems. The OCV and CCV of the two systems connected in series were not equal to the sum of the voltages produced in each individual MFC unit. Voltage drops – in the two systems – occurred due to the architecture of the systems (shared reactor) and due to the use of the same anolyte in all MFC units. The maximum power density of stacked MFC 1 (connected in series) was 79 ± 0.65 mW/m² (current density of 1.3 ± 0.4 mA/m²). For the individual MFC unit (not connected in series), the maximum power density was 1,106 ± 1.2 mW/m² (current density of 5.5 0.6 mA/m²) at a HRT of 3 days. The power production of stacked MFC 2 (4.2 ± 0.6 mW/m²) and its current density (0.04 ± 0.006 mA/m²) were lower compared to the power generated by stacked MFC 1. The results showed that the COD removal increased when the HRT was increased from 0.5 to 3 d.

Financial support was provided by SEP-CONACYT, project CB-2013/221433.