Membrane aeration as an energy-efficient method for supplying oxygen to microbial fuel cells

Abstract

An active supply of oxygen will help with electricity generation in microbial fuel cells (MFCs). Although air-cathode MFCs have been developed to eliminate aeration, the MFCs with active aeration could have their niches in system scaling up and/or removal of certain contaminants that require oxygen in the cathode. In this study, an alternative aeration method based on a gas-transfer membrane has been investigated for MFC applications, in comparison with diffused aeration. The membrane-aerated microbial fuel cell (MAMFC) achieves a maximum coulombic efficiency (CE) of 55.4%, a current density of 17.3 A m⁻³ and COD removal efficiency of >61%. The CE of the MAMFC is higher than that of the diffused aeration MFC (DAMFC), indicating a higher conversion efficiency of substrate to electricity with membrane aeration. At the similar dissolved oxygen level of 6.61 mg O₂ L⁻¹, the MAMFC requires an energy input of 0.05 kW h m⁻³, significantly lower than 1.76 kW h m⁻³ by the DAMFC. Although both MFCs have negative energy balances under the testing conditions, the MAMFC could theoretically save 588–3485% of energy compared with the DAMFC. This study demonstrates that membrane aeration could be an energy efficient method for providing an active oxygen supply for MFC applications.

---

1. Introduction

Microbial fuel cells (MFCs) are an emerging technology that can convert chemical energy in organic compounds to electrical energy through catalytic reactions of microorganisms under anaerobic conditions.¹ This technology can be potentially applied to treat wastewater with simultaneous contaminant removal and energy recovery.² In addition, the low sludge production and low energy consumption by MFCs will help to create a cost-effective approach for sustainable wastewater treatment.³ Given those potential advantages, MFCs have attracted a great attention from the research community with significant advancement in the aspects of microbiology, electrochemistry, materials, system configuration and operation, and modularization.^4–6 The challenges for further developing MFC technology include system scaling up, improving energy recovery, decreasing energy consumption, and reducing material cost. Among those, the optimization for reducing energy consumption have not been well investigated, likely related to the fact that energy consumption by MFC systems is not often reported in the literature. Some prior studies have identified that the main energy consumers in an MFC system are pumping for feeding/recirculation and aeration.⁷

Aeration does not seem to be a major issue, because the development of the air-cathode MFCs has eliminated the active aeration. However, given the difficulty of scaling up air-cathode MFC systems at this moment and a potential need for aeration in the cathode (e.g., for nitrification when nitrogen removal is encorporated as a part of the MFC system⁸), there remains possibility that aerated MFC systems may need to be further developed. In this way, precise control of aeration for the purpose of optimized treatment performance and minimized energy consumption will become important. Conventional air supply in MFCs is based on diffused aeration, similar to that of activated sludge processes with a high demand for energy. Alternative aeration methods have been developed, and among them, membrane aeration is of strong interest because of significantly high oxygen utilization efficiencies with energy saving.⁹ For membrane aeration, bubble-free oxygen supply is achieved by placing a thin film of synthetic polymer at the interface of the liquid and gas phase. Thus, high oxygen-mass transfer efficiency can be achieved, which confers an economic advantage in terms of aeration energy requirement.^10,11 In diffused aeration systems, only 5–25% of oxygen supplied by blowers is transferred to the liquid phase.¹² Membrane aeration systems allow the oxygen to be continuously vented in the membrane lumen (in a flow-through mode), achieving high oxygen transfer efficiency up to 100%.¹³

Membrane aeration has been applied to biological wastewater treatment, for example, membrane-aerated biofilm reactors (MABR) can effectively remove contaminants with precise control of oxygen supply.¹⁴ Membrane aeration was also employed to improve nitrogen removal via simultaneous nitrification and denitrification¹⁵ and nitritation–anammox process by facilitating mass transfer.¹⁶ However, there is very limited information about membrane-aerated MFCs (MAMFC). A previous study of an MAMFC focused on nitrogen removal and found that the nitrogen removal efficiency was improved to 52% compared with 24% with a diffused aeration MFC (DAMFC), likely due to the enhanced denitrification with low dissolved oxygen (DO) of 0.5 mg L⁻¹ in the cathode chamber.¹⁷ The energy performance of the MAMFCs, such as production, consumption and balance, has not been reported before. In this study, membrane aeration was evaluated in a tubular MFC for its effects on treatment performance and energy production/consumption, and compared with diffused aeration. A series of lumen pressures (i.e., intra-membrane pressures) and flow rates have been examined under membrane aeration and diffused aeration, respectively. Energy production and consumption was expressed in kW h m⁻³, and eventually energy balances under the selected conditions were established for comparing membrane aeration with diffused aeration.

2. Materials and methods

2.1. MFC setup

The MFC used in this study consisted of two compartments, physically separated by a cation exchange membrane (CEM) (Membrane International Inc., Ringwood, NJ, USA). The anode compartment (370 mL) was placed in a tubular container, which created a cathode compartment of 1220 mL. The anode electrode was made of non-wet proofed carbon brush, which was pretreated by being immersed in acetone overnight and then heated at 450 °C for 0.5 h, while the cathode electrode was wet-proofed carbon cloth (13.5 cm × 30 cm, Zoltek Corporation, St. Louis, MO, USA) containing 4 mg cm⁻² activated carbon powder (Thermo Fisher Scientific, Bridgewater, NJ, USA) as a catalyst for oxygen reduction. The activated carbon powder was coated to the carbon cloth by using a 15% PTFE solution as a binder agent and heat-treated at 370 °C for 0.5 h. A gas-permeable silicone membrane module (5 m active length, outer diameter 3.18 mm, and inner diameter 1.98 mm, Silastic laboratory tubing #508-009, Dow Corning Corp., Midland, MI, USA) wrapped around the carbon cloth. By adjusting the lumen pressure (1, 2.5, 10, 15, 20, 25, 28 and 30 psi) in the MAMFC, the DO concentration in the cathode compartment could be controlled. In the DAMFC, air was pumped via a diffuser (a flexible air curtain with a diameter of 1.37 cm, pore size of 1 mm, and a length of 60 cm, Active Aqua, USA) adjacent to the cathode electrode to provide oxygen. Meanwhile, a flow meter was used to control the air flow rate (10, 20, 30, 40 and 50 mL min⁻¹).

2.2. MFC operation

The MFC was operated in a continuous mode at room temperature (21 ± 1 °C) and with its anode being fed with a synthetic domestic wastewater, which was composed of 300 ± 18 mg COD L⁻¹, 70 ± 5 mg L⁻¹ NH₄⁺–N, 15 ± 3 mg L⁻¹ MgSO₄, 20 ± 2 mg L⁻¹ CaCl₂, 500 ± 10 mg L⁻¹ NaCl, 100 ± 8 mg L⁻¹ NaHCO₃, 5.35 ± 0.20 mg L⁻¹ K₂HPO₄, 2.65 ± 0.60 mg L⁻¹ KH₂PO₄, and trace elements.¹⁸ The anode hydraulic retention time (HRT) was set at 17.6 h throughout the experiment. The MFC anode chamber was inoculated with 50 mL sludge from the anaerobic digester in a local wastewater treatment plant (Pepper's Ferry Wastewater Treatment Authority, Radford, VA, USA).

2.3. Measurement and analysis

The MFC voltage was recorded by a digital multimeter (2700, Keithley Instruments, Inc., Cleveland, OH, USA) every 5 min. The concentration of chemical oxygen demand (COD) was measured by using a DR/890 colorimeter (Hach Company, Loveland, CO, USA). Current density (A m⁻³) was calculated based on the liquid volume of the anode compartment. The statistical error bars are given as the standard deviations on the principle that each current generation within 24 hours act as an independent count. The coulombic efficiency (Q_eff) was determined by integrating the amount of coulombs produced over time, divided by the amount of coulombs present in the influent:where COD_inf is the COD concentration of the influent, n is the number of electrons per mole of O₂ (equal to 4), MM⁻¹ is the molar mass of O₂ (32 g mol⁻¹), F is the Faraday constant (96 485 C mol⁻¹ electrons), Q_f is the flow rate of the influent, and Δt is the time of operation. The energy balance was energy production subtracting total energy consumption, including energy requirement for the pumping system and aeration. Energy production was expressed as normalized energy recovery (NER) in kW h m⁻³ and calculated based on energy recovery and the volume of the treated wastewater.¹⁹ The theoretical power requirement for the pumping system was estimated as:where P is the pumping system power requirement (kW), Q is the flow rate (m³ s⁻¹), γ is specific weight of water (9800 N m⁻³), and E is the hydraulic pressure head (m).²⁰ In this study, the hydraulic pressure head of 0.02 m was estimated for the recirculation pump. For the membrane aeration, adiabatic compression energy was calculated following the standard method:²¹where R is gas law constant (8.314 J K⁻¹ mol⁻¹), T is absolute temperature (294 K), k is ratio of specific heat of air at constant volume (1.395), P₁ is absolute pressure before compression (101 325 Pa), P₂ is absolute pressure after compression (Pa), e_b is blower efficiency (0.7), e_d is drive efficiency (0.95), e_m is motor efficiency (0.92), and W_m is adiabatic compression energy (J mol⁻¹ air). Energy consumption by the diffused aeration was estimated as:²²where P₁ is standard atmospheric pressure (101 325 Pa), P₂ is blower inlet pressure (Pa), T is air temperature (294 K), ζ is blower efficiency (0.8), λ is aerator constant (1.4), and ρ₀ is air density in standard conditions (1.29 kg m⁻³).

3. Results and discussion

3.1. MFC performance with two aeration methods

After about one month of start-up, the MFC was then operated for 180 days through two stages: in the stage I (120 days, MAMFC), membrane aeration was examined by delivering the air to the cathode chamber with controlled lumen pressure; in the stage II (60 days, DAMFC), diffused aeration was applied. The DO was directly affected by aeration strength and a higher aeration intensity (either a higher lumen pressure or a higher flow rate) resulted in a higher DO (Fig. 1). For example, the DO dramatically increased from 0.92 mg O₂ L⁻¹ to 6.96 mg O₂ L⁻¹ with increasing the lumen pressure from 1 psi to 20 psi in the MAMFC, but this increase became much slower when the lumen pressure was higher than 20 psi (Fig. 1A). A similar trend of the DO was observed in the DAMFC; however, the amount of air supply in DAMFC could be 11 to 12 times more than that of MAMFC to create a similar DO level. For example, to reach the DO of 4.56 mg O₂ L⁻¹, 10 mL of air was supplied by the diffuser while 0.9 mL of air was permeated through the membrane.

Fig. 1 The concentrations of DO and current density: (A) MAMFC system and (B) DAMFC system. Error bars represent the standard deviation, n = 5.

Current generation was strongly related to the DO concentration with both aeration methods (Fig. 1). In the MAMFC, the current density increased from 0.4 to 17.3 A m⁻³ when the DO increased from 0.41 to 6.41 mg L⁻¹, and further increase in the DO did not significantly improve the current density, likely related to the insufficient supply of protons from the anode to match the increased demand for the cathode reaction.²³ In the DAMFC, the current density increased from 5.9 to 14.2 A m⁻³ with the DO increased from 4.98 to 8.03 mg L⁻¹. Although the DAMFC had higher DO than the MAMFC, its current density was generally lower at the similar DO level. Such disparity could be attributed to higher oxygen transfer efficiency of membrane aeration than that of diffused aeration. Due to the lack of agitation in the cathode chamber, the primary obstruction for oxygen transfer lies in the liquid interface. Membrane aeration offers an advantage over diffused aeration as it can be operated at a high pressure, which increases the concentration gradient and hence the oxygen transfer rate.²⁴ Another possible reason for lower current generation in the DAMFC is that the diffusion aeration caused more oxygen transport from the cathode into the anode, thereby inhibiting the activities of electrogenic bacteria in the anode chamber; however, this could also occur with the membrane aeration and is difficult to quantify at this moment.

Both MFCs had similar COD removal efficiency, varying from 61.1% to 72.0% in the MAMFC and from 64.0% to 71.7% in the DAMFC (Fig. 2). In the absence of aeration, the MFC achieved 34.3% of COD removal, mostly because of anaerobic oxidation of organic compounds (Fig. 2A). Aeration significantly increased COD removal, due to current generation that stimulated organic oxidation.²⁵ Oxygen intrusion from the cathode into the anode may also contribute to COD removal. Coulombic efficiency (CE), which was determined by both COD input and current generation, exhibited different profiles between the two aeration methods. The CE of the MAMFC increased from 29.8% to 54.4% when the lumen pressure increased from 1 to 20 psi, but further increase in the lumen pressure led to a decline of CE (Fig. 2A). Previous study reported that the presence of oxygen in the anode chamber could lead to deterioration of the MFC performance.²⁶ The oxygen permeated via CEM from the cathode chamber could be more under a higher lumen pressure, resulting in a higher organic substrate utilization rate directed toward the pathway of aerobic degradation. The CE of the DAMFC persistently increased from 14.2% to 28.0% when the air flow was improved from 10 to 50 mL min⁻¹ (Fig. 2B). The maximum CE of the MAMFC was 94.5% higher than that of the DAMFC, indicating that the potential effects of oxygen in the anode chamber might be more serious in the DAMFC. A more detailed comparison of the CE between the two aeration methods was conducted under three selected DO conditions (4.56 mg O₂ L⁻¹, 5.41 mg O₂ L⁻¹, and 6.61 mg O₂ L⁻¹), which were achieved in the MAMFC at three lumen pressures of 15, 20 and 28 psi, or in the DAMFC at three air flow rates of 10, 20 and 30 mL min⁻¹. As shown in Fig. 2C, the CE of the MAMFC was 75.7–246.3% higher that of the DAMFC under the similar DO conditions, indicating that the MAMFC could achieve better conversion of organics to electric energy. Hence, an optimum balance between the improvement of the cathode reaction and the deterioration of the anode reaction due to oxygen migrations must be established.

Fig. 2 COD removal and coulombic efficiency: (A) MAMFC system, (B) DAMFC system, and (C) the comparisons of coulombic efficiency in MAMFC and DAMFC at three selected DO levels. Error bars represent the standard deviation, n = 5.

3.2. Energy production, consumption and balance

Energy production and consumption in the MFC system was expressed as kilowatt hour per cubic meter of the treated wastewater (kW h m⁻³). The energy consumption by the DAMFC was much higher than that by the MAMFC, and this difference is significant (Fig. 3A and B). For example, the maximum energy consumption by the MAMFC with the lumen pressure of 20 psi was 0.054 kW h m⁻³, much lower than the minimal energy requirement of 0.20 kW h m⁻³ by the DAMFC at the flow rate of 10 mL min⁻¹. Given that the energy consumption by the recirculation pump was same at 0.0059 kW h m⁻³ in both MFCs, the aeration energy contributed to 1.0–91.6% in the MAMFC and 97.1–99.9% in the DAMFC, suggesting aeration energy was the major contributor to energy consumption in two MFCs. Energy production in the MAMFC was 0.00037–0.019 kW h m⁻³, which was in the similar range of that in the DAMFC (0.0022–0.013 kW h m⁻³). The highest energy production of 0.019 kW h m⁻³ was obtained in the MAMFC at 25 psi, which slightly decreased to 0.017 kW h m⁻³ when the lumen pressure was >25 psi. On the other hand, the energy production in the DAMFC demonstrated a positive relationship with the increased air flow rate; however, the highest energy production of 0.013 kW h m⁻³ in the DAMFC was 32.0% lower than that of the MAMFC.

Fig. 3 Energy production and energy consumption: (A) MAMFC system, (B) DAMFC system, and (C) the comparisons of energy balance in MAMFC and DAMFC under three selected DO levels.

Neither of the MFCs could theoretically generate self-sufficient energy, and thus the energy balance was negative under all the conditions. However, the energy balance for the DAMFC was much more negative than that of the MAMFC, because of its much higher energy consumption. Fig. 3C showed the comparison of energy balance between the two MFCs under three DO conditions (4.56, 5.41, and 6.61 mg O₂ L⁻¹). With the DO of 4.56 mg O₂ L⁻¹, the energy balance was −0.029 kW h m⁻³ for the MAMFC, while it decreased by 26.3% (more negative) with a higher DO of 5.41 mg O₂ L⁻¹ (−0.037 kW h m⁻³). Further increasing the DO to 6.61 mg O₂ L⁻¹ resulted in an energy balance of −0.049 kW h m⁻³. Similar energy balance tendency was observed in DAMFC but with much more negative values (−0.20, −0.79, −1.76 kW h m⁻³ for 4.56, 5.41 and 6.61 mg O₂ L⁻¹, respectively), which corresponded with the large amount of energy consumption by diffused aeration (Fig. 3B). The MAMFC would save 588–3485% of energy, compared with the DAMFC. It should be noted that this study focused on aeration methods, instead of optimizing MFC systems for maximal energy production. It is possible to increase energy production by optimizing reactor configuration and operation, catalysts, and substrates, towards a possible energy balance in the MAMFC.

4. Conclusions

The results of this study have demonstrated the advantages of membrane aeration in terms of energy consumption and production, compared with diffused aeration for MFC applications. Despite the negative energy balances under all the testing conditions, the membrane aerated MFC could theoretically save a significant amount of energy without obvious influence on organic removal. Further research is warranted to increase energy production for achieving positive energy balances and understanding of potential fouling of membranes when biological processes are involved in the cathode. This study provides implications to special application niches for MFCs where aeration is needed in the cathode for the purpose of facilitating system scaling up and/or accomplishing removal of a certain contaminants such as nitrogen that requires oxygen.

Acknowledgements

This work was financially supported by a grant from National Science Foundation (#1358145). Yuli Yang was supported by Preponderant Discipline of Southeast University (#CE02-1/2-0×) and National Natural Science Foundation of China (#41571476).

Notes and references

1. B. E. Logan, B. Hamelers, R. Rozendal, U. Schröder, J. Keller, S. Freguia, P. Aelterman, W. Verstraete and K. Rabaey, Environ. Sci. Technol., 2006, 40, 5181–5192.
2. D. Pant, G. Van Bogaert, L. Diels and K. Vanbroekhoven, Bioresour. Technol., 2010, 101, 1533–1543.
3. M. Sun, L.-F. Zhai, W.-W. Li and H.-Q. Yu, Chem. Soc. Rev., 2016, 45, 2847–2870.
4. J. Arends and W. Verstraete, Microb. Biotechnol., 2012, 5, 333–346.
5. B. E. Logan and K. Rabaey, Science, 2012, 337, 686–690.
6. Z. Ge, L. Wu, F. Zhang and Z. He, J. Power Sources, 2015, 297, 260–264.
7. Z. Ge, Q. Ping, L. Xiao and Z. He, Desalination, 2013, 312, 52–59.
8. B. Virdis, K. Rabaey, R. A. Rozendal, Z. Yuan and J. Keller, Water Res., 2010, 44, 2970–2980.
9. A. Terada, K. Hibiya, J. Nagai, S. Tsuneda and A. Hirata, J. Biosci. Bioeng., 2003, 95, 170–178.
10. E. Syron and E. Casey, Environ. Sci. Technol., 2008, 42, 1833–1844.
11. X. Wei, B. A. Li, S. Zhao, L. Wang, H. Y. Zhang, C. Li and S. C. Wang, Bioresour. Technol., 2012, 122, 189–195.
12. M. Aybar, G. Pizarro, K. Martin, J. Boltz, L. Downing and R. Nerenberg, Proceedings of the Water Environment Federation, 2012, 2012, 5458–5485.
13. E. Syron, M. J. Semmens and E. Casey, Chem. Eng. J., 2015, 273, 120–129.
14. K. J. Martin and R. Nerenberg, Bioresour. Technol., 2012, 122, 83–94.
15. X. Wei, B. A. Li, S. Zhao, C. C. Qiang, H. Y. Zhang and S. C. Wang, J. Membr. Sci., 2012, 389, 257–264.
16. X. Li, S. Sun, B. D. Badgley, S. Sung, H. Zhang and Z. He, Water Res., 2016, 94, 23–31.
17. C.-P. Yu, Z. Liang, A. Das and Z. Hu, Water Res., 2011, 45, 1157–1164.
18. J. Hou, Z. Liu and P. Zhang, J. Power Sources, 2013, 224, 139–144.
19. L. Xiao, Z. Ge, P. Kelly, F. Zhang and Z. He, Bioresour. Technol., 2014, 157, 77–83.
20. J. Kim, K. Kim, H. Ye, E. Lee, C. Shin, P. L. McCarty and J. Bae, Environ. Sci. Technol., 2010, 45, 576–581.
21. P. Cote, J.-L. Bersillon, A. Huyard and G. Faup, J.–Water Pollut. Control Fed., 1988, 1986–1992.
22. B. Verrecht, S. Judd, G. Guglielmi, C. Brepols and J. Mulder, Water Res., 2008, 42, 4761–4770.
23. J. R. Varcoe, P. Atanassov, D. R. Dekel, A. M. Herring, M. A. Hickner, P. A. Kohl, A. R. Kucernak, W. E. Mustain, K. Nijmeijer and K. Scott, Energy Environ. Sci., 2014, 7, 3135–3191.
24. A. Hussain, S. R. Guiot, P. Mehta, V. Raghavan and B. Tartakovsky, Appl. Microbiol. Biotechnol., 2011, 90, 827–836.
25. F. Zhang, K. S. Jacobson, P. Torres and Z. He, Energy Environ. Sci., 2010, 3, 1347–1352.
26. M. Rodrigo, P. Canizares, J. Lobato, R. Paz, C. Sáez and J. Linares, J. Power Sources, 2007, 169, 198–204.

---