Analysis of the effect of biofouling distribution on electricity output in microbial fuel cells

Abstract

Single chamber air-cathode microbial fuel cells (MFCs) were operated for 6 months to demonstrate the impact of biofouling distribution on cathode performance. Total biofouling decreased the maximum power density by 38% from 892 ± 8 mW m⁻² to 549 ± 16 mW m⁻². Cleaning surface biofouling slightly enhanced the power density by 12%, but additional removing of the biofouling inside the catalyst layer further increased power output by 30% to 802 ± 14 mW m⁻², indicating that inner biofouling aggressively inhibits cathode activity. Compared with surface biofouling, the inner biofouling clogged a portion of pores in the catalyst layer, which severely reduced oxygen permeability, conductivity and reaction sites. Consequently, the kinetic activity of the cathode was impaired as the exchange current density declined and the charge transfer resistance increased. Thus, it is shown that the biofouling within the catalyst layer plays a more crucial role for air cathodes over long-term operation.

---

1 Introduction

Microbial fuel cells (MFCs) are devices that convert chemical energy from organic matter into electricity.^1–3 Air-cathode single chamber MFCs are widely used in research due to their simple configuration and readily available oxygen from the air as a sustainable electron acceptor.^4–6 Generally, oxygen reduction reaction (ORR) occurs on three-phase boundaries which depend on exposure area of catalyst and framework of catalyst layer. Traditional Pt catalyst efficiently reduces the overpotential of ORR, however, the expensive cost limits the extensive application in MFCs. Since activated carbon based cathode acquires comparable performance to Pt/C due to abundant pores and larger specific area,^7,8 it has attracted increasing attention. However, one of the drawbacks of the porous structure in the catalyst layer lies in the faint resistance to biological fouling. Even through some ionic membranes and separators are used in MFCs, the adverse effect of biofouling on proton exchange layers^9,10 and separators¹¹ are also confirmed to impair the MFCs performance.

With removing the ionic membranes and separators in single chamber MFCs, the catalytic layer is exposed directly to the substrate. Bacteria in the substrate attach to the surface of catalyst layer and release extra-cellular polymeric substances (EPSs), which develop the biofouling. Just like the influence of biofilm on membranes and separators, it is noteworthy to verify the effect of biofouling on catalyst layer. Previous result showed that the thick biofilm developed on carbon based cathode probably increased the diffusion resistance,¹² which was responsible for the declined performance after running for one year. Coincidentally, the activity of catalyst layer with different dopants (Fe, heat treatment, and carbon black) also decreased due to the mixture fouling of biofilms and salt deposition. After removing the fouling by weak hydrochloric acid, cathode performances increased as power density improved by 14–29%.¹³ Despite several studies observe the biofouling during long-term operation, these works mainly focus on reporting the impact of entire biofouling on MFCs performance. However, there are not clear and enough demonstrations regarding the individual effect of biofouling distribution on cathode performance, i.e., biofouling located on surface of catalyst layer and inside the layer.

This work focused on investigating the effect of biofouling distribution on the performance of catalyst layer. The comprehensive research was conducted with carbon cathode in MFCs to estimate the contribution of biofouling location to the porous structure, oxygen permeability, electrochemical activity of cathode. In addition, the interrelation ship between biofouling behavior and property of carbon catalyst layer were proposed.

2 Materials and methods

2.1 Cathode preparation

Cathode was prepared by rolling-press method which successively pressed gas diffusion layer and catalyst layer on each side of stainless steel mesh (60 mesh) as previously described.¹⁴ The activated carbon powder (Carbosino Material Co., Ltd, Shanghai, China) and NH₄HCO₃ as pore former (mass ratio of 5 : 1)¹⁵ were stirred homogeneously, followed by dropping PTFE with ratio of carbon powder : binder at 6 : 1 and then rolled to be catalyst layer. The carbon black powders (Hesen Electrical Co., Ltd, Shanghai, China) and PTFE with mass ratio of 3 : 7 were mixed evenly and used to form gas diffusion layer which further sintered at 340 °C. Finally, the prepared cathodes were heated at 160 °C in a muffle furnace to remove pore former.

2.2 MFC construction and operation

The single chamber MFC was constructed of cylindrical plastic chamber with 4 cm in length and 3 cm in diameter (a liquid volume of 28 mL) as described.¹⁶ The prepared cathode and carbon fiber brush anode pretreated by heating 30 minutes at 450 °C (ref. 17) were placed vertically with distance of 1 cm. The MFCs were fed with 20% (v/v) domestic wastewater collected from a municipal pipe network (Harbin, China) and 80% medium which was 1 g L⁻¹ glucose, 50 mM phosphate buffer solution (PBS) contained vitamins 5 mL L⁻¹ and minerals 12.5 mL L⁻¹ during inoculation as described.^18,19 All reactors were operated in at least duplicate in fed-batch mode at 30 °C with external resistor of 1000 Ω except as noted.²⁰

2.3 Morphology of catalyst layer

Scanning electron microscope (SEM, S-4700, Hitachi Ltd.) equipped with energy dispersive X-way spectrum (EDX) was used to observe the surface and cross-section morphology and chemical element of the catalyst layer. For SEM analysis, the fouled samples were immersed in 2.5% glutaraldehyde for 30 min, followed by washing with PBS (50 mM, pH 7.2) for three times and deionized water once. Then, the samples were dehydrated by adding 50%, 70%, 90% and 100% ethanol successively and replaced by tert-butanol. After removing organic solvents by nitrogen blow, the samples were placed in cryogenic freeze dryer overnight.

2.4 Electrochemical activity of cathode

Electrochemical measurements, linear sweep voltammetry (LSV), exchange current density (i₀) and electrochemical impedance spectroscopy (EIS), were tested with Auto Lab PGSTAT128N (Metrohm, Swiss) in 50 mM PBS at ambient temperature (25 °C) to evaluate the effect of biofouling location on cathode activity. All electrochemical measurements were conducted using three electrodes abiotic electrochemical cell which was constructed by cylindrical plastic chamber (3 cm diameter, 4 cm long) with a net volume of 28 mL. Carbon cathode was used as the working electrode (projected surface area of 7 cm²), Pt sheet (1 cm²) as the counter electrode and Ag/AgCl electrode as the reference electrode (saturated KCl, +197 mV versus standard hydrogen electrode; SHE).

LSV was performed from 0.3 V to −0.3 V with a scan rate of 1 mV s⁻¹.²¹ Tafel plot was recorded by sweeping the overpotential (|η|) from 0 to 100 mV at 1 mV s⁻¹,²² where η = 0 is the open circuit potential of the cathode versus Ag/AgCl reference electrode, and exchange current density (i₀, A cm⁻²) was calculated based on Tafel plots. EIS was conducted at constant potential of −0.1 V vs. Ag/AgCl over a frequency range of 100 kHz to 10 mHz with the amplitude of 10 mV to further explain the change of catalytic activity.²³ This polarized potential was chosen because the maximum power of MFCs was obtained around −0.1 V in most cases based on electrode potential curve. The experimental data were fitted with equivalent circuit R(Q(RW)) by Zsimpwin software 3.10 to obtain the resistance distribution.

2.5 Oxygen permeability of cathode

Oxygen diffusion through cathode was measured using a reactor configuration same to MFC reactor but without anode, which was filled with 50 mM PBS. Before measurement, the PBS was sparged with nitrogen gas to remove dissolved oxygen and then sealed by rubber plug. A magnetic stirring rotor was used to agitate buffer to facilitate even oxygen distribution in the chamber. After fixing the non-consumptive dissolved oxygen probe (FOXY, Ocean Optics, Inc., Dunedin, FL) and temperature sensor in the chamber with 0.5 cm distance to cathode, oxygen concentration was recorded continuously, the mass transfer coefficient was calculated based on formula.²⁴
where v is the net volume of the reactor, A the cross-sectional area of cathode, C the oxygen concentration in the solution at time t, and C_s the oxygen concentration at the air side of the cathode which was assumed to be the saturation concentration of oxygen in water.

2.6 Calculations

Cell voltages across external resistor were recorded at a time interval of 30 minutes using a data acquisition board (PISO-813, ICP DAS Co., Ltd.) connected to a personal computer. Power density and polarization curves were obtained by varying the external resistor over a range from 1000 to 50 Ω. Current density and power density were calculated from I = U/RA and P = U²/RA, where U (V) is the steady-state voltage, R (Ω) is the external resistance, and A (m²) is the projected surface area of the cathode. Anode and cathode potentials were measured using Ag/AgCl as reference electrode.

3 Results and discussion

3.1 MFCs performance with biofouling

To better understand the impact of biofouling on MFCs performance, maximum power density (P_max) and electrode potentials were tested (Fig. 1A). The original cathode generated P_max of 892 ± 8 mW m⁻². After 6 months' operation, the biofouled cathode decreased P_max by 38% to 549 ± 16 mW m⁻², indicating the severe impairment of MFCs performance due to biofouling. Biofouling of cathode surface was removed, followed by rinsing repeatedly with deionized water to clear surface residues (denoted as Ext-cleaned cathode) to evaluate the impact of surface biofilm on cathode activity. As a result, the P_max was enhanced by 12% after cleaning surface biofilm than that of the biofouled cathode. Compared with stable anode potentials (Fig. 1B), the overall performance of the Ext-cleaned cathode was hardly comparable to that of the original cathode, which could be because the biofouling accumulation inside the catalyst layer blocked the gas pores and reduced catalytic activated interfaces.

Fig. 1 Power density (A) and electrode potential (B) as a function of current density in MFCs. Measurements were conducted with original cathode, biofilm cathode and Ext-cleaned cathode (cathode with cleaning surface biofouling). Error bars ± SD were based on averages.

3.2 Surface and cross-section morphology of catalyst layer

The effect of surface and inner biofouling on porous structure of catalyst layer was examined by SEM, and the salt deposition mixed with biofilm was investigated by EDX. The SEM images clearly showed that the porous structure of original catalyst layer on water-facing side (Fig. 2A) was covered by biofilm layer after running for 6 months (Fig. 2B). The thick biofouling was developed mainly because the large and rough specific surface of original catalyst layer favored bacteria attachment. Some inorganic salt precipitations, including kalium, sodium, calcium and iron, were deposited on biofilm layer due to cations migration from the medium according to the results of EDX (Fig. 2E).²⁵ Compared with water-facing side, the catalyst layer on mesh-facing side still maintained the porous structure (Fig. 2C), indicating the thickness of catalyst layer could obstruct the total penetration of biofouling.

Fig. 2 Surface morphology and elements distribution of catalyst layer. SEM images of surface morphology and EDX analysis for elements distribution of original catalyst layer (A and D), biofilm catalyst layer faced water (B and E) and stainless steel mesh side (C and F).

Fig. 3 showed the microphotographs of cross-section of layer. Apparent biofouling was detected inside the catalyst layer. This was caused by the larger pore size of abundant open pores than microorganisms. As a result, EPSs surrounded the rugged carbon skeleton and filled partial pores (Fig. 3D). These clogged pores would reduce the porosity in catalyst layer, which impaired the conductivity and reduced the reaction interfaces. Additionally, the confirmed inner biofouling would lower the oxygen permeability of carbon cathode.

Fig. 3 SEM images of cross-section morphology for original and biofilm catalyst layer. Lower (×200, A and C) and higher (×1000, B and D) magnification were showed. The representative area remarked by red square was showed with higher magnification.

3.3 Effect of biofouling distribution on oxygen permeability

Fig. 4 showed the variation of oxygen transfer for the original and biofouled cathode. The oxygen transfer coefficient (K_O2) reduced due to biofouling on the cathode surface and pores. The original cathode had K_O2 1.3 ± 0.07 × 10⁻⁴ cm s⁻¹, whereas it decreased to 0.21 ± 0.04 × 10⁻⁴ cm s⁻¹ for the biofouling cathode. The lower oxygen permeability was probably attributed to two aspects: one reason was due to the oxygen consumption by aerobic microorganisms in biofilm,^26,27 the other was the diffusion resistance of blocked pores inside catalyst layer as a result of biofouling. The big difference in the oxygen permeability indicated that the biofouling decreased oxygen diffusion over times which would obstruct the oxygen reduction. When surface biofilm was physically removed, the Ext-cleaned cathode slightly enhanced K_O2 (0.4 ± 0.05 × 10⁻⁴ cm s⁻¹). Meanwhile, the P_max was also increased but only by 12% increment. Considering power density and oxygen permeability, the positive relation between the two parameters confirmed the biofouling decreased MFCs performance by reducing oxygen permeability. However, the oxygen transfer of Ext-cleaned cathode was still much lower than that of original cathode, which could be due to the inner biofouling.

Fig. 4 The relationship between oxygen diffusion coefficient and maximum power density. Measurements were conducted with original cathode, biofouled cathode, Ext-cleaned cathode (cathode with cleaning surface biofouling) and Int-cleaned cathode (cathode with further cleaning biofouling inside the catalyst layer by heating and ultrasonic concussion). Error bars ± SD were based on averages.

In order to evaluate the impact of inner biofouling on oxygen permeability, biofilm inside the cathode was eliminated as follows: the cathode removed surface biofouling was heated at 120 °C in a muffle furnace for 1 hour to completely dehydrate the biofilm inside the cathode, followed by ultrasonic concussion in deionized water for another 1 hour to remove residues stuck in pores as much as possible (denoted as Int-cleaned cathode). The results revealed that oxygen diffusion of the Int-cleaned cathode was improved by 2.3 times compared to that of Ext-cleaned cathode, and P_max further increased by 30% to 802 ± 14 mW m⁻². As the sole electron acceptor, oxygen concentration was one of the important factor for higher cathode performance especially during long-term operation.²⁸ Since biofouling inside the catalyst layer blocked pores, oxygen transported along activated reaction boundaries decelerated and caused deficient reactant to reduction reaction. The result indicated inner rather than surface biofouling was the limiting factor to impact the oxygen permeability of cathode.

3.4 Effect of biofouling distribution on electrochemical activities

The electrochemical characters of cathode were tested by linear sweeping voltage (LSV), exchange current density (i₀), and electrochemical impedance spectroscopy (EIS). By comparison, the maximum and minimum current responses were obtained with the original and biofouled cathode, respectively (Fig. 5A). Current density was enhanced after removing biofouling, especially for the Int-cleaned cathode compared to Ext-cleaned cathode. High exchange current density (i₀) could reduce the activation energy barrier of the cathode reaction. Based on the Tafel plots (Fig. 5B), the i₀ was calculated and listed in Table 1. The i₀ of the Int-cleaned cathode was 0.78 ± 0.01 A m⁻² which was 1.3 and 1.2 times higher than that of biofouled cathode and Ext-cleaned cathode, and the kinetic activity (KA) of Int-cleaned cathode also increased by 23% than that of Ext-cleaned cathode.

Fig. 5 Linear sweeping voltage (A), Tafel plot (B), and electrochemical impedance spectroscopy (C) for different cathodes. Measurements were conducted with original cathode, biofouled cathode, Ext-cleaned cathode (cathode with cleaning surface biofouling) and Int-cleaned cathode (cathode with further cleaning biofouling inside the catalyst layer by heating and ultrasonic concussion) in 50 mM phosphate buffer solution. The inserted figure in (B) indicated the linear fit for the Tafel plots of overpotential from 80 mV to 100 mV. The electrochemical impedance spectroscopy was conducted at −0.1 V (versus Ag/AgCl), for Nyquist plot, symbols represented the experimental data, lines represented the fitting data with the equivalent circuit.

Table 1 Exchange current density and internal resistance distributions and capacitance of different cathodes

Cathode	Original	Biofouled	Ext-cleaned	Int-cleaned
i0 (A m^−2)	0.96 ± 0.06	0.58 ± 0.05	0.63 ± 0.02	0.78 ± 0.01
KA	1	0.59	0.70	0.85
Rohm (Ω)	22.3 ± 0.8	29.0 ± 0.1	24.1 ± 1.6	26.1 ± 0.4
Rct (Ω)	50.4 ± 0.6	119.3 ± 12.8	91.7 ± 3.5	66.8 ± 3.0
Zw (Ω s^−1/2)	0.50 ± 0.1	3.3 ± 1.2	2.6 ± 0.4	1.4 ± 0.2
Cdl (×10^−5 Ω^−1 s^N cm^−2)	2.5	0.69	0.82	2.2

---

Electrochemical impedance spectroscopy was conducted at cathode potential of −0.1 V vs. Ag/AgCl electrode (Fig. 5C) to illustrate the resistance by biofouling. This potential was chosen because the maximum power of MFCs was obtained among the potential range between −0.06 V and −0.11 V based on electrode potential curve. From the Nyquist plot, the semicircle corresponded to the charge transfer resistance (R_ct).^29,30 The R_ct was the major kinetic limitation³¹ and increased after biofilm developed from Nyquist plots. The biofouled cathode had the maximum R_ct which was 2.4 times higher than that of the original, indicating the biofouling hindered the electron transfer and increased the polarization loss. When biofouling was removed from the surface, R_ct of Ext-cleaned cathode decreased by 23.1% from 119.3 ± 12.8 Ω to 91.7 ± 3.5 Ω (Table 1). With removing biofouling inside the catalyst layer, the blocked the three-phase boundaries were once “opened”. As a result, the conductivity of cathode increased and promoted electrons transfer rate. Therefore, R_ct of Int-cleaned cathode further decreased by 27.2% to 66.8 ± 3.0 Ω. The variation of dominant R_ct was consistent with LSV and exchange current density, which demonstrated the cathode kinetics of ORR was mainly limited by inner biofouling.

The double layer capacitance (C_dl) described the non-ideal capacitive property. Table 1 showed C_dl decreased after biofouling. The lowest value of C_dl was for biofouled cathode, which decreased by 72.4% compared to original cathode. It was reported that the capacitance was closely related to the pore structure.^32,33 For the biofouled cathode, the biofouling inside the catalyst layer inhibited the formation of electric double layer at the interfaces between reaction sites and electrolyte. In addition, the capacitance greatly depended on the internal resistance.³⁴ The biofouled cathode increased R_ct which indicated more difficult for charge transfer along cathode interfaces. The capacitance increased after cleaning surface biofouling, and almost recovered to the original value after further removing the internal biofouling. This was mainly because the blocked pores were “opened” again, the capability of electrons storage and release of catalyst layer was gradually recovered. Therefore, higher cathode performance was output.

Based on these results, the effect of biofouling distribution on cathode performance was proposed. The decreased oxygen permeability of cathode was attributed to both surface and internal biofouling. In contrast with surface biofouling, internal biofouling was mainly responsible for the lowered oxygen transfer coefficient because when oxygen diffused from ambient environment to bulk solution, the oxygen channels formed by hydrophobic PTFE and carbon aggregates were severely clogged as described in Fig. 6. The situation would be more deteriorative as a result of further extension of biofouling. Since cross-linked pore structure not only provided convenient channels for oxygen transport but also abundant reaction sites. Once pore structure of water-facing side was blocked, the exposure of these electrochemical activated sites decreased which caused depressed kinetic activity as depicted by declined exchange current density. In addition, biofouling attached on carbon skeleton impaired the conductivity and hindered the electron transport along three-phase interfaces which also enhanced the polarization resistance as EIS test showed. Therefore, biofouling inside the catalyst layer was the main reason for the impaired cathode activity.

Fig. 6 The schematics of transport process for biofouled and original cathode in MFCs.

The inevitably biofouling was confirmed because the cathode was directly exposed to the bulk solution. Further research should be focused on the effective methods to avoid the biofouling. The undesired biofouling was effectively reduced by incorporating catalyst with broad spectrum antibacterial chemicals like silver nanoparticles,²⁷ enrofloxacin²⁶ and vanillin³⁵ which could efficiently restrain the bacteria growth. In addition, controlling the hydrophobicity and charge property³⁶ of catalyst layer as well as applying separators³⁷ could also alleviate biofouling.

4 Conclusions

The entire biofouling of cathode impaired the maximum power density by 38%. Compared to the slight increment of cathode activity after removing surface biofouling, cathode performance with further cleaning inner biofouling increased rapidly as maximum power density increased by 30%. Compared to the surface biofouling, the accumulation of biofouling inside catalyst layer caused more oxygen scarcity along reaction sites due to the clogged pores. Meanwhile, the blocked porous structure of catalyst layer reduced kinetic activity and capacitance. Therefore, inner biofouling was the limiting factor to air cathode during long-term operation.

Acknowledgements

This work was supported by State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (Grant No. 2015DX05) and by the National Natural Science Fund for Distinguished Young Scholars (Grant No. 51125033) and National Natural Science Fund of China (Grant No. 51408156, 51209061). The authors also acknowledged the supports from the International Cooperating Project between China and European Countries (Grant No. 2014DFE90110) and Project funded by China Postdoctoral Science Foundation (2014T70355).

Notes and references

1. B. E. Logan, B. Hamelers, R. A. Rozendal, U. Schrorder, J. Keller, S. Freguia, P. Aelterman, W. Verstraete and K. Rabaey, Environ. Sci. Technol., 2006, 40, 5181–5192.
2. B. Min and B. E. Logan, Environ. Sci. Technol., 2004, 38, 5809–5814.
3. D. R. Lovley, Curr. Opin. Biotechnol., 2008, 19, 564–571.
4. Y. Fan, H. Hu and H. Liu, J. Power Sources, 2007, 171, 348–354.
5. H. Liu and B. E. Logan, Environ. Sci. Technol., 2004, 38, 4040–4046.
6. H. Rismani-Yazdi, S. M. Carver, A. D. Christy and I. H. Tuovinen, J. Power Sources, 2008, 180, 683–694.
7. J. C. Wei, P. Liang and X. Huang, Bioresour. Technol., 2011, 102, 9335–9344.
8. F. Zhang, S. A. Cheng, D. Pant, G. Van Bogaert and B. E. Logan, Electrochem. Commun., 2009, 11, 2177–2179.
9. K. J. Chae, M. Choi, F. F. Ajayi, W. Park, I. S. Chang and I. S. Kim, Energy Fuels, 2007, 22, 169–176.
10. J. Xu, G. P. Sheng, H. W. Luo, W. W. Li, L. F. Wang and H. Q. Yu, Water Res., 2012, 46, 1817–1824.
11. X. Y. Zhang, S. A. Cheng, X. Wang, X. Huang and B. E. Logan, Bioresour. Technol., 2009, 43, 8456–8461.
12. F. Zhang, D. Pant and B. E. Logan, Biosens. Bioelectron., 2011, 30, 49–55.
13. X. Y. Zhang, D. Pant, F. Zhang, J. Liu, W. H. He and B. E. Logan, ChemElectroChem, 2014, 1, 1859–1866.
14. H. Dong, H. Yu, X. Wang, Q. Zhou and J. Feng, Water Res., 2012, 46, 5777–5787.
15. D. Li, Y. Qu, J. Liu, W. He, H. Wang and Y. Feng, J. Power Sources, 2014, 272, 909–914.
16. J. Liu, J. Liu, W. He, Y. Qu, N. Ren and Y. Feng, J. Power Sources, 2014, 265, 391–396.
17. Y. Feng, Q. Yang, X. Wang and B. E. Logan, J. Power Sources, 2010, 195, 1841–1844.
18. S. A. Cheng, H. Liu and B. E. Logan, Environ. Sci. Technol., 2006, 40, 364–369.
19. D. R. Lovley and E. J. Phillips, Appl. Environ. Microbiol., 1988, 54, 1472–1480.
20. J. Liu, Y. Feng, X. Wang, X. Shi, Q. Yang, H. Lee, Z. Zhang and N. Ren, J. Power Sources, 2011, 196, 8409–8412.
21. F. Zhang, T. Saito, S. A. Cheng, M. A. Hickner and B. E. Logan, Environ. Sci. Technol., 2010, 44, 1490–1495.
22. D. A. Lowy, L. M. Tender, J. G. Zeikus, D. H. Park and D. R. Lovley, Biosens. Bioelectron., 2006, 21, 2058–2063.
23. X. Wang, N. S. J. Gao, Q. X. Zhou, H. Dong, H. B. Yu and Y. J. Feng, Bioresour. Technol., 2013, 144, 632–636.
24. S. Cheng, H. Liu and B. E. Logan, Electrochem. Commun., 2006, 8, 489–494.
25. R. A. Rozendal, H. V. M. Hamelers and C. J. N. Buisman, Environ. Sci. Technol., 2006, 40, 5206–5211.
26. W. Liu, S. Cheng, D. Sun, H. Huang, J. Chen and K. Cen, Biosens. Bioelectron., 2015, 72, 44–50.
27. J. An, H. Jeon, J. Lee and I. S. Chang, Environ. Sci. Technol., 2011, 45, 5441–5446.
28. X. Zhang, H. Sun, P. Liang, X. Huang, X. Chen and B. E. Logan, Biosens. Bioelectron., 2011, 30, 267–271.
29. K. Ding, H. Yang, Y. Cao, C. Zheng, L. Liu, L. Liu, Y. Wang, X. Yan and Z. Guo, Int. J. Electrochem. Sci., 2013, 8, 5343–5358.
30. Y. Wang, Q. He, J. Guo, J. Wang, Z. Luo, T. D. Shen, K. Ding, A. Khasanov, S. Wei and Z. Guo, ACS Appl. Mater. Interfaces, 2015, 7, 23920–23931.
31. N. Sekar and R. P. Ramasamy, J. Microb. Biochem. Technol., 2013, 6.
32. Z. Chen, K. Li, P. Zhang, L. Pu, X. Zhang and Z. Fu, Chem. Eng. J., 2015, 259, 820–826.
33. H.-R. Yu, S. Cho, M.-J. Jung and Y.-S. Lee, Microporous Mesoporous Mater., 2013, 172, 131–135.
34. A. Matsuda, H. Honjo, K. Hirata, M. Tatsumisago and T. Minami, J. Power Sources, 1999, 77, 12–16.
35. P. Chatterjee and M. Ghangrekar, Water Sci. Technol., 2014, 69.
36. C. Liu, D. Zhang, Y. He, X. Zhao and R. Bai, J. Membr. Sci., 2010, 346, 121–130.
37. V. J. Watson, T. Saito, M. A. Hickner and B. E. Logan, J. Power Sources, 2011, 196, 3009–3014.

---