A novel and high performance activated carbon air-cathode with decreased volume density and catalyst layer invasion for microbial fuel cells

Abstract

To enhance the performance of activated carbon air cathodes for microbial fuel cells (MFCs), scattered and dense air-cathodes are fabricated by the rolling-press method with the gas diffusion layer or the catalyst layer (CL) invaded to stainless steel meshes (SSM). The maximum power density with scattered cathodes (2503 ± 61 mW m⁻²) is 29% higher than those of dense cathodes because of the decreased internal resistance and increased oxygen reduction reaction activity. The decrease in the electrode volume density and invasion of CL into SSM has been demonstrated to be an optimal structure, with the highest exchange current density up to 1.26 A m⁻² and the lowest charge transfer resistance as low as 1.5 Ω. The increase in performance is believed to result from the enhanced mass transport through extra pores and the contact of activated carbon to the current collector. This novel structure of an air-cathode has a promising future as an application in MFCs.

---

Microbial fuel cells (MFCs) are devices that use bacteria as catalysts to oxidize organic or inorganic matter and produce electrical energy.^1–5 Their ability to generate power from wastewater makes MFCs a promising and “green” approach for wastewater treatment,^6,7 as a lactate sensor and as a toxicity biosensor.⁸ The single chambered air-cathode system is a promising configuration of MFCs which could be scaled up for wastewater treatment.^9–11 Air-cathodes provide an efficient way for transferring oxygen to the cathode catalytic sites. Although the power output has been enhanced in recent years, it is still a long way from being used in real applications. In order to make the process cost-effective, the power density needs to be further increased.¹²

The power densities produced by MFCs are mainly limited by the cathodes.^13,14 The cathode design is challenging because of the relatively poor kinetics of the oxygen reduction reaction (ORR) under neutral pH conditions. The design of the gas diffusion layer (GDL) is essential for the gas transfer and water management, while the catalyst layer (CL) is crucial to ORR kinetics and three-phase interfaces (TPIs).¹⁵ Additionally, a well-designed CL can improve the uniform gas diffusion and favor the reaction processes.¹⁶

The ORR mechanism at the activated carbon (AC) is not well understood, especially in neutral pH and in the phosphate buffered solutions (PBS) used in many MFC studies.^17,18 Zhang et al. observed a good agreement between pyridinic-N content and ORR activity, indicating that pyridinic-N is the most active site for ORR in the nitrogen-doped AC.¹⁹ Both the total volume of pore areas and pore size distributions of the AC catalyst are important to its performance because they can establish more active ORR sites and provide more proton transfer channels.^20,21 Therefore, changing the pressure during the fabrication of cathodes by the rolling-press method can alter and control the volume densities of GDL and CL, thus influencing the total volume of pore areas and pore size distributions. In addition, the contact of CL to the current collector is another problem that may affect the electron transfer from AC to the external circuit. Furthermore, the heating of stainless steel meshes (SSMs) at 340 °C may cause corrosion during long-term operation. Here we firstly pressed the CL into the SSM followed by pressing the GDL on to top (Fig. S1†), which was a reverse procedure to that used in our previous report (pressing the GDL on first and following it by the CL). The changes in the procedure of making the cathodes as well as the densities of both CL and GDL were investigated in terms of electrochemical and MFC performance.

The air-cathodes consisted of a GDL and a CL with a SSM as the matrix. The GDL (0.5 mm) and the CL (0.5 mm) films were made using the rolling-press method according to procedures previously described.²² The GDL was made by rolling a mixture of carbon black (Jinqiushi Chemical Co. Ltd., Tianjin, China) and polytetrafluoroethylene (PTFE, 60%, Horizon, Shanghai, China) with a mass ratio of 3 : 7 into a film and the CL consisted of AC (1500 m² g⁻¹, Xinsen Carbon Co. Ltd., Fujian, China) and PTFE suspension with a AC/PTFE mass ratio of 6 : 1. Using the same GDL and CL, the volume density (Table S1†) was controlled by the final thickness of the whole electrode, either as dense (G-D; 0.8 mm) or scattered (G-S; 1.0 mm). For two GDL invaded cathodes, GDL was pressed in to a SSM and heated at 340 °C for 25 min. The CL was then pressed on to the opposite side to the 0.8 mm (G-D) or the 1.0 mm (G-S). Two other CL invaded cathodes were made by rolling CLs into SSM followed by rolling a heated GDL to the opposite side to final thickness of 0.8 (C-D) and 1.0 mm (C-S). The specific area of AC was 2161 m² g⁻¹ as described previously,²³ which contained 328 m² g⁻¹ of micropores (<2 nm) and 1932 m² g⁻¹ of mesopores (2 nm < pore diameter < 50 nm).

Single chambered membrane-less MFCs were operated in batch mode as previously described.²⁴ Anodes were acetone cleaned carbon fiber brushes in all the MFCs whereas the air-cathodes were the G-D, G-S, C-D and C-S (projected surface area of 7 cm²). MFCs were inoculated using the effluent from MFCs operated under similar conditions. The medium contained acetate (1.0 g L⁻¹) in 50 mM of PBS.²⁴ All the reactors were operated in a room at constant temperature (30 ± 1 °C) under 1000 Ω of external resistance and refreshed when voltages decreased to below 50 mV, forming one complete cycle.

Polarization curves were obtained as previously described.²⁴ Anode and cathode potentials were simultaneously recorded with Ag/AgCl as the reference electrode (+197 mV, 3.5 M KCl, vs. standard hydrogen electrode). Linear sweep voltammetry (LSV) and electrochemical impedance spectroscopy (EIS) were conducted using a potentiostat (Metrohm Autolab PGSTAT 302N, Switzerland), with the cathode as the working electrode, the platinum sheet as the counter electrode and an Ag/AgCl close to the cathode as the reference electrode. Abiotic cathodes were soaked in 50 mM PBS for 24 h before electrochemical tests and all the electrochemical tests were performed in the same reactor. For LSV tests, the cathode was operated over a potential range of −0.3 to 0.2 V at a rate of 0.1 mV s⁻¹. Tafel plots (log|current density|, A m⁻² versus |overpotential|, V) were recorded by sweeping the overpotential (|η|) from 0 to 100 mV at 0.1 mV s⁻¹, where η = 0 is the open circuit potential (OCP). EIS was conducted at the OCP over a frequency range of 100 kHz to 10 mHz with a sinusoidal perturbation of 10 mV amplitude. The Nyquist plots were simulated by ZSimpwin (ver. 3.10, Applied Princeton Research) with errors of fit of ≤10%. The equivalent circuit model (Fig. S2†) was R_s(Q(R_ctZ_w)) with the assumption that the cathode reaction was affected by both reaction kinetics and diffusion.

According to the polarization and power density curves shown in Fig. 1A, MFCs with C-S (1.06 g cm⁻³ GDL, 0.65 g cm⁻³ CL) achieved the maximum power density (MPD) of 2503 ± 61 mW m⁻² at 8.45 ± 0.10 A m⁻², which was comparable to that produced with G-S (1.04 g cm⁻³ GDL, 0.65 g cm⁻³ CL) of 2392 ± 34 mW m⁻² at 8.27 ± 0.06 A m⁻². The MPD of 1936 ± 17 mW m⁻² at 7.44 ± 0.03 A m⁻² was the lowest and obtained in G-D (1.46 g cm⁻³ GDL, 0.76 g cm⁻³ CL), which was comparable to that produced with C-D (1.57 g cm⁻³ GDL, 0.77 g cm⁻³ CL) of 2011 ± 53 mW m⁻² at 7.58 ± 0.10 A m⁻². The MPD of scattered cathodes was ∼25% higher than those of dense cathodes, demonstrating that the volume density of the GDL and CL of the AC air-cathode were vital to the cathodic performance. The anode potentials of four samples were similar, indicating that the differences in MPD were attributed to the cathodes but not anodes (Fig. 1B).

Fig. 1 Polarization and power density curves (A), electrode potentials (B) of the four different air-cathodes in MFCs (error bars ± SD based on the measurements of three duplicate reactors) and LSVs in abiotic reactors (C).

Based on the linear fitting correlation between the volume density of GDL or CL and the MPD (Fig. S3 and S4†), volume densities of GDLs (R² = 0.8834) were less interrelated with MPD than those of CLs (R² = 0.9371), showing that the volume density of CL was more important to cathodic performance than that of GDL.

LSV tests

MFCs with both scattered cathodes (G-S and C-S) exhibited higher ORR current densities than those of dense cathodes (G-D and C-D) in the whole scan range according to the LSVs shown in Fig. 1C. For example, at the potential of 0 V versus Ag/AgCl, the highest current density of 5.87 A m⁻² was achieved by G-S, with a value 3.3% higher than that of C-S (5.62 A m⁻²). It was 17% higher than those of G-D (4.66 A m⁻²) and C-D (4.67 A m⁻²). According to the Tafel plots (Fig. S5†), the corresponding exchange current densities (j₀) were determined over the linear region where the overpotential was from 60 to 80 mV. The j₀ decreased as follows: C-S (1.26 A m⁻²) >G-S (1.11 A m⁻²) >C-D (0.73 A m⁻²)> G-D (0.62 A m⁻²), which illustrated that the C-S cathode had the best ORR activity (Table S2†).

Scattered cathodes exhibited higher current densities and exchange current densities in abiotic LSVs and Tafel plots, which was consistent with the trend of power densities in MFCs. The possible reason for the improvement was that the scattered GDL and CL increased the total pore volume. These extra pores supplied more paths for the diffusion of oxygen and promoted the migration of electrolyte further into the scattered CL at the same time. So the scattered CL possibly produced more catalytic contacts and additional TPIs between the oxygen, protons and the electrons to achieve a better ORR performance.

The total internal resistances of the MFCs with four different air-cathodes are compared in Fig. 2. The highest total internal resistances was observed in G-D (26.8 Ω), and the resistances were decreased as follows: C-D (23.5 Ω) >G-S (21.7 Ω) >C-S (12.7 Ω). The total internal resistances of dense cathodes were higher than those of scattered cathodes, which was consistent with the polarization and power density data. Most of the differences in the total internal resistance were from charge transfer resistance because the ohmic resistance (R_s) was the same (∼10 Ω).

Fig. 2 Nyquist plots of four different air-cathodes at the OCP (A) and the total internal resistances of the four different air-cathodes (B).

As indicated in previous studies, the charge transfer resistances (R_ct) was the dominant internal resistance despite the ohmic resistance.²⁴ The decrease of CL and GDL volume density led to a decrease in R_ct, with values decreased by 87% from 11.4 Ω (C-D) to 1.5 Ω (C-S) for CL invaded cathodes and 37% from 13.3 Ω (G-D) to 8.4 Ω (G-S) for GDL invaded cathodes (Table S1†). Comparing the R_ct of G-S (8.4 Ω) and G-D (13.3 Ω) to those of C-S (1.5 Ω) and C-D (11.4 Ω), the invasion of CL into SSM prior to GDL decreased the R_ct by 82% and 14%, respectively. Although the R_ct of G-S (8.4 Ω) was 4.6 times larger than that of C-S (1.5 Ω), the MPDs of MFCs with C-S only increased by 4.7%, indicating that R_ct had a limited effect on the cathodic performance when the cathodes had a low volume density in the GDL and CL. Despite the difference in equipment and methods, the lowest R_ct of 1.5 Ω obtained in this study was lower than those reported recently at these OCPs (5–800 Ω).

A straight long line with an angle of approximately 45° shown in Fig. 2A is a characteristic of the semi-finite diffusion (Warburg impedance, Z_w) on the flat electrode. The diffusion resistance R_d was obtained from Z_w as previously described.²⁵ The R_d was not affected by the volume density until CL was invaded into SSM (Fig. 2B). It was decreased by 62% from 1.3 Ω (C-D) to 0.5 Ω (C-S) when the scattered cathode was applied. The R_d of C-S was even 82% lower than those of GDL invaded cathodes (2.75 ± 0.05 Ω), indicating that both the invasion of CL and the decrease of electrode volume density had a solid contribution to accelerating the mass transport of AC air-cathodes.

The decrease of volume density and the invasion of catalyst layer to the current collector effectively increased the performance of an AC air-cathode, with a maximum power density of 2503 ± 61 mW m⁻² and the lowest charge transfer resistance of 1.5 Ω. The correlation between the volume density of GDL or CL and MPD was determined and the results showed there was a higher linear fitting correlation between the volume densities of CL and MPD, indicating that CL had a significant influence on the performance of the cathodes. Optimizing the ORR interfaces of the AC air-cathode and increasing the total pore volume of the GDL and CL were good strategies to enhance cathode performance in MFCs without heating the current collector.

Acknowledgements

This research was supported by the MOE Innovative Research Team in University (IRT13024), the National Natural Science Foundation of China (no. 21107053 and 21037002), the Ministry of Science and Technology as an 863 major project (2013AA06A205), the Beijing Green Future Environment Foundation and the Tianjin Research Program of Application Foundation and Advanced Technology (13JCQNJC08000).

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. X. Wang, Y. Feng, H. Wang, Y. Qu, Y. Yu, N. Ren, N. Li, E. Wang, H. Lee and B. E. Logan, Environ. Sci. Technol., 2009, 43, 6088–6093.
3. D. R. Lovley, Nat. Rev. Microbiol., 2006, 4, 497–508.
4. K. Rabaey and W. Verstraete, Trends Biotechnol., 2005, 23, 291–298.
5. H. Wang and Z. J. Ren, Biotechnol. Adv., 2013, 31, 1796–1807.
6. B. Logan, S. Cheng, V. Watson and G. Estadt, Environ. Sci. Technol., 2007, 41, 3341–3346.
7. S. Cheng and B. E. Logan, Electrochem. Commun., 2007, 9, 492–496.
8. X. Wang, N. Gao and Q. Zhou, Biosens. Bioelectron., 2013, 43, 264–267.
9. H. Liu and B. E. Logan, Environ. Sci. Technol., 2004, 38, 4040–4046.
10. D. H. Park and J. G. Zeikus, Biotechnol. Bioeng., 2003, 81, 348–355.
11. P. Liang, J. Wei, M. Li and X. Huang, Front. Environ. Sci. Eng., 2013, 7, 913–919.
12. A. Dekker, A. T. Heijne, M. Saakes, H. V. Hamelers and C. J. Buisman, Environ. Sci. Technol., 2009, 43, 9038–9042.
13. S. Cheng, H. Liu and B. E. Logan, Environ. Sci. Technol., 2006, 40, 364–369.
14. B. H. Kim, I. S. Chang and G. M. Gadd, Appl. Microbiol. Biotechnol., 2007, 76, 485–494.
15. X. Wang, S. Cheng, Y. Feng, M. D. Merrill, T. Saito and B. E. Logan, Environ. Sci. Technol., 2009, 43, 6870–6874.
16. H. Wang, Z. Wu, A. Plaseied, P. Jenkins, L. Simpson, C. Engtrakul and Z. Ren, J. Power Sources, 2011, 196, 7465–7469.
17. V. J. Watson, C. Nieto Delgado and B. E. Logan, Environ. Sci. Technol., 2013, 47, 6704–6710.
18. F. Zhang, S. A. Cheng, D. Pant, G. Van Bogaert and B. E. Logan, Electrochem. Commun., 2009, 11, 2177–2179.
19. B. Zhang, Z. Wen, S. Ci, S. Mao, J. Chen and Z. He, ACS Appl. Mater. Interfaces, 2014, 6, 7464–7470.
20. H. Dong, H. Yu and X. Wang, Environ. Sci. Technol., 2012, 46, 13009–13015.
21. H. Dong, H. Yu, H. Yu, N. Gao and X. Wang, J. Power Sources, 2013, 232, 132–138.
22. H. Dong, H. Yu, X. Wang, Q. Zhou and J. Feng, Water Res., 2012, 46, 5777–5787.
23. X. Wang, C. Feng, N. Ding, Q. Zhang, N. Li, X. Li, Y. Zhang and Q. Zhou, Environ. Sci. Technol., 2014, 48, 4191–4198.
24. X. Li, X. Wang, Y. Zhang, N. Ding and Q. Zhou, Appl. Energy, 2014, 123, 13–18.
25. F. Zhang, M. D. Merrill, J. C. Tokash, T. Saito, S. Cheng, M. A. Hickner and B. E. Logan, J. Power Sources, 2011, 196, 1097–1102.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra07078j

---