Note
Performance of a Fe-N-C Catalyst in Single-chamber MFC Air-cathode at Neutral Media
2024 年 92 巻 2 号 p. 022016
詳細
2024 年 92 巻 2 号 p. 022016
Microbial fuel cells (MFC) are technologies that use microorganisms that transfer electrons to the anode, which flows to the cathode to find a final electron acceptor. Oxygen (O2) is the most widely used electron acceptor as it can diffuse through air-cathodes in single-chamber MFCs. However, microorganisms need neutral to slightly acid pH to survive, which is detrimental to the oxygen reduction reaction (ORR). Therefore, catalysts are needed at the air-cathodes to sustain a stable operation of single-chamber MFCs. Here, we report that the use of small amount (0.15 mg cm−2) of a Fe-N-C catalyst with carbon black in air-cathodes promote the ORR in neutral media and can sustain a stable MFC operation, keeping cell voltages of 0.3 V for 8 days.
Microbial fuel cells (MFC) are a promising technology as their potential applications are attractive to a circular economy. The coupling of wastewater treatment to energy production by this technology at a large scale is still challenging, and improvements are needed at different levels as the MFC involves different components such as an anode, a cathode, and an electrolyte.1,2 In MFC the microorganism function as a catalyst, which oxidize various substrates depending on the species. The electrons from the substrate oxidation are transferred to the anode, which later flow towards the cathode, where they meet an electron acceptor. Oxygen is widely used as electron acceptor due to its availability and high efficiency.1,3,4 Air-cathodes in single-chamber MFCs have been an alternative to reduce costs of MFC, due to the passive oxygen diffusion for the ORR and the elimination of the exchange membrane.4 However, as the anolyte and catholyte are no longer different within a single-chamber MFC, the air-cathode is exposed to the neutral or slightly acid pH media required for the optimal microbial survival and operation.1,4 The ORR activity is highly dependent on the electrolyte pH and it is optimal in alkaline electrolytes.5 Thus, the single-chamber MFC media has a negative impact in the ORR kinetics, slowing it down, so a catalyst is needed.6
Four main groups of electrocatalysts have been used at the cathode for MFC application: the platinum-group metal (PGM), transition metal-oxide-based catalysts, carbon-based metal-free, and transition metal-heteroatom-carbon catalysts. The PGM catalyst exhibits the best performance, either in acid or alkaline media, but it is not a scalable option.7 Pt is a scarce and expensive metal, the MFCs electrical power output is still low that would not justify its use,8 and it gets deactivated easily by poisoning due to CO or the attachment of other metabolites to the Pt centers.6 The carbon-based metal-free type can achieve high ORR activity by the functionalization of its surface by an acid oxidation process,9 or N-doping,10 but this varies depending on the pH.
The transition metal-heteroatom-carbon catalysts are the most promising group. The transition metals, like Fe, Mn, Ni, Cu, Co, etc., can be atomically dispersed, and the heteroatoms, such as N, S, P, B, etc., can be incorporated into carbon frameworks.11 One way of obtaining carbon-based materials is by the pyrolysis of polyacrylonitrile (PAN). PAN-based carbon has high content of nitrogen in its polymer, which makes them also an attractive alternative for commercialization of carbon-based catalysts.12 Within this metal-nitrogen-carbon (M-N-C) catalysts, the Fe-N-C catalysts have shown an attractive ORR catalytic activity.11,13 Some Fe-N-C catalysts exhibit comparable performances to Pt-based catalysts.14 However, the Fe-N-C catalysts are dependent on the electrolyte pH, exhibiting high ORR activity in alkaline media.15
The ORR at the cathode involves, depending on the catalyst, a 4e− pathway (EORR of 1.23 V vs. RHE), which is more efficient and produces water or hydroxide; or a 2e− pathway (EORR of 0.68 V vs. RHE) producing intermediates like hydrogen peroxides.11,13,16 Wang et al. (2007)17 demonstrated that the iron-based catalyst promotes the 4e− reduction pathway, whereas cobalt-based catalyst can promote both pathways. Cobalt, manganese, and nickel were proven to enhance a sequential pathway that involves 2e− that forms peroxide, followed by the final reduction of the peroxide with two additional electrons, making a 2e− + 2e− reduction pathway.7,18 In metal-nitrogen-doped carbon (NDC) with Fe molecules there is an enhancement of the direct 4e− reduction pathway mainly due to the presence of Fe-Nx sites.19 The specific origin of the ORR activity in Fe-N-C catalysts is still unclear but some attempts have been made to explain it, providing a model emphasizing the importance of the Fe-N4 sites and the hybridization between Fe (3dz2, 3dyz, 3dxz) and O2 (π*) orbitals.20 In spite of the debate about whether the pyridinic N or the pyrrolic N contributes to the ORR, the ORR mechanisms in alkaline and acid media have been explained.21 The Fe-N-C catalyst facilitates a pure 4e− reduction pathway in an alkaline environment. This is attributed to the stable adsorption of the peroxide intermediates, which can immediately be reduced and ensure a complete a 4e− pathway.13 On the contrary, in an acidic environment, the catalyst promotes a mixed reduction pathway involving both 4e− and 2e− mechanisms. Given the higher efficiency associated with the 4e− pathway, attributed to an EORR of 1.23 V vs. RHE, it is considered more desirable.
Most of the research done about carbon-based electrodes are trying to obtain more active surface areas and long term stability,22 especially in neutral media. In single-chamber MFCs, the cathode is in constant contact with the microorganism and some metabolites may attach onto the surface, which may decrease the active surface area and stability of the cathode. Moreover, the microorganisms can form a thick biofilm, which acts as a diffusion barrier for the ions and causes biofouling.9,23,24 In this study we analyzed the performance of a highly ORR active Fe-N-C catalyst produced from PAN facile pyrolysis process reported as a good cathode alternative for Polymer Electrolyte Fuel Cells (PEFCs).12 The Fe-N-C catalyst at small amount of 10 % in air-cathodes exhibit a stable performance in the operation of single-chamber MFC.
Shewanella oneidensis MR-1 (RIKEN, Japan) was used as the electroactive microorganism in this study. Initial cultures were prepared in Luria-Bertani broth (Miller) tubes that were grown for 20 hours under 30 °C and constant shaking. Erlenmeyer flasks were inoculated with the initial cultures and put in a water bath under 30 °C with constant shaking. After 20 hours the cells were collected by centrifugation at 3000 rpm in centrifuge tubes for 20 minutes. Later, the cells were washed three times with 0.1 M (= mol L−1) phosphate buffer (pH 7) containing 80 mM KCl. The final pellet was resuspended in the same solution and kept under refrigeration until it was inoculated into the MFC.
2.2 Preparation of air-cathodesSheets of carbon cloth (thickness = 1 mm, Nippon Carbon, Japan) were cut into 4 cm × 4 cm and coated on one side with a layer of polytetrafluoroethylene (PTFE), and the other side with the catalyst ink (Scheme 1). The Fe-N-C catalyst (CC) was kindly provided by Teijin Ltd. (Japan) for the purposes of this study.12 First, one side of each carbon cloth sheet was immersed in a PTFE 40 % dilution, which was prepared from 60 % PTFE (31-JR, Chemours-Mitsui Fluoroproducts Co., Ltd., Japan) diluted in water, and let dry for 30 minutes at room temperature (approximately 25 °C). The dried sheets were placed inside a furnace and left to dry at 370 °C for 15 minutes. The sheets were taken out of the furnace to cool down at room temperature. Once cooled down, they were dipped into the 40 % PTFE solution again, dried and cooled down the same way described. Later, the catalyst ink was prepared by suspending PTFE powder (24 mg, 6-J, Chemours-Mitsui Fluoroproducts Co., Ltd., Japan) in isopropanol (1.44 mL) and homogenized with a tip-type ultrasonicator (3.0 mmφ, UH-50, SMT Co., Ltd.) for about 3 minutes in water bath to avoid overheating (room temperature around 25 °C). Then, 24 mg of carbon black (KB, Ketjenblack EC-300J) and 2.4 mg of the Fe-N-C catalyst (CC)12 was added to the solution and homogenized with the ultrasonicator again for another 2 minutes in the water bath. The carbon cloth sheets were painted with this final ink using a brush (KB+CC-MFC). Another group of air-cathodes was prepared using the same method but containing only 24 mg KB without CC (KB-MFC), or 24 mg CC without KB (CC-MFC).
Schematic overview of the preparation of the air-cathode and its placement within the single-chamber MFC.
The air-cathodes (KB+CC and KB) were evaluated in single-chamber MFCs. Carbon felts (CF, 5 cm × 5 cm × 1 cm) were used as anodes and were previously cleaned with 0.5 M HCl for 5 minutes, 50 % ethanol for another 5 minutes, and sonicated for 5 minutes. Then, the CF were electrochemically polished in a three-electrode system using a 0.1 M KCl solution as the electrolyte, Ag|AgCl (Sat. KCl) as reference electrode, and Pt wire as the counter electrode. Later, the anodes and air-cathodes were placed inside the MFCs, and 0.1 M phosphate buffer (pH 7) with 80 mM KCl solution was poured in each MFC as a medium. The MFCs were placed inside an incubator set at 30 °C and 5.1 KΩ resistances were connected to each of them. All the MFCs systems were connected to eight channel potentiostat (SP-08, Toho Technical Research, Japan). Shewanella oneidensis MR-1 (>108 CFU mL−1) was added as an inoculum after the cell voltage became stable, and after 2 hours sodium lactate (50 %) was added to each reactor to make a final concentration of 5 mM. The power densities curves were obtained by linear sweep voltammetry at a scan rate of 10 mV s−1 when the cell voltages were stable.
2.4 Electrode characterizationThe performance of the catalyst was evaluated by rotating ring-disk electrode (RRDE) under Argon (Ar) or Oxygen (O2) saturated conditions. Briefly, the ink with and without the Fe-N-C catalyst was prepared by suspending PTFE powder (3 mg) in isopropanol (180 µL) and homogenized with a tip-type ultrasonicator for about 1 minute in water bath to avoid overheating (room temperature around 25 °C). For KB+CC, 3 mg of carbon black (KB, Ketjenblack EC-300J) and 0.3 mg of the Fe-N-C catalyst (CC) was added to the solution and homogenized for another 2 minutes in the water bath. In the case of KB or CC ink, only 3 mg of either KB or CC were added to the solution and homogenized the same way. 2 µL of the ink was dropped onto the RRDE disk, and after it dried at room temperature, 2 µL more was dropped and dried again. Cyclic voltammograms (CVs) and linear sweep voltammograms (LSVs) were taken using Ag|AgCl (Sat. KCl) as reference electrode, Pt wire as counter electrode, and the electrolyte was the same medium used for the MFC (without sodium lactate and Shewanella). The catalyst was also evaluated in air-cathodes within the single-chamber MFCs by cyclic voltammetry as working electrode, Ag|AgCl (Sat. KCl) as reference electrode, and Pt wire as counter electrode.
The electrocatalytic activity of KB and KB+CC was analyzed by RRDE. Figure 1, show the LSV curves under O2-saturated conditions at the ring (interrupted lines) and the disk (solid lines). At the ring, KB current density was 0.12 mA cm−2, higher than KB+CC and CC that were close to 0 mA cm−2. The high current density at the ring represents the oxidation of the intermediates that were generated from the ORR reaction, like H2O2 in the case of acidic media, or HO2− in the case of alkaline media.25 Thus, the presence of intermediates can also represent a trend towards a 2e− pathway. Leaning towards a 2e− pathway may result in a significant energy loss, which is led by a low electron transfer and more negative onset potential.26 The presence Fe-Nx sites in KB+CC may be contributing to a lower generation of intermediates, very similar to the Fe-N-C catalyst alone (CC), because of a dominant 4e− pathway. Thus, the addition of Fe-N-C catalyst to KB significantly reduced the generation of intermediates and improved the ORR.
LSV curves at disk (solid lines) or ring electrode (dashed lines) with the dried ink at the disk containing carbon black with Fe-N-C catalyst (KB+CC, red), carbon black (KB, black), or only the Fe-N-C catalyst (CC, blue). Rotation speed was 500 rpm, and the electrolyte was 0.1 M phosphate buffer with 80 mM KCl in oxygen (O2) saturated conditions. The catalyst loading was 0.58 mg cm−2. Scan rate was 2 mV s−1.
CC exhibit the highest limiting current density −3.0 mA cm−2 at −600 mV. KB+CC exhibit a current density of −1.67 mA cm−2, which also demonstrate a higher ORR activity due to the Fe-N-C catalyst compared to KB that exhibit −1.09 mA cm−2. Even though KB has electrocatalytic properties,27 it is significantly lower than the metal-based catalysts, but its functionalization by acid oxidation for example is a way other studies used to increase its electrocatalytic activity.9 In our case, the addition of a low amount (10 %) of a Fe-N-C catalyst to KB (KB+CC) significantly improved the current densities of KB. This improvement can be attributed to the Fe-Nx sites present with the addition of CC to KB. The Fe-Nx sites in Fe-N-C catalysts can promote a 4e− reduction pathway,19 thus the addition of CC to KB improved the current density by decreasing the generation of intermediates, as observed in the ring LSV curves. However, the CC current density was only 1.8-fold higher than KB+CC. This result does not correlate to the amount of the Fe-N-C catalyst added on each electrode, being 3 mg in CC and 0.3 mg in KB+CC. It is possible that the oxygen supply became rate-limiting in the conditions that the experiment was conducted due to the large amount of the catalyst, and the limit current reached its maximum. Based on these results it is possible that the application of Fe-N-C catalyst alone is not suitable for air-cathodes. Hence, the use of KB as a carbon support become relevant to overcome the rate-limiting conditions by increasing the surface area28 and porosity29 of the catalyst layer and improve the oxygen diffusion. It becomes crucial to find an ideal balance between the catalyst and carbon supports. In CC, the reduction current density at the disk started to increase as soon as the LSV started at a positive potential of 300 mV, from which KB+CC (230 mV) was very close. The onset potential for KB+CC aligns with earlier reported onset potentials of a catalyst ink containing PTFE as a catalyst binder, and 50 wt% unpyrolyzed iron phthalocyanine (FePc) supported on Ketjen black EC300J.30 However, the main difference was with KB that exhibit a negative potential of −26 mV. The positive overpotential of KB+CC and CC represent a faster and higher ORR activity over KB. The amount of the Fe-N-C catalyst (CC vs. KB+CC) mainly affect the maximum current density at −600 mV, but it does not significantly affect the onset potential. Even though KB+CC current density was lower than CC, it is significantly higher than the anode current density, which justifies its use in air-cathodes. Especially, the small amount of Fe-N-C catalyst used in KB+CC becomes a strong reason for its application in MFC.
In Fig. 2 the cyclic voltammograms of the air-cathodes containing carbon black with the Fe-N-C catalyst (KB+CC) or without (KB) exhibit differences compared to the RRDE results. The KB air-cathode onset potential switched to a more negative potential (−215 mV), while the KB+CC air-cathode onset potential remained very close to the onset potentials described in the RRDE results, around 200 mV. The oxygen diffusion at the air-cathodes occur in a passive way, while in RRDE the oxygen supply is constant to give saturated conditions. These results indicate that the Fe-N-C catalyst can perform the ORR even in passive diffusion conditions. The current density at −400 mV was −1.90 mA cm−2 and −2.74 mA cm−2 for KB and KB+CC, respectively. The poor electrocatalytic activity of KB towards the ORR is evident in these results. It is observed to be unstable compared to KB+CC. The results consistently shows that the addition of Fe-N-C catalyst to KB+CC improved the ORR activity, even at the small amount of 10 : 1, especially in the air-cathodes. In previous studies, the integration of carbon black with a catalyst (such as Pt) led to notable enhancements, including an increase in surface area, a reduction in charge transfer resistance, improved distribution of the catalyst, and promoted more contacts within the composite components.28 Furthermore, it was determined that the optimal ratio for the mixture of carbon black with activated carbon for example is 10 %.31 It is likely that KB+CC ORR activity was due to some of these improved characteristics.
Cyclic voltammogram of carbon black (KB-MFC) and Fe-N-C type (KB+CC-MFC) air-cathodes before the addition of Shewanella and lactate. The exposed electrode area to the liquid phase was 6.25 cm2. Scan rate 1 mV s−1.
The cell voltage of single-chamber MFC in Fig. 3 increased as soon as lactate was available for Shewanella, which started to oxidize it and transfer the electrons to the anode. The electron continues to the air-cathode, where it finally reduces the oxygen and forms water or hydroxide in a 4e− pathway. Although the cell voltage increased for both cases, the KB-MFC gave a maximum voltage of 0.25 V, while the KB+CC-MFC gave a maximum voltage of 0.39 V, the highest between them. By day 4, the cell voltage decreased as Shewanella consumed the lactate. By day 5, sodium lactate was added to the MFCs, and the voltage increased for both MFCs. However, the cell voltage in KB-MFC continued decreasing while in the KB+CC-MFC continued increasing until day 7. This difference shows a better stability of the KB+CC-MFC air-cathode than the KB-MFC air-cathode. Biofouling represents a problem during the operation of single-chamber MFC.23,24 Biofouling could be one of the reasons for the continued decrease of KB-MFC cell voltage. Upon the conclusion of our experiments, the removal of air-cathodes from the microbial fuel cells (MFCs) revealed the presence of a biofilm on both KB and KB+CC surfaces. Biofouling emerges as a significant challenge in single-chamber MFC operations, where microorganisms have the propensity to adhere to the cathode surface in contact with the liquid phase, forming a biofilm.23,24 This phenomenon may account for the observed sustained decline in cell voltage for KB-MFC and the gradual increase in cell voltage for KB+CC-MFC from day 5, contrasting the initial rise at the onset of MFC operation. Biofouling typically leads to an increase in internal resistance, a reduction in ORR activity, diminished active area, and pH imbalance at the cathode.32 The presence of this microbial layer or biofilm can impede the diffusion of H+ to the catalyst layer, depending on its thickness.32 Notably, removing the biofilm from the cathode surface has been shown to ameliorate the slow cell voltages and recover the initial values.33
Cell voltage of the MFCs with Fe-N-C catalyst air-cathode (KB+CC-MFC, red line) and carbon black air-cathode (KB-MFC, black line) with bare-CF as anodes. Arrows at approximately day 0 and day 5 indicates lactate addition to the MFCs medium.
In Fig. 4, KB+CC-MFC exhibits a maximum power density of 17.3 µW cm−2, at 41.6 µA cm−2. While KB-MFC maximum power density was 8.2 µA cm−2 at 30.6 µA cm−2. These results confirms that CC contributes to the power increase and improvement of MFC performance due to its catalytic activity to ORR. The catalyst in KB+CC play an important role in the reduction of oxygen, even at a ratio of 10 : 1, enabling a higher cell voltage (Fig. 3) and power density (Fig. 4). Although, the values were not as high as other studies,34 we consider that in our experiments the anode was also a limiting factor. The use of bare-CFs may not have promoted a high electron transfer, affecting in this case the power density. Further improvements at the anode may increase the power densities. Carbon cloth exhibit a lower resistivity compared to CFs, which makes it a better conductive current collector34 at the air-cathode. Some of the factors that can affect the durability of the Fe-N-C catalysts are the carbon oxidation, water flooding, the biofilm growth,35 salt precipitation (scaling), and local alkalization (inorganic fouling).32 By day 8, none of these factors represented a problem in the single-chamber MFCs.
Power density curves of the MFCs with the Fe-N-C catalyst air-cathode (KB+CC-MFC, red line) and carbon black air-cathode (KB-MFC, black line). The values were normalized to the air-cathode surface area.
Some of the characteristics that are considered when choosing an ORR catalyst are its durability, scalability, catalytic activity, and its cost-effectiveness.9 Especially in systems like MFC, a good performance at neutral pH is desirable.18 In this study, KB alone had higher overpotentials and lower catalytic activity compared to KB+CC in neutral media. The low performance of KB is explained by the generation of intermediates detected in the RRDE experiments. The amount of Fe-N-C catalyst does not affect the onset potential, but it significantly affects the current density. Still, the current density obtained by KB+CC with only 10 % of Fe-N-C catalyst was high enough for its application in air-cathodes. The ORR capacity of KB+CC was confirmed in single-chamber MFCs. The cell voltages and power densities improved using KB+CC air-cathode. Its stability during the MFC operation was confirmed for at least 8 days, demonstrating its usefulness in this kind of systems.
The authors gratefully acknowledge Teijin Ltd. (Japan) for providing the catalyst for this study. This work was supported by JSPS KAKENHI Grant Number 22K18912.
S.S-S. designed and performed the experiments, analyzed data, and drafted the manuscript. S.S. designed and performed the experiments, reviewed, and edited the manuscript S.T. designed, supervised, managed the project, reviewed, and edited the manuscript. All the authors contributed to the discussion of the results.
Silvia Sato-Soto: Investigation (Lead), Methodology (Lead), Visualization (Lead), Writing – original draft (Lead)
Shota Sato: Investigation (Equal), Writing – review & editing (Supporting)
Seiya Tsujimura: Conceptualization (Lead), Project administration (Lead), Supervision (Lead), Writing – review & editing (Lead)
The authors declare no conflict of interest.
Japan Society for the Promotion of Science: 22K18912
S. Tsujimura: ECSJ Active Member
