- Email: [email protected]
Mitigation of the effect of catholyte contamination in microbial fuel cells using a wicking air cathode
Biosensors and Bioelectronics 24 (2009) 3144–3147
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Short communication
Mitigation of the effect of catholyte contamination in microbial fuel cells using a wicking air cathode Christian J. Sund, Michael S. Wong, James J. Sumner ∗ US Army Research Laboratory, Sensors and Electron Devices Directorate, Adelphi, MD 20783, USA
a r t i c l e
i n f o
Article history: Received 4 February 2009 Received in revised form 11 March 2009 Accepted 12 March 2009 Available online 24 March 2009 Keywords: Microbial fuel cell Cathode Microbial contamination Current generation
a b s t r a c t Cathode design greatly affects microbial fuel cell (MFC) performance. Cathode contamination is inevitable in a single-chamber MFC yet it is impossible to study the magnitude of this effect in the single-chambered format. Therefore to study the effect of contamination at the cathode two-chamber MFCs must be used. The advantages of the two-chamber MFC design used in this study include: the assembled and filled fuel cell is autoclavable and the cathode can easily be moved from the submerged to air exposed position while maintaining sterility. This study was performed with the cathode in two positions: completely submerged in the catholyte and raised to a point where wicking action was used to coat the cathode with catholyte. When the cathode was submerged and the catholyte was inoculated with Bacillus megaterium, Shewanella oneidensis or Escherichia coli current generation was greatly decreased as compared to sterile. When the cathodes were raised, allowing contact with the catholyte by wicking, the current rose to levels comparable with sterile cathode MFCs. The reduced performance of submerged cathodes is most likely due to the microbial culture in the cathode greatly reducing the available oxygen for completion of the cathode reaction. This shows simple designs with low-cost materials can be used to mitigate effects of cathode contamination. Published by Elsevier B.V.
1. Introduction Microbial fuel cells are biological fuel cells that utilize microorganisms to degrade organic molecules while producing an electrical current (Logan and Regan, 2006; Kim et al., 2007; Bullen et al., 2006). MFCs can use a wide variety of organics from small molecules such as acetate and sugars to large biopolymers such as cellulose as substrates. This ability makes MFCs an attractive option for future applications in environmentally friendly electricity generation, waste remediation, and biological oxygen demand sensors as previously reviewed (Logan and Regan, 2006; Kim et al., 2007; Bullen et al., 2006). The cathode is a significant performance-limiting component in MFCs and low power densities make use of traditional fuel cell catalysts such as platinum cost-prohibitive. Thus there have been many cathode designs that try to optimize cathode performance with relatively inexpensive materials (Logan and Regan, 2006; Tender et al., 2002; Chaudhuri and Lovley, 2003; Ieropoulos et al., 2005; Min and Logan, 2004; Gil et al., 2003; Park and Zeikus, 2003; Rabaey et al., 2005; He et al., 2005, 2006; Ringeisen et al., 2006; Biffinger et al., 2007). Protons and oxygen are reduced at the cathode with electrons from the MFC circuit. Therefore, oxygen availability can
∗ Corresponding author. Tel.: +1 301 394 0252; fax: +1 301 394 0310. E-mail address: [email protected] (J.J. Sumner). 0956-5663/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.bios.2009.03.019
be the limiting factor of cathode performance. Cathodes that are submerged in a catholyte which is bubbled with air to maintain oxygen saturation have been reported however this method requires energy to aerate the cathode which may result in a net energy loss of the system (Bergel et al., 2005; Liu and Logan, 2004). Another common design of air cathodes involve the use of wet-proofed carbon paper in which one side is exposed to the catholyte while the other is exposed to air (Liu and Logan, 2004; Zhang et al., 2007; Cheng and Logan, 2007; Zhao et al., 2006; Cheng et al., 2006a,b; Liu et al., 2004; Fan et al., 2007). This design allows air to contact the cathode in a passive manner increasing the overall efficiency of the system however, these cathodes are costly when compared to untreated graphite. Upflow MFCs have also been constructed using untreated graphite cathodes where the catholyte flowing over the cathode presumably allows for high catholyte oxygen concentrations (Rabaey et al., 2005; Clauwaert et al., 2007). MFC cathodes are often subjected to conditions that are detrimental to their performance and are subject to bio-fouling (Logan et al., 2007) which can come from multiple sources. Many MFC designs are difficult to sterilize for experimental purposes and in real world applications are subject to contamination directly either on the cathode or in the cathode compartment. There are some reports that indicate that cellular growth on the cathode can increase MFC performance by the organisms acting as an electron acceptor, however this could be highly organism-specific (Freguia et al., 2008; Chen et al., 2008; Rabaey et al., 2008; Rabaey and Keller,
C.J. Sund et al. / Biosensors and Bioelectronics 24 (2009) 3144–3147
3145
2008). Since contamination of cathodes could greatly affect MFC performance we investigated the effect of cathode contamination on current generation in two-chamber microbial fuel cells utilizing three organisms, Bacillus megaterium, Shewanella oneidensis and Escherichia coli as contaminants. The advantage of the two-chamber MFC design used in this study is that the assembled and filled fuel cell is autoclavable allowing sterility to be maintained. While this design is not optimized for power production, we are able to use this system to study effects such as cathode contamination without directly interfering with processes at the anode, which is not practical in more power dense single-chambered designs. 2. Experimental 2.1. Biological materials and reagents S. oneidensis MR-1 ATCC BAA-1096, E. coli MG1655 ATCC 47076 and B. megaterium (Raven Laboratories, Omaha, NE, USA) were used in all studies and were cultured in trypticase soy broth (TSB) (Becton Dickinson, Sparks, MD, USA) in an aerobic 30 ◦ C incubator shaking at 250 RPM. S. oneidensis and E. coli were maintained as frozen stocks (−80 ◦ C) in TSB containing 30% glycerol. B. megaterium was maintained as spores in 40% ethanol. In select fuel cells, the redox indicator resazurin was added to final concentration of 80 M to determine if the catholyte was oxidized. 2.2. MFC design/data collection Two-chamber MFCs were constructed in a similar fashion as previously described (Sund et al., 2007; Milliken and May, 2007) with some modifications and pictured in Scheme 1. The anodes consisted of graphite rods (Ultra Carbon, Bay City, MI, USA), 0.6 cm in diameter, cut to a length of 2.5 cm and placed in direct contact with 4 g of lump graphite (Goodfellow Cambridge Limited, Huntingdon, England). The cathodes consisted of graphite rods as described above; however, no graphite lump was added to the cathode compartment.
Scheme 1. Drawing of MFC design constructed of glass 90◦ elbow, ball-and-socket joints. The anode and cathode are separated by a Nafion 117 proton exchange membrane. The cathode is pictured in the two arrangements, submerged and wicking air. The dotted line represents the electrolyte height. The lump graphite, septa on the anode, and glass cover on the cathode are not pictured because of clarity.
In both cases, the graphite rods were attached to wires with silver epoxy (H20E, Epoxy Technology, Billerica, MA, USA) which was later insulated by coating with Devcon® 5 min fast drying epoxy (ITW Performance Polymers, Riveria Beach, FL, USA) leaving 1.25 cm of graphite exposed. A new piece of Nafion® 117 (DuPont, Wilmington, DE, USA) was used for each fuel cell. Anode wires were passed through silicone rubber septa used to seal the anode compartment following inoculation while the cathode compartment was covered with a loose-fitting glass vial. Each fuel cell compartment was filled with 20 mL of media and autoclaved. The anode compartment and in some cases the cathode compartment of the MFCs were inoculated with 200 L of overnight cultures. MFCs were incubated in an aerobic 30 ◦ C incubator and the potential across a 10 k resistor was measured and recorded every 10 s via a DAQPad-6016 and a custom LabView® VI (National Instruments, Austin, TX, USA). Open circuit potentials of the cells were recorded using the same National Instruments hardware and custom software while AC impedance
Fig. 1. Potential traces of MFCs inoculated with S. oneidensis in anode compartment. Initially, the cathodes were placed just below the surface of the catholyte which was either sterile (A) or inoculated with S. oneidensis (B), B. megaterium (C), or E. coli (D). The control (E) was sterile in both the cathode and anode compartments. After four days the cathodes where repositioned so that only the tip of the cathode was submerged in the catholyte. The addition of resazurin did not affect the voltage which was demonstrated with a set of fuel cells comparing sterile and inoculated cathodes with 80 M resazurin added to both the anode and the cathode compartments. An example is shown in (F) where the cathode was inoculated with B. megaterium and can be directly compared with (C).
3146
C.J. Sund et al. / Biosensors and Bioelectronics 24 (2009) 3144–3147
Table 1 Average and standard deviation of the open circuit potentials in volts and impedance in ohms for each condition of microbial fuel cells. Cell description (anode–cathode)
OCP (V) (cathode up)
OCP (V) (cathode down)
Impedance () (cathode up)
Impedance () (cathode down)
Sterile–sterile S. oneidensis–sterile S. oneidensis–S. oneidensis S. oneidensis–E. coli S. oneidensis–B. megaterium
0.05 (±0.02) 0.43 (±0.05) 0.29 (±0.05) 0.43 (±0.07) 0.45 (±0.01)
0.08 (±0.03) 0.41 (±0.05) 0.06 (±0.06) 0.19 (±0.05) 0.32 (±0.02)
500 (±46) 450 (±10) 410 (±35) 410 (±29) 360 (±40)
330 (±12) 350 (±10) 330 (±43) 350 (±25) 300 (±31)
measurements were obtained with amplitude of 5 mV from 10,000 to 1 Hz using a Model 660a Electrochemical Workstation from CH Instruments (Austin, TX). 3. Results and discussion 3.1. Potential versus time traces for the MFCs MFCs were constructed and the anode compartments were inoculated with S. oneidensis with the exception of the control MFCs. Initially, the cathodes were completely submerged below the surface of the catholyte as seen in Scheme 1. Between 24 and 96 h the potential of the MFCs which contained S. oneidensis in the anode compartments and sterile cathode compartments had reached a potential of approximately 60 mV, as seen in the three replicate potential versus time traces in Fig. 1A. All of the MFCs in which the cathode compartment had been inoculated had much lower potentials, as seen in Fig. 1B–D, which were similar to background levels. The limited potentials seen in the MFCs with inoculated catholyte were most likely due to the reduction of oxygen via microbial metabolism. After approximately 96 h, the cathodes were raised to a point where only the tip of the cathode maintained contact with the catholyte. This allowed the catholyte to wick up the sides of the cathode which, in effect, created a novel air cathode design. When the cathodes were repositioned, there was an increase in potential of all of the MFCs (as shown in Fig. 1) however this increase was less dramatic in the controls where the cathode was not inoculated (as shown in Fig. 1A). Most notable were the large potential increases seen in the MFCs with inoculated cathode compartments. Inoculation of the anode and cathode were performed at the same time. Since the organisms added to the cathode were either facultative or aerobic and the cathode compartments were open to air these cultures reproduce more rapidly than Shewanella in the anaerobic anode compartments. Therefore, the performance is low at inoculation, may rise slightly but, is then quenched until the cathode is raised to the wicking position. 3.2. Characterization of MFCs via OCP and impedance Open circuit potentials (OCPs) of the cells and total cell impedance data were collected on each series of MFCs and are displayed in Table 1. The control fuel cells performed as expected where the OCP of the sterile–sterile (negative control) fuel cells show essentially no electrochemical activity and the S. oneidensis–sterile (positive control) had significant OCPs. In both cases, the OCP did not change significantly whether the cathode was in the wicking or in the submerged position. Contaminated cathodes that are submerged always showed significantly lower potentials than the positive control cathodes. When these cathodes were raised to the wicking position the effects of contamination were mitigated as exemplified by both the plots in Fig. 1 and the OCPs in Table 1, where the potentials returned to within error of the sterile cathode MFCs. The only case where OCP decreased significantly for contaminated cathodes in the wicking position
over the sterile cathode was in fuel cells inoculated with S. oneidensis in both the anode and the cathode. This is logical as those two compartments would have been the most chemically identical based on similar metabolic processes, especially after oxygen was consumed in the catholyte, which would have decreased the potential difference between the compartments. The differences in total cell impedance between submerged and wicking cathodes were only significant in the control fuel cells. Total impedance takes several factors into account including: membrane resistance, electrolyte resistance and contact impedances. The contamination in the cathodes probably decreased the electrolyte resistance and possibly the electrode–electrolyte contact impedance; however, the change observed in all of the fuel cells was small enough to have little effect on overall performance. The increase in potential seen in MFCs with wicking cathodes was most likely due to an increase of oxygen availability as oxygen saturation at the cathode has been shown to greatly enhance MFC performance (Rabaey et al., 2008). The wicking cathodes reduced the microorganisms’ ability to scavenge all of the oxygen prior to reaching cathode suggesting that this design may be beneficial in mitigating the effects of cathode contamination. In order to determine the redox status of the catholyte, as an indication of oxygen availability, identical MFCs to those in Fig. 1A–E were assembled with the redox indicator, resazurin, added to both the anode and the cathode compartments (an example of this data is shown in Fig. 1F). One of the fuel cells was inoculated with E. coli and the other is sterile. The pink color of resazurin in the sterile catholyte indicate that it was more oxidized when compared to the inoculated catholyte which lacked the pink hue, indicating that the resazurin was in a reduced state. An image of the resazurin containing cathode compartment of two fuel cells, one inoculated with E. coli (A) and the other sterile (B) has been included in the Supplementary Material. The oxidation state of the resazurin could have been effected by cellular metabolism however it is reversible and therefore if any oxygen was present in the cathode it would have reoxidized the resazurin. This strongly suggests that there was a high concentration of oxygen in the sterile cathodes while the inoculated cathodes were oxygen poor. The lower redox potential of the contaminated catholyte due to microbial oxygen reduction was most likely responsible for the decrease in MFC potentials seen in Fig. 1 prior to repositioning the cathodes. Repositioning of the cathodes allowed the catholyte to wick up the sides of the cathode which in effect created a novel air cathode design. 3.3. Redox indicator results A previous report has indicated that resazurin can act as a mediator between the bacteria and the anode in MFCs (Sund et al., 2007). In this study, resazurin was added at the same concentration in both the anode and cathode compartments preventing discharge due to chemical potential difference. Also, there was no effect on potential traces in the presence of resazurin (as shown in Fig. 1C and F) indicating that either the MFCs were cathode limited or the S. oneidensis cannot utilize the resazurin as a mediator.
C.J. Sund et al. / Biosensors and Bioelectronics 24 (2009) 3144–3147
4. Conclusions Since most proposed applications of MFCs require that they be located in environments where the cathodes are likely to become contaminated, the effects of contamination on MFC design can be significant. If microorganisms consume oxygen in the catholyte before it reaches the cathode then the MFCs’ performance will be severely limited. Current solutions to this problem are to bubble the catholyte with air, utilize wet-proofed air cathodes or flowing designs where the catholyte flows over the cathode. Since bubbling the catholyte can be energy intensive, the air cathodes and flowing designs are more favorable because they will be less affected by contamination and can be passively exposed to air. Data presented here provide further evidence that cathode design needs to be taken into account. Specifically, designs must take into account oxygen availability in environments that contain oxygen-reducing microorganisms. We have also demonstrated a simple low-cost air cathode design which relies on the wicking action of the catholyte to coat the cathode. Finally, since cathode contamination can dramatically alter MFC characteristics, this must be considered when comparing MFC designs and performance. Acknowledgement The authors thank Prof. Scott Crittenden for helpful discussions in this research. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2009.03.019. References Bergel, A., Feron, D., Mollica, A., 2005. Electrochemistry Communications 7 (9), 900–904. Biffinger, J.C., Ray, R., Little, B., Ringeisen, B.R., 2007. Environmental Science and Technology 41 (4), 1444–1449. Bullen, R.A., Arnot, T.C., Lakeman, J.B., Walsh, F.C., 2006. Biosensors and Bioelectronics 21 (11), 2015–2045.
3147
Chaudhuri, S.K., Lovley, D.R., 2003. Nature Biotechnology 21 (10), 1229–1232. Chen, G.W., Choi, S.J., Lee, T.H., Lee, G.Y., Cha, J.H., Kim, C.W., 2008. Water Research 42 (6–7), 1387–1396. Cheng, S., Liu, H., Logan, B.E., 2006a. Electrochemistry Communications 8 (3), 489–494. Cheng, S., Liu, H., Logan, B.E., 2006b. Environmental Science and Technology 40 (1), 364–369. Cheng, S., Logan, B.E., 2007. Electrochemistry Communications 9 (3), 492–496. Clauwaert, P., Van der Ha, D., Boon, N., Verbeken, K., Verhaege, M., Rabaey, K., Verstraete, W., 2007. Environmental Science and Technology 41 (21), 7564–7569. Fan, Y., Hu, H., Liu, H., 2007. Journal of Power Sources 171 (2), 348–354. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2008. Water Research 42 (6–7), 1387–1396. Gil, G.C., Chang, I.S., Kim, B.H., Kim, M., Jang, J.K., Park, H.S., Kim, H.J., 2003. Biosensors and Bioelectronics 18 (4), 327–334. He, Z., Wagner, N., Minteer, S.D., Angenent, L.T., 2006. Environmental Science and Technology 40 (17), 5212–5217. He, Z., Minteer, S.D., Angenent, L.T., 2005. Environmental Science and Technology 39 (14), 5262–5267. Ieropoulos, I.A., Greenman, J., Melhuish, C., Hart, J., 2005. Enzyme and Microbial Technology 37 (2), 238–245. Kim, B.H., Chang, I.S., Gadd, G.M., 2007. Applied Microbiology and Biotechnology 76 (3), 485–494. Liu, H., Logan, B.E., 2004. Environmental Science and Technology 38 (14), 4040–4046. Liu, H., Ramnarayanan, R., Logan, B.E., 2004. Environmental Science and Technology 38 (7), 2281–2285. Logan, B., Cheng, S., Watson, V., Estadt, G., 2007. Environmental Science and Technology 41 (9), 3341–3346. Logan, B.E., Regan, J.M., 2006. Environmental Science and Technology 40 (17), 5172–5180. Milliken, C.E., May, H.D., 2007. Applied Microbiology and Biotechnology 73 (5), 1180–1189. Min, B., Logan, B.E., 2004. Environmental Science and Technology 38 (21), 5809–5814. Park, D.H., Zeikus, J.G., 2003. Biotechnology and Bioengineering 81 (3), 348–355. Rabaey, K., Clauwaert, P., Aelterman, P., Verstraete, W., 2005. Environmental Science and Technology 39 (20), 8077–8082. Rabaey, K., Keller, J., 2008. Water Science and Technology 57 (5), 655–659. Rabaey, K., Read, S.T., Clauwaert, P., Freguia, S., Bond, P.L., Blackall, L.L., Keller, J., 2008. ISME Journal 2 (5), 519–527. Ringeisen, B.R., Henderson, E., Wu, P.K., Pietron, J., Ray, R., Little, B., Biffinger, J.C., Jones-Meehan, J.M., 2006. Environmental Science and Technology 40 (8), 2629–2634. Sund, C.J., McMasters, S., Crittenden, S.R., Harrell, L.E., Sumner, J.J., 2007. Applied Microbiology and Biotechnology 76 (3), 561–568. Tender, L.M., Reimers, C.E., Stecher, H.A., Holmes, D.E., Bond, D.R., Lowy, D.A., Pilobello, K., Fertig, S.J., Lovley, D.R., 2002. Nature Biotechnology 20 (8), 821–825. Zhang, T., Zeng, Y., Chen, S., Ai, X., Yang, H., 2007. Electrochemistry Communications 9 (3), 349–353. Zhao, F., Harnisch, F., Schroeder, U., Scholz, F., Bogdanoff, P., Herrmann, I., 2006. Environmental Science and Technology 40 (17), 5193–5199.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Short communication
Mitigation of the effect of catholyte contamination in microbial fuel cells using a wicking air cathode Christian J. Sund, Michael S. Wong, James J. Sumner ∗ US Army Research Laboratory, Sensors and Electron Devices Directorate, Adelphi, MD 20783, USA
a r t i c l e
i n f o
Article history: Received 4 February 2009 Received in revised form 11 March 2009 Accepted 12 March 2009 Available online 24 March 2009 Keywords: Microbial fuel cell Cathode Microbial contamination Current generation
a b s t r a c t Cathode design greatly affects microbial fuel cell (MFC) performance. Cathode contamination is inevitable in a single-chamber MFC yet it is impossible to study the magnitude of this effect in the single-chambered format. Therefore to study the effect of contamination at the cathode two-chamber MFCs must be used. The advantages of the two-chamber MFC design used in this study include: the assembled and filled fuel cell is autoclavable and the cathode can easily be moved from the submerged to air exposed position while maintaining sterility. This study was performed with the cathode in two positions: completely submerged in the catholyte and raised to a point where wicking action was used to coat the cathode with catholyte. When the cathode was submerged and the catholyte was inoculated with Bacillus megaterium, Shewanella oneidensis or Escherichia coli current generation was greatly decreased as compared to sterile. When the cathodes were raised, allowing contact with the catholyte by wicking, the current rose to levels comparable with sterile cathode MFCs. The reduced performance of submerged cathodes is most likely due to the microbial culture in the cathode greatly reducing the available oxygen for completion of the cathode reaction. This shows simple designs with low-cost materials can be used to mitigate effects of cathode contamination. Published by Elsevier B.V.
1. Introduction Microbial fuel cells are biological fuel cells that utilize microorganisms to degrade organic molecules while producing an electrical current (Logan and Regan, 2006; Kim et al., 2007; Bullen et al., 2006). MFCs can use a wide variety of organics from small molecules such as acetate and sugars to large biopolymers such as cellulose as substrates. This ability makes MFCs an attractive option for future applications in environmentally friendly electricity generation, waste remediation, and biological oxygen demand sensors as previously reviewed (Logan and Regan, 2006; Kim et al., 2007; Bullen et al., 2006). The cathode is a significant performance-limiting component in MFCs and low power densities make use of traditional fuel cell catalysts such as platinum cost-prohibitive. Thus there have been many cathode designs that try to optimize cathode performance with relatively inexpensive materials (Logan and Regan, 2006; Tender et al., 2002; Chaudhuri and Lovley, 2003; Ieropoulos et al., 2005; Min and Logan, 2004; Gil et al., 2003; Park and Zeikus, 2003; Rabaey et al., 2005; He et al., 2005, 2006; Ringeisen et al., 2006; Biffinger et al., 2007). Protons and oxygen are reduced at the cathode with electrons from the MFC circuit. Therefore, oxygen availability can
∗ Corresponding author. Tel.: +1 301 394 0252; fax: +1 301 394 0310. E-mail address: [email protected] (J.J. Sumner). 0956-5663/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.bios.2009.03.019
be the limiting factor of cathode performance. Cathodes that are submerged in a catholyte which is bubbled with air to maintain oxygen saturation have been reported however this method requires energy to aerate the cathode which may result in a net energy loss of the system (Bergel et al., 2005; Liu and Logan, 2004). Another common design of air cathodes involve the use of wet-proofed carbon paper in which one side is exposed to the catholyte while the other is exposed to air (Liu and Logan, 2004; Zhang et al., 2007; Cheng and Logan, 2007; Zhao et al., 2006; Cheng et al., 2006a,b; Liu et al., 2004; Fan et al., 2007). This design allows air to contact the cathode in a passive manner increasing the overall efficiency of the system however, these cathodes are costly when compared to untreated graphite. Upflow MFCs have also been constructed using untreated graphite cathodes where the catholyte flowing over the cathode presumably allows for high catholyte oxygen concentrations (Rabaey et al., 2005; Clauwaert et al., 2007). MFC cathodes are often subjected to conditions that are detrimental to their performance and are subject to bio-fouling (Logan et al., 2007) which can come from multiple sources. Many MFC designs are difficult to sterilize for experimental purposes and in real world applications are subject to contamination directly either on the cathode or in the cathode compartment. There are some reports that indicate that cellular growth on the cathode can increase MFC performance by the organisms acting as an electron acceptor, however this could be highly organism-specific (Freguia et al., 2008; Chen et al., 2008; Rabaey et al., 2008; Rabaey and Keller,
C.J. Sund et al. / Biosensors and Bioelectronics 24 (2009) 3144–3147
3145
2008). Since contamination of cathodes could greatly affect MFC performance we investigated the effect of cathode contamination on current generation in two-chamber microbial fuel cells utilizing three organisms, Bacillus megaterium, Shewanella oneidensis and Escherichia coli as contaminants. The advantage of the two-chamber MFC design used in this study is that the assembled and filled fuel cell is autoclavable allowing sterility to be maintained. While this design is not optimized for power production, we are able to use this system to study effects such as cathode contamination without directly interfering with processes at the anode, which is not practical in more power dense single-chambered designs. 2. Experimental 2.1. Biological materials and reagents S. oneidensis MR-1 ATCC BAA-1096, E. coli MG1655 ATCC 47076 and B. megaterium (Raven Laboratories, Omaha, NE, USA) were used in all studies and were cultured in trypticase soy broth (TSB) (Becton Dickinson, Sparks, MD, USA) in an aerobic 30 ◦ C incubator shaking at 250 RPM. S. oneidensis and E. coli were maintained as frozen stocks (−80 ◦ C) in TSB containing 30% glycerol. B. megaterium was maintained as spores in 40% ethanol. In select fuel cells, the redox indicator resazurin was added to final concentration of 80 M to determine if the catholyte was oxidized. 2.2. MFC design/data collection Two-chamber MFCs were constructed in a similar fashion as previously described (Sund et al., 2007; Milliken and May, 2007) with some modifications and pictured in Scheme 1. The anodes consisted of graphite rods (Ultra Carbon, Bay City, MI, USA), 0.6 cm in diameter, cut to a length of 2.5 cm and placed in direct contact with 4 g of lump graphite (Goodfellow Cambridge Limited, Huntingdon, England). The cathodes consisted of graphite rods as described above; however, no graphite lump was added to the cathode compartment.
Scheme 1. Drawing of MFC design constructed of glass 90◦ elbow, ball-and-socket joints. The anode and cathode are separated by a Nafion 117 proton exchange membrane. The cathode is pictured in the two arrangements, submerged and wicking air. The dotted line represents the electrolyte height. The lump graphite, septa on the anode, and glass cover on the cathode are not pictured because of clarity.
In both cases, the graphite rods were attached to wires with silver epoxy (H20E, Epoxy Technology, Billerica, MA, USA) which was later insulated by coating with Devcon® 5 min fast drying epoxy (ITW Performance Polymers, Riveria Beach, FL, USA) leaving 1.25 cm of graphite exposed. A new piece of Nafion® 117 (DuPont, Wilmington, DE, USA) was used for each fuel cell. Anode wires were passed through silicone rubber septa used to seal the anode compartment following inoculation while the cathode compartment was covered with a loose-fitting glass vial. Each fuel cell compartment was filled with 20 mL of media and autoclaved. The anode compartment and in some cases the cathode compartment of the MFCs were inoculated with 200 L of overnight cultures. MFCs were incubated in an aerobic 30 ◦ C incubator and the potential across a 10 k resistor was measured and recorded every 10 s via a DAQPad-6016 and a custom LabView® VI (National Instruments, Austin, TX, USA). Open circuit potentials of the cells were recorded using the same National Instruments hardware and custom software while AC impedance
Fig. 1. Potential traces of MFCs inoculated with S. oneidensis in anode compartment. Initially, the cathodes were placed just below the surface of the catholyte which was either sterile (A) or inoculated with S. oneidensis (B), B. megaterium (C), or E. coli (D). The control (E) was sterile in both the cathode and anode compartments. After four days the cathodes where repositioned so that only the tip of the cathode was submerged in the catholyte. The addition of resazurin did not affect the voltage which was demonstrated with a set of fuel cells comparing sterile and inoculated cathodes with 80 M resazurin added to both the anode and the cathode compartments. An example is shown in (F) where the cathode was inoculated with B. megaterium and can be directly compared with (C).
3146
C.J. Sund et al. / Biosensors and Bioelectronics 24 (2009) 3144–3147
Table 1 Average and standard deviation of the open circuit potentials in volts and impedance in ohms for each condition of microbial fuel cells. Cell description (anode–cathode)
OCP (V) (cathode up)
OCP (V) (cathode down)
Impedance () (cathode up)
Impedance () (cathode down)
Sterile–sterile S. oneidensis–sterile S. oneidensis–S. oneidensis S. oneidensis–E. coli S. oneidensis–B. megaterium
0.05 (±0.02) 0.43 (±0.05) 0.29 (±0.05) 0.43 (±0.07) 0.45 (±0.01)
0.08 (±0.03) 0.41 (±0.05) 0.06 (±0.06) 0.19 (±0.05) 0.32 (±0.02)
500 (±46) 450 (±10) 410 (±35) 410 (±29) 360 (±40)
330 (±12) 350 (±10) 330 (±43) 350 (±25) 300 (±31)
measurements were obtained with amplitude of 5 mV from 10,000 to 1 Hz using a Model 660a Electrochemical Workstation from CH Instruments (Austin, TX). 3. Results and discussion 3.1. Potential versus time traces for the MFCs MFCs were constructed and the anode compartments were inoculated with S. oneidensis with the exception of the control MFCs. Initially, the cathodes were completely submerged below the surface of the catholyte as seen in Scheme 1. Between 24 and 96 h the potential of the MFCs which contained S. oneidensis in the anode compartments and sterile cathode compartments had reached a potential of approximately 60 mV, as seen in the three replicate potential versus time traces in Fig. 1A. All of the MFCs in which the cathode compartment had been inoculated had much lower potentials, as seen in Fig. 1B–D, which were similar to background levels. The limited potentials seen in the MFCs with inoculated catholyte were most likely due to the reduction of oxygen via microbial metabolism. After approximately 96 h, the cathodes were raised to a point where only the tip of the cathode maintained contact with the catholyte. This allowed the catholyte to wick up the sides of the cathode which, in effect, created a novel air cathode design. When the cathodes were repositioned, there was an increase in potential of all of the MFCs (as shown in Fig. 1) however this increase was less dramatic in the controls where the cathode was not inoculated (as shown in Fig. 1A). Most notable were the large potential increases seen in the MFCs with inoculated cathode compartments. Inoculation of the anode and cathode were performed at the same time. Since the organisms added to the cathode were either facultative or aerobic and the cathode compartments were open to air these cultures reproduce more rapidly than Shewanella in the anaerobic anode compartments. Therefore, the performance is low at inoculation, may rise slightly but, is then quenched until the cathode is raised to the wicking position. 3.2. Characterization of MFCs via OCP and impedance Open circuit potentials (OCPs) of the cells and total cell impedance data were collected on each series of MFCs and are displayed in Table 1. The control fuel cells performed as expected where the OCP of the sterile–sterile (negative control) fuel cells show essentially no electrochemical activity and the S. oneidensis–sterile (positive control) had significant OCPs. In both cases, the OCP did not change significantly whether the cathode was in the wicking or in the submerged position. Contaminated cathodes that are submerged always showed significantly lower potentials than the positive control cathodes. When these cathodes were raised to the wicking position the effects of contamination were mitigated as exemplified by both the plots in Fig. 1 and the OCPs in Table 1, where the potentials returned to within error of the sterile cathode MFCs. The only case where OCP decreased significantly for contaminated cathodes in the wicking position
over the sterile cathode was in fuel cells inoculated with S. oneidensis in both the anode and the cathode. This is logical as those two compartments would have been the most chemically identical based on similar metabolic processes, especially after oxygen was consumed in the catholyte, which would have decreased the potential difference between the compartments. The differences in total cell impedance between submerged and wicking cathodes were only significant in the control fuel cells. Total impedance takes several factors into account including: membrane resistance, electrolyte resistance and contact impedances. The contamination in the cathodes probably decreased the electrolyte resistance and possibly the electrode–electrolyte contact impedance; however, the change observed in all of the fuel cells was small enough to have little effect on overall performance. The increase in potential seen in MFCs with wicking cathodes was most likely due to an increase of oxygen availability as oxygen saturation at the cathode has been shown to greatly enhance MFC performance (Rabaey et al., 2008). The wicking cathodes reduced the microorganisms’ ability to scavenge all of the oxygen prior to reaching cathode suggesting that this design may be beneficial in mitigating the effects of cathode contamination. In order to determine the redox status of the catholyte, as an indication of oxygen availability, identical MFCs to those in Fig. 1A–E were assembled with the redox indicator, resazurin, added to both the anode and the cathode compartments (an example of this data is shown in Fig. 1F). One of the fuel cells was inoculated with E. coli and the other is sterile. The pink color of resazurin in the sterile catholyte indicate that it was more oxidized when compared to the inoculated catholyte which lacked the pink hue, indicating that the resazurin was in a reduced state. An image of the resazurin containing cathode compartment of two fuel cells, one inoculated with E. coli (A) and the other sterile (B) has been included in the Supplementary Material. The oxidation state of the resazurin could have been effected by cellular metabolism however it is reversible and therefore if any oxygen was present in the cathode it would have reoxidized the resazurin. This strongly suggests that there was a high concentration of oxygen in the sterile cathodes while the inoculated cathodes were oxygen poor. The lower redox potential of the contaminated catholyte due to microbial oxygen reduction was most likely responsible for the decrease in MFC potentials seen in Fig. 1 prior to repositioning the cathodes. Repositioning of the cathodes allowed the catholyte to wick up the sides of the cathode which in effect created a novel air cathode design. 3.3. Redox indicator results A previous report has indicated that resazurin can act as a mediator between the bacteria and the anode in MFCs (Sund et al., 2007). In this study, resazurin was added at the same concentration in both the anode and cathode compartments preventing discharge due to chemical potential difference. Also, there was no effect on potential traces in the presence of resazurin (as shown in Fig. 1C and F) indicating that either the MFCs were cathode limited or the S. oneidensis cannot utilize the resazurin as a mediator.
C.J. Sund et al. / Biosensors and Bioelectronics 24 (2009) 3144–3147
4. Conclusions Since most proposed applications of MFCs require that they be located in environments where the cathodes are likely to become contaminated, the effects of contamination on MFC design can be significant. If microorganisms consume oxygen in the catholyte before it reaches the cathode then the MFCs’ performance will be severely limited. Current solutions to this problem are to bubble the catholyte with air, utilize wet-proofed air cathodes or flowing designs where the catholyte flows over the cathode. Since bubbling the catholyte can be energy intensive, the air cathodes and flowing designs are more favorable because they will be less affected by contamination and can be passively exposed to air. Data presented here provide further evidence that cathode design needs to be taken into account. Specifically, designs must take into account oxygen availability in environments that contain oxygen-reducing microorganisms. We have also demonstrated a simple low-cost air cathode design which relies on the wicking action of the catholyte to coat the cathode. Finally, since cathode contamination can dramatically alter MFC characteristics, this must be considered when comparing MFC designs and performance. Acknowledgement The authors thank Prof. Scott Crittenden for helpful discussions in this research. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2009.03.019. References Bergel, A., Feron, D., Mollica, A., 2005. Electrochemistry Communications 7 (9), 900–904. Biffinger, J.C., Ray, R., Little, B., Ringeisen, B.R., 2007. Environmental Science and Technology 41 (4), 1444–1449. Bullen, R.A., Arnot, T.C., Lakeman, J.B., Walsh, F.C., 2006. Biosensors and Bioelectronics 21 (11), 2015–2045.
3147
Chaudhuri, S.K., Lovley, D.R., 2003. Nature Biotechnology 21 (10), 1229–1232. Chen, G.W., Choi, S.J., Lee, T.H., Lee, G.Y., Cha, J.H., Kim, C.W., 2008. Water Research 42 (6–7), 1387–1396. Cheng, S., Liu, H., Logan, B.E., 2006a. Electrochemistry Communications 8 (3), 489–494. Cheng, S., Liu, H., Logan, B.E., 2006b. Environmental Science and Technology 40 (1), 364–369. Cheng, S., Logan, B.E., 2007. Electrochemistry Communications 9 (3), 492–496. Clauwaert, P., Van der Ha, D., Boon, N., Verbeken, K., Verhaege, M., Rabaey, K., Verstraete, W., 2007. Environmental Science and Technology 41 (21), 7564–7569. Fan, Y., Hu, H., Liu, H., 2007. Journal of Power Sources 171 (2), 348–354. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2008. Water Research 42 (6–7), 1387–1396. Gil, G.C., Chang, I.S., Kim, B.H., Kim, M., Jang, J.K., Park, H.S., Kim, H.J., 2003. Biosensors and Bioelectronics 18 (4), 327–334. He, Z., Wagner, N., Minteer, S.D., Angenent, L.T., 2006. Environmental Science and Technology 40 (17), 5212–5217. He, Z., Minteer, S.D., Angenent, L.T., 2005. Environmental Science and Technology 39 (14), 5262–5267. Ieropoulos, I.A., Greenman, J., Melhuish, C., Hart, J., 2005. Enzyme and Microbial Technology 37 (2), 238–245. Kim, B.H., Chang, I.S., Gadd, G.M., 2007. Applied Microbiology and Biotechnology 76 (3), 485–494. Liu, H., Logan, B.E., 2004. Environmental Science and Technology 38 (14), 4040–4046. Liu, H., Ramnarayanan, R., Logan, B.E., 2004. Environmental Science and Technology 38 (7), 2281–2285. Logan, B., Cheng, S., Watson, V., Estadt, G., 2007. Environmental Science and Technology 41 (9), 3341–3346. Logan, B.E., Regan, J.M., 2006. Environmental Science and Technology 40 (17), 5172–5180. Milliken, C.E., May, H.D., 2007. Applied Microbiology and Biotechnology 73 (5), 1180–1189. Min, B., Logan, B.E., 2004. Environmental Science and Technology 38 (21), 5809–5814. Park, D.H., Zeikus, J.G., 2003. Biotechnology and Bioengineering 81 (3), 348–355. Rabaey, K., Clauwaert, P., Aelterman, P., Verstraete, W., 2005. Environmental Science and Technology 39 (20), 8077–8082. Rabaey, K., Keller, J., 2008. Water Science and Technology 57 (5), 655–659. Rabaey, K., Read, S.T., Clauwaert, P., Freguia, S., Bond, P.L., Blackall, L.L., Keller, J., 2008. ISME Journal 2 (5), 519–527. Ringeisen, B.R., Henderson, E., Wu, P.K., Pietron, J., Ray, R., Little, B., Biffinger, J.C., Jones-Meehan, J.M., 2006. Environmental Science and Technology 40 (8), 2629–2634. Sund, C.J., McMasters, S., Crittenden, S.R., Harrell, L.E., Sumner, J.J., 2007. Applied Microbiology and Biotechnology 76 (3), 561–568. Tender, L.M., Reimers, C.E., Stecher, H.A., Holmes, D.E., Bond, D.R., Lowy, D.A., Pilobello, K., Fertig, S.J., Lovley, D.R., 2002. Nature Biotechnology 20 (8), 821–825. Zhang, T., Zeng, Y., Chen, S., Ai, X., Yang, H., 2007. Electrochemistry Communications 9 (3), 349–353. Zhao, F., Harnisch, F., Schroeder, U., Scholz, F., Bogdanoff, P., Herrmann, I., 2006. Environmental Science and Technology 40 (17), 5193–5199.