Abstract:The use of air-cathodes in microbial fuel cells (MFCs) has been considered sustainable for large scale applications, but the performance of most current designs is limited by the low efficiency of the three-phase oxygen reduction on the cathode surface. In this study we developed carbon nanotube (CNT) modified air-cathodes to create a 3-D electrode network for increasing surface area, supporting more efficient catalytic reaction, and reducing the kinetic resistance. Compared with traditional carbon cloth cathodes, all nanotube modified cathodes showed higher performance in electrochemical response and power generation in MFCs. Reactors using carbon nanotube mat cathodes showed the maximum power density of 329 mW m −2 ; more than twice that of the peak power obtained with carbon cloth cathodes (151 mW m −2 ). The addition of Pt catalysts significantly increased the current densities of all cathodes, with the maximum power density obtained using the Pt/carbon nanotube mat cathode at 1118 mW m −2 . The stable maximum power density obtained from other nanotube coated cathodes varied from 174 mW m −2 to 522 mW m −2 . Scanning electron micrographs showed the presence of conductive carbon nanotube networks on the CNT modified cathodes that provide more efficient oxygen reduction. Keywords Microbial fuel cell Carbon nanotube Cathode Electricity 1 Introduction Microbial fuel cells (MFCs) are renewable energy systems that employ bacteria to convert chemical energy stored in biodegradable materials to electrical energy. The MFC technology carries great potential as it concurrently removes pollutants from the environment and produces energy. An MFC reactor generally consists of an anode, a cathode, and sometimes a separator between the two electrodes [1] . In order to provide good access for bacteria and improve power generation, the electrodes in MFCs need to have high surface area, high conductivity, and be resistant to physical and chemical corrosion. Many anode materials have been tested, including woven graphite mat [2] , carbon paper [3] , carbon cloth [4] , and activated carbon [5] . A recent development of graphite brush anodes provides a solution for scaling up because it has very high specific surface area (7170–18,200 m 2 m −3 ) and an open structure to prevent fouling problems [6] . Moreover, this makes the anode no longer the main limitation on power production, but instead brings up the challenge of effective cathode development [7,8] . Even though ferricyanide or permanganate can provide a higher cathode open circuit potential, oxygen is considered as the electron acceptor for eventual MFC applications due to its availability and high redox potential ( E 0 = 1.23 V) [8,9] . However, the tri-phase reaction among oxygen gas, solid catalyst, and liquid electrolyte does not make satisfactory reaction kinetics. Moreover, the specific surface area (∼100 m 2 m −3 ) of current popular air-cathode designs are orders of magnitude lower compared with the brush anode, making the cathode the primary limiting factor for improved power production in MFCs. Previous results showed that MFC power output improved along with the increase of cathode specific surface area. For example, Deng et al., found the power density increased from 67 mW m −2 to 315 mW m −2 by replacing carbon paper cathodes with high specific surface area activated carbon felt cathodes [10] . Cheng and Logan recently demonstrated that the volumetric power density has a linear relationship with the cathode specific surface area [11] . Carbon nanotubes (CNTs) have exhibited great potential as electrode materials in fuel cell applications due to their high surface-to-volume ratio and unique electrical and mechanical properties. Several studies investigated the performance of CNT-modified anodes in MFCs and found that the anode biofilm activity was not affected by the carbon nanotubes but the power density was improved [12–14] . Though the CNT-modified anodes do not provide direct large enough macro-scale porosity for more microbial colonization, the highly conductive nanotube network serves as nanowires to facilitate the electron transfer between the microbes to the electrodes. The reported power densities obtained from CNT-modified anodes varied from 22 mW m −2 (multi-walled CNT) [15] to 1098 mW m −2 (3-D CNT-textile) [13] . However, to date little effort has been reported on CNT-modified cathodes, even though the benefits can be more significant because the CNT-modified cathodes provide higher conductivity for improved electron transfer efficiency and allow more nano-sized catalyst particles to be deposited in a 3D texture to facilitate the tri-phase reaction kinetics. Therefore, in this study we used carbon nanotubes to modify the air-cathode using several different methods and characterized the MFC performance under each condition. 2 Materials and methods 2.1 Cathode construction Different coating methods were used to modify MFC air-cathodes using carbon nanotubes. In addition, carbon nanotube mat and carbon cloth modified with Pt nanoparticles were tested for comparison. Table 1 shows the list of 8 different electrodes that were characterized and the specifications of their modification. Carbon cloth cathodes (CC) were made by applying one layer of carbon black nanoparticles and four PTFE diffusion layers on the air side of the carbon cloth according to Cheng et al. [16] . In some cases, a catalyst layer containing 0.5 mg cm −2 platinum nanoparticles was applied on the water side of the carbon cloth to improve reaction kinetics (CC-Pt). Carbon nanotube mat (CNTM) that contains more than 90% carbon nanotubes was donated by Nanocomp Technologies Inc. (OH, USA) and was used directly as the MFC cathode. The same amount of Pt catalyst was applied in some of the CNTM cathodes for comparison (CNTM-Pt). Single-walled carbon nanotubes (SWNTs) were synthesized by laser vaporization methods [17,18] at the National Renewable Energy Laboratory (NREL) (CO, USA). The as-prepared nanotubes were purified using a method introduced by Dillon, which consisted of treatment in 3 M nitric acid flux for 16 h, followed by an acetone wash and air burn at 525 °C [17] . Commercially available SWNTs and multi-walled carbon nanotubes (MWNTs) were purchased from CheapTubes Inc. (NJ, USA) and purified by the same procedure. The purified carbon nanotubes were suspended in 1% sodium dodecyl sulfate (SDS) surfactant water solution to prevent agglomeration. The suspended carbon nanotubes were then deposited onto a PTFE membrane (Advantec MFS, Inc., CA, USA) by vacuum filtration to form a compact cathode. Sufficient DI water washing was applied after to remove the extra surfactant. In some cases, Pt nanoparticles (0.5 mg cm −2 ) were uniformly coated on the surface of carbon nanotubes by using a controlled temperature microwave method (250 W, 140 °C) for 90 s [19] . The cathodes containing nanotubes synthesized at NREL with and without Pt coatings were labeled as SWNT n -Pt and SWNT n , respectively. Similarly, the cathodes using commercial SWNTs and MWNTs from CheapTubes Inc. with Pt coating were labeled as SWNT c -Pt MWNT c -Pt, respectively ( Table 1 ). 2.2 MFC construction and operation Single chamber cubic-shaped reactors were constructed as previously described [6,20] . The total volume of each reactor was 28 mL. Graphite fiber brushes were used as the anodes in all of the experiments. Prior to the tests, the brushes were treated by soaking in acetone overnight and then heated at 450 °C for 30 min [21] . All reactors were initially equipped with CC-Pt cathodes and inoculated using effluent from an air-cathode MFC operated for nearly one year. When exoelectrogenic bacteria acclimated and all reactors showed repeatable and comparable voltages (510 ± 10 mV), nanotube modified air-cathodes were transferred to MFC reactors for performance characterization. Medium solution was prepared containing 1.25 g L −1 CH 3 COONa, 0.31 g L −1 NH 4 Cl, 0.13 g L −1 KCl, 3.321 g L −1 NaH 2 PO 4 ·2H 2 O, 10.317 g L −1 Na 2 HPO 4 ·12H 2 O, 12.5 mL L −1 mineral solution, and 5 mL L −1 vitamin [22] . Reactors were operated in fed-batch mode at room temperature and refilled with new medium solution when voltages reduced below 30 mV forming one cycle of operation. 2.3 Electrochemical and microscopy analysis Voltage across an external resistor (1000 Ω) was recorded at 10-min intervals using a data acquisition system (Keithley Instrument, OH, USA) connected to a computer. Polarization power density curves were obtained by altering external resistances from 50,000 Ω to 50 Ω during the stable power production stage of each batch. The calculations of power density were performed according to Ren et al. [23] . Linear sweep voltammetry (LSV) was applied using a potentiostat (PC4/300, Gamry Instruments, NJ, USA) to characterize the abiotic electrochemical oxygen reduction behavior on different cathodes. LSV tests were conducted in the same reactor filled with same media solution but without carbon source and bacterial inoculum [24] . The working electrode was the cathode (4.9 cm 2 projection area), the counter electrode was a titanium wire (diameter: 0.081 cm, length: 40 cm), and the reference electrode was an Ag/AgCl electrode (RE-5B, BASi, IN, USA). The potential was scanned from open circuit potential to −0.2 V at a rate of 1.0 mV s −1 . Reactor internal resistances were measured by electrochemical impedance spectroscopy (EIS) with the anode as the working electrode, and the cathode as the counter electrode and reference electrode. The scan range was from 10 5 Hz to 0.005 Hz with a small sinusoidal perturbation of ±10 mV [3] . All tests were repeated at least three times to verify consistency. Selected cathode samples were examined using a dual beam focused ion beam scanning electron microscope (FIB/SEM, NOVA 600i, FEI Company). Samples were fixed overnight at 4 °C by Karnovsky's fixative (Electron Microscopy Sciences, CA, USA), washed three times in phosphate buffer (0.2 M, pH 7.2), and then dehydrated stepwise in a series of water/ethanol solutions with increasing ethanol concentration (50, 70, 80, 90, 100%). Samples were then kept in a desiccator prior to Pd/Pt sputtering and SEM observation. 3 Results and discussion 3.1 Electrochemical performance The electrochemical performance of the cathodes with respect to current density was evaluated using LSV tests in the absence of bacteria ( Fig. 1 ). Among the cathodes without Pt catalyst coating, those electrodes consisting of or modified with carbon nanotubes (e.g., CNTM and SWNT n ) showed much larger current responses than the carbon cloth cathode (CC). SWNT n cathodes showed the most positive onset potential, followed by the CNTM cathode, while the CC showed minimal current response across the potential scan range. The larger current response from the nanotube cathodes indicates a higher limiting current density and better electrochemical performance, presumably due to the higher specific surface area generated by nanotube modification [25] . In comparison, the flat current response of CC cathodes over the range of voltages examined indicates that the carbon cloth itself did not catalyze oxygen reduction. The addition of Pt catalyst on the cathodes significantly increased the current densities of all cathodes ( Fig. 1 ). The current densities of CNTM-Pt, SWNT n -Pt, and CC-Pt cathodes all increased substantially compared with their non-Pt counterparts, and the SWNT n -Pt showed the best performance across the potential range. Considering the same amount of catalyst was applied on the electrodes, the LSV results suggest that the 3-D structure created by the carbon nanotubes on SWNT n -Pt allowed more Pt nanoparticles to be deposited inside the electrode space rather than on the surface, leading to increased reaction kinetics [26] . The electrochemical performance of cathodes modified by commercial single-walled nanotubes (SWNT c -Pt) and multi-walled nanotubes (MWNT c -Pt) with Pt coatings were also evaluated using the same LSV tests. In general, both cathodes showed lower current densities compared to the other three electrodes described previously. The reason could be attributed to the different nanotube synthesis approaches. The laser vaporization method used for SWNT n -Pt has been known to produce higher quality nanotubes than the chemical vaporization method used for SWNT c -Pt. The higher crystallization and electric conductivity of SWNT n resulted in higher electron transfer efficiency thus higher performance. The SWNT c -Pt showed larger current response than the MWNT c -Pt across the scanned potentials. This is presumably due to the high resistance thus more electrochemical loss of MWNTs compared with SWNTs. The resistance of the reactor with the MWNT c -Pt cathode was 94 Ω, which is nearly 3 times higher than with the SWNT c -Pt cathode (35 Ω). 3.2 Performance of MFCs with nano-modified air-cathodes The nano-modified air-cathodes were transferred to pre-acclimated single chamber MFCs after LSV tests. Rapid voltage generation was observed in all reactors. Fig. 2 shows the power density and polarization curves obtained from each MFC reactor during steady power production stages. Fig. S1 shows the voltage profiles as a function of time ( Supporting information ). For MFC cathodes without Pt catalyst coating, CNTM reactor showed the maximum power density of 329 mW m −2 , more than twice that of the peak power obtained from CC reactor (151 mW m −2 ). The SWNT n reactor showed a lower power density (117 mW m −2 ) than the other two reactors without Pt. Similar to the results observed during LSV tests, the reactor performance was greatly improved by the application of Pt catalyst in the cathodes. During the first several batches, the power densities of SWNT n -Pt and CNTM-Pt reactors, calculated from voltages at 1000 Ω, were 735 mW m −2 and 723 mW m −2 , respectively, which were higher than that of CC-Pt reactor (672 mW m −2 ). The voltages of CNTM-Pt and CC-Pt MFCs kept stable during multiple batches, but the voltage of the SWNT n -Pt reactor decreased gradually from the first batch and stabilized after the fourth batch, resulting in a voltage drop from more than 600 mV to around 300 mV. As shown in Fig. 2 , the power density curves obtained at the steady state operation of each reactor show the maximum power densities from CNTM-Pt and CC-Pt MFCs of 1118 mW m −2 and 1071 mW m −2 , respectively, while the maximum power density of SWNT n -Pt after stabilizing was 302 mW m −2 . A similar voltage decline was observed in SWNT c -Pt and MWNT c -Pt MFCs in the first batches. The maximum power density produced by SWNT c -Pt was 522 mW m −2 , which was about two times higher than that from MWNT c -Pt (174 mW m −2 ), confirming that SWNT material performs better as MFC cathodes due to its lower ohmic resistance compared with MWNT material. The electrochemical performance of CC-Pt and SWNT n -Pt cathodes were analyzed again by LSV tests after MFC operation in order to understand the chemical and microbial effects on the cathode performance. Fig. 3 shows the electrochemical performance of the used CC-Pt cathode decreased slightly compared with the new cathode, while the current response of the used SWNT n -Pt cathode declined significantly compared with the new SWNT n -Pt. This drop may explain why the steady state power density of the SWNT n -Pt reactor was lower than the CNTM-Pt and CC-Pt reactors despite a better LSV performance with new SWNT n -Pt cathodes. Such finding could also be confirmed by the changes of open circuit potential (OCP). The OCP of the SWNT n -Pt cathode dropped by 35%, from 357 mV in batch 1 to 233 mV in batch 4, while the OCP of CC-Pt only dropped by 9%, from 345 mV to 310 mV during the same period. Similar findings in air-cathode performance decline were reported by several other studies, and the reasons were believed to be due to the adherence of biofilm and chemical deposits on the cathode surfaces that reduced charge transfer and catalyst activity [27,28] . For the SWNT n -Pt cathode, the large amount of pores, cavities, and curving paths among carbon nanotubes that increase the specific surface area also enables adsorption of impurities that may cover the surfaces of the cathode and reduce charge transfer. Adding a separator on the nanotube modified electrode to block adsorption or switching the nanotube layer to the air-face side could be possible solutions [29,30] . It may also be possible that the SWNTn-Pt cathode performance decreased due to the loss of the Pt catalyst overtime since the microwave deposition method has not been optimized for MFC operation. However, this was not confirmed using energy dispersive spectroscopy element percentage test during FIB/SEM characterization. 3.3 FIB/SEM analysis Scanning electron micrographs were taken for the CC-Pt, CNTM-Pt, and SWNT n -Pt cathode materials before and after MFC operation. Fig. 4 shows a distinct difference in morphology between the new CC-Pt cathode and new CNTM-Pt and SWNT n -Pt cathodes. The CC-Pt cathode surface was covered by aggregated particles or short fibers ( Fig. 4 A), while the new CNTM- Pt and SWNT n -Pt electrode surfaces clearly showed the carbon nanotube network ( Fig. 4 C and E). The deposited Pt catalyst shows up in the images as bright nanoparticles deposited across the nanotube network. Denser microbial biofilms were formed on the surface of the CC-Pt cathode ( Fig. 4 B) compared with the CNTM-Pt and SWNT n -Pt cathodes ( Fig. 4 D and F), but more chemical deposit covered a large portion of the surface of the two CNT-modified electrodes. Studies showed that the excess accumulation of biofilm and chemical scales [27,28] could adversely affect the system performance due to the decrease in active cathode specific surface area and increase in diffusion resistance in oxygen. Such observations may also explain the reduced electrochemical activities in certain MFCs after a period of operation, where the presence of cavities and the adsorption of impurities played a major role as compared to cathode biofilms. 4 Conclusions Microbial fuel cells provide direct and efficient electricity generation from renewable sources, but the current 2-D air-cathode configuration limits system performance due to the low kinetics of oxygen reduction. Carbon nanotubes were used in this study to modify air-cathodes in single chamber MFCs to create a 3-D structure for improved surface area and reaction kinetics. Compared with traditional carbon cloth cathodes, all nanotube modified cathodes showed greater electrochemical performance as well as higher power density in MFCs. The maximum power density obtained from carbon nanotube mat cathodes (329 mW m −2 ) was more than double the power output from traditional carbon cloth cathodes (151 mW m −2 ). The addition of Pt catalyst on the cathodes increased the current densities of all cathodes, with the maximum power density achieved by a CNTM-Pt of 1118 mW m −2 . Electrodes made from commercial single wall carbon nanotubes (SWNT c -Pt) have much lower ohmic resistance than those made from multiwall nanotubes (MWNT c -Pt) and showed larger current response and higher power. The customized SWNT n electrode showed great electrochemical responses, but its performance declined gradually due to the deposition of chemical and microbial impurities which blocked reaction surfaces. Scanning electron micrographs demonstrated different electrode surface morphologies, with the CNT modified cathodes showing carbon nanotube networks and carbon cloth cathodes showing aggregated particles on the electrode surface. Acknowledgements This research was supported by the Office of Naval Research (ONR) Grant N000140910944 . The authors thank Nanocomp Technologies Inc. for donating carbon nanotube mat, and Dr. Paul Rice at University of Colorado Nanomaterials Characterization Facility (NCF) for helping with FIB/SEM operation. Appendix A Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jpowsour.2011.05.005 . Appendix A Supplementary data References [1] B.E. Logan Microbial Fuel Cells 2008 John Wiley & Sons New York [2] K. Rabaey P. Clauwaert P. Aelterman W. Verstraete Environ. Sci. Technol. 39 2005 8077 8082 [3] Z. Ren R.P. Ramasamy S.R. Cloud-Owen H. Yan M.M. Mench J.M. Regan Bioresour. Technol. 102 2011 416 421 [4] S. Cheng B.E. Logan Electrochem. Commun. 9 2007 492 496 [5] Z. He N. Wagner S.D. Minteer L.T. Angenent Environ. Sci. Technol. 40 2006 5212 5217 [6] B. Logan S. Cheng V. Watson G. Estadt Environ. Sci. Technol. 41 2007 3341 3346 [7] Y.Z. Fan E. Sharbrough H. Liu Environ. Sci. Technol. 42 2008 8101 8107 [8] H. Rismani-Yazdi S.M. Carver A.D. Christy I.H. Tuovinen J. Power Sources 180 2008 683 694 [9] F. Zhao F. Harnisch U. Schroder F. Scholz P. Bogdanoff I. Herrmann Environ. Sci. Technol. 40 2006 5193 5199 [10] Q. Deng X. Li J. Zuo A. Ling B.E. Logan J. Power Sources 195 2010 1130 1135 [11] S. Cheng B.E. Logan Bioresour. Technol. 102 2011 4468 4473 [12] P. Liang H. Wang X. Xia X. Huang Y. Mo X. Cao M. Fan Biosens. Bioelectron. 26 2011 3000 3004 [13] X. Xie L. Hu M. Pasta G.F. Wells D. Kong C.S. Criddle Y. Cui Nano Lett. 11 2011 291 296 [14] J.L. Lamp J.S. Guest S. Naha K.A. Radavish N.G. Love M.W. Ellis I.K. Puri J. Power Sources 196 2011 5829 5834 [15] H.-Y. Tsai C.-C. Wu C.-Y. Lee E.P. Shih J. Power Sources 194 2009 199 205 [16] S. Cheng H. Liu B.E. Logan Electrochem. Commun. 8 2006 489 494 [17] A.C. Dillon T. Gennett K.M. Jones E.L. Alleman P.A. Parilla M.J. Heben Adv. Mater. 11 1999 1354 1358 [18] A. Thess R. Lee P. Nikolaev H. Dai P. Petit J. Robert C. Xu Y.H. Lee S.G. Kim A.G. Rinzler D.T. Colbert G.E. Scuseria D. Tomanek J.E. Fischer R.E. Smalley Science 273 1996 483 487 [19] C.-M. Chen M. Chen H.-W. Yu S.-C. Lu C.-F. Chen Jpn. J. Appl. Phys. 47 2008 2324 2329 [20] H. Liu B.E. Logan Environ. Sci. Technol. 38 2004 4040 4046 [21] X. Wang S. Cheng Y. Feng M.D. Merrill T. Saito B.E. Logan Environ. Sci. Technol. 43 2009 6870 6874 [22] H. Luo P.E. Jenkins Z. Ren Environ. Sci. Technol. 45 2011 340 344 [23] Z. Ren H. Yan W. Wang M. Mench J. Regan Environ. Sci. Technol. 45 2011 2435 2441 [24] F. Zhang T. Saito S. Cheng M.A. Hickner B.E. Logan Environ. Sci. Technol. 44 2010 1490 1495 [25] J.R. Kim J.-Y. Kim S.-B. Han K.-W. Park G.D. Saratale S.-E. Oh Bioresour. Technol. 102 2011 342 347 [26] S. Cheng P. Kiely B.E. Logan Bioresour. Technol. 102 2011 367 371 [27] P.D. Kiely G. Rader J.M. Regan B.E. Logan Bioresour. Technol. 102 2011 361 366 [28] K. Chung I. Fujiki S. Okabe Bioresour. Technol. 102 2011 355 360 [29] O. Lefebvre W.K. Ooi Z. Tang M. Abdullah-Al-Mamun D.H. Chua H.Y. Ng Bioresour. Technol. 100 2009 4907 4910 [30] S. Yang B. Jia H. Liu Bioresour. Technol. 100 2009 1197 1202

Abstract:To investigate the effects of external resistance on the biofilm formation and electricity generation of microbial fuel cells (MFCs), active biomass, the content of extracellular polymeric substances (EPS) and the morphology and structure of the biofilms developed at 10, 50, 250 and 1000 Ω are characterized. It is demonstrated that the structure of biofilm plays a crucial role in the maximum power density and sustainable current generation of MFCs. The results show that the maximum power density of the MFCs increases from 0.93 ± 0.02 W m −2 to 2.61 ± 0.18 W m −2 when the external resistance decreases from 1000 to 50 Ω. However, on further decreasing the external resistance to 10 Ω, the maximum power density decreased to 1.25 ± 0.01 W m −2 because of a less active biomass and higher EPS content in the biofilm. Additionally, the 10 Ω MFC shows a highest maximum sustainable current of 8.49 ± 0.19 A m −2 . This result can be attributed to the existence of void spaces beneficial for proton and buffer transport within the anode biofilm, which maintains a suitable microenvironment for electrochemically active microorganisms. Keywords External resistance Microbial fuel cell Power density Sustainable current generation Biofilm structure 1 Introduction Microbial fuel cell (MFC) is an economical pathway to a sustainable energy future. In a typical MFC, electron donors, such as organic materials in waster, are oxidized by the electrochemically active bacteria mostly growing as a biofilm on the anode surface [1,2] . Although experiencing a significant development in recently years, the power generation is still insufficient for the practical applications. In order to improve the MFC performance, efforts have been made to enrich more electrochemically active bacteria [3–5] , to improve reactor configuration [6–10] , to identify better electrode materials [6,8–12] , as well as to optimize process parameters [6,13] . Optimizing the growth conditions for the electrochemically active bacteria in the anode was also an important consideration for improving the performance of MFCs. Many factors, such as nutrient supply, flow rate, pH, temperature [2,6,7,14,15] , have been found to strongly affect the MFC performance and start-up time. One of the most important and most investigated factors is the anode potential at which the MFC is operated, as it controls the theoretical energy gain for microorganisms [16] . Finkelstein et al. [17] reported that a larger and earlier maximum current were obtained at a more positive applied potential due to the increased energy yield for microbial colonization. In addition, Aelterman et al. [4] found that the anode potential had no influence on the start-up time but an improved current and power generation can be obtained at an optimal anode potential of −200 mV vs. Ag/AgCl. Wang et al. [18] reported that a poised potential of 200 mV (vs. Ag/AgCl) accelerated the start-up of MFCs. Recently, Wei et al. [19] investigated the current and biomass production of G. sulfurreducens growing on anodes at a potential of −160, 0, and 400 mV (vs. SHE). The results indicated that faster start-up times, higher current generation and greater biomass production were achieved when the anode set to 0 and 400 mV. Besides the anode potential, the effect of external resistance applied to the electrical circuit also received wild attention since controlling the growth condition for the electrochemically active bacteria by changing the external resistance is more feasible than poising the anode potential in MFC applications. Generally speaking, MFC performance increases with decreasing the applied external resistance. Liu et al. [20] demonstrated that the lower external resistance was applied, the higher maximum power output was obtained. Aelterman et al. [12] reported a significant increase both in continuous current generation and in power generation when the external resistance of MFCs was lowered from 50 to 25 and then to 10.5 Ω when the MFC was operated at each resistance for 9–11 days. In addition, the changes in the microbial diversity and metabolism were also noted when different external resistances were applied to the MFCs. Lyon et al. [21] observed large differences in power generation and structure of microbial community when MFCs were operated at different external resistances. Rismani-Yazdi et al. [22] investigated the effect of external resistance on bacterial diversity and metabolism in MFCs. They found that changes in the external resistance affected not only the power generation, but also the microbial metabolism. Mclean et al. [23] described how difference in external resistances affect cellular electron transfer rates on a per cell basis and overall biofilm development in Shewanella oneidensis strain MR-1 by monitoring the real-time microscopic imaging of anode population. They found that the anode of a MFC started up at a low resistance (100 Ω) has a thinner biofilm and higher current per cell, compared with that at a high external resistance (1 MΩ). Moreover, the power production and coulombic efficiency of MFCs can be substantially improved using an automatic control strategy of external resistance as reported by Premier et al. [24] . The external resistance may also influence the MFC performance by affecting the structure of the biofilm in the anode. For an MFC using organic matters as substrate, a lower external resistance usually leads to a higher amount of electrons per unit of time from the biofilm to electrode, and therefore a higher proton production rate in the interior of the biofilm. As a result, when the MFC is operated under a low external resistance, the higher proton production rate will cause a pH drop within the biofilm, which in turn would lead to unfavorable microenvironments for the growth of the electrochemically active bacteria. On the other hand, it is well known that the architecture of the biofilm adapts in response to environmental stresses, such as low nutrient availability, high shear forces, unfavorable pH value and toxic compounds. Therefore, it was reasonable to postulate that the morphology and structure of the biofilm developed at various external resistances would be different dues to the different proton producing rates within the biofilm. In the present study, four biofilm samples were established in the anode of a two-chamber MFC using acetate as substrate employing anaerobic mixed consortia as anodic inoculums under an external resistance of 10, 50, 250 and 1000 Ω. The objectives of the present study were to assess the effects of external resistance on the start-up process, biofilm formation and structure, as well as the electricity generation of the MFCs. 2 Materials and methods 2.1 MFC configuration and operation The experiments were conducted using four two compartment MFCs. Each MFC consisted of a proton exchange membrane (PEM) (Nafion 117, Dupont), two carbon cloth electrodes (E-TEK, B-1A, America) and two plexiglass plates with a flow channel holding a volume of 2.7 mL. The PEM and electrodes had an apparent surface area of 25 cm 2 . The anode compartment was inoculated with the effluent from a MFC running on acetate. Both anode and cathode compartments were equipped with Ag/AgCl reference electrodes. All MFC tests were conducted in a temperature-controlled room at 25 °C and were repeated three times to check reproducibility of the data. In order to evaluate the impact of external resistances, the MFCs were started at an external resistance of 10, 50, 250, and 1000 Ω (denoted as MFC-10, MFC-50, MFC-250 and MFC-1000, respectively). After 7 days of operation, fresh anolyte, containing sodium acetate 2.7 g L −1 , 6 g L −1 Na 2 HPO 4 , 3 g L −1 KH 2 PO 4 , 0.1 g L −1 NH 4 Cl, 0.5 g L −1 NaCl, 0.1 g L −1 MgSO 4 ·7H 2 O, 15 mg L −1 CaCl 2 ·2H 2 O and 1.0 mL L −1 of a trace elements solution, was fed at a rate of 1.5 L day −1 . To minimize the influence of the cathode potential on the overall cell performance, a 50 mM potassium ferricyanide solution was continuously supplied to the cathode at a rate of 1 mL min −1 during the start-up stage and steady-state operation. 2.2 Data acquisition and analysis Cell voltages ( U ), currents ( I ), anode and cathode potentials of each MFC were collected every 15 s via an Agilent 34970A data acquisition unit connected to a PC. In the measurement, the external resistance was varied in a range of (5–1.0) × 10 5 Ω to control the condition of discharging and record the voltage–current curves. After each change in the electrical resistance, the MFCs were held until the current and cell voltage reached steady-state. The power density of the MFCs was calculated according to Eq. (1) : (1) P = U × I A where A is the surface area of the electrode, I and U are the sustainable current and cell voltage after the MFC reached steady-state, respectively. The cyclic voltammetry (CV) experiments were carried out in a range from −0.6 to +0.37 V vs. SHE at a scan rate of 5 mV s −1 using a Zennium electrochemical workstation (Zahner, Germany) with the anode as the working electrode. The cathode was used as the counter electrode and the reference electrode was an Ag/AgCl reference electrode. The metabolic energy gain (MEG) for the electrochemically active bacteria was calculated with the following equation described by Wei et al. [19] : (2) M E G = ∫ 0 t ( E andoe ( t ) − E 0 ′ ) I ( t ) d t where E anode ( t ) and I ( t ) are the anode potential and current at time t during start-up period respectively. E 0 ′ is the standard potential of CO 2 /acetate (−0.29 V). The unavailable part of MEG for the electrochemically active bacteria was neglected because this portion make up less than 20% of the MEG when the MFC was started up by applying an external resistance to the electrical circuit [19] . 2.3 SEM characterization The morphology of biofilms formed under different external resistances was analyzed by a scanning electron microscopy (SEM) (3400N, HITACHI instrument). At the removal from the anode compartment, the samples was immediately sliced by a sterilized bistoury, and fixed by phosphate buffer saline (PBS) with 2.5% glutaraldehyde (pH 7.2–7.4) for 1 h. The samples were then washed three times with the same buffer and dehydrated stepwise with a graded series of ethanol solutions (30, 50, 70, 80, 90 and 100%). The electrode samples were finally critical-point dried with tert-butyl ethanol and sputter coated with a thin layer of gold. The biofilm was observed in high vacuum using SEM in the secondary electron imaging mode. 2.4 Quantization of biofilm components 2.4.1 Biomass quantification by phospholipid method For the evaluation of total active biomass, small pieces of anode electrode with biofilms were taken from the MFCs and analyzed according to the procedure reported by Verstraete [4] as soon as the current generation was stable. The phosphate concentration released from phospholipids was determined using a Leng Guang 756mc spectrophotometer at 610 nm. Phosphate concentration was then correlated to total active biomass using the conversion factor of 191.7 μg of biomass-C per 100 mmol of phospholipid [25] . 2.4.2 EPS characterization The characteristics of biofilm EPS (extracellular polymeric substances) in each reactor were analyzed according to Fang et al. [26] . Four biofilm-coated samples were removed from each MFC after the start-up and immersed in 50 mL of a pH 7.5 TE buffer containing 10 mM Trizma base, 10 mM EDTA and 2.5% NaCl. Biofilm was scrapped from each sample, and centrifuged at 3500 rpm for 20 min. The concentrated biomass was then re-suspended in 10 ml of a 0.85% sodium chloride solution containing 0.22% formaldehyde at 80 °C for 30 min for EPS extraction. The EPS dissolved in the formaldehyde solution was recovered by further centrifugation at 3500 rpm for 30 min. The carbohydrate content of EPS (EPS c ) in the extracted solution was measured using the phenol/sulfuric-acid method [27] and the protein content (EPS p ) using the Lowry method [28] . The sum of EPS p and EPS c represented the total EPS of the samples. Lactose and bovine serum albumin were used to make calibration curves for the determination of carbohydrates and proteins EPS, respectively. 3 Results and discussion 3.1 Influence of external resistance on the start-up process During the start-up period, the anode compartment was inoculated with the effluent from a MFC running on acetate in a continuous feed mode. As shown in Fig. 1 a , both current generation, anode and cathode potential of the MFCs were significantly affected by the external resistance. The lag period for MFC-10 was 2.2 day, reaching 14.6 mA after 6.3 days ( Fig. 1 ). Correspondingly, the potential of the anode and the cathode evolved from their initial potentials to −0.14 V and +0.41 V (vs. SHE), respectively. For MFC-50, a steady state current of about 11.2 mA and anode potential of −0.19 V (vs. SHE) were obtained after a 4.3-day start-up period. Fig. 1 also shows a steady current of 2.8 and 0.7 mA, a steady anode potential of −0.23 and −0.25 V (vs. SHE) and a start-up time of 3.0 and 2.6 days for MFC-250 and MFC-1000, respectively. From these results, it is clear that the MFC with a lower resistance showed a higher steady anode potential and current generation after start-up. During start-up, the power of MFC-10, MFC-50, MFC-250 and MFC-1000 increased from 0 to 1.96, 6.05, 1.95 and 0.64 mW, respectively. Fig. 1 also illustrates that there was no obvious change in current generation and anode potential for the four MFCs after the replacement of fresh anolyte, suggesting successful biofilm establishment on the surface of carbon cloth in the anode chamber. 3.2 Influence on active biomass and energy gain for microorganisms The influence of external resistance on the accumulated active biomass and the energy gain of the biofilms formed at 10, 50, 250 and 1000 Ω during the period of start-up is presented in Fig. 2 a . It can be seen that the energy gain for the electrochemically active bacteria decreased with the increase of external resistance from 10 to 1000 Ω. Such a result was expected because a lower external resistance induced a higher current generation and a higher potential difference between the anode potential and the redox potential of acetate, both leading to a higher energy gain for the microorganisms (Eq. (2) ). It is also interesting to note that the energy gain for the electrochemically active bacteria grown at 10 Ω (1112.4 J) was significantly higher than the previously reported range of 0–180 J [19] . This result can be mainly attributed to the higher current densities (6.0 A m −2 ) obtained in this study compared with the previous study (0–2.0 A m −2 ). From Fig. 2 a, it also can be seen that the active biomass of the MFCs increased with the decrease of external resistance from 1000 to 50 Ω, and then decreased with external resistance from 50 to 10 Ω. This trend is inconsistent with previous reports in which a higher biomass production was observed at a lower external resistance because of the increased energy gain for the microorganisms at low resistances [3,12] . This probably related to the fact that a higher portion of energy was consumed for the synthesis of EPS rather than for bacterial growth in the case of MFC-10, as will be shown in the following paragraph. It is commonly accepted that the energy for microorganism is essential for the maintenance of the bacterial vitality [29] . However, the microorganisms may also produce EPS during growth on the anode surface, consuming energy which is available for bacterial growth [30] . EPS could contribute up to 50–90% of the total organic carbon of biofilms and serves many functions such as promoting the initial attachment of cells to solid surfaces, maturing biofilm structure, and enhancing biofilm resistance to environmental stress and disinfectants. The existence of EPS on the anode surface of bio-electrochemical systems fed with various substrates has been noted in recent studies [31–33] . In this study, the EPS content of the anode biofilms developed at various external resistances after start-up was also estimated and elucidated in Fig. 2 b. It appears that the EPS content of biofilm increased as the external resistance increased. For instance, the EPS content of biofilm established during start-up decreased from 296.8 ± 42.9 to 51.9 ± 4.2 mg g carbon-cloth − 1 when the external resistance was increased from 10 to 1000 Ω. Considering that the active biomass of the biofilm developed at 10 Ω was lower than that of 50 Ω, whereas the energy gain was higher, this result suggested that a greater portion of energy was consumed for EPS synthesis in the case of the biofilm developed at 10 Ω. 3.3 Influence on electrochemical behavior of biofilms In order to evaluate the importance of mediators produced by the microorganisms for electricity generation, CV tests were conducted on the anode during stable current generation and with depleted substrate. From the voltammograms in Fig. 3 , all the biofilms during stable current generation ( Fig. 3 a) show catalytic behavior for the bioelectrooxidation of acetate by the appearance of an oxidation current at an onset potential of −0.26 V (vs. SHE) during the anodic scan. In the cathodic scan, additional oxidation peaks can be observed around −0.16 V (vs. SHE) and no reduction peaks were found. It is obvious from Fig. 3 a that the voltammograms of all the samples exhibited no significant difference, except that the oxidation peak (0.11 V vs. SHE) of the biofilms established at 50 Ω was significantly higher than that obtained from the other samples. This improvement can be attributed to the higher active biomass within the biofilm of MFC-50, as indicated in Fig. 2 a. In comparison, all the curves for the anodes with depleted substrate ( Fig. 3 b) exhibited oxidation peaks at −0.1 V (vs. SHE) and reduction peaks at −0.18 V (vs. SHE). These results are consistent with previous studies [20,34] , which were indicative of the existence of mediators produced by the mixed culture. The similarities among these results also suggest that the dominant species in the biofilms were slightly influenced by changes in the external resistance from 10 to 1000 Ω [35] . In addition, considering that the current responses of the anodes during stable current generation were much stronger than that of the redox peaks with depleted substrate, it is reasonable to postulate that direct electron transfer was the main mechanism for current generation by the biofilms. 3.4 Influence on biofilm formation Fig. 4 displays SEM microphotographs of the surface of the anode biofilms grown under different external resistances. Although the morphology of the biofilm may have been altered by the dehydration process, the SEM results provided good comparative information demonstrating clear differences in the architectures of the biofilm established at various external resistances. It can be clearly seen from Fig. 3 a–d that the biofilms did not totally cover the electrode surface except for the biofilm developed at 10 Ω. For MFC-10, the entire anode surface was covered by the biofilm, converting the porous, rough electrode surface to a smooth, uniform one. The SEM images at higher magnification ( Fig. 4 e–h) provide further insight into the morphology of the biofilms formed at different external resistances. Compared with Fig. 4 f–h, Fig. 3 e indicates that the biofilm established at 10 Ω contained greater amounts of EPS than that of the biofilm developed at 50, 250 and 1000 Ω. This observation corroborates the EPS results seen in Fig. 2 b. The cross sectional morphology of the biofilm samples was also obtained in this study by SEM ( Fig. 5 ). SEM images show that the biofilms were only developed on the surface of the electrode and only few bacteria grew into the porous carbon cloth matrix. This could be attributed to both the small pore size and the limited substrate supply within the carbon cloth matrix. Similar phenomenon was also observed when reticulated vitreous carbon was used as anode support [31] . In addition, as can be seen in Fig. 5 , all the biofilms exhibited substantial heterogeneity in thickness. The average thicknesses of the biofilms formed under 10 ( Fig. 5 a) and 50 Ω ( Fig. 5 b) was significantly higher than that at 250 ( Fig. 5 c) and 1000 Ω ( Fig. 5 d). This result can be mainly attributed to the higher active biomass and EPS content in the biofilm samples established at 10 and 50 Ω. It is also interesting to note that although the biofilm formed at 10 Ω contained less active biomass than that at 50 Ω, both biofilms were of similar thickness. Putting these results together and recalling that the biofilm formed at 10 Ω contained more EPS than that at 50 Ω, one could assumed that the bacteria was more loosely packed in the biofilm developed at 10 Ω. In order to further confirm this point, the microstructures of the biofilms formed at different external resistances were observed in higher magnification ( Fig. 5 e–h). It can be clearly seen from these figures that the extensive voids were found to be distributed within biofilm established at 10 Ω. These voids were beneficial for mass transport within the biofilm by forming water channels that facilitate the substrate and buffer supply, as well as product removal. It should be pointed out, however, that the existence of void spaces within the biofilms would also lead to imperfect contact between the electroactive microorganisms and the anodes, causing a decrease in the electrical conductivity of the biofilm matrixes and consequently the lowered performance of the MFCs. By contrast, the biofilm formed at 50, 250 and 1000 Ω ( Fig. 5 f–h) appeared quite homogeneous throughout its thickness and few voids can be discerned. Based on the results from the surface and cross-sectional SEM images, it is apparent that the biofilms developed at 50, 250 and 1000 Ω showed a compact structure, in which cells are tightly linked together by visible polymeric viscous materials. 3.5 Influence on electricity generation Fig. 6 compared the polarization curves of the MFCs (10 kΩ to 5 Ω) started up at different external resistances. The details of the MFC performance were summarized in Table 1 . As illustrated in Fig. 6 and Table 1 , the maximum power density of the MFCs decreased in the following order: MFC-50 > MFC-250 > MFC-10 > MFC-1000. The best performance of MFC-50 may come from the highest active biomass production ( Fig. 2 a) on the surface of the carbon paper. The lower performance of MFC-10, despite the higher active biomass, might be due to the significant ohmic loss ( Table 1 ) resulting from the existence of void spaces in the interior of the biofilm. More importantly, it is clear from Fig. 6 b that there were steep falls of cell voltage at relatively high current densities in the case of MFC-50, MFC-250 and MFC-1000, whereas a gradually decrease in the cell voltage with the increase of current density can be noted for MFC-10. The experiments were repeated three times with similar results. The rapid fall of cell voltage at high current densities is a common phenomenon for both single [36–38] and stacked MFCs [39] . In order to further study this phenomenon, the variation of the voltage of the MFCs with step-change of external resistance during performance evaluation were recorded and shown in Fig. 7 . As can be seen in Fig. 7 , a slow decrease of voltage with time followed by a steep fall of voltage output was observed when an external resistance of 150, 50 and 20 Ω was connected to the electrical circuit of MFC-1000, MFC-250 and MFC-50, respectively. Whereas in the case of MFC-10, a stable voltage was reached within 20 min even when the external resistance as low as 5 Ω was applied. As a consequence, MFC-10 exhibits a highest maximum sustainable current density ( I max ) of 8.49 ± 0.19 A m −2 ( Table 1 ), while MFC-50, MFC-250 and MFC-1000 reached a limit of 5.78 ± 0.39, 2.46 ± 0.02 and 1.25 ± 0.01 A m −2 , respectively. The observed differences can be explained by the improved buffer access and proton removal in the case of the biofilm developed at 10 Ω. It is well known that the electrochemical oxidation of acetate is an acid-producing process. In the anode compartment, acetate is oxidized according to the following reaction: C H 3 COO − + 2 H 2 O → 2 C O 2 + 7 H + + 8 e − As shown in the above equation, the oxidation of every mole of acetate produces 7 mol of proton, which can result in a pH decrease within the biofilm microenvironment. This can lead to an increase in the potential for acetate oxidation and an inhabitation of microbial metabolism, both negatively influencing the MFC performance [40] . When a MFC was operated at a high current density or a low resistance, this problem becomes more serious due to the higher proton production rate. For MFC-50, MFC-250 and MFC-1000, the compact structure of the biofilms ( Fig. 5 f–h) hindered the buffer access and proton removal, leading to gradual accumulation of proton in the interior of the biofilm. When the proton concentration within the biofilm exceeded a critical value, the current generation of the MFCs became unsustainable over time due to the inhibition of the electrochemically active bacteria. Correspondingly, steep drops in cell voltage were observed ( Fig. 7 a–c). By contract, the biofilm formed at 10 Ω has a looser structure ( Fig. 5 e) with void spaces within the biofilm. These characteristics facilitated water and buffer access as well as proton removal. As a result, the highest sustainable current can be obtained by MFC-10 due to the maintenance of the favorable microenvironment for the microorganisms within the anode biofilm. 4 Conclusions In the present study, the effects of external resistance on the biofilm formation and electricity generation were investigated. Structural and morphological characterizations of the biofilms formed at various external resistances indicate that a suitable biofilm structure is crucial for the maximum power density and sustainable current generation of the MFCs. The experimental results showed that the biofilm established at a higher external resistance tended to form a compact structure with a less active biomass and a lower EPS content. This led to an increase of the maximum power density with the decrease of external resistance. However, when the external resistance became lower than the optimum value, a loose biofilm structure with a less active biomass and a higher EPS content was developed, resulting in more void spaces beneficial for proton and buffer transport and reduction of the electrical conductivity within the biofilm. Therefore, a lower maximum power density and higher maximum sustainable current generation can be obtained. Acknowledgements This work was supported by the National Natural and Science Foundation of China (Nos. 50806087 and 50825602 ), and Fundamental Research Funds for the Central Universities (No. CDJXS11142230 ). The authors also wish to thank Ms. Lu at the Central laboratory of Third Military Medical University for assistance in SEM measurements. References [1] B.E. Logan B. Hamelers R. Rozendal U. Schröder J. Keller S. Freguia P. Aelterman W. Verstraete K. Rabaey Environ. Sci. Technol. 40 2006 5181 5192 [2] K. Rabaey W. Verstraete Trends Biotechnol. 23 2005 291 298 [3] S. Freguia K. Rabaey Z. Yuan J. Keller Environ. Sci. Technol. 41 2007 2915 2921 [4] P. Aelterman S. Freguia J. Keller W. Verstraete K. Rabaey Appl. Microbiol. Biotechnol. 78 2008 409 418 [5] X. Cao X. Huang N. Boon P. Liang M. Fan Electrochem. Commun. 10 2008 1392 1395 [6] H. Liu R. Ramnarayanan B.E. Logan Environ. Sci. Technol. 38 2004 2281 2285 [7] S. You Q. Zhao J. Zhang J. Jiang C. Wan M. Du S. Zhao J. Power Sources 173 2007 172 177 [8] Q. Deng X. Li J. Zuo A. Ling B.E. Logan J. Power Sources 195 2010 1130 1135 [9] Y. Feng Q. Yang X. Wang B.E. Logan J. Power Sources 175 2010 1841 1844 [10] H. Liu S. Cheng L. Huang B.E. Logan J. Power Sources 179 2008 274 279 [11] U. Schröder J. Nießen F. Scholz Angew. Chem. Int. Ed. 115 2003 2986 2989 [12] P. Aelterman M. Versichele M. Marzorati N. Boon W. Verstraete Bioresour. Technol. 99 2008 8895 8902 [13] G.C. Gil I.S. Chang B.H. Kim M. Kim J.K. Jang H.S. Park H.J. Kim Biosens. Bioelectron. 18 2003 327 334 [14] G.S. Jadhav M.M. Ghangrekar Bioresour. Technol. 100 2009 717 723 [15] T. Catal P. Kavanagh V. O’Flaherty D. Leech J. Power Sources 196 2011 2676 2681 [16] R.C. Wagner D.F. Call B.E. Logan Environ. Sci. Technol. 44 2010 6036 6041 [17] D.A. Finkelstein L.M. Tender J.G. Zeikus Environ. Sci. Technol. 40 2006 6990 6995 [18] X. Wang Y. Feng N. Ren H. Wang H. Lee N. Li Q. Zhao Electrochim. Acta 54 2009 1109 1114 [19] J. Wei P. Liang X. Cao X. Huang Environ. Sci. Technol. 44 2010 3187 3191 [20] H. Liu S. Cheng B.E. Logan Environ. Sci. Technol. 39 2005 658 662 [21] D.Y. Lyon F. Buret T.M. Vogel J.M. Monier Bioelectrochemistry 78 2010 2 7 [22] H.R. Rismani-Yazdi A.D. Christy S.M. Carver Z. Yu B.A. Dehority O.H. Tuovinen Bioresour. Technol. 102 2011 278 283 [23] J.S. Mclean G. Wanger Y.A. Gorby M. Wainstein J. Mcquaid S.I. Ishii O. Bretschger H. Beyenal K.H. Nealson Environ. Sci. Technol. 44 2010 2721 2727 [24] G.C. Premier J.R. Kim I. Michie R.M. Dinsdale A.J. Guwy J. Power Sources 196 2011 2013 2019 [25] R.H. Findlay G.M. King L. Watling Appl. Environ. Microbiol. 55 1989 2888 2893 [26] H.H.P. Fang L.C. Xu K.Y. Chan Water Res. 36 2002 4709 4716 [27] A.F. Gaudy Ind. Water Wastes 7 1962 17 22 [28] O.H. Lowry N.J. Rosebrough A.L. Farr R.J. Randall J. Biol. Chem. 193 1951 265 275 [29] U. Schröder Phys. Chem. Chem. Phys. 9 2007 2619 2629 [30] N. Qureshi B.A. Annous T.C. Ezeji P. Karcher I.S. Maddox Cell. Fact 4 2005 24 44 [31] C.I. Torrres R. Krajmalnik-brown P. Parameswaran A. Kato Marcus G. Wanger Y.A. Gorrby B.E. Rittmann Environ. Sci. Technol. 43 2009 9519 9524 [32] H. Yi K.P. Nevin B.C. Kim A.E. Franks A. Klimes L.M. Tender D.R. Lovely Biosens. Bioelectron. 24 2009 3498 3503 [33] A. Larrosa-Guerrero K. Scott K.P. Katuri C. Godinez I.M. Head T. Curtis Appl. Microbiol. Biotechnol. 87 2010 1669 1713 [34] J.R. Kim S.H. Jung J.M. Regan B.E. Logan Bioresour. Technol. 98 2007 2568 2577 [35] S.A. Patil F. Harnisch B. Kapadnis U. Schröder Biosens. Bioelectron. 26 2010 803 808 [36] B. Min Ó.B. Román I. Angelidaki Biotechnol. Lett. 30 2008 1213 1218 [37] D.P.B.T.B. Strik H. Terlouw H.V.M. Hamelers C.J.N. Buisman Appl. Microbiol. Biotechnol. 81 2008 659 668 [38] A. Heijne H.V.M. Hamelers M. Saakes C.J.N. Buisman Electrochim. Acta 53 2008 5697 5703 [39] P. Aelterman K. Rabaey H.T. Pham N. Boon W. Verstraete Environ. Sci. Technol. 40 2006 3388 3394 [40] C.I. Torres A.K. Marcus B.E. Rittmann Biotechnol. Bioeng. 100 2008 872 881

A graphene modified anode to improve the performance of microbial fuel cells

Performance of microbial fuel cell with graphene doped bioanode

Carbon nanotube modified air-cathodes for electricity production in microbial fuel cells

Capacitance-enhanced 3D Graphene Anode for Microbial Fuel Cell with Long-Time Electricity Generation Stability

A graphene/poly(3,4-ethylenedioxythiophene) hybrid as an anode for high-performance microbial fuel cells

Power Generation and Electrochemical Analysis of Biocathode Microbial Fuel Cell Using Graphite Fibre Brush As Cathode Material

Modification of carbon based cathode electrode in a batch-type microbial fuel cells

Eighteen-Month Assessment of 3d Graphene Oxide Aerogel-Modified 3d Graphite Fiber Brush Electrode As A High-Performance Microbial Fuel Cell Anode

Biotic Cathode of Graphite Fibre Brush for Improved Application in Microbial Fuel Cells

Biofilm formation and electricity generation of a microbial fuel cell started up under different external resistances

Performance evaluation of microbial fuel cell fabricated using green nano-graphene oxide as coating anode material

Facile Fabrication of Graphene-Containing Foam as a High-Performance Anode for Microbial Fuel Cells

High-performance Anode Material Based on S and N Co-Doped Graphene/iron Carbide Nanocomposite for Microbial Fuel Cells

Performance Analysis of different Anode Materials of a Double Chambered Microbial Fuel Cell

Three-dimensional Graphene/pt Nanoparticle Composites As Freestanding Anode for Enhancing Performance of Microbial Fuel Cells

Optimizing microbial fuel cells performance: An innovative approach integrating anode materials, dual-pollutant treatment, and long-term operation

Performance of Two Different Types of Anodes in Membrane Electrode Assembly Microbial Fuel Cells for Power Generation from Domestic Wastewater

Thermally Expanded Graphite Incorporated with PEDOT:PSS Based Anode for Microbial Fuel Cells with High Bioelectricity Production

Enhanced electroactive bacteria enrichment and facilitated extracellular electron transfer in microbial fuel cells via polydopamine coated graphene aerogel anode

Carbon nanofiber/graphene hybrids anchoring Fe, N, S heteroatoms simultaneously enhancing extracellular electron transfer and biofilm adhesion in microbial fuel cells

Tubular Membrane Cathodes for Scalable Power Generation in Microbial Fuel Cells.