Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Daniel R. Noguera, Department of Civil and Environmental Engineering, University of Wisconsin-Madison, 3216 Engineering Hall, 1415 Engineering Drive, Madison, WI 53706, USA. E-mail email@example.com
Aims: To understand factors that impact solar-powered electricity generation by Rhodobacter sphaeroides in a single-chamber microbial fuel cell (MFC).
Methods and Results: The MFC used submerged platinum-coated carbon paper anodes and cathodes of the same material, in contact with atmospheric oxygen. Power was measured by monitoring voltage drop across an external resistance. Biohydrogen production and in situ hydrogen oxidation were identified as the main mechanisms for electron transfer to the MFC circuit. The nitrogen source affected MFC performance, with glutamate and nitrate-enhancing power production over ammonium.
Conclusions: Power generation depended on the nature of the nitrogen source and on the availability of light. With light, the maximum point power density was 790 mW m−2 (2·9 W m−3). In the dark, power output was less than 0·5 mW m−2 (0·008 W m−3). Also, sustainable electrochemical activity was possible in cultures that did not receive a nitrogen source.
Significance and Impact of the Study: We show conditions at which solar energy can serve as an alternative energy source for MFC operation. Power densities obtained with these one-chamber solar-driven MFC were comparable with densities reported in nonphotosynthetic MFC and sustainable for longer times than with previous work on two-chamber systems using photosynthetic bacteria.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
A predominantly untapped resource for powering future generations of MFC is sunlight. With more solar energy striking the earth in 1 h (4·3 × 1020 J) than all the energy consumed on the planet in a year (4·1 × 1020 J) (US Department of Energy 2005b), photosynthetic microbes highly adapted to capture this solar energy, and the sheer abundance of photosynthetic microbes on the planet the technological development of solar-driven MFC has potential as an alternative way to capture energy from renewable resources. In principle, the ability of cyanobacteria to produce hydrogen via water bio-photolysis (Melis 2002) or hydrogen production by direct electron transfer to protons in photosynthetic purple nonsulfur bacteria (Gest and Kamen 1949; Koku et al. 2002) are attractive alternatives for the development of solar-powered MFC. Strategies have been described in which the hydrogen produced in such bioreactors can be collected (Nath and Das 2004) before sending the collected gas to a separate fuel cell (He et al. 2005a). However, a direct coupling of hydrogen production and electricity generation could also be achieved within an MFC; in such a system the hydrogen produced will react at a catalytic anodic surface. Such a direct coupling has been demonstrated with dark fermentations (Niessen et al. 2005) as well as with photosynthetic reactors (Rosenbaum et al. 2005a; b). More recently, evidence for the presence of nanowires in cyanobacteria has also been presented (Gorby et al. 2006), suggesting the possibility of developing photosynthetic MFC reactors that do not depend on hydrogen production for electricity generation.
The purpose of this study was to develop an understanding of factors that impact solar-powered electricity generation by Rhodobacter sphaeroides strain 2·4·1, which is a well-studied facultative phototroph with extensive structural, functional, and genomic information on solar energy capture (Blankenship et al. 1995).
Materials and methods
Description of photosynthetic MFC system
Experiments were conducted in single-chamber MFC constructed in 30-ml glass test tubes or scaled-up 500- or 700-ml reactors (Fig. 1). In the simplest configuration (Fig. 1a), the anode was submerged in the microbial culture and the cathode sealed the top of a test tube. A second configuration (Fig. 1b,c) used a side arm sealed with the cathode, while the top of the vessel was sealed with a rubber stopper.
The anode was a rectangular piece of either platinum-coated phosphoric acid fuel cell electrode on Toray carbon paper (0·35 mgPt cm−2; E-Tek, Somerset, NJ, USA) or plain Toray carbon paper (E-Tek) that did not contain platinum. The cathode was also made of platinum-coated Toray carbon paper. In most experiments, the anode and cathode were connected through a 10 000 Ohm external resistance. The surface areas of the anode and cathode were 5 cm2 (unless noted otherwise) and 1·7 cm2 for test tube MFC, and 16 cm2 and 20 cm2 for the scaled-up MFC, respectively.
The biogas produced by the cultures was vented out through a needle placed at the top of the reactors and connected to either a U-shaped tube filled with water to prevent oxygen from diffusing back into the MFC (in test tube MFC) or collected in airtight gas bags (Jesen Inert Products, Coral Springs, FL, USA). When necessary, sterile media without any organic carbon was added to the reactors to maintain a constant culture volume.
All experiments were conducted with R. sphaeroides strain 2·4·1. Cultures were grown inside the MFC reactors under anaerobic photosynthetic conditions, using Sistrom’s minimal medium (Sistrom 1960) with succinate or propionate as carbon sources, and ammonia, nitrite, nitrate, or glutamate as potential nitrogen sources. The cultures were placed in front of an incandescent light source (10 W m−2, as measured with a Yellow–Springs–Kettering model 6·5-A radiometer through a Corning 7–69, 620- to 110-nm filter). To test the effect of light on function of the MFC, parallel cultures were pregrown photosynthetically, and then amended with additional carbon, placed in the dark, and monitored for power output.
The voltage drop across the external resistance (V) was measured and logged at 5-min intervals using a computer-controlled digital multimeter (DMM PCI-4070; National Instruments, Austin, TX, USA) combined with a data input/output card (PCI-6518; National Instruments) and a relay system that facilitated on-line measurements of up to eight MFC operated in parallel. LabVIEW®- based software (National Instruments) was used as a graphical interface for data handling. The response variables derived from these measurements were current (I) and power (P) generated through the circuit, as well as current and power densities calculated per unit area of anode surface (A), or per unit volume of microbial culture (VL). Current was calculated according to Ohm’s law (I = V/R, where R is the external resistance), and power was estimated as P = V2/R.
To generate polarization curves, the external circuits were disconnected and the MFC devices were allowed to stabilize to an open circuit potential. Next, the external resistance was varied from 100 000 Ohms to 10 Ohms at discrete intervals. At each condition, voltage readings were taken once the voltage drop had reached an equilibrium condition, which occurred a few minutes after the replacement of the external resistance. The internal resistance in MFC devices was calculated from the slope of the linear region of the polarization curves (Logan et al. 2006).
Other analytical methods
Ammonium was measured by the salicylate method using a Test N’ Tube kit (Hach, Loveland, CO, USA). Succinate was measured by high-performance liquid chromatography (HPLC), with a Prevail organic acid column (Altech, Flemington, NJ, USA) and ultraviolet (UV) detection at 210 nm. The composition of the biogas was analysed by gas chromatography using a Shimadzu GC-8A system equipped with a thermal conductivity detector and a stainless steel column packed with Carbosieve SII (Supelco, Bellefonte, PA, USA). Helium was used as the carrier gas and the temperatures for the injector, column, and detector were 150, 100 and 150°C, respectively. Bacterial growth was measured by optical density (OD) at 600 nm.
Results and discussion
Preliminary experiments with MFC in which ammonia was used as the nitrogen source (3·8 mmol l−1) showed minimum power generation in batch cultures, but when the cultures entered stationary phase (after c. 2 days of incubation) and an organic carbon source was added, biogas formation and power generation were noticeable, suggesting a correlation between biogas and electricity generation. Analysis of the biogas indicated that hydrogen and carbon dioxide were the main gases produced, with hydrogen corresponding to 68–78% of the total. MFC placed in the dark immediately after the addition of organic substrate to the stationary phase culture resulted in insignificant power densities (less than 0·5 mW m−2) in comparison with the power densities observed when the cultures were exposed to light (c. 55–65 mW m−2). In addition, experiments in the dark failed to accumulate biogas, providing further evidence for the connection between biogas production and electricity generation by these solar-powered MFC. These preliminary observations were consistent with light-dependent hydrogen formation, which occurs in R. sphaeroides and related photosynthetic purple nonsulfur bacteria under nitrogen-limiting conditions, which is reported to be predominantly catalysed by nitrogenases (Zurrer and Bachofen 1979; Nandi and Sengupta 1998). Measurements of ammonium concentration in these cultures revealed that nitrogen-limiting conditions (i.e. [N] < 0·2 mmol l−1) were established after 2 days of growth.
The coupling of hydrogen production with electricity generation is also consistent with in situ hydrogen oxidation in the anode of the MFC, as described by Rosenbaum et al. (2005a,b) with photosynthetic cultures. To confirm this mechanism of electricity generation, we performed light-exposed experiments in which the platinum-coated anode was replaced by a similar-sized piece of plain carbon paper. Under these conditions, the power output was less than 0·01 mW m−2, which is insignificant compared with the power densities obtained when the anode-contained platinum.
In the MFC literature, other reported mechanisms for the transfer of electrons from micro-organisms to the anodic surface include membrane complex-mediated electron transfer (Bond and Lovley 2003), the presence of mobile redox mediators (Rabaey et al. 2005), and the direct transfer of electrons through nanowires (Reguera et al. 2005). It is also reported that in many MFC acclimation is achieved by the development of a biofilm directly at the anodic surface (Kim et al. 2005). In our experiments with R. sphaeroides strain 2·4·1, we did not observe any biofilm development either on the electrodes or at the glass surfaces. Thus, as a direct contact between the cells and the anode was not achieved, and our experiments in which the platinum-coated anode was replaced by a similar-sized piece of plain carbon paper resulted in insignificant power outputs, our experiments do not show any evidence for the existence of electron transfer mechanisms other than hydrogen production and its in situ oxidation.
Effect of nitrogen source on power output
As the preliminary experiments with ammonia as a nitrogen source suggested that nitrogen-limited conditions were needed for electricity generation, and this observation agrees with the current understanding of hydrogen production by photoheterotrophic bacteria (Vignais et al. 1985; Yokoi et al. 1998; Zhu et al. 2001; Takabatake et al. 2004; Kapdan and Kargi 2006), we then investigated the effect of the nitrogen source on MFC performance. Figure 2 shows a comparison of voltage generation when nitrogen was added as ammonium, nitrite, nitrate, or glutamate (3·8 mmol l−1 initial concentration), and when a nitrogen source was not provided. In these experiments, succinate (12 mmol l−1) was the organic electron donor. When ammonium was provided in the culture media, cells grew photosynthetically, but electricity was not generated over the duration of the experiment. This is consistent with our preliminary observations described before. In contrast, when other nitrogen sources were used or when an external nitrogen source was absent, a voltage surge occurred after a lag phase. The period of high electrochemical activity was somewhat dependent on the nitrogen source. With glutamate, high voltage was maintained for about 32 h, while with nitrate or when nitrogen was not added to the growth media, the period of high electrochemical activity could be maintained for up to 48 h. When nitrite was provided in the medium, the lag phase was the longest, but the period of high electrochemical activity was about 30 h. In all cases, the voltage during the period of high electrochemical activity was relatively constant, and suddenly dropped, corresponding to the depletion of the organic carbon source.
Figure 2 also shows an abiotic control experiment in which an MFC was assembled but not inoculated. Instead, hydrogen gas was continuously bubbled inside the MFC chamber. This experiment shows that electrochemical activity was initiated as soon as the hydrogen gas was introduced, and that the activity was maintained for as long as the hydrogen was being bubbled into the reactor.
Parallel to the measurements of voltage shown in Fig. 2, we also measured microbial growth on the electrochemically active MFC, as shown in Fig. 3. This figure shows that glutamate was used as a nitrogen source and it supported relatively rapid growth compared with the other MFC. In addition, the OD curves show that when nitrite or nitrate was present in the medium, the culture density increase was similar to that observed when no external nitrogen source was added.
The participation of nitrogenases on hydrogen formation by purple nonsulfur bacteria has been well established (Gest and Kamen 1949). Nitrogenases are required for nitrogen fixation, and therefore, their activity is inhibited by the availability of ammonium. Yokoi et al. (1998) reported that hydrogen production by R. sphaeriodes was completely inhibited at ammonium concentrations above 2 mmol l–1. The presence of a poor nitrogen source like glutamate has been shown not to affect nitrogenase expression and not to inhibit hydrogen production (Takabatake et al. 2004). The growth curve in Fig. 3 shows that high cell densities where achieved when glutamate was the nitrogen source, and the MFC experiment (Fig. 2) demonstrated that glutamate was an effective alternative nitrogen source to allow sustained photosynthetic MFC performance.
The observations when nitrate or nitrite was added to the medium were surprising, as R. sphaeroides 2·4·1 is reported not to contain an assimilatory nitrate reductase and is not capable of performing denitrification (Mackenzie et al. 2001). Thus, we anticipated that cultures with nitrate or nitrite would behave similarly to cultures not receiving any nitrogen in the media. However, although the growth curves reflected similar patterns of growth in the presence of nitrate, nitrite, and in the absence of a nitrogen source, the presence of nitrite and nitrate resulted in longer lag phases, which suggest a heretofore uncharacterized effect of oxidized nitrogen forms on hydrogen production by R. sphaeroides.
Evaluating the degradation in catalytic activity in the anode
A common feature of the R. sphaeroides MFC experiments shown in Fig. 2 is the very slow decline in voltage during the period of high power production, while such decline is not observed in the abiotic control. This phenomenon has been reported to occur in MFC that use platinum-based anodes for in situ hydrogen utilization (Schroder et al. 2003; Rosenbaum et al. 2005a). Schroder et al. (2003) attributed this activity decrease to the poisoning of the catalyst by microbial byproducts and products of electrocatalytic oxidation, and described the development of polymer-coated platinized electrodes to overcome the problem. In addition, they suggested a regenerative procedure in which voltage pulses were applied to the electrodes to reverse their deactivation (Schroder et al. 2003). In R. sphaeroides MFC experiments performed by Rosenbaum et al. (2005a) using glutamate as a nitrogen source, polymer coating was essential to achieve sustainable electrocatalytic activity of the MFC. However, anodes coated with polyaniline showed rapid decline in electrocatalytic activity presumably because polyaniline degradation released ammonium into the medium. With an anode coated with poly(3,4-ethylenedioxythiophene), they were able to slow down the decline in electrocatalytic activity and achieve a somewhat sustainable power generation until the depletion of the organic substrate c. 30 h after the initiation of the experiment. Curiously, the latter observations are similar to the glutamate experiments shown in Fig. 2 although a polymer coating was not needed in our case.
Thus, the discrepancies in the two studies with respect to potential catalyst poisoning during in situ hydrogen oxidation prompted further experiments to investigate whether the steady decline in voltage was because of catalyst degradation or the aging of the photosynthetic culture. Figure 4 shows two 3-week experiments that provide additional insights into the catalytic activity during in situ hydrogen oxidation within the MFC. First, an abiotic control experiment was conducted, as described before, by bubbling hydrogen gas into an uninoculated MFC reactor. Hydrogen bubbling was implemented during the first 5 days, stopped for 7 days, and then resumed for an additional period of 5 days, and as the figure shows, high electrochemical activity was observed during the hydrogen bubbling periods. During the first bubbling period, a slow decline in voltage was evident, but this observation was not reproduced during the second period of gas bubbling, which occurred 12 days after setting up the uninoculated MFC reactor. Thus, we conclude from this experiment and from the abiotic experiments described in Fig. 2 that there was no catalytic poisoning of the electrodes in the absence of bacterial activity.
In the second experiment shown in Fig. 4, we first operated a photosynthetic MFC for 5 days and observed the typical pattern during one cycle of operation, with an initial lag phase, followed by a rapid increase in voltage when hydrogen formation started, a period of high voltage, and then a final voltage reduction when the organic substrate was depleted. The period of high voltage was not as defined as in Fig. 2, likely because we employed the larger vessel MFC for these experiments (Fig. 1c). To test whether irreversible catalytic poisoning occurred in the anode during the initial cycle, a second cycle was performed by transferring the used anode to a new MFC vessel that contained a new cathode, fresh medium, and new R. sphaeroides inoculum. This second cycle of operation revealed the same pattern as in the initial cycle, with voltage peaking at similar levels as during the first cycle, hence showing no evidence of catalytic poisoning of the anode. After the conclusion of the second cycle (end of day 12), the culture was replaced with fresh medium containing no organic substrate, and hydrogen gas bubbling was implemented. Under these conditions, the system mirrored the abiotic experiment, reaching voltages higher than that observed during the earlier biotic cycles, and this voltage was stable until the conclusion of the experiment. Thus, opposite from the findings of Rosenbaum et al. (2005a), we conclude that catalytic poisoning of the anode is not a reason for the observed decline in performance in R. sphaeroides MFC, and therefore, that protecting the anode with polymer coatings is not needed. Alternative reasons for the slow degradation in catalytic activity during high performance periods could be aging of the culture, which can result in lower metabolic activity possibly because of photo-induced damage of the photosynthetic reaction center (Anthony et al. 2005), and inefficient light penetration by shadowing at high cell densities, both of which will eventually result in decreasing hydrogen production rates. Possible evidence for the shadowing effect might be seen when comparing Figs 2 and 4, which depict MFC experiments conducted in narrow test tube MFC and the larger and wider MFC vessels, respectively. The rate of voltage decline in the test tube MFC was lower compared with that observed in the larger vessels, which agrees with efficient light penetration in the test tubes, and a less than optimal light exposure in the larger MFC design. However, the scaling up of the reactors may introduce additional factors beyond the shadowing effect, so a firm conclusion about shadowing cannot be stated.
Effect of MFC configuration on power output
In order to investigate the range of power densities achievable with R. sphaeroides MFC, experiments were conducted with varying distances between the electrodes and with electrodes differing in anode size. In single-chamber MFC, the distance between the anode and the cathode has been shown to significantly affect power output. When the electrodes are further apart, ohmic losses restrict performance, but when they are very close to each other MFC performance can be compromised if oxygen diffusing through the cathode reaches the anode. For instance, Cheng et al. (2006) found that reducing the spacing between electrodes to less than 2 cm negatively impacted the power output in their MFC. To test the effect of electrode spacing, experiments were conducted in test tube photosynthetic MFC by placing the anode at different distances from the cathode. The polarization curves presented in Fig. 5 (obtained when the cultures were under nitrogen-limiting conditions and exhibiting their maximum voltage output), demonstrate increased power generation as the spacing between the electrodes was reduced, with a maximum power density point of 170 mW m−2 obtained when the spacing between the centre of the electrodes was 3 cm and the external resistance was 510 Ohms. On a volumetric basis, the maximum power density in this configuration was 2·8 W m−3. As the shape of the polarization curves in Fig. 5 shows a clear differentiation between the slopes representative of the activation and ohmic losses (Logan et al. 2006), the internal resistance in each MFC was calculated from the slope of the linear region representing the ohmic losses. For the configuration with the largest distance between electrodes (i.e. centre of electrodes 12·5 cm apart) the internal resistance was calculated to be 1750 Ohms, but with the smallest electrode spacing (i.e. 3 cm apart), the internal resistance reduced to 510 Ohms.
Thus, the gain in power output can be attributed to the decrease in the internal resistance as the spacing between electrodes was reduced, while the potentially negative effect of oxygen diffusion through the cathode was not observed. As our experiments were performed in pure culture and R. sphaeroides 2·4·1 did not form biofilms, oxygen diffusion into the MFC could have been minimized by aerobic respiration of planktonic R. sphaeroides cells located near the cathode.
Another factor to consider is the relative ratio of anode to cathode surface areas, as this has also been shown to affect power generation in MFC (Oh and Logan 2006). The effect was demonstrated in dual-chamber MFC designs, where the surface area of the proton exchange membrane also had a significant impact on power output (Oh and Logan 2006). However, the effect of the surface area ratio in single-chamber MFC systems with air cathodes and without a proton exchange membrane has not been reported. Thus, to explore the impact of the anode to cathode surface area ratio in the power output of our single-chamber solar-powered MFC devices, we performed experiments with anodes having surface areas of 1·25, 2·5, or 5 cm2, while maintaining the surface area of the cathode constant (Fig. 6). In these experiments, the spacing between electrodes was kept as small as possible to minimize ohmic losses, and the polarization curves were constructed when the MFC were exhibiting their maximum voltage output, as described for the experiments in Fig. 5. Figure 6 shows that the maximum point power density in the solar-powered MFC systems increased as the size of the anode was reduced, which suggests that the anodic reaction is not the limiting step in these devices. The combination of the smallest anode surface and the shortest distance between the electrodes produced the best power density outputs observed so far with any solar-driven MFC design. The maximum power density point obtained was 700 mW m−2, which occurred with an external resistance of 510 Ohms. On a volumetric basis, this maximum output was 2·9 W m−3. In addition, the internal resistance in this MFC was reduced to 130 Ohms (based on the slope of the polarization curve), which is an improvement over the internal resistance calculated from the experiments shown in Fig. 5. These internal resistance values are orders of magnitude higher than observed in optimized MFC configurations (He et al. 2005b; Cheng et al. 2006), suggesting that power output in solar-powered MFC devices could be enhanced with next generation reactor designs that maximize proton mass transport.
Comparison with other MFC systems
With the large variety of MFC reactor designs, electrode materials, microbial catalysts, and operational conditions described in the literature, a unifying metric for comparison of MFC performance is difficult to extract. In the contemporary MFC literature, maximum point power densities normalized per projected surface area of the anode, or per total reactor volume (Logan et al. 2006) have been typically used. In our experiments, the best performance obtained corresponded to normalized power densities of 700 mW m−2 and 2·9 W m−3. We have demonstrated that power in these MFC devices is light-driven, as units incubated in the dark did not produce more than 0·5 mW m−2 (i.e. 0·008 W m−3 on a volumetric basis). Available data on the performance of other solar-powered MFC devices is particularly scarce (Yagishita et al. 1998; He et al. 2005a; Rosenbaum et al. 2005a). The only direct comparison that can be made is with the study of Rosenbaum et al. (2005a), who reported a volumetric power density of 7·3 W m−3 obtained on a two-chambered MFC, in which R. sphaeroides was used as the microbial catalyst and a solution of ferricyanide was used in the cathodic chamber. This volumetric power density is 2·5 times higher than that obtained in our MFC devices. However, Rosenbaum et al. (2005a) used a surface area to reactor volume ratio 10 times higher than what we used [i.e. 0.4 cm2 ml−1 in Rosenbaum et al. vs 0.04 cm2 ml−1 in this study), and when the maximum point power density is expressed on a surface area basis (i.e. 182 mW m−2), the performance of their system is 3·8 times lower than what we observed. Other studies have reported the use of purple nonsulfur bacteria for hydrogen generation, and the subsequent use of the hydrogen in an abiotic fuel cell (He et al. 2005a), or the use of cyanobacteria in MFC devices (Yagishita et al. 1998), but these studies did not characterize fuel cell performance in terms of maximum point power densities.
It is also important to compare the performance of our solar-powered MFC devices with other nonphotosynthetic single-chambered units reported in the literature. In the initial designs of this type of MFC devices by Logan and coworkers (Liu and Logan 2004; Liu et al. 2004, 2005), maximum point power densities of 26 mW m−2 were reported when domestic wastewater was used as an energy source (Liu et al. 2004), and as high as 494 mW m−2 when glucose was used as the carbon source (Liu and Logan 2004). More recently, power densities as high as 1540 mW m−2 (51 W m−3) have been reported for single-chambered cells in which advective flow through a porous anode and a minimal internal resistance were achieved (Cheng et al. 2006). Others have reported the optimization of a single-chamber MFC that used Mn4+-based anodes and Fe3+-based cathodes, reaching maximum power densities of up to 788 mW m−2 (Park and Zeikus 2003). As the performance of the solar-powered MFC devices described here compares well with other single-chambered units, it is reasonable to expect that modifications of the solar-powered MFC devices aimed at reducing the internal resistance, maximizing the rates of in situ hydrogen oxidation without affecting light penetration effectiveness, and improving electrode performance, as others, will likely result in a significant increase in the power density output.
Considerations for future applications of solar-driven MFC
Many alternative energy sources are being considered to offset society’s dependence on fossil fuels. Some of these may be viable options in the near future, while major technological advances are needed for others to make a significant impact on the overall energy budget. Harvesting solar energy is a long-term attractive strategy for meeting the global energy challenge. Solar energy use is also a carbon-neutral process that poses no known threat from pollution or greenhouse gases. Despite these advantages, sunlight provided less than 0·1% of the world’s electricity in 2001 (US Department of Energy 2005b). Another attractive alternative to fossil fuels is the capture of energy from renewable resources, and MFC offer a microbe-mediated technology with significant potential (Lane 2006). Our research on the development of solar-powered MFC is aimed at developing efficient reactors for simultaneously harvesting power from solar energy and renewable resources.
Purple nonsulfur bacteria such as R. sphaeroides are efficient in capturing solar energy when grown photosynthetically under anaerobic conditions and in the presence of an external carbon source. These metabolic requirements are consistent with the operation of MFC, in which the anodic chamber requires that anaerobic conditions be maintained for the efficient transfer of electrons from the micro-organisms to the anode, and that an external electron donor be provided to induce the biological activity that fuels the system. The main difference with respect to typical MFC described in the literature is that the reactor vessels need to allow sufficient light penetration. In our experiments, light penetration was maximized in the test-tube MFC by using tubes with a small diameter, and by restricting the size and location of the electrodes so that they did not block light.
In our experiments, the rate of hydrogen production was significantly higher than the rate of in situ hydrogen utilization, and therefore, most of the hydrogen produced was vented out of the reactors. Consequently, a calculation of Coulombic efficiencies was not relevant, as most of the energy recovered as hydrogen was wasted with the biogas. To increase in situ hydrogen oxidation, it would be necessary to substantially increase the surface area of the anode per unit of reactor volume. However, the material used in the anode was based on black carbon paper, and therefore, increasing anode surface area would result in a decrease in light penetration with the consequent decrease in light-driven hydrogen production. Thus, from a materials science perspective, improving the efficiency of photosynthetic MFC requires the development of anode materials that allow penetration of the near infrared light needed for photosynthesis by purple nonsulfur bacteria.
In addition, from a systems biology perspective, there is room for improving the microbial catalysts used in solar-powered MFC to maximize hydrogen production. When R. sphaeroides generates excess reducing power, it passes these electrons to one of the several pathways (Richardson et al. 1988), including polyhydroxybutyrate synthesis (which can account for up to 60% of the cellular dry weight), the Calvin cycle [assimilating carbon dioxide even if other carbon sources are present (Richaud et al. 1991; Paoli et al. 1998; Tichi and Tabita 2001)], hydrogen evolution (Gest and Kamen 1949), reduction of other electron acceptors (McEwan et al. 1987), or other uncharacterized pathways (Tavano et al. 2005). It may be possible to optimize solar-driven MFC function by removing systems that compete for reducing power, such as polyhydroxyalkanoate synthesis, which can be blocked by mutation (Kim et al. 2006).
The perspective of using solar-driven MFC for electricity generation from renewable resources has been advocated (Rosenbaum et al. 2005a). In particular, the use of a combination of dark and photo-fermentations holds promise for maximizing hydrogen production from complex organic compounds (Kapdan and Kargi 2006). In such processes, dark fermentation would be used for hydrolysis of complex organics and volatile fatty acid production, while the photo-fermentation would use the acids produced during dark fermentation for hydrogen production (Fascetti et al. 1998; Rosenbaum et al. 2005a; Kapdan and Kargi 2006). A significant limitation in this scheme is the inhibition of nitrogenase expression by high ammonium concentrations, as in many cases the anaerobic degradation of complex organic material will release high levels of ammonium (Fascetti et al. 1998). Thus, engineering reactor configurations that reduce the ammonium loads to solar-driven MFC are essential, some of which have been proposed for photosynthetic hydrogen production systems (Fascetti and Todini 1995; Fascetti et al. 1998). One strategy to reduce ammonium loads is the use of reactor configurations that include an aerobic stage for nitrification downstream of the anaerobic photosynthetic reactor, and the recycle of the nitrified stream back to the photosynthetic reactor, similar to conventional approaches for biological nitrogen removal during wastewater treatment, where nitrification is implemented downstream of the denitrification process (Rittmann and McCarty 2001). In this case, the photosynthetic reactor would receive the organic substrates from the upstream processes and nitrate from the downstream nitrification process, which according to our experiments, would not inhibit biological hydrogen production by R. sphaeroides. Alternatively, the development of photosynthetic mutants with constitutive hydrogen production in the presence of ammonium (Rey et al. 2007) is also a promising approach.
In conclusion, we have demonstrated that it is possible to operate single-chamber MFC that capture solar energy and simultaneously utilize organic renewable resources. Using a single chamber configuration was an important consideration, as molecular oxygen is ultimately the preferred electron acceptor for practical applications of MFC, because of its virtually unlimited availability, low cost, high oxidation potential, and lack of undesirable by-products (Logan et al. 2006; Zhao et al. 2006). In our single-chamber MFC, hydrogen was produced by R. sphaeroides and oxidized in situ on an anodic surface containing platinum as the catalyst, without the need for protective polymer coatings. In the initial reactor designs presented here, the rate of in situ hydrogen oxidation was much lower than the rate of hydrogen production, and therefore, most of the biogas produced was vented out of the system. In situ hydrogen oxidation could be maintained for up to 48 h, without any evidence of inhibition of the electrocatalytic anodic reactions. Power densities obtained with these solar-driven MFC were comparable with densities reported by others using either photosynthetic or nonphotosynthetic cultures. Based on the steady improvements reported for nonphotosynthetic MFC over the last few years (Rabaey and Verstraete 2005; Logan et al. 2006), it is reasonable to expect that improvements in reactor configuration, electrode materials, and microbial catalysts specifically geared at maximizing solar energy capture and rates of electrocatalytic activities, while at the same time minimizing ohmic resistances, could lead to significant improvements on the performance of solar-driven MFC. Finally, the long-term goal of using solar-driven MFC to tap energy resources in organic waste materials necessitates the development of strategies to overcome nitrogenase inhibition by ammonium or engineering reactor configurations that minimize ammonium input to the MFC.
The authors recognize Rodolfo Perez for helping in the configuration and automation of the data acquisition system used in this research. Partial funding was provided by the Wisconsin Alumni Research Foundation (WARF) under project grant 135-GH56, and by a gift from Associated Engineering (Burnaby, British Columbia, Canada) to D.R. Noguera.