Article pubs.acs.org/est
Temporal-Spatial Changes in Viabilities and Electrochemical Properties of Anode Biofilms Dan Sun,† Shaoan Cheng,*,† Aijie Wang,‡ Fujian Li,† Bruce E. Logan,§ and Kefa Cen† †
State Key Laboratory of Clean Energy Utilization, Department of Energy Engineering, Zhejiang University, Hangzhou 310027, P.R. China ‡ Key Laboratory of Environmental Biotechnology, Research Center for Eco-Environmental Sciences, China Academy of Sciences, Beijing, China § Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States S Supporting Information *
ABSTRACT: Sustained current generation by anodic biofilms is a key element for the longevity and success of bioelectrochemical systems. Over time, however, inactive or dead cells can accumulate within the anode biofilm, which can be particularly detrimental to current generation. Mixed and pure culture (Geobacter anodireducens) biofilms were examined here relative to changes in electrochemical properties over time. An analysis of the three-dimensional metabolic structure of the biofilms over time showed that both types of biofilms developed a live outer-layer that covered a dead inner-core. This two-layer structure appeared to be mostly a result of relatively low anodic current densities compared to other studies. During biofilm development, the live layer reached a constant thickness, whereas dead cells continued to accumulate near the electrode surface. This result indicated that only the live outer-layer of biofilm was responsible for current generation and suggested that the dead inner-layer continued to function as an electrically conductive matrix. Analysis of the electrochemical properties and biofilm thickness revealed that the diffusion resistance measured using electrochemical impedance spectroscopy might not be due to acetate or proton diffusion limitations to the live layer, but rather electron-mediator diffusion.
■
of ∼5 mS/cm measured for G. sulf urreducens biofilms.10,11,13 There is currently no consensus on which of these two models is correct. Therefore, additional insights into the structure of the biofilm relative to its electrochemical properties may help further our understanding of the mechanism of electron transfer in these biofilms, as well as lead to improvements in BES performance.12 The growth of bacteria in thick electrogenic biofilms can be affected by biofilm conductivity and rates of diffusion of substrates and end products.6,9,14−18 Three different conditions have been considered important relative to bacteria location within the biofilm (i.e., nearer the anode interface or solution interface) and metabolic activity. First, if the biofilm electrical conductivity limits growth, then the metabolic activity of the outer cells (at the solution interface) could become restricted by long distance EET due to charge transfer resistances. In this case, the metabolic activity would decrease with distance from
INTRODUCTION Bioelectrochemical systems (BESs), such as microbial fuel cells (MFCs) used for electricity generation and microbial electrolysis cells (MECs) for hydrogen production, are technologies in which exoelectrogens generate electrical current by oxidizing organic matter and using the anode as their terminal electron acceptor.1 Anodic microorganisms, such as Geobacter sulfurreducens and Geobacter anodireducens,2 as well as other bacteria in mixed microbial communities, attach to the anode surface and form a thick biofilm.3 Long-distance extracellular electron transfer (EET) occurs in these thick biofilms at distances of tens or even hundreds of microns distant from the electrode surface.4−6 There are two competing models proposed for long-distance EET through the electrically conductive biofilms: electron hopping; and metallic-like conductivity.7−12 According to the electron hopping model, electrons are relayed through the biofilm by electron hopping/ tunneling using mediators positioned in the exopolymer matrix and on pili.8 In contrast, the metallic-like conduction model describes electron transport as occurring by delocalized electrons in the pili (nanowire) network within the biofilm.10 Both of these models support the high electronic conductivity © 2015 American Chemical Society
Received: Revised: Accepted: Published: 5227
January 12, 2015 March 15, 2015 March 26, 2015 March 26, 2015 DOI: 10.1021/acs.est.5b00175 Environ. Sci. Technol. 2015, 49, 5227−5235
Article
Environmental Science & Technology the anode.19−21 Second, if accumulation of metabolic end products limited growth, for example proton accumulation leading to low pH,14 then growth at the solution interface would be favored. Third, if diffusion of substrate to the biofilm was important, then bacterial growth at the solution interface would again be favored for improved metabolic activity.6,17 So far, most previous studies have generally shown that the entire biofilm remained metabolically active.4,5,10,16,22−25 However, metabolic activities relative to temporal changes and spatial structure, and current densities and electrochemical properties have not been previously investigated. The activity of the biofilm can change over time, and thus it is possible that biofilm viability might change as well. Changes in either EET pathways or inherent biofilm conductivity could affect its electrochemical activity, and thus current densities over time. In this study, spatial metabolic structure, biological properties (morphology, biomass, and microbial community), and electrochemical properties (current generation, linear sweep voltammetry, nonturnover cyclic voltammetry, electrochemical impedance spectroscopy analyses for charge transfer and diffusion resistance) were simultaneously investigated throughout the biofilm development process. Both mixed culture and pure culture biofilms were examined. A pure culture of G. anodireducens was studied because this strain was isolated from a mature biofilm of an MFC anode, as opposed to other Geobacter strains that have been isolated under iron reducing conditions.26 G. anodireducens SD-1 forms thicker biofilms than G. sulf urreducens PCA, a species which has previously been studied and indicated to be predominant in several mixedculture biofilms.2,26−28 However, these two species share more than a 98% similarity based on 16S rRNA sequences, so there is no certainty about which strain was really predominant in previous studies in anode biofilms. Mini-BESs were used here to cultivate biofilms.27 These reactors have low internal resistance, they have shown to be useful for studying both pure and mixed cultures,27,29 and their simple construction and operation enabled us to simultaneously operate dozens of reactors in order to independently examine multiple biological and electrochemical properties of the biofilms over real time. As a large number of reactors can be started up at the same time, individual reactors could be sacrificed and analyzed over time to avoid changes in the biofilms that would occur when repeatedly analyzing reactors using disruptive chemicals (such as fluorescent dyes) or exposing the biofilms to different potentials during various electrochemical analyses.5,6
were used to record electrode potentials in some tests. All electrode potential values are reported in this study vs SHE. Before being inserted into the BESs, the Ag/AgCl electrodes were immersed in 70% ethanol for 6 h to sterilize them. Two-chamber BESs32 and air-cathode BESs33 were constructed as previously described for additional tests of the metabolic structure of anode biofilms in BESs with different architectures. The electrodes were the same with those in miniBES except that an air cathode was used in the single-chamber, air-cathode BES.34 A two-chamber BES was created by separating the electrode using an anion exchange membrane (AMI-7001S; Membrane International, NJ), and it was operated at a slightly higher applied voltage (0.9 V) to increase the rate of biofilm growth. The air-cathode BES produced voltage, and was operated with a 1000 Ω external resistance for half a month. Mini-BES Operation and Biofilm Growth. The initial inocula were the effluents from the two mini-BESs that were inoculated with G. anodireducens SD-1 or domestic wastewater, and operated for a 1 week period as previously described.27 These two inocula were diluted to the same cell density (based on cell counts), and then used to inoculate BESs (10% v/v). Except as noted (in electrochemical tests without an electron donor), BESs were operated with a 50 mM phosphate buffer solution (PBS) nutrient medium that contained (per liter): 1 g acetate, 2.45 g NaH2PO4·H2O, 4.58 g Na2HPO4, 0.31 g NH4Cl, 0.13 g KCl, 12.5 mL metal salts and 5 mL vitamins (pH 7; conductivity = 7.5 mS/cm).27 All mini-BESs were filled with 100% N2 in the headspace, and sterilized by autoclaving before being inoculated. All reactors were operated in fed-batch mode at 30 °C. The cycle time of the mini-BES was 1 day, except for the first cycle (2−3 days). For each culture, three mini-BESs were used at each time-point to examine the metabolic structure, protein content of the anode, and electrochemical properties. Based on the current generation and visual biofilm, the BESs were examined at the end points of cycles 1, 2, 3, 5, 12, 30, 45, 60 for both types of inocula, and also at cycle 78 for the mixed culture. In addition, other BESs used in the experiments (mini-BES with 0.9 V applied voltage, mini-BES with small anode, mini-BES inoculated with G. sulf urreducens PCA, a two-chamber BES and an air-cathode BES) were also examined once they had produced stable and reproducible cycles of current generation. Standard anaerobic techniques were used throughout all tests. Biofilm growth was monitored based on the protein content extracted from the entire anode using a Bicinchoninic acid protein assay kit. Metabolic Structure of the Anode Biofilm. Anodes were stained using a LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen, CA)10,35 and examined with a confocal laser scanning microscope (CLSM) (LSM710 NLO, ZEISS) with a water objective (LD LCI Plan-Apochromat 25 × /0.8 Imm Korr DIC). The three-dimensional biofilm structure (z-stack) was reconstructed and analyzed using the software ZEN 2009 Light Edition. Anode samples were rinsed in sterile 50 mM PBS to eliminate the original medium, stained for 20 min, and then rinsed in sterile 50 mM PBS twice to eliminate excess dye. The anode was then placed in a dish, immersed in sterile 50 mM PBS, and observed using the CLSM. For each sample, at least two CLSM images were taken. Electrochemical Analysis. Linear sweep voltammetry (LSV), electrochemical impedance spectroscopy (EIS) and nonturnover cyclic voltammetry (CV) were used to examine the biofilm electrochemical characteristics with a potentiostat
■
MATERIALS AND METHODS Mini-BES Construction. Seventy mini-BESs were constructed as previously described.29 Anodes were 1.5 × 1 × 0.3 cm graphite plates (4.5 cm2 surface area, except as noted). All anodes were successively polished using grit type P400 and P1500 sandpaper and 0.05 μm Al particles. Anodes with a smaller available surface area of 0.75 cm2 were produced by covering part of the anode surface using an insulating layer of epoxy. The cathodes were 1.5 × 1 cm stainless steel mesh (Type 304, mesh size 90 × 90). An added voltage of 0.7 V was applied here to each BES using a power supply (except as noted), which is typical of other BES studies.30,31 All MECs were connected in parallel to the power supply, with each circuit containing a 10 Ω resistor to monitor voltage. Current was calculated using Ohm’s law (I = U/R), and current density was normalized by the anode area. Ag/AgCl reference electrodes (+200 mV vs a standard hydrogen electrode, SHE) 5228
DOI: 10.1021/acs.est.5b00175 Environ. Sci. Technol. 2015, 49, 5227−5235
Article
Environmental Science & Technology (CHI660D; Chenhua, China; EC-Lab V10.02 software). The anode was the working electrode, the cathode was the counter electrode, and an Ag/AgCl electrode was used as the reference electrode. Before tests, the reactors were emptied, sparged with N2, and then refilled with fresh medium for EIS and LSV tests, or 50 mM PBS for nonturnover CV tests. EIS was conducted 2 h after connecting the circuit at a set potential equal to the potential of −0.15 V, over a frequency range of 200 kHz to 10 mHz, with a sinusoidal perturbation of 5 mV amplitude. The EIS spectra (Supporting Information Figure S1) with two circles were fitted to an equivalent circuit containing a solution resistance (Rs), two charge transfer resistances (Rct1 and Rct2), a diffusion resistance (Rd), and double layer capacitance (Q) (Supporting Information Figure S2).27 For LSV analyses, the reactors were set at open circuit conditions for 1 h, and then scanned from −0.50 to +0.30 V at a rate of 1 mV/s. For CV tests, the reactors were rinsed with 50 mM PBS for 10 min (twice), refilled with 50 mM PBS, set at open circuit conditions for 1 h, and then scanned from −0.50 to +0.30 V at a rate of 1 mV/s. After all electrochemical tests were finished, electrode potentials were adjusted for the accuracy of the Ag/AgCl electrode, and the corrected potentials adjusted (vs Ag/AgCl) and presented in all results. Community Analysis of the Mixed Culture Biofilm. The mixed culture biofilm grew in the 30th cycle was analyzed the bacterial community by the MiSeq Illumina sequencing technology. DNA was extracted, amplified and purified as previously described,26 except that paired primers in the variable regions V4 (F: 5′-AYTGGGYDTAAAGNG-3′ and R: 5′-TACNVGGGTATCTAATCC-3) were used for the PCR amplification. The MiSeq Illumina sequencing was conducted and analyzed as described previously.36
Figure 1. Current densities of G. anodireducens SD-1 and mixed cultures (MC). (a) Maximum current density recorded for each fed batch cycle with an applied voltage of 0.7 V. The data indicate standard deviations (error bars) based on the mean of triplicate reactors. (b) Peak current densities recorded for different cycles in LSVs (1 mV/s; Supporting Information Figure S3 for details).
cells. The biofilm structure showed a consistent evolution of a live outer-layer on top of a dead inner-core layer for both types of biofilms (Figure 2 and Supporting Information Figure S4). This structure indicated that live cells were only present on the top of the biofilm, rather than throughout the entire biofilm, showing that the entire biofilm was not metabolically active. The lack of a viable biofilm near the electrode cannot be due to incomplete penetration of the dye as the thickness of the live biofilm varied over time, and a dead layer was observed even in very thin biofilms. This finding of the lower part of the biofilm being metabolically inactive was unexpected given published biofilm models and previous observations using live/dead staining of electroactive biofilms.4,5,10,16,19−25 The two-layer structure of the biofilm with dead cells at the electrode-biofilm interface demonstrated that continued metabolic activity of the outer layer cells, and therefore that respiration even by cells on the outer layer of thick biofilms was not limited by electrical conductivity.6,37 The presence of this dead inner core layer also showed that long-distance EET within biofilms did not rely on metabolic activity of all cells in the biofilm since electrons had to have been transported through the dead layer of biofilm to the electrode. This inactive base layer did not appear to result from any diffusion limitation of chemicals into or out of the biofilm6 since this same deadlive two layer structure was observed in the thin and newly developing biofilms (cycle 1, Figure 2a and b) as well as thicker and more mature biofilms. The thickness of the live outer-layer reached an apparent limit of ∼10−15 μm based on comparison of the maximum current densities in fed-batch tests as well as peak electrochemical activities in LSVs. As biofilm thicknesses increased, the size of the live layer remained constant, while the thickness of the dead inner layer continued to increase from 15 to 30 μm for G. anodireducens biofilms (Figure 2e and g), and
■
RESULTS AND DISCUSSION Operation and Maximum Current Densities over Time. Current generation of G. anodireducens SD-1 biofilms changed over time, showing four different stages relative to current generation: a rapid increase in current over the first few cycles; a peak current after around 12 cycles; a decline in current over the next 10−15 cycles; and a final stable current over the last 30 cycles (Figure 1). Current generation by the mixed culture increased more slowly, and reached a stable current after ∼15 cycles. In the last 30 cycles, the average peak current densities measured for the mixed culture averaged 2.58 ± 0.09 A/m2 over cycles 30−60, which was significantly larger than that obtained for the pure culture of G. anodireducens of 2.07 ± 0.08 A/m2 (t test, p < 0.001) (Figure 1a). The highest current densities that these biofilms could produce were evaluated over time based on peak currents measured in LSV tests (Figure 1b and Supporting Information Figure S3). The same general trends in peak current densities relative to maximum current densities were observed over time between the two different cultures, with the pure culture producing the highest peak current density of 5.29 A/m2 in the earlier fed batch cycles, and then decreasing to lower peak current densities in the later cycles. The G. anodireducens biofilms produced an average peak current density 4.25 ± 0.33 A/m2 in cycles 30−60 (3 LSVs), that was similar to 4.47 ± 0.14 A/m2 for the mixed cultures (2 LSVs for cycles 60 and 78) (t test, p = 0.46). Spatial and Real-Time Metabolic Structure of Anode Biofilm. The changes in biofilm viability were examined over time using fluorescent staining to distinguish live versus dead 5229
DOI: 10.1021/acs.est.5b00175 Environ. Sci. Technol. 2015, 49, 5227−5235
Article
Environmental Science & Technology
Figure 2. 3D metabolic-structure images of anode biofilms operated with anodic current density of 3.8 A/m2. (a) G. anodireducens SD-1 (∼3.8 A/ m2, 0.9 V applied voltage), (b) the mixed culture (∼4.1 A/m2, 0.9 V applied voltage), (c) G. anodireducens SD-1 (∼5.2 A/m2, smaller anodic area), and (d) the mixed culture (∼5.6 A/m2, smaller anodic area). Live cells were imaged as green, whereas dead cells were imaged here as red.
that the differences in metabolic structure were due to changes in current density and did not result from differences in biofilm growth or the rate of change in the current. Hydrogen gas could have been a factor in the development of a two-layer biofilm as hydrogen gas is produced on the cathode and released into solution. Hydrogen gas is sparingly soluble, but it can be used by the anodic biofilm for current generation. Thus, hydrogen gas utilization might benefit microbes on the water-side of the biofilm compared to those near the electrode. To test whether hydrogen gas evolution was responsible for the development of a dead inner layer on the anode, biofilms were examined in two other reactors: a two-chamber BES, which is the most common BES construction in anode biofilm studies, with the two chambers separated by an anion exchange membrane that minimizes hydrogen gas crossover from the cathode to the anode; and a single-chamber, air-cathode MFC. Even with these two different architectures, the same two-layer metabolic structure was observed for the biofilms (Supporting Information Figure S5a, b and c). In the two-chamber BES studies, biofilms of G. anodireducens were thicker than those of the mixed culture, and the current densities reached a maximum of
