Introduction

Extracellular electron transfer (EET) is an intriguing biological event, in which the electroactive microorganisms (exoelectrogens) transfer electrons to the external acceptors after metabolizing the organic pollutants/fuels1,2. EET is the core process in a wide diversity of biological phenomena in nature such as bioelectrochemical synthesis, global elemental cycles and bioremediation3,4,5,6,7. It is also critical to the sustainable clean energy device, namely microbial fuel cell (MFC) that converts chemical energy in the fuels into electricity8,9,10,11. EET proceeds via two pathways, namely direct electron transfer (DET) and mediated electron transfer (MET)12,13. The former is achieved through redox active moiety cytochromes on the outer membrane of the bacteria and/or produced conductive nanowire appendages that are in direct contact with the solid/electrode14,15,16, while MET is mediated by soluble redox shuttles secreted by the bacteria17,18,19. Generally, both DET and MET occur simultaneously and contribute to the output current in the operation of MFC. Their corresponding contributions depend on various factors and evolve dynamically in response to the local biological/electrochemical environment. The underlying molecular mechanisms of MET and DET pathways are different while related to one another13. Therefore, differentiating the partial currents from DET and MET pathways is of great significance not only for optimizing the design of MFCs, but also for fundamental understanding of the complicated EET process20,21,22.

The conventional method for EET investigation is to evaluate the discharge ability of bacteria with electrochemical techniques under MFC configuration23,24. However, it is difficult to reliably differentiate MET and DET currents as they are integrated and collected simultaneously. Based on the different redox potentials of electron mediators in two pathways, electrochemical techniques were applied to disentangle the MET and DET currents25. By removing the soluble redox mediators in the supernatant or suppressing the secretion of mediators, the MET process was found to be dominant in EET of Shewanella oneidensis strains12,26. Later, it is found that the endogenous Flavins accelerate EET from Shewanella putrefaciens as cytochrome-bound cofactors to facilitate the DET rather than free soluble shuttles for MET18,19. Recently, a broad diversity of methods including ultramicroelectrode, plasmonic microscopy, dark-field scattering microscopy and fluorescence microscopy were recruited for study of EET mechanism with high sensitivity and spatial resolution in subpopulation even single cell level27,28,29,30,31,32. However, it remains a tremendous challenge to reliably measure the DET and MET currents and to reveal their possible interplay under the practical operation condition of MFCs as they are highly entangled spatiotemporally on the electrode.

Oblique-incidence reflection difference (OIRD) is a novel dielectric constant sensitive technology emerged in past three decades. By detecting the relative difference in s- and p-polarized reflectivity of an elliptically polarized incident light on the surface of interest, it is capable of monitoring a broad diversity of surface/interface processes in a label-free and real-time manner33,34,35,36,37. Further, under two-dimensional scanning mode, it is able to simultaneously probe the spatiotemporal information of the reaction occurred across the whole surface. In our previous work, the spatially-resolved EET flux in a Shewanella putrefaciens CN32 based bioanode was successfully imaged by using OIRD38, but differentiating the DET/MET partial currents was yet to be achieved.

In this work, we report a novel method to disentangle the DET and MET partial currents in bacterial anodes by using OIRD imaging coupled with polyaniline (PANI)-based dual electrode. The PANI film serves as electrochromic electron-acceptor to translate the current information into OIRD optical signals, enabling the mapping of EET currents. Upon complete reduction of PANI, the local EET current is switched off, and the soluble mediators secreted by the bacteria are forced to discharge on the nearby counterpart PANI electrode, enabling imaging of MET current. In such a way, the DET and MET currents from adhering bacteria are successfully measured on a population level. The average DET and MET currents from individual bacterium are further quantified to elucidate the principle of EET. This work thus provides a powerful tool for DET/MET deconvolution and offers valuable scientific insights into the EET process.

Results

OIRD response to PANI reduction and ΔI-Q correlation

The diagram of OIRD instrument used in this study was shown in Supplementary Fig. 1. PANI film was electrochemically polymerized on transparent Fluorine-doped tin oxide (FTO) coated glass with cyclic voltammetry (CV, Supplementary Fig. 2) for bacteria adhesion. As-grown PANI film shows dense nanowires with 100–200 nm diameters to form ~6.3 μm thick inter-entangled porous network on FTO (Supplementary Fig. 3) and exhibits characteristic CV curve and FTIR spectrum (Supplementary Fig. 4). To image the EET current, this PANI film serves as an acceptor to integrate EET electrons released from the microorganisms under open external circuit; as a consequence, it is reduced to arouse in situ OIRD signal change (ΔI, in the unit of arbitrary unit (a.u.)). As shown in Fig. 1a, the spatially-resolved EET flux over a course of discharge was mapped by imaging the entire bioanode in an addressable manner during discharge and the differential image was translated to the charge distribution pattern according to the correlation between ΔI and passed charges (Q). The dynamic EET current (J) of every specific position on the surface were further obtained from the in situ ΔI signals-derived Q curve by using J=dQ/dt.

Fig. 1: Principle of OIRD imaging of EET current.
Fig. 1: Principle of OIRD imaging of EET current.
Full size image

a Schematic illustration of the OIRD setup coupled with a PANI-underlying bioanode, where the EET electron injection in PANI film results in OIRD response to enable quantitative imaging of the EET current; b LSV curve and in situ OIRD response of PANI-FTO electrode from 0.1 to −0.4 V at 1 mV s−1 in M9 buffer with 18 mM lactate. The PANI-FTO electrode was pre-conditioned at 0.1 V for 5 min before sweeping; c Correlation between averaged ∆I and passed charge density (Q) derived from (b) and its fitting curve. The applied potential was also plotted against Q for reference. Data were presented as mean ± standard deviation. Error bar indicates the standard deviation, n = 1000.

The electrochemical behavior and simultaneous OIRD response of PANI-FTO electrode were first investigated by linear sweep voltammetry (LSV) in M9 buffer with 18 mM lactate nutrient. As shown in Fig. 1b, when the PANI-FTO electrode is negatively scanned from 0.1 V, distinct cathodic current appears immediately and reaches a peak at −0.05 V, indicating the reduction of PANI film from its emeraldine (partially reduced) state to leucoemeraldine (completely reduced) state39. Further negative polarization (up to −0.8 V) results in gradual decay and disappearance of cathodic Faradaic current. Only weak charging current is recorded because at completely reduced leucoemeraldine state the PANI film is insulating40. The CV curve (Supplementary Fig. 5) confirms this unique electrochemical behavior of PANI film.

On the in situ OIRD response, ΔI signal exhibits continuous decrease from 0.1 to −0.1 V (stage 1, Fig. 1b) because the reduction of electrochromic PANI changes its dielectric constant change41,42. At stage 2 (potential lower than −0.1 V), ΔI stops decreasing and maintains relatively stable until slight increase at more negative potential. This is because the PANI was completely reduced at stage 2 and its conjugation configuration does not change even more negative potential was applied.

The transferred charges (Q) in the linear sweeping were integrated and plotted against the ΔI to build their quantitative correlation. Each ΔI was obtained by averaging 1000 values collected along a straight line at 4 μm resolution on the surfaces and presented with an error bar. As in Fig. 1c, ΔI positively correlates with Q from 0.1 to −0.05 V (linear range), following a linear relationship of ΔI = −20.79* Q (in the unit of a.u. and mC cm−2, respectively). At lower potential (<−0.05 V), the ΔI deviates from the linear relationship. Together, it is unveiled that the PANI film demonstrates unique potential (redox state) dependent ΔI-Q correlation; there is a linear ΔI-Q correlation at the potential range from 0.1 to −0.05 V, which will be used to convert the collected ΔI into the transferred charge density (Q) in subsequent investigation.

EET behaviors of adherent Shewanella putrefaciens CN32 on dual electrode

A dual electrode was constructed by cutting a complete PANI-FTO electrode into two isolated parts, namely W1 and W2, as shown in Fig. 2a. Specifically, the electroactive Shewanella putrefaciens CN32 bacteria were planted on W1 while no cells on W2. As shown in Supplementary Fig. 6, scattered bacteria were tightly loaded on W1, showing cylindrical shape with a length of ~2 μm, consistent with the morphology of Shewanella putrefaciens CN3243,44. On W2, however, no microbe was found (Supplementary Fig. 7).

Fig. 2: OIRD responses on dual electrode upon EET of adherent Shewanella putrefaciens CN32.
Fig. 2: OIRD responses on dual electrode upon EET of adherent Shewanella putrefaciens CN32.
Full size image

a Schematic illustration showing the dual electrode setup; b OIRD image of the dual electrode at the potential of 0.1 V; c In situ ΔI signals collected along the dashed line as indicated in (b) on the dual electrode in 140 min discharge in M9 buffer containing 18 mM nutrient. At t0 = 0 the potential of both W1 and W2 were removed from 0.1 V to open circuit state; d Representative in situ ΔI signals collected at W1 (top) and W2 (bottom), and the OCP evolution of W1 (middle) during 140 min discharge; e OIRD image of the dual electrode after 140 min discharge; f Charge distribution and intensity profile along the marked line of the dual electrode during 140 min discharge, obtained from the differential OIRD image (Supplementary Fig. 8). The scale bars in (b, e, f) are 3 mm.

For OIRD measurement, both W1 and W2 were first conditioned at 0.1 V for 5 min and imaged with OIRD. As shown in Fig. 2b, the OIRD image exhibits homogeneous OIRD signals on W1 and W2. The average OIRD signal of the former is slightly lower due to the presence of adherent bacteria. At t0 = 0 the potential of 0.1 V on both electrodes was removed to open circuit state. The dynamic ∆I signals collected on multiple positions along a straight line across W1 and W2 (as indicated by the dashed line in Fig. 2b) were presented in Fig. 2c, showing ΔIs on W1 decrease continuously as a result of EET-induced reduction of PANI film, while on W2 ΔIs remain stable with slight fluctuation. A representative in situ ΔI signal on W1 was shown in top panel of Fig. 2d. The approximate linear decrease implies that the electrons were injected to W1 at constant rate, i.e., at constant current according to the linear ΔI-Q correlation in Fig. 1c. At the same time, the open circuit potential (OCP) of W1 continuously shifts negatively (middle panel in Fig. 2d), indicating the gradual reduction of PANI film on W1. On W2, on the other hand, the representative ΔI curve (bottom panel in Fig. 2d) remains stable at the initial stage (from t0 to t1 = ~ 60 min), suggesting no electrons accepted. Therefore, it is signified that the redox mediators secreted by Shewanella putrefaciens CN32 on W1 are confined to discharge locally and cannot be transported to nearby W2 at the initiate stage. Only at stage 2 (at t1 = 60 min onwards), OIRD signal on W2 starts to decrease slowly, implying the overflowing of redox mediators from W1 area, which will be further investigated in the following section.

After 140 min discharging, the dual electrode surface was imaged with OIRD again (Fig. 2e), showing much lower OIRD signals of W1 than W2. By subtracting this image from the first image collected at t0 = 0, an OIRD differential image was obtained (Supplementary Fig. 8) and further converted into charge distribution image by using the linear Q-ΔI correlation in Fig. 1c. As shown in Fig. 2f, EET electrons obtained on W1 is much higher than W2. According to the representative Q profile, the entire W2 exhibits uniform and low charge distribution. On W1, more electrons are accepted near the W1/W2 boundary because of higher bacteria density at the boundary area (Supplementary Fig. 9), which was near to the solution/air interface during bacteria inoculation, thus promoting the attachment and growth of bacteria45. This charge distribution image further highlights the localization of EET current.

Switch off of PANI film for EET

We then focused on the interesting dynamic behavior of W2. It is well recognized that at partially reduced emeraldine state, the PANI film exhibits metal-like conductivity while at completely reduced leucoemeraldine state, it is insulating46. The dynamic EET behavior on W2 is speculated to be associated with the redox state-dependent electron-accepting ability of PANI film on W1. As in Fig. 3a, at 0.1 V PANI on W1 is conducting to accept all EET electrons while upon reduction (as a consequence of EET electron injection, note that the OCP of W1 was decreased to ~ −0.02 V at this moment), it gradually loses its electron-accepting ability; the soluble redox mediators are not able to fully discharge on W1 and therefore a small proportion of them are transported to vicinal W2 and discharged on W2.

Fig. 3: Electron-injection induced switch off of PANI film.
Fig. 3: Electron-injection induced switch off of PANI film.
Full size image

a Schematic illustration showing the reduction-induced switch off of EET on PANI-FTO electrode and the interplay of EET electrons on the dual electrode; b Discharge curve recorded on W1 (top) and in situ ΔI signals collected on W1 and W2, respectively (bottom) in M9 buffer containing 18 mM nutrient; W1 with bacteria at 0.1 V while W2 w/o bacteria at OCP; c Discharge curves of W1 and W2, both at 0.1 V in M9 buffer containing 18 mM nutrient; W1 with bacteria while W2 w/o bacteria; d Discharge curve recorded on Bacteria-PANI-FTO at 0.1 V and −0.3 V in M9 buffer containing 18 mM nutrient; e CV curves of CC, sterile PANI-FTO and PANI-FTO pre-conditioned at −0.3 V in M9 buffer containing 2 μM of FMN, scan rate: 1 mV s−1.

Several experiments were carried out to verify this hypothesis. As shown in Fig. 3b, when W1 was held at 0.1 V while W2 was kept at open circuit state, stable discharge current was obtained on W1 with stable ΔI signals over 150 min discharge as expected. The ΔI signals on W2 at open circuit were also stable, suggesting no any electrons accepted on W2. In a separated experiment, both W1 and W2 were held at constant 0.1 V and their discharge curves were collected. As in Fig. 3c, during 20 h discharge, the anodic current on W1 is continuously increases due to the electrochemical activation and proliferation of bacteria; while no any noticeable current is recorded on W2. A bioanode with Shewanella putrefaciens CN32 inoculated on PANI-FTO was held at 0.1 V, and it delivers increased anodic current, as shown in Fig. 3d. When being polarized to −0.3 V (at t = 9.7 h), however, the anodic current almost disappears, again suggesting the switch off of PANI film under negative potential.

The CV curve of sterile PANI-FTO electrode in M9 buffer containing Flavin mononucleotide (FMN) further confirms the redox state-induced switch off of PANI film. FMN is one of Flavin compounds in oxidized form to mimic the soluble redox mediator secreted by Shewanella bacteria26. As in Fig. 3e (top panel), carbon cloth (CC) shows a pair of well-defined redox peaks centered at ~ -0.42 V from FMN. On PANI-FTO electrode, a wide cathodic peak at −0.38 V is observed in the first negative scanning from 0.1 V (middle panel in Fig. 3e), which originates from the superposed reductive currents from PANI and FMN. On the following scanning, the cathodic current gradually decreases and eventually disappears. Moreover, when the PANI-FTO electrode was pre-conditioned at −0.3 V for 5 min, no any redox current from FMN is generated below −0.1 V (bottom panel in Fig. 3e). Together, these results unambiguously confirm the redox-state dependent electron-accepting ability of PANI film. At positive potential range (i.e., 0.1 ~ −0.02 V), PANI is able to fully accept the EET electrons while at more negative potential (e.g., below -0.02 V), its electron-accepting ability gradually decreases and finally inhibits EET electron injection at fully reduced insulating state (i.e., at −0.3 V). As a result, all EET electrons are highly localized so long as the PANI on W1 remains conducting state, and no current could be collected on the vicinal W2. Upon EET electron injection, the PANI film on W1 is reduced to lose its electron-accepting ability and the redox mediators are transported to W2 achieving MET process.

Measuring DET and MET currents of Shewanella putrefaciens CN32 bioanode

The unique switchable electron-accepting ability of PANI film enables us to switch off the EET on W1, and measure the MET current on W2 with OIRD imaging. In details, the EET (DET + MET) electrons were first collected and measured on W1; then W1 was intentionally turned off, and the soluble redox mediators were then collected by W2 for measuring MET flux, as illustrated in Fig. 4a. To avoid possible disturbance from the proliferation and migration of bacteria, the surface density of adherent bacteria on W1 was improved with prolonged inoculation to accelerate the measurement (see discharging profiles in Supplementary Fig. 10 and SEM images in Supplementary Fig. 11).

Fig. 4: Measuring DET and MET currents of Shewanella putrefaciens CN32 bioanode by controlling on/off state of PANI film.
Fig. 4: Measuring DET and MET currents of Shewanella putrefaciens CN32 bioanode by controlling on/off state of PANI film.
Full size image

a Schematic illustration of strategy for measuring the DET and MET currents; b Representative in situ ΔI signals collected at W1 (top) and W2 (bottom), and the OCP evolution of W1 (middle) during 147 min discharge; at t = 0, the potential of 0.1 V was simultaneously removed on W1 and W2; at t = 147 min and 184 min, the potential of W1 was stepped to −0.5 and −0.3 V, respectively while W2 was kept at open circuit state; c ΔI vs. time (t) plots collected on W1 and W2, respectively. Data were presented as mean ± standard deviation. Error bar indicates the standard deviation, n = 1000; d Corresponding EET, DET, and MET currents; e Corresponding DET and MET currents of individual bacterium.

As shown in Fig. 4b, ΔI on W1 with inoculated Shewanella putrefaciens CN32 bacteria immediately decreases (top panel) after the removal of 0.1 V at t = 0, accompanied by slow decrease of its OCP (middle panel). After a period of discharging (~ 60 min), ΔI decrease of W1 starts to decelerate and almost stops at t = 147 min. At the same time, its OCP reaches ~ -0.13 V at t = 147 min. ΔI on W2 remains stable during the initial 60 min (bottom panel in Fig. 4b) because no electrons are accepted at this stage; after that it starts to decrease slightly, suggesting the discharge of the redox mediators overflowed from W1. Negative potential was then applied to W1 to completely switch off it from local EET current. When stepped to −0.5 V at t = 147 min, ΔI of W1 shows a sudden jump and gradually reaches a stable value (top panel in Fig. 4b), implying its full reduction and switch off. Simultaneously, ΔI on W2 decreases linearly at higher rate, indicating higher MET flux as a result of switch off of W1 (bottom panel in Fig. 4b). Further polarizing W1 to −0.3 V at t = 184 min, ΔI on W1 remains unchanged while that on W2 keeps linear decrease.

The partial currents from DET and MET pathways are calculated from the in situ OIRD responses on this dual electrode. As shown in Fig. 4c, averaged ΔI values collected on W1 (total 1000 points, from t = 0 to 50 min) were plotted against time t, showing a good linear correlation, representing the EET (DET + MET) current. Using the quantitative correlation between ΔI and Q in Fig. 1c, the average EET current was calculated to be ~6.54 μA cm−2. Similarly, the linearly decreased ΔI signals on W2 after the switch off of W1 are associated with the MET current (Fig. 4c), which was calculated to be ~0.74 μA cm−2 and the DET current was ~5.80 μA cm-2, as plotted in Fig. 4d. It is found that for adherent Shewanella putrefaciens CN32, the DET pathway is dominant, accounting for 88% of total EET current.

The EET current of individual bacterium was further quantified and differentiated by normalizing to the bacteria density (obtained by statistical analysis of their SEM image as in Supplementary Fig. 11b). It is found that each adherent Shewanella putrefaciens CN32 cell delivers EET current of 54.05 fA (Fig. 4e), which agrees well with previous results47,48. Meanwhile, the DET and MET partial currents are 47.93 and 6.12 fA, respectively.

Parallel measurements were carried out with three independent Shewanella putrefaciens CN32 bioanodes to check the detection reproducibility. The results are shown in Supplementary Figs. 12-14. As summarized in Supplementary Table S1, the DET and MET currents exhibit considerable deviations (5.56 ± 0.21 and 0.75 ± 0.02 μA for DET and MET, respectively) among three bioanodes due to the variation in cell density. However, the DET and MET partial currents per bacterium are both highly reproducible with low variations (48.32 ± 0.2 and 6.51 ± 0.18 fA for DET and MET, respectively), suggesting high reliability of as-reported method for disentangling DET and MET currents. As this method relies on the optical response after integrating the EET electrons in certain duration over a macroscopic surface, it thus provides high reliability. It is capable of reliably disentangling DET and MET currents at population and single-cell level in a non-invasive and spatial-resolved manner even though it fails to detect the EET process of single cell with high spatiotemporal resolution as accomplished in previous microscopic-based methods28,29,30,31.

OIRD signal exhibits similar response toward potential sweeping on various PANI films with more or less thickness (Supplementary Fig. 15), therefore as-reported method is not depend on PANI film with specific thickness, although that grown by 5 potential cycles shows the highest sensitivity as the charge-to-optical translator. As shown in Supplementary Fig. 16, a new linear ΔI-Q correlation was built on PANI film polymerized with 4 potential cycles (less thickness). The DET/MET currents from attached Shewanella putrefaciens CN32 were successfully measured as in Supplementary Fig. 17, highlighting the great flexibility of as-reported method.

Validation and applicability of as-reported method

To validate as-reported method, the concentration of soluble redox FMN shuttles in the supernatant was measured to be ~75 nM after the switch off of W1 by using fluorescene spectrophotometry (Supplementary Fig. 18), which is much lower than that reported in previous work (250–500 nM)26, consistent with a lower MET contribution of ~12% observed in present work. At the same time, 2 μM FMN mediators were supplemented to the supernatant after the switch off of W1. As shown in Fig. 5a, the OIRD signal on W2 accelerates to decrease quickly, suggesting improved MET current, in line with previous finding that the redox mediator could enhance the MET current by wiring-up the periplasm/cytoplasm conduits18,19,23. Therefore, the MET contribution to the total EET current depends on the shuttle concentration in the supernatant, and as-reported method is reliable for measuring DET and MET currents.

Fig. 5: Validation and applicability of as-reported method.
Fig. 5: Validation and applicability of as-reported method.
Full size image

a In situ OIRD responses on W1 and W2 upon adding 2 μM FMN mediator into the supernatant after the switch off of W1; b Plots of measured integrated DET/MET currents of various bioanodes; c corresponding DET/MET currents from individual bacterium.

As the soluble redox shuttles usually migrate only in a limited range in MFC operation, it raises a reasonable concern on the collection efficiency of W2 in present work. The MET current will be considerably underestimated if not all electron carriers produced by bacteria are collected on W2 after switch off of W1. The collection efficiency of W2 was determined expereimentally by using a graphitic plate mimicking the bacteria on W1 to electrochemically reduce the FMN mediators in the solution. As illustrated in Supplementary Fig. 19, the results suggest a high collection efficiency close to 91% of W2. It is not suppressing as the whole electrochemical cell was driven by the motorized stage to continuously move forward and backward during OIRD measurement; the mass transport of the redox species was greatly accelerated by the forced convection to efficiently shuttle electrons. Meanwhile, the experiment was carried out in anaerobic condition and no other oxidant (e.g., dissolved oxygen) could oxidize the reduced FMN before their transportation to W2.

This method was applied to other bioanodes to evaluate its effectiveness. The influence of cell density on the EET was investigated. The population of adherent Shewanella putrefaciens CN32 was adjusted by inoculation duration (Supplementary Fig. 20), and as expected the measured integrated EET current increases with the cell density (Fig. 5b, Supplementary Fig. 21 and Supplementary Fig. 22). Remarkably, the increase mainly comes from DET current while the MET current increase is insignificant. Meanwhile, the DET partial current from a single cell increases significantly with their density on the electrode, while the MET currents slightly decreases (Fig. 5c and Supplementary Figs. 21, 22). The biological implication of this observation deserves further investigation.

Shewanella oneidensis MR-1another typical electroactive stains were tested and the results were summarized in Fig. 5b, c (details in Supplementary Figs. 23 and  24), showing lower EET ability compared to Shewanella putrefaciens CN32 in both population and single-cell level. A S. oneidensis CN32 mutant (CN32-ΔmtrC/undA) with in-frame deletions of mtrC and omcA gene clusters encoding two outer membrane c-type cytochromes, shows greatly depressed DET and MET currents (Fig. 5b, c, see details in Supplementary Fig. 25 and 26), possibly due to the c-type cytochromes in outer membrane are simutaneously involved in MET and DET pathways18,19,49. These results confirm the effectiveness of this method to various bioanodes with different electroactive strains.

Measuring EET Currents of Adherent and Suspended Shewanella putrefaciens CN32

Under the practical operation of a MFC, there are abundant bacteria suspended in the supernatant as well. These planktonic bacteria also directly contribute to the current/power output by secreting soluble mediators, i.e., MET pathway50,51,52. The contribution of suspended bacteria was further investigated by using OIRD imaging coupled with PANI-based dual electrode. As shown in Fig. 6a, a cellulose membrane with pore size of ~200 nm and thickness of ~200 μm was covered on the dual working electrode to separate the suspended Shewanella putrefaciens CN32 bacteria from the electrode, allowing for only MET pathway. On the dual electrode, Shewanella putrefaciens CN32 bacteria were selectively inoculated on W1, leaving W2 without cells.

Fig. 6: Measuring the partial currents from adherent and suspended Shewanella putrefaciens CN32.
Fig. 6: Measuring the partial currents from adherent and suspended Shewanella putrefaciens CN32.
Full size image

a Schematic illustration showing the bioanode setup; b Representative in situ ΔI signal collected at W1 (top) and W2 (bottom); at t = 0, the potential of 0.1 V was removed on W1 and W2; c ΔI vs. time (t) plots collected on W1 and W2, respectively. Data were presented as mean ± standard deviation. Error bar indicates the standard deviation, n = 1000; d Corresponding DET and MET currents from adherent bacteria (a-DET, a-MET) and MET current from suspended bacteria (s-MET); e Percentage of corresponding current densities from a-DET, a-MET and s-MET pathways; f Corresponding a-DET, a-MET currents of individual adherent bacterium and s-MET current of individual suspended bacterium.

Figure 6b shows the representative in situ ΔI signals synchronously collected from W1 and W2. When the pre-conditioned potential of 0.1 V was removed, ΔI on W1 rapidly decreases and reaches a plateau after 35 min discharge, indicating the PANI film on W1 is rapidly reduced to insulating state as a consequence of EET electron injection. On W2, the ΔI signal exhibits approximately linear decrease at the initial stage, suggesting a constant MET current exclusively from the suspended bacteria. The averaged ΔI values collected on W1 and W2 (each with total 1000 points) at the initiate 15 min and on W2 (total 1000 points) at t = 35–50 min were plotted against t, respectively (Fig. 6c). The total current (including DET and MET currents of adherent bacteria, denoted as a-DET and a-MET, respectively and MET current from suspended bacteria, denoted as s-MET) and the partial current of the suspended bacteria was calculated and presented in Fig. 6d. As shown in Fig. 6e, Shewanella putrefaciens CN32 bacteria attached on the electrode contribute up to 67.5% current via DET and MET pathways during the discharge process; meanwhile, the suspended bacteria contribute 32.5% current. Of course, these portions are believed to be variable depending on their population/density on the electrode and in the suspension.

The current contributions of individual cell were further quantified. The population of adherent bacteria was obtained by statistical analysis of the SEM image (Supplementary Fig. 27) while that of the suspended bacteria was obtained by the optical density (OD) value of their suspension. As shown in Fig. 6f, each adherent bacterium delivers a total current of 54.35 fA, of which 48.2 fA is via DET while the remaining 6.15 fA is from MET pathway. Every suspended bacterium is able to produce an average MET current of 1.21 fA. The adherent bacteria are much more efficient to generate current compared with the suspended ones. Therefore, it is a feasible way to enhance a MFC performance by improving the density of adherent bacteria. Meanwhile, the suspended bacteria, despite their low discharge efficiency, also contribute to a considerable portion of current, especially at high population density.

Discussion

In summary, we successfully measured the extracellular currents via direct and mediated pathways of the electroactive bacteria by using OIRD imaging technique coupled with PANI-based dual electrode. As a switchable EET-electron acceptor, the PANI film translated the current information into dynamic OIRD signals at its conducting state, enabling mapping of EET currents under open circuit condition. When it was electrochemically switched off, the MET current was measured on the nearby counterpart PANI electrode with OIRD imaging. In such a way, the DET and MET currents from adhering bacteria and suspended bacteria in the electrolyte were measured. The average DET and MET currents from a single bacterium were further quantified. As-reported method enables successful deconvolution of the DET and MET currents and may offer valuable insights to the research areas of MFCs among others.

Methods

Chemicals and materials

FTO glass (conductivity of 14 ± 1 Ω/sq) was purchased from OPV Tech New Energy Co., Ltd. Sulfuric acid (H2SO4, 96%), toluene, acetone, and ethanol were purchased from Sangon Biotechnology Co., Ltd (Shanghai, China). Polydimethylsiloxane (PDMS) was obtained from Dow Corning Investment Co., Ltd. Aniline (ACS, ≥99%) was purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd. Sodium chloride, tryptone, and yeast extract were purchased from OXOID Co., Ltd. All chemicals were commercially available and used without further purification.

Preparation of PANI-FTO dual electrode

PANI-FTO electrodes were prepared by electrochemical polymerization of PANI film on FTO53. FTO was ultrasonically cleaned in acetone and deionized water for 15 min, respectively and then dried with N2. In a three-electrode system, the clean FTO electrode was swept using CV in 0.5 M H2SO4 containing 0.1 M aniline with an Ag/AgCl/saturated KCl electrode as the reference electrode and a graphite sheet as the counter electrode. The potential was scanned between −0.25 to 0.85 V at 6 mV s−1 for 5 cycles. After that, the PANI-FTO electrode was washed with copious deionized water and dried with N2. Then, the FTO layer and grown PANI film was cut to two electrically isolated parts, denoted as W1 and W2 for further use.

Bacterial culture

A single bacterial colony of S. putrefaciens CN32 was first cultured in 100 mL of Luria Bertani (LB) broth medium containing 10 g L−1 NaCl, 10 g L−1 tryptone, and 5 g L−1 yeast extract on a shaker for 12 h (220 rpm, 30 °C) to obtain the bacteria solution49. Then, 1 mL of as-obtained primary bacterial solution was inoculated into 200 mL fresh LB medium and incubated on a shaker for 12 h to obtain the secondary bacterial suspension. The secondary bacteria were collected by centrifugation and re-suspended in 50 mL of M9 buffer containing 18 mM sodium lactate. M9 buffer contains Na2HPO4 (6 g L−1), KH2PO4 (3 g L−1), NaCl (0.5 g L−1), NH4Cl (1 g L−1), MgSO4 (1 mM) and CaCl2 (0.1 mM). Other strains were cultured with the same procedure.

Preparation of bacteria-PANI-FTO bioanode

The bacterial suspension in M9 buffer containing 18 mM sodium lactate was bubbled with N2 for 30 min. The W1 of the PANI-FTO dual electrode was carefully immersed into the bacterial suspension and connected to the external circuit as the working electrode. An Ag/AgCl/saturated KCl electrode and a graphite sheet were placed in the suspension as the reference electrode and counter electrode, respectively. After sealing the electrochemical cell from the atmosphere, the potential of W1 was positioned at 0.1 V for 12 h to stimulate the bacteria inoculation on W1. After that, the electrode was rinsed with M9 buffer to remove bacteria that are not firmly adhered on the electrode.

Electrochemical measurements

A bottomless PDMS cell was fixed on the bioanode to form an electrochemical cell where W1 (Bacteria-PANI-FTO) and W2 (PANI-FTO, without bacteria) were simultaneously exposed to the electrolyte. M9 buffer containing 18 mM sodium lactate was used as the electrolyte, which was pre-bubbled with N2 for 30 min to remove the dissolved O2. A graphite sheet was used as the counter electrode, and an Ag/AgCl/saturated KCl was used as the reference electrode. The potentials at W1 and W2 were independently controlled with a bipotentiostat. All potentials reported in present work are reference to this Ag/AgCl/saturated KCl electrode unless specified otherwise.

SEM analysis

The bacteria-PANI-FTO bioanodes were immersed in a 2.5% glutaraldehyde solution overnight for cell fixation and then dehydrated in increasing concentrations of ethanol solution54. Finally, the cell-attached electrodes were dried for SEM observation using a JSM-7800F system from JEOL.

OIRD instrument and measurements

A homemade OIRD device was used in this study. The diagram of the OIRD device is shown in Supplementary Fig. 1. In brief, a He-Ne laser (632.8 nm, 3 mW in power) passes through a polarizer to ensure its high polarization. The laser is modulated with a PEM 100 photoelastic modulator system from Hinds Instruments to periodically change between s- and p-polarized states at 50 kHz frequency. A phase shifter is used to introduce an adjustable phase difference between s- and p-beams. After being focused with a lens, the laser illuminates on the bottom surface of FTO at a constant incident angle of 60° and the amplitude of alternating component of reflected light at 100 kHz (I2Ω) is monitored with a photodiode and recorded with a lock-in amplifier (SR830, dual-phase DSP lock-in amplifier from Stanford Research Systems Inc.). For scanning mode, a motorized stage (KS202-100 stage with a DS102MS controller, from SURUGA SEIKI Co., Ltd.) controls the sample stage moving in the horizontal direction and the OIRD image of the whole surface is collected.

All OIRD measurements were carried out at constant temperature in present work to avoid possible disturbance from temperature fluctuation (see in situ OIRD curve collected on M9 buffer immersed FANI/FTO surface showing negligible drift in Supplementary Fig. 28). At 0.1 V, the entire electrode was imaged with OIRD after the I2Ω signal reaches a stable baseline (defined as ΔI = 0). Then the polarization was removed and the in situ ΔI signals across the surface were continuously collected by moving the slide. After certain duration of discharge under open circuit state, the entire electrode surface was imaged again. The differential OIRD image was obtained by subtracting the OIRD image after discharge from the one collected at 0.1 V.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.