Introduction

Biophotovoltaics (BPVs) harness photosynthetic biocatalysts—such as enzymes, thylakoid membranes, oxygenic cyanobacteria, anoxygenic purple sulphur and non-sulphur bacteria, or algae—to convert solar energy into electrical power1,2,3,4,5. In particular, living BPVs that employ intact microorganisms like cyanobacteria and algae have gained substantial attention due to their inherent self-renewing capabilities6,7,8,9,10. Despite the theoretically high quantum efficiency of intracellular light-harvesting photosystems, i.e. greater than 90%, the overall solar-to-electricity conversion efficiency in whole-cell devices remains relatively low, typically below 1%11,12,13. This limitation arises in part from the lack of efficient electron transport pathways from the photosynthetic microorganisms to the electrode surface, which ultimately hampers the effective extraction of electrons generated during photosynthesis14,15. Notably, exoelectrogenic cyanobacteria like Synechocystis sp. PCC 6803 can engage in both direct electron transfer (e.g., via microbial nanowires16) and mediated electron transfer, which involves endogenous or exogenous redox mediators6,17, or even through microbial consortia18. Such processes emphasize the diversity of electron flow strategies in BPVs and highlight the need for more effective ways to harness the electrons generated by photosynthesis. Improving interfacial electron transfer (IET) at the microbe–electrode interface—whether via direct contact or mediator-facilitated pathways—remains a key challenge for realizing higher photocurrent and overall BPV performance19.

Substantial efforts have been dedicated to facilitating the IET process across abiotic/biotic interfaces by employing artificial redox-active materials and electrode engineering strategies2,20. For instance, exogenous electron mediators like lipid-soluble quinones6,21 and phenazines22 can diffuse beyond the periplasmic space and interact closely with the photosynthetic apparatus, facilitating electron extraction from cells14,23,24. Ferri/ferrocyanide [Fe(CN)6]3−/4−, which is relatively lipid-insoluble, primarily accepts electrons from the periplasmic space of cyanobacteria25. Meanwhile, coating electrodes with interfacial nanomaterials has emerged as an effective approach to reduce electron transport resistance26. For example, studies have demonstrated that polymers containing electroactive osmium centers can serve as an electron transport network between photosynthetic microorganisms and electrodes, thereby enhancing photocurrent27. Similarly, quinone-28 and ferrocene-containing polymers29, carbon nanotubes30, as well as organic semiconductors31,32,33, have been found to enhance photocurrent by improving the electron transfer pathway at the interface. In particular, polypyrrole (PPy)-34, polyaniline (PANI)-35,36, polydopamine (PDA)-5,37, and poly(3,4-ethylenedioxythiophene) (PEDOT)-functionalized38,39,40,41 matrices have provided stable and conductive interfaces that enhance current extraction from photosynthetic cells and membranes. In these cases, the electrode is modified with polymers through electropolymerization, vapor-phase deposition, or simple coating, followed by subsequent addition of the biological element. Alongside these electrode surface modifications, engineering the material and architecture of electrodes offers yet another approach to enhancing IET42. Electrodes with rough surfaces, such as carbon felt and graphite, exhibit greater affinity for microorganism attachment compared to smooth surfaces like copper, platinum, and tin oxide2. Recent advancements also include the development of porous or hierarchical indium tin oxide (ITO) electrodes with high dimensionality and optical transparency for increasing the electroactive surface area available to cells without hindering their photosynthetic activity6,43,44.

Conjugated polyelectrolytes (CPEs)—characterized by electronically delocalized backbones bearing ionic functionalities—offer a distinct and promising strategy to enhance electron transfer efficiency and overall device performance45,46,47. Previous studies have shown that CPEs can spontaneously self-assemble with electroactive Shewanella oneidensis MR-1 to form three-dimensional (3D) living composites that not only facilitate efficient charge transport for charge extraction or injection in microbial electrochemical cells but also accommodate higher bacterial loading45,48. These abiotic/biotic composites leverage a synergetic effect, with the encapsulated bacteria upregulating energy metabolism genes and their extracellular biomaterials enhancing the structural cohesion of the CPE49.

In this work, we demonstrate a method to improve the photocurrent of living BPV systems by encapsulating the photosynthetic cyanobacterium Synechococcus elongatus PCC 7942 (S. elongatus) within a conductive and biocompatible CPE matrix (Fig. 1a). A highly p-doped CPE, namely PEDOT-S, is employed to form a 3D network with inter-chain connectivity (Fig. 1b) in Blue Green medium (BG11, bacteria culture medium; see the Methods for the detailed composition), while maintaining sufficient optical transparency for photosynthesis. The self-assembled biocomposite system amplifies photocurrent output in a photoelectrochemical cell by enhancing extracellular electron transfer to the electrode (Fig. 1c). Furthermore, photocurrent mapping down to the single cell level demonstrates the enhanced electron transfer capabilities of individual cells facilitated by the PEDOT-S matrix.

Fig. 1: S. elongatus/PEDOT-S biocomposites form a 3D conductive matrix for efficient electron extraction.
figure 1

a Schematic of the S. elongatus/PEDOT-S biocomposite, depicting a conductive matrix (PEDOT-S, in blue color) encapsulating cyanobacterium S. elongatus (in green color). The PEDOT-S forms a 3D conductive network, allowing for efficient electron transport throughout the system. The zoomed-in schematic highlights the formation of a polymer network via interchain interactions. b Molecular structure of PEDOT-S, with the p-doped conjugated backbone (colored in blue) and the negatively charged chains (colored in orange). The PEDOT-S gel is expected to form through interchain connections via hydrophobic effects and ionic interactions between the negatively charged sulfonate groups and positively charged counter ions. c Cartoon schematic representation of the electron transport pathway in the S. elongatus/PEDOT-S biocomposites. Electrons are generated by the photosynthetic activity of S. elongatus and are transferred to a redox mediator (K3[Fe(CN)6], yellow spheres), which facilitates electron shuttling between the biological system and the PEDOT-S matrix. The PEDOT-S polymer then serves as the conductive medium that transports these electrons to an external electrode. The cartoon schematic represents some key components within the cyanobacterium: carboxysomes (orange hexagons), chromosomes (grey tangled lines), phycobilisomes (grey spheres), and thylakoids (light green concentric curves).

Full size image

Results

Characterizations and properties of PEDOT-S

PEDOT-S was obtained via a one-step polymerization from the monomer sodium 4-((2,3-dihydrothieno[3,4-b][1,4]dioxin-2-yl)oxy)butane-1-sulfonate (EDOT-S) using iron sulfate (FeSO4) and ammonium persulfate ((NH4)2S2O8) as oxidants50. Gel permeation chromatography (GPC) revealed a molecular weight of approximately 4700 Da, corresponding to an average of ≈16 repeat units per polymer chain, aligning with values reported in the literature51. As shown in Fig. 2a, PEDOT-S as obtained immediately after synthesis demonstrates low absorption in the visible spectrum (400–700 nm) and significant polaronic absorption above 1000 nm (Figs. 2a and  S6), indicative of a high degree of doping50.

Fig. 2: Optical and mechanical properties of PEDOT-S at various concentrations.
figure 2

a Absorption spectra in the visible to near-infrared range (400–1300 nm). b CV traces at a scan rate of 5 mV s−1 with BG11 as the electrolyte. c Transmittance (%) in the wavelength range of 400–900 nm. Absorption, transmittance, and CV spectra were measured atop ITO electrodes with 600 μm spacer thickness (optical pathlength) d Rheological measurements in BG11 depicting the elastic (G′) and loss moduli (G″) as a function of frequency (ω) at a constant strain of 0.01% and a temperature of 23 °C.

Full size image

The relationship between the optical absorption of PEDOT-S and the degree of doping was elucidated by interactions with acidic and basic environments52,53. As shown in Fig. S7, aqueous sodium hydroxide (NaOH) treatment led to distinct absorption peaks at 585 nm (neutral) and 990 nm (polaronic), with a decrease in absorption beyond 1200 nm (bipolaron absorption), in accordance with a de-doping process. Subsequent addition of H2SO4 increased the absorption beyond 1000 nm, confirming that the doping process is reversible. These findings are supplemented by spectroelectrochemical experiments54 that imposed decreasing potential from −0.6 V to 0.6 V, leading to decreased absorption of the neutral state (Fig. S8), illustrating the doping process. Electron paramagnetic resonance (EPR) spectroscopy of pristine PEDOT-S revealed the presence of radical species (Fig. S9), further supporting a p-doped state51. In this system, not only does p-doping facilitate low absorption in the visible light range, but also increases the presence of charge carriers, thereby facilitating electron transport47 within biophotovoltaic systems.

The electrochemical properties of PEDOT-S were assessed using a three-electrode electrochemical cell. Cyclic voltammetry (CV) was performed on PEDOT-S in BG11 (medium for cyanobacterium S. elongatus), using an Ag/AgCl reference electrode. CV profiles displayed reversible and quasi-rectangular traces, indicative of a pseudocapacitive characteristic (Fig. 2b)55. The onset oxidation potential at −0.24 V corresponds to the polaron formation within the PEDOT-S structure. The low oxidation potential implies that the CPE will be predominantly p-doped at positive potentials, which makes PEDOT-S an effective electron acceptor56, ultimately facilitating efficient electron transfer from the photosynthetic bacteria to PEDOT-S. Furthermore, the electrochemical stability of PEDOT-S was confirmed over 200 cycles (Fig. S10), demonstrating its potential for sustained electron transport in bioelectronic systems.

Optical transmittance is a critical factor as light penetration is essential for the photosynthetic activity of cells44. To quantify the optical transmittance, spectrophotometric analysis was conducted on PEDOT-S in BG11. An optical path length of 600 µm was used to approximate the thickness of biocomposites in the photocurrent measurements. We focused on the 400–500 nm (Soret band) and 600–700 nm (Qy band) regions, corresponding to key chlorophyll absorption bands (Fig. S11)57. From Figs. 2c and  S12, 0.5 mg mL−1 PEDOT-S allowed for over 80 % transmission across the key absorption ranges. Increasing the concentration to 2 mg mL−1 and 5 mg mL−1 reduced transmittance to approximately 45% and 25%, respectively. These results quantify the optical semitransparency of PEDOT-S, particularly within a spectral region vital for photosynthesis.

Of particular interest is that PEDOT-S exhibits viscoelastic properties when dispersed in BG11. This feature aligns with the properties of other CPEs that assemble into 3D conductive hydrogel networks through interchain ionic cross-links and hydrophobic contacts (Fig. 1b)45,58,59. To assess the impact of concentration on the physical properties of PEDOT-S, we performed rheology measurements using a plate-cone geometry60. For these tests, PEDOT-S in BG11 at concentrations of 1, 2, and 5 mg mL−1 were deposited onto the testing plate, and a frequency sweep was conducted in the linear viscoelastic region. From Fig. 2d, one observes that the elastic modulus (G′) and loss modulus (G″) display a linear and parallel relationship across the measured frequency (ω) range, with G′ consistently exceeding G″. In contrast, PEDOT-S in DI water at 2 and 5 mg mL−1 displayed a divergent trend in G’ and G”, with G” exceeding G’ over some of the frequency ranges, indicating a viscous-dominating behaviour (Fig. S13). This comparison indicates the gel-like behavior of PEDOT-S in BG11 and confirms the formation of a cohesive and conductive matrix for encapsulating living cells onto the electrode (Fig. S14)61,62.

The biocompatibility of PEDOT-S with S. elongatus was investigated before the formulation of biocomposites. We cultured S. elongatus (OD730 = 0.2, corresponding to ≈1.0 × 108 cells mL−1) with PEDOT-S at concentrations from 0.5 mg mL−1 to 5 mg mL−1 in BG11 medium. These cultures were incubated at 30 °C under illumination for 4 days. Bacterial viability was subsequently quantified using colony-forming unit (CFU) counts on agar plates. The results revealed that after 4 days of illumination, CFU counts of all PEDOT-S concentrations exceeded the initial seeding levels, suggesting that the introduction of PEDOT-S does not compromise cell viability. Moreover, 2 mg mL−1 and lower exhibited no significant growth inhibition, whereas 5 mg mL−1 led to a 51 % CFU reduction compared to the bare S. elongatus, likely due to reduced light penetration (Fig. S15). Previous studies have also demonstrated that PEDOT-based materials do not inhibit PSII photosynthetic efficiency of green algae and cyanobacteria over 72 h and 48 h, respectively63. Therefore, PEDOT-S in concentrations of 5 mg mL−1 and lower were deemed suitable for subsequent photoelectrochemical experiments with S. elongatus.

Fabrication of PEDOT-S/S. elongatus biocomposites and morphological studies

Living biocomposites were prepared by combining various concentrations of PEDOT-S with the same initial S. elongatus loading of OD730 = 0.9 (corresponding to 4.5 ×108 cells mL−1) in BG11 (pH = 7.1). To assess the electrochemical properties and photocurrent output, a three-electrode electrochemical cell was designed (Fig. S16). The PEDOT-S/S. elongatus biocomposites were confined atop an ITO glass working electrode using a silicone gasket with an exposed area of 2 cm2 and an optimized thickness of 600 µm (Supplementary Note 1). The assembly was covered with a dialysis membrane (MWCO of 3.5 kDa) to ensure confinement while allowing ion exchange with the electrolyte.

The morphology of the bare S. elongatus and PEDOT-S/S. elongatus biocomposites at different concentrations (1, 2, and 5 mg mL−1) atop ITO electrodes were examined using scanning electron microscopy (SEM) and confocal laser scanning microscopy (CLSM). SEM images of bare S. elongatus, as depicted in Fig. 3a and b, show non-uniform coverage of S. elongatus cells on the electrode surface. The S. elongatus cells tended to form small clusters in certain areas (Fig. S17), which may be associated with biofilm formation. Neat PEDOT-S shows a sheet-like and porous structure with irregular textures on the surface (Fig. S18) suggesting the three-dimensional network of the polymer. In the PEDOT-S biocomposites, cells were embedded within the polymeric matrix (Figs. 3c, d, and S17), forming a thick agglomerated structure (Fig. 3e). As the PEDOT-S concentration increased, the polymer framework became progressively thicker (Fig. S19). CLSM analysis provides further insights into the in-situ spatial arrangement and distribution of S. elongatus utilizing the intrinsic fluorescence of the cells57. Specifically, S. elongatus cells exhibit intrinsic fluorescence from chlorophyll pigments at wavelength 650–750 nm when excited by 620 nm or 635 nm laser, whereas PEDOT-S does not emit fluorescence in this wavelength range (Fig. S20 and S21), allowing selective imaging of the cells. As shown in Fig. 3f, bare S. elongatus exhibited a thin and uneven biofilm layer with a limited vertical (z-axis) organization of less than 10 μm. For the PEDOT-S biocomposites, S. elongatus were dispersed within a larger 3D framework (Fig. 3g) with an enhanced z-axis spread around 30 μm, indicating a spatial distribution of the cells into the conductive PEDOT-S matrix. These observations suggest that the PEDOT-S matrix facilitates cell immobilization and supports a 3D architecture, which improves cell-electrode wiring and provides charge transfer pathways between cells and the electrode as shown in Fig. 3e and represented in Fig. 1a.

Fig. 3: Microstructure of S. elongatus and PEDOT-S/S. elongatus biocomposites.
figure 3

a, c Representative scanning electron microscopy (SEM) images of S. elongatus and PEDOT-S/S. elongatus biocomposites after photocurrent collection in a photoelectrochemical experiment, respectively. The imaging experiments were independently repeated three times with similar results. b, d Cross-sectional SEM images of S. elongatus and the PEDOT-S/S. elongatus biocomposites, respectively. e Schematic representation of the microstructure of S. elongatus and PEDOT-S/S. elongatus biocomposites. f, g Z-stack imaging of S. elongatus and PEDOT-S/S. elongatus biocomposites at a depth of 30 μm performed using confocal laser scanning microscopy (CLSM), utilizing the intrinsic fluorescence of chlorophyll within the S. elongatus cells (λexc = 620 nm, λemi = 650–750 nm), which are shown as red dots.

Full size image

Photocurrent output facilitated by the PEDOT-S matrix

To evaluate the photocurrent production of PEDOT-S/S. elongatus biocomposites, we conducted chronoamperometry (CA) in electrochemical cells by applying a potential (ECA) of 0.3 V (vs Ag/AgCl) under alternating dark and light conditions using a white LED light (1 mW cm−2, approximately 50 μmol photons m−2 s−1). The living biocomposites were prepared by combining various concentrations of PEDOT-S with S. elongatus in BG11 media. The photocurrent density (Jph) is defined as the difference in current density (J) between the endpoints of illuminated and dark conditions over a light on-off cycle44. Before measuring the photocurrent, biocomposites were kept in the dark at an ECA of 0.3 V for 3 hours to facilitate cell attachment and minimize capacitive charges.

Photocurrent was first measured without the addition of an exogenous electron mediator (Fig. 4a). S. elongatus alone displayed a Jph of 0.004 ± 0.001 μA cm−2. After incorporating PEDOT-S at concentrations of 1 mg mL−1, 2 mg mL−1, and 5 mg mL−1, the Jph increased to 0.046 ± 0.004 μA cm−2, 0.049 ± 0.015 μA cm−2, and 0.091 ± 0.019 μA cm−2, respectively (Fig. 4c). Notably, the 5 mg mL−1 PEDOT-S biocomposite demonstrated an over 20-fold Jph increase compared to S. elongatus alone, suggesting enhanced electron transfer from S. elongatus to electrode through the conductive PEDOT-S matrix45. We then performed electrochemical impedance spectroscopy (EIS) to elucidate the electrochemical properties. In the Nyquist plots obtained from EIS measurements, characteristic semicircles at high frequencies denote charge transfer resistance (RCT), with smaller diameters signifying more efficient charge transfer and current collection at the abiotic-biotic interface64. S. elongatus exhibits large impedance without a distinct semicircular shape (Fig. 4d), suggesting limited interfacial charge transfer. Increasing concentrations of PEDOT-S led to decreased impedance and smaller semicircle diameters (Fig. 4d), suggesting that PEDOT-S serves as a conductive matrix to decrease the kinetic barriers for charge extraction.

Fig. 4: Photocurrent generation and electrochemical properties of PEDOT-S/S. elongatus biocomposites.
figure 4

a, b Chronoamperometry (CA) plots for S. elongatus and PEDOT-S/S. elongatus biocomposites under alternating light and dark conditions without and with the presence of 1 mM K3[Fe(CN)6]. The solid line shows the average current density while the shaded area represents the standard error of the mean (n = 3, biologically independent samples). The photocurrent density (Jph) is defined as the difference in current between the light-on and light-off states. c Comparison of Jph for S. elongatus alone and PEDOT-S/S. elongatus biocomposites under mediator-less conditions. The p-value for the comparison between bare cells and 5 mg mL−1 PEDOT-S/S. elongatus biocomposites is 0.009 (*p < 0.05, **p < 0.01). d Representative Nyquist plots for S. elongatus and PEDOT-S/S. elongatus biocomposites under mediator-less conditions. e Electroactive surface areas (ESAs) of bare ITO and PEDOT-S at various concentrations determined by CV analysis. The p value for the comparison between bare cells and 2 mg mL−1 PEDOT-S/S. elongatus biocomposites is 0.005 (*p < 0.05, **p < 0.01). f Comparison of Jph for S. elongatus and PEDOT-S/S. elongatus biocomposites under mediator conditions. The DCMU control experiments were performed on the 2 mg mL−1 PEDOT-S biocomposite samples. The p-value for the comparison between bare cells and 5 mg mL−1 PEDOT-S/S. elongatus biocomposites is 0.006 (*p < 0.05, **p < 0.01). g Representative Nyquist plots for S. elongatus and PEDOT-S/S. elongatus biocomposites under mediator conditions. The average charge transfer resistance (RCT) values annotated were determined from equivalent circuit fitting. h Current density (J) over time from CA measurements for S. elongatus and PEDOT-S/S. elongatus biocomposites under continuous illumination over 12 hours. The data are shown as a percentage of the initial current output (J %). In all figures the error bars represent the standard error of the mean (n = 3, except 2 mg mL−1 in (c) n = 4, all are biologically independent samples). Statistical analysis was performed using one-way ANOVA followed by Tukey’s Honestly Significant Difference (HSD) post-hoc test. All tests were two-sided, and p values were adjusted for multiple comparisons.

Full size image

To further investigate the mechanisms underlying photocurrent enhancement in the PEDOT-S biocomposites, we evaluated the electroactive surface areas (ESAs) to quantify the extent of the electrode that participates in electrochemical reactions (detailed calculations provided in Supplementary Note 2)65. For different PEDOT-S concentrations, we extracted the peak current responses of the K3[Fe(CN)6] redox probe from CV profiles at varying scan rates, and then analyzed the data using the Randles-Sevcik equation66. Notably, 5 mg mL−1 PEDOT-S possesses an ESA of 6.20 ± 0.81 × 10−3 cm2, over three times higher than the bare ITO control at 1.62 ± 0.01 × 10−3 cm2 (Fig. 4e). This implies that the conductive network formed by PEDOT-S increases the number of active sites for redox processes, thereby enhancing charge extraction from the cells45.

With the addition of 1 mM potassium ferricyanide (K3[Fe(CN)6], midpoint potential at 0.20 V) as an exogenous redox mediator, the Jph of S. elongatus was elevated to 0.14 ± 0.11 μA cm−2 (Fig. 4f), reflecting enhanced mediated electron transfer from the cells to the electrode surface25. Moreover, incorporating 1 mg mL−1 and 2 mg mL−1 PEDOT-S increased photocurrent to 0.21 ± 0.03 and 0.83 ± 0.07 μA cm−2, respectively, indicating that PEDOT-S can also enhance mediated electron transfer processes39. However, further increasing the PEDOT-S concentration to 5 mg mL−1 lowered the Jph to 0.16 ± 0.03 μA cm−2.

To elucidate the behavior of the PEDOT-S biocomposites with the added mediator, EIS was performed to quantify the RCT and the restricted diffusion resistance (Rd) by fitting the impedance spectra (Z) as a function of frequency (f) to an equivalent circuit model (Fig. 4g, see Methods for details)59. While the S. elongatus alone had an RCT of 2500 ± 300 Ω cm−2, introducing 1 mg mL−1, 2 mg mL−1, and 5 mg mL−1 PEDOT-S reduced RCT to 500 ± 50 Ω cm−2, 350 ± 50 Ω cm−2, and 400 ± 50 Ω cm−2, respectively (Figs. S22 and S23). Although 2 mg mL−1 and 5 mg mL−1 PEDOT-S biocomposites show relatively comparable RCT, the Rd for the 2 mg mL−1 biocomposite is 2 ± 1 Ω cm−2, significantly lower than the 150 ± 30 Ω cm−2 for 5 mg mL−1 biocomposites, signifying less diffusional hindrance. CV measurement provides further insights. S. elongatus and 1 mg mL−1 PEDOT-S biocomposites displayed similar redox peaks centered around 0.20 V originating from K3[Fe(CN)6] (Fig. S24). The 2 mg mL−1 biocomposite shows higher and broader redox waves, attributed to enhanced pseudocapacitance characteristics resulting from higher PEDOT-S loading. This increased polymer loading fosters more continuous interchain contacts55,67. However, the 5 mg mL−1 PEDOT-S/S. elongatus biocomposite showed a lower peak current and narrower redox wave compared to 2 mg mL−1, consistent with the ion diffusion impedance. Collectively, these results suggest that while higher concentrations of PEDOT-S enhance the formation of a more conductive matrix, excessive concentrations can hinder the diffusional transport of redox mediators68. With the presence of 1 mM K3[Fe(CN)6], 2 mg mL−1 PEDOT-S emerges as the optimal concentration to integrate with S. elongatus with the benefits of increased active sites for electron transfer while maintaining low diffusional resistance.

Cell loading capacity was investigated to examine the interaction between S. elongatus cells and the electrode by varying the bacterial cell density within the biocomposites. While bare ITO exhibited a clear saturation in photocurrent when the S. elongatus loading reached 3.0 × 108 cell mL−1, the 2 mg mL−1 PEDOT-S biocomposites displayed a linear relationship between photocurrent and bacterial loading up to 6.0 × 108 cell mL−1 (Fig. S25). This behavior indicates that cells within the biocomposite receive adequate light, and the polymer matrix effectively integrates the cells and facilitates a robust electronic connection between them and the electrode.

The specific involvement of photosynthetic activity in photocurrent production was further verified by the addition of 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU), an inhibitor of photosynthetic electron transport32, to the 2 mg mL−1 PEDOT-S/S. elongatus biocomposite. This led to a significant decrease in Jph from 0.83 ± 0.07 to 0.05 ± 0.01 μA cm−2 (Fig. 4b and f). The controls of PEDOT-S + K3[Fe(CN)6] and K3[Fe(CN)6] without cells generated negligible photocurrent (Figs. 4b and S26), indicating that the abiotic components do not contribute to the photocurrent. Collectively, these findings confirm that the observed photocurrent was generated by the photosynthetic activity of S. elongatus and that the PEDOT-S conveys an advantage for electron extraction, leading to photocurrent improvement.

BPV systems are typically evaluated not only based on power output but also on operational stability under illumination69,70. Of relevance is that long-term operational stability is limited due to the inherent complexities of integrating living organisms with electronic components25. We monitored the output current density (J) every 2 hours, comparing it to the initial current output under light exposure (Fig. S27). Both S. elongatus and the 2 mg mL−1 PEDOT-S biocomposite retained over 70 % of their performance after continuous illumination for 12 hours (Fig. 4h), indicative of a sustained photocurrent output comparable with previous studies6,7. Sustained cell growth and ATP production after the experiment suggest that cell viability is maintained (Fig. S28).

Microscopic cell-electrode interactions demonstrated by single-cell photocurrent mapping

We performed electrochemical experiments down to single-cell levels to further demonstrate the electrochemical interaction between cells and the electrodes. We employed a multimodal microscopy platform to image and map the photocurrent of the biocomposites (Fig. S29)71. The use of single-cell photocurrent mapping allows us to disentangle the cell-electrode interactions in complex biological-electrode systems and provides a microscopic evaluation of the performance of biophotovoltaics. While methods such as CFU estimation provide an overall view of biocatalyst performance, they mask the heterogeneity in cell–electrode interactions—especially in polymeric networks. As normalizing photocurrent by bacterial number or chlorophyll content remains challenging for accurate performance evaluations, single-cell photocurrent mapping assisted by fluorescence microscopy offers an alternative method to evaluate the performance by directly measuring per-cell current outputs, revealing which cells are effectively “wired” to the electrode and how the polymer network influences electron generation/collection. Furthermore, single-cell measurements provide insight into the maximum electron transfer capabilities of individual cells, revealing design parameters such as interfacial materials or cell density that can optimize the performance of the system.

To perform single-cell photocurrent measurements, suspensions of 2 mg mL−1 PEDOT-S biocomposite and bare S. elongatus cells in the presence of 1 mM K3[Fe(CN)6] were dispersed onto a fluorine-doped tin oxide (FTO) substrate within a photoelectrochemical microfluidic cell. The samples were confined using a transparent coverslip with a 100 μm spacer (Figs. 5a and  S29). Before the photocurrent experiment, a soft agarose pad was gently pressed on top of the samples to ensure close contact between the bacteria and the electrode. This procedure was to flatten the sample into a thin layer, preventing cells from stacking above or below the focal plane and allowing accurate single-cell current measurements. Fluorescence imaging confirmed that both the bare bacteria and PEDOT-S biocomposite samples were arranged nearly as monolayers (Figs. S30 and S31), with no significant out-of-focus cell clusters in the detected areas.

Fig. 5: Multimodal imaging and photocurrent mapping of S. elongatus and PEDOT-S/S. elongatus biocomposites.
figure 5

a Schematic illustration of the multimodal imaging platform integrating a photoelectrochemical microfluidic cell with a fluorescence microscope. The setup utilizes multiple modulated lasers to enable both wide-field and focused excitation. The FTO working electrode is connected to a potentiostat for photoelectrochemical current mapping. b, c Fluorescence images of S. elongatus cells and the representative zoomed-in images of the regions indicated by dotted circles. Excitation at 405 nm, emission at 650–750 nm. d, e Fluorescence images of PEDOT-S/S. elongatus biocomposites and the representative zoomed-in images of the regions indicated by dotted circles. Each dotted area contains 15 bacteria. Scale bars indicate the magnification level. f Local current–time (J–t) profiles for bare S. elongatus (red curve) and PEDOT-S/S. elongatus biocomposites (blue curve) under a 405 nm light on-off cycle, corresponding to the regions in (c, e). g Plot of Jph as a function of bacterial count for bare S. elongatus (red data points) and PEDOT-S/S. elongatus biocomposites (blue data points). The error bars represent the standard error of the mean. For bare S. elongatus, replicates (n) of 3, 2, and 3 for bacterial numbers of 10, 15, and 26, respectively. For PEDOT-S/S. elongatus biocomposites, replicates (n) of 5, 2, 2, 11, 3, 8, 9, 5, and 4 for bacterial numbers of 3, 4, 5, 6, 8, 9, 10, 11, and 14, respectively. h Jph per cell for bare S. elongatus (red bars, n = 9) and PEDOT-S/S. elongatus biocomposites (blue bars, n = 49). Box plots show the distribution of the data: the center line indicates the median, the box bounds represent the lower to upper quartile values of the data, and the whiskers extend to the minimum and maximum values within 1.5× the interquartile range from the box. Individual data points in g, h represent both biological and technical replicates. Statistical analysis was performed using one-way ANOVA followed by Tukey’s Honestly Significant Difference (HSD) post-hoc test. All tests were two-sided, and p values were adjusted for multiple comparisons. The p-value for the comparison between bare cells and biocomposites is 0.001 (*p < 0.05, **p < 0.01).

Full size image

To quantify the photocurrent at localized regions, we utilized a focused 405 nm laser to excite cells within areas approximately 20–30 μm in diameter. The J versus time (J-t) response under on-off light illumination was probed at ECA = 0.5 V (versus Ag/AgCl). The focused light ensured that only the bacteria within the illuminated regions contributed to the photocurrent. Upon light excitation, bare S. elongatus (bacteria number of 15) generated a Jph of 0.28 ± 0.07 nA. In contrast, PEDOT-S biocomposites with the same cell count exhibited a Jph approximately ten times higher, measuring 2.13 ± 0.19 nA (Fig. 5f). This increase suggests that PEDOT-S effectively facilitates charge transfer at the abiotic-biotic interface. By randomly illuminating different regions of the FTO substrate, Jph measurements were taken for both bare S. elongatus and biocomposites across varying cell numbers (Figs. S32–S45). The PEDOT-S biocomposites showed a nearly linear increase in photocurrent with an increasing number of bacteria (Fig. 5g). This linearity suggests that the PEDOT-S conductive matrix promotes interfacial electron transfer capabilities of individual cells as well as consistent charge transfer across multiple cells. In contrast, the bare S. elongatus shows a much lower photocurrent with less dependence on cell count. This result is consistent with the bulk experiments, in which we observed a nearly linear increase of the photocurrent with increasing bacterial loading in the PEDOT-S matrix, while the bare S. elongatus shows an obvious saturation (Fig. S24). Normalizing the photocurrent to the number of cells shows that the per-cell Jph of PEDOT-S biocomposites (0.18 ± 0.06 nA) was approximately tenfold higher than that of bare S. elongatus cells (0.018 ± 0.004 nA) (Fig. 5h). This enhancement in single-cell photocurrent suggests that the PEDOT-S biocomposite facilitates efficient interfacial electron transfer of individual cells.

Comparison of PEDOT-S matrix properties with previous studies

It is worth pointing out that a variety of PEDOT-based materials have been used previously in bioelectronics with the intent of improving the ability to harness the electrons of bacteria or photosynthetic microorganisms38,39,41,61. While many of these studies employ PEDOT:PSS or electrochemically deposited PEDOT films, our work utilizes a self-doped PEDOT-S, characterized by delocalized π-backbone bearing ionic alkyl chains. Such a structure not only provides water solubility and biocompatibility but also allows the formation of a soft 3D matrix through interchain ionic cross-links and hydrophobic contacts. The mixed ionic-electronic conduction of the gel facilitates charge transport under the mildly ionic conditions of photosynthetic cultures. Such properties are particularly relevant when compared to prior PEDOT coatings that rely on additional dopants or additives that can potentially reduce biocompatibility and electrochemical properties of the polymer.

In previous studies, coating electrodes with conductive polymers such as polyaniline, polypyrrole, or PEDOT-based materials has typically yielded improvements in photocurrent outputs on the order of two- to tenfold, depending on the microorganisms and specific polymer formulation13,34,38,39,72. In this study, we observed a more than 20-fold increase in photocurrent in the absence of a mediator and an approximately 6-fold increase in the presence of ferricyanide when PEDOT-S was integrated with S. elongatus compared to the bare cells. The increment in photocurrent is attributed to the increased ESA and low charge transfer resistance of the PEDOT-S network thus improving the cell–electrode interface and harnessing of photosynthetically generated electrons.

Compared to strategies focused on electrode architectures—such as inverse opal–indium tin oxide (IO-ITO)6,73 or hierarchical pillar array ITO structures44—that enhance photocurrent output by engineering electrodes with porous or pillar-based surfaces, our approach focuses on improving cell–electrode communication. This is achieved by forming a soft, 3D polymer matrix that intimately connects individual cells to the electrode. These approaches are distinct yet complementary: incorporating conductive polymer and CPEs like PEDOT-S into high-surface-area architectures such as IO-ITO could further enhance cell loading and reduce charge transfer resistance, potentially leading to even greater photocurrent. Given the straightforward synthesis of PEDOT-S and the scalability and cost considerations of the biocomposites (Supplementary Note 3), merging these methods may unlock opportunities for high-performance, scalable biophotovoltaic devices.

Discussion

In summary, we developed a photosynthetic 3D living biocomposite by integrating a self-doped conjugated polyelectrolyte PEDOT-S with the photoactive microorganism S. elongatus. The living biocomposite amplified the photocurrent output by more than 20-fold without a mediator relative to S. elongatus alone and by approximately 6-fold with a mediator present. Electrochemical characterization revealed that PEDOT-S enhanced interfacial electron transfer from S. elongatus cells to the electrode. Furthermore, the photosynthetic function of the microorganism is maintained in the biocomposite under continuous illumination. Single-cell level experiments offer further insights into microscopic cell-electrode interactions facilitated by the PEDOT-S matrix. From a broader perspective, the approach presented here integrates the individual properties and advantages of both biotic and abiotic components, leveraging the natural photosynthetic capabilities of microorganisms and the characteristic electronic properties of conjugated polyelectrolytes. The potential applications of such biocomposites extend to various renewable energy technologies with the perspect of providing a sustainable solution for green energy generation. Meanwhile, single-cell experiments will bridge the gap between understanding individual cell behavior and applying it to larger systems, driving innovation across bioelectrochemical technologies.

Methods

Culture of cyanobacterium Synechococcus elongatus PCC 7942

S. elongatus was purchased from Chemoscience Pte Ltd for this study. The concentration of the bacterial cells was quantified using a spectrophotometer to measure the optical density at 730 nm (OD730). The cell culture medium, BG11, was prepared following the American Type Culture Collection (ATCC) guidelines, comprising 1.5 g L−1 NaNO3, 0.04 g L−1 K2HPO4, 0.075 g L−1 MgSO4 × 7H2O, 0.036 g L−1 CaCl2 × 2H2O, 6.0 mg L−1 citric acid, 6.0 mg L−1 ferric ammonium citrate, 1.0 mg L−1 EDTA, 0.02 g L−1 Na2CO3, 1.0 mL Trace Metal Mix A5 containing 2.86 g L−1 H3BO3, 1.81 g L−1 MnCl2 × 4H2O, 0.222 g L−1 ZnSO4 × 7H2O, 0.039 g L−1 Na2MoO4 × 2H2O, 0.079 g L−1 CuSO4 × 5H2O, 49.4 mg L−1 Co(NO3)2 × 6H2O. The pH of the medium is adjusted to 7.1 to minimize abiotic pseudo-photocurrent from manganese ions in weakly alkaline environments (pH ~8)15.

For culturing, a 10 mL aliquot of BG11 was sterilely inoculated with a cryopreserved stock of S. elongatus and revived under continuous white light illumination at 0.4 mW cm−2 (approximately 20 μmol photons m−2 s−1), at a temperature of 30 °C within a 1% CO2 environment. This preculture was then streaked onto BG11 agar plates and incubated at 30 °C for colony formation. Once colonies were confirmed on agar, a small portion of the biomass was inoculated into a liquid BG11 medium. The growth of these liquid cultures was regularly monitored by measuring their absorbance at 730 nm. Before biophotovoltaic experiments, a mild ethylenediaminetetraacetic acid (EDTA) treatment was conducted to remove extracellular polymeric substances (EPS) of S. elongatus to facilitate extracellular electron transfer74,75,76. Briefly, approximately 5 mg wet-cell-biomass mL−1 was mixed with an equal volume of 2% EDTA in BG11 medium and incubated at 4 °C for 3 h77. Subsequently, cultures were centrifuged at 5724 × g for 5 mins and washed with fresh BG11 medium three times. The final cultures, adjusted to an OD730 of approximately 0.9 (equivalent to about 4.5 × 108 cells mL−1), were used for further experimental studies involving photosynthetic activity and biophotovoltaic experiments.

Fabrication of microbial three-electrode electrochemical cell

As shown in Fig. S16, an ITO glass substrate with a size of 5 × 2.5 cm and with a sheet resistance of 6–8 Ω served as the working electrode in the setup. To define a precise active area for the electrochemical reactions, a rubber gasket was placed on the electrode surface, creating a confined region of 2 cm2 with a depth of 600 μm. This ensured a consistent and controlled environment for the biocomposite interface.

Before the application of the biocomposite, the ITO electrodes underwent a thorough cleaning process and were treated with ultraviolet (UV) ozone to render the surface hydrophilic. This treatment removes organic contaminants and introduces hydroxyl groups onto the electrode surface, enhancing the adhesion of the biocomposite and ensuring intimate contact between the electrode and the PEDOT-S/S. elongatus layer. Subsequently, 100 μL of the PEDOT-S/S. elongatus biocomposite was carefully dispensed onto the defined working area of the ITO electrode. The biocomposite was evenly spread to form a uniform layer, which is crucial for reproducible electrochemical measurements and efficient wiring between the biological and electronic components. The same procedure is used to prepare bare S. elongatus samples.

To facilitate ion exchange and redox mediator transport while retaining the cells near the electrode, a dialysis membrane with a molecular weight cut-off (MWCO) of 3.5 kDa was placed over the biocomposite or liquid culture layer. The assembled electrochemical cell was filled with 15 mL of BG11 medium, incorporating 1 mM K3[Fe(CN)6] to act as an electron transfer mediator. Carbon felt (area of 1 cm2) and an Ag/AgCl electrode (3 M KCl) were employed as counter and reference electrodes, respectively.

Bioelectrochemical experiments

Electrochemical measurements were performed using a VMP-300 potentiostat, Bio-Logic SAS, under 30 °C incubation, ambient air conditions. Chronoamperometry (CA) was performed at apotential (ECA) of 0.3 V versus Ag/AgCl. Before monitoring the microbial photocurrent, the electrochemical cells were incubated in darkness for 3 hours to stabilize the system and eliminate residual capacitive currents. A white LED light (1 mW cm−2, approximately 50 μmol photons m−2 s−1) controlled by a timer switch was used to affect alternating dark and light conditions.

Cyclic voltammetry (CV) experiments were designed to characterize the electrochemical behaviors and electron transfer at the biocomposite-electrode interface. The working electrode potential was swept from an initial potential (Einitial) of −0.5 V to a vertex potential (Evertex) of 0.6 V, and then returned to a final potential (Efinal) of −0.5 V. This potential sweep was conducted at a scan rate of 0.005 V s−1.

In electrochemical impedance spectroscopy (EIS) experiments, the working electrode potential was maintained at a direct current (DC) potential (EDC) of 0.3 V versus Ag/AgCl. A small sinusoidal alternating current (AC) potential with an amplitude (EAC) of 5 mV was applied spanninga frequency range from 100 kHz down to 100 MHz. EIS data were analyzed using equivalent circuit models to extract parameters including charge transfer resistance and diffusion resistance.

Calculation of charge transfer resistance and restricted diffusion resistance

The charge transfer resistance (RCT) and the restricted diffusion resistance (Rd) were determined following established literature protocols using an equivalent circuit model to fit the impedance spectra (Z) as a function of frequency (f)59. This model is represented by the equation:

$$Z\left(f\right)={R}_{s}+\frac{{R}_{{CT}}}{Q}+{Ma}$$
(1)

In this model, all resistances indicated in the Nyquist plots are normalized by the area of the working electrode (2 cm2). The solution resistance (RS) in the Nyquist plot denotes the equivalent series resistance in Ω cm−2 and corresponds to the initial intersection with the real axis on the high-frequency side of the Nyquist plots. The RCT, in Ω cm−2, is identified at the second intersection with the real axis in the mid-frequency section of the plots. The element Q represents a constant phase element that accounts for potential deviations from an ideal capacitor.

Additionally, the presence of a finite space Warburg (FSW) impedance, indicated by a linear section terminating with an interfacial capacitor in the Nyquist plots, is modeled by the modified restricted diffusion element Ma, defined as:

$${Ma}={R}_{d}\frac{\coth \left(\root {0.5a} \of{\tau j\omega }\right)}{\root {0.5a} \of{\tau j\omega }}$$
(2)

Here Rd is the restricted diffusion resistance (Ω cm−2), τ is relaxation time (s) and Ω is angular frequency (rad s−1). From this modeling, we can obtain the RCT and Rd values for PEDOT-S and biocomposites. The fit quality was evaluated using the normalized Chi-squared (χ2/|Z|) value, with values lower than 0.01 indicating a reliable agreement between the experimental data and the model.

Scanning electron microscopy measurement

After the photoelectrochemical experiments, the medium in the electrochemical chambers was replaced with 4% glutaraldehyde prepared in fresh BG11 medium to fix the electrode-associated cells. The electrodes were incubated in this fixative solution for 24 hours at room temperature to ensure thorough fixation of the cellular structures. Following fixation, the electrodes were sequentially rinsed twice with BG11 medium and deionized water to remove residual glutaraldehyde and culture medium components.

Subsequent dehydration of the samples was performed using a graded ethanol series to gradually remove water while minimizing cellular distortion. The electrodes were immersed twice in 35% ethanol, followed by twice in 70% ethanol, each step lasting 1.5 hours. Finally, the electrodes were treated with 100% ethanol, with the first immersion lasting 1.5 hours and the second immersion extended overnight to ensure complete dehydration.

After dehydration, the electrodes underwent CO2 critical point drying to prevent structural collapse during drying45. At the critical point, CO2 transitions from liquid to gas phase without crossing a liquid–gas boundary, thus avoiding surface tension effects that can damage delicate structures or porosity in electrodes. This method preserves the microstructure of the biocomposites. Specifically, after disassembling the electrochemical cells, the working electrodes were gently rinsed with ethanol and then subjected to CO2 critical point drying using a Leica EM CPD300 critical point dryer. The electrodes underwent sequential solvent exchange from ethanol to liquid CO2. Once the chamber is filled with liquid CO2 and all ethanol has been replaced, increase the temperature and pressure in the chamber to reach the critical point of CO2 (31.1 °C, 7.38 MPa). The CO2 gas was then slowly vented from the chamber, allowing the pressure to drop gradually to atmospheric levels while maintaining the chamber above the critical temperature. Finally, the dried electrodes were stored under inert atmosphere to avoid moisture uptake.

The dried electrodes were sputter-coated with a thin layer of platinum using an Anatech Hummer Sputtering System to enhance electrical conductivity and reduce charging effects during imaging. Scanning electron microscopy (SEM) imaging was conducted using a Hitachi FlexSEM1000. Imaging parameters were set with an accelerating voltage of 5 kV, a working distance of approximately 6 mm, and a beam current of 80 pA. Secondly, electron detection mode was employed to obtain high-resolution images of the microstructures of the PEDOT-S/S. elongatus biocomposites.

Confocal laser scanning microscopy

The microstructure and spatial distribution of S. elongatus cells within the PEDOT-S/S. elongatus biocomposites on ITO glass electrodes were examined using a confocal laser scanning microscope (Olympus FV3000). A ×60 oil immersion objective lens was utilized to image the samples. The intrinsic red fluorescence of chlorophyll within the S. elongatus cells was captured by exciting the samples at a wavelength of 620 nm and collecting emissions between 650 and 750 nm.

To investigate the three-dimensional architecture of the biocomposites, z-stack imaging was performed over a depth of 30 μm. This approach allowed for the reconstruction of the biocomposite structure along the z-axis, providing insights into the distribution and integration of S. elongatus cells within the PEDOT-S matrix. Image acquisition parameters were carefully optimized to minimize photobleaching and ensure an accurate representation of the cellular structures.

The acquired image stacks were processed and analyzed using Imaris Viewer software. This software facilitated three-dimensional visualization, rendering, and quantitative analysis of the cell distribution within the biocomposites.

Single-cell multimodal microscopy platform setup

The bacterial-cluster photocurrent measurement is based on an inverted microscope equipped with epifluorescence illumination and a customized three-electrode photoelectrochemical microfluidic cell. A 405 nm laser beam was expanded 15-fold using a beam expander (Thorlabs, GBE15-A), reflected by 425 nm long-pass dichroic mirror (Thorlabs, DMLP425R) and subsequently focused through a ×60 water immersion objective lens (Olympus, UPLSAPO60XW). An adjustable iris was positioned at the back port of the microscope, before focusing the light onto the back aperture of the objective, to precisely control the illuminated area. This setup enabled selective excitation of specific regions containing only a few isolated S. elongatus or PEDOT-S/S. elongatus biocomposites, facilitating the measurement of photocurrents originating exclusively from these targeted bacterial clusters. The illumination area was defined by a laser spot with a diameter of approximately 10–20 μm and a power density of 8.18 W cm−2. The number of bacteria within the illuminated area was determined by detecting their intrinsic fluorescence, which was filtered through a 700 ± 25 nm emission filter (Edmund) and captured using a scientific Complementary Metal–Oxide–Semiconductor (sCMOS) camera (ANDOR SONA). The imaging process was controlled by ANDOR SOLIS software.

In the customized three-electrode photoelectrochemical microfluidic cell, FTO electrodes were used as the substrate, onto which 1–2 µL of either bare S. elongatus or PEDOT-S/S. elongatus biocomposites were deposited. These were subsequently covered with an agarose pad approximately 100 µm thick to ensure close contact between the bacteria and the electrode. BG11 medium containing 1 mM K3Fe(CN)6 was introduced into the microfluidic cell using a syringe pump (Chemyx Incorporation), after which the flow was halted to prevent disturbances in the current. The FTO electrode with the deposited cells acted as the working electrode, controlled by a potentiostat (Bio-Logic SP 300), while a platinum wire and an Ag/AgCl electrode served as the counter and reference electrodes, respectively. Chronoamperometric measurements were performed at 0.5 V during light on–off cycles, typically 10 seconds of illumination followed by 20 seconds of darkness, repeated over five cycles. The resulting current-time (i-t) responses were recorded. To quantify the photocurrent, the currents before and after the onset of illumination were fitted with linear functions. The difference between these fitted functions at the time point was used to determine the photocurrent for each light on/off cycle.

Compared to 0.3 V applied in bulk experiments, the use of 0.5 V in single-cell photocurrent mapping is mainly for enhancing sensitivity and improving signal-to-noise ratio. Single-cell measurements inherently produce much smaller currents compared to bulk measurements due to the limited number of cells contributing to the signal. To reliably detect these low current signals, a higher applied potential is necessary to increase the driving force for electron transfer and help to distinguish the photocurrent generated by individual cells from the background noise and capacitive currents.

Statistics analysis

Results are presented as mean ± standard error of the mean (SEM) taken from distince samples (n ≥ 3) unless otherwise stated. Statistical analysis was performed using one-way ANOVA followed by Tukey’s Honestly Significant Difference (HSD) post-hoc test. Asterisk (*) denotes statistical significance between bars (**P < 0.01, and *P < 0.05). The values of P < 0.01 were considered to be statistically significant. The statistical analyses and graphs were performed using OriginLab 2018.

Reporting summary

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