Introduction

Over the past several decades, the exploration of various renewable energy to reduce the consumption of fossil fuels and sustain the global energy supply has acquired much interest1,2. Analogous to natural photosynthesis, where green plants convert solar energy into chemicals, the artificial photosynthesis or photocatalysis process can directly utilize semiconductor materials to harvest solar energy towards the synthesis of various chemical fuels, offering a promising and straightforward approach to addressing this challenge3,4,5. However, despite notable breakthroughs and advancements throughout the years, the low solar-to-chemical efficiency and inadequate materials stability continue to hinder the industrial application of photocatalysts6,7. Additionally, poor selectivity for specific products further limits their practical use in the future8,9,10.

Recently, the combination of semiconductor materials with good light-harvesting and electroactive microorganisms with complete intracellular biocatalytic pathways has emerged as a potential strategy to address these challenges11,12,13. In biotic-abiotic hybrid systems, semiconductor photoelectrodes or photocatalysts can effectively facilitate photogenerated charge separation and promote electron transfer at the microbe-semiconductor interfaces, thus enhancing metabolic activities of the microorganisms14,15,16. Up till now, a range of biotic-abiotic hybrid systems have been successfully constructed for the solar-driven production of chemical fuels, such as H2 and CH417,18,19. Regrettably, unfavorable contacts and slow diffusion of redox mediators such as flavins at the microbe-semiconductor interfaces often result in low electron transfer rate and suboptimal solar energy conversion efficiency20,21. Moreover, the complex interactions between semiconductors and microorganisms as well as their heterogeneities complicate the elucidation of interfacial electron transfer mechanisms derived from bulk-level measurements (Fig. 1a)22,23. Therefore, it is essential to explore atomically precise interfacial connections and quantitatively probe interfacial electron transfer at the single- or sub-cellular levels towards efficient solar-to-chemical conversion17,24,25. In the field of artificial photosynthesis, single-atom catalysts have gained attention due to their maximized atomic utilization and abundant active sites26,27. A single atom can regulate the electronic structure and valence of semiconductor materials28,29, offering great possibilities to enhance interfacial contact and reduce interfacial impedance between photocatalysts and microorganisms. Furthermore, single-atom catalysts with unique structures provide a valuable platform to uncover electron transfer processes and catalytic mechanisms at the biological-material interfaces20,30,31.

Fig. 1: Schematic illustration of single-atom bridges across biotic-abiotic interfaces facilitating electron transfer for solar-driven H2 production.
figure 1

a Conventional EET mechanism through direct contact with cytochromes or diffusing flavins: understanding EET at the semiconductors- or electrodes-microbe interface from bulk-level measurement. b Single-atom bridges across the PCN/Ru-SHE interfaces facilitate charge transfer via cytochromes C-mediated direct and flavin-mediated indirect pathways from PCN/Ru to intracellular hydrogenases for solar-driven H2 production, with the major enhancement from direct pathway: understanding EET at the semiconductors-/electrodes-microbe interface from operando single-cell measurement. Created in BioRender. Jiajing, G. (2025) https://BioRender.com/fomstig.

Full size image

In this work, we report a general strategy of facilitating direct electron uptake via building single-atom bridges across biotic-abiotic interfaces to enhance solar-driven H2 production (Fig. 1b). Ru-decorated porous C3N4 (denoted as PCN/Ru) and Shewanella oneidensis (abbreviated as SHE) were selected to construct a biotic-abiotic biohybrid system, leveraging the biocompatibility of the organic semiconductor and the complete extracellular electron transfer (EET) pathways32,33,34. Aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and synchrotron radiation X-ray absorption spectra (XAS) confirm the successful incorporation of Ru single atoms onto PCN, featuring a unique Ru-N4 structure35. After the construction of the PCN/Ru-SHE biohybrid system, optical and photoelectrochemical characterizations reveal that Ru single atoms promote effective charge separation and reduce the electron transfer barrier at the biohybrid interfaces. Notably, operando single-cell photocurrent measurement demonstrates that the PCN/Ru-SHE with a Ru-N4 interfacial configuration achieves a 3.0-fold increase in inward EET compared to PCN-SHE22. Cyclic voltammetry (CV) analysis further unravels that the PCN/Ru-SHE biohybrids could enhance both cytochromes C-mediated direct and flavin-mediated indirect electron transfer, with the majority of the improvement stemming from the direct EET process. Additionally, density functional theory (DFT) calculations also provide insight into the interfacial connection between PCN/Ru and SHE, showing that the distinct Ru-N4 structure ensures close contact and promotes rapid electron transfer from PCN to SHE through both direct and indirect EET processes. In stark contrast to SHE and PCN-SHE, the optimized PCN/Ru-SHE hybrid system exhibits 47.5- and 14.2-fold improvements in H2 production, respectively, achieving a superior quantum yield of 8.46% at 450 nm. Proteomic analysis suggests that the single-atom-induced modulation of microbial metabolism favors effective charge transfer and H2 production, with similar results observed in PCN/Cu-SHE biohybrids. This work highlights a strategic optimization of single atom facilitating direct charge transfer and provides a quantitative understanding of the electron transfer mechanisms at the microbe-semiconductor interfaces from an atomic perspective for enhanced solar energy utilization.

Results

Synthesis and properties of PCN/Ru single atom nanosheets

Using melamine as a raw material, we firstly synthesized bulk C3N4 (donated as BCN) by calcination at 550 °C with the rate of 3 °C/min (see details in Methods and Supplementary Fig. 1), and the morphology of BCN was characterized by transmission electron microscopy (TEM) (Supplementary Fig. 2). Then, the porous C3N4 (donated as PCN) nanosheets were obtained by second calcination at 500 °C and followed by water exfoliation (see details in Methods)36. TEM images and Energy-dispersive X-ray spectroscopy (EDX) mapping clearly exhibit its porous morphology (Supplementary Fig. 3 and Supplementary Fig. 4). Next, Ru single atoms were loaded onto PCN nanosheets (donated as PCN/Ru) through adding RuCl3 in the PCN methanol solution and further annealing at 160 °C for 1 h under Ar atmosphere (see details in Methods)37. As depicted in TEM and high-resolution TEM (HRTEM) images (Supplementary Fig. 5), there is no obvious morphological change of PCN before and after loading Ru single atoms, while the EDX mapping (Supplementary Fig. 5 and Supplementary Table 1) and EDX spectrum (Supplementary Fig. 6) clearly show the uniform distribution of Ru element on PCN nanosheets. Aberration-corrected HAADF-STEM was further conducted to directly confirm the existence of Ru single atoms onto PCN nanosheets. HAADF-STEM images show that the Ru species are atomically dispersed in PCN nanosheets (Fig. 2a and Supplementary Fig. 7), and no observable Ru nanoparticles are observed. HAADF-STEM EDX mapping also reveals that the Ru atoms are distributed uniformly on PCN nanosheets (Supplementary Fig. 7), further verifying the successful synthesis of Ru single atoms. Besides, we also synthesized Ru nanoparticles on PCN nanosheets for comparison. It can be clearly seen that TEM and HRTEM images depict an interplanar spacing of 0.24 nm (Supplementary Fig. 8), corresponding to (100) facet of Ru nanoparticles, providing another evidence for the successful synthesis of Ru single atoms in our work.

Fig. 2: Preparation of PCN/Ru single atom nanosheets.
figure 2

a Aberration-corrected HAADF-STEM image of PCN/Ru nanosheets. The bright dots are Ru single atoms on PCN nanosheets. b Normalized XANES spectra at the Ru K-edge of PCN/Ru nanosheets. Inset: absorption edge positions of Ru XANES spectra. c Fourier transforms of EXAFS spectra at the Ru K-edge. d Fitting of Ru coordination environment in PCN/Ru. Inset: Theoretical model of Ru-N4 configuration. e XANES spectra at the N K-edge of PCN/Ru nanosheets. f EIS spectra of BCN, PCN, and PCN/Ru. WT-EXAFS of the Ru K-edge of g Ru foil, h RuO2, and i PCN/Ru.

Full size image

X-ray diffraction (XRD) patterns and Fourier transform infrared (FTIR) spectra were also performed to ascertain the formation of PCN/Ru nanosheets. The XRD peaks at 13.3° and 27.5° (Supplementary Fig. 9), corresponding to (100) and (002) planes of standard g-C3N438, and the FTIR peaks at 1250–1750 cm−1 belongs to the characteristic vibration of C-N and C=N groups (Supplementary Fig. 10)39. Notably, electron spin resonances (ESR) spectra show that the characteristic peak at g = 2.002 represents the emergence of N vacancy (Supplementary Fig. 11)40. The decreased N vacancy signal indicates that Ru atoms are attached to the void center of PCN. X-ray photoelectron spectroscopy (XPS) test was further conducted to examine the chemical interaction and elemental composition of PCN/Ru (Supplementary Figs. 12–14). The distinct peaks of C 1s at 284.5 and 288.3 eV, as well as N 2p at 398.3, 398.8, and 400.3 eV prove the synthesis of PCN (Supplementary Fig. 14)41. The high-resolution Ru 3p peaks are at around 463.5 and 485.6 eV (Supplementary Fig. 14), while the typical binding energies of metallic Ru0 are at 461.2 and 483.7 eV, indicating that the oxidation state of Ru in PCN/Ru is between Ru0 and Ru4+42. Then, synchrotron radiation XAS spectra were performed to ascertain the local coordination environment of PCN/Ru. The normalized X-ray absorption near-edge structure (XANES) spectra of Ru K edge show that the absorption threshold of PCN/Ru is situated between Ru foil and RuO2 (Fig. 2b), further demonstrating that the valence state of Ru in the PCN/Ru is between 0 and +442. The linear fitting of the absorption edge position in XANES spectra reveals that the valence state of the Ru species in PCN/Ru is +2.9 (Supplementary Fig. 15). We also analyzed the Fourier transforms of the Ru K-edge extended X-ray absorption fine structure (EXAFS) spectra (Fig. 2c). In contrast to Ru-Ru bond in Ru foil and Ru-O bond in RuO2, the peak at 1.84 Å can be attributed to Ru-N bond35. No Ru-Ru bond shows up in PCN/Ru, indicating that the Ru element exists in the form of single atoms. The fitting result of EXAFS curve reveals that the coordination number of Ru atom is 3.7 (Fig. 2d), indicating that each Ru atom is coordinated with four N atoms in PCN. Afterwards, the theoretical model also shows that Ru atoms is simulated with four N atoms (inset of Fig. 2d), denoted as a single Ru-N4 site. Wavelet transform (WT)-EXAFS analysis at the Ru K-edge was also performed to provide both K- and R-space resolution of PCN/Ru (Fig. 2i) to show the coordination environment of Ru atoms, using Ru foil (Fig. 2g) and RuO2 (Fig. 2h) as comparison. This result shows that an obvious intensity peak at 5.7 Å−1 in K-space and 1.84 Å in R-space further elucidates the emergence of Ru−N configuration. In addition, as shown in N K-edge XANES spectra (Fig. 2e), the typical π* resonances at 403.4 and 406.3 eV can be ascribed to C=N-C in the tri-s-triazine heteroring and N–C3 bridging among three tri-s-triazine moieties, respectively. In contrast to PCN, the C=N-C peak of PCN/Ru shows an obvious positive shift (0.3 eV), which can be attributed to the strong interaction between PCN and Ru atoms43, resulting in electron transfer from PCN to Ru atoms.

Besides, the optical and photoelectrochemical characterizations were also measured to unveil the charge transfer behavior of PCN/Ru nanosheets. The photoluminescence (PL) spectra show that the PL intensity of PCN/Ru is greatly decreased (Supplementary Fig. 16), indicating that Ru single atoms can reduce the photogenerated charge recombination and facilitate effective electron transfer44. Time-resolved photoluminescence (TRPL) decay spectra were further carried out to calculate the charge recombination lifetimes (Supplementary Fig. 17). In sharp contrast to BCN and PCN, the PCN/Ru shows the shortest PL lifetime (Supplementary Table 2), representing the superior charge transfer ability after loading Ru single atoms onto PCN44. Additionally, compared to BCN and PCN, the smallest arc radius of PCN/Ru observed in electron impedance spectra (EIS) indicates that PCN/Ru has the lowest charge transfer resistance and the fastest charge migration (Fig. 2f). Transient photocurrent curves also show that PCN/Ru has the highest transient photocurrent density, further supporting its superior carrier separation and transfer abilities (Supplementary Fig. 18). Therefore, these characterizations elucidate that the photogenerated electrons could be effectively transferred from PCN to Ru atoms through Ru-N4 site under light irradiation.

Single-atom bridges facilitating direct electron transfer

We then assembled the PCN/Ru-SHE hybrid systems by immersing PCN/Ru nanosheets into the SHE medium. Upon light irradiation, the PCN/Ru could produce photogenerated electrons, which are transferred to SHE, serving as the electron donor for H2 production. The scanning electron microscopy (SEM) images and SEM-EDX mapping clearly show that SHE grows on the PCN/Ru nanosheets (Supplementary Figs. 19 and 20). TEM images (Supplementary Fig. 21) and TEM-EDX mapping further prove the assembly of PCN/Ru-SHE hybrid systems. Notably, after loading Ru atoms on PCN, dynamic light scattering (DLS) reveals that no obvious size changes are observed (Supplementary Fig. 22), but the Zeta potential test of PCN/Ru shows a positive charge (Supplementary Fig. 23), facilitating effective electrostatic interactions between the negatively charged SHE and PCN/Ru nanosheets. Atomic force microscopy (AFM) was also performed to characterize the topography and height of both the PCN/Ru (Supplementary Fig. 24) and the PCN/Ru-SHE hybrid system (Fig. 3a). The images clearly reveal the presence of SHE cells distributed on the PCN/Ru, supporting the successful assembly of PCN/Ru-SHE biohybrid systems.

Fig. 3: Operando single-cell interfacial electron uptake.
figure 3

a AFM of PCN/Ru-SHE biohybrids. b Histograms of single-cell/single-particle photocurrent distribution of PCN and PCN-SHE as well as photocurrent density difference Δi1 obtained of PCN-SHE and PCN at −0.4 V vs Ag/AgCl. c Histograms of single-cell/single-particle photocurrent distribution of PCN/Ru and PCN/Ru-SHE as well as photocurrent density difference Δi2 obtained of PCN/Ru and PCN/Ru-SHE at −0.4 V vs Ag/AgCl. d The differences of Δi1 and Δi2. This difference highlights the function of Ru single-atom bridges at the biohybrid interfaces. Inset: schematic illustration of operando single-/sub-cell photocurrent test and optical microscope images of cells. Created in BioRender. Jiajing, G. (2025) https://BioRender.com/dd89tnd.

Full size image

The charge transfer dynamics at the PCN/Ru-SHE biohybrid interfaces was further investigated. We then examined UV−vis diffuse reflectance spectra (UV−vis DRS) and Mott-Schottky curves to determine the band structures of PCN/Ru (Supplementary Figs. 25–27). The conduction band (CB) at −0.9 V (vs RHE) ensures the prerequisite of inward EET from PCN/Ru to SHE through both cytochrome C-mediated direct and flavin-mediated indirect EET pathways (Supplementary Fig. 28). To directly quantify the capability of SHE cells to take up electrons from PCN/Ru, we employed the operando single-cell photocurrent technique to measure electron transfer at the microbe-semiconductor biohybrid interfaces from our previous report (inset of Fig. 3d)22. SHE cells were dispersed onto PCN/Ru-SHE films, which were placed on a transparent conductive indium tin oxide (ITO) substrate in a photoelectrochemical microfluidic cell. The dispersed SHE cells on the PCN/Ru nanosheets could be easily observed in a wide-field optical microscope (Supplementary Figs. 29–31). Under a focused 405 nm laser excitation, the single-cell photocurrent response (I-t) of PCN/Ru-SHE was clearly shown (Supplementary Fig. 31). The single-cell/single-particle photocurrents of different biohybrid systems were measured for multiple cycles, and the histograms of the single-cell/single-particle photocurrent density for each system were recorded (Supplementary Fig. 32). Among these biohybrid systems, the PCN/Ru-SHE biohybrids clearly exhibited the largest single-cell photocurrent density (Supplementary Fig. 32 and Supplementary Table 3). We then compared photocurrent density distributions between PCN and PCN-SHE and defined their differences as Δi1 (Fig. 3b), which represents the capability of SHE to uptake electrons from PCN. Meanwhile, the Δi2 measures the photocurrent density differences of PCN/Ru and PCN/Ru-SHE (Fig. 3c), demonstrating the ability of SHE to capture electrons from PCN/Ru. In sharp contrast to Δi1, the Δi2 showed a 3.0-fold increase in photocurrent response (Fig. 3d), indicating that the photogenerated electrons from PCN/Ru could be effectively transported to SHE cells. This result straightforwardly highlights that Ru atoms at the biohybrid interfaces can enable effective electron transport from PCN/Ru to SHE. Notably, the absence of a significant photocurrent response from SHE suggests that it cannot directly convert light into electrons (Supplementary Fig. 30), supporting that SHE is not a photoactive bacterium but rather a typical electroactive bacterium with complete EET pathways.

In addition, we surveyed the optical properties to explore the charge transfer dynamics at the microbe-semiconductor interfaces. The lowest photoluminescence (PL) intensity observed for PCN/Ru-SHE indicates that charge transfer occurs from PCN/Ru to SHE (Fig. 4a). We also measured the TRPL spectra and calculated PL lifetime of PCN/Ru-SHE biohybrid system (Fig. 4b and Supplementary Table 4). In view of PL lifetimes of PCN/Ru (6.52 ns) and PCN-SHE (6.81 ns), PCN/Ru-SHE has the shortest PL lifetime (5.14 ns), further supporting effective electron transfer from PCN/Ru to SHE. Besides, EIS spectra were also tested to investigate the charge transfer resistance from PCN/Ru to SHE cells. The EIS results show that PCN/Ru-SHE biohybrid systems exhibit the lower charge transfer impendence and the faster electron migration dynamics compared to SHE (Supplementary Fig. 33), further indicating rapid charge transfer at the PCN/Ru-SHE biohybrid interfaces.

Fig. 4: Charge transfer dynamics at the biotic-abiotic interfaces.
figure 4

a PL spectra of PCN/Ru-SHE biohybrid systems. b PL lifetimes of PCN, PCN-SHE, PCN/Ru, and PCN/Ru-SHE hybrid systems. c Calculated indirect transfer and direct transfer current density of PCN-SHE and PCN/Ru-SHE biohybrids from CV curves. d ELF analysis of PCN/Ru and Heme 10. e Charge density differences at the interface of PCN/Ru and Heme 10. f ELF analysis of PCN/Ru and flavin. g Charge density differences at the interface of PCN/Ru and flavin. The yellow and green regions in (e) and (g) represent the charge accumulation and charge depletion, respectively.

Full size image

Next, we conducted CV analysis to investigate the influence of PCN/Ru on the EET pathways (Fig. 4c and Supplementary Fig. 34)45,46. The peaks in the CV curves observed around −0.2 V and 0.1 V correspond to the flavin-mediated indirect catalytic current and the cytochrome C-mediated direct catalytic current. The calculated indirect and direct EET current densities for PCN-SHE are 0.04 and 0.01 mA/cm2, while the PCN/Ru-SHE biohybrids achieve 0.07 and 0.11 mA/cm2, reflecting 1.75- and 11.0-fold increases in indirect and direct transfer pathways, respectively. This result indicates that the unique Ru-N4 structure at the PCN/Ru and SHE interfaces primarily enhances the cytochrome C-mediated direct EET pathway. DFT calculations were further conducted to build interface models of PCN/Ru-SHE biohybrids and confirm the mechanism of electron transfer from PCN/Ru to SHE. The primary components of cytochrome C and flavin contacting with the extracellular materials are Heme 10 at the surface of MtrC and the electron shuttling molecule, riboflavin30. We build the interface model of Heme 10 and PCN/Ru nanosheets to investigate direct EET pathway (Supplementary Fig. 35)47. Electron localization function (ELF) analysis is a position-dependent function ranging from 0 to 1, which can reflect the delocalization and delocalization of electrons. ELF analysis reveals the emergence of strong bonding interaction between PCN/Ru and Heme 10, a prerequisite for interfacial electron transfer (Fig. 4d and Supplementary Fig. 36). Based on the calculated charge density differences for PCN/Ru-Heme 10 (Fig. 4e), it can be observed that charge accumulation occurs at Heme 10 while charge depletion occurs at Ru-N4. Bader charge analysis further confirms that 0.16 e- is transferred from PCN/Ru to Heme 10 through direct EET pathway (Supplementary Fig. 37 and Supplementary Table 5). Similarly, we also perform DFT calculations to investigate charge transfer dynamics between PCN/Ru and flavin (Supplementary Fig. 38). ELF analysis (Fig. 4f and Supplementary Fig. 39) reveals the ionic bond would show up between Ru and N. Charge density differences (Fig. 4g) and Bader charge analysis (Supplementary Fig. 40 and Supplementary Table 6) of PCN/Ru-flavin also indicate that 0.54 e- migrates from PCN/Ru to flavin through the Ru-N4 site, suggesting the effective electron transfer from PCN/Ru to flavin molecule through indirect EET pathway. These theoretical calculations also offer valid evidence for direct and indirect EET processes from PCN/Ru to SHE cells.

Solar-to-chemical conversion and proteomic analysis

To assess whether the enhanced charge transfer at the PCN/Ru-SHE interfaces could improve solar energy conversion and facilitate bacterial metabolism, we conducted solar-driven H2 production experiments under simulated solar irradiation (see details in Methods). The H2 production yield of PCN/Ru-SHE biohybrids reaches 18.6 µmol g−1h−1 (Fig. 5a), representing 47.5-fold and 14.2-fold increases compared to SHE and PCN-SHE, respectively. This result indicates that the Ru-N4 structure at the PCN/Ru-SHE interfaces significantly promotes efficient H2 production through rapid electron transfer. Furthermore, all the biotic-abiotic hybrid systems exhibit improved H2 evolution in contrast to their pristine counterparts, underscoring that the semiconductors under excitation could empower SHE with improved inward EET for H2 metabolic activity. Notably, the quantum yield (QY) of the PCN/Ru-SHE biohybrid system (Supplementary Fig. 41) reaches 8.46% at 450 nm, a remarkable efficiency among all the reported works (Fig. 5b and Supplementary Table 7). We then measured the concentrations of ATP and NADH/NAD+ before and after H2 production (Fig. 5c, d). The results show that photogenerated electron injection could facilitate the production of intracellular ATP and NADH/NAD+ in SHE, highlighting their role in generating a proton motive force that enhances the activity of hydrogenases17. Control experiments confirm that light has no obvious effect on SHE metabolism for H2 production, reinforcing that SHE is a non-photoactive bacterium (Supplementary Fig. 42), consistent with our single-cell photocurrent tests. Meanwhile, after three testing cycles, H2 production remains nearly unchanged (Supplementary Fig. 43), demonstrating the stability of the PCN/Ru-SHE biohybrids, which was also validated by XRD patterns and SEM images (Supplementary Fig. 44). Meanwhile, the solar-driven H2 production (Supplementary Fig. 45) over 3 days shows that the H2 production yield slightly drops with time, but after we refresh the medium and further conduct H2 production, the H2 production yield over PCN/Ru-SHE biohybrid could be almost recovered, further supporting the good stability of PCN/Ru-SHE biohybrids. Besides, the cell growth curves reveal that the number of SHE cells with the addition of PCN/Ru is nearly the same as in the control groups after 24 h (Supplementary Fig. 46), indicating its excellent biocompatibility.

Fig. 5: PCN/Ru-SHE hybrid systems for H2 production and proteomic analysis.
figure 5

a H2 production of BCN, PCN, and PCN/Ru with and without assembly with SHE (n  =  3 independent samples; mean ± SD). b The quantum yields of PCN/Ru-SHE and reported biohybrid systems. c ATP level of SHE and PCN/Ru-SHE before and after 3 h light irradiation (n  =  3 independent samples; mean ± SD). d NADH/NAD+ level of SHE and PCN/Ru-SHE before and after 3 h light irradiation (n  =  3 independent samples; mean ± SD). e PCA and f Volcano plot of the results of proteomic analysis comparing SHE and PCN/Ru-SHE biohybrid after solar-driven H2 production. Circles in (e) mean 95% confidence; Lines in (f) represent the gating of FC of photocatalytic PCN/Ru-SHE biohybrid/SHE > 1.5 or <0.66 and P < 0.05 from a two-tailed t-test. g Heatmaps of significantly up-regulated proteins. The proteins and relevant Gene ID: Protein related to flavin synthesis: Q8EC43, Q8E9V9; Outer membrane: Q8EAK6, Q8EFZ3, Q8EBH3, Q8EHD1, Q8EJQ9, Q8EAD7, Q8EJZ4, Q8EC06, Q8EBD3; ATPase: Q8EGB6, Q8E8B6; Cytochrome: Q8E8Y5, Q8E8Y9; NAD(P)H reductase: Q8EID5, Q8EAZ9; Ferredoxin: Q8EJI2.

Full size image

To further explore microbial metabolism and charge transfer pathways at the biohybrid interfaces, we conducted proteomic analyses for the PCN/Ru-SHE biohybrid systems following photocatalytic H2 production. We identified a total of 2,622 proteins and compared their relative abundances between SHE and PCN/Ru-SHE biohybrids. Principal component analysis (PCA) score plots (Fig. 5e) show significant differences in protein profiles between the two systems (95% confidence). The volcano plot for the biohybrids (Fig. 5f) reveals that 210 proteins are significantly up-regulated (fold-change (FC) > 1.5 or < 0.66, P < 0.05) compared to SHE. We analyzed these up-regulated proteins and their associated biological processes using the UniProt and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases (Fig. 5g). The proteomic analysis shows that proteins related to EET and hydrogen metabolism, such as proteins involved in flavin synthesis, outer membrane proteins, cytochromes, ATPase, and NADPH reductase, are markedly expressed in the PCN/Ru-SHE biohybrids, facilitating effective electron uptake and intracellular hydrogen production. These significantly enhanced biological processes suggest that single atoms at the interfaces promote both direct and indirect electron transfer in the microbe-semiconductor biohybrid systems, further unraveling that a well-designed microbe-material hybrid can regulate microbial metabolic activity.

Generality of single atom facilitating direct electron transfer

Additionally, to confirm that single atoms can generally facilitate direct electron transfer at the biohybrid interfaces, we also constructed PCN/Cu-SHE hybrid systems towards solar-to-chemical conversion. STEM image and EDX mapping (Fig. 6a and Supplementary Fig. 47) show the obvious distribution of Cu single atoms on PCN nanosheets. XANES spectra of Cu K edge (Fig. 6b) reveal that the valence of Cu is +2, and the only Cu-N bond in the EXAFS spectra of PCN/Cu (Fig. 6c) also confirms the successful synthesis of Cu single atoms. High-resolution XPS peaks of Cu 2p (Supplementary Fig. 48) also support the incorporation of Cu2+ on PCN nanosheets. We then assembled the PCN/Cu-SHE biohybrids (Supplementary Figs. 49 and 50), and positive Zeta potential of PCN/Cu contributes to the effective contact with negatively charged SHE (Supplementary Fig. 51). CV results (Fig. 6d and Supplementary Fig. 52) show that the Cu atoms at the biohybrid interfaces mainly promote direct EET pathway, further indicating the function of single atom mediating direct electron transfer pathway in the PCN/Cu-SHE hybrid system. This superior electron uptake capability of PCN/Cu-SHE biohybrids also improves solar-driven H2 production (Fig. 6e).

Fig. 6: Generality of single atom facilitating direct electron transfer.
figure 6

a Aberration-corrected HAADF-STEM image of PCN/Cu. b Normalized XANES spectra at the Cu K-edge of PCN/Cu nanosheets. Inset: absorption edge positions of Cu XANES spectra. c Fourier transforms of EXAFS spectra at the Cu K-edge. d Calculated indirect transfer and direct transfer current density of PCN-SHE and PCN/Cu-SHE biohybrids from CV curves. e H2 production of SHE, PCN-SHE, and PCN/Cu-SHE biohybrids. (n  =  3 independent samples; mean ± SD).

Full size image

Discussion

In summary, we have successfully constructed the PCN/Ru-SHE hybrid system to improve the metabolic activity of solar-driven H2 production. Comprehensive experiments and ultrafast spectroscopy demonstrate that incorporating Ru atoms onto PCN-SHE facilitates effective charge separation and reduces electron transfer impedance at the biohybrid interfaces. Operando single-cell photocurrent measurements and CV analysis reveal that the direct electron uptake of PCN/Ru-SHE biohybrid exhibits an 11.0-time increase than that of PCN-SHE, highlighting that the unique Ru-N4 structure at the biohybrid interfaces primarily enhances direct electron transfer, further investigated by theoretical calculations. As a result, the superior electron transfer at the PCN/Ru-SHE interfaces displays an obvious improvement of H2 evolution compared to both SHE and PCN-SHE. Additionally, proteomic analysis supports that the single atoms at the interfaces could facilitate electron transfer and modulate biological metabolism. Further confirmed via Cu atoms across biotic-abiotic interfaces, we discover that the single-atom bridges primarily facilitate direct electron uptake while little contribute is flavin-mediated indirect transfer. This thoughtful design unlocks charge transfer dynamics at the biotic-abiotic interfaces and sheds light on the optimization of the biohybrid systems from the perspective of atom-level for enhanced solar-to-chemical conversion. We can also build single-atom bridges in different biological-material hybrid systems to promote electron uptake for enhanced bio-photosynthesis and bio-electrosynthesis. We hope that our multidisciplinary approach and understanding will provide valuable insights into the future development of microbe-materials biohybrid systems.

Methods

Synthesis of bulk g-C3N4 nanosheets (denoted as BCN)

The BCN was prepared by annealing melamine at 550 °C for 2 h with a heating rate of 3 °C min−1.

Synthesis of porous g-C3N4 nanosheets (denoted as PCN)

The PCN nanosheets were obtained by twice calcination of as-prepared bulk g-C3N4 at 500 °C for 2 h with a heating rate of 2 °C min−1. Then, the powder was dispersed in 100 ml of deionized water and ultrasound for 2 h, and further dried at 60 °C to obtain PCN nanosheets.

Synthesis of porous g-C3N4/Ru single atom nanosheets (denoted as PCN/Ru)

The 92 mg PCN nanosheets were added into 10 mL methanol solution. Then, 13.1 mg RuCl3·3H2O was added into PCN methanol solution. Finally, the solution was continued to be rotary evaporated and further calcined at 160 °C for 1 h in an Ar atmosphere.

Synthesis of porous g-C3N4/Ru nanoparticles (denoted as PCN/Ru nanoparticles)

The 92 mg PCN nanosheets were added into 10 mL methanol solution. Then, 13.1 mg RuCl3·3H2O was added into PCN methanol solution. Finally, the 5 mg NaBH4 was added for 10 min to synthesize PCN/Ru nanoparticles.

Characterization

The crystal phases of all samples were characterized by X-ray diffraction (XRD, Ultima III, Rigaku Corp.) with a Cu Kα radiation source (λ = 1.5406 Å). The absorption of samples was performed on an UV-visible-near-infrared (UV-Vis-NIR) spectrophotometer (UV-1700 spectrometer, Shimadzu). The steady-state and transient photoluminescence (PL) spectra were characterized by Edinburgh Instruments FLS 1000. The X-ray photoelectron spectra (XPS) were recorded with the ARL 9800. Fourier transform infrared spectra (FTIR) were performed with a Jasco FTIR-4000. The transmission electron microscopy (TEM) was recorded using Tecnai G2 F20. Field emission scanning electron microscopy (FE-SEM) was conducted via JEOL JSM6700F. High-angle annular dark-field scanning transmission electron microscope (HAADF-STEM) was performed via JEM-ARM200F at 200 kV. The atomic force microscopy (AFM) was carried out on a Bruker Dimension Icon atomic force microscope. The evolved gas was determined using a gas chromatograph with Argon carrier (GC2014, Shimadzu). The active species were analyzed using JEOL (FA200) electron spin resonance (ESR) Spectrometer. X-ray absorption fine structure (XAFS) spectra of Ru and Cu K-edge were performed at Shanghai Synchrotron Radiation Facility (SSRF). Zeta potential was tested using Malvern Zetasizer Nano ZS90 (ZEN3600).

Solar-driven H2 production test

100 μL of the SHE liquid cultures were transferred to a 30 mL of LB medium and incubated (30 °C, 200 rpm) until the OD600 of the culture increased to 1.0. The SHE cells were obtained by centrifugation and re-suspended into 30 mL hydrogen-production medium (M9:LB = 95:5) with the addition of 20 mM of sodium lactate and 10 mM of cysteine. The H2 measurements were performed within a 50 mL sealed Pyrex bottle filled with 5 mg of PCN/Ru nanosheets. The whole bottle of bacteria and catalysts in the culture media was deoxidized by purging N2 for 1 h. Finally, the bottle was strictly sealed and anaerobically incubated for hydrogenase expression. A 300 W Xe lamp (Perfectlight PLS-SXE300+) with the light intensity of 30 mW·cm−2 was equipped to simulate the solar light illumination for H2 production. The gaseous product was detected by an online gas chromatograph (GC 2014, Shimadzu) equipped with a thermal conductivity detector (TCD) to quantify the produced H2 during the reaction. We also used 450 nm excitation wavelength with the light intensity of 10 mW·cm−2 to measure the quantum yield of H2 production at 450 nm.

Single-cell/single-particle photocurrent test

Semiconductor nanoparticles (PCN and PCN/Ru) suspended in water were spin-coated onto an ITO electrode and annealed at 160 °C for 1 h, and further assembled into the photoelectrochemical microfluidic cell using double-sided tapes and a cell-deposited agarose pad sandwiched between a microscope coverslip and the ITO electrode. Because the PCN and PCN/Ru nanoparticles were sparsely dispersed on ITO, we were able to identify single particles with and without SHE cells (OD600 = 0.5) in contact, and even single particles with a single SHE cell attached to them. The sizes of those particles and SHE cells range from 1 to 10 μm, much larger than the size of a focused 405 nm laser (power density: 50 mW cm−2). Hence, we could direct the focused laser onto different single-cell/sub-cell or single-particle/sub-particle locations.

Additional details regarding the materials, methods, and DFT calculations can be found in the Supplementary Methods in the Supplementary Information.

Statistics and reproducibility

All statistical analyses were carried out by using Origin 2021 and GraphPad Prism 10.0.0. All data are expressed as the mean ± standard deviation (SD). All the representative experiments of micrographs are repeated with three times.

Reporting summary

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