Designing 2D carbon dot nanoreactors for alcohol oxidation coupled with hydrogen evolution

Abstract

The coupled green energy and chemical production by photocatalysis represents a promising sustainable pathway, which poses great challenges for the multifunction integration of catalytic systems. Here we show a promising green photocatalyst design using Cu-ZnIn₂S₄ nanosheets and carbon dots as building units, which enables the integration of reaction, mass transfer, and separation functions in the nano-space, mimicking a nanoreactor. This function integration results in great activity promotion for benzyl alcohol oxidation coupled H₂ production, with H₂/benzaldehyde production rates of 45.95/46.47 mmol g⁻¹ h⁻¹, 36.87 and 36.73 times to pure ZnIn₂S₄, respectively, owning to the enhanced charge accumulation and mass transfer according to in-situ spectroscopies and computational simulations of the built-in electrical field. Near-unity selectivity of benzaldehyde is achieved via the effective separation enabled by the Cu(II)-mediated conformation flipping of the intermediates and subsequent π-π conjugation. This work demonstrates an inspiring proof-of-concept nanoreactor design of photocatalysts for coupled sustainable systems.

Cu-doped ZnIn₂S₄ nanosheets decorated with carbon dots are reported for the photocatalytic oxidation of organic alcohols coupled with H₂ evolution with enhanced yields due to efficient charge accumulation and mass transfer at the catalyst surface.

Introduction

The development in modern catalysis needs to address the prominent issues in energy and chemical coproduction, cleanness and sustainability, as well as the complexity of catalytic research itself^1–3. Photo- and electrocatalysis have the remarkable environment-benign and green nature, which are regarded as powerful tools towards a more sustainable future^4–6. Meanwhile, the design of new catalysts is expected to achieve multifunction simultaneously for integrating the reactant introduction, reaction, product concentrating and separation in one system^7,8. In photocatalysis, a core concept of the integrating is to transplant the knowledge of chemical reaction, mass transfer and product separation to the comprehensive design of new photocatalysts⁹. Along this line, it needs to address two aspects specifically: one is the strategy consulting from electrocatalysis to photocatalysis; the other one is the realization of efficient mass transfer and product separation in the function-oriented delicate nanostructures beyond light harvesting and charge separation¹⁰. For the first aspect, it has been widely debated that photocatalysis is equal to the electrocatalysis initiated and driven by the photo-induced potential shift on the photocatalyst upon irradiation in certain circumstances, which consists of a series of complex photophysical and electrocatalytic processes and requires special interest on the charge and mass transfer in an electric field^11,12. For the second aspect, in terms of chemical engineering, the mass transfer in nanoscale space is quite different from that in the bulk, which has been one of the most inspiring topics in nanoscience and is expected to play a prominent role in catalysis research¹³. Usually, a strong driving force in the nanoscale is needed to achieve the fast diffusion and efficient separation of the reactants and products¹⁴.

Thanks to the rapid development of nanoscience and nanotechnology, tremendous complex and delicate semiconductor nanostructures are available for the design of the highly demanding photocatalysts^15,16. Semiconductor heterostructures and composites with specified photoelectrochemical properties have shown the potential on providing strong driving force under light irradiation for efficient charge separation and subsequent conversion^17,18, such as p-n junctions¹⁹, Schottky junctions and semiconductors/carbon heterostructures^20,21. Carbon dots (C-Dots), as an emerging multifunctional carbon-based species, have received much attention in photo/electrocatalysis, especially the multifunction integration^22,23. In terms of photoelectrochemical functions, C-Dots have the characteristic properties of charge transfer and storage owning to the conjugated sp² domain, which provides great opportunity to improve the extraction and distribution of the photogenerated charge carriers^24,25. In terms of chemical engineering, C-Dots often play twofold roles in catalyst design: the construction of photoreduction active centers via the surface groups and the potential driving forces for mass transfer^26,27. Structurally, the abundant surface groups (often negatively charged) and the conjugated carbon core of C-Dots can provide strong intermolecular forces together, which are conducive to the mass transfer and separation of the reactants and products²⁸. By now, the semiconductor nanostructures, C-Dots, and their rich photoelectrochemical properties are ready for integrating the functions of reaction, mass transfer, and product separation²⁹. And thus, it is time to demonstrate the great potential of this kind of catalyst design strategies with multifunction integration in more demanding and valuable reactions.

Here, we use Cu-doped ZnIn₂S₄ (Cu-ZIS, Unit-1, light harvesting and reaction) nanosheets and carbon dots (C-Dots, Unit-2, charge and mass transfer) as the basic building units to achieve this integrated photocatalyst design, e.g. the composite nanocatalyst itself has multifunction of reaction, mass transfer, and separation, making it a nanoreactor (Fig. 1). More importantly, the mass transfer of protons was prompted by the strong internal electrical field between the adjacent Cu(II)/C-Dots, while the timely product separation was achieved by the Cu(II)-mediated conformation flipping of the adsorbed substrates towards C-Dots within the nanoscale catalytic space. As a result, the photocatalyst with the integrated multifunction brings large catalytic performance promotion, mainly reflected in an activity enhancement on benzyl alcohol (BA) oxidation for coupled H₂/benzaldehyde (BAD) production (45.95/46.47 mmol g⁻¹ h⁻¹), which are 36.87 and 36.73 times to that of pure ZIS, respectively. A series of experimental and theoretical results show that the high activity is mainly due to the dramatically enhanced transfer of the electrons and protons within the strong electric field of the neighboring Cu(II)/C-Dots sites. At the same time, the conformation transformation of the adsorbed intermediates on Cu(II) (an almost 90°-degree flip) and the π-π conjugation with C-dots jointly determine the efficient product separation and high selectivity by avoiding excessive oxidation. This work clearly demonstrates an inspiring linkage between renewable energy and green chemical engineering driven by photocatalysis, as well as a proof-of-concept route for the integrating design of multi-unit photocatalysts in terms of reaction, mass transfer, and separation.

Fig. 1. Reaction, mass transfer and product separation of Cu-ZIS/C-Dots.

Conceptual process diagram for the coupled production of H₂/BAD from water and BA with a multifunction integration of reaction, mass transfer and separation, mimicking a nanoreactor.

Results and discussions

Photocatalyst characterizations

The size and morphology of the composites were characterized by transmission electron microscopy (TEM) and high-resolution TEM (HRTEM). The TEM image of Cu-ZIS nanosheets reveals an ultrathin 2D layered structure with a lateral dimension of 400–500 nm (Supplementary Fig. 1a), while C-Dots show a quasi-spherical morphology with a size of 2.6 ± 0.5 nm (Supplementary Fig. 1b). After combination, the Cu-ZIS/C-Dots composite shows a 0D/2D lamellar structure with embedded small dots, as shown in Fig. 2a. The HRTEM image of Cu-ZIS/C-Dots (Fig. 2b) further shows two lattice spacings of 0.321 nm attributed to the ZIS (102) plane³⁰, and 0.210 nm attributed to the graphite (100) plane in C-Dots²⁷. Supplementary Fig. 1c–i show the elemental mapping images of Cu-ZIS/C-Dots, where the Cu, S, Ni, In, Zn and C elements can be observed with uniform distribution. In addition, the TEM and HRTEM images of 0D quantum dots (QDs) and 1D nanowires of Cu-ZIS are also provided in Supplementary Fig. 2 for comparison. And the SEM images further demonstrate the successful preparation of the Cu-ZIS nanowires (Supplementary Fig. 3). The structure of all samples was further revealed by X-ray diffraction (XRD) and Raman spectroscopy. In Supplementary Fig. 4a, diffraction peaks at 21.6°, 27.7°, 47.2° and 55.6° of Cu-ZIS/C-Dots match well with (006), (102), (110) and (202) planes of chalcopyrite phase ZnIn₂S₄ (JCPDS No. 72–0773)³⁰, respectively, similar to that of pure ZIS and Cu-ZIS nanosheets. Supplementary Fig. 4b shows the four Raman vibrational modes located at 250, 301, 346, and 370 cm⁻¹, attributed to the longitudinal optical mode (LO₁), transverse optical mode (TO₂), longitudinal optical mode (LO₂), and A_1g mode of crystalline ZIS³¹, respectively. Nevertheless, the peak corresponding to LO₁ in Cu-ZIS shifts to a lower frequency, and all peaks exhibit a relatively weaker intensity, suggesting a reduction in crystalline symmetry compared to pristine ZIS. In the case of Cu-ZIS/C-Dots, these four peaks are further diminished and barely observable, indicating a more pronounced distortion in its structure³². Two peaks near 1305 and 1570 cm⁻¹ are observed that come from the D band of sp³ carbon and the G band in-plane vibration of sp² carbon, respectively, proving the existence of C-Dots in the composite³³. In addition, the XRD of Cu-ZIS QDs and Cu-ZIS nanowires are also given in Supplementary Fig. 5.

Fig. 2. Structure characterizations of Cu-ZIS/C-Dots.

a, b TEM and HRTEM images of Cu-ZIS/C-Dots. c Calculated chemical potentials of surface energy of different Cu-interstitial sites in Cu-ZIS. d EPR spectra of ZIS and Cu-ZIS. e Optimized structure of Cu-ZIS (The yellow, gray, pink, and blue spheres represent the S, Zn, In, and Cu atoms, respectively). f Schematic structure of the neighboring Cu(II)/C-Dots sites in the catalyst. g UV−vis absorption spectra of ZIS, Cu-ZIS, C-Dots and Cu-ZIS/C-Dots. h Schematic band alignment of Cu-ZIS/C-Dots.

The C-Dots (Unit-2) used here were actually prepared by photo-induced deposition of Ni(II) on the original C-Dots to better perform the function of reduction center. The X-ray absorption near-edge spectroscopy (XANES) was carried out to study the electronic structure of C, N and O in the Ni(II)-modified C-Dots. The original C-Dots prepared from citric acid and ethylenediamine were also tested for comparison. The C K-edge spectrum of Ni(II)-C-Dots (Supplementary Fig. 6a) exhibits three characteristic peaks of π* C–C (286.8 eV), π* C–N–C (289.8 eV), and σ* C–C (293.8 eV), which all shift to the low energy direction compared to those of the original C-Dots^34,35. The N K-edge spectra of both Ni(II)-doped and original C-Dots (Supplementary Fig. 6b) show two peaks of 1 s  →  π* at C = N (401.1 eV) and 1 s  →  σ* in the hexagonal-graphitic units (408.5 eV)^36–38. The above left shifts of the peaks in Ni(II)-C-Dots suggest a strong electrostatic interaction after the complexation of C-Dots and metal ions, which indicates the successful incorporation of Ni(II)³⁹. The O K-edge spectra of both C-Dots (with Ni(II)) and original C-Dots (Supplementary Fig. 6c) exhibit a peak at 539.4 eV attributed to 1 s  →  σ* of C=O/–COOH, proving the abundant surface groups in C-Dots^40,41.

The valence states of the photocatalysts were studied by X-ray photoelectron spectroscopy (XPS, Supplementary Figs. 7 and 8), which suggested the existence of Cu(II), In(III), Zn(II), S²⁻ and Ni(II)^42,43. Besides the Ni(II) signals in Ni 2p spectra, the peak for Ni-S was observed with a binding energy around 852.0 eV (Supplementary Fig. 8e)⁴⁴, indicating the close interaction of Ni(II) in C-Dots and Cu-ZIS. Compared with ZIS and Cu-ZIS, the binding energies of the corresponding Cu 2p, In 3d, Zn 2p and S 2p peaks in Cu-ZIS/C-Dots slightly shifted to the lower binding energy (Supplementary Fig. 8a–d). In addition, it worths nothing that In 3d and Zn 2p show a similar decrease of the binding energy. On the other hand, compared with C-Dots, the binding energies of Ni 2p and N 1s in Cu-ZIS/C-Dots slightly shifted to the higher binding energy (Supplementary Fig. 8e, f). Over all, these observations suggest a partial electron transfer from C-Dots to Cu-ZIS, revealing a close interfacial contact in favor of photogenerated charge transfer. The XPS of Cu-ZIS QDs and Cu-ZIS nanowires are also given in Supplementary Figs. 9 and 10.

For the position of Cu(II) sites, we employed DFT calculation to assess the Cu(II) energies at six different sites as depicted in Fig. 2c and Supplementary Fig. 11, including replaced Zn site, replaced S site on Zn-S plane, S hollow site on Zn-S plane, replaced In site, replaced S site on In-S plane, and S hollow site on In-S plane. In the first case, when the surface chemical potential energy of S is from −7.1 to −6.6 eV, the Cu(II) replaced S site on Zn-S plane exhibits the lowest surface energy, which aligns with the green region, representing sulfur deficient state. Secondly, the Cu(II) occupied S hollow site on Zn-S plane shows the lowest surface energy, which aligns with the red region, representing the suitable stoichiometric state when the surface chemical potential energy of S is from −6.6 to −4.1 eV. Finally, the copper replaced indium site on Zn-S plane displays the lowest surface energy, which aligns with the blue region, representing the S rich state when the surface chemical potential energy of S is from −4.1 to −2.0 eV. Electron paramagnetic resonance (EPR) measurement further proves the existence of S vacancy (V_s). In Fig. 2d, no signal was detected in ZIS, whereas Cu-ZIS shows a strong EPR response of 2.003 g at 3515 G³⁰, accounting for V_s. In addition, as displayed in Supplementary Fig. 12, the signal of V_s becomes higher with increasing Cu doping and the 5% Cu−ZIS exhibits the largest V_s concentration. The experimental results show the S deficient state in the Cu-ZIS/C-Dots system, demonstrating the stable site as the Cu(II) replaced S site on Zn-S plane, which indicates a propensity for Cu(II) to occupy the original V_s site in the substantial chemical potential region of Cu-ZIS. The observation of EPR signals of V_s in Cu(II)-doped ZIS, despite the theoretical propensity of Cu(II) to occupy such vacancies, suggests a complex defect chemistry, potentially involving additional vacancy formation or incomplete vacancy filling, influenced by thermodynamics and lattice dynamics^30,32. According to the Boltzmann distribution, there is a competing relationship between the three positions that Cu (II) may occupy, S site on Zn-S plane, S hollow site on Zn-S plane and Cu replaced S site on In-S plane, which are very close. Based on the formation energy of S and DFT-optimized structure, the simulation result is in good agreement with the experimental EPR data, demonstrating that the Cu(II) dopant prefers to chemisorb on the V_s site of Zn-S in the ZIS basal plane (Fig. 2e).

DFT calculations were further carried out to dive fundamental insight into the effect of Cu(II) doping. The charge density difference isosurface images of Cu-ZIS (Supplementary Fig. 13) reveal a strong charge redistribution in the Cu-bonding region after the incorporation of Cu(II) in the basal plane of ZIS, where Cu(II) further occupies the original V_s site³⁰. This charge redistribution confirms the strong interaction between Cu(II) and ZIS owning to the lattice reconstruction in Cu-ZIS, which facilitates strong Lewis acid-base interactions with electron-rich sites in the ZIS basal plane and preferred coordination environments, with optimal ionic sizes and charge compatibility within the crystal lattice. In addition, the adsorption energy of C-Dots and Cu(II) site were estimated through DFT calculations in Supplementary Fig. 14^45–47. The results show that the adsorption energy of C-Dot near Cu(II) site was much higher with a shorter distance (5.7 Å), compared to that with a larger distance (8.9 Å), indicating that the adjacent C-Dot and Cu(II) site are thermodynamically more favorable. By now, we can see that the Cu-ZIS/C-Dots catalyst tends to have the neighboring C-Dots and Cu(II) sites on the 2D Cu-ZIS matrix and a schematic illustration is shown in Fig. 2f.

The UV-vis absorption spectroscopy was used to explore the light-harvesting capability and calculate the band gaps of the photocatalysts. In Fig. 2g, the ZIS nanosheets display an absorption edge at approximately 510 nm. After complexation, the absorption capability of Cu-ZIS/C-Dots is dramatically enhanced in the long-wavelength range compared to individual Cu-ZIS²⁴. According to the absorption spectra, we can obtain the Tauc plots (Supplementary Fig. 15a)⁴⁸, from which the band gaps of ZIS and Cu-ZIS are estimated to be 2.45 and 1.83 eV, respectively. Next, the valence band (VB) position of ZIS and Cu-ZIS were measured to be 1.46 and 1.32 eV in Supplementary Fig. 15b. Supplementary Fig. 16 further shows the band gap narrowing of ZIS nanosheets gradually upon Cu introduction^30,32. In addition, the cyclic voltammetry (CV) curve (Supplementary Fig. 17) was used to estimate the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) levels of C-Dots with ferrocene redox system as the standard, for which the detailed calculation is shown in the Supplementary Information^49,50. Two pairs of redox peaks corresponding to ferrocene and the catalyst can be seen clearly from the CV curve. The HOMO and LUMO energy levels of C-Dots are calculated to be 1.23 and −0.61 V vs. NHE, respectively. The steady-state photoluminescence (PL) spectra of Cu-ZIS/C-Dots (Supplementary Fig. 18) reveal a prominent PL peak at 450 nm, which is significantly quenched by 82.60% upon the introduction of C-Dots, indicating the enhanced charge separation in the Cu-ZIS/C-Dots system. In the time resolved PL curves in Supplementary Fig. 19, Cu-ZIS/C-Dots exhibits a PL lifetime of 4.67 ns, considerably shorter than that of ZIS (17.21 ns) and Cu-ZIS (13.63 ns), which highlights the important synergistic effect of Cu(II) sites and C-Dots for improved charge separation.

Work function (ϕ) is another powerful parameter to indicate the electron transfer between two components (Supplementary Fig. 20)⁵¹. The work function of C-Dots was obtained from ultraviolet photoelectron spectroscopy (UPS) spectra in Supplementary Fig. 21^52,53. Therefore, as illustrated in Supplementary Fig. 22, the electrons in Cu-ZIS can spontaneously migrate toward C-Dots at the interface. Based on the above data, the band alignment of the Cu-ZIS/C-Dots composite is given in Fig. 2h, showing suitable energy levels for H₂ production and potentially efficient electron and hole transfer.

Photocatalytic performance

Here, BA oxidation was employed to evaluate the catalytic performance of the Cu-ZIS/C-Dots system for H₂/BAD coproduction. As shown in Figs. 3a, b, the photo-generated electrons and holes by Cu-ZIS/C-Dots can reduce proton to H₂ and oxidize BA to BAD, respectively, where the H₂ and BAD evolution on all catalysts show continuous increase with time. The coupled H₂ and BAD production rates over Cu-ZIS/C-Dots reach 45.95 and 46.47 mmol g⁻¹ h⁻¹, which are 36.87 and 36.73 times higher than that of pure ZIS (1.24/1.26 mmol g⁻¹ h⁻¹) (Fig. 3c), respectively. Control experiments indicate that C-Dots itself only produces very little hydrogen in all plots. Moreover, Cu-ZIS shows 97.5% selectivity of BA oxidation to BAD, significantly superior than that of ZIS (82.4%), indicating that the introduction of Cu(II) can significantly improve the selectivity along high H₂/BAD yields. And the near-unity selectivity is achieved with Cu-ZIS/C-Dots, which may be attributed to the Cu(II)-mediated substrate adsorption and product separation facilitated by π-π conjugation with C-Dots to prevent excessive oxidation. The apparent quantum efficiencies (AQE) of H₂ over Cu-ZIS/C-Dots reached 39.99%, 26.62%, 18.13%, 8.02%, 4.21% and 0.61% at 420, 450, 485, 520, 595 and 630 nm, respectively (Fig. 3d and Supplementary Tables 2 and 3).

Fig. 3. Catalytic performances of Cu-ZIS/C-Dots.

a, b Time-dependent H₂ and BAD production profile. c Comparison of photocatalytic H₂/BAD production rates of C-Dots, ZIS, Cu-ZIS and Cu-ZIS/C-Dots. d Wavelength-dependent AQE of H₂ production over Cu-ZIS/C-Dots in comparison with the absorption spectrum. e Cycle stability of H₂/BAD production over ZIS and Cu-ZIS/C-Dots. f H₂ production rates for Cu-ZIS/C-Dots in this work compared with recently reported representative photocatalysts.

Moreover, the Cu-ZIS/C-Dots composite show good cycle stability for both H₂ and BAD production with no significant decrease in four photocatalytic cycles (Fig. 3e). The XRD patterns and TEM images before and after reaction indicate no obvious change or aggregation of Cu-ZIS/C-Dots, indicating its structure robustness (Supplementary Figs. 24 and 25). It worths noting that the H₂ production rate of Cu-ZIS/C-Dots is far beyond the most of the state-of-the-art results over the recently reported photocatalysts^54–63 (see the comparisons in Fig. 3f and Supplementary Table 4) on BA oxidation-coupled H₂ production with a high selectivity of BAD. In addition, as C-Dot is a typical photothermal catalyst⁶⁴, we also conducted Infrared thermography tests on Cu-ZIS/C-Dots at different temperatures using the temperature control function to investigate the impact of the photothermal effect through an all glass automatic on-line trace gas analysis system. Supplementary Fig. 26 shows the Infrared thermal imaging distribution of Cu-ZIS/C-Dots under visible-light irradiation at 5–20 °C, indicating that the heat distribution of the solution is uniform with a negligible temperature variation of ± 1.0 °C with the cooling system of the photoreactor. There is no obvious heat accumulation on the catalyst and no photothermal effect on the reaction.

In addition, for better comparison, the photocatalytic performance of the catalysts was also estimated with the existence of L-ascorbic acid (AA), a commonly used hole sacrificial reagent for QDs, showing a more advanced H₂ production rate of 98.84 mmol g⁻¹ h⁻¹, among the highest activity of metal chalcogenides in the literature (Supplementary Table 5). The related AQE, activity and stability results are also provided in Supplementary Figs. 27–32. It is worth noting that the ZIS and Cu-ZIS nanosheets actually show lower hydrogen production rate using AA compared to those using BA, which may be attributed the more severe reverse hole transfer process⁶⁵ due to its electron-withdrawing groups, along with the rare hydrogen evolution reaction (HER) site in ZIS and Cu-ZIS. Further introduction of C-Dots with HER sites results in the dramatic activity promotion to 98.84 mmol g⁻¹ h⁻¹. In this system, C-Dots can effectively transfer the photo-generated electrons to its own HER sites, which is essential for charge separation and stabilization in this efficient sacrificial reagent system.

Mechanism insights of charge extraction and active species

Efficient charge separation plays an import role in photocatalysis. Here, the intrinsic polarized internal electric field (IEF) in ZIS and Cu-ZIS is revealed through both theoretical calculations and experiments. As shown in Fig. 4a, the asymmetrical [S–In]–[S–In–S]–[Zn–S] unit cell in ZIS can generate a bulk-IEF (B-IEF) along the Z-axis of [001] direction according to the calculated electrostatic potential and charge density distribution. The imbalanced charge distribution in [Zn–S], [S–In–S], and [In–S] layers results in different electrostatic potentials, and subsequently a vertical B-IEF, which can drive efficient inner charge separation within the nanosheets. For a semi-quantitative analysis, an obvious increase of the local electrostatic potential difference (∆E) was observed between [Zn–S] and [In–S] slices after Cu-doping (1.58 eV), indicating a stronger B-IEF in Cu-ZIS than that of ZIS (1.40 eV). This enhanced B-IEF was also proved by the larger dipole moment of 10.85 Debye in Cu-ZIS nanosheets than that of ZIS (0.84 Debye) (Supplementary Fig. 33), confirming the advantage of the Cu(II)-induced structure variation on charge separation.

Fig. 4. Charge extraction and active species mechanism.

a DFT calculated electrostatic potential of the B-IEF for ZIS and Cu-ZIS. b Charge attenuation constants (τ) from TPV spectra of ZIS (τ₁), Cu-ZIS (τ₂) and Cu-ZIS/C-Dots (τ₅). c The in-situ TPV spectra of Cu-ZIS/C-Dots with the introduction of ACN and ACN/AA (both containing 5‰ water) for estimating hole extraction efficiency. d The in-situ EPR spectra of photogenerated holes by TEMPO spin-label with different additives over Cu-ZIS/C-Dots. e EPR detection of the formed DMPO-C_α radicals from BA splitting under light irradiation over Cu-ZIS/C-Dots. All the EPR spectra were recorded after 10 min of irradiation.

The B-IEF was further demonstrated via the surface potential and surface charge density of ZIS and Cu-ZIS tested by atomic force microscopy (AFM) under the Kelvin probe force mode (KPFM) (Supplementary Figs. 34, 35), which indicate a largely increased surface potential in Cu-ZIS (17.20 mV) compared to ZIS (4.28 mV)⁶². In addition, AFM images and the corresponding height profiles of ZIS and Cu-ZIS are provided in Supplementary Fig. 36. According to the following equation:

$$E={\left(\frac{-2V_{sp}\rho }{\epsilon {\epsilon }_0}\right)}^{1/2}$$ 1

E is the B-IEF value, $V_{sp}$ is the surface potential, ρ is the surface charge density, ε is the low-frequency dielectric constant, and ε₀ is the vacuum dielectric constant. ε and ε₀ are regarded as two constants for similar crystal lattices here. And ρ is determined according to the Gouy–Chapman equation⁶⁶:

$$\rho =\sqrt{8kT\epsilon {\epsilon }_{0}n}\sin h\left(\frac{z{e}_0{\phi }_0}{2kT}\right)$$ 2

whereby k is the Boltzmann constant, T is absolute temperature, n is the number of electrolytes per unit volume, z is electrolyte valence, e₀ is the electron charge and ${\phi }_0$ is surface potential in concordance to Eq. (3):

$${\phi }_0=\zeta \left(1+\frac{D}{a_1}\right)e^{KD}$$ 3

in which ζ is the zeta potential, a₁ is the particle Stokes radius, $K$ is the Debye length and D is the distance from the sliding layer to the particle surface. Hence, the surface charge densities for ZIS and Cu-ZIS were calculated to be 18.10 and 23.56 mC m², respectively. Using the zeta potentials (ζ_ZIS = −22.0 mV, and ζ_Cu-ZIS = −28.3 mV), the B-IEF intensity (E) can be approximately estimated as a function of surface potential $V_{sp}$ and Zeta potential ζ following the equation:

$$E={\left[A{V}_{sp}\sin h\left(\frac{z{e}_0\zeta \left(1+\frac{D}{a_1}\right)e^{KD}}{2kT}\right)\right]}^{1/2}$$ 4

Then the relative B-IEF magnitude was obtained experimentally, which is increased 2.28 times in Cu-ZIS than that of ZIS after normalization (Supplementary Fig. 37). Overall, the above theoretical and experimental results unveil the polarized B-IEF feature in ZIS and Cu-ZIS, which plays an important role on photocatalyst development.

Transient photovoltage (TPV) spectroscopy was employed to investigate the Cu(II) and C-Dots-promoted charge transfer and subsequent surface reactions (Supplementary Figs. 38, 39)^49,67,68. TPV spectra indicate a slightly prolonged charge extraction time (t_max) after Cu-doping (0.091 ms) and C-Dots-introduction (0.108 ms) compared to pristine ZIS (0.069 ms) owning to the Cu-induced structure deformation and defect states. More importantly, the charge extraction amount (A) was increased more dramatically from ZIS (A₁ = 0.003) to Cu-ZIS (A₂ = 0.017) and Cu-ZIS/C-Dots (A₅ = 0.051). At last, the charge recombination of Cu-ZIS/C-Dots (τ₅, 0.30 ms) was largely restricted compared to Cu-ZIS (τ₂, 0.14 ms) and pure ZIS (τ₁, 012 ms) in Fig. 4b, attributing to the effective improvement of charge storage after the introduction of C-Dots. The number of effective charges (A_eff) remaining on the surface (after the charge extraction and recombination) is a more crucial parameter for catalysis, which can be calculated by

$$A_{eff}=\frac{A\cdot \tau }{t_{\max }}$$ 5

Here, the A_eff was promoted from ZIS (A_eff1 = 0.005) and Cu-ZIS (A_eff2 = 0.026) to Cu-ZIS/C-Dots (A_eff5 = 0.142) owning to the charge storage effect of C-Dots. Other TPV data of comparing samples are also displayed in Supplementary Figs. 40–42⁴⁹. Ultrafast transient absorption spectroscopy (TAS) is another powerful technique for providing direct evidence for charge carrier kinetics inside the photocatalysts^5,6,69,70. To figure out how Cu doping and C-Dots modification bring superior photocatalytic activity to ZIS, TAS was performed on ZIS, Cu-ZIS and Cu-ZIS/C-Dots for the charge separation and transfer dynamics. Supplementary Fig. 43 present the 3D color-map plots of TAS pumped at 320 nm, which show a prominent negative ∆OD (change in optical density) in the range of 350–400 nm, corresponding to the photoinduced bleaching (PIB) signal of ZIS. Based on steady-state absorption results and previous reports⁷¹, such PIB signal originated from state filling of the intrinsic excitonic state dominated by photoelectrons at CB of ZIS rather than holes at VB^72,73. It worths noting that the bleach maxima observed in ZIS often originate from higher energy excitonic transitions, rather than the band-edge excitons⁷², which makes the ultrafast hot electron transfer (HET) from higher energy states play a dominant role in the ZIS-derived systems⁷⁴. Here the introduction of Cu enhances the signal of PIB and the introduction of C-Dots reduces the signal of PIB, which can be attributed to the intrinsic HET process in Cu-ZIS and the strong electron accepting capability of C-Dots^75,76. As displayed in Supplementary Fig. 44, the TAS curves indicate a bi-exponential decay behavior with an average lifetime of 1.61 ps at 380 nm in Cu-ZIS/C-Dots, which is only half to than that of Cu-ZIS nanosheets (3.20 ps), indicating the efficient electron transfer from Cu-ZIS to C-Dots inside the nanoreactor⁷⁷. Here TAS indicates the promoted charge separation within the Cu-ZIS/C-Dot composite in ps, which facilitates the following charge extraction and conversion later. With TPV, we are able to monitor the decay of the effective surface charges, which can be considered as a reactant in photocatalysis with much longer lifetime (up to 0.30 ms). Along this line, the ps kinetics of TAS and ms kinetics of TPV provide an interesting insight into the photophysical charge separation and surface conversion processes, respectively.

To further simulate the photocatalytic environment in part and explore the photocatalytic mechanism, the in-situ TPV was further performed over ZIS, Cu-ZIS and Cu-ZIS/C-Dots with the introduction of AA/ACN vs. pure ACN⁷⁸. Here AA can consume the photogenerated holes and thus result in the change of photovoltage intensity (∆Q) when introducing AA to promote hole transfer. Cu-ZIS/C-Dots (0.14923) shows much larger ∆Q than that of Cu-ZIS (0.09504) and ZIS (0.04559) (Fig. 4c and Supplementary Fig. 45), which indicates that loading C-Dots effectively promotes the transfer of photogenerated electrons and holes over the photocatalyst/solution interface to further improve the photocatalytic activity⁴⁹.

We also directly tested the existence and conversion of the photoinduced holes and electrons during light irradiation in the presence of different hole sacrificial reagents through the in-situ EPR spectroscopy, by using 2,2,6,6-tetramethylpiperidine-1-oxyl/acetonitrile (TEMPO/ACN) and TEMPO/water to trap holes and electrons, respectively⁶². As shown in Fig. 4d, the EPR spectra show a triple characteristic peak of TEMPO itself under dark, which is lowered upon light irradiation due to that TEMPO undergoes oxidation by photogenerated holes to form its oxidized state (TEMPO⁺)⁷⁹. Here, TEMPO captures a hole, resulting in the formation of TEMPO⁺. This process represents the radical cation form of TEMPO, where the unpaired electron is localized on the nitrogen atom. When AA and BA are applied for hole scarifications, the lowest EPR signal was observed. This observation suggests a more efficient hole transfer from the photocatalyst to the reactive molecules of AA and BA. Similarly, the photoinduced electrons were also tested by using TEMPO/water as the trapping agent (Supplementary Fig. 46)⁸⁰. For a more comprehensive mechanism understanding, we further employed the in-situ EPR technique for the detection of BA radicals by utilizing 5,5-dimethyl-1-pyrrolidine N-oxide (DMPO) as a spin trapping reagent. As depicted in Fig. 4e, the values of α_N = 16.2 and α_H = 23.4 correspond to the hydrogen and nitrogen hyperfine splitting for the nitroxide nitrogen in DMPO–BA*. The presence of six distinct peaks signifies the generation of α-BA radicals (•BA) under light irradiation⁸¹, indicating the favorable dehydrogenation of the C_α–H by the surface holes on the Cu-ZIS/C-Dots photocatalyst. Combined with DFT calculation results, when Cu(II) serve as a hole trapping site, it facilitates the transfer of photogenerated holes by the preferential dehydrogenation of the α-C–H bond, a process that is challenging with other catalysts⁸². Therefore, the Cu(II) site as the oxidation unit enables efficient BA oxidation through the rapid trapping and subsequent consumption of holes with the adsorbed substrates.

Supplementary Fig. 47a presents the linear sweep voltammetry (LSV) curves for Cu-ZIS/C-Dots in a 1 M PBS aqueous solution, demonstrating superior electrocatalytic hydrogen evolution activity with an overpotential of 110 mV at 10 mA cm⁻² (η₁₀). This is notably lower than pure ZIS (414 mV) and Cu-ZIS (242 mV) under light irradiation. Tafel plots in Supplementary Fig. 47b reveal a significant decrease in the Tafel slope of Cu-ZIS/ C-Dots to 39 mV dec⁻¹, in contrast to that of pure ZIS (106 mV dec⁻¹) and Cu-ZIS (81 mV dec⁻¹). Next, we calculated |∆G_H*| of H* on Cu-ZIS and Cu-ZIS/C-Dot (by putting H* on C-Dot), respectively. As shown in Supplementary Fig. 48, Cu-ZIS/C-Dots (−0.149 eV) shows a much lower |∆G_H*| compared to that Cu-ZIS of (−0.996 eV), which suggests that the effectively weakened adsorption of H* on the C-Dots with boosted HER activity. Compared with pure ZIS, Cu-ZIS and Cu-ZIS/C-Dots also show much higher photocurrent (Supplementary Fig. 49a), indicating a more effective charge separation⁸³. The electrochemical impedance spectra (EIS) curves in Supplementary Fig. 49b also show the promoted charge transfer in Cu-ZIS/C-Dots with the smallest arc radius.

Photocatalytic mechanism

Proton transfer is another crucial factor in the BA oxidation process, especially on facilitating the intermediate formation and influencing reaction kinetics and selectivity. Thus, profound understanding and control of the proton transfer process are essential for designing effective catalysts for organic oxidation coupled H₂ production. To investigate the synergistic mass transfer process, here we simulated the electric field intensity and dimensionless proton concentration distribution between the neighboring Cu(II) sites and C-Dots in the composite system by COMSOL Multiphysics^84,85. The charge densities of all sites in Cu-ZIS/C-Dots were calculated by DFT (Supplementary Table 7), which were used to estimate the charge densities on Cu(II) and C-Dots (0.28 × 10⁻¹⁹ and −1.46 × 10⁻¹⁹C, respectively) for simulation. Other detailed parameters of COMSOL Multiphysics simulation are given in Supplementary Table 8, including material densities, size and diffusion coefficient of proton. First, we established the positively and negatively charged unit models to simulate the Cu(II) sites and C-Dots on the 2D nanoreactor, respectively. Supplementary Fig. 50 reveals that the highest potential is located around the positively charged Cu(II) sites, with a fast potential decrease towards the negatively charged C-Dots. This giant potential gradient in the nano-space induces a great electric field strength (~10³ V/m, Fig. 5a), which is favorable for subsequent proton transfer between the Cu(II) and the neighboring C-Dots. Figure 5b illustrates a continuous proton concentration gradient from Cu(II) to C-Dots (varying from 1 to 0), which signifies the ongoing proton transfer process from the Cu(II) sites (for BAD adsorption and proton generation) to C-Dots (for H₂ production and BAD separation), emphasizing the critical role of mass transfer in mediating the coupled reactions with high activity and selectivity.

Fig. 5. Mass transfer and product separation.

a, b Distribution of electric field strength and dimensionless proton concentration over the neighboring Cu(II) sites and C-Dots by COMSOL Multiphysics simulation. c DFT-calculated free energy diagram for BA to BAD conversion on the adsorption sites of Cu-ZIS (Insets: the optimized structures of the active intermediates adsorbed on the Cu(II) sites, indicating an accompanying conformation evolution). d Schematic diagram of the flipped conformation evolution process favoring BAD separation. e Time-dependent BA consumption, H₂ generation and BAD production profile.

The above experimental and simulation data show that the adjacent Cu(II)/C-Dots on the 2D Cu-ZIS matrix constitute an efficient nanoreactor that enables reaction, mass transfer and separation. This means that the parallel up and synergy of these nanoreactors could make a big difference on the overall activity. Along this line, we further compared different shaped Cu-ZIS matrixes (0D QDs, 1D nanowires and 2D nanosheets) that can allow paralleling up different amount of nanoreactors, as shown in Supplementary Fig. 51. As a result, the nanoreactors on 2D Cu-ZIS/C-Dots system allows great activity promotion of H₂/BAD production rates (45.95/46.47 mmol g⁻¹ h⁻¹), compared to those composites using 0D Cu-ZIS QDs (0.66/0.72 mmol g⁻¹ h⁻¹) and 1D nanowires (11.33/13.94 mmol g⁻¹ h⁻¹) as displayed in Supplementary Fig. 52a. Moreover, as shown in Supplementary Fig. 52b, the 2D Cu-ZIS/C-Dots system shows much higher BAD selectivity (near unity) than those of the 0D (16.3 %) and 1D (32.7 %) systems. It can be seen that, due to the size and space constraints, the 2D nanosheets show great advantages on low curvature and better synergy of multiple nanoreactors, compared to those on 0D Cu-ZIS QDs and 1D nanowires. The schematic coupling of Cu-ZIS QDs and Cu ZIS nanowires with C-Dots is shown in Supplementary Figs. 53, 54.

Next, to further demonstrate the synergy between Cu(II) and C-Dots, Gibbs free energies of the elementary steps were calculated for investigating the contribution of Cu(II) sites on the ZIS-catalyzed BA oxidation reaction. As illustrated in Fig. 5c, BA is initially adsorbed onto Cu(II) to form a Cu-O bond on the Cu-ZIS nanosheets, with an endothermic energy of −0.34 eV. Subsequently, the O-H bond on RCH₂OH* is cleaved at a large down-hill free energy of −1.07 eV to form RCH₂O* and proton. This process results in the formation of the alkoxide intermediate (RCH₂O*) and simultaneous proton extraction from BA adsorbed on Cu(II). Subsequently, the photoinduced holes in Cu-ZIS preferentially attack the C_α–H in RCH₂O*, which facilitates the formation of the carbon-centered radicals (RCHO*), producing another proton adsorbed on Cu(II)⁶¹. This methyl hydrogen atom was then transferred to adjacent V_s site, encountering an up-hill free energy barrier of 0.16 eV. Finally, the Cu-O bond breaks and leads to the formation of BAD on Cu(II). This last step, as the rate-determining one, underscores the distinctive nature of our 2D nanoreactor system in BA oxidation. The requirement to surmount an energy barrier of 1.40 eV distinguishes this system from most other reports, indicating that the product separation step is crucial in this system rather than the C_α–H cleavage step.

In addition, to further determine how the product separation was assisted by π-π conjugation with C-Dots, in-situ XANES spectra were obtained in the dark and illuminated conditions⁸⁶. As shown in Supplementary Fig. 55a, the C K-edge spectrum of C-Dots exhibits the characteristic peaks of π* C–C (286.7 eV), π* C–N–C (289.7 eV) and σ* C–C (293.7 eV) upon illumination, which shows an enhanced π* characteristic and a weakened σ* characteristic compared to the dark state, indicating that the π* states are modified during photocatalysis³⁴. The N K-edge spectrum of C-Dots (Supplementary Fig. 55b) demonstrates two strong peaks of π* C=N (401.1 eV) and σ* at the hexagonal-graphitic units (408.1 eV), where only the former one is enhanced under light irradiation, indicating that the π-conjugated structure in C-Dots involves the C 2p and N 2p-hybridized orbitals and enables the π electrons to be delocalized across C=N bonds instead of being localized. As a result, these light-promotable π* characteristics of C-Dots render them great capability to undergo π-π conjugation with the π electrons in the aromatic ring of BAD during photocatlaysis^41,87. More importantly, a Cu(II)-mediated conformation flip of the adsorbed intermediates was found in DFT calculation, which actually pushes them from the oxidation unit (Cu(II)) to the reduction unit (C-Dots). Therefore, upon the conversion of BA, the final product BAD is close to C-Dots to facilitate the π-π conjugation, which not only affects the electronic structure of BAD, but also plays a key role in enhancing the product separation. Thus, the synergy between strong π* characteristics of C-Dots and Cu(II)-mediated conformation flip enables the final step of efficient separation of BAD to avoid extra oxidation, which finally fulfils the required three functions of a nanoreactor along with reaction and mass transfer, as summarized in Fig. 5d. In addition, the curves of BA consumption, BAD formation and H₂ generation over time are also shown in Fig. 5e for visualizing the Cu-ZIS/C-Dots coupled nanoreactor system.

Here the integration of these catalytic functions yields improved BA oxidation activity in the Cu-ZIS/C-Dots system. Experimental and theoretical findings consistently highlight the significance of efficient charge accumulation and mass transfer for high H₂/BAD yields, while the near-unity selectivity of BAD is attributed to the substrate conformation evolution and product separation facilitated by C-Dots, preventing its excessive oxidation. The catalytic system design achieves efficient mass transfer through a strong electric field within neighboring Cu(II)/C-Dots sites. It also illustrates a discernible conformation transition of the adsorbed substrate, manifesting as an almost 90°-degree conformation flip to facilitate the π-π conjugation with C-Dots. The Cu(II)-mediated conformation shift assumes a critical role in both the reaction and separation processes in our nanoreactor. First, the gradual substrate turnover process facilitates the controlled release of hydroxyl hydrogen and methyl hydrogen through C_α–H cleavage, followed by subsequent mass transfer of proton to C-Dots. Secondly, this gradual transition signifies a transformation from the oxidation state on Cu(II) to the reduced state on C-Dots. Last, following the conformational flip, the favorable π-π conjugation fosters the synergistic separation of H₂/BAD. Moreover, in the 2D system, the low curvature conferred a substantial advantage of better synergy and parallel arrangement of multiple nanoreactors.

In summary, we exemplified an effective design of photocatalysts as a nanoreactor using Cu-ZIS nanosheets and C-Dots by combining functions of reaction, mass transfer, and separation within the nano-space. Our C-Dots/Cu-ZIS system exhibits improved BA oxidation catalytic activity, with a coupled H₂/BAD production rate of 45.95/46.47 mmol g⁻¹ h⁻¹, surpassing pure ZIS by 36.87 and 36.73 times, respectively. The high yields mainly come from efficient charge accumulation and efficient mass transfer. The near-unity BAD selectivity comes from the conformation evolution of the adsorbed substrate on the Cu(II) center, which is pushed away from the oxidation unit to the reduction unit and facilitates the product separation assisted by π-π conjugation with C-Dots to avoid excessive oxidation. Our work demonstrates an effective compositing strategy, which not only enables highly efficient coproduction of solar fuels and value-added chemicals, but also provides inspiration for the function integration design simulating nanoreactors for other challenging reactions, such as CO₂ reduction and nitrogen fixation.

Methods

Chemicals

All reagents were obtained from Sinopharm Chemical Reagent Co. Ltd without any further purification, including copper(II) acetate monohydrate (Cu(OAc)₂ ∙ H₂O), zinc acetate (Zn(OAc)₂ ∙ 2H₂O), zinc(II) sulfate heptahydrate (ZnSO₄ ∙ 7H₂O), indium(III) chloride tetrahydrate (InCl₃ ∙ 4H₂O), nickel(II) chloride hexahydrate (NiCl₂ ∙ 6H₂O), thioacetamide (TAA), N,N-dimethylformamide (DMF), ethylene glycol (EG), hexadecyl trimethyl ammonium bromide (CTAB), L-ascorbic acid (AA), ethanol, citric acid, ethylenediamine, and benzyl alcohol (BA). The dialysis membrane (MWCO: 200 Da) was purchased from Shenzhen Bymerson Scientific Instrument Co., Ltd.

Synthesis of Cu-ZIS nanosheets

The Cu-ZIS nanosheets were synthetized referring to previous work with modifications³⁰. Zn(Ac)₂·2H₂O (0.219 g, 1 mmol), InCl₃·4H₂O (0.586 g, 2 mmol), TAA (0.301 g, 4 mmol) and different amount of Cu(OAc)₂ ∙ H₂O were added into a mixed solvent of DMF and EG (30 mL, vol % = 1:1) with stirring for 4 h. The surfactant CTAB (0.650 g, 1.8 mmol) was also added slowly under stirring. Then the orange mixture was transferred to a 50 mL Teflon-lined autoclave and maintained at 180 °C for 12 h. After cooling down, the obtained orange precipitates were washed with deionized water three times. After freeze-drying for 12 h, the Cu-ZIS nanosheets with different amounts of Cu/(Cu + Zn) (x% Cu−ZIS, x = 1, 3, 5, 7 and 9) were obtained. Pure ZIS nanosheets were also synthesized following the same procedure but without Cu-doping.

Synthesis of C-Dots

The original C-Dots were synthesized from hydrothermal treatment of citric acid and ethylenediamine according to previous researches with modifications^26,53. Citric acid (1.05 g, 5.5 mmol) and ethylenediamine (335 µL, 5.0 mmol) were dissolved in deionized water (10 mL), which was then transferred to a 20 mL Teflon-lined autoclave and heated to 180 °C and kept for 5 h. It is worth noting that C-Dots here refers to Ni(II)-doped carbon dots, to differ from the C-Dots. For the preparation of original C-Dots used in this work, 50 µL of Ni(II) solution (25 mM) was added to the pre-synthesized original C-Dots (10.0 mg) dispersion (5.0 mL) under 300 W Xenon lamp irradiation and continuous stirring for 60 min to enable the in-situ attachment of Ni(II) onto C-Dots. The obtained suspension was then filtered through a 0.22 μM filter membrane to remove large particles and dialyzed (MWCO: 200 Da, for 1–3 days) to remove free Ni(II). Finally, the dialysate was freeze-dried under vacuum over night to obtain the powder product of C-Dots.

Synthesis of Cu-ZIS/C-Dots

For modification of Cu-ZIS with C-Dots, 20 mg of Cu-ZIS nanosheets and different amounts of C-Dots dispersion (4 mg mL⁻¹) were dispersed in an aqueous solution containing 45 mL of H₂O. Subsequently, the suspension was bubbled with argon for 30 min to completely remove the dissolved oxygen and ensure the anaerobic condition. The dispersion was kept stirring under visible light irradiation (λ > 420 nm) for 60 min to ensure the combination of C-Dots and Cu-ZIS nanosheets through the photo-induced coulomb force. The samples were centrifuged, washed with deionized water twice and then freeze-dried overnight. The C-Dots content relative to Cu-ZIS was adjusted at 1, 3, 5, 7, and 9 wt%.

Characterizations and measurements

TPV was measured on a home-made system, which was excited by a nanosecond laser radiation pulse (355 nm, 5 Hz) from a third harmonic Nd: YAG laser (Beamtech Optronics Co., Ltd). In the test, 5 mg of the sample was covered on a Pt mesh (1 cm × 1 cm) as the working electrode with a blank Pt mesh as the counter electrode and the TPV signal was amplified by the amplifier and recorded by an oscilloscope. The in-situ TPV test was conducted on the sample film (1 mg, 1 × 1.5 cm²) deposited on an indium-tin oxide (ITO) glass substrate, which was used as the working electrode with the Pt wire as the counter electrode. During the measurement, the sample film was wetted with N₂-saturated ACN, N₂-saturated water/ACN (0.1 vol%) solution and N₂-saturated water/AA + ACN (0.1 vol%) solution, respectively, for the measurements of different situations. All the measurements were performed at room temperature under ambient pressure.

The EPR and in-situ EPR tests were measured by a Bruker A300 EPR spectrometer. For the in-situ EPR test for photogenerated holes, 10 mg of the sample was dispersed in 10 mL of ACN, into which 5 mg of TEMPO was added as a spin-trapping agent under ultrasonification. Then 100 μL of CH₃CN, 0.25 M AA aqueous solution, or BA were added into 1 mL of the above catalyst/TEMPO suspension to test the hole transfer efficiencies with different hole scarification situations. After that, 80 μL of the mixtures were injected into a glass capillary and then sealed in a glass tube under argon (Ar) atmosphere with or without light irradiation by a 100 W LED lamp for EPR test. In this manner, a strong signal of TEMPO was observed under dark, which would be weakened after illumination since TEMPO was used to trap h⁺ and produce TEMPO⁺. Thereby the weaker signal intensity of TEMPO indicates a higher concentration of h⁺, where further reaction with the additives (AA or BA) with a different reactivity can also reduce the signal. The in-situ EPR test for photogenerated electrons was similar to the above process, except that the photocatalyst and TEMPO were dispersed in 10 mL of water. In this situation, the reduction of TEMPO by electrons produced an EPR silent molecule TEMPOH and led to the reduction of intensity and flatting of EPR spectra. Moreover, 2 mg of the photocatalyst was dispersed in 2 mL of BA to form a uniform dispersion, the 200 μL of which was mixed with 20 μL of DMPO were mixed for the test of C_α radical after light irradiation of 15 min (100 W LED lamp).

Photocatalytic H₂ evolution and BA oxidation

The photocatalytic performance of the samples was evaluated for using a multi-channel photoreactor system (PCX-50B Discover, Beijing Perfectlight, China) under visible light^78,88. Specifically, 5 mg of the photocatalysts were dispersed in 15 mL of aqueous BA solution (10 vol%) with an ultrasonic bath, which was then transferred to a 40 mL quarts vial and bubbled with nitrogen for 30 min to ensure the anaerobic condition. Next, the mixture was irradiated for 5 h (white LED, 100 mW/cm²) under vigorous stirring and air cooling. A gas chromatography (GC-7900) with a TCD detector was used to analyzed the evolved gas products, which were extracted (1.5 mL) every one hour. After that, the reaction suspension was filtered through a 0.22 µm nylon syringe filter for the liquid products, which were analyzed by the high-performance liquid chromatography (HPLC, Agilent 1260 Infinity II Prime Online LC System, C18 column). The conversion of BA and the selectivity of BAD were calculated as follows:

$$\mathrm{Conversion}\left(\%\right)=\left[\left(c_0-c_{BA}\right)/c_0\right]\times 100\%$$ 6

$$\mathrm{Selectivity}\left(\%\right)=\left[\left(c_{BAD}/\left(c_0-c_{BA}\right)\right)\right]\times 100\%$$ 7

where $c_0$ is the initial concentration of BA, and $c_{BA}$ and $c_{BAD}$ are the concentrations of BA and BAD after a certain time of photocatalytic reaction, respectively.

The photocatalytic hydrogen evolution in AA system was performed almost the same as that in BA system, except for using 0.2 M AA instead of BA.

COMSOL Multiphysics simulations

For the simulation of the electric field distribution and proton concentration in mass transfer process, COMSOL Multiphysics was used to build the computational domain according to the experimental system. The simulation setup in COMSOL Multiphysics is designed to model the simplified behavior of proton, Cu(II), and C-Dots within a nanosheet structure, i.e. simulating the mass transfer in a 2D nanoreactor as a whole system. The size of the Cu(II) and C-Dots were set as 1.28 Å and 2 nm, respectively. The physical parameters used in the simulation including the material densities, relative permittivity, porosity, and diffusion coefficient of proton are provided in Supplementary Table 8. Positive charge condition was set for the Cu(II) with an initial proton concentration as the boundary condition, while negative charge condition was set for the C-Dots with a proton concentration of zero. The detailed setting for all the boundary conditions and governing equations can be seen in Supplementary Table 8 and Supplementary Fig. 56. The diffusion coefficient of proton (9.31 × 10⁻⁹ m²/s) is taken from Cussler’s book⁸⁹. The other physical parameters are all sourced from COMSOL database⁹⁰. Specifically, the potential V = 0 and no flux are used as the boundary conditions for boundaries 1–4 to mimic an electrically neutral and isolated system, as common by adopted in similar studies⁹¹. Furthermore, the charge amounts for Cu(II) (0.28 × 10⁻¹⁹ C) and C-Dots (−1.46 × 10⁻¹⁹ C) are obtained from DFT calculation. It was assumed that the Cu(II) gives the maximum proton concentration, while the proton is consumed immediately as it reaches C-Dots. Therefore, the dimensionless concentration of proton for Cu(II) and C-Dots were set as 1 and 0, respectively.

Author contributions

Q.C. performed catalyst synthesis, photocatalytic testing, catalyst characterization and wrote-original draft. B.M. and Y.L. conceived the experiments, planned synthesis, analysed results, revised-original draft, and obtained Funding. Y.Z. and H.H. helped with catalyst synthesis, photocatalytic testing, and catalyst characterization. S.W. contributed to TAS characterization. L.L. contributed to the DFT calculations. W.Y. contributed to the computational simulations of the electrical field and mass transfer. W.S. and Z.K. contributed to the data analysis through discussions and obtained Funding.

Peer review

Peer review information

Nature Communications thanks Chuanbiao Bie and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The data supporting the findings of this study are available within the article and its Supplementary Information files. All other relevant source data are available from the corresponding author upon request following the data management specifications of Jiangsu University.  Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-024-52406-2.